|Home | About | Journals | Submit | Contact Us | Français|
Over the past decade, there has been much interest in the biology and diagnostic applications of cell-free nucleic acids in plasma. Research in this area started with the demonstration of tumour-derived DNA and RNA molecules in the plasma of cancer patients. My group has a longstanding interest in using this approach for the detection and monitoring of cancer common in China, including nasopharyngeal carcinoma and lung cancer. My interest has since expanded into other species of circulating nucleic acids and has led to the discovery of cell-free fetal DNA and RNA in the plasma of pregnant women. The latter has allowed the non-invasive prenatal diagnosis of a number of fetal genetic diseases and traits (e.g. RhD blood group) from maternal blood. Progress in this area has been accelerated with the development of a number of powerful analytical tools. Recent examples include microfluidics digital PCR and next-generation DNA sequencing.
Disruptive technologies are described as ‘innovations that create an entirely new market through the introduction of a new product or service’, or ‘innovations that improve a product or service in ways that the market does not expect, typically by lowering price or designing for a different set of consumers’. They have also been described as ‘innovations that challenge the status quo’. Whilst they may be generic descriptions they can equally be applied to health care insofar as they represent a means of changing the way that health care is delivered e.g. care closer to home (new market), being more patient centred (new customer) and reducing costs (lowering price).
Reviews of health care in recent years have found systems to be fragmented and wasteful, with poor delivery and constraints in the use of new technologies, including informatics. Solutions have focused on being more patient-centred, safe, efficient and effective. Care has increasingly gravitated to the hospital setting, which cannot always accommodate patient needs in remote areas, or the increasing burden of chronic disease in the population, as well as not always being appropriate for the needs of public health.
Laboratory medicine has evolved in much the same way, with early testing at the bedside moving to an increasingly centralised mode of delivery with increasing levels of automation. However this is at the expense of the needs of rapid turnaround of results to support a more patient centred approach - whether it be for acute clinical needs, enabling better dialogue between patient and carer, or to improve the whole process (e.g. reducing length of stay).
Point-of-care testing addresses these shortcomings. The technologies being developed reflect the experience seen in other service areas e.g. telecommunications and computing, enabling testing to be performed by less skilled people, and ultimately the patient as the customer.
Strategies to address the future health of Australians and the sustainability of the health professional work force have been the subject of recent Government funded reports. These reports acknowledge the current and future shortfall in the number of active workers in professional groups required to address increasing health service demand. Incentives to address recruitment, training and retention of health workers, in particular medical practitioners, nurses and other registered allied professional groups have been activated or proposed. To ensure the success of the health service reforms the same attention should extend to allied health professionals who are not bound by registration legislation and whose activities are essential to the quality of health care in Australia. Pathology testing underpins Australians’ healthcare being the way most diagnoses are confirmed and being integral to the safe management of most diseases. Recent research into the composition of the Pathology workforce has confirmed that the demand for medical scientists will exceed supply in the near future. Drivers for the undersupply of medical scientists are a combination of demographic, technological and regulatory processes. Pathology is the most highly regulated of the health professions however the pressure from these processes may change the position of the traditional laboratory. Innovations such as point of care testing, emerging applications in molecular pathology and convergence of diagnostic technologies will impact on traditional diagnostic testing but are unlikely to reduce demand for laboratory scientists. Purposeful and innovative action must be taken now by government, industry, the profession and secondary and tertiary education facilities to address the impending crisis.
During the last decade LC-MS/MS has expanded from drug/toxicology testing and newborn screening into measurement of other low-molecular targets, which often tax the performance limits of current LC-MS/MS technology. In particular, steroid hormones and related compounds might circulate at very low concentrations in some patients and are often poor ionisers. The resultant low detection-signals can be further obscured by matrix interferences or by the presence of structurally similar precursors and metabolites. The consequent analytical challenges are compounded by the recent dramatic increases in sex-steroid and vitamin D testing.
The challenge for the next years will be to reconcile the requirement for better detection sensitivity and specificity with this simultaneous need for increased sample throughput. While LC-multiplexing and analyte, or patient-sample, multiplexing can be used to improve sample throughput to ~1,000,000 samples per year per LC-MS/MS instrument, improved sample clean up and better instrumentation are necessary to improve analytical sensitivity and specificity. MS systems with better mass accuracy and -resolution are good candidates for the latter.
These instruments are also potentially better suited to expand clinical quantitative MS into the realm of peptides and proteins, a field that is moving from basic science/discovery into translational and, most recently, clinical applications. The focus of clinical peptide/protein MS has at first been to show proof of principle of quantitative and specific detection of defined protein/peptide targets. More recently, this work has been translated into clinical assays, with the initial focus being the establishment of reference methods. In addition, early clinical applications focus on targets which (i) are difficult or impossible to measure specifically and accurately with immunoassay, (ii) are subject to frequent interferences in immunoassays, or (iii) exist as isoforms or metabolites that are impossible or difficult to distinguish from each other by immunoassay.
Pain is an important component of life. In an acute setting, it serves protective, even beneficial purposes – alerting us to trauma or disease – resolving when the injury or pathology heals. Unfortunately, >50 million patients within the USA are estimated to be experiencing pain persisting beyond the healing period. Pain of this nature is considered to be one type of chronic pain.
Chronic pain is a complex phenomenon that may also develop de novo, without any evidence of a precipitating event. Its presence frequently exerts a negative impact on a patient’s overall health and quality of life as current treatment is less than optimal and often extends over the course of weeks to months. Moreover, the costs incurred by the health care system, workplace, and society are significant.
Relying on medications that have a high potential for misuse, pain management clinics have rapidly become the major clients of our toxicology services. Their use of urine drug testing as a means to determine compliance, as well as detecting misuse of prescriptive and illicit drugs by their patients, provides numerous opportunities for additional analytic and consultative services. For example, confirmation of both positive and negative screening results is warranted in many instances, particularly for those tests representing classes of drugs, i.e. the opiates and benzodiazepines. Unexpected and inappropriate urine drug tests results may be explained by additional serum drug concentrations and, occasionally, pharmacogenetic testing.
Serious cognitive impairment, involving memory and other cognitive abilities, that interferes with an older person’s social or occupational activities, fulfils criteria for dementia. It is now realised that cognitive impairment, not fulfilling criteria for dementia (CIND), may be more common than dementia. In one community-based study of older people, the prevalence of CIND was found to be 17%, whereas the prevalence of all types of dementia combined was 8%. Conditions commonly associated with the presence of CIND include delirium, alcohol use, drug intoxication, depression, psychiatric disorders, memory impairment associated with aging and intellectual disability. Recently this category of CIND has undergone reclassification with many people within this category labelled as mild cognitive impairment (MCI). The diagnosis of MCI has not been fully operationalised and the prognosis of this condition is still very uncertain. In view of these limitations the utility of this diagnosis is questionable.
The World Health Organisation’s ICD 10 classification system has classified the common types of dementia manifest in older people as dementia in Alzheimer disease (AD) (either early or late onset), vascular dementia (VD), and dementia in other diseases. Dementia with Lewy bodies is also now commonly diagnosed. Recent neuropathologic studies of older people who have died with dementia, have revealed that pure forms of these disease entities are relatively uncommon, and many older people manifest multiple neuropathologies that lead to cognitive dysfunction. In particular for people with dementia who are older than 80 years, neuropathologic abnormalities are not as clear cut, and many older people who die with dementia, do not have severe Alzheimer or vascular pathology. The future massive increase of cases of dementia is largely due to the increase in the number of people of such age. i.e. over the age of 80 years and the focus should be on addressing the problems of cognitive dysfunction in this age group.
Alzheimer’s disease (AD) is a burgeoning epidemic throughout the world particularly in aging Western world nations such as Australia. It is estimated that by 2050 in Australia over 1 million people will be suffering from dementia. The pathognomonic marker of AD is the deposition of beta-amyloid within the brain with subsequent neurotoxicity and neuronal cell death. A major problem in the medical management of AD is early and accurate diagnosis. It is possible to derive a clinical diagnosis, but only after there has already been significant degeneration of the brain. The diagnosis of AD historically has been made with the use of neuropsychological testing, assessment of functional ability and monitoring changes over time. Anatomical imaging modalities such as CT and MRI have by and large been utilised to exclude other pathologies rather than accurately diagnosing AD. With new potential treatments it is critical to diagnose AD early, accurately and preferably non-invasively. Functional imaging techniques appear to be able to not only give insights into the alteration in brain function that occurs over time with AD but now with potential specific imaging markers for AD for the first time we are able to define groups of patients at risk of developing AD well before the development of clinical symptoms. Three major functional imaging techniques have been utilised to assess brain function and more recently pathogenesis in AD. These include brain perfusion single photon emission computed tomography/positron emission tomography (SPECT/PET) which utilise tracers that map brain perfusion, 18-fluorodeoxyglucose (FDG) PET which measures glucose metabolism of the brain and C-11 Pittsburgh protein B (PIB) which measures beta-amyloid deposition in the brain. Brain perfusion agents and FDG PET appear useful in differentiating dementia types by the pattern of perfusion or metabolism demonstrated. Loss of perfusion and metabolism however indicates neuronal loss of function and likely neuronal death. C-11 PIB allows accurate assessment of cerebral amyloid burden and appears to predict the development of the clinical syndrome of AD. These techniques are allowing earlier diagnosis and neuronal function assessment. They will be important tools in assessing treatment response until other simpler biomarkers are developed.
The evolution of the concept of the metabolic syndrome has been of interest in medical and scientific circles over the last several years. At its core, the syndrome identifies a cluster of disturbances, primarily those of glucose and lipid metabolism, and blood pressure, which together with central adiposity, predicts an increased future risk of diabetes mellitus and cardiovascular disease. Further interest in clinical research has necessarily generated a number of different definitions over the years and more recently, stimulated attempts at unifying the different confusing criteria into consensus definitions (the latest one by a number of different organisations in late 2009). The use of categorical cut-offs for each criterion remains a weakness of these definitions. Whilst the clinical utility of recognition of the syndrome in patients is an ongoing source of debate and controversy, the concept has nevertheless appropriately driven much clinical and basic research into our understanding of metabolic and cardiovascular disease over recent years.
Nonalcoholic fatty liver disease (NAFLD) refers to accumulation of hepatic fat in the absence of excessive alcohol ingestion. NAFLD is now the commonest chronic liver condition in the world, affecting up to 9% of people in developing countries and up to 30% in the developed world. NAFLD is pathogenically associated with central obesity and insulin resistance, with recent evidence highlighting a direct role of toxic lipids and oxidative stress in causing progressive disease. Histologically, NAFLD may be divided into simple steatosis and nonalcoholic steatohepatitis (NASH), with the latter being the more aggressive sub-type. In a few individuals, NAFLD may progress to cirrhosis with sequelae of liver failure and hepatocellular carcinoma which may lead to death or requirement for liver transplantation. Subjects with NAFLD may also be at increased risk of developing metabolic complication such as diabetes as well as dying from cardiovascular disease. Currently NAFLD is diagnosed by confirming the presence of hepatic steatosis by imaging and excluding secondary causes of hepatic steatosis clinically. Traditional liver biochemistry is insensitive and non-specific for diagnosing NAFLD or evaluating its severity, however, novel serum markers reflecting apoptosis and fibrogenesis are promising as non-invasive clinical tools for the future. Treatment currently revolves around reversing weight gain, correcting metabolic abnormalities such as insulin resistance and targeting oxidative stress.
Polycystic ovary syndrome (PCOS) is the most common form of endocrine disorder in premenopausal women. Reported prevalence of PCOS among women of reproductive age varies according to the mode of diagnosis and is estimated to be between 5 and 18%. The diagnostic hallmarks are hyperandrogenism and ovulatory dysfunction leading to hirsutism, acne, male pattern alopecia and infertility. The diagnostic criteria for PCOS are broadly inclusive and there are undoubtedly subsets of PCOS with different pathophysiology. However, insulin resistance and subsequent hyperinsulinaemia are thought to be central to the pathophysiology in the majority of PCOS with high circulating insulin amplifying ovarian androgen production via insulin receptors on theca cells and increases bioavailable testosterone by suppression of SHBG. There is a high prevalence of obesity, dyslipidaemia, glucose intolerance and diabetes in women with PCOS and women with PCOS have a higher prevalence of the metabolic syndrome compared with age-matched population, implying increased cardiovascular risk. No long-term studies, using accepted diagnostic criteria, have been done to confirm this prediction.
Nevertheless, there have been many studies in PCOS - biochemical, functional and structural - identifying an increased prevalence of conventional and non-conventional cardiovascular risk markers, impaired endothelial and vascular function and structural changes in arterial walls compared with controls. Lifestyle modification, weight control and exercise and insulin sensitisers have been demonstrated to ameliorate hyperandrogenism and menstrual irregularity. More importantly, these interventions are likely to reduce the cardiometabolic consequences of the syndrome.
Osteoporosis is a major clinical problem in Australia and worldwide, and its prevalence is increasing exponentially. Fracture is the important clinical outcome of osteoporosis; the lifetime risk of sustaining an osteoporotic fracture for a 50 year old Australian woman is >40%. While bone density measurement is used for diagnosing osteoporosis, age and other risk factors also significantly impact on, and are used in, fracture risk calculation in individual patients. The 10 year absolute fracture risk calculation in individual patients, incorporating all identified risk factors, is a useful tool for targeting those at high risk of developing osteoporotic fractures in order to make cost effective treatment decisions.
The availability of several bone turnover markers in urine and in serum has led to different markers being used in different studies. There is currently inadequate data on any bone turnover marker for their inclusion in fracture risk calculation; however, they are useful in monitoring therapy. There is a need for an internationally agreed reference bone marker each for bone formation and for bone resorption for use in clinical trials. This would help accumulate adequate data on the designated reference bone markers in prospective studies for them to be then included in fracture risk calculation algorithms. IFCC is working towards this goal.
Apart from the established antiresorptive agents such as bisphosphonates which are widely used for treatment of osteoporosis, the recent availability of anabolic agents such as injected PTH analogue (Teriparatide) has broadened choice in treatment. In addition, the elucidation of molecular pathways that control bone remodelling and influence bone loss has led to development of novel treatments that selectively target those pathways with precision, and these agents are now becoming available for effective use in treatment of osteoporosis.
Vitamin D contributes to the maintenance of calcium, and phosphate homeostasis as well as exerting a wider range of biological activities including regulation of cellular differentiation and proliferation. The endocrine action of vitamin D acts through its renal metabolism, producing 1,25 dihydroxyvitamin D (1,25D) in the circulation for which the intestine is the major responsive organ controlling absorption of calcium and phosphate. Serum 25-hydroxyvitamin D (25D) levels below 20 nmol/L decrease the endocrine action of vitamin D resulting in hypocalcaemia, hypophosphataemia, secondary hyperparathyroidism and osteomalacia in bone. 1,25D is also synthesised in a wide range of tissues including bone cells where it is under investigation as an autocrine or paracrine agent. Vitamin D insufficiency in the elderly increases the risk of hip fracture due to osteoporosis. A major question is the cellular and molecular mechanisms by which depleted levels of vitamin D produce osteoporosis. Preclinical studies with low vitamin D diets demonstrate that serum 25D levels between 20 and 80 nmol/L result in trabecular and cortical bone loss without any evidence of osteomalacia. This bone loss is largely due to increased bone resorption. No relationship is evident between bone volume and either serum 1,25D or parathyroid hormone in these animals. 25D is metabolised to 1,25D by each of the major bone cell types which is essential for its anabolic bone cell activities particularly related to promotion of bone cell maturation. These preclinical data indicate that osteomalacia is resolved with serum 25D levels below 20 nmol/L while osteoporosis is resolved at serum 25D levels above 80 nmol/L.
Whilst vitamin D has well recognised effects on musculoskeletal health, the effects on non-skeletal health have been equally interesting but more controversial. Vitamin D may exert effects on both innate and adaptive immune function resulting in a possible role in the pathogenesis of a number of autoimmune diseases including Multiple sclerosis, type 1 diabetes, Rheumatoid arthritis and Crohns disease. Indirectly, but possibly through effects on immune function, vitamin D may also play a role in infection and asthma incidence rates. These effects appear to be mediated in a paracrine or autocrine fashion via local synthesis of active vitamin D metabolites. These outcomes are based largely on observational and epidemiological data. Whether dietary fortification or supplementation has the same effect as endogenous vitamin D also needs to be examined. Two small randomised clinical studies to date have demonstrated an effect of cholecalciferol supplementation on infection rates. The dose of vitamin D required to exert beneficial effects on immune function are not known and in the absence of large randomised clinical trial data, it is premature to recommend universal supplementation for the prevention of non musculoskeletal disease.
Profound advances in our knowledge of hereditary haemochromatosis (HH) have occurred during the last 150 years resulting in two distinct iron ages; the pre-HFE gene era and post-HFE gene era. During these periods, family studies, HLA association studies and ultimately HFE gene studies in various populations have informed us of the genotypic prevalence, as well as the clinical and biochemical penetrance of HH. We have learned that HH has a highly variable clinical penetrance in susceptible individuals of Northern European ancestry. Further, we now recognise that the natural history of HH is not as discrete as previously believed, as genetic and environmental modifiers of disease penetrance are increasingly identified as influencing the clinical expression of HH. The key genetic mutations which result in HH include those affecting HFE, transferrin receptor 2, ferroportin, haemojuvelin and hepcidin. These primarily result in impaired sensing of iron status and production of hepcidin, the key regulator of iron absorption from the gastrointestinal tract and iron release from bone marrow macrophages.
Neuromuscular disorders, diseases of the nervous system and/or muscle, include devastating adult diseases such as Alzheimer’s, and motor neuron disease, but also infant diseases causing paralysis at birth. In 1987, molecular diagnosis for neuromuscular disorders was based on the only gene identified, the gene for Duchenne muscular dystrophy. Now >260 genes are known and molecular diagnosis could be provided for all. However, molecular diagnosis is difficult because neuromuscular disorders involve many of the largest human genes (e.g. titin) and because genetic heterogeneity (multiple genes causing the same disease) is high. Molecular diagnosis for neuromuscular disorders is currently performed one gene at a time and is thus expensive and time consuming. New technologies offer the possibility to analyse multiple genes in parallel: copy number variation chips to detect deletions and duplications, gene capture and new generation sequencing to analyse multiple genes for small-scale mutations. It is now possible to sequence the entire exome or whole genome for a reasonable price. The importance of identifying the disease-causing mutation in a patient is that it allows accurate diagnosis, determination of inheritance and prenatal diagnosis. Prenatal diagnosis in turn allows disease prevention, saving millions of dollars in health budget. Another aim of researching the genetic cause of diseases was to develop treatments. However, many of the treatments are very expensive, for example enzyme replacement therapy at >$500,000 pa. If the patient achieves a normal lifespan through treatment, the lifetime cost could be >$35m. This is not affordable. What is perhaps needed is prevention rather than cure. Many disorders are recessive. New technologies, for example genome sequencing when it is less than $1,000, may permit population screening to detect carriers before they find out they are carriers by having affected children. Maximising the possible benefit from genetic information, could take society to interesting places.
The clinical aspects and diagnosis of classical cystic fibrosis are sufficiently well known that it is enough to say that it is an autosomal recessive syndrome of suppurative endobronchitis often with sinusitis, more or less pancreatic insufficiency, and many other manifestations and late complications. It is due to absence or malfunction of an epithelial chloride channel called cystic fibrosis transmembrane conductance regulator (CFTR) from more than 1500 documented mutations in the CFTR gene.
Recently, patients with incomplete phenotypes, and others with intermediate sweat test values but only one documented disease-causing mutation, have been labelled as having ‘CFTR-related disorder’ and ‘CFTR-related metabolic syndrome’ respectively. The use of such terms, while not universally accepted, at least emphasise that this is a spectrum disorder, as is becoming clear with most genetic disorders.
With widespread newborn screening especially in the developed west, the overwhelming number of patients are diagnosed pre-symptomatically. However we are now seeing an increasing awareness of intermediate and adult presentations of the disease which is presenting new challenges in testing and in interpretation of results. Sweat testing, the subject of this workshop, is well documented in children in particular but has only recently been applied to adults in a systematic way.
The analysis of the electrolytes themselves, often at low concentrations, is clearly important but would not justify an AACB working party. What makes the sweat test special is the method of collection of the sweat. Quite clearly, variations between laboratories in this collection cannot but be reflected in the precision of the results, and these variations have been well documented in the literature. It is therefore the intention of the working party to standardise the sweat test to the extent that it is necessary. Local and international publications have clearly documented this procedure.
The collection of sweat by pilocarpine iontophoresis, followed by electrolyte analysis for the diagnosis of Cystic Fibrosis, is based on the technique developed by Lewis Gibson and Robert Cooke in 1959. Associated literature reports highlight both false positive and false negative results. Whilst a small proportion of these are associated with simultaneous pathologies the majority have been attributed to preventable collection and analytical issues. This has provided the impetus for the development of guidelines to improve standards.
The Clinical and Laboratory Standards Institute (CLSI) published such a guideline in 1994 and then a second edition in 2000. In 2003 a UK multidisciplinary group produced an extensive evidence based guideline. Using this UK guideline as a reference, the AACB Sweat Test Working Party (STWP) developed ‘Australian guidelines for the performance of the sweat test for the diagnosis of Cystic Fibrosis’ in 2005. Together these three sets of guidelines have served to improve performance.
In late 2009 the third edition of the CLSI guideline was published; ‘Sweat Testing: Sample Collection and Quantitative Chloride Analysis; Approved Guidelines – Third Edition.’ Some important new features are the inclusion of additional information relating to the preferred site of collection, applied iontophoresis current, specimen stability and storage, age related reference intervals, and increased emphasis reinforcing chloride as the diagnostic analyte.
The production of the CLSI guideline 3rd edition improves alignment of the three sets of guidelines. The Australian (AACB) sweat testing guideline is based on the UK guideline. The CLSI guideline should however continue to be used as an additional resource. Ideally the development of a freely available CLSI-UK-AACB combined guideline will be a future joint initiative.
Cystic fibrosis (CF) is the most common inherited life-shortening genetic disorder affecting Australians. Other than genetic testing, the sweat test remains the major diagnostic test for the confirmation of CF. The sweat test measures the concentration of electrolytes that are excreted in sweat. An elevated concentration of chloride is the diagnostic analyte for CF.
The laboratory plays a major role in the early diagnosis of CF, and with this comes the responsibility of reporting reliable and accurate results to ensure proper patient management. Firstly, is the patient suitable to test, as there are a number of factors than can affect sweating and therefore the result. It is also important that the staff collecting the sweat are well informed, experienced and trained as they are the first point of contact for the patient after their initial consultation with the doctor. Secondly, the laboratory needs to select the appropriate method for sweat stimulation and collection. A good quality sweat sample is the first step to accurate results. As for the sweat analysis, the methodology must be able to measure chloride in the concentration ranges for both normal and patients with cystic fibrosis. The internal quality control levels should be at the decision points and the analysis performed by trained scientists. Finally sweat test results should be reported with reference intervals and with the option to include interpretative comments on the patient report.
To optimise sweat tests the laboratory must have the appropriate procedural processes for sweat collection, sweat analysis and reporting of results with ongoing assessment of competency to minimise the potential for errors or misdiagnosis.
Cystic Fibrosis (CF) is a serious life-limiting disease. The reporting of false negative, and false positive, sweat chloride results can have detrimental consequences. Total intra-individual variation (CVt) for sweat chloride has been demonstrated to be as high as ≈30% in the non-CF population. The collection of sweat is a significant contributor to CVt. To minimise pre-analytical error, rigorous attention is required to maintain high levels of control over the collection process.
The performance of sweat testing is operator dependent. Stimulation and collection of sweat should only be performed on either the inner forearm (preferred site) or standardised alternative sites such as the inner thigh. The site selected is based on eccrine sweat gland density, safety considerations and the electrode surface sitting flat on the skin. Each trained collector should perform at least 10 collections per annum and ideally >95% of their collections should provide sufficient sweat based on the prescribed average sweat rate of one gram per meter squared per minute over a total sweat collection time of 25+/−5 minutes. Thorough training of collection staff is important to ensure optimisation of these components.
Internal and external quality control procedures are valuable for analytical competency but do not test the ability of the operator to collect the sweat accurately. The area of sweat testing most susceptible to incorrect performance, operator competency, remains effectively uncontrolled. The Australian Sweat Test Guidelines support the repeat testing of a volunteer at regular intervals as a control. This approach provides the best assessment of staff competency, controls the overall sweat testing process and allows for the laboratory to assess their CVt.
Sweat testing is important in the diagnosis of Cystic Fibrosis (CF). Sweat chloride results <40 mmol/L are considered normal and >60 mmol/L suggestive of CF. The analytical imprecision of the test is quantifiable by internal and external quality assurance programs but the variation due to the collection procedure and the biological variation of the patient is not well defined.
In a study of healthy adults sweat was collected on eight occasions in three subjects and on twelve occasions in one subject over a 2 year period using the Wescor Macroduct collector system. A separate study assessed sweat conductivity weekly for 5 weeks in 15 healthy adult volunteers, 20 healthy infants and 20 known CF patients. A third study investigated test variability by comparing sweat electrolyte values from two sites collected simultaneously by the Gibson and Cooke method.
The simultaneous measurement study showed a coefficient of variation (CVt) between the two sites of 20.2% for chloride (n = 269). The CVt of the adult volunteer group ranged from 14.9 to 32.9%. This may have been higher due to temporal variation. The sweat conductivity study showed CVt of 18% for the healthy infant group. These uncertainties of measurement may lead to misclassification in diagnosis, with false positives up to 15% and false negatives up to 12%.
These results indicate that great care must be taken in the interpretation of sweat chloride results. Defined cut-off values do not take account of the CF spectrum, which is likely to be continuous, nor the wide intra-individual variation. The diagnosis of CF must include a critical evaluation of the clinical presentation of the patient.
