Perioperative parenteral nutrition: a meta-analysis.

We used a meta-analysis protocol to evaluate the results of 18 controlled trials that measured the effectiveness of perioperative total parenteral nutrition. The pooled results of 11 trials that were randomized or quasi-randomized showed trends suggesting that total parenteral nutrition reduced the risk for complications from major surgery (p = 0.21) and fatalities (p = 0.21). Consideration of these pooled estimates of the effectiveness of this intervention must be offset by the poor quality of the trials' methodologies and the iatrogenic complications (pneumothorax, septicemia) that occurred at a pooled rate of 0.067. Alternately, other design flaws, such as the failure to exclude patients who were not malnourished from the trials, may have limited the ability of these trials to show the effectiveness of total parenteral nutrition. The evidence available up to August 1986 shows that the routine use of perioperative total parenteral nutrition in unselected patients having major surgery is not justified; however, this intervention may be helpful in subgroups of these patients who are at high risk.