|Home | About | Journals | Submit | Contact Us | Français|
This study examined heel ultrasonography, levels of 25-OHD, and levels of urinary NTX-I in patients with a recent osteoporotic hip fracture to see if they were clinically useful.
T scores of stiffness index(SI) obtained from quantitative heel ultrasound(QUS), levels of serum 25-hydroxyvitamin D(25-OHD) and levels of urine N-terminal cross linking telopeptide of type I collagen(NTX-I) were obtained in 53 female and 32 male hip fracture patients. Sixty five female patients and five male patients attending our geriatric clinic were used for comparison.
The T scores of SI of the hip fracture patients were less than those of the geriatric clinic patients. The difference was signficant in females(p = 0.0001), but not in males (p = 0.1). Serum levels of 25-OHD were less than 28 ng/ml in 50 out of 59 patients, and less than 5 ng/ml in two patients. Levels of urinary NTX-I were variable, and were not correlated with other parameters.
Patients who have sustained a hip fracture have a low SI determined by QUS; this is easy to perform and it provides a baseline T score from which to assess treatment effects. The majority of these patients are vitamin D deficient and measurement of the level of 25-OHD would enable physicians to prescribe an appropriate dose of vitamin D. Urine NTX-I measured shortly after a hip fracture is not clinically helpful.
There is widespread inertia among physicians concerning screening for and treatment of osteoporosis. A survey done in 2000 of 490 physicians in Ontario with 275 respondents illustrates this. Ninety two per cent of the respondents were family phsycians; 28.7% were caring for more than 100 patients in long term care. Most (85.8%) saw from one to 10 hip fractures yearly in their practices. Half of the respondents estimated the prevalence of osteoporosis to be 40–80% among their long term care patients; 45.5% said that they did not routinely assess their patients for the disease and 26.8% did not routinely treat it. Half(50.9%) of physicians would treat patients at high risk based on clinical history; 47.9% if patients had a vertebral compression fracture on plain X-ray examination; 43.8% if patients were highly functional; 42.0% if osteoporosis were confirmed with bone mineral densitometry; and 30.0% if patients had a recent fracture. Perceived barriers to initiating treatment included cost of therapy, patient or family reluctance to accept therapy, and time or cost of diagnosis .
Port et al reported that a study involving 348 hip fracture patients, in which only 18% of women and 8% of men with a prior hip fracture were on any specific anti-osteoporosis therapy . In a survey of 2804 members of a health maintenance organization consisting of women aged 50 to 89 years and men aged 65 to 89 years, who had sustained an osteoporotic fracture, only 4.6% had treatment initiated after the fracture. Women sustained 80.7% of these fractures, but only 8.4% had BMD measurements, and 42.4% of the latter received any treatment during the 2 years of the study .
The gold standard for diagnosing osteoporosis is dual energy Xray absorptiometry (DXA), but it cannot be used in the immediate aftermath of a hip fracture, because pain prevents patients from lying immobile on a metal table, which this technique requires. When the patients are discharged, their restricted mobility makes it difficult for them to travel to a facility which performs DXA. Prospective studies involving a total of 16,000 patients comparing QUS performed at the heel with DXA performed at the spine, hip, and wrist, show that they are similar or identical in their ability to predict osteoporotic fractures of the spine, hip, and wrist [4, 5, 6]; we believed that QUS would overcome the difficulties of screening with DXA in this context. Heel QUS does not require a licensed technician. It is comparitively cheap, does not require special training to use, and can be used at bedside. The actual measurement takes one minute to perform and the patient can sit or lie in any comfortable position. These advantages are considerable when compared to DXA.
There is evidence that vitamin D deficiency may contribute to hip fracture [7, 8, 9, 10, 11] and we thought that the simple blood test required for measuring 25-OHD, a surrogate of vitamin D status, would be a simple and valuable screening procedure in hip fracture patients.
We also explored screening with urine NTX, an index of bone turnover, because increased bone turnover has been reported to be predictive of osteoporotic fracture [12, 13]. However, Gerdhem et al reported that high levels of bone turnover markers predict fractures that engage trabecular bone, but that they did not predict hip fracture .
The objective of this study was to see if these three simple screening tests would provide information for guiding physicians taking care of patients who had sustained a hip fracture. To our knowledge there are no previous reports of QUS or NTX-I after hip fracture although there are several reports of measurements of levels of 25-OHD.
