Abstract
Background
The Lok index is a composite laboratory marker derived from the international normalized ratio (INR), the aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio, and platelet count. In critically ill patients, this index may reflect not only hepatic dysfunction but also coagulation disturbances, hematologic imbalances, and overall systemic illness burden. However, the association between the Lok index and mortality risks in critically ill patients with coronary artery disease (CAD) remains poorly understood. This study aims to assess the association between the Lok index and short- and long-term all-cause mortality (ACM) risks in this population.
Methods
Relevant data were extracted from the MIMIC-Ⅳ (v3.1) database. Kaplan–Meier curves were generated based on the quartiles of the Lok index to compare ACM risks across groups, with differences assessed using the Log-Rank test. Restricted cubic spline (RCS) analysis and Cox proportional hazards models were employed to examine the association between the Lok index and ACM risks.
Results
In total, 7,594 patients were included, with mortality rates of 20.71% at 30 days, 27.88% at 90 days, and 38.83% at 365 days. Across increasing Lok quartiles, absolute mortality increased progressively: from 16.01% to 27.28% at 30 days, from 22.17% to 35.65% at 90 days, and from 32.23% to 47.13% at 365 days. In the fully adjusted Cox model, when analyzed as a continuous variable, the Lok index was significantly associated with ACM risks at 365 days (HR = 1.011, 95% CI: 1.005–1.016, P < 0.001). When analyzed as a categorical variable, compared with Q1, a high Lok index (Q4) was significantly associated with ACM risks at 365 days (HR = 1.237, 95% CI: 1.107–1.381, P < 0.001). RCS analysis revealed a nonlinear relationship between the Lok index and mortality outcomes. Subgroup analyses, sensitivity analyses, and interaction tests confirmed the robustness of these associations.
Conclusion
In critically ill CAD patients, a higher Lok index is significantly associated with an elevated ACM risk at 30 days, 90 days, or 365 days, indicating that a higher Lok index is independently associated with both short- and long-term ACM risks. Nevertheless, given the modest per-unit effect size and the retrospective design, the Lok index should not be used in isolation to guide individual treatment decisions.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12872-026-06194-w.
Keywords: Coronary artery disease, Lok index, All-cause mortality risk, ICU, MIMIC-IV
Introduction
Coronary artery disease (CAD) carries a persistently elevated risk of all-cause mortality (ACM), representing a major public health burden worldwide. This is particularly evident in critically ill patients, including those with acute myocardial infarction, cardiogenic shock, or refractory heart failure [1]. Traditional risk scores (e.g., GRACE and TIMI), which are based primarily on indicators specific to the cardiovascular system, are valuable but may have limited predictive performance in assessing critically ill patients with CAD complicated by multiple organ dysfunction [2]. In recent years, multiple large cohort studies have identified non-alcoholic fatty liver disease (NAFLD) as an independent risk factor for cardiovascular events. The increasing prevalence rate of NAFLD parallels the epidemic trend of metabolic syndrome or obesity [3, 4]. Moreover, the concept of the ‘liver-heart axis’ has laid a theoretical foundation for understanding the bidirectional interactions between the liver and the heart [5]. Acute cardiovascular events cause a sharp decline in cardiac output, leading to hepatic hypoperfusion and passive cardiac-origin liver injury, which manifests as elevated transaminases and impaired synthesis function [6]. Furthermore, chronic liver diseases, particularly progressive liver fibrosis, may exacerbate CAD severity and worsen outcomes via multiple mechanisms, including systemic inflammation [7], coagulation disorders [8], neurohormonal activation, and metabolic dysregulation [9]. Therefore, assessing the degree of liver fibrosis may be crucial for assessing the overall pathophysiological burden in individuals with CAD.
In this study, it is hypothesized that a higher Lok index is associated with an increased ACM risk in ICU patients with CAD. The Lok index is a non-invasive, clinically validated tool for liver fibrosis assessment that integrates the international normalized ratio (INR), the aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio, and platelet count [10]. The Lok index is selected because it comprises routinely available ICU variables and may capture the combined burden of hepatic dysfunction, coagulation abnormalities, and hematologic derangement. Crucially, the aim of this study is not to demonstrate that the Lok index outperforms other fibrosis-related scores, but to assess whether it provides complementary prognosis-related information for critically ill patients with CAD. Accordingly, in this study, the Lok index is calculated based on the Medical Information Mart for Intensive Care Ⅳ (MIMIC-Ⅳ) (v3.1) database, and its association with adverse outcomes in this population is analyzed.
Methods and materials
Study population
In this retrospective study, data on patients with CAD were extracted from the MIMIC-Ⅳ database, a large-scale, publicly available repository developed and maintained by the Laboratory for Computational Physiology at the Massachusetts Institute of Technology (MIT). The database contains comprehensive clinical data from patients admitted to the intensive care unit (ICU) at Beth Israel Deaconess Medical Center. The author has obtained the required certification (ID: 66647897) and has been granted access to the dataset and relevant data. The database was approved for research purposes by the institutional review boards of MIT and Beth Israel Deaconess Medical Center, and the need for informed consent was waived.
Patients were eligible if they had an ICD-9 or ICD-10 diagnosis of CAD and a first ICU admission recorded in the MIMIC-Ⅳ database. Exclusion criteria were: ICU stay < 24 h; missing values for ALT, AST, platelet count, or INR; or survival time shorter than 24 h. Because the cohort was defined based on ICD-coded CAD diagnoses, detailed information on CAD subtype, revascularization status, vasoactive therapy, and cardiac imaging parameters was not uniformly incorporated into this analysis. To assess potential selection bias, baseline characteristics and mortality risks were compared between included patients and those excluded due to missing Lok index components.
