Nutrition Support Blog: Wet sidewalks cause it to rain
February 28, 2012
If I were to stand at a podium during a national or international nutrition conference and state that based on the results of a large observational study1, changing your diet to eat less whole grain foods would result in adults becoming less educated, earn less money each year, and start smoking, it is very likely I would be laughed off the stage. Nevertheless, I have sat at national and international conferences while someone stood at the podium and based on observational studies, stated that critically ill patients who do not receive a certain amount of enteral nutrition (EN) within several days are more likely to die, thus we should add parenteral nutrition (PN) early in their admission. To my amazement, the audience, instead of breaking into uproarious laughter, took notes.
There is general agreement among learned experts (a rare thing in itself) that observational studies should never be used to imply cause and effect. Observational studies only document associations – things that occur together – and there is no way to establish causality from an observational study regardless of what factors are controlled for, such as severity of illness, age, etc. We have learned that it is impossible to statistically control for (or even know) all the factors that influence outcomes, and that it is necessary to do randomized studies to begin to understand causality. In the case of the amount of EN received by critically ill patients, we know that injury severity or APACHE scores do not consider GI dysfunction, and we also know that GI dysfunction has been independently associated with compromised ICU outcomes.2 Inadequate EN provision in the ICU may simply be an indirect marker for GI dysfunction, rather than the lack of nutrition directly influencing outcome. There is a myriad of reasons why patients that will have a poor outcome receive less EN, and we know it is impossible to identify, or statistically account, for all of these factors.
Medical professionals have no difficulty accepting the limitations of observational studies when it is intuitively improbable that one factor could cause another – such as whole grain intake and smoking. Unfortunately, when observational data agrees with our preconceived notions, the mental transmission seems to slip gears and there appears to be a tendency to accept that associations can be cause and effect. Unfortunately we have learned the hard way that making treatment decisions based on observational studies can actually harm patients. Observational studies suggested that women who took hormone replacement therapy had significantly less cardiovascular disease, but randomized studies demonstrated that hormone replacement therapy actually increased cardiovascular disease risk. Adults with increased serum levels of β-carotene had decreased incidence of lung cancer, but large randomized studies of β-carotene supplementation demonstrated a significant increase in cancer incidence compared to placebo. Now we have data that adding early supplemental PN to critically ill patients that are not achieving EN goals resulted in significantly increased infections, time in the ICU, and hospital costs, without any apparent outcome advantage3; demonstrating once again the problems of basing treatment on observational data.
Of course, it is likely that a large cumulative nutrition deficit will eventually have a negative impact on the patient. However, it will take further randomized studies before we understand at what point the benefits of supplemental parenteral nutrition might outweigh the risks.
1) Lang R, Thane CW, Bolton-Smith C, Jebb SA. Consumption of whole-grain foods by British adults: findings from further analysis of two national dietary surveys. Public Health Nutr. 2003 Aug;6(5):479-84.
2) Ukleja A. Altered GI motility in critically Ill patients: current understanding of pathophysiology, clinical impact, and diagnostic approach. Nutr Clin Pract. 2010 Feb;25(1):16-25.
3) Casaer MP, Mesotten D, Hermans G, Wouters PJ,et al. Early versus Late Parenteral Nutrition in Critically Ill Adults. N Engl J Med. 2011 Aug 11;365(6):506-17.
“He uses statistics as a drunken man uses lamp posts - for support rather than illumination.”
- Andrew Lang