Correlation between albumin–bilirubin score and comorbidity of cardiovascular disease and chronic kidney disease: insights from NHANES 2011–2018

Abstract

Background

Growing evidence indicates that cardiovascular disease (CVD), chronic kidney disease (CKD), and liver dysfunction share common pathophysiological pathways and constitute an interconnected syndrome. The albumin–bilirubin (ALBI) score, originally developed to assess liver function, has shown prognostic value in various nonhepatic diseases. However, its association with the comorbidity of CVD and CKD remains unclear. This study aimed to investigate the association between ALBI score and CVD–CKD comorbidity and to explore potential intermediate factors in a nationally representative US adult population.

Methods

A cross-sectional analysis of 15,681 adults from the National Health and Nutrition Examination Survey (NHANES) 2011–2018 was conducted. The ALBI score was calculated as: (log10 bilirubin [μmol/L] × 0.66) + (albumin [g/L] × − 0.085). Weighted multivariable logistic regression, restricted cubic spline, threshold effect, subgroup, and mediation analyses were performed.

Results

After fully adjusting for confounders (age, sex, race, education level, poverty-income ratio, body mass index, hypertension, smoking, alcohol use, and diabetes), each one-unit increase in the ALBI score was associated with more than a 4.38-fold higher prevalence of cardiovascular disease–chronic kidney disease (CVD–CKD) comorbidity (OR = 5.38, 95% CI 3.84–7.54; P < 0.001). A nonlinear relationship was observed (nonlinearity test P < 0.001), with an inflection point at approximately − 3.436. Exploratory mediation analyses indicated that diabetes and the body roundness index (BRI) statistically accounted for about 5%–6% and 0.5% of the association, respectively. However, the proportion explained by reverse mediation (diabetes → ALBI → CVD–CKD) was larger, reaching up to 6.3%.

Conclusion

Higher ALBI score is independently associated with increased odds of CVD–CKD comorbidity. The association appears to be largely driven by shared metabolic disturbances rather than liver dysfunction being upstream of diabetes or obesity. ALBI may serve as a simple, objective marker for early risk stratification in the cardiovascular–kidney–metabolic axis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-026-03889-w.

Introduction

Cardiovascular disease (CVD) and chronic kidney disease (CKD) are major global public health concerns [1]. A growing body of research has shown that CVD and CKD are closely linked through shared biological and social risk factors, referred to as cardiorenal syndrome [2, 3]. CKD has been identified as a separate predictor of myocardial infarction, stroke, and mortality in men under 55 years of age and women under 65 years of age [4]. In addition, CVD remains the primary cause of mortality among patients with CKD. Therefore, it is important to look for potential risk indicators for CVD–CKD comorbidity.

The albumin–bilirubin (ALBI) score, a new index for evaluating liver function, has been shown to correlate with prognosis in patients with cirrhosis and hepatocellular carcinoma [5, 6]. In addition, the ALBI score has been utilized in nonhepatic diseases. It has been found to be strongly associated not only with hospital death in patients with chronic heart failure [7] but also with mortality following heart valve surgery [8]. More notably, the ALBI score is a significant risk factor for acute kidney injury and long-term mortality [9]. However, the correlation between ALBI score and comorbid CVD and CKD remains uncertain.

In the past, obesity, diabetes, CVD, CKD, and metabolic dysfunction-associated steatotic liver disease (MASLD) were considered distinct conditions. However, growing evidence now suggests that they interact with each other by sharing pathophysiologic mechanisms, finally leading to multiorgan dysfunction and an elevated risk of adverse cardiovascular outcomes [10, 11]. This linked syndrome has been named Cardiovascular–Renal–Hepatic–Metabolic (CRHM) syndrome [10]. The liver, as a central metabolic organ, plays a crucial role in this network, yet little research has been conducted on the relationship between liver function and CVD–CKD comorbidity.