Data extraction
Relevant data was extracted using Structured Query Language (SQL) from the PostgreSQL (v17.2.2). The following potential confounding variables were considered: (ⅰ) baseline demographic information: age, sex, and race; (ⅱ) comorbidities: hypertension, diabetes mellitus (DM), atrial fibrillation, heart failure, respiratory failure, sepsis (identifying sepsis in MIMIC-Ⅳ using previously validated Sepsis-3 extraction algorithms), malignancy, cirrhosis, and renal dysfunction; (ⅲ) treatment approaches: invasive ventilation and vancomycin administration (including the latter as a clinically relevant covariate as it may partially reflect infection burden and treatment intensity among critically ill patients); (ⅳ) laboratory parameters: red blood cell (RBC) count, white blood cell (WBC) count, platelet (PLT) count, hemoglobin (HGB), fibrinogen (FBG), INR, serum creatinine, ALT, AST, and serum sodium; and (ⅴ) disease severity scores: Simplified Acute Physiology Score Ⅱ (SAPS Ⅱ), Glasgow Coma Scale (GCS), and Sequential Organ Failure Assessment (SOFA). Physiological variables included heart rate. Albumin and body mass index were not included in the final multivariable models due to substantial missingness and potential collinearity with illness-severity indicators. For all laboratory parameters, only the first measured values within 24 h after ICU admission were selected. Patients with missing ALT, AST, platelet count, or INR were excluded, as these variables are required to calculate the Lok index. After cohort selection, remaining continuous covariates with missingness below 20% were imputed using multiple imputation based on a random forest algorithm.
Exposure variables and outcomes
The Lok index was calculated for all participants using a previously established formula [10]:
 |
The primary outcome was ACM risks at 365 days after admission, while secondary outcomes were ACM risks at 30 days and at 90 days after admission.
Statistical analysis
Participants were stratified into quartiles of the Lok index for descriptive and survival analyses. This categorical approach was employed for descriptive stratification and visualization; however, continuous-variable and spline analyses were also performed to avoid overreliance on arbitrary categorization. Normality was tested for all continuous variables, none of which followed a normal distribution; accordingly, they were summarized as median with interquartile range (IQR). Between‑group comparisons were performed using the Kruskal–Wallis rank-sum test. Categorical variables were expressed as frequencies and percentages, with differences between groups assessed using Pearson’s chi-square test. Time zero was defined as ICU admission. Follow-up for the 30-, 90-, and 365-day analyses was right-censored at the corresponding time horizon or at the last available survival time. Because only the first ICU admission was included, left truncation was not present, and interval censoring was not applicable. Covariates were selected a priori based on clinical relevance and their established associations with the hazards of mortality in critically ill patients. Multicollinearity was assessed for all variables included in the models, and those with a variance inflation factor (VIF) ≥ 5 were removed to reduce collinearity. Detailed results are shown in Supplementary Table 1. Cox proportional hazards models were used to estimate hazard ratios (HRs) for the Lok index, treated both as a continuous and a categorical variable. Three hierarchical Cox models were constructed. Model 1 was unadjusted. Model 2 was adjusted for age and sex as a minimal demographic benchmark. Model 3 (the primary model) was additionally adjusted for prespecified comorbidities, treatments, laboratory indices, and validated ICU severity scores (including SAPS Ⅱ and SOFA score) and served as the basis for our main interpretation. The proportional hazards assumption was tested using Schoenfeld residuals, with a P‑value > 0.05 indicating no violation (Supplementary Fig. 1). Kaplan–Meier survival curves, stratified by Lok index quartiles, were generated to compare outcomes related to mortality risks across groups. Restricted cubic spline (RCS) analysis was performed to examine potential nonlinear relationships between the Lok index and ACM risks among critically ill patients with CAD. Subgroup analyses were conducted, and forest plots were constructed to assess the consistency of outcomes across different groups. Sensitivity analysis was performed by excluding patients with cirrhosis, end-stage renal disease, or cancer to verify the robustness of the findings. To assess the incremental prognostic value of the Lok index beyond established ICU severity scores, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) analyses were performed by adding the Lok index to models containing SOFA score and SAPS Ⅱ. All statistical analyses were carried out using RStudio (v4.2.3). A two-sided P-value < 0.05 was considered statistically significant.
Results
Baseline characteristics
In total, 7,594 critically ill patients with CAD were included (Fig. 1). Of these, 4,953 (65.22%) were male, and 2,641 (34.78%) were female. During the 365-day follow-up period, 2,949 patients (38.83%) died. Baseline characteristics stratified by Lok index quartiles are summarized in Table 1. In group Q4, a higher proportion of males was observed, along with a greater prevalence rate of several comorbidities, including atrial fibrillation, heart failure, respiratory failure, sepsis, and cirrhosis. Higher values were also recorded for laboratory parameters such as AST, ALT, INR, WBC count, and serum creatinine. Both SOFA and SAPS Ⅱ values were higher in group Q4. Comparisons between included and excluded patients are shown in Supplementary Table 2.
Fig. 1.
Flow diagram for participants
Table 1.
