Introduction

Chronic kidney disease (CKD) is notably more prevalent in patients with rheumatoid arthritis (RA) than in the general population.1 This heightened prevalence results from a multitude of factors, such as RA-associated AA amyloidosis,2 kidney damage due to the use of non-steroidal anti-inflammatory drugs (NSAIDs) for joint pain management3 and chronic inflammation in RA that promotes atherosclerosis4 and anaemia.5

However, the management of RA in patients with CKD presents unique challenges. Among patients with CKD, the incidence of infection is significantly higher.6 The use of methotrexate (MTX), an anchor drug, is restricted due to CKD.7 On the other hand, inadequate management of RA promotes deterioration of kidney function.8 The use of NSAIDs requires special attention in terms of NSAID-induced kidney damage. In this context, we hypothesised that biological disease-modifying antirheumatic drugs (bDMARDs) could be potential promising cornerstones for RA management in patients with CKD.

Nonetheless, the efficacy and safety of bDMARDs in patients with RA with CKD, particularly in those undergoing haemodialysis (HD), have yet to be adequately established. In patients with RA with CKD, several case reports and case series involving limited patient populations have demonstrated the safety and efficacy of single bDMARDs.9–14 However, to our knowledge, no studies have explored the efficacy and safety of comprehensive modalities of bDMARDs in patients with RA with CKD through extended follow-up periods and among a certain number of patients.

The drug retention rate is regarded as a valuable indicator of drug efficacy and safety in observational studies since the most common reasons for cessation of bDMARDs have been reported to be lack/loss of effectiveness and adverse events.15 The primary objective of this study was to investigate the 36-month drug retention rates of bDMARDs to validate their efficacy and safety in a real-world setting among patients with RA with CKD.

MethodsStudy design and population

In order to validate our hypothesis that bDMARDs effectively and safely manage RA in patients with CKD, a retrospective cohort study was planned to evaluate the efficacy and safety of bDMARDs in patients with RA with CKD.

We included patients with RA who had been prescribed their first-line bDMARDs at Toranomon Hospital (Tokyo, Japan) or Toranomon Hospital Kajigaya (Kanagawa, Japan) between 1 January 2004 and 31 December 2021. Since we adopt the retention rate of bDMARDs as the overall indicator of the efficacy and safety, the analysis was limited to first-line bDMARDs, excluding second-line and later bDMARDs, which generally have a lower retention rate. The diagnosis of RA was made according to the 2010 American College of Rheumatology/European Alliance of Associations for Rheumatology classification criteria.16 Patients were excluded if they lacked data on age, sex, initiation and discontinuation dates, or reasons for discontinuation of bDMARDs. The majority of the study population, mainly living in the Tokyo metropolitan area, was under the care of each hospital or its satellite clinics every 1–3 months.

Patients were followed up with from the date of first bDMARDs prescription until the earliest date of (1) discontinuation of the first bDMARDs; (2) all-cause death or (3) censoring (censoring for loss to follow-up or administrative censoring occurring on the end of December 2021).

Data collection and variables

Clinical data were retrieved from medical records and validated via patient interviews using the interview form (online supplemental material 1). The obtained data included baseline patient demographics (sex, height, body weight, age at diagnosis, disease duration, comorbidities, symptoms, HD), blood examination findings (creatinine (Cr), estimated glomerular filtration rate (eGFR), anti-cyclic citrullinated peptide antibody, C reactive protein (CRP), erythrocyte sedimentation rate (ESR), matrix metalloproteinase 3 (MMP3), rheumatoid factor) and medication at the time of the initiation of bDMARDs. eGFR was calculated by the CKD Epidemiology Collaboration equation modified by a Japanese coefficient.17 For patients with CKD (eGFR<60 mL/min/1.73 m2), the cause of CKD was also searched. Furthermore, disease activity of RA based on Disease Activity Score 28 (DAS28)-CRP and DAS28-ESR, and the dosage of prednisolone (PSL) were assessed over a period of 6 months following the initiation of bDMARDs.

