Article Text
PDF
Abstract
Objectives To determine the cumulative incidence and identify the factors associated with difficult-to-treat axial spondyloarthritis (D2T-axSpA) in French patients newly benefiting from the French ‘long-term illness’ (LTI) social security scheme for axial spondyloarthritis (axSpA).
Methods This national cohort study was based on the French National Medico-Administrative Database, SNDS, which contains data on hospitalisation, LTI and outpatient care consumption. All French patients newly receiving LTI benefits for ankylosing spondylitis (AS) between 2010 and 2013 were included in the study. In France, LTI is required to access biological/targeted synthetic DMARDs (b/tsDMARDs). The follow-up period ended on 31 December 2018. So-called D2T-axSpA was defined as the failure of three b/tsDMARDs or of two b/tsDMARDs with different modes of action. Comorbidities and extra-musculoskeletal manifestations were identified using previously described algorithms. Characteristics were compared between patients with D2T-axSpA and patients with non-D2T-axSpA who had received at least one b/tsDMARD with bivariate and multivariate analysis using logistic regression. Incidence rates of major cardiovascular event (MACE) and death were compared using competitive risk analysis.
Results 22 932 patients were included. 10 798 (47.08%) patients received at least one bDMARD. None received tsDMARD. During follow-up, 2115 patients were classified as having D2T-axSpA, representing 19.59% of patients who received at least one bDMARD. In multivariate analysis, D2T-axSpA was significantly associated with female gender, peripheral involvement, psoriasis, hypertension and depression (p<0.001 for each case). There was no difference in the incidence of MACE (p=0.92) or death (p=0.87).
Conclusion D2T-axSpA affects one in five patients exposed to bDMARDs in this national cohort. D2T-axSpA is more common in women and patients with peripheral involvement and/or comorbidities.
- ankylosing spondylitis
- epidemiology
- biological therapy
Data availability statement
Data are available upon reasonable request. All relevant data are reported in the article. Additional details can be provided by the corresponding author upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
There is a proportion of patients with axial spondyloarthritis (axSpA) for whom the disease can be considered difficult to treat (D2T).
WHAT THIS STUDY ADDS
D2T-axSpA affects one in five patients with axSpA who have received at least one biological/targeted synthetic DMARD.
The main factors associated with D2T-axSpA were female sex, peripheral symptoms, psoriasis, hypertension and depression.
Patients with D2T-axSpA do not have a higher risk of major cardiovascular event or death compared with patients with non-D2T-axSpA.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study can help clinicians identify individuals who are more likely to have D2T-axSpA and may require more aggressive treatment or closer monitoring, especially as this concerns a significant proportion of patients.
Further research using clinical data is needed to confirm the findings of this study.
Introduction
Axial spondyloarthritis (axSpA) has benefited from significant therapeutic progress with the availability of targeted therapies, anti-TNF, anti-IL17 and more recently, JAK inhibitors. However, efficacy is not present in all cases; some patients have failed several treatments directed against the same or different targets. This is a special population that, as has been proposed for rheumatoid arthritis,1 could correspond to a disease that is difficult to treat (D2T) and could correspond to a special patient profile. This notion is being discussed within the Assessment of SpondyloArthritis international Society (ASAS) group to propose a definition of these patients, knowing that a partial transposition of the definition of D2T from rheumatoid arthritis to axial spondyloarthritis could be used as an exploratory measure.2 This approach has recently been implemented for psoriatic arthritis, with the definition tested in a single-centre cohort.3 The importance of analysing the cause of therapeutic failure (primary or secondary) as well as the duration of exposition should be emphasised to reach a consensus definition of therapeutic non-response.4 Phenotypic presentation may also play a role, as targeted therapies may have different efficacies for the predominant symptoms (axial, peripheral and enthesitic).
