Frequency of Low HDL Cholesterol and the Dose-Dependent Risk of Type 2 Diabetes: A Longitudinal Cohort Study

Abstract

Objective

Although low high-density lipoprotein cholesterol (HDL-C) is an established risk factor for type 2 diabetes (T2D), most studies have relied on single-point measurements. We investigated whether persistent low HDL-C levels were associated with an increased risk of incident T2D in young adults.

Methods

We included 1,840,251 individuals from the Korean National Health Insurance Service database (2009–2012) who underwent 4 consecutive annual health examinations. Participants were categorized into 5 groups according to the frequency of low HDL-C exposure (0–4 times; <40 mg/dL in men and <50 mg/dL in women). The risk of incident T2D was assessed using multivariable Cox proportional hazards models over a mean follow-up of 6.5 years.

Results

The risk of T2D increased significantly with increasing frequency of low HDL-C exposure (p<0.0001). The hazard ratios (95% confidence intervals [CIs]) for incident T2D were 1.472 (95% CI, 1.433–1.512), 1.786 (95% CI, 1.730–1.844), 1.990 (95% CI, 1.919–2.063), and 2.021 (95% CI, 1.941–2.105) for the groups with low HDL-C exposure 1, 2, 3, and 4 times, respectively, compared with the 0-time group (p for trend <0.0001). Subgroup analyses revealed consistent associations across all strata, with stronger associations among participants aged <30 years, women, and those with body mass index <25 kg/m².

Conclusion

A clear dose-response relationship was observed, indicating that cumulative exposure to low HDL-C levels was strongly associated with an increased risk of T2D. This association remained robust regardless of metabolic status, although it was more pronounced in younger adults, women, and non-obese individuals.

INTRODUCTION

Type 2 diabetes (T2D) has emerged as a major global health threat, with its incidence increasing alarmingly among young adults aged <40 years.¹,² Compared with late-onset diabetes, early-onset T2D is characterized by a more aggressive clinical course, more rapid deterioration of β-cell function, and a higher lifetime risk of debilitating microvascular and macrovascular complications.³ Despite these risks, metabolic health in young adults is often overlooked, underscoring the need for early and reliable markers to identify high-risk individuals in this age group.⁴

The prevalence of dyslipidemia among young adults is steadily increasing; however, awareness and treatment rates in this group remain the lowest across all age groups.⁵ Although low-density lipoprotein cholesterol and other metabolic markers have been extensively studied, high-density lipoprotein cholesterol (HDL-C) also plays a pivotal role in glucose homeostasis.⁶,⁷ Beyond its traditional role in reverse cholesterol transport, HDL-C has been shown to exert insulin-sensitizing, anti-inflammatory, and antioxidant effects, as well as direct protective effects on pancreatic β-cells.⁸ Numerous epidemiological studies have reported an inverse association between HDL-C levels and T2D.⁹

However, most of these studies have relied on a single baseline assessment, which may not accurately reflect the long-term metabolic burden of low HDL-C. Lipid profiles are dynamic and may fluctuate because of lifestyle changes, pharmacological interventions, or aging.¹⁰ Therefore, a 1-time measurement may fail to capture the cumulative impact of prolonged exposure to low HDL-C. Despite the potential clinical importance of longitudinal lipid patterns, evidence quantifying the cumulative burden of low HDL-C and its specific association with T2D risk remains limited, particularly in young adults.

Accordingly, this study aimed to investigate the association between the cumulative frequency of low HDL-C exposure over a 4-year period and the risk of incident T2D in a large nationwide cohort of young adults. In addition, we examined whether this relationship was independent of traditional metabolic risk factors.

MATERIALS AND METHODS

1. Data source

We conducted a nationwide population-based cohort study using data from the Korean National Health Insurance Service (NHIS) database. The NHIS is a single-payer system that covers the entire Korean population, including 97.1% of residents through the national health insurance program and approximately 3% through the Medical Aid program.¹¹ It provides comprehensive longitudinal data, including sociodemographic characteristics, medical diagnoses coded according to the International Classification of Diseases, 10th Revision (ICD-10), prescription records, and the results of annual or biennial health examinations. These screenings are mandatory for employed individuals aged 20 years or older and for all other subscribers beginning at age 40 years. The study protocol was approved by the Institutional Review Board (IRB) of Myongji Hospital (IRB File No. MJH 2022-09-030). The requirement for informed consent was waived because the database was de-identified and the study adhered to confidentiality guidelines. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.

