Parenteral nutrition in cancer patients undergoing chemotherapy: a meta-analysis.

Despite numerous controlled clinical trials, the indications for parenteral nutritional support of cancer patients undergoing chemotherapy remain controversial. The results of the published trials have therefore been subjected to meta-analysis. The pooled data show that total parenteral nutrition (TPN) has a detrimental effect: best estimates associate TPN use with trends to decreased survival (relative risk 0.81; 95% confidence limits [CL] 0.62, 1.0) and poorer tumor response (odds ratio 0.68; 95% CL 0.40, 1.1), with a significant increase in infectious complications (odds ratio 4.1; 95% CL 2.4, 6.9), and no clinically significant effect on gastrointestinal or hematologic toxicity (data not amenable to quantitative analysis). The increased risk of infection persists when catheter-related septicemia is excluded, suggesting that TPN itself may increase susceptibility to infection. In summary, for the patients in trials to date, TPN was associated with net harm. Routine use of TPN in patients undergoing chemotherapy should be strongly discouraged, and trials involving specific groups of patients or modifications in TPN should be undertaken with caution.