Association of life’s essential 8 with prevalence and all-cause mortality of chronic kidney disease among US adults: Results from the National Health and Nutrition Examination Survey (2015–2018)

Data source and study design

The NHANES is a series of independent, nationally representative, cross-sectional surveys designed to collect nationally representative health and nutrition data from the US household population.^[17] The samples used in our study were obtained from 2 cycles of the NHANES spanning 2015–2018. The NHANES is approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided informed consent.

We initially included data from 11,288 adult participants (age ≥ 20 years) from the NHANES 2015–2018 (n = 19,225). Among the adult participants, those with missing information of CKD (n = 1224) and CVH (n = 1157) were excluded. Participants with missing information of others characteristics (the missing rates are shown in Table S1 as supplementary materials) were not excluded; further, we used the “NOMCAR” option in estimation procedures to treat missing values as not missing completely at random as suggested by software tips of the NHANES (https://wwwn.cdc.gov/nchs/nhanes/tutorials/softwaretips.aspx).

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline (Supplemental Checklist).^[18]

Measurement of cardiovascular health

The assessment of CVH was based on eight essential metrics, including four health behaviors (diet, physical activity, nicotine exposure, and sleep) and four health factors (BMI, blood lipids, blood glucose, and blood pressure). Each CVH metric was scored ranging from 0 to 100 points, and detailed algorithms for calculating the scores for each metric of the NHANES data have been published previously (Table S2).^[9,10,19] Diet metric was evaluated by the Healthy Eating Index (HEI) 2015, and calculated by using the official SAS code provided by the National Cancer Institute.^[20,21] The information of physical activity, nicotine exposure, and sleep, diabetes history, and medication treatment history was obtained by self-reported questionnaires. The weight, height, blood pressure, blood lipids, plasma glucose, and hemoglobin A1c were collected from “Examination Data” and “Laboratory Data” sections of the NHANES.

The overall LE8 score was calculated for each individual by summing the scores for each of the eight metrics and dividing the total score by 8. The LE8 score ranged from 0–100 points. The health behaviors score and health factors score were calculated by summing the scores for the corresponding metrics and dividing the total score by 4. The CVH score was divided into three groups using the cut points suggested by the AHA: high CVH (80–100), moderate CVH (50–79), and low CVH (0–49).^[9]

Assessment of renal function

The estimated glomerular filtration rate (eGFR) was calculated based on the CKD-EPI 2009 (Chronic Kidney Disease Epidemiology Collaboration) equation.^[22,23] The primary objective of our study was to investigate the association between CVH scores and stage 3–5 CKD, which was defined as an eGFR of less than 60 ml per minute per 1.73 m². Moreover, we evaluated the association between CVH scores and stage 1–5 CKD, which was defined as a GFR of less than 60 ml per minute per 1.73 m² or persistent albuminuria (urinary albumin-to-creatinine ratio > 30 mg/g).^[24]

Follow-up data

The mortality status and time of NHANES CKD participants was collected from public-use linked mortality files updated with mortality follow-up data through 26 April 2022 (https://www.cdc.gov/nchs/data-linkage/mortality-public.htm).

Demographic characteristics

Demographic characteristics of the NHANES were collected, including age, gender, race (Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, Non-Hispanic Asian, Other Race), poverty ratio, education level (high school or less, some college or associate degree, college graduate or above), marital status (married, single or separated), and drink status (yes or no). Age was stratified into three strata (20–39 years, 40–59 years, or ≥ 60 years); and the poverty ratio was stratified into three strata (< 1.3 low income, 1.3–3.5 middle income, and > 3.5 high income).

Statistical analysis

The NHANES uses a complex multi-stage sampling method to select the study population. Therefore, it is necessary to assign corresponding weights, clustering and stratification to the data to correct for sampling errors during the analysis process. We applied the weighting and stratification methods suggested by the NHANES instructions. We calculated the weighted means and weighted percentages with corresponding 95% confidence intervals (CI) for the continuous and categorical variables. We applied the t-test and chi-square test to conduct univariate analyses to compare demographic characteristics between the non-CKD and CKD groups.

To investigate the association of CVH scores with the prevalence of CKD, univariable and multivariable survey logistic regression models, which considered weights, cluster, and strata, were used to obtain crude and adjusted odds ratios (AORs) and corresponding 95% CIs of the comparison between the CKD and non-CKD groups. Logistic regression models were adjusted for all demographic characteristics (only using age and poverty ratio as a continuous variable), and the adjusted prevalence rate of CKD was estimated with logistic regression model. We further performed logistic regression models with restricted cubic splines (RCS) with 3 knots (the 25th, 50th, and 75th percentiles) for CVH scores adjusting for all covariates above to evaluate the dose–response relationship. Moreover, we also conducted subgroup analyses based on the demographic characteristics, and the p values for the production terms between CVH scores and the stratified factors were used to estimate the significance of interactions.

Survey-weighted Cox proportional hazards models were used to analyze the association between CVH scores and the risk of all-cause mortality among participants with CKD. The models were adjusted for all demographic characteristics, and adjusted hazards ratios (AHRs) and 95% CIs were reported.

