Sage Journals: Discover world-class research

Abstract

Background:

Switching between therapies is a recommended strategy for rheumatoid arthritis (RA) and psoriatic arthritis (PsA) patients who experience treatment failure; moreover, data on switching due to primary non-response and subsequent failures are limited.

Objectives:

To obtain information from clinical practice regarding failures due to primary non-response in patients on biologic and target synthetic disease-modifying antirheumatic drugs (ts/bDMARDs), assessing the incidence rate (IR) of switching due to primary non-response and risk of subsequent treatment failure by cycling compared to swapping.

Design:

A longitudinal retrospective study, spanning from 2007 to 2022, was conducted on patients with RA or PsA treated with ts/bDMARDs at an outpatient rheumatology clinic.

Methods:

The main outcomes were as follows: (1) ts/bDMARD failure due to primary non-response and (2) subsequent discontinuation of prescribed ts/bDMARD due to lack of efficacy. The independent variable was switching between classes compared to switching within class. As covariates, clinical, sociodemographic, clinical, and treatments were considered. To estimate ts/bDMARDs switching rates, survival techniques were used, expressing the IR per 100 patients * year with their 95% confidence interval. Cox multivariate regression analyses were run to assess the role of switching between/within class in the subsequent treatment failure.

Results:

In total, 327 patients were included. Of these, 50 patients in 77 treatment courses developed primary non-response with an IR of 4.25 (3.4–5.3). The IR was quite similar between RA and PsA, higher in women, and in those who started ts/bDMARDs after 2018. In those with primary non-response, there were 42 subsequent treatment failures with an IR of 26.38 (19.49–35.69). The multivariate model showed that cycling increased the risk of subsequent failure compared to swapping (hazard ratio: 2 (1.1–3.77), p = 0.03).

Conclusion:

This study provides support to consider swapping a better alternative rather than cycling after primary non-response.

Plain language summary

Initial lack of efficacy for biologic therapies in rheumatoid arthritis or psoriatic arthritis in daily clinical practice

Introduction:

The studied drugs are biological therapies that are arthritis-modifying drugs designed early in the last decade to prevent or reduce inflammation caused by the disease. This study focuses on initial lack of efficacy, in patients with rheumatoid arthritis or psoriatic arthritis who are treated with biologic therapies. We wanted to know how often these treatments fail to work right from the start and how it impacts future treatment choices in real-life practice.

Methods:

We included patients from 2007 to 2022 in which their consultant rheumatologist had decided to commence them on biologic therapy. We studied the changes due to initial lack of efficacy, we also included sociodemographic, clinical and treatments information.

Results:

The results showed that, of the 327 patients, 50 experienced primary initial lack of efficacy. This means about 4 out of every 100 patients had this treatment failure due, with a higher rate seen in women and those starting treatment after 2018. Among those who didn’t respond to their initial treatment, there were 42 further treatment failures, with changes within the same drug class (cycling) more often than changes between classes (swapping).

Conclusion:

This study provides support to consider swapping (changes between drug classes) a better alternative rather than cycling (changes within the same drug class) after initial lack of efficacy. We consider these finding useful for the management of patients. Implementation of more effective strategies may change the disease course.

Keywords

biologic and target synthetic disease-modifying antirheumatic drugs primary non-response psoriatic arthritis rheumatoid arthritis switching

Introduction

Rheumatoid arthritis (RA) and psoriatic arthritis (PsA) are chronic inflammatory joint diseases that can impair patients’ ability to work and perform daily activities, reducing quality of life.^1,2

Advances in early diagnosis, targeted treatment strategies, and the development of biologic and synthetic therapies (target synthetic disease-modifying antirheumatic drugs, ts/bDMARDs) have significantly improved their prognosis.^3,4

Biologic agents and small molecules target specific pathways involved in disease pathogenesis, aiming to achieve remission or low disease activity.^5–12

However, 30%–40% of patients still experience poor outcomes, adverse events, or lack of response, leading to treatment withdrawal.^8,13–16 This variability reflects differences in study designs, populations, healthcare practices, and systems. For these patients, successive biologic switching—whether within the same class (cycling strategy) or to a different mechanism of action (swapping strategy)—is advocated by all major international treatment guidelines.^8,11,12,17 Nevertheless, the evidence for directing switching from one agent to another is limited with most switches being based on clinical criteria and guidelines algorithms are not clear enough and can differ from one another.

Focusing on inefficacy, treatment response to ts/bDMARDs in RA or PsA can be classified as primary or secondary non-response, based on evidence of an initial response. Primary non-response is considered if the drug is ineffective, with no clinical response and secondary non-response would be considered if, after an initial response, the effectiveness is lost over time.^18,19 While primary non-response may be due to a mechanistic failure, secondary non-response may be driven by immunogenicity.²⁰ We believe that studying the type of non-response separately is a key requirement to improve the current knowledge on patient approach and management. Focusing on primary non-response, it seems reasonable to consider a different mode of action for the next treatment rather than switching within class, but studies addressing the best possible strategy in rheumatic diseases are controversial,^{14,16,21–27} and except for a few, mainly are related to tumor necrosis factor-alpha inhibitors (TNFi).^28,29

A recent study of 12 European registries found similar retention rates for first, second, and third TNFi treatments in ankylosing spondylitis, though rates were lower for second TNFi in cases of primary non-response compared to secondary non-response. Both second and third TNFi treatments were deemed potentially beneficial.²¹ Another real-world study in PsA showed low 3-year persistence with bDMARDs as second-line therapy after TNFi, with higher persistence rates for IL-17 and IL-12/23 inhibitors compared to TNFi, though reasons for discontinuation were not analyzed.³⁰

In RA, both cycling and swapping strategies for b/tsDMARDs are supported by clinical trials and real-world data, with no clear preference in international recommendations. A 2012 meta-analysis by Schoels et al.³¹ included one trial on cycling (GO-AFTER) and three on swapping (ATTAIN, REFLEX, and RADIATE), showing both strategies to be more effective than placebo. A 2021 meta-analysis suggested swapping might be superior to cycling after first-line TNFi failure, though it often did not distinguish between primary and secondary failures.^26,32 Trials focusing on primary non-responder support cycling as effective compared to placebo.^14,33,34 Recently, two head-to-head trials comparing TNFi to non-TNFi suggest swapping may be more beneficial than cycling to another TNFi for primary non-responders.^35,36