Baseline characteristics of participants grouped by quartiles of Lok index
| Variable | Overall | Q1 (Lok < 11.5) | Q2 (11.5 ≤ Lok < 13) | Q3 (13 ≤ Lok < 15.2) | Q4 (Lok ≥ 15.2) | P-value |
|---|---|---|---|---|---|---|
| N | 7594 | 1899 | 1898 | 1898 | 1899 | |
| Age (years) | 72.61(63.67,80.98) | 69.80(60.74,78.69) | 73.49(64.54,81.73) | 73.60(64.75,81.81) | 73.49(64.79,81.43) | < 0.001 |
| Sex (%) | < 0.001 | |||||
| Male | 4953(65.22%) | 1156(60.87%) | 1196(63.01%) | 1310(69.02%) | 1291(67.98%) | |
| Female | 2641(34.78%) | 743(39.13%) | 702(36.99%) | 588(30.98%) | 608(32.02%) | |
| Race (%) | 0.176 | |||||
| Black Individuals | 701(9.23%) | 183(9.64%) | 186(9.80%) | 165(8.69%) | 167(8.79%) | |
| White Individuals | 5101(67.17%) | 1232(64.88%) | 1280(67.44%) | 1282(67.54%) | 1307(68.83%) | |
| Other | 1792(23.60%) | 484(25.49%) | 432(22.76%) | 451(23.76%) | 425(22.38%) | |
| Diabetes Mellitus (%) | 0.009 | |||||
| No | 3814(50.22%) | 900(47.39%) | 953(50.21%) | 1004(52.90%) | 957(50.39%) | |
| Yes | 3780(49.78%) | 999(52.61%) | 945(49.79%) | 894(47.10%) | 942(49.61%) | |
| Hypertension (%) | < 0.001 | |||||
| No | 3115(41.02%) | 686(36.12%) | 753(39.67%) | 798(42.04%) | 878(46.23%) | |
| Yes | 4479(58.98%) | 1213(63.88%) | 1145(60.33%) | 1100(57.96%) | 1021(53.77%) | |
| AF (%) | < 0.001 | |||||
| No | 3654(48.12%) | 1158(60.98%) | 984(51.84%) | 870(45.84%) | 642(33.81%) | |
| Yes | 3940(51.88%) | 741(39.02%) | 914(48.16%) | 1028(54.16%) | 1257(66.19%) | |
| Cirrhosis (%) | < 0.001 | |||||
| No | 6937(91.35%) | 1832(96.47%) | 1778(93.68%) | 1705(89.83%) | 1622(85.41%) | |
| Yes | 657(8.65%) | 67(3.53%) | 120(6.32%) | 193(10.17%) | 277(14.59%) | |
| Malignancy (%) | 0.758 | |||||
| No | 6141(80.87%) | 1545(81.36%) | 1541(81.19%) | 1520(80.08%) | 1535(80.83%) | |
| Yes | 1453(19.13%) | 354(18.64%) | 357(18.81%) | 378(19.92%) | 364(19.17%) | |
| Heart Failure (%) | < 0.001 | |||||
| No | 2713(35.73%) | 795(41.86%) | 702(36.99%) | 699(36.83%) | 517(27.22%) | |
| Yes | 4881(64.27%) | 1104(58.14%) | 1196(63.01%) | 1199(63.17%) | 1382(72.78%) | |
| Respiratory Failure (%) | 0.009 | |||||
| No | 3996(52.62%) | 1010(53.19%) | 1034(54.48%) | 1015(53.48%) | 937(49.34%) | |
| Yes | 3598(47.38%) | 889(46.81%) | 864(45.52%) | 883(46.52%) | 962(50.66%) | |
| Renal Dysfunction (%) | 0.119 | |||||
| No | 6809(89.66%) | 1710(90.05%) | 1712(90.20%) | 1712(90.20%) | 1675(88.20%) | |
| Yes | 785(10.34%) | 189(9.95%) | 186(9.80%) | 186(9.80%) | 224(11.80%) | |
| Sepsis (%) | < 0.001 | |||||
| No | 2382(31.37%) | 658(34.65%) | 665(35.04%) | 560(29.50%) | 499(26.28%) | |
| Yes | 5212(68.63%) | 1241(65.35%) | 1233(64.96%) | 1338(70.50%) | 1400(73.72%) | |
| Administration of Vancomycin (%) | < 0.001 | |||||
| No | 2048(26.97%) | 580(30.54%) | 550(28.98%) | 481(25.34%) | 437(23.01%) | |
| Yes | 5546(73.03%) | 1319(69.46%) | 1348(71.02%) | 1417(74.66%) | 1462(76.99%) | |
| MV (%) | 0.033 | |||||
| No | 6193(81.55%) | 1521(80.09%) | 1561(82.24%) | 1582(83.35%) | 1529(80.52%) | |
| Yes | 1401(18.45%) | 378(19.91%) | 337(17.76%) | 316(16.65%) | 370(19.48%) | |
| GCS score | 15.00(14.00,15.00) | 15.00(14.00,15.00) | 15.00(14.00,15.00) | 15.00(14.00,15.00) | 15.00(14.00,15.00) | 0.082 |
| SAPS II | 38.00(31.00,47.00) | 35.00(28.00,42.00) | 37.00(30.00,45.00) | 39.00(31.00,48.00) | 41.00(33.00,51.00) | < 0.001 |
| SOFA score | 5.00(3.00,8.00) | 4.00(2.00,6.00) | 4.00(2.00,7.00) | 5.00(3.00,8.00) | 6.00(4.00,9.00) | < 0.001 |
| HR (bmp) | 84.00(73.00,97.00) | 83.00(71.00,96.00) | 83.00(72.00,96.00) | 84.00(73.00,97.00) | 85.00(74.00,99.00) | < 0.001 |
| ALT (IU/L) | 26.00(16.00,56.00) | 27.00(17.00,54.00) | 25.00(15.00,50.00) | 25.00(15.00,51.00) | 29.00(16.00,72.00) | < 0.001 |
| AST (IU/L) | 40.00(23.00,90.00) | 28.00(19.00,52.00) | 35.00(22.00,67.00) | 45.00(26.00,100.00) | 66.00(34.00,210.00) | < 0.001 |
| PLT (×109/L) | 189.00(138.00,250.00) | 248.00(198.00,314.00) | 189.00(148.00,240.00) | 164.00(121.00,215.00) | 156.00(109.00,218.00) | < 0.001 |
