Abstract
Background
Acute kidney injury (AKI) is a common and life-threatening complication among critically ill patients with cirrhosis, yet early prediction remains challenging. The C-reactive protein-triglyceride-glucose index (CTI), a marker reflecting systemic inflammation and insulin resistance, may serve as a useful prognostic tool. However, its association with AKI in this population has not been previously investigated.
Methods
A cohort of patients was selected from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The primary outcome was the incidence of AKI during intensive care unit (ICU) stay. Secondary endpoints included the ICU length of stay and the hospital length of stay. Least Absolute Shrinkage and Selection Operator (LASSO) regression was used to identify key predictors of AKI. Logistic regression models were then applied to assess the association between CTI and the risk of AKI, treating CTI both as a continuous and categorical variable. A restricted cubic spline model was established to evaluate the dose–response relationship. Finally, subgroup analyses were performed across strata defined by the results of multivariate logistic regression analysis.
Results
A total of 887 ICU patients with cirrhosis were included. AKI exhibited an incidence of 67.5% and exhibited a positive association with increasing CTI quartiles (from 59.5% in Q1 to 73.3% in Q4; P = 0.009). Higher CTI levels were associated with more severe AKI stages (P = 0.015), greater renal replacement therapy (RRT) use (1.8% to 7.7%; P = 0.016), and longer ICU and hospital stays (P = 0.028 and P = 0.002, respectively). Multivariate logistic regression confirmed CTI as an independent predictor of AKI (OR = 1.26, 95% CI: 1.03–1.55; P = 0.027). A linear relationship between CTI and AKI was observed using restricted cubic spline models. Subgroup analyses showed consistent associations between AKI and CTI across different clinical strata.
Conclusion
CTI is independently associated with AKI risk, severity, and adverse outcomes in critically ill patients with cirrhosis. As a readily available marker with physiological relevance, CTI may aid in early risk stratification for AKI and guide preventive strategies in intensive care settings.
Clinical trial number
Not applicable.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12882-025-04613-2.
Keywords: C-reactive protein, Triglyceride-glucose index, Acute kidney injury, Cirrhosis, Critical care, Insulin resistance
Introduction
Cirrhosis remains a major global public health concern, often progressing to decompensated stages that necessitate admission to the intensive care unit (ICU) due to life-threatening complications [1, 2]. Among critically ill patients with cirrhosis, acute kidney injury (AKI) has emerged as one of the most prevalent and severe complications. The onset of AKI in this patient cohort is strongly correlated with increased in-hospital mortality [3], extended hospital stays, and an increased requirement for renal replacement therapy (RRT) [4–6]. Despite the complex interplay of systemic inflammation, hypoperfusion, and nephrotoxic exposures in the pathophysiology of AKI in cirrhosis, the early identification of patients at risk remains a significant clinical challenge. Traditional biomarkers, such as serum creatinine (Scr), exhibit limited ability in identifying AKI during its early and most treatable phase [7–9]. This underscores the necessity for more sensitive and reliable prognostic indicators in this high-risk population.
The C-reactive protein-triglyceride-glucose index (CTI) is an emerging biomarker that encapsulates the interaction between systemic inflammation and insulin resistance [10]. This index is derived from three readily accessible laboratory parameters: C-reactive protein, triglycerides, and glucose. CTI has been reported to be a surrogate marker for inflammatory-metabolic burden. CTI has shown considerable clinical utility in predicting both the onset and prognosis of cardiovascular diseases and various malignancies, underscoring its potential as a non-invasive, cost-effective tool for risk stratification [10–12]. In patients with cirrhosis, systemic inflammation and insulin resistance frequently coexist and play a significant role in the development of complications such as AKI. The pathogenesis of AKI in cirrhosis is multifactorial, encompassing systemic inflammation, hemodynamic disturbances, and metabolic dysfunction. Despite this complexity, there is a notable absence of reliable, integrated biomarkers for the early prediction of AKI in this context [9, 13, 14]. Given its capacity to capture two principal mechanisms implicated in AKI, CTI may serve as a practical and physiologically relevant tool for early risk assessment of AKI. Accordingly, investigating its prognostic utility in this high-risk population is both timely and clinically significant. Nonetheless, no studies have directly explored whether elevated CTI levels are associated with an increased risk of AKI or adverse renal outcomes in this specific setting.
In this study, we sought to address the existing knowledge gap by utilizing the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, a comprehensive and publicly accessible dataset focused on critical care, to assess the correlation between CTI and the risk of AKI among ICU patients with cirrhosis. Beyond evaluating the primary outcome of AKI incidence, we also explored the associations between CTI and clinically significant secondary outcomes, such as the duration of ICU and hospital stays.
Methods
Data source
This investigation employed a retrospective case-control design utilizing clinical data extracted from the MIMIC-IV (version 2.0) database, a comprehensive open-access repository containing anonymized medical records for approximately 60,000 intensive care unit admissions from 2008 to 2019, curated by MIT’s Laboratory for Computational Physiology [15]. The database encompasses multiple clinical parameters, including demographic characteristics, admission details, therapeutic interventions, pharmacological treatments, diagnostic codes (ICD-9/ICD-10), laboratory results, physiological measurements, fluid management records, clinical documentation, and patient outcomes.
Participants
Patients were considered eligible for inclusion in the study if they satisfied the following criteria: (1) aged 18 years or older at the time of admission to the ICU; and (2) confirmed diagnosis of cirrhosis, as identified by the International Classification of Diseases, Ninth Revision (ICD-9), or Tenth Revision (ICD-10) diagnostic codes. Exclusion criteria were as follows: (1) a diagnosis of AKI prior to ICU admission; or (2) Missing baseline laboratory parameters required for CTI calculation.
Data extraction
Data extraction was performed using Structured Query Language (SQL) query tools (version 1.13.1). Based on clinical relevance, current scientific understanding, and predictive factors identified in previous studies [16–18], the following parameters were systematically gathered and assessed within the first 24 h after ICU admission, prior to the occurrence of AKI.
