Factors influencing bone loss in rheumatoid arthritis: A longitudinal study

Transcription

1 Factors influencing bone loss in rheumatoid arthritis: A longitudinal study B. Cortet 1, M.-H. Guyot 1, E. Solau 1, P. Pigny 2, F. Dumoulin 1, R.-M. Flipo 1, X. Marchandise 3, B. Delcambre 1 1 Department of Rheumatology, CHRU Lille, Hôpital R. Salengro; 2 Department of Biochemistry, CHRU Lille, Clinique Marc Linquette; 3 Department of Nuclear Medicine, CHRU Lille, Hôpital R. Salengro, Lille, France. Abstract Objectives To assess the occurrence of bone loss in rheumatoid arthritis (RA) and to determine the factors influencing bone loss (particularly the usefulness of bone turnover markers) over an 18-month period. Methods A total of 51 patients were studied, 6 men and 45 females (of whom 35 were menopausal). Their mean age was 56 ± 10 years and the mean RA duration was 12 ± 10 years. Twenty-eight (55%) were receiving corticosteroids (10 mg/day for a mean duration of 6 ± 5 years). Several clinical and biological parameters reflecting disease activity or severity were recorded both at the 0 and 18-month investigations. Bone turnover was assessed at baseline by measuring the serum levels of 4 biological markers. Three of them reflected bone formation, i.e., procollagen type I C-terminal propepeptide (PICP), procollagen type I N-terminal propeptide (PINP) and osteocalcin (OC). The fourth, procollagen type I-C terminal telopeptide (ICTP), reflected bone resorption. Bone mineral density (BMD) was measured by dual energy X-ray absorptiometry both at the lumbar spine (LS) and femoral neck (FN) at baseline and 18 months later. Results Bone loss occurred both at the LS: 2.1%, [95% CI: 0.8% - 3.4%, P < 0.005] and femoral neck: 3.1%, [95% CI: 1.1% - 5.1%, P < 0.005]. Bone loss was markedly increased for postmenopausal women at the FN: 5.3% [95% CI: 2.9% - 7.6%, P < 0.005]. Bone loss was not statistically significantly different between users and non-users of steroids. Bone loss at the LS was significantly correlated with both osteocalcin (r = 0.51, P < 0.01) and ICTP levels (r = 0.32, P < 0.05). FN bone loss was correlated with the osteocalcin level only (r = 0.34, P < 0.05). Fast losers (bone loss at the LS above the median) had higher OC (P < 0.01) and ESR (P < 0.05) levels at baseline as compared with slow losers (bone loss at the LS below the median). Conclusion Bone loss occurs in RA particularly at the FN and seems to be influenced by increased bone turnover and high levels of inflammation. Key words Rheumatoid arthritis, bone loss, markers of bone turnover, dual-energy X-ray absorptiometry. Clinical and Experimental Rheumatology 2000; 18:

2 Bone loss in RA / B. Cortet et al. Bernard Cortet, MD, Hospital Practitioner; Marie-Hélène Guyot, MD, Head Clinical Assistant; Elisabeth Solau, MD, Head Clinical Assistant; Pascal Pigny, MD, PhD, University Conference Master; Florence Dumoulin, Registrar; René-Marc Flipo, MD, Professor of Rheumatology; Xavier Marchandise, MD, PhD, Professor of Nuclear Medicine; Bernard Delcambre, MD, Professor of Rheumatology. Please address correspondence and reprint requests to: Bernard Cortet, Department of Rheumatology, CHRU Lille, Hôpital R. Salengro, 2 Avenue Oscar Lambret, Lille Cedex, France. address: bcortet@chru-lille.fr Received on February 9, 2000; accepted in revised form on August 21, Copyright CLINICAL AND EXPERIMENTAL RHEUMATOLOGY Introduction Although generalized osteoporosis is a well-known extra-articular manifestation of rheumatoid arthritis (1-4) leading to increased fracture risk, few studies have assessed the occurrence of bone loss (longitudinal studies) in rheumatoid arthritis (5-10). Moreover, data available on this issue are conflicting, perhaps due to variations in bone mass measurement techniques, study population characteristics, and follow up duration. Thus, some authors found that patients with rheumatoid arthritis (RA) lost bone (7, 9-11), while others did not find any bone loss in RA or only in patients for whom the disease duration was shorter than 6 months (5, 6, 8, 12). Furthermore, although corticosteroids are well known to cause bone loss, there are some discrepancies regarding this issue in RA patients (8, 9, 10, 13). In fact, in RA corticosteroids are usually used at low doses which should theoretically be safe for bone tissue. Is is also important to note that corticosteroids improve disease activity which is associated with accelerated