| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Internal Medicine, University of Turin, Turin, Italy; 2 Clinical Chemistry Laboratory, San Giovanni Battista Hospital of Turin, Turin, Italy;
aaddress correspondence to this author at: Clinical Chemistry Laboratory, San Giovanni Battista Hospital of Turin, Corso Bramante, 88, 10126 Turin, Italy; fax 39-011-676052, e-mail gmengozzi{at}molinette.piemonte.it
There is evidence that bone turnover in women is more rapid during the first years after menopause than in subsequent years. The assessment of deoxypyridinoline (DPD) cross-links in a fasting urine sample is considered a specific index of bone resorption by osteoclasts and also can be used for monitoring the response to pharmacologic antiresorption treatment. The interpretation of results, however, is hampered by biological and other preanalytical variability (1)(2)(3)(4).
Specific degradation products of the bone matrix, such as DPD and pyridinoline (PYD), closely reflect the rate of bone metabolism. Vesper et al. (1) reported mean within-day variabilities of 67% for DPD (range, 53%–75%) and 71% for PYD (57%–78%). The mean between-day variability was 16% for both PYD and DPD (ranges, 5%–24% and 12%–21% for DPD and PYD, respectively). The mean between-person variabilities across different studies were 34% for DPD (8%–98%) and 26% for PYD (12%–63%) in healthy premenopausal women and 40% (27%–54%) and 36% (22%–61%), respectively, in postmenopausal women. Specimen instability and errors in creatinine measurements were additional sources of variability(1). Some authors have reported that the variability can be reduced by collecting urine for 24 h (or longer) instead of using single voids and by expressing the results as ratios to creatinine(5)(6).
The usefulness of urinary markers of bone turnover in monitoring therapy depends on the within-person variability of these markers compared with their changes in response to treatment (7). Thus, the biological and laboratory variabilities of DPD cross-links are important considerations for clinical evaluation.
We evaluated the biological variability of DPD cross-link concentrations in fasting morning urine specimens collected during a 2-week period from women in their first years of menopause. We examined 64 postmenopausal women (mean age, 53 years; range, 49–57 years) between 1 and 5 years after menopause (mean time after menopause, 39 months; range, 11–58 months). Written informed consent was obtained from all participants in the study.
Each participant had a lumbar bone mineral density t-score (dual-energy X-ray absorptiometry technique) in the range of –1 to –2 SD and a body mass index between 21 and 30 kg/m2. None of the participants smoked or was affected by diseases known to cause secondary osteoporosis, nor had they been treated previously with therapy acting on bone calcium and phosphorus metabolism. Before starting the study, we measured each woman’s serum total and ionized calcium, total alkaline phosphatase, and 25-hydroxyvitamin D concentrations. During the 2-week follow-up period, we measured DPD in fasting first morning urine samples at baseline and thereafter repeated this measurement 3 times each day at the same times and under the same conditions at 7-day intervals. Intra- and interperson variability was reduced by collecting specimens at a specific time of the day, i.e., the first morning void before 1000 in the morning, to avoid any possible influence of diurnal variation and by maintaining similar patient status at each specimen collection, taking into account circumstances such as intake of medications and dietary supplements.
The DPD assay was performed with a solid-phase chemiluminescent enzyme-labeled immunoassay [Pyrilinks-D on a multianalyte automated analyzer (Immulite 2000; manufactured by DPC and purchased from Medical Systems SpA)]. Data are expressed as the ratio to creatinine concentration [modified Jaffe colorimetric method performed on a Hitachi 917 analyzer (Modular Analytics, Roche Diagnostics)]. The within- and between-run imprecision for DPD cross-link measurements at 25 nmol/L were 4.5% and 8.7%, respectively. The intra- and interseries variations in urine creatinine assay were 1.3% and 1.7%, respectively. An equilibrium RIA procedure was used to determine 25-hydroxyvitamin D concentrations after rapid extraction of this and other hydroxylated metabolites from serum with acetonitrile (DiaSorin Inc.). The reference interval given by the manufacturer is 9.0–37.6 µg/L with a detection limit of 2.5 µg/L.
