Evaluation of the association between history of open chest or abdominal surgery and cardiovascular risks: an NHANES study, January 2007 – March 2020

Study sample

The NHANES program consists of a series of cross-sectional surveys designed to assess and represent the overall health of civilian, noninstitutionalized adults and children across the United States. The NHANES program began in the early 1960s and continues to be operated by the National Center for Health Statistics (NCHS) in association with the Centers for Disease Control and Prevention (CDC). The program employs stratified multistage probability cluster sampling to examine a biennial sample of 5,000 individuals from all regions of the United States and conducts interviews and physical examinations to assess the health and nutritional status of each participant [5]. The interview portion includes questions regarding demographics, socioeconomics, diet, and health, whereas the examination component encompasses medical, dental, and physiological measurements. Additionally, there are laboratory tests administered by highly trained medical personnel and technicians in mobile examination centers. All of the collected NHANES data are de-identified and, therefore, Institutional Review Board (IRB) approval was not required. The inclusion criteria consisted of NHANES participants aged 20 years or older, whereas the exclusion criteria eliminated individuals with missing data values.

Primary data

This study utilized respondent survey answers and lab values from the NHANES database from January 2007 to March 2020 to conduct a population-level analysis. The subjective qualitative data points were collected as binary yes/no survey responses, in contrast to the quantitative data points, which were collected as continuous values [5]. The history of open-chest or abdominal surgical intervention was captured by a binary yes/no subjective survey question. Seven quantitative variables were identified as indicators of metabolic and cardiovascular health: hemoglobin A_1c (HbA_1c) level, low-density lipoprotein (LDL) level, total triglycerides level, total cholesterol level, systolic blood pressure, diastolic blood pressure (DBP), and high-density lipoprotein (HDL) level. The selection of these variables stemmed from the inputs of the American College of Cardiology (ACC) atherosclerotic cardiovascular disease (ASCVD) risk score, which calculates an individual’s 10-year risk of cardiovascular sequelae, such as heart attack or stroke [6].

Covariates (NHANES-designated subgroups)

Race and ethnicity (Non-Hispanic White, Non-Hispanic Black, Mexican American, Other Hispanic, Other Race), gender (male, female), education (less than 9th grade, 9th to 12th grade, high school graduate/General Education Development [GED] equivalent, any degree higher than high school), health insurance status (no insurance, government insurance, private insurance), and income-to-poverty ratio. The income-to-poverty ratio is a proxy index utilized by the Department of Health and Human Services (HHS) that is calculated by dividing family income by the poverty guidelines specific to family size, as well as the appropriate year and state. If family income is reported as a range, the midpoint is utilized to compute the variable [5].

Analysis

The association for each variable was quantified utilizing an odds ratio (OR) obtained from a logistic regression model run in STATA 16, adjusting for potential confounding factors such as race and ethnicity, insurance status, income-to-poverty ratio, gender, and education level, in order to minimize the impact of bias on the relationship. Statistical significance was determined by a two-sided p value of less than 0.05, and all models were adjusted for the complex survey design utilizing NHANES examination sample weights, primary sampling units, and strata [7]. Given the use of publicly available, nonidentifiable data, IRB approval was not required. Upon calculating the primary outcome, additional subgroup analyses were performed to reveal the differential impacts of HbA_1c levels. A subpopulation logistic regression model, as implemented in STATA 16, was utilized to compute point and variance estimates for the selected subgroups.

Race/ethnicity

It is well established that type 2 diabetes affects racial and ethnic minorities disproportionately, leading to a pattern of increased risk and rates of diabetes complications and mortality [13]. However, our results demonstrated that Non-Hispanic White (p=0.001, OR=1.22) had a statistically significant likelihood of surgical history as HbA_1c levels increased. Given that the literature supports this assertion regarding the disproportionate impact of diabetes on minorities, our results must have a multifactorial explanation; while noninsured and low-income patients may be more likely to undergo surgery, White patients may have greater access to elective surgical care. This complex relationship necessitates a deeper examination and further analysis to unravel the various elements at play.

Insurance

Individuals without insurance were significantly more likely to have a history of surgery as HbA_1c levels increased (p<0.001, OR=1.22), further illustrating that individuals with health insurance diagnosed with type 2 diabetes have better glycemic control, lower diabetes-related mortality, and greater healthcare utilization [14]. Additional literature suggests that the influence of insurance may be exerted through the coverage of comprehensive diabetes metrics such as HbA_1c tests and lipid profiles [15].

