Article Text
PDF
Abstract
Objectives To assess the potential impact of targeted therapies for psoriatic arthritis (PsA) on symptomatic treatments (non-steroidal anti-inflammatory drugs (NSAIDs), corticosteroids, opioid analgesics), methotrexate and mood disorder treatments and on hospitalisation and sick leave.
Methods Using the French health insurance database, this nationwide cohort study included adults with PsA who were new users (not in the year before the index date) of targeted therapies for ≥9 months during 2015–2021. Main endpoints were difference in proportion of users of associated treatments, hospitalisations and sick leaves between 3 and 9 months after and 6 months before targeted therapy initiation. Logistic regression models adjusted for sex, age, psoriasis, inflammatory bowel disease and Charlson Comorbidity Index compared the impact of biologics initiation (tumour necrosis factor inhibitor (TNFi)/interleukin 17 inhibitor (IL17i)/IL12/23i) on associated treatment discontinuation.
Results Among 9793 patients initiating targeted therapy for PsA (mean age: 51±13 years, 47% men), 62% initiated TNFi, 14% IL17i, 10% IL12/23i, 1% Janus kinase inhibitor, 12% phosphodiesterase-4 inhibitor. After treatment initiation, the proportion of treatment users was significantly reduced for NSAIDs (−15%), opioid analgesics (−9%), prednisone (−9%), methotrexate (−15%) and mood disorder treatments (−2%), along with decreased hospitalisations (−12%) and sick leaves (−4%). TNFi had a greater sparing effect on NSAIDs and prednisone use than IL17i (ORa=1.04, 95% CI=1.01 to 1.07; 1.04, 1.02 to 1.06) and IL12/23i (1.07, 1.04 to 1.10; 1.06, 1.04 to 1.09). Odds of methotrexate discontinuation was reduced with TNFi versus IL17i (0.96, 0.94 to 0.98) and IL12/23i (0.94, 0.92 to 0.97).
Conclusions Targeted therapy initiation for PsA reduced the use of associated treatment and healthcare, with TNFi having a slightly greater effect than IL17i and IL12/23i, except for methotrexate discontinuation.
- Antirheumatic Agents
- Arthritis, Psoriatic
- Epidemiology
Data availability statement
Data are available on reasonable request. All relevant data are reported in the article. Additional details can be provided by the corresponding author on 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
While targeted therapies are widely recognised for their efficacy in treating psoriatic arthritis (PsA), concerns over their safety profile and cost persist. Emerging data suggest a potential to reduce the need for other drugs, especially symptomatic ones, and certain costly cares. This potential beneficial impact still needs to be specifically studied.
WHAT THIS STUDY ADDS
We observed a significant sparing effect of targeted therapies on symptomatic treatments, and in particular on the use of non-steroidal anti-inflammatory drugs and prednisone (reduction in both the prevalence of users and the mean dosage), as well as lower rates of hospitalisations and sick leave. This effect was slightly more pronounced with tumour necrosis factor inhibitor than interleukin 17 inhibitor (IL17i) and IL12/23i.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
These results underscore the potential for optimising treatment strategies for PsA, suggesting that targeted therapies reduce the burden of associated treatments and healthcare utilisation, which can be particularly interesting given the potential for side effects and cost overruns.
