Abstract
Purpose
Though metabolic diseases increasingly affect younger populations, their mental health implications remain poorly understood. We examined associations between age at metabolic disease diagnosis and mental health conditions, comparing early-onset (age <40 years) and late-onset (age ≥40 years) cases.
Materials and Methods
We analyzed the 2014–2023 Korea National Health and Nutrition Examination Survey. Participants reported physician-diagnosed hypertension, diabetes, or dyslipidemia, with age at diagnosis. Mental health conditions included depression, perceived stress, and suicidality. Depression was assessed using the Patient Health Questionnaire-9, perceived stress through self-reported daily experiences, and suicidality as the presence of suicidal ideation, planning, or attempts within the past year. Associations were examined using multiple logistic regression with Bonferroni correction.
Results
Early-onset hypertension [odds ratio (OR)=2.67; 95% confidence interval (CI): 1.91–3.74] and dyslipidemia (OR=2.55; 95% CI: 1.71–3.79) showed higher odds of depression compared to undiagnosed and late-onset cases (p-diff.<0.001 and p-diff.=0.001, respectively). Early-onset hypertension (OR=1.41; 95% CI: 1.21–1.66) and dyslipidemia (OR=1.90; 95% CI: 1.59–2.26) showed higher odds of perceived stress compared to undiagnosed and late-onset cases (both p-diff.<0.001). Early-onset diabetes (OR=2.76; 95% CI: 1.79–4.25) showed higher odds of suicidality compared to undiagnosed and late-onset cases (p-diff.<0.001).
Conclusion
Early-onset metabolic diseases showed stronger associations with adverse mental health conditions than late-onset cases, with patterns suggesting these associations were not primarily explained by disease duration. While the cross-sectional design limits causal inference, mental health screening in early-onset metabolic disease care could be considered.
Keywords: Age of onset; metabolic diseases; hypertension; depression; stress, psychological; suicide
Graphical Abstract
INTRODUCTION
The relationship between metabolic diseases and mental health conditions is being actively investigated, beyond their well-defined physical complications.1,2,3 In hypertension, a study from the UK reported that higher systolic blood pressure was associated with fewer depressive symptoms, while prehypertension was associated with poorer mental health.4 For diabetes, mood and anxiety symptoms, as well as eating, sleeping, and stress-related problems, were common.5 A systematic review found that among diabetes patients, 14% had generalized anxiety disorder, while 40% showed elevated symptoms of anxiety.6 For dyslipidemia, a cross-sectional study found associations between high levels of HDL-C and depression in adult men, and high levels of triglycerides and depression in adult women.7
These mental health associations warrant particular attention in early-onset metabolic diseases, as recent epidemiological data indicate their increasing prevalence among younger populations. Insurance data from Korea demonstrated that early-onset cases (defined as onset before 40 years of age) increased significantly over a 4-year period between 2019 and 2023: hyperlipidemia by 38.15%, diabetes by 17.35%, and hypertension by 11.73%.8 Early-onset diabetes is associated with poorer mental health conditions, particularly higher levels of depressive symptoms and lower levels of self-compassion.9,10 Furthermore, a prospective cohort study in Hong Kong found that 36.8% of the bed-days of early-onset diabetes patients before age 40 were due to mental illnesses.11
While these findings suggest important mental health implications, research on early-onset metabolic diseases has focused primarily on diabetes. The associations of early-onset hypertension and dyslipidemia with mental health conditions remain largely unexplored, despite their rising prevalence. Given that early-onset patients face longer disease exposure, understanding these relationships is crucial. Therefore, this study aims to examine associations between timing of metabolic disease diagnosis and mental health conditions including depression, perceived stress, and suicidality, comparing early-onset and late-onset patients across all three metabolic conditions: hypertension, diabetes, and dyslipidemia.
MATERIALS AND METHODS
Data source
Our study utilized data from the Korea National Health and Nutrition Examination Survey (KNHANES), an annual nationwide cross-sectional study conducted by the Korea Disease Control and Prevention Agency (KDCA).12 This comprehensive survey evaluates the health and nutritional conditions of the Korean population through health interviews, examinations, and nutritional assessments. The survey employs a multi-stage clustered probability design. Enumeration districts (EDs), which serve as primary sampling units, are stratified by region and housing type. From approximately 200000 EDs, 192 districts are selected annually, with 20 to 25 sample households selected per ED.13 This sampling approach ensures representativeness of the Korean population while maintaining a large sample size.
Participant selection
Different KNHANES survey years were used for each mental health condition in adult participants based on the availability of mental health measures. For depression, data from KNHANES years 2014, 2016, 2018, 2020, and 2022 were analyzed. From an initial sample of 30302 adult participants, we excluded those who lacked information on depression (n=3480), metabolic diseases (n=42), or covariates (n=427). The final analytic sample for depression comprised 26353 participants. For perceived stress, data from KNHANES years 2014 to 2023 were analyzed. From an initial sample of 61230 adult participants, we excluded those who lacked information on perceived stress (n=3638), metabolic diseases (n=682), or covariates (n=3424). The final analytic sample for perceived stress comprised 53486 participants. For suicidality, data from KNHANES years 2015, 2017, 2019, 2021, 2022, and 2023 were analyzed. From an initial sample of 36250 adult participants, we excluded those who lacked information on suicidality (n=1888), metabolic diseases (n=384), or covariates (n=2253). The final analytic sample for suicidality comprised 31725 participants. A detailed flowchart of the participant selection process is shown in Fig. 1.
Fig. 1. Flowchart of the participant selection process. KNHANES, Korea National Health and Nutrition Examination Survey.