Income

Socioeconomic status (SES) is a multifaceted representation that includes factors such as education and economic security; SES has proven to be a consistent predictor of type 2 diabetes onset and progression [16]. SES encompasses an individual’s access to material resources like healthcare, housing, and transportation; therefore, income remains a core component [17]. Consequently, individuals with a lower income-to-poverty ratio (p<0.001, OR=1.15) likely have a lower SES and a greater likelihood of type 2 diabetes onset and progression, ultimately leading to a higher likelihood of having a history of surgery as HbA_1c increases. This finding aligns with a study that utilized data from the National Health Interview Survey (NHIS), revealing an increasing prevalence of diabetes at lower income levels, as reflected in the ratio of income to poverty level [18].

Education

Individuals with less than a ninth-grade education (p=0.03, OR=1.09) or a high school/GED equivalent (p=0.006, OR=1.18) (Table 3) were more likely to have a history of surgery alongside increases in HbA_1c. Education has been proposed as a proxy for overall health literacy and knowledge of diabetes management, providing a plausible explanation for this association. In one study, the prevalence of type 2 diabetes was significantly greater in communities with lower high school graduation rates, and it was inversely associated with education in a stepwise pattern across the adult US population [19]. Nonetheless, higher levels of health literacy have not been associated with increased adherence to diabetes management practices and behavior [16].

Gender

A previous study found that female patients have a decreased willingness to undergo surgery due to fear of postoperative risks and sequelae, which may lead to familial burden. As a result, female patients may delay surgery and experience further disease progression. In this case, a preference for nonsurgical type 2 diabetes management may increase the likelihood of surgical intervention, thereby clarifying our finding that females (p=0.001, OR=1.17) have a significant association between a history of surgery and HbA_1c level.

Public health implications

This study’s statistically significant association between elevated HbA_1c levels and a history of major surgery, particularly among socioeconomically disadvantaged populations, emphasizes the necessity of targeted public health interventions. Although further research is necessary to elucidate the temporal dynamics and causal mechanisms, these findings indicate action to minimize risks and address disparities.

The observed risk among uninsured individuals and those with lower income-to-poverty ratios underscores the pivotal role of financial constraints in shaping associated health outcomes. Policy actions should prioritize expanding access to affordable healthcare, such as through Medicaid expansion or legislation that guarantees comprehensive coverage for specific, evidence-based preventive and treatment services for diabetes [20]. This initiative could enhance access to comprehensive diabetes care, encompassing regular HbA_1c assessments and lipid profiles, which are closely associated with improved glycemic management and reduced diabetes-related mortality. Community-level interventions should prioritize neighborhoods characterized by low income and low levels of education. This encompasses the deployment of resources such as community health centers, which are essential for providing comprehensive primary care to medically disadvantaged and underserved individuals, typically on a sliding-scale fee structure [21], 22]. Interventions involving community health workers (CHWs) can also be very effective. CHWs are frontline public health professionals who connect underserved communities with healthcare systems. They provide essential services, such as education, coaching, and social support, that collectively improve blood glucose management for individuals with diabetes [23], 24].

Limitations

The primary limitations of this study stem from the scope of data available within the NHANES model. Although the dataset includes a large, diverse sample, the context and specificity of the recorded variables are constrained, limiting the depth of analysis. In our study, a key limitation was the method utilized to record surgical history: open-chest and abdominal surgical interventions were grouped together and assessed utilizing a binary yes/no survey question. This subjective, broad categorization reduces the ability to differentiate among distinct surgical procedures. A secondary limitation of this study is the relatively limited number of clinical markers available in the quantitative NHANES data. Therefore, there is little to no context regarding the timing of lab value collection, the general trends of these markers over the participants’ lifetimes, or the temporal relationship to participants’ surgical history. This lack of temporal specificity makes it challenging to interpret the true clinical causative significance of the available data. Additionally, the study cannot adjust for many potential confounders, such as unspecified comorbidities, due to the limited scope of the recorded variables. Furthermore, much of the dataset relies on subjective self-reported survey responses, which introduces the potential for recall bias and variability in data accuracy, further limiting the study’s ability to draw definitive conclusions and potentially leading to overreporting or false associations that threaten internal validity. Ideally, validation through associated medical records would allow for reduction in bias; however, the sampling technique and inherent anonymity of the NHANES surveys make this problematic.