Abstract
Objective. Spondyloarthritis (SpA) is the second most frequent inflammatory rheumatic disease, characterized by spinal involvement, peripheral arthritis, or enthesitis with marked pain, stiffness, and fatigue. Fibromyalgia (FM) may be associated with SpA, and shares some common symptoms. We aimed to determine how FM influences assessment of SpA disease activity, which is mainly dependent on patient-based outcome measures such as the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) or the Ankylosing Spondylitis Disease Activity Score (ASDAS).
Methods. This single-center cross-sectional study included consecutive patients with SpA according to the Assessment of SpondyloArthritis International Society criteria. FM was diagnosed according to the 1990 American College of Rheumatology criteria. Patient characteristics, BASDAI, ASDAS/C-reactive protein (CRP), Bath Ankylosing Spondylitis Functional Index, Bath Ankylosing Spondylitis Metrology Index, and the Medical Outcomes Study Short Form-36 questionnaire were recorded and compared.
Results. The study included 103 patients with SpA; 81 with axial and 22 with peripheral forms. Eighteen patients presented with concomitant FM, of whom 12 had axial SpA and 6 peripheral SpA. Demographic characteristics did not differ except for sex, with a female predominance in the FM group that was more marked in peripheral forms. BASDAI was higher in patients with FM [median (IQR): 4.2 (4.2) vs 2.2 (3.1); p = 0.0068], whereas ASDAS-CRP was not significantly different [median (IQR): 2.7 (2) vs 2 (1.3); p = 0.1264]. Nevertheless, median ASDAS-CRP corresponded to high disease activity in patients with SpA or FM compared with moderate activity in non-FM patients.
Conclusion. FM is a frequent comorbidity in patients with SpA, especially in peripheral forms. In patients with SpA-FM, disease activity may be overestimated when measured by BASDAI and to a lesser extent by ASDAS-CRP, and this overestimation could lead to inappropriate treatment escalation.
Spondyloarthritis (SpA) is the second most frequent inflammatory rheumatic disease after rheumatoid arthritis (RA), with an estimated prevalence of 0.5–1.9%1. In recent decades, the pattern of the disease has moved from the well-known male-predominant ankylosing spondylitis (AS) to descriptions of more sex-balanced forms. These include on the one hand nonradiographic axial forms and on the other, peripheral forms in which joint and enthesis involvement is predominant and still dependent on a common elementary enthesitis process2. This led to the need for new classification criteria, published in 2011 by the Assessment of SpondyloArthritis International Society (ASAS)3. These criteria focus on clinical presentation and distinguish axial SpA (axSpA), divided into radiographic forms (previously known as ankylosing spondylitis, AS) and nonradiographic forms, and peripheral SpA (pSpA), i.e., without axial involvement but with peripheral arthritis and/or enthesitis.
Although sex influence on the severity of SpA is still a matter of debate, women generally present with worse self-reported outcome and less radiographic damage in early and advanced SpA4,5,6,7,8,9. These studies, which postulated the potential effect of a particular behavior with regard to pain discomfort in women to explain the paradoxical discrepancy between objective and subjective presentation, unfortunately did not consider the presence of concomitant fibromyalgia (FM). Yet this frequent diffuse painful syndrome of indeterminate etiology10, largely seen in women11 and often associated with inflammatory diseases12, could be one of the factors that underlie this discrepancy. In a first preliminary study comparing 18 women and 18 men with AS, Aloush and colleagues diagnosed 50% of women with concomitant FM. These patients reported higher disease activity13. Three further studies focused on FM in axial SpA, reporting a higher prevalence than in the general population (from 4.1% to 15.5%)14,15,16.
Disease activity measurement in SpA is principally based on indexes reporting the intensity of subjective discomfort. The most frequently used index, the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), is calculated based on six 10-cm horizontal visual analog scales for self-measurement of subjective discomfort such as fatigue, spinal pain, peripheral joint pain and swelling, localized tenderness suggesting enthesitis, and morning stiffness17. The more recent Ankylosing Spondylitis Disease Activity Score (ASDAS) aims at a more objective approach by adding the erythrocyte sedimentation rate (ESR; ASDAS-ESR) or the C-reactive protein (CRP) level (ASDAS-CRP) to 3 items of the BASDAI (questions 2, 3, and 6 concerning spinal pain, peripheral joint pain, and duration of morning stiffness) and a patient global assessment of the activity of the disease18.