Measurements
Metabolic disease diagnoses were assessed using self-reported questionnaires. Participants were asked whether they had been diagnosed with hypertension, diabetes, or dyslipidemia by a physician. For participants reporting a diagnosis, their age at first diagnosis was surveyed. Based on diagnosis age, each exposure variable was categorized into three groups: early-onset (defined as onset age <40 years), late-onset (defined as onset age ≥40 years), and undiagnosed.
The early-onset age threshold at age 40 was selected, acknowledging various thresholds (35 to 45 years) used in previous studies. As reviewed by Misra, et al.,14 while any age cutoff may oversimplify risk patterns, this threshold identifies a cohort with distinct healthcare needs, including women of childbearing age who represent a high-risk group. Sensitivity analysis using alternative thresholds (35 and 45 years) was conducted to test the robustness of this cutoff.
Mental health conditions included depression, perceived stress, and suicidality. Depression was assessed using the Patient Health Questionnaire-9 (PHQ-9), a validated self-reported screening instrument that measures depression severity.15 The study employed the validated Korean version of the PHQ-9,16 using a cut-off score of 10 to indicate clinically significant depression.17 For perceived stress assessment, levels were evaluated through self-reported experiences in daily life. Participants responded to the question: “How much stress do you typically feel in your daily life?” Responses were dichotomized into two groups: a stressed group (comprising those who responded “Feel it very strongly” or “Tend to feel it strongly”) and an unstressed group (those who responded “Tend to feel it a little” or “Hardly feel it”). Suicidality was defined as the presence of suicidal ideation, planning, or attempts within the past year. Participants responded to three questions: “Have you seriously considered suicide in the past year?,” “Have you made specific plans to commit suicide in the past year?,” and “Have you actually attempted suicide in the past year?” A positive response to any of these questions was classified as indicating the presence of suicidality.
Covariates
Several covariates were included in the analysis. Demographic factors consisted of age and sex. Socioeconomic factors included household income, education level, and marital status. Personal health factors included body mass index (BMI), alcohol consumption, cigarette smoking, and physical activity. The survey year was included as an additional covariate.
We categorized household income into annual quartiles using equivalized income, defined as monthly household income divided by the square root of household size. Education level was grouped into four levels by highest completed education: elementary school or lower (≤8 years), middle school graduate (≥9 years), high school graduate (≥12 years), and college graduate or higher (≥14 years). BMI was calculated using weight and height measured during the survey by dividing weight by height squared (kg/m2). Both alcohol consumption and cigarette smoking status were categorized into three groups: never (those who never drank/smoked), former (previous drinkers/smokers), and current (current drinkers/smokers). Physical activity was evaluated using the validated Korean version of the Global Physical Activity Questionnaire.18,19
Statistical analysis
Baseline characteristics of the study population were assessed separately by metabolic disease status (undiagnosed, early-onset, and late-onset) for hypertension, diabetes, and dyslipidemia, respectively. Continuous variables (age, BMI, and physical activity) were presented as weighted means (weighted sample standard deviation). Categorical variables were presented as crude N (weighted %). To compare baseline characteristics across groups, survey-weighted analysis of variance (PROC SURVEYREG) and Rao-Scott chi-square tests (PROC SURVEYFREQ) were used for continuous and categorical variables, respectively. Additional assessment of the baseline characteristics was performed based on adverse mental health conditions.
Multiple logistic regression (PROC SURVEYLOGISTIC) was performed to examine the association between early-onset metabolic diseases and adverse mental health conditions. All analyses were performed separately for depression, perceived stress, and suicidality. Stepwise adjustments were conducted starting with demographic factors (age and sex). The next model additionally adjusted for socioeconomic factors (household income, education level, and marital status). The subsequent model further adjusted for personal health factors (BMI, alcohol consumption, cigarette smoking, and physical activity). The final model additionally adjusted for survey year. For each metabolic disease, odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated comparing both early-onset and late-onset groups to the undiagnosed group. Additionally, to analyze the differences between early-onset and late-onset groups, p for difference (p-diff.) was computed. To account for multiple comparisons, the Bonferroni correction was applied. With three metabolic diseases and three mental health conditions, the significance threshold was set at p<0.05/9. Results are presented with 95% CIs, and only comparisons that remained significant after Bonferroni correction are reported.
Three subgroup analyses were performed. For the first subgroup analysis, disease duration for each metabolic disease was calculated as the difference between current age and self-reported age at diagnosis. The study population was stratified by disease duration using the median as the threshold. Within each stratum, ORs were calculated comparing early-onset to late-onset groups. For the second subgroup analysis, the exposure was categorized into five groups: 1) undiagnosed, 2) early-onset and uncontrolled, 3) early-onset and controlled, 4) late-onset and uncontrolled, and 5) late-onset and controlled. Metabolic disease was defined as controlled if systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg for hypertension, HbA1c <6.5% for diabetes, and total cholesterol <200 mg/dL for dyslipidemia based on examination data. We compared the odds of mental health outcomes for early-onset versus late-onset participants, separately for controlled and uncontrolled groups. The third subgroup analysis stratified participants by BMI, using a cut-off of 25 kg/m2.
For sensitivity analyses, the first approach used alternative age thresholds (35 and 45 years) for defining early-onset. The second approach additionally adjusted for major cardiovascular and renal comorbidities including stroke, myocardial infarction, angina, and kidney disease, which are common consequences of metabolic diseases. Comorbidity was defined as a self-reported diagnosis and currently having one or more of these conditions at the time of the survey. Since data on stroke, myocardial infarction, and angina were only available for KNHANES years 2014–2021, this analysis was restricted to those years. The third approach restricted the analysis to participants aged 40 years and above.
To account for the complex survey design of KNHANES, which employs a two-stage stratified cluster sampling method, all analyses incorporated sampling weights, strata, and clusters. This approach was applied consistently across baseline comparisons, main analyses, subgroup analyses, and sensitivity analyses. The Taylor series linearization method was used to calculate standard errors and variances to account for the complex sampling design.20
All statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA).
