PURPOSE: The objective of this study was to retrospectively analyze the association of mean glucose level (MGL) and glycemic lability index (GLI; as a measure of glucose variability) with intensive care unit (ICU) mortality in patients with severe acute pancreatitis (SAP). MATERIALS AND METHODS: Paper-based medical records of patients with SAP who were admitted to the ICU of West China Hospital between July 1, 2005, and July 1, 2010, were analyzed. Glucose measurements, demographic characteristics, clinical features, data on the first and second 24-hour Acute Physiology and Chronic Health Evaluation (APACHE) II scores, and outcomes were obtained. Time-weighted glucose parameters were used. We statistically analyzed the relationship between these variables and both ICU and hospital mortality. RESULTS: A total of 294 patients with 34,796 glucose measurements were included in the final analysis. The time-weighted MGL was 9.31 ± 1.91 mmol/L, and the median of GLI was 55.27 (mmol/L)(2) h-(1) wk-(1). Intensive care unit mortality was 43.5% and increased progressively as GLI increased, reaching 62.5% of patients with GLI above 115.89 (mmol/L)(2) h-(1) wk-(1). The highest odds ratio for ICU death was found in patients with the highest quartile of GLI: odds ratio, 3.47 (95% confidence interval, 1.76-6.86; P < .000). No such relationship could be found with MGL. Glycemic lability index was better able to predict ICU death than was MGL (the area under the curves were 0.642 vs 0.561, respectively; z test was 2.677; P = .0074). The logistic regression analysis showed that GLI, the second 24-hour APACHE II score, and the number of organ failures upon ICU admission contributed independently to the risk of mortality. CONCLUSIONS: We observed that GLI was a better predictor of ICU and hospital mortality than was MGL. Together with the second 24-hour APACHE II score and the number of organ failures upon ICU admission, GLI is an independent predictor of mortality in patients with SAP.
PURPOSE: The objective of this study was to retrospectively analyze the association of mean glucose level (MGL) and glycemic lability index (GLI; as a measure of glucose variability) with intensive care unit (ICU) mortality in patients with severe acute pancreatitis (SAP). MATERIALS AND METHODS: Paper-based medical records of patients with SAP who were admitted to the ICU of West China Hospital between July 1, 2005, and July 1, 2010, were analyzed. Glucose measurements, demographic characteristics, clinical features, data on the first and second 24-hour Acute Physiology and Chronic Health Evaluation (APACHE) II scores, and outcomes were obtained. Time-weighted glucose parameters were used. We statistically analyzed the relationship between these variables and both ICU and hospital mortality. RESULTS: A total of 294 patients with 34,796 glucose measurements were included in the final analysis. The time-weighted MGL was 9.31 ± 1.91 mmol/L, and the median of GLI was 55.27 (mmol/L)(2) h-(1) wk-(1). Intensive care unit mortality was 43.5% and increased progressively as GLI increased, reaching 62.5% of patients with GLI above 115.89 (mmol/L)(2) h-(1) wk-(1). The highest odds ratio for ICU death was found in patients with the highest quartile of GLI: odds ratio, 3.47 (95% confidence interval, 1.76-6.86; P < .000). No such relationship could be found with MGL. Glycemic lability index was better able to predict ICU death than was MGL (the area under the curves were 0.642 vs 0.561, respectively; z test was 2.677; P = .0074). The logistic regression analysis showed that GLI, the second 24-hour APACHE II score, and the number of organ failures upon ICU admission contributed independently to the risk of mortality. CONCLUSIONS: We observed that GLI was a better predictor of ICU and hospital mortality than was MGL. Together with the second 24-hour APACHE II score and the number of organ failures upon ICU admission, GLI is an independent predictor of mortality in patients with SAP.
Authors: Chirag J Jivanji; Varsha M Asrani; Sayali A Pendharkar; Melody G Bevan; Nicola A Gillies; Danielle H E Soo; Ruma G Singh; Maxim S Petrov Journal: Dig Dis Sci Date: 2017-03-14 Impact factor: 3.199