Study design and study participants
In pursuit of comprehending the prevalence of CKD within the demographic parameters of United States military veterans and their civilian counterparts, we availed ourselves of the extensive dataset provided by NHANES. This invaluable dataset encompasses a span of 19 years, starting from 1999 and ending in 2018, inclusive of 10 cyclical periods. This affords us a broad and deep temporal perspective on the evolving landscape of CKD incidence.
Our investigation ventured further, aiming to unravel potential correlations between the various stages of CKD and the veteran status of the individuals. Our goal was to determine if being a veteran inherently predisposes an individual to a heightened risk of CKD.
Our study's sample size was considerable, encompassing a pool of 54,481 adults aged 20 years or older. The data underpinning our analysis was derived from a combination of demographic information, as well as health status markers, procured from both household interviews and laboratory measurements as part of the NHANES data collection protocol.
The sampling method employed by NHANES is particularly noteworthy, being a stratified multistage probability sampling approach. This robust technique ensures the collection of representative samples from a broad spectrum of U.S. states, thus amplifying the generalizability of our findings. To ensure the continuation of this representativeness in our subsequent analysis, we judiciously incorporated the NHANES sampling weights.
The primary characteristics of the study population were summarized using weighted means and standard deviations for continuous variables, while for binary or categorical variables, unweighted proportions were computed. These descriptive statistics are neatly encapsulated in Table 1 for easy cross-reference.
Assessment of Veteran Status
In the systematic appraisal of veteran status, a methodical approach was adopted during each iteration of NHANES. The chosen methodology involved conducting household interviews, where queries were formulated concerning individuals' historical involvement in the U.S. Armed Forces, military Reserves, or National Guard. This strategy permitted us to directly ascertain veterans, thus creating an avenue for the comparative analysis of the prevalence of CKD among veteran and non-veteran participants. The respondents who provided feedback in the form of "Refused" or "Do not know" were categorized as missing data.
CKD stages
We conducted a comprehensive analysis of serum creatinine and albumin-creatinine ratio (ACR) measurements, which are recognized as crucial markers of kidney function. The precise measurement of these biological parameters facilitated the estimation of the glomerular filtration rate (eGFR), a paramount indicator of renal function. To calculate eGFR, we employed the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, an established formula that primarily relies on serum creatinine levels as input (8). This widely accepted equation offers an accurate and dependable estimation of kidney function, which is essential in the diagnosis and staging of CKD.
To ensure consistency with existing clinical guidelines, we delineated the stages of CKD into three categories. The first group comprised individuals with no CKD, characterized by an eGFR ≥60 mL/min/1.73 m², an ACR <30 mg/g, and an absence of historical reports indicating weak or failing kidneys. The subsequent category, designated as Stage G1-G3, included individuals who exhibited an eGFR ≥60 mL/min/1.73 m² and an ACR ≥30 mg/g or those with an eGFR ranging between 30–59 mL/min/1.73 m². The final and most severe category, Stage G4-G5, encompassed individuals with an eGFR <30 mL/min/1.73 m² or those who had undergone dialysis within the past 12 months.
This systematic categorization enabled an accurate assessment of the prevalence of various CKD stages among U.S. veterans and non-veterans. Consequently, our approach provided a nuanced understanding of the CKD landscape within these populations.
Covariates
Covariates included demographic and clinical characteristics, demographic characteristics included age (years), gender, race (Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other Race), current smoking status (yes vs no). Current smoking status, as a two categorical variable, was derived by asking whether participants were currently smoking, with the answer "every day" or "some days" indicating the respondent smoke cigarettes now.
Clinical characteristics included hypertension, diabetes, body mass index (BMI), atherogenic index of plasma (AIP). Hypertension was defined as systolic/diastolic blood pressure of 140/90mmHg or higher(9). Diabetes was defined as a physician-confirmed diagnosis, or glycated hemoglobin (A1c) level of 6.5% or above, or FPG level of 126 mg/dL or higher, or 2-hour plasma glucose (2HPG) level of 200 mg/dL or higher [10]. BMI (kg/m2) is calculated as weight in kilograms divided by height in meters squared. The World Health Organization defines normal/underweight as a BMI lower than 25, overweight as a BMI greater than or equal to 25, and obesity as a BMI greater than or equal to 30 (10).
AIP is a vital metric indicating the equilibrium between protective and atherogenic lipoproteins. AIP, known for its association with pre- and anti-atherogenic lipoprotein particle size (11), was a dependent biomarker in prognosticating atherosclerosis risk (12–18), so could culminate in renal failure at severe stages [19]. We systematically procured triglyceride (TG) and high-density lipoprotein (HDL) measurements throughout each NHANES cycle. AIP was then deduced utilizing the logarithmic formula log(TG/HDL), in adherence to time-tested methodologies, thereby generating a valuable index for cardiovascular risk appraisal (12–14,19,20). AIP was classified as three categories, according to prior study(11,21). “Low risk of cardiovascular disease” was defined as AIP level below 0.11, “Intermediate risk of cardiovascular disease” as AIP level ranging between 0.11 and 0.21, and “Elevated risk of cardiovascular disease” for AIP level exceeding 0.21.
Statistical Analysis
For descriptive statistics, mean ± SD or the median with the interquartile range were used for continuous variables, and numbers (percentage) for categorical variables. Differences in characteristics between groups were tested using a t-test, the Wilcoxon rank test, or the chi-square test. Furthermore, the logistic regression models were used to estimate the association between CKD stages and veteran status.
Statistical analysis was conducted using R software (version 4.1.2), as well as the package “Survey”. Two-tailed P<0.05 was referred to as a statistically significant level.