Albumin-corrected anion gap and mortality risk in cardiorenal syndrome: Insights from a nationally representative cohort study Authors: Hongman Li, Xing Huang, Shengyong Si, Zhijing Li, Chun Zhao, Xue Zhao, Mengdie Liu, Yuan Deng, Xue Zhao Published: 2026 ABSTRACT The prognostic significance of the albumin-corrected anion gap (ACAG) in patients with cardiorenal syndrome (CRS) remains unclear. This study aimed to examine the association between ACAG and all-cause and cardiovascular mortality in a nationally representative cohort. We identified 1783 patients with CRS from the National Health and Nutrition Examination Survey 2005 to 2018 cycles and categorized them into terciles based on ACAG levels. Associations between ACAG and mortality were assessed using Cox regression, Kaplan–Meier survival curves, and restricted cubic spline analyses, with subgroup analyses across demographic and clinical strata. Among 1783 CRS patients, with a median follow-up time of 72 months, 859 deaths occurred, of which 36% were cardiovascular. Higher ACAG was independently associated with increased risks of all-cause mortality (hazard ratio [HR] = 1.14, 95% confidence interval [CI]: 1.09–1.18) and cardiovascular mortality (HR = 1.14, 95% CI: 1.08–1.21). A linear dose–response relationship was observed, and associations were consistent across subgroups. Elevated ACAG is significantly associated with higher risks of all-cause and cardiovascular mortality in CRS patients. To our knowledge, this is the first study to demonstrate the prognostic value of ACAG in CRS, highlighting its potential as a simple, inexpensive, and widely available marker for risk stratification in nephrology practice. 1. Introduction Cardiorenal syndrome (CRS) describes the bidirectional dysfunction of the heart and kidneys, where impairment of one organ precipitates injury in the other through neurohormonal and hemodynamic mechanisms.[1–3] This vicious cycle contributes to systemic inflammation, oxidative stress, and metabolic disturbances, ultimately leading to poor outcomes.[4] With population aging and the increasing burden of chronic diseases, the prevalence of CRS has risen markedly,[5] and patients with CRS face substantially elevated risks of mortality and healthcare utilization.[6–8] Metabolic acidosis is common in CRS and has been associated with adverse outcomes in both kidney and cardiovascular disease. Acid–base imbalance exacerbates renal and cardiac dysfunction and contributes to electrolyte disturbances, inflammation, and progression of chronic kidney disease (CKD).[9–12] The anion gap (AG) is commonly used to assess metabolic acidosis but can be confounded by hypoalbuminemia.[13] The albumin-corrected anion gap (ACAG) accounts for serum albumin levels and provides a more accurate assessment.[14] Elevated ACAG has been linked to higher mortality in acute kidney injury,[15] sepsis, myocardial infarction, and diabetes,[16] and has also been associated with impaired kidney function in population studies.[17] However, no prior study has investigated the prognostic significance of ACAG in patients with CRS. To address this knowledge gap, we analyzed data from a large, nationally representative U.S. cohort to evaluate the association between ACAG and both all-cause and cardiovascular mortality in CRS. 2. Methods 2.1. Study population Data for this study were derived from the National Health and Nutrition Examination Survey (NHANES) (https://www.cdc.gov/nchs/nhanes/), conducted by the National Center for Health Statistics (NCHS) in the United States. NHANES collects nationally representative health and nutrition data to evaluate health status, nutritional factors, and risk behaviors, thereby informing public health policy, guiding disease prevention strategies, and improving healthcare services. All study protocols were approved by the NCHS Research Ethics Review Board. The present analysis included data from 2005 to 2018, encompassing 70,190 participants. We excluded individuals younger than 20 years, pregnant women (n = 31,152), participants not meeting CRS diagnostic criteria (n = 37,152), and those with incomplete ACAG data, mortality information, or missing covariates required for the analyses (n = 103). Ultimately, 1783 participants were included in the final analysis (Fig. 1). The NHANES protocol was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided written informed consent. A flow diagram of eligible participant selection in the National Health and Nutrition. ACAG = albumin-corrected anion gap, CRS = cardiorenal syndrome. 