The outcomes

The primary outcome was the 36-month drug retention rate of first-line bDMARDs. The drug retention rate was defined as the duration from the initiation to the discontinuation of a specific bDMARD modality.

The secondary outcomes included the trends in the DAS28-CRP, DAS28-ESR and PSL dose, and the proportion of reasons for discontinuation. The reasons for discontinuation were classified into one of the following four categories by the attending rheumatologists: (1) ineffectiveness; (2) infection; (3) adverse effects and (4) others. Adverse effects refer to adverse events other than infections.

Statistical analysis

All patients were categorised on the basis of kidney function (eGFR of ≥60, 30–60, <30 mL/min/1.73 m2) and the modality of bDMARDs (tumour necrosis factor α inhibitors (TNFαis), interleukin-6 inhibitors (IL-6is) and cytotoxic T-lymphocyte antigen-4 immunoglobulin (CTLA4-Ig)). Further subanalysis involved dividing patients with an eGFR of <30 mL/min/1.73 m2 into HD (non-HD or HD) groups, as well as dividing them by the modality of bDMARDs, for analysis.

First, baseline characteristics were quantified using medians and interquartile ranges for continuous variables and percentages for categorical variables. Then, these characteristics were compared across groups, with p values calculated by the Wilcoxon test, Kruskal-Wallis test and Cuzick’s test for trends in continuous variables and χ2 test for categorical variables.

Second, for the evaluation of overall effectiveness and safety, we compared the 36-month drug retention rates of bDMARDs across the above-mentioned categories. For the analysis of the drug retention rate, survival curves were generated via Kaplan-Meier analysis. A Cox proportional hazards regression model was stratified according to the above-mentioned categories and that was adjusted for the possible confounders to estimate the HR and 95% confidence intervals (CIs) for the discontinuation of first-line bDMARDs. Prespecified possible confounders included age, CRP level, PSL dose and MTX dose, which may influence drug retention rate according to the previous reports.18 19

Third, we compared the trends in the DAS28-CRP, DAS28-ESR and PSL dose over a period of 6 months following the initiation of bDMARDs.

Fourth, we investigated whether the proportion of reasons for discontinuation within 36 months varied according to the kidney function.

Furthermore, as a subanalysis exclusively focusing on patients with an eGFR of <30 mL/min/1.73 m2, we compared and examined the 36-month drug retention rates of first-line bDMARDs between the non-HD and HD groups, as well as among different modalities of bDMARDs.

All statistical analyses were performed using STATA software version 16.1 (StataCorp LLC, College Stone, TX, USA).

ResultsBaseline characteristics

Of the 437 patients initially included, 12 were excluded due to incomplete data, resulting in a total of 425 patients being eligible for analysis. The study flow and selection of the study population are summarised in online supplemental material 2. Median observation period was 40.8 weeks. No patient died during the observation period. Table 1 presents the baseline characteristics of the patients. The distributions of CKD stages were as follows: G1, 165 patients; G2, 140 patients; G3a, 36 patients; G3b, 14 patients; G4, 27 patients and G5, 43 patients. The first-line bDMARDs used were TNFαis in 347 patients (infliximab in 112 patients, etanercept in 98 patients, certolizumab in 65 patients, golimumab in 45 patients and adalimumab in 27 patients), IL-6is in 36 patients (tocilizumab in 34 patients and sarilumab in 2 patients) and CTLA4-Ig (abatacept) in 42 patients. The baseline eGFR was generally lower in the IL-6i group.

Table 1

Baseline characteristics

Among patients with CKD, the most common clinical or pathological diagnosis of the primary kidney disease causing CKD was lifestyle-related diseases such as nephrosclerosis (NS) and diabetic kidney disease (DKD). Most of the patients classified as ‘others’ had a clinical diagnosis of chronic glomerulonephritis or CKD.