In all these studies, it is important to consider the heterogeneity in terms of design, study duration, inclusion criteria, treatment regimens, comparators, covariables, and definition of main outcomes (many of them do not specifically focus on primary non-responders but on inefficacy or even persistence as a whole). This makes comparing studies and concluding difficult. Most studies emphasize the failure of the first biologic treatment, especially TNFi, yet in real-world practice, patients often experience multiple therapeutic failures. Thus, it is crucial to examine switches across the follow-up period for all types of ts/bDMARDs. This study aims to analyze long-term data on treatment failures due to primary non-response in PsA and RA patients, comparing the risk of subsequent failure between cycling and swapping strategies.

Materials and methods

Setting, study design, and patients

The study was conducted at a public reference tertiary hospital, HCSC, in Madrid, Spain. It was an observational, retrospective longitudinal study involving patients with chronic inflammatory arthritis who attended the HCSC rheumatology outpatient clinic from January 2007 to April 2021. The study included all patients whose data were recorded in the hospital’s electronic health records and followed them until they were either lost to follow-up or until the study concluded on December 30, 2022.

In the present study, we selected patients with the following criteria: (a) 16 years of age or older; (b) had at least two medical claims with an International Classification of Diseases, Ninth or Tenth Revision, Clinical Modification diagnosis of PsA or RA, and fulfilled the PsA and RA diagnosis according to The CASPAR criteria,³⁷ ACR/EULAR 2010 classification criteria,³⁸ respectively; (c) PsA patients were included from January 2007 to April 2021 and followed up until December 2022. RA patients were included from January 2007 to December 2015 and followed up until December 2022; (d) had to be on ts/bDMARDs in the inclusion period.

Exclusion criteria: (a) Patients with a concomitant diagnosis of inflammatory bowel disease and (b) missing data related to the study’s primary outcomes.

The study was conducted in accordance with the Declaration of Helsinki and the principles of Good Clinical Practice and ethics review board approval was obtained. This research is executed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.³⁹

Data sources

Data were obtained from our EHR, which is used in all outpatient interactions between health professionals and patients.

Variables

Dependent variable

Two main outcomes were established: (a) switching between ts/bDMARDs due to primary non-response defined if the duration of the ts/bDMARDs course was <6 months. When the duration was longer (6–12 months), researchers (D.F-N., M.R-L., and L.A.) reviewed health records and classified patients as primary non-response according to their information and medical criteria. In the case of patients on rituximab, they were reviewed individually even though their treatment duration was longer than 12 months. (b) In those patients who achieved switching between ts/bDMARDs due to primary non-response, the main outcome was subsequent discontinuation of prescribed ts/bDMARD due to lack of efficacy.

The independent variable was switching between classes compared to switching within class. As a class, we included the following exposures: (a) Tumor necrosis factor-alpha inhibitors (TNFi); (b) Janus kinase inhibitors (JAKi); and (c) other biologics: Anti-IL6, Anti-CD20, CTLA4-Ig, Anti-IL17, Anti-IL12/23, and Anti-IL23.

The following covariates were considered including (1) demographic; (2) disease-related; (3) baseline comorbidities; (4) laboratory parameters at baseline: erythrocyte sedimentation rate (ESR), C-reactive protein, rheumatoid factor (RF), and HLAB27; and (5) treatment-related: current use of glucocorticoids, nonsteroidal anti-inflammatory drugs (NSAIDs), and concomitant conventional synthetic DMARDs (csDMARDs) during the whole follow-up of the study. (6) Calendar time: Dividing the starting time of each ts/bDMARDs into year intervals based on treatment strategies and the commercialization of anti-IL17 or JAKi (cutoff in January 2015 and January 2019).

Statistical analysis

A descriptive analysis was conducted. Frequency distribution was used for qualitative variables, mean and standard deviation, or median and percentiles for quantitative variables. Bivariate associations were studied using the Student’s t-test for normally distributed continuous variables and the Mann–Whitney or Kruskal–Wallis tests for non-normally distributed continuous variables. Categorical variables were analyzed with the Chi-square or Fisher test.

Switching rates between ts/bDMARDs due to primary nonresponse were estimated by including all PsA or RA patients, measuring from the baseline visit to loss of follow-up, the main outcome, or the end of the study. Discontinuation due to lack of efficacy was estimated for patients with primary nonresponse, measuring from the date of primary nonresponse to loss of follow-up, the main outcome, or the end of the study. Incidence rates (IR) of switching due to primary nonresponse and subsequent failure due to lack of efficacy were estimated by survival techniques (allowing for multiple failures per patient), expressing the IR per 100 patient-years with their respective 95% confidence interval (CI).

Cox bivariate analyses were done to assess the differences between demographic, clinical covariates, and the risk of subsequent failure. Cox multivariate regression analyses were run to assess the role of switching between/within class in the subsequent treatment failure. We included age, sex, calendar time, other known related factors, and all variables with a p < 0.2 in the bivariate analysis. The mode in which drug prescription was done in real-life conditions hampered the categorization of therapeutic options including csDMARDs, NSAIDs, and glucocorticoids, being analyzed in a time-dependent manner (at the beginning of each ts/bDMARDs). ESR was analyzed in a time-dependent manner as well. All these variables were divided into periods according to the retention rate of each ts/bDMARDs that determined the presence or not of an event in that time frame. In every period, if a patient was taking any csDMARD for more than 6 months was considered exposed. Glucocorticoids and NSAIDs were considered exposed if the patient was on this medication for at least 3 months during the period analyzed.