Ethics statement
All participants provided written informed consent to participate in the KNHANES. The survey was conducted under the approval of the Institutional Review Board (IRB) of KDCA (IRB numbers: 2013-12EXP-03-5C for 2014, 2018-01-03-P-A for 2018, 2018-01-03-C-A for 2019, 2018-01-03-2C-A for 2020, 2018-01-03-5C-A for 2021, 2018-01-03-4C-A for 2022, and 2022-11-16-R-A for 2023). The study was conducted in accordance with the Declaration of Helsinki.
RESULTS
Baseline characteristics
Baseline characteristics of the study population by metabolic disease status are presented in Table 1. Compared with late-onset patients, participants with early-onset hypertension, diabetes, and dyslipidemia were more likely to be younger, male, have higher household income, have a higher education level, be unmarried, have higher BMI, have higher physical activity levels, drink alcohol, and be current cigarette smokers. Baseline characteristics of the study population by mental health conditions are presented in Supplementary Table 1 (only online).
Table 1. Baseline Characteristics of the Study Population by Metabolic Disease Status in KNHANES 2014–2023 (n=53504).
| Hypertension* | Diabetes* | Dyslipidemia* | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Undiagnosed (n=40078) | Early-onset† (n=1138) | Late-onset‡ (n=12288) | p | Undiagnosed (n=48097) | Early-onset† (n=501) | Late-onset‡ (n=4906) | p | Undiagnosed (n=42682) | Early-onset† (n=747) | Late-onset‡ (n=10075) | p | |||
| Demographic factors | ||||||||||||||
| Age (yr) | 43.41 (15.10) | 46.39 (13.09) | 64.90 (10.09) | <0.001 | 45.90 (16.13) | 47.80 (12.73) | 64.76 (9.91) | <0.001 | 44.61 (16.04) | 41.35 (9.32) | 62.48 (9.69) | <0.001 | ||
| Sex | <0.001 | <0.001 | <0.001 | |||||||||||
| Male | 17031 (48.91) | 772 (75.12) | 5567 (49.46) | 20627 (49.10) | 310 (68.65) | 2433 (53.63) | 19216 (50.61) | 478 (70.07) | 3676 (41.88) | |||||
| Female | 23047 (51.09) | 366 (24.88) | 6721 (50.54) | 27470 (50.90) | 191 (31.35) | 2473 (46.37) | 23466 (49.39) | 269 (29.93) | 6399 (58.12) | |||||
| Socioeconomic factors | ||||||||||||||
| Income§ | <0.001 | <0.001 | <0.001 | |||||||||||
| Highest (Q1) | 13244 (34.64) | 355 (33.57) | 2210 (21.97) | 14840 (33.33) | 125 (26.60) | 844 (21.01) | 13344 (33.54) | 273 (37.37) | 2192 (25.49) | |||||
| High (Q2) | 12019 (31.54) | 308 (28.14) | 2425 (21.84) | 13642 (30.34) | 129 (26.93) | 981 (22.60) | 12237 (30.61) | 233 (32.12) | 2282 (24.79) | |||||
| Low (Q3) | 9458 (22.70) | 281 (24.58) | 3292 (25.69) | 11569 (23.04) | 134 (26.73) | 1328 (25.81) | 10156 (22.92) | 150 (19.42) | 2725 (25.64) | |||||
| Lowest (Q4) | 5357 (11.12) | 194 (13.71) | 4361 (30.51) | 8046 (13.30) | 113 (19.74) | 1753 (30.58) | 6945 (12.93) | 91 (11.09) | 2876 (24.08) | |||||
| Education∥ | <0.001 | <0.001 | <0.001 | |||||||||||
| College or higher (≥14 yr) | 17676 (47.33) | 462 (44.78) | 1926 (19.48) | 19138 (44.22) | 165 (36.07) | 761 (19.35) | 17672 (45.49) | 419 (57.80) | 1973 (23.53) | |||||
| High school (12–13 yr) | 14315 (38.12) | 402 (38.06) | 3103 (28.26) | 16365 (36.90) | 191 (40.91) | 1264 (29.08) | 14710 (37.36) | 256 (34.14) | 2854 (31.30) | |||||
| Middle school (9–11 yr) | 3247 (6.58) | 102 (6.86) | 1996 (15.69) | 4474 (7.61) | 58 (9.00) | 813 (15.63) | 3613 (6.89) | 39 (4.75) | 1693 (15.72) | |||||
| Elementary or lower (≤8 yr) | 4840 (7.96) | 172 (10.30) | 5263 (36.56) | 8120 (11.26) | 87 (14.02) | 2068 (35.95) | 6687 (10.26) | 33 (3.31) | 3555 (29.45) | |||||
| Marital status | <0.001 | <0.001 | <0.001 | |||||||||||
| Married | 26971 (62.99) | 796 (66.47) | 8773 (73.18) | 32690 (64.20) | 339 (67.15) | 3511 (73.01) | 28602 (62.80) | 532 (70.43) | 7406 (75.57) | |||||
| Divorced or bereaved | 3972 (7.49) | 128 (8.95) | 3284 (24.59) | 6013 (9.45) | 84 (13.07) | 1287 (24.18) | 4888 (8.58) | 52 (5.26) | 2444 (21.84) | |||||
| Unmarried | 9135 (29.52) | 214 (24.57) | 231 (2.24) | 9394 (26.35) | 78 (19.78) | 108 (2.81) | 9192 (28.61) | 163 (24.31) | 225 (2.60) | |||||
| Personal health factors | ||||||||||||||
| BMI (kg/m2) | 23.68 (3.67) | 26.93 (4.53) | 25.19 (3.38) | <0.001 | 23.91 (3.71) | 26.23 (4.69) | 25.06 (3.39) | <0.001 | 23.80 (3.73) | 26.34 (4.33) | 24.96 (3.33) | <0.001 | ||