Demographic data included age, sex, and race. Clinical comorbidities comprised type 2 diabetes mellitus (T2DM), hypertension, chronic kidney disease (CKD), chronic obstructive pulmonary disease (COPD), acute coronary syndrome (ACS), acute heart failure, acute pancreatitis, and the etiology of liver disease (alcoholic vs. nonalcoholic).
Laboratory data encompassed albumin (ALB), C-reactive protein (CRP), Fasting blood glucose (FBG), triglycerides, total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoprotein (HDL), hemoglobin (Hb), glycated hemoglobin A1c (HbA1c), serum creatinine, blood urea nitrogen (BUN), and white blood cell count (WBC).
Scoring indices included the Model for End-Stage Liver Disease (MELD) score and the Systemic Inflammatory Response Syndrome (SIRS) score. Therapeutic interventions were also recorded, including vasoactive drug use, ventilator support, aminoglycoside administration, and RRT. Outcome variables included the length of ICU stay and the total length of hospital stay.
Missing data handling
Several laboratory variables exhibited missing values, including blood urea nitrogen (11.3%), TC (12.5%), LDL (16.0%), HDL (16.5%), HbA1c% (25.6%), and Hb (11.2%). Assuming data were missing at random (MAR), multiple imputation by chained equations (MICE) was implemented to address the issue of missing data and mitigate potential bias using the ‘mice’ package in R [19]. The imputation model incorporated all clinically relevant variables, encompassing exposures, outcomes, and covariates employed in the final models. Five imputed datasets were generated, and the results of subsequent analyses were pooled using Rubin’s rules. In addition, a complete-case sensitivity analysis restricted to patients without missing values was conducted to assess the robustness of the findings. All descriptive statistics, regression analyses, and model estimations were conducted based on the multiply imputed, combined dataset.
Definitions and outcomes
The primary outcome of this study was the occurrence of AKI during ICU admission. AKI was defined and staged according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria [20], based on changes in serum creatinine. Urine output criteria were not applied due to the well-recognized limitations of urine data availability and reliability within the MIMIC-IV database. The secondary outcomes included: length of ICU stay, defined as the total number of days from ICU admission to ICU discharge, and length of hospital stay, defined as the number of days from hospital admission to hospital discharge, regardless of ICU duration. The administration of vasoactive drugs, ventilator support, and aminoglycosides was recorded as any instance of their use for any indication during the patient’s ICU stay. CTI was defined as 0.412* Ln (CRP [mg/L]) + Ln (Triglyceride [mg/dl] × FBG [mg/dl])/2 [10, 21].
Statistical analysis
All statistical analyses were conducted using R software (version 4.2.1). Continuous variables were reported as medians with interquartile ranges (IQRs), and categorical variables as counts and percentages. Group comparisons were performed using the Mann–Whitney U test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables, as appropriate. A two-sided P value < 0.05 was considered statistically significant for all analyses.
To explore the association between the CTI and the primary outcome of AKI, patients were stratified into quartiles (Q1–Q4) based on CTI values. The incidence of AKI and the proportion of patients requiring RRT were compared across CTI quartiles using the chi-square test. To assess for potential dose–response relationships, a trend test was conducted by modeling CTI quartiles as an ordinal variable in logistic regression. For secondary outcomes, including length of ICU stay and total length of hospital stay, comparisons across CTI quartiles were conducted using the Kruskal-Wallis test due to non-parametric data distribution.
To identify independent predictors of AKI, least absolute shrinkage and selection operator (LASSO) regression was applied with 10-fold cross-validation to select the most informative variables. These variables were then included in univariate and multivariate logistic regression models to estimate the odds ratio (OR) and 95% confidence intervals (CI) for AKI risk. CTI was evaluated both as a continuous variable and as a categorical variable by quartiles. Three multivariable models were constructed with progressively increasing levels of covariate adjustment. P values for trend across CTI quartiles were calculated from these models.
To examine whether the association between CTI and AKI followed a linear or non-linear pattern, restricted cubic spline (RCS) regression with four knots was performed. RCS is a flexible regression technique that models potential non-linear relationships by fitting piecewise cubic polynomials joined smoothly at prespecified knots, with the function constrained to be linear beyond the boundary knots [22]. This approach reduces instability at the extremes of the data and allows for a more accurate characterization of the dose–response relationship between CTI and AKI. Subgroup analyses were performed to evaluate the consistency of the CTI-AKI association across clinically relevant strata by the results of multivariate logistic regression analysis. Interaction terms were included in the multivariable logistic models, and P values for interaction were reported.
Results
A definitive study cohort comprising 887 patients was established and stratified into four groups based on CTI quartiles measured within the first 24 h of ICU admission (Fig. 1). The cohort had a median age of 60 years, with an interquartile range (IQR) of 52 to 67 years. Among the participants, 366 individuals (41.26%) were male. The median CTI value recorded was 5.85 (IQR: 5.79 to 5.93). AKI was identified in 599 patients (67.53%), distributed as follows: stage 1 (n = 105, 11.84%), stage 2 (n = 262, 29.54%), and stage 3 (n = 232, 26.16%). The median time from ICU admission to AKI onset was 2.6 days (IQR: 1.5–3.8 days). Stratification by CTI quartiles revealed significant differences in baseline characteristics (Table 1). The incidence of AKI increased progressively across quartiles, from 59.5% in Q1 to 66.1% in Q2, 71.3% in Q3, and 73.3% in Q4 (P = 0.009). Furthermore, CTI demonstrated a positive correlation with AKI severity (P = 0.015). Notably, patients with higher CTI values were more likely to require renal replacement therapy, with the proportion increasing from 1.8% in Q1 to 3.2% in Q2, 4.5% in Q3, and 7.7% in Q4 (P = 0.016).
Fig. 1.
Patient selection flowchart. Adult patients with cirrhosis admitted to the ICU were included, and baseline laboratory parameters for CTI calculation were obtained within 24 h of ICU admission. After applying exclusion criteria, 887 patients were included in the final analysis cohort and stratified into quartiles according to admission CTI values
Table 1.
