Is BMI a Risk Factor for Enteral Nutrition‐Related Complications?

Abstract

Background/ObjectiveEnteral nutrition (tube feeding) is standard of care for patients who are unable to tolerate oral intake. Although generally regarded as safe, enteral nutrition devices can result in complications and death. More evidence is needed to clarify enteral nutrition risks. Body mass index (BMI; kg/m2) is a poorly studied risk factor in this regard. With the increasing prevalence of obesity in the US, it is important to better understand the relationship of BMI and enteral nutrition risk. In this retrospective study of electronic medical records, we explored the role of BMI in in‐hospital mortality and length of stay (LOS) among patients receiving enteral nutrition via nasogastric tube (NGT) feeding or percutaneous endoscopic gastrostomy (PEG), enteral access devices inserted via the nasal passages or through the abdominal wall, respectively.MethodsAdult patients (n=1204) who had a PEG or NGT placed in our hospital 2010–2015 were identified by ICD‐9 codes. Age, gender and race‐adjusted linear and logistic regression analyses determined associations between PEG or NGT placement, mortality, log‐transformed LOS, and continuous and categorical BMI. Categorical BMI was divided into underweight, normal weight, overweight, class I, II and III obesity.ResultsThere were 91 deaths (11.29%) in PEG patients (n=806) and 38 deaths (9.5%) in NGT patients (n=398). BMI was not significantly associated with mortality in PEG or NGT patients. Mean LOS for PEG patients increased across the BMI categories (28 days, 33 days, 32 days, 33 days, 42 days and 41 days, respectively) while mean LOS varied across BMI categories for NGT patients. The overall association between BMI and LOS was insignificant (β= 0.00502; p =0.2582); however, when stratified by NGT vs PEG, the association between BMI and LOS was significant for PEG patients (β= 0.0146, p=0.0013 for continuous BMI and β= 0.0939, p=0.0009 for categorical BMI) but was not significant for NGT patients (β= 0.00124, p=0.8654; β= 0.00752, p=0.8566). When BMI was examined categorically, the odds of a PEG patient with class III obesity to have a LOS in the 4th quartile was 2.1 (95% CI 1.083–4.090) times that of a PEG patient with normal weight.ConclusionEnteral nutrition device type modified the association between BMI and LOS (p‐interaction=0.0376 on the multiplicative scale). There was an exponential association between LOS and BMI for PEG patients: for every kg increase in BMI, LOS increased by one percent. For categorical BMI, PEG patients with class III obesity had significantly greater odds of having the longest LOS and, on average, had a LOS two times longer than normal BMI patients. No association was found between BMI and mortality in either PEG or NGT patients. Future research should examine BMI risk in equivalent groups of PEG and NGT patients as well as seek a mechanism for this potential risk factor. The current findings suggest that clinicians should take weight and device type into account when administering enteral nutrition to morbidly obese patients.Support or Funding InformationThis study was funded by National Institutes of Health Training Grant #HL007