| INR | 1.30(1.10,1.60) | 1.10(1.00,1.20) | 1.20(1.10,1.30) | 1.40(1.30,1.50) | 1.90(1.60,2.60) | < 0.001 |
| FBG (mg/dL) | 134.00(108.00,181.00) | 138.00(108.00,192.00) | 135.50(108.00,180.00) | 131.00(107.00,174.00) | 133.00(106.00,181.00) | 0.003 |
| HGB (g/L) | 10.60(9.00,12.30) | 11.40(9.70,13.00) | 10.70(9.20,12.30) | 10.20(8.70,12.00) | 9.90(8.40,11.80) | < 0.001 |
| WBC (×109/L) | 11.00(7.90,15.40) | 11.00(8.20,15.30) | 10.40(7.50,14.30) | 11.20(7.90,15.60) | 11.30(8.10,15.90) | < 0.001 |
| RBC (×109/L) | 3.55(3.02,4.13) | 3.86(3.32,4.40) | 3.59(3.08,4.14) | 3.40(2.90,3.98) | 3.34(2.81,3.94) | < 0.001 |
| Cr (mg/dL) | 1.20(0.90,1.80) | 1.10(0.80,1.50) | 1.10(0.80,1.70) | 1.20(0.90,1.70) | 1.40(1.00,2.10) | < 0.001 |
| Na (mmol/L) | 138.00(135.00,141.00) | 139.00(136.00,141.00) | 138.00(136.00,141.00) | 138.00(135.00,141.00) | 138.00(135.00,141.00) | 0.006 |
| Status_30d | 1573(20.71%) | 304(16.01%) | 336(17.70%) | 415(21.87%) | 518(27.28%) | < 0.001 |
| Time_30d | 30.00(30.00,30.00) | 30.00(30.00,30.00) | 30.00(30.00,30.00) | 30.00(30.00,30.00) | 30.00(21.32,30.00) | < 0.001 |
| Status_90d | 2117(27.88%) | 421(22.17%) | 459(24.18%) | 560(29.50%) | 677(35.65%) | < 0.001 |
| Time_90d | 90.00(58.62,90.00) | 90.00(90.00,90.00) | 90.00(90.00,90.00) | 90.00(49.37,90.00) | 90.00(21.32,90.00) | < 0.001 |
| Status_1year | 2949(38.83%) | 612(32.23%) | 676(35.62%) | 766(40.36%) | 895(47.13%) | < 0.001 |
| Time_1year | 365.00(58.62,365.00) | 365.00(131.97,365.00) | 365.00(109.10,365.00) | 365.00(49.37,365.00) | 365.00(21.32,365.00) | < 0.001 |
AST aspartate aminotransferase, ALT alanine aminotransferase, FBG fasting blood glucose, Cr creatinine, GCS Glasgow Coma Scale, AF atrial fibrillation, INR international normalized ratio, HGB hemoglobin, PLT platelet, SOFA Sequential Organ Failure Assessment, SAPS II: Simplified Acute Physiology Score II, WBC white blood cell, RBC red blood cell, MV mechanical ventilation, HR hear rate
Survival analysis
Kaplan–Meier survival curves stratified by Lok index quartiles were plotted to compare the incidence rate of the primary outcome across groups (Fig. 2a and b, and c). A significantly higher mortality risk was observed in group Q4 at 30, 90, and 365 days (Log-Rank P < 0.001). These findings indicated that a high Lok index was significantly associated with a hazard of ACM in this population.
Fig. 2.
Kaplan-Meier curves for ACM across Lok index quartiles at (a) 30, (b) 90, and (c) 365 days, with group comparisons performed using the log-rank test. Numbers at risk are shown below each panel
Association between the lok index and outcomes
The proportional hazards assumption was verified using Schoenfeld residuals for all variables included in the Cox regression model, and no violations were detected (Global test P = 0.1405). In baseline comparisons, the mortality risk of group Q1 was found to be the lowest compared with the other groups. Using group Q1 as the control, Cox proportional hazards models were fitted to analyze the relationship between the Lok index and ACM risks. In the fully adjusted model, when the Lok index was analyzed as a continuous variable, a one-unit increase was associated with a 1.1% increase in 365-day ACM risks, a 0.7% increase in 90-day ACM risks, and a 0.9% increase in 30-day ACM risks. When analyzed as a categorical variable, a high Lok index (group Q4) was significantly associated with ACM risks at 365 days (HR = 1.237, 95% CI: 1.107–1.381, p < 0.001), 30 days (HR = 1.269, 95% CI: 1.090–1.478, p < 0.01), and 90 days (HR = 1.265, 95% CI: 1.110–1.442, p < 0.001). Accordingly, a high Lok index was identified as a risk factor for ACM in this population (Table 2). Addition of the Lok index to models based on SOFA and SAPS Ⅱ scores significantly enhanced risk reclassification and discrimination for short-term mortality outcomes (Supplementary Table 3).
Table 2.