bone loss in RA (7, 11). The mechanisms leading to bone loss in RA are not yet well documented, but are hypothesized to involve an increase of bone resorption possibly due to an increase in pro-inflammatory cytokines which play a role in the pathogenesis of the disease, a decrease in bone formation or both (13-22). The most recent data suggest however that osteoclastic activation is the principal mechanism leading to osteoporosis in RA (4, 11). In three recent cross-sectional studies (3, 4, 23) we showed that patients with RA had decreased values of bone mineral density (BMD) at the femoral neck and elevated levels of bone turnover markers, namely procollagen type I C-terminal propeptide (PICP), procollagen type I N-terminal propeptide (PINP) which reflects bone formation, and procollagen type I C-terminal telopeptide (ICTP) which reflects bone resorption. The main purpose of the present study was to determine whether or not bone loss occurs in RA over an 18-month period. The second objective was to assess the factors influencing bone loss in RA and, particularly, the role of bone turnover markers in predicting bone loss. Patients and methods Patients The patients who participated in this longitudinal study fulfilled the American College of Rheumatology criteria for RA (24). They have been partly described in three previous cross-sectional studies (3, 4, 23). Eighty-five unselected, consecutive Caucasian patients were enrolled in this study. After the exclusion of patients for whom bone markers data were not collected and patients who were receiving drugs which could affect bone mass or bone metabolism (i.e., calcium, vitamin D, bisphosphonates, hormone replacement therapy or fluoride salts), 51 patients were eligible for the study. Several patients were receiving corticosteroids and/or methotrexate and/or cyclosporin. None of the patients received intramuscular injections of corticosteroid during the follow up, while 5 patients received intra-articular injections during the follow up. For the latter patients these injections were taken into account when calculating the cumulative dose of corticosteroids. The study was conducted from January 1995 to December It was approved by our local ethics committee and all patients gave their informed consent to participate. Bone densitometry BMD was measured by dual-energy X- ray absorptiometry (DEXA, Sophos L- XRA, Buc/Yvette, France) at the lumbar spine (L2-L4) and the non-dominant femoral neck at the 0 and 18-month clinical examinations. The in vivo reproducibility of this measurement, expressed as the coefficient of variation, was 1% for the lumbar spine and 1.6% for the femoral neck. The individual values for the BMD were also expressed as a fraction of the SD of the mean of the normal value for the patients s sex and decade of age (Z-score), but also as a fraction of the SD of the normal values for young adults of the same sex (T-score). Clinical data collection The following parameters reflecting activity or severity were determined twice (at baseline and at the 18-month evaluation): duration of morning stiffness, num- 684

3 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo Bone loss in RA / EDITORIAL B. Cortet et al. ber of painful and swollen joints, Ritchie articular index (25), Lee s Index (26), and the health assessment questionnaire (HAQ) score (27). In steroid-treated patients, the duration of treatment, the mean daily dose and the cumulative dose were also recorded (prednisone-equivalent). Finally, the number of patients who were taking second-line drugs and, in particular, methotrexate was also recorded. The erythrocyte sedimentation rate (ESR, by the Westergren method) and C-reactive protein level (CRP, by laser immunonephelometry) were measured after an overnight fast at baseline and at the 18- month evaluation. Multiple regression models (stepwise multiple regression analysis) were constructed to identify independent factors influencing bone loss. Several models were constructed based on the results of simple regression analysis and the best model for explaining bone loss either at the lumbar spine or femoral neck was chosen. Two subgroups were also identified according to BMD changes: (i) fast losers whose bone loss was above the median [both for the lumbar spine (1.9%) and the femoral neck (3.1%)]; and (ii) slow losers whose bone loss was below the median (both for the lumbar spine and femoral neck). P values < 0.05 were