Home > Full Text Reviews > Adalimumab, Etanercept, Infliximab,... > Withdrawals from treatment with tumour...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Malottki K, Barton P, Tsourapas A, et al. Adalimumab, Etanercept, Infliximab, Rituximab and Abatacept for the Treatment of Rheumatoid Arthritis After the Failure of a Tumour Necrosis Factor Inhibitor: A Systematic Review and Economic Evaluation. Southampton (UK): NIHR Journals Library; 2011 Mar. (Health Technology Assessment, No. 15.14.)

Appendix 12Withdrawals from treatment with tumour necrosis factor inhibitors

Withdrawal from treatment with second-line tumour necrosis factor inhibitor (British Society for Rheumatology Biologics Registry data)

Updated BSRBR model data163 provided Kaplan–Meier (K–M) plots for survival in treatment for four groups of patients receiving second-line TNF inhibitors as follows: (i) withdrew from first-line TNF inhibitor for lack of efficacy and from second-line TNF inhibitor for lack of efficacy; (ii) withdrew from first-line TNF inhibitor for lack of efficacy and from second line TNF inhibitor for AEs; (iii) withdrew from first-line TNF inhibitor for AEs and from second-line TNF inhibitor for lack of efficacy; (iv) withdrew from first-line TNF inhibitor for AEs and from second-line TNF inhibitor for AEs.

The proportion lost to treatment at 3-month time points in each category was read from the graphs in the BSRBR submission and the absolute number lost calculated using N = 995 for first-line withdrawal through lack of efficacy and N = 1,882 for first-line withdrawal due to AEs. The proportion of patients withdrawing for any reason was then estimated and the proportion remaining in treatment plotted (data points in Figure 155). A Weibull distribution (time in years) was fitted to the data [scale parameter (lambda) 0.441555; standard error (SE) 0.00958300], shape parameter (gamma) 0.7008 (SE 0.033681) labelled BSRBR Weibull fit in Figure 155 (extrapolation to 25 years is shown in the inset).

FIGURE 155. Continuation in second-line TNF inhibitor.

FIGURE 155

Continuation in second-line TNF inhibitor.

Comparison with manufacturers' submissions

The Schering-Plough Ltd (infliximab) submission165 provided Weibull parameters for treatment withdrawal that were also based on BSRBR data; the parameters are shown below.

Log(scale)3.529 (time in months)
Log(shape)−0.19 (time in months)

Assuming log(scale) in the table above refers to ‘log β’ where β = (1/λ) ^ [1/γ], and survival = exp[–(t × β) ^ γ], then lambda = 0.054 and gamma = 0.827 and the fitted curve labelled Schering-Plough Ltd in Figure 155 is generated (and can be seen to be very similar to the review group's fit).

The Wyeth Pharmaceuticals submission226 modelled withdrawal from treatment using a ‘shared frailty’ model and this is also represented in Figure 155.

Withdrawal from second-line treatment according to tumour necrosis factor inhibitor

According to analysis of Danish Registry for Biologic Therapies in Rheumatology (DANBIO) data withdrawal from first line TNF inhibitor occurs at rates that are statistically significantly different between the three TNF inhibitors, Table 112 provides the reported HRs and 95% CIs (Hetland et al.227).

TABLE 112. Hazard ratios for withdrawal from first-line TNF inhibitors (DANBIO data).

TABLE 112

Hazard ratios for withdrawal from first-line TNF inhibitors (DANBIO data).

It may be reasonable to expect that similar differences might apply for second line TNF inhibitors.

Data were extracted from the K–M graph for each TNF inhibitor published for the Danish registry.227 These were fitted with Weibull distributions (Figure 156) and survivors then combined for each drug (according to number of patients given each TNF inhibitor) so as to provide overall survival (N = 2,935), and this in turn was fitted with a Weibull distribution.

FIGURE 156. Withdrawal from first-line TNF inhibitors (DANBIO data with Weibull fits).