The dose of PSL was higher in the IL-6i group, whereas the dose of MTX was higher in the TNFαi group. The proportion of patients receiving concomitant MTX significantly decreased with lower kidney function and was almost zero in the eGFR<30 mL/min/1.73 m2 group. The proportion was significantly lower in the IL-6i group. The proportion of patients receiving NSAIDs was lower in patients with CKD but higher in those on HD, and overall, there was no significant difference in kidney function. The proportion of patients receiving acetaminophen significantly increased with lower kidney function. The proportion of patients receiving renin–angiotensin system (RAS) inhibitors significantly increased with lower kidney function but was lower in those on HD. Among the patients with HD, the proportion of patients receiving RAS inhibitors was higher in the CTLA4-Ig group for clinically unknown reasons.

Among the 70 patients who had an eGFR of <30 mL/min/1.73 m2, 40 were on HD. The first-line bDMARDs used were TNFαis in 50 patients (etanercept in 31 patients, certolizumab in 7 patients, adalimumab in 5 patients, golimumab in 4 patients and infliximab in 3 patients), IL-6is in 14 patients (tocilizumab in 12 patients and sarilumab in 2 patients) and CTLA4-Ig (abatacept) in 6 patients. In the comparison between the non-HD and HD groups, the non-HD group tended to be older and to have lower CRP and DAS28-CRP levels. The MMP-3 level was higher in the HD group. In the comparison by modality of bDMARDs, the duration of RA was longer in the IL-6i group and shorter in the CTLA4-Ig group; the CRP and DAS28-CRP levels were lower in the IL-6i group; and PSL use was slightly higher in the IL-6i group.

Drug retention rates of first-line bDMARDs stratified by kidney function

Figure 1A–D shows the Kaplan-Meier survival curves of patients treated with first-line bDMARDs stratified by kidney function group. When the patients were divided into three groups according to kidney function (eGFR of ≥60, 30–60, <30 mL/min/1.73 m2), the 36-month drug retention rates were 45.2%, 32.0% and 41.4% for all bDMARDs; 45.3%, 28.2% and 34.0% for TNFαis; 47.4%, 66.7% and 71.4% for IL-6is and 42.9%, 37.5% and 33.3% for CTLA4-Ig, respectively.

Figure 1

36-month drug survival curves of (A) all bDMARDs, (B) TNFαis, (C) IL-6is or (D) CTLA4-Ig stratified by kidney function. (E) HRs and 95% CIs for bDMARD discontinuation stratified by kidney function and drug modality. †Indicates a p<0.05 compared with the group with an eGFR of ≥60 mL/min/1.73 m2. bDMARDs, biological disease-modifying antirheumatic drugs; CKD, chronic kidney disease; CTLA4-Ig, cytotoxic T-lymphocyte antigen-4 immunoglobulin; eGFR, estimated glomerular filtration rate; IL-6i, interleukin-6 inhibitor; TNFαi, tumour necrosis factor α inhibitor.

According to the Cox proportional hazards model, the HR for the discontinuation of first-line bDMARDs according to kidney function status was 1.08 (95% CI 0.91 to 1.28), p=0.37, and the adjusted HR according to age, CRP level, PSL dose and MTX dose was 1.18 (95% CI 0.96 to 1.45), p=0.11 for all bDMARDs.

For TNFαis, the crude HR was 1.17 (95% CI 0.97 to 1.41), p=0.10, and the adjusted HR was 1.30 (95% CI 1.04 to 1.63), p=0.02.

For IL-6is, the crude HR was 0.76 (95% CI 0.38 to 1.40), p=0.42, and the adjusted HR was 0.91 (95% CI 0.35 to 2.38), p=0.85.

For CTLA4-Ig, the crude HR was 1.11 (95% CI 0.63 to 1.96), p=0.72, and the adjusted HR was 1.20 (95% CI 0.58 to 2.46), p=0.62.

Figure 1E summarises the HRs and 95% CIs for bDMARD discontinuation stratified by kidney function and drug modality. We compared HRs for the discontinuation of first-line bDMARDs among three groups stratified according to kidney function using the Cox proportional hazards model, using the group with an eGFR of ≥60 mL/min/1.73 m2 as the reference.