As our understanding of Cystic Fibrosis and its manifestations grows, it is sensible that our understanding of sweat chloride reference intervals grows as well.
In classical disease, the traditional reference intervals of <40 mmol/L- normal, 40–60 mmol/L- equivocal, and >60 mmol/L- supporting a diagnosis of CF, work well. The more subtle disease associated with varying ages and mutations are best illuminated by the changes in reference intervals recently described showing a gradual increase in sweat chloride concentration throughout life.
Particularly in subtle disease, one must be aware of the non-CFTR mutation related causes of increase in sweat chloride including adrenal disorders and malnutrition.
While investigation of mutations in people with suggestive chloride values can be helpful, it is important to recognise that standard panels of mutations often do not include the mutations associated with difficult sweat chloride results.
In typical CF, reporting sweat chlorides is straightforward. In atypical CF, reporting sweat chlorides is challenging and calls for close communication between the laboratory and the clinician.
Chronic kidney disease (CKD) is an important health issue with some 2 million Australians estimated to be affected. Early detection requires input from both clinicians and laboratorians. The effectiveness of the implementation of eGFR in Australasian laboratories was in part due to the co-operative activities of professional Nephrology and Clinical Biochemistry bodies. In September 2009 a working group convened, seeking to achieve similar uniform practice in screening for proteinuria, the hallmark of CKD. The Proteinuria Albuminuria working group (PAWG) has members from Australasian Association of Clinical Biochemists (AACB), Australian and New Zealand Society of Nephrology (ANZSN), Kidney Health Australia (KHA), the Royal College of Pathologists of Australasia (RCPA), the Royal Australian College of General Practitioners (RACGP) and the Australian Diabetes Association (ADA). The purpose of the group is to provide recommendations for the screening of proteinuria in the adult Australasian population to clinicians and laboratorians. A draft document has been produced that recommends urine albumin/creatinine ratio, ideally on a first morning void as the preferred sample to screen for proteinuria. Dipstick testing is not recommended, nor is urine protein/creatinine ratio. Current reporting practice amongst laboratories is far from uniform and there are a number of opportunities for improvement; for example the Group recommends the adoption of gender based reference intervals, use of S.I units and repeat testing to confirm positive results. Possible interpretative comments are also offered.
The Endocrine Working Party (EWP) was formed in 2009. There was an overwhelming response to the call for expressions of interest for EWP membership published in the September 2009 Clinical Biochemist Newsletter. To maintain the valuable expertise of all those who nominated two levels of membership were initiated; full and corresponding. There are representatives from most states and New Zealand with a wide breadth of knowledge spanning paediatric, clinical and laboratory endocrinology.
The main function of the EWP is to provide advice to the Chemical Pathology QAP on the Endocrine, IGF-1/C-peptide, BNP and Tumour Marker programs. Levels of the low and high pools used to produce the QAP samples are reviewed annually and a number have been adjusted to reflect changes in clinical needs and improvements in assay technology. Allowable limits of performance are currently under review. Discussion has been initiated with a number of relevant professional bodies regarding growth hormone standardisation and the EWP will be assisting the RCPA in defining preferred units for endocrine tests which is a necessary part of the National e-Health initiative. Macroprolactin has been proposed as a potential new analyte. The EWP is investigating the suitability of a synthetic macroprolactin material and has drafted a macroprolactin survey will be sent to QAP participants. Red Cell folate (RCF) is new to the Endocrine program and results indicate a high degree of within and between method variability. The EWP has established a project aimed at improving the measurement of RCF which will include establishing a reference method.
In 2008 and 2010 the Royal College of Pathologists (RCPA) Quality Assurance Program (QAP) introduced two additional programs for vitamins B1 (thiamine diphosphate, TDP) and B6 (pyridoxal phosphate) and then vitamin C. A review of results submitted for the vitamin B1/B6 program revealed a wide dispersion, particularly for TDP. This led to the circulation of a questionnaire to participants. The pilot vitamin C program was introduced for the first cycle of 2010. Here we provide an update of these programs.
Both sub-programs are conducted in the established protocol widely adopted for programs in the QAP. Each cycle contains 6 linearly related samples in duplicate. Medians of results are used as the expected values. The acceptable ranges are determined by clinically based Allowable Limits of Performance.
In the first cycle of 2010 results for TDP (11 participants) demonstrated a CV of up to 46% between and 37.6% within laboratories. For pyridoxal phosphate (5 participants) the variation is up to 28% between laboratories. Results for Vitamin C (13 participants) appear to be linearly related with a wide dispersion, as expected for the first cycle.
These new vitamins programs offer a vital service that is of an international standard. The wide variation of results between laboratories is currently being addressed by the working party with a view to better understanding the source of variation and to assist in limiting it. The vitamin C was successful. While not large, it is a niche program that provides an international service and will continue as a full program from the second half of 2010.
SolarisCare Cancer Support Centre (originally Brownes Dairy Cancer Support Centre) opened at Sir Charles Gairdner Hospital in September 2001 with the purpose of providing support, information and complementary therapy (CT) sessions for individuals with cancer and their carers. SolarisCare provides innovative community-based support services through its volunteers, both meet and greet and complementary therapists.
The Centre has over 120 visitors weekly, half will attend CT sessions. Inpatients can also access sessions. Patients completed a symptom distress scale (SDS) and a quality of life (QoL) scale immediately before and after the first, third and sixth sessions. Results reported here are for 1244 patients treated between February 2002 and November 2007.
Of 1244 patients, 77% (n=952) were female and the mean age was 55 (range 19–85 years). Outpatients accounted for 81.7% of patients (n=970) and 32% (n=846) received mainstream cancer care at another hospital. In total, 2724 CT sessions were performed. These are categorised into: touch based therapies e.g. reflexology (51%); energy-based e.g. reiki (39%); mind-based e.g. meditation (7%) and supportive counselling (3%). Over a course of 6 CT sessions, individual SDS improved for fatigue, nausea and pain, bowels, breathing, sleep and appetite. The improvement in overall SDS (calculated from mean of individual scores) was statistically significant over time (p<0.0001). Mean QoL also showed significant improvement (p<0.001). Due to marked attrition over time (1303 first sessions, 977 third sessions and 444 sixth sessions) analysis was made for ‘missingness’ using a multi-variate random-effects model, which confirmed the statistical significance of the results.
Data presented demonstrates that offering complementary therapies in a teaching hospital through a community-based cancer support centre has a positive impact upon QoL and can reduce symptom distress in cancer patients.
Mesenchymal stromal cells (MSCs) can be isolated and expanded, then be utilised for a variety of therapeutic applications including inflammatory, cardiovascular, autoimmune and orthopaedic diseases. MSC were first discovered in bone marrow by their adherence to plastic in tissue culture. They self renew and can differentiate into cells of mesodermal lineage including bone, cartilage, tendon, muscle, fat and neuronal cells.
MSC are important cells regulating immunity in the body because they respond to inflammatory cytokines, becoming activated and expressing anti-inflammatory cytokines and other factors such as Indoleamine 2,3-dioxygenase (IDO). This results in suppression of cytotoxic T lymphocyte activity, suppression of B cell antibody production. They express low levels of HLA class I and no class II surface molecules and are immuno-privileged and escape detection by natural killer cells. MSCs also possess regenerative and reparative properties and have been employed in areas such as repair of myocardium after ischaemic injury.
Since 2007 we have been conducting a phase I study of MSCs in steroid refractory graft-versus-host-disease (GVHD) after allogeneic bone marrow transplantation. MSCs were cultured from bone marrow from HLA-identical or related ‘third party’ normal donors. Initially, patients were treated with 2 x 106 MSC / kg intravenously twice weekly for four weeks and subsequently modified to once weekly for two weeks. To date 102 infusions have been administered with no adverse safety effects. Of the ten patients evaluable with acute GVHD there were seven complete responses and four deaths, three due to sepsis and one to a haemorrhagic stroke. There have been no recurrences of haematological malignancy. Of the seven chronic GVHD patients, two have had CR and two PR, with four deaths due to ongoing progression of chronic GVHD. We are proceeding to a phase II study of GVHD.
Other MSC therapy trials are being developed. Ethics approval has been obtained at Royal Perth Hospital for a multi-centre Australian study in biologic-refractory Crohn’s disease and for a study in collaboration with Prince Charles Hospital in Brisbane for progressive obliterative bronchiolitis after lung transplantation. Further studies are being instigated in renal transplantation, initially to see if MSC growth from patients with end stage renal failure is normal and will support immunomodulatory activity.
Newly introduced Federal initiatives, including the MBS Quality Review Framework and Review of Funding Arrangements for Pathology Services, have highlighted the government’s intention to make explicit the role of ‘evidence’ in funding decisions of new and existing MBS items. This project explores how this policy context might impact existing pathology services such as vitamin B12 and folate tests. In accordance with the review frameworks, the diagnostic accuracy of these services will form a critical component of MBS assessment. Therefore, this study aimed to investigate the evidence base of the diagnostic accuracy of serum B12, serum folate and red cell folate tests through a systematic literature review.
Based on a protocol, a standardised search strategy was applied to six databases. Primary studies and systematic reviews that met inclusion criteria were selected. Included studies underwent appropriate quality assessment and data extraction.
The search identified 57 articles comparing (any of) serum B12, serum folate and red cell folate tests with a reference standard(s). Results demonstrated highly variable diagnostic accuracy for serum B12 and serum folate, and a limited evidence base for red cell folate tests. Extensive heterogeneity existed, particularly in test methods, threshold values, clinical indications and study quality. The discriminatory power of these tests is often low across various subgroups.
Estimates of the diagnostic accuracy of serum B12, serum folate and red cell folate tests suggest these services could be candidates for assessment in the context of the Federal Quality Assessment framework. This work demonstrates the nuanced approach required for evaluating diagnostic tests in the context of public reimbursement.
A generic definition of commissioning is ‘a means securing the services that most appropriately address the needs and wishes of the individual service user, making use of market intelligence and research, and planning accordingly’. In health care this means securing the best value for patients and payers, with the best possible health outcomes, the best possible health care experience, which is affordable.
There are generally considered to be four stages in commissioning a service (i) assessing the local health care needs – and the expected outcomes, (ii) specifying the services required - and identifying the resources required, (iii) securing the services required – and any clinical practice change and resource reallocation, and (iv) monitoring against the contract issued, and evaluating the outcomes against the expectations. There are similarities to the so-called evidence-based laboratory medicine cycle, of (a) identifying the clinical problem or need, (b) acquiring the evidence, (c) critically appraising the evidence, (d) applying the evidence, and (e) auditing the process. These attributes are also embodied in continuous quality improvement. Commissioning has generally been applied to care pathways, the implication being that all of the elements of the pathway operate effectively - a major challenge in a complex service such as health care.
The evidence of impact of laboratory medicine on health outcomes is poor, and inappropriate utilisation is acknowledged. A diagnostic test can be employed in a number of ways (screening, diagnosis, monitoring etc) and consequently should be evaluated and commissioned in the context of a ‘diagnostic pathway’ rather than merely as a ‘test’, reversing the increasing commoditisation of laboratory services. Effective implementation of a test requires a change in clinical process and reallocation of resources; these aspects of service integration and transformation represent major challenges in health care if laboratory services are to be used effectively.
Evidence based medicine (EBM) can be described as the integration of clinical expertise with the best available evidence from systematic research. Information technology (IT) allows the rapid accumulation of research findings in readily accessible formats. It also provides a vehicle to deliver evidence-based practice through electronic decision support to improve the quality of health care. In this way IT can be seen as EBM’s underlabourer. The aim of this presentation is to critically review the evidence about the development and progress of health IT and to identify the major challenges to its successful uptake.
A review of the research literature about the role of health IT with particular reference to pathology services, drawing upon an extensive 6-year body of Australian-based research into the impact of electronic ordering systems on hospital pathology services.
Despite the considerable international enthusiasm for IT systems in health, their diffusion remains slow, barely reaching 20% across major OECD countries. This experience has prompted concerns about the generalisability of the current evidence base. The pioneering randomised controlled trials from the last two decades may have highlighted the huge potential of IT, but a significant proportion of these trials came from relatively few sites that developed their own “home grown” systems with the support of clinical and managerial enthusiasts. This experience is not necessarily replicable across other sites, particularly with the increasing market for commercially-developed ‘off-the-shelf’ systems. The literature shows that there are complex organisational and communication processes involved with the implementation of IT in health involving both social and technical dimensions.
The challenging task of incorporating IT into the infrastructure of healthcare requires a broader and more inclusive evidence base that is premised on the notion that one size may not necessarily suit all.
Incidence rates of differentiated thyroid carcinoma are increasing in most countries. Since thyroid cancer is associated with low mortality and long survival, this has led to a dramatic increase in living thyroid cancer patients. Given current trends, thyroid cancer will be third most common diagnosis in living cancer patients/survivors in the USA within the next 5–6 years, behind breast and prostate cancer and ahead of colon cancer. Since tumour recurrence affects about 1/3 of patients at some stage of their illness, this creates a huge and growing need for follow-up testing.
Because of its organ-specificity, serum thyroglobulin (Tg) has long been the key thyroid cancer tumour marker. Recent improvements in Tg assay sensitivity have largely obviated the need for stimulated Tg measurements, while Tg testing in body fluids and biopsy needle washes has improved definitive diagnosis of suspected recurrences. However, Tg measurements are hampered in 15–20% of patients by the presence of anti-Tg autoantibodies (TgAB). In addition, if current trends of increasing incidence and falling mortality continue, a growing subset of future patients might neither receive total thyroidectomy, nor radioiodine remnant ablation. Consequently, serum Tg measurements will become less useful for detecting recurrence, as Tg is organ-specific, but not tumour-specific.
Identification of strategies to overcome TgAB interferences and to identify other thyroid tumour markers will therefore become increasingly important in the future. We have shown recently proof of principle for the latter in a prospective multi-year study of measurement of circulating BRAFT1799A in the follow up of papillary thyroid cancer patients. We detected BRAFT1799A in the blood of 20 of 173 patients, and its presence correlated with persistent/recurrent disease. We are now expanding this approach to other known molecular thyroid cancer markers and are investigating the diagnostic potential of circulating micro-RNA based markers.
Hypertension is common and is a major contributor to vascular morbidity and mortality. Endocrine hypertension, where hormonal derangements cause significant elevation of blood pressure, accounts for 3% of secondary forms of hypertension. The most common causes of endocrine hypertension are excessive production of mineralocorticoids, for example in primary hyperaldosteronism, catecholamines in phaeochromocytoma or glucocorticoids in Cushing’s syndrome. The likelihood of a secondary cause of hypertension is increased when patients are in a younger age group, have hypertension which is refractory to treatment or a family history of an endocrinopathy.
Identification of an endocrine cause of hypertension is critically important as appropriate management often leads to partial or complete normalisation of blood pressure. Biochemical confirmation is usually required for diagnosis, as demonstrated in primary hyperaldosteronism, Cushing’s syndrome, phaeochromocytoma or the 11- and 17-hydroxylase deficiency disorders of cortisol/aldosterone biosynthesis.
Laboratory test results must be interpreted in comparison to expected reference intervals (normal ranges) to determine if patients are healthy or subject to a pathological condition. It’s a challenge for laboratories and IVD manufacturers to conduct meaningful reference interval studies because they do not have ready access to large numbers of suitable healthy subjects, the studies are difficult to perform and resource intensive, and they should be tailored to suit the patient population (e.g., partitioned by age, gender, ethnicity, geography).
The CLSI and IFCC jointly published CLSI C28-A3 (Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory) in 2008. C28 provides reference interval theory, but individual laboratories, regional initiatives, or national programs must apply the theory to generate meaningful intervals. Reference interval studies are direct (subjects preselected for inclusion prior to testing) or indirect (data collected from subjects previously tested but not subject to prescreening). Direct studies are either a priori (subjects vetted prior to sampling) or a posteriori (subjects selected after sampling).
Good examples of direct studies include the national Canadian Laboratory Initiative on Paediatric Reference Intervals (CALIPER), the ‘Aussie Norms’ project in Canberra, and the Hong Kong Blood Donors study. Studies using the indirect approach also provide useful data.
The direct approach is preferable as it allows for fine partitioning of the data, producing highly specific intervals well suited for various patient populations. The indirect approach requires subjects to be screened after the fact, but it has the advantage of using data that has already been collected and that is readily available. Both approaches have strengths and weaknesses. Clinical laboratory scientists must understand reference interval theory to make judicious decisions when selecting normal ranges.
The appropriateness of using the reference intervals supplied by diagnostic reagent and equipment manufacturers for clinical decision making in Australia is questionable. The manufacturers are predominantly North American, quote small sample populations do not include health status data and sometimes draw those populations from hospital-based groups. The Aussie Normals Study aims to establish reference intervals for haematological, biochemical and immunological metrics for healthy Australians using community sampling.
In this ongoing study, blood is collected from volunteers into tubes containing no anticoagulant, ACD, EDTA, and lithium heparin. 116 biomarkers including LFT, iron studies, electrolytes, TFTs, lipids, cancer, cardiovascular and glycaemia markers, immunoglobulins, complement and acute phase proteins, rheumatic markers, vitamin D and bone markers, endocrinology and reproductive hormones, FBC, T and B cell markers are analysed with additional biomarkers added as they become available. Ethics has been approved by ANU, UNSW, ACT and AEC HRECs.
Results generated include age-related data sets from young adults to geriatric subsets, which will provide the basis of appropriate reference intervals and assist us in identifying novel gender specific differences in some analytes and other unique information. We have identified novel associations between phenotype and analytes such as BMI and complement (C3 and C4).
There is a substantial gap between theory and practice with respect to the application of reference intervals and closing that gap, overturning current reference intervals, moving us towards a national standard is not easy. The data generated by the Aussie Normals project will be used to determine appropriate reference intervals for the Australian population and as a validation source for existing Bhattacharya data. To assess analyser bias a cohort of Aussie Normals samples will be analysed on variety of platforms.
Suitable reference intervals are the keys to unlocking the benefits of results of laboratory testing. Even for common analytes, children’s reference intervals have been sorely lacking. There are a number of good reasons for this.
Over the last few years a number of projects around the world have been undertaken to help improve this state of affairs. The CALIPER project from Canada, Norip from Scandinavia, the American National Children’s study, and CHILDx from ARUP Laboratories are projects underway.
In addition to using projects designed to determine reference intervals, it is good to keep an eye open for unexpected but usable data.
Novel approaches to data manipulation have been developed to better describe reference intervals. TJ Coles’ smoothed percentiles (lmsChartmaker) are among useful tools for evaluating data.
For use today, the Mayo Medical Laboratories and Pediatric References Intervals, 6th Edition by SJ Soldin et al. are readily available offshore resources. Sonic Network reference intervals for children are new local resources.
Whole scale changes in our ability to provide useful information to enhance the care of children are becoming available to everyone involved in the care of children.
Point-of-care testing (PoCT) provides a practical option for the provision of pathology services in rural and remote communities, as tests are conducted on-site on a small blood sample, results are available in less than 15 minutes, and clinical management can be initiated ‘on the spot’.
The Community Point-of-Care Services unit at Flinders University has introduced innovative education, training and competency, quality management and support service frameworks to underpin and implement a range of PoCT models using a variety of different devices for the prevention and management of chronic and acute diseases in Indigenous and non-Indigenous rural and remote community settings across Australasia. These include the QAAMS Program (involving 115 Aboriginal medical services), the Diabetes Management Along the Mallee Track Program (at the rural town of Ouyen, Victoria), the Northern Territory i-STAT POCT Program (involving 33 remote health services), the KEY Study (with Kidney Health Australia) and the Ngati Porou Hauora Warfarin Management Program (in 6 Maori health services, New Zealand).
Analytical quality has consistently met profession based analytical goals and/or state of the art laboratory performance across these models. Clinical effectiveness has been proven using case studies, statistically significant improvements in clinical outcome measures (e.g. glycaemic control, time in target range, changes in clinical categories), stabilisation of acute conditions and reduction in medical retrievals. Community satisfaction with PoCT has been validated using qualitative surveys of device operators and patients. Partnerships with Governments, professional organisations and community health services have been pivotal to success and sustainability.
There is a strong evidence base that PoCT has facilitated community engagement, enhanced service delivery, and improved patient outcomes in rural and remote Australasia.
External Quality Assessment (EQA) is an integral part of the quality assurance of laboratory performance. Laboratories are accredited against specific international standards, and medical diagnostic laboratories in Australia are accredited to the standard ISO15189. This standard requires that laboratories enrol in a suitable EQA program for all tests for which they release results, if such a program is available. Medicare rebates for pathology tests are only available for tests performed in accredited laboratories.
RCPA Quality Assurance Programs (QAP) provides a comprehensive range of EQA for virtually all pathology disciplines, incorporating both proficiency testing and educational components. Originally serving laboratories in Australia and New Zealand, the Company initially expanded to providing programs for laboratories in SE Asia, but soon found that there were a small number of laboratories in a wide range of other countries that wanted to use some of the programs. More recently there has been a concerted effort to market the programs into other jurisdictions, notably in India and the Middle East, using local agents. This has presented a number of logistic challenges, and required an understanding of the laboratory systems in these countries and the regulatory requirements or aspirations of their Governments.
The QAP has developed into a highly-respected and professional organisation which occupies a unique position in the maintenance and improvement of laboratory quality. It has recently diversified into areas of financial benchmarking (BiP), pre- and post-analytical error analysis (KIMMS), remote indigenous health testing (QAAMS), point-of-care testing (POCT) and the broader area of poorly performing laboratories. The QAP has established an enviable position of independence, and has highly trained professional staff that enable the organisation to make a significant contribution to the pathology industry in many areas.
Interest from within the industry and accreditation bodies concerning the absence of data on the accuracy of results produced by operators of the CSL SVDK (Snake Venom Detection Kit) prompted the RCPA Transfusion QAP to investigate the technical aspects of establishing a suitable external quality assurance program. The aim of the program would be to assess kit performance as well as provide on-going operator training and assessment for a test that may be performed infrequently in many facilities.
A letter was sent to targeted institutions throughout Australia informing interested personnel of the possibility of a snake venom detection QAP commencing in 2011. 111 respondents were interested in participating in a survey and 106 indicated that they would be willing to participate in the pilot. The test sample for the pilot was a standard dacron microbiology swab dipped in a weak solution of brown snake venom that was provided by CSL. A total of 82 pilot participants returned results for analysis. Participants were asked to identify the immunotype together with providing other information such as type of wash solution, reaction strength and timing of test reaction. Participants were also asked to comment on the clinical interpretation.
All participants correctly identified the test swab as brown immunotype. One participant failed to get a reaction but commented that it was used as a training exercise with an expired kit.
The test was well performed by the pilot group with all participants correctly identifying the test venom as brown immunotype.
The RCPA BioSecurity QAP has been established by a request from the Department of Health and Ageing to provide a proficiency testing program for Security Sensitive Biological Agents (SSBAs). Commencing in 2009 the focus has been on the preparation of surveys for Bacillus anthracis. This bacterium has been used in bioterrorism incidents and is an organism that laboratories may not encounter in the normal course of their work. The surveys are available to laboratories within Australia that have facilities for handling SSBAs.
Program purpose is to:
Performing these functions quickly and accurately would result in a significant reduction in loss of life.
Complies with ISO/IEC 17043:2010 Conformity assessment - General requirements for proficiency testing. Specimens are cultures (pure or mixed), blood, cerebral spinal fluid, swabs, white powders, environmental specimens, virtual microscopic images or genetic material.
Five surveys have been sent in 2009/2010.
The first surveys indicated that participants were not always aware of the unique characteristics of B.anthracis e.g. morphology and tenacity of colonies and they varied in their ability to deliver accurate results for those tests that they knew were indicative of B.anthracis e.g. motility, haemolysis and penicillin sensitivities.
In performing these surveys the participants have become increasingly consistent and confident in their ability to confirm or rule out B.anthracis.
This controversial program has been accepted by participants and has developed a sound reputation within the biosecurity community.
The explosion of POC-INR monitors onto the market place has raised some challenging EQA issues. Due to the importance of the INR result in monitoring of oral anticoagulation for prevention of recurrent thrombosis or bleeding, it is essential to ensure that these instruments and their operators provide correct results. Although the procedure is relatively simple for an experienced operator there can still be unmonitored erroneous results. Most monitors have built-in internal QC checking the quality of the strips or cartridge before testing. However, there is no procedural check equivalent to EQA to check the operator/instrument overall and ongoing performance. Although most POC monitors require a unique sample type for EQA surveys, several EQAs, such as the RCPA QAP have successfully provided survey programs for select monitors including the Roche Coaguchek XS and the ABBOTT i-STAT. Currently there is no Australian government regulation to encourage users of these instruments to participate in EQA.
The POC INR program is dispatched 5 times a year with 2 samples in each batch. The samples are lyophilised artificially depleted plasma with a range of INRs between 1.0 and 4.0.
118 users participated in the 2010 program including laboratories, doctors’ surgeries, operating theatres and pharmacists. Analysis of results is by robust statistics and participants are sent an interim report every 2 months and an End of Cycle report at the completion of the year. The reports include peer group assessments and lot/instrument/cartridge comparisons.
External quality assurance is important for users of POC INR devices. Despite specific challenges, a successful EQA program has been developed for two commonly used devices to help ensure that results are clinically safe.
Virtual Microscopy is rapidly gaining momentum as a valuable diagnostic and educational tool in laboratories and educational institutions. Although not yet widespread in diagnostic laboratories its use is increasing particularly in networks with remote laboratories. Digital images comprise a significant component of RCPA QAP Anatomical Pathology surveys and are produced with a virtual microscope in their facility. This enables a much wider range of material to be used because it only requires a small amount of diagnostic sample. The technical requirements are higher for producing scans of blood films, cytology and microbiology smears, and the RCPA QAP is now producing high quality scans of blood and other diagnostic material with a more ‘manual’ higher resolution scanner which is located within the RCPA Haematology QAP.