2. Study population

A total of 2,111,470 young adults aged 20–39 years who completed 4 consecutive annual health screenings from 2009 to 2012 were initially identified. Of these, 167,310 individuals with missing data for at least 1 variable were excluded. We further excluded 102,248 participants with a preexisting diagnosis of T2D. Finally, after applying a 1-year lag period, 1,661 participants were excluded to ensure that all identified T2D cases were newly diagnosed. Ultimately, 1,840,251 participants (1,306,408 men and 533,843 women) were included in the final analysis (Fig. 1).

Fig. 1. Flow diagram of the study population.

HDL-C, high-density lipoprotein cholesterol.

3. Assessment of cumulative low HDL-C exposure

Serum HDL-C levels were measured as part of routine health examinations. Low HDL-C was defined as <40 mg/dL in men and <50 mg/dL in women, according to criteria established by the National Cholesterol Education Program Adult Treatment Panel III and the International Diabetes Federation.¹²

To estimate the cumulative impact of persistently low HDL-C, we assessed the frequency of low HDL-C over 4 consecutive annual health examinations. Each participant was assigned a cumulative exposure score ranging from 0 to 4, representing the total number of years in which the low HDL-C criteria were met (men, <40 mg/dL; women, <50 mg/dL). Specifically, a score of 0 indicated that the participant never had low HDL-C during the 4-year period, whereas scores of 1–4 indicated the number of positive screenings.

4. Covariates and definitions

The fourth health examination in 2012 was designated as the index date (baseline), at which all baseline covariates were assessed. Detailed information on anthropometric measurements, laboratory findings, and lifestyle factors was obtained from the health screening records. Blood samples were collected after an overnight fast. The estimated glomerular filtration rate was calculated using the Modification of Diet in Renal Disease formula. Height, weight, waist circumference, and blood pressure (BP) were measured according to standard protocols. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m²). Smoking status was categorized as never smoker, former smoker, or current smoker. Heavy alcohol consumption was defined as >28 standard drinks per week (equivalent to 210 g of alcohol per week). Regular exercise was defined as moderate-to-vigorous physical activity performed at least 3 times per week. Low income was defined as the lowest 25th percentile.

We identified hypertension on the basis of ICD-10 codes (I10 or I11) accompanied by antihypertensive medication use or recorded BP levels ≥140/90 mmHg. Dyslipidemia was defined as a total cholesterol concentration ≥240 mg/dL or at least 1 annual claim for a lipid-lowering drug with an ICD-10 code of E78. Participants were classified as statin users if they had received a prescription within the 12 months preceding their final health screening.

5. Outcome and follow-up

The primary outcome was incident T2D. T2D was defined as either 1) a fasting plasma glucose (FPG) level ≥126 mg/dL recorded during a health examination or 2) the presence of ICD-10 diagnosis codes E11–E14 accompanied by at least 1 annual prescription for antidiabetic medication.¹³

Participants were followed from baseline in 2012 until the date of incident T2D or December 31, 2020, whichever came first. The mean follow-up duration was 6.53 years.

6. Statistical analysis

Baseline characteristics are presented as mean ± standard deviation for continuous variables, geometric mean (95% confidence interval [CI]) for heavily skewed distributions, and number (percentage) for categorical variables. Differences across groups defined by the cumulative frequency of low HDL-C were compared using 1-way analysis of variance and the chi-square test. The incidence of T2D was calculated as the number of incident cases per 1,000 person-years of follow-up. Cumulative incidence curves for T2D were generated using the Kaplan–Meier method and compared with the log-rank test. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% CIs. Model 1 was unadjusted. Model 2 was adjusted for age and sex. Model 3 was further adjusted for BMI, systolic BP, smoking status, alcohol consumption, and physical activity. Model 4 was additionally adjusted for FPG, and Model 5 included a final adjustment for statin use. Subgroup analyses were performed after stratification by age (<30 vs. ≥30 years), sex, BMI (<25 vs. ≥25 kg/m²), FPG (<100 vs. ≥100 mg/dL), hypertension, and statin use. P for interaction was calculated to assess the consistency of the associations across subgroups. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), and statistical significance was defined as p<0.05.