For all bDMARDs, the crude HR for the group with an eGFR of 30–60 mL/min/1.73 m2 was 1.29 (95% CI 0.88 to 1.90), p=0.20, and the HR adjusted for age, CRP level, PSL dose and MTX dose was 1.42 (95% CI 0.94 to 2.16), p=0.20. The crude HR for the group with an eGFR of <30 mL/min/1.73 m2 was 1.05 (95% CI 0.88 to 1.26), p=0.59, and the adjusted HR was 1.18 (95% CI 0.95 to 1.47), p=0.13.

For TNFαis, the crude HR for the group with an eGFR of 30–60 mL/min/1.73 m2 was 1.33 (95% CI 0.87 to 2.04), p=0.19, and the adjusted HR was 1.51 (95% CI 0.94 to 2.41), p=0.09. The crude HR for the group with an eGFR of <30 mL/min/1.73 m2 was 1.14 (95% CI 0.94 to 1.39), p=0.19, and the adjusted HR was 1.32 (95% CI 1.04 to 1.67), p=0.02.

For IL-6is, the crude HR for the group with an eGFR of 30–60 mL/min/1.73 m2 was 0.89 (95% CI 0.11 to 7.41), p=0.91, and the adjusted HR was 0.77 (95% CI 0.06 to 9.95), p=0.84. The crude HR for the group with an eGFR of <30 mL/min/1.73 m2 was 0.75 (95% CI 0.38 to 1.50), p=0.42, and the adjusted HR was 1.34 (95% CI 0.45 to 4.01), p=0.60.

For CTLA4-Ig, the crude HR for the group with an eGFR of 30–60 mL/min/1.73 m2 was 1.25 (95% CI 0.45 to 3.48), p=0.67, and the adjusted HR was 0.95 (95% CI 0.29 to 3.06), p=0.93. The crude HR for the group with an eGFR of <30 mL/min/1.73 m2 was 1.05 (95% CI 0.56 to 1.96), p=0.87, and the adjusted HR was 1.29 (95% CI 0.59 to 2.80), p=0.52.

Even in groups with lower kidney function, the drug retention rate of bDMARDs was generally maintained. However, the HRs and 95% CIs for the discontinuation of TNFαis were significantly lower in the group with an eGFR of <30 mL/min/1.73 m2 than in the group with an eGFR of ≥60 mL/min/1.73 m2. Although not statistically significant, a similar trend was observed for CTLA4-Ig.

Drug retention rates of first-line bDMARDs among patients with an eGFR of <30 mL/min/1.73 m2

Figure 2A shows the Kaplan-Meier survival curves of patients treated with first-line bDMARDs stratified into non-HD and HD groups. The retention rates of first-line bDMARDs at 36 months were as follows: non-HD (40.0%) and HD (42.5%). The crude HR for discontinuation of first-line bDMARDs, comparing non-HD patients to HD patients, was 1.00 (95% CI 0.52 to 1.90), yielding a p value of 0.99, as determined by the log-rank test. The age-adjusted HR was 1.03 (95% CI 0.50 to 2.11), with a p value of 0.93.

Figure 2

36-month drug survival curves of first-line bDMARDs among patients with an eGFR of <30 mL/min/1.73 m2, stratified by (A) non-HD or HD status and (B) modality of bDMARD therapy. bDMARDs, biological disease-modifying anti-rheumatic drugs; CTLA4-Ig, cytotoxic T-lymphocyte antigen-4 immunoglobulin; eGFR, estimated glomerular filtration rate; HD, hemodialysis; IL-6i, interleukin-6 inhibitor; TNFαi, tumour necrosis factor α inhibitor.

Figure 2B shows the Kaplan-Meier survival curves of first-line bDMARDs stratified by drug modality. The retention rates of first-line bDMARDs were as follows: TNFαis (34.0%), IL-6is (71.4%) and CTLA4-Ig (33.3%). The crude HR comparing patients receiving TNFαis to those receiving IL-6is was 0.27 (95% CI 0.08 to 0.88), with a p value of 0.03. The age-adjusted HR was 0.27 (95% CI 0.08 to 0.88), with a p value of 0.03. The crude HR for the comparison of patients treated with TNFαis and CTLA4-Ig was 0.93 (95% CI 0.28 to 3.07), with a p value of 0.91. After adjusting for age and HD status, the HR was 1.02 (95% CI 0.29 to 3.61), with a p value of 0.97.