The Virtual Microscopy project officer works with scientists from different branches of the RCPA QAP to produce digital images of a range of samples. These images are being used in some QAP surveys, e.g. Serology, Cytology (trial survey) and Malaria, as the basis for educational packages (Malaria) and as an adjunct to glass slide surveys (Haematology Morphology). Digital images are provided either on DVDs or online. The additional features such as the ability to insert annotations, record measurements, track the viewing of the scan and take snapshots will also be described.
Although digital images cannot replace the use of glass slides, we anticipate they will be used increasingly in diagnosis, education and proficiency testing, not just because they can be transferred electronically, but also because of the small amount of material required to provide cases for hundreds of laboratories and the range of value-adding features that can be included.
The RCPA Cytopathology QAP encompasses both gynaecological and non gynaecological cytology and provides testing material to more than 200 laboratories within Australia and overseas. Testing material comprises glass slide surveys, staining exercises and questionnaires on laboratory practice. Laboratories may enrol in any or all of four cytology modules: conventional gynaecological, liquid based gynaecological, general non-gynaecological and fine needle aspiration.
The provision of a program comprising solely of glass slides raises specific challenges for cytology proficiency testing. Gynaecological slides cannot be manufactured or purchased. Often the paucity of cellular material in non-gynaecological specimens means preparing sufficient slides is may not be possible. The Cytopathology QAP is entirely dependent on members of its advisory committee and other participants to donate suitable material for surveys. In recent years, the program has encountered some difficulty in obtaining enough high quality cases with sufficient slides to ensure the timely distribution of both gynaecological and non gynaecological slide surveys. The use of Virtual Microscopy in proficiency testing is not new but presents some specific challenges to cytology proficiency testing. In contrast to histological sections or haematological smears, the combination of a high degree of depth of focus and the presence of hyperchromatic crowded epithelial cells in cytology preparations can make digitising the whole slide difficult. The challenge to reliably create a virtual proficiency test for Cytopathology that closely mirrors the actual work environment will be discussed as well as alternative strategies currently being considered for the use of virtual microscopy in the Cytopathology QAP.
The primary reason for performing serum protein electrophoresis (SPEP) and urine protein electrophoresis (UPEP) is to detect monoclonal immunoglobulins associated with plasma cell dyscrasias and lymphoproliferative disorders. As any laboratory assay report needs to deliver clear information to the clinician to assist them in the management of their patients, an understanding of the clinical requirements of a protein electrophoresis report is essential. Clinicians are primarily interested in whether a paraprotein is present or not, and if present, how large it is. Given the role of protein electrophoresis in monitoring plasma cell dyscrasias (and current information technology), the ability to view a cumulative report is also essential. In cumulative reports each paraprotein description should be stable and consistent, particularly if there is more than one paraprotein. Reports must contain adequate information to enable assessment of disease response to therapy. Several recent publications have defined and highlighted the clinical importance of response criteria in the monoclonal gammopathies including partial response, very good partial response, near complete response and complete response. For example, the latter two response categories require performance of immunofixation electrophoresis on samples where a paraprotein was previously detected in order to demonstrate its absence. Other specific circumstances where the approach to SPEP reporting varies considerably and may be a source of confusion to clinicians include paraproteins migrating in the non-gamma regions, small abnormal bands appearing after haematopoietic stem cell transplantation and first presentations of small abnormal bands in patients without a known paraprotein. Commenting also needs to take into account the level of knowledge of the requesting clinicians and their patient demographics. For example, commenting for general practitioners may differ from commenting for specialist physicians in tertiary level hospitals. Occasional samples can be more complex and require individualised commenting. A uniform approach to the reporting of serum and urine protein electrophoresis addressing these clinical requirements would be beneficial.
First trimester screening (FTS) markers of aneuploidy have been associated with other adverse pregnancy outcomes. Pregnancy-associated plasma protein A (PAPPA) levels have been linked to small or large for gestational age babies, gestational diabetes and preterm deliveries. Maternal BMI and ethnicity are also reported to be associated with gestational diabetes, hypertension and macrosomia. We examined the relationship between maternal BMI, ethnicity and serum PAPPA levels as the first step to investigate the potential combined predictive values of these markers.
Serum was collected from women presenting for routine FTS in our laboratory. PAPPA was measured using Immulite assay and multiples of the medians (MoMs) were calculated using the PRISCA software. Height, weight, BMI and ethnicity were recorded for each patient. Fetal outcomes data is being obtained from the Obstetrix database.
Of 1450 pregnancies, 1241 and 209 were Caucasian and Asian respectively. BMI was significantly higher in Caucasian women (23.61 ± 0.11 vs 22.43 ± 0.27; p<0.001). The significant differences in PAPPA levels between Caucasian and Asian women (1.92 ± 0.04 vs 2.39 ± 0.11; p<0.001) resolved when PAPPA MoMs were corrected for weight and ethnicity. The median PAPPA MoMs of the same Caucasian population (1.03) were not significantly different (p=0.552) to the Asian population (1.01). The median for the weight corrected PAPPA MoMs for the total 1450 pregnancies was 1.02; the median PAPPA MoMs for BMI <20.0 was 1.05 (n=203) and for BMI>30.0 was 1.04 (n=96).
If weight and ethnic corrections of PAPPA MoMs have any predictive value for adverse pregnancy outcomes it will be separate and independent of BMI. The utility of PAPPA and BMI in predicting maternofetal outcomes is being further investigated.
HbA1c is recommended for screening/diagnosis of diabetes in non-pregnant adults in the US. Here, we aimed to examine whether HbA1c can be used for screening/diagnosis of gestational diabetes mellitus (GDM), classified using both the current and the new international criteria.
Data from pregnant women (n=96) who had a 75g-OGTT and an HbA1c (Bio-Rad Variant-II Turbo analyser) performed within 2-weeks of OGTT were reviewed. GDM was classified using the current criteria (FPG≥5.5 or GLU-2h≥8.0 mmol/L) and the new criteria (FPG≥5.5; or GLU-1h ≥10.0; or GLU-2h ≥8.5 mmol/L). Prevalence of GDM with either criterion was calculated for each defined HbA1c group.
In this group, 53 (55%) had GDM based on the current criteria and 52 (54%) with the new criteria. People with GDM had significantly higher HbA1c than those without (GDM-current: 5.69 (mean) ±0.60% (mean±SD) vs 5.32±0.38%, respectively, p<0.01; GDM-new: 5.74±0.58% vs 5.37±0.38%, respectively, p<0.01). Prevalence of GDM by HbA1c is listed below:
|HbA1c||Prevalence of GDM, n (%)|
|%||n||Current Criteria||New Criteria|
|≤5.0||14||3 (21%)||0 (0%)|
|5.1||7||0 (0%)||0 (0%)|
|5.2||5||0 (0%)||0 (0%)|
|5.3||10||7 (70%)||6 (60%)|
|5.4||14||7 (50%)||6 (43%)|
|5.5||10||10 (100%)||10 (100%)|
|5.6–6.0||22||17 (77%)||22 (100%)|
|>6.0||14||14 (100%)||14 (100%)|
There is a significant difference in the mean HbA1c from those with or without GDM. Using the new criteria, GDM is unlikely when HbA1c is ≤5.2% but is highly probable when HbA1c is ≥5.5%.
This preliminary study, although based on small sample size, suggests that HbA1c may have a clear potential for screening/diagnosis of GDM. Further studies are needed to explore this issue including linking HbA1c to pregnancy outcomes and the future development of diabetes in the mother.
The short Synacthen© test (SST) is used for assessing the integrity of the hypothalamic–pituitary–adrenal axis. Measuring basal serum cortisol and adrenocorticotrophin hormone (ACTH) simultaneously and using the cortisol: ACTH ratio as a screening test may be a more convenient alternative to performing a SST. Our aim was to assess the reliability of the cortisol:ACTH ratio for screening for primary hypoadrenalism.
We performed a retrospective review of patients who had a SST together with a simultaneous baseline ACTH performed from April 2006 to July 2009. Their baseline cortisol:ACTH ratio was calculated and their final diagnosis was determined by examining their clinical notes. We also retrospectively reviewed the pre-treatment cortisol:ACTH ratio of 10 patients with known primary hypoadrenalism to validate its reliability.
Eighty-five patients were included. Five patients had primary hypoadrenalism, 26 patients had secondary hypoadrenalism and 54 patients had normal adrenal function. All the patients with primary hypoadrenalism had a cortisol:ACTH ratio <25. All 54 patients with normal adrenal function had a cortisol:ACTH ratio >25. Of the 26 patients with secondary hypoadrenalism 17 had a cortisol:ACTH ratio >25 and nine had a ratio <25. Therefore, a cortisol:ACTH ratio <25 had a 100% sensitivity and 89% specificity for the diagnosis of primary hypoadrenalism in this population. All 10 patients with known primary hypoadrenalism also had a pre-treatment cortisol:ACTH ratio <25.
Cortisol:ACTH ratio >25 reliably excluded primary hypoadrenalism in our study population. Cortisol:ACTH ratio <25 very likely indicates primary hypoadrenalism although secondary hypoadrenalism can not be excluded. Larger studies are needed to confirm whether cortisol:ACTH ratio is sufficiently reliable to replace SST for the diagnosis of primary hypoadrenalism.
α-1 antitrypsin (AAT) is a serine protease inhibitor encoded by SERPINA1 gene (also known as PI). AAT deficiency increases the risk of emphysema and liver disease. Severe AAT deficiency mostly arises when two copies of the PI*Z allele have been inherited. Very low AAT levels may also occur when a single PI*Z allele has been co-inherited with a PI*Null allele, however these compound heterozygous cases are not readily distinguished from PI*Z homozygotes on the basis of serum protein levels or isoelectricfocusing electrophoresis. We report here a 43 year old male with early onset degenerative lung disease in whom a very low AAT level was identified together with an unusual MZ genotype.
TaqMan PCR was used to screen for the common PI*Z (GLU342LYS) and PI S (GLU264VAL) gene variants. DNA sequencing of the AAT gene locus (14q32.1) was performed to identify rarer gene variants.
The AAT level was 0.2g/L (0.9–2.0). TaqMan PCR yielded a PI*MZ result. DNA sequencing demonstrated a PI*Z allele along with a three base pair deletion (c.486_488del3), which resulted in loss of a phenylalanine at codon 52. This variant is known as PI*M(Malton).
Sequencing revealed an unusual PI*M(Malton)Z genotype in a middleaged male with severe pulmonary disease and a very low AAT level. The PI*M(Malton) has been previously shown to be associated with hepatocyte inclusions and impaired secretion as a result of AAT protein misfolding, akin to the misfolding associated with the PI*Z allele. This variant would have been missed if the gene had not been sequenced.
The BCHE gene mediates synthesis of butyrylcholinesterase (BChE) in the liver. While the precise physiological role is yet to be established, BChE is important in the metabolism of numerous natural and synthetic drugs (pesticides, acetylcholine, cocaine, mivacurium and succinylcholine). Impaired enzyme activity is usually of little consequence, but is associated with serious risk of prolonged apnoea following administration of the muscle relaxant suxamethonium.
Reduced BChE activity is strongly associated with the Atypical (Asp70Gly), FLUORIDE1 (Thr243Met), FLUORIDE2 (Gly390Val) and quantitative K-(Ala539Thr) gene variants. Compound heterozygotes for the Atypical and K-variants, exhibit lower activity than simple Atypical heterozygotes. We report a novel BCHE genotype in a 42 year old female with very low BChE activity.
DNA sequencing of the BCHE gene locus, located on chromosome 3(3q26.1–26.2), was performed to identify rare gene variants.
A middle-aged female demonstrated consistently low BChE activity over 12 years (2.0kU/L [5.2–12.0kU/L]). Genotyping yielded compound heterozygosity for a novel missense mutation, Phe43Ile (NM_000055.2:c.211T>A;NP_000046.1:p.Phe71Ile), in combination with the K-variant. Further studies revealed a son, with borderline BChE activity (5.3 kU/L), heterozygous for Phe43Ile; and a daughter, with BChE of 6.0 kU/L, homozygous for the Kvariant. Testing revealed that maternal variants were located on alternate alleles and that, in the heterozygous state, the novel variant exerted minimal influence on total enzyme activity.
The novel Phe43Ile variant in the simple heterozygous state does not cause a significant reduction in BChE activity. However, when co-inherited with the K-variant, the Phe43Ile variant is associated with a marked reduction in enzyme activity.
A wide range of neurotransmitters, metabolites, cofactors, amino acids, vitamins and inflammatory markers were measured in cerebrospinal fluid (CSF) from children and adults using HPLC analysis.
Deficiencies in Tyrosine Hydroxylase (TH), Aromatic Amino Acid Decarboxylase (AADC), and Tetrahydrobiopterin levels associated with severe movement disorders were identified by CSF metabolite analysis and confirmed by genetic testing. In inflammatory diseases neopterin is elevated and tryptophan may be lowered due to cytokine induction of GTP-cyclohydrolase, and indolamine dioxygenase. CSF analysis was compared with MRI brain scans and neurological assessments.
CSF neurotransmitters and biopterin are low in metabolic diseases causing neurotransmitter deficiencies and are also low in neurodegenerative conditions where there is loss of the presynaptic terminals which produce neurotransmitters. Neurotransmitter abnormalities were found in 44% of 67 infantile spasms cases with little evidence for inflammatory involvement, suggesting this disease is due to pre-existing brain structural abnormalities. Of seven severe movement disorder patients with CSF HVA/5-HIAA ratio below 1.5, four had normal MRI and TH deficiency, while the remaining three showed brain stem destruction on MRI. Ongoing elevated neopterin levels and changing neurotransmitter levels over time have been used to follow the progression of Encephalitis Lethargica, an autoimmune disease targeting NMDA receptors. Severe movement disorders in conjunction with normal CSF neurotransmitters are rare and may indicate postsynaptic signal losses.
CSF trace-level analysis helps elucidate the biochemistry and physiology of the brain, allows identification of genetic diseases, helps distinguish between metabolic and neurodegenerative brain diseases, is more sensitive and specific than MRI in detecting and identifying neuronal losses and is useful for diagnosis and monitoring inflammatory brain diseases.
The Critical Difference (CD) is the smallest difference between two results which is likely to indicate a true change in the patient. Current calculations use the initial result as the best estimate of the patient’s homeostatic set point, with an equal chance that a following result will be higher or lower due to random changes. However when the initial result is above the population mean, it is most likely to be a high result from a patient with a set point closer to the population mean. Population and result distributions are used to estimate the patient mean and this approach is confirmed with patient data.
For a given initial result, the patient set point was estimated by combining the probability distributions of the result (initial result as centre; CVtot as SD, CVtot=√(CVa2+CVi2)) and the population (population mean as center; CVg as the SD). The CD calculated as +/− 2.77 x CVtot from the estimate of the patient set point. This theoretical approach was modelled in Excel and validated using data from pathology data mining as previously described (Jones Clin Biochem Rev 2007,S28).
The model showed that the estimated patient set point was always closer to the population mean than the initial result. The CD is smaller moving away from the population center than when moving towards. The model provided a good prediction of results seen for analytes with a Gaussian distribution.
The approach described here shows that it is more probable for a second result on a patient to be closer to the population mean than further away. CD calculations should be developed to take this effect into account.
Folate is primarily ordered to investigate macrocytic anaemia. The prevalence of folate deficiency (FD) has declined substantially in Australia since volunteer folate fortification from 1995 and is expected to decline further with mandatory fortification introduced in 2009. Folate can be measured in red cells (RCF) or serum (SF). RCF is routinely measured in Australia while SF is used in the UK. We examined FD prevalence and the concordance of RCF and SF in identifying FD in community and hospital settings.
Community setting: FD Prevalence was calculated using 104,080 consecutive RCF results from Melbourne Pathology. Concordance was assessed with paired RCF & SF (n=215). Reference intervals (RI) (Roche E170) were: RCF 800–3000 and SF 6–45 nmol/L.
Hospital setting: FD prevalence was determined on 100 serial adult admissions to Monash Medical Centre. Concordance of RCF & SF was evaluated in separate 114 patients. RI (Beckman DXI800) were >380 and >6.8 nmol/L, respectively.
Community setting: FD prevalence was 0.2%. The RCF & SF gave concordant results in classifying people with or without FD except in one patient. Hospital setting: Prevalence of FD was 1%. Of the pairs results, all had normal RCF but 4 had slightly low SF.
The prevalence of FD is now extremely low in both community and hospital populations as compared to the 8.5% prior to folate fortification. RCF assays are more costly, labour intensive and have much higher imprecision (RCPA/QAP data) than SF assays. Furthermore, RCF results can be misleading at high haematocrit levels probably due to incomplete in vitro haemolysis prior to analysis. In view of low prevalence of FD, the good concordance of RCF and SF and the analytical issues, we propose that SF be the first-line screening test for FD and reserve RCF for confirmation only.
Pathology Queensland consists of 33 laboratories spread throughout Queensland. Between 2007 and 2008, a common instrument platform was rolled out in all Chemical Pathology laboratories across the state. Tests now had the same collection procedures and could be reported with the same reference ranges and units across the State, a distinct advantage for mobile clinicians and mobile patients. A test result from Thursday Island was now directly comparable with a test result from Toowoomba or the Prince Charles Hospital. Or was it? How could we demonstrate consistency in accuracy and precision across all laboratories when each managed their own internal quality control program?
The answer was to move to the same quality control material and establish state-wide limits of performance for this material based on clinical need rather than traditional statistical analysis. Initially 28 analytes were moved to the new quality control system. The running mean and CV of each analyte at each site was calculated monthly (pre- and post- change) and analysed against a set of performance indicators.
Analytes at some sites initially fell outside the new limits. Priorities were set to determine and resolve the causative issues. Reagent variability was expected to be a major cause but was not. Overall an improvement in accuracy was seen in 10 analytes. No analytes demonstrated a loss of precision. Statewide limits have been extended to all quality control material in use.
Statewide standardisation is achievable. It provides a means of demonstrating accuracy and precision in our testing across the State. It improved laboratory processes, fostered a sense of unity and cooperation amongst laboratories and most importantly, improved the quality of our service.
Common liver diseases such as chronic viral hepatitis types B (HBV) and C (HCV) and non-alcoholic fatty liver disease (NAFLD) often require staging of liver fibrosis to guide treatment decisions. Currently, staging involves an invasive liver biopsy. Hepascore is a model which uses four serum biochemical markers to calculate a composite score which is predictive of liver fibrosis.
HBV (194 patients) and HCV (104 patients) were confirmed by virology testing and NAFLD (119 patients) by liver biopsy. All patients had liver biopsies staged for fibrosis and significant fibrosis was defined as stages F2 to F4. The prevalences of significant fibrosis were 42% in HBV and NAFLD patients and 57% in HCV patients. Serum taken at liver biopsy was analysed for bilirubin, γ-glutamyl-transferase, hyaluronic acid and α2-macroglobulin. Hepascore (range 0.0–1.0) was calculated and a 0.50 cut-off applied as follows: scores ≥0.50 indicated significant fibrosis was present and scores <0.50 indicated significant fibrosis was absent.
In both HCV and NAFLD patients, a Hepascore ≥0.50 had a positive predictive value of 88% for significant fibrosis. Hepascores <0.50 were less accurate for excluding fibrosis; the negative predictive values were 65% for HCV and 75% for NAFLD. In HBV patients, Hepascores <0.50 excluded fibrosis with a negative predictive value of 83%, whereas the positive predictive value of a score ≥0.50 was 68%.
Hepascore had a very good positive predictive value for predicting the presence of significant hepatic fibrosis in NAFLD and HCV patients. In HBV patients it was more effective for the exclusion of fibrosis.
A rise and/or fall in troponin concentration (delta troponin) is a recognised tool to assist in separating acute myocardial infarction (AMI) from other causes of troponin elevation. The factors which contribute to background variation are generally considered to be biological and analytical variation. In this study we investigate the effect of some pre-analytical factors on delta troponin.
The SydPath database was searched for samples with troponin and albumin results collected in the Emergency Department within 8 hours of each other on the same patient. Changes in albumin concentration were investigated as a marker of changes in sample hydration due to pre-analytical factors.
Over 5 years a total of 15,639 samples with both troponin and albumin from the ED were identified. Of these only 900 pairs fulfilled the complete selection criteria. The median albumin on the first sample was 41 g/L. For samples separated by between 4 and 12 hours there was a median fall in albumin concentration of 9% (10th and 90th centiles +4% and −24%). The extent of the fall increased with the time between collections up to 4 hours and then remained stable.
The fall in albumin concentration is likely to reflect a combination of artefactual haemoconcentration of the first sample due to prolonged use of the tourniquet, such as commonly occurs in our institution when an intravenous cannula is inserted, and intravascular fluid increase in the second sample due to recumbency or iv fluid administration. The study is limited by the small number of data points and lack of systematic data collection. The effect described may be relevant when assessing values for delta troponin.
Since 1999, the RCPA Quality Assurance Program (QAP) has conducted an assessment of analytical performance for retinol, α-tocopherol and β-carotene. Previous target setting exercises have been carried out with National Institute of Standards and Technology (NIST) standard reference material (SRM) 968c, provided at two levels; with targets successfully set for retinol but not α-tocopherol or β-carotene. In 2009 NIST released a new, single level, SRM 968d. Here we describe the outcome of the 2010 target setting exercise with NIST SRM 968d.
Laboratories participating in this QAP utilise reverse phase HPLC with spectrophotometric detection. Target setting is conducted using a standardised procedure where selected laboratories are supplied with material for the forthcoming cycle along with the SRM, in this case NIST SRM 968d. The QAP target values are set using the SRM as calibrator.
Assigned values for NIST SRM 968d are: total retinol 1.09±0.17μmol/L; α-tocopherol 13.77±1.28μmol/L; β-carotene 0.145±0.013μmol/L. The linear regression analysis of the NIST SRM 968d target values against all results returned by laboratories in cycle 22 of the RCPA Vitamin QAP are: retinol y=1.0032x–0.0099, r2=0.9987; α-tocopherol y=0.8020x+1.5772, r2=0.9964; β-carotene y=0.8031x–0.1065, r2=0.9914.
Target setting was successful for retinol, but unsuccessful for α-tocopherol and β-carotene. The concentration of β-carotene in NIST SRM 968d is too low. α-Tocopherol is assessed in the NIST material by HPLC with fluorometric detection, which is different to the detection method of QAP participants. Furthermore, the allowable error reported by NIST (≈10%) approximates the total allowable limits for the QAP.
NIST SRM 968d is not usable as a primary calibrator for α-tocopherol and β-carotene.
First trimester screening involves measuring nuchal translucency (ultrasound), free beta human chorionic gonadotropin (Free βhCG) and pregnancy-associated plasma protein A (PAPP-A). Biochemical data is corrected for gestation and weight to produce Multiple of Median (MoM) using specific populations. The aims of this study were (a) determine twin specific medians and compare them to singleton pregnancies and (b) determine how twin specific medians in this population compared to published data.
In the period 25/10/2005 to 30/12/2009, 283 twin pregnancies and 45913 singleton pregnancies were measured for Free βhCG and PAPP-A using immunoassay (Siemens Immulite 2000). Gestation medians were calculated and compared at each week (10–14) for twin and singletons pregnancies. This data was used to derive the twin specific regression equations and then MoMs calculated. This MoM data was then compared to published studies.
Serum concentrations of Free βhCG and PAPP-A were higher in twin compared to singleton pregnancies. Median Free βhCG concentrations were 87.0, 77.6, 69.6 and 64.6 in twin and 52.6, 45.1, 38.2 and 32.3 IU/L in singleton pregnancies at 10, 11,12 and 13 weeks whilst median PAPP-A concentrations were 1.90, 2.65, 3.82 and 7.64 in twin and 0.86, 1.33, 1.93, 2.82 IU/L in singleton pregnancies at 10,11, 12 and 13 weeks respectively. MoMs calculated for both markers using the derived equations correlated well compared to published studies using different twin populations.
Free βhCG and PAPP-A concentrations in twins were higher compared to singletons pregnancies. Derived equation Free βhCG and PAPP-A MoM data correlated well with published studies. Twin specific regression equations developed in our population will be used for future calculations of individual MoM data.
The presence of heterophile antibodies in serum can induce false elevations of analytes measured via immunoassays. Commercial automated immunoassays are formulated to reduce the effect of heterophile antibodies, however the potential for heterophile antibody interactions is not completely eliminated. We present a case where interference in the 1–84 parathyroid hormone (PTH) assay led to falsely elevated PTH levels and an incorrect diagnosis of primary hyperparathyroidism. A 50 year old man had a PTH of 26 pmol/L (reference range 1.0–6.8) measured on the Siemens Healthcare Diagnostics Immulite 2500. The corrected calcium was 2.37 nmol/L (2.20–2.55). He underwent surgical exploration; biopsies on the parathyroid glands confirmed normal parathyroid tissue.
A further sample was collected on this patient. Plasma samples were analysed on the Abbott Diagnostics Architect i2000SR and Siemens Healthcare Diagnostics Immulite 2500. The nature of the interferant was investigated via serial dilutions analysis and treatment with Heterophile Blocking Tubes (HBT, Scantibodies Laboratory).
Neat plasma PTH level measured on the Architect i2000 SR was 4.8 pmol/L whilst on the Immulite 2500, PTH was >263 pmol/L. The discrepancies were investigated via serial dilutions on the Immulite 2500 and showed a non-linear correlation, consistent with positive interference. Treatment of the plasma with HBT gave a PTH level of 5.6 pmol/L on the Immulite 2500, confirming the presence of heterophile antibodies.
