Combined effects of glycemic status and adiposity on cardiovascular risk in chronic kidney disease: A nationwide population-based study

Abstract

BACKGROUND

Obesity and diabetes are well-established risk factors for cardiovascular disease (CVD), and their coexistence is particularly detrimental in chronic kidney disease (CKD). However, the interactions between various adiposity patterns and glycemic status in influencing CVD outcomes in CKD remain inadequately defined.

AIM

To evaluate the combined effects of diabetes, body mass index (BMI), and waist circumference (WC) on CVD risk.

METHODS

We analyzed data from 1714859 adults with CKD sourced from the Korean National Health Insurance database. Participants were classified into three glycemic groups: Normoglycemia, impaired fasting glucose (IFG), and diabetes mellitus (DM). BMI and WC were further categorized into five and six levels, respectively. Incident CVD events and all-cause mortality were assessed across the combined categories of glycemic status and adiposity. Incidence rates and adjusted hazard ratios were computed using Cox proportional hazards models.

RESULTS

A significant interaction was identified between glycemic status and adiposity indices concerning CVD risk (P for interaction < 0.001). Among normoglycemic individuals, both underweight (BMI < 18.5 kg/m²) and central obesity (WC ≥ 100/95 cm in men/women) were associated with increased CVD risk and mortality. In individuals with IFG, underweight remained a consistent risk factor, while WC displayed a linear relationship with CVD but not with mortality. In those with DM, the highest CVD risk was observed in individuals who were underweight (BMI < 18.5 kg/m²) and had low WC (< 80 cm in men/< 75 cm in women).

CONCLUSION

Cardiovascular risk is jointly influenced by glycemic status and adiposity, with diabetes consistently elevating risk across all BMI and WC categories, underscoring the importance of their assessment in CKD.

Key Words: Chronic kidney disease; Cardiovascular disease; Diabetes mellitus; Waist circumference; Body mass index; Obesity

Core Tip: This nationwide cohort study of over 1.7 million individuals with chronic kidney disease (CKD) demonstrates that cardiovascular (CV) risk is strongly modified by both glycemic status and patterns of adiposity. Diabetes consistently amplified CV risk across all body mass index and waist circumference categories, negating any protective effect of higher adiposity. Conversely, underweight and centrally lean individuals with diabetes exhibited the greatest vulnerability, underscoring that leanness is not universally protective in CKD. These findings highlight the importance of integrating both glycemic status and obesity indices into individualized CV risk assessment and prevention strategies for patients with CKD.

INTRODUCTION

Cardiovascular disease (CVD) is the leading cause of morbidity and mortality globally, imposing a particularly heavy burden on patients with chronic kidney disease (CKD)[1]. Both diabetes mellitus (DM) and obesity are well-established risk factors for adverse cardiovascular (CV) outcomes, and their coexistence with CKD exacerbates vascular damage and worsens prognosis[2]. As the prevalence of CKD continues to rise in aging populations, understanding the interplay of metabolic risk factors in influencing CV health has become a critical clinical priority.

Obesity is traditionally assessed using body mass index (BMI), which reflects overall adiposity, and waist circumference (WC), which serves as a surrogate for central obesity. While both indices are associated with CVD in the general population, evidence suggests that their predictive values may vary across different metabolic states[3]. Notably, BMI does not account for body fat distribution, while WC more accurately reflects visceral adiposity, which is closely linked to cardiometabolic risk[4]. In patients with CKD, the interpretation of anthropometric data is further complicated by factors such as malnutrition, sarcopenia, and fluid overload, raising concerns regarding the reliability of obesity indices in predicting CV outcomes in this population.

DM significantly heightens CV risk; however, the extent to which glycemic status modifies the relationship between obesity indices and CVD in CKD remains inadequately understood. Studies in the general population have reported paradoxical associations, including the so-called “obesity paradox”, as well as varying effects based on glycemic status[5,6]. Nevertheless, few large-scale investigations have systematically examined the joint effects of diabetes status and adiposity indices on CVD outcomes in CKD.

Considering the limited evidence regarding the mechanism by which diabetes status modifies the relationship between adiposity and CV outcomes in CKD, we conducted a nationwide cohort study using the Korean National Health Insurance Service (KNHIS) database. We evaluated the combined effects of diabetes status and obesity indices—BMI and WC—on the incidence of CVD and all-cause mortality among patients with CKD. By simultaneously stratifying patients according to glycemic status and adiposity measures, our study aimed to not only refine CV risk stratification in CKD but also provide clinically actionable insights that may guide individualized prevention strategies and optimize long-term management in this high-risk population.

MATERIALS AND METHODS

KNHIS data source

In this study, we utilized the national health insurance claims database established by the KNHIS, which includes all claims data from the KNHIS and Medical Aid programs. The KNHIS database is considered representative of the entire South Korean population, and its details have previously been described[7,8]. Consequently, data extracted from the KNHIS database represent approximately 50 million insured Koreans. All insured individuals aged over 40 years undergo a biannual health checkup supported by the KNHIS, while employees aged over 20 years are mandated to undergo a health checkup annually. The KNHIS databases generated through these checkups provide wide-ranging information, including anthropometric measurements, health questionnaire data, and laboratory findings. Anonymized data are publicly available from the KNHIS and can be accessed at https://nhiss.nhis.or.kr/bd/ab/bdaba000eng.do. This study was approved by the Chonnam National University Hospital (Approval No. CNUH-EXP-2025-219) and conducted in accordance with the principles of the Declaration of Helsinki. Our review board waived the requirement for written informed consent.

Study design and population

All participants who underwent KNHIS health checkups from 2012 to 2017 (n = 2219145) were initially enrolled in the study and followed up until the end of 2022. Participants aged under 20 years (n = 2398), those with missing data on baseline characteristics and covariates (n = 85117), and individuals with pre-existing end-stage renal disease (n = 22641) or underlying CVD (n = 369430) at baseline were excluded from the study. Additionally, participants excluded during the 1-year lag period (n = 24700) were also removed. The final sample comprised 1714859 participants who were evaluated based on their DM status and obesity indices (Figure 1).