Therefore, we conducted a cross-sectional study using nationally representative NHANES data (2011–2018) to: (1) investigate the independent association between ALBI score and CVD–CKD comorbidity, (2) explore nonlinear patterns and threshold effects, and (3) perform exploratory mediation and reverse-mediation analyses to better understand possible inter-relationships among liver function, metabolic factors, and cardio-renal outcomes.

Methods

Data source

Publicly accessible data from the NHANES conducted in the United States were employed in this cross-sectional study. The findings derived from this dataset are of substantial significance for estimating the prevalence of various diseases and identifying associated risk factors. Moreover, the results provide critical insights into the assessment of nutritional status and its contribution to health promotion and disease prevention. Written informed consent was obtained from all NHANES participants.

Study participants

This study utilized data from 4 NHANES cycles from 2011 to 2018 involving 39,156 participants. The subsequent exclusion criteria were implemented: (1) individuals with incomplete information on cardiovascular disease or chronic kidney disease; (2) individuals with missing data on ALBI scores; and (3) individuals lacking data on key covariates. The analysis ultimately included 15,681 individuals (Fig. 1).

Fig. 1.

Flowchart of participant selection

Definition of ALBI

We obtained bilirubin and albumin data from NHANES “Laboratory Data” and calculated ALBI scores utilizing the following formula: ALBI = (log10 bilirubin μmol/L × 0.66) + (albumin g/L × − 0.085) [12, 13]. ALBI data could be analyzed as a continuous or categorical variable. ALBI scores were divided into quartiles (Q1: ≤ − 3.16; Q2:− 3.16 to − 2.96; Q3: − 2.96 to − 2.78; Q4: >  − 2.78) for categorical analysis.

Diagnosis of CVD and CKD

A standardized medical condition questionnaire was employed to diagnose cardiovascular disease. Participants were asked, “Have you ever been informed by a physician or other health professional that you have CHD/CHF/MI/angina pectoris/stroke?” and those who answered positively were considered to have CVD [14]. Chronic kidney disease was characterized by a structural or functional abnormality of the kidneys for more than 3 months, diagnosed based on proteinuria (UACR ≥ 30 mg/g) or an eGFR of less than 60 mL/min/1.73 m² [15]. Estimated glomerular filtration rate (eGFR) was derived based on serum creatinine measurements, applying the equation developed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) [16].

Covariates

The study incorporated a range of demographic and clinical factors that may influence the incidence of CVD and CKD. Demographic data were collected through standardized questionnaires, including family income to poverty, race, sex, age, and level of education. Participants who answered “no” to the following question were nondrinkers: at least 12 alcoholic beverages in their lives or during any given year. Smokers were characterized as individuals who had consumed a minimum of 100 cigarettes throughout their lifetime. Diabetes mellitus and hypertension were identified based on self-reported questionnaire responses. We obtained physical measures directly from the NHANES database, including data on height, weight, waist circumference (WC), and BMI (kg/m²). We calculated body roundness index (BRI) [17], conicity index (C index) [18], and relative fat mass (RFM) [19] using the following formula:

$$BRI=364.2-365.5\times \sqrt{1-{\left(\frac{\frac{\mathit{WCm}}{2\mathrm{\Pi }}}{0.5\times heightm}\right)}^2}$$

$$C-\text{index}=0.{109}^{-1}\text{WCm}\frac{{\text{weightkg}}^{-\frac{1}{2}}}{\text{heightm}}$$