The overlap between SpA and FM symptoms such as pain, fatigue, or sleep disturbance suggests that SpA disease activity assessed with these indices may be overestimated in patients with concomitant FM. This has been demonstrated for the 28-joint DAS index in RA19,20 and it could eventually lead to an inappropriate increase in antiinflammatory treatment. Consequently, the aim of our study was to assess the effect of FM on SpA disease activity assessment measured with BASDAI and ASDAS-CRP. Functional disability and quality of life were also assessed as secondary endpoints.
MATERIALS AND METHODS
Our single-center cross-sectional study considered for inclusion consecutive patients with SpA from the department of one of the authors (JGT) between March 2010 and May 2011. Inclusion criteria were age over 18 years, fulfillment of ASAS classification criteria for axSpA3 and of Amor21 or European Spondylarthropathy Study Group (ESSG) classification criteria for peripheral SpA22. ASAS classification criteria for peripheral SpA, which were published after the beginning of our study, were assessed retrospectively, using items of the previous criteria3. FM was diagnosed when patients fulfilled the 1990 American College of Rheumatology (ACR) classification criteria at the time of examination or during a previous consultation10. To meet these criteria, patients must present widespread pain for more than 3 months with 11 or more tender points at clinical examination of 18 anatomically defined sites. A consecutive FM control population defined by the 1990 ACR criteria but without inflammatory rheumatic disease was included in our study.
A standardized case report form was completed for each patient for demographic data, current treatment, disease activity reported with BASDAI17 and patient global assessment (“How active was your rheumatism on average during the last week?”, as defined in the ASDAS-CRP index)18, disability assessed by the Bath Ankylosing Spondylitis Functional Index (BASFI; consisting of 10 self-reported questions about functional disability)23, and quality of life assessed by the Medical Outcomes Study Short Form-36 questionnaire (SF-36)24. A single examiner (JGT) completed the form for the Bath Ankylosing Spondylitis Metrology Index (BASMI), consisting of 5 standardized measures25, number of FM tender points, number of tender and swollen joints, sacroiliac radiographic assessment, and laboratory results.
One missing value was permitted for the BASDAI, BASMI, BASFI, and ASDAS-CRP questionnaires and 2 missing values for the SF-36; if more were missing, the patient was excluded from analysis. Missing values were replaced by the mean value of the remaining patients of each of the 2 groups (patients with SpA, and FM control patients).
Observational analysis of the effect of FM was carried out on the whole population and on axial and peripheral subgroups as defined by the ASAS classification criteria by measuring primarily the differences in BASDAI and ASDAS-CRP scores with and without this disorder. Secondary endpoints were also measured such as BASFI, BASMI scores, and the SF-36 score, to assess physical and psychological effect.
The ethical review board of the Hospice Civil de Lyon checked and approved the protocol. Because of the exclusive observational design of our study, consent of the consultative committee for the protection of persons was not required. Informed consent was obtained from all patients.
Data were analyzed using GraphPad Prism version 5.03 (GraphPad Software) and SPSS software version 17.0 (SPSS). Groups of patients were compared using the nonparametric Mann-Whitney test for continuous variables and the Fisher’s exact test for categorical variables as appropriate. Continuous variables are presented as medians (interquartile range, IQR) and categorical variables as numbers (percentages). In percentage calculation for each variable, the number of missing values was excluded from the denominator. The Spearman nonparametric test was used to assess the correlation between scores. For all analyses, a p value less than 0.05 was considered statistically significant.
RESULTS
Populations
The study included 103 patients with SpA, of whom 81 presented with axSpA and 22 with pSpA. All patients diagnosed with pSpA on the basis of the Amor or ESSG criteria retrospectively fulfilled ASAS classification criteria. Eighteen patients (17.5%) had concomitant FM, including 12 patients with axSpA (14.8%) and 6 patients with pSpA (27.3%). Thirteen of the 18 patients fulfilled the ACR 1990 criteria for FM at the time of examination, whereas 5 of the 18 had fulfilled them in a previous consultation. A control group of 18 consecutive patients with FM seen during the same period agreed to complete the same form and were similarly examined.