Figure 1 Study design. CKD: Chronic kidney disease; ESRD: End-stage renal disease; CVD: Cardiovascular disease; DM: Diabetes mellitus; BMI: Body mass index; WC: Waist circumference; IFG: Impaired fasting glucose.

Measurements and definitions

The diagnosis of diabetes at baseline was determined based on a fasting blood glucose level ≥ 126 mg/dL or the presence of International Classification of Diseases 10 (ICD-10) codes E11-14 along with a claim for anti-diabetic medication[7]. Hypertension (systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg; ICD-10 codes I10-I15) and hyperlipidemia (total cholesterol ≥ 240 mg/dL; ICD-10 code E78) diagnoses were confirmed using laboratory data, accompanied by a claim for medication specific to each condition. Ischemic heart disease was identified using ICD-10 codes I21-25. CKD was defined as an estimated glomerular filtration rate (eGFR) < 60 mL/minute/1.73 m², calculated using the Modification of Diet in Renal Disease equation[8]. Participants were categorized into three groups based on their smoking status: Non-smokers, current smokers, and former smokers. Additionally, participants were grouped according to alcohol consumption into non-drinkers, moderate drinkers, and heavy drinkers (defined as consuming > 30 g of alcohol per day). Regular exercise was defined by responses to the question: “During the last week, did you exercise vigorously for over 20 minutes on more than 3 days until you were almost out of breath?” Income was stratified into quartiles (Q): Q1 (the lowest), Q2, Q3, and Q4 (the highest), with the lowest income level defined as the bottom 25% of income. Urban residency was also assessed. The quality of laboratory tests is upheld by the Korean Association for Laboratory Medicine, while the KNHIS certifies hospitals participating in the National Health Insurance health check-up programs. Proteinuria was assessed using the dipstick method, with results classified as negative, trace, or graded from 1+ to 4+. Energy expenditure was denoted as the “minimum” amount of energy consumption calculated from a self-reported survey question regarding the frequency of each exercise intensity over a specified duration.

Body weight (kg) and height (cm) were measured using an electronic scale, while WC (cm) was measured at the midpoint between the rib cage and iliac crest by trained examiners. Obesity was defined as a BMI ≥ 25 kg/m², and patients were classified according to the World Health Organization’s recommendations for Asians as follows: Underweight (BMI < 18.5 kg/m²), normal weight (BMI = 18.5-23 kg/m²), overweight (BMI = 23-25 kg/m²), stage I obesity (BMI = 25-30 kg/m²), and stage II obesity (BMI ≥ 30 kg/m²)[9]. Abdominal obesity was defined as a WC ≥ 90 cm for men and 85 cm for women, in accordance with the criteria established by the Korean Society for the Study of Obesity[10]. WC was divided into six categories based on 5-cm increments: (1) Level 1: < 80 cm in men and < 75 cm in women; (2) Level 2: 80-85 cm in men and 75-80 cm in women; (3) Level 3: 85-90 cm in men and 80-85 cm in women; (4) Level 4: 90-95 cm in men and 85-90 cm in women; (5) Level 5: 95-100 cm in men and 90-95 cm in women; and (6) Level 6: ≥ 100 cm in men and ≥ 95 cm in women.

Study outcomes and follow-up

The study endpoints included incident CVD, myocardial infarction (MI), stroke, and all-cause mortality. CVD was defined as a composite of MI and stroke, based on hospitalization records with relevant ICD-10 codes. MI and stroke were defined as the first hospitalization with a primary discharge diagnosis of I21-I22 for MI and I63-I64 for ischemic or hemorrhagic stroke. All-cause mortality was ascertained based on national mortality records linked to the KNHIS database[11].

The follow-up period for each outcome commenced 365 days after the index health checkup date to minimize reverse causation (1-year lag period). The time to each event (CVD, MI, stroke, and death) was calculated as the interval from the index date + 365 days to the first occurrence of the corresponding outcome. Participants were censored at the date of death (for non-fatal outcomes) or at the conclusion of the follow-up period, whichever preceded. The study termination date for follow-up was December 31, 2022.

The mean follow-up duration ± SD was 7.42 ± 1.96 years for CVD, 7.54 ± 1.83 years for MI, 7.51 ± 1.87 years for stroke, and 7.64 ± 1.72 years for all-cause mortality.

Statistical analysis

Data are presented as the mean ± SD for continuous variables and expressed as proportions for categorical variables. Non-normally distributed data are reported as median values (25^th, 75^th percentile). Comparisons between groups defined by anthropometric indices, including BMI, WC, and diabetes-specific indices, were performed using Student’s t-test for continuous variables and the χ² test for categorical variables.

The incidence rates of CVD, MI, stroke, and all-cause mortality were calculated per 1000 person-years. To determine whether the association between glycemic status and CV outcomes varied by adiposity level, we examined statistical interactions between glycemic status [normal, impaired fasting glucose (IFG), and diabetes] and each adiposity index (BMI and WC). Multiplicative interaction terms (glycemic status × adiposity index) were incorporated into multivariable Cox proportional hazards models, adjusting for covariates including age, sex, income, smoking, alcohol consumption, regular exercise, hypertension, and dyslipidemia. The significance of the interactions was assessed using the likelihood ratio test, which compared models with and without the interaction terms. Hazard ratios (HRs) and 95% confidence intervals for each combination of glycemic status and adiposity category were subsequently estimated.

Analyses were conducted in three models: Model 1 (unadjusted), model 2 (adjusted for age and sex), and model 3 (adjusted for model 2 plus income, smoking, alcohol consumption, regular exercise, hypertension, and dyslipidemia). Analyses were performed using SAS software (version 9.4; SAS Institute Inc., Cary, NC, United States).

