Journal of NeuroEngineering and Rehabilitation

Background Individuals with amputations due to dysvascular causes are at increased risk of cardiovascular disease, which is associated with increased peri- and post-operative mor- tality [ 1 – 3 ] , is one of the leading causes of death [ 4 ] , and is associated with increased disability. [ 2 , 5 ] In 2015, there were two million Americans living with limb loss, most commonly due to diabetes and peripheral arterial disease [ 1 ] . The age-adjusted rate of transfemoral ampu- tation (TFA) reaches 40 per 100,000 patients with dia- betes [ 6 ] . Due in part to the aging population and increase in prevalence of those living with diabetes, the number of American amputees is projected to double by the year 2050. [ 7 , 8 ] This growing population of dysvascular amputees has a higher prevalence of cardiovascular disease than the general U.S. adult population: up to 75% have coronary artery disease, 60 – 80% have hypertension, 15 – 25% have cerebrovascular disease, and 20 – 50% have congestive heart failure. [ 4 , 9 – 12 ] In comparison, only approxi- mately 37% of U.S. adults have at least one type of car- diovascular disease. [ 13 ] Cardiovascular disease remains the leading cause of death and health care expenditure in the U.S., with direct and indirect costs reaching $316.1 billion between 2012 and 2013 [ 13 ] . Expenditures are expected to nearly triple by the year 2030, with a large proportion of these growing costs attributable to modifiable risk factors [ 13 – 16 ] . To date studies evaluating major cardiovascular events (MACE), including cardiac death or non-fatal myocardial infarction among individuals with TFA, have been cross-sectional. There has not been a longitudinal evalu- ation of MACE risk in a population-based dataset. This study was undertaken to examine the association between TFA status and the long-term risk of experiencing a major cardiac event for those who underwent an amputation due to either a dysvascular or traumatic cause. The associ- ation between receiving a prosthesis and the risk of ex- periencing a major cardiac event was also examined. Methods Data source and study population Individuals with TFA residing in Olmsted County, MN, were identified using the resources available through the Rochester Epidemiology Project (REP). The REP was de- signed to take advantage of the unique circumstances within Olmsted County: being relatively isolated from other urban areas and having only a few health care pro- viders including Mayo Clinic, Olmsted Medical Center, and their affiliates. [ 17 ] The Olmsted County population is similar to that of the Upper Midwest but is less diverse, wealthier, and more highly educated than the general U.S. population, yet results have been found to be generalizable to populations outside the Upper Midwest. [ 18 ] Using the resources of the REP, TFA (both incident and prevalent TFA patients) were identified using the ICD-9 diagnostic & procedure codes for amputations (84.17 for a TFA procedure or V49.76 indication and individual has a TFA). Each adult with TFA was matched (1:10 ratio) with adults without TFA on age, sex, and duration of residency in Olmsted County. All of the controls also resided in Olmsted County during the same period as the individuals with TFA and were identified in the REP. The 10:1 matching was used to capture the largest possible representative sample of controls and it allowed for an equal number of controls per subject. Patients who had denied research authorization for use of their medical records in research were excluded. This study was approved by both the Mayo Clinic and Olmsted Medical Center Institutional Review Boards. Medical records for TFA individuals were reviewed to confirm their amputation status and level. Additional data obtained included gender, race, amputation etiology, year of amputation (index date), pre and post-amputation comorbidities, and use of prosthesis. Comorbidities were extracted from administrative data and classified using modified Charlson comorbidities via the icd9 package in R [ 19 , 20 ] . The outcome of interest was whether an individual had a major cardiovascular event (MACE), defined as cardiac death, non-fatal myo- cardial infarction, or coronary revascularization, while residing in Olmsted County. Events were identified using pre-existing REP scripts for MACE including Berkson, Hospital Adaptation of the International Classification of Diseases (HICDA), and ICD-9/ − 10 diagnostic codes. Statistical analysis Due to the relatively high initial rates of mortality among individuals with TFA, a competing risk Cox proportional hazard model was used that account for the risk of death due to other cause and the risk of experiencing a MACE in a given time period [ 21 ] . The relative risk (hazard ratio) of individuals with a TFA experiencing a MACE was compared to matched controls. The cohort was divided by amputation etiology: dysvascular vs trauma/cancer. Additional analysis was performed on only TFA to look at the relationship between prosthesis receipt and MACE. Because the prosthesis receipt occurs after the index date for most subjects, prosthesis status was treated at a time-dependent covariate. Differences throughout all of the analysis were considered statistically significant at p < .05. Simulated survival curves for each sub-cohort were per- formed for a man of average age and Charlson Comorbidity Index value to provide a visualization of model results. All statistical analysis was conducted using R version 3.3.2. [ 22 ] Results The study population included at total of 162 individuals with TFA; 107 with amputation due to dysvascular etiology Mundell et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):58 Page 4 of 72