Antibody interference should be considered when PTH results are inconsistent with clinical findings. While the potential for antibody interference in immunoassay is well known, there have been few reports of this problem in PTH assays. Close clinician-laboratory communication is essential to present incorrect diagnosis and potentially harmful consequences for patients.
Blood has been the accepted routine specimen worldwide for assessing the secretion rate of progesterone. Whether salivary and urine progesterone assays are similarly useful needs to be investigated, as these specimens are non-invasive and more convenient for frequent collections. In this study, blood, saliva and urine progesterone were compared to see whether they exhibit similar patterns during the menstrual cycle.
Serum, urine and saliva samples were collected simultaneously from 12 premenopausal females, twice, two weeks apart, for the measurement of progesterone. Serum progesterone was assayed using Architect ci8200 Chemiluminescent Microparticle Immunoassay technology (Abbott Diagnostics). Salivary progesterone was analysed using the DRG progesterone ELISA kit. Urine pregnanediol (the main progesterone metabolite) was assayed using an Agilent 6890 Gas Chromatograph coupled to a 5973 Mass Selective Detector. Statistical analysis included the Wilcoxon sign rank test, Friedman test, paired Student t test, and McNemar test, performed with SPSS version 16.
The concentration of progesterone found in the follicular phase was: serum <0.30–2 nmol/L; saliva 70–150 pmol/L; and urine (pregnanediol) 0.02–0.26 μmol/mmol creatinine. Comparatively the concentration of progesterone in the luteal phase was: serum 4.8–26 nmol/L; saliva 250–780 pmol/L; and urine (pregnanediol) 0.4–1.1 μmol/mmol creatinine.
A strong correlation between serum, saliva and urinary progesterone levels was demonstrated. Statistically, there was no difference in the proportion of results interpreted as follicular phase (or similarly, luteal phase) between saliva or urine versus serum; all p values >0.05.
Both saliva and urine are good alternative specimens to serum in assessing progesterone production and stage of the menstrual cycle.
This study aimed to assess glucose homeostasis in a healthy paediatric population and observe changes between ages 8, 10 and 12 years.
Using a population of 753 8y.o. in 2005 enrolled in the LOOK Study, a baseline analysis of glycated haemoglobin A1c (HbA1c), fasting serum glucose and insulin was performed with a calculation of Homeostasis Model assessment for insulin resistance (HOMA-IR). Measurements were repeated from the original population on 594 10y.o. in 2007 and on 501 12y.o. in 2009.
Significant changes were seen in fasting glucose levels between all age groups and between genders at ages 8 and 10 years, but not 12. A significant difference was also observed for HOMA-IR as both sexes grew older. Girls had a two fold increase compared with boys. At age 12 more girls had a HOMA-IR >2 than boys aged 12 or girls aged 10.
By comparing serial changes in HOMA-IR and fasting glucose with the lack of any change in HbA1c we show that insulin sensitivity decreases from age 8 to 12 in boys and girls and the greater decrease in girls aged 12 may reflect the subtle physiological moderations associated with growth and more latterly puberty.
New generation troponin assays have emerged, with progressively lower detection limits for which we have evaluated their ability to predict clinical outcomes up to 2 years.
We studied 441 patients admitted with chest pain suggestive of acute myocardial infarction (AMI). Blood samples for troponin (Abbott Architect TNI 2nd generation) were taken at 0 and 6–24 hours. Frozen samples were later analysed for Architect TNI 3rd generation (TNI), Roche TNT 4th generation (TNT), Roche high sensitivity TNT (hsTNT) and OCD Vitros TNI (TNI-V). Patients were followed up for 2 years major adverse cardio-vascular events (MACE), being a composite of AMI, revascularisation procedure, and/or cardiac death.
156 (35%) were diagnosed with AMI and MACE was seen in 92 (20.9%) patients at 2 years. Comparing ROC curve AUC at baseline, for 2 year MACE, hsTNT (0.75; 95%CI 0.67–0.83) and TNI-V (0.72; 95%CI 0.64–0.81) out performed TNT (0.63; 95%CI 0.53–0.64) (p=0.003 and 0.024 respectively) but were equivalent to each other and TNI (0.69; 95%CI 0.60–0.78), which did not out-perform TNT. ROC curve optimum cut-points for prediction of MACE were 0.016, 0.020, 0.011, 0.01 μg/L for hsTNT, TNI-V, TNI, and TNT respectively, all below the 99th percentile. All troponins above the 99th percentile were predictive of MACE, in multi-variate analysis, even when adjusted for age and history of ischaemic heart disease.
Detectable troponins are prognostically important, even at levels below the 99th percentile.
Red cell folate concentration may be calculated either with or without a correction factor for serum folate. The correction factor is used to account for the contribution of folate from serum to the whole blood haemolysate prepared for red cell folate measurement. Among Australian laboratories there appears to be uncertainty as to the need for this correction factor, with 26% of laboratories considering a correction factor appropriate. This study sought to clarify the need for employing a correction factor.
Data was collected from 15,485 patients requiring red cell folate determinations over a 15 month period. Red cell folate concentrations were calculated both with and without a serum folate correction factor. The difference incurred by not correcting for serum folate was assessed against the optimum criterion for bias based on biological variation. Also examined was the incidence of patient misclassification, that is the calculated red cell folate concentration falling within the reference interval only when the correction factor was omitted.
The mean difference due to not correcting for serum folate was 2.4%, easily meeting the optimum criterion for bias of 8.4%. In fact 99.7% of all individual differences fell within the criterion for optimal performance. Not correcting for serum folate resulted in 0.2% of patients being misclassified. This compares favourably to the 1.5% misclassification rate inherent to the definition of the optimum criterion for bias. Most significantly, among all patient results there was only a single instance where a difference of greater than 8.4% resulted in misclassification of the patient.
These data support the calculation of red cell folate without the need for a correction factor for the concomitant serum folate.
Opiates are a class of drugs comprising both prescribed (codeine containing medications) and illicit agents (heroin). Gas chromatography-mass spectrometry (GCMS) for opiates in urine is commonly performed to confirm heroin use for work safety or legal purposes or to monitor compliance during/after drug rehabilitation. Thus, incorrect interpretation of the results has serious consequences. GCMS is designed to detect metabolites of heroin, namely 6-monoacetylmorphine (6-MAM), codeine and morphine. Presence of 6-MAM is diagnostic for heroin use, however, the urine needs to be collected within hours of heroin use as 6-MAM has a short half-life. Here, we examined the pattern of codeine to morphine ratio (Cod/Morph-ratio) from the people who were taking either panadeine or heroin.
Over two years, 796 GCMS for opiates were performed. Medication list was obtained from the chain of custody declaration form. Overall, 153 people were reported to be taking codeine-containing medications alone – the ‘Panadeine group’; 40 people had detectable 6-MAM – the ‘Heroin group’. The GCMS cut-off: 300 μg/L for morphine or codeine each.
The median (25th to 75th percentile) Cod/Morph-ratio and the median (25th to 75th percentile) morphine were calculated based on the codeine cut-offs for each group. Results are below:
|Cod/morph ratio||0.0 (0.0–0.1)||0.1 (0.1–0.4)||0.1 (0.1–0.2)||0.2 (0.2–0.4)|
|Morphine (μg/L)||352 (176–1830)||1703 (773–2579)||3621 (2834–4385)||9908 (4238–16970)|
|Cod/morph ratio||0.3 (0.1–0.9)||0.9 (0.6–2.1)||2.0 (1.0–3.5)||3.4 (1.8–6.1)|
|Morphine (μg/L)||150 (64–243)||141 (66–190)||357 (178–621)||809 (478–2359)|
The overall results usually show a clear separation. In the heroin group, the ratio was usually very low (<0.5). When codeine was >300 μg/L, the morphine level was very high (above 2800 μg/L). In the panadeine group, the ratio was typically >0.9 with the corresponding morphine level commonly <800 μg/L. When the codeine level was >300 μg/L, the ratio is commonly >2.0.
HbA1c is widely accepted as requiring further international standardisation and calibrator traceability to provide consistency in long-term monitoring of diabetic control. In 2009 the Western Australia (WA) QC Subcommittee surveyed methods across public and private pathology providers in WA using pooled patient samples.
Seven linearly related EDTA preserved pooled whole blood samples were distributed to nine participating laboratories for duplicate HbA1c assay. Three commercial QC samples and two samples containing Hb variants (HbS and HbE) were included. Methods were either immunoassay (IA) (Roche Cobas Integra 800, n=4; Siemens Advia 1650, n=1; Siemens DCA Vantage, n=1) or HPLC (Bio-Rad Variant 2, n=2; (Bio-Rad Variant 2, n=2; Primus Diagnostics Ultra 2, n=1).
All labs quote HbA1c ranges as: non-diabetic <6% (<42 mmol/mol), good control 6–7% (42–53 mmol/mol) and action required >8% (>64 mmol/mol). Inter-laboratory imprecision across all nine labs for the seven samples ranged from 4.9% to 9.2%. HbA1c on the patient pool sample with a mean of 7.75% (61 mmol/mol) gave an average HPLC result of 7.1% (54 mmol/mol) and an average IA result of 8.1% (65 mmol/mol). This difference was not consistent with the QC material results (mean difference <2%). Results were more discordant at higher HbA1c concentrations. HbA1c results for the Hb variant samples were also highly method dependent.
HbA1c results in Western Australia fall into two distinct method dependent subgroups. This is reflected in the RCPA QAP scheme with immunoassay methods giving higher results than HPLC. An average 13% difference in results between methods is significant in long term monitoring of diabetic control. The improved standardisation of methods and calibrators should significantly reduce these differences.
The CKD Epidemiology Collaboration (CKD-EPI) group recently published a new equation to overcome the limitation of the Modification of Diet in Renal Disease (MDRD) equation, which systematically underestimates glomerular filtration rates (GFRs) above 60 mL/min/1.73 m2. Our study was aimed to assess whether the new CKD-EPI equation performs better than the MDRD equation.
102 samples were collected from adult patients undergoing gold-reference radionuclide GFR measurements. Plasma samples were analysed for creatinine using the modified Jaffe and enzymatic reactions, by Abbott Laboratories and Roche Diagnostics. The eGFR values for the 4 methods were then compared with radionuclide GFR with respect to bias, correlation and accuracy in predicting GFR.
CKD-EPI equation is more accurate when using enzymatic reactions to measure plasma creatinine. Laboratories would need to change from the traditional Jaffe methods to enzymatic methods to fully benefit the use of the CKD-EPI equation.
In children with cancer, an accurate method for assessing glomerular filtration rate (GFR) is essential during treatment with nephrotoxic chemotherapeutic drugs. Determination of GFR is time-consuming and difficult to perform. Plasma creatinine is inaccurate due to low muscle mass in children, and potential assay interferences. We assess the potential utility of cystatin C as a renal marker in children with cancer, with comparisons to radio-isotope GFR and plasma creatinine.
Samples from paediatric patients undergoing radionuclide GFR assessments were collected. Plasma samples were analysed for cystatin C by two methods; nephelometry and turbidimetry (Siemens and Abbott), and creatinine by two enzymatic creatinine methods (Roche and Abbott). Calculations of bias, correlation coefficients and percentage estimates between cystatin C equations, creatinine equations were compared with measured GFR.
Mean radionuclide GFR was 128.62 (range = 19.79 to 248.78) mL/min/1.73m2. Preliminary results on 45 paediatric patients showed that cystatin C equation such as Filler et al, and creatinine equations by Counahan-Barratt and Leger et al did not produce any significant bias compared with radionuclide GFR. When comparing the proportion of results within 30% of GFR, the most accurate prediction equation was that of Zapitelli et al (82.2%), where both cystatin C and creatinine were used in conjunction to estimate GFR. When comparing cystatin C and creatinine methods, results by the Siemens method (nephelometry) and Roche Creatininase Plus method, both had higher proportion of patients within 30% of GFR.
Using plasma cystatin C in conjunction with creatinine to predict GFR may improve GFR prediction in the paediatric population. A larger sample size and more research are needed to warrant its use before adaptation into operational use.
Measurement of TPMT enzyme activity identifies genetically-determined deficiencies that strongly influence the toxicity of thiopurine drugs. A lack of consensus on a TPMT reference method, reference intervals and cut-offs has led to varying criteria for TPMT deficiency. Introduction of this assay into Western Australia necessitated the comparison of TPMT results with genotype analysis.
TPMT activity was determined in 209 people by a phase partition method, using 6- mercaptopurine as a substrate. Genotyping was conducted by initial screening of the TPMT gene for the common variants *2, *3A, *3B, *3C and *8. Multiplexed Ligase-Dependant Probe Analysis (MLPA) and direct sequencing of coding regions was performed in cases with low activity in the absence of those common variants.
The TPMT enzyme activity assay gave an intra- and inter-assay precision of 2.3% and 12.9% respectively. Correlation coefficient and least square regression of 100 TPMT samples assayed by two independent laboratories were r2 = 0.807 with a slope of ±0.869 0.064 and y intercept of 0.049. There were 24 (11%) subjects found to have TPMT activities in the lower quartile of the population range (<0.57 nmol/gHb/min). Twelve of these subjects were heterozygous for a low activity allele and 12 subjects homozygous for normal activity alleles.
TPMT activity measurement, followed by selective genotyping, allowed laboratory results to be usefully linked to guidelines on thiopurine drug dosing. Reporting rules, based on the predictive value of TPMT activity and genotype for adverse drug response, have been developed to assist clinicians with drug dosage choice.
Spot urine measurement of albumin is now the most commonly accepted approach to assessing proteinuria. Exertion prior to the collection may potentially influence the result of spot urine albumin estimation. We evaluated the effect of exercise on albuminuria in subjects at various stages of diabetic nephropathy and healthy control volunteers.
Thirty-Five diabetic subjects (19 with normoalbuminuria (NA), 9 with microalbuminuria (MA) and 7 with overt proteinuria (OP)) and nine control subjects were assessed. A 1km treadmill walk was performed. Four Spot urine specimens were collected; first morning void, immediately prior to exercise, and 1 and 2 hours after of exercise. A random effect linear regression mixed model was used to assess the effect of exercise on the urine albumin to creatinine ratio (uACR). Results are presented separately for males and female diabetic subjects due to a significant exercise/gender interaction (p<0.05).
No significant effect of exercise on uACR was seen in control subjects. In NA diabetic males no effect of exercise was seen while in females uACR 1hr post exercise was significantly higher than the early morning sample (3.55 mg/mmol, 95% CI 0.27–6.83). uACR increased 1hour post exercise in MA subjects although the effect was more marked in females (87.8, −24.3–199.4) compared to males (6.7, 2.1–11.3). For both males and females with OP, uACR was significantly increased 1 hour post exercise (67.5, 22–113 and 21.6, 8.4–34.8 respectively). In all groups uACR at 2hours post exercise was not significantly different to the early morning sample.
Exercise increased uACR estimation in subjects with diabetes with a larger effect in females. Whether exercise unmasks early diabetic nephropathy in NA subjects requires further study.
There is evidence that intravenous fat emulsion (IFE) may be helpful in amelioration of lipid soluble drug toxidromes. We describe here a procedure to determine which drugs may be amenable to IFE therapy.
Sodium valproate, a relatively water soluble drug and carbamazepine, a relatively water insoluble drug were selected to establish the protocol. Tablets were dissolved in water (valproate) or ethanol (carbamazepine). A serum sample was then ‘spiked’ to make a primary working solution. From this 4 different concentrations of the drug extending well beyond the upper therapeutic limit were prepared. Aliquots were treated with 2 different volumes of fat emulsion (20% and 33%) or saline. The lipid-free infranatant was separated by centrifugation and then further centrifuged at high speed in a Millipore Ultrafree-MC 10,000 MW cutoff filter unit to get the protein-free fraction of the drug. At each step the concentration of the drug was measured in Beckman CoulterDxC-800 autoanalyser.
Free sodium valproate was approximately 10% at therapeutic levels (<700 μmol/L) but increased dramatically (>70%) when valproate exceeded 2,000 μmol/L. However there was minimal partitioning of valproate into the lipid phase. Carbamazepine was approximately 80% protein bound but showed modest partitioning into the lipid phase. Total plasma carbamazepine was reduced by approximately 30% at all drug concentrations using the 33% lipid addition while the free fraction was reduced by approximately 25%.
Our proposed method for determining total and free drug fraction after treatment with IFE was a simple and practical way to screen drugs that may be amenable to such detoxification treatment.
Re-formulated reagents for Roche folate assays were introduced onto the Australian market in 2008. These ‘Folate III’ reagents were observed to exhibit a proportional positive bias of approximately 22% in patient results for red cell folate compared to the previous, ‘Folate II’, reagents. This study sought to re-establish reference intervals for red cell folate using ‘Folate III’ reagents on both the Roche E170 and e411 analysers.
EDTA-anticoagulated whole blood and serum specimens were collected from 240 healthy volunteers and subjects presenting for outpatient blood collection over a 4 month period who met the following criteria: aged 18–65 years, not pregnant, not consuming vitamin supplements and had a measured plasma homocysteine concentration less than 15 μmol/L. Red cell and serum folate measurements were made on the Roche E170 and e411 analysers and central 95% reference intervals were calculated using non-parametric statistics are per CLSI document C28-A2. Reference intervals for red cell folate were established both with and without a correction factor for serum folate.
The reference intervals thus established were: E170 red cell folate uncorrected for serum folate 1135–3026 nmol/L; E170 red cell folate corrected for serum folate 1105–2808 nmol/L; e411 red cell folate uncorrected for serum folate 1423–3295 nmol/L; e411 red cell folate corrected for serum folate 1403–3219 nmol/L.
These data re-establish reference intervals for Roche analysers using ‘Folate III’ reagents. Furthermore, they allow laboratories to employ red cell folate reference intervals appropriate to whether or not they utilise a correction factor for serum folate.
The measurement of red cell folate is a relatively time-consuming, and thus costly, endeavour. The stages of folate deficiency as described by Herbert portray low serum folate preceding a decrease in red cell folate. This study sought to establish whether serum folate would therefore have utility in identifying patients at increased likelihood of having a red cell folate concentration within the deficient range.
Data was collected from 15,485 patients requiring a red cell folate determination over a 15 month period. Serum and red cell folate were measured using the Roche E170 analyser. The ability of serum folate to predict a low red cell folate was established both at the cut-off routinely used by the laboratory, derived from a population-based lower reference limit, as well as at a cut-off chosen to maximise sensitivity.
A poor concordance between serum and red cell folate was demonstrated. The serum folate cut-off routinely used by the laboratory showed poor sensitivity (51%) and positive predictive value (11%) for identifying individuals with low red cell folate concentrations. Specificity (95%) and negative predictive value (99%) were reasonable. Conversely, using a serum cut-off to maximise sensitivity (<25nmol/L, to give sensitivity of 94%) lead to a concomitant decrease in specificity (59%) and positive predictive value (2%).
For a hospital-based population, serum folate appears to have limited utility in screening for patients at increased likelihood of having a low red cell folate.
Unusually high urine iodine concentrations have been observed at our laboratory. Iodine levels as high as 24 g/L have been measured (WHO guidelines: urine iodine >100 μg/L Not iodine deficient). This study tested possible sources of laboratory contamination as well as other potential sources of iodine e.g. seafood.
The baseline iodine level was measured in a urine pool before the addition of contaminants. Potential contaminants were added separately to aliquots of the pool and the iodine level was measured. Iodine in seafood was investigated following the consumption of prawns. A urine sample was collected prior to prawn consumption as well as a first morning urine sample after the prawns. Urine iodine was measured using an Agilent 7500ce ICPMS.
The urine iodine level increased following contamination of the urine pool with Multistix, potassium iodide and Betadine, but did not increase following pH strips. Following the consumption of both wild prawns and farmed prawns, the urine iodine level decreased. The urine iodine of a patient on amiodarone was 8000 μg/L. One patient with a urine iodine concentration of 8 g/L was found to have had x-rays involving iodine radiocontrast media prior to testing.
Patient preparation is essential when collecting a sample for urine iodine. Collection of urine following Xray procedures involving radiocontrast media should be avoided. The patient must avoid contact of the urine sample with Betadine antiseptic. There is potential for laboratory errors in iodine estimation if the samples are contaminated with dipsticks or other reagents in the laboratory prior to analysis. The consumption of seafood and patients on amiodarone do not appear to be the cause of the unusually high urine iodine levels observed.
The RCPA Chemical Pathology Quality Assurance Programs (RCPA QAP) provides a PoCT quality assurance program (QAP) for HbA1c testing on the Axis-Shield AfinionTM AS100 Analyser System. As this instrument requires the red blood cell membrane to be intact, the material being provided is patient whole blood samples.
Identification of suitable patients, sample collection, pre-screening, aliquoting, packing and transport to testing sites is carried out by Point of Care Diagnostics to the RCPA QAP specifications. Every four weeks two samples are distributed to participating sites. The sites analyse the samples and return the results to the RCPA QAP. An extra aliquot is sent to the European Reference Laboratory for Glycohaemoglobin to assign a reference value which is used as the target value. The acceptable range is determined from the HbA1c allowable limits of performance, ±0.5 for target values up to 10.0% or ±5% for target values >10%.
A total of 13 sites are submitting results. The target values for the samples range from 5.1–12.1%. The mean deviation from the target values for 10 samples is 0.1% with a range of 0–0.4. The mean between run %CV is 3.8% with a range of 2.5–6.2. 93% of the results are within the acceptable range, 6% are in the review area and 1% are unacceptable.
Providing a whole blood sample for the PoCT QAP participants presents a significant logistical challenge compared with providing a lyophilised sample. However, the advantage is that there are no matrix issues. The Afinion instrument has shown excellent accuracy and a suitable number of acceptable results for the point-of-care environment.
Assessment of proteinuria and albuminuria has traditionally been done using 24 h urine rather than spot urine collections with the expectation that a longer collection period will be more informative. A number of expert bodies have now recommended spot urine, especially from first morning void, as a preferred specimen and with creatinine correction to standardise the urine concentration. This, however, has not been universally accepted by clinicians. We conducted a study examining the correlation between 24 h and spot urine for protein and albumin.
All the requests over three months for 24 h urine for albumin and/or protein excretion were searched for concurrent request for mid-stream (spot) urine for microscopy and culture. Three measurements were made on the paired specimens: creatinine (Roche Jaffe), albumin (Roche immuno-turbidimetry) and protein (Roche Benzethonium Chloride). The correlations were assessed non-parametrically using Spearman’s correlation.
After excluding infection using the spot urine, 121 paired 24 h and spot urine specimens were used in the analysis. There were highly significant correlations (all p<0.001) between albumin concentration in the spot urine and albumin concentration in the 24 h urine (r=0.88) or 24 h albumin excretion (r=0.84). The highest correlation, however, was found between the albumin/creatinine ratio in the spot and 24 h urine samples (r=0.93). Similarly, there were highly significant correlations (all p<0.001) between the spot urine protein concentration and protein concentration in the 24 h urine (r=0.69) or the 24 h protein excretion (r=0.58). Again, the highest correlation was found between the protein/creatinine ratio in the spot and 24 h urine samples (r=0.85).
The results confirm the high correlation of a spot albumin/creatinine ratio or protein/creatinine ratio with all albumin or protein measurements made in 24 h urine samples. This suggests that 24 h collections usually add little more information.
A teenage male suffered severe burns after a domestic explosion accident in Tahiti. He was transferred two days later to the National Burn Centre at Middlemore Hospital, Auckland. Unphysiologically high levels of whole blood chloride were noted by Radiometer ABL 800 blood gas analyser. Paired plasma chlorides measured by Abbott Architect Ci 8200 autoanalyser were much lower. We suspected interference to the chloride electrode of ABL 800 and investigated for the possible interferent(s).
Patient’s plasma nitrate, bromide and thiocyanate were measured by ion-paired reversed phase HPLC with low UV detection at 205 nm. Plasma iodide was measured by ICPMS. The in-vitro effect of nitrate on ABL 800 and Abbott Architect Ci 8200 chloride measurements was tested by spiking blank pooled plasma with increasing concentrations of sodium nitrate.
Patient’s whole blood chloride levels (ABL 800) were 170 mmol/L, 137 mmol/L, and 119 mmol/L (from post-burn day 3,4 and 5 respectively) with corresponding plasma nitrate at 6.7 mmol/L, 4.9 mmol/L and 1.1 mmol/L. (reference limit <0.08 mmol/L). The decrease in chloride paralleled the decrease in nitrate over the five days of observation. In-vitro spiking with 7 mmol/L nitrate gave an apparent plasma chloride level of 180 mmol/L from ABL 800. Thiocyanate was not detected and other potential interferents were normal with bromide 0.045 mmol/L and iodide fluctuating <150 nmol/L.
Endogenous nitrate levels this high are not achievable by activation of NO synthase following inflammation. Enquiries with the Tahiti hospital confirmed that cerium nitrate (Flammacerium) had been applied topically to his burn wounds. No nitrate containing medications had been given in New Zealand throughout the admission till his death. We suggest systemic absorption of nitrate from cerium nitrate has caused the positive interference on chloride measurement on Radiometer ABL 800 blood gas analyser.
Instrumentation Laboratory is developing a cartridge for the GEM Premier 4000 blood gas analyser for measurement of electrolytes, glucose, lactate, haematocrit, urea, creatinine and tCO2. Measurement of urea and creatinine in patient samples are compared to current laboratory methods.
103 arterial blood gas samples obtained from ITU, Renal and Emergency wards were analysed on the GEM 4000, and plasma on a c8000 analyser (Abbott) using the Abbott urea method, the Abbott Jaffe and Roche enzymatic creatinine methods. Plasma was also analysed for urea and creatinine on the GEM 4000 to compare with blood.
Blood creatinine = 1.06 (plasma creatinine) + 11.96, with positive bias for plasma measurement, increased at higher concentrations.
