Osteoarthritis (OA) is a common age-related disabling locomotor disease characterized by degradation of articular cartilage. The most commonly used radiographic method to assess cartilage damage is measurement of the joint space width (JSW). However, a limitation of using plain radiographs for detecting cartilage destruction is that significant cartilage degradation must have occurred in order to be visible on a radiograph (1). Therefore, cartilage degradation that is detectable on radiographs is already considered irreversible joint damage. Because of its relatively insensitive reflection of the disease process, it also takes at least 1 or 2 years to detect progression of damage that has been visualized on radiographs.
To overcome this, biochemical markers have been developed with the aim of detecting changes in OA with more reliability and sensitivity, preferably in an early stage of OA (1–4). Biochemical markers are molecules derived from connective tissue matrices that are released into biologic fluid during the process of tissue turnover (1, 2). Such biochemical markers might be useful for the early identification of patients with OA or patients at high risk for progression, for monitoring disease progression, and for assessing therapeutic response in OA, all because of their greater sensitivity compared with radiographs (2, 4).
One approach to the identification of such a marker could involve the analysis of cartilage metabolism. Proteoglycans and type II collagen are the major constituents of cartilage (4). Type II collagen is localized almost exclusively in cartilage, where it is a major structural component of the tissue. Hence, measurements of fragments derived from this protein may potentially represent a specific marker for cartilage degradation (1, 3). Recently, a specific marker of cartilage degradation, measured as the urinary concentration of C-telopeptide fragments of type II collagen (CTX-II), was recognized (2, 3, 5). Mouritzen et al described slightly increased concentrations of CTX-II with increasing age and higher CTX-II concentrations in women (both after age 55 years) as well as in subjects with a higher body mass index (BMI) (5). Some evidence supporting the use of CTX-II as a marker has already been obtained. Urinary CTX-II levels are elevated in diseases in which there is increased cartilage turnover, such as OA (1, 2) and rheumatoid arthritis (6). Garnero et al reported weak associations of CTX-II with prevalent radiographic knee OA (1) and modest associations with progression of radiographic knee OA (2). However, those studies were small, and it remains uncertain to what extent CTX-II is an independent marker for radiographic OA as well as which factors could modify the relationship between CTX-II and radiographic OA.
We were therefore interested in exploring the extent to which the CTX-II marker could be considered to be independent of known risk factors for radiographic OA, such as age, sex, and BMI. Because of the limited numbers of subjects included to date in studies of CTX-II, there is a clear need to examine larger populations to obtain more accurate estimates. Furthermore, it is conceivable that a dynamic change in cartilage metabolism could be detected in combination with factors that might reflect an ongoing OA process, such as the presence of joint pain. As such, joint pain might then be considered a potential effect modifier of the relationship between CTX-II and radiographic OA.
Therefore, we undertook the present study to investigate the association between CTX-II and the prevalence and progression of radiographic OA of the knee and hip in a large population of men and women ages ≥55 years. Additionally, we stratified the baseline associations between CTX-II and radiographic OA of the knee or hip for the presence or absence of knee or hip pain at baseline.
SUBJECTS AND METHODS
- Top of page
- SUBJECTS AND METHODS
The study population consisted of participants in the Rotterdam Study, a prospective cohort of men and women ages ≥55 years. The objective of the Rotterdam Study is to investigate the incidence of, and risk factors for, chronic disabling diseases. The rationale and study design have been described previously (7). The focus is on neurogeriatric, cardiovascular, ophthalmologic, and locomotor diseases. All 10,275 inhabitants of Ommoord, a district in Rotterdam, were invited to participate. The response rate was 78%, resulting in 7,983 subjects participating in the present study. Written informed consent was obtained from each participant. The Medical Ethics Committee of the Erasmus Medical Center has approved the Rotterdam Study.
