Hyperglycemia Was Not a Risk Factor for Mortality in Malnourished Critically Ill Patients

Presentation Number: MON 586
Date of Presentation: April 3rd, 2017

Luiza de Azevedo Gross*1, Marina Vercoza Viana1, Ana Laura Jardim Tavares1, Vicente Lobato Costa1, Rafael Baberena Moraes1, Tiago Antonio Tonietto1, Luciana Vercoza Viana2 and Mirela J Azevedo3
1Hospital de Clinicas de Porto Alegre, 2HCPA - Hospital de Clinicas de Porto Alegre, PORTO ALEGRE, 3Hospital de Clinicas de Porto Alegre, Brazil


Background: Critically ill patients with low BMI (<20 kg/m2) present higher mortality than eutrophic and overweight patients (1). Optimizing nutrition support can improve their outcomes but frequently is associated with hyperglycemia, a condition linked to an increased morbi-mortality. Also, glycemic variability has been associated with high mortality rates (2). The aim of this study was to evaluate in critical ill patients with low BMI the association of glycemic control with nutritional support and mortality.


Methods: In this prospective cohort, critically ill patients with low BMI had their caloric and protein intakes evaluated at 48-72h after have been admitted in a general ICU. In the same occasion patients were divided according to glycemic status as estimated by capillary glucose (CG) measurements. CG measurements was performed as prescribed by the assistant physician, at minimum every 6 hours. Maximum and minimum registered CG values were used to classify patients as hyperglycemic (CG>180 mg/dl), hypoglycemic (CG<70 mg/dl), or normoglycemic (CG=70-180 mg/dl). Glycemic variability was defined as the difference between maximum and minimum CG value in the same day. The cohort was followed until death or hospital discharge.


Results: Prevalence of low BMI in 2810 screened patients was 6%. Therefore, 167 patients were included (age 54.4±17.6 years, 60.8% males, BMI 17.7±2.2 kg/m2) and followed during 21 days (3-217). Their maximum and minimum CG (median) were 160 mg/dl (86-443) and 106 mg/dl (87-130), respectively. About nutritional support, median caloric/protein intakes were 912 kcal/day (0-2313) and 46 g/day (0-106). Cohort mortality was 53.3%. Hyperglycemia occurred in 33.7%, hypoglycemia in 8.4%, and normoglycemia in 40.1% of patients. Hyperglycemia was not correlated with caloric intake (r=0.091; P=0.250). Hyperglycemia was associated with older age (59.5±15 vs.52.2±18.4 years; p=0.007), higher glycemic variability (79[56-132.5] vs. 38[23-53] mg/dl; p<0.001), but not with diabetes mellitus (21.8 vs. 10.7 %, P=0.059). In multivariate analyses, hyperglycemia was associated with age (RR 1.03, 95%CI 1.004-1.050), but not with SAPS3, NUTRIC, or previous diabetes. Mortality was not associated with hyperglycemia or hypoglycemia.


Conclusion: In critically ill malnourished patients, different from eutrophic or overweigh patients, hyperglycemia or hypoglycemia were not associated with caloric intake or mortality.


Nothing to Disclose: LDAG, MVV, ALJT, VLC, RBM, TAT, LVV, MJA