Baseline characteristics of patients stratified by CTI quartiles (Q1–Q4)
| Variables | Q1, N = 222 | Q2, N = 222 | Q3, N = 222 | Q4, N = 221 | P value |
|---|---|---|---|---|---|
| Race, n (%) | |||||
| White | 152 (68.5%) | 159 (71.6%) | 142 (64.0%) | 146 (66.1%) | 0.300 |
| Other | 70 (31.5%) | 63 (28.4%) | 80 (36.0%) | 75 (33.9%) | |
| Gender, n (%) | |||||
| Male | 126 (57%) | 141 (64%) | 125 (56%) | 129 (58%) | 0.400 |
| Female | 96 (43%) | 81 (36%) | 97 (44%) | 92 (42%) | |
| Age, years | 60 (52, 67) | 60 (52, 67) | 59 (52, 67) | 59 (52, 67) | 0.910 |
| ALB, g/dL | 32 (27, 37) | 31 (26, 36) | 32 (27, 36) | 32 (27, 36) | 0.200 |
| CRP, mg/dL | 2 (1, 4) | 8 (5, 16) | 29 (15, 49) | 82 (48, 132) | < 0.001 |
| Glucose, mg/dL | 109 (90, 132) | 118 (99, 152) | 126 (101, 158) | 148 (115, 209) | < 0.001 |
| Triglycerides, mg/dL | 83 (60, 111) | 92 (67, 126) | 102 (72, 146) | 135 (95, 193) | < 0.001 |
| Serum creatinine, mg/dL | 0.90 (0.70, 1.38) | 1.10 (0.80, 1.60) | 1.30 (0.80, 1.80) | 1.30 (0.90, 2.10) | < 0.001 |
| Blood urea nitrogen, mg/dL | 20 (13, 31) | 25 (15, 43) | 27 (15, 46) | 29 (17, 50) | < 0.001 |
| White blood cell, K/uL | 6.9 (4.2, 10.4) | 7.3 (4.8, 10.5) | 7.2 (4.7, 10.9) | 7.7 (5.2, 13.5) | 0.015 |
| Hemoglobin, g/dL | 9.40 (8.30, 10.90) | 9.05 (7.90, 10.80) | 9.05 (8.03, 10.58) | 9.40 (8.10, 10.50) | 0.300 |
| HbA1c % | 5.50 (4.90, 6.08) | 5.60 (4.90, 6.40) | 5.60 (5.10, 6.68) | 5.90 (5.30, 7.20) | < 0.001 |
| HDL, mg/dL | 47 (31, 61) | 46 (30, 61) | 43 (28, 57) | 39 (27, 53) | 0.006 |
| LDL, mg/dL | 70 (47, 94) | 70 (48, 99) | 71 (56, 104) | 72 (50, 102) | 0.400 |
| TC, mg/dL | 143 (107, 173) | 137 (101, 172) | 139 (112, 176) | 141 (108, 175) | 0.700 |
| MELD score | 18 (12, 26) | 21 (15, 27) | 22 (15, 28) | 23 (16, 29) | < 0.001 |
| SIRS score | 2(2, 3) | 3(2, 3) | 3(2, 3) | 3(2, 3) | 0.912 |
| AKI, n (%) | |||||
| No | 90 (40.5%) | 75 (33.9%) | 64 (28.7%) | 59 (26.6%) | 0.009 |
| Yes | 132 (59.5%) | 147 (66.1%) | 158 (71.3%) | 162 (73.3%) | |
| Length of hospital stay, days | 8 (5, 15) | 11 (6, 21) | 9 (5, 21) | 11 (6, 23) | 0.002 |
| CKD, n (%) | |||||
| No | 213 (95.9%) | 205 (92.3%) | 203 (91.4%) | 210 (95.0%) | 0.200 |
| Yes | 9 (4.1%) | 17 (7.7%) | 19 (8.6%) | 11 (5.0%) | |
| COPD, n (%) | |||||
| No | 219 (98.6%) | 218 (98.2%) | 221 (99.5%) | 210 (95%) | 0.006 |
| Yes | 3 (1.4%) | 4 (1.8%) | 1 (0.5%) | 11 (5.0%) | |
| ACS, n (%) | |||||
| No | 216 (97.3%) | 211 (95.0%) | 216 (97.3%) | 211 (95.5%) | 0.500 |
| Yes | 6 (2.7%) | 11 (5.0%) | 6 (2.7%) | 10 (4.5%) | |
| T2DM, n (%) | |||||
| No | 213 (95.9%) | 207 (93.2%) | 209 (94.1%) | 187 (84.6%) | < 0.001 |
| Yes | 9 (4.1%) | 15 (6.8%) | 13 (5.9%) | 34 (15.4%) | |
| Hypertension, n (%) | |||||
| No | 198 (89.2%) | 200 (90.1%) | 198 (89.2%) | 190 (86.0%) | 0.500 |
| Yes | 24 (10.8%) | 22 (9.9%) | 24 (10.8%) | 31 (14.0%) | |
| Use of vasoactive drug, n (%) | |||||
| No | 151 (68.0%) | 150 (67.6%) | 142 (64.0%) | 135 (61.1%) | 0.400 |
| Yes | 71 (32.0%) | 72 (32.4%) | 80 (36.0%) | 86 (38.9%) | |
| Use of ventilator support, n (%) | |||||
| No | 68 (30.6%) | 45 (20.3%) | 46 (20.7%) | 40 (18.1%) | 0.008 |
| Yes | 154 (69.4%) | 177 (79.7%) | 176 (79.3%) | 181 (81.9%) | |
| Use of aminoglycosides, n (%) | |||||
| No | 203 (91.4%) | 202 (91%) | 193 (86.9%) | 179 (81.0%) | 0.002 |
| Yes | 19 (8.6%) | 20 (9.0%) | 29 (13.1%) | 42 (19.0%) | |
| Acute heart failure, n (%) | |||||
| No | 210 (94.6%) | 205 (92.3%) | 207 (93.2%) | 205 (92.8%) | 0.800 |
| Yes | 12 (5.4%) | 17 (7.7%) | 15 (6.8%) | 16 (7.2%) | |
| Acute pancreatitis, n (%) | |||||
| No | 218 (98.2%) | 221 (99.5%) | 219 (98.6%) | 216 (97.7%) | 0.400 |
| Yes | 4 (1.8%) | 1 (0.5%) | 3 (1.4%) | 5 (2.3%) | |
| RRT, n (%) | |||||
| No | 218 (98.2%) | 215 (97.8%) | 212 (95.5%) | 204 (92.3%) | 0.016 |
| Yes | 4 (1.8%) | 7 (3.2%) | 10 (4.5%) | 17 (7.7%) | |
| Basic liver diseases, n (%) | |||||
| Nonalcoholic | 134 (60.4%) | 129 (58.1%) | 123 (55.4%) | 134 (60.6%) | 0.700 |
| Alcoholic | 88 (39.6%) | 93 (41.9%) | 99 (44.6%) | 87 (39.4%) | |