2.2. Assessment of CRS Consistent with previous studies, CRS was defined as the coexistence of cardiovascular disease and CKD.[18] Cardiovascular disease is identified through self-reported diagnoses. CKD is defined by an estimated glomerular filtration rate below 60 mL/min/1.73 m2. For estimating estimated glomerular filtration rate, we applied the CKD-epidemiology collaboration formula.[19] 2.3. Measurement of ACAG The AG was calculated as: AG (mmol/L) = [Sodium (mmol/L) + Potassium (mmol/L)] − [Chloride (mmol/L) + Bicarbonate (mmol/L)]. To account for variations in serum albumin, the ACAG was calculated using the following formula: ACAG (mmol/L) = AG + [4.4 − Albumin (g/dL)] × 2.5.[14] 2.4. Mortality data collection Mortality information for study participants was obtained from the U.S. NCHS, with follow-up completed on December 31, 2019. Deaths from cardiovascular disease were identified according to the National Death Index linked mortality files using International Classification of Diseases, Tenth Revision codes, including diseases of the heart (I00–I09, I11, I13, I20–I51), cerebrovascular diseases (I60–I69), and other major diseases of the circulatory system. All-cause mortality included deaths from any cause. Follow-up time was calculated for each participant from the date of the baseline interview until either death or the end of the follow-up period. 2.5. Covariables The NHANES survey obtained comprehensive demographic and health information, such as age, sex, race/ethnicity, marital status, education level, poverty income ratio, obesity, smoking, alcohol intake, hypertension, diabetes, and hyperlipidemia. Detailed definitions and grouping criteria for these variables are provided in Table S1, Supplemental Digital Content 1. 2.6. Statistical analyses Descriptive analyses were performed for demographic characteristics, with continuous variables expressed as mean ± standard deviation and categorical variables summarized as counts and percentages. All analyses accounted for the complex multistage sampling design of NHANES by incorporating appropriate sampling weights, strata, and primary sampling units to obtain nationally representative estimates. Group differences were evaluated using weighted t-tests for continuous data and weighted chi-square tests for categorical data. Associations between ACAG and both all-cause and cardiovascular mortality were examined using Cox proportional hazards models. Covariates included in the adjusted models were selected a priori based on established clinical relevance, previous literature, and their potential roles as confounders. Model 1 was unadjusted; Model 2 was adjusted for age, sex, education, marital status, poverty income ratio, and race/ethnicity; and Model 3 was further adjusted for obesity, alcohol consumption, smoking, hypertension, diabetes, and hyperlipidemia. Restricted cubic spline (RCS) modeling was used to examine dose–response associations between ACAG and mortality. Kaplan–Meier curves were applied to estimate survival, with group differences tested by log-rank analysis. As Kaplan–Meier analyses are unadjusted, multivariable Cox proportional hazards models were used to assess the independent association of ACAG with mortality after adjustment for age and other covariates. Subgroup analyses were also performed to explore interactions across stratified variables. Statistical analyses were carried out in R software (version 4.2.2), and significance was defined as a two-sided P < .05. Because the primary analyses were conducted using Cox proportional hazards models, hazard ratios and 95% confidence intervals were used as the principal effect measures. Conventional R2 statistics were not reported, as they are not standard measures for time-to-event models. 