Enteral Nutrition In Critically Iii Patients With Hyperglycemia: A Meta-Analysis Of Randomized Controlled Trials With Trial Sequential Analysis

This systematic review with meta-analysis of randomized controlled trials (RCT) aimed to compare the effect of glycemic control-specific enteral formula specific and standard enteral formula on clinical outcomes in critically ill patients with hyperglycemia. The protocol was performed according to Cochrane Handbook and registered under PROSPERO-CRD42016036495. Databases were searched up to July 2018 using medical subject headings related to enteral nutrition, critically ill patients, hyperglycemia and RCT. The quality of studies was assessed using the six GRADE domains. Two reviewers performed all steps independently (M.C.C.K. and O.F.) and extracted the data of eligible studies through a standard form. A third researcher (F.M.S.) also settled any disagreements in all steps. Meta-analyses were performed at RevMan version 5.3 using the random effect model. The I2 test was performed to evaluate the magnitude of heterogeneity. Furthermore, we performed trial sequential analysis (TSA) of the identified studies (TSA software). In the initial search literature, 178 studies were identified and four of them were eligible. Diabetes-specific enteral formula reduced capillary glycemia (I2=62%; WMD=-25.68 mg/dL, 95% CI -47.21 to -4.17, p=0.02) in comparison to standard enteral formula. Plasma glucose, duration of mechanical ventilation, length of hospital stay, and incidence of death did not differ between groups. There was not publication bias identified. TSA meta-analysis showed a benefit trend, but it is not yet determined as the analysis points to insufficient information to confirm or refute the use of diabetes-specific enteral formula. Specific enteral formula for glycemic control reduced capillary glycemia when compared to the standard formula, but it did not have significant effect in others clinical outcomes in critically ill patients. Disclosure M. Kumbier: None. F.M. Silva: None. L.C. Pinto: None. J. Almeida: None. F. Oellen: None.