Obesity in critical illness: what weight or why weight?
B. Erstad
DOI: https://doi.org/10.1097/CCM.0b013e3182411720
IF: 8.8
2012-05-01
Critical Care Medicine
Abstract:The relationship between a size descriptor, such as body mass index (BMI), and patient outcomes is complex and requires clinical context. Pooled data from prospective, long-term studies of nonsmoking, healthy, white adults demonstrate that the relationship between mortality and BMI is nonlinear (i.e., U shaped or J shaped), since mortality is higher in more underweight (BMI subsets below 20 kg/m2) and overweight/obese (BMI subsets above 24.9 kg/m2) subjects compared to more normal-sized (BMI of 20 to 24.9 kg/m2) subjects (1). Furthermore, there is a progressive increase in all-cause mortality with each increase in BMI subset (i.e., BMI of 25 to 29.9 kg/m2, 30 to 34.9 kg/m2, 35 to 39.9 kg/m2, and 40 to 49.9 kg/m2). In critically ill patients, the relationship between increasing body size and outcomes is more complex and controversial. Three meta-analyses have been published in the last 3 yrs that attempted to elucidate the impact of obesity on various outcome measures, including mortality (2–4). One meta-analysis did not find an association between obesity and intensive care unit (ICU) mortality based on a dichotomous BMI (30 vs. 30 kg/m2), although duration of mechanical ventilation and length of ICU stay were longer in the obese patients (p .04 and p .009, respectively) (2). A subgroup analysis demonstrated decreased mortality in patients with BMIs of 30 to 39.9 kg/m2 compared to patients with BMIs 30 kg/m2. The second metaanalysis found decreased mortality (combined hospital and ICU) in patients in BMI subsets of 25 to 29.9 kg/m2 and 30 to 39.9 kg/m2 compared to patients with BMIs of 18.5 to 24.9 kg/m2, but the results must be interpreted with caution since statistical heterogeneity was found (3). The third and most recent meta-analysis with the largest number of evaluated studies and patients found no association between obesity and ICU mortality, duration of mechanical ventilation, or length of ICU stay (4). Lower hospital mortality was found in patients with higher BMIs (30 to 39.9 kg/m2 and ≥40 kg/m2) compared to those with BMIs in the range of 18.5 to 24.9 kg/m2 (risk ratio 0.76, 95% confidence interval 0.59 to 0.92 and risk ratio 0.83, 95% confidence interval 0.66 to 1.04, respectively), but patients with BMIs 40 kg/m2 were found to have significantly longer ICU and hospital stays when studies without estimates of SD were removed from the analysis (1.2 days, 95% confidence interval 0.5 to 2.2 days and 2.9 days, 95% confidence interval 1.1 to 4.3 days, respectively). The study by O’Brien and colleagues (5) in this issue of Critical Care Medicine attempted to control for some of the factors that might account for some of the disparate results obtained in previous investigations of obesity and critical illness. This group of experienced researchers has published a number of papers investigating obesity-related issues in critically ill patients. Their current prospective study involved patients in three medical ICUs of two hospitals that were part of an academic medical center. The purpose of the study was to determine the association between fluid-adjusted excess weight and processes of care and outcomes for mechanically ventilated critically ill adults. Patients were enrolled over a 2-yr study period. Processes of care were evaluated at four time points during hospitalization (1, 3, 7, and 28 days). Additionally, study coordinators contacted patients for information every 90 days after discharge until study conclusion, which occurred 2 months after the final day of enrollment. Patient weight was obtained a median time of 22.7 hrs (interquartile range 13.3 to 44.8) from ICU admission by bed scales used for routine clinical care. In contradistinction to most previous studies, a fluid-adjusted BMI was obtained by adjusting for fluid weight gained or lost during the period of hospitalization before weight measurement. The initial BMI subsets compared in the analyses were 25 kg/m2 (referent group), 25 to 30 kg/m2 (overweight), and 30 kg/m2 (obese). However, based on suggestions from peer review of the paper, the obese group was further subdivided into patients with BMIs of 30 to 40 kg/m2 and 40 kg/m2 for additional analyses. The primary study end points were successful extubation and mortality. A competing risks analysis was used during the modeling procedures to account for death before extubation that would preclude evaluation of the latter end point. For results, no significant association was found between fluid-adjusted BMI category and the competing end points for models adjusted by patient demographics, severity of illness, or comorbidities. Compared to the referent group (BMI 25 kg/m2), obese patients (i.e., BMIs 30 kg/m2) were more likely to be successfully extubated when adjustments were made for sedative use (p .047), as defined by continuous infusions of benzodiazepines and/or opioids at enrollment, and ventilator parameters (p .011), as defined by highest tidal volume, positive end-expiratory pressure, and peak airway pressure. The successful extubation results for obese patients remained significant (p .005) in a final model that included sedative use and ventilator parameters in addition to Simplified Acute Physiology Score II score, PaO2/FIO2 at enrollment, and history of obstructive sleep apnea. Further, the successful extubation results were significant when patients were subcategorized based on a BMI of 30 to 40 kg/m2 (p .02) and a BMI 40 kg/m2 (p .02). With respect to mortality, overweight patients (i.e., BMI 25 to 30 kg/m2) were less likely to die compared to patients in the referent group in the model that adjusted for sedative use (p .039) and the final model (p .044) with the parameters described above. Compared to the referent group, patients with BMIs 40 kg/m2 were also less likely to die (p .03) in the final model. Finally, longer-term mortality that included posthospitalization deaths was significantly lower in patients *see also p. 1456.