3. Results 3.1. Characteristics of the participants A total of 1783 individuals met the inclusion criteria and were included in the final analysis. The median follow-up time was 72 months. During follow-up, 924 participants survived and 859 died. Most participants were older than 65 years (73%), with an approximately equal distribution of sex, and the majority were non-Hispanic White (76%). Overall, 52% lived with a partner, 28% had less than a high school education, and 28% resided in low-income households. Compared with survivors, nonsurvivors were more likely to be older, non-Hispanic White, unmarried, better educated (at least high school), nonobese, and more frequently reported smoking, alcohol consumption, and hyperlipidemia (all P < .05) (Table 1). Table 1The demographic characteristics of the CRS population in the present study were stratified by survival status, weighted.CharacteristicOverall (weighted N = 7,414,449)Survivor (weighted N = 4,037,239)Non-survivors (weighted N = 3,377,210) P Unweighted sample size, n1783924859–Age (%) 20–652,011,816 (27%)1,419,835 (35%)591,981 (18%)<.001 >655,402,633 (73%)2,617,404 (65%)2,785,229 (82%)Sex (%) Female3,670,261 (50%)1,974,108 (49%)1,696,153 (50%).67 Male3,744,188 (50%)2,063,131 (51%)1,681,057 (50%)Race/ethnicity (%) Non-Hispanic White5,632,150 (76%)2,878,565 (71%)2,753,585 (82%)<.001 Non-Hispanic Black905,552 (12%)562,080 (14%)343,472 (10%) Other567,294 (7.7%)405,299 (10%)161,995 (4.8%) Mexican American309,453 (4.2%)191,295 (4.7%)118,158 (3.5%)Married/live with partner (%) No3,548,684 (48%)1,752,467 (43%)1,796,217 (53%).001 Yes3,865,765 (52%)2,284,772 (57%)1,580,993 (47%)Education level (%) Below high school2,085,961 (28%)989,129 (25%)1,096,833 (32%)<.001 High school or above5,324,733 (72%)3,045,803 (75%)2,278,930 (68%)PIR (%) Poor1,952,660 (28%)1,049,472 (28%)903,188 (29%).59 Not Poor4,901,310 (72%)2,725,860 (72%)2,175,450 (71%)Obesity (%) No3,620,705 (51%)1,823,671 (46%)1,797,034 (57%)<.001 Yes3,496,661 (49%)2,150,565 (54%)1,346,097 (43%)Smoking (%) Never3,107,567 (42%)1,768,284 (44%)1,339,282 (40%).013 Former3,098,721 (42%)1,525,689 (38%)1,573,031 (47%) Current1,208,162 (16%)743,266 (18%)464,896 (14%)Drinking (%) Never992,178 (15%)456,019 (13%)536,159 (18%)<.001 Former2,107,863 (33%)868,794 (26%)1,239,069 (41%) Mild2,365,201 (37%)1,427,070 (42%)938,130 (31%) Moderate471,912 (7.3%)326,528 (9.6%)145,384 (4.8%) Heavy489,613 (7.6%)325,712 (9.6%)163,901 (5.4%)Hypertension (%) No1,005,350 (14%)579,468 (14%)425,882 (13%).33 Yes6,406,416 (86%)3,457,771 (86%)2,948,645 (87%)Diabetes (%) No3,706,071 (50%)2,040,881 (51%)1,665,190 (49%).68 Yes3,708,378 (50%)1,996,358 (49%)1,712,020 (51%)Hyperlipidemia (%) No702,410 (9.5%)329,201 (8.2%)373,208 (11%).045 Yes6,712,039 (91%)3,708,038 (92%)3,004,002 (89%)ACAG (mean (SD))16.51 (2.58)16.58 (2.82)16.43 (2.27).82ACAG, Tertile (%) T12,431,487 (33%)1,371,158 (34%)1,060,330 (31%).022 T22,497,453 (34%)1,210,824 (30%)1,286,629 (38%) T32,485,509 (34%)1,455,258 (36%)1,030,252 (31%)Cardiovascular mortality (%)1,220,539 (16%)0 (0%)1,220,539 (36%)<.001Note: T1–T3: Grouped by tertile according to ACAG; Mean (SD) for continuous variables: the P value was calculated by the weighted Student’s t test; Percentages (weighted N, %) for categorical variables: the P value was calculated by the weighted chi-square test.ACAG = albumin-corrected anion gap, CRS = cardiorenal syndrome, PIR = ratio of family income to poverty. The demographic characteristics of the CRS population in the present study were stratified by survival status, weighted. Note: T1–T3: Grouped by tertile according to ACAG; Mean (SD) for continuous variables: the P value was calculated by the weighted Student’s t test; Percentages (weighted N, %) for categorical variables: the P value was calculated by the weighted chi-square test. ACAG = albumin-corrected anion gap, CRS = cardiorenal syndrome, PIR = ratio of family income to poverty. 3.2. ACAG and all-cause mortality, cardiovascular mortality Table 2 summarizes findings from 3 Cox regression models assessing the association between ACAG and mortality. As a continuous variable, higher ACAG was significantly related to all-cause mortality (Model 1: hazard ratio [HR] = 1.12, 95% confidence interval [CI]: 1.08–1.15, P < .001) and cardiovascular mortality (HR = 1.12, 95% CI: 1.06–1.18, P < .001). These associations persisted after full multivariable adjustment (Model 3: all-cause mortality, HR = 1.14, 95% CI: 1.09–1.18; cardiovascular mortality, HR = 1.14, 95% CI: 1.08–1.21). When participants were categorized into tertiles, those in the highest tertile (T3) had significantly greater risks of both all-cause mortality (HR = 1.78, 95% CI: 1.41–2.26) and cardiovascular mortality (HR = 1.96, 95% CI: 1.31–2.93) compared with the lowest tertile, indicating a clear dose–response relationship. RCS analyses further demonstrated a linear positive association between ACAG and mortality risk (Fig. 2). Kaplan–Meier curves showed clear survival differences among ACAG tertiles (P < .001), indicating that elevated ACAG was linked to higher all-cause and cardiovascular mortality (Fig. 3). Table 2HRs (95% CIs) for all-cause mortality and