Article Text
Abstract
Objectives: We examined the effects of four different treatment strategies on bone mineral density (BMD) in patients with recently diagnosed, active rheumatoid arthritis (RA) and the influence of disease-related and demographic factors on BMD loss after 1 year of follow-up in the BeSt trial.
Methods: BMD measurements of the lumbar spine and total hip were performed in 342 patients with recent onset RA at baseline and after 1 year. Multivariable regression analyses were performed to determine independent associations between disease and demographic parameters and BMD loss after 1 year.
Results: Median BMD loss after 1 year was 0.8% and 1.0% of baseline in the spine and the hip, respectively. No significant differences between the treatment groups, including corticosteroids and the anti-tumour necrosis factor-α infliximab, were observed with regard to BMD loss after 1 year of treatment. Joint damage at baseline and joint damage progression according to the Sharp–van der Heijde score were independently associated with more BMD loss after 1 year. The use of bisphosphonates independently protected against BMD loss.
Conclusions: After 1 year of follow-up in the BeSt study, we did not find differences in BMD loss between the four treatment strategies, including high doses of corticosteroids and anti-tumour necrosis factor-α. Joint damage and joint damage progression are associated with high BMD loss, which emphasises that BMD loss and erosive RA have common pathways in their pathogenesis.
Statistics from Altmetric.com
Accelerated generalised bone loss is regarded as a complication in rheumatoid arthritis (RA)1 2 and an increased risk of hip3–5 and vertebral fractures5–7 has been demonstrated in patients with RA.
Previous studies showed that treatment of RA with disease-modifying antirheumatic drugs (DMARDs) generally do not have any effect on bone mineral density (BMD).8–11 On the other hand, it is well known that treatment with corticosteroids may negatively influence the rate of BMD loss in patients with RA; however, this effect might be neutralised by the suppression of inflammatory activity, an important risk factor for high BMD loss in RA.12–17 In contrast, treatment with tumour necrosis factor-α antagonist (anti-TNF-α) might protect against BMD loss.18–20
It is thought that the pathogenesis of generalised osteoporosis and local bone erosions share common pathways.21 This hypothesis is strengthened by the finding that osteoclasts, stimulated mainly by the receptor activator of nuclear factor κB ligand (RANKL), play a central part in both pathophysiological mechanisms.22–25 In accordance with this, high BMD loss in patients with RA was associated with joint damage progression, disease activity, functional disability and immobility in previous longitudinal studies, even in early RA.26–29 While the efficacy of calcium and vitamin D supplements is not very clear, the use of bisphosphonates is protective against (especially corticosteroid induced) BMD loss.30–34
The extent and the risk factors of BMD loss have not been studied in patients with recent onset RA while treated aggressively with DMARDs, corticosteroids or anti-TNF-α. In this longitudinal study with a large cohort of patients with recently diagnosed, active RA, we evaluated the changes in BMD in the lumbar spine and the total hip after 1 year of treatment. We also studied the influence of disease-related and demographic factors on BMD loss in recent onset RA that might indicate a common pathway between rheumatoid inflammation and generalised osteoporosis.
PATIENTS AND METHODS
Patients
All measures were performed in the setting of the BeSt study.35 Five hundred and eight patients, who met the definition of RA as defined by the American College of Rheumatology 1987 revised criteria, were included in the trial. Other inclusion criteria were symptom duration <2 years and active disease with ⩾6 of 66 swollen joints and ⩾6 of 68 tender joints. Patients with previous treatment with DMARDs and increased serum liver enzymes and creatinine (>150 mmol/l) were excluded. All inclusion and exclusion criteria were reported previously.35
The BeSt study was conducted by rheumatologists participating in the Foundation for Applied Rheumatology Research (FARR) in 18 peripheral and two university hospitals in the western part of the Netherlands. The medical ethics committee at each participating centre approved the study protocol and all patients gave written informed consent before participation in the study.
Study design
The BeSt study is a multicentre, randomised clinical trial designed to compare the effectiveness of four different treatment strategies in patients with recently diagnosed, active RA. In case of an insufficient response to therapy, based on the disease activity score (DAS) in 44 joints >2.4, the medication was immediately adjusted by proceeding to the next step according to the pharmacoprotocol.
The patients assigned to sequential monotherapy (group 1) started with 15 mg/week methotrexate (MTX), which could be increased to 25–30 mg/week and subsequent steps were sulfasalazine (SSZ) and leflunomide monotherapies. The patients assigned to step-up combination therapy (group 2) also started with 15 mg/week MTX, which could be increased to 25–30 mg/week and in the next steps SSZ and hydrochloroquine could be added. The patients assigned to initial combination therapy with prednisone (group 3) started with the combination 7.5 mg/week MTX, 2000 mg/day SSZ and 60 mg/day prednisone quickly tapered to 7.5 mg/day in 7 weeks and MTX could be increased to 25–30 mg. The patients assigned to initial combination therapy with infliximab started with infliximab (group 4) 3 mg/kg at weeks t = 0, 2 and 6, and every 8 weeks thereafter and MTX 25–30 mg/week, the dose of infliximab could be increased step by step to 6 mg/kg, 7.5 mg/kg and finally 10 mg/kg. If the clinical response was consistently adequate (DAS ⩽2.4 for at least 6 months), medication was tapered until one drug remained at maintenance dose, which was MTX 10 mg/week, SSZ 2000 mg/day or leflunomide 10 mg every other day. The complete treatment protocol has previously been published.35
In case of calcium intake <1000 mg/day and serum vitamin D level below the local reference value, suppletion of 500–1000 mg/day calcium and 400 IE/day vitamin D (colecalciferol) was advised at baseline. Alendronate, 10 mg/day or 70 mg/week, or risedronate, 5 mg/day or 35 mg/week, was advised if the baseline BMD measurement showed a T score, number of standard deviations (SD) from the mean BMD of young, healthy persons, ⩽−2.5 SD in the spine and/or hip in non-corticosteroid users or a T score ⩽−1 SD; this is the American College of Rheumatology recommended threshold for corticosteroid-induced osteoporosis, in corticosteroid users.
Bone mineral density measurements
A total of 342 of the 508 patients received BMD measurements of the lumbar spine L2–4 anteroposterior view and the total hip at baseline and after 1 year in 14 of 20 centres where dual energy X-ray absorptiometry (DEXA) was available. The BMD measurements were carried out using a Hologic 4500 QDR (Hologic, Waltham, MA, USA) in eight centres and a Lunar DPX (Lunar, Madison, WI, USA) in six centres. T scores were determined according to references provided by the manufacturers (Lunar: UK or US references, Hologic: Hologic’s spine and NHANES femur references). Changes in BMD were expressed as the change after 1 year in absolute BMD values compared with baseline BMD. The rates of change after 1 year of follow-up calculated from serial measurements, assessed by the same machine, measurement procedure and references for each patient, on different machine types are comparable.36 Changes in BMD after 1 year of treatment were divided into tertiles with high and low BMD loss, with the median percentage BMD loss used as a cut-off point, and no BMD loss/BMD increase.
Demographic and clinical variables
Demographic and clinical data were obtained by specially trained research nurses who were blinded to the allocated treatment group. The DAS, based on the number of swollen joints, the Ritchie articular index for pain in tender joints, the visual analogue scale for patient’s global assessment of disease activity (0–100 mm) and the erythrocyte sedimentation rate, and the Dutch validated health assessment questionnaire (HAQ), measuring functional disability, were obtained 3-monthly. Radiographic joint damage at baseline and after 1 year was assessed according to the Sharp–van der Heijde score (SHS) scored independently by two blinded assessors. The intra-observer coefficients were 0.93 and 0.94 and the interobserver coefficient was 0.93. The mean of the scores of the two assessors was used for analysis. A patient was classified as having erosive disease at baseline if the mean erosion score was >0.5. Progression of joint damage after 1 year was defined as SHS progression greater than the smallest detectable change (SDC = 4.18 points).
Statistical analysis
Changes in BMD between the four treatment strategies were compared by regression analyses adjusted for use of bisphosphonates, vitamin D and calcium supplements, hormone replacement therapy (HRT) and intra-articular steroids. Potential contribution of the demographic and disease-related variables as independent risk factors of BMD loss was evaluated by multivariable regression analyses performed as forward (conditional) procedures and adjusted for age, gender, menopausal status, race, smoking and alcohol status. Additionally, symptom duration, joint damage, rheumatoid factor (RF) status, DAS and HAQ were adjusted for each other and disease-related variables after 1 year were corrected for their baseline values. Further, use of bisphosphonates, vitamin D, calcium, HRT and intra-articular steroids were adjusted for each other. The adjusted odds ratios obtained by regression analyses were corrected into relative risks with the formula of Zhang and Kai37 to interpret the magnitude of the associations more appropriately. All tests were two tailed and p⩽0.05 were considered statistically significant.
RESULTS
Patient characteristics
A total of 342 patients received BMD measurements in the lumbar spine and the total hip at baseline and after 1 year. In 16 patients no hip measurement was performed at baseline and/or after 1 year, two of these due to bilateral hip prosthesis, others due to logistic reasons. Two patients did not receive spine measurements at both time points.
Table 1 shows the baseline characteristics of the 342 patients. The group of patients in whom BMD was measured did not differ significantly from the patients in whom BMD was not measured (n = 166) with regard to the distribution in treatment strategies, demographic and disease variables (data not shown). The baseline demographic and disease variables were not significantly different between the four treatment groups (table 1).
In table 2, the treatment with anti-resorptive agents and intra-articular steroid injections during follow-up is listed. Patients in group 3 used significantly more bisphosphonates, calcium and vitamin D supplements than patients in the other groups (p<0.0001, p = 0.004, p = 0.01, respectively). More patients received at least once intra-articular steroid injection in groups 1 and 2 than in groups 3 and 4 (p = 0.01).
Patients who used bisphosphonates were more often present in group 3, postmenopausal women and older, had a lower BMI and used more often calcium and vitamin D suppletion (data not shown). There were no significant differences in disease-related factors between bisphosphonate users and non-users (data not shown).
Effect of treatment strategies on bone mineral density changes
After 1 year of treatment, the median (interquartile range) BMD change was −0.8% (−3.0 to 1.5) of baseline in the spine and −1.0% (−3.3 to 1.4) in the hip. Patients who received bisphosphonates and/or HRT had less BMD loss after 1 year in the spine (−0.9% (−3.1 to 1.1) vs 0.0% (−2.3 to 2.3), p = 0.012), but not in the hip (−1.0% (−3.5 to 1.2) vs −0.9% (−3.0 to 2.1), p = 0.4), than patients who did not receive these agents.
There were no statistically significant differences in the BMD changes between the four treatment groups (table 3, overall p = 0.9).
In group 3 the mean (SD) cumulative dose prednisone was 2.5 (0.5) g/patient during follow-up, equivalent to 9.3 mg/day. In group 4, the mean cumulative dose of infliximab was 34.5 (10.9) mg/kg per patient, equivalent to approximately 4.5 mg/kg per infusion. Subanalyses between current prednisone users and non-prednisone users and current infliximab users and non-infliximab users showed no significant differences in BMD changes after 1 year (data not shown).
Determinants of bone mineral density loss
We divided the changes in BMD after 1 year in tertiles and the third tertile, with no BMD loss, was used as reference. Univariable regression analyses showed that SHS at baseline and SHS progression >SDC after 1 year were associated with high BMD loss in the hip and the spine, respectively. Other disease-related factors, such as high DAS or HAQ or a positive RF, were not associated with more BMD loss (table 4, data of the hip are not shown). Of the demographic variables listed in table 1, higher age and smoking were associated with BMD loss in the hip and patients with BMD loss in the spine included more men than women and more postmenopausal than premenopausal women.
Osteoporosis treatment
Of the patients who were advised to use bisphosphonates, based on T score ⩽−2.5 SD in non-corticosteroid users or T score ⩽−1.0 SD in corticosteroid users, 43% were actually prescribed oral bisphosphonates (67% alendronate and 33% risendronate). Thirty-five per cent of the patients with low calcium intake (<1000 mg/day) received calcium supplement and 33% of the patients with low 25(OH)vitamin D levels received vitamin D supplement. Seventy per cent of the patients with calcium and vitamin D supplements did not use bisphosphonates. Patients using bisphosphonates, with or without calcium or vitamin D supplements, had less BMD loss in the spine. The use of calcium and/or vitamin D, without bisphosphonates, had no effect on bone loss after 1 year of treatment.
Multivariable analyses
Variables that showed significant associations with BMD loss in the univariable analyses where entered in a multiple regression analyses, adjusted for possible confounders. Total SHS at baseline and SHS >SDC were independently associated with BMD loss in the hip and spine, respectively (table 5). Bisphosphonates, adjusted for other anti-resorptive treatment, independently protected against BMD loss in the spine.
DISCUSSION
In this large longitudinal study we examined the changes in BMD in 342 patients with recent onset, active RA after 1 year of treatment in the BeSt study. The main finding of this study is that there are no differences between the treatment groups, including treatment strategies using a quickly tapered high-dose of corticosteroids and the anti-TNF-α infliximab, considering changes in BMD after 1 year of follow-up. The independent disease-related determinants of BMD loss in these patients were joint damage at baseline, joint damage progression and non-use of bisphosphonates.
As BMD change is dependent of many population-specific factors, such as genetic factors, race, environmental and dietary influences, age and gender, it is important to compare our data on BMD loss with data in the general Dutch population. The Rotterdam study, a population-based longitudinal cohort study of 4333 men and women, aged ⩾55 years, with BMD measured in the femoral neck at baseline and after 2 years, showed a BMD change of −0.4% in men and −0.6% in women.38 In our population, patients aged ⩾55 years showed a BMD change of −1.4% in men and −1.5% in women in the hip after 1 year. This suggests that the BMD loss in our patients with recent onset RA treated for 1 year in the BeSt trial is higher compared with the general Dutch population; however, comparisons should be done with caution due to demographic differences between the two populations.
In previous longitudinal studies, annual BMD changes ranged from −0.3% to −2.4% in the lumbar spine and from −0.1% to −4.3% in the hip.27–29 39–41 The magnitude of BMD loss demonstrated in the present study was generally less. An explanation for the relatively small BMD loss in our population is the strict monitoring of the disease activity with adequate adjustments in treatment during disease flares, DAS >2.4, to suppress the inflammation, which is a risk factor for bone loss, immediately. Further, the use of anti-resorptive agents was not allowed in some of the previous studies.
Numerous previous studies had shown glucocorticoid-induced high bone loss in patients with RA.16 39 In contrast we did not find more BMD loss in the initial combination group with high dose prednisone compared with the other treatment groups. In line with our results, Verhoeven et al13 did not find differences in BMD loss between patients treated with SSZ monotherapy and combination therapy with high-dose prednisone, almost similar to group 3 in our study, in patients with early (disease duration <2 years) active RA who were treated for a period of 56 weeks. A plausible explanation for this is that the beneficial effect of corticosteroids on BMD by quick, aggressive suppression of inflammation outweighs the negative influence.
Several studies showed some inconsistencies concerning the influence of anti-TNF-α agents on bone loss.18–20 After 1 year of treatment with infliximab, Vis et al18 showed in 102 patients with RA an unchanged BMD in the spine and hip, while Lange et al19 even showed a significant increase in BMD in 26 patients with RA. In the present study, infliximab was not associated with less BMD loss compared with the other treatment strategies after 1 year.
The determinants of BMD loss in this study were joint damage at baseline, joint damage progression and the non-use of bisphosphonates. Our results are in line with the results of Forslind et al,26 who showed in female patients with recent onset RA that lower BMD was associated with more radiographic damage after 2 years of treatment with DMARDs. Jensen et al42 found in patients with polyarthritis with a disease duration <2 years that patients with erosive disease had more BMD loss in the hand/forearm than patients with non-erosive disease. The association between joint damage and BMD loss suggests that pathophysiological processes involved in generalised osteoporosis appear to be common to local destructive joint involvement. This hypothesis is strengthened by the finding that osteoclasts, stimulated mostly by RANKL, play a central part in both pathophysiological mechanisms.22–24
Other disease-related factors such as DAS were not associated with BMD loss in our population, probably due to the aggressive suppression of the DAS with the DAS-steered treatment.
Unfortunately, the guidelines for anti-osteoporotic treatment in osteopenic or osteoporotic patients were poorly implemented by the rheumatologists in our study. Only 43% of the patients who needed bisphosphonates were actually prescribed oral bisphosphonates. In line with previous studies, the use of bisphosphonates independently protected against BMD loss.43–45 Our data support that calcium and/or vitamin D alone might be ineffective to reduce BMD loss.34
In conclusion, in patients with recently diagnosed RA, suppression of inflammation with any aggressive, effective treatment strategy is essential for preserving BMD. Because of the relationship between joint damage and BMD loss, disease activity should be suppressed as early and aggressively as possible, not only to prevent future disability due to erosive joint damage but also to prevent high BMD loss and associated risks. Furthermore, treatment with bisphosphonates is necessary in order to prevent further BMD loss.
Acknowledgments
We would like to thank all the patients as well as the following rheumatologists (other than the authors) who participated in the Foundation for Applied Rheumatology Research (all locations are in The Netherlands): MHW de Bois, MD (Medical Center Haaglanden, The Hague); G Collée, MD (Medical Center Haaglanden, The Hague); JAPM Ewals, MD (Haga Hospital, The Hague); BAM Grillet, MD (De Honte Hospital, Terneuzen); JHLM van Groenendael, MD (Franciscus Hospital, Roosendaal); JMW Hazes, MD (Erasmus Medical Center, Rotterdam); MH de Jager, MD (Albert Schweitzer Hospital, Dordrecht); JM de Jonge-Bok, MD (Groene Hart Hospital, Gouda); PJSM Kerstens, MD (Jan van Breemen Institute, Amsterdam); MV van Krugten, MD (Walcheren Hospital, Vlissingen); H van der Leeden, MD (retired); MF van Lieshout-Zuidema, MD (Spaarne Hospital, Hoofddorp); A Linssen, MD (Kennemer Gasthuis, Haarlem); PAHM van der Lubbe, MD (Vlietland Hospital, Schiedam); C Mallée, MD (Kennemer Gasthuis, Haarlem); HK Markusse, MD (deceased); AJ Peeters, MD (Reinier de Graaf Hospital, Delft); HK Ronday, MD (Haga Hospital, The Hague); D van Schaardenburg, MD (VU Medical Center, Amsterdam and Jan van Breemen Institute, Amsterdam); PEH Seys, MD (Lievensberg Hospital, Bergen op Zoom); RM van Soesbergen, MD (retired); PBJ de Sonnaville, MD (Oosterschelde Hospital, Goes); I Speyer, MD (Bronovo Hospital, The Hague); JPh Terwiel, MD (Spaarne Hospital, Hoofddorp); AE Voskuyl, MD (VU Medical Center, Amsterdam); ML Westedt, MD (Bronovo Hospital, The Hague); S ten Wolde, MD (Kennemer Gasthuis, Haarlem); D van Zeben, MD (Sint Franciscus Gasthuis, Rotterdam). We would also like to thank all other rheumatologists and trainee rheumatologists who enrolled patients in this study, AH Zwinderman (MD, Academic Medical Center, Amsterdam) for his statistical advice and JC Roos (MD, VU Medical Center, Amsterdam) for his advice concerning the BMD measurement methods.
REFERENCES
Footnotes
Funding: The BeSt trial was supported by a grant of the Dutch College for Health Insurance Companies (CVZ). Schering Plough and Centocor provided additional funding. The funding sources were not involved in the design of the study, the collection, analysis and interpretation of the data, the writing of the report or the decision to submit the paper for publication.
Competing interests: FCB did a paid expert testimony for Centocor in 1996 and was a paid speaker in a Schering Plough sponsored symposia. CFA was a paid speaker in a Schering Plough sponsored symposium in 2006.