$$\text{RFMmen}=64-20\times \frac{\text{heightm}}{\text{waistm}}$$

$$\text{RFMwomen}=76-20\times \frac{\text{heightm}}{\text{waistm}}$$

Statistical analysis

All statistical analyses were performed using R software (version 4.3.3). Sample weights were used for all analyses to ensure the national representativeness of the estimated data. For continuous variables, t tests were used to determine the mean ± standard errors and p values. Percentages were used to represent categorical variables, and chi-square tests were used to determine p values. Three logistic regression models were constructed to adjust for different confounders. Model 1 included no covariate adjustments. Model 2 was partially adjusted for age, sex, race, education level, PIR, and BMI. Model 3 further adjusted for hypertension, alcohol consumption, smoking status, and diabetes. We used ROC curves and the area under the curve to compare the effects of the composite ALBI index with those of albumin and bilirubin individually on CVD–CKD comorbidity. Restricted cubic splines (RCS) were employed to investigate the presence of a nonlinear relationship between ALBI score and CVD–CKD comorbidity. Segmented regression analysis was employed to investigate threshold effects and inflection points, and log-likelihood ratios were applied to determine the statistical significance of these inflection points. Exploratory mediation analyses were performed using the R package ‘mediation’, and bootstrap methods were used to calculate indirect effects, direct effects, total effects, and P values. Given the cross-sectional design and uncertain temporality, reverse mediation models (e.g., diabetes → ALBI → CVD–CKD) were also conducted.

Results

Participant characteristics

The cross-sectional study had 15,681 participants, of whom 704 were diagnosed with CVD and CKD. The sample comprised 7861 males and 7820 females, with a mean age of 47.45 ± 16.79 years. On average, patients with CVD–CKD comorbidity were older (mean age 69.06 vs. 46.76 years), had lower levels of education (22.7% vs. 12.1% below high school), and exhibited lower family poverty income ratios (mean ratio 2.45 vs. 3.05). They also showed a higher prevalence of diabetes (44% vs. 11%) and hypertension (78% vs. 29%). In addition, significant differences were found for BMI, BRI, C Index, RFM, and ALBI scores (all p < 0.001). No significant differences were found for smoking and drinking (p > 0.05) (Table 1).

Table 1.

