Nutrition Support Traineeship
Hi folks! This month we combined our Journal club with our bi-annual Nutrition Support Forum (we invite a small group of local dietitians along with our own two nutrition support teams and trainees) and discussed these two articles along with some clinical cases.
I will put the abstracts from Medline below. I will not do a full review, just a short note about what we thought was interesting, and our “take home” message. As always, our goal is not to get you to take our word for it, only to get your interest up to read and dissect it yourself.
The first article (Austrums) on postoperative enteral feeding in pancreatitis was a retrospective review of patients who received nasojejunal feedings (n=33) compared to patients who were NPO with IV fluids after surgery (n= 30) for acute pancreatitis. The major results reported were a significant decrease in postoperative complications and mortality in the patients who received jejunal feeding, compared to those who received the standard therapy (NPO X 10 days).
This study is interesting because:
1. It helps to address the concern raised by some, that previous studies have only “confirmed that TPN is bad”. This study suggests that not only does enteral feeding avoid infections caused by TPN, but that it may actually provides benefit via feeding the gut.
2. The feeding formula was an intact protein, “polymeric” formula.
3. The jejunal feeding group only received an average of 980 calories/day from tube feeding. The authors made no attempt to meet full calorie goals, and only wanted to provide at least 300 mL of formula per day for a minimum of 5 days for gut stimulation. This could suggest that the benefits of enteral feeding in pancreatitis are not related to meeting nutrition needs, but are a direct benefit of feeding the gut.
The down side of this study is that since it is a retrospective review there is an incredible potential for bias to be introduced. The groups appeared reasonably matched, but there was no data provided, or control for factors such as weight status, nutrition status on admission, or antibiotic use for example. Especially with group sizes this small, there is a real chance that some other factor besides the enteral nutrition was the real reason for the outcome differences between the two groups.
Our take home message is that these are interesting preliminary results, and support the need for a larger prospective, randomized trial of polymeric enteral feeding in complicated pancreatitis or in the postoperative setting.
Finney SJ, Zekveld C, Elia A, Evans TW. Glucose control and mortality in critically ill patients. JAMA. 2003 Oct 15;290(15):2041-7.
The second article (Finney) on glucose control and mortality in critically ill patients is a great example of how “data” can be misleading when you do not randomize patients to a therapy! It will be one of the primary articles I will use to teach our dietetic interns about how “bad” research can give you misleading answers if you do not read more than the abstract.
The study was prospective, but it was only observational…it has a similar potential for bias to be introduced as a retrospective study. The authors recorded the outcomes and insulin needs and glucose control of 531 patients admitted to an ICU. They reported that increased administration of insulin was significantly associated with increased mortality. They attempted to control for degree of illness by factoring in APACHE II and SOFA (Sequential Organ Failure Assessment) scores with multivariable logistic regression.
The problem of course, with this sort of thing is that there are variables that could affect outcome that you just did not consider. It would not be a surprise to anyone that the same people who were “sicker” in ways not considered by the scores (had resistant bugs, received less or the wrong antibiotic, etc, etc..) happened to have the worse outcomes AND because they were “sicker” they also required more insulin.
Our take home message is that this study does raise the point that it is not necessarily insulin itself that improves outcome, it is likely the improved blood glucose that improves outcome. It should not, however be used to suggest that insulin drips can compromise outcome, or that high doses of insulin can cause worsen outcome.
"It is astonishing with how little reading a doctor can practice medicine, but it is not astonishing with how badly he may do it." -Sir William Osler