The results of the regression models were expressed by hazard ratio (HR) and 95% CI. We considered the influence of covariates if the prevalence was >10%, and if the frequency was lower (in comorbidities or several treatments), categories were grouped based on affinity.⁴⁰ The number of variables in the multivariate model followed the rule of Freeman.⁴¹ Missing values of more than 10% were not used in the multivariate analysis. The proportional hazard assumption was tested using Schoenfeld residuals and the scaled Schoenfeld residuals. All statistical analyses were performed using STATA software. A two-tailed p-value under 0.05 was considered to indicate statistical significance.

Results

A total of 327 patients were included (141 PsA and 186 RA patients), with a total follow-up of 1807.9 patients-year, a median of 5.1 (2.3–8.9) years, and a maximum follow-up of 16 years. Table 1 includes a detailed description of the baseline characteristics of the patients. As expected, individuals with RA were older and had a higher prevalence of females compared to PsA. The median lag time from diagnosis to first ts/bDMARD was 2.2 (0.45–5.26) years, being higher for RA compared to PsA.

Table 1.

Description of baseline (at first ts/bDMARD) demographic and clinical characteristics of patients.

Variable	Total (N = 327)	RA (N = 186)	PsA (N = 141)	p
Gender, female; n (%)	226 (69)	152 (81.7)	74 (52.48)	0.000
Age, years	52.7 ± 14.2	56.3 ± 13.9	48.0 ± 13.21	0.000
Smoking habit (n = 283); n (%)
Current	64 (22.6)	33 (18.6)	31 (30)	0.000
No-former	219 (77.4)	144 (81.4)	75 (70)
Lag time from symptoms to diagnosis; mean ± SD (years)	3.64 ± 5.65	–	3.64 ± 5.65	–
Lag time from diagnosis to first ts/bDMARD; mean ± SD (years)	3.39 ± 3.26	4.09 ± 3.13	2.32 ± 3.16	0.000
Comorbidity; n (%)
Dyslipidemia	135 (42.1)	82 (44.1)	53 (39.26)	0.4
Arterial hypertension	76 (23.7)	51 (27.4)	25 (18.52)	0.07
Diabetes mellitus	35 (10.9)	28 (15)	7 (5.19)	0.005
Obesity (n = 135)	18 (13.3)	–	18 (13.3)	–
Ischemic heart disease	7 (2.2)	5 (5.7)	2 (1.48)	0.4
Ischemic CV disease	4 (1.25)	0	4 (2.96)	–
COPD (n = 321)	19 (5.8)	16 (8.6)	3 (2.22)	0.001
Liver disease (n = 321)	13 (4.1)	7 (3.7)	6 (4.44)	0.7
Thyroid disease (n = 321)	53 (16.5)	37 (19.9)	16 (11.85)	0.055
Peptic ulcer disease (n = 321)	14 (4.36)	9 (4.8)	5 (3.70)	0.6
Cancer history (n = 321)	18 (5.6)	14 (7.5)	4 (2.96)	0.08
Depression (n = 321)	41 (14)	27 (14.5)	14 (10.37)	0.01
Psoriasis history (n = 135)	108 (80.6)	–	108 (80.60)	–
RF + (n = 279); n (%)ACPA + (n = 257); n (%)	126 (43.9)102 (39.7)	119 (64)101 (56)	7 (6.80)1 (1.3)	0.0000.000
CRP (n = 130); mean ± SD mg/dL	1.28 ± 2.8	–	1.28 ± 2.8	–
ESR (n = 313); mean ± SD mg/dL	22.2 ± 20.9	26.4 ± 22.8	16.0 ± 16.1	0.000
DAS 28; mean ± SD	4.93 ± 1.6	–	4.93 ± 1.6	–
Concomitant glucocorticoids; n (%)	209 (63.9)	156 (83.9)	53 (37.59)	0.000
Concomitant NSAIDs; n (%)	138 (42.2)	66 (35.5)	72 (51.1)	0.001
ts/bDMARDs; n (%)
TNFi	260 (79.5)	135 (72.5)	125 (88.65)	–
Adalimumab	105	42	63
Etanercept	77	52	25
Certolizumab	51	30	21
Golimumab	12	4	8
Infliximab	15	0	8
Abatacept	7 (2.1)	7 (3.8)	0
Rituximab	31 (9.5)	31 (16.7)	0
Anti-IL6	3 (0.9)	3 (1.6)	0
Tocilizumab	3	3	0
Anti-IL17	12 (3.7)	0	12 (8.51)
Secukinumab	11	0	11
Ixekizumab	1	0	1
Anti-IL12/IL23	1 (0.3)	0	1 (0.71)
Ustekinumab	1	0	1
JAKi	13 (4)	10 (5.4)	3 (2.12)
Tofacitinib	2	0	2
Baricitinib	8	7	1
Upadacitinib	3	3	0
Therapeutic regimen; n (%)
ts/bDMARDs in Monotherapy	163 (25.5)	39 (21)	38 (27)
ts/bDMARDs combined with 1				0.000
csDMARDs	334 (50.3)	81 (43.5)	89 (63.1)
ts/bDMARDs combined with 2
csDMARDs	167 (24.5)	64 (35.5)	14 (9.9)
Concomitant csDMARDs; n (%)	250 (76.45)	147 (79)	103 (73)
Methotrexate	177 (54.1)	95 (51.1)	82 (58.16)	0.2
Sulfasalazine	79 (24.2)	11 (5.9)	15 (10.64)	0.12
Leflunomide	26 (7.9)	65 (34.9)	14 (9.93)	0.000

ACPA, anti-citrullinated peptide antibodies; Anti-CD20, anti-CD20 monoclonal antibody; Anti-IL12/23, anti-interleukin 12/23; Anti-IL17, anti-interleukin 17; Anti-IL6, anti-interleukin 6; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; csDMARDs, conventional synthetic disease modifying anti-rheumatic drugs; CV, cardiovascular; ESR, erythrocyte sedimentation rate; JAKi, Janus kinase inhibitors; NSAIDs, non-steroidal anti-inflammatory drugs; PsA, psoriatic arthritis; RA, rheumatoid arthritis; RF, rheumatoid factor; SD, standard deviation; TNFi, TNF-alpha inhibitors; ts/bDMARDs, targeted synthetic and biologic disease-modifying anti-rheumatic drugs.