Baseline characteristics of participants

Characteristic	N	Overall N = 179,071,885	No cardiovascular disease-chronic kidney disease N = 173,585,486	Cardiovascular disease-chronic kidney disease N = 5,486,400	p value
Sexb, n (%)
Female	15,681	7861 (51%)	7563 (51%)	298 (49%)	0.503
Male		7820 (49%)	7414 (49%)	406 (51%)
Age(year)a, Mean ± SD	15,681	47.45 ± (16.79)	46.76 ± (16.49)	69.06 ± (11.02)	< 0.001
Race/ethnicityb, n (%)
Mexican	15,681	2082 (8.1%)	2024 (8.3%)	58 (4.1%)	< 0.001
Other hispanic		1590 (5.8%)	1534 (5.9%)	56 (3.8%)
Non-hispanic white		6248 (68%)	5880 (67%)	368 (73%)
Non-hispanic black		3359 (10%)	3187 (10%)	172 (13%)
Other race		2402 (8.2%)	2352 (8.3%)	50 (6.4%)
Educationb, n (%)
Less Than 9th	15,681	1215 (3.9%)	1132 (3.8%)	83 (7.7%)	< 0.001
9-11th		1842 (8.5%)	1716 (8.3%)	126 (15%)
High school		3533 (23%)	3337 (22%)	196 (28%)
Some college		5020 (33%)	4826 (33%)	194 (28%)
College graduate		4071 (32%)	3966 (33%)	105 (22%)
Family income to poverty ratioa, Mean ±	15,681	3.03 ± (1.65)	3.05 ± (1.65)	2.45 ± (1.56)	< 0.001
BMIa, Mean ± SD	15,681	29.27 ± (6.85)	29.21 ± (6.82)	31.36 ± (7.66)	< 0.001
BRIa, Mean ± SD	15,681	5.50 ± (2.37)	5.45 ± (2.35)	6.96 ± (2.56)	< 0.001
C Indexa, Mean ± SD	15,681	1.31 ± (0.09)	1.31 ± (0.09)	1.39 ± (0.08)	< 0.001
RFMa, Mean ± SD	15,681	35.51 ± (8.61)	35.41 ± (8.60)	38.70 ± (8.20)	< 0.001
Hypertensionb, n (%)
No	15,681	10,193 (69%)	10,045 (71%)	148 (22%)	< 0.001
Yes		5488 (31%)	4932 (29%)	556 (78%)
Smoking statusb, n (%)
No	15,681	12,596 (81%)	12,032 (81%)	564 (80%)	0.528
Yes		3085 (19%)	2945 (19%)	140 (20%)
Diabetes b, n (%)
No	15,681	11,546 (77%)	11,242 (78%)	304 (45%)	< 0.001
Yes		2402 (12%)	2067 (11%)	335 (44%)
Borderline		1733 (12%)	1668 (12%)	65 (11%)
Drinking statusb, n (%)
No	15,681	2156 (10%)	2070 (10.0%)	86 (11%)	0.447
Yes		13,525 (90%)	12,907 (90%)	618 (89%)
Albumina, Mean ± SD	15,681	42.64 ± 3.47	42.70 ± 3.46	40.66 ± 3.20	< 0.001
Bilirubina, Mean ± SD	15,681	10.23 ± 5.22	10.23 ± 5.24	10.43 ± 4.74	0.132
Proteinuriaa, Mean ± SD	15,681	32.26 ± 232.36	26.96 ± 202.16	203.46 ± 669.64	< 0.001
HbA1ca, Mean ± SD	15,681	5.64 ± 0.94	5.61 ± 0.91	6.45 ± 1.47	< 0.001
ASTa, Mean ± SD	15,681	24.91 ± 17.79	24.98 ± 17.77	22.63 ± 18.41	< 0.001
ALTa, Mean ± SD	15,681	24.80 ± 15.44	24.77 ± 15.36	26.01 ± 17.61	0.229
LDLa, Mean ± SD	15,681	112.98 ± 35.46	113.64 ± 35.24	93.65 ± 36.52	< 0.001
eGFRa, Mean ± SD	15,681	96.25 ± 20.75	97.38 ± 19.65	60.39 ± 22.70	< 0.001
ALBIa, Mean ± SD	15,681	− 2.99 ± (0.29)	− 3.00 ± (0.28)	− 2.81 ± (0.28)	< 0.001
ALBI groupb, n (%)
Q1[− 4.42,− 3.16]	15,681	4069 (29%)	3974 (30%)	95 (12%)	< 0.001
Q2(− 3.16,− 2.96]		3831 (25%)	3728 (25%)	103 (14%)
Q3(− 2.96,− 2.78]		3995 (25%)	3800 (25%)	195 (30%)
Q4(− 2.78,− 0.745]		3786 (21%)	3475 (20%)	311 (44%)

Association of ALBI with CVD–CKD

Table 2 summarizes the outcomes of the weighted multivariate logistic regression analyses evaluating the association between the ALBI score and comorbid CVD and CKD. Three models with progressive adjustments for covariates consistently demonstrated a positive association between ALBI and CVD–CKD comorbidity (all p < 0.001). In the fully adjusted model (Model 3), each one-unit increase in ALBI was associated with a 4.38-fold higher prevalence of CVD–CKD comorbidity (OR = 5.38, 95% CI 3.84–7.54). When ALBI scores were categorized into quartiles, participants in the highest quartile (Q4) had a 3.31-fold higher prevalence of CVD–CKD comorbidity compared with those in the lowest quartile (Q1) (OR = 3.31, 95% CI 2.47–4.42). The prevalence in the Q3 group was 1.10-fold higher (OR = 2.10, 95% CI 1.49–2.95). No statistically significant association was observed for participants in the Q2 group. Restricted cubic spline analysis demonstrated a nonlinear relationship between ALBI and CVD–CKD comorbidity in both unadjusted and fully adjusted models (p for nonlinear < 0.001) (Fig. 2). Segmented regression analysis further supported the existence of a threshold effect, identifying an inflection point at an ALBI score of − 3.436 (log-likelihood ratio test, p = 0.002) (Table 3). Threshold effect analysis showed that when the ALBI score was below − 3.436, ALBI was negatively associated with CVD–CKD comorbidity [OR = 0.03, 95% CI 0.01–0.42; p = 0.003]. However, this trend needs to be interpreted with caution due to the wide confidence intervals. Conversely, when ALBI exceeded − 3.436, a notable positive correlation was identified [OR = 3.08, 95% CI 2.30–4.12; p < 0.001]. The ROC curve indicated that ALBI had the highest diagnostic ability for CVD–CKD, with an AUC of 0.864 (Fig. 3).

