Total parenteral nutrition and cancer clinical trials.

Twenty-eight prospective randomized controlled clinical trials evaluating the use of total parenteral nutrition (TPN) in cancer patients were identified through a search of major indexing sources. The data were pooled across studies to increase the ability to detect therapeutic effects. The impact of publication bias and the quality of reporting each trial were used to critically assess the conclusions drawn from the pooled analysis. The authors conclude that TPN may be useful when used preoperatively in patients with gastrointestinal tract cancer. It appears to be beneficial in reducing major surgical complications (pooled P = 0.01) and operative mortality (pooled P = 0.02). No statistically significant benefit from TPN could be demonstrated in survival, treatment tolerance, treatment toxicity, or tumor response in patients receiving chemotherapy or radiotherapy. An increase in the risk of developing an infection in chemotherapy patients given TPN (pooled P less than 0.0001) underscores the importance of demonstrating significant benefits in randomized trials before TPN is used routinely in these patients.