considered statistically significant. Results RA findings The main characteristics of the 51 patients are summarized in Table I. There were 45 women (91%) and 6 men. Thirty-five of the women were menopausal (77%). The mean age of the patients was 56 ± 10 years and the mean disease duration was 12 ± 10 years. Forty-one patients (80%) had positive tests for rheumatoid factors and 28 (55%) were receiving corticosteroid therapy with a mean treatment duration of 6 ± 5 years and a mean daily dose of 10 ± 4 mg/day. Patients who were receiving corticosteroids did not differ at baseline from patients who were on corticosteroids (Table II). Bone turnover markers The following serum assays were undertaken at baseline only: PICP, PINP and osteocalcin (OC) which reflect bone formation, and ICTP which reflects bone resorption. Blood samples were collected between 7:30 am and 9:00 am after an overnight fast. Serum samples were stored at -80 until assayed. Radioimmunologic assays were used to measure serum PICP, PINP, ICTP (Orion Diagnostica, Espoo, Finland) and osteocalcin (Cis-Bio International, Gif/Yvette, France). The within run and run-to-run coefficients of variation were 2.8% and 5.1%, respectively, for the PICP assay. The corresponding figures were 8.75% and 5.1% for the PINP assay, 4.8% and 5.7% for the ICTP assay and 3.7% and 6.6% for the osteocalcin assay. Statistical analysis All analyses were performed using the Statview program (version 5, SAS Institute, Cary, NC, USA). Group data are expressed as the mean ± standard deviation (SD) or mean ± 95% confidence interval (CI). The incremental percentage change in BMD was calculated for each patient at both the lumbar spine and femoral neck. Statistical comparisons were made using Student s-t-test for paired data or Wilcoxon s test as appropriate. The comparisons of BMD changes according to sex or menopausal status were performed using analysis of variance or the Kruskal-Wallis test. Simple linear regression was used to determine the coefficients of correlation. Table I. Main features of the 51 rheumatoid arthritis patients at baseline. Age (years) 56 ±10 Weight (kg) 69 ± 15 Height (cm) 163 ± 9 Number of women 45 (88%) Number of postmenopausal women 35 (80%) Rheumatoid arthritis duration (years) 12 ± 10 Positive rheumatoid factor 41 (80%) Steroid therapy 28 (55%) Duration of steroid therapy (years) 6 ± 5 Cumulative steroid dose (grams) 20 ± 20 Mean daily dose of steroids (mg/day) 10 ± 4 Morning stiffness duration (mn) 82 ± 106 Number of painful joints 11 ± 10 Number of swollen joints 7 ± 5 Lee s index 13 ± 10 Ritchie articular index 14 ± 9 Health assessment questionnaire score 1.6 ± 0.7 Table II. Comparison at baseline of patients according to their corticosteroid status. Cortico+: patients with corticosteroids, Cortico-: patients without corticosteroids. There was no significant difference between the 2 subgroups (cortico+ versus cortico-). Cortico+ Cortico- Age (years) 57 ± 9 55 ± 13 Weight (kg) 66 ± ± 16 Height (cm) 165 ± ± 8 Number of women 24/28 (86%) 21/23 (91%) Number of postmenopausal women 21/35 (60%) 7/10 (70%) RA duration (years) 13 ± ± 7 Morning stiffness duration (mn) 121 ± ± 63 Number of painful joints 13 ± 12 10± 10 Number of swollen joints 7 ± 5 5 ± 5 Lee s index 15 ± ± 10 Ritchie articular index 14 ± 9 14 ± 9 Health assessment questionnaire score 1.7 ± ± 0.7 Erythrocyte sedimentation rate (mm/h) 50 ± ± 32 C-reactive protein level (mg/l) 42 ± ±

4 Bone loss in RA / B. Cortet et al. Table III. Bone mineral density (BMD) with Z- and T-scores and markers of bone turnover in 51 patients with rheumatoid arthritis (baseline and 18 month evaluations). 0 months 18 months LS BMD (g/cm 2 ) 0.94 ± ± 0.13* Lumbar spine Z-score ± ± 0.64* Lumbar spine T-score ± ± 0.95 FN BMD (g/cm 2 ) 0.73 ± ± 0.11** Femoral neck Z-score ± ± 0.70* Femoral neck T-score ± ± 0.93 PICP (ng/ml) ± NP PINP (ng/ml) ± NP ICTP (ng/ml) 5.90 ± 2.61 NP Osteocalcin (ng/ml) 5.8 ± 2.62 NP LS: lumbar spine; FN: femoral neck; PICP: procollagen type I C-terminal propeptide; PINP: procollagen type I N-terminal propeptide; ICTP: procollagen type I C-terminal telopeptide. *P < 0.01, **P < (versus baseline). NP: not performed. Fig. 1. Bone mineral density (BMD) changes at the lumbar spine and femoral neck in 51 patients with rheumatoid arthritis (error bars = SEM). Fig. 2. Bone mineral density (BMD) changes at the femoral neck in 51 patients with rheumatoid arthtritis according to sex