For the present study, we selected 1,235 subjects who were enrolled in the Rotterdam Study. The selection was based on the availability of radiographs of the hip and knee (both at baseline and at followup) and on the availability of urine samples at baseline. The fact that subjects had to be mobile enough to visit the research center at baseline and followup (and well enough to survive the followup period) led to the selection of a relatively younger and healthier population. Compared with the total Rotterdam Study population, the present study population was indeed younger (ages 66.6 years versus 70.6 years), had a lower prevalence of lower limb disability at baseline (an index score of ≥0.5; 11.4% versus 35.5%), and had a somewhat lower percentage of women (58.4% versus 61.1%). The baseline measurements were conducted between April 1990 and July 1993, and the followup measurements were conducted between 1996 and 1999, with a mean followup time of 6.6 years (range 5.1–9.4 years).
Weight-bearing anteroposterior radiographs of the knee and hip were obtained at 70 kV, a focus of 1.8, and a focus-to-film distance of 120 cm, using Fuji High Resolution G 35 × 43–cm film (Fuji Photo Film Company, Kanagawa, Japan). Radiographs of the pelvis were obtained with both feet in 10° internal rotation and the x-ray beam centered on the umbilicus; those of the knee were obtained in extended position with the patella in central position. Two trained readers who were unaware of the clinical status of the patients independently evaluated the radiographs of the knee and hip at baseline and followup. All radiographs were grouped by patient and read as chronologically ordered pairs, the chronological order being known to the reader (chronologically ordered reading procedure) (8).
At baseline, radiographic OA of the knee and hip was quantified by measurements according to the Kellgren/Lawrence (K/L) grading system (9–12) (atlas-based) using 5 grades (from 0 to 4). A person was considered to have radiographic OA of the knee or hip if the K/L score of one or both joints was ≥2.
At baseline and followup, the minimal JSW of the knee and hip joints was measured using a 0.5-mm graduated magnifying glass laid directly over the radiograph (13). For the knee, the medial and lateral compartments were measured, and for the hip, the lateral, superior, and axial compartments were measured, as described previously by Croft et al (13). Joint space narrowing (JSN) was defined as the JSW at baseline minus the JSW at followup. Because of the absence of consensus regarding the cutoff point for JSN, we used different cutoff points for JSN: 1.0-, 1.5-, and 2.0-mm decreases in the JSW between baseline and followup. JSN was evaluated in each compartment; for the knee, JSN in at least 1 of 2 compartments (medial and lateral ) was defined as positive progression, and for the hip, JSN in at least 1 of 3 compartments (lateral, superior, and axial ) was defined as positive progression. Additionally, we used JSN of the medial compartment of the knee as a definition of progression. Radiographic progression of JSN can be regarded as the most reliable measurement of OA progression (15).
The radiographs of the knee were scored for OA by two independent observers who were blinded to all data for the participant, as described previously (14, 16). After each set of 150 radiographs, the scores of the two readers were evaluated. Whenever the K/L score differed, the two readers met to read the radiographs together, and a consensus score was determined. Two independent readers tested the interrater reliability of scoring of the hip in a random set of 148 radiographs. We determined the interrater reliability for K/L scoring to be 0.68 (kappa statistic), and for measurement of minimal JSW, we obtained an intraclass correlation coefficient of 0.85 (12).
Subsequent to overnight fasting, urine samples were obtained from all subjects at baseline and kept frozen at –20°C. Monoclonal antibody mAbF46, specific for CTX-II fragments, was used in a competitive enzyme-linked immunosorbent assay developed for measurement of urine samples, as described previously (3). In order to ensure the reproducibility and performance of the assay, 3 genuine urine samples were added as controls on each microtiter plate, and the entire plate was rerun if any of the genuine controls were determined to have a concentration >20% of the predetermined value. The concentration of CTX-II (in ng/liter) was standardized to the total urine creatinine (mmoles/liter), and the units for the corrected CTX-II concentration are ng/mmole.
Potential confounders and effect modifiers.
At baseline, trained interviewers performed an extensive home interview on demographic characteristics, medical history, risk factors for chronic diseases, and use of medications. Lower limb disability was assessed using a modified version of the Stanford Health Assessment Questionnaire (16). A lower limb disability index (LDI) was obtained by calculating the mean score of answers to 6 questions, as described previously (12). We used the LDI to measure the participant's mobility. The subject was asked about the presence of knee and hip pain (“Did you have joint complaints in your right/left knee/hip during the last month?”) during the home interview at baseline.