RESULTS

Baseline characteristics

Table 1 presents the baseline characteristics of participants stratified by glycemic status. Among 1714859 participants, those with DM were older, predominantly male, and exhibited a higher prevalence of hypertension, dyslipidemia, obesity, smoking, and heavy alcohol consumption compared with those with normal glucose levels or IFG. Additionally, they demonstrated poorer metabolic and renal profiles, characterized by higher values for BMI, WC, blood pressure, triglyceride content, and fasting glucose, alongside lower levels of high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, and eGFR, as well as a greater incidence of proteinuria. Table 2 summarizes baseline characteristics categorized by BMI. Participants with elevated BMI were generally younger, tended to be male, and displayed a higher prevalence of hypertension, dyslipidemia, and other cardiometabolic risk factors. In contrast, underweight individuals tended to be older, with lower eGFR values and a higher prevalence of proteinuria. Table 3 details baseline characteristics according to WC categories. Higher WC values correlated with older age, male predominance, and a substantially higher prevalence of hypertension, dyslipidemia, obesity, and associated cardiometabolic risk factors. Blood pressure, fasting glucose, triglycerides, and BMI progressively increased across WC categories, while HDL-C levels decreased. Participants exhibiting the highest WC values also displayed elevated rates of proteinuria, although eGFR values appeared to be lower in the mid-range groups. Conversely, individuals with the lowest WC values were younger, had fewer cardiometabolic comorbidities, yet demonstrated a relatively higher prevalence of proteinuria.

Table 1 Baseline characteristics of subjects according to diabetes mellitus status.

Group	Total (n = 1714859)	Normal (n = 857846)	IFG (n = 457410)	DM (n = 399603)	P value
Age (year)	57.78 ± 14.68	54.81 ± 15.65	59.25 ± 13.66	62.50 ± 11.89	< 0.0001
39	198542 (11.58)	149867 (17.47)	35887 (7.85)	12788 (3.20)	< 0.0001
40-64	902922 (52.65)	449993 (52.46)	249689 (54.59)	203240 (50.86)
≥ 65	613395 (35.77)	257986 (30.07)	171834 (37.57)	183575 (45.94)
Sex (male)	815109 (47.53)	350599 (40.87)	236377 (51.68)	228133 (57.09)	< 0.0001
Smoking					< 0.0001
Non	1119098 (65.26)	599967 (69.94)	286581 (62.65)	232550 (58.2)
Former	284742 (16.60)	118485 (13.81)	85402 (18.67)	80855 (20.23)
Current	311019 (18.14)	139394 (16.25)	85427 (18.68)	86198 (21.57)
Heavy drinker1					< 0.0001
Non	1068788 (62.33)	544658 (63.49)	271279 (59.31)	252851 (63.28)
Mild	534295 (31.16)	272101 (31.72)	149782 (32.75)	112412 (28.13)
Heavy	111776 (6.52)	41087 (4.79)	36349 (7.95)	34340 (8.59)
Physical activity-regular	351280 (20.48)	174600 (20.35)	94593 (20.68)	82087 (20.54)	< 0.0001
Income-low2	373049 (21.75)	183523 (21.39)	97205 (21.25)	92321 (23.1)	< 0.0001
Hypertension	883385 (51.51)	336506 (39.23)	249054 (54.45)	297825 (74.53)	< 0.0001
Dyslipidemia	613846 (35.8)	229687 (26.77)	165923 (36.27)	218236 (54.61)	< 0.0001
eGFR < 60 mL/minute/1.73 m2	991664 (57.83)	472872 (55.12)	279726 (61.15)	239066 (59.83)	< 0.0001
Proteninuria, positive	799235 (46.61)	411203 (47.93)	195049 (42.64)	192983 (48.29)	< 0.0001
Height, cm	161.43 ± 9.46	161.22 ± 9.36	161.73 ± 9.62	161.53 ± 9.48	< 0.0001
Weight, kg	63.98 ± 12.43	62.00 ± 11.90	65.46 ± 12.63	66.52 ± 12.60	< 0.0001
BMI, kg/m2	24.44 ± 3.55	23.75 ± 3.41	24.90 ± 3.49	25.39 ± 3.61	< 0.0001
Waist circumference, cm	82.70 ± 9.76	80.17 ± 9.60	83.97 ± 9.25	86.68 ± 9.05	< 0.0001
Fasting glucose, mg/dL	108.24 ± 36.02	88.85 ± 7.41	108.53 ± 6.79	149.52 ± 53.73	< 0.0001
SBP, mmHg	126.37 ± 16.40	123.28 ± 15.91	128.38 ± 16.05	130.71 ± 16.47	< 0.0001
DBP, mmHg	77.69 ± 10.69	76.40 ± 10.49	79.16 ± 10.70	78.78 ± 10.75	< 0.0001
Total cholesterol, mg/dL	198.27 ± 41.31	197.97 ± 38.72	205.44 ± 41.04	190.72 ± 45.40	< 0.0001
HDL-C, mg/dL	53.87 ± 16.26	55.79 ± 16.69	53.81 ± 15.82	49.82 ± 15.00	< 0.0001
LDL-C, mg/dL	116.05 ± 37.74	117.34 ± 35.65	121.51 ± 38.03	107.02 ± 40.10	< 0.0001
eGFR, mL/minute/1.73 m2	68.52 ± 24.02	71.03 ± 25.23	66.63 ± 22.03	65.30 ± 22.94	< 0.0001
Triglyceride, mg/dL	122.65 (122.54-122.75)	107.31 (107.18-107.43)	132.22 (132.01-132.43)	149.92 (149.67-150.18)	< 0.0001

¹Alcohol consumptions ≥ 30 g/day.

²Low income 25%.