Introduction
Psoriatic arthritis (PsA) is a complex chronic inflammatory rheumatic disease characterised by articular and periarticular involvement as well as extramusculoskeletal manifestations. This condition can be severe, potentially leading to irreversible joint damage and impaired quality of life.1 Over the past decades, the availability of biologic and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) has rapidly expanded, resulting in substantial advancements in the treatment landscape for PsA. Inhibitors of not only tumour necrosis factor (TNFi) but also interleukin 12/23 (IL12/23i), IL23i, IL17i and Janus kinase (JAKi) are now recommended for moderate-to-severe PsA when conventional synthetic DMARDs (csDMARDs) fail to adequately control disease or are not tolerated.2–5 In patients with mild disease and an inadequate response to at least one csDMARD, in whom neither a bDMARD nor a JAKi is appropriate, apremilast, a phosphodiesterase-4 inhibitor (PDE4i), may be considered.2
The efficacy of these therapies is widely acknowledged, but is often contrasted by their safety profile and cost.6 Nonetheless, some emerging evidence suggests that they may reduce the consumption of other drugs, particularly symptomatic treatments, which are frequently poorly tolerated. For instance, although non-steroidal anti-inflammatory drugs (NSAIDs) effectively alleviate pain and stiffness in PsA, their daily use can be problematic because of potential adverse effects on the digestive, cardiovascular and renal systems.7 Therefore, they should be used at the minimum effective dose and for the shortest possible duration.2 8 9 Similarly, systemic corticosteroid therapy has undeniable symptomatic efficacy but should be used cautiously to mitigate adverse effects and the risk of dermatosis rebound on abrupt discontinuation or dose reduction.2 10 Although analgesics are useful as adjunct therapy in chronic inflammatory rheumatism, their chronic use must be limited because of serious risks.11–13 Thus, as part of comprehensive patient management, a decrease or even discontinuation of those treatments can be considered as an objective in itself. Such a ‘sparing effect’ of targeted therapies seems to be common practice but has been little evaluated, and to our knowledge, no specific studies have been performed in PsA.14 The introduction of these second-line therapies also seems promising in reducing sick leaves and the frequency of certain costly care, such as hospitalisations.15 However, these potential benefits require validation in real-world settings. Furthermore, whether these effects are uniform across all targeted therapies or are preferentially observed with certain classes remains uncertain.
Thus, the main objective of this study was to assess the potential impact of targeted therapies in PsA on the consumption of symptomatic treatments (NSAIDs, opioid analgesics, prednisone, corticosteroids injections), methotrexate and mood disorder treatments as well as their influence on the need for hospitalisation or sick leave.
Methods
Data source and study design
This nationwide cohort study was based on analysis of the French national health insurance database (Système National des Données de Santé (SNDS)).16 The database contains individualised anonymous health data and covers 99% of the French population (>67 million people). As previously described,17 the SNDS contains exhaustive data on all reimbursements for health-related expenditures and outpatient medical and nursing cares prescribed/performed by healthcare professionals, together with sociodemographic data. The data on all pharmacy-dispensed medications include the date of prescription delivery, the type of formulation and the quantity delivered. The database also contains information on the date and nature of medical/paramedical interventions and information on patient eligibility for fully reimbursed care related to severe, costly chronic diseases, such as moderate-to-severe PsA, encoded according to the International Classification of Diseases, 10th Revision (ICD-10). It provides detailed medical information concerning all admissions to French hospitals, including the dates of hospital admission and discharge, the ICD-10 code at discharge, the medical procedures performed in the hospital and costly drugs (such as targeted therapies) administered in the hospital.16 18 This large database has been used for several pharmacoepidemiological studies.19–22
This study followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.
Study population and exposure definition
All adults (≥18 years old) with PsA who were registered in the SNDS from 1 January 2015 to 31 March 2021 were eligible for inclusion. Adults with PsA were identified by a specific ICD-10 code (M07, except M07.4 and M07.5) according to a previously published algorithm.23 Then, patients with at least one prescription for a targeted therapy were identified. Targeted therapies studied were adalimumab, certolizumab, etanercept, golimumab and infliximab as TNFi; ustekinumab as an IL12/23i; ixekizumab and secukinumab as IL17i; tofacitinib and upadacitinib as JAKi; and apremilast as a PDE4i. We did not evaluate abatacept, brodalumab or IL23i because they had not received marketing authorisation and/or reimbursement for PsA in France before the end of the study. Next, we selected previously targeted therapy-naive patients (ie, ‘new users’), defined as those who had not filled a prescription for one of these drugs for 1 year. Finally, we excluded patients who did not maintain treatment for at least 9 months, during which outcomes were assessed. We defined discontinuation of treatment as (1) a period of more than 60 days after the period of coverage (28 days for all targeted therapies, except infliximab (56 days) and ustekinumab (84 days)) by the last delivery of the molecule or (2) the date of delivery of a new targeted therapy in case of a therapeutic switch.