AKI, acute kidney injury; ACS, acute coronary syndrome; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; ALB, albumin; CRP, C-reactive protein; TC, total cholesterol; LDL, low-density lipoprotein; HDL, high-density lipoprotein; CTI, C-reactive protein-triglyceride glucose index; T2DM, Type 2 diabetes; RRT, renal replacement therapy; Q1–Q4 represent increasing quartiles of the CTI
Baseline characteristics
As illustrated in Table 2, baseline demographic characteristics, including age, sex, and race, were comparable between the AKI and non-AKI groups. Nonetheless, patients who developed AKI demonstrated a significantly more compromised clinical profile. Inflammatory and metabolic markers, such as CRP, white blood cell count, and glucose levels, were notably elevated in the AKI group, whereas serum albumin levels were significantly lower (P = 0.001), indicating an enhanced inflammatory state and diminished nutritional status. Correspondingly, the CTI, the primary exposure of interest, was higher among AKI patients (median: 5.92 vs. 5.67; P < 0.001). Renal function indicators, including serum creatinine and blood urea nitrogen, were also significantly elevated (both P < 0.001) in the AKI cohort. Besides, AKI patients exhibited higher MELD and SIRS scores (both P < 0.001), experienced prolonged hospital stays and were more likely to receive intensive interventions such as vasoactive agents, ventilator support, aminoglycoside antibiotics, and renal replacement therapy (all P < 0.001). Comorbidities, including type 2 diabetes and hypertension, were more prevalent in the AKI group (P = 0.046 and P = 0.036, respectively).
Table 2.
Baseline characteristics of the AKI and Non-AKI groups
| Variables | Non-AKI (n = 289) | AKI (n = 599) | P value |
|---|---|---|---|
| Race, n (%) | |||
| White | 201 (69.8) | 398 (66.4) | 0.319 |
| Other | 87 (30.2) | 201 (33.6) | |
| Gender, n (%) | |||
| Male | 166 (57.6) | 355 (59.3) | 0.698 |
| Female | 122 (42.4) | 244 (40.7) | |
| Age, years | 60 (52, 66) | 60 (52, 68) | 0.658 |
| ALB, g/dL | 33.00 (28.00, 38.00) | 31.00 (27.00, 36.00) | 0.001 |
| CRP, mg/dL | 10.75 (3.08, 34.62) | 16.20 (5.30, 54.85) | 0.001 |
| Glucose, mg/dL | 116.50 (96.00, 158.50) | 125.00 (104.00, 162.00) | 0.029 |
| Triglycerides, mg/dL | 98.00 (69.00, 135.00) | 100.00 (73.50, 147.50) | 0.224 |
| Serum creatinine, mg/dL | 1.00 (0.70, 1.30) | 1.20 (0.80, 2.00) | < 0.001 |
| Blood urea nitrogen, mg/dL | 19.00 (12.00, 29.00) | 29.00 (16.00, 47.00) | < 0.001 |
| White blood cell, K/uL | 6.55 (4.30, 9.50) | 8.00 (4.90, 12.50) | < 0.001 |
| Hemoglobin, g/dL | 9.20 (8.10, 10.83) | 9.20 (8.00, 10.70) | 0.453 |
| HBA1C% | 5.60 (5.00, 6.82) | 5.70 (5.10, 6.50) | 0.481 |
| HDL, mg/dL | 45.50 (29.00, 60.00) | 42.00 (29.00, 57.00) | 0.129 |
| LDL, mg/dL | 72.50 (55.00, 99.25) | 70.00 (49.50, 101.00) | 0.324 |
| TC, mg/dL | 142.00 (113.00, 174.25) | 139.00 (105.00, 174.00) | 0.257 |
| MELD score | 16 (10, 23) | 23 (17, 29) | < 0.001 |
| SIRS score | 2 (2, 3) | 3 (2, 3) | < 0.001 |
| CTI | 5.67 (5.12, 6.26) | 5.92 (5.33, 6.42) | < 0.001 |
| Length of hospital stay, days | 5.80 (3.63, 9.68) | 13.00 (7.22, 24.88) | < 0.001 |
| CKD, n (%) | |||
| No | 275 (95.5) | 556 (92.8) | 0.167 |
| Yes | 13 (4.5) | 43 (7.2) | |
| COPD, n (%) | |||
| No | 283 (98.3) | 585 (97.7) | 0.74 |
| Yes | 5 (1.7) | 14 (2.3) | |
| ACS, n (%) | |||
| No | 281 (97.6) | 573 (95.7) | 0.223 |
| Yes | 7 (2.4) | 26 (4.3) | |
| T2DM, n (%) | |||
| No | 273 (94.8) | 543 (90.7) | 0.046 |
| Yes | 15 (5.2) | 56 (9.3) | |
| Hypertension, n (%) | |||
| No | 265 (92.0) | 521 (87.0) | 0.036 |
| Yes | 23 (8.0) | 78 (13.0) | |
| Use of vasoactive drug, n (%) | |||
| No | 242 (84.0) | 336 (56.1) | < 0.001 |
| Yes | 46 (16.0) | 263 (43.9) | |
| Use of ventilator support, n (%) | |||
| No | 108 (37.5) | 91 (15.2) | < 0.001 |
| Yes | 180 (62.5) | 508 (84.8) | |
| Use of aminoglycosides, n (%) | |||
| No | 283 (98.3) | 494 (82.5) | < 0.001 |
| Yes | 5 (1.7) | 105 (17.5) | |
| Acute heart failure, n (%) | |||
| No | 273 (94.8) | 554 (92.5) | 0.256 |
| Yes | 15 (5.2) | 45 (7.5) | |
| Acute pancreatitis, n (%) | |||
| No | 285 (99.0) | 589 (98.3) | 0.667 |
| Yes | 3 (1.0) | 10 (1.7) | |
| RRT, n (%) | |||
| No | 287 (99.7) | 562 (93.8) | < 0.001 |
| Yes | 1 (0.3) | 37 (6.2) | |