Data were showed as mean ± SD, n (%) or median (range). Abdominal obesity: Waist circumference ≥ 90 in male, ≥ 85 in female. DM: Diabetes mellitus; IFG: Impaired fasting glucose; eGFR: Estimated glomerular filtration rate; BMI: Body mass index; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol.

Table 2 Baseline characteristics of participants by level of body mass index.

Group	BMI (kg/m2)					P value
	< 18.5 (n = 56566)	< 23 (n = 535450)	< 25 (n = 410244)	< 30 (n = 602426)	≥ 30 (n = 110173)
Age (year)	51.24 ± 20.28	56.65 ± 15.99	59.59 ± 13.39	58.95 ± 13.19	53.53 ± 14.69	< 0.0001
39	19502 (34.48)	80030 (14.95)	30517 (7.44)	47728 (7.92)	20765 (18.85)	< 0.0001
40-64	19602 (34.65)	270107 (50.44)	220489 (53.75)	332408 (55.18)	60316 (54.75)
≥ 65	17462 (30.87)	185313 (34.61)	159238 (38.82)	222290 (36.9)	29092 (26.41)
Sex (male)	16627 (29.39)	210349 (39.28)	207643 (50.61)	326356 (54.17)	54134 (49.14)	< 0.0001
Smoking						< 0.0001
Non	41252 (72.93)	374889 (70.01)	263541 (64.24)	370438 (61.49)	68978 (62.61)
Former	4484 (7.93)	67386 (12.58)	74651 (18.20)	120838 (20.06)	17383 (15.78)
Current	10830 (19.15)	93175 (17.40)	72052 (17.56)	111150 (18.45)	23812 (21.61)
Drinking1						< 0.0001
Non	36289 (64.15)	345181 (64.47)	256710 (62.57)	364598 (60.52)	66010 (59.91)
Mild	17634 (31.17)	162628 (30.37)	127973 (31.19)	191763 (31.83)	34297 (31.13)
Heavy	2643 (4.67)	27641 (5.16)	25561 (6.23)	46065 (7.65)	9866 (8.96)
Physical activity-regular2	7235 (12.79)	106891 (19.96)	91849 (22.39)	126601 (21.02)	18704 (16.98)	< 0.0001
Income-low3	13067 (23.10)	120206 (22.45)	87657 (21.37)	127458 (21.16)	24661 (22.38)	< 0.0001
Hypertension	14261 (25.21)	206166 (38.5)	212345 (51.76)	372258 (61.79)	78355 (71.12)	< 0.0001
Dyslipidemia	8264 (14.61)	143192 (26.74)	150791 (36.76)	258843 (42.97)	52756 (47.88)	< 0.0001
eGFR < 60 mL/minute/1.73 m2	23040 (40.73)	298062 (55.67)	255719 (62.33)	363721 (60.38)	51122 (46.40)	< 0.0001
Proteninuria, positive	35622 (62.97)	259211 (48.41)	172597 (42.07)	267338 (44.38)	64467 (58.51)	< 0.0001
Height, cm	160.74 ± 8.76	160.7 ± 8.93	161.42 ± 9.38	161.95 ± 9.74	162.48 ± 10.69	< 0.0001
Weight, kg	45.11 ± 5.50	54.95 ± 6.92	62.64 ± 7.45	70.7 ± 9.21	85.8 ± 13.26	< 0.0001
BMI, kg/m2	17.4 ± 0.92	21.21 ± 1.2	23.95 ± 0.57	26.86 ± 1.33	32.35 ± 2.53	< 0.0001
Waist circumference, cm	66.44 ± 6.20	75.24 ± 6.61	82.12 ± 5.83	88.33 ± 6.37	98.71 ± 8.13	< 0.0001
Fasting glucose, mg/dL	98.56 ± 34.86	103.29 ± 34.63	107.86 ± 34.75	111.82 ± 36.18	119.11 ± 41.66	< 0.0001
SBP, mmHg	116.46 ± 16.62	122.15 ± 16.2	126.51 ± 15.6	129.54 ± 15.56	134.12 ± 16.85	< 0.0001
DBP, mmHg	72.68 ± 10.47	75.16 ± 10.27	77.49 ± 10.13	79.56 ± 10.42	83.07 ± 11.85	< 0.0001
Total cholesterol, mg/dL	185.47 ± 37.79	194.51 ± 39.8	199.27 ± 41.18	201.09 ± 42.17	204.02 ± 43.11	< 0.0001
HDL-C, mg/dL	63.82 ± 19.06	57.73 ± 17.3	53.03 ± 15.56	50.85 ± 14.68	49.66 ± 14.08	< 0.0001
LDL-C, mg/dL	102.76 ± 33.77	113.4 ± 36.16	117.81 ± 37.74	118.01 ± 38.69	118.47 ± 39.65	< 0.0001
eGFR, mL/minute/1.73 m2	78.73 ± 29.01	70.14 ± 25.38	66.19 ± 22.44	66.76 ± 22.43	73.77 ± 25.75	< 0.0001
Triglyceride, mg/dL	82.17 (81.84-82.52)	100.83 (100.68-100.98)	124.19 (123.99-124.4)	142.81 (142.61-143)	162.06 (161.55-162.58)	< 0.0001

¹Alcohol consumptions ≥ 30 g/day.

²Regular exercise: Mid-term exercise ≥ 5 days or vigorous exercise ≥ 3 days in a week.

³Low income 25%.

Data were showed as mean ± SD, n (%) or median (range). eGFR: Estimated glomerular filtration rate; BMI: Body mass index; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol.

Table 3 Baseline characteristics of participants by level of waist circumferences.