| Physical activity (METs) | 1121.90 (1845.49) | 1148.70 (1994.32) | 740.47 (1424.23) | <0.001 | 1084.00 (1817.40) | 960.78 (1670.16) | 699.73 (1327.40) | <0.001 | 1100.76 (1837.97) | 1319.66 (2442.93) | 777.14 (1343.91) | <0.001 | ||
| Alcohol consumption | <0.001 | <0.001 | <0.001 | |||||||||||
| Never | 3473 (6.82) | 83 (5.45) | 2386 (17.06) | 4913 (7.87) | 57 (9.71) | 972 (17.68) | 4082 (7.38) | 42 (4.97) | 1818 (15.64) | |||||
| Past only | 6393 (14.26) | 210 (17.70) | 2769 (21.25) | 7989 (14.87) | 109 (20.26) | 1274 (24.03) | 6858 (14.33) | 123 (15.29) | 2391 (22.43) | |||||
| Current | 30212 (78.92) | 845 (76.86) | 7133 (61.68) | 35195 (77.26) | 335 (70.02) | 2660 (58.29) | 31742 (78.29) | 582 (79.74) | 5866 (61.93) | |||||
| Cigarette smoking | <0.001 | <0.001 | <0.001 | |||||||||||
| Never | 25608 (60.86) | 506 (41.42) | 7376 (56.56) | 30559 (60.41) | 224 (41.63) | 2707 (52.29) | 26572 (59.73) | 337 (42.56) | 6581 (61.12) | |||||
| Past only | 7282 (18.37) | 308 (27.51) | 3277 (28.06) | 9334 (19.57) | 131 (25.79) | 1402 (28.84) | 8382 (19.32) | 190 (26.99) | 2295 (24.85) | |||||
| Current | 7188 (20.77) | 324 (31.08) | 1635 (15.38) | 8204 (20.02) | 146 (32.58) | 797 (18.87) | 7728 (20.95) | 220 (30.45) | 1199 (14.03) | |||||
| Mental health conditions | ||||||||||||||
| Depression¶ | <0.001 | <0.001 | <0.001 | |||||||||||
| Not depressed | 18931 (95.06) | 492 (88.99) | 5486 (94.36) | 22539 (95.06) | 213 (85.53) | 2157 (92.53) | 20213 (95.04) | 315 (89.09) | 4381 (94.01) | |||||
| Depressed | 1013 (4.94) | 62 (11.01) | 369 (5.64) | 1219 (4.94) | 31 (14.47) | 194 (7.47) | 1073 (4.96) | 42 (10.91) | 329 (5.99) | |||||
| Perceived stress** | <0.001 | <0.001 | <0.001 | |||||||||||
| Not stressed | 29043 (71.45) | 767 (64.75) | 9820 (78.78) | 35453 (72.30) | 329 (64.22) | 3848 (77.42) | 31429 (72.25) | 419 (55.49) | 7782 (76.27) | |||||
| Stressed | 11022 (28.55) | 371 (35.25) | 2463 (21.22) | 12630 (27.70) | 172 (35.78) | 1054 (22.58) | 11237 (27.75) | 328 (44.51) | 2291 (23.73) | |||||
| Suicidality†† | 0.002 | <0.001 | 0.002 | |||||||||||
| No suicidality | 22405 (95.56) | 645 (94.00) | 7103 (94.42) | 27074 (95.55) | 273 (88.03) | 2806 (93.49) | 23677 (95.53) | 434 (93.16) | 6042 (94.46) | |||||
| Suicidality | 1083 (4.44) | 40 (6.00) | 449 (5.58) | 1341 (4.45) | 33 (11.97) | 198 (6.51) | 1154 (4.47) | 31 (6.84) | 387 (5.54) | |||||
BMI, body mass index; METs, metabolic equivalents.
Age, BMI, and physical activity are presented as weighted mean (weighted sample standard deviation). Categorical variables (all other variables) are presented as n (weighted %).
*Metabolic diseases (including hypertension, diabetes, and dyslipidemia): Assessed based on self-reported physician diagnosis; †Early-onset: Defined as diagnosis age <40 years; ‡Late-onset: Defined as diagnosis age ≥40 years; §Income: Categorized into quartiles based on equivalized income, defined as monthly household income divided by the square root of household size; ∥Education: Categorized by highest completed education level; ¶Depression: Defined as a Patient Health Questionnaire-9 score ≥10. Data from KNHANES years 2014, 2016, 2018, 2020, and 2022 were used with a total of 26353 participants; **Perceived stress: Assessed based on participants’ self-reported perception of stress in daily life. Data from KNHANES years 2014 to 2023 were used with a total of 53486 participants; ††Suicidality: Defined as experiencing suicidal ideation, planning, or attempts within the past year. Data from KNHANES years 2015, 2017, 2019, 2021, 2022, and 2023 were used with a total of 31725 participants.
Association between early-onset metabolic diseases and adverse mental health conditions
Stepwise analyses for adverse mental health conditions according to early-onset and late-onset metabolic diseases are presented in Supplementary Table 2 (only online). The models were progressively adjusted for demographic, socioeconomic, personal health factors, and survey year. The significant differences between early-onset and late-onset groups were maintained across adjustment steps, particularly for hypertension and dyslipidemia with depression and perceived stress, and for diabetes with suicidality.
Adjusted ORs and 95% CIs for adverse mental health conditions according to early-onset and late-onset metabolic diseases using the final model are presented in Fig. 2.