| Basic liver diseases, n (%) | |||
| Nonalcoholic | 172 (59.7) | 348 (58.1) | 0.698 |
| Alcoholic | 116 (40.3) | 251 (41.9) | |
AKI, acute kidney injury; ACS, acute coronary syndrome; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; ALB, albumin; CRP, C-reactive protein; TC, total cholesterol; LDL, low-density lipoprotein; HDL, high-density lipoprotein; CTI, C-reactive protein-triglyceride-glucose index; T2DM, Type 2 diabetes; RRT, renal replacement therapy
Primary endpoint
A total of 14 clinical variables were identified by LASSO regression as being associated with AKI (Fig. 2; Table 3). In subsequent univariate logistic regression analyses, CTI, SIRS score, albumin, type 2 diabetes mellitus, hypertension, use of vasoactive drugs, ventilator support, and aminoglycoside administration were significantly associated with AKI. In the multivariate logistic regression analysis (Table 3), the CTI (OR = 1.26, 95% CI: 1.03–1.55, P = 0.027) and SIRS score (OR = 1.24, 95% CI: 1.05–1.48, P = 0.014) remained significantly associated with AKI. Among the treatment-related variables, the use of vasoactive drugs (OR = 2.81, 95% CI: 1.94–4.11, P < 0.001), ventilator support (OR = 2.38, 95% CI: 1.68–3.38, P < 0.001), and aminoglycoside antibiotics (OR = 7.61, 95% CI: 3.32–22.05, P < 0.001) also retained statistical significance. In contrast, albumin, type 2 diabetes mellitus, and hypertension did not remain significant in the multivariate model.
Fig. 2.
LASSO regression for selecting variables associated with AKI. (A) This plot depicts how the coefficients of 28 candidate variables change with increasing regularization strength (log λ). As the penalty increases, less important variables shrink toward zero, allowing identification of the most relevant predictors of AKI. (B) Ten-fold cross-validation was used to determine the optimal λ value. The red dots represent the average binomial deviance (prediction error), and the vertical bars indicate standard error. The left dashed line marks the λ value that minimizes prediction error (λ_min), while the right dashed line marks the more conservative λ value (λ_1se) that offers a simpler model with comparable performance
Table 3.
Logistic regression analysis of the variables influencing the incidence of AKI among the study population
| Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|
| Variable | OR (95% CI) | P value | OR (95% CI) | P value | |
| Age | 1.00 (0.99–1.02) | 0.584 | Not applicable | ||
| ALB | 0.97 (0.95–0.99) | 0.003 | 0.98 (0.96–1.00) | 0.077 | |
| Triglycerides | 1.00 (1.00–1.00) | 0.075 | Not applicable | ||
| Serum creatinine | 1.62 (1.37–1.96) | 0.064 | Not applicable | ||
| Hemoglobin | 0.96 (0.90–1.03) | 0.284 | Not applicable | ||
| HBA1C% | 1.01 (0.93–1.10) | 0.844 | Not applicable | ||
| MELD score | 1.09 (1.07–1.11) | < 0.835 | Not applicable | ||
| SIRS score | 1.42 (1.21–1.66) | < 0.001 | 1.24 (1.05–1.48) | 0.014 | |
| CTI | 1.48 (1.23–1.79) | < 0.001 | 1.26 (1.03–1.55) | 0.027 | |
| T2DM | 1.88 (1.07–3.50) | 0.036 | 1.43 (0.73–2.90) | 0.311 | |
| Hypertension | |||||
| No | Reference | 0.029 | Reference | ||
| Yes | 1.72 (1.08–2.87) | 1.41 (0.82–2.48) | 0.220 | ||
| Use of vasoactive drug | |||||
| No | Reference | < 0.001 | Reference | ||
| Yes | 4.12 (2.91–5.92) | 2.81 (1.94–4.11) | < 0.001 | ||
| Use of ventilator support | |||||
| No | Reference | < 0.001 | Reference | ||
| Yes | 3.35 (2.42–4.65) | 2.38 (1.68–3.38) | < 0.001 | ||
| Use of aminoglycosides | |||||
| No | Reference | < 0.001 | Reference | ||
| Yes | 12.03 (5.37–34.36) | 7.61 (3.32–22.05) | < 0.001 | ||
ALB, albumin; CTI, C-reactive protein-triglyceride glucose index; T2DM, Type 2 diabetes; HBA1C, Hemoglobin A1c; MELD, Model for End-Stage Liver Disease; SIRS, Systemic Inflammatory Response Syndrome
A multivariable RCS model, fully adjusted for clinically relevant covariates, revealed a linear association between the CTI and the incidence of AKI in critically ill patients with cirrhosis (P for non-linearity = 0.482). The analysis indicated a positive association between the CTI and the incidence of AKI, as illustrated in Fig. 3.
Fig. 3.