84/103 patients were classified into the same CKD stage group (after eGFR MDRD4 calculation) across each method. Negative interference by creatine on creatinine measurement was observed in a septic patient with serum creatine = 1334 μmol/L by LC-MS/MS. Creatine interference is detected and flagged by the GEM 4000.
Good correlation was observed between the IL GEM 4000 and current laboratory methods for analysis of urea and creatinine in blood and plasma. GEM Premier 4000 shows promise for measurement of urea and creatinine in laboratory and POCT environments.
Total bilirubin is traditionally measured from plasma or serum samples on chemistry analysers. Recently, whole blood total bilirubin measured on blood gas and CO-Oximetry analysers is gaining attention due to the elimination of plasma/serum separation step and clinically equivalent performance to chemistry analysers. GEM Premier 4000 is a critical care analyser offering blood gas, electrolyte and metabolite measurements together with integrated CO-Oximetry. Whole blood or plasma total bilirubin measurement has been added to the menu of assays on GEM Premier 4000. Analytical performance of whole blood based GEM Premier 4000 total bilirubin measurement is compared to the plasma/serum assay on the Olympus AU680 chemistry analyser.
Heparinised blood samples were collected from neonatal and adult patients in capillary or syringe for testing on the Instrumentation Laboratory GEM Premier 4000 and simultaneously in collection tubes (Li-heparin) for plasma testing on the Olympus AU680 chemistry analyser.
Total bilirubin levels from 11–617 μmol/L (0.6–36.1 mg/dL) were tested on both GEM and Olympus AU680 for method correlation. Total bilirubin measurements on GEM Premier 4000 showed good agreement to the measurements on Olympus AU680 (y = 1.05x – 6.7, R² = 0.987, n = 50).
GEM Premier 4000 showed equivalent performance to plasma measurements on the Olympus AU680 chemistry analyser for assessing the total bilirubin levels in newborns and adults. In addition, elimination of a separation step and the smaller volume requirements for the neonatal sample makes the GEM Premier 4000 suitable for total bilirubin measurements in POCT settings.
Assessment of low levels of plasma haemoglobin is a recognised method for the assessment of acute in-vitro red blood cell damage. The practical value of the measurement is virtually limited to circumstances in which acute intravascular haemolysis is believed to have occurred, as in haemolytic transfusion reactions, or in evaluation of the degree of haemolysis occurring by mechanical damage in medical equipment. Because free plasma haemoglobin is quickly bound by haptoglobin and removed from the blood and any non-haptoglobin bound haemoglobin is promptly excreted in the urine, measurement of free haemoglobin in plasma is of no practical value in the diagnosis of chronic haemolytic disorders.
Plasma from 97 patient samples collected in lithium heparin were correlated using an empirical polynomial (AII) method using a Beckman DU®70 Spectrophotometer and the Ortho Clinical Diagnostics Vitros 5,1 FS Chemistry Analyser’s spectrophotometer used for measuring haemolysis index.
Correlation of the two methods shows R2 = 0.98 where PlHb = 0.0097x – 0.04, where x is the haemoglobin index value derived form the spectro scan. A detailed two tailed t-test shows p = 0.0054.
Analysis of plasma haemoglobin by spectrophotometry on the Vitros 5,1 FS shows a positive bias when compared to an empirical polynomial (AII) method using a Beckman DU®70 Spectrophotometer. Limitations of plasma haemoglobin analysis by spectrophotometry derived haemolysis index on the Vitros 5,1 FS do not allow for measurement of plasma haemoglobin levels <1.0 g/L.
The Bio-Rad TURBO 2.0 method is an HPLC method capable of analysing a sample for HbA1c in 97 seconds. The method was compared with an established method (Bio-Rad VARIANT™ II HbA1c) for precision and potential interference from haemoglobin variants and carbamylated haemoglobin.
The correlation study was performed by comparing 181 non-variant whole blood samples over a 5-day period by both methods. Fifty samples with eGFR <60 ml/sec were analysed by both methods and sent to a National Glycohemoglobin Standardization Program laboratory for analysis. Samples containing haemoglobin variants were analysed by both methods. Both intra- and interday precision studies were performed.
The interday precision at HbA1c levels of 5.8 and 9.7% were 0.1 and 0.1% respectively.
The intraday precision at HbA1c levels of 5.8 and 9.7% were 0.7 and 0.4% respectively.
The regression equation was %HbA1c ‘TURBO 2.0’ = 0.9651 x %HbA1c
‘VARIANT II’ + 0.3938 (n=181, r=0.9946, Sy/x=0.15).
For carbamylated haemoglobin samples, the regression equation was %HbA1c ‘VARIANT II’ = 1.054 x %HbA1c ‘NGSP’ +0.0853 (n=50, r= 0.9911, Sy/x =0.19) and %HbA1c ‘TURBO 2.0’ = 1.0402 x %HbA1c ‘NGSP’ + 0.1491 (n=50, r= 0.9901, Sy/x=0.19).
For haemoglobin variant samples, the regression equation of the combined variant haemoglobins E, D, S, and C was %HbA1c ‘TURBO 2.0’ = 0.9498 x %HbA1c ‘VARIANT II’ + 0.3518 (n=105, r=0.9564, Sy/x=0.33).
The precision is comparable with other HPLC based HbA1c methods. The TURBO 2.0 correlates well with an established method and is not affected by carbamylated haemoglobin. The method identifies haemoglobin variants and the retention time. A presumptive identification of the haemoglobin variant can be made via the retention time.
This study aims to establish serum reference ranges for procollagen type I N propeptide (P1NP) and type I collagen C-telopeptide (CTx) in Australian men.
Fasting sera from 1145 males (mean age 60 years; range 20–97 years) who participated in the Geelong Osteoporosis Study were analysed for CTx and P1NP using the automated Roche Modular Analytics E170 analyser. None of the subjects had any medical condition known to affect bone metabolism.
The reference ranges took into account values for the geometric mean ± 2 standard deviations as well as the 2.5th and 5th percentiles and the 95th and 97.5th percentiles.
The CTx reference range is divided into three age groups. For men aged 25 to 40 years, the range is 170–600 ng/L; 40 to 60 years, the range is 130–600 ng/L; and for men aged greater 60 years the range is 100–600 ng/L. For P1NP the reference range is 15–80 μg/L for men aged between 25 to 70 years.
In men greater than 70 years of age values are usually higher possibly due to increased bone turnover.
We also found that high values are frequently seen for both CTx and P1NP in males aged younger than 25 years. This is probably due to bone growth that is not completely finalised.
Values obtained from this large study provide sound age related reference ranges for serum P1NP and CTx in Australian men.
We have evaluated the Roche Troponin T high-sensitivity assay (TnT-hs) as a replacement for the Siemens Centaur Troponin I Ultra Assay (TnI-ultra).
Measurements were made using the Siemens Centaur TnI-ultra and the Roche TnT-hs assays. Studies performed included: precision, method comparison, clinical comparison and effect of haemolysis.
Within and between-run CVs of the TnT-hs were below 5% at all levels tested. Method comparison with the Centaur analyser showed a poor correlation overall with TnT-hs results approximately 0.1 times the TnI results (both expressed in μg/L) across 3 orders of magnitude with a scatter of ratios between 0.02 and 2. 20% more patients with troponin requests from the Emergency Department were positive with TnT-hs compared with TnI-ultra, however the majority of these patients were admitted to the hospital with only the negative TnI value available for clinical assessment. A delta troponin >30% provided reasonable separation of stable from rising troponin concentrations with most low positive TnT-hs patients showing a stable pattern. ROC analysis suggested a TnT-hs value of 30 ng/L would have the same sensitivity and specificity as a TnI-ultra decision point of 0.06 μg/L. The effect of haemolysis was shown to vary between patients with an average fall in measured troponin of 10% per 100 mg/dL Hb (range 0–20% fall). Methods for reporting haemolysed samples are developed. Clinical consultation supported the use of ng/L rather than μg/L for reporting.
Changing from the Centaur TnI to TnT-hs involves new units, decision points and numerical values. There will be a higher number of positive results, many of whom are clinically unwell but may have a normal delta troponin.
Quantitation of Bence Jones protein (BJP) is used to monitor the size of the population of malignant cells in an individual patient. BJP is quantitated by densitometric measurement of the band from the urine protein electrophoresis (UPEP) and urine BJP concentration is calculated as a percent of the urine total protein (UTP). RCPA Immunology QAP Urine Paraprotein Specimen 14–01 was analysed by 59 laboratories and BJP measurement assessed.
UPEP methods included Helena, Sebia (Acid Violet stain), Sebia (Amido Black stain), Capillary Zone Electrophoresis (CZE), cellulose acetate, and inhouse. UTP methods were pyrogallol red-molybdate (PRM), benzethonium chloride (BTCl) and pyrocatechol violet-molybdate. Half of laboratories did not specify their UTP method.
Lambda BJP was the predominant band identified by immunofixation or immunosubtraction (56 of 59 laboratories). Mean (SD, range) for %BJP from UPEP was: All methods (N=49): 60 (19, 0.7–97); Helena (N=10): 62 (15, 43–82); Sebia (Acid Violet stain, N=11): 59 (22, 18–84); Sebia (Amido Black stain, N=15): 58 (12, 36–80). Mean (SD, range) for UTP in mg/L was: All methods (N=48): 1240 (410, 100–2000); PRM: 1530 (240, 1300–1900); BTCl: 1230 (300, 800–1600). Mean (SD, range) for BJP concentration in mg/L was 760 (280, 310–1540, N=39).
Between-laboratory variation for BJP quantitation was approximately 5-fold. Factors that may contribute to this variation are: 1) lower UTP concentration by BTCl assay compared with PRM, and the lesser reactivity of tubular proteins, e.g. BJP; 2) differences in BJP dye-binding by different stains; and 3) use of concentrated versus unconcentrated urine with 2-fold variation for %BJP noted within individual UPEP methods. It is recommended to use the same methods to monitor BJP concentration in individual patients.
Salivary measurements have the potential to provide a convenient and non-invasive assessment of serum ‘free’ cortisol. When assessed against strict criteria, midnight salivary cortisol may be an effective screen for Cushing’s syndrome.
Samples (n=33) were collected both morning and evening, by salivating into a 50ml spot urine container and then centrifuged at 2000 RPM for 10 minutes. The saliva was then divided into 2 aliquots and stored at −20°C until analysed. One set was analysed at Monash Medical Centre on the Beckman Access using a ‘Research only’ direct salivary cortisol method supplied by Beckman Coulter USA. This direct method, analysed saliva without extraction. The second set was analysed by Melbourne Pathology on an Immulite 2000 (Siemens Diagnostics) after extraction into ethylacetate and reconstitution into 1/10th volume normal saline.
Salivary cortisol values ranged from 1.46 to 36.65 nmol/L. There was a strong correlation between the two methods. The linear regression was Beckman = 0.8393xImmulite + 0.6136, r2 = 0.9749.
The Beckman direct salivary cortisol assay is very simple and convenient to use and has the sensitivity to detect midnight cortisol in the normal range (<9) as well as in hypercortisolism (Cushings syndrome or acute stress). The potential to supplement or replace 24 hour urinary free cortisol for the diagnosis of Cushing’s disease is attractive but requires further validation. The method uses the serum/urine cortisol reagent kit available on board 24 hours a day. All three tests (serum, urine and saliva) can be calibrated at the same time, using the one kit, with one set of calibrators. One kit on board minimises wastage. Currently there is no EQA for salivary cortisol and this is clearly desirable.
A predominance of small-dense low density lipoprotein (sdLDL) increases cardiovascular risk, and has been called the atherogenic lipoprotein phenotype ‘pattern B’. Gradient gel electrophoresis is considered a ‘gold standard’ method for identifying this phenotype, though it is impractical for routine laboratory practice. The LDL/Apolipoprotein-B ratio and recently introduced direct assays of sdLDL could potentially simplify LDL sub-fraction analysis and help recognise this atherogenic phenotype. We aimed to compare a direct assay for sdLDL and the LDL/Apolipoprotein-B ratio with more accessible lipid parameters, fasting triglyceride and triglyceride/HDL ratio, to predict the presence of ‘pattern B’ phenotype.
Blood samples were collected from 97 fasted subjects on up to three separate occasions. Total cholesterol, triglyceride, Apolipoprotein-B (Abbott Architect c8000) and sdLDL (sLDL-EX, Denka Seiken) were measured; LDL and HDL-cholesterol were determined after lipid ultracentrifugation. The predominant LDL particle size and phenotype were assigned by gradient gel electrophoresis.
‘Pattern B’ phenotype was identified in 36 % of samples. Peak particle size had a positive correlation with HDL-cholesterol, and a negative correlation with triglyceride and Apolipoprotein-B. Receiver operating curve (ROC) analysis showed triglycerides/HDL ratio and triglyceride alone to be the best predictors of ‘pattern B’ phenotype, area under the curve (AUC) 0.87 and 0.84 respectively. LDL/Apolipoprotein-B ratio and sdLDL gave significantly lower AUCs, 0.74 and 0.71 respectively (p<0.05). A raised sdLDL concentration had the highest specificity, 95 % and positive predictive value, 74 %, for predicting ‘pattern B’ phenotype, but low sensitivity, 43 %. The negative predictive values for sdLDL, triglyceride and triglyceride/HDL ratio were all similar at ~87 %.
Direct measurement of sdLDL provided the most specific predictor of ‘pattern B’ phenotype with the highest positive predictive value; but triglyceride/HDL ratio or triglycerides alone, which are readily available in most laboratories, were the best predictors by ROC analysis.
Pathology Queensland uses the Beckman Coulter DxI800 platform for the quantification of the common tumour markers AFP, CEA, CA125, CA15.3, CA19.9, and PSA. The aim of this study was to compare the analytical imprecision (CVa) of these analytes against the desirable (DG) and optimal (OG) goals for imprecision, derived from biological variation (CVi), and the allowable limits of performance (ALP) as defined in the RCPA Tumour Markers Quality Assurance Program. From this data we will establish guidelines for result critical difference (CD) to facilitate better definition of delta check limits for these analytes.
CVa were calculated for each analyte from internal quality control (IQC) data, collected over the last 5–8 months, and compared to corresponding DG (0.5CVi) and OG (0.25 CVi) and the RCPA ALP. Critical difference (CD) was calculated using the formula 2.77(CVi2 + CVa2)½.
All analytes showed similar CVa profiles across all levels of IQC. The mean and CVa (DG, ALP and CD in parentheses) for each biomarker, at IQC level nearest the clinically significant decision point, were as follows: AFP, 31.5 μg/L, 5.0% (6.0%, ±20%, 36%); CEA, 3.0 μg/L, 7.7% (6.4%, ±2 μg/L, 41%); CA125, 41.1 kU/L, 4.7% (12.4%, ±10 kU/L, 70%); CA15.3, 28.0 kU/L, 5.0% (3.1%, ±15%, 22%); CA19.9, 26.4 kU/L, 5.3% (8.0%, ±15%, 47%); PSA, 3.14 μg/L, 5.1% (9.1%, ±1.5 μg/L, 52%).
Our data shows, that in our laboratory, all tumour markers perform well within the RCPA ALP; CEA and CA15.3 do not meet DG as defined by biological variation; and only CA125 performs at the OG. Empirically derived delta check limits of ±50% are inappropriate for CA125 and CA15.3 due to respectively high and low CVi.
Through a partnership between Ngati Porou Hauora, Roche Diagnostics New Zealand and the Flinders University Community Point-of-Care Services unit, point-of-care pathology testing (POCT) for INR on the CoaguCheck XS is being conducted for warfarin management across a network of 6 remote Maori health centres along the East Cape region in the North Island of New Zealand. A quality management framework has been established to monitor the safety and analytical quality of INR testing.
Clinic nurses were trained on-site in the theory and practice of INR POCT. A set of laminated posters provided a simple step-by-step guide for nurses on patient and quality testing. All nurses underwent a written and practical assessment to obtain a competency certificate as a qualified POCT operator in the program. Since 2008 nurses have conducted monthly quality control (QC) testing (Roche PT Control) and split patient sample testing (where a patient sample was tested by both POCT and the local laboratory).
Between-site imprecision for INR QC testing across three reagent lot numbers averaged 6.6% in 2008 (n=43), 7.3% in 2009 (n=49) and 7.9% to May 2010 (n=15). This compares favourably with the average between-site imprecision of 8.0% observed for INR POC testing in the PoCT in General Practice Trial. For split patient sample testing, the mean difference between POCT and laboratory results were 0.23 (limits of agreement −0.42 to 0.88) (n=74).
INR testing on the CoaguCheck XS has been conducted safely and to an acceptable analytical standard.
Analysis of sweat chloride remains the gold standard for the diagnosis of cystic fibrosis (CF). Since about 1988 testing has been performed as follow up of the newborn heel prick IRT screening test, or in older symptomatic children and very occasionally for the investigation of male infertility. We have reviewed our data base of >1600 serial sweat analyses with a view to clarifying:
Sweat was collected by pilocarpine iontophoresis, chloride measured by microtitration method of Shales and Shales and sodium by flame photometry. Sweat chloride results of 40–60 were followed up to ascertain final clinical diagnosis. Concordancy with low sweat yields was based on results interpreted against the cutoffs; <40 normal and >60 CF.
Of all adequate sweat collections (>75mg) 3.6% had [Cl] between 40 and 60 i.e. equivocal. 104 patients had repeat testing. Based on repeat testing, measurement of uncertainty of an adequate sweat analysis was estimated. However all yields >40 mg gave concordant results. The incidence of equivocal results based on initial indication for testing was: Failure to thrive and respiratory symptoms 2.1%, heterozygote 4.2%, family history CF 9.3%. The incidence of confirmed CF with [Cl] 40–60 will be discussed.
Sweat testing is a skilled procedure which remains the gold standard for diagnosis of a serious lifelong and life threatening disease. Auditing results provides valuable insights to aid reporting and interpretation.
Free PSA (fPSA) is known to be an unstable analyte which potentially creates problems when requested as an add-on test. We evaluated the stability of fPSA in serum and EDTA plasma when stored at 4oC for up to 7 days.
10 paired serum with gel separator tubes and EDTA tubes were collected from patients with a total PSA between 0.2–10 μg/L. fPSA was measured on the day of collection (Day 1), Day 3 and Day 7 using the Roche free PSA assay on the Roche E170. Serum samples kept in the primary tube and whole blood EDTA aliquots were stored at 4oC in the same conditions as usual patient samples.
Baseline serum fPSA ranged between 0.087–1.53 μg/L with a mean EDTA plasma/serum ratio of 0.985 and linear regression equation: EDTA plasma= 1.0485x serum − 0.0089 (r2=0.9999). fPSA in serum decreased an average 11.7±12.6% and 31.2±19.1% by Day 3 and Day 7, respectively. fPSA in EDTA plasma decreased an average 2.8±1.8% by Day 3 and 5.9±7.6% (7/10 samples <10%) by Day 7.
There was no difference in the baseline fPSA concentration obtained in serum and EDTA plasma. Roche claims a stability of 5 days in both tube types using recovery criteria of 90–110%. We found that only EDTA plasma fPSA is stable by Day 3 and would therefore be the preferred specimen for fPSA add-ons. While stable in most samples, fPSA measured in EDTA plasma at Day 7 should be interpreted with caution.
Haemoglobinopathies and thalassaemias are inherited conditions, occurring with increasing prevalence in Australia. Identification of haemoglobin (Hb) abnormalities is important in diagnosis of microcytic anaemias and in prenatal counselling of high risk pregnancies. Haemoglobinopathy testing at The Alfred was previously referred to another pathology provider. The Clinical Biochemistry Unit commenced in-house testing in 2009, using two techniques currently performed in our laboratory; HPLC screening using a Primus CLC385 and gel electrophoresis (acid and alkaline) using the Sebia Hydrasys system. These results in conjunction with haematology results and clinical information were used by a consultant haematologist for interpretation.
EDTA whole blood was analysed for Hb A2, Hb F and for any Hb variants present, using 3 different methods; Primus CLC385 HPLC compared with Primus Ultra 2 HPLC (n=30) and compared with a Biorad Variant I HPLC (n = 71). Structural Hb variants were confirmed using the Sebia Hydrasys system. We reviewed our internal QC (6 months), and the incidence of haemoglobinopathies diagnosed since in-house testing commenced.
We have demonstrated that haemoglobinopathy testing can be successfully performed by a Clinical Biochemistry Unit (with the appropriate expertise in both HPLC and gel electrophoresis), in collaboration with a consultant haematologist.
The Western Australian branch of the AACB QC subcommittee compared the performance of homogeneous high density lipoprotein cholesterol (HDL-C) methodologies across WA. The aim of the study was to determine if the matrix used for the RCPA QAP samples was affecting the HDL-C results using different methods. This was done by comparing HDL-C results for serum pools and 2010 End-of-Cycle RCPA QAP General Serum Chemistry results (EOC QAP).
Five serum pools were analysed for direct HDL-C by fifteen laboratories using four methods (Siemens ADVIA 1650, n=1; Dade Behring Dimension RXL-MAX, n=1; OCD Vitros 250, n=10; Abbott Architect, n=4; Roche Integra, n=1). The samples were produced by linear dilution of high and low HDL-C pools at 0, 25, 50, 75 and 100 % dilutions.
The inter-laboratory precision (%CV) ranged from: 3.8 to 13.1% for HDLC values of 0.85 to 2.26 mmol/L. At a mean HDL-C value of 1.21 mmol/L the CV was 7.9% with values from 1.05 to 1.33 mmol/L. The Architect and Integra assays produced the lowest results with a mean of 1.07 mmol/L and the other assays gave similar results with a mean of 1.26 mmol/L. Whereas, the Architect assay produced high results on the EOC QAP with a group mean of 1.73 mmol/L versus the target of 1.37 mmol/L; the group means for other assays ranged from 1.30 to 1.87 mmol/L.
The HDL-C results from the EOC QAP compared to the patient serum pooled samples gave conflicting results despite most laboratories using the same cut-off values. It is likely that the QAP matrix and the HDL-C methods both contribute to these differences.
The Gem Premier 4000 is a cartridge based blood gas analyser which measures total bilirubin by multi-wavelength spectrophotometry between 480 and 650nm using multi-variate regression. Sample size is 100μl. We evaluated this whole blood bilirubin and compared it with the Abbott Architect c8000 assay to assess its suitability for use in a Neonatal ICU.
Heparinised whole blood samples from neonates and adults were analysed for bilirubin using the Gem Premier 4000. These samples were then centrifuged and the plasma analysed on the Architect c8000 using the Abbott diazonium salt method. Between run imprecision was determined for the whole blood assay.
Regression analysis produced the equation: y = 0.92x + 8.1, r = 0.99, n = 82 (range <10 to 650 μmol/L) when comparing whole blood samples (Gem Premier) and plasma (Abbott Architect C8000). Excellent correlation was demonstrated in the range 50–650 μmol/L, however significant differences were observed below 50 μmol/L. Precision studies showed a CV of 2.8% at a concentration of 333 μmol/L (SD = 9.47, n = 34) and 13.3% at a concentration of 73 μmol/L (SD = 9.73, n = 34).
Monitoring of jaundice in the newborn is vital for the prevention of encephalopathy. Loss of accuracy is acknowledged by the manufacturer for whole blood bilirubin concentrations below 34 μmol/L due to high haemoglobin background. Inaccuracy at this concentration when monitoring jaundice in neonates is of limited clinical significance. The Gem Premier 4000 provides a rapid, small volume testing mode for measuring whole blood bilirubin and is suitable for monitoring clinically significant bilirubin concentrations in neonates.
Paracetamol (Acetaminophen) overdose is an important presentation at emergency departments (ED). Samples from our ED are often haemolysed. Product Information for the Roche paracetamol assay indicates significant interference at haemoglobin concentration (H index) above 150 mg/dL. Our aim was to develop a method to allow release of results on haemolysed samples.
Haemolysate was made from human red blood cells and added to 9 patient samples, 3 BioRad Liquichek™ Immunoassay Plus Controls and 2 aqueous standards. The paracetamol concentrations in the patient samples ranged from 0 to 84 mg/L, 7 to 100 mg/L in the controls, and were 105 and 524 mg/L in the aqueous standards. The H index ranged from 0–1000 mg/dL. The samples were analysed on a Roche P module for paracetamol and H index before and after adding the haemolysate.
A significant positive interference was seen on the haemolysed patient samples, which showed a constant linear relationship with the amount of haemoglobin. The haemoglobin effect can be described as 0.03 mg/L change in measured paracetamol per mg/dL haemoglobin (range 0.022–0.047). In contrast no significant interference was seen in the controls and aqueous standards.
Due to the constant relationship between paracetamol concentration and H index we are reporting corrected paracetamol results for patient samples with an H index up to 600 mg/dL using the following equation: reported paracetamol = measured paracetamol – (0.03 * H index). The lower reporting limit was increased to 5 mg/L to allow for the uncertainty associated with the correction. We noted that haemolysis interference in patient samples is different than in QC samples and aqueous standards.
Ethanol ingestion is an exclusion criterion for patients on the liver transplant waiting list. However, monitoring compliance is difficult, due to the rapid elimination of ethanol from both blood and urine, and because ethanol may be produced in vitro by fermentation. Ethyl glucuronide (ETG) is a direct metabolite of ethanol, formed in the liver by enzymatic conjugation with glucuronic acid. ETG may be detected in urine up to 80 hours after ethanol ingestion, depending on the amount consumed, and the cut-off used. We evaluated the Microgenics DRI Ethyl Glucuronide Immunoassay on a CDx90 analyser.
Imprecision was evaluated above and below the cut-off of 500 ng/ml, and at a level of 1250 ng/ml. Spot urine samples were collected from patients attending the liver transplant outpatient clinic. Samples were assayed for ETG, ethanol, creatinine and CEDIA sample check. Confirmation of ETG results was performed by liquid chromatography-mass spectrophotometry (LCMS).
