Search tips
Search criteria 


Logo of jclinpathJournal of Clinical PathologyVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
J Clin Pathol. 2007 May; 60(5): 584.
PMCID: PMC1994539

Limitations of transferability of absolute cut‐points in non‐standardised assays

We read the paper of Matull et al1 on the biochemical markers of acute pancreatitis with much interest. We would, however, like to comment on some aspects that could be misleading.

First, absolute cut‐off points as part of diagnosis or risk prediction need to be employed with great care.2 This is especially true for enzyme assays such as alanine aminotransferase, aspartate aminotransferase, lactate dehydrogenase and amylase. For example, more than 200 methods of activity determination have been described for amylase.3 Amylase analytical methods vary with respect to the type of substrate and substrate concentration, use of auxiliary enzymes as well as assay conditions such as pH, temperature, incubation period and protein concentration.3,4

Accordingly, reference intervals are method dependent, and it could be confusing to imply that values such as 600 or 1000 U/l have certain clinical sensitivities and specificities1 without mentioning which analytical methods were employed, what the reference intervals were or that the figures will vary between laboratories owing to a lack in assay standardisation.2,4 Reference intervals for amylase range from 90–300 U/l (Phadebas, Magle AB, Lund, Sweden) through 25–115 U/l (Dimension RxL, Dade Behring, Milton Keynes, UK) to 22–89 U/l (Olympus, Clare, Ireland). Not surprisingly, there are large differences between different assays when results from external quality assessment schemes are examined. The difference between the intermethod reference ranges is so significant that they have been flagged up as a clinical governance issue in some hospitals when changing routine clinical chemistry analysers.

The second, unrelated, aspect on which we wished to comment relates to interferences. Significant increase in serum triglyceride concentrations is a recognised interferent in some assays; however, it is unlikely to be an interferent or a competitive interferent in all 220+ amylase assays3,4,5 due to the inherent differences in assay formulation. Most modern amylase assays are not subject to such problems when the triglyceride concentration is <10 mmol/l.

It also needs to be noted that prognostic criteria such as Ranson and Glasgow are also affected by the above transferability limitations. Direct transfer without validation, especially when assays have not been standardised will lead to avoidable errors in diagnosis and assessing prognosis.


Competing interests: None declared.


1. Matull W R, Pereira S P, O'Donohue J W. Biochemical markers of acute pancreatitis. J Clin Pathol 2006. 59340–344.344 [PMC free article] [PubMed]
2. Bossuyt P M, Reitsma J B, Bruns D E. et al Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ 2003. 32641–44.44 [PMC free article] [PubMed]
3. Foo A Y. Amylase measurement‐which method? Ann Clin Biochem 1995. 32239–243.243 [PubMed]
4. Panteghini M. In: Tietz textbook of clinical chemistry and molecular diagnostics. 4th edn. WB Saunders: Philadelphia, 2006. 616–619.619
5. Junge W, Wortmann W, Wilke B. et al Development and evaluation of assays for the determination of total and pancreatic amylase at 37°C according to the principle recommended by the IFCC. Clin Biochem 2001. 34607–615.615 [PubMed]

Articles from Journal of Clinical Pathology are provided here courtesy of BMJ Publishing Group