Outcomes
The primary endpoints were the difference between the proportion of patients (1) using symptomatic treatment, (2) using methotrexate, (3) using mood disorder treatments, (4) being hospitalised and (5) requiring a sick leave during the 3–9 months after initiation of the first line of targeted therapy (ie, the ‘index date’) as compared with the 6 months before initiation of the first line of targeted therapy. Thus, we considered a neutral period’ of 3 months after the initiation of the specific molecule. The symptomatic treatments studied were NSAIDs, opioid analgesics (weak and strong), prednisone and corticosteroids injections. Hospitalisations (excluding scheduled hospitalisations, lasting <24 hours) included those (1) for all causes, (2) specifically in a rheumatology department and (3) for PsA relapse (defined as a hospitalisation with an ICD-10 for PsA as the main diagnosis). We assessed all sick leaves, excluding maternity and paternity leaves. Mood disorder treatments included antidepressants and anxiolytics.
The difference in consumption of symptomatic treatments and methotrexate before and after targeted therapy initiation was estimated for each patient. Specifically, we calculated the following. (1) The difference in Assessment of SpondyloArthritis international Society-NSAID (ASAS-NSAID) scores, which reflects overall NSAID consumption, considering the type of NSAID, the average daily dose and the proportion of days with intake over the period considered.24 We also report the proportion of patients achieving a 50% reduction in ASAS-NSAID Score after targeted therapy initiation and the proportion attaining an ASAS-NSAID Score≤10 after targeted therapy initiation among those with a score>10 before therapy initiation. (2) The difference between total doses of weak/strong opioid analgesics (calculated using the dosage of each tablet, number of tablets per box and number of boxes dispensed), after conversion of doses to an oral morphine equivalent.25 26 (3) The difference in total prednisone doses and average daily dose between the two periods. (4) The difference in total methotrexate doses and average weekly doses between the two periods (using similar methods as above). Additionally, we evaluated the time to methotrexate discontinuation after targeted therapy initiation among patients with previous methotrexate treatment. The period of methotrexate exposure was estimated with a methodology similar to that used for defining exposure to targeted therapies.
Similarly, we determined the difference in proportion of hospitalisations and sick leaves between the two periods and reported the most frequent causes of hospitalisation (after grouping ICD-10 codes for main diagnoses by specialty).
Covariates
We collected basic demographics, including age, sex, French Deprivation Index (a geographic indicator of social disadvantage specifically adapted for population health studies in France)17 and complementary universal health coverage. We identified inflammatory diseases associated with PsA (active skin psoriasis, defined by at least four deliveries of topical vitamin D derivatives or topical corticosteroids in the 2 years before the index date, and inflammatory bowel disease (IBD)) and variables used to calculate the Charlson Comorbidity Index.27 These covariates are defined in online supplemental table S1. Furthermore, we also collected data on treatments of interest other than targeted therapies (csDMARDs, NSAIDs or prednisone) at the index date and during the 2 years before the index date. The combination of an add-on therapy and a targeted therapy at baseline was defined as a 30-day period between reimbursements for the two treatments. Vital status was also recorded during follow-up.
Supplemental material
Statistical analysis
Patient characteristics at baseline are described overall and for each targeted therapy class. Categorical variables are reported with frequencies and quantitative variables with mean and SD. Percentages before and after targeted therapy initiation were compared with the McNemar test and mean score/doses with a paired Student’s t-test for quantitative variables. Given the multiplicity of tests performed, we applied a Bonferroni correction for these analyses: p<0.008 (=0.05/6 corresponding to the five therapeutic classes plus the overall cohort) was considered significant. Kaplan-Meier survival curves were used to model the time to discontinuation of methotrexate in patients with previous treatment.