Table 2.

Logistic regression analysis of the association between ALBI score with CVD–CKD comorbidity

Exposure variable	Model 1		Model 2		Model 3
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
ALBI	8.08 (6.24 ~ 10.45)	< .001***	3.74 (2.63 ~ 5.30)	< .001***	5.38 (3.84 ~ 7.54)	< .001***
ALBI group
Q1	1.00 (Reference)	–	1.00 (Reference)	–	1.00 (Reference)	–
Q2	1.33 (0.90 ~ 1.96)	0.153	0.85 (0.57 ~ 1.26)	0.407	1.04 (0.71 ~ 1.51)	0.841
Q3	2.91 (2.07 ~ 4.10)	< .001***	1.57 (1.09 ~ 2.26)	0.016*	2.10 (1.49 ~ 2.95)	< .001***
Q4	5.29 (3.88 ~ 7.22)	< .001***	2.26 (1.64 ~ 3.10)	< .001***	3.31 (2.47 ~ 4.42)	< .001***
P for trend	< .001***	–	< .001***	–	< .001***	–

Model 1: Crude

Model 2: Adjust: Age, Sex, Race/ethnicity, Education, Family income to poverty ratio, BMI

Model 3: Adjust: Age, Sex, Race/ethnicity, Education, Family income to poverty ratio, BMI, Hypertension, Drinking status, Smoking status

Significance: * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 2.

Restrictive cubic spline analysis for the association between ALBI and CVD–CKD comorbidity. A Model 1; B Model 3

Table 3.

Utilize the two-segment piecewise linear regression model for the analysis of threshold effects between ALBI score and CVD–CKD comorbidity

	OR (95%CI)	p value
One—line linear regression model	2.73 (2.05 ~ 3.63)	< .001
Two—piecewise linear regression model
Inflection point	− 3.436
< − 3.436	0.03 (0.01 ~ 0.42)	0.003
> − 3.436	3.08 (2.30 ~ 4.12)	< .001
Log—likelihood ratio test		0.002

Age, Sex, Race/ethnicity, Education, Family income to poverty ratio, BMI, Hypertension, Drinking status, Smoking status

Fig. 3.

ROC curves and area under the curves compare the identification performance of ALBI, albumin, and bilirubin for CVD–CKD comorbidities

Subgroup analysis

We conducted subgroup analyses to better assess the reliability and consistency of the association between ALBI and CVD–CKD comorbidity. Age, gender, race, education, diabetes, hypertension, smoking, alcohol consumption, BMI, BRI, C Index, and RFM were included in the analysis as shown in Fig. 4. Correlations between ALBI scores and CKD–CVD comorbidity in various subgroups were generally consistent with our preliminary analyses, but statistical significance was not observed in a few subgroups. Notably, in subgroups of people aged > 49, higher ALBI scores were significantly associated with a higher prevalence of CVD–CKD comorbidity (p < 0.001). Moreover, no notable interactions were detected among any of the subgroups (p for interaction > 0.05), suggesting that the association between ALBI and CVD–CKD comorbidity is largely independent of these stratifying variables. This supports the role of ALBI as a stable and independent risk factor for CVD–CKD comorbidity.

Fig. 4.

