Response to Viana MB et al

Paolo Di Bartolomeo (1)

1 Dipartimento di Ematologia

TO THE EDITOR American Journal of Hematology Prof Carlo Brugnara There is no doubt that the observations made by Viana MB et al (Kaplan-Meyer method is inappropriate for estimating cumulative incidence of graft-versus-host disease in the presence of competing events) (American Journal of Hematology, 2010, in press) are relevant. Due to these suggestions, we have estimated again the graft-versus-host data as cumulative incidence and have compared them to those obtained with the Kaplan-Meyer method reported in our paper (Am J Hematol 2008;83:528-530). We found the following differences: Kaplan-Meyer methodCumulative incidence Acute GvHD grade I to IV37%35.6% Acute GvHD – CSA alone27%27% Acute GvHD – CSA + MTX54%P=0.001*51.2% P=0.002** Chronic GvHD17%14.8% *Log-rank test; ** Gray test The overestimation is minimal due the low number of failures of competing risks (early death before the possible occurrence of GvHD and graft rejection). Paolo Di Bartolomeo Stella Santarone Unità di Terapia Intensiva Ematologica per il Trapianto Emopoietico, Ospedale Civile, Pescara, Italy Marta Di Nicola Laboratorio di Biostatistica, Dipartimento di Scienze Biomediche, Università degli Studi, Chieti, Italy

Mots clés: Medicine

Paolo Di Bartolomeo. Response to Viana MB et al. American Journal of Hematology, 2010, ⟨10.1002/ajh.21700⟩. ⟨hal-00552315⟩