The study was approved by the hospital Institutional Review Board. It was conducted from January 2003 until September 2005. All subjects gave informed consent. Risks to confidentiality were minimized by coding the subjects with numbers rather than with names. The criteria for participation were: any woman aged 50 or over, or any man aged 60 or over who had been hospitalized for a hip fracture. Patients who were incompetent were excluded as were patients with high energy trauma or pathologic fractures or those who had secondary osteoporosis, EG hyperparathyroidism, renal failure, non-skin cancer, or bone disease other than osteoporosis. We were informed of admissions for hip fracture by the orthopedic department. Patients were interviewed within one to six days after fracture repair. Patients were weighed in a chair scale, and they were measured lying supine with a tape measure to the nearest half inch. Body mass index(BMI) was calculated from the formula BMI = weight in kg/(height in meters)2.
Serum levels of 25-OHD were measured using a radioimmunassay(Diasorin, USA). Sensitivity was measured as 1.5 ng/ml, intra-assay precision 8.2%, and interassay precision 10.5%, with a laboratory reference range of 9–54 ng/ml.
Urine concentrations of NTX-I were measured by competitive immunoassay(Vitros Immunodiagnostic, USA). Sensitivity was measured as 10 nanomoles bone collagen equivalents per liter per millimole creatinine per liter(nM BCE/(mM creatinine), intra-assay precision 3.9%, inter-assay precision 9.4%. The reference range was 10–110 nmol BCE/mmol creatinine in women and 11–103 nmol BCE/mmol creatinine in men.
As far as possible a second morning urine was obtained for measurement of urine NTX-1.
SI was measured in the calcaneus contralateral to the side of the fracture, using an Achilles Express ultrasonometer. The measurement was performed with the patient either lying or sitting. The precision of this machine is ± 2%. SI combines both speed of sound(SOS) and broadband ultrasound attenuation(BUA) in the following formula: SI = (0.67 × BUA + 0.28 SOS) − 420 . The SI is reported as a T score which is the number of SDs away from the mean T score of the database of 314 women aged 20–35 . No database exists for males and thus the T score of the male patients used the same database of females.
Table 1 shows the characteristics of the female hip fracture and office patients.
The mean age(81.3) of the hip fracture group was older(p = 0.0343) than the mean age(78.3) of the office patients. The mean T score of the hip fracture group was −2.5 which was lower(p = 0.0001) than the score of the office group(−1.7).
In females T score was correlated with BMI(r = 0.41, p < 0.001) and with age(r = −0.31, p < 0.001). Age/BMI correlated with T score(r = −0.44, p < 0.001). This ratio(age/BMI) was significantly different in patients with hip fracture vs office patients(p < 0.0001). 37 female patients had osteoporosis defined by a T score of ≤ −2.5, 13 patients had osteopenia defined by a T score of > −2.5 < −1.0, and three patients had a normal T score, defined as a score ≥ −1.0.
Table 2 shows the characteristics of the male hip fracture and office patients.
The number of male office patients without hip fracture was five, too small to be a useful comparison group for the 32 hip fracture patients. Thus there were no statistically significant differences between the male hip fracture patients and the office patients for age and T score; however there was a statistical trend suggesting that the mean T score for the hip fracture patients was lower than the mean T score for the office patients(p < 0.1). The mean BMI of the male hip fracture patients was 23.8 whereas that of the office patients was 27.1(p = 0.16). In males, T score did not correlate with BMI(r = 0.3, p = 0.08) or with age (r = −.0.009, p = 0.96). The mean of the ratio age/BMI in the male hip fracture patients(3.4 ± 0.7) was not siginificantly different from that of the office patients(3.1 ± 1.0)(p = 0.35). There was no correlation between T score and the ratio age/BMI among males, although there was a trend(r = −0.26, p = 0.13).
Four patients had levels of 25-OHD above or equal to 32 ng/ml. The mean T score for the females was significantly lower(p = 0.002) than the mean T score for the males. The mean age(81.3 years) of the female hip fracture patients was higher than the mean age(79.0 years) of the male hip fracture patients(p < 0.07). There were no other significant differences between the male and female patients.
Among the male hip fracture patients, as among the females, there were no correlations between the levels of 25-OHD, NTX, and T score.