RESULTS

1. Baseline characteristics

The baseline characteristics of the study population are presented in Table 1. The mean age of the overall population was 34.08±4.98 years, and 1,306,408 participants (70.99%) were men. As the cumulative frequency of low HDL-C increased, mean BMI increased from 23.21±3.37 kg/m² to 25.43±3.90 kg/m². Waist circumference, systolic and diastolic BP, triglyceride levels, and FPG levels were also positively associated with the frequency of low HDL-C. The prevalence of hypertension increased from 7.71% in the 0-frequency group to 10.51% in the 4-frequency group. In contrast, the proportions of heavy drinkers and participants who exercised regularly decreased as the frequency of low HDL-C increased. The prevalence of dyslipidemia was 9.33% in the 0-frequency group and 7.32% in the 4-frequency group. Smoking status remained relatively constant across groups at approximately 36%.

Table 1. Baseline characteristics of participants according to the cumulative frequency of low HDL-C over 4 years.

Characteristics	Total	Cumulative frequency of low HDL-C
		0	1	2	3	4
Participants	1,840,251	1,335,652	253,570	120,500	76,511	54,018
HDL-C (mg/dL)	56.1±18.18	60.15±18.21	48.95±13.82	44.37±11.82	41.12±11.23	36.72±5.75
Mean age (yr)	34.08±4.98	33.88±4.98	34.33±4.97	34.64±4.92	34.96±4.87	35.23±4.82
Male sex	1,306,408 (70.99)	975,145 (73.01)	170,018 (67.05)	78,744 (65.35)	48,850 (63.85)	33,651 (62.3)
Body mass index (kg/m2)	23.56±3.53	23.21±3.37	24.1±3.65	24.63±3.75	25.02±3.8	25.43±3.9
Waist circumference (cm)	79.58±9.73	78.87±9.47	80.63±10.06	81.74±10.16	82.54±10.13	83.36±10.19
SBP (mmHg)	119.02±12.77	118.92±12.7	119.04±12.89	119.28±12.93	119.55±13.02	119.97±13.04
DBP (mmHg)	74.97±9.24	74.87±9.19	75.04±9.31	75.25±9.39	75.44±9.41	75.71±9.48
Current smoker	664,552 (36.11)	482,578 (36.13)	91,370 (36.03)	43,429 (36.04)	27,674 (36.17)	19,501 (36.1)
Heavy alcohol drinker	165,736 (9.01)	132,927 (9.95)	19,150 (7.55)	7,380 (6.12)	3,917 (5.12)	2,362 (4.37)
Regular exerciser	336,034 (18.26)	250,155 (18.73)	44,462 (17.53)	20,142 (16.72)	12,563 (16.42)	8,712 (16.13)
Low income	100,053 (5.44)	66,049 (4.95)	16,122 (6.36)	8,425 (6.99)	5,515 (7.21)	3,942 (7.3)
Total cholesterol (mg/dL)	190.18±33.93	180.38±29.62	193.37±33.34	200.41±33.51	206.53±34.49	213.91±36.27
Triglycerides (mg/dL)	105.27 (105.18–105.36)	97.1 (97–97.19)	119.38 (119.09–119.67)	133.2 (132.73–133.68)	145.49 (144.85–146.13)	160.87 (160.02–161.72)
LDL-C (mg/dL)	109.88±32.5	109.64±32.29	110.87±32.88	111.1±32.88	110.21±33.8	107.86±33.02
FPG (mg/dL)	90.92±10.21	90.77±10.12	91.03±10.38	91.34±10.48	91.63±10.47	92.06±10.5
eGFR (mL/min)	100.67±67.77	101.04±70.38	100.06±61.17	99.53±59.32	99.34±60.44	98.77±58.2
Hypertension	149,275 (8.11)	102,990 (7.71)	21,985 (8.67)	11,104 (9.21)	7,519 (9.83)	5,677 (10.51)
Dyslipidemia	167,639 (9.11)	124,632 (9.33)	22,361 (8.82)	10,456 (8.68)	6,236 (8.15)	3,954 (7.32)
Statin use	49,141 (2.67)	32,384 (2.42)	7,789 (3.07)	4,245 (3.52)	2,702 (3.53)	2,021 (3.74)
Mean follow-up (yr)	6.53±1.23	6.53±1.22	6.53±1.24	6.51±1.28	6.49±1.32	6.43±1.38

Data are presented as mean ± standard deviation, proportion (%), or geometric mean (95% confidence interval).