We then compared the effect of first-line bDMARDs (TNFi/IL17i/IL12/23i) on the possibility of discontinuing symptomatic treatment or methotrexate and limiting the number of hospitalisations or sick leaves by using logistic regression models adjusted for age, sex, active psoriasis, IBD and Charlson Comorbidity Index to estimate the adjusted OR (ORa) and 95% CIs. We did not include apremilast, given its different place in the recommendations as compared with other treatments, and JAKi, the most recent molecules introduced, with a limited number of patients in the cohort, in the subsequent analyses.2 We performed predefined subgroup analyses among patients with and without active psoriasis and among women and men. To assess the robustness of our results, we conducted the following sensitivity analyses: (1) modifying the ‘new user’ definition to include patients who had not filled a prescription for a targeted therapy for 5 years before the index date; (2) defining treatment discontinuation as >90 days without filling a prescription for the same treatment after the period covered by the previous prescription; and (3) extending the ‘neutral period’ to 6 months.
Results
Description of the cohort population
During the study period, we identified 15 889 patients initiating targeted therapy for PsA. After excluding 6096 patients who had not maintained treatment for at least 9 months, we finally included 9793 patients in the cohort (mean age 51±13 years, 47% men, 37% with active skin psoriasis): 8537 in the bDMARDs cohort, including 6107 (62%) initiating TNFi, 1408 (14%) IL17i, and 1022 (10%) IL12/23i, and 1256 in the tsDMARDs cohort including 99 (1%) initiating JAKi and 1157 (12%) PDE4i (table 1 and figure 1). Details by molecule are given in online supplemental table S2.
Flow chart of the patients included in the analysis. bDMARDs, biologic disease-modifying antirheumatic drugs; IL, interleukin; JAKi, Janus kinase inhibitor; PDE4i, phosphodiesterase-4 inhibitor; PsA, psoriatic arthritis; TNFi, tumour necrosis factor inhibitors; tsDMARDs, targeted synthetic disease-modifying antirheumatic drugs.
Characteristics of the included patients
At the time of targeted therapy initiation, 4200 (43%) patients had a coprescription of a csDMARD, and 3300 (34%) and 1524 (16%) had a coprescription of an NSAID or prednisone, respectively. Characteristics of the overall cohort and by therapeutic class are presented in table 1. The proportion of patients with active skin psoriasis was higher in the IL12/23i group (55%) than JAKi group (14%). Additionally, patients initiating a JAKi or PDE4i were older than those initiating a bDMARD (mean 58 and 57 vs 49–51 years) and had a higher prevalence of comorbidities (50% and 43% vs 37%–39% with Charlson Comorbidity Index≥1).
Use of associated treatments and healthcare resources before/after initiation of targeted therapy
After targeted therapy initiation, the proportion of users significantly decreased for NSAIDs (−15%; decrease in mean ASAS-NSAID Score among users: 25 vs 14, p<10−4), weak (−9%) and strong (−1%) opioid analgesics (p<10−4), prednisone (−9%; reduction in mean daily dose among users: 11 vs 5 mg/day, p<10−4), corticosteroids injections (−1%, p<10−3) and methotrexate (−15%; decrease in mean weekly dose among users=12 vs 9 mg, p<10−4) (table 2). However, we found variations in the magnitude of effect among therapeutic classes. The reduction in NSAID users after targeted therapy initiation ranged from −18% for TNFi to −6% for PDE4i new users, with a corresponding mean reduction in ASAS-NSAID Score of 50% and 13%, respectively (for detailed effects of targeted therapy initiation on ASAS-NSAID Score, see online supplemental table S3. Similarly, the reduction in opioid analgesic users varied: −11% for TNFi to −2% for PDE4i new users. Moreover, we found a more substantial decrease in prednisone consumption after TNFi initiation as compared with other therapeutic classes (other bDMARDs and tsDMARDs), both in number of users (−11% vs −4% for IL12/23i to −6% for IL17i) and average daily dose (−56% vs −29% to −47%, respectively). Conversely, methotrexate discontinuation and dose reduction among users were more frequent after the initiation of IL12/23i (−22%; mean dose reduction: 45%) than TNFi (−12% and −20%, respectively). The time to discontinuation of methotrexate after the initiation of each targeted therapy class is provided in online supplemental figure S1.