Height and weight were measured with participants wearing indoor clothing without shoes. BMI was calculated as weight in kilograms divided by height in square meters.
Differences in baseline characteristics were evaluated by analysis of variance for continuous variables and by chi-square test for categorical variables. Distribution analysis by the Shapiro-Wilk test showed that biochemical markers were not normally distributed; therefore, concentrations of markers were log transformed to obtain a normal distribution before statistical analysis. Hereafter, CTX-II concentrations are the log-transformed values. Influences of age, sex, and BMI on the baseline CTX-II concentration were tested by independent t-tests.
Cross-sectional associations between the CTX-II concentration and radiographic OA of the knee or hip were assessed using logistic regression analysis to calculate odds ratios (ORs) by means of generalized estimating equations (GEEs) (cross-sectional design). This is a procedure of repeated measurements, which is used here to take into account the correlation between the left and right hip, while using each joint (left or right) as the observation unit (17). ORs and 95% confidence intervals (95% CIs) were calculated for each quartile (also shown is the range of values within each quartile) of log-transformed CTX-II concentration, as well as for the change in risk per SD of the mean log-transformed value. The first quartile was used as the referent. For the baseline associations, we calculated crude ORs and adjusted them for age, sex, BMI, and LDI. Additionally, we stratified these associations for the presence or absence of pain in the knee or hip (during the last month).
Longitudinal associations between the baseline CTX-II concentration and progression of radiographic OA of the knee or hip were assessed using logistic regression analysis to calculate ORs to estimate the relative risk for progression, by means of GEEs (longitudinal design). ORs were calculated for each quartile (also shown is the range of values within each quartile) of log-transformed CTX-II concentration, as well as for the change in risk per SD of the mean log-transformed value. For the associations between baseline CTX-II concentration and progression of radiographic OA of the knee or hip, we calculated crude ORs and adjusted for age, sex, BMI, LDI, baseline K/L score, and followup time. The baseline K/L score is a known risk factor for radiographic progression (18, 19). Additionally, we assessed the longitudinal associations between CTX-II concentration and incident osteophytes of the knee or hip at followup. A (2-sided) P value of 0.05 was considered significant.
We estimated the magnitude of confounding by the degree of discrepancy between the unadjusted and adjusted estimates (the change-in-estimate criterion) (20). We chose a cutoff point of 10% as representing important change in the estimate. We used SPSS software, version 11.0 (SPSS, Chicago, IL) and SAS software, version 8.2 (SAS Institute, Cary, NC) for all analyses.
- Top of page
- SUBJECTS AND METHODS
Table 1 presents the baseline characteristics of the total study population stratified for the absence or presence of radiographic OA of the knee or hip. In this study population, with a mean age of 66.6 years, 19.2% of the subjects had radiographic OA of the knee and 10.0% had radiographic OA of the hip (K/L grade ≥2). During the last month before the baseline interview, 12.3% of all subjects had knee pain and 18.1% had hip pain. The median CTX-II concentration (not log transformed) of the study population was 177.0 ng/mmole. Participants with radiographic knee OA were 3.1 years older, more frequently female (70.5% versus 50.2%), 3.9 kg heavier, 2.2 cm shorter, and had a higher BMI (mean difference of 2.1 kg/m2) than those without radiographic knee OA. Subjects with radiographic hip OA were 3.8 years older than those without radiographic hip OA. Compared with those with radiographic hip OA, persons with radiographic knee OA were more frequently female (70.5% versus 58.5%), were 2.1 kg heavier, and had a higher BMI (mean difference of 1.3 kg/m2).