Variations in DPD values around the means in all 3 series of measurements did not follow a gaussian distribution (P <0.2, Kolmogorov–Smirnov test); thus, logarithmic transformation was applied (8)(9). The within-subject biological variation (CVI), based on ANOVA, was calculated with the following formula:
 |
in which SI + A2 is the experimental variation in the results of the 3 determinations for each patient; SA2 is the analytical variation, including both DPD and creatinine imprecision; and M is the mean DPD concentration of the 3 values obtained for each patient studied.
The between-person biological variation (CVG) was calculated with the following formula:
 |
in which ST2 is the total variation derived from all values from all patients.
The index of individuality (II) was calculated according to the formula:
 |
where
 |
For the dynamic assessment of DPD values, to define the difference between 2 consecutive results that may indicate a change in patient status, we calculated the least significant change (LSC) according to the formula (10):
 |
where CVI and CVA represent the calculated within-subject and analytical variabilities, respectively.
Before we performed these calculations, we applied 3 levels of outlier tests to remove outlier points (9), and because an outlier point was identified, 1 patient was excluded from the subsequent analysis.
We calculated correlation coefficients with Spearman rank analysis. P <0.05 was considered statistically significant.
Results of the DPD variability analysis are summarized in Table 1
. Serum total calcium, ionized calcium, and total alkaline phosphatase were within the reference intervals in all patients (data not reported); interestingly, 25-hydroxyvitamin D concentrations were <12 µg/L in 12 individuals (18.7%) and <15 µg/L in 23 individuals (35.9%). CVs from the 3 daily determinations of DPD cross-links were 1.6%–84%, with a median of 18%. We observed a statistically significant correlation between the means of the 3 determinations and CVs for each individual (rS = 0.44; P <0.01). 25-Hydroxyvitamin D concentrations inversely correlated (P <0.05) with the mean DPD cross-links concentrations (rS = –0.34). After we divided the study population into 2 subgroups on the basis of 25-hydroxyvitamin D status, the correlation remained statistically significant (r = –0.36; P <0.05) only in the group with concentrations >15 µg/L (n = 41). On the other hand, vitamin D data did not correlate with the CVs for DPD cross-links.
|
The LSC, calculated from the within-person variability and the analytical imprecision, for DPD results was 70.3%. It is well known that vitamin D deficiency is common in older individuals (11). Our data suggest that in the early postmenopausal years it is not unusual to observe 25-hydroxyvitamin D concentrations below the threshold of 15 µg/L in osteopenic patients. Although inversely correlated to DPD, 25-hydroxyvitamin D concentrations did not appear to affect DPD excretion variability in our study population.
Standardization of urine sampling allowed us to assess variability in DPD excretion in a cohort of women. We found a high biological variability in early postmenopausal women. It is known that the first years of the postmenopausal period are characterized by rapid bone turnover in most cases, and the biological variability of DPD excretion increases with the rate of bone turnover. Increased biological DPD cross-link variability seems to be correlated with bone turnover and represents an important clinical measurement to include in the evaluation of laboratory results.
The index of individuality has been used as a measure of the likely utility of population-based reference intervals (12), with the reference intervals being more useful if the ratio is 
1.4. Our current results show that population-based reference values may be of more limited use for the correct interpretation of DPD concentrations because women in the first 5 years of the postmenopausal period may have values that are unusual for them but still lie within the population-based reference limits.
For biochemical markers of bone turnover, there are no uniform criteria establishing how large a difference between 2 consecutive measurements indicates progression of the disease. Individual responses can be interpreted only in relation to the within-person variability of the marker. In our study, the estimate of within-person variability over a short time did not reflect current clinical practice. In a routine clinical setting in which conditions are not strictly controlled, as in regular patient monitoring, it is likely that the variability will be greater because samples are not collected at the same time of day at all visits. Therefore, the LSC values described here probably underestimate the values that would be found in clinical practice.
Monitoring of bone turnover through DPD assessment in the first 5 years of menopause should take into account variation in the marker concentrations above calculated thresholds of indexes such as the LSC. Our data suggest that points of reference for detection or exclusion of disease progression be established specifically for other bone markers. Relatively large changes between values from sequential samplings should be expected for those analytes such as DPD that display high intraindividual variation during the early menopausal period.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