Use of symptomatic treatments and methotrexate by targeted therapy initiation (before and after initiation)
Additionally, we found a decrease in hospital admissions (−12%), sick leave (−4%) and users of mood disorder treatments (−2%) after targeted therapy initiation (table 3). Further details on hospitalisations specifically in rheumatology departments and related to PsA relapse are presented in online supplemental table S4.
Care consumption (hospitalisations, stop working and consumption of treatments for mood disorders) by targeted therapy initiation (before and after initiation)
Possibility of discontinuing symptomatic treatment/methotrexate and limiting the number of hospitalisations/sick leaves by bDMARD class initiation
The sparing effect of NSAIDs and prednisone was slightly more pronounced after the initiation of TNFi than IL17i (ORa=1.04, 95% CI=1.01 to 1.07 and 1.04, 1.02 to 1.06) and IL12/23i (1.07, 1.04 to 1.10; 1.06, 1.04 to 1.09) (table 4). Conversely, the odds of methotrexate discontinuation were reduced after the initiation of TNFi versus IL17i (0.96, 0.94 to 0.98) and IL12/23i (0.94, 0.92 to 0.97). The odds of discontinuation of weak opioid analgesics were marginally increased after the initiation of TNFi (1.07, 1.04 to 1.10) and IL17i (1.05, 1.02 to 1.09) versus IL12/23i. We found no statistically significant differences in other outcomes after initiation of IL17i versus IL12/23i, except for discontinuation of mood disorder treatments, which was slightly less frequent with IL17i (0.97, 0.96 to 0.99).
Odds of discontinuing symptomatic treatments, methotrexate and mood disorder treatments and limiting the number of hospitalisations and sick leaves after initiation of a biologic disease-modifying antirheumatic drug
Increasing age was significantly associated with a reduction in use of associated treatments. Further factors associated with the discontinuation of each coprescription are presented in online supplemental table S5. Subgroup analyses of patients with and without psoriasis (online supplemental table S6) and women and men (online supplemental table S7) yielded results similar to those of the main analysis. Sensitivity analyses supported the robustness of our findings (online supplemental table S8).
Discussion
In this nationwide study involving 9793 new users of targeted therapies for PsA, we investigated the impact of each therapeutic class on the use of symptomatic treatments, methotrexate and mood disorder treatments as well as the need for hospitalisation or sick leave. After the introduction of targeted therapy, we found a significant overall decrease in the use of NSAIDs, opioid analgesics, corticosteroids, methotrexate and mood disorder treatments as well as hospitalisations and sick leaves. However, the magnitude of these effects varied among certain therapeutic classes. The odds of discontinuation of NSAIDs and prednisone were slightly increased with TNFi versus IL17i or IL12/23i initiation. Conversely, the odds of methotrexate discontinuation were increased with ILi versus TNFi initiation. Finally, reduction in the use of associated treatments was more pronounced with advancing age.