Cox proportional hazard models for ACM risks at 30 days, 90 days, and 365 days
| Variable | Non-Adjusted | Adjustment I | Adjustment II |
|---|---|---|---|
| All-cause mortality risks at 30 days | |||
| Lok index as a continuous variable | 1.029 (1.022,1.035) ※※※ | 1.026 (1.019,1.032) ※※※ | 1.007 (1.000,1.014) ※ |
| Lok index (Q1) | Ref | Ref | Ref |
| Lok index (Q2) | 1.130 (0.968,1.320) | 1.036 (0.887,1.210) | 0.988 (0.845,1.156) |
| Lok index (Q3) | 1.430 (1.233,1.658) ※※※ | 1.313 (1,131,1.523) ※※※ | 1.100 (0.943,1.282) |
| Lok index (Q4) | 1.859 (1.614,2.142) ※※※ | 1.712 (1.485,1.974) ※※※ | 1.269 (1.090,1.478) ※※ |
| All-cause mortality risks at 90 days | |||
| Lok index as a continuous variable | 1.028 (1.022,1.033) ※※※ | 1.025 (1.019,1.031) ※※※ | 1.009 (1.003,1.015) ※※ |
| Lok index (Q1) | Ref | Ref | Ref |
| Lok index (Q2) | 1.116 (0.977,1.273) | 1.019 (0.892,1.164) | 0.995 (0.844,1.103) |
| Lok index (Q3) | 1.407 (1.240,1.597) ※※※ | 1.287 (1.133,1.461) ※※※ | 1.094 (0.959,1.248) |
| Lok index (Q4) | 1.792 (1.587,2.024) ※※※ | 1.647 (1.457,1.861) ※※※ | 1.265 (1.110,1.442) ※※※ |
| All-cause mortality risks at 365 days | |||
| Lok index as a continuous variable | 1.027 (1.022,1.032) ※※※ | 1.025 (1.020,1.030) ※※※ | 1.011 (1.005,1.016) ※※※ |
| Lok index (Q1) | Ref | Ref | Ref |
| Lok index (Q2) | 1.134 (1.016,1.265) ※ | 1.044 (0.935,1.165) | 0.974 (0.872,1.089) |
| Lok index (Q3) | 1.344 (1.209,1.495) ※※※ | 1.242 (1.116,1.382) ※※※ | 1.076 (0.963,1.202) |
| Lok index (Q4) | 1.679 (1.515,1.861) ※※※ | 1.560 (1.407,1.730) ※※※ | 1.237 (1.107,1.381) ※※※ |
Adjustment I: Adjustments are made based on age and sex. Adjustment II: Adjustments are made based on age, sex, race, atrial fibrillation, creatinine value, hypertension, DM, FBG, GCS score, heart rate, heart failure, respiratory failure, renal dysfunction, sepsis, SAPS Ⅱ, SOFA score, serum sodium, WBC count, RBC count, malignant tumor, cirrhosis, mechanical ventilation, and administration of vancomycin
P value: ※ < 0.05; ※※ < 0.01; ※※※ < 0.001
RCS analysis
RCS curves were used to investigate potential nonlinear relationships between the Lok index and adverse outcomes. A nonlinear association was observed (P for nonlinearity = 0.003). In the higher range of the Lok index, the risk curve rose more steeply, possibly reflecting the concurrent presence of coagulation dysfunction, thrombocytopenia, AST/ALT imbalance, and broader multiorgan derangement in critically ill patients with CAD (Fig. 3a and b, and c).
Fig. 3.
RCS curves for the hazard ratio of ACM risks at (a) 30 days, (b) 90 days, and (c) 365 days. Vertical dashed lines indicate inflection points, dark gray lines demonstrate fully adjusted risk ratios, shaded areas show 95% confidence intervals, and a horizontal dashed line represents a hazard ratio = 1. The nonlinear pattern and apparent inflection points are exploratory and should not be interpreted as clinical cut-offs
Subgroup analysis
Subgroup analyses were performed according to age, sex, DM, hypertension, sepsis, and heart failure (Fig. 4a and b, and c). For 365-day ACM risks, a high Lok index (Q4) was significantly associated with ACM risks in all subgroups except those not suffering from heart failure. For 30-day ACM risks, a significant association was observed in all subgroups except males aged < 70 years and patients without DM, hypertension, sepsis, or heart failure. For 90-day ACM risks, the association remained significant in all subgroups except patients aged < 70 years and those without sepsis or heart failure. Interaction tests revealed significant differences in 90‑day and 365‑day ACM risks between age groups (P < 0.05), and between patients with and without sepsis for 30‑day and 365‑day mortality risks (P < 0.05). Some subgroup strata may have been underpowered, and all subgroup analyses should be regarded as exploratory and hypothesis-generating, with interpretation based primarily on interaction testing rather than on significance within individual strata.
Fig. 4.
Exploratory subgroup analyses of ACM at (a) 30 days, (b) 90 days, and (c) 365 days
Sensitivity analysis
A sensitivity analysis based on the fully adjusted model (Model 3) was performed after excluding patients with renal failure (n = 785), cirrhosis (n = 560), or malignancy (n = 1,174). The results showed that a high Lok index (Q4) remained significantly associated with 365‑day ACM risks (HR = 1.214, 95% CI: 1.047–1.407, p < 0.01).
Incremental prognostic value
NRI and IDI analyses were performed to assess the added prognostic value of the Lok index. Addition of the Lok index to models based on SOFA and SAPS II scores significantly enhanced risk reclassification for both 30‑day and 90‑day mortality risks (all P < 0.001), suggesting that the Lok index may provide supplementary prognostic information beyond conventional ICU severity scores (Supplementary Table 3).