Fig. 2. Association between early-onset metabolic diseases and adverse mental health outcomes. CI, confidence interval; PHQ-9, Patient Health Questionnaire-9.
For depression, early-onset patients with hypertension (OR=2.67; 95% CI: 1.91–3.74), diabetes (OR=2.92; 95% CI: 1.82–4.68), and dyslipidemia (OR=2.55; 95% CI: 1.71–3.79) showed higher odds compared to the undiagnosed group. Among late-onset patients, those with diabetes (OR=1.58; 95% CI: 1.28–1.96) showed higher odds compared to the undiagnosed group.
For perceived stress, early-onset patients with hypertension (OR=1.41; 95% CI: 1.21–1.66), diabetes (OR=1.44; 95% CI: 1.14–1.81), and dyslipidemia (OR=1.90; 95% CI: 1.59–2.26) showed higher odds compared to the undiagnosed group. Among late-onset patients, those with dyslipidemia (OR=1.20; 95% CI: 1.12–1.28) showed higher odds compared to the undiagnosed group.
For suicidality, early-onset patients with diabetes (OR=2.76; 95% CI: 1.79–4.25) showed higher odds compared to the undiagnosed group.
When comparing early-onset to late-onset groups, early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.=0.001) showed higher odds of depression. Early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress. Early-onset patients with diabetes (p-diff.<0.001) showed higher odds of suicidality.
Subgroup analyses
Adjusted ORs and 95% CIs in stratified groups according to the median disease duration of metabolic diseases are presented in Fig. 3. Among those with below-median disease durations, early-onset patients with hypertension (p-diff.=0.001) showed higher odds of depression, and early-onset patients with diabetes (p-diff.=0.004) showed higher odds of suicidality compared to late-onset patients.
Fig. 3. Subgroup analysis according to disease duration. Mdn., median; OR, odds ratio; CI, confidence interval; PHQ-9, Patient Health Questionnaire-9.
Analysis comparing early-onset to late-onset patients by disease control status is presented in Supplementary Fig. 1 (only online). Among those with controlled disease, early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of depression, early-onset patients with dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress, and early-onset patients with diabetes (p-diff.<0.001) showed higher odds of suicidality compared to late-onset patients.
Subgroup analysis stratified by BMI (cut-off 25 kg/m2) is presented in Supplementary Fig. 2 (only online). Among participants with BMI <25 kg/m2, early-onset patients with hypertension (p-diff.<0.001) showed higher odds of depression compared to late-onset patients. Among participants with BMI ≥25 kg/m2, early-onset patients with hypertension (p-diff.=0.004) showed higher odds of depression, early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress, and early-onset patients with diabetes (p-diff.<0.001) showed higher odds of suicidality compared to late-onset patients.
Sensitivity analyses
Sensitivity analysis using alternative thresholds (35 and 45 years) is presented in Supplementary Fig. 3 (only online). Using a 35-year threshold, early-onset patients with hypertension (p-diff.=0.002) showed higher odds of depression compared to late-onset patients. Early-onset patients with hypertension (p-diff.=0.003) and dyslipidemia (p-diff.=0.002) showed higher odds of perceived stress compared to late-onset patients. Using a 45-year threshold, early-onset patients with hypertension (p-diff.<0.001) showed higher odds of depression compared to late-onset patients. Early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress compared to late-onset patients. Early-onset patients with diabetes (p-diff.<0.001) showed higher odds of suicidality compared to late-onset patients.
Sensitivity analysis with additional adjustment for comorbidities including stroke, myocardial infarction, angina, and kidney disease is presented in Supplementary Fig. 4 (only online). Early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of depression compared to late-onset patients. Similarly, early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress compared to late-onset patients.
We conducted an additional sensitivity analysis restricted to participants aged 40 years and above, presented in Supplementary Fig. 5 (only online). The analysis evaluated the age distribution difference between groups, as patients aged 20–39 were present only in the early-onset group. In this older subset, early-onset patients with hypertension (p-diff.<0.001) and dyslipidemia (p-diff.=0.002) showed higher odds of depression compared to late-onset patients. Early-onset patients with dyslipidemia (p-diff.<0.001) showed higher odds of perceived stress compared to late-onset patients.
DISCUSSION
In our study, early-onset hypertension and dyslipidemia showed associations with increased odds of depression and perceived stress, while early-onset diabetes was associated with increased odds of suicidality. These associations remained consistent across alternative age thresholds (35 and 45 years) used to define early-onset metabolic disease. Our subgroup analysis stratified by median disease duration showed that early-onset patients had elevated odds of adverse mental health outcomes compared to late-onset patients in both duration subgroups. The timing of diagnosis, rather than the duration of disease exposure, may be important in these associations.
Our findings align well with previous studies investigating associations between early-onset diabetes and mental health conditions. A cohort study in Sweden reported that participants with early-onset type 2 diabetes had a higher risk of depression, bipolar disorder, anxiety, and stress-related disorders.21 Another study in the UK showed that being diagnosed with type 2 diabetes at a younger age was associated with lower psychological well-being, including depressive symptoms, diabetes distress, and lower levels of self-compassion.9 Although the associations between early-onset diabetes with depression and perceived stress did not reach statistical significance after Bonferroni correction in our study, we observed elevated odds for these associations, consistent with prior research. Our finding of significantly elevated suicidality among early-onset diabetes patients extends this literature in an important direction. A meta-analysis reported that suicidal ideation and suicide attempts were prevalent in adolescents and young adults with type 1 diabetes.22 While we could not distinguish between type 1 and type 2 diabetes in our analysis, we found significantly elevated odds of suicidality among early-onset diabetes patients compared to late-onset patients, suggesting this association may extend beyond type 1 diabetes.