Demographic and disease characteristics are detailed in Table 1. The proportion of women was significantly higher in the FM group than in the whole population (66.7% vs 29.4%; p = 0.0056). The sex difference was not significant in the axial subgroup (50.0% vs 26.1%; p = 0.1669), but patients with FM had less radiographic sacroiliitis. Detailed data are not presented for the subgroup with peripheral involvement but only 2 characteristics were different: women were also predominant among the patients with pSpA FM (100% vs 43.7%; p = 0.0461) and were shorter [height 159.3 cm (IQR 9.8) vs 173 cm (IQR 18); p = 0.0386]. Prescriptions did not differ between FM and non-FM patients, whether for disease-modifying antirheumatic drugs, tumor necrosis factor (TNF) inhibitors, analgesics, antidepressant agents, or hypnotics (Table 1).
Population characteristics. Data presented as n (%) except as otherwise indicated.
Missing values
Concerning BASDAI, 3 patients had 1 missing value (item 4 for 1 patient with SpA, item 6 for 1 patient with SpA-FM, and item 4 for 1 FM control patient). Concerning ASDAS-CRP, global patient evaluation was missing in 2 patients with SpA (1 patient with SpA-FM and 1 FM control patient); CRP was missing in 4 patients with SpA and 1 FM control patient. In the axSpA subgroup, BASFI and BASMI were not analyzed in 6 and 5 patients, respectively, while 2 and 5 patients had 1 missing value. Concerning SF-36, 6 patients were excluded and 1 or 2 missing values were replaced in 13 patients.
Assessment of activity
Comparisons of BASDAI according to FM status are presented in Table 2. In the whole population, total score (p < 0.01) and questions 2, 3, and 4 (p < 0.05) were higher in the SpA-FM group. In the axial subgroup, only questions 3 and 4 were significantly higher in FM patients (p < 0.05). In the peripheral subgroup all questions except number 6 were higher (p < 0.05), but interpretation should be cautious owing to the low number of patients. ASDAS-CRP was not significantly higher in SpA FM and non-FM patients. However, median ASDAS-CRP corresponded to high disease activity in patients with SpA-FM compared with moderate activity in non-FM patients.
Comparison of BASDAI, ASDAS-CRP, and secondary criteria in patients with spondyloarthritis according to fibromyalgia status. Data presented as median (IQR).
The number of painful joints was higher in patients with SpA-FM, whereas the number of swollen joints was not (Table 2). The same results were observed in both the axial and peripheral subgroups.
In the axSpA subgroup, there was no difference according to FM status for the BASFI [median (IQR) axSpA-FM (n = 10): 1.4 (2.7); axSpA without FM (n = 65): 0.9 (2.5), p = 0.2921], and for the BASMI [median (IQR) axSpA-FM (n = 12): 1.5 (1.8); axSpA without FM (n = 64): 1.5 (3.8), p = 0.3526; Table 2]. Correlation according to Spearman’s nonparametric test between BASFI and BASMI scores was found only in patients with axSpA without FM (n = 60; r = 0.5496; p < 0.0001) and not in patients with concomitant FM (n = 10; r = 0.2733; p = 0.4448). There was no correlation between BASDAI and BASMI in either the FM or the non-FM group (data not shown).
Quality of life assessed by the SF-36 did not differ between non-FM (n = 82) and FM patients (n = 15; data not shown), except for 1 of the 8 concepts, “physical health,” which was lower in patients with SpA-FM [median (IQR): 70 (45) vs 85 (25); p = 0.0156].
Patients with SpA-FM were compared with 18 FM control patients (Table 3). There were no differences in demographic characteristics, total BASDAI, and ASDAS-CRP, or in the tender and swollen joint counts.
Comparison of demographics and disease activity assessment in spondyloarthritis patients with concomitant fibromyalgia and fibromyalgia control patients. Data presented as median (IQR) except where indicated.
Because female predominance is a major epidemiological characteristic of FM, secondary cross-sectional analysis of the whole SPA population was performed according to sex (Table 4). Women presented significantly less axial involvement (clinical and radiographic), more peripheral manifestations, more uveitis, and were more affected by concomitant FM. BASDAI and ASDAS-CRP did not differ between men and women.
Comparison of demographics, spondyloarthritis profile, and fibromyalgia status between men and women. Data presented as median (IQR) except where indicated.