Recent registry and cohort data have highlighted the low percentage of remission or low disease activity that is maintained over time. For example, in a Spanish prospective cohort, one-third of patients on biological therapy for axSpA achieved a low activity status over 18 months, and less than 20% achieved remission.5 In a report from the Scandinavian registries, analysis of targeted therapy switches in more than 6000 patients revealed that 8%, 3% and 1% of patients received at least three, four or five targeted therapies, respectively, during the 3 years of follow-up. There was no relationship between the number of rotations and calendar period. Baseline characteristics associated with multiple rotations (at least three) were female sex, short disease duration, high patient score, presence of comorbidities and psoriasis, but not uveitis.6
Moreover, we have previously shown in a study based on the same database that the incidence of major cardiovascular events (MACE) is low in patients with ankylosing spondylitis, as are the death rates.7 Nevertheless, the effect of the ‘difficult-to-treat’ nature of the disease on the risk of MACE or death is unknown. This study aimed to determine the cumulative incidence, identify the factors associated with D2T-axSpA, and compare the risk of MACE and death between patients with D2T-axSpA and patients with non-D2T-axSpA.
Methods
Data source and study design
This cohort study was based on data from the National Healthcare Data System (SNDS), a French medico-administrative database containing administrative data, data on long-term illnesses, outpatient care and medication consumption, and hospitalisations of all beneficiaries of the French health insurance system. We had access to an extraction of this database provided by the French National Health Insurance in accordance with the expression of needs corresponding to the inclusion and non-inclusion criteria of the study. The data analysed covered the period from 1 January 2009 (ie, at least 1 year before the inclusion date) to 31 December 2018. Therefore, there was a minimum follow-up period of 5 years and a maximum of 9 years.
Data cleaning by applying filters to the inclusion and non-inclusion criteria and deleting duplicates was performed before statistical analysis.
This article is consistent with the REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement for observational studies using routinely collected health data.8
Ethics statement
This study was approved by the Ethics and Scientific Committee for Health Research, Studies, and Evaluations (CESREES) and the French National Commission for Information Technology and Civil Liberties (CNIL) (authorisation number: DR-2021-231). Written consent was not obtained from patients in accordance with French law because the study was based solely on routinely collected, anonymised medico-administrative data. Patients were informed of the study via the French Health Data Hub website and Besançon University Hospital website.
Study population
Patients who were newly enrolled in the long-term illness (LTI) scheme with a diagnosis of ‘severe spondyloarthritis’ between 1 January 2010 and 31 December 2013, aged 18 years or older and with an International Classification of Diseases 10th Revision (ICD-10) code related to the enrolment of M45 were included. The choice of code complies with recommendations from the Network for Better Use of National Health Data System (REDSIAM).9 Although the name attached to the ICD-10 code is ‘ankylosing spondylitis’, we have made the assumption, based on our clinical experience and use of the database, that it is used in France for both radiographic and non-radiographic axSpA. The criteria for non-inclusion were the identification of treatment with abatacept or apremilast during the study period, indicating a diagnosis of psoriatic arthritis rather than axSpA. The use of LTI allowed us to build a cohort of patients with a presumed active disease, the SNDS, being a medico-administrative database with no data on the activity of the disease. In France, because they are costly, almost all patients using biological/targeted synthetic DMARDs (b/tsDMARDs) are enrolled in the LTI scheme, which covers all healthcare costs related to illness. Thus, all patients with active SpA despite Non-steroidal anti-inflammatory drug (NSAID) treatment and therefore requiring DMARDs are theoretically enrolled in the LTI scheme for the first time.
Only patients covered by the general scheme of the French health insurance, that is, most employees, students, social benefit recipients and ordinary residents (> 85% of all beneficiaries in France), were included to limit missing data because LTI data were not available for the other insurance schemes for the study period. The year 2010 was chosen as the inclusion start date because the LTI data in the SNDS were not reliable and/or comprehensive before that date. The inclusion period was arbitrarily set at 4 years (ie, up to and including 2013) in order to create a large cohort and allow at least 5 years of follow-up with the data available at the time the study was carried out.
The date of inclusion in the study corresponded to the date of enrolment in LTI, or to the date of first prescription of the biotherapy in cases where this preceded enrolment in LTI due to administrative delays.
Outcomes
The primary endpoint was the incidence rate of D2T-axSpA. As there is currently no conventional definition of D2T-axSpA, we chose a definition similar to that of difficult-to-treat rheumatoid arthritis (D2T-RA) as defined by European Alliance of Associations for Rheumatology (EULAR),1 that is, failure of three b/tsDMARDs in total or two b/tsDMARDs with different modes of action. The second and third criteria of the EULAR definition of D2T-RA, namely ‘Existence of active disease, that is, signs suggestive of active/progressive disease’ and ‘The management of signs and/or symptoms is perceived as problematic by the rheumatologist and/or the patient’, cannot be studied in medico-administrative databases, and therefore have not been considered in this study. Subgroups of patients with D2T-axSpA were formed according to whether they met the definition in less or more than 2 years of exposure to b/tsDMARDs.