Subgroup analysis of the association between ALBI score and CVD–CKD comorbidity. Continuous variables are demarcated by medians

Mediation analysis

We further explored the mediating role of diabetes, BRI, C Index, and RFM in the association between ALBI scores and CVD–CKD comorbidity (Fig. 5 and Table 4). After adjusting for all covariates, BRI was found to partially mediate this association, with a mediation proportion of 0.5%. The direct effect of ALBI remained significant after adjustment for BRI (observed coefficient = 0.234, p < 0.001). In addition, diabetes also demonstrated a statistically significant mediating effect. The indirect effect size of ALBI on CVD–CKD comorbidity via diabetes was 0.007 (95% CI 0.003–0.010; p < 0.001). Even after adjusting for diabetes, ALBI score and CVD–CKD comorbidity showed a statistically significant correlation, with a direct effect size of 0.120 (95% CI 0.070–0.170; p < 0.001). The results suggest that approximately 6% (95% CI 2.4–11%) of the effect of ALBI score on CVD–CKD comorbidity was mediated by diabetes (p < 0.001). Although the direct effect was significant in all other mediation models, no mediation was detected for the C Index and RFM.

Fig. 5.

Mediation effects of potential mediators in the associations of ALBI score with CVD–CKD comorbidity. ALBI albumin-bilirubin, BRI roundness index, C Index conicity index, RMF relative fat mass

Table 4.

Mediation analysis for the associations between ALBI score and CVD–CKD comorbidity

Paths	Observed coefficient	Bootstrap test		LLCI	ULCI
		Bootstrap standard error	P
BRI
Total effect	0.235	0.033	< .001	0.135	0.250
Indirect effect	0.001	0.001	0.020	0.001	0.010
Direct effect	0.234	0.033	< .001	0.135	0.250
C Index
Total effect	0.233	0.032	< .001	0.117	0.240
Indirect effect	0.001	0.001	0.068	− 0.001	0.010
Direct effect	0.232	0.031	< .001	0.115	0.240
RFM
Total effect	0.234	0.031	< .001	0.122	0.240
Indirect effect	0.001	0.001	0.160	− 0.001	0.002
Direct effect	0.233	0.030	< .001	0.121	0.240
Diabetes
Total effect	0.127	0.129	< .001	0.077	0.180
Indirect effect	0.007	0.009	< .001	0.003	0.010
Direct effect	0.120	0.126	< .001	0.070	0.170

LLCI lower level for confidence interval, ULCI upper level for confidence interval

Age, Sex, Race/ethnicity, Education, Family income to poverty ratio, BMI, Hypertension, Drinking status, Smoking status were adjusted

Sensitivity analysis

We further examined the associations of ALBI with the risk of CVD and with the risk of CKD separately. After adjusting for potential confounders, including sex, age, race/ethnicity, education, family income-to-poverty ratio, BMI, drinking status, smoking status, and hypertension, ALBI showed a significant positive association with CVD risk (OR = 3.24, 95% CI 2.49–4.20; p < 0.001) (Supplementary Table 1). A similarly significant positive association was observed between ALBI and CKD risk (OR = 2.10, 95% CI 1.70–2.59; p < 0.001) (Supplementary Table 2).

Next, we excluded participants outside the 1st and 99th percentiles to avoid unstable results caused by overly wide confidence intervals in the RCS analysis. After adjustment for potential confounders, the nonlinear relationship between ALBI and CVD-CKD risk remained stable (p for nonlinear = 0.009; p for overall < 0.001) (Supplementary Fig. 1). Threshold effect analysis indicated that, when restricting the sample to the 1st through 99th percentiles, the turning point for the nonlinear association between ALBI and CVD–CKD risk shifted to − 3.113. When ALBI exceeded − 3.113, a positive association was observed between ALBI and CVD–CKD risk (OR = 4.31, 95% CI 2.86–6.48; p < 0.001). When ALBI was below –3.113, no statistically significant association was detected.