Unlike the substantial literature on early-onset diabetes and mental health, previous evidence on associations between early-onset hypertension and dyslipidemia with psychological outcomes is limited. Our findings that patients with early-onset hypertension and dyslipidemia showed significantly higher odds of depression and perceived stress compared to late-onset patients represent novel contributions.
Several factors may contribute to the observed associations. Prior longitudinal studies have shown that patients with early-onset metabolic diseases may be more likely to develop life-threatening complications compared to late-onset patients. A previous cohort study in the U.S. found that early-onset hypertension was associated with hypertensive end-organ damage already by midlife.23 Another study reported that patients with early-onset type 2 diabetes had a higher risk of cardiovascular disease incidence and all-cause mortality.24 To examine whether the worse psychological outcomes among early-onset patients might be explained by increased complications, we performed sensitivity analyses adjusting for the presence of stroke, myocardial infarction, angina, and kidney disease. The elevated odds among early-onset patients generally persisted after this adjustment, suggesting that factors beyond current disease complications may play important roles in the observed mental health burden.
One possible mechanism relates to increased psychological burden due to stigma, self-blame, or stereotypes. Our subgroup analysis stratified by median disease duration showed that early-onset patients had elevated odds of adverse mental health outcomes compared to late-onset patients regardless of disease duration. This suggests that the timing of diagnosis, rather than cumulative disease exposure, may be more important in explaining the observed mental health burden. A multi-center study in the U.S. reported that patients with early-onset hypertension felt they were older than their peers due to needed medications and behavior changes, and faced the prospect of lifelong disease management.25 Additionally, experiences of stigma and self-blame from early diagnosis of type 2 diabetes were common.26 Specifically, patients with early-onset type 2 diabetes felt embarrassed about having the disease or felt like failures.27 These psychosocial factors associated with early timing of diagnosis may contribute to the higher psychological stress observed among early-onset patients in our study.
In addition to psychosocial pathways, biological mechanisms may also contribute to the observed associations. One possible mechanism relates to autonomic dysfunction. A cohort study in Korea reported that while low heart rate variability (HRV) was associated with increased risk of hypertension in both younger (<40 years) and older (≥40 years) adults, the association was stronger in younger adults.28 Another cohort study in the U.S. found reduced HRV with parasympathetic loss and sympathetic overdrive among patients with youth-onset type 2 diabetes.29 While the association between autonomic imbalance and mental health is being actively investigated, it has been proposed that alterations in HRV might precede the onset of depression, and HRV has been associated with proximal suicidal risk.30,31 Further research is needed to elucidate whether autonomic dysfunction or other biological pathways mediate the associations between early-onset metabolic diseases and adverse mental health outcomes.
Our study has several strengths. First, while prior studies have focused on diabetes, our study adds new evidence about the association of early-onset hypertension and dyslipidemia with adverse mental health conditions. Second, our study utilized the KNHANES, a large, nationally representative sample, which enhances the generalizability of our findings. Third, our comprehensive subgroup analyses by disease duration, disease control status, and BMI provide insights into potential mechanisms and effect modifiers underlying these associations. Finally, our findings remained consistent across multiple sensitivity analyses. These analyses included different age thresholds for defining early-onset, additional adjustment for major comorbidities, and analyses restricted to older participants, demonstrating the robustness of our results.
However, our study has several limitations. First, the cross-sectional design limits our ability to establish causal relationships. While the measurements were collected cross-sectionally, the timeframe for psychological outcomes assessed the recent 2 weeks or past year, whereas for metabolic diseases, the age of onset was surveyed retrospectively, providing some temporal sequence. Nevertheless, causal inference between exposure and outcome remains limited due to the cross-sectional design. Importantly, we cannot rule out reverse causation, particularly for hypertension and dyslipidemia, where pre-existing depression or psychological distress may have preceded or contributed to the development of metabolic diseases. Longitudinal studies are needed to establish temporal relationships and causal mechanisms. Second, age distribution differs between early-onset and late-onset groups, with patients aged 20–39 years present only in the early-onset group; however, our sensitivity analysis restricted to participants aged 40 years and above showed similar results. Third, mental health conditions reflect only recent status rather than longitudinal patterns. Fourth, self-reported age of diagnosis may be subject to recall bias, particularly for participants with longer disease duration. Fifth, using age of diagnosis as a proxy for age of onset may not accurately reflect disease onset timing. Sixth, the potential presence of undiagnosed metabolic diseases in the undiagnosed control group could lead to misclassification bias; however, this would likely bias our results toward the null, making our findings more conservative. Seventh, we were unable to distinguish between type 1 and type 2 diabetes in our analysis. This is particularly important for interpreting early-onset diabetes findings, as early-onset cases may include a higher proportion of type 1 diabetes, which has distinct etiological factors including autoimmune processes, family history, and insulin dependence that may independently affect mental health outcomes. Eighth, since the survey years differed across mental health outcomes, period effects across survey years may have influenced our findings; however, survey year was adjusted to minimize this potential effect. Ninth, the sensitivity analysis adjusting for comorbidities was restricted to KNHANES years 2014–2021 due to limited data availability for cardiovascular disease variables in other survey years. This restriction reduced the sample size and may have decreased statistical power to detect some associations. Finally, while the Bonferroni correction effectively controls type 1 error, it is a conservative approach that may increase the risk of type 2 errors, potentially missing true associations. For example, although early-onset diabetes patients showed elevated odds for depression and early-onset dyslipidemia patients showed elevated odds for suicidality compared to late-onset patients, the differences did not reach statistical significance after Bonferroni correction.