Group	WCs (male/female)
	< 80/< 75 (n = 445617)	< 85/< 80 (n = 358075)	< 90/< 85 (n = 378139)	< 95/< 90 (n = 273818)	< 100/< 95 (n = 149269)	≥ 100/≥ 95 (n = 109941)	P value
Age (year)	52.37 ± 16.17	58.13 ± 13.73	60.06 ± 13.20	60.82 ± 13.16	61.02 ± 13.66	58.81 ± 15.12	< 0.0001
39	198542 (11.58)	96433 (21.64)	32726 (9.14)	26123 (6.91)	17981 (6.57)	11316 (7.58)	< 0.0001
40-64	902922 (52.65)	239284 (53.70)	202415 (56.53)	201019 (53.16)	138486 (50.58)	70970 (47.55)
≥ 65	613395 (35.77)	109900 (24.66)	122934 (34.33)	150997 (39.93)	117351 (42.86)	66983 (44.87)
Sex (male)	815109 (47.53)	160554 (36.03)	184340 (51.48)	198169 (52.41)	147268 (53.78)	73968 (49.55)	< 0.0001
Smoking							< 0.0001
Non	321941 (72.25)	226439 (63.24)	236728 (62.60)	167768 (61.27)	95150 (63.74)	321941 (72.25)
Former	48723 (10.93)	62843 (17.55)	72346 (19.13)	55303 (20.20)	27670 (18.54)	48723 (10.93)
Current	74953 (16.82)	68793 (19.21)	69065 (18.26)	50747 (18.53)	26449 (17.72)	74953 (16.82)
Drinking1							< 0.0001
Non	278821 (62.57)	218854 (61.12)	234984 (62.14)	169578 (61.93)	95478 (63.96)	71073 (64.65)
Mild	145938 (32.75)	116100 (32.42)	117063 (30.96)	83123 (30.36)	42164 (28.25)	29907 (27.20)
Heavy	20858 (4.68)	23121 (6.46)	26092 (6.90)	21117 (7.71)	11627 (7.79)	8961 (8.15)
PA-regular2	91577 (20.55)	80060 (22.36)	81067 (21.44)	54497 (19.90)	26872 (18.00)	17207 (15.65)	< 0.0001
Income-low3	100487 (22.55)	76710 (21.42)	80359 (21.25)	58281 (21.28)	32372 (21.69)	100487 (22.55)	< 0.0001
Hypertension	133696 (30.00)	169039 (47.21)	215195 (56.91)	176686 (64.53)	105284 (70.53)	83485 (75.94)	< 0.0001
Dyslipidemia	97624 (21.91)	120857 (33.75)	150220 (39.73)	119632 (43.69)	70589 (47.29)	54924 (49.96)	< 0.0001
eGFR < 60 mL/minute/1.73 m2	217826 (48.88)	212775 (59.42)	236063 (62.43)	171542 (62.65)	92273 (61.82)	61185 (55.65)	< 0.0001
Proteninuria, positive	243137 (54.56)	160187 (44.74)	159386 (42.15)	116057 (42.38)	65197 (43.68)	55271 (50.27)	< 0.0001
Height, cm	160.3 ± 8.52	161.43 ± 9.18	161.64 ± 9.58	162.18 ± 9.91	162.02 ± 10.24	162.57 ± 10.75	< 0.0001
Weight, kg	54.07 ± 7.75	61.23 ± 8.52	65.2 ± 9.39	69.53 ± 10.31	73.53 ± 11.45	82.1 ± 15.06	< 0.0001
BMI, kg/m2	21 ± 2.17	23.42 ± 1.95	24.86 ± 2.05	26.33 ± 2.19	27.88 ± 2.40	30.88 ± 3.55	< 0.0001
WC, cm	70.92 ± 4.97	79.6 ± 2.91	84.52 ± 2.9	89.45 ± 2.82	94.14 ± 2.88	101.99 ± 5.56	< 0.0001
FBS, mg/dL	99.68 ± 31.06	106.53 ± 34.58	109.96 ± 35.92	112.9 ± 37.22	115.71 ± 38.98	120.79 ± 43.36	< 0.0001
SBP, mmHg	119.97 ± 15.86	125.62 ± 15.69	127.97 ± 15.59	129.87 ± 15.71	131.38 ± 15.91	133.72 ± 16.80	< 0.0001
DBP, mmHg	74.46 ± 10.25	77.24 ± 10.26	78.42 ± 10.29	79.42 ± 10.52	80.29 ± 10.79	81.91 ± 11.65	< 0.0001
TC, mg/dL	193.52 ± 38.91	198.98 ± 40.9	200.18 ± 41.8	200.21 ± 42.38	200.46 ± 43.09	200.89 ± 43.59	< 0.0001
HDL-C, mg/dL	59.99 ± 17.79	53.94 ± 16.02	51.8 ± 14.89	50.52 ± 14.48	50.03 ± 14.51	49.54 ± 14.12	< 0.0001
LDL-C, mg/dL	112.21 ± 35.32	117.59 ± 37.46	117.95 ± 38.22	117.2 ± 38.90	116.67 ± 39.66	116.33 ± 39.70	< 0.0001
eGFR, mL/minute/1.73 m2	74.08 ± 26.44	67.72 ± 23.19	65.91 ± 22.22	65.51 ± 22.03	65.77 ± 22.74	68.84 ± 25.00	< 0.0001
Triglyceride, mg/dL	92.04 (91.9-92.19)	119.72 (119.51-119.93)	133.83 (133.6-134.06)	144.27 (143.98-144.56)	151.22 (150.81-151.62)	158.00 (157.51-158.48)	< 0.0001

¹Alcohol consumptions ≥ 30 g/day.

²Regular exercise: Mid-term exercise ≥ 5 days or vigorous exercise ≥ 3 days in a week.

³Low income 25%.

Data were showed as mean ± SD, n (%) or median (range). PA: Physical activity; eGFR: Estimated glomerular filtration rate; BMI: Body mass index; WC: Waist circumference; FBS: Fasting blood glucose; TC: Total cholesterol; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol.