Focusing on therapy at the beginning of the study, 63.9% and 42.2% of the patients were on glucocorticoids and NSAIDs, respectively, with differences between RA and PsA. The most frequent drugs used as first ts/bDMARDs were TNFi in both groups (79.5%), followed by rituximab in RA and by Anti-IL17 in PsA. Concomitantly, 76.45% of the patients used csDMARDs, and the most frequent was methotrexate in both groups.

As shown in Figure 1, 327 patients received 609 courses of ts/bDMARDs therapy during the study period (405 TNFi, 50 RTX, 42 JAKi, 43 Anti-IL17, 33 Abatacept, 25 Anti-IL6, and 11 Anti-IL23), of which 190, 66, 37, 13, and 11 patients received one, two, three, four, and five courses ts/bDMARDs, respectively. Of them, 126 patients (38.5%) presented 263 switches between ts/bDMARDs over time (70 patients (37.6%) and 142 events in RA; 56 patients (39.7%) and 121 events in PsA). Regarding inefficacy (191 events), 50 of all patients developed 77 switches (12.6%) due to primary non-response (25 patients and 37 events in RA; 25 patients and 40 events in PsA).

Figure 1.

Therapy flow chart during the study period.

The estimated IR due to primary non-response was 4.25 (3.4–5.3) per 100 patients-year (Table 2). It was quite similar between RA and PsA and among different patient ages. However, it seemed to be lower in men compared to females and in those with negative RF. The calendar time effect seems to be important as well, it was quite low in the first two periods, increasing after 2018. Regarding the ts/bDMARDs, anti-TNF was one of the drugs with the smallest IR, together with rituximab, JAKi, or Anti-IL6. In relation to other therapeutic co-variables, IR was lower in those on concomitant csDMARDs or in those using NSAIDs, but higher in those on glucocorticoids.

Table 2.

IR of switching due to primary non-response.

Variable	Patient-years	Events	IR per 100	CI (95%)
Primary non-response	1807.9	77	4.25	3.4–5.32
Group
RA	914.9	37	4.04	2.93–5.58
PsA	893.0	40	4.47	3.28–6.10
Gender
Male	629.5	12	1.9	1.1–3.35
Female	1178.4	65	5.5	4.32–7.03
Age, years
<40	313.8	13	4.14	2.4–7.13
41–50	511.2	33	4.30	2.8–6.53
51–60	483.2	17	3.5	2.2–5.65
61–70	298.2	15	5.02	30.3–8.34
⩾71	201.5	10	4.96	2.67–9.22
Smoking habit
Yes	398.36	20	5.02	3.23–7.78
No-former	1155.9	50	4.32	3.27–5.71
Comorbidity
Yes	1014.1	37	3.64	2.64–5.0
No	753.1	40	5.31	3.89–7.24
Cardiovascular comorbidity
Yes	937.9	27	2.87	1.97–4.19
No	829.2	50	6.02	4.56–7.95
Calendar time, years
2001–2014	1017.4	31	3.04	2.14–4.33
2015–2018	571.9	16	2.79	1.71–4.5
⩾2019	218.6	30	13.72	9.59–19.62
RF
Positive	648.9	24	3.69	2.47–5.18
Negative	892.2	46	5.15	3.86–6.88
ESR
<40	1470.7	64	4.35	3.4–5.55
⩾40	240.4	11	4.57	2.53–8.26
Concomitant glucocorticoids
Yes	1117.2	58	5.19	4.01–6.71
No	690.7	19	2.75	1.75–4.31
Concomitant NSAIDs
Yes	921.5	29	3.14	2.18–4.52
No	886.4	48	5.41	4.1–7.18
ts/bDMARDs
Anti-TNF alpha	1408.1	51	3.6	2.75–4.76
Anti-IL17	69.6	10	14.37	7.73–26.7
Anti-IL23	10.3	3	29.17	9.14–90.46
Anti-CD20	148.4	5	3.36	1.4–8.09
Anti-IL6	57.5	2	3.47	0.87–13.91
JAKi	65.4	2	3.05	0.76–12.21
Abatacept	48.6	4	8.21	3.08–21.89
Combined csDMARDs
Yes	1435.5	55	3.83	2.94–4.99
No	372.4	22	5.9	2.88–8.97

Anti-CD20, anti-CD20 monoclonal antibody; Anti-IL17, anti-interleukin17; Anti-IL23, anti-interleukin23; Anti-IL6, anti-interleukin 6; CI, confidence interval; csDMARDs, conventional synthetic disease-modifying anti-rheumatic drugs; ESR, erythrocyte sedimentation rate; IR, incidence rate; JAKi, Janus kinase inhibitors; NSAIDs, non-steroidal anti-inflammatory drugs; PsA, psoriatic arthritis; RA, rheumatoid arthritis; RF, rheumatoid factor; TNFi, TNF-alpha inhibitors; ts/bDMARDs, targeted synthetic and biologic disease-modifying anti-rheumatic drugs.

In those patients with primary non-response (n = 50), there were 42 subsequent ts/bDMARD discontinuation due to lack of efficacy (18 events in RA and 24 events in PsA patients) with an IR of 26.38 (19.49–35.69). This IR was somewhat higher in females compared to males, larger in PsA compared to RA, over 60 years of age, in those that started on ts/bDMARDs after 2018, or in those patients without concomitant csDMARDs (Table 3).

Table 3.

Incidence rate of subsequent ts/bDMARD discontinuation due to lack of efficacy in patients with primary non-response (n = 50).