12 male patients had a T score of ≤ −2.5, 13 males had a T score of > −2.5 < −1.0, 7 males had a T score of ≥ −1.0.
In our study, the parameter which best discriminated between fracture or no fracture among the females was the ratio age/BMI(p < 0.0001). However this ratio(age/BMI) was not significantly different in males with and without hip fracture. There are no previous reports of this ratio as a predictor or as an association of osteoporotic fracture, although it is well known that increased age , and decreased BMI  are both associated with an increased risk of osteoporotic fracture. The importance of this ratio which is a function only of height, weight, and age of the patient, is that it is easily accessible but will still provide better(statistical) prediction of osteoporotic fracture than measuring 25-OHD or QUS - in females.
Wildner et al identified 959 postmenopausal women from the NHANES III study to assess the relative contribution of risk predictors for low BMD. Age and weight were by far the most informative predictors for low BMD out of 20 candidate risk predictors .
Two recent studies agree that a low BMI is a risk factor for osteoporosis and fracture. One study of over 15,000 men and women, that of Welch et al , found that the effect of BMI on the linear regression coefficient of BMI with broadband ultrasound attenuation (BUA) was one and a half times greater in women than in men(p < 0.001). That is, BMI has a greater effect on T score and therefore bone mass in women than in men. Welch at al also found that the age-related decline in BUA was five times greater in women than in men; they stated that, ‘it is possible that in postmenopausal women, in whom gonadal production of sex hormones is low, endogenous estrogen levels are related to conversion of precursors to estrogen, which occurs in fat cells, so …BMI may indicate not just loading forces but also endocrine status.’ A meta-analysis by De Laet et al  of more than 60,000 men and women reported that the risk ratio for osteoporotic fracture per unit change in BMI was very similar in men and women((p > 0.30).
Our findings are a reminder that the combination of advanced age and low BMI in women signify a high risk of osteoporotic fracture.
There have been no previous studies of QUS in patients who have sustained a hip fracture. Although they varied greatly, the T scores of QUS in the female hip fracture patients were significantly lower than those of the non-fracture patients. This can be seen graphically in Figure 1. Previous studies of QUS have focused on T score for assessment of fracture risk. Our study demonstrates that besides being very simple to do, a low T score is indeed associated with fracture. Our data also show that hip fractures occur in patients with osteopenia or a normal T score. This indicate that although risk assessment is meaningful for large groups, is of limited value for individual patients
For QUS to be as useful as DXA, it needs to be sufficiently sensitive to be able to show responses to treatment for osteoporosis. There are conflicting reports about the usefulness of QUS when used for follow up. In a study of 18 patients with osteoporosis treated with an anti-resorptive agent, Ingle et al found that finger ultrasound was similar in clinical utility to DXA at the femoral neck for monitoring treatment . In a four year longitudinal study, Gonnelli S et al compared BMD at the lumbar spine with QUS in 150 menopausal women, for monitoring response to alendronate. They found that SI showed a sensitivity which was only slightly lower than BMD. They concluded that although spinal BMD remains the optimal method, QUS at the heel and in particular SI seems to be a sensitive tool for monitoring the response to alendronate . Frost ML et al did a 2 year longitudinal study on 195 postmenopausal women to monitor the response to antiresorptive therapy comparing QUS with BMD measurements at the lumbar spine, femoral neck, and total hip. They found that calcaneal QUS showed a highly significant response to antiresorptive therapy, but their conclusion was that the precision of QUS was not good enough to allow QUS to be used for monitoring response to treatment . Weiss et al found that ultrasound performed in several skeletal sites increased after treatment of postmenopausal osteoporosis with alendronate. They concluded that peripheral QUS measurements could be used for following skeletal changes in response to alendronate . Therefore more study is needed for a definitive conclusion regarding the utility of QUS for monitoring osteoporosis management.
We found that 50 out of 59 patients in the present study were vitamin D insufficient using a levels of 25-OHD of 28 ng/ml as a cut point. Two patients had osteomalacia defined by a level of < 5 ng/ml. One of these was an 86 year old male with a T score of −3.1, and a BMI of 21.2. The other was an 86 year old female with a T score of −4.0 and a BMI of 18.6. In addition to vitamin D deficiency(osteomalacia), these two patients had two major risk factors for osteoporosis, namely advanced age and low BMI.
