Background: Psoriasis affects about 2-3% of the Caucasian population. Biologics such as infliximab, etanercept, adalimumab and ustekinumab are efficacious treatments of plaque-type psoriasis. Critical to monitoring drug efficacy and safety is availability of long-term data. Despite the chronic nature of psoriasis, to date limited long-term clinical data have been available, as challenges are inherent in conducting a long-term analysis. With increasing time, it is more likely that the number of patients discontinuing treatment will also increase, due to loss of efficacy, adverse events or loss to follow-up. Interpretation of these data becomes confounded when one must consider missing data. Several approaches to analysing long-term data exist, and each accounts for missing data differently.
Objective: To demonstrate that the choice of a particular analysis method to account for missing data has great impact on the assessed response rate.
Methods: We used data from an open-label study over 3 years of continuous treatment with infliximab in patients with plaque-type psoriasis. These data were analysed by three methods--last observation carried forward, observed values and non-responder imputation--to account for missing data.
Results: The 3-year PASI 75 responses varied from 41 to 75%, depending on the method of analysis; this shows that the response rate can almost double when a more liberal analytical approach is used.
Conclusions: While it is clear that the need for long-term data on biologics in psoriasis is great, considering the analysis undertaken is important when designing long-term studies and interpreting the resulting data. When analysis methods such as observed values only or last observation carried forward are used, the results of the more conservative non-responder imputation should also be presented to give a fair overview of the long-term efficacy of a treatment for plaque-type psoriasis.
Copyright © 2010 S. Karger AG, Basel.