Sep 17, 2019
Biologics in rheumatoid arthritis: better evidence, but direct comparisons still lacking
Only few differences between biologics detectable / Operationalization of outcomes can influence results
In the summer of 2018, the German Institute for Quality and Efficiency in Health Care (IQWiG) presented its preliminary report on the benefit assessment of biotechnologically produced drugs – so-called biologics – in the treatment of rheumatoid arthritis and asked for comments. The comprehensive final report has now been published. According to the findings, long-term studies and, most notably, studies in which at least two of the nine biologics to be assessed were compared with each other are still lacking - even though these drugs have been on the market for almost two decades and there is no shortage of potential study participants. Nevertheless, there are important new findings, not least because upon request, pharmaceutical companies submitted further analyses of older studies to IQWiG.
Drugs from cell cultures
Rheumatoid arthritis is an autoimmune disease and the most common form of chronic inflammatory joint disease. Patients suffer from pain, fatigue and exhaustion, depressive moods, and limitations of physical function. They also suffer from a loss of quality of life, independence and involvement in social and professional life. The primary aim of treatment is to relieve patients of most of the symptoms of disease and to prevent joint destruction (remission). Where this is not possible, at least disease activity should be reduced.
Treatment includes the use of biotechnologically produced drugs – so-called biologics. They are produced by cell cultures and influence various parts of the inflammatory process. Under certain conditions, biologics are approved for both first-line treatment and further lines of treatment, partly in combination with the drug methotrexate and partly as monotherapy. This results in seven treatment situations for which IQWiG was to investigate the advantages and disadvantages of nine biologics on behalf of the Federal Joint Committee (G-BA).
Even after almost two decades, hardly any comparative studies available
As stated, long-term studies and, most notably, studies in which at least two of the nine biologics to be assessed were compared with each other are still lacking – even though these drugs have been on the market for almost two decades and there is no shortage of potential study participants.
In order to investigate the advantages and disadvantages of biologics, so-called network meta-analyses can be used in which pairwise comparisons that are clinically and methodologically sufficiently similar are combined to form a network that ideally enables the comparison of all biologics with each other.
Little evidence of differences between biologics
In first-line treatment in combination with methotrexate, there is no hint of greater or lesser benefit of one biologic versus another for the primary treatment goal, clinical remission. Low disease activity was better achieved with adalimumab and etanercept than with certolizumab pegol or tocilizumab; for the former two, this leads to a hint of greater benefit. No further differences were shown.
After failure of methotrexate monotherapy, patients may additionally be treated with a biologic. In this treatment situation, all drugs except etanercept showed greater benefit compared with anakinra. For certolizumab pegol, there is a hint of greater harm due to more side effects compared with all other biologics. In addition, more side effects were shown under golimumab and tocilizumab than under infliximab. In addition, more patients stopped treatment due to side effects under tocilizumab than under abatacept. In each case, these results were interpreted as a hint of greater harm.
For combination therapy after failure of a biologic, the submission of data on the corresponding subpopulations allowed conclusions on several biologics and outcomes. Moreover, with regard to monotherapy with a biologic after methotrexate intolerance, for the first time ever, results could be determined for outcomes that were originally not included in the study documents.
No conclusion was possible for the three remaining treatment situations of the benefit assessment due to the insufficient data situation.
New analyses improved basis for network meta-analyses
In addition to the results on the benefit and harm of biologics, this assessment provides an example of how data from older studies can also be used for new methods such as network meta-analyses. This method can only be applied if the patient populations included in the analysis are sufficiently similar. This was achieved for many studies because companies provided new analyses of the relevant patient populations on the basis of old study data. Further re-analyses of the old data concerned the benefit outcomes “clinical remission” and “low disease activity”: Based on the current definitions of these outcomes, the companies recalculated results for those studies where older definitions had been used. This important information is now also available for the first time for future assessments of drugs for the treatment of rheumatoid arthritis.
Operationalization of outcomes can influence conclusions on benefit
For example, for tocilizumab, the results analysed with a current operationalization excluding inflammation parameters were consistently less positive than with an older operationalization including inflammation parameters. These differences in results are presumably caused by the fact that compared with other biologics, tocilizumab has a stronger effect on inflammation parameters. Only the use of the current operationalization excluding inflammation parameters allows a fair comparison of the biologics with each other.
Discussion about prerequisites for meaningful network meta-analyses
If the network had contained insufficient data, it would not have been possible to derive reliable conclusions on the objective of the IQWiG commission. IQWiG chose a 50-percent threshold: Network meta-analyses were only calculated for treatment situations in which data were available for at least half of the biologics approved. Therefore, some smaller study pools, individual outcomes or less frequent operationalizations of outcomes were not investigated.
In the commenting procedure on the preliminary report, several participants criticized this threshold. However, in the discussion, no consistent line of argument was recognizable. Furthermore, no adequate alternative proposal was made that would have come close to fulfilling the objective of the IQWiG commission, namely to compare the biologics with each other.