and menopausal status (error bars = SEM) Forty-eight patients (94%) were receiving a slow-acting anti-rheumatic drug and among them 20 (42%) were taking methotrexate. The others were receiving intra-muscular gold (n = 10, 20%), sulphasalazine (n = 5, 10%), hydroxychloroquine (n = 4, 8%), cyclosporin A (n = 4, 8%), tiopronin (sulphydryl compound, n = 3, 6%) and finally azathioprin (n = 2, 6%). Bone mineral density Table III shows the BMD values, Z- scores and T-scores at baseline and at the 18-month investigation. BMD changes expressed as a percentage are also represented in Figure 1. Significant bone loss occurred both at the lumbar spine: 2.1%, [95% CI: 0.8% - 3.4%, P < 0.005] and femoral neck: 3.1% [95% CI: 1.1% - 5.1%, P < 0.005]. Bone loss was not statistically significantly different in steroid-treated patients as compared with nonsteroid-treated patients both at the lumbar spine and femoral neck: 2.1% ± 4 versus 2.1% ± 5 (P = 0.9) and 2.7% ± 6 versus 4.1% ± 8 (P = 0.5) respectively. Bone loss at the lumbar spine was: 1.2% ± 3.4 for males, 3.4% ± 4.3 for premenopausal women and 1.8% ± 4.9 for postmenopausal women. Bone loss at the lumbar spine was statistically significant for premenopausal women only [95% CI: 0.7% %, P < 0.05]. Bone loss at the lumbar spine (expressed as percentage) was not statistically significantly different, however, in the 3 subgroups previously defined (men, and premenopausal and postmenopausal women). Bone loss at the femoral neck was increased in postmenopausal women as compared with premenopausal women: 5.3% ± 6 versus 0.7% ± 6.6 (P < 0.05). In men a non-significant bone gain was observed (+2.5% ± 7). BMD changes at the femoral neck were significantly different between postmenopausal women and men (P < 0.01). Bone loss at the femoral neck was statistically significant for postmenopausal only [95% CI: 2.9% - 7.6%, P < ]. Bone loss was not different between those patients taking and those not taking methotrexate, both at the lumbar spine and at the femoral neck (P =

5 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo and P = 0.9 respectively). Fast losers at the lumbar spine were characterized at baseline by elevated levels of both osteocalcin and ESR as compared with slow losers: 62 mm ± 34 versus 42 mm ± 23 (P < 0.05) and 6.9 ng/ml ± 2.3 versus 4.7 ng/ml ± 2.6 (P < 0.01) respectively. Determinants of bone loss Simple regression analysis. The changes in BMD both at the lumbar spine and femoral neck were not correlated with either the clinical and biological data at the first visit (disease duration, morning stiffness duration, number of painful and swollen joints, Ritchie articular index, Lee s Index, HAQ score, ESR and CRP levels) or with the changes in clinical and biological data beween the 0 and the 18- month evaluations (Table IV). Moderate correlations were found between bone loss and the levels of some bone turnover markers. Bone loss at the lumbar spine was significantly correlated with both osteocalcin (r = 0.51, P < 0.005) and ICTP levels (r = 0.32, P < 0.05). Bone loss at the femoral neck was correlated with osteocalcin levels only (r = 0.34, P < 0.05). No other correlations were found between the level of bone turnover markers at baseline and the changes in BMD both at the lumbar spine and femoral neck (Table IV). Stepwise multiple regression analysis. Several models were constructed to explain both the lumbar spine and femoral neck bone loss by stepwise multiple regression analysis. At the lumbar spine the only significant predictor was the osteocalcin level which, however, explained only 22% of the variation in BMD changes (adjusted R 2 ). None of the models constructed to explain BMD changes at the femoral neck achieved an adjusted R 2 greater than 0.15, showing the weak predictive effect for this population. Discussion Although generalized osteoporosis is a well-known extra-articular manifestation of RA (1-4), few studies have focused on the occurrence of bone loss in RA (5-12). Moreover the most recent studies addressing this issue have yielded conflicting results (6, 7, 10-12). Thus Gough et al. in 2 recent studies (7, 11) found that patients with early RA had faster bone loss both at the lumbar spine and at the hip than that observed in a control group. Table IV. Correlation between the first visit clinical and biological data (including markers of bone turnover) and bone loss both at the lumbar spine and femoral neck. Lumbar spine bone loss Femoral neck bone loss Age Weight Height RA duration Lee s index Ritchie articular index HAQ score Morning stiffness duration ESR CRP level Steroid duration Cumulative dose of steroids PICP level PINP level Osteocalcin level 0.51** 0.34* ICTP level 0.32* 0.19 HAQ: health assessment questionnaire; ESR: erythrocyte sedimentation rate; CRP: c-reactive protein, PICP: procollagen type I C-terminal propeptide; PINP: procollagen type I N-terminal propeptide; ICTP: procollagen type I C-terminal telopeptide.