Changes in RA activity and dosage of PSL

Figure 3A–L illustrates the changes in the DAS28-CRP, DAS28-ESR and PSL dosage over 6 months after the initiation of first-line bDMARDs stratified by drug modality and kidney function. In all modalities of bDMARDs and across all kidney function groups, there was a significant improvement in the DAS-CRP and DAS-ESR at 3 and 6 months after the initiation of bDMARDs compared with baseline, and the dosage of PSL was significantly reduced.

Figure 3

Changes in (A–D) DAS28-CRP, (E–H) DAS28-ESR and (I–L) PSL dosage over 6 months after the initiation of first-line bDMARDs, stratified by drug modality. The background colours A–H indicate the following disease activities: red for high, orange for moderate, green for low, and blue for remission. * and † indicate p values <0.01 and <0.05, respectively, compared with the month=0 group. bDMARDs, biological disease-modifying antirheumatic drugs; CRP, C reactive protein; CTLA4-Ig, cytotoxic T-lymphocyte antigen-4 immunoglobulin; DAS28, Disease Activity Score 28; ESR, erythrocyte sedimentation rate; IL-6i, interleukin-6 inhibitor; PSL, prednisolone; TNFαi, tumour necrosis factor α inhibitor.

Reasons for discontinuation of first-line bDMARDs

Table 2 shows the reasons for discontinuation of first-line bDMARDs. The most common reason for discontinuation was ineffectiveness in all subgroups. No patient died within the observation period.

Table 2

Reasons for discontinuation of first-line bDMARDs

Compared with patients with an eGFR of ≥60 mL/min/1.73 m2, the crude HR for the discontinuation of bDMARDs due to ineffectiveness in patients with an eGFR of <30 mL/min/1.73 m2 was 1.10 (95% CI 0.89 to 1.37), p=0.38, and the age-adjusted HR was 1.11 (95% CI 0.89 to 1.39), p=0.35. Compared with patients with an eGFR of ≥60 mL/min/1.73 m2, the crude HR for the discontinuation of bDMARDs due to infection in patients with an eGFR of<30 mL/min/1.73 m2 was 1.20 (95% CI 0.72 to 2.00), p=0.47, and the age-adjusted HR was 1.20 (95% CI 0.72 to 2.01), p=0.49. Cox hazard analysis revealed that an eGFR of <30 mL/min/1.73 m2 was not a risk factor for the discontinuation of bDMARDs due to ineffectiveness or infection.

Compared with those of the TNFαi group, the crude HRs for discontinuation of IL-6is and CTLA4-Ig due to ineffectiveness were 0.48 (95% CI 0.21 to 1.08), p=0.08, and 0.90 (95% CI 0.50 to 1.64), p=0.74, respectively, and the age-adjusted HRs were 0.48 (95% CI 0.21 to 1.08), p=0.08, and 0.91 (95% CI 0.50 to 1.65), p=0.75, respectively. Compared with those for TNFαis, the crude HRs for the discontinuation of IL-6is and CTLA4-Ig due to infection were 0.80 (95% CI 0.19 to 3.40), p=0.76, and 1.63 (95% CI 0.56 to 4.73), p=0.37, respectively, and the age-adjusted HRs were 0.80 (95% CI 0.19 to 3.41), p=0.76, and 1.58 (95% CI 0.54 to 4.62), p=0.41, respectively. The results of the Cox hazard analysis indicated that compared with TNFαis, IL-6is and CTLA4-Ig are not associated with an increased risk of bDMARD discontinuation due to ineffectiveness or infection.