Table 1. Baseline characteristics of the study population, stratified by the absence or presence of radiographic OA of the knee or hip*
| ||Study population (n = 1,235)||Radiographic knee OA||Radiographic hip OA|
|Subjects without (n = 998)||Subjects with (n = 237)||Subjects without (n = 1,112)||Subjects with (n = 123)|
|Age, mean ± SD years||66.6 ± 6.8||66.0 ± 6.6||69.1 ± 6.9†||66.2 ± 6.7||70.0 ± 6.7†|
|Weight, mean ± SD kg||73.8 ± 11.5||73.1 ± 11.2||77.0 ± 12.0†||73.7 ± 11.6||74.9 ± 10.8|
|Height, mean ± SD cm||167.5 ± 9.1||167.9 ± 9.2||165.7 ± 8.4‡||167.5 ± 9.1||167.5 ± 8.8|
|Body mass index, mean ± SD kg/m2||26.3 ± 3.6||25.9 ± 3.4||28.0 ± 3.9†||26.3 ± 3.6||26.7 ± 3.4|
|Presence of knee pain, %||12.3||15.1||30.8†||17.6||23.0|
|Presence of hip pain, %||18.1||11.1||17.1||10.3||29.9†|
|Lower limb disability, %||11.4||9.0||21.7||8.5||34.6|
|CTX-II concentration, median ng/mmole§||177.0||167.0||228.0†||172.0||231.5†|
The CTX-II concentration was 72.3 ng/mmole higher in women than in men (P < 0.0001), increased by 1.1 ng/mmole per year with age (P for trend = 0.03) (Figure 1), and increased by 3.3 ng/mmole per kg/m2 with higher BMI (P for trend < 0.0001). When we excluded participants with radiographic OA of the knee or hip at baseline and those with incident radiographic OA of the knee or hip at followup, only the sex difference in CTX-II concentration remained.
Figure 1. Distribution of median concentrations of C-telopeptide fragments of type II collagen (CTX-II) by age and sex. The concentrations are not log transformed. ⋄ = women; ▪ = men.
Download figure to PowerPoint
Baseline CTX-II concentration (by quartiles, with range of values shown within each quartile) was higher in subjects with baseline radiographic OA of the knee or hip than in those without baseline radiographic OA of the knee or hip (Table 2). The crude data showed a stronger association for the highest quartile of CTX-II concentration with radiographic knee OA than with radiographic hip OA, but the difference was not significant. After adjustment for age and sex, the risk estimate increased for the hip and decreased for the knee, resulting in similar ORs for the hip and knee. Additional adjustment for BMI and lower LDI did not essentially change the risk estimates for the knee and hip. Overall, we observed a clear trend with increasing CTX-II concentration toward a stronger association with prevalent radiographic knee and hip OA.
Table 2. Cross-sectional association between baseline CTX-II concentration and baseline radiographic OA of the knee and/or hip*
| ||Radiographic knee OA (n = 237)||Radiographic hip OA (n = 123)|
|Crude OR (95% CI)||Adjusted OR (95% CI)†||Crude OR (95% CI)||Adjusted OR (95% CI)†|
|Quartile (range of CTX-II values)‡|| || || || |
| First (1.49–2.10)||1||1||1||1|
| Second (2.11–2.25)||1.7 (1.0–2.9)||1.7 (1.0–2.9)||1.3 (0.7–2.5)||1.5 (0.8–2.9)|
| Third (2.26–2.39)||3.2 (2.0–5.2)||2.8 (1.6–4.6)||1.7 (0.9–3.1)||2.1 (1.1–4.0)|
| Fourth (2.40–3.11)||5.2 (3.3–8.4)||4.2 (2.5–7.0)||3.6 (2.0–6.2)||4.2 (2.2–7.8)|
|P for trend||<0.0001||<0.0001||<0.0001||<0.0001|
|Change in risk per SD§||1.9 (1.6–2.2)||1.7 (1.4–2.0)||1.8 (1.5–2.1)||1.8 (1.4–2.2)|
Table 3 shows the associations between baseline CTX-II concentration (by quartiles, with range of values shown within each quartile) and progression of radiographic knee OA using different cutoff points for JSN. We found significant crude associations between decreases in joint space of ≥1.5 mm or ≥2.0 mm and the highest quartile of CTX-II concentration. After adjustment for age, sex, BMI, LDI, baseline radiographic knee OA, and followup time, the risk estimates changed importantly, and only the association between JSN ≥2.0 mm and the fourth quartile of CTX-II concentration reached significance, with an OR of 6.0. Especially for a JSN of ≥2.0 mm, but also for a JSN of ≥1.5 mm, we observed that with an increasing CTX-II concentration there was a clear trend toward a stronger association with progression of radiographic knee OA. Additionally, we assessed the association between the CTX-II concentration and OA progression in the medial compartment. This association did not essentially differ from the above-mentioned associations (for JSN ≥1.5 mm and fourth-quartile CTX-II concentration, adjusted OR 2.0 [95% CI 0.8–5.1]).