Effects of DM on incident CV outcomes

Table 4 summarizes the multivariate Cox analysis results for incident CV outcomes based on glycemic status. Compared with participants with normal glucose, those with IFG exhibited modestly increased risks of CVD, MI, stroke, and all-cause mortality after multivariable adjustment (model 3: HRs = 1.04-1.05 for CVD, MI, and stroke; HRs = 1.03 for mortality). In contrast, participants with diabetes demonstrated significantly higher risks, yielding HRs = 1.57, 1.60, 1.58, and HRs = 1.60 for CVD, MI, stroke, and mortality in fully adjusted models, respectively. Overall, diabetes was strongly associated with an increased incidence of adverse CV outcomes and mortality, while IFG conferred a smaller yet significant excess risk.

Table 4 Multivariate cox analysis for incident cardiovascular outcome by diabetes mellitus status.

Outcome	DM status	n	Events (n)	Duration, PY	IR, per 1000	Adjusted HR (95%CI)
						Model 1	Model 2	Model 3
CVD	Normal	857846	47590	6508730.39	7.31	1 (reference)	1 (reference)	1 (reference)
	IFG	457410	32092	3393918.30	9.46	1.30 (1.28-1.32)	1.07 (1.05-1.08)	1.04 (1.03-1.06)
	DM	399603	48151	2824691.28	17.05	2.35 (2.32-2.38)	1.69 (1.67-1.72)	1.57 (1.55-1.59)
MI	Normal	857846	22657	6586843.49	3.44	1 (reference)	1 (reference)	1 (reference)
	IFG	457410	15083	3446873.19	4.38	1.28 (1.25-1.30)	1.06 (1.04-1.09)	1.05 (1.02-1.07)
	DM	399603	23193	2901757.43	7.99	2.35 (2.30-2.39)	1.73 (1.70-1.76)	1.60 (1.57-1.63)
Stroke	Normal	857846	27725	6568204.74	4.22	1 (reference)	1 (reference)	1 (reference)
	IFG	457410	19065	3433118.71	5.55	1.32 (1.30-1.34)	1.07 (1.05-1.09)	1.05 (1.03-1.07)
	DM	399603	28972	2881142.97	10.06	2.40 (2.36-2.44)	1.70 (1.67-1.73)	1.58 (1.56-1.61)
Death	Normal	857846	66149	6652038.98	9.94	1 (reference)	1 (reference)	1 (reference)
	IFG	457410	45975	3490148.71	13.17	1.33 (1.32-1.35)	1.03 (1.02-1.04)	1.03 (1.02-1.05)
	DM	399603	70520	2965646.58	23.78	2.41 (2.39-2.44)	1.61 (1.60-1.63)	1.60 (1.58-1.62)

Model 1: Non-adjusted; Model 2: Adjusted for age, sex; Model 3: Adjusted for model 2 plus smoking, alcohol drinking, regular exercise, income, hypertension, dyslipidemia. 95%CI: 95% confidence interval; CVD: Cardiovascular disease; MI: Myocardial infarction; DM: Diabetes mellitus; IFG: Impaired fasting glucose; PY: Person years; IR: Incidence rate; HR: Hazard ratio.

Joint impact of glycemic status and BMI on CV outcomes

Table 5 and Figure 2 illustrate the synergistic associations of glycemic status and BMI categories with incident CV outcomes. Across all BMI groups, participants with diabetes exhibited higher risks of CVD, MI, stroke, and all-cause mortality than their normoglycemic counterparts within the corresponding BMI category. Both underweight individuals (< 18.5 kg/m²) and those classified as obese (≥ 30 kg/m²) displayed increased risks, with these associations being most pronounced among those with diabetes. Among participants without diabetes, those who were overweight, along with individuals with mild obesity, yielded the lowest risk of all-cause mortality; however, the presence of diabetes was consistently associated with an elevated risk, irrespective of BMI. These findings underscore the additive impact of diabetes and abnormal BMI on adverse CV outcomes.

Figure 2 Subgroup analysis by diabetes mellitus status and body mass index. A: Cardiovascular disease; B: Myocardial infarction; C: Stroke; D: Death. DM: Diabetes mellitus; BMI: Body mass index; IFG: Impaired fasting glucose; MI: Myocardial infarction; CVD: Cardiovascular disease; IR: Incidence rate; HR: Hazard ratio; 95%CI: 95% confidence interval.

Table 5 Multivariate cox analysis for incident cardiovascular outcome by level of body mass index.