Discussion
In this study, a higher Lok index was independently associated with increased short- and long-term ACM risks after adjustment for major confounders and established ICU severity scores. Nevertheless, given the observational design, the modest effect sizes, and the considerable clinical heterogeneity of the cohort, these findings should be interpreted with caution. Due to the modest per-unit effect size, statistical significance should not be equated with substantial individual-level predictive utility.
The robustness of this relationship was supported by subgroup and sensitivity analyses. Although NRI and IDI analyses suggested incremental prognostic value, the Lok index should be viewed as a supplementary marker rather than a standalone risk predictor. The observed association may partly reflect global illness severity and multiorgan dysfunction rather than isolated liver fibrosis. These findings suggested that a higher Lok index was associated with increased mortality risks among critically ill patients with CAD. Given the large sample size, statistical significance should not be overinterpreted as evidence of clinically useful individual prediction. Instead, the Lok index should be considered an auxiliary risk indicator, not a standalone prognostic marker.
Association and mechanisms between Lok index and critical CAD
Of note, in the ICU setting, the Lok index may capture the overall burden of systemic illness rather than liver fibrosis in isolation. By integrating INR, the AST/ALT ratio, and platelet count [11–13], the Lok index potentially reflects hepatic dysfunction, coagulation abnormalities, hematologic disturbances, systemic inflammation [14], and the multiorgan dysfunction commonly observed in critically ill patients [15, 16]. Thus, the observed association may in part arise from acute physiological derangement and illness severity rather than purely liver-specific pathology.
Despite the fact that several biologically plausible mechanisms may contribute to this association, including inflammation, neurohormonal activation [17, 18], metabolic dysregulation [19], and impaired coagulation [20], these interpretations remain speculative and cannot be directly verified by this study. Accordingly, the Lok index should be viewed as an adjunctive risk signal rather than a standalone prognosis marker.
Notably, although the continuous-variable analyses reached statistical significance, the per-unit HRs were relatively small, limiting their interpretability at the individual patient level. Therefore, statistical significance in this large cohort should not be equated with substantial clinical predictive utility.
Furthermore, the subgroup and RCS analyses should be considered exploratory. The CAD cohort remained clinically heterogeneous, and key cardiovascular variables, such as CAD subtype and revascularization status, were unavailable, which may have influenced effect estimates and constrained subgroup interpretations. In the future, prospective studies with external validation and more detailed cardiovascular characterization are needed to further clarify the prognostic performance of the Lok index.
In summary, among critically ill patients, the Lok index may in part reflect acute systemic severity and hepatic dysfunction rather than liver fibrosis in isolation, serving as a composite indicator of systemic physiological derangement that complements traditional severity scores.
Differences in the association between Lok Index and ACM risks across subgroups
Subgroup analyses indicated that a high Lok index was associated with a higher ACM risk. Specifically, a high Lok index was significantly associated with long-term (at 365 days) mortality risks in most subgroups. Patients with heart failure inherently face higher risks of hepatic congestion and abnormal liver function, which may obscure the relationship between the Lok index and mortality risks among such patients [21, 22]. Furthermore, in the subgroup analyses of short-term mortality risks (at 30 days and at 90 days), a high Lok index was more significantly associated with mortality risks among patients aged ≥ 70 years, as well as those with DM, sepsis, or heart failure. The high prevalence rate of sepsis observed in this cohort is probably explained by the exclusive focus of this study on critically ill ICU patients, a population in which severe infection and multiorgan dysfunction are common.
The subgroup analyses were exploratory and primarily intended to assess the consistency of the observed association across clinically relevant strata. Interpretation was therefore based mainly on interaction testing rather than significance within individual subgroups. Although the associations appeared more significant in older patients and those with sepsis, DM, or heart failure [23, 24], these findings should be interpreted with caution due to multiple comparisons, residual confounding, and limited power in some strata.
The observed subgroup patterns may reflect differences in systemic inflammation, metabolic dysfunction, or vulnerability to multiorgan injury among critically ill patients [25–30]. However, these interpretations remain biologically plausible hypotheses rather than conclusions directly supported by these observational data. Overall, the subgroup findings should be viewed as hypothesis-generating and not overinterpreted. In the future, prospective studies with more detailed cardiovascular phenotyping and external validation are needed to determine whether these subgroup differences are reproducible and clinically meaningful.
Limitations
Several limitations should be acknowledged. First, as a retrospective single-center study based on the MIMIC-Ⅳ database, the findings in this study are subject to selection bias, information bias, and limited generalizability to other ICU populations and healthcare systems. Given the large sample size, statistically significant differences do not necessarily correspond to clinically meaningful effects. Second, CAD was defined by ICD codes, and several significant cardiovascular variables, including CAD subtype, revascularization status, vasoactive support, and cardiac imaging parameters, were unavailable. These omissions may have biased the estimated effect sizes and limited the generalizability of the findings in this study. In addition, patients with missing ALT, AST, platelet count, or INR values were excluded because these variables are required for calculating the Lok index. These factors may have introduced selection bias, contributed to clinical heterogeneity, and affected the representativeness of the study population. Third, although multiple imputation and extensive multivariable adjustment were performed, residual confounding cannot be completely excluded. The Lok index may in part reflect overall illness severity, coagulation abnormalities, and multiorgan dysfunction rather than isolated liver fibrosis. Accordingly, the observed associations should not be interpreted as causal. Finally, the Lok index was measured only at baseline, and the exploratory RCS and subgroup analyses need to be validated in independent prospective cohorts.