Table 3. Associations between baseline CTX-II concentration and radiographic progression of knee OA*
| ||JSN ≥1.0 mm (n = 233)||JSN ≥1.5 mm (n = 73)||JSN ≥2.0 mm (n = 26)|
|Crude OR (95% CI)||Adjusted OR (95% CI)†||Crude OR (95% CI)||Adjusted OR (95% CI)†||Crude OR (95% CI)||Adjusted OR (95% CI)†|
|Quartile (range of CTX-II values)‡|| || || || || || |
| First (1.49–2.10)||1||1||1||1||1||1|
| Second (2.11–2.25)||1.0 (0.7–1.5)||0.9 (0.6–1.5)||1.5 (0.7–3.3)||1.3 (0.6–2.9)||3.7 (0.8–17.7)||4.1 (0.8–20.5)|
| Third (2.26–2.39)||1.2 (0.8–1.8)||1.1 (0.7–1.7)||1.9 (0.9–4.1)||1.5 (0.6–3.3)||3.6 (0.7–17.5)||4.5 (0.9–23.0)|
| Fourth (2.40–3.11)||1.2 (0.8–1.8)||1.1 (0.7–1.7)||2.5 (1.2–5.2)||1.8 (0.8–4.1)||5.2 (1.1–23.8)||6.0 (1.2–30.8)|
|P for trend||0.219||0.730||0.009||0.120||0.033||0.064|
|Change in risk per SD§||1.1 (1.0–1.3)||1.1 (0.9–1.3)||1.5 (1.1–1.8)||1.4 (1.0–1.8)||1.5 (1.0–2.2)||1.6 (1.0–2.5)|
Table 4 shows the associations between baseline CTX-II concentration (by quartiles, with range of values shown within each quartile) and progression of radiographic hip OA using different cutoff points for JSN. The results for the JSN ≥2.0-mm cutoff point are not presented because of insufficient statistical power (n = 11 subjects). At the hip, we found a trend similar to that for prevalent radiographic OA of the knee and hip (i.e., the higher the CTX-II concentration, the stronger the association with progression of radiographic OA). After adjustment for age, sex, BMI, LDI, baseline radiographic hip OA, and followup time, only the association between JSN ≥1.5 mm and the fourth quartile of CTX-II concentration reached significance, with an OR of 8.4. When we compared the association between the different aspects of radiographic OA as measured by the K/L score (i.e., osteophytes and JSN), we observed no association with incident osteophytes of the knee and the hip. The ORs for the fourth quartile of CTX-II concentration were 0.3 for both the knee and the hip (P = 0.288 and P = 0.232, respectively).