Next, Cox hazard analysis was conducted only among patients with an eGFR of <30 mL/min/1.73 m2. Compared with those in the non-HD patients, the crude HRs for discontinuation of bDMARDs due to ineffectiveness and infection in the HD patients were 1.15 (95% CI 0.53 to 2.50), p=0.73, and 0.58 (95% CI 0.10 to 3.46), p=0.55, respectively, and the age-adjusted HRs were 1.20 (95% CI 0.51 to 2.83), p=0.68, and 0.41 (95% CI 0.06 to 2.89), p=0.37, respectively. The results of the Cox hazard analysis revealed that, in patients with an eGFR of <30 mL/min/1.73 m2, HD was not a risk factor for discontinuation of bDMARDs due to ineffectiveness or infection.

Compared with those for patients with an eGFR of <30 mL/min/1.73 m2, the crude HRs for the discontinuation of IL-6is due to ineffectiveness and infection were 0.11 (95% CI 0.02 to 0.85), p=0.03, and 0.64 (95% CI 0.07 to 5.79), p=0.69, respectively, and the age-adjusted HRs were 0.11 (95% CI 0.02 to 0.85), p=0.03, and 0.64 (95% CI 0.07 to 5.78), p=0.69, respectively. The results of the Cox hazard analysis revealed that among patients with an eGFR of <30 mL/min/1.73 m2, the use of IL-6is was associated with a lower risk of bDMARD discontinuation due to ineffectiveness than that for TNFαis.

Discussion

The management of RA in patients with CKD is crucial not only for alleviating joint symptoms but also for preventing the progression of CKD. RA accelerates CKD worsening, not only due to factors including RA-related AA amyloidosis and kidney damage caused by NSAIDs but also due to inflammation-induced atherosclerosis and anaemia.1–5 There are an estimated one million patients suffering from RA in Japan; Tokoroyama et al reported that the prevalence of CKD among patients with RA is 24.5%,1 which is significantly higher than that among the general Japanese population (12.9%).20 Given the substantial number of patients affected, it is crucial to establish robust evidence-based strategies for managing patients with RA with CKD.

However, managing RA in patients with CKD presents unique challenges. In patients with CKD, the incidence of infection is higher than that in non-CKD patients, even in patients with mild kidney dysfunction.6 Especially among patients with an eGFR of <30 mL/min/1.73 m2, the therapeutic options are critically limited; MTX, an anchor drug for RA treatment, is contraindicated,7 and the use of NSAIDs requires careful consideration. In this context, bDMARDs may offer potential breakthroughs for these patients.

Nonetheless, there is a significant lack of evidence regarding the efficacy and safety of these drugs in patients with RA with CKD. First, the population of patients with RA with CKD was excluded from clinical trials of bDMARDs.21–28 Several case reports and case series with limited patient populations have highlighted the successful application of infliximab,9 etanercept10–12 and tocilizumab13 in the management of patients with RA with CKD. Sumida et al specifically investigated the safety of adalimumab for 65 patients with RA with CKD, including two patients receiving HD.14 Their findings suggested that adalimumab does not exacerbate kidney dysfunction and is not associated with any significant adverse events. Thus, the authors proposed that adalimumab might serve as a viable therapeutic option for patients with RA with CKD. However, there is a lack of studies that evaluate a substantial number of patients or investigate the long-term drug retention rate of bDMARDs among patients with RA who also have an eGFR of <30 mL/min/1.73 m2. Moreover, comprehensive analyses encompassing multiple modalities of bDMARDs are also noticeably absent from the literature.

Our study was conducted to fill this evidence gap. A key strength of this study lies in its pioneering approach, as it is the first to explore the real-world efficacy and safety of comprehensive bDMARD modalities. Additionally, this study offers an extended follow-up and involves an adequate sample size of patients with RA with CKD. This report represents the first comprehensive evaluation of the efficacy and safety of bDMARDs in patients with an eGFR of <30 mL/min/1.73 m2, particularly those receiving HD, across a large patient cohort and encompassing a broad range of bDMARD modalities.

Among patients with CKD included in this study, the most common primary kidney disease causing CKD was lifestyle-related diseases such as NS and DKD. We have previously reported that the frequency of RA-associated kidney diseases, such as AA amyloidosis and bucillamine-related membranous nephropathy, decreased among patients with RA while the proportion of NS and DKD increased in the 2000s.29 A high proportion of NS and DKD in the current study, which included cases since 2004, coincides with the increase of NS and DKD in patients with RA.