The secondary objectives were to compare the incidence of major cardiovascular events, defined as the first occurrence of myocardial infarction (MI) or stroke, and death between patients with ‘D2T’-axSpA and patients with non-‘D2T’-axSpA.
Covariates
Regarding clinical phenotype and extra-articular manifestations (EMMs), peripheral symptoms were defined as having an ICD-10 code related to the LTI enrolment of M07 (psoriatic and enteropathic arthropathies) in addition to M45, and/or having at least one prescription of a conventional synthetic disease modifying antirheumatic drug (csDMARD) at baseline or during follow-up. EMMs (psoriasis, inflammatory bowel disease (IBD) and severe uveitis) were identified from algorithms previously used in the literature based on LTI benefits, and/or diagnoses during hospital stays (online supplemental table S1).
Supplemental material
Regarding comorbidities, the existence of diabetes, hypertension, dyslipidaemia, smoking, obesity, chronic kidney disease and depression was identified using previously described algorithms.10–15 These were based on LTI benefits, reimbursement of drug treatments, and diagnoses or technical procedures during a hospital stay (online supplemental table S2). For example, a patient is considered to be diabetic if he or she benefits from an LTI for diabetes, if he or she has been hospitalised for the discovery or decompensation of diabetes, or if he or she has benefited from reimbursements for specific antidiabetic drugs.
Regarding the treatments of interest, we collected outpatient reimbursements for NSAIDs (differentiating between COX2-selective and non-selective inhibitors) and three csDMARDs (methotrexate, leflunomide and sulfasalazine). Outpatient reimbursements and inpatient prescriptions were recorded for b/tsDMARDs, including biosimilars: five anti-TNF drugs (adalimumab, certolizumab pegol, etanercept, infliximab and golimumab) and two anti-IL17 drugs (secukinumab and ixekizumab). The treatment period was defined as the period from the date of first reimbursement to the last available date. There were no indications for tsDMARDs in axial spondyloarthritis during the study period, and none of the patients received them. Data on comorbidities and treatments were collected from 1 January 2009, at least 1 year before the date of inclusion in the study.
Two socioeconomic indicators were used as proxies for social deprivation. At the individual level, there are indicators of assistance in accessing complementary health insurance (free complementary health coverage (CMU-C) or aid in the payment of complementary health insurance (ACS)). State medical aid (AME) provides emergency coverage for undocumented aliens. At the ecological level, the French Social Deprivation Index (Fdep)16 at the city of residence scale was collected. Social deprivation is considered to be an accumulation of material and social deprivation on a geographical scale. The score is derived from a principal component analysis of the four parameters and then standardised. Therefore, its mean value is 0, and the higher the value, the greater the social deprivation.
MACE was defined as the first occurrence of stroke or MI. These events were identified using previously described algorithms based on the ICD-10 diagnoses of hospital discharge abstracts (online supplemental table S3). The incidences of stroke, MI and death were also assessed separately.
Statistical analysis
Descriptive analysis was performed by calculating the means and SDs for quantitative variables, and frequencies and percentages for qualitative variables. To limit selection bias, comparisons were restricted to patients with D2T-axSpA and patients with non-D2T-axSpA who had received at least one b/tsDMARD. Multivariate analysis using stepwise logistic regression was performed, considering variables that were significant in the bivariate analysis. The model was adjusted for age and the duration of exposure to b/tsDMARDs. A sensitivity analysis restricting the analysis to patients who met the D2T-axSpA definition within 2 years of initiating the first bDMARD was performed.
The incidence rates of MACE and their two components (stroke and MI) were calculated by computing the crude rate and cumulative incidence functions (CIFs), considering the competing risk of death. The CIF curves were compared using Gray’s test. Survival curves were constructed using the Kaplan-Meier method and compared using the log-rank test.
All tests were bilateral, with the threshold for statistical significance set at p<0.05. All analyses were performed using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA).
