BACKGROUND/RATIONALE:
Obesity is fast becoming the leading health problem in the United States. The prevalence of obesity and class 3 obesity (BMI > 40) in VA were 33.1% and 3.4%, respectively, in 2000. Obesity is an independent risk factor for type 2 diabetes, coronary heart disease, obstructive sleep apnea, stroke, and several types of cancer. Based on samples comprised primarily of middle-aged women, bariatric surgery is effective in reducing weight and BMI of obese individuals and improving cardiovascular risk factors (e.g., diabetes, hypertension, hyperlipidemia) sufficiently to allow reduction or discontinuation in lipid-lowering, diabetes, hypertension and gastroesophageal reflux (GERD) medications within six-twelve months after surgery. It is unknown whether the weight loss and medication discontinuation induced by bariatric surgery would result in similar improvements in more diverse populations, including older men.

OBJECTIVE(S):
In this short-term research project, we examined the impact of bariatric surgery on weight change and medication discontinuation between 860 obese veterans who had bariatric surgery and 68,168 obese veterans who did not have bariatric surgery.

METHODS:
In this retrospective cohort study of weight change and medication discontinuation, the sample of 860 obese veterans who had bariatric surgery in one of twelve VA bariatric surgical centers in 2000-2006 was drawn from National Surgical Quality Improvement Program (NSQIP) data. Vital statistics and comorbidity data were obtained from data being used in an ongoing evaluation of bariatric surgery on survival, VA health care utilization and expenditures (IIR 05-205-1). The post-surgical weight change of these 860 surgical cases were compared to weight change of 68,168 obese veterans who did not have bariatric surgery, but had valid height and weight data in 2000 from the National Center for Health Promotion (NCP) and valid longitudinal height and weight data from the Corporate Data Warehouse (CDW).

For the Aim 1 analysis in bariatric cases, we examined weight change at 3, 6, 12, 18, 24, 30 and 36 months after baseline in the 860 bariatric cases with univariate analyses to examine the distribution of weight change (for regression specification). We also conducted bivariate analyses with covariates to understand how percentage change in weight varies by covariates and over time. We estimated a mixed model of weight change to determine if weight differed by gender, race, comorbidity burden, smoking status, medication taking, site and time. For the Aim 2 analysis of the difference between 860 surgical cases and 68,168 non-surgical controls in percentage change in weight and BMI over 24 months, we generated similar univariate and bivariate statistics to examine how weight and BMI change differs between surgical cases and non-surgical controls. We then illustrated how weight change differs between these groups. For the Aim 3 analysis of medication discontinuation in bariatric cases, we examined discontinuation of oral hypoglycemic agents and insulin in the 856 bariatric cases at 12 months after surgery. We also examined discontinuation of lipid-lowering medications (e.g., fibrates and statins). We used univariate analyses to examine the distribution of discontinuation by medication and conduct bivariate analyses with covariates to understand how weight change varies by covariate and over time. We then estimated a logistic regression model to understand factors associated with discontinuation.

FINDINGS/RESULTS:
We found that veterans undergoing bariatric surgery in 2000-2006 lost an average of 72 pounds by 6 months after surgery and 101 pounds by 12 months after surgery. In regression analysis that examined weight change over time and controlled for patient factors, super obese veterans (BMI 50) lost significantly more weight over 3 years than veterans with lower BMI levels (p<0.0001). Super-obese veterans also had significantly more weight regain within three years after surgery than veterans with lower BMI levels (p<0.0001). In an analysis comparing surgical cases and non-surgical controls, surgical cases lost an average of 101 pounds in the 12 months after surgery while non-surgical controls gained an average of 5-10 pounds over a similar timeframe.

In medication discontinuation analyses of surgical cases, there were 284 veterans with complete longitudinal data who were taking diabetes medications (oral hypoglycemic agents (OHAs), insulin or both) and 298 veterans taking lipid-lowering medications (fibrates, statins or both). Fifty-three percent of veterans with diabetes taking OHAs discontinued one year after surgery. We also found that 41% of veterans with hyperlipidemia discontinued lipid-lowering medications (fibrates or statins) one year after surgery. In regression analysis that controlled for demographics, baseline super-obesity, smoking status and other factors, veterans with diabetes taking OHAs were more likely to discontinue their medication within a year after surgery than veterans with diabetes taking insulin or veterans taking OHAs and insulin (odds ratio=3.1, p<0.001). In a second regression examining veterans with hyperlipidemia that controlled for patient factors, we found that veterans taking fibrates were more likely to discontinue within a year after surgery than veterans taking statins (odds ratio=5.41, p<0.001).

IMPACT:
This Short-Term Project enabled us to examine the extent to which bariatric surgery for obese veterans results in sustained weight loss and medication discontinuation, which has not be examined extensively in predominantly older, male populations. We found that bariatric surgery provides significant health improvements within one year for older men and enables veterans with less severe chronic conditions to discontinue medication. When coupled with our larger ongoing analysis of survival, VA health care utilization and VA expenditures, this short-term project will provide a comprehensive assessment of the health and economic impacts of bariatric surgery in VA.

The methodology that we have developed for using vital signs data from the VHA Corporate Data Warehouse will be useful to other investigators within VA R&D and VHA Central Office and Operations personnel who may want to use this rich data source for health care operations evaluation. Given that veteran demand for bariatric surgery vastly outpaces its availability, these results can also help the VA Bariatric Surgery Workgroup and VA leadership modify clinical protocols and policies about access to bariatric surgical services. This study can also provide necessary input parameters for future cost-effectiveness assessments of bariatric surgery relative to other obesity treatments within VA settings. Data systems and future research will be needed to monitor the longer-term health outcomes of patients who undergo bariatric surgery in VA.

---

External Links for this Project

---

PUBLICATIONS:

Journal Articles

1. Maciejewski ML, Livingston EH, Smith VA, Kavee AL, Kahwati LC, Henderson WG, Arterburn DE. Survival among high-risk patients after bariatric surgery. JAMA. 2011 Jun 15; 305(23):2419-26. [view]

DRA: Health Systems
DRE: Treatment - Observational
Keywords: Cost effectiveness, Outcomes, Obesity
MeSH Terms: none