Thirty-four RCTs were included in the review (number of patients unclear). Four trials compared ertapenem versus piperacillin/tazobactam. One trial compared cefepime versus imipenem/cilastatin. Twenty six trials compared imipenem/cilastatin versus meropenem. Three trials compared imipenem/cilastatin versus piperacillin/tazobactam. There were no trials that compared: ertapenem versus cefepime, imipenem/cilastatin or meropenem; meropenem versus cefepime or piperacillin/tazobactam; and cefepime versus piperacillin/tazobactam.
Meropenem had the highest clinical response rate, with a probability of 91.6% that this was the best treatment in the network of comparisons (OR 1.52, 95% credible interval 1.23 to 1.87). Meropenem was also the best treatment in terms of bacteriological response (OR 1.45, 95% credible interval 1.15 to 1.80) and overall reduced risk of serious adverse events (OR 0.88, 95% credible interval 0.76 to 1.02), including those that led to withdrawal (OR 0.73, 95% credible interval 0.42 to 1.20) and those that were gastro-intestinal related (OR 0.76, 95% credible interval 0.56 to 1.02). Ertapenem was preferable in terms of all-cause mortality (OR 1.15, 95% credible interval 0.39 to 2.67). The authors also presented probability rankings as part of the analysis (reported in the paper).
Sensitivity analysis produced similar results using the random-effects model.