This study utilized data from obese adult participants in the National Health and Nutrition Examination Survey (NHANES) conducted between 1999 and 2018. NHANES is a nationally representative survey conducted by the National Center for Health Statistics (NCHS), using a stratified, multistage probability sampling method to assess the health and nutritional status of the non-institutionalized U.S. population. Adults aged ≥20 years with obesity, defined as a body mass index (BMI) ≥30 kg/m², were included in the study. Exclusions were made for participants who were lost to follow-up, those with missing laboratory data required to calculate phenotypic age (PA), those with missing covariate data, and pregnant women. Ultimately, the study sample included 9,925 obese adults (Figure 1). The NHANES study protocol was approved by the NCHS Research Ethics Review Board, and all participants provided written informed consent at enrollment. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting observational studies. The Ethics Committee of First Hospital of Quanzhou Affiliated to Fujian Medical University approved the study (decision no. 2024-184).

PA was estimated using multiple physiological markers obtained from laboratory analysis of blood samples, including 10 aging-related variables: chronological age (CA), albumin, creatinine, glucose, C-Reactive Protein (CRP), lymphocyte percentage, mean cell volume, red cell distribution width, alkaline phosphatase, and white blood cell count (12). The detailed method for calculating PA is provided in Supplementary material. The residual of PAA was calculated through linear regression of PA on CA. PAA reflects the difference between an individual's BA and CA, with those having a positive residual classified as experiencing “PAA,” and those with a negative residual classified as having “without PAA” (13).

The survival status of participants was determined through the NHANES publicly available linked mortality files, with a cutoff date of December 31, 2019. Causes of death were classified based on the International Classification of Diseases, 10th Revision (ICD-10), focusing on all-cause mortality, cardiovascular disease mortality (ICD-10: I00-I09, I11, I13, I20-I51), and cancer mortality (ICD-10: C00-C97).

Based on previous studies, the potential covariates included age, gender, race, marital status, poverty income ratio (PIR), educational level, Healthy Eating Index (HEI-2015), physical activity, smoking status, alcohol intake, cardiovascular disease (CVD), hypertension, hyperlipidemia, diabetes, and cancer diagnosis (9, 14, 15). Specific definitions were as follows: age was treated as a continuous variable, recording the participants' actual age. Gender was self-reported; race was categorized as non-Hispanic White, others (non-Hispanic Black, Mexican American, other Hispanic, and other races). Marital status was categorized as married/cohabitating or unmarried/other (including widowed, divorced, or separated); PIR was divided into 1–1.3, 1.31–3.50, and >3.50 (16). Educational levels were divided into less than high school, high school or equivalent, and more than high school; HEI-2015 score evaluated diet quality based on 13 components, summing to a 100-point score indicating adherence to 2015–2020 Dietary Guidelines (17). Physical activity time was a continuous variable indicating time spent on activities such as walking, cycling, work, and recreational activities in a week, and PA was categorized into three levels: inactive (0 MET-min/week), insufficiently active (1-599 MET-min/week), and sufficiently active (≥600 MET-min/week) (18).Smoking status was categorized as never smokers (fewer than 100 cigarettes in their lifetime), former smokers (over 100 cigarettes but not currently smoking), and current smokers (over 100 cigarettes and currently smoking sometimes or daily) (19). Participants were categorized by alcohol intake into never drinkers (< 12 drinks in a lifetime), former drinkers (≥12 drinks in a year but not in the past year), and current drinkers (≥12 drinks in any given year and drank in the past year) (20). CVD is defined as a self-reported diagnosis of coronary heart disease, angina, stroke, myocardial infarction, or congestive heart failure (21). Hypertension was defined as a mean systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg, self-reported diagnosis, or use of antihypertensive medications (19). Hyperlipidemia was defined by any of the following: (1) use of lipid-lowering drugs; (2) triglycerides ≥150 mg/dl; or (3) high cholesterol (total cholesterol ≥200 mg/dl, or LDL ≥130 mg/dl, or HDL < 40 mg/dl) (22). Diabetes was defined by a doctor's diagnosis, HbA1c ≥6.5%, fasting glucose ≥7.0 mmol/L, random glucose/Oral Glucose Tolerance Test (OGTT) ≥11.1 mmol/L, or use of diabetes medication/insulin (23). Cancer diagnosis was based on an affirmative response to the question, “Have you ever been told by a doctor or other health professional that you had cancer or any type of malignancy?” (18).

