Authors' reply.

Sir,

We appreciate the comments by Nast[1] on our recent article[2] and we are happy to respond to queries raised by him on the statistical analysis. We are grateful for his comments that this study has to a certain extent filled a lacuna in the literature comparing a conventional topical therapy with a new topical modality.

In response to the query raised on the reason for the drop out, we would like to clarify that all except five dropouts were patients who did not return for follow-up visits despite instructions to follow-up. These patients could not be contacted despite repeated attempts. Five patients; three in Group A (coal tar group) and two in Group B (calcipotriol + betamethasone group), were withdrawn from the study at week 8 because they either failed to achieve at least 25% reduction in psoriasis area and severity index (PASI) at week 8 or worsened from their baseline PASI score. We would like to state that the analysis of mean PASI reduction and physician global assessment (PGA) scores at each visit was performed by per protocol analysis while the comparison of percentage of patients achieving PASI 50 and PASI 75 are a result of intention to treat analysis. We agree with the comments that because of the high dropout rate, it is important to account for these patients and hence we hereby present the analysis with nonresponder imputation.

As per this analysis, median percentage reduction in PASI (mPASIr) was significantly higher in Group B (calcipotriol + betamethasone ointment) compared with Group A (coal tar) at week 2 (9.3% [range: −46.7–73%)] in Group A vs. 30.2% [range: −33.3–100%]) in Group B (P = 0.036), but not at subsequent follow-up visits. There was no significant difference in mean PASI reduction at week 4 (mPASIr A = 8.82% [range: 0–89%] and mPASIr B = 18.2% [range: −34.6–100%]; P = 0.29), week 8 (mPASIr A = 0% [range: −4–100%] and mPASIr B = 8.33% [range: −33.3–100%] P = 0.98) and week 16 (mPASIr A = 0% [range: 0–100%] and mPASIr B = 0% [range: −66.7–100%]; P = 0.31). As per intention to treat analysis, mean PGA scores were significantly lower in Group B compared with Group A at week 2 (7.3 and 8.6 respectively) and week 4 (6.4 and 8.8 respectively); (P = 0.019 and P = 0.004 respectively), but not at subsequent follow-up visits. Thus, the significance of results did not differ with respect to the previously performed analysis in the study.

We do hope that our reply clarifies the statistical remarks presented in the letter by Nast.