Finally, to minimize the influence of reverse causation on the mediation analysis of ALBI in the association between diabetes and CVD–CKD, we found that ALBI accounted for 6.3% of the effect of diabetes on CVD–CKD risk (Fig. 6).

Fig. 6.

Mediation effects of ALBI score in the associations of diabetes with CVD-CKD comorbidity

Discussion

This study analyzed data from 15,681 participants in the 2011–2018 U.S. National Health and NHANES and confirmed a significant positive association between higher ALBI scores and the prevalence of comorbid CVD and CKD. In the fully adjusted model, each one-unit increase in ALBI was associated with more than a 4.38-fold higher risk of CVD–CKD comorbidity (OR = 5.38, 95% CI 3.84–7.54), following a nonlinear pattern with an inflection point at approximately − 3.436. Exploratory mediation analyses indicated that diabetes and the body roundness index (BRI) statistically accounted for a small portion of this association (5–6% and 0.5%, respectively), whereas reverse mediation analyses showed a larger effect, up to 6.3%, highlighting the potential for bidirectional or reverse relationships in this cross-sectional study.

The present study is, to our knowledge, the first to demonstrate a strong, independent, and nonlinear association between the ALBI score and combined CVD–CKD comorbidity in a large, nationally representative sample. When compared with traditional liver function markers, ALBI (AUC 0.864) showed superior discrimination over albumin alone (AUC 0.650). The observed inflection point around -3.436 corresponds physiologically to the transition from fully preserved hepatic synthetic function (normal albumin > 41–42 g/L) to early impairment, at which compensatory mechanisms may fail and systemic inflammation, oxidative stress, and endothelial dysfunction accelerate cardio-renal damage. Our exploratory mediation findings, especially the substantially larger effect in the reverse direction, strongly suggest that liver dysfunction reflected by higher ALBI lies downstream of longstanding metabolic disturbances (diabetes, visceral obesity) rather than being their primary cause. This is biologically more plausible and aligns with the evolving concept of Cardiovascular–Kidney–Metabolic (CKM) syndrome.

These findings align with emerging evidence on the prognostic utility of ALBI beyond its original hepatic applications. For instance, in a recent NHANES-based analysis (2005–2018), higher ALBI scores were independently associated with increased all-cause and cardiovascular mortality in individuals with metabolic-associated fatty liver disease (MAFLD), with hazard ratios (HR) ranging from 1.5 to 2.0 for the highest tertile [20]. The magnitude of association in our study appears stronger, possibly due to our focus on combined CVD–CKD comorbidity rather than isolated outcomes or mortality. Similarly, another NHANES study reported elevated ALBI as a predictor of depression severity, with odds ratios around 1.8–2.5 for higher quartiles [21], suggesting ALBI’s broader role in systemic conditions involving inflammation and metabolic dysregulation. In non-NHANES cohorts, ALBI has been linked to acute kidney injury (AKI) post-percutaneous coronary intervention [9] and ICU mortality in heart failure patients (HR ~ 1.7–2.0) [7]. Compared to these, our OR of 5.38 is notably higher, which may reflect the synergistic risk amplification in CVD–CKD comorbidity, where hepatic dysfunction exacerbates cardiorenal interactions via shared pathways like oxidative stress and endothelial injury.