Outcome	BMI status (kg/m2)	n	Events (n)	Duration, PY	IR, per 1000	Adjusted HR (95%CI)
						Model 1	Model 2	Model 3
CVD	< 18.5	56566	3383	397648.60	8.51	0.91 (0.88-0.94)	1.10 (1.07-1.14)	1.12 (1.08-1.16)
	< 23	535450	37165	3947307.38	9.42	1 (reference)	1 (reference)	1 (reference)
	< 25	410244	31995	3058693.76	10.46	1.11 (1.09-1.13)	0.99 (0.97-1.00)	0.97 (0.96-0.98)
	< 30	602426	47427	4504619.24	10.53	1.12 (1.10-1.13)	1.03 (1.02-1.04)	0.98 (0.96-0.99)
	≥ 30	110173	7863	819070.98	9.60	1.02 (1.00-1.05)	1.25 (1.22-1.28)	1.11 (1.08-1.14)
MI	< 18.5	56566	1682	401800.32	4.19	0.96 (0.92-1.01)	1.18 (1.12-1.24)	1.20 (1.14-1.26)
	< 23	535450	17569	4005904.07	4.39	1 (reference)	1 (reference)	1 (reference)
	< 25	410244	14925	3112148.57	4.80	1.09 (1.07-1.12)	0.97 (0.95-0.99)	0.95 (0.93-0.97)
	< 30	602426	22864	4583820.63	4.99	1.13 (1.11-1.16)	1.04 (1.02-1.06)	0.98 (0.96-1.00)
	≥ 30	110173	3893	831800.52	4.68	1.07 (1.03-1.11)	1.27 (1.23-1.32)	1.12 (1.08-1.16)
Stroke	< 18.5	56566	1928	401035.13	4.81	0.87 (0.83-0.91)	1.04 (0.99-1.09)	1.05 (1.00-1.10)
	< 23	535450	22294	3988702.59	5.59	1 (reference)	1 (reference)	1 (reference)
	< 25	410244	19280	3097139.88	6.23	1.11 (1.09-1.13)	1.00 (0.98-1.01)	0.98 (0.96-1.00)
	< 30	602426	27770	4565772.03	6.08	1.09 (1.07-1.11)	1.01 (0.99-1.03)	0.96 (0.94-0.98)
	≥ 30	110173	4490	829816.79	5.41	0.97 (0.94-1.00)	1.22 (1.18-1.26)	1.09 (1.05-1.12)
Death	< 18.5	56566	10755	405566.28	26.52	1.65 (1.62-1.68)	1.86 (1.82-1.90)	1.83 (1.79-1.87)
	< 23	535450	66001	4052235.98	16.29	1 (reference)	1 (reference)	1 (reference)
	< 25	410244	42492	3154949.89	13.47	0.82 (0.81-0.83)	0.75 (0.74-0.76)	0.76 (0.75-0.77)
	< 30	602426	54969	4651464.79	11.82	0.72 (0.71-0.73)	0.71 (0.71-0.72)	0.70 (0.70-0.71)
	≥ 30	110173	8427	843617.33	9.99	0.62 (0.60-0.63)	0.92 (0.90-0.94)	0.87 (0.85-0.89)

Model 1: Non-adjusted; Model 2: Adjusted for age, sex; Model 3: Adjusted for model 2 plus smoking, alcohol drinking, regular exercise, income, hypertension, dyslipidemia. BMI: Body mass index; 95%CI: 95% confidence interval; CVD: Cardiovascular disease; MI: Myocardial infarction; PY: Person years; IR: Incidence rate; HR: Hazard ratio.

Collaborative impact of glycemic status and BMI on CV outcomes

Table 6 and Figure 3 depict the synergistic influence of WC categories and glycemic status on incident CV events. Higher WC values were linked to progressively increasing risks of CVD, MI, stroke, and all-cause mortality, with the strongest associations observed among individuals with diabetes. Within each WC category, those with diabetes exhibited higher risks than those without, with the combination of high WC values and diabetes conferring the greatest hazard. Notably, even within “lower WC value” groups, the presence of diabetes was associated with significantly elevated event rates and mortality, emphasizing the combined adverse effects of central obesity and diabetes on CV outcomes.

Figure 3 Subgroup analysis by diabetes mellitus status and waist circumference. A: Cardiovascular disease; B: Myocardial infarction; C: Stroke; D: Death. DM: Diabetes mellitus; WC: Waist circumference; IFG: Impaired fasting glucose; MI: Myocardial infarction; CVD: Cardiovascular disease; IR: Incidence rate; HR: Hazard ratio; 95%CI: 95% confidence interval.

Table 6 Multivariate cox analysis for incident cardiovascular outcome by level of waist circumference.

Outcome	WC status (cm)	n	Events (n)	Duration, PY	IR, per 1000	Adjusted HR (95%CI)
						Model 1	Model 2	Model 3
CVD	< 80/< 75	445617	23223	3336354.36	6.96	0.64 (0.63-0.65)	0.91 (0.90-0.93)	0.96 (0.95-0.98)
	< 85/< 80	358075	25492	2672239.66	9.54	0.87 (0.86-0.89)	0.95 (0.94-0.97)	0.97 (0.96-0.99)
	< 90/< 85	378139	30709	2808044.85	10.94	1 (reference)	1 (reference)	1 (reference)
	< 95/< 90	273818	23991	2020208.22	11.88	1.09 (1.07-1.11)	1.04 (1.02-1.06)	1.02 (1.00-1.04)
	< 100/< 95	149269	13995	1093690.46	12.80	1.17 (1.15-1.20)	1.12 (1.10-1.14)	1.07 (1.05-1.09)
	≥ 100/≥ 95	109941	10423	796802.43	13.08	1.20 (1.18-1.23)	1.26 (1.23-1.29)	1.17 (1.15-1.20)
MI	< 80/< 75	445617	11098	3372362.79	3.29	0.64 (0.63-0.66)	0.90 (0.88-0.92)	0.95 (0.93-0.98)
	< 85/< 80	358075	12119	2713709.57	4.47	0.87 (0.85-0.89)	0.95 (0.92-0.97)	0.97 (0.94-0.99)
	< 90/< 85	378139	14642	2858902.81	5.12	1 (reference)	1 (reference)	1 (reference)
	< 95/< 90	273818	11379	2060078.54	5.52	1.08 (1.05-1.11)	1.04 (1.01-1.07)	1.02 (0.99-1.04)
	< 100/< 95	149269	6665	1116627.75	5.97	1.17 (1.14-1.20)	1.12 (1.09-1.16)	1.07 (1.04-1.11)
	≥ 100/≥ 95	109941	5030	813792.65	6.18	1.21 (1.18-1.25)	1.28 (1.23-1.32)	1.18 (1.15-1.22)
Stroke	< 80/< 75	445617	13666	3362891.68	4.06	0.64 (0.62-0.65)	0.93 (0.91-0.95)	0.98 (0.95-1.00)
	< 85/< 80	358075	15068	2703325.42	5.57	0.87 (0.85-0.89)	0.96 (0.94-0.98)	0.98 (0.96-1.00)
	< 90/< 85	378139	18173	2846368.13	6.39	1 (reference)	1 (reference)	1 (reference)
	< 95/< 90	273818	14340	2049514.23	7.00	1.10 (1.07-1.12)	1.05 (1.03-1.07)	1.03 (1.00-1.05)
	< 100/< 95	149269	8355	1110933.10	7.52	1.18 (1.15-1.21)	1.12 (1.09-1.15)	1.07 (1.05-1.10)
	≥ 100/≥ 95	109941	6160	809433.86	7.61	1.20 (1.16-1.23)	1.25 (1.22-1.29)	1.17 (1.13-1.20)
Death	< 80/< 75	445617	41741	3401800.39	12.27	0.88 (0.87-0.90)	1.35 (1.33-1.37)	1.36 (1.34-1.38)
	< 85/< 80	358075	36435	2748088.03	13.26	0.95 (0.94-0.97)	1.06 (1.05-1.08)	1.07 (1.06-1.09)
	< 90/< 85	378139	40422	2901383.41	13.93	1 (reference)	1 (reference)	1 (reference)
	< 95/< 90	273818	31625	2092744.55	15.11	1.09 (1.07-1.10)	1.02 (1.01-1.04)	1.01 (1.00-1.03)
	< 100/< 95	149269	18297	1135942.79	16.11	1.16 (1.14-1.18)	1.08 (1.06-1.10)	1.05 (1.04-1.07)
	≥ 100/≥ 95	109941	14124	827875.1	17.06	1.24 (1.21-1.26)	1.30 (1.27-1.32)	1.24 (1.22-1.27)