DISCUSSION
In our population of 103 patients with SpA, 18 (17.5%) were diagnosed with concomitant FM according to the ACR 1990 classification criteria. Prevalence was higher in pSpA (27.3%) than in axSpA (14.8%). Only a few previous studies have focused on FM in SpA: prevalence was 4.1% among 462 patients with AS (New York criteria) from 10 Spanish centers14, 15.5% among 71 patients with AS (New York criteria) from 1 Brazilian center15, and 14.9% among 402 Italian patients with axSpA [AS (New York criteria) and axial psoriatic arthritis (ASAS classification criteria for axSpA and psoriasis)]16. For FM classification, the first 2 studies used ACR 1990 criteria while the last used ACR 2010 criteria.
It may be supposed that some primary FM may have been misclassified as SpA. Our study was not designed to compare the accuracy of different sets of criteria. However, Baraliakos, et al in a preliminary study compared 93 FM and 91 axSpA cases26. No patient with FM (median age 50.7 yrs, 93.4% women, 7.5% HLA-B27) fulfilled ASAS classification criteria for axial SpA. Doubts may remain for peripheral SpA where prevalence of FM was higher, but patients were considered as having SpA by the treating physician (JGT) even before they were considered for inclusion, and fulfilled at least 2 sets of classification criteria (either ESSG or Amor criteria and retrospectively ASAS criteria). Moreover, ASAS criteria require 1 objective symptom (arthritis, enthesitis or dactylitis) to be fulfilled3. Nevertheless, a lack of specificity of the new ASAS criteria has been disputed, particularly for nonradiographic axSpA and peripheral SpA27,28. Widespread pain and tenderness of the body in FM could be mistaken for enthesitis. Indeed, Marchesoni, et al found a Maastricht Ankylosing Spondylitis Enthesitis Score higher among 120 patients with FM compared to 266 patients with psoriatic arthritis, which was correlated to the number of FM tender points29.
The frequencies of FM in our population of SpA (17.5%) and especially of axSpA (14.8%) are close to those of the Brazilian study of Azevedo, et al (15.5%)15 and of the Italian study of Salaffi, et al (14.9%)16. Still, prevalence was lower in the Spanish study by Almodóvar, et al14 concerning 462 AS cases from 12 centers. There may be some recruitment bias in tertiary care centers such as in our study and in the first 2 studies cited that could lead to overestimation of prevalence. Another explanation for this discrepancy may be the criteria used for SpA. Most previous studies used the New York modified criteria for AS, which represented only 55 patients in our study, with an FM prevalence of 9.1%. Still, even in the study by Almodóvar, et al, prevalence was higher than expected in the general population, as has been reported for other chronic inflammatory diseases12. Finally, this study confirms that FM prevalence is higher in SpA than in the general population.
We found a female predominance in patients with concomitant FM, as has previously been described in SpA. The sex ratio was 2 in our study, similar to the results of Almodóvar, et al in Spain (2.2)14 and Salaffi, et al in Italy (2.3)16, and higher than those of Azevedo, et al in Brazil (1.2)15, but still lower than most estimations in the general population10,11. With the exception of height, female predominance was the only significant difference between FM and non-FM patients in the whole population and in the peripheral subgroup. Female predominance in patients with SpA-FM was less marked in the axial subgroup. This is in agreement with the male predominance classically observed in radiographic axSpA, whereas the sex ratio is usually more balanced in peripheral SpA30,31,32.
However, FM prevalence in men was unexpectedly high in this clinical cohort (9.1%) compared with the frequency found in other diseases such as RA19 or in the general population11. It is however consistent with the results reported by Salaffi, et al16 and more generally with the increased incidence of FM in other chronic rheumatic diseases12. This observation may give new insight into the debate on the relationship between FM and chronic diseases such as RA or SpA: chronic inflammatory pain may induce alterations in central pain regulation mechanisms, known as central sensitization, and finally lead to chronic widespread pain and other somatic symptoms, even eventually FM, in a person who might not have developed such a condition in the absence of inflammatory disease33,34.