Variable	Patient-years	Events	IR per 100	CI (95%)
IR of subsequent ts/bDMARD discontinuation due to lack of efficacy	159.2	42	26.38	19.49–35.69
Group
RA	93.9	18	19.2	12.1–30.4
PsA	65.3	24	36.7	24.6–54.8
Gender
Male	28.61	7	24.4	11.6–51.3
Female	130.59	35	26.8	19.24–37.32
Age, years
<40	17.4	6	34.5	15.4–76.4
41–50	64.2	17	26.5	16.4–42.6
51–60	23.7	7	29.5	14.1–61.9
61–70	13.1	9	52.5	27.3–100.9
⩾71	36.8	3	8.1	2.6–25.26
Smoking habit
Yes	42.32	13	30.7	17.8–52.9
No-former	89.05	27	30.3	20.79–44.21
Comorbidity
Yes	91.9	18	19.6	12.3–31.1
No	67.3	24	35.7	23.9–53.2
Calendar time, years
2001–2014	86.3	18	20.8	13.1–33.1
2015–2018	48.8	9	18.44	9.6–35.4
⩾2019	24.1	15	62.3	37.5–103.4
RF
Positive	75.8	12	15.8	8.9–27.8
Negative	73.2	27	36.8	25.3–53.7
ESR
<40	143.9	37	25.7	18.6–35.5
⩾40	15.3	5	32.7	13.6–78.6
Concomitant glucocorticoids
Yes	132.1	34	25.7	18.38–36.01
No	27.1	8	29.5	14.7–59.1
Concomitant NSAIDs
Yes	71.5	19	26.2	16.9–41.6
No	87.7	23	26.2	17.4–39.5
Subsequent ts/bDMARDs
Same family (within class)	76.6	20	26.1	16.8–40.5
Different family (between class)	82.6	22	26.6	17.5–40.4
Combined csDMARDs
Yes	134.5	32	24.7	13.52–39.8
No	23.7	10	42.2	22.7–78.49

CI, confidence interval; csDMARDs, conventional synthetic disease-modifying anti-rheumatic drugs; ESR, erythrocyte sedimentation rate; NSAIDs, non-steroidal anti-inflammatory drugs; PsA, psoriatic arthritis; RA, rheumatoid arthritis; ts/bDMARDs, targeted synthetic and biologic disease-modifying anti-rheumatic drugs.

To evaluate differences between demographic, clinical covariates, and the risk of subsequent failure among patients who experienced primary non-response, we initially conducted bivariate analyses (Table 4). These analyses revealed no statistically significant differences between the principal variable and swapping versus cycling, nor other covariates. However, after adjusting for age, gender, disease group (RA vs PsA), and the year of ts/bDMARD initiation, cycling was associated with an increased risk of subsequent failure compared to swapping (HR: 2 (1.1–3.77), p = 0.03; Table 5). Gender and disease groups did not have a significant impact on the model and were dropped (p = 0.6 and p = 0.8, respectively). The final model also highlighted the impact of the year of prescription, with an increased risk observed in 2019 compared to earlier years. The proportionality of this regression model was confirmed with a p-value of 0.7.

Table 4.

Bivariate analysis: risk of subsequent failure in patients who have experienced primary non-response.

Variable	HR	CI (95%)	p
Gender, female	1.04	0.51–2.12	0.9
Age at first ts/bDMARDs, years	0.86	0.67–1.1	0.23
PsA vs RA	1.59	0.83–3.02	0.154
Smoker	1.03	0.6–1.77	0.9
Calendar time, years
2001–2014	1	–	–
2015–2018	0.8	0.43–1.49	0.49
⩾2019	1.9	0.88–4.34	0.126
Comorbidity	0.6	0.31–1.12	0.11
ERS >40	1.46	0.42–5.1	0.5
Concomitant glucocorticoids	0.88	0.41–1.87	0.7
Concomitant NSAIDs	1.18	0.66–2.15	0.56
Subsequent ts/bDMARDs: cycling vs swapping	1.15	0.63–2.1	0.6
Combined csDMARDs	0.84	0.5–1.3	0.3

CI, confidence interval; csDMARDs, conventional synthetic disease-modifying anti-rheumatic drugs; HR, hazard ratio; NSAIDs, non-steroidal anti-inflammatory drugs; PsA, psoriatic arthritis; RA, rheumatoid arthritis; ts/bDMARDs, targeted synthetic and biologic disease-modifying anti-rheumatic drugs.

Table 5.

Multivariate analysis: risk of subsequent failure in patients who have experienced primary non-response.

Variable	HR	CI (95%)	p
Age at 1st ts/bDMARDs, ⩾60	0.48	0.19–1.19	0.11
Calendar time, years, ⩾2019	2.81	1.19–6.63	0.01
Subsequent ts/bDMARDs: cycling vs swapping	2.01	1.07–3.77	0.03
Combined csDMARDs	0.72	0.45–1.14	0.16

CI, confidence interval; HR, hazard ratio; ts/bDMARDs, targeted synthetic and biologic disease-modifying anti-rheumatic drugs.

Discussion

This longitudinal real-world study examines primary non-response in chronic inflammatory joint diseases, enabling a direct comparison of switching strategies. The findings support swapping as a more effective approach than cycling after primary non-response, regardless of the ts/bDMARD type, treatment course, or other factors. In addition, the study provides descriptive insights, including the incidence of ts/bDMARD failures due to primary non-response in these patients. This study included patients with PsA and RA treated with all available ts/bDMARDs in a single center. Sociodemographic and clinical data were comparable to other national PsA and RA populations studied.^42,43

It is reasonable to consider that the underlying mechanisms for treatment failure may differ between primary or secondary non-response. But nonetheless, there are limited scientific publications distinguishing between them.¹⁶ In this study, primary non-response was defined as no improvement observed within 6 months, prompting a switch in ts/bDMARD therapy.¹⁹ To accurately identify all cases of switching due to primary non-response, and given the real-world nature of the data, treatments lasting longer than 6 months were individually reviewed, particularly for patients on rituximab and infliximab. Considering these factors, it is essential to account not only for the 6-month cutoff but also for the specific type of b/tsDMARD and the context from which the data were derived.

In accordance with previously published results,^14,16 around 40% of our patients required switching between ts/bDMARDs. This percentage was similar between RA and PsA. As in other studies, the main reason for switching was drug inefficacy, being somewhat less marked in RA compared to PsA, at the expense of more adverse events in RA, maybe because of sex or age differences, more comorbidities rather than the disease per se.