Our study is important because it is the first to investigate the sparing effect of the different classes of targeted therapies specifically in PsA. Previous research has predominantly focused on demonstrating the effect of TNFi on NSAIDs consumption in axial spondyloarthritis (axSpA). Consistent with our findings, NSAID consumption was reduced with TNFi treatment in both dedicated clinical trials and analyses of axSpA cohorts.14 28 29 Of note, the baseline ASAS-NSAID Score was higher in these studies (median 55 in the DESIR (DEvenir des Spondylarthropathies Indifférenciées Récentes) cohort; mean 98 in the SPARSE trial (NSAIDs sparing effect of etanercept in axial spondyloarthritis)) than in our study (mean 27). This difference can be attributed in part to disparities in study populations (studies of axSpA populations, involving younger patients with a known propensity for increased NSAID consumption).30 However, our results remained consistent with previously documented trends, such as the 50% reduction in ASAS-NSAID Score observed in 57%–67% of patients receiving TNFi (69% in our study) and the achievement of an NSAID Score≤10 observed in 46%–58% of cases (62% in our study). These effects have also been reported from studies of secukinumab, an IL17i, and IL12/23i in axSpA.31 32
Targeted therapies used in PsA also demonstrated a significant sparing effect on other treatments, including analgesics, corticosteroids and methotrexate. Of note, the introduction of targeted therapies led to a substantial reduction in corticosteroids use, with an average decrease of 6 mg/day over 6 months post initiation. Furthermore, hospital admissions and sick leaves declined in general after targeted therapy initiation. Although the extent of the reduction varies across studies (in particular because of differences in study populations and assessment time points) and treatment modalities, our findings align with those from other studies of chronic inflammatory diseases33–36 and agree with those observed in daily practice. As expected, the maintenance of methotrexate in association with TNFi was more prevalent and continued over a longer duration than with the other molecules, which reflects common practice in rheumatology aimed at preventing the development of antidrug antibodies particularly against infliximab and adalimumab.37 In addition, it is important to note that tofacitinib, unlike upadacitinib (both JAKi), requires coprescription with methotrexate as part of its European marketing authorisation. This requirement may affect the proportion of patients using methotrexate at initiation of treatment and the number of patients discontinuing it after starting tofacitinib. These sparing effects hold significant promise, both from the perspective of the patient and society because they offer the prospect of enhanced tolerance, with fewer adverse events, and a potential long-term reduction in healthcare costs.
TNFi had a slightly higher sparing effect in PsA than other classes of biologics, particularly concerning the use of NSAIDs and prednisone, with no difference between ILi agents. Although direct comparisons between molecules are lacking, the literature suggests that, in line with our results, sparing effects may be slightly less pronounced with other therapies than with TNFi.31 32 36 Of note, IL17i seemed to facilitate more frequent discontinuation of weak opioid analgesics as compared with IL12/23i.38 Neutralisation of IL17 may lead to a reduction in hyperalgesia and somatic signs induced by opioids discontinuation. Although several studies suggest a pivotal role for TNF and IL agents in the pathogenesis and treatment of depression, we noted a significant reduction in the use of mood disorder treatments in our study within 6 months of targeted therapy initiation.39 This effect seemed more pronounced with IL12/23i agents and among patients with active skin psoriasis, as shown in some studies,40–42 which supports the hypothesis of a major psychological impact of psoriasis in certain patients with PsA. However, it should be borne in mind that the overall magnitude of the effect of biologics on some studied outcomes, such as the treatment of mood disorders (reduction of around 2%), remains limited during this 6 month observation period and would require longer-term observation. Additionally, age seemed to play a role in the discontinuation of coprescribed drugs. This finding could be attributed to a heightened concern regarding adverse events associated with these prescriptions in older patients, for whom targeted therapies have been found to have comparable efficacy and safety profiles as in a younger population in PsA.43 44
