Home
|
Inbox
|
Search
|
View Abstract
Challenging the Obesity Paradox in the Medical Intensive Care Unit
Description
.abstract img { width:300px !important; height:auto; display:block; text-align:center; margin-top:10px } .abstract { overflow-x:scroll } .abstract table { width:100%; display:block; border:hidden; border-collapse: collapse; margin-top:10px } .abstract td, th { border-top: 1px solid #ddd; padding: 4px 8px; } .abstract tbody tr:nth-child(even) td { background-color: #efefef; } .abstract a { overflow-wrap: break-word; word-wrap: break-word; }
A6053 - Challenging the Obesity Paradox in the Medical Intensive Care Unit
Author Block: A. Chowdhury, X. Han, X. Wang, A. Duggal; Respiratory Institute, Cleveland Clinic, Cleveland, OH, United States.
Rationale:
The prevalence of obesity in the United States is 36.5% and rising. Obesity is associated with higher rates of chronic illnesses (stroke, diabetes, coronary artery disease), but has been thought to be protective in critical illness (acute respiratory distress syndrome and sepsis). We aim to evaluate the effect of obesity on mortality and to evaluate its impact on the accuracy of Acute Physiology and Chronic Health Evaluation (APACHE) IV mortality prediction at a multicenter hospital system.
Methods:
We conducted a multicenter retrospective study in five medical ICUs during 2014-2015. Variables in the APACHE IV database were combined with median inpatient body mass index (BMI) and median income by ZIP code (a potential confounder). Multiple regression models using BMI as both a continuous and categorical variable (World Health Organization [WHO] classes) were developed. We compared the accuracy of our model's mortality prediction against APACHE IV.
Results:
We analyzed 6,753 patients. The median age 65 [18-105 years], acute physiology score (APS) 38 [0-187 points], BMI 27.4 [11.2-117.4 kg/m2], and income $51,121 [$12,416-156,595] were representative of most critically ill cohorts. The overall unadjusted hospital mortality was 7.6%. In a univariate analysis, there was a statistically significant difference in the observed mortality in each WHO BMI class (p = 0.0074). Our logistic regression model using BMI and income as predictor variables had an accuracy of 0.936 (sensitivity 0.307, specificity 0.986, concordance index 0.884). This compares to APACHE IV well, which had a concordance index of 0.885 with an empirical p-value of 0.28 between the two models. Neither BMI nor income were significant predictor variables (p = 0.113 and 0.068 respectively). Similarly, the accuracy of the APACHE IV mortality estimate was similar in all groups when stratified by both BMI and income (median accuracy 0.933, IQR 0.867-0.971).
Conclusions:
Similar to other recent analyses of critically ill patients, our study demonstrated that the effects of obesity and socioeconomic status were not significant predictors of hospital mortality after adjusting for acute and chronic illnesses. Furthermore, incorporating these terms into a multivariate logistic regression model did not improve its accuracy over the baseline APACHE IV model. This suggests that the “obesity paradox” may be an epiphenomenon that is abolished when granular data about acute and chronic illnesses are available. We plan to subsequently evaluate the effects of BMI and median income on length of stay in this group.
Author Block: A. Chowdhury, X. Han, X. Wang, A. Duggal; Respiratory Institute, Cleveland Clinic, Cleveland, OH, United States.
Rationale:
The prevalence of obesity in the United States is 36.5% and rising. Obesity is associated with higher rates of chronic illnesses (stroke, diabetes, coronary artery disease), but has been thought to be protective in critical illness (acute respiratory distress syndrome and sepsis). We aim to evaluate the effect of obesity on mortality and to evaluate its impact on the accuracy of Acute Physiology and Chronic Health Evaluation (APACHE) IV mortality prediction at a multicenter hospital system.
Methods:
We conducted a multicenter retrospective study in five medical ICUs during 2014-2015. Variables in the APACHE IV database were combined with median inpatient body mass index (BMI) and median income by ZIP code (a potential confounder). Multiple regression models using BMI as both a continuous and categorical variable (World Health Organization [WHO] classes) were developed. We compared the accuracy of our model's mortality prediction against APACHE IV.
Results:
We analyzed 6,753 patients. The median age 65 [18-105 years], acute physiology score (APS) 38 [0-187 points], BMI 27.4 [11.2-117.4 kg/m2], and income $51,121 [$12,416-156,595] were representative of most critically ill cohorts. The overall unadjusted hospital mortality was 7.6%. In a univariate analysis, there was a statistically significant difference in the observed mortality in each WHO BMI class (p = 0.0074). Our logistic regression model using BMI and income as predictor variables had an accuracy of 0.936 (sensitivity 0.307, specificity 0.986, concordance index 0.884). This compares to APACHE IV well, which had a concordance index of 0.885 with an empirical p-value of 0.28 between the two models. Neither BMI nor income were significant predictor variables (p = 0.113 and 0.068 respectively). Similarly, the accuracy of the APACHE IV mortality estimate was similar in all groups when stratified by both BMI and income (median accuracy 0.933, IQR 0.867-0.971).
Conclusions:
Similar to other recent analyses of critically ill patients, our study demonstrated that the effects of obesity and socioeconomic status were not significant predictors of hospital mortality after adjusting for acute and chronic illnesses. Furthermore, incorporating these terms into a multivariate logistic regression model did not improve its accuracy over the baseline APACHE IV model. This suggests that the “obesity paradox” may be an epiphenomenon that is abolished when granular data about acute and chronic illnesses are available. We plan to subsequently evaluate the effects of BMI and median income on length of stay in this group.
|
Home
|
Inbox
|
Search
|