Association of Preoperative Body Weight and Weight Loss With Risk of Death After Bariatric Surgery

Key Points Question Are preoperative body mass index and weight loss associated with 30-day mortality after bariatric surgery? Findings In a cohort study of 480 075 patients who underwent bariatric surgery from 2015 to 2017, even modest weight loss before bariatric surgery was associated with lower risk of 30-day mortality after the procedure. Compared with patients with no preoperative weight loss, patients with weight loss greater than 0% to less than 5.0%, 5.0% to 9.9%, and 10.0% and greater had 24%, 31%, and 42%, respectively, lower risk of 30-day mortality. Meaning In this study of patients who underwent bariatric surgery, even moderate weight loss (ie, >0% to <5%) before the procedure was associated with lower risk of 30-day mortality; these findings may help to inform future updates of clinical guidelines regarding bariatric surgery.


Introduction
Obesity is a rising epidemic in the US and worldwide. Obesity is associated with increased all-cause mortality, and more than 300 000 deaths annually in the US are attributed to obesity. 1,2 In addition to the health consequences of obesity, the direct medical costs related to adult obesity in the US has been estimated at more than $315 billion yearly. 3 Bariatric surgery has been shown to be the most effective and durable treatment for clinically morbid obesity (ie, body mass index [BMI; calculated as weight in kilograms divided by height in meters squared] Ն35.0 with comorbidities), which is difficult to reverse through traditional approaches such as lifestyle intervention. 4 There has been long-standing uncertainty and debate regarding the value of preoperative weight loss as a requirement for bariatric surgery. Perception of preoperative weight loss requirements varies among patients, physicians, hospitals, and health insurance payers. Current clinical guidelines by the American Society for Metabolic and Bariatric Surgery do not recommend preoperative weight loss because of a lack of scientific support regarding its benefits. 5 Clearly, there is an urgent and critical need to clarify the potential influence of preoperative weight loss on postoperative outcomes after bariatric surgery.
One of the major indicators for surgical outcomes is 30-day mortality after the procedure. While it is important to aim for reducing the death rate to 0 for any surgical procedure, this is especially important for bariatric surgery because most bariatric operations are voluntary and patients are not expected to die in the short term if the surgery is not performed. Several previous studies have examined the association of preoperative weight loss with postoperative weight loss, and they have yielded inconsistent findings. [6][7][8] So far, the association of preoperative weight loss with 30-day mortality after bariatric surgery remains unclear. 9 Previous studies pertaining to bariatric surgery outcomes usually have limited sample sizes or limited procedure types, which has compromised their ability to detect an association with a rare outcome, such as death. 9,10 Therefore, we aimed to examine the associations of preoperative BMI and weight loss with 30-day mortality following bariatric surgery using a large database of approximately 500 000 patients who underwent bariatric surgery in the United States and Canada.

Study Population
We used data from the Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP) Participant Use Data File (PUF) from 2015 to 2017. 11 The PUF includes data that were collected prospectively from 832 participating centers between January 1, 2015, and December 31, 2017, in the United States and Canada. These data are abstracted from medical records by certified metabolic and bariatric surgical clinical reviewers who undergo database-specific trainings and must pass an annual certification exam. Preoperative and demographic variables, comorbidities, information on the procedure, and 30-day outcomes are collected. Data integrity is ensured by an audit process requiring a disagreement rate of less than 5%. The MBSAQIP covers more than 90% of all bariatric surgery programs in the United States and Canada. 11 The PUF from 2015 to 2017 includes 555 239 cases performed in all MBSAQIP-participating institutions during that period. [12][13][14][15] The University of Iowa institutional review board determined that this analysis was exempt from review and informed consent given the use of deidentified data. The present study was performed and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.
For this analysis, we initially identified 505 287 patients aged 18 years or older with complete information on preoperative weight measurement and 30-day vital status. We excluded 22 277 patients (4.4%) whose highest recorded preoperative BMI was less than 35.0 and 2935 participants (0.6%) with weight gain before surgery, resulting in 480 075 patients (95.0%) in our analysis.

Exposure Assessment
The highest recorded preoperative weight of the patient was measured at the bariatric center within 1 year before the principal operation. The preoperative weight closest to the day of surgery was defined as the most recent weight documented in the medical record within 30 days before the principal operation or at the time the patient was being considered a candidate for surgery. Height was the patient's most recent height documented in the medical record within 30 days before the principal operation or at the time the patient was being considered a candidate for surgery. Weight loss (in pounds) was defined as the difference between the highest recorded preoperative weight and preoperative weight recorded closest to the procedure. The BMI reduction was the difference between the highest recorded preoperative BMI and preoperative BMI closest to the day of the operation.

Outcome Assessment
The outcome was death within 30 days of bariatric surgery. This information was retrieved from the reporting of mortality up to and including the 30th day after the operation in all MBSAQIPparticipating institutions.

Covariate Assessments
Demographic details were collected for all patients, including age, sex, and race/ethnicity. Patient race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Hispanic, and other.
Patients were categorized as currently smoking tobacco or not currently smoking tobacco based on information about smoking cigarettes at any point within the 12 months before admission for surgery. Patient comorbidities included diabetes; hypertension; hyperlipidemia; obstructive sleep apnea; chronic obstructive pulmonary disease; previous percutaneous coronary intervention or percutaneous transluminal coronary angioplasty; previous major cardiac surgery; renal insufficiency, receiving dialysis, or requiring dialysis 2 weeks before the principal operation; deep vein thrombosis requiring therapy; presence of inferior vena cava filter; and therapeutic anticoagulation. The number of comorbidities was further categorized as 0, 1, or 2 or more.

