Statistical aspects in studies of preoperative fluid intake and gastric content

Abstract

Twelve randomised studies of preoperative intake of clear fluids and gastric content were reviewed and presented together with a separate study performed by the authors. A common deficiency in design was the omission of detailed information concerning randomisation procedures, distribution of data, and choice of statistical methods. Only one reviewed study included power calculations and only 3 reported 95% C.I. (Confidence Intervals). Deficiencies in design and data collection rendered the conclusion of two studies invalid. Incorrect use of statistical tests in some of the remaining studies may have resulted in non-significant differences becoming statistically significant. This, however, did not change the main conclusion of a meta-analysis that intake of clear fluids up to 2 hours before general anaesthesia was safe as fluid intake either decreased or had no effect on gastric fluid volume.