The relationship between CTI and the risk of AKI is shown using a restricted cubic spline curve. The red line indicates how the risk of AKI changes with increasing CTI, and the shaded area shows the range of the 95% confidence interval. The dashed line at 1.0 represents the baseline risk used for comparison. The analysis did not find evidence of a non-linear relationship (P = 0.482), suggesting that the risk of AKI rises steadily as CTI increases
The results of the hierarchical multivariate logistic regression models are shown in Table 4. In Model 1 (adjusted for ALB and SIRS score), CTI was significantly associated with AKI. This association persisted after further adjustment for vasoactive drug use, ventilator support, and aminoglycoside exposure in Model 2, and remained statistically significant, though slightly attenuated, in the fully adjusted Model 3. Similar results were obtained in the complete-case sensitivity analysis (Table 4), with effect estimates closely aligning with those from the multiply imputed datasets, supporting the robustness of our findings. Specifically, after controlling for all clinical covariates, each 1-standard deviation (SD) increase in CTI (SD = 0.77) was associated with a 20% higher likelihood of developing AKI during ICU stay. (OR = 1.20, 95% CI: 1.02–1.40, P = 0.027). Moreover, when CTI was treated as a categorical variable divided into quartiles, with Q1 serving as the reference, the adjusted odds ratios for AKI in Model 3 were 1.18 (95% CI: 0.77–1.80) for Q2, 1.40 (95% CI: 0.90–2.18) for Q3, and 1.50 (95% CI: 0.98–2.31) for Q4. A significant dose–response relationship was observed across quartiles (Ptrend = 0.040).
Table 4.
Multivariable logistic regression models evaluating the association between CTI index and AKI
| Model 1 | Model 2 | Model 3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | Patients (n) | OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | ||
| CTI | 887 | 1.45(1.20–1.76)a | < 0.001 | 1.28(1.05–1.58) a | 0.017 | 1.26 (1.03–1.55) a | 0.027 | ||
| 1.48 (1.23–1.79)b | < 0.001 | 1.39 (1.13–1.71) b | 0.002 | 1.21 (0.97–1.51) b | < 0.001 | ||||
| Quartile of CTI | |||||||||
| Q1 | 222 | Reference | Reference | Reference | |||||
| Q2 | 221 | 1.26 (0.85–1.87) | 0.244 | 1.17 (0.77–1.79) | 0.458 | 1.18 (0.77–1.80) | 0.453 | ||
| Q3 | 223 | 1.67 (1.12–2.50) | 0.012 | 1.45 (0.94–2.26) | 0.043 | 1.40(0.90–2.18) | 0.044 | ||
| Q4 | 221 | 1.83 (1.22–2.74) | 0.004 | 1.50 (0.98–2.31) | 0.094 | 1.50 (0.98–2.31) | 0.137 | ||
| P value for trend | 0.001 | 0.048 | 0.040 | ||||||
Model 1: adjusted for ALB and SIRS score
Model 2: adjusted for Model 1 plus use of vasoactive drugs, ventilator support, and aminoglycosides
Model 3: adjusted for Model 2 plus type 2 diabetes mellitus and hypertension
a: Results based on multiple imputation with chained equations (m = 5) and pooled using Rubin’s rules
b: Results based on complete-case analysis, including only observations without missing data
Subgroup analysis
To further investigate the robustness of the relationship between the CTI and AKI, subgroup analyses were conducted across several clinically pertinent strata (Fig. 4). Overall, a higher CTI was significantly correlated with an increased risk of AKI (OR = 1.48, 95% CI: 1.23–1.78, P < 0.001). When stratified by the SIRS score, the association remained significant among patients with a SIRS score of 2 (OR = 1.42, 95% CI: 1.05–1.93, P = 0.024) and 3 (OR = 1.42, 95% CI: 1.04–1.95, P = 0.027), while it was not significant in other SIRS categories. Notably, among patients not receiving vasoactive drugs, the association persisted (OR = 1.35, 95% CI: 1.09–1.68, P = 0.007) and was more pronounced in those receiving vasoactive drugs (OR = 1.79, 95% CI: 1.17–2.80, P = 0.008). A similar trend was observed in relation to ventilator support: CTI was associated with AKI in both patients without (OR = 1.51, 95% CI: 1.07–2.16, P = 0.019) and with ventilator use (OR = 1.34, 95% CI: 1.06–1.69, P = 0.014).
Fig. 4.
Subgroup and interaction analysis between CTI and the risk of AKI. Subgroup analyses were conducted across various clinical characteristics, including SIRS scores, use of vasoactive drugs, ventilator support, and aminoglycoside exposure. The odds ratios (ORs) and 95% confidence intervals (CIs) for each subgroup are shown. The association between higher CTI and increased risk of AKI remained statistically significant in most subgroups. No significant interactions were observed (P for interaction > 0.05), indicating that the effect of CTI on AKI was generally consistent across different patient conditions
In patients not administered aminoglycosides, the association remained statistically significant (OR = 1.40, 95% CI: 1.15–1.70, P = 0.001). Conversely, no significant association was detected among patients who received aminoglycosides (OR = 1.04, 95% CI: 0.33–3.11, P = 0.949). Furthermore, no significant interactions were identified between CTI and any of the examined subgroups (all P interaction >0.05), suggesting that the relationship between elevated CTI and an increased risk of AKI was generally consistent across various clinical conditions, including the use of ventilator support, administration of vasoactive drugs, and differing SIRS scores.
Secondary outcome
Within the study cohort, patients categorized into higher CTI quartiles exhibited a tendency for prolonged ICU and hospital stays. Notably, the median ICU length of stay demonstrated a progressive increase across quartiles, increasing from 2.0 days in Q1 to 2.6 days in Q4 (P = 0.028). A similar pattern was observed for the median length of hospital stay, which increased from 8.2 days in Q1 to over 11 days in Q4 (P = 0.002).