The coefficient of variation (CV) of the three control samples was 8.0% at 250 ng/ml, 3.9% at 750 ng/ml, and 2.7% at 1250 ng/ml. Fourteen patient samples were analysed for ETG by both DRI ETG and LCMS. Eleven of the fourteen were greater than the 500 ng/ml cut-off by both methods. Ethanol was detected in one sample.
The immunoassay and LCMS methods correlated well, with all elevated results by immunoassay being confirmed by LCMS. ETG levels ranged from undetectable to over 200,000 ng/ml. One patient assayed on four occasions over a period of three weeks returned values between undetectable and 21,000 ng/ml, despite insisting that he had not ingested any ethanol. We believe ETG analysis is valuable in monitoring ethanol abstinence.
The Chempaq XBC is a point of care instrument for the quantitative determination of leucocytes, haemoglobin and three part differential count (granulocytes, lymphocytes and monocytes) in capillary or venous whole blood. Approximately 20 μL of blood is applied to the test cartridge and automatically diluted by a reagent solution. The Chempaq XBC dilutes the sample and uses a sensor to count the blood cells and a result is displayed within 2 minutes.
Accuracy was evaluated by testing K-EDTA venous whole blood samples on both the Chempaq XBC instrument and hospital laboratory analyser (Beckman Coulter LH750). Precision was evaluated by performing replicate tests on a control solution.
Samples tested (n=104) showed good correlation between the Chempaq XBC instrument and the laboratory analyser for leucocytes, haemoglobin, granulocytes and lymphocytes (r≥0.90). Poor correlation was observed for monocytes (r=0.35). Precision analysis of the Chempaq XBC gave CV levels of 3.5%, 2.4%, 6.5%, 21.2% and 10.6% respectively for leucocytes, haemoglobin, granulocytes, monocytes and lymphocytes (n=20). A high error rate was obtained in the first batch of test cartridges (17%, n=69). However, this appeared to be batch specific as a second batch of test cartridges tested displayed a much lower error rate (4%, n=50).
The Chempaq XBC displayed adequate analytical performance for all analytes except monocytes. It may be clinically useful if the laboratory cannot provide appropriate turnaround times. The instrument should not be used for determining a patient’s monocyte count due to the poor accuracy and precision displayed. Further investigation into correlations with the laboratory using capillary samples would be recommended as this would be the most common blood type used at the point of care.
The diagnostic and prognostic significances of borderline elevations of Troponin I (TnI) are controversial. Although guidelines for diagnosis of myocardial infarction (MI) include a TnI rise or fall, this does not indicate a primary coronary thrombotic aetiology (type 1 MI). This is of particular relevance in patients admitted with other serious conditions who may have coincident coronary artery disease or risk factors.
TnI results in the 6 months of 2007 Quarter 4 and 2008 Quarter 1 were retrieved from a common Pathology database of two Queensland teaching hospitals. Computerised hospital discharge summaries were also searched for a diagnosis code of MI and matched with grouped TnI episodes, which included TnI baseline, peak and delta values. The study also included a further prospective period of 15 months. The same TnI Beckman® methodology and 99% cut-off (0.04 μg/L) were in use at each hospital. MI incidence was assessed in three TnI episode groups: Negative, Grey, Positive i.e. peak TnI ≤0.04, >0.04 and <0.15, ≥0.15 μg/L). Statistical significances were calculated from differences between the cut-off or serial TnI values.
Of the 8373 TnI episodes, 1145 (21.3%) were from admissions with episodes in the Grey area. MI occurred in 233/1145 (20%), with 88/233 (38%) in the period after the relevant Quarter and 134/233 (58%) within the relevant Quarter. 21/134 (16%) were associated with episodes in the Positive Group with or without the diagnosis of a second MI. Since the TnI changes in the Grey area Group are small the incidence and final classification of MI are likely to depend largely on supporting ECG or clinical findings and the influences of coexistent comorbidities.
Spurious hyperkalaemia due to in vitro EDTA contamination is well documented, and can also lead to artefactually low Ca, Mg, Zn and ALP. We evaluated the prevalence of K3EDTA contamination of patient specimens referred for routine biochemistry tests from a tertiary level hospital.
Plasma EDTA concentrations were measured using a previously published method (Davidson DF, Ann Clin Biochem 2007;44:294–6) implemented on a Roche Modular® analyser. The method utilises the ability of EDTA to extract Cu from PAN-Cu (pyridylazonaphthol-Cu) followed by colourimetric quantitation of Cu free PAN. EDTA was measured in consecutive specimens received over 1 month for electrolyte analysis. EDTA concentrations >0.2 mmol/L were considered significant as this level causes an artefactual increase in K of approximately 0.5 mmol/L, confirmed by spiking plasma samples with EDTA. Haemolysed samples were excluded from the analysis.
The limit of EDTA detection was 0.157 mmol/L (mean of blank + 3SD); reproducibility CV% was 0.27%. Eight-five of 11,896 patient specimens had EDTA >0.2 mmol/L; 68 of the 85 were excluded due to haemolysis, which causes significant positive spectrophotometric interference in the EDTA method. The daily percentage of EDTA-contaminated specimens ranged up to 0.9%, 70% being from wards and 30% from the Emergency Department.
Automated analysis of EDTA in plasma is a rapid method to confirm contamination in a suspected sample and to explain moderately raised K levels and low Ca levels in non-haemolysed specimens. We now routinely measure EDTA in all non-haemolysed samples with a K >6.0 mmol/L, or in samples with unexpectedly low Ca, Mg or ALP.
‘Critical’ results may be defined as results sufficiently abnormal as to require urgent and sometimes ‘invasive’ intervention. This raises the dilemma as to what extent results should be repeated and confirmed to the highest level of analytical certainty at the expense of urgent reporting and effective, and possibly life saving, intervention. In this study we have examined the effect of repeat analysis on the quality of results as well as on the turn around time to report them.
Critical potassium results (<3 and >6 mmol/L) were measured by indirect potentiometry on a Beckman Coulter DxC 800 while critical troponin I (>0.08 μg/L) was measured by double antibody immunometric assay on the Beckman Coulter DxI 800. All critical potassium results were repeated prior to reporting while troponin I was only reanalysed after repeat centrifugation of the serum. Results (326 for potassium and 161 for troponin I) were audited over a 1 month period.
Using a 2 sided t-test the duplicate potassium results were within the measurement uncertainty (MU) and troponin I was also within MU but 13.4% were not concordant in that the original result was marginally above the positive cut off of >0.08. Repeat troponin I testing prolonged TAT approximately 45 min.
Using the particular Beckman Instruments in this study we were able to satisfy the analytical goal of all ‘first time’ potassium results providing reliable basis for immediate medical management. The time, reagents and labour of confirming results could not be justified within the limitations of this 1 month audit. Repeat troponin I testing however was essential and was a significant contributor to the unacceptable (>4hr) management time of chest pain patients in our Emergency Department.
Unusual immunoassay results due to interferences are a well recognised problem and may cause unnecessary investigations and treatment. Here we present our experience on a rare but known cause of interference in immunoassays using biotinylated antibodies.
Baby WN, one of non-identical twins born with features of liver failure and lactic acidosis. Thyroid function tests (TFT, Beckman DxI) at 1week of age: highly elevated FT4 (>77.7*) and FT3 (24.9*) but normal TSH (3.75). Mother and other twin had normal TFT’s excluding any interference from maternal antibodies, however all three individuals had elevated TPO antibodies.
Heterophile antibody interference was excluded using antibody blocking tubes (Scantibodies). Furthermore, TFT’s were normal when measured on Abbott Architect and Siemens Advia Centaur analysers. All results were reproducible at one month and 6 weeks of age.
Increasingly assay interference was suspected either from metabolites related to the underlying liver failure and organic acidosis or drugs. Further questioning revealed the organic academia was being treated with biotin 10mg 8hrly since day 2 of life.
Review of the TFT assay formats indicated potential interference in the FT3 and FT4 assays due to use of biotinylated antibodies. Biotin was then discontinued and all TFTs rapidly normalised. Ingestion of 30mg of biotin by one of the authors reproduced the observed interference.
Beckman FT3 and FT4 assays utilise biotinylated antibodies and streptavidin coated paramagnetic particles. Excess of biotin in blood competes with biotinylated antibodies for the binding site on streptavidin resulting in apparently high FT3 and FT4 measurements. A number of manufacturers employ biotinylated antibodies that may be subject to both positive and negative interference from therapeutic doses of biotin depending on the exact assay format.
Biochemical assessment of the extent of damage after traumatic brain injury (TBI) might help in the management of these patients. Heart type- fatty acid binding protein (h-FABP) was initially detected in cardiac myocytes but is also found in the brain. In this pilot study we investigated the value of h- FABP measured in cerebrospinal fluid (CSF) and blood of patients with severe TBI as a potential marker of brain damage and predictor of neurological outcome.
We measured h-FABP in CSF and serum samples of 31 patients with severe TBI (GCS≤ 10) and 15 patients without TBI. For control h-FABP was quantified in serum samples of 16 healthy volunteers. Patient outcomes were assessed according to the extended Glasgow outcome scale (GOSE) at 6 months post-injury.
Patients with severe TBI showed significantly high h-FABP in CSF and serum up to day 3 post TBI compared to patients without TBI(Mann-Whitney U test: p<0.001 for CSF and p< 0.05 for serum). Serum h-FABP values of TBI patients were significantly increased up to day 5 compared to healthy volunteers (p<0.001). TBI patients were categorised into 2 groups. (Poor outcome- GOSE 1 to 5, Good outcome – GOSE 6 to 8) Significant inverse correlation was found between peak levels of h-FABP (highest value D0–D2) detected in serum and patient outcome (Spearman r = −0.4, p < 0.05). CSF peak values in contrast did not predict outcome.
This pilot study shows that h-FABP in serum might have the potential to be used as a marker of brain damage severity and patient outcome in TBI patients.
Ayurvedic medicine is a system of alternative medicine native to the Indian subcontinent. Today it is manufactured on an industrial scale and formulated into pills and capsules with sophisticated packaging. However while exported products are subjected to some regulation the local market within India is essentially unregulated.
28 yr old male presented to the GP with one month history of epigastric pain and constipation. There was a past history of low back pain for which he had been taking three Indian Ayurvedic tablets (Vatyog, Sahacharadi, Gandharvahastadi), two of each tablets per day for three months.
FBE: Hb 106 g/L (RR: 130–180) with basophilic stippling. Blood lead: 4.14 μmol/L (<0.48). After stopping all Ayurvedic medicines and undergoing oral dimercaptosuccinic acid (DMSA) chelation therapy blood lead level was 0.88 μmol/L. The Vatyog tablets contained 448 μg (2.16 μmol) of lead per tablet although this was not listed in the ingredients. The Sahacharadi tablet was lead free. No Gandharvahastadi tablets were available for testing.
There are numerous reports in the medical literature about the dangers of heavy metal poisoning from lead, mercury and arsenic in Indian Ayurvedic medicine. Sophisticated packaging is no guarantee that such preparations are not highly toxic. It is unclear whether the lead in Vatyog tablets is due to deliberate formulation or contamination of the raw materials. With the increased influx of ethnic Indians in Australia it is likely that cases of heavy metal poisoning will increase until such time as the manufacturing of such preparations is brought under regulatory control in the country of manufacture. Australian doctors should be aware of this potential source of poisoning.
Acute porphyrias are rare metabolic disorders associated with enzyme defects in the heme synthetic pathway and can present with diverse symptoms. Although the classic presentation is with acute abdominal pain, other less common manifestations may predominate.
62 year old lady presented with 3 month history of upper and lower limb weakness, severe pain in the finger tips and feeling generally unwell. She had ‘band like’ sensation around costal margin. After several admissions and normal gastroscopy she was referred for psychiatric counselling. No similar family history. She had 40 yr history of rheumatoid arthritis managed with methotrexate for past few years.
General biochemistry normal. Urine porphobilinogen (PBG) screening: strongly positive. Urine porphyrins: PBG 1,629 μmol/L (RR<10), total porphyrin 5,762 nmol/L (<300) and uroporphyrin 4,358 nmol/L (<40). Faecal porphyrins: total porphyrin 256 μmol/kg (<200), isocoproporphyrin 13 μmol/kg (<2), coproporphyrin, protoporphyrin and C3/C1 ratio -normal. Plasma total porphyrin: 62 nmol/L (<10).
The very high urine PBG confirmed the acute attack of porphyria. Normal faecal coproporphyrin, protoporphyrin and C3/C1 ratio excludes variegate porphyria and hereditary coproporphyria. Therefore, the diagnosis of acute intermittent porphyria (AIP) was made. However, red cell porphobilinogen deaminase (PBGD) was normal indicating that the AIP is probably due to PBGD defect restricted to the non-erythroid cells (AIP-Type II). In addition, the raised isocoproporphyrin in faeces with raised plasma porphyrin and uroporphyrin in urine is suggestive of porphyria cutanea tarda (PCT). Uroporphyrinogen decarboxylase estimation will clarify whether this is congenital or acquired.
Acute porphyria is commonly considered but seldom diagnosed in cases of recurrent abdominal pain. Acute porphyria can present late in life and, as in this case, without a history of abdominal pain. The peripheral neuropathy and psychiatric overlay were the clues to a correct clinical diagnosis. The co-existing but asymptomatic PCT findings added an additional twist to the biochemical findings.
Environmental lead exposure is now a rare occurrence as a result of removal of lead from both petrol and paint and stricter workplace and environmental policies. Nevertheless lead exposure and toxicity still occurs but usually from more exotic sources.
We audited lead results from 1990–2010 to establish the yield from this investigation as well as haematological features that may assist in establishing whether lead analysis is justified. The sources of lead poisoning were reviewed.
Approximately 32 patients per year were tested for lead toxicity. Over 20 years 12 individuals (1.9%) had lead levels above 0.48 μmol/L (NH&MRC recommended cut-off for further clinical assessment). Only 7 symptomatic patients (1.1%) with lead level above 2.4 μmol/L (upper limit for occupational exposure) were found.
Basophilic stippling was always seen with lead >4.0 μmol/L but absent if lead ≤3.0 μmol/L. There was significant negative correlation between lead level and Hb concentration among lead poisoning patients (Spearman correlation: r = −0.839, p <0.001). All patients with lead >2.4 μmol/L had Hb <12.0 g/dL. At a cost of $32/lead analysis and yield of 1.1% the cost per case of lead poisoning is approximately $2900.
Ayurvedic and herbal medicine were identified as the cause of poisoning in 3 patients, 2 were related to the occupational exposure, one domestic exposure from lead soldered antique copper kettle and a child case was due to ingestion and retention of a lead curtain weight.
Lead poisoning is commonly investigated in patients with unexplained abdominal pain or neurological deficits however it is rarely the cause. We saw no cases of lead poisoning (lead >2.4 μmol/L) when Hb was >12 g/dL suggesting that investigation for lead poisoning is not cost effective unless this criteria is met. The addition of basophilic stippling makes lead poisoning much more likely. Asymptomatic lead exposure (lead 0.48–2.4 μmol/L) may be present without haematological manifestations.
Rifampicin (3-(4-Methyl-1-piperazinyliminomethyl)rifamycin SV) is a semisynthetic antibiotic used in the treatment of tuberculosis in combination with other drugs such as isoniazid, ethambutol and pyrazinamide. Therapeutic drug monitoring of rifampicin can be useful for patients with poor treatment response, suspected non-compliance and malabsorption. We developed an HPLC-UV method to assay serum rifampicin.
The stability of rifampicin is well known to be a problem and this was addressed in both specimens and extracts. Serum and plasma aliquots were measured repeatedly for the first few hours and thereafter until 24 hours. Sample preparation was by simple protein precipitation. Several organic solvents were tested and the stabilities of the resulting extracts were examined. The extracts were chromatographed on a Phenomenex Prodigy ODS3 column and monitored by photodiode array detection at 334nm wavelength. Results from the HPLC method and our existing spectrophotometric method (adapted from McConnel et al 1979) were compared.
Serum and plasma rifampicin stored at room temperature, 4oC or −20oC showed no deterioration until at least 24 hours, 1week and 3months respectively. Specimens precipitated with acetonitrile were stable for at least 24 hours at room temperature which is entirely sufficient for our purposes because batches are short and are run quickly. Extracts in methanol or 50% acetonitrile/methanol mixtures are far less stable. The deterioration rate in methanolic solutions is too fast to be of use despite being utilised in several published rifampicin methods.
Comparison of HPLC vs. spectrophotometric method (n=14) revealed slope of 0.9994, r2 = 0.9447.
This HPLC-UV assay is a robust method for the measurement of serum or plasma rifampicin. The specimens and extracts are more stable than anticipated from the literature.
After routine evaluation, GEM Premier 4000 blood gas analysers were installed into the Canberra and Calvary Hospitals in 2009. The only reagents required are disposable cartridges, GEM Premier 4000 PAK, comprising all components necessary for analysis, including quality controls. Of the first 62 cartridge replacements 32 were unscheduled indicating a reliability of <50%. This failure rate caused angst among our clients and required significant scientist time in cartridge replacement and client liaison. The manufacturer (Instrument Laboratory) acknowledged our concerns and agreed that the system/cartridge performance experienced was unacceptable. The manufacturing processes were reviewed and changes were made to improve the performance of the sensors. We report on those changes and subsequent cartridge reliability.
Gold Epoxy was used to seal the back of the sensor wires to prevent air leakage and new cold formed pins were used to eliminate gaps between the sensor well and sensor wire. Visual inspection of all cards prior to their assembly into the cartridges was also implemented. Comparison of failure rates before and after modifications was undertaken.
In the first 8 months since the introduction of the modifications to the sensors the average reduction in cartridge failure rate fell from 52% to 19%. This failure rate was reduced to 12% in those areas where high numbers of blood gas analyses are performed. Failure to detect process control solutions was the main cause of error after the modifications.
Improvements to the sealing of the sensor wires to prevent air leakage and the elimination of gaps between the sensor wells and sensor wires has reduced the failure rate of cartridges used on the Gem premier 4000.
Free PSA (fPSA) and its ratio to total PSA (FTR) are useful in determining a man’s risk for prostate cancer and hence the necessity for a prostate biopsy. Until recently there was no assay available for the direct measurement of fPSA on the Advia Centaur. Instead users could calculate the fPSA by measuring complexed PSA (cPSA) and subtracting this from the tPSA, measured separately. This work compares the newly available fPSA assay with the previously available calculation.
Sera from 364 patients aged between 40 and 87 years and without known prostate cancer were stored frozen until they could be concurrently analysed for fPSA, cPSA and tPSA. All methods were run according to manufacturer specification on the Advia Centaur.
Measured and calculated FTR (mFTR, cFTR) were compared using linear regression and clinical concordance grids.
Correlation between measured tPSA and tPSA calculated from (cPSA+fPSA), r2 = 0.99 was excellent, suggesting no significant loss of fPSA due to storage. The correlation between calculated fPSA and measured fPSA was r2 = 0.77 and r2 = 0.80 for mFTR vs cFTR. Quartile and tertile risk categories based on FTR are in common use; overall risk category concordance was 61/70 % respectively.
Within and between run precision based on QC replicates was excellent, no CV value being >4.1%.
In 2004 the central biochemistry laboratory at Monash Medical Centre installed a system of total laboratory automation (TLA) from Beckman Coulter – sample inlet, centrifugation, aliquotter, three analytical instruments, outlet and sample storage and retrieval. This resulted in the Dandenong and Moorabbin laboratories sending the non-urgent routine samples, out-patients and non-urgent ward patients, to the central laboratory. The benefits of TLA envisaged were improved turn around times, decreased dependence on the central reception staff knowledge base, predictable process times, elimination of turn around time outliers, assistance to the evening and night shifts, improved efficiencies on the day shift, coping with increased work across all shifts without an increase in staff, decreased number of analytical platforms.
Data extracted from the pathology laboratory information system for the turn around time for potassium, workload increase and for result validation time outliers.
Turn around time for potassium from the emergency department was substantially improved: 50% of the pre-automation and 88% of post-automation of samples were validated within the 60 minute window.
The number of time outliers decreased (sample collection to result validation): pre-automation 44 + 18 minutes, post-automation 27 + 4 minutes.
The workload increase for biochemistry since the introduction of total laboratory automation has been 59%. The staff numbers in the central clinical biochemistry laboratory have remained constant, whereas the staff numbers at the Dandenong and Moorabbin laboratories has decreased.
Turn around times for automated chemistries has decreased, outliers have been substantially reduced, greater efficiency in the utilisation of staff has been achieved. The aliquotter has enabled the biochemistry laboratory to change work procedures to a parallel, rather than a serial, analytical process.
Many common genetic disorders arise from a single base mutation where there is substitution of one nucleotide for another, resulting in a single nucleotide polymorphism (SNP). The polymerase chain reaction (PCR) can be used in conjunction with a variety of techniques to detect commonly occurring SNPs. An alternate technique for SNP detection is mass-array SNP genotyping. This method, using a Sequenom MALDI-TOF Spectrometer, was compared with PCR, to test for the commonly occurring SNPs associated with the disorders of haemochromatosis and thrombophilia.
A single-well assay was set-up on the Sequenom Mass-array Spectrometer to simultaneously detect the common SNPs associated with haemochromatosis (C282Y, H63D, S65C), and thrombophilia, namely Factor V Leiden G1691A mutation, Prothrombin gene G20210A mutation, Methylene TetraHydrofolate Reductase (MTHFR) C677T and A1298C.
150 EDTA whole blood samples sent for routine genetic testing for haemochromatosis and thrombophilia were run on the Sequenom Massarray Spectrometer, and the results were compared with those obtaind by the routine real-time PCR performed on the Corbett (Qiagen) Rotorgene 3000 (haemochromatosis) and the Roche LightCycler 1 (thrombophilia).
There was 100% concordance between the Sequenom Mass-array platform and the routine real-time PCR methods, in the detection of commonly identified Haemochromatosis, Factor V Leiden, Prothrombin gene and MTHFR SNPs.
We demonstrate a high-throughput, efficient and sensitive assay, that simultaneously determines the presence of the commonly occurring SNPs in the genes associated with the genetic disorders of haemochromatosis and thrombophilia. The Sequenom Mass-array Spectrometer is now in routine use for genetic testing in our laboratory, replacing multiple PCR tests.
A number of studies have established that smoking is associated with hypertension, however few studies have looked at genetic factors that influence this relationship. Our aim was to determine whether the acute effects of smoking on blood pressure (BP) translate into an appreciable effect on out-of-office or central BP at a population level, and whether this effect is genetically pre-determined.
In 689 randomly recruited participants of African descent with 24-hour BP and central hemodynamic measurements, we assessed the relationship between smoking and 24-hour or central BP and whether the angiotensin-converting enzyme (ACE) insertion-deletion (D/D) variant modifies these effects.
Independent of age, sex, body mass index (BMI), treatment for hypertension, diabetes mellitus or an HbA1c >6.1%; smoking was associated with 24-hour (p<0.0005) blood pressure. This relationship BP was only present in participants with the DD genotype of the ACE gene (p<0.001), but not in participants with the I allele (p=0.26). The adjusted day diastolic blood pressure (DBP) in smokers (81.8±1.1 mm Hg) was greater than that in nonsmokers (76.9±0.41 mm Hg, p<0.0001) a difference that was enhanced in participants with the DD genotype (smokers day DBP=83.3±1.7 mm Hg; non-smokers day DBP=76.1±0.59 mm Hg, p<0.0005). Although smoking was independently related to the aortic reflected pressure wave, this effect did not translate into an association between smoking and central (p=0.24) or 24-hour (p=0.35) pulse pressure.
Smoking contributes to ambulatory blood pressure in populations with a high prevalence of the ACE DD genotype.
Measurement of oestradiol (E2) levels in postmenopausal women and children requires an assay sensitivity not achieved by most automated systems. The Diasorin E2 RIA assay was used for this purpose by a number of laboratories in Australia until the assay was unexpectedly withdrawn in 2009. The aim of this study was to find a suitable replacement assay.
The following E2 assays were studied; Orion Spectria Sensitive RIA (O), Pantex RIA (P), DSL Ultra-sensitive RIA (D), Elecsys Estradiol II (E) and results compared to the DiaSorin RIA (DS) where possible. Assay bias was investigated by comparing sample analyte levels using regression analysis and difference plot. Standardisation was assessed by testing kit and in-house standards in all assays. Between and within run precision were determined using pooled sera of appropriate E2 levels.
Although results correlated well for all assays (n=31, O-r2= 0.96; P-r2=0.99, D-r2=0.90, E-r2=0.97), three of the assays displayed considerable bias. P and E had a positive bias (mean ratio P/DS-1.4, E/DS-1.7) and D a negative bias (mean ratio D/DS-0.77). In contrast the mean ratio O/DS was 0.99 indicating good comparison of values for O. Assay of kit and in-house standards indicated that the bias differences could be accounted for by standardisation. Correction for the standardisation differences improved the comparability in results (mean ratio P/O-1.1, E/O-1.1, D/O-1.1, DS unavailable). Between and within run precision were acceptable for all assays and functional sensitivity (pmol/L) was found for O<39(CV 12%), P<49 (CV 6.2%), D<33 (CV 4.6%), E=43 (CV 20%), DS <33 (CV 14%).
The O assay is the most suitable replacement assay for the DS because of the comparability of sample values eliminating the need to re-establish and change reference intervals.
The 75g oral glucose tolerance test (OGTT) has been the ‘gold-standard’ for diagnosis of diabetes for decades although it is recognised that its reproducibility in people with diabetes is poor (~70%) due to high intraindividual variability (fasting glucose (FPG): CVi=7%; 2-hr glucose (GLU-2h): CVi=15%). We examined potential factors causing discordant classifications in repeat OGTTs available in a large private pathology.