Conclusion
In summary, the Lok index is independently associated with an increased risk of ACM in critically ill patients with CAD. Rather than serving as a standalone prognostic instrument, the Lok index possibly reflects a composite burden of liver dysfunction, coagulation abnormalities, hematologic derangements, and systemic illness. Although the association is statistically significant, its immediate clinical utility remains limited. In the future, external validation studies are needed to clarify whether the Lok index is of incremental value for risk stratification. At present, the Lok index should not be used in isolation to guide clinical decisions or therapeutic interventions.
Supplementary Information
Supplementary Material 1: Supplementary Table 1. Multicollinearity Diagnostic Table. Supplementary Table 2. Comparison of Baseline Characteristics Between Included and Excluded Patients. Supplementary Table 3. Improvement in Risk Reclassification with the Addition of Lok to SOFA and SAPS II Scores: Net Reclassification Index and Integrated Discrimination Improvement. Supplementary Fig. 1. Residual Plot.
Acknowledgements
Not applicable.
Abbreviations
- CAD
Coronary artery disease
- ACM
All-cause mortality
- NAFLD
Non-alcoholic fatty liver disease
- INR
International normalized ratio
- AST
Aspartate aminotransferase
- ALT
Alanine aminotransferase
- ICU
Intensive care unit
- SQL
Structured Query Language
- DM
Diabetes mellitus
- RBC
Red blood cell
- WBC
White blood cell
- PLT
Platelet
- HGB
Hemoglobin
- FBG
Fibrinogen
- SAPS II
Simplified Acute Physiology Score II
- GCS
Glasgow Coma Scale
- SOFA
Sequential Organ Failure Assessment
- IQR
Interquartile range
- VIF
Variance inflation factor
- HR
Hazard ratio
- FBG
Fasting blood glucose
- RCS
Restricted cubic spline
Authors’ contributions
All authors contributed to the study conception and design. Writing - original draft preparation: Tao Yang; Writing - review and editing: Tao Yang, Shaning Yang; Conceptualization: Tao Yang; Methodology: Tao Yang; Formal analysis and investigation: Tao Yang; Resources: Tao Yang, Shaning Yang; Supervision: Tao Yang, Shaning Yang, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Data availability
The datasets analysed during the current study are available in the MIMIC-Ⅳ repository, https://physionet.org/content/mimiciv/3.1/.
Declarations
Ethics approval and consent to participate
This study was conducted under the ethical principles of the Declaration of Helsinki using the MIMIC-Ⅳ database. Its use was approved by the review boards of MIT and BIDMC, with a waiver of individual informed consent.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Tsao CW, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, Baker-Smith CM, Beaton AZ, Boehme AK, Buxton AE, et al. Heart Disease and Stroke Statistics-2023 Update: A Report From the American Heart Association. Circulation. 2023;147(8):e93–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fox KA, Fitzgerald G, Puymirat E, Huang W, Carruthers K, Simon T, Coste P, Monsegu J, Gabriel Steg P, Danchin N, et al. Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score. BMJ Open. 2014;4(2):e004425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Duell PB, Welty FK, Miller M, Chait A, Hammond G, Ahmad Z, Cohen DE, Horton JD, Pressman GS, Toth PP. Nonalcoholic Fatty Liver Disease and Cardiovascular Risk: A Scientific Statement From the American Heart Association. Arterioscler Thromb Vasc Biol. 2022;42(6):e168–85. [DOI] [PubMed] [Google Scholar]
- 4.Mantovani A, Csermely A, Petracca G, Beatrice G, Corey KE, Simon TG, Byrne CD, Targher G. Non-alcoholic fatty liver disease and risk of fatal and non-fatal cardiovascular events: an updated systematic review and meta-analysis. Lancet Gastroenterol Hepatol. 2021;6(11):903–13. [DOI] [PubMed] [Google Scholar]
- 5.Samsky MD, Patel CB, DeWald TA, Smith AD, Felker GM, Rogers JG, Hernandez AF. Cardiohepatic interactions in heart failure: an overview and clinical implications. J Am Coll Cardiol. 2013;61(24):2397–405. [DOI] [PubMed] [Google Scholar]
- 6.Samsky MD, Dunning A, DeVore AD, Schulte PJ, Starling RC, Tang WH, Armstrong PW, Ezekowitz JA, Butler J, McMurray JJ, et al. Liver function tests in patients with acute heart failure and associated outcomes: insights from ASCEND-HF. Eur J Heart Fail. 2016;18(4):424–32. [DOI] [PubMed] [Google Scholar]
- 7.Younossi ZM. Non-alcoholic fatty liver disease - A global public health perspective. J Hepatol. 2019;70(3):531–44. [DOI] [PubMed] [Google Scholar]
- 8.Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med. 2011;365(2):147–56. [DOI] [PubMed] [Google Scholar]
- 9.Møller S, Bernardi M. Interactions of the heart and the liver. Eur Heart J. 2013;34(36):2804–11. [DOI] [PubMed] [Google Scholar]
- 10.Udell JA, Wang CS, Tinmouth J, FitzGerald JM, Ayas NT, Simel DL, Schulzer M, Mak E, Yoshida EM. Does this patient with liver disease have cirrhosis? JAMA. 2012;307(8):832–42. [DOI] [PubMed] [Google Scholar]