cardiovascular mortality according to ACAG in the CRS population.ACAGModel 1 [HR (95% CI)]P-valueModel 2 [HR (95% CI)]P-valueModel 3 [HR (95% CI)]P-valueAll-cause mortalityContinuous1.12 (1.08, 1.15)<.0011.12 (1.08, 1.15)<.0011.14 (1.09, 1.18) <.001 Tertile T11 (ref.)1 (ref.)1 (ref.) T21.39 (1.12, 1.72).0031.31 (1.04, 1.64).0201.39 (1.13, 1.72) .002 T31.71 (1.40, 2.11)<.0011.68 (1.37, 2.06)<.0011.78 (1.41, 2.26) <.001 P for trend<.001<.001<.001Cardiovascular mortalityContinuous1.12 (1.06, 1.18)<.0011.11 (1.05, 1.17)<.0011.14 (1.08, 1.21) <.001 Tertile T11 (ref.)1 (ref.)1 (ref.) T21.54 (1.09, 2.18).0101.46 (0.99, 2.14).0601.42 (1.00, 2.01).050 T31.76 (1.20, 2.56).0031.72 (1.17, 2.53).0101.96 (1.31, 2.93) .001 P for trend.002.005.001Bold values indicates statistically significant results (P < .05).Model 1: no covariates were adjusted; Model 2: age, sex, education level, marital status, PIR, and race/ethnicity were adjusted; Model 3: age, sex, education level, marital status, PIR, race/ethnicity, obesity, smoking, drinking, hypertension, diabetes, and hyperlipidemia were adjusted.ACAG = albumin-corrected anion gap, CI = confidence interval, CRS = cardiorenal syndrome, HR = hazard ratio, PIR = ratio of family income to poverty. HRs (95% CIs) for all-cause mortality and cardiovascular mortality according to ACAG in the CRS population. Bold values indicates statistically significant results (P < .05). Model 1: no covariates were adjusted; Model 2: age, sex, education level, marital status, PIR, and race/ethnicity were adjusted; Model 3: age, sex, education level, marital status, PIR, race/ethnicity, obesity, smoking, drinking, hypertension, diabetes, and hyperlipidemia were adjusted. ACAG = albumin-corrected anion gap, CI = confidence interval, CRS = cardiorenal syndrome, HR = hazard ratio, PIR = ratio of family income to poverty. The association of ACAG with all-cause (A) and cardiovascular mortality (B) among CRS visualized by restricted cubic spline. ACAG = albumin-corrected anion gap, CRS = cardiorenal syndrome. Kaplan–Meier analyses for mortality among the 3 groups. T1–T3, Tertiles 1–3; ACAG, albumin-corrected anion gap. (A) All-cause mortality; (B) Cardiovascular mortality. ACAG = albumin-corrected anion gap. 3.3. Subgroup analysis Figure 4 shows the relationship of ACAG with all-cause and cardiovascular mortality in various subgroups. No significant interactions were detected, suggesting that ACAG’s impact on mortality risk was stable across demographic strata. Notably, individuals aged 20 to 65 years appeared to be more vulnerable to elevated ACAG levels compared with those older than 65 years. Subgroup analysis between ACAG and mortality. (A) ACAG – all-cause mortality; (B) ACAG – cardiovascular mortality. HRs were calculated per 1-unit increase in ACAG. Analyses were adjusted for age, sex, education level, marital status, PIR, race/ethnicity, obesity, smoking, drinking, hypertension, diabetes, and hyperlipidemia. ACAG = albumin-corrected anion gap, HR = hazard ratio, PIR = poverty income ratio. 4. Discussion In this large, nationally representative cohort, we investigated the prognostic value of ACAG in patients with CRS. We found that elevated ACAG was independently associated with higher risks of both all-cause and cardiovascular mortality. To our knowledge, this is the first study to evaluate ACAG specifically in CRS patients, filling an important gap in the literature and suggesting its potential role as a simple and inexpensive marker for risk stratification in nephrology practice. The AG largely reflects negatively charged albumin, and in cases of hypoalbuminemia, a normal AG may misleadingly suggest a normal acid–base balance. Correcting AG for serum albumin levels therefore provides a more accurate assessment, and the ACAG is widely regarded as a superior indicator of metabolic acidosis and its associated pathophysiology.[20] Previous studies have consistently demonstrated the prognostic significance of elevated ACAG across diverse populations. Analyses from the MIMIC database linked higher ACAG to short- and long-term mortality in acute kidney injury,[21] myocardial infarction,[22] and asthma patients,[23] while NHANES data revealed its association with cardiovascular outcomes in diabetes.[24] Similarly, elevated ACAG predicted adverse prognosis in patients with severe heart failure[25] and in those undergoing coronary artery bypass surgery.