OGTT data (n=63,972) were reviewed. Glucose (fluoride-preserved, Roche Hexokinase method) had CVa<3%. OGTT interpretations (mmol/L): ‘Normal’ (FPG≤6.0 and GLU-2h≤7.7); ‘Impaired’ (FPG 6.1–6.9 and/or GLU-2h 7.8–11.0) or ‘Diabetes’ (FPG ≥7.0 or GLU-2h ≥11.1).
Overall, 9,577 (15%) had repeat OGTTs but majority of the repeats (n=6,626, 69%) were performed after one-year. Only 101 (1.1%) were repeated within one-month (results are below):
|Normal||34 (79%)||8 (19%)||1 (2%)|
|Impaired||17 (53%)||15 (47%)||0 (0%)|
|Diabetes||5 (19%)||10 (38%)||11 (42%)|
Normal: A ‘Normal’ OGTT is likely to be ‘Normal’ when repeated (79%). The only ‘Diabetic’ on repeat was due to error: the glucose load was not given initially! Impaired: An ‘Impaired’ OGTT was as likely to repeat as ‘Impaired’ or ‘Normal’ but not ‘Diabetic’. Diabetes: Only 42% of the initial ‘Diabetic’ OGTTs remained ‘Diabetic’ on repeat and all had significantly lower FPG (9.5 vs 7.1, p<0.03) with no difference in GLU-1h (15.6 vs 16.0) or GLU-2h levels (14.0 vs 14.6), indicating better patient preparation prior to repeat. Three out of five ‘Normal’ repeats had initial FPG higher than the initial GLU-1h or GLU-2h values, indicating probable initial sample mixed up.
OGTT is poorly reproducible in routine practice due to not only the known biological variability of glucose levels but also a mixture of other factors including improper patient preparation; insufficient glucose load and sample mix up.
The continued global demand for 25-hydroxy vitamin D (25-OHD) assays has created significant pressure for laboratories to seek more automated reliable platforms for analysis. The recently released IDS iSYS 25-OHD automated chemiluminescence assay was evaluated and compared to an ID-SEIA assay and a validated LC-MS/MS method.
Serum samples were compared against both LC-MS/MS (n = 115) and ID-SEIA (n = 64) methods. Precision was assessed by analysis of 3 levels of QC material and an in-house serum pool, twice a day for 5 days. Linearity was assessed by diluting a serum pool with demonstrated high 25-OHD in a serum pool with low 25-OHD. Samples from 3 cycles of the DEQAS proficiency survey were also analysed to establish accuracy.
Regression between LC-MS/MS and iSYS methods gave a Passing-Bablok fit of −8.0 + 1.05x, with a bias of −4.4 nmol/L (R2 0.82). This was an acceptable deviation with 5.2% of the population falling outside the 95% limit of agreement. Regression between IDS-EIA and iSYS methods gave −1.61 + 1.07x, with a bias of 1.9 (R2 0.86). Acceptable reproducibility was obtained for 3 levels of the IDS controls (9.4% at 23.6 nmol/L, 9.0% at 65.8 nmol/L, 6.9% at 146.7 nmol/L), with excellent recovery (100%, 97%, 97%). The in-house pool also showed acceptable precision (5.5%CV) and a recovery of 92% from the mean established by LC-MS/MS. Linearity gave 91–117 percent recovery between 24 and 217 nmol/L. The average percentage difference of the DEQAS samples from the all laboratory trimmed mean (ALTM) was −2%, ranging from −20% to 7% difference.
The iSYS assay has performance characteristics suitable for routine estimation of vitamin D status, and enables our laboratory to process approximately 400 patient samples per day.
A recent study investigated seasonal variation in the diagnosis of Type 1 diabetes Mellitus (T1DM) in children worldwide. The study concluded that the incidence of T1DM showed a clear seasonal pattern. The investigators were able to demonstrate an increase in the number of T1DM diagnosed during the winter months compared to the summer months. This finding held true for both Southern and Northern Hemisphere countries. We were curious to see if we could demonstrate a similar pattern, looking at the number of Type 2 Diabetes Mellitus (T2DM) over a five year period.
We analysed 65,572 (27,466 male and 38,106 female) glucose tolerance tests (GTT) done between January 2001 and December 2005. The data were grouped according to gender, month and the year the test was performed. Using WHO diagnostic criteria, we constructed a time series graph looking at the occurrence rate (%T2DM diagnosed) for each month. We then tested for seasonality at 3 locations (Darwin, Perth and Albany), looking specifically at GTTs done on adult males before 10 AM.
The mean fasting glucose levels did not differ significantly between summer and winter months (P= 0.3)
Surprisingly, the mean post load (2 hour) glucose was 0.5 mmol/L lower during the winter months (P= <0.001). We also noticed a 6.9% rise in abnormal GTTs (impaired glucose tolerance and diabetic levels) during the summer months. Reactive hypoglycaemic events seem to increase during winter compared to summer months (P = 0.03).
These unexpected findings suggest a significant seasonal pattern in the occurrence of T2DM (opposite seasonality to T1DM) and redistribution of post load glucose .This may have several implications.
Serum free cortisol is a better indicator of cortisol bioactivity than total cortisol. Recently there has been increased interest in methods for measuring serum free cortisol. We have compared three of these methods to measure serum free cortisol.
Serum samples were assayed for total and free cortisol using LC-MS/MS. Serum free cortisol was separated from protein-bound cortisol using temperature controlled ultra-filtration. Calculated free cortisol incorporated transcortin measurement by radioimmunoassay. Free cortisol was also estimated using tracer dilution of 1,2,6,7-3H(N) cortisol.
Total cortisol levels ranged from 23–1162 nmol/L (median 554 nmol/L) whilst the range of free cortisol levels was <1.0–95 nmol/L. The free cortisol levels measured by LC-MS/MS equated to 1.4–11.1 % (median 3.6%) of total cortisol. The calculated percent free cortisol ranged from 2.2–13.2% (median 5.3%) and the tracer dilution free cortisol ranged from 7.5–17.5% (median 10.8%), purified tracer, 4.6–13.5% (median 9.7%). There was a strong correlation between LC-MS/MS % free cortisol and calculated % free cortisol (r2 = 0.910) with regression parameters: Calc % FC = 0.96* LC-MS/MS % FC + 1.6. The difference between the median values of the LC-MS/MS and tracer dilution groups of % free cortisol was significant (p<0.004) with regression parameters: Tracer dilution % FC = 0.89* LC-MS/MS % FC + 5.8 (r2 = 0.832).
There was an excellent correlation between free cortisol levels measured by LC-MS/MS and calculated levels but levels of free cortisol estimated from tracer dilution were significantly different to levels measured by LC-MS/MS. Ease of isolation of free cortisol by controlled ultra-filtration and measurement by LC-MS/MS has made the assay of free cortisol levels much more accurate and accessible.
Macroprolactin is prolactin complexed to IgG. While macroprolactin is not biologically active it is recognised in all common prolactin assays with the potential to provide misleading results. We describe our experience with routine determination of macroprolactin and compare results with published reference intervals for monomeric prolactin.
Prolactin and PEG precipitation techniques were performed according to manufacturer’s guidelines on the Siemen’s Centaur and Roche E170. A positive macroprolactin control was obtained from a known positive patient with ongoing supply of material. The calculation for the monomeric prolactin is provided in the respective macroprolactin manufacturer’s protocols. Macroprolactin was measured in patients with prolactin > 400 mIU/L (Centaur) and >500 mIU/L (E170).
Siemen’s Centaur data set: n=342, positive=6.4%, borderline=6.1%, negative=87.5%. E170 data set: n=127, positive=4%, negative=96%. QC - Centaur: n=45, mean=28% recovery, SD=5.9%, CV=21%, Monomeric Ref. Ints – Centaur: 1 / 20 patients outside reference intervals. QC – E170: n=24, mean=36% recovery, SD=6.1%, CV=16.8%. Monomeric Ref.Ints – E170: 1 / 5 patients outside reference intervals.
The positive rate in each system is similar although the Centaur also classified additional patients as borderline. Our detection rate is similar to a published figure of 4%. The positive QC allows control and assessment of the PEG step. The use of monomeric reference intervals is an alternative method of reporting, however, the available literature shows variation in the reference interval data.
Measurement of 25-hydroxyvitamin D is the accepted clinical indicator of vitamin D status. However standardisation of results between methods and laboratories remains a significant problem. Significant discrepancies were found between results obtained using our method (Roche Elecsys E170) with a previous measurement (DiaSorin radioimmunoassay). Subsequently we measured 25-hydroxyvitamin D in the discrepant samples using the gold standard method (liquid chromatography-tandem mass spectrometry- LC-MS/MS) and compared the results.
Fasting sera from a subset (n=55) of female participants (mean age 57 years; range 21–92 years) from the Geelong Osteoporosis Study (n=1265) were analysed for 25-hydroxyvitamin D using three different methods; the DiaSorin radioimmunoassay, the Roche Elecsys Vitamin D3 (25-OH) assay, and the LC-MS/MS method (RMIT Drug Discovery Technologies). Serum aliquots were stored for up to 8 years at −70oC before analysis.
In a selected subset of discrepant samples the Roche method correlated poorly with the DiaSorin method, however, both methods correlated similar with the LC-MS/MS method (Passing-Bablok regression; Roche = 0.46 DiaSorin + 4.62, r2 = 0.39; LC-MS/MS = 1.59 Roche + 3.12, r2 = 0.66; and LC-MS/MS = 0.92 DiaSorin + 6.74, r2 = 0.69).
Only one specimen (1.8%) contained measurable amounts (>4.0 nmol/L) of 25-hydroxyvitamin D2.
When using 50nmol/L as the cut-off for vitamin D sufficiency the Roche immunoassay determined that 83% of subjects were vitamin D insufficient. This was compared to 55% and 52% of subjects testing vitamin D insufficient using the DiaSorin and LC-MS/MS assays respectively.
In regards to vitamin D, these findings showed the DiaSorin radioimmunoassay has slightly better agreement with LC-MS/MS than the Roche assay. The Roche assay seems to underestimate vitamin D levels.
Bias assessment between methods requires patient samples or commutable material. Material is commutable between two systems when it behaves in the same manner as patient samples. In this study we assess the commutability of Quality Assurance (QA) and Quality Control (QC) material for 6 commonly used systems for Luteinising Hormone (LH) measurements
LH measurements were made in patient samples, QA material from the RCPA QAP Endocrine Program and Liquicheck and Lyphocheck QC material from BioRad. Measurements were made in duplicate in single runs on the Abbott AxSYM, Abbott Architect, Siemens Centaur, Roche E170, Beckman-Coulter Dxi, Ortho Clinical Diagnostics ECi and Siemens Immulite analysers. Data was analysed by comparison between each instrument pair for each material. Assessment was for linearity, and for bias at 10 and 20 IU/L using an acceptance criteria of +/− 10%.
The QAP material was linear in all 21 analyser pairs but only unbiased for AxSYM v Centaur; AxSYM v E170, Centaur c E170 and Immulite v OCD. Liquicheck was only linear compared to patient samples for Centaur v E170, OCD v Immulite and OCD v Dxi but only Centaur v E170 also passed bias criteria. Lyphocheck was linear in all systems except the Immulite v E170 and AxSYM v E170 with the bias criteria met by only 1/3 of the analyser pairs.
The QA and QC material tested was not commutable for the majority of analyser pairs. This may indicate differences in assay specificity and differences between the molecular species of LH in the materials or the presence of some cross-reacting substances. This work indicates that commutability of materials for immunoassay must be confirmed with experimental data.
Hyperprolactinaemia can indicate a prolactin secreting tumour, investigation of which is expensive and invasive. Significant levels of prolactin also circulate as macroprolactin (MP) which is not biologically active and accumulates because of its long half life. The presence of high monomer prolactin levels must be confirmed before initiating unnecessary investigation. We reviewed 7,000 prolactin assays over 5 years at RPAH and report on the analysis for MP and monomeric prolactin (P) in these samples.
Total prolactin (TP) levels were determined on the Immulite 2000. MP was determined after precipitation of high molecular weight proteins with 12.5% final polyethylene glycol according to standard protocols. P levels were determined in the supernatant and the difference between TP and P determined MP.
An average of 150 samples exceeded 30 ng/ml/year and including macroprolactin requests but excluding repeat samples resulted in 684 individual samples investigated for MP between January 2005 and May 2010 (9.7% of total assays). Females (F) comprised 81% and males (M) 19% of investigated patients. Macroprolactin was present in 154 patients, 72% f and 28% m and was even present in 19 patients in the absence of hyperprolactinaemia (<30 ng/ml TP, <15 ng/ml P). Of 116 MP patients, 72 had P values in the normal Range (<16 male,-<20 female ng/ml) but 44 had elevated biologically active P. 25 Patients were investigated for MP that did not have MP or elevated P.
It is critical to examine and report on samples for P even in the presence of high MP as it is the critically active component producing the clinical condition.
HbA1c is now recommended for screening and diagnosis in the USA although not in Australia yet. Recently we have also reported using our own data that HbA1c ≤5.5% confidently rule-out while HbA1c ≥7.0% rule-in diabetes in adults. Here, we aimed to examine whether these HbA1c cut-off values for screening and diagnosis of diabetes are age or gender related.
Data from patients (n=6,743) who had a 75g-OGTT and an HbA1c (Bio-Rad Variant II Turbo analyser) performed within 2-weeks was extracted from our pathology database. Diabetes was classified as FPG ≥7.0 mmol/L and/or GLU-2h ≥11.1mmol/L. Patients were grouped by their HbA1c levels.
Of the 6,743 (47.2% females) individuals, 45.7% were diabetic and 18.7% normal. The age and gender-related prevalence of diabetes are below:
The prevalence of diabetes rises exponentially as HbA1c increases in all the groups. To rule-in diabetes using HbA1c≥7.0%, almost all the people in each group would have diabetes. Confidence in rule-out diabetes using HbA1c ≤5.5% is much higher in young (<40y) than in the older adults (≥60y) of both genders.
The utility of HbA1c≥7.0% for diabetes diagnosis is independent of age and gender. However, confidence in using HbA1c ≤5.5% to rule-out diabetes is age-dependent.
Macroprolactin is a high molecular weight complex of prolactin bound to an immunoglobulin molecule and is a source of immunoassay interference that may lead to misdiagnosis of hyperprolactinaemia. Polyethylene glycol (PEG) is widely used to precipitate these biologically inactive complexes from the biologically active monomeric prolactin. Results of PEG precipitation have traditionally been reported as percent total prolactin recovery with a 40% cutoff to distinguish macroprolactinaemia from true hyperprolactinaemia. Recent publications have called for a change in reporting to one based on monomeric prolactin levels.
Prolactin was measured on the Siemens Immulite 2000 platform. High molecular weight complexes of prolactin in serum were precipitated using a final concentration of PEG of 12.5%. Samples with a normal level of prolactin were precipitated with PEG in order to establish a monomeric prolactin reference range. Samples with levels of prolactin >500 mIU/L were also subjected to PEG precipitation and their resulting monomeric prolactin levels assessed against this reference range to determine the presence of macroprolactinaemia.
A monomeric prolactin reference range was established as being <500 mIU/L (n=70). This correlated to prolactin recovery levels greater than 80%. Hyperprolactinaemia was established at monomeric prolactin levels >500 mIU/L (prolactin recovery >80%) whilst macroprolactin precipitated into the monomeric reference range with prolactin recoveries less than 40%. Combined Hyperprolactinaemia and Macroprolactinaemia was seen in those samples with monomeric prolactin above the reference range (prolactin recovery >40<80%).
A report that includes the initial prolactin level, the recovered monomeric prolactin level, the percent recovery together with appropriate reference ranges and an interpretative comment provides an easy to understand format.
Immunoassays used to estimate levels of free thyroid hormones are problematic and subject to possible interference from a number of sources-particularly drugs and binding proteins. The measurement of Free T3 is appropriate in pregnancy when TSH is suppressed and Free T4 normal as antithyroid drugs may affect the foetus.
TSH, Thyroxine Binding Globulin (TBG) and thyroid hormone measurements, both free and total, were made using a Siemens Immulite 2000 platform.
In a population of pregnant subclinical hyperthyroid women, elevated levels of Free T3 were present far more frequently than expected. Comparisons of normal pregnant women and those with subclinical hyperthyroidism using measures of TBG, Total T3 and Free T3 clearly established erroneous elevations of Free T3 in pregnancy.
A population of neonates (n=78) were all clinically euthyroid with normal levels of both TSH and Free T4. Thirty four (44%) showed distinct elevations of Free T3 levels. There was no indication of increased levels of TBG but 29 % of the babies with elevated Free T3 levels also had elevated levels of Total T3 (when measured). In the elevated Free T3 group (n=34), 23 had clinical comments on the request form. Of these 23 clinical comments, 21 (91%) indicated some form of maternal thyroid disease with “Maternal Hypothyroidism on Thyroxine” the most prevalent.
Two clinical populations showed erroneous elevations of Free T3. Elevated levels of TBG appear to be the primary cause of the false elevation of Free T3 in pregnancy. The mechanism for the falsely elevated levels of Free T3 in euthyroid neonates admitted to the NIC is less clear.
New diagnostic criteria for gestational diabetes mellitus (GDM) using a 75g-oral glucose tolerance test (OGTT) have recently been recommended. We examined the impact of these on the GDM prevalence.
Data from women who had a 75g-OGTT performed during pregnancy (n=12,409) were reviewed. GDM was classified using the current criteria (FPG≥5.5 or GLU-2h≥8.0 mmol/L) and the new criteria (FPG≥5.1 or GLU- 1h ≥10.0 or GLU-2h ≥8.5 mmol/L). Age-specific GDM prevalence by both criteria was calculated.
Overall, 3,218 women had GDM either by current or new criteria. Of these, 55.6%, 14.3% and 30.2% met both criteria, current criteria alone and new criteria alone, respectively.
|Age||Number||Prevalence of GDM|
|(years)||current criteria % (95% CI)||new criteria % (95% CI)||Difference (new-current/current)|
|all||12,409||18.1 (17.4–18.8)||22.2 (21.5–23.0)||23%|
|<20||30||13.3 (0.4–26.2)||10.0 (0.0–21.4)||−25%|
|20–24.9||292||11.6 (7.9–15.3)||15.8 (11.6–20.0)||36%|
|25–29.9||1974||15.4 (13.8–17.0)||19.5 (17.7–21.2)||27%|
|30–34.9||5011||18.2 (17.1–19.2)||21.4 (20.3–22.5)||18%|
|35–39.9||4085||19.0 (17.8–20.2)||23.7 (22.4–25.0)||25%|
|40–44.9||953||21.6 (19.0–24.2)||28.3 (25.5–31.2)||31%|
|≥45||64||15.6 (6.5–24.8)||23.4 (12.8–34.1)||50%|
The new criteria identified 23% more women as having GDM compared to the current criteria. This increase is age-dependent as the new criteria identified 25% less in the young (<20y) but 50% more in the older group (≥45y) as having GDM. Moreover, the foregoing in the new guidelines of the glucose challenge test (GCT) as an initial screening would only increase the GDM prevalence further, given GCT is associated with a 18% false negative rate.
The increased prevalence of GDM by the new criteria has workforce implications not only in antenatal/postnatal care by the medical profession, but also in pathology services for the provision of the increased number of OGTTs required during pregnancy as well as post-partum follow-up.
Human Epididymis Protein 4 (HE4) alone is not recommended as a screening test for malignancy, however, we wished to evaluate HE4 was a useful adjunct test to CA125 in assessing malignancy risk.
For 2 weeks samples requested for CA125 analysis at RPAH were tested for HE4 and the risk of ovarian malignancy algorithm (ROMA) value calculated (menopausal status considered). HE4 assay performance was also evaluated.
The HE4 assay had an intra-assay CV of 3.30% and 1.96% at 25.1 pmol/L and 357.7 pmol/L respectively. The inter-assay CV at 3 levels was 4.7%, 2.7% and 2.7%. An average 98% recovery was found with samples serially diluted up to 64 fold.
73 of 118 samples were from O&G and Oncol/Gynae wards and likely to have the highest incidence of malignancy. Of these 17 were positive and 36 were negative for all 3 parameters (CA125, HE4 & ROMA). When CA125 alone was used 31 out of 73 exceeded the reference range, HE4 alone had 19 samples above the reference range and when both were combined 37 of the 73 patients exceeded the reference ranges. In the whole group 50.8% were normal, 22.9% were distinctively positive, 26.3% were found to be in the grey-zone. On a test-by-test basis, 39% had CA125 above the normal, 26.3% had HE4 above the normal but when combined 44.1% were found to be above the reference.
HE4 and CA125 may be useful in assessing malignancy or the therapeutic response in cancer patients. This finding is preliminary and the validity of its use requires clinical validation to determine true effectiveness.
In 2007, a survey was undertaken to determine the variability of testosterone results in Western Australian (WA) laboratories. The study focussed on the clinically critical lower end of the assay and results were compared to a gold standard method (LC-MS/MS, Mayo Clinic, Rochester, Minnesota, USA).
Seven linearly related male serum samples (series A) and four linearly related female serum samples (series B) were assayed by all eight public and private pathology laboratories in WA offering testosterone measurement. Immunoassay platforms included Roche Elecsys and E170 Modular, Siemens Immulite 2000, Centaur and coat-a-count RIA, Abbott Architect and OCD Vitros Eci. In 2009 a local LC-MS/MS method became available and was retrospectively included in the survey.
For male series A, mean testosterone ranged from 0.6–11.9 nmol/L (between-method CV 64.9%–16.7%). Immulite overestimated testosterone concentration compared to LC-MS/MS (mean difference +18%), whereas Architect, Roche and Vitros platforms recovered lower results (mean difference −13% to −44%). For female series B, mean testosterone ranged from 0.3–4.1 nmol/L (between-method CV 94.1%–27.4%). For samples 2–4, all assays except Vitros overestimated testosterone when compared to LC-MS/MS (mean difference +9% to +146%).
The wide range of these results for these 11 samples demonstrates the need for improved harmonisation of testosterone assays to enable laboratories to provide comparable patient results. For example, a sample with a mean result of 8.4 nmol/L had a range of 5.3–11.0 nmol/L. Further investigation is required to determine the causes of poor between-method agreement for low-level testosterone measurement which may indicate inadequate calibration or non-optimisation of assays.
Everolimus is an immunosuppressant for the prevention of transplant rejection. The current TDx assay has a limited life and a high positive bias against the HPLC-MS/MS. We evaluated the new Thermo Scientific QMS Everolimus turbidimetric assay on the CDx90 analyser against HPLC-MS/MS.
Two pooled samples, a high control and a non-transplant patient sample were used to determined assay performance (CLSI guidelines). Both stored and fresh patient samples were quantitated with HPLC-MS/MS (SydPath). Onboard reagent stability and interference by Cyclosporine metabolites were determined.
The correlation between the two methods was high (R=0.976) with a linearity coefficient of 0.994. Bland Altman plots showed a negative bias for the QMS assay 18.7% with an absolute difference at zero concentration of 17.4%. Intra-assay precision (CV) at levels of 3.99 ng/mL, 8.02 ng/mL and 16.18 ng/mL were 5.87%, 3.66% and 3.77% and inter-assay precision was 6.42%, 3.64% and 2.99% respectively.
Onboard reagent stability increased when capped after each run. Cyclosporine metabolites did not interfere with the assay and fresh samples correlated better (R=0.970) than stored refrigerated samples. (R=0.939).
The new QMS Everolimus assay, yet to be released worldwide, had a constant bias but correlated well with HPLC-MS/MS. The increased reliability and accuracy of the QMS assay compared to the TDx technology provides a timely replacement. The random access assay will reduce the need to batch samples and provides a suitable alternative method to LCMS/MS to monitor Everolimus concentrations.
Kits and consumables subsidised by Thermo Scientific.
Common anti-convulsants such as Phenytoin and Carbamezapine lower FreeT4 (FT4) however the effects of Valproate and Lithium are less clear. The aim of this study was to quantify these effects.
Subjects with simultaneous requests for thyroid function and therapeutic drug (TD) were extracted from the database from over 5 years (n=12,969). Cases then removed included any with undetectable drug levels, known thyroid or anti-thyroid therapy and those on more than one TD. In subjects with more than one record over the timeframe results were averaged. Thyroid function tests were then compared to a reference group. Four groups were considered, those on Carbamezapine (n=841), Phenytoin (n=616), Lithium (n=1830) and Valproate (n=1989). A secondary data extraction examined subjects on Lithium and Valproate (n= 325).
Those on Phenytoin, Carbamezapine and combination Valproate/Lithium had reduced FT4 levels of 1.2, 1.1 and 0.7 pmol/L on average respectively (p <0.01) compared to the reference group. Lithium had no significant effect on FT4. Valproate alone elevated FT4 levels by 0.9 pmol/L on average (p<0.01). All 5 groups had mean and median TSH values above the reference group (p<0.01), both when all TSH data was examined or when analysis was limited to subjects with TSH within the reference interval (0.40–4.00 mIU/L). Compared to a reference population, those taking a TD had increased likelihood (14.7%) of an abnormal TSH. Similarly they were more likely (9.5%) to have an abnormal FT4 with the exception of patients on Lithium alone.
FT4 was only modestly reduced by Phenytoin and Carbamezepine in this population. Subjects on Valproate had variable FT4 effects. The rate of abnormal TSH results in this group was increased in comparison to a reference population.
Literature suggests that the biological variation of TSH is ~20% for euthyroid patients. We undertook a retrospective review of patients euthyroid over 10 years to investigate factors which effect TSH levels and to verify the biological variation.
Records of patients with a TSH in the reference interval (0.40–4.00 mIU/L) between 1st Jan and 30th June 2000 were extracted from the database. All subsequent thyroid function records and thyroid antibodies were then extracted for a 10 year period until 31st Dec 2009. Cases with more than 3 TSH records were retained (n=17,171). Any subject with a thyroid result outside the reference interval or with history which could effect TSH in any records was removed. The remaining data set (n = 12,569) was then examined for population distribution, sex, age, time of collection trends and biological variation.