In relation to the total number of switches due to inefficacy, 40% were due to primary non-response, with an estimated IR of 4.25 per 100 patient-years, similar between disease groups or age, but higher in females, and differed between ts/bDMARDs. The role of sex has been reported previously in both diseases and other rheumatic diseases.^44,45 In this sense, physiological differences, but also differences in subjective assessment of patient-reported outcomes, could influence the response to pharmacotherapy in men and women.^46,47 Regarding the ts/bDMARDs, in other studies, the efficacy shown was similar between drugs,^22,48,49 but in our study, TNFi together with rituximab, JAKi, or Anti-IL6 had the lowest incidence of primary failure. However, the data are crude and linked to bias. It must be taken into consideration that this study has included patients with ts/bDMARD since 2007 when there were fewer therapeutic alternatives, conditioning the small patients-year in those more recent ts/bDMARDs, unable to provide robustness to these results.

In our study, the trend to switch between different mechanisms of action became notably more pronounced in 2018. This difference is shown in the crude incidences of primary non-response and subsequent failures, and it has been confirmed in the multivariate analysis. This change is also found in other publications^27,50 and can be attributed to the increased availability of treatment options and the full implementation of a Treat-to-Target approach in clinical practice.

Interestingly, in those patients with primary non-response, their subsequent failures due to lack of inefficacy increased by sixfold reflecting the challenges in managing treatment failures in chronic inflammatory arthritis.^19,51 Although the incidence of lack of inefficacy was higher in the PsA group compared to RA, in those with high ESR, or in those without concomitant csDMARDs, after adjustment in the multivariate analysis none of these variables achieved statistical significance.

The IR was quite similar between switching within and between classes; nevertheless, after the adjustment in the multivariate analysis, the cycling strategy increased independently the risk of subsequent failure compared to the swapping strategy. There has been scientific literature in RA^{14,16,22,23,26,29} and PsA²⁷ with controversial results. In more recent observational studies, data supporting the evidence of swapping strategy are increasing, and just two observational studies did not find differences in both strategies.^23,29 This fact is also demonstrated in two head-to-head clinical trials.^35,36

Bearing in mind our results, prolonged treatment with the same mechanism of action might not be recommended for primary non-responders. It might be associated with disease progression, clinical complications, and a reduced quality of life. In addition, it may incur unnecessary costs and expose patients to potentially harmful side effects or the development of blocking auto-antibodies.³²

Nowadays, switching between ts/bDMARDs and mainly TNFi due to primary non-response, albeit controversial, still occurs because of confidence and experience with these drugs. In addition, it is not yet clearly established in the guidelines,^5,11,12 although the trend seems to be downward. Besides, it cannot be forgotten that certain individual characteristics of the patients such as their preferences, their comorbidities, and/or extra-articular manifestations.^11,14,52

It is important to note that all TNFi target the same molecular and inflammatory pathways; thus, it is not surprising that most patients who are primary non-responders to their initial TNFi fail to achieve their treatment targets when cycled through alternative TNFi. The same reasoning may apply to other families such as Anti-IL17 or Anti-IL6. On the other hand, available JAKi differ in terms of affinity for the receptor-associated tyrosine kinases of the JAK family. Maybe in the case of JAKi, cycling to other JAKi after primary failure could be considered an alternative as shown in a recent study by Pombo-Suarez et al.²⁹; however, our study was not able to differentiate by types of JAKi due to the small number of courses. Even so, this group was too small to influence the overall result of this study.

The main limitations of this study are those that affect any retrospective observational study. On one hand, prescription bias cannot be completely ruled out. On the other, there was missing data; however, those included in the analysis have less than 10% missing data. Besides, given the extended follow-up, clinical practice has changed over time as more therapeutic options have become available. This reflects the small sample of the most recent ts/bDMARDs but also has conditioned different therapeutic management, prevailing the switch between TNFi mostly in the first years of the study. To alleviate this imbalance, we have included the calendar time variable in the analysis. Third, we did not report direct parameters of disease activity, but we used ESR as a subrogate variable. Finally, the number of patients and patient-years for the final model is somewhat modest, which might reduce the strength of the analysis. With this in mind, we still consider that these results add to the current knowledge on the real-world management of patients with chronic inflammatory arthritis.

The main strength of this study lies in its longitudinal design with a median follow-up of 5 years and a maximum follow-up of 16 years; the inclusion of a broad non-selected real-world patient spectrum and a wide variety of treatment options. Moreover, all the data were available for analysis, allowing adjustment for confounders to elude possible bias.

In summary, this study provides valuable data on the course of treatment of patients with PsA and RA, as well as on the switch pattern of ts/bDMARDs in the long term. This study can directly compare swapping versus cycling of ts/bDMARDs. We consider that our results add to current knowledge on the real-world management of chronic inflammatory arthritis including the treatment decision-making of these complex patients.

Conclusion

In our study, the primary non-response to ts/bDMARDs had an IR of 4.25 (3.4–5.3), being quite similar between RA and PsA, higher in women, and in those that started ts/bDMARDs after 2018. In those with primary non-response, there were 42 subsequent treatment failures with an IR of 26.38 (19.49–35.69). Cycling increases the risk of subsequent failure. This study provides support to consider swapping a better alternative rather than cycling after primary non-response.

Footnotes

Acknowledgements

None.

Declarations

ORCID iDs

Leticia Leon

Benjamín Fernández-Gutiérrez

Lydia Abasolo

References

Geijer

Alenius

André

, et al. Health-related quality of life in early psoriatic arthritis compared with early rheumatoid arthritis and a general population. Semin Arthritis Rheum 2021; 51(1): 246–252.

Smolen

Aletaha

Barton

, et al. Rheumatoid arthritis. Nat Rev Dis Primers 2018; 4(1): 18001.

Aletaha

Smolen

. Diagnosis and management of rheumatoid arthritis. JAMA 2018; 320(13): 1360.

Watanabe

Okano

Gon

, et al. Difficult-to-treat rheumatoid arthritis: current concept and unsolved problems. Front Med (Lausanne) 2022; 9: 1049875.

Ogdie

Coates

Gladman

. Treatment guidelines in psoriatic arthritis. Rheumatology (Oxford) 2021; 59: i37–i46.