**p < 0.005, *P < Bone loss in RA / EDITORIAL B. Cortet et al. Moreover, bone loss was slightly greater in the first year as compared with the second year of follow up. Conversely Shenstone et al. (6), who studied 67 patients with RA of less than 5 years duration and 72 controls over a 12-month period, found that bone loss was low (about 1%) and comparable in patients and controls. However, they also found that bone loss at the femoral neck was faster in those patients who had experienced onset of their disease less than 6 months earlier, as compared with the controls and the patients who had more long-standing disease. Finally Aman et al. (12), who studied 52 patients with early RA (i.e., of less than 5 years duration), over a 24 month-period did not find any bone loss. The reasons for these discrepancies are unclear since these 4 recent studies used the same method, DEXA, to measure BMD and the study populations were very comparable (early RA). It is, however, possible that the patients who were early referrals by primary care physicians to the district rheumatism hospital and who therefore represented a community-based population had less severe disease compared with patients who were attending a secondary or tertiary care hospital. Our study suggests that bone loss occurs even in patients with late RA. Although bone loss was statistically significant both at the lumbar spine and at the femoral neck, it was faster at the femoral neck as compared with the lumbar spine: 3.1% versus 2.1%. Bone loss was also significant both at the lumbar spine and at the femoral neck after adjustment for age, i.e. when taking into account the mean Z-score changes. These findings are in agreement with the recent study by Gough et al. (11), who found that the mean percentage change in BMD in patients with RA was -1.8% and -2.8% over a 12- month period at the lumbar spine and the femoral neck, respectively. They also found that bone loss occurred during the second year of follow up: 1.4% (lumbar spine) and 2.4% (femoral neck) respectively. Finally they showed that patients with RA lost bone significantly faster than controls both at the 12-month and 24-month investigations. Nevertheless, we cannot strictly compare our findings with those of Gough et al. since ours was 687

6 Bone loss in RA / B. Cortet et al. an open study (i.e., there was no control group). Our study suggests that bone loss was the same in steroid and non-steroid treated patients: 2.1% ± 4 versus 2.1% ± 5 at the lumbar spine and 2.7% ± 6 versus 4.1% ± 8 at the femoral neck, respectively. Although steroids are known to cause bone loss, they also improve disease activity which is associated with accelerated bone loss in patients with RA (7, 11). Moreover, the low doses usually used in RA ( 10 mg/day of prednisoneequivalent) might be safe for bone tissue. Our findings are in agreement with the metaanalysis of Verhoeven and Boers (13) who found that patients with RA and treated by steroids (mean daily of prednisone: 7 mg) did not lose bone. In fact, the authors found that the mean annual change in BMD was 0% [95% CI: -0.6% to +0.7%] at the lumbar spine. They also found significant bone loss at the lumbar spine (mean annual change) in non- RA patients taking a mean dose of 20 mg prednisone/day: -4.7% [95% CI: -5.2% to -4.3%], suggesting that the steroid dose used may have a major effect on bone loss. Moreover, as indicated in Table II, the 2 subgroups of patients defined according to corticosteroid status did not differ at baseline and this finding emphasizes our conclusion that bone loss is not higher in patients who are treated with corticosteroids as compared with those who are not. On the other hand, we recently showed (28) that patients with RA for whom steroids were given for less than 3 months lost bone significantly at the lumbar spine over a 12-month period despite the fact that they also received 500 mg of supplemental calcium