Original article
Results of three analytical approaches on long-term efficacy of etanercept for psoriasis in daily practice

https://doi.org/10.1016/j.jaad.2012.05.040Get rights and content

Background

A problem encountered when analyzing long-term efficacy is that the number of patients in follow-up decreases with time for different reasons. The method used to account for missing observations for the therapy under analysis has a great influence on the inference of efficacy.

Objective

To describe the long-term efficacy of etanercept for psoriasis in daily practice using 3 analytical approaches.

Methods

Prospective data from a cohort of patients with psoriasis treated with etanercept for at least 24 weeks were analyzed using 3 analytical approaches: as treated analysis, intention-to-treat analysis (ITT) with last observation carried forward (LOCF) and intention-to-treat analysis with modified nonresponder imputation (modified NRI).

Results

One hundred thirty-one patients were treated with etanercept during 134 treatment episodes with a mean treatment duration of 2.7 years. The maximum follow-up was 6.0 years. The methodological approach chosen had a great influence. Psoriasis Area and Severity Index (PASI) 75 response rates varied from 60% in the as-treated approach to 34% in LOCF and to 29% in modified NRI at week 264.

Limitations

All analytical methods applied have limitations. Other outcome measures could be used to overcome the bias introduced by each method of analysis, such as drug survival.

Conclusions

The methodological approach chosen to analyze long-term efficacy data has a great influence on the inferences that may be drawn regarding the degree of efficacy. Therefore we support the use of different methods to present long-term efficacy data.

Introduction

The efficacy and safety of etanercept for psoriasis have been studied in randomized controlled trials with open label extension studies for up to 4 years of treatment.1, 2, 3 Open-label extension studies have shown some loss of efficacy of etanercept after week 48 to week 52.2, 3

However, the results from randomized controlled trials (RCTs) may not reflect the daily clinical practice situation, as patients included in RCTs are highly selected and treated according to a predefined treatment schedule.

There are few reports on the management of unselected patients in day-to-day practice, especially in the long term. This study provides complementary information to RCTs about the efficacy and safety of etanercept for psoriasis in daily practice up to 300 weeks of treatment.

Observational cohort studies can be analyzed according to the intention-to-treat (ITT) principle and the as-treated principle. The intention-to-treat principle was originally designed for the analysis of RCTs. In an intention-to-treat analysis, all patients are analyzed according to the initial treatment intent following randomization, irrespective of the treatment actually received. This is different from an as-treated analysis, in which the analysis is based on the treatment that patients actually received.

A problem encountered when analyzing long-term efficacy data is that the number of patients with available efficacy data for the therapy under analysis decreases with time. In observational studies, inclusion of patients is continuously ongoing. At the time of analysis a data lock is performed, including patients with a short follow-up for the therapy under analysis. Furthermore, in some patients, biologic therapy is discontinued because of insufficient efficacy or intolerance. In that case, efficacy data for the therapy analyzed are not available anymore and can only be measured when patients are on another therapy or no therapy. In addition, in case of loss of follow-up, outcome measures cannot be measured anymore.

In an ITT analysis, none of the patients are excluded from the analysis. This is different from an as-treated analysis, in which patients with insufficient follow-up for the therapy under analysis are excluded from the analysis. In order to include patients with missing efficacy data for the therapy under analysis in the ITT analysis, the last observation available for the therapy analyzed can be extrapolated, which is also applied in the RCTs with etanercept for psoriasis. This involves making assumptions about the outcomes, which can be done with the last observation carried forward (LOCF) method or the (modified) nonresponder imputation ([modified] NRI) method.

The method used has a great influence on the inference of efficacy.4, 5 The ITT analysis with (modified) nonresponder imputation approach may give a too-negative view of the efficacy of etanercept. On the other hand, the as-treated analysis introduces a bias towards a too-positive outcome. The ITT analysis with LOCF approach produces intermediate results.

The primary objective of this study was to describe the long-term efficacy of etanercept therapy for psoriasis in daily practice and to compare 3 analytical approaches. Efficacy was expressed as the percentage of patients reaching a reduction of the Psoriasis Area and Severity Index6 of 50% (PASI 50), 75% (PASI 75) or 90% (PASI 90). Results are described as follows: as-treated analysis, LOCF, and modified NRI.4

Section snippets

Patients

This prospective cohort study involved all consecutive patients with psoriasis treated with etanercept between February 2005 and February 2011 for at least 24 weeks. Efficacy data were extracted from a prospective patient registry, containing data from all patients starting biological treatment for psoriasis in the Dermatology outpatient department of the Radboud University Nijmegen Medical Centre.7 In the Netherlands, biological treatment was approved for the treatment of patients with

Methods

Efficacy and safety evaluations were scheduled at baseline, week 6 of treatment, week 12, and subsequently every 12 weeks. At baseline, demographic information and information about the medical history, previous medication use for psoriasis and concomitant medication was collected. PASI scores were collected at baseline and at each subsequent visit, as well as information on concomitant medication use and adverse events.

The long-term efficacy of etanercept was investigated in patients with a

Patients

The cohort consisted of 131 patients treated with etanercept during 134 treatment episodes lasting for 24 weeks or longer. The number of patient-years of follow-up was 362. Eighty-one patients were male (62%), the mean age at the start of etanercept treatment was 47.5 (±11.4) years, and the mean duration of psoriasis was 22.4 (±10.6) years (Table I). The mean body mass index (BMI) was 28.7 (±5.6) kg/m2 and 40 patients (31%) suffered from psoriatic arthritis. The mean duration of an etanercept

Discussion

As shown in this study, the methodological approach chosen to analyze long-term efficacy has a great influence on the efficacy results. As an example, the PASI 75 response rate varied from 60% in the as-treated approach to 34% in LOCF and 29% in modified NRI at week 264. This means that efficacy doubled when the as-treated approach was used instead of modified NRI. Therefore, when comparing efficacy data from different studies, it is important to consider the analysis method used.

With the

References (13)

There are more references available in the full text version of this article.

Cited by (19)

View all citing articles on Scopus

Funding sources: The Radboud University Nijmegen Medical Centre was supported in part by UMC St Radboud Foundation, which received funding from Pfizer and Abbott for the project.

Disclosure: Dr van Lümig carries out clinical trials for Abbott and Janssen-Cilag, has received speaking and consulting fees from Wyeth and Schering-Plough as well as receiving reimbursement for attending a symposium from Schering-Plough and Pfizer. Dr Driessen has received funding from Merck Serono and carried out clinical trials for Wyeth, Schering-Plough, Centocor, Abbott, Merck Serono and Barrier Therapeutics; in addition, Dr Driessen has received speaking and consulting fees from Wyeth and Schering-Plough as well as reimbursement for attending a symposium from Merck Serono, Wyeth, and Janssen-Cilag. Dr van de Kerkhof serves as a consultant for Schering-Plough, Celgene, Centocor, Allmirall, UCB, Wyeth, Pfizer, Sofinnova, Abbott, Actelion, Galderma, Novartis, Janssen-Cilag and LEO Pharma. In addition, Dr van de Kerkhof receives research grants from Centocor, Wyeth, Schering-Plough, Merck Serono, Abbott, and Philips Lighting. Dr de Jong served as a consultant for Biogen, Merck Serono, Wyeth, and Abbott and has received research grants from or was involved in clinical trials from Schering-Plough, Abbott, Merck Serono, Wyeth, Centocor, and Janssen-Cilag. Authors Kievit and Boezeman have no conflicts of interest to declare.

Reprints not available from the authors.

View full text