Low HDL-C levels were defined as <40 mg/dL in men and <50 mg/dL in women.

HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; FPG, fasting plasma glucose; eGFR, estimated glomerular filtration rate.

2. Risk of T2D according to the cumulative frequency of low HDL-C

During follow-up, 40,286 incident cases of T2D were identified. The cumulative incidence of T2D increased significantly with increasing cumulative frequency of low HDL-C (log-rank p<0.001; Fig. 2). The incidence of T2D per 1,000 person-years increased from 2.4755 in the 0-frequency group to 9.2817 in the 4-frequency group. In the fully adjusted model (model 5), the HRs for T2D relative to the reference group were 1.472 (95% CI, 1.433–1.512), 1.786 (95% CI, 1.730–1.844), 1.990 (95% CI, 1.919–2.063), and 2.021 (95% CI, 1.941–2.105) for the 1-, 2-, 3-, and 4-frequency groups, respectively (Table 2).

Fig. 2. Cumulative incidence of type 2 diabetes according to the cumulative frequency of low high-density lipoprotein cholesterol.

Table 2. Risk of type 2 diabetes according to the cumulative frequency of low high-density lipoprotein cholesterol.

Frequency	Number	Cases	Follow-up duration (person-years)	Incidence rate (per 1,000 person-years)	HR (95% CI)
					Model 1	Model 2	Model 3	Model 4	Model 5
0	1,335,652	21,598	8,724,545.43	2.4755	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)
1	253,570	7,292	1,656,631.96	4.4017	1.776 (1.73–1.824)	1.816 (1.768–1.865)	1.528 (1.487–1.569)	1.485 (1.445–1.525)	1.472 (1.433–1.512)
2	120,500	4,631	784,068.09	5.9064	2.386 (2.311–2.463)	2.438 (2.362–2.517)	1.905 (1.845–1.967)	1.815 (1.757–1.874)	1.786 (1.73–1.844)
3	76,511	3,540	496,588.32	7.1286	2.88 (2.779–2.984)	2.933 (2.83–3.039)	2.145 (2.069–2.224)	2.012 (1.941–2.086)	1.99 (1.919–2.063)
4	54,018	3,225	347,457.15	9.2817	3.76 (3.624–3.902)	3.829 (3.689–3.974)	2.041 (1.957–2.129)	2.013 (1.932–2.097)	2.021 (1.941–2.105)

Model 1 was unadjusted. Model 2 was adjusted for age and sex. Model 3 was adjusted for age, sex, body mass index, systolic blood pressure, smoking status, alcohol consumption, and physical activity. Model 4 was adjusted for the variables in model 3 plus fasting plasma glucose. Model 5 was adjusted for the variables in model 4 plus statin use.

HR, hazard ratio; CI, confidence interval; Ref., reference.

3. Subgroup analyses

In subgroup analyses, the positive association between cumulative low HDL-C exposure and T2D risk remained consistent across all strata, including age, sex, BMI, FPG, hypertension, and statin use (Table 3). Significant interactions were observed for all variables (all p-values for interaction <0.001), with more pronounced associations among participants aged <30 years, women, and those with a BMI <25 kg/m². At a frequency of 4, the HRs (95% CIs) were 3.319 (2.960–3.722) for participants aged <30 years, 2.988 (2.767–3.227) for women, and 2.747 (2.518–2.996) for those with a BMI <25 kg/m². In contrast, the corresponding HRs for participants aged ≥30 years, men, and those with a BMI ≥25 kg/m² were 1.918, 1.779, and 1.792, respectively.

Table 3. Subgroup analysis for the risk of type 2 diabetes.