All analyses were conducted using R statistical software (version 4.2.1; http://www.R-project.org, The R Foundation) and Free Statistics software (version 1.7; Beijing, China, http://www.clinicalscientists. cn/freestatistics). Categorical variables were expressed as percentages, and continuous variables as means (standard deviation). Group differences were compared using the Chi-square test and two-sample independent t-test. Survival analyses were conducted using Kaplan-Meier survival curves to estimate all-cause, CVD, and cancer mortality in different PAA groups, with group differences assessed by the Log-rank test. To further analyze the relationship between PAA and mortality, Cox proportional hazards regression models were used to calculate hazard ratios (HRs) and their 95% confidence intervals (CIs). The covariates included in the model were determined based on clinical expertise, univariate screening results (P-value < 0.05), or whether they led to a change in the effect estimate >10%. The multivariable models were progressively adjusted for covariates: Model 1 adjusted for age and gender; Model 2 further adjusted for race, marital status, PIR, educational level, HEI-2015, physical activity, smoking status, and alcohol intake; Model 3 further adjusted for CVD, hypertension, hyperlipidemia, diabetes, and cancer. Restricted cubic spline (RCS) models were used to estimate and visualize the dose-response relationship between age acceleration residual and risks of all-cause and cause-specific mortality.

To assess the consistency of the relationship between PAA and all-cause mortality across different populations, subgroup analyses were performed for age, gender, smoking, alcohol intake, CVD, hypertension, hyperlipidemia, and diabetes. Sensitivity analyses were conducted to validate the robustness of the results: (1) excluding participants who died within 2 years of follow-up; (2) excluding participants with a baseline history of CVD; and (3) excluding participants with a baseline history of cancer. Cox regression models were re-fitted for each sensitivity analysis to evaluate the relationship between PAA and all-cause mortality; (4) Following NHANES analytic guidelines, we incorporated the complex sampling design and mobile examination center (MEC) exam sample weights into our study. Detailed information regarding the weighted analysis can be found in the Supplementary material. Specifically, in the Cox regression analysis, we used a weighted data model to conduct an in-depth analysis of the relationship between PAA and all-cause mortality, as well as cause-specific mortality, in the obese population.

In Table 1, among the 9,925 obese participants, 5,496 (55.4%) were classified as the without PAA group, while 4,429 (44.6%) were in the PAA group. The age acceleration residual for the without PAA group was −5.5 ± 3.5, compared to 8.4 ± 9.3 in the PAA group (P < 0.001). In terms of covariates, the PAA group had a higher mean age (52.1 vs. 50.0 years, P < 0.001). The proportion of non-Hispanic Whites was higher in the without PAA group (48.9 vs. 41.9%, P < 0.001), and a greater percentage of individuals in the without PAA group were married or cohabiting with a partner (65.2 vs. 58.3%, P < 0.001). Additionally, the PAA group had a higher proportion of low-income individuals and current smokers, while the without PAA group had a higher Healthy Eating Index score (P < 0.001).

During the median follow-up of 10.6 years (interquartile range: 4–14.7 years), a total of 1,537 participants died, including 419 deaths from CVD and 357 deaths from cancer. Kaplan-Meier survival analysis showed that participants with PAA had significantly higher all-cause mortality, CVD mortality, and cancer mortality compared to those without PAA (Figure 2, all P-values < 0.001). Table 2 shows that, after adjusting for age, gender, race, marital status, PIR, educational level, physical activity, smoking status, alcohol intake, CVD, hypertension, diabetes, and cancer, PAA remained significantly associated with all-cause mortality, CVD mortality, and cancer mortality. Compared to participants without PAA, those with PAA had an 84% increased risk of all-cause mortality (Model 3, HR = 1.84, 95%CI: 1.64–2.06, P < 0.001), an 86% increased risk of CVD mortality (Model 3, HR = 1.86, 95%CI: 1.50–2.31, P < 0.001), and a 47% increased risk of cancer mortality (Model 3, HR = 1.47, 95%CI: 1.17–1.85, P < 0.001). Age acceleration residual (measured in years) significantly increased the risk of all-cause mortality (HR = 1.04 per 1-year increase), CVD mortality (HR = 1.04–1.05 per 1-year increase), and cancer mortality (HR = 1.02–1.03 per 1-year increase), with all models reaching statistical significance (P < 0.001). As shown in Figure 3, the restricted cubic spline regression model indicated a linear positive correlation between PAA and all-cause mortality, CVD mortality, and cancer mortality (all P-values > 0.05).

Figure 4 shows that the association between PAA and all-cause mortality was consistent across different subgroups based on age, gender, smoking, alcohol intake, CVD, hypertension, hyperlipidemia, and diabetes. Specifically, the all-cause mortality rate was significantly higher in the PAA group across all subgroups. A significant interaction was observed in the age subgroup and gender subgroup.

Several sensitivity analyses were conducted to validate the robustness of the results. First, after excluding participants who died within the first 2 years of follow-up, the association between PAA and all-cause mortality remained significant (HR = 1.80, 95% CI: 1.60–2.03, P < 0.001). Second, after excluding participants with a baseline history of CVD or cancer, the relationship between PAA and all-cause mortality remained stable (CVD exclusion group: HR = 1.80, 95% CI: 1.57–2.07, P < 0.001; cancer exclusion group: HR = 1.84, 95% CI: 1.58–2.14, P < 0.001, Table 3). In Supplementary Table 1, weighted analysis showed that PAA significantly increased the risk of all-cause mortality (HR = 1.90), CVD mortality (HR = 2.03), and cancer mortality (HR = 1.47).