Interpretation of results of pooled analysis of individual patient data – Authors' reply