Patient and public involvement
Patients were not involved in this study as it was based solely on medico-administrative data.
Results
A total of 117 892 observations of patients with LTI for SpA were present in the extraction provided by French Health Insurance. After various stages of database verification and application of the inclusion criteria, 22 932 distinct patients were included in the study (figure 1). The mean (SD) follow-up time was 6.43 (1.00) years (min 0.03 – max 9 years). During the study period, 10 798 (47.08%) patients received at least one b/tsDMARD. The distribution of patients according to the number of b/tsDMARDs prescribed during follow-up is presented in figure 2. During follow-up, 2115 patients were classified as having D2T-axSpA, representing 9.22% of all patients and 19.59% of patients who received at least one b/tsDMARD.
Flow chart of study subject selection. b/tsDMARDs, biological/targeted synthetic DMARDs.
Distribution of patients according to the number of b/tsDMARDs prescribed during follow-up. b/tsDMARDs, biological/targeted synthetic DMARDs; D2T-axSpA, difficult-to-treat axial spondyloarthritis.
In bivariate analysis, peripheral involvement and psoriasis were significantly more frequent in the D2T-axSpA group (table 1). A significant increase in severe smoking, severe obesity, hypertension and depression prevalence was observed in the D2T-axSpA group. These results were confirmed by multivariate analysis, which showed the highest OR for peripheral symptoms (table 2). Analyses restricted to patients fulfilling the definition of D2T-axSpA over a maximum period of 2 years (n=1055 (0.05%)) showed similar results, except for severe obesity, for which the comparison became non-significant, and severe uveitis, which was significantly higher in the D2T-axSpA group, although with very small numbers: 11 (1.04%) versus 43 (0.44%), p=0.009 (online supplemental table S4).
Comparison of sociodemographic characteristics, extra-musculoskeletal manifestations and comorbidities between patients with D2T-axSpA and patients with non-D2T-axSpA who received at least one bDMARD
Bivariate and multivariate analysis of factors associated with D2T-axSpA
During follow-up, 124 MACE and 96 deaths occurred in patients who received at least one b/tsDMARD. The repartitions between groups are presented in table 3. There was no significant difference in the incidence of MACE when the competitive risk of death was considered (8-year cumulative incidence: 1.32% in the D2T-axSpA group vs 1.27% in the non-D2T-axSpA group, p=0.92) (figure 3). The risk of death was not significantly higher in patients with D2T-axSpA after adjustment for age, sex and duration of exposure to b/tsDMARDs (HR: 0.95 (0.50–1.79), p=0.87).
Comparison between patients with D2T-axSpA and patients with non-D2T-axSpA and crude incidences of major cardiovascular events (MACE) and deaths
Cumulative incidence functions of MACE in patients with ‘D2T’-axSpA and patients with non-‘D2T’-axSpA. D2T-axSpA, difficult-to-treat axial spondyloarthritis; MACE, major cardiovascular event.
Discussion
In this study, based on the medico-administrative data of 22 932 patients newly enrolled in the spondyloarthritis LTI scheme, D2T-axSpA affected almost one in five patients who received at least one b/tsDMARD.
The characteristics at inclusion of all patients were similar to those of the population in our previous study using the same database.7 The predominance of women among LTI patients, already demonstrated in France,17 may be explained by the fact that women have a higher average disease activity.