This study has yielded two noteworthy insights.

First, we demonstrated that the drug retention rate of first-line bDMARDs does not significantly decrease even as CKD progresses. Although the proportion of NSAIDs users decreased in the population with lower kidney function, the retention rate was maintained. This reinforces the efficacy of bDMARDs among patients with lower kidney function. Furthermore, regardless of kidney function, the disease activity of RA and the dose of PSL notably decreased following the initiation of bDMARDs. Reducing the dosage of PSL while maintaining well-controlled disease activity is also desirable for lowering the risk of infection. Particularly among patients with an eGFR of <30 mL/min/1.73 m2, we found no significant difference in the retention rates of first-line bDMARDs between the HD and non-HD groups. This finding is important because it suggests that bDMARDs could be used tolerably even in patients with RA on HD. Note that the proportion of NSAIDs users was higher in HD group than in non-HD group. The maintained retention rate of bDMARDs among patients with HD may be at least partially attributed to pain control by NSAIDs.

Second, however, it was suggested that the drug retention rate of TNFαis may be lower in the group with an eGFR of <30 mL/min/1.73 m2. IL-6is group had a higher proportion of monotherapy without MTX, and the retention rate of IL-6is was higher than that of other bDMARDs. This is a reasonable outcome, considering that prior reports have indicated that TNFαis tend to decrease the retention rate in the absence of MTX,30 whereas IL-6is are able to maintain the retention rate, even when administered as monotherapy.31 Supporting these findings, the HR for discontinuation of TNFαis due to ineffectiveness was significantly higher than that for discontinuation of IL-6is in the group with an eGFR of <30 mL/min/1.73 m2. Considering these findings, IL-6is may be a promising first-line therapeutic option for the management of RA in patients with an eGFR of <30 mL/min/1.73 m2. Tocilizumab also has the secondary benefit of reducing anaemia associated with chronic inflammation in patients with RA32 and contributes to the prevention of CKD progression.5

The ANSWER study reported the 36-month drug retention rates of the following first-line bDMARDs: etanercept (57.4%), adalimumab (71.6%), tocilizumab (76.5%) and abatacept (81.7%).33 Note that most of the patients in the ANSWER study received concomitant MTX, and the amount of concomitant PSL was higher than that in our study. In terms of bDMARD monotherapy, the French 12-month ACT-SOLO study reported a 12-month drug retention rate of 67% for first-line tocilizumab monotherapy,34 a finding that is comparable to the 36-month retention rate of tocilizumab monotherapy observed in our study. The rates of TNFαis and CTLA4-Ig were lower than those of IL-6is but still practically acceptable.

This study has several limitations. Its relatively small sample size and retrospective design restrict the robustness and the generalisability of the findings, especially when stratified according to kidney function categories. The larger and prospective studies are warranted to validate our findings. In particular, the number of IL-6i-treated patients was small in the group with an eGFR of 30–60 mL/min/1.73 m2. Furthermore, the number of CTLA4-Ig-treated patients included in this study was insufficient, necessitating further research and compilation of experiences regarding the use of CTLA4-Ig. Our study design limiting the scope only to first-line therapy allowed an accurate evaluation of the retention rate of bDMARDs but also resulted in the lack of Janus kinase inhibitor (JAKi) users. Further exploration is needed to evaluate the tolerability of JAKis. It is also crucial to highlight the importance of patient-centred care in this setting. Clinicians must, therefore, consider patients’ preferences, lifestyles and psychosocial factors when deciding on treatment options.

In conclusion, our study provides initial evidence suggesting that first-line bDMARDs can be used effectively and safely in patients with RA with CKD, even among patients on HD. In particular, IL-6is had a significantly higher drug retention rate in patients with an eGFR of <30 mL/min/1.73 m2 and fewer drug discontinuations due to its ineffectiveness. Further research is needed to confirm these findings, elucidate the precise mechanisms underlying the observed outcomes and optimise the use of bDMARDs in patients with RA with CKD.