per day. In fact, the mean bone loss per year (± SEM) at the lumbar spine was 2.70% ± Nevertheless, in that study the parameters reflecting the severity and activity of RA were not measured and we cannot exclude the possibility that they had a major influence on bone loss (perhaps greater than the influence of steroid therapy). Although methotrexate osteopathy is a well-known entity (29, 30), the present study suggests that methotrexate does not seem to influence bone loss in patients with RA. This finding is in agreement with one recent study on this issue (31). Our study suggests that bone loss occurs at the femoral neck in postmenopausal women: 5.3% [95% CI: 2.9% - 7.6%, P < ]. Our findings are in agreement with the study of Eggelmeijer et al. (32) who found after 36 months of follow up that postmenopausal women with RA lost bone at the femoral neck. In fact, they found that the mean change in the femoral neck BMD for postmenopausal women was -8.1%. Gough et al. (7) also found that bone loss was greater for postmenopausal women as compared with premenopausal women at the femoral neck over a 1-year period (2.6% versus 1.6%). We found that bone loss was statistically significant in premenopausal women solely at the lumbar spine: 3.4% [95% CI: 0.7% %, P < 0.05]. We do not have any satisfying explanation for this finding. Gough et al. (19) also found in a small sample of premenopausal women with RA that bone loss over 1 year was increased at the lumbar spine as compared with postmenopausal women (1.8% versus 0.1%). Finally, it is also important to note that the number of premenopausal women in the present study was low (n = 10, against 35 postmenopausal women) and therefore the above finding must be interpreted cautiously. None of the first visit clinical and biological data (disease duration, morning stiffness duration, number of painful and swollen joints, Ritchie articular index, Lee s Index, HAQ score, ESR and CRP levels) was correlated with the changes in BMD either at the lumbar spine or the femoral neck. In contrast Gough et al. (7, 11) found that the initial HAQ score was moderately correlated with BMD changes over one year both at the lumbar spine (P < 0.05) and femoral neck (P < 0.01). Furthermore, in a previous cross-sectional study we showed a negative correlation between the HAQ score and BMD (r = for both the lumbar spine and femoral neck). Data (not shown) were very similar for the present study: r = (lumbar spine) and r = (femoral neck), respectively. Hence disability seems to be a relevant factor for explaining osteoporosis in patients with RA. However, in the present study the duration of the follow up was probably too short to detect an influence of the HAQ score on bone loss. Although the ESR at baseline was not correlated with the change in BMD, we found that ESR might be a determinant of bone loss. In fact, fast bone losers had increased ESR levels at baseline compared with slow losers. This finding is in agreement with the two studies by Gough et al. (7, 11) who found that patients whose mean CRP over 1 year was > 20 mg/l showed significantly greater bone loss compared to patients with inactive disease. In the same manner Eggelmeijer et al. (32) found that losses in the femoral neck were more marked in patients with active disease (ESR at baseline 20 mm/hr). Our study supports the hypothesis that increased bone turnover is a determinant of bone loss in patients with RA. In fact, bone loss at the lumbar spine was significantly correlated with both osteocalcin (r = 0.51, P < 0.01) and ICTP levels (r = 0.32, P < 0.05). Bone loss at the femoral neck was correlated with osteocalcin levels only (r = 0.34, P < 0.05). There was no significant correlation, however, with the other markers of bone formation (i.e., PINP and PICP). Although osteocalcin is a well-known marker of bone formation, it has been clearly demonstrated that osteocalcin levels are increased when bone turnover (and not only bone formation) is increased. For instance, Garnero et al. (33) in the OFELY study showed that osteocalcin levels increased by approximately 50% in postmenopausal women as compared with premenopausal women. It is clear, however, that theoretically the correlation between bone loss and the level of markers of bone turnover should be stronger for markers of bone resorption than for markers of bone formation. This apparent discrepancy could be due to the lack of sensitivity of ICTP as compared with other markers of bone resorption and particularly the urinary type I collagen C-telopepetide breakdown products (CTX). It is also important to note that data are limited and conflicting regarding the relationship between the level of bone turnover as assessed by the measurement of markers of bone turnover and bone loss 688