Table 4. Associations between baseline CTX-II concentration and radiographic progression of hip OA*
| ||JSN ≥1.0 mm (n = 73)||JSN ≥1.5 mm (n = 24)|
|Crude OR (95% CI)||Adjusted OR (95% CI)†||Crude OR (95% CI)||Adjusted OR (95% CI)†|
|Quartile (range of CTX-II values)‡|| || || || |
| First (1.49–2.10)||1||1||1||1|
| Second (2.11–2.25)||1.1 (0.5–2.6)||1.0 (0.4–2.4)||4.2 (0.5–37.4)||3.9 (0.4–36.9)|
| Third (2.26–2.39)||2.3 (1.1–4.7)||2.1 (0.9–4.6)||8.5 (1.1–68.5)||8.3 (1.0–72.2)|
| Fourth (2.40–3.11)||2.8 (1.4–5.8)||1.7 (0.7–4.0)||12.3 (1.6–95.5)||8.4 (1.0–72.9)|
|P for trend||<0.0001||0.05||<0.0001||0.005|
|Change in risk per SD§||1.6 (1.3–2.0)||1.3 (1.0–1.8)||2.2 (1.5–3.2)||1.9 (1.2–3.0)|
Figure 2 shows the baseline associations between high CTX-II concentrations (fourth quartile) and radiographic OA of the knee and hip, stratified for the absence or presence of knee or hip pain. For this analysis, we compared subjects with a high CTX-II concentration (fourth quartile) with those with a low concentration (first quartile), resulting in a lower number of subjects, as reported before. We observed a substantially stronger association between CTX-II levels and radiographic OA for subjects with hip pain (OR 20.4) than for those without hip pain (OR 3.0). Adjustment for potential confounders changed the risk estimates importantly (from 17.1 to 20.4 and from 2.3 to 3.0 for subjects with and without hip pain, respectively). The difference between the ORs for subjects with and those without hip pain just failed to reach significance (P = 0.105). In the case of radiographic knee OA, the differences in ORs between participants with and those without knee pain were similar but smaller than those found for the hip. After adjustment for potential confounders, the risk estimates changed importantly for participants with (OR from 7.1 to 6.3) and those without (OR from 4.3 to 3.6) knee pain.
Figure 2. Associations between baseline concentration of C-telopeptide fragments of type II collagen (CTX-II; log transformed, highest quartile) and baseline radiographic osteoarthritis (OA) of the knee or hip (Kellgren/Lawrence score ≥2), stratified for the absence or presence of knee or hip pain. Associations are adjusted for age, sex, body mass index, and lower limb disability index. For subjects without knee pain (n = 495), odds ratio (OR) 3.6, 95% confidence interval (95% CI) 2.0–6.5; for subjects with knee pain (n = 115), OR 6.3, 95% CI 2.0–20.0; for subjects without hip pain (n = 531), OR 3.0, 95% CI 1.5–6.0; and for subjects with hip pain (n = 71), OR 20.4, 95% CI 2.3–185.2.
Download figure to PowerPoint
- Top of page
- SUBJECTS AND METHODS
We report the findings of our analysis of urinary CTX-II levels in a large population-based prospective cohort study, indicating a strong relationship between CTX-II levels and the risk of radiographic OA. For persons with a CTX-II level in the highest quartile, we observed a >4-fold increased risk of having prevalent radiographic OA of the knee or hip and 6-fold and 8.4-fold increased risks for progression of radiographic OA at the knee and the hip, respectively. All these associations were found to be independent of known risk factors for radiographic OA, including age, sex, BMI, LDI, baseline K/L score, and followup time. Furthermore, CTX-II seems to be a specific marker for cartilage degradation, since CTX-II is associated with JSN, but not with incident osteophytes.
The baseline associations seemed stronger for the participants with hip pain than for those without hip pain. Because of the low numbers of subjects with hip pain, the 95% CI of the association for those with hip pain is huge (2.3–185.2) and overlaps with the 95% CI of the association for those without hip pain. We confirm that women had higher CTX-II concentrations than men, and we found that this was not explained by prevalent or incident radiographic OA of the knee or hip. Thus, the present study shows that a single degradation marker (CTX-II) can identify patients who are at high risk for rapid progression of joint destruction.
The distribution of CTX-II concentration by age, sex, and BMI in the present study was similar to that described by Mouritzen et al (5). We found a slight rise in urinary CTX-II concentration with increasing age (>55 years), a significantly higher level for women than for men (after age 55 years), and a significantly higher CTX-II concentration in subjects with higher BMI. The increased concentration with age seems to reflect the increase in prevalence of radiographic OA with increasing age. However, the higher concentration found in women remained after we excluded participants with prevalent (at baseline) and incident (at followup) radiographic OA of the knee or hip.
