Bye and coworkers1 have provided extremely valuable data on the efficacy of corticosteroid therapy in treating Pneumocystis carinii pneumonia (PCP) in children infected with human immunodeficiency virus. However, I believe that some of the statistical analyses shown in their Table are incorrect and that the errors have influenced some of their conclusions. Using their data, I performed χ2 analysis using Epi Info, Version 5,2 the statistical software package distributed by the Centers for Disease Control and Prevention, Atlanta, Ga. The results are shown in the Table.
This reanalysis shows that there is not a significant difference in the incidence of adverse reactions to a combination of trimethoprim and sulfamethoxazole between the two groups or a difference in the proportion of patients who presented with PCP as the initial manifestation of human immunodeficiency virus infection. The reduction in the proportion of patients requiring mechanical ventilation is of marginal