Model 1: Non-adjusted; Model 2: Adjusted for age, sex; Model 3: Adjusted for model 2 plus smoking, alcohol drinking, regular exercise, income, hypertension, dyslipidemia. WC: Waist circumference; 95%CI: 95% confidence interval; CVD: Cardiovascular disease; MI: Myocardial infarction; PY: Person years; IR: Incidence rate; HR: Hazard ratio.

DISCUSSION

In this extensive nationwide CKD cohort, both glycemic status and measures of adiposity indices were found to jointly influence CV risk. The principal findings can be summarized as follows: Diabetes conferred the most prominent incremental risk of CVD and mortality across all BMI and WC categories, underscoring its central role in CV pathology in patients with CKD. Both abnormal BMI and WC were independently associated with adverse outcomes; nonetheless, their impact varied with glycemic status. In participants with normal glucose levels or IFG, underweight status and a high WC value were risk-enhancing factors, whereas in those with diabetes, paradoxically, low BMI and WC values were associated with the highest hazards. These results highlight the heterogeneity of CVD risk patterns in CKD and demonstrate the necessity of interpreting adiposity measures within the context of glycemic status.

Our findings reinforce and extend previous observations regarding the “obesity paradox” in CKD, where overweight individuals and those with mild obesity—particularly those without diabetes—occasionally exhibit lower mortality rates than their leaner counterparts[12-14]. Consistent with this phenomenon, our cohort demonstrated that overweight individuals and those with class I obesity, in the absence of diabetes, experienced the lowest mortality. Notwithstanding, the protective effect of a higher BMI diminished in the presence of diabetes, as CV risk remained elevated across all BMI and WC groups. These data support the notion that diabetes negates the benefits associated with increased adiposity, possibly through mechanisms involving insulin resistance, inflammation, and accelerated atherosclerosis[15,16].

The varying predictive value of BMI and WC was also evident in our cohort, underscoring the complementary roles of total and central adiposity. While BMI reflects overall adiposity, WC more accurately estimates visceral fat, which is metabolically active and strongly linked to CVD[17,18]. In CKD patients without diabetes, a high WC value proved to be a significant risk marker for CVD events. Conversely, among participants with diabetes, low BMI and WC values were both associated with an increased risk, likely reflecting malnutrition, sarcopenia, and unintentional weight loss—conditions prevalent in advanced CKD and diabetes[19-21]. These findings suggest that variations in body composition, including fat and muscle distribution, rather than simplistic measures of body weight or central obesity, are critical prognostic factors in diabetic CKD.

Clinically, our results underscore the necessity for individualized CV risk stratification in CKD, incorporating both glycemic status and adiposity patterns. Underweight or centrally lean CKD patients with diabetes should not be regarded as low risk, as they may represent a particularly frail and vulnerable subgroup. In contrast, targeting visceral obesity remains a key preventive strategy among CKD patients without diabetes owing to its strong association with adverse CV outcomes. These insights advocate for the tailoring of CVD prevention and management strategies according to both metabolic and body composition profiles.

This study, based on a nationwide CKD cohort of over two million individuals with comprehensive anthropometric and CV outcome data, provides robust evidence that both glycemic status and adiposity patterns jointly influence CV risk. The findings underscore the significance of personalized risk stratification, reflecting that underweight or centrally lean CKD patients with diabetes are particularly vulnerable and that visceral obesity remains a crucial focus in individuals without diabetes. Nevertheless, several limitations must be acknowledged. Anthropometric measures were assessed solely at baseline, detailed body composition data were unavailable, and residual confounding from unmeasured factors, such as diet, physical activity, and muscle mass, cannot be excluded. Furthermore, sex-related differences and longitudinal alterations in body composition over time were not accounted for, providing room for further refinement of risk prediction in future research. Finally, causal inference is constrained by the study’s observational design. Despite these limitations, the results offer clinically relevant insights for tailoring preventive strategies based on both metabolic status and body composition in CKD.

CONCLUSION

In conclusion, the impact of adiposity on CVD risk was significantly modified by glycemic status. Notably, underweight individuals with diabetes exhibited the greatest vulnerability, indicating that leanness is not universally protective in CKD. These findings emphasize the imperativeness of incorporating both glycemic status and obesity phenotypes into individualized CVD risk assessments and prevention strategies for patients with comorbid diabetes and CKD.