7 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo Bone loss in RA / EDITORIAL B. Cortet et al. in RA patients (11, 12, 19). More generally, the usefulness of markers of bone turnover for predicting bone loss in postmenopausal women is still debated and a recent paper by Marcus et al. (34) showed that the levels of several markers of bone turnover did not correlate with bone loss in postmenopausal women. Furthermore, we found that fast losers of bone had, at baseline, higher levels of oseocalcin as compared with slow losers. In a previous study (4) we showed that both osteocalcin and ICTP levels were correlated with BMD, particularly at the femoral neck (r = and r = -0.29, respectively). In the present study (data not shown), ICTP levels were also correlated with baseline BMD both at the lumbar spine (r = -0.5) and femoral neck (r = -0.39). Conversely Gough et al. (11) found no correlation at baseline between pyridinoline or deoxypyridinoline excretion and BMD. However, they studied patients with early RA (less than 2 years duration) whereas we studied patients with late RA (mean duration 12 ± 10 years). They also showed that both pyridinoline and deoxypyridinoline excretion correlated closely with BMD changes. Unfortunately, pyridinoline and deoxypyridinoline excretion were not assessed in the present study. However it is important to note that these 2 bone turnover markers have a poor precision (intra-assay variation 15% for deoxypyridinoline) compared with the markers measured in the present study (intra-assay variation 5%). Furthermore, it is well known that the timing of urine sampling can significantly alter excretion rates. Finally the gold-standard method, i.e. high performance liquid chromatography for measuring pyridinoline and deoxypyridinoline, is neither simple nor readily available. In conclusion, this study suggests that bone loss occurs in patients with RA, particularly at the femoral neck. Bone loss is associated with increased bone turnover as assessed by high osteocalcin and ICTP levels, but also with persistent inflammation as indirectly measured by the ESR. References 1. CORTET B, FLIPO RM, DUQUESNOY B, DELCAMBRE B: Bone tissue in rheumatoid arthritis (1). Bone mineral density and fracture risk. Rev Rhum [Engl. Ed.] 1995; 62: CORTET B, FLIPO RM, DUQUESNOY B, DELCAMBRE B: Bone tissue in rheumatoid arthritis (2). Pathophysiologic data, pathologic findings and therapeutic implications. Rev Rhum [Engl. Ed.] 1995; 62: CORTET B, FLIPO RM, BLANCKAERT F, DUQUESNOY B, MARCHANDISE X, DEL- CAMBRE B: Evaluation of bone mineral density in patients with rheumatoid arthritis. Influence of disease activity and glucocorticoid therapy. Rev Rhum [Engl Ed] 1997; 64: CORTET B, FLIPO RM, PIGNY P, DUQUESNOY B, BOERSMA A, MARCHANDISE X, DEL- CAMBRE B: Is bone turnover a determinant of bone mass in rheumatoid arthritis? J Rheumatol 1998; 25: SAMBROOK PN, ANSELL BM, FOSTER S, GUMPEL JM, HESP R, REEVE J: Bone turnover in early rheumatoid arthritis. 2. Longitudinal bone density studies. Ann Rheum Dis 1985; 44: SHENSTONE BD, MAHMOUD A, WOODWARD R, ELVINS D, PALMER R, RING EF, BHALLA AK: Longitudinal bone mineral density changes in early rheumatoid arthritis. Br J Rheumatol 1994; 33: GOUGH AKS, LILLEY J, EYRE S, HOLDER RL, EMERY P. Generalised bone loss in patients with early rheumatoid arthritis. Lancet 1994; 344: SAMBROOK PN, COHEN ML, EISMAN JA, POCOCK NA, CHAMPION GD, YEATES MG: Effects of low dose corticosteroids on bone mass in rheumatoid arthritis: a longitudinal study. Ann Rheum Dis 1989; 48: REID DM, KENNEDY NS, SMITH MA, TOTHILL P, NUKI G: Bone loss in rheumatoid arthritis and primary generalized osteoarthrosis: Effects of corticosteroids, suppressive antirheumatic drugs and calcium supplements. Br J Rheumatol 1986; 25: EVANS WD, EVANS C: Long-term follow-up of spinal trabecular bone mineral density in women with non-steroid treated rheumatoid arthritis. Ann Rheum Dis 1994; 53: GOUGH A, SAMBROOK P, DEVLIN J et al.: Osteoclastic activation is the principal mechanism leading to secondary osteoporosis in rheumatoid arthritis. J Rheumatol 1998; 25: AMAN