Current issues of ACP Journal Club are published in Annals of Internal Medicine


Total parenteral nutrition did not reduce major complications of nonemergent major abdominal or thoracic surgery in malnourished patients

ACP J Club. 1992 Jan-Feb;116:11. doi:10.7326/ACPJC-1992-116-1-011

Source Citation

The Veterans Affairs Total Parenteral Nutrition Cooperative Study Group. Perioperative total parenteral nutrition in surgical patients. N Engl J Med. 1991 Aug 22;325:525-32.



To investigate if perioperative total parenteral nutrition (TPN) decreases the incidence of serious complications after major nonemergency abdominal or thoracic surgery in malnourished patients.


Randomized, unblinded study of perioperative TPN with 30-day follow-up.


10 Veterans Affairs hospitals.


3259 patients ≥ 21 years of age, scheduled for nonemergency laparotomy or thoracotomy, were reviewed. Exclusion criteria included an expected survival < 90 days, recent TPN use (within 15 days) or surgery (within 30 days), contraindications to the experimental treatment, and major concurrent illness. Nutritional status of 2448 patients was evaluated. 459 of 782 malnourished patients eligible for randomization consented to participate (mean age, 62.9 years; 99% men), and 395 of them had surgery. Follow-up was 100%.


The control group (n = 203) received no TPN or forced enteral feedings before surgery or for 72 hours after surgery. The TPN group (n = 192) received TPN preoperatively (7 to 15 days) and postoperatively (72 hours) through a central venous catheter in doses increasing for 72 hours to a daily caloric goal of 1000 kcal above the resting metabolic expenditure. 550 kcal were supplied as lipid and the rest as dextrose. Crystalline amino acids were provided at a ratio of 150 kcal:1 g of nitrogen.

Main outcome measures

Major operative complications at 30 days. Secondary outcomes were mortality, all complications (major and minor), infectious and noninfectious complications, and major complications stratified by severity of malnutrition.

Main results

Major complication rates in the 2 groups were similar 30 days (TPN group 26% vs control group 25%) and 90 days (both groups 28%) after surgery. More patients had infectious complications during the first 30 days in the TPN group than in the control group (14% vs 6% P = 0.01). {This absolute risk increase of 8% means that 1 additional infectious complication occurred within 30 days for every 13 patients who received TPN (compared with no forced feedings or TPN), 95% CI 7 to 57; the relative risk increase was 120%, CI 18% to 310%.}* The groups did not differ for rates of mortality or noninfectious complications. However, severely malnourished patients (by the Nutrition Risk Index) receiving TPN had fewer noninfectious complications than those in the control group (5% vs 43%, CI 0.02 to 0.91; P = 0.03).


Perioperative total parenteral nutrition did not reduce major complications of nonemergent, major abdominal or thoracic surgery in malnourished patients. More infectious complications occurred within the first 30 days of surgery. Major noninfectious complications may be reduced in severely malnourished patients.

Sources of funding: Department of Veterans Affairs Cooperative Studies Program; Armour Pharmaceutical; KabiVitrum Laboratories; Kendall McGaw Laboratories.

Address for article reprint: Dr. W.O. Williford, Department of Veterans Affairs Medical Research Service, Cooperative Studies Program Coordinating Center, DVA Medical Center, Perry Point, MD 21902, USA.

*Numbers calculated from data in article.


This is a well-designed study addressing an important clinical question regarding the use of TPN. However, there are important qualifiers to the conclusions. Of the 782 patients eligible for randomization, 323 did not participate. Of these, 305 had surgery and 51% had benign disease. Mortality and complication rates were low in this group. The study group was composed mainly of patients with cancer (66%). The TPN study period was stated as 7 to 15 days, but the mean duration of "adequate" caloric intake was only 7.9 days. Although an appropriate intention-to-treat analysis was done, 32% of the TPN group received suboptimal or no TPN (49 and 13, respectively), and 13% of the control group received TPN (3 preoperatively, 24 postoperatively).

Several questions still remain. Did a significant selection bias occur? Do the conclusions drawn apply to surgical patients without cancer? Is 1 week of preoperative TPN sufficient to make a clinical impact (that is, to make up large nutritional deficits)? Did the lack of adequate TPN in the experimental group and addition of postoperative TPN in the control group diminish the measured effect of the therapeutic intervention?

Although the difference was not significant, the control group had double the TPN group's rate of anastomotic leaks and bronchopulmonary fistulas. This finding could be interpreted as evidence of a "healing" effect with TPN. Inexplicably, the bulk of the difference in "major, infectious" complications was in the rate of pneumonia/empyema, despite the higher number of thoracotomies in the control group.

The conclusion to be drawn from this study is that in a group of patients, most of whom had cancer, preoperative TPN for an average of 8 days had no benefit except in severely malnourished patients. Additional data are needed to draw conclusions about other patient groups and longer nutritional support.

Patrick L. Ergina, MD, MPH
McGill UniversityMontreal, Quebec, Canada