Total parenteral nutrition vs enteral nutrition: no difference in critically ill? (CALORIES trial)

Feeding patients enterally (nasogastric or nasojejunal tube feedings) has been the standard of care for critically ill patients, based on weak evidence that it reduces infection rates; hence the adage "feed the gut, if you can." That last caveat is included because so many critically ill patients have gastric motility impairment (with inability to achieve necessary throughput of feedings) or are in vasopressor-dependent shock (making their doctors fearful of prescribing full feeding rates). In these patients, some believe that early total parenteral nutrition could be beneficial because it more often delivers enough calories to meet a theoretical target, or could reduce the risk of aspiration (this has never been shown).

Prior evidence has not borne out these suppositions: both randomized trials and observational evidence have suggested there is no benefit of early TPN in critical illness, even if that means enterally-fed patients "go hungry"(fail to achieve theoretical caloric targets) for up to 7 days.

The CALORIES trial, recently published in the New England Journal of Medicine, does not change that, and in fact does not give us much information at all, making me wonder why it's in NEJM.

Authors enrolled 2,400 patients in 33 ICUs in England, expected to stay for 3+ days, and randomized them to either enteral or parenteral nutrition for up to 5 days. The hypothesis was that more of the patients given TPN would get their full nutrition needs early, and have a 20% relative risk reduction in mortality -- a pretty optimistic hope given the preexisting literature on the subject.

If you read the headline, you already know that didn't happen. There was no difference in 30-day mortality between groups (33% vs 34%). Since no randomized trial I know of has ever shown any measurable improvement in outcomes from any nutritional strategy in critically ill patients, that's not that surprising.

But what's most vexing here in trying to conclude anything is that the patients receiving TPN got almost the same total calories as those getting enteral nutrition. Less than half the patients in both groups achieved their caloric goals. Since providing more calories was central to the hypothesized benefit of early TPN -- what other conceivable advantage could it have? -- this is a startling result. How could this happen? Because this was a pragmatic trial design, authors can't say. Pragmatic trials have a lower level of centralized command-and-control -- in this case there were no standardized preparations shipped out, or unified protocols to be adhered to, etc.; the participating sites were just supposed to follow the general principles of the trial using their local processes and practices.

Possible contributory reasons as to why the parenteral route did not meet its caloric target include lack of availability of nutritional product, content (the use of commercially available rather than individually titrated product), delivery (delays or interruptions in delivery for procedures, transfers, patient factors, etc.), and clinical preference.

In other words, who knows?

This seems like a problem on par with not enrolling enough patients to reach the necessary statistical power to draw a conclusion. As such, it doesn't alter the equipoise on enteral vs parenteral feedings in critical illness.

There were no more infections in the parenteral nutrition group than in the enteral feeding group. However, as suggested in previous randomized trials, infections may occur in a dose-response relationship to the amount of TPN infused. Since the TPN patients didn't get more calories in CALORIES, it's impossible to conclude the real-world infection risk from "full dose" TPN is lower than previously supposed. Fewer patients receiving TPN vomited or had hypoglycemia, but there was no difference noted in aspiration events or aspiration pneumonia.

The CALORIES trial is a good example of the risk of "brand bias" when reading the literature. Because this article made it into the NEJM, many people will assume it is an epic landmark trial establishing the equivalency of enteral and parenteral nutrition, when (because of the lack of divergence in calories achieved between the 2 arms) it actually shows no such thing. Because of that fundamental shortcoming, I'm really not sure what it shows.

Clinical Takeaway: PulmCCM periodically updates a wider survey of the literature in nutrition during critical illness; CALORIES will take its place in that inconclusive muddle. The general absence of signal of benefit or harm with any particular approach to nutrition in critical illness suggests that nutritional strategy doesn't influence outcomes much, our intuitions and preoccupation with food notwithstanding.

Harvey SE et al for the CALORIES investigators. Trial of the Route of Early Nutritional Support in Critically Ill Adults. The New England Journal of Medicine, October 1, 2014.