BASDAI was higher in patients with FM, especially in peripheral SpA, where female predominance was the highest. In axSpA, only questions 3 and 4 relating to peripheral pain were higher in patients with FM, whereas responses to questions about fatigue and morning stiffness did not differ. Because FM is a diffuse painful syndrome, we can suppose that this difference may reflect the effect of FM rather than a more severe inflammatory process. Almodóvar, et al and Azevedo, et al reported higher BASDAI in patients with axSpA-FM14,15, whereas Salaffi, et al found no significant difference in the BASDAI score between FM and non-FM patients with SpA16. Because BASDAI assesses symptoms, it may reflect the sum of the manifestations of both SpA and FM. The high BASDAI score in our control population of FM with similar demographics underlines its lack of specificity, in accordance with previous studies reporting median BASDAI (IQR) of 6.0 (2.7) in 70 patients with FM and of 4.5 (1.6) in 248 patients with FM16,35. This lack of specificity in measuring SpA inflammatory disease activity in patients with concomitant FM needs to be borne in mind when antiinflammatory therapy is discussed, with the risk of inefficacy, occurrence of adverse events, and excessive cost if TNF blockers are considered.
The ASDAS may then be of interest because it includes the ESR or CRP level as a single objective item. Salaffi, et al concluded that ASDAS may distinguish axSpA patients with disease activity from those with functional impairment related to FM, but values were similar in the axSpA and axSpA-FM groups and considerably overlapped with those of the FM group. We found no significant difference in ASDAS score between patients with SpA-FM and patients with SpA. However, with regard to clinical therapeutic decision making, median ASDAS-CRP in patients with SpA-FM was considered to indicate high disease activity, whereas the SpA non-FM group was considered to present with moderate activity, which may still lead to inappropriate treatment. The relative lack of specificity of ASDAS may be related to the usually low ESR and CRP levels in patients with SpA, as reported here and in other cohorts30,31,32. This drawback could be overcome by using high-sensitivity CRP36 or another more pertinent marker such as calprotectin37. Nevertheless, in the Spanish cohort CRP was significantly lower in 19 patients with SpA-FM than in 443 patients with SpA, whereas BASDAI was higher14. In any case, compared with the BASDAI, ASDAS-CRP still includes 4 patient-reported outcomes (patient’s global assessment and BASDAI items 2, 3, and 6), but excludes the 2 questions concerning fatigue, which are the most unspecific, and local tenderness suggestive of enthesitis, a diagnosis that may be challenging in patients who have concomitant FM.
The influence of FM on subjective assessment is also illustrated by the discrepancy observed between a high count of tender joints and a low count of swollen joints in patients with SpA-FM, as highlighted in RA19,20, and by the lack of correlation between the BASFI and the BASMI in patients with axSpA-FM. Similarly, Almodóvar, et al described higher BASFI and lower Bath Ankylosing Spondylitis Radiographic Index in patients with AS-FM compared with non-FM patients14.
Interestingly, quality of life assessed with the SF-36 questionnaire revealed little difference between patients with FM and non-FM, except for one of the 8 concepts, “physical health,” which was poorer in SpA-FM, illustrating the complexity of such evaluation.
Our study has a number of limitations. The relatively small sample required the use of nonparametric tests that, although robust, limited the drawing of conclusions in some subgroup analyses, especially in the pSpA subgroup. However, the epidemiological characteristics and presentation of the SpA profile of this population, such as the more frequent axial involvement in men and the more frequent peripheral involvement in women (Table 4), were similar to the findings of larger studies, suggesting this sample was nevertheless representative4,5,6,7,8,9. Clinical assessment was not blinded but this did not affect our main results concerning patient-based outcome measures, and assessment by a single experienced examiner ensured accuracy38.
FM is often associated with SpA, especially in peripheral forms and in women. This concomitant FM should be considered as an important contextual factor that may account for inaccuracy in assessing disease activity in SpA, particularly pSpA. Inflammatory disease activity may be overestimated in such patients if assessed using BASDAI, and to a lesser extent using ASDAS-CRP. When interpreting these indices, physicians should be aware of the presence of concomitant FM to avoid inappropriate escalation of antiinflammatory treatment.
Acknowledgment
The authors acknowledge with gratitude the contribution to this work of their greatly missed co-author, Marie-Claude Letroublon, who died August 30, 2015.
Footnotes
M.C. Letroublon died August 30, 2015.
- Accepted for publication July 21, 2016.
