Biologic monotherapy in the biologic naïve patient with rheumatoid arthritis (RA): results from an observational study

Up to a third of rheumatoid arthritis (RA) patients on biologic therapy do not take concurrent conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) contrary to the recommendations from the most recent EULAR guidance. [1, 2] There are many reasons for this including adverse effects, poor drug adherence and contraindications to csDMARDs [3]. Despite this commonplace real-world practice, there are few head-to-head randomised control trials (RCTs) comparing biologic monotherapies to guide our prescribing and those published have limited follow-up periods of typically 24 weeks [3]. The ADACTA and MONARCH trials showed superiority of IL6 receptor inhibitor (IL6Ri) monotherapy to adalimumab (a TNFi) monotherapy in reduction of DAS28 score at 24 weeks [4, 5]. Furthermore, large commercial RCTs frequently exclude patients with complex comorbidity, thus results are not always directly applicable in real-world settings. Aiming to address this knowledge gap, we compared the efficacy and persistence of multimodal biologic agents as monotherapy in biologic naïve patients with RA in the real-world setting focusing specifically on the use of TNFi, IL6Ris and CTLA-4is.

We conducted a retrospective observational study across four UK Rheumatology units. Ethical approval for the study protocol was sought and gained with the UK Health Research Authority following research ethics committee review. Data were extracted from clinical biologic drug monitoring databases at each participating unit. Data were anonymised prior to analysis. All patients with a diagnosis of RA (as defined by their rheumatologist) initiating a biologic therapy without concurrent csDMARD therapy between 2012 and 2016 were included. In line with NHS prescribing policies, to be applicable for funding for biologic monotherapy, all patients were either intolerant or had contraindications to csDMARDs. Only subcutaneous biologic agents were included. DAS28 scores had to be greater than 5.1 on two occasions to be applicable for funding through the National Health Service (NHS) for biologic therapy. To continue to receive NHS funding for biologic therapy, patients must show initial and continued response to treatment as demonstrated by a reduction in DAS28 score. Patients are assessed routinely at baseline and then at every 6 months thereafter to comply with national guidance. Due to the recognised effect of IL6-i on CRP, we only included DAS28 score calculated using the ESR.

Patients on concurrent csDMARDs, long-term steroids or previous bDMARD exposure were excluded. During the study, some patients may have received short courses of oral steroids or intraarticular steroids if they had a flare of disease. Due to the ad hoc nature of these prescriptions, it was not possible to collect these data. However, at a patient’s 6-month review, routine questioning identifying repeated disease flares would be an indication of failure of biologic treatment prompting change in treatment which would be recorded.

Baseline demographical data was collected including age, sex, seropositivity status using rheumatoid factor (RF) or presence of anti-citrullinated protein antibodies (ACPA) where available and presence of comorbidities (specifically including malignancy, diverticulitis, severe infection, heart failure, ischaemic heart disease, demyelinating disease, tuberculosis, diabetes or interstitial lung disease). Rationale for prescription of biologic agent as monotherapy was also recorded.

We compared biologic classes TNFi, IL6Ri and CTLA-4 inhibitors. The primary outcome was absolute DAS28 at 6, 12 and 18 months.

Covariates were pre-defined according to their potential association with the outcome, including age, sex, seropositivity status using ACPA or RF, comorbidities (including malignancy, demyelinating disease diverticulitis, severe infection, heart failure, ischaemic heart disease, tuberculosis, diabetes or interstitial lung disease), previous DMARD history and the reason for discontinuation; baseline DAS28-ESR was recorded with individual components. Drug exposure duration was recorded.

In this study using real-world data, IL6Ri led to lower DAS28 at 6 and 12 months compared to TNFi monotherapy, although differences were attenuated by 18 months. Similarly, CTLA-4i monotherapy resulted in lower DAS28 at 6 months than TNFi, although superiority was lost by 12 months. Drug persistence at 18 months was great in the CTLA-4i and IL6Ri groups than those on TNFi.

To date, there is only one head-to-head RCT of CTLA-4i with the TNFi adalimumab, however, all patients were on concurrent methotrexate [7]. Our study demonstrated significant reduction of SJC at 18 months and comparably low DAS28 score for those on CTLA-4i versus TNFi. Notably, those on CTLA-4i had more comorbidities (a typically more challenging group to treat), the high drug retention of 80% at 18 months would support the use of CTLA-4i in this cohort. These real-world data are particularly useful as the BSR biologics registry does not include CTLA-4i monotherapy so there is little published data on this cohort. The observational study by Jorgensen et al. compared biologic monotherapy (including those on TNFi, CTLA-4i and IL6Ri) and disease activity at 6 months only [8]. The study did not adjust for potential confounding factors. It included patients both newly started on the drug during the study period and those started on it prior to study initiation, although post hoc statistical analysis found there was no difference in drug adherence between groups or CDAI remission rates. Furthermore, 22% of all patients included were on long-term prednisolone complicating the analysis. Like the study by Bachkaus et al. (which compared TNFi and Tocilizumab mono therapy), Jorgensen et al. included patients who had been on multiple biologic therapies previously which could indicate harder-to-treat disease [9]. In a further study by Jorgenson et al., no adjustment for potential confounders has been included therefore the results are open to confounding by indication [10].

By contrast, our study design including only biologic naive patients and adjustment for potential confounding factors including comorbidities, seropositivity and age amongst others makes the results more easily interpretable. Due to the study design relying on retrospective data collection, there were some data points missing particularly in regard to seropositivity status of patients, however, our statistical analysis took account of this.

Key strengths of this study included examination of DAS28 components. We showed that IL6Ri had profound impact on ESR, as expected, but also a lower SJC for IL6i and CTLA-4i monotherapies. Our duration of follow-up was also longer than typical RCTs, allowing us to see CTLA-4i may be more efficacious initially, but not at long-term follow-up. Our study was limited by the lack of other outcomes including CRP. Patients enrolled in this study had high disease activity with baseline median DAS28 scores 6.1–6.3 across the drug groups. We recognise that in some clinical contexts, patients maybe commenced on biologic monotherapy with less active disease. Furthermore, caution should be taken drawing conclusions from our small sample size in particular relating to the IL6i and CTLA-4i groups. However, the effect size of TNFi vs IL6Ri comparison was similar to randomised trials [4].

Our findings support EULAR guidance that IL6Ri should be considered as a first-line biologic in the biologic naive patient requiring monotherapy [2]. However, CTLA-4i may also have a role in this patient population and further studies of larger patient numbers are needed to investigate this further.