The limitations of this study include the lack of availability of certain data, particularly concerning disease activity/severity and phenotype. Although we used ‘proxies’, such as therapies of interest for PsA in the 2 years preceding the index date, to approximate these parameters and reduce confounding bias, some residual bias may still remain. Moreover, because of the different profiles of treated patients, especially between bDMARDs and tsDMARDs, and the low number of patients exposed to certain first-line therapeutic classes (especially JAKi), direct comparison of the effect of these targeted therapies on healthcare consumption must be considered with caution. However, we adjusted our analyses for several confounders to accurately estimate the differential impact of bDMARD classes. We defined drug exposure based on healthcare reimbursement data, which are not necessarily equivalent to days of use, and the drugs within each class were not separated on analysis. The definition of PsA population was based on either ICD-10 diagnostic codes for PsA (M07 except M07.4 and M07.5, which correspond to arthropathy in Crohn’s disease and ulcerative colitis, respectively) applied to in-patients or out-patients with fully reimbursed PsA-related care procedures.23 We did not exclude patients with ICD-10 code M07.6. However, within the total population included, we found only 6 (<0.1%) patients identified solely via an ICD-10 code M07.6. As this study was based on real-life data, variations in sample size across different treatment groups were to be expected. Indeed, due to the greater use experience, the majority of patients initiate a TNFi. Although unequal sample sizes may affect statistical power and precision under certain conditions (smaller groups may reduce power, and larger groups may dominate comparative analyses), the use of logistic regressions to compare effects of biologics generally allows these imbalances to be managed effectively. Moreover, this risk is considerably reduced when sample sizes exceed 500, which was the case in our study where the smallest group (IL12/23i) included 1022 patients.45 46 Finally, the sparing effect emphasised in this study was demonstrated only in the short-term/medium-term, and its long-term validity still requires confirmation. Previous findings indicated a moderate persistence of first-line targeted therapies, decreasing from 73% at 1 year to 36% at 3 years, which raises concerns about a possible reduction in effect over time.47 Nevertheless, even in the short-term/medium-term, the effect remains of significant interest both for the patient and society.
This study has several strengths. Our cohort included a large number of patients in real world settings from a national exhaustive database providing health insurance data with a quality and consistency plan ensuring homogeneous data processing.16 This framework minimises selection bias. To mitigate channelling bias (ie, a result of confusion in assessing certain treatments in specific subgroups) and confounding by indication bias, we focused exclusively our analyses on patients who were naïve to targeted therapy. It is important to note that the targeted therapies studied are all recommended treatments for moderate-to-severe PsA, and that in France, each physician is free to choose the treatment labelled for PsA.3 Except in a minority of cases where an extramusculoskeletal manifestation (very active psoriasis, IBD, severe or repeated acute anterior uveitis) guides the prescriber’s choice, no factor is today likely to influence this prescription at the population level. In addition, outcomes were based on prospectively collected data, thus eliminating the risk of information bias, and involved standardised and validated tools such as the ASAS-NSAID Score, developed to evaluate the magnitude of NSAID intake and the NSAID-sparing effect of treatments. At last, to test the robustness of the results, we performed sensitivity analyses, which supported the integrity of our findings.
Conclusion
First-line targeted therapy for PsA resulted in a significant sparing effect for symptomatic treatments and methotrexate, leading to reductions in both the prevalence of users and the mean dosage, along with decreased rates of hospitalisations and sick leave. This effect seemed slightly more pronounced with TNFi than IL17i and IL12/23i, except for methotrexate, with odds of discontinuation greater with ILi agents. These findings highlight the potential for optimising PsA treatment strategies, suggesting that targeted therapies also reduce the burden of associated treatments and healthcare utilisation, which can be particularly of interest given the potential for side effects and cost overruns.
Data availability statement
Data are available on reasonable request. All relevant data are reported in the article. Additional details can be provided by the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
Specific approval was obtained to conduct this study from the French data protection agency (Commission nationale de l’informatique et des libertés: MLD/MFI/AR2010413), and patients were collectively informed about the use of pseudonymised data.
Acknowledgments
We thank Mrs. Laura Smales for English editing.
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
X @LauraPnVg
Contributors Study idea and concept: LPV, ES and PC. Data collection: SI and LPV. Data analysis: SI and LPV. LPV drafted the article and acts as guarantor of the study. All authors interpreted the results, critically revised the manuscript and approved the final version.
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 LPV received a subsidy from Novartis to attend a congress. SI and ES have no conflict of interest to declare. PC has received consulting fees from AbbVie, Amgen, Biogen, Celltrion, Galapagos, Janssen, Lilly, MSD, Novartis, Pfizer and UCB (less than US$10 000 each) and has been an investigator for Abbvie, Janssen, Lilly, MSD, Novartis and Pfizer.
Patient and public involvement statement Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
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.