Gossec

Baraliakos

Kerschbaumer

, et al. EULAR recommendations for the management of psoriatic arthritis with pharmacological therapies: 2019 update. Ann Rheum Dis 2020; 79(6): 700–712.

Smolen

Aletaha

Bijlsma

JWJ

, et al. Treating rheumatoid arthritis to target: recommendations of an international task force. Ann Rheum Dis 2010; 69(4): 631–637.

Smolen

Landewé

RBM

Bergstra

, et al. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2022 update. Ann Rheum Dis 2023; 82(1): 3–18.

Coates

Soriano

Corp

, et al. Group for research and assessment of psoriasis and psoriatic arthritis (GRAPPA): updated treatment recommendations for psoriatic arthritis 2021. Nat Rev Rheumatol 2022; 18(8): 465–479.

10.

Rombach

Tillett

Jadon

, et al. Treating to target in psoriatic arthritis: assessing real-world outcomes and optimising therapeutic strategy for adults with psoriatic arthritis—study protocol for the MONITOR-PsA study, a trials within cohorts study design. Trials 2021; 22(1): 185.

11.

Fraenkel

Bathon

England

, et al. 2021 American College of Rheumatology guideline for the treatment of rheumatoid arthritis. Arthritis Care Res (Hoboken) 2021; 73(7): 924–939.

12.

Tucker

Allen

Chandler

, et al. The 2022 British Society for Rheumatology guideline for the treatment of psoriatic arthritis with biologic and targeted synthetic DMARDs. Rheumatology (Oxford) 2022; 61(9): e255–e266.

13.

Rubbert-Roth

Szabó

Kedves

, et al. Failure of anti-TNF treatment in patients with rheumatoid arthritis: the pros and cons of the early use of alternative biological agents. Autoimmun Rev 2019; 18(12): 102398.

14.

Taylor

Matucci Cerinic

Alten

, et al. Managing inadequate response to initial anti-TNF therapy in rheumatoid arthritis: optimising treatment outcomes. Ther Adv Musculoskelet Dis 2022; 14: 1759720X221114101.

15.

Karlsson

Kristensen

Kapetanovic

, et al. Treatment response to a second or third TNF-inhibitor in RA: results from the South Swedish Arthritis Treatment Group Register. Rheumatology 2007; 47(4): 507–513.

16.

Johnson

Sanchez

Schoenbrunner

. Defining response to TNF-inhibitors in rheumatoid arthritis: the negative impact of anti-TNF cycling and the need for a personalized medicine approach to identify primary non-responders. Clin Rheumatol 2019; 38: 2967–2976.

17.

Leung

Korotaeva

Candia

, et al. Management of peripheral arthritis in patients with psoriatic arthritis: an updated literature review informing the 2021 GRAPPA treatment recommendations. J Rheumatol 2023; 50(1): 119–130.

18.

Keystone

Rampakakis

Movahedi

, et al. Toward defining primary and secondary nonresponse in rheumatoid arthritis patients treated with anti-TNF: results from the BioTRAC and OBRI Registries. J Rheumatol 2020; 47(4): 510–517.

19.

Vallejo-Yagüe

Keystone

Kandhasamy

, et al. Primary and secondary non-response: in need of operational definitions in observational studies. Ann Rheum Dis 2021; 80(8): 961–964.

20.

Jamnitski

Bartelds

Nurmohamed

, et al. The presence or absence of antibodies to infliximab or adalimumab determines the outcome of switching to etanercept. Ann Rheum Dis 2011; 70(2): 284–288.

21.

Linde

Ørnbjerg

Heegaard Brahe

, et al. Second and third TNF inhibitors in European patients with axial spondyloarthritis: effectiveness and impact of the reason for switching. Rheumatology 2024; 63(7): 1882–1892.

22.

Pappas

St John

Etzel

, et al. Comparative effectiveness of first-line tumour necrosis factor inhibitor versus non-tumour necrosis factor inhibitor biologics and targeted synthetic agents in patients with rheumatoid arthritis: results from a large US registry study. Ann Rheum Dis 2021; 80(1): 96–102.

23.

Todoerti

Favalli

Iannone

, et al. Switch or swap strategy in rheumatoid arthritis patients failing TNF inhibitors? Results of a modified Italian Expert Consensus. Rheumatology 2018; 57(Suppl. 7): vii42–vii53.

24.

Mease

Karki

Liu

, et al. Discontinuation and switching patterns of tumour necrosis factor inhibitors (TNFis) in TNFi-naive and TNFi-experienced patients with psoriatic arthritis: an observational study from the US-based Corrona registry. RMD Open 2019; 5(1): e000880.

25.

Merola

Lockshin

Mody

. Switching biologics in the treatment of psoriatic arthritis. Semin Arthritis Rheum 2017; 47: 29–37.

26.

Migliore

Pompilio

Integlia

, et al. Cycling of tumor necrosis factor inhibitors versus switching to different mechanism of action therapy in rheumatoid arthritis patients with inadequate response to tumor necrosis factor inhibitors: a Bayesian network meta-analysis. Ther Adv Musculoskelet Dis 2021; 13: 1759720X211002682.

27.

Costa

Perricone

Chimenti

, et al. Switching between biological treatments in psoriatic arthritis: a review of the evidence. Drugs R D 2017; 17(4): 509–522.

28.

D’Agostino

Carron

Gaillez

, et al. Effects of secukinumab on synovitis and enthesitis in patients with psoriatic arthritis: 52-week clinical and ultrasound results from the randomised, double-blind ULTIMATE trial with open label extension. Semin Arthritis Rheum 2023; 63: 152259.

29.

Pombo-Suarez

Sanchez-Piedra

Gómez-Reino

, et al. After JAK inhibitor failure: to cycle or to switch, that is the question—data from the JAK-pot collaboration of registries. Ann Rheum Dis 2023; 82(2): 175–181.

30.