Subgroup		Frequency	Incidence rate (per 1,000 person-years)	HR (95% CI)
Age
	<30	0	1.23473	1 (Ref.)
		1	2.49249	1.607 (1.478–1.747)
		2	3.91582	2.239 (2.031–2.469)
		3	5.00394	2.49 (2.226–2.785)
		4	7.51425	3.319 (2.96–3.722)
	≥30	0	2.79253	1 (Ref.)
		1	4.81788	1.457 (1.417–1.499)
		2	6.29083	1.743 (1.685–1.803)
		3	7.48627	1.944 (1.872–2.019)
		4	9.54814	1.918 (1.837–2.003)
Sex
	Male	0	3.0299	1 (Ref.)
		1	5.5234	1.461 (1.42–1.504)
		2	7.2212	1.738 (1.677–1.8)
		3	8.4969	1.923 (1.846–2.002)
		4	10.6002	1.779 (1.695–1.867)
	Female	0	0.9596	1 (Ref.)
		1	2.116	1.583 (1.474–1.7)
		2	3.4326	2.104 (1.949–2.272)
		3	4.7147	2.369 (2.188–2.564)
		4	7.1118	2.988 (2.767–3.227)
BMI
	<25	0	1.0131	1 (Ref.)
		1	1.5476	1.465 (1.387–1.547)
		2	2.107	1.916 (1.789–2.052)
		3	2.4259	2.128 (1.962–2.309)
		4	3.2594	2.757 (2.527–3.008)
	≥25	0	6.3797	1 (Ref.)
		1	9.2189	1.345 (1.304–1.387)
		2	10.9972	1.555 (1.5–1.613)
		3	12.4944	1.691 (1.624–1.76)
		4	15.0693	1.778 (1.703–1.857)
FPG
	<100 mg/dL	0	1.3995	1 (Ref.)
		1	2.5454	1.531 (1.472–1.591)
		2	3.489	1.928 (1.841–2.02)
		3	4.3876	2.23 (2.119–2.347)
		4	5.6289	1.82 (1.709–1.938)
	≥100 mg/dL	0	7.5833	1 (Ref.)
		1	12.6998	1.417 (1.366–1.47)
		2	16.1191	1.665 (1.593–1.74)
		3	18.3149	1.79 (1.703–1.882)
		4	23.2776	2.171 (2.062–2.287)
Hypertension
	No	0	2.0083	1 (Ref.)
		1	3.5866	1.506 (1.461–1.554)
		2	4.9279	1.875 (1.807–1.945)
		3	5.998	2.122 (2.036–2.211)
		4	7.9322	2.067 (1.972–2.166)
	Yes	0	8.1049	1 (Ref.)
		1	13.1623	1.359 (1.289–1.433)
		2	15.8278	1.539 (1.443–1.642)
		3	17.8332	1.626 (1.512–1.748)
		4	21.2577	1.935 (1.792–2.09)
Statin use
	No	0	2.2815	1 (Ref.)
		1	4.0506	1.487 (1.446–1.53)
		2	5.4983	1.844 (1.782–1.907)
		3	6.6714	2.05 (1.974–2.129)
		4	8.7	2.028 (1.942–2.117)
	Yes	0	10.3632	1 (Ref.)
		1	15.8054	1.34 (1.235–1.454)
		2	17.5451	1.37 (1.239–1.514)
		3	20.2004	1.533 (1.367–1.719)
		4	25.1532	1.954 (1.734–2.202)

All p-values for interaction values across subgroups were <0.001.

HR, hazard ratio; CI, confidence interval; Ref., reference; BMI, body mass index; FPG, fasting plasma glucose.

DISCUSSION

In this large longitudinal study of approximately 1.8 million young adults, we found that cumulative exposure to low HDL-C was strongly and independently associated with T2D risk. The risk of T2D increased stepwise as the frequency of low HDL-C exposure increased from 1 to 4 years, demonstrating a significant cumulative dose-response relationship. This association was particularly pronounced among participants aged <30 years, women, and those with a BMI <25 kg/m². These findings suggest that a chronic burden of low HDL-C may substantially increase susceptibility to T2D in young adults.

Although inverse associations between a single HDL-C measurement and T2D risk have been reported in various populations, our study extends this literature by focusing on chronic exposure in a young population. Most previous studies have used cross-sectional designs or single baseline measurements, which may not capture the long-term metabolic effects of persistent dyslipidemia.¹⁴,¹⁵ The duration of exposure to metabolic derangements has emerged as a more reliable predictor of diabetes than single-point assessments, highlighting the importance of evaluating long-term HDL-C trajectories.¹⁶ By using 4 consecutive years of health examination data, our study shows that persistent low HDL-C is an important determinant of metabolic health in young adulthood, a period during which early intervention may yield substantial long-term benefits.