There were two populations contained within the reference interval, the first with a median TSH of 1.50 mIU/L. There were analytically significant changes in TSH mean/median values based on age and time of collection. The biological variation for any individual lay within a large range (6–54%) and varied with the mean/median TSH value. The average biological variation remained relatively constant between TSH of 0.5 and 2.75 mIU/L at ~ 24%. Thereafter it fell and at an average TSH of 3.75 mIU/L was 12.2%.
The use of a generic reference interval of 0.4–4.0 has limitations. Age and time of collection should be considered when reviewing TSH especially near the upper limits of a reference interval. Biological variation of TSH is markedly individual and may not be constant across the whole reference interval.
Prolactin is not commonly recognised as a hormone that changes significantly within the menstrual cycle or menopause. The aim of this study was to determine the variability of prolactin and establish prolactin reference intervals for our laboratory.
Prolactin levels obtained from 6540 subjects between January 2006 and November 2008 were divided into 5 groups - males, postmenopausal females and premenopausal females in follicular/non-cycling, ovulatory and luteal phases. The 97.5th centile defined the upper reference limit of prolactin for each group.
The prolactin upper limit was not significantly different in males and postmenopausal females (430 and 440 mU/L respectively). It was significantly higher in premenopausal females (590 mU/L) compared to males and postmenopausal females. Within premenopausal females, the prolactin upper limit was significantly higher in ovulatory phase (680 mU/L) compared to follicular/non-cycling (540 mU/L) and luteal phases (580 mU/L), and in luteal phase compared to follicular/non-cycling phase. Women in ovulatory phase had an upper reference limit 54% higher than postmenopausal women.
Prolactin varied significantly throughout the menstrual cycle. The utility and accuracy of testing may be improved by applying specific reference intervals for each phase of the menstrual cycle. Alternatively, a single reference interval could be used in premenopausal women if prolactin was always measured in the follicular phase, well before midcycle. A premenopausal woman may have a prolactin result up to 50% higher than a postmenopausal woman and use of a similar reference interval for all women may not be optimal. Prolactin levels in postmenopausal females and males were not significantly different and a common prolactin reference interval could be applied for these groups.
Sex hormone-binding globulin (SHBG) is produced in the liver and levels regulated by insulin. Low levels of SHBG are correlated with hyperinsulinaemia. We assessed proportions of men and women with insulin resistance, using criteria of Kidson et al (MJA 2005), at varying levels of SHBG.
Moderate insulin resistance was defined as fasting insulin >14 mIU/L with normal or elevated glucose if requested. Retrospective analysis of 5 years data for concurrent fasting insulin and SHBG was performed. SHBG results were stratified in deciles for men and women and proportions of patients with insulin >14 mIU/L determined for each decile.
3,416 records with concurrent fasting insulin and SHBG were available, 2386 in females and 1130 in males. The rate of moderate insulin resistance in men with lowest SHBG results (0–10 nmol.L) was 26.7% compared to a rate of 22.7% in men with highest SHBG levels (>50 nmol/L). In women the rate of moderate insulin resistance increased progressively with lower SHBG. Women with SHBG results over 50 nmol/L had an average rate of insulin resistance of 11.6%. In women with lowest SHBG results (0–10 nmol/L) the rate of insulin resistance was 75.0%, a much higher rate than that in men with similarly low SHBG levels. In subjects with an SHBG in the lowest decile (0–10 nmol/L) the average insulin was 21.5 mU/L in women and 12.7 mU/L in men.
In this population the relationship of SHBG and insulin was different in men and women. The rate of moderate insulin resistance and the average insulin were higher in women than men for a given decile of low SHBG. The reasons for this are unclear but further study may be warranted.
Varying morning cortisol cut points (550, 500, 300, 100 nmol/L) and their implications for assessment of adrenal function have been reviewed in the literature. Cortisol has a well-known diurnal variation and time of morning collection may vary. This study assessed the effect of time collection on cortisol levels.
Cortisol results (n=13,953) were extracted from the WDP database over a 4 year period. Dexamethasone suppression, Synacthen tests and results >1,000 or <30 nmol/L were excluded. The proportion of morning cortisol results falling below 100, 300 and 500 nmol/L were determined in 15 minute intervals from 7.30–9.45 am.
The proportion of cortisol levels over 500 nmol/L fell from 40% to 21% from 7.30–9.45 with an average of 5% more patients below the cut-point every 30 minutes. Including cortisol levels at all times of the morning, most results (69%) fell in an intermediate range (100–500 nmol/L). The proportion of cortisols below 100 nmol/L remained relatively constant with a mean of 1.3% (range 0.2–2.6%) results below this cut point. The percentage of patients with levels <300 nmol/L increased from 12.3% to 31.7% over the morning and more than doubled between 8.00–9.15 am.
A cortisol of <100 nmol/L only occurs in a small proportion of patients and is relatively consistent over the morning. Not many patients would be considered as likely adrenal sufficient (level >500 nmol/L) using morning cortisol. The percentage of patients with levels >500 nmol/L is halved over the morning from ~40 to 20%. Follow-up of all intermediate results (100–500) would require Synacthen testing in up to 79% of patients. Careful attention to time of collection is important in interpretation of morning cortisol levels.
The aim was to evaluate the new Roche Testosterone II assay (Testo II) for DHEA-S cross-reactivity, precision and accuracy compared to LC/MS.
The Roche Testo-II & I, performed on the Roche Cobas e601, were compared to LC/MS (Royal Perth Hospital). Male and female samples (n=42) were analysed. Female patients (n=16) with varying DHEA-S were measured on the three assays. Two female patient pools (n = 76) and QC data were used to assess precision of Testo II. Routine patient testosterone data at WDP was collected for January 2009 and January 2010. The two data sets were compared to one another.
Females (n = 25) with testosterone concentrations 0 – 5 nmol/L demonstrated ~ 26% negative bias with Testo II compared to Testo I. Female testosterone population data for January 2009 (n = 848) and January 2010 (n = 603) were compared using a z-test for the sample means. Two-tail p-value at α = 0.05 (95%) was 3.8 x 10−13 confirmed a significant difference. Male testosterone population data for January 2009 (n= 680) and January 2010 (n = 574) were similarly compared using a z-test for the sample means. Two-tail p-value at α = 0.05 (95%) was 0.55, no significant difference. Testo I and II, compared to LC/MS, demonstrated correlation r2 = 0.990 and r2 = 0.995. Female patients with varying levels of DHEA-S were assayed on Testo I, II and LC/MS. For the same sample Testo I gave higher results than Testo II. Testo II precision was 3.3%, 2.5 %, 4.1 % and 6.7 % for concentrations of 21.8, 8.3, 1.6 and 0.6 nmol/L.
Testo II has provided flexibility, precision and a good solution to previous interference and performance issues.
Tandem mass spectrometry (LC-MS/MS) is a highly specific detection method that can profile many analytes simultaneously. The aim of this work was to develop a rapid ‘dilute-and-shoot’ method for the quantitation of the tumour markers vanillyl-mandelic (VMA or HMMA), homovanillic (HVA) and 5-hydroxy-indole acetic (HIAA) acids in urine.
Urines (20 μL), calibrators (Recipe) and QCs (BioRad) were diluted with water containing 3 deuterated internal standards. Samples were injected onto an Atlantis T3 column (50 x 2 mm) at 40oC and analysed by negative electrospray MS/MS with multiple reaction monitoring on an AB SCIEX 3200 Q-Trap instrument with a run time of 5 min. Analytes were measured using peak area ratios with Analyst 1.4 software.
Under the LC conditions used, elution times for VMA, HIAA and HVA were 1.0, 2.0 and 3.2 min respectively, and assay linearity was established over the range 1–1000 μmol/L for each analyte. Between-run imprecision of the assay over 3 months was 5.9, 4.9 and 5.3% for normal urinary concentrations of VMA, HIAA and HVA. Method comparison studies were carried out against an existing HPLC method with fluorescence detection (FLD) using over 30 urines. Passing-Bablok regression equations relating LC-MS/MS (y) to HPLC-FLD (x) were y=0.97x−1.7 for VMA (r=0.99), y=0.96x+0.5 for HIAA (r= 1.00) and y=0.59x+0.2 for HVA (r=0.61). Chromatographic interferences from other urinary fluorescent compounds were problematic for HVA, which was often over-estimated by FLD even after solvent extraction.
A rapid and highly specific isotope-dilution LC-MS/MS method has been developed for the simultaneous measurement of VMA, HVA and HIAA in urine with no extraction step required.
The measurement of plasma free metanephrines has become an important diagnostic test to exclude a diagnosis of phaeochromocytoma. Tandem mass spectrometry (LC-MS/MS) has high analytical specificity for the assay of metanephrines, but the sensitivity of detection can be limiting and is instrument-dependent.
A simple method involving deuterated (d4)-acetaldehyde and reductive amination for 30 min at 37oC was used to convert metanephrine, normetanephrine and 3-methoxytyramine, and their deuterated internal standards, to d4-monoethyl or d8-diethyl derivatives. Precursor and product ion mass spectra were recorded to allow multiple reaction monitoring methods for both derivatised and non-derivatised plasma metanephrines to be compared under reverse-phase LC-MS/MS conditions. Positive electrospray ionisation was used on an AB SCIEX 3200 Q-Trap instrument.
Conversion of metanephrines to less polar deuterated-ethyl derivatives increased their mass by 32 or 64 amu, and greatly enhanced the intensity of their molecular ions and fragments. Derivatisation also improved the chromatographic properties of the metanephrines, and mobile phase containing 20% methanol was used for separation on reverse-phase HPLC columns in under 5 min. The sensitivity of tandem mass spectrometric detection of metanephrine, normetanephrine and 3-methoxytyramine was increased up to 20, 30 and 5-fold respectively compared to underivatised compounds.
The LC-MS/MS detection of plasma free metanephrines is greatly enhanced by ethyl derivatisation, which is easy and rapid to perform. Advantages are a smaller plasma volume requirement and lower limits of quantitation to improve the measurement of normal plasma concentrations of free metanephrines under reverse-phase HPLC conditions.
Urine cortisol is frequently analysed during the investigation of clinically suspected hypercortisolism. Urine contains cortisol metabolites and related steroids so analysis is frequently performed after manual extraction. The Abbott Architect® urine cortisol assay does not specify if extraction is required, but reference a poster by Hoffman et al which suggests that urine cortisol results from the Architect urine cortisol assay compare favourably to LC-MS/MS, with a slope of 1.06, intercept of 23.1 nmol/L and r = 0.997. We examined if extraction was required prior to urine cortisol analysis by comparing results from non-extracted samples with dichloromethane extracted samples on the Architect analyser.
40 patient samples of non-extracted and extracted urine were analysed using a competitive chemiluminescent microparticle immunoassay. Urine cortisol competes with acridinium labelled cortisol conjugate for binding to mouse monoclonal anti-cortisol coated microparticles. 1.0 mL of urine was extracted with 2.0 mL of dichloromethane, evaporated to dryness, and reconstituted with a cortisol free diluent.
Linear regression analysis of extracted and non-extracted urine cortisol results demonstrated a slope of 0.9852, intercept of 24.3 nmol/L and R2 of 0.9934.
Our data confirms the Abbott Architect® urine cortisol assay provides comparable results without the need for extraction. Analysis of non-extracted urine cortisol may be appropriate for the investigation of suspected hypercortisolism. Confirmation using LC-MS/MS to verify our findings, including analysis of urine samples from cases with hypercortisolism is suggested prior to recommending the analysis of non-extracted urine for the routine investigation of suspected hypercortisolism.
Immunoassays used to estimate levels of free thyroid hormones are problematic and subject to possible interference from a number of sources-particularly drugs and binding proteins. The measurement of Free T3 is appropriate in pregnancy when TSH is suppressed and Free T4 normal as antithyroid drugs may affect the foetus.
TSH, Thyroxine Binding Globulin (TBG) and thyroid hormone measurements, both free and total, were made using a Siemens Immulite 2000 platform.
In a population of pregnant subclinical hyperthyroid women, elevated levels of Free T3 were present far more frequently than expected. Comparisons of normal pregnant women and those with subclinical hyperthyroidism using measures of TBG, Total T3 and Free T3 clearly established erroneous elevations of Free T3 in pregnancy.
A population of neonates (n=78) were all clinically euthyroid with normal levels of both TSH and Free T4. Thirty four (44%) showed distinct elevations of Free T3 levels. There was no indication of increased levels of TBG but 29 % of the babies with elevated Free T3 levels also had elevated levels of Total T3 (when measured). In the elevated Free T3 group (n=34), 23 had clinical comments on the request form. Of these 23 clinical comments, 21 (91%) indicated some form of maternal thyroid disease with ‘maternal hypothyroidism on thyroxine’ the most prevalent.
Two clinical populations showed erroneous elevations of Free T3. Elevated levels of TBG appear to be the primary cause of the false elevation of Free T3 in pregnancy. The mechanism for the falsely elevated levels of Free T3 in euthyroid neonates admitted to the NIC is less clear.
Tests may be added to a request using algorithms. This practice is not well established or defined. Critical results may be divided into truly critical required to be phoned 24/7 and those required to be phoned during the day. Requestors based in the community would prefer not to be contacted unless a result was truly critical and contacting a community based requestor out of usual office hours requires substantial staff resource. We have derived algorithms to attempt to define that a result is truly critical.
The literature on critical results was reviewed and a critical phoning list was agreed. Algorithms with reflex tests added, were devised to help define a truly critical result. Each algorithm was costed, so as to justify the reflex tests, by audit of results in the laboratory database and test costs.
This affects 75 tests per month and the total cost of the reflex tests added is $48 per month. Tests are only reflex added if not previously assayed. If s-creatine kinase >1999 μ/L, then s-creatinine and s-potassium are added. If glucose >24.9 mmol/L, then bicarbonate is added. If s-lithium >1.49 mmol/L, s-creatinine is added. If s-magnesium <0.31 mmol/L then s-potassium & calcium are added. If an isolated raised s-ALT is present with normal GGT and ALP, then s-creatine kinase is added. If s-digoxin >2.49 nmol/L, s-potassium, calcium and magnesium are added.
The use of reflex tests with critical results has helped to inform a two tier critical result phoning system. This system both avoids disturbing requestors and conserves staff time at night.
An essential prerequisite for electronic ordering or reporting of pathology results to electronic health records (EHR) is clear identification of each test. Numerous terminologies now exist to perform this role including:
In Australia, most laboratories electronically reporting results utilise LOINC. The National eHealth Transition Authority (NEHTA) recommends a future transition to SNOMED-CT, so it is important that laboratories are both aware of developments in this field and using an optimal code for test identification.
The AACB LOINC Working Party (WP) is a group of scientists, pathologists and IT professionals, representing both public and private pathology institutions throughout Australia, working on the following LOINC related issues:
The WP has reviewed approximately 80 common tests representing the majority of high volume biochemistry performed in Australia. Approximately 10% of the surveyed entries for each lab/test combination showed unnecessary variation that would complicate the task of comparing results between different laboratories, or worse, potentially leading to the misidentification of a result in an EHR.
The WP has worked to establish relationships with NEHTA, Standards Australia and the Medical Software Industry Association to clarify what codes should be in routine use. The WP recommends the ‘Austpath’ codes (http://www.ahml.com.au/austpath) and continues to work with Standards Australia (IT 014-06-05 committee) to ensure these codes represent ‘state of the art’. The makeup of the WP, with very wide representation, is an important component in this process.
Although high performance liquid chromatography (HPLC) methods for HbA1c analysis show superior precision and detect the presence of haemoglobin variants, there is a reluctance to use these methods in laboratories analysing large volumes of HbA1c tests because of the time involved in manually checking the chromatograms for sampling errors or the presence of haemoglobin variants. The recently introduced Bio-Rad Hb- Advisor™ System, a middleware software, eliminates the need to manually review normal chromatograms while allowing review, based on pre-established criteria, of abnormal chromatograms.
Studies were performed to compare the time needed to review chromatograms from both the Bio-Rad VARIANT™ II TURBO and VARIANT™ II HbA1c methods manually and using the Bio-Rad Hb-Advisor System. The ability of Hb-Advisor to correctly identify samples that failed to meet established criteria due to sampling errors or the presence of a variant was tested by submitting samples containing a haemoglobin variant or samples with low haemoglobin.
For both Bio-Rad systems, the time needed (technologist dependent) to manually review and release a normal chromatogram varied from 8 to 20 seconds and 25 to 50 seconds for those chromatograms with sampling errors, and 60 to 120 seconds for a haemoglobin variant chromatogram. With use of the Hb-Advisor, only the latter 2 categories (1.9 to 2.4% of total chromatograms) are manually reviewed. Following review of Hb-Advisor flagged chromatograms, release to the laboratory information system takes 8 seconds independent of batch size. The Hb-Advisor detected all sampling errors and identified all haemoglobin variant samples.
The Bio-Rad Hb-Advisor System increases productivity in HbA1c analysis without compromising quality.
There are no quantitative reports on the effect of low levels of haemolysis on the stability of ACTH although it is often recommended that samples should be chilled immediately. Additionally, no inhibitors of the haemolysis-induced loss of ACTH have been reported. We have examined the effect of haemolysis, rapid chilling and a protease inhibitor on the stability of ACTH.
Whole blood (EDTA anti-coagulated) was collected from 9 volunteers and an aliquot from each was haemolysed and added back to the same sample at 37º to produce 0, 0.1, 0.25, 0.5 or 1.0% haemolysed blood. Four treatments were then carried out at each level of haemolysis. Aliquots were either A) immediately centrifuged at 4º and the plasma stored frozen for later analysis, or B, C and D), treated as for A, but incubated in ice/water for 30’ before centrifugation. Post-centrifugation D had 2 mM N-phenyl maleimide (NPM) added. After thawing, samples were analysed for ACTH (Roche Elecsys 2010 analyser) either immediately (A and B), or after one hour at 21º (C and D).
Median levels of haemolysis for 10% loss of ACTH were 0.19, 0.32, 0.11 and 0.86% (0.26, 0.44, 0.15 and 1.2 g/L Hb) for sample treatments A, B, C and D respectively (p = 0.001 and 0.0003 for difference between A and B, and between C and D respectively).
For investigating relationships between vitamin D and the risk of diabetes and kidney disease, we planned to analyse 25-hydroxy-vitamin D (25OHD) and enzymatic-creatinine (enzCr) in the serum from 11,247 people who participated in the Australian national survey in 1999–2000, the Australian Diabetes, Obesity and Lifestyle (AusDiab) cohort, for prevalence of diabetes. However, 5.6% of specimens had insufficient serum but had fluoride oxalate (FO) plasma available. We thus evaluated whether FO plasma was suitable for 25OHD and enzCr analysis.
Bloods from 101 staff were collected into serum separator tubes (SST) and FO tubes. Both tubes were processed and analysed simultaneously for 25OHD (DiaSorin Liaison) and enzCr (Roche Modular). Reportable limit for 25OHD was 10 nmol/L. Differences were analysed using Bland-Altman plot and paired T-test.
25OHD data: Range of serum 25OHD was from <10 to 197 nmol/L. Correlation was done only for levels between 10 and 120 nmol/L as only two people had values above 120 nmol/L. There was no difference in 25OHD (mean±SD) between the FO plasma and the paired serum (60±23 vs 61±24 respectively, n=98, p=0.32). The correlation for FO vs serum was: y=0.97x +2.5 (r²=0.89). Bland-Altman plot showed a 1.4% positive bias for 25OHD from the FO plasma.
EnzCr data: Serum enzCr range: 36–108 μmol/L. Mean enzCr values were 62±13 vs 59±12, respectively, for FO vs serum (n=90, p<0.01). Bland-Altman plot showed a negative bias of 3 μmol/L (95%CI: −2.3 to −3.8 μmol/L) for enzCr from the FO plasma. The equation was: y=0.89x+3.8 (r² =0.93) for FO vs serum.
For 25OHD, results from serum or FO plasma were very similar. For enzCr, there was a small negative bias when measured in FO plasma. This small bias, however, will have minimal clinical significance. Thus, FO plasma was accepted for 25OHD and enzCr analysis.
Homocysteine (Hcy) concentration can rise after blood collection in certain tubes due to on going release by the red blood cells in vitro. Our routine practice was to collect into EDTA tubes, spin and separate the plasma immediately then to send the frozen plasma on dry ice to the laboratory. This minimises the problem, however, not always practical especially when the blood was collected by referring doctors at their clinics where a centrifuge is not available. Here, we aimed to evaluate the effectiveness of the Vacuette® Hcy (V-Hcy) tube for Hcy analysis.
Seven tubes of blood each were collected from eight laboratory staff: six in V-Hcy tubes and one in EDTA tube (routine). One V-Hcy and the EDTA tube was processed within 30–60 min after collection (reported as 0 h) and the remaining V-Hcy tubes were kept at room temperature for further 2, 4, 6, 8 and 24 hours before centrifugation. All the plasma from the tubes were separated from the cells after centrifugation and analysed in duplicates immediately. The average values of the duplicates were used in the final analysis.
The mean Hcy at 0 h from the V-Hcy tubes was 11.1±7.2 (mean±SD) μmol/L) which was significantly lower than that from the EDTA tubes (12.7±8.0 μmol/L) (p<0.01). The levels of Hcy in V-Hcy tubes remain the same when kept at room temperature for the entire duration (24 h prior to centrifugation) of the study. At the 24 h time point, the mean Hcy was 10.8±6.3 μmol/L which was not significantly different than that from the 0 h (p=0.94).
The Vacuette® Hcy tube adequately stabilises Hcy in vitro for 24 h at room temperature before centrifugation. It minimises pre-analytical errors for Hcy analysis thus permits blood collections in doctors’ clinics and remote areas for assessment of patient’s Hcy status.
In 2006 Sullivan Nicolaides Pathology (SNP) began working towards ISO 14001:2004 Environmental Management Systems (EMS) implementation and certification at the Coffs Harbour and Grafton Laboratories. SNP had recognised it’s corporate responsibility to respect our environment and to manage any environmental impacts associated with our pathology work practices. The ISO 14001 standard provided a framework by which this could be achieved.
Local and corporate environmental policies, procedures and initiatives were introduced concurrently, with elements of the existing Quality Management System, Safety Management System and LEAN policies and procedures included in the EMS framework. EMS implementation has included the provision of training for staff and contractors; initial environmental reviews, the determination of significant environmental aspects and impacts; identification of legal requirements and establishment of a legal register; setting of objectives, targets and programmes; establishment of documentation; corporate audit / system review; management review.
The Grafton and Coffs Harbour laboratories were certified to ISO 14001:2004 at an audit held in May 2009. Using the new EMS framework, implementation is continuing at a number of sites across the Practice. The aim is to establish an integrated management system, encompassing Environment, Quality, Safety and LEAN principles.
The benefits of the established EMS to the organisation have been improved environmental awareness and performance (e.g. decreased usage of energy, water, consumables, increased recycling) cost savings, business efficiencies, compliance with regulations, improved corporate image, marketing opportunities, reduced risk of disaster and improved relationships with the public and community. Since the implementation process began, SNP has seen a cultural change, with Environmental Management being at the forefront of many business decisions.
TAT is the time from sample receipt in the laboratory to result release to a Clinical Information System. In our laboratory transfer is usually complete within 1 minute of release. This presentation demonstrates a graphical method of calculating the key performance parameters which allows them to be easily monitored over an extended period of time.
TAT is calculated as 3 or 5 minutes segments, or hours or days, together with various categories. The number of TATs per time slot is counted and from this the cumulative total up to the various time slots is calculated thus allowing the cumulative percentage of tests resulted to the time slots to be determined. Resultant data plotted as ‘% completed tests’ versus ‘time slots’ enables percentile values of TAT to be read from this chart. This method is suitable for a limited number of category plots with 5 being the maximum number such that trends may not be easily seen. An application has been developed, using a bar-chart format, allowing large amounts of data to be assessed with up to 50 category plots visible per screen and trends being more discernable. This presentation can show overall performance for an analyte and also of subsets based on urgency, source, time and day of the week resulted.
The data presented for potassium and D-dimer tests demonstrate the effect of changed workflows and new instruments on result performance.
A graphical method of summarising vast amounts of performance data is presented with improved applications for management and technical reviews. It can be used to assess past performance, set realistic targets and measure on-going performance as a Performance Index as well as provide other useful data, e.g. samples outside specified limits. This form of presentation would be best applied directly on-line.
The effect of the pathology workforce crisis on the pathology industry has been widely discussed. Sullivan Nicolaides Pathology (SNP), a comprehensive central laboratory service, provides 24 hour stat laboratory services via peripheral laboratories to metropolitan and regional private hospitals throughout Queensland. The challenge is to maintain this service in the face of a change in the skill base of the workforce.
Routine laboratory testing in each laboratory was reviewed and the competencies needed to perform each step were identified. A minimum standard of competency was defined to allow non-scientifically trained staff to perform all pre-verification steps in the analytical process and training programs implemented. The LIS allows remote access to a laboratory’s test and quality control results. Test results which meet the criteria set by SNP auto-verify, meaning these results are not viewed by laboratory staff. Test results that did not auto-verify are able to be verified locally or remotely by scientists with appropriate authorisation privileges.
Specially designed training programs were developed to equip staff with the appropriate skills which included equipment troubleshooting and quality control interpretation. Testing at the peripheral laboratory is performed by a competent person, and results either auto-verify, or are remotely verified by a scientist either locally or at the remote (verifying) laboratory. This remote verification has been successfully employed in five peripheral laboratories. The role of Remote Verification Coordinator was created to manage the process.
Substituting non-scientifically qualified staff for scientifically qualified staff and employing remote verification allows SNP to continue to provide essential 24 h 365 day/year service to metropolitan and regional hospitals despite the pathology workforce crisis.