Pina Vegas

Hoisnard

Bastard

, et al. Long-term persistence of second-line biologics in psoriatic arthritis patients with prior TNF inhibitor exposure: a nationwide cohort study from the French health insurance database (SNDS). RMD Open 2022; 8(2): e002681.

31.

Schoels

Aletaha

Smolen

, et al. Comparative effectiveness and safety of biological treatment options after tumour necrosis factor α inhibitor failure in rheumatoid arthritis: systematic review and indirect pairwise meta-analysis. Ann Rheum Dis 2012; 71(8): 1303–1308.

32.

Emery

Gottenberg

Rubbert-Roth

, et al. Rituximab versus an alternative TNF inhibitor in patients with rheumatoid arthritis who failed to respond to a single previous TNF inhibitor: SWITCH-RA, a global, observational, comparative effectiveness study. Ann Rheum Dis 2015; 74(6): 979–984.

33.

Smolen

Burmester

Combe

, et al. Head-to-head comparison of certolizumab pegol versus adalimumab in rheumatoid arthritis: 2-year efficacy and safety results from the randomised EXXELERATE study. Lancet 2016; 388(10061): 2763–2774.

34.

Bessette

Khraishi

Kivitz

, et al. Single-arm study of etanercept in adult patients with moderate to severe rheumatoid arthritis who failed adalimumab treatment. Rheumatol Ther 2017; 4(2): 391–404.

35.

Humby

Durez

Buch

, et al. Rituximab versus tocilizumab in anti-TNF inadequate responder patients with rheumatoid arthritis (R4RA): 16-week outcomes of a stratified, biopsy-driven, multicentre, open-label, phase 4 randomised controlled trial. Lancet 2021; 397(10271): 305–317.

36.

Gottenberg

Brocq

Perdriger

, et al. Non-TNF-targeted biologic vs a second anti-TNF drug to treat rheumatoid arthritis in patients with insufficient response to a first anti-TNF drug. JAMA 2016; 316(11): 1172–1180.

37.

Taylor

Gladman

Helliwell

, et al. Classification criteria for psoriatic arthritis: development of new criteria from a large international study. Arthritis Rheum 2006; 54(8): 2665–2673.

38.

Britsemmer

Ursum

Gerritsen

, et al. Validation of the 2010 ACR/EULAR classification criteria for rheumatoid arthritis: slight improvement over the 1987 ACR criteria. Ann Rheum Dis 2011; 70(8): 1468–1470.

39.

Von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Epidemiology 2007; 18(6): 800–804.

40.

Peduzzi

Concato

Feinstein

, et al. Importance of events per independent variable in proportional hazards regression analysis II. Accuracy and precision of regression estimates. J Clin Epidemiol 1995; 48(12): 1503–1510.

41.

Freeman

. Applied categorical data analysis. Statistics: textbooks and monographs. Marcel Dekker Inc, New York, Vols. 79 and 26, 1988, 737 p.

42.

Carmona

Gómez-Reino

; BIOBADASER Group. Survival of TNF antagonists in spondylarthritis is better than in rheumatoid arthritis. Data from the Spanish registry BIOBADASER. Arthritis Res Ther 2006; 8(3): R72.

43.

Glintborg

Østergaard

Krogh

, et al. Clinical response, drug survival, and predictors thereof among 548 patients with psoriatic arthritis who switched tumor necrosis factor α inhibitor therapy: results from the Danish Nationwide DANBIO Registry. Arthritis Rheum 2013; 65(5): 1213–1223.

44.

Intriago

Maldonado

Cárdenas

, et al. Clinical characteristics in patients with rheumatoid arthritis: differences between genders. ScientificWorldJournal 2019; 2019: 1–6.

45.

Lorenzin

Ortolan

Cozzi

, et al. Predictive factors for switching in patients with psoriatic arthritis undergoing anti-TNFα, anti-IL12/23, or anti-IL17 drugs: a 15-year monocentric real-life study. Clin Rheumatol 2021; 40(11): 4569–4580.

46.

Lorenzin

Cozzi

Scagnellato

, et al. Relationship between sex and clinical and imaging features of early axial spondyloarthritis: results from a 48 month follow-up (Italian arm of the SPondyloArthritis Caught Early (SPACE) study). Scand J Rheumatol 2023; 52(5): 519–529.

47.

Lubrano

Perrotta

Manara

, et al. The sex influence on response to tumor necrosis factor-α inhibitors and remission in axial spondyloarthritis. J Rheumatol 2018; 45(2): 195–201.

48.

Lauper

Iudici

Mongin

, et al. Effectiveness of TNF-inhibitors, abatacept, IL6-inhibitors and JAK-inhibitors in 31 846 patients with rheumatoid arthritis in 19 registers from the ‘JAK-pot’ collaboration. Ann Rheum Dis 2022; 81(10): 1358–1366.

49.

Wang

Huang

, et al. Comparison of the efficacy and safety indicators of DMARDs for rheumatoid arthritis. Medicine 2021; 100(29): e26524.

50.

Gómez-Reino

Rodríguez-Lozano

Campos-Fernández

, et al. Change in the discontinuation pattern of tumour necrosis factor antagonists in rheumatoid arthritis over 10 years: data from the Spanish registry BIOBADASER 2.0. Ann Rheum Dis 2012; 71(3): 382–385.

51.

Manica

Sepriano

Pimentel-Santos

, et al. Effectiveness of switching between TNF inhibitors in patients with axial spondyloarthritis: is the reason to switch relevant? Arthritis Res Ther 2020; 22(1): 195.

52.

Song

Cho

Kim

, et al. Factors associated with selection of targeted therapy in patients with rheumatoid arthritis. PLoS One 2023; 18(1): e0280234.

Primary non-response in inflammatory arthritis treated with biologics and targeted therapies in daily clinical practice

Abstract

Background:

Objectives:

Design:

Methods:

Results:

Conclusion:

Plain language summary

Introduction:

Methods:

Results:

Conclusion:

Keywords

Introduction

Materials and methods

Setting, study design, and patients

Data sources

Variables

Dependent variable

Statistical analysis

Results

Discussion

Conclusion

Footnotes

Acknowledgements

Declarations

ORCID iDs

References