The biological mechanisms linking chronically low HDL-C to T2D are likely multifactorial.¹⁷ HDL-C is known to protect pancreatic β-cells by inhibiting apoptosis and promoting insulin secretion through the ABCA1/G1 pathways.¹⁸ Persistent exposure to low HDL-C may gradually reduce β-cell mass and function over time, thereby accelerating progression to overt diabetes.¹⁹ In addition to its role in reverse cholesterol transport, HDL-C enhances glucose uptake in skeletal muscle through activation of AMP-activated protein kinase.²⁰ Furthermore, chronically low HDL-C is often associated with a proinflammatory state and increased oxidative stress, both of which are key drivers of systemic insulin resistance.²¹,²² Our study demonstrated a cumulative effect, whereby T2D risk increased stepwise as the duration of low HDL-C exposure increased from 1 to 4 years. These findings suggest that a chronic, rather than transient, burden of low HDL-C may be particularly detrimental because of the sustained loss of protective metabolic functions over time.

Subgroup analyses showed that the positive association between cumulative low HDL-C exposure and T2D risk remained consistent across all strata, further supporting the robustness of our findings. The association was stronger in participants younger than 30 years, suggesting that early-life exposure to chronic dyslipidemia may have a greater effect on long-term metabolic health.²³ This pattern may reflect greater relative β-cell sensitivity to metabolic stress in younger individuals, as well as the absence of competing risk factors that are more common in older populations.²⁴ In addition, although the absolute incidence of T2D was lower in women than in men, the relative risk associated with cumulative low HDL-C exposure was significantly greater in women. This disparity may reflect the lower baseline risk in women, partly because of the protective effects of estrogen, thereby amplifying the relative impact of chronic metabolic insults.²⁵ Notably, the association was also stronger in participants with a BMI <25 kg/m², indicating a greater incremental effect of cumulative low HDL-C exposure in non-obese individuals. These findings illustrate the residual metabolic risk associated with the “metabolically unhealthy lean” phenotype and may reflect an intrinsic susceptibility to dyslipidemia-induced β-cell dysfunction.²⁶

These findings have important clinical and public health implications. Although current screening guidelines for T2D often focus on FPG or glycated hemoglobin levels, our data suggest that monitoring long-term HDL-C patterns may provide additional prognostic value, particularly in young adults who might otherwise be considered at low risk. Early lifestyle or pharmacological interventions aimed at correcting persistently low HDL-C may be important for reducing the growing burden of T2D in younger populations.

The main strengths of this study include its large sample size and the use of repeated longitudinal HDL-C measurements, which enabled estimation of cumulative exposure. However, several limitations should be noted. First, although 4 consecutive health examinations were used to assess cumulative exposure and reduce the variability inherent in single-point measurements, annual screening data may not capture short-term fluctuations in HDL-C. Second, despite the high follow-up rate, the incidence of T2D may have been underestimated because cases were identified primarily through insurance claims and FPG levels. Third, because of the observational study design, residual confounding cannot be excluded, and causal relationships cannot be established definitively. Fourth, triglycerides were excluded from the models to avoid multicollinearity arising from their strong inverse correlation with HDL-C. Instead, we adjusted for other major metabolic components to minimize confounding. Finally, the requirement for 4 consecutive annual measurements may have introduced selection bias, although we attempted to mitigate this by adjusting for baseline comorbidities and lifestyle factors in the models. In addition, the study population consisted of young Korean adults, which may limit the generalizability of the findings to older populations and other ethnic groups.

In conclusion, this large nationwide cohort study demonstrated that cumulative exposure to low HDL-C was strongly and independently associated with T2D in young adults. The relative impact of persistently low HDL-C was more pronounced in individuals traditionally considered to be at lower risk, including women, younger adults, and those with a BMI <25 kg/m². These findings underscore that the long-term trajectory of HDL-C may provide important prognostic information beyond single-point measurements or traditional glycemic markers. Accordingly, clinical strategies should emphasize early identification and sustained management of persistently low HDL-C in this population. Such proactive approaches may help mitigate the growing global burden of T2D.