[26] Collectively, these findings underscore ACAG as a robust prognostic marker across multiple clinical contexts. However, no prior study has specifically evaluated ACAG in patients with CRS, a high-risk group characterized by the interplay between cardiac and renal dysfunction. Our study is therefore the first to address this knowledge gap, highlighting the potential of ACAG as a simple and inexpensive marker for risk stratification in nephrology practice. Metabolic acidosis plays a central role in the progression of CRS by exacerbating acid–base imbalance and amplifying the pathological interaction between the heart and kidneys.[27,28] It contributes to adverse outcomes through multiple mechanisms, including electrolyte disturbances (e.g., hyperkalemia predisposing to arrhythmias and hyponatremia worsening renal dysfunction),[28] impaired myocardial contractility due to disrupted calcium handling,[29] and tubular injury that accelerates renal deterioration.[30,31] In addition, metabolic acidosis aggravates chronic inflammation[32] and oxidative stress,[33] promoting vascular damage, atherosclerosis, and heightened cardiovascular risk. These pathophysiological pathways provide a mechanistic explanation for our findings that elevated ACAG is associated with increased mortality in CRS. As ACAG reflects both albumin levels and unmeasured anions, it offers a more accurate estimate of metabolic acidosis than the traditional anion gap. Thus, monitoring ACAG may help nephrologists identify high-risk CRS patients and improve risk stratification in clinical practice. As ACAG is derived from routinely available biochemical parameters and does not require additional cost or specialized testing, it may serve as a practical tool for identifying high-risk patients with CRS in everyday clinical settings. Incorporating ACAG into routine assessment may improve early risk stratification, guide closer follow-up, and support timely preventive interventions. To our knowledge, this is the first study to investigate the association between ACAG and outcomes in CRS patients using a large, nationally representative cohort. We found that higher ACAG levels were consistently associated with increased risks of all-cause and cardiovascular mortality, suggesting its potential role as a prognostic indicator in CRS management. Several limitations should be acknowledged. First, the observational design precludes causal inference, and future studies using approaches such as Mendelian randomization are warranted. Second, some potential confounders—including inflammatory markers, electrolyte disturbances, and medication use—were not available in NHANES, which may have introduced residual confounding. Third, the cross-sectional nature of NHANES limited our ability to assess longitudinal changes in ACAG and their impact on outcomes. Finally, the study population was predominantly non-Hispanic White from the United States, which may restrict generalizability to other ethnicities and regions. Despite these limitations, our study provides the first evidence that elevated ACAG is strongly associated with adverse outcomes in CRS patients. These findings highlight the need for prospective, multicenter studies to validate ACAG as a simple and inexpensive marker for risk stratification in nephrology practice. 5. Conclusion In this nationally representative cohort of adults with CRS, higher ACAG levels were significantly associated with increased risks of all-cause and cardiovascular mortality. These findings suggest that ACAG may serve as a simple, inexpensive, and widely available marker to improve risk stratification in CRS patients, particularly in nephrology practice. While further prospective and interventional studies are needed to confirm its clinical utility, our study provides the first evidence supporting the prognostic value of ACAG in this high-risk population. Acknowledgments We express our gratitude to the participants of the National Health and Nutrition Examination Survey for their dedication to continuous research and the integration of electronic health records. Author contributions Conceptualization: Hongman Li, Xing Huang. Data curation: Hongman Li, Xing Huang. Formal analysis: Hongman Li, Chun Zhao, Yuan Deng. Methodology: Hongman Li, Xing Huang, Shengyong Si, Mengdie Liu. Resources: Hongman Li, Xing Huang. Validation: Hongman Li, Shengyong Si, Chun Zhao, Xue Zhao, Yuan Deng, Xue Zhao. Visualization: Hongman Li. Writing – original draft: Hongman Li, Shengyong Si, Zhijing Li. Writing – review & editing: Hongman Li, Xing Huang, Shengyong Si.