These associations suggest clinical relevance for early-onset metabolic disease care. The observed patterns indicate that early-onset metabolic disease patients experience higher rates of mental health challenges. Given the consistent associations across multiple analyses and substantial effect sizes, these findings suggest potential value in considering mental health screening in this population, though longitudinal research is needed to establish causal relationships and evaluate intervention effectiveness. Such screening could potentially be incorporated into routine follow-up visits using validated assessment tools. From a public health perspective, awareness of these associations may inform clinical guidelines and care protocols for younger patients with metabolic diseases.
In conclusion, this large, nationally representative study demonstrates that early-onset metabolic diseases are associated with higher odds of adverse mental health conditions compared to late-onset cases. While causal relationships cannot be established from cross-sectional data, these consistent associations suggest important clinical implications for early-onset metabolic disease care, particularly regarding potential mental health screening. Routine mental health screening using validated instruments could be considered for patients diagnosed before age 40, and could be integrated into standard follow-up visits. Future research should evaluate intervention strategies including depression screening protocols and stress management programs tailored for early-onset patients. Longitudinal cohort studies are essential to establish temporal and causal relationships, distinguish the independent effects of disease onset timing and duration, and identify mediating mechanisms linking early diagnosis to adverse mental health outcomes.
ACKNOWLEDGEMENTS
This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: RS-2024-00440320).
Footnotes
An abstract of this article has been presented in part as an oral presentation at the Korean Society of Epidemiology Fall Conference 2024.
The authors have no potential conflicts of interest to disclose.
- Conceptualization: Changhyun Kim, Dongkyu Lee, Yeon Woo Oh, and Sun Jae Jung.
- Formal analysis: Changhyun Kim.
- Funding acquisition: Sun Jae Jung.
- Investigation: Changhyun Kim.
- Methodology: Changhyun Kim, Dongkyu Lee, Yeon Woo Oh, and Sun Jae Jung.
- Supervision: Sun Jae Jung.
- Visualization: Changhyun Kim and Jaeyong Lee.
- Writing—original draft: Changhyun Kim.
- Writing—review & editing: all authors.
- Approval of final manuscript: all authors.
SUPPLEMENTARY MATERIALS
Baseline characteristics of the study population by adverse mental health conditions in KNHANES 2014-2023 (N=53,504)
Stepwise analysis of association between early-onset metabolic diseases and adverse mental health conditions in KNHANES 2014-2023
Subgroup analysis according to disease control status.
Subgroup analysis according to body mass index (BMI).
Sensitivity analysis according to different early-onset age thresholds.
Sensitivity analysis with additional adjustment for comorbidities (stroke, myocardial infarction, angina, and kidney disease) in KNHANES 2014-2021.
Sensitivity analysis restricted to participants aged 40 years and above.
References
- 1.Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International Society of Hypertension global hypertension practice guidelines. Hypertension. 2020;75:1334–1357. doi: 10.1161/HYPERTENSIONAHA.120.15026. [DOI] [PubMed] [Google Scholar]
- 2.Freiberg JJ, Tybjaerg-Hansen A, Jensen JS, Nordestgaard BG. Nonfasting triglycerides and risk of ischemic stroke in the general population. JAMA. 2008;300:2142–2152. doi: 10.1001/jama.2008.621. [DOI] [PubMed] [Google Scholar]
- 3.American Diabetes Association Professional Practice Committee. 4. Comprehensive medical evaluation and assessment of comorbidities: standards of care in diabetes—2024. Diabetes Care. 2024;47(Suppl 1):S52–S76. doi: 10.2337/dc24-S004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Schaare HL, Blöchl M, Kumral D, Uhlig M, Lemcke L, Valk SL, et al. Associations between mental health, blood pressure and the development of hypertension. Nat Commun. 2023;14:1953. doi: 10.1038/s41467-023-37579-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Diabetes Canada Clinical Practice Guidelines Expert Working Group. Robinson DJ, Hanson K, Jain AB, Kichler JC, Mehta G, Melamed OC, et al. Diabetes and mental health. Can J Diabetes. 2023;47:308–344. doi: 10.1016/j.jcjd.2023.04.009. [DOI] [PubMed] [Google Scholar]
- 6.Grigsby AB, Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. Prevalence of anxiety in adults with diabetes: a systematic review. J Psychosom Res. 2002;53:1053–1060. doi: 10.1016/s0022-3999(02)00417-8. [DOI] [PubMed] [Google Scholar]
- 7.Oh J, Kim TS. Serum lipid levels in depression and suicidality: the Korea National Health and Nutrition Examination Survey (KNHANES) 2014. J Affect Disord. 2017;213:51–58. doi: 10.1016/j.jad.2017.02.002. [DOI] [PubMed] [Google Scholar]
- 8.Health Insurance Review & Assessment Service. [Statistics on diseases and medical treatment in everyday life] Wonju: Health Insurance Review & Assessment Service; 2023. Korean. [Google Scholar]
- 9.Barker MM, Davies MJ, Zaccardi F, Brady EM, Hall AP, Henson JJ, et al. Age at diagnosis of type 2 diabetes and depressive symptoms, diabetes-specific distress, and self-compassion. Diabetes Care. 2023;46:579–586. doi: 10.2337/dc22-1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kim S, Kim G, Cho SH, Oh R, Kim JY, Lee YB, et al. Increased risk of incident mental disorders in adults with new-onset type 1 diabetes diagnosed after the age of 19: a nationwide cohort study. Diabetes Metab. 2024;50:101505. doi: 10.1016/j.diabet.2023.101505. [DOI] [PubMed] [Google Scholar]
