INTRODUCTION. Malnutrition is widely implicated as an etiologic factor in a variety of chronic illnesses. Whether malnutrition affects survival following critical care is unknown.
OBJECTIVES. We hypothesized that malnutrition at the time of critical care would be associated with all cause mortality.
METHODS. We performed a two center observational study of patients treated in medical and surgical ICUs in Boston, Massachusetts. All data were obtained from the Research Patient Data Registry at Partners Healthcare. We studied 51,581 patients, age = 18 years, who received critical care between 1997 and 2007. The exposure of interest, malnutrition, was considered to be present if ICD-9 code 260-263.1, 263.8, or 263.9 was assigned to the patient three days prior to or up to seven days after critical care initiation. The primary outcome was all cause 30-day mortality determined by the US Social Security Administration Death Master File. The secondary outcomes were bloodstream infection, as well as 90 and 365 day mortality. Adjusted odds ratios were estimated by multivariable logistic regression models with inclusion of covariate terms thought to plausibly interact with both nutrition status and mortality.
RESULTS. 58.2% of the cohort was male, 79.6% was white, 50.5% were medical patients, 13.6% were septic. The mean ± SD age was 61.7 ± 18.4. The 30-day mortality rate was 14.2%. 1,692 cohort patients were diagnosed with malnutrition. Patients with a malnutrition diagnosis have higher odds of mortality 30 days following critical care initiation, (OR 1.54; 95%CI, 1.36-1.74;P< .0001) relative to patients without a diagnosis of malnutrition. Following adjustment for age, race, gender, Deyo-Charlson Index, and medical vs surgical patient type, the adjusted OR for 30-day mortality was 1.48 (95%CI, 1.31-1.68;P< .0001) relative to patients without a diagnosis of malnutrition. The multivariable adjusted OR for 90-day mortality was 2.03 (95%CI, 1.82-2.27;P< .0001) and the adjusted OR for 365-day mortality was 2.37 (95%CI, 2.14-2.64;P< .0001) all relative to patients without a diagnosis of malnutrition. Further adjustment for creatinine, hemotocrit and white blood count did not alter the effect sizes or significance of the mortality data. In a subgroup analysis of patients who had blood cultures drawn 48 hours prior to or after critical care initiation (n=18,596), diagnosis of malnutrition was associated with increased risk of bloodstream infection (adjusted OR for bloodstream infection was 1.47; 95%CI, 1.25-1.72; P< .0001 relative to patients without a diagnosis of malnutrition).
CONCLUSIONS. In a large population of critically ill adults, a diagnosis of malnutrition is significantly associated with all-cause patient mortality and the risk of bloodstream infection. Further studies are required to confirm our findings, extend the observations to the severity of malnutrition present, and establish mechanisms underlying these observations.