Figures (excluding the flowchart) were generated using the R software package (version 4.1.2). The remaining statistical analyses were performed using the SAS software (version 9.4; SAS Institute Inc., Cary, NC). All reported p values were two-sided, and a P value < 0.05 was regarded as statistically significant.

Participants and demographic characteristics

Based on the inclusion and exclusion criteria, 8907 participants were included in the study out of 19,225 participants in the 2015–2018 NHANES database (Figure 1). Among the included participants, 789 had stage 3–5 CKD (CKD group), and 8118 were in the non-CKD group.

Figure 1

The flow chart of the study population selection process.

The demographic characteristics of the two groups are shown in Table 1. Compared with the non-CKD group, participants in the CKD group were older (weighted mean: 70.2 vs. 46.7), and had lower educational levels. The weighted percentage of females may have been higher in the CKD group. The current drink rate was lower in the CKD group, probably due to their physical conditions. For the CVH scores and eight metrics, we found the weighted means of scores were significantly lower in the CKD group excluded the HEI-2015 diet score. The weighted means of LE8 score for the non-CKD group and the CKD group were 69.0 (68.2–69.9) vs. 61.0 (59.8–62.2), and of the health factors score were 69.2 (68.3–70.0) vs. 56.5 (54.6–58.3). However, the health behaviors score was slightly lower in CKD group [68.9 (68.0–69.9) vs. 65.5(64.0–67.1)]. The distribution of CVH scores in the total and subgroups of participants are shown in Figure 2B.

Figure 2

The prevalence rate of CKD and distribution of CVH score. (A) the bar plots of adjusted prevalence rate of CKD in different CVH score groups. (B) the violin plots of CVH scores distribution for all study population and subgroup population in CKD and non-CKD groups.

Association between cardiovascular health and prevalence of chronic kidney disease

The adjusted prevalence rates of CKD were 10.7% (95% CI: 7.9%–13.5%) for low LE8 score, 7.9% (95% CI: 6.7%–9.1%) for moderate LE8 score, and 7.7% (95% CI: 5.9%–9.5%) for high LE8 score (Figure. 2A). In the multivariable logistic model, compared with the low LE8 score, moderate LE8 score (AOR 0.628, 95% CI: 0.463–0.853, P = 0.004) and high LE8 score (AOR 0.328, 95% CI: 0.142–0.759, P = 0.011) were associated with lower prevalence rates of CKD (Table 2). No nonlinear association was observed between the LE8 score and CKD (P = 0.308 for nonlinearity, Figure S1), and AOR for per 10 points of LE8 score increase was 0.814 (95% CI: 0.727–0.913, P < 0 .001) in association with prevalence rates of CKD. For health factors score, we found a trend similar to that of the LE8 score, and the AOR for per 10 points of health factors score increase was 0.867 (95% CI: 0.796–0.944, P = 0.002). However, there was no significant association between health behaviors score and prevalence of CKD (AOR 0.930, 95% CI: 0.861–1.004, P = 0.061). We also evaluated the association between CVH and stage 1–5 CKD, and found results similar to those of the primary analyses. Higher LE8 score and health factors score were associated with lower prevalence rates of stage 1–5 CKD (Table S3).

Among the eight CVH metrics, we found per 10 points increase of sleep health score (AOR 0.946, 95% CI: 0.908–0.984, P = 0.008), body mass index score (AOR 0.930, 95% CI: 0.897–0.965, P < 0.001), and blood glucose score (AOR 0.913, 95% CI: 0.874–0.953, P < 0.001) was associated with lower prevalence rates of CKD (Table S4). Although there was no significant association between the HEI-2015 diet score and CKD in univariable model, per 10 points of score increase was potentially associated with a lower prevalence of CKD in the multivariable model (AOR 0.879, 95% CI: 0.810–0.955, P = 0.003).

Association between cardiovascular health and all-cause mortality among chronic kidney disease participants

During a median follow-up of 35 months, 108 all-cause mortalities were recorded among the 789 participants with CKD. In multivariable Cox proportional hazards models, increase of LE8 score (AHR 0.665, 95% CI: 0.504–0.876, P = 0.005) and health behaviors score (AHR 0.702, 95% CI: 0.594–0.829, P < 0.001) was associated with a lower risk of all-cause mortality. Due to the low health factors score of the CKD participants, we did not find a significant association between health factors score and all-cause mortality (Table 3). Similar results were observed among stage 1–5 CKD participants (Table S5)

Subgroup analyses

Subgroup analyses were based on the seven strata factors. There were lower prevalence rates of CKD (AOR < 1) with increased LE8 score and health factors score in most of the subgroup analyses (Figure 3). We found significant interactions between the LE8 score and gender (P = 0.028, Table S6). A higher LE8 score was associated with a lower prevalence of CKD among male participants, but not among female participants. A similar association was also found between health behaviors score and health factors score. Additionally, higher CVH scores were associated with lower CKD prevalence rates among older (age ≥ 60 years old) and married participants.

Figure 3

The association of CVH score with prevalence rate of CKD in subgroup analyses.

Supplementary Information

Supplementary materials are only available at the official site of the journal (www.intern-med.com).