- 11.Ke C, Lau E, Shah BR, Stukel TA, Ma RC, So WY, et al. Excess burden of mental illness and hospitalization in young-onset type 2 diabetes: a population-based cohort study. Ann Intern Med. 2019;170:145–154. doi: 10.7326/M18-1900. [DOI] [PubMed] [Google Scholar]
- 12.Kweon S, Kim Y, Jang MJ, Kim Y, Kim K, Choi S, et al. Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES) Int J Epidemiol. 2014;43:69–77. doi: 10.1093/ije/dyt228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Statistics Korea. [2020 population and housing census] Daejeon: Statistics Korea; 2021. Korean. [Google Scholar]
- 14.Misra S, Ke C, Srinivasan S, Goyal A, Nyriyenda MJ, Florez JC, et al. Current insights and emerging trends in early-onset type 2 diabetes. Lancet Diabetes Endocrinol. 2023;11:768–782. doi: 10.1016/S2213-8587(23)00225-5. [DOI] [PubMed] [Google Scholar]
- 15.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–613. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Han C, Jo SA, Kwak JH, Pae CU, Steffens D, Jo I, et al. Validation of the patient health questionnaire-9 Korean version in the elderly population: the Ansan Geriatric study. Compr Psychiatry. 2008;49:218–223. doi: 10.1016/j.comppsych.2007.08.006. [DOI] [PubMed] [Google Scholar]
- 17.Lee S, Huh Y, Kim J, Han C. [Finding optimal cut off points of the Korean version of the patient health questionnaire-9(PHQ-9) for screening depressive disorders] Mood Emot. 2014;12:32–36. Korean. [Google Scholar]
- 18.Armstrong T, Bull F. Development of the world health organization global physical activity questionnaire (GPAQ) J Public Health. 2006;14:66–70. [Google Scholar]
- 19.Lee J, Lee C, Min J, Kang DW, Kim JY, Yang HI, et al. Development of the Korean global physical activity questionnaire: reliability and validity study. Glob Health Promot. 2020;27:44–55. doi: 10.1177/1757975919854301. [DOI] [PubMed] [Google Scholar]
- 20.Binder DA. On the variances of asymptotically normal estimators from complex surveys. Int Stat Rev. 1983;51:279–292. [Google Scholar]
- 21.Liu S, Leone M, Ludvigsson JF, Lichtenstein P, Gudbjörnsdottir S, Landén M, et al. Early-onset type 2 diabetes and mood, anxiety, and stress-related disorders: a genetically informative register-based cohort study. Diabetes Care. 2022;45:2950–2956. doi: 10.2337/dc22-1053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Renaud-Charest O, Stoljar Gold A, Mok E, Kichler J, Nakhla M, Li P. Suicidal ideation, suicide attempts, and suicide deaths in adolescents and young adults with type 1 diabetes: a systematic review and meta-analysis. Diabetes Care. 2024;47:1227–1237. doi: 10.2337/dc24-0411. [DOI] [PubMed] [Google Scholar]
- 23.Suvila K, McCabe EL, Lehtonen A, Ebinger JE, Lima JAC, Cheng S, et al. Early onset hypertension is associated with hypertensive end-organ damage already by midlife. Hypertension. 2019;74:305–312. doi: 10.1161/HYPERTENSIONAHA.119.13069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhao M, Song L, Sun L, Wang M, Wang C, Yao S, et al. Associations of type 2 diabetes onset age with cardiovascular disease and mortality: the Kailuan study. Diabetes Care. 2021;44:1426–1432. doi: 10.2337/dc20-2375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Johnson HM, Warner RC, LaMantia JN, Bowers BJ. “I have to live like I’m old.” Young adults’ perspectives on managing hypertension: a multi-center qualitative study. BMC Fam Pract. 2016;17:31. doi: 10.1186/s12875-016-0428-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wong SKW, Soon W, Griva K, Smith HE. Identifying barriers and facilitators to self care in young adults with type 2 diabetes. Diabet Med. 2024;41:e15229. doi: 10.1111/dme.15229. [DOI] [PubMed] [Google Scholar]
- 27.Hagan J, Davies MJ, Speight J, Hadjiconstantinou M. Stigma associated with early-onset type 2 diabetes: a secondary qualitative analysis. Br J Health Psychol. 2026;31:e70042. doi: 10.1111/bjhp.70042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kang J, Chang Y, Kim Y, Shin H, Ryu S. Ten-second heart rate variability, its changes over time, and the development of hypertension. Hypertension. 2022;79:1308–1318. doi: 10.1161/HYPERTENSIONAHA.121.18589. [DOI] [PubMed] [Google Scholar]
- 29.Shah AS, El Ghormli L, Vajravelu ME, Bacha F, Farrell RM, Gidding SS, et al. Heart rate variability and cardiac autonomic dysfunction: prevalence, risk factors, and relationship to arterial stiffness in the treatment options for type 2 diabetes in adolescents and youth (TODAY) study. Diabetes Care. 2019;42:2143–2150. doi: 10.2337/dc19-0993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Galin S, Keren H. The predictive potential of heart rate variability for depression. Neuroscience. 2024;546:88–103. doi: 10.1016/j.neuroscience.2024.03.013. [DOI] [PubMed] [Google Scholar]
- 31.Lee D, Baek JH, Cho YJ, Hong KS. Association of resting heart rate and heart rate variability with proximal suicidal risk in patients with diverse psychiatric diagnoses. Front Psychiatry. 2021;12:652340. doi: 10.3389/fpsyt.2021.652340. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Baseline characteristics of the study population by adverse mental health conditions in KNHANES 2014-2023 (N=53,504)
Stepwise analysis of association between early-onset metabolic diseases and adverse mental health conditions in KNHANES 2014-2023
Subgroup analysis according to disease control status.
Subgroup analysis according to body mass index (BMI).
Sensitivity analysis according to different early-onset age thresholds.
Sensitivity analysis with additional adjustment for comorbidities (stroke, myocardial infarction, angina, and kidney disease) in KNHANES 2014-2021.
Sensitivity analysis restricted to participants aged 40 years and above.