FIGURE 156

Withdrawal from first-line TNF inhibitors (DANBIO data with Weibull fits).

The shape parameters for the Weibull fits were similar and therefore it was considered reasonable to average these and apply the same shape parameter for each drug and for overall survival. Because the BSRBR first-line withdrawal data were derived using equal numbers of patients (∼ 4,000) treated with each TNF inhibitor the shape parameters for the DANBIO data were combined to give an unweighted average. Using this ‘common’ shape parameter (0.5595) the data were again fitted with Weibull distributions, providing the fits shown in Figure 156; the overall survival then assumed that equal numbers received each of the three TNF inhibitors; this allows a comparison of DANBIO and BSRBR first-line withdrawal data (see below).

The HRs (ratio of scale parameters) for comparison of TNF inhibitors using these Weibull fits were within the HR 95% CIs reported for the Danish registry data (note: contact with the lead author confirmed that the published HRs were reversed for ADA versus ETN and IFX versus ADA; this has been corrected in Table 112. Relative to all patients (equal mixture) the HRs for each TNF inhibitor were calculated as follows: ETN versus all, 0.751; ADA versus all, 0.958; IFX versus all, 1.353.

When these HRs are applied to the Weibull fit of BSRBR data163 for continuation of second-line treatment, the drug-specific rates of withdrawal over 25 years are as shown in Figure 157.

FIGURE 157. Estimated continuation of second-line treatment according to TNF inhibitor.

FIGURE 157

Estimated continuation of second-line treatment according to TNF inhibitor.

The Danish National Registry and the British Society for Rheumatology Biologics Registry withdrawal rates from first-line tumour necrosis factor inhibitor

Data for first-line withdrawal were extracted from the UK BSRBR submission163 and fitted with Weibull distributions in which the shape parameter was or was not fixed to that for overall survival derived from the DANBIO data (0.5595, see above). Extrapolations to 25 years were compared between UK and Danish first-line treatments and between first-line and second-line treatments (Figure 158).

FIGURE 158. Modelled survival in treatment with first- and second-line TNF inhibitors.

FIGURE 158

Modelled survival in treatment with first- and second-line TNF inhibitors.

Additional sources of evidence

Several additional sources were identified with potentially relevant information on withdrawal from the different TNF inhibitors; these are listed in Table 113.

TABLE 113. Studies reporting withdrawal rates from TNF inhibitors.

TABLE 113

Studies reporting withdrawal rates from TNF inhibitors.

Except for the DANBIO registry data227 the studies do not provide the information required (K–M plots) to easily compare withdrawal rates between different TNF inhibitors, the main reasons being mixed analysis of first- and second-line withdrawal, mixed populations [rheumatoid arthritis (RA) only a subpopulation, or outcome measure a combination of switching and of dose escalation.218 The German study229 does provide information for ETN and ADA but follow-up was insufficient to see any difference developing. Wolfe and Michaud230 reported median survival on second-line TNF inhibitor. These results (Table 114) compare reasonably well with the median survival for each TNF inhibitor calculated as described above and shown in Figure 157.

TABLE 114. Median survival for second-line TNF inhibitors.

TABLE 114

Median survival for second-line TNF inhibitors.

In general, the data from these studies are consistent with the DANBIO study in that continuation with ETN appears to be superior to that with infliximab and continuation with ADA treatment being intermediate.

© 2011, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Cover of Adalimumab, Etanercept, Infliximab, Rituximab and Abatacept for the Treatment of Rheumatoid Arthritis After the Failure of a Tumour Necrosis Factor Inhibitor: A Systematic Review and Economic Evaluation
Adalimumab, Etanercept, Infliximab, Rituximab and Abatacept for the Treatment of Rheumatoid Arthritis After the Failure of a Tumour Necrosis Factor Inhibitor: A Systematic Review and Economic Evaluation.
Health Technology Assessment, No. 15.14.
Malottki K, Barton P, Tsourapas A, et al.
Southampton (UK): NIHR Journals Library; 2011 Mar.

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...