Statistical Analysis
Comparisons of covariates among different groups were performed using analysis of variance for continuous variables and χ 2 tests for categorical variables. Dummy categories were created for missing values. We used multivariable logistic regression to estimate odds ratios (ORs) and 95% CIs for 30-day mortality associated with preoperative BMI and preoperative weight loss or BMI reduction. We fit 3 different adjusted models. Model 1 was adjusted for age, sex, and race/ethnicity. Model 2 was adjusted for covariates included in model 1 plus Current Procedural Terminology principal operation, history of bariatric surgery, whether the procedure was a revision or conversion, and whether the patient underwent emergency surgery during the hospital admission. Model 3 was adjusted for covariates included in model 2 plus preoperative smoking status and comorbidities. For the analysis of preoperative weight loss or BMI reduction with mortality risk, the highest recorded preoperative BMI was also adjusted in the model. Tests for linear trends were conducted across categories of preoperative BMI or weight loss by assigning the median value for each category and fitting this continuous variable in the models.
We performed sensitivity analyses by excluding patients who had previous bariatric surgery, revision bariatric surgery, or emergency surgery and restricting the analysis to 2 main bariatric types (ie, Roux-en-Y gastric bypass and sleeve). Data analysis was performed from December 2018 to November 2019, using survey procedures in SAS statistical software version 9.4 (SAS Institute). P < .05 was considered statistically significant, and all tests were 2-tailed.  Table 3).

JAMA Network Open | Surgery
Greater weight loss before bariatric surgery was associated with lower risk of 30-day mortality.

Discussion
In this large binational cohort study of patients who underwent bariatric procedures, we found that higher preoperative BMI was associated with a higher risk of 30-day mortality following bariatric surgery after adjusting for a variety of covariates. Furthermore, weight loss, even at a moderate degree (ie, >0% to <5% weight loss), was associated with significantly lower risk of 30-day mortality following bariatric surgery.
To our knowledge, this is the largest data set analyzed to investigate the dose-response association of preoperative weight and weight loss with risk of 30-day mortality. Previous studies have examined the association of preoperative weight loss with several surgical outcomes, such as intraoperative and postoperative complications, operating time, and length of hospital stay. 6,7,16,17 The association of preoperative weight loss with 30-day mortality was investigated in several studies; however, the findings have been inconclusive. 9 A 2019 meta-analysis 9 identified 4 randomized clinical trials and 12 cohort studies, but the pooled sample size (6060 patients) was insufficient to  have a robust estimate for the association of preoperative weight loss with risk of mortality because the mortality event rate was low. 9 Another observational study 10 did not find a significant association of preoperative weight loss as a binary category (ie, yes or no) with 30-day mortality, probably because of restricted surgery types and thus restricted sample size. 10 There are several potential explanations for our results. Previous studies have confirmed the beneficial association of weight loss with comorbid medical conditions, including predisposition to thromboembolism. 18,19 Venous thromboembolism, including deep vein thrombosis and pulmonary embolism, is a leading cause of postoperative mortality in patients undergoing bariatric surgery 20 ; thus, improvements in blood coagulopathy and fibrinolysis and an increase in antithrombin III associated with weight loss may be important factors in reducing postoperative mortality. 21 Furthermore, modest weight loss has been shown to significantly reduce apneic episodes and improve gas exchange, 22,23 which may in turn reduce the risk of perioperative death. 24 Another possibility is that morbidity and mortality are associated with greater technical difficulty and complexity of surgery in patients with obesity. [25][26][27] Preoperative weight loss leads to a reduced liver volume and thus a reduction in technical complexity owing to better visualization of target abdominal structures during the procedure. 6 Preoperative weight loss under the guidance of the bariatric team and bariatric dietitian also improves nutrient status, which contributes to improved outcomes and reduced mortality. 28 Finally, hyperglycemia has been shown to increase in-hospital mortality, 29 which can be attenuated by better glycemic control as a result of weight loss. 30 Our study has important clinical implications. Each year, hundreds of patients in the US die after undergoing bariatric surgery. Using a binational data set, our findings showed that even a small reduction in preoperative weight (ie, >0% to <5%) was associated with a 24% reduction in the odds of 30-day mortality following bariatric surgery. Although current clinical guidelines do not require preoperative weight loss and a decision to perform bariatric surgery should not be based on whether and how much preoperative weight loss is achieved, it may be beneficial for patients with obesity to be referred to an established weight loss program before surgery to reduce the risk of mortality.

Strengths and Limitations
The strengths of our study include a large sample size, a diverse population, and rich information that allowed adjustment for potential confounders. Nevertheless, our study has several limitations. First, despite all the advantages of MBSAQIP, our study could not establish causality between preoperative weight loss and postoperative mortality because of the observational nature of the study design.
However, because the mortality rate among patients who underwent bariatric surgery is low, it is less feasible to conduct a randomized clinical trial with a sufficiently large sample size to evaluate the effects of weight loss on mortality risk. Second, institution-specific variables (such as type of intervention used to promote preoperative weight loss and implementation of enhanced recovery after surgery protocols) and surgeon-specific variables (such as experience and surgical technique) may affect mortality outcomes, but such information is not available in the MBSAQIP to protect the privacy of participating institutions, practitioners, and patients. Third, information about whether the weight loss program was voluntary, mandated by insurance, or unintentional was not available; thus, it was impossible to evaluate if the associations would differ by programs or reasons for weight loss. Fourth, our results were limited to short-term mortality after bariatric surgery. Long-term risk of mortality associated with preoperative weight loss warrants further investigation.

Conclusions
Our study showed that preoperative weight loss, even at a moderate degree (ie, >0% to <5%), was associated with lower risk of 30-day mortality following bariatric surgery. Further investigation is needed to replicate our findings in a setting with further information on how weight loss was achieved and to inform future updates of clinical guidelines regarding bariatric surgery.