The mean levels of 25-OHD of 20.2 ng/ml and 16.1 ng/ml in the male and female hip fracture patients, are somewhat higher than the range of 8.9–16 ng/ml found in previous studies of hip fracture patients[7, 8, 9, 10, 11]. We do not know the levels of 25-OHD of free living subjects from the same community as those with hip fracture. We also looked at levels of serum albumin which were significantly lower in the patients who had sustained a hip fracture than in the office patients. We have not included the data because serum albumin is a reverse acute phase reactant and it has been found that a fracture causes it to decrease.
A recent editorial pointed out that there is currently no standard definition of optimal vitamin D status for the skeleton, but there is a common opinion that the optimal serum 25-OHD level for bone health is between 50 and 80 nmol/l, with five of six authorities recommending between 70 and 80 nmol/l(equivalent to between 28 and 32 ng/ml) .
However, the levels that we and others are reporting are at one point in time and it is not known how they reflect vitamin D status over an extended period. The same comments apply to previous studies of vitamin D levels in hip fracture patients. The levels of 25-OHD in subjects who have not sustained a hip fracture from the same communities as those who have had hip fractures have not been reported. Levels of 25-OHD among free living and institutional communities have been lower than the currently recommended threshold mentioned in the last paragraph. It is not known if there is a fracture threshold for 25-OHD, nor is it known what is the relationship between bone strength and 25-OHD levels in humans. In rats with undetectable levels of 25-OHD, there was a reduction in BMD in the distal tibial trabecular bone; the femoral necks were weaker than those of non non-vitamin D deficient animals. Another way in which vitamin D deficiency may contribute to hip fracture is by causing proximal muscle weakness and a consequent tendency to fall; the threshold level of 25-OHD at which this happens has not been defined.
The studies which show that osteoporotic fracture is less frequent in subjects treated with vitamin D[26, 27, 28], have also reported levels of 25-OHD at one point in time, as have the studies which report that vitamin D does not prevent osteoporotic fracture[29, 30, 31]. As it is not known how one measurement of 25-OHD reflects the vitamin D status of an individual over time, these studies may give an incomplete picture of vitamin D sufficiency or insufficiency. But in the light of current recommendations, our study suggests that patients who have sustained a hip fracture should have their levels of 25-OHD measured, and that the majority will require supplementation with vitamin D.
The levels of NTX in our study showed considerable variation and did not correlate with the levels of 25-OHD, T score, or age in hip fracture patients. Veitch et al studied several bone markers sequentially after tibial shaft fractures in 18 patients. For the bone resorption marker, serum C-telopeptides of type 1 collagen(β-CTX), there was a decrease of around 40% from baseline by the third day of fracture, rising 100% above baseline by the seventh day and reaching a maximum of 150% of baseline by two weeks after fracture . These results demonstrate the wide fluctuation in bone resorption which occurs during the week following fracture, which was when we measured urine NTX in our patients. Thus measuring bone markers during this time likely reflects the effects of fracture rather than baseline bone turnover. Even when not used immediately after a fracture, Gillett and Vasikaran reported that urinary NTX results rarely alter the clinical management of patients with osteoporosis.
One limitation was the lack of a control group; however finding a suitable control group may not be possible, because hip fractures are more common with increasing age, thinness, and malnutrition. Our study would have been improved if we had measured the levels of 25-OHD in our geriatric office patients as well as in the hip fracture patients. A nutritional assessment of the hip fracture patients would have been useful, given that previous studies have indicated a significant prevalence of malnutrition in patients with a hip fracture. There was no database of young males from which to calculate the true T score of the male patients.
This study showed a high prevalence of vitamin D deficiency in hip fracture patients. The T scores of QUS and BMI were significantly lower, whereas age and the ratio age/BMI, were significantly higher, among the female hip fracture patients than in patients attending our geriatric clinic. There were similar changes among the male patients but not all reached statistical significance. There was a wide range of urinary NTX levels in the hip fracture patients which may reflect an acute response to the fracture; it is not helpful to measure bone makers in the immediate aftermath of a fracture.
A database of QUS in young males is needed.
This study received financial support from the Alliance for Better Health(Procter & Gamble Pharmaceuticals Inc and Sanofi-Aventis).