Ref (type)	Population	Outcome, Interventions	Results and statistical analysis	Effect size	Favours
Adverse effects
RCT Crossover design	41 people with a proven first episode of PCP	Major adverse effects 10/20 (50%) with intravenous pentamidine 11/20 (55%) with TMP–SMX	P = 0.9		Not significant
RCT Crossover design	187 people with suspected PCP; 163 people with confirmed diagnoses (only those with confirmed diagnoses and who gave consent were evaluated further)	Proportion of people changing treatment because of adverse effects 17/68 (25%) with intravenous pentamidine 31/92 (34%) with TMP–SMX	P = 0.23		Not significant
RCT Crossover design	70 people with confirmed or presumed PCP	Adverse effects requiring cessation of treatment with intravenous pentamidine with TMP–SMX