S, HAKALA M, SILVENNOINEN, MAN- ELIUS J, RISTELI L, RISTELI J: Low incidence of osteoporosis in a two year follow-up of early community based patients with rheumatoid arthritis. Scand J Rheumatol 1998; 27: VERHOEVEN AC, BOERS M: Limited bone loss due to corticosteroids; A systematic review of prospective studies in rheumatoid arthritis and other diseases. J Rheumatol 1997; 24: KRÖGER H, RISTELI J, RISTELI L, PENTTILÄ I, ALHAVA E: Serum osteocalcin and carboxyterminal propeptide of type I procollagen in rheumatoid arthritis. Ann Rheum Dis 1993; 52: HALL GM, SPECTOR TD, DELMAS PD: Markers of bone metabolism in postmenopausal women with rheumatoid arthritis. Effects of corticosteroids and hormone replacement therapy. Arthritis Rheum 1995; 38: FRANCK H, ITTEL TH, TASCH O, HERBORN G, RAU R: Osteocalcin in patients with rheumatoid arthritis. Effects of anatomical stages, inflammatory activity and therapy. Rheumatol Int 1992; 12: EGGELMEIJER F, PAPAPOULOS SE, WESTEDT ML, VAN PAASSEN HC, DIJKMANS BAC, BREEDVELD FC: Bone metabolism in rheumatoid arthritis; relation to disease activity. Br J Rheumatol 1993; 32: 387, RICO H, REVILLA M, ALVAREZ DE BUERGO M, VILLA LF: Serum osteocalcin and calcitropic hormones in a homogeneous group of patients with rheumatoid arthritis: its implication in the osteopenia of the disease. Clin Exp Rheumatol 1993; 11: GOUGH AKS, PEEL NFA, EASTELL R, HOLDER RL, LILLEY J, EMERY P: Excretion of pyridinium crosslinks correlates with disease activity and appendicular bone loss in early rheumatoid arthritis. Ann Rheum Dis 1994; 53: HAKALA M, RISTELI L, MANELIUS J, NIE- MINEN P, RISTELI J: Increased type I collagen degradation correlates with disease severity in rheumatoid arthritis. Ann Rheum Dis 1993; 52: HAKALA M, AMAN S, LUUKKAINEN R, RISTELI L, KAUPPI M, NIEMINEN P, RISTELI J: Application of markers of collagen metabolism in serum and synovial fluid for assessment of disease process in patients with rheumatoid arthritis. Ann Rheum Dis : COMPSTON JE, VEDI S, CROUCHER PI, GARRAHAN NJ, O SULLIVAN MM: Bone turnover in non-steroid treated rheumatoid arthritis. Ann Rheum Dis 1994; 53: CORTET B, FLIPO RM, PIGNY P, DUQUESNOY B, RACADOT A, BOERSMA A, DELCAMBRE B: How useful are bone turnover markers in rheumatoid arthritis? Influence of disease activity and corticosteroid therapy. Rev Rhum [Engl Ed] 1997; 64: ARNETT FC, EDWORTHY SM, BLOCH DA, et al.: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: RITCHIE DM, BOYLE JA, MCINNES JM, JAS- ANI MK, DALAKOS TG, GRIEVESON P, BUCHANAN VW: Clinical studies with an articular index for the assessment of joint tenderness in patients with rheumatoid arthritis. Q J Med 1968; 37: LEE P, JASANI MK, WILLIAM HJ and the Cooperative Systematic Studies of Rheumatic Diseases Group: Analysis of improvement in individual rheumatoid arthritis patients treated with disease-modifying antirheumatic drugs, based on the findings in patients treated with placebo. Scand J Rheumatol 1973; 2: GUILLEMIN F, BRIANÇON S, POUREL J: Mesure de la capacité fonctionnelle dans la polyarthrite rhumatoïde: Adaptation française du Health Assesssment Questionnaire. Rev Rhum Mal Ostéoartic 1991; 58: CORTET B, HACHULLA E, BARTON I, BON- VOISIN B, ROUX C: Evaluation of the efficacy of etidronate therapy in preventing glucocorticoid-induced bone loss in patients with in- 689

8 Bone loss in RA / B. Cortet et al. flammatory rheumatic diseases. A randomized study. Rev Rhum [Engl. Ed.] 1999; 66: PRESTON SJ, DIAMOND T, SCOTT A, LAURENT MR: Methotrexate osteopathy in rheumatic disease. Ann Rheum Dis 1993; 52: MAENAUT K, WESTHOVENS R, DEQUEKER J: Methotrexate osteopathy, does it it exist? J Rheumatol 1996; 23: BUCKLEY LM, LEIB ES, CARTULARO KS, VACEK PM, COOPER SM: Effects of low dose methotrexate on the bone mineral density of patients with rheumatoid arthritis. J Rheumatol 1997; 24: EGGELMEIJER F, PAPAPOULOS SE, VAN PAASSEN HC et al.: Increased bone mass with pamidronate treatment in rheumatoid arthritis. Results of a three-year randomized, doubleblind trial. Arthritis Rheum 1996; 39: GARNERO P, SORNAY-RENDU E, CHAPUY MC, DELMAS PD: Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 1996; 11: MARCUS R, HOLLOWAY L, WELLS B, GREENDALE G, JAMES MK, WASILAUSKAS C, KELAGHAN J: The relationship of biochemical markers of bone turnover to bone density changes in postmenopausal women: Results from the postmenopausal estrogen/progestin intervention (PEPI) trial. J Bone Miner Res 1999; 14: