Article Text
PDF
Abstract
Aim To assess whether in early (rheumatoid) arthritis (RA) patients, metacarpal bone mineral density (BMD) loss after 4 months predicts radiological progression after 1 year of antirheumatic treatment.
Methods Metacarpal BMD was measured 4 monthly during the first year by digital X-ray radiogrammetry (DXR-BMD) in patients participating in the IMPROVED study, a clinical trial in 610 patients with recent onset RA (2010 criteria) or undifferentiated arthritis, treated according to a remission (disease activity score<1.6) steered strategy. With Sharp/van der Heijde progression ≥0.5 points after 1 year (yes/no) as dependent variable, univariate and multivariate logistic regression analyses were performed.
Results Of 428 patients with DXR-BMD results and progression scores available, 28 (7%) had radiological progression after 1 year. Independent predictors for radiological progression were presence of baseline erosions (OR (95% CI) 6.5 (1.7 to 25)) and early DXR-BMD loss (OR (95% CI) 1.5 (1.1 to 2.0)). In 366 (86%) patients without baseline erosions, early DXR-BMD loss was the only independent predictor of progression (OR (95% CI) 2.0 (1.4 to 2.9)).
Conclusions In early RA patients, metacarpal BMD loss after 4 months of treatment is an independent predictor of radiological progression after 1 year. In patients without baseline erosions, early metacarpal BMD loss is the main predictor of radiological progression.
- Early Rheumatoid Arthritis
- Bone Mineral Density
- Outcomes research
Statistics from Altmetric.com
Introduction
Early treatment of patients with rheumatoid arthritis (RA) improves disease outcomes, including radiological joint damage.1–,3 Identification of patients who will have a more severe disease course may steer early treatment strategies. Since predicting disease outcome is currently not possible in a reliable way for all patients, there is a need for new predictors to improve existing prediction models.4–,7
Periarticular osteopenia is one of the earliest radiological manifestations in RA and may already be found in the phase of undifferentiated arthritis (UA).8 ,9 Metacarpal bone mineral density (BMD) loss may therefore be a potentially new predictor of disease outcome in patients with early (rheumatoid) arthritis. Previous research showed that metacarpal BMD loss is associated with disease activity10 and metacarpal BMD loss in the first year after diagnosis is predictive of radiological damage up to 5 years in patients with early RA.11–,13 For clinical practice, however, any predictive value of metacarpal BMD loss would be greater if it can be measured earlier in the disease course.
Therefore, we investigated whether metacarpal BMD loss after 4 months of treatment, as measured by digital X-ray radiogrammetry (DXR-BMD), may be a predictor of radiological joint damage progression after 1 year in patients with undifferentiated or early RA treated according to a tight control, remission-steered treatment strategy in the IMPROVED (Induction therapy with Methotrexate and Prednisone in Rheumatoid Or Very Early arthritic Disease) study.
Patients and methods
Patients and study design
Data from the IMPROVED study were used; it is a multicenter, randomised clinical trial in 610 patients, including 479 (80%) patients with recent onset RA (according to the 2010 classification criteria for RA14 with a symptom duration <2 years), 122 patients with UA (having at least one joint clinically assessed as ‘arthritis’ and one other painful joint, clinically suspected of having early RA regardless of symptom duration) and 9 patients that could not be classified because of missing data. Patients were treated according to a tight control strategy, aimed at achieving remission, defined as a disease activity score (DAS)<1.6 (DAS remission).15 All patients started with 4 months of methotrexate (MTX) 25 mg/week and prednisone 60 mg/day tapered to a stable dose of 7.5 mg/day in 7 weeks. Patients who achieved DAS remission after 4 months (early DAS remission group), first tapered prednisone to zero and if still in remission after 8 months, also tapered MTX to zero. Patients not in early DAS remission were randomised either to MTX 25 mg/week plus hydroxychloroquine 400 mg/day, sulfasalazine 2000 mg/day and prednisone 7.5 mg/day (arm 1) or to MTX 25 mg/week plus adalimumab (ADA) 40 mg/2 weeks (arm 2). Some patients who were not in DAS remission after 4 months were not randomised and treated outside of protocol (OP group). Full details about the IMPROVED study protocol were published previously.16 ,17
In the current analysis, we included all patients participating in the IMPROVED study whose radiological progression data after 1 year and at least one DXR-BMD result during the first year were available.
Demographic and clinical variables
At baseline, the following variables were collected: age, gender, symptom duration, body mass index, current smoking status and alcohol use, calcium intake, postmenopausal status, previous fractures, family history on osteoporosis, anticitrullinated protein antibodies (ACPA) and rheumatoid factor (RF) status. At baseline and every 4 months, the following clinical and laboratory variables were collected: DAS, including Ritchie Articular Index, swollen joint count, erythrocyte sedimentation rate (ESR, mm/h) and visual analogue scale (VAS) for global health, and C-reactive protein (CRP). During the first year, X-rays of hand and feet were made 4 monthly by digital radiography in all patients. Radiological progression, scored using the Sharp/van der Heijde scoring method, was assessed by two independent readers blinded for patient identity and time order of the radiographs.18 Progression was defined as an increase in Sharp/van der Heijde Score (SHS) of ≥0.5 points. Details on inter-reader reliability were published previously.17
Metacarpal BMD measurements
Suitable routine digital X-rays of both hands were used to measure metacarpal BMD using DXR-BMD measured by DXR online (Sectra, Linköping, Sweden), a computerised method that automatically recognises three regions of interest on the second, third and fourth metacarpal bones. At each region, DXR-BMD is estimated from multiple measurements of cortical thickness, bone width and porosity.19 The mean value of both hands was used in all analyses to avoid bias induced by hand dominance. ‘DXR-BMD loss’ was defined as a loss in DXR-BMD of ≥1.5 mg/cm2/4 months.10
Statistical analysis
Almost half of the available X-rays were found unsuitable for DXR measurements. This resulted in missing DXR-BMD values in 141/428 patients (33%) at baseline, 73/428 (17%) after 4 months, 148/428 (35%) after 8 months and 140/428 (33%) after 1 year. To avoid possible bias induced by missing data and to increase power, multiple imputation was performed. Ten datasets were created, in which missing DXR-values were imputed based on a linear regression model fitting available patient and disease characteristics and DXR values.20 Estimates obtained from regression analyses were automatically pooled by SPSS, and other multiple estimates were averaged.
Median (IQR) DXR-BMD changes were shown because of a skewed distribution. Mann–Whitney U test was used for comparisons of DXR-BMD changes between patients with and without radiological progression. To identify independent predictors of radiological progression, we performed univariate followed by multivariate regression analyses. From previous literature, the following potential predictors for (rapid) radiological progression were identified and entered in a univariate logistic regression model with radiological progression (yes/no) as a dependent variable: presence of ACPA and/or RF, baseline swollen joint count, baseline ESR and CRP levels, baseline total SHS, baseline erosion score and treatment.4 ,6 ,7 In addition, we selected age, gender, fulfilling the 2010 ACR/EULAR criteria for RA and achieving DAS remission after 4 months. Next to baseline erosion score we also entered presence of erosions, defined as ≥1 erosions, as covariate. Because only 28 (7%) patients included in this analysis had radiological progression, multivariate regression in the total study population was powered for about three variables.21 ,22 Therefore, in addition to DXR-BMD loss from baseline to 4 months, those two univariate significant predictors (using a significance level of 0.10) with the highest effect size were selected for multiple regression. As radiological progression was present in <10% of the patients and therefore can be classified as ‘rare’, we argued that ORs obtained from all logistic regression analyses can be interpreted as relative risks.23
All statistical analyses were conducted with SPSS for Windows V.20.0.
Results
Clinical characteristics
We included 428 patients in the current analysis. Baseline characteristics of these patients did not differ significantly from those participating in the IMPROVED study, where no SHS or DXR data were available (data not shown). Twenty-eight (7%) patients had radiological progression after 1 year and 400 (93%) had no radiological progression. For those patients with radiological progression, the median (IQR) progression score was 0.5 (0.5–1.4). One patient had rapid radiological progression (progression score ≥5 points)24 after 1 year (18 points).
When compared with patients without progression, patients with progression were older, more often postmenopausal, ACPA-positive and more often fulfilled the 2010 criteria for RA. Furthermore, they had more often ≥1 erosions at baseline and a higher median total baseline SHS and, only at 8 months, a slightly higher DAS (table 1).
Clinical characteristics at baseline and during 1 year follow-up of the total study group and separate for patients with and without radiological progression
DXR-BMD change
Table 2 shows absolute DXR-BMD values and DXR-BMD changes during the first year. When compared with patients without radiological progression after 1 year, patients with radiological progression had lower absolute DXR-BMD values at baseline and after 4, 8 and 12 months follow-up. From baseline to 4 months, median DXR-BMD changes were significantly larger in patients with radiological progression (median (IQR) −9.6 (−15.2;−2.7 mg/cm2)) than in patients without radiological progression (−2.0 (−7.2;2.5 mg/cm2), p=0.007). Twenty-four (86%) patients with radiological progression had DXR-BMD loss within the first 4 months compared with 212 (53%) patients without radiological progression (p=0.01). One patient with rapid radiological progression (18 points after 1 year) had DXR-BMD change within the first 4 months of −27.4 mg/cm2.
DXR-BMD measurements and changes in DXR-BMD during the first study year of the total study population and separate for patients with and without radiological progression
Treatment steps
Seventeen (61%) patients with radiological progression after 1 year had been in early DAS remission after 4 months and subsequently had started tapering prednisone to zero, nine (32%) had not achieved early remission and were randomised, and two were treated outside of protocol. Of the 17 patients in early DAS remission, 5 relapsed after tapering prednisone and restarted it, whereas 12 remained in remission and started tapering MTX to zero. Six patients relapsed after tapering MTX and restarted it, and six did not relapse and were in drug-free remission after 1 year. The median (IQR) early DXR-BMD change of all 17 patients was −10.9 (−14.5;−2.5) mg/cm2 (corresponding to −2.7 (−3.6;−0.6 mg/cm2/month)) compared with −1.8 (−7.3;2.4) mg/cm2 (corresponding to −0.5 (−1.8;0.6 mg/cm2/month)) in 258 patients who achieved early DAS remission and had no radiological progression after 1 year (p=0.02). DXR-BMD loss after 4 months was present in 14/17 (82%) patients in early DAS remission who had radiological progression after 1 year compared with 134 (52%) patients in early DAS remission without radiological progression after 1 year (p=0.053).
Predictors of radiological progression
Univariate predictive variables for radiological progression after 1 year were fulfilling the 2010 criteria for RA (p=0.07), presence of baseline erosions (yes/no) (p<0.001), presence of both ACPA and RF (p=0.03), early DXR-BMD loss after 4 months (p=0.008), baseline total SHS score (p=0.07), age (p=0.01), baseline ESR (p=0.06) and baseline tender joint count (p=0.05). Female gender, presence of either ACPA or RF, symptom duration, baseline erosion score and CRP level and treatment group were not predictive. Achieving DAS remission after 4 months was also not predictive of radiological progression after 1 year (table 3).
Univariate logistic regression analysis with radiological progression as binomial-dependent variable in the total study population
Together with early DXR-BMD loss, presence of baseline erosions and fulfilling the 2010 criteria for RA were selected for inclusion in the multivariate regression analysis. Both presence of baseline erosions and early DXR-BMD loss were predictive of radiological progression after 1 year independent of each other and independent of fulfilling the 2010 criteria for RA (table 4).
Multivariate logistic regression with radiological progression as binomial dependent variable in the total study population
In an additional multivariate model including early DXR-BMD loss, presence of baseline erosions and presence of both ACPA and RF, presence of both ACPA and RF was not an independent predictor of radiological progression, whereas both DXR-BMD loss and presence of baseline erosions were independent predictors for radiological progression (data not shown).
After leaving out one patient with rapid radiological progression (18 points after 1 year), the above results did not change significantly (data not shown).
Patients without baseline erosions
In 366 (86%) patients, no baseline erosions were present; of these 366 patients, 17 (5%) showed radiological progression after 1 year (61% of all 28 patients with radiological progression) and 349 (95%) did not. Median DXR-BMD change from baseline to 4 months was −11.8 (−16.7;−4.7 mg/cm2) in patients with progression and −2.0 (−7.0;2.4) mg/cm2 in patients without progression (corresponding to −2.9 (−4.2;−1.2) and −0.5 (−1.7;0.6) mg/cm2/months, respectively). Univariate significant predictors for progression after 1 year in patients without baseline erosions were age (p=0.004), baseline total SHS (in these patients reflecting baseline joint space narrowing) (p=0.009), baseline ESR level (p=0.096) and early DXR-BMD loss (p=0.02) (table 5).
Univariate logistic regression analysis with radiological progression as binomial-dependent variable in patients without baseline erosions
Early DXR-BMD loss and total baseline SHS were selected for inclusion in the multivariate regression analysis. Early DXR-BMD loss was predictive of radiological progression after 1 year independent of baseline total SHS in patients without baseline erosions (table 6).
Multivariate logistic regression with radiological progression as binomial-dependent variable in patients without baseline erosions
Discussion
In patients with early RA or UA, metacarpal BMD loss measured by DXR after 4 months of treatment with MTX and a tapered high dose of prednisone is predictive of future joint damage after 1 year of remission-steered treatment. In patients without baseline erosions (86%), metacarpal BMD loss was the main predictor of future joint damage.
These data suggest that DXR measurements over a period of 4 months from baseline can help to decide which patients with early arthritis should start antirheumatic treatment to prevent joint damage or damage progression, one of the main goals in the treatment of RA.25 Early treatment and suppression of disease activity has been shown to be associated with better suppression of radiological damage progression.1–,3 To facilitate this, in 2010, new classification criteria for RA were formulated.14 In the IMPROVED trial, we not only included patients with RA (according to the 2010 classification criteria) but also patients with UA, who were judged to represent RA in an early phase of the disease by the treating rheumatologist. Starting treatment so early in disease course carries the risk of overtreatment of patients who are misdiagnosed as RA, but a delay in treatment means risking irreversible joint damage progression.
To individualise treatment, predictive factors for damage progression have been identified and prediction models built.4 ,6 ,7 In particular in patients without baseline damage, predicting which patient will develop joint damage may be difficult. We predicted metacarpal BMD loss since this was linked with both disease activity and joint damage progression in patients with early and established RA, and metacarpal BMD loss after 1 year has been shown to have predictive value in addition to known predictors.11 ,12 Our paper is the first to report metacarpal BMD changes after 4 months, and we found that changes do occur.
Ideally, an outcome predictor can be identified already at baseline. In this early arthritis population, presence of baseline erosions was the only independent baseline predictor of radiological progression after 1 year besides metacarpal BMD loss after 4 months. Another obvious outcome after 4 months, remission yes or no, was not predictive for progression after 1 year. Some patients who had radiological joint damage after 1 year were in remission throughout the whole year and tapered all medication according to the study protocol. Our results indicate that after 4 months a strong predictor of progression may help to decide if adjustments of the chosen treatment strategy should be made in patients with early arthritis.
One limitation of this study is the fact that due to the inclusion of patients with early and relatively mild disease, progressively treated with the aim of achieving remission, only a few patients had radiological damage progression. Our results however reached statistical significance, although we acknowledge that the damage scores are hardly of clinical relevance this early in the disease phase. But as RA treatment aims at achieving total disease and damage control in an early phase of the disease, we think that our findings may be relevant for daily practice.
Another limitation was that we found many of the ‘routinely’ acquired radiographs to be unsuitable for DXR. To handle missing metacarpal BMD data, we performed multiple imputation20 to account for potential bias caused by data ‘missing at random’, meaning that missingness depends on other observed patient characteristics rather than on the fact whether metacarpal BMD measurements were done or not.
A third possible limitation may be that as DXR measurements in this study were done in retrospect on X-rays taken in 12 different hospitals using imaging protocols not adjusted to DXR, precision of the method may be lower than that previously published. DXR-BMD has been shown to have a very high short-term and long-term precision in both in vitro cadaver studies (coefficients of variation (CV) of 0.22–1%) and in one cohort study and one clinical trial (CV of 0.25–0.46%).26–29 However, supported by the consistency of our results, precision in this study may still be considered high.
If metacarpal BMD is to be applied in clinical practice using the DXR online method, neither low precision nor missing values may be problematic as X rays will be taken according to a predefined protocol (Sectra, Sweden). Precision may reach values described above, and in case of mal positioning, direct feedback will be given, which makes it more suitable for use in clinical practice.
In conclusion, we showed that loss of metacarpal BMD measured by DXR after the first 4 months of treatment is an independent predictor of future bone damage in patients with early RA. This suggests that 4-monthly metacarpal BMD measurements can help take treatment decisions in individual patients or may be added to improve the predictive value of existing prediction models for disease outcome in RA.
Acknowledgments
We would like to thank all patients as well as the following rheumatologists (other than the authors) who participated in the IMPROVED study (all locations are in The Netherlands): W M de Beus (Medical Center Haaglanden, Leidschendam); C Bijkerk (Reinier de Graaf Gasthuis, Delft); M H W de Bois (Medical Center Haaglanden, The Hague); M de Buck (Medical Center Haaglanden, Leidschendam); G Collée (Medical Center Haaglanden, The Hague); J A P M Ewals (Haga Hospital, The Hague); R J Goekoop (Haga Hospital, The Hague); Y P M Goekoop-Ruiterman (Haga Hospital, The Hague); B A M Grillet (Zorgsaam, Terneuzen); J H L M van Groenendael (Franciscus Hospital, Roosendaal); L R Lard (Medical Center Haaglanden, Leidschendam); E T H Molenaar (Groene Hart Hospital, Gouda); M van Oosterhout (Groene Hart Hospital, Gouda); A J Peeters (Reinier de Graaf Gasthuis, Delft); N Riyazi (Haga hospital, The Hague); H K Ronday (Haga hospital, The Hague); A A Schouffoer (Groene Hart Hospital, Gouda); P E H Seys (Lievensberg hospital, Bergen op Zoom); P B J de Sonnaville (Oosterschelde Hospital, Goes); I Speyer (Bronovo Hospital, The Hague); G M Steup-Beekman (Bronovo Hospital, The Hague); M L Westedt (Bronovo Hospital, The Hague). We would also like to thank all other rheumatologists and trainee rheumatologists who enrolled patients in this study, all research nurses for their contributions and the Sectra Company (Sweden) for estimating BMD of the metacarpals by online DXR.
References
Footnotes
Handling editor Tore K Kvien
Contributors KVCW-dB performed statistical analysis, interpreted the data and drafted the manuscript. LH, KV and JK contributed to the acquisition of the data, were involved in analysing and interpreting the data, and revised the manuscript. RJG, MvO, JBH, CB, MS-B, MdB and PdS participated in the study design and contributed to the acquisition of the data. TWJH participated in the study design, contributed in the acquisition of the data and was involved in revising the manuscript by critically revising the content. CFA participated in the study design, contributed to the acquisition of the data, was involved in analysing and interpreting the data, and helped to draft the manuscript. All authors read and approved the final version of the manuscript to be published.
Funding This study was designed by the investigators and financially supported by Abbott.
Competing interests None.
Patient consent Obtained.
Ethics approval Medical Ethical Committees of participating hospitals.
Provenance and peer review Not commissioned; externally peer reviewed.