EMM rates at inclusion were slightly higher than those previously described in the literature for IBD and psoriasis.18 19 This could be explained by the fact that the subjects were included on the basis of LTI and therefore had a more severe disease with more EMMs than all patients with axSpA. Moreover, the comparisons made were only in patients exposed to b/tsDMARDs, who have more active disease and therefore more comorbidities and EMM than unexposed patients. In contrast, uveitis was only observed in 82 patients (0.36%), which was much lower than expected. This is because the algorithm used only detects hospitalised patients, leading to a significant lack of sensitivity. With regard to comorbidities, there was a similar lack of sensitivity in the algorithms for detecting smoking and obesity, which explains the low rates observed. For the rest of the comorbidities, rates were broadly similar to what has already been reported in the literature.20 21
We found a significant association between psoriasis, female sex, severe smoking, severe obesity, depression and D2T-axSpA. This is consistent with the study by Di Giuseppe et al, based on several Nordic registries.6 In that study, female sex, comorbidities (unspecified) and psoriasis were associated with use of ≥3 b/tsDMARDs. Interestingly, a history of uveitis was not associated with D2T-axSpA, as observed in our results. This may be explained by an indication bias, since only anti-TNF monoclonal antibodies have so far shown efficacy in the treatment of uveitis.22 23 Zhao et al also showed that patients with multiple comorbidities were at a greater risk of discontinuing the first TNF inhibitor (TNFi) than those with only one.24
Redeker et al showed that psoriasis was associated with higher disease activity and functional impairment, which may help explain this association.25 Similar associations have been reported between depression26 27 and smoking.28 It is also known that female patients have significantly lower response rates to TNFi therapy and significantly lower adherence to treatment.29 30 Peripheral involvement also seems to play an important role, and cluster analysis of the Be-Giant cohort showed that patients with peripheral involvement more often had persistently high disease activity.31 The main factors associated with persistently high disease activity are therefore correlated in our study with the occurrence of D2T-axSpA.
We also observed a higher prevalence of hypertension in patients with D2T-axSpA. The occurrence of hypertension has been shown to correlate with disease duration, particularly in axial disease.32 33 As we have no information on disease duration in our cohort, it is possible that this significant association may be explained by confounding bias.
We found no excess mortality or MACE in patients with D2T-axSpA. This result may seem surprising, given that mortality is linked to disease activity34 and comorbidities,21 two factors themselves associated with D2T-axSpA. The prevalence of cardiovascular comorbidities was similar in both groups, apart from hypertension. Smoking and obesity rates were higher in the D2T-axSpA group, but this only concerned the most severe cases (hospitalisation or prescription of substitution treatments). Nevertheless, it has already been shown that these factors are associated with lower therapeutic maintenance of anti-TNF drugs.35–37 Further studies are needed to clarify the influence of D2T-axSpA on mortality and MACE.
Our study has several strengths: it was based on a large cohort and used a national and exhaustive database. The follow-up period was relatively long. Our study is the first to describe the characteristics and factors associated with D2T-axSpA in a large cohort of patients. A large number of comorbidities could be studied using most of the algorithms already validated in the literature.
However, our study had several limitations. Because there is no consensus definition of D2T-axSpA, we relied on the D2T-RA definition. Only the number of treatment criteria was used because we had no clinical, biological or radiographic data. We therefore have no data on disease activity, and no information on the reasons for treatment switches. However, we performed sensitivity analyses, taking into account patients who reached the D2T-axSpA definition rapidly (less than 2 years), and obtained similar results. Moreover, taking into account the failure of multiple modes of action makes our definition more than just a failure of multiple lines of treatment. Finally, our hypothesis of selecting patients with LTI to represent those with active disease was not verified, since 50% of patients ultimately received no bDMARDs. The specificity of the algorithm used, although not studied, probably remains high, since LTI is required for the prescription of a bDMARD in France. It can be assumed that only patients who received another, older LTI and for whom a new LTI for SpA was not obtained were not included in this study.
In conclusion, D2T-axSpA affects one in five patients exposed to bDMARDs in this national cohort. D2T-axSpA is more common in women and patients with peripheral involvement, psoriasis, hypertension and depression. Although D2T-axSpA is associated with comorbidities, it does not appear to increase the risk of MACE or death. Studies on large cohorts with clinical, biological and radiographic data are important to clarify the various characteristics of D2T-axSpA.
Supplemental material
Data availability statement
Data are available upon reasonable request. All relevant data are reported in the article. Additional details can be provided by the corresponding author upon reasonable request.
Ethics statements
Patient consent for publication
Acknowledgments
We wish to thank Mrs. Robert for her help with the administrative procedures necessary to access the data.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
FV and DW contributed equally.
Contributors OF: design, acquisition and interpretation of the work; drafting of the work; revision and approval of the final version. MD: Design, acquisition, revision and approval of the final version. BM: Acquisition, revision of the manuscript and approval of the final version. CP and FV: Design, interpretation of the work, revision and approval of the final version. EM: Acquisition of the work, revision and approval of the final version. DW: Design, interpretation, revision and approval of the final version. OF acts as the guarantor of this study.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.