High rates of continuation for RA patients switched to second TNF inhibitor
January 2007 – Jon Giles, M.D.
Clinical trials of TNF inhibitors in rheumatoid arthritis (RA) typically show minimal clinical response rates (i.e. ACR20) of 60 to 70%. However, clinically meaningful response rates, such as an ACR50 or ACR70 response, are substantially lower. This translates to more than half of patients who are treated with a first TNF inhibitor not achieving an ideal response. Clinical observation and a handful of small, underpowered studies have suggested that switching to an alternative TNF inhibitor can achieve greater efficacy. Here, Hyrich et al (Arthritis Rheum 2007; 56: 13) investigate the efficacy of switching TNF inhibitors in a large, registry-based cohort of RA patients in the United Kingdom (UK).
Methods
Subjects derived from the British Society for Rheumatology Biologics Register, a prospective, observational cohort of all subjects initiating TNF inhibitor therapy in the UK. Clinical and laboratory data are collected from the patient and treating rheumatologist at baseline and every 6 months. For this study, Continuation rates were compared for subjects initiating TNF inhibitor therapy, and for subjects switching to an alternate TNF inhibitor after inefficacy or toxicity of a first TNF inhibitor.
Results
6739 subjects were analyzed. Most were female (77%) with a mean age of 55 years and a mean disease duration of 14 years prior to initiation of their first TNF inhibitor. The median number of previous DMARDs was 4 and the patients had high baseline disease activity (mean DAS28 6.6). The proportions of subjects receiving etanercept, infliximab, and adalimumab were 42%, 45%, and 13%, respectively. Of these, 30%, 85%, and 40%, respectively, received concomitant methotrexate.
Approximately 28% of subjects discontinued their initial TNF inhibitor, with relatively equal proportions discontinuing because
of inefficacy or adverse events. Of the subjects discontinuing for inefficacy, 60% switched to an alternate TNF inhibitor while
only 35% those discontinuing for adverse events switched to an alternate TNF inhibitor. Subjects first receiving etanercept
were just as likely to be switched to infliximab or adalimumab if switching for inefficacy, but more likely to be switched to
adalimumab for adverse events. Subjects receiving monoclonal antibody TNF inhibitors were more likely to be switched to
etanercept than the alternate monoclonal antibody for both inefficacy and adverse events.
Of the subjects who switched to a second TNF inhibitor, 73% were still receiving the second agent at close of the observation
period (mean 6 months, longest period 32 months). Of the 37% who discontinued the second TNF inhibitor, the frequency for
discontinuations for inefficacy and adverse events were similar. Subjects who discontinued the first TNF inhibitor because of
inefficacy were almost 3 times more likely to discontinue the second TNF inhibitor secondary to inefficacy (RR 2.7 (95%CI 2.1 – 3.4)), but there was no increased risk of discontinuing the second agent due to an adverse event if the first agent was
discontinued due to inefficacy. Similarly, Subjects who discontinued the first TNF inhibitor because of an adverse event were
more than 2 times more likely to discontinue the second TNF inhibitor secondary to an adverse event (RR 2.3 (95%CI 1.9 –
2.9)), but there was no increased risk of discontinuing the second agent due to inefficacy if the first agent was discontinued
due to an adverse event. If the reason for discontinuing the second TNF inhibitor was also an adverse event, only rarely was
the second adverse event a recurrent one.
Conclusions
Rates of continuation of a second TNF inhibitor were high after discontinuation a first TNF inhibitor because of inefficacy or adverse events. Of those who discontinued the second agent, reasons for discontinuation were frequently similar to the reason for discontinuing the first agent.
Editorial Comment
This investigation is the largest and best available to confirm a common clinical practice. Though no head
to head studies have been performed, the TNF inhibitors, as a class, appear to have similar efficacy. However, these data
suggest that all TNF inhibitors are not necessarily alike and a class effect for all outcomes cannot be assumed. This is
particularly interesting for the finding that recurrent adverse events of the same type are relatively uncommon. However, it is
important to note that those patients with particularly severe adverse events with the first TNF inhibitor were likely not switched
to a second TNF inhibitor and thus excluded form the later analyses. Extended follow-up of this informative cohort will
establish longer-term continuation rates and, potentially, establish whether there is any value in considering a third TNF
inhibitor in patients who have failed two previous anti-TNF agents.