Body Mass Index and Mortality Among Adults Undergoing Cardiac Surgery

Abstract

I n an apparent paradox, obesity, an important risk factor for cardiovascular death, 1,2 is associated with reduced mortality after cardiac surgery. 3 Similar observations have been described in patients with acute coronary syndromes 4 or heart failure 5 and those requiring dialysis. 6 It is unclear whether this simply reflects the limitations of epidemiological analyses or whether there may be actual protective factors associated with obesity that contribute to improved outcomes. The obesity paradox has been attributed to reverse epidemiology (causation) or collider bias whereby the survival benefit associated with obesity actually reflects worse outcomes in underweight patients who also have frailty, cachexia, or severe chronic disease. 7 Alternative hypotheses are that obese patients are selected for surgery only if they are subjectively at lower risk and have high body mass index (BMI) but no metabolic syndrome with its related complications. 8 We report the results of 2 related studies in patients undergoing cardiac surgery: a cohort study of UK and Ireland cardiac surgery audit data and a systematic review with meta-analysis of this and other similar studies that have considered the relationship between BMI and mortality. The aim of these studies was to assess whether the obesity paradox in cardiac surgery can be attributed to reverse epidemiology, bias and confounding , or other mechanisms. METHODS Observational Study Cohort Prospectively collected data were extracted from the National Adult Cardiac Surgery Audit (NACSA) registry (version 4.1.2) of the National Institute for Cardiovascular Outcomes Research on December 1, 2014, for all cardiac operations performed in the United Kingdom and Ireland. These data are collected prospectively and undergo robust validation and checking procedures to maintain data quality. 9-12 Duplicate records and nonadult cardiac surgery entries were removed; transcriptional discrepancies were harmonized; and clinical and temporal conflicts and extreme values were corrected or removed. 11 No attempt to replace missing values was made. The need to obtain informed consent was waived because patients' identifiable information was either removed or pseudonymized. The study was approved by the National Institute for Cardiovascular Outcomes Research NACSA Research Board (study reference 14-ACS-29) and complies with the Strengthening the Reporting of Observational Studies in Epidemiology reporting requirements for observational studies (Appendix I in the online-only Data Supplement). 13 Study Design We performed a retrospective, observational cohort study encompassing all adult cardiac surgical procedures performed in the United Kingdom and Ireland between April 1, 2002, and March 31, 2013. For each operation, data were recorded on patient characteristics and demographics, comorbidities, intraoperative factors, and postoperative outcomes. Administrative data were also extracted. The analysis data set was obtained by including all cases with complete data on a set of key preoperative, intraoperative, and post-operative variables as follows: age, BMI, sex, left ventricular ejection fraction category, history of myocardial infarction, renal impairment, diabetes mellitus on medication, previous cardiac surgery, operation type, and cardiopulmonary bypass use. Patients undergoing salvage surgical procedures (car-diac arrest before induction), patients with critical preopera-tive state (ventilated, cardiogenic shock, inotropic support, intra-aortic balloon pump), and patients with stage 5 chronic kidney disease (dialysis) were excluded. Patients for whom it was not possible to calculate the BMI or for whom the sex of the patient, operation type, or discharge status was missing were also excluded. Study Outcomes, Exposures, and Confounding BMI was defined as the weight in kilograms divided by the square of the height in meters. 14 According to the World Health Organization (WHO) classification, 15 BMI was further categorized into 6 classes: underweight (BMI <18.5 kg/m 2), normal weight (BMI 18.5-<25 kg/m 2), overweight (BMI 25-<30 kg/ m 2), obese class I (BMI 30-<35 kg/m 2), obese class II (BMI 35-<40 kg/m 2), and obese class III (BMI ≥40 kg/m 2). The primary end point was in-hospital mortality, defined as death in hospital after the index surgical procedure and before transfer from the cardiac surgery unit as per the definition used in the national audit. Potential confounders prespecified in our analyses included severe chronic disease: chronic lung Clinical Perspective What Is New? • In a nationwide cohort study of 401 227 adult patients and a systematic review of 557 720 patients from 13 countries, we demonstrated that overweight and obese patients had improved outcomes after cardiac surgery compared with normal-weight patients. • Subgroup and sensitivity analyses designed to mitigate the effects of likely sources of bias and confounding did not affect our estimates that demonstrated reductions in mortality with increasing levels of obesity. • Analysis of secondary outcomes indicated that obesity also had divergent associations with important causes of death. What Are the Clinical Implications? • The present findings do not support common practice whereby weight loss is recommended before surgery in the morbidly obese or very obese patients are refused surgery. • These results suggest a new area for research into strategies that may minimize organ failure in cardiac surgery and other clinical settings characterized by acute surgical metabolic stress.