Discussion
This study presents novel evidence indicating that the CTI serves as a significant and independent predictor of AKI in critically ill patients with cirrhosis. Our analysis demonstrated a consistent and linear association between elevated CTI levels and increased AKI risk, even after adjusting for key clinical and biochemical confounders. Importantly, this relationship was observed across various subgroups and exhibited a clear dose–response trend, suggesting that CTI reflects a cumulative burden of inflammatory and metabolic stress pertinent to renal vulnerability in this high-risk population. In addition to its association with AKI incidence, CTI was correlated with more severe clinical outcomes, including an increased need for RRT and extended ICU and hospital stays. Collectively, these findings underscore the potential utility of CTI not only as a risk stratification marker but also as a proxy indicator of disease trajectory in patients with cirrhosis admitted to the ICU.
The observed correlation between elevated CTI and the risk of AKI in critically ill patients with cirrhosis can be attributed to the intricate interplay of systemic inflammation, insulin resistance, and hemodynamic instability, all of which are well-recognized contributors to renal dysfunction in this demographic [23–25]. CTI, as a composite index, encompasses CRP, an indicator of the acute-phase inflammatory response, along with two metabolic parameters, glucose and triglycerides, that are intimately associated with insulin resistance. These components represent two fundamental pathophysiological axes in cirrhosis-associated AKI. First, systemic inflammation serves as a primary driver of AKI in cirrhosis. Elevated CRP levels signify increased cytokine release (e.g., IL-6, TNF-α), endothelial dysfunction, and the activation of both vasodilatory and vasoconstrictive mediators, collectively undermining effective renal perfusion [13, 26, 27]. Besides, insulin resistance, which is prevalent in both cirrhosis and critical illness, contributes to oxidative stress, endothelial damage, and impaired renal tubular function [28–30]. It is now understood that insulin resistance contributes to renal injury by promoting sympathetic nervous system and renin-angiotensin-aldosterone system activation, leading to elevated angiotensin II levels that cause systemic hypertension and intrarenal vasoconstriction, ultimately reducing renal perfusion [31]. This condition consequently lowers the threshold for AKI when additional stressors such as hypovolemia, infection, or nephrotoxic medications are present. Importantly, the incorporation of glucose and triglycerides into the CTI captures this metabolic susceptibility. When combined with the inflammatory burden, CTI provides a more comprehensive reflection of renal stress and vulnerability than single biomarkers. In cirrhotic patients, who often lack reliable indicators of renal reserve due to reduced muscle mass and altered creatinine kinetics [32], CTI may more accurately represent underlying subclinical risks that precede overt renal injury. Therefore, the ability of CTI to integrate both inflammatory and insulin resistance pathways may account for its strong and consistent association with AKI observed in this study, supporting its physiological plausibility as a predictive marker in cirrhotic critical care settings.
Although CTI has attracted significant attention in recent years, existing research has predominantly concentrated on its utility in predicting cardiovascular events, insulin resistance, and outcomes related to malignancies. For example, elevated CTI levels have been associated with an increased risk of myocardial infarction, adverse outcomes following percutaneous coronary intervention, and mortality in patients with solid tumors [11, 33, 34]. Collectively, these studies overlap in their assertion that CTI represents a reliable marker of systemic inflammatory and metabolic burden; however, its applicability has been largely restricted to relatively stable, non-critical populations. In contrast, our study provides hitherto undocumented evidence of the utility of CTI in critically ill patients with cirrhosis, a cohort marked by significant hemodynamic instability, immune dysregulation, and rapidly progressing multi-organ dysfunction. This distinction is critical, as AKI in this context arises not merely from chronic metabolic stress but from a complex interplay of systemic inflammation, portal hypertension-related circulatory collapse, and nephrotoxic exposures. This pathophysiological context is fundamentally different from that of previously studied populations. Furthermore, while previous research has often considered the CTI as a predictor of long-term outcomes, such as stroke incidence in a hypertensive population over several years [35], our findings underscore its acute predictive value in a high-risk ICU setting. In this environment, early identification of AKI risk may facilitate time-sensitive interventions.
In this context, CTI presents several practical advantages as a risk assessment tool. Indeed, CTI is derived from routinely available laboratory parameters, CRP, triglycerides, and glucose, which are cost-effective, rapidly obtainable, and widely accessible in ICU settings. This distinguishes the CTI index from more specialized renal biomarkers such as Neutrophil Gelatinase-Associated Lipocalin (NGAL) [36], Kidney Injury Molecule-1 (KIM-1) [37], or Interleukin-18 (IL-18) [38]. Although these novel markers have demonstrated potential in AKI prediction, their performance in cirrhotic and critically ill populations has been inconsistent, partly attributable to variations in AKI diagnostic criteria, patient selection, and assay methodologies, thereby complicating the establishment of reliable diagnostic thresholds. Furthermore, their levels may be affected by non-AKI conditions such as pulmonary infection, sepsis, or chronic kidney disease, which can confound interpretation and reduce specificity in practical settings. Consequently, while NGAL, KIM-1, IL-18, and related markers hold promise, their clinical application in the context of cirrhosis necessitates careful consideration [39]. By contrast, CTI leverages routine laboratory data and may therefore serve as a feasible and pragmatic adjunct for AKI risk stratification in clinical practice. Although our study did not directly compare CTI with these biomarkers, its simplicity and feasibility may facilitate broader clinical application, particularly in resource-constrained environments. Furthermore, by utilizing a large, high-quality critical care database and employing a comprehensive analytical approach, including multivariate adjustment, restricted cubic spline modeling, and subgroup validation, we provide robust evidence supporting a consistent, dose-dependent association between CTI and AKI in critically ill patients with cirrhosis. Through the application of a metabolic-inflammatory composite index to a cohort of critically ill patients with cirrhosis, this study broadens the clinical applicability of the CTI beyond diseases previously investigated and presents an innovative tool for renal risk stratification within the fields of hepatology and intensive care.