- 11.Wai CT, Greenson JK, Fontana RJ, Kalbfleisch JD, Marrero JA, Conjeevaram HS, Lok AS. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology. 2003;38(2):518–26. [DOI] [PubMed] [Google Scholar]
- 12.Lok AS, Ghany MG, Goodman ZD, Wright EC, Everson GT, Sterling RK, Everhart JE, Lindsay KL, Bonkovsky HL, Di Bisceglie AM, et al. Predicting cirrhosis in patients with hepatitis C based on standard laboratory tests: results of the HALT-C cohort. Hepatology. 2005;42(2):282–92. [DOI] [PubMed] [Google Scholar]
- 13.Sterling RK, Lissen E, Clumeck N, Sola R, Correa MC, Montaner J, Torriani MSS, Dieterich FJ, Thomas DT. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology. 2006;43(6):1317–25. [DOI] [PubMed] [Google Scholar]
- 14.Iba T, Levy JH, Warkentin TE, Thachil J, van der Poll T, Levi M. Diagnosis and management of sepsis-induced coagulopathy and disseminated intravascular coagulation. J Thromb Haemost. 2019;17(11):1989–94. [DOI] [PubMed] [Google Scholar]
- 15.Helleberg J, Sundelin A, Soltani N, Rooyackers O, Mårtensson J. Predictive validity of daily sequential organ failure assessment (SOFA)-2 score for 30-day mortality. Crit Care. 2026;30(1):280. [DOI] [PMC free article] [PubMed]
- 16.Kramer L, Jordan B, Druml W, Bauer P, Metnitz PG. Incidence and prognosis of early hepatic dysfunction in critically ill patients–a prospective multicenter study. Crit Care Med. 2007;35(4):1099–104. [DOI] [PubMed] [Google Scholar]
- 17.Wolf D, Ley K. Immunity and Inflammation in Atherosclerosis. Circ Res. 2019;124(2):315–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Braunwald E. The path to an angiotensin receptor antagonist-neprilysin inhibitor in the treatment of heart failure. J Am Coll Cardiol. 2015;65(10):1029–41. [DOI] [PubMed] [Google Scholar]
- 19.North BJ, Sinclair DA. The intersection between aging and cardiovascular disease. Circ Res. 2012;110(8):1097–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Br VK, Sarin SK. Acute-on-chronic liver failure: Terminology, mechanisms and management. Clin Mol Hepatol. 2023;29(3):670–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Poelzl G, Ess M, Mussner-Seeber C, Pachinger O, Frick M, Ulmer H. Liver dysfunction in chronic heart failure: prevalence, characteristics and prognostic significance. Eur J Clin Invest. 2012;42(2):153–63. [DOI] [PubMed] [Google Scholar]
- 22.Fortea JI, Puente Á, Cuadrado A, Huelin P, Pellón R, González Sánchez FJ, Mayorga M, Cagigal ML, García Carrera I, Cobreros M, et al. Congestive Hepatopathy. Int J Mol Sci. 2020;21(24):9420. [DOI] [PMC free article] [PubMed]
- 23.Simon TG, Roelstraete B, Hagström H, Sundström J, Ludvigsson JF. Non-alcoholic fatty liver disease and incident major adverse cardiovascular events: results from a nationwide histology cohort. Gut. 2022;71(9):1867–75. [DOI] [PubMed] [Google Scholar]
- 24.Hagström H, Nasr P, Ekstedt M, Hammar U, Stål P, Hultcrantz R, Kechagias S. Fibrosis stage but not NASH predicts mortality and time to development of severe liver disease in biopsy-proven NAFLD. J Hepatol. 2017;67(6):1265–73. [DOI] [PubMed] [Google Scholar]
- 25.Ridker PM, Everett BM, Thuren T, MacFadyen JG, Chang WH, Ballantyne C, Fonseca F, Nicolau J, Koenig W, Anker SD, et al. Antiinflammatory Therapy with Canakinumab for Atherosclerotic Disease. N Engl J Med. 2017;377(12):1119–31. [DOI] [PubMed] [Google Scholar]
- 26.Libby P. The changing landscape of atherosclerosis. Nature. 2021;592(7855):524–33. [DOI] [PubMed] [Google Scholar]
- 27.Targher G, Byrne CD. Treatment algorithm in patients with type 2 diabetes and atherosclerotic cardiovascular disease or high/very high cardiovascular risk. Eur Heart J. 2020;41(2):331. [DOI] [PubMed] [Google Scholar]
- 28.Low Wang CC, Hess CN, Hiatt WR, Goldfine AB. Clinical Update: Cardiovascular Disease in Diabetes Mellitus: Atherosclerotic Cardiovascular Disease and Heart Failure in Type 2 Diabetes Mellitus - Mechanisms, Management, and Clinical Considerations. Circulation. 2016;133(24):2459–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Cecconi M, Evans L, Levy M, Rhodes A. Sepsis and septic shock. Lancet. 2018;392(10141):75–87. [DOI] [PubMed] [Google Scholar]
- 30.Ferrucci L, Fabbri E. Inflammageing: chronic inflammation in ageing, cardiovascular disease, and frailty. Nat Rev Cardiol. 2018;15(9):505–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1: Supplementary Table 1. Multicollinearity Diagnostic Table. Supplementary Table 2. Comparison of Baseline Characteristics Between Included and Excluded Patients. Supplementary Table 3. Improvement in Risk Reclassification with the Addition of Lok to SOFA and SAPS II Scores: Net Reclassification Index and Integrated Discrimination Improvement. Supplementary Fig. 1. Residual Plot.
Data Availability Statement
The datasets analysed during the current study are available in the MIMIC-Ⅳ repository, https://physionet.org/content/mimiciv/3.1/.