The ALBI score relies on only two objective serum markers, bilirubin and albumin, to assess liver function in patients. Several studies have partially revealed the common pathophysiologic mechanisms of hepatic dysfunction in CVD and CKD. In individuals with significant hepatic dysfunction, the overproduction of vasodilatory mediators such as NO, along with hypoproteinemia and reduced osmolality, results in diminished effective arterial blood volume, potentially causing renal damage [22]. Furthermore, compensatory activation of the sympathetic nervous system (SNS)and the renin–angiotensin–aldosterone system (RAAS) is initiated [23]. Activation of the RAAS causes vasoconstriction and elevates afterload, resulting in or exacerbating left ventricular hypertrophy and myocardial fibrosis [24]. Chronic activation of SNS exacerbates renal vasoconstriction, which further reduces GFR and contributes to ischemic kidney injury [23]. In patients with hepatic failure, the liver is unable to efficiently eliminate ammonia and other metabolites, leading to their accumulation. These accumulated toxins induce endothelial dysfunction and promote renal vasoconstriction, resulting in the impairment of renal function [25]. These toxins also promote systemic inflammation and oxidative stress, elevate the incidence of cardiac arrhythmias, and accelerate atherosclerosis [26]. Chronic liver injury stimulates Kupffer cells and various immune cells, leading to the secretion of pro-inflammatory cytokines (e.g., TNF-alpha, IL-6) and oxidative stress. These factors damage cardiomyocytes and renal tubular cells directly, thereby promoting both atherosclerosis and renal dysfunction [27].

A high BRI indicates significantly abnormal body fat and visceral fat distribution, typically resulting from lipid metabolic disorders [28]. Our research indicated that diabetes and BRI partially mediated the influence of the ALBI score on the comorbidity of CVD and CKD. This indicates a possible interrelationship among hepatic dysfunction, lipid metabolism, obesity, and diabetes in the development of CVD–CKD comorbidity. Dysfunctional adipose tissue, particularly visceral adipose tissue, releases pro-inflammatory cytokines that harm cardiac and renal tissues [29–31]. Furthermore, surplus free fatty acids (FFA) released from adipose tissue are promptly oxidized, producing reactive oxygen species (ROS), resulting in myocardial and hepatic injury and dysfunction [10]. Hyperglycemia precipitates glomerular hyperfiltration and hypertension, potentially resulting in renal damage and atherosclerosis [32, 33]. Moreover, hyperglycemia initiates several intracellular pathways that promote renal and vascular damage through inflammation and fibrosis [32–34].

This study’s strengths include its large, weighted NHANES sample ensuring national representativeness, comprehensive covariate adjustment, and novel explorations of nonlinearity and bidirectional mediation. Nonetheless, limitations persist: (1) cross-sectional design precludes causality—longitudinal data could clarify temporal sequences; (2) self-reported CVD and covariates may introduce recall bias, though NHANES validation studies show high accuracy (> 80%); (3) unmeasured confounders (e.g., viral hepatitis, medication effects) were not fully addressed; (4) the nonlinear threshold’s wide CIs at extremes indicate instability, meriting replication; (5) substantial exclusions for missing data, though comparisons revealed only modest biases.

Conclusion

Higher ALBI score is robustly and independently associated with increased prevalence of CVD–CKD comorbidity. The association is nonlinear, with a clinically interpretable threshold, and appears largely attributable to shared upstream metabolic risk factors. The simplicity and objectivity of ALBI support its potential use as an additional risk marker in the cardiovascular–kidney–metabolic continuum. In the future, we hope that prospective clinical trials will be conducted to further substantiate the causal relationship between the ALBI score and CVD–CKD comorbidity.

Author contributions

All authors contributed to the study conception and design. Liang Fang and Xingjiang Li designed the study. Liang Fang, Xingjiang Li, and Jianfeng Yin collated and analyzed the data. Liang Fang, Xingjiang Li, Anna Dai, and Maijun Niu wrote the manuscript. All authors approved the final manuscript. All authors have accepted responsibility for the entire content of this manuscript and approved its submission.

Funding

The author(s) declare that no financial support was received for the research, authorship, and/or publication of this article.

Data availability

Publicly available datasets were analyzed in this study. This data can be found here: (https://wwwn.cdc.gov/nchs/nhanes/default.aspx).

Declarations

Ethics approval and consent to participate

All protocols were approved by the ethics review board of the National Center for Health Statistics (#2011-17, #2018-01), and written informed consents were obtained from the participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.