RESUMO
BACKGROUND: Gastrointestinal (GI) motility disorders may be directly associated with the intensity of acute brain injury, edema of the brainstem, and opioid use in neurosurgical patients. METHODS: In this retrospective study, patient demographic characteristics, computed tomography (CT) scans, the occurrence of gastroparesis, constipation, and opioid use were registered during the intensive care unit (ICU) stay and correlated with days of mechanical ventilation, length of ICU stay, and survival. Gastroparesis was defined as residual gastric volume > 250 mL per day, and constipation was defined as the absence of stool for 3 days or more during the ICU stay. RESULTS: Of 207 neurosurgical patients screened, 69 adult patients who spent more than 4 days in the ICU were included in the study. Gastroparesis was observed in 48 (69.6%) patients, constipation was observed in 67 (97.1%) patients, and stress ulcers were observed in 4 (5.8%) patients. Patients with brainstem edema (n = 57, 82.6%) had the first stool evacuation later compared with patients with no edema (8 [interquartile range (IQR) 5.25-9.75] vs. 3.5 [IQR 2.25-4] days; P < 0.001). In the logistic regression analysis, factors that were associated with GI dysmotility were central nervous system (CNS) bleeding (odds ratio [OR] 5.1, 95% confidence interval [CI] 1.26-20.8, P = 0.02), opioid use > 19.3 morphine equivalents (ME) per day (OR 5.37, 95% CI 1.1-27.1, P = 0.04), and brainstem edema (OR 4.9, 95% CI 1.1-21.6, P = 0.04). A receiver operating characteristic curve analysis confirmed that the cutoff value of > 6.78 ME per day was a good predictor determining GI dysmotility, with 89.5% sensitivity and 72.7% specificity (95% CI 0.67-0.88, area under the curve 0.784, Youden index 0.62, P = 0.001). Poor survival correlated with lower Glasgow Coma Score values (ρ = - 520, P < 0.001), CNS bleeding (ρ = 0.393, P < 0.001), associated cardiac diseases (ρ = 0.279, P < 0.001), and cardiorespiratory arrest on admission (ρ = 0.315, P < 0.001), but not with GI dysmotility (ρ = 0.175, P = 0.402). CONCLUSIONS: Significant correlation was registered between brainstem edema, gastrointestinal dysmotility, and opioids. CNS bleeding was the most important single factor influencing GI dysmotility. Further studies with opioid and nonopioid sedation may distinguish the influence of acute brain lesions versus drugs on GI dysmotility.