Consistent with this finding, Mouritzen et al (5) reported a sudden and marked increase in CTX-II concentration after menopause. This observation may be explained by a higher turnover rate for cartilage in women after menopause. Indeed, a recent study in cynomolgus monkeys showed that ovariectomy induced OA lesions of articular cartilage (21). Furthermore, in a cross-sectional observational study, Wluka et al (22) reported that the use of estrogen replacement therapy (ERT) for >5 years is associated with greater knee cartilage volume. Similarly, a number of retrospective and observational studies indicated that ERT is associated with decreased prevalence of OA, but this finding is not universal (23). Finally, polymorphisms in the estrogen receptor α gene have been identified as genetic risk factors for knee OA (14). Together, these data suggest that estrogen can prevent cartilage erosion, and they thereby identify the estrogen endocrine system as a significant regulator of cartilage turnover and structural integrity (21). However, the exact mechanism by which estrogen influences cartilage metabolism needs further investigation (21, 23–26).
Type II collagen markers are probably a specific tool for detecting changes in OA (27). Other proposed markers of OA, such as collagen crosslinks, proteoglycan, cartilage oligomeric matrix protein, matrix metalloproteinases, and inflammatory markers (27), reflect general remodeling of the various tissues of the cartilage, bone, and synovium. Up to now, increased serum or urine levels of the different markers have been obtained in small cross-sectional studies (27, 28). This is the first large followup study in which the use of CTX-II as a biomarker for cartilage degradation and disease progression has been investigated.
The strengths of the present study are its size, its population-based prospective design, and the clinically meaningful followup period of 6.6 years. A potential limitation of the present study is that the results are based on a single determination of CTX-II at baseline. Because of a possible diurnal variability of the CTX-II level, we obtained urine samples from all subjects after overnight fasting. However, we found no indication of the presence of a systematic bias due to the inherent variability of the measurements of CTX-II.
Another limitation is a potential health-based selection bias. The subjects in the present study had to be mobile enough to visit the research center at baseline and followup, and they had to be expected to survive the followup period (mean 6.6 years). Overall, participants were generally healthier than nonparticipants. In other words, patients with the most severe symptoms were most likely not included. Therefore, it seems probable that in this younger and healthier population with less frequent lower limb disability and (knee and hip) pain, the prevalence of radiographic knee and hip OA at baseline and the number of subjects with progression of radiographic OA at followup may have been underestimated. This could have resulted in an underestimation of the reported associations.
Another limitation is the procedure used for knee radiography (i.e., the obtaining of serial anteroposterior radiographs). The reliability of radiographic JSW measurements in the knee increases when an anteroposterior radiograph of the knee in 20–30° flexion is used (28, 29), and therefore, this procedure has been recommended for longitudinal studies (30, 31). The procedure used in the present study could have resulted in an under- or overestimation of the reported associations for the knee. The reliability of the JSW measurements on the knee radiographs in the present study was not assessed. As reported by Günther and Sun (32) and by Sun et al (33), the lateral JSW measurement is less reliable than the medial JSW measurement. Additionally, we repeated the analyses of the association between the CTX-II concentration and progression of radiographic knee OA using another definition of progression, namely, JSN of only the medial compartment. The associations we found did not differ essentially from the associations reported here.
Based on the results of the present study, we conclude that the CTX-II concentration is markedly associated with the prevalence and progression of radiographic OA of the knee and hip, and that these associations are independent of known risk factors for radiographic OA. The presence of joint pain seems to augment this relationship, which might reflect the effects of an ongoing OA process. The increase of CTX-II in women after menopause may reflect a protective effect of estrogen against cartilage loss. Further research is necessary to establish the clinical utility of this novel biomarker for OA.
- Top of page
- SUBJECTS AND METHODS
We are very grateful to Dr. E. Odding, Prof. H. A. Valkenburg, and Dr. A. P. Bergink for scoring the radiographs of the knee, F. van Rooij, E. van der Heijden, R. Vermeeren, and L. Verwey for collection of followup data, and we thank Dr. S. C. E. Schuit and F. Imani for help with collection and transfer of urine samples. Moreover, we thank the participating general practitioners, the pharmacists, the many field workers at the research center in Ommoord, and, of course, all of the participants.