The findings of this study have several significant clinical implications. Importantly, the CTI emerges as a straightforward, accessible, and physiologically relevant tool for the early identification of patients with cirrhosis who are at an elevated risk for AKI upon admission to the ICU. Given that the three components of CTI, CRP, triglycerides, and glucose, are routinely measured in standard critical care panels, their integration into clinical workflows would incur no additional costs or require additional infrastructure, unlike more specialized renal biomarkers such as NGAL or cystatin C. This enhances its practicality as a point-of-care predictor, particularly in resource-constrained settings. Furthermore, our results indicate that CTI could play a crucial role in risk stratification and the development of individualized management strategies. Patients in the highest CTI quartiles exhibited not only a higher risk of AKI but also more severe AKI stages, an increased need for renal replacement therapy, and prolonged hospital stays. These findings suggest that CTI could be incorporated into early warning systems or AKI prevention protocols, prompting closer hemodynamic monitoring, early nephrology consultations, or more conservative use of nephrotoxic agents in high-risk patients. Anticipatory methodologies are essential in the context of cirrhosis, where conventional indicators such as serum creatinine frequently fail to accurately reflect early renal impairment due to diminished muscle mass and altered creatinine dynamics [40]. Furthermore, the integration of CTI-based alerts into the electronic health record systems of the ICU may facilitate automated, real-time risk profiling for AKI. This capability would be particularly beneficial for prioritizing interventions in resource-constrained ICUs or for informing the selection process in clinical trials focused on preventive therapies. Given its independent and dose-dependent correlation with adverse renal outcomes, CTI represents a promising adjunctive tool for AKI prediction; however, it warrants further validation in prospective studies before routine clinical implementation.
In addition, the robustness of the observed associations was supported by the complete-case sensitivity analysis (Table 4), which yielded effect estimates consistent with those obtained from the multiply imputed datasets. Nonetheless, the timing of sepsis onset relative to ICU admission and AKI events could not be reliably determined, and detailed data on contrast exposure were unavailable, precluding their evaluation in sensitivity analyses. Future studies with more comprehensive clinical data are warranted to explore these aspects.
This study is subject to several limitations that should be acknowledged. First, its retrospective design limits the ability to infer causality. Although multivariable adjustments were employed, the potential for residual confounding due to unmeasured variables remains considerable. Importantly, we were unable to account for the severity of portal hypertension (e.g., presence of varices, hepatic venous pressure gradient), detailed etiological subtypes of liver disease beyond the alcoholic versus non-alcoholic classification, or exposures to nephrotoxic agents other than aminoglycosides (e.g., NSAIDs, contrast media). These unmeasured factors may have influenced both CTI values and the risk of AKI. Besides, the reliance on a single-center U.S. database (MIMIC-IV) may constrain the external validity of our findings. The patient population in this database may differ from cirrhotic cohorts in other geographic regions, healthcare systems, and etiological backgrounds (e.g., higher prevalence of viral hepatitis in Asia). Therefore, the observed associations may not be directly generalizable to broader cirrhotic populations, and future validation in multicenter and ethnically diverse cohorts is warranted. Moreover, although subgroup analyses indicated that the association between CTI and AKI was generally consistent, the effect appeared to be attenuated among patients receiving aminoglycosides. Given the small sample size of this subgroup, the absence of statistical significance may reflect limited statistical power and the possibility of a type II error, suggesting that aminoglycoside exposure could potentially modify the CTI–AKI relationship. Finally, the present study did not investigate the biological mechanisms underlying the association between CTI and AKI. Indeed, future prospective studies that incorporate biomarker profiling, dynamic CTI monitoring, and larger subgroup samples are essential to elucidate the pathophysiological pathways, assess effect modification, and evaluate clinical utility. Despite these limitations, the findings of this study identify CTI as a promising, accessible, and clinically relevant marker for early AKI risk stratification in critically ill patients with cirrhosis, thereby supporting further research into its potential role in guiding preventive strategies and personalized care.
Conclusion
Overall, this study identifies the CTI as a novel and independent predictor of AKI in critically ill patients with cirrhosis, highlighting its potential as a simple and accessible tool for early risk stratification in intensive care settings.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Not applicable.
Abbreviations
- MIMIC-IV
Medical information mart for intensive care IV
- AKI
Acute kidney injury
- ACS
Acute coronary syndrome
- COPD
Chronic obstructive pulmonary disease
- CKD
Chronic kidney disease
- ALB
Albumin
- CRP
C-reactive protein
- TC
Total cholesterol
- LDL
Low-density lipoprotein
- HDL
High-density lipoprotein
- CTI
C-reactive protein triglyceride glucose index
- T2DM
Type 2 diabetes
- RRT
Renal replacement therapy
Author contributions
Study design: Lu-Huai Feng and Zhenhua Yang; data collection: Yang Lu, Lina Huang, Tingting Su and Tianbao Liao; manuscript preparation: Tingting Su and Tianbao Liao; data analysis and interpretation: Lu-Huai Feng, Tianbao Liao and Zhenhua Yang; all authors confirm that they contributed to manuscript reviews—revising it critically for important intellectual content—and read and approved the final draft for submission.
Funding
Not Applicable.
Data availability
The data analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The MIMIC-IV database was established with the approval of the institutional review boards of both the Beth Israel Deaconess Medical Center (Boston, MA) and the Massachusetts Institute of Technology (Cambridge, MA), with a waiver of informed consent due to the use of de-identified health data. As this study involved a secondary analysis of publicly available and fully anonymized data, no additional ethical approval and consent to participate was required. Access to the MIMIC-IV database was granted to the corresponding author (L-H F) upon successful completion of the required online training course provided by the National Institutes of Health (NIH) (certification number: 35897462). All methods were carried out in accordance with relevant guidelines and regulations, and the study was conducted in compliance with the principles of the Declaration of Helsinki.
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.
Lu-Huai Feng and Tianbao Liao contributed equally to this work and share first authorship.
References
- 1.Ma AT, Solé C, Juanola A, Escudé L, Napoleone L, Avitabile E, Pérez-Guasch M, Carol M, Pompili E, Gratacós-Ginés J, et al. Prospective validation of the EASL management algorithm for acute kidney injury in cirrhosis. J Hepatol. 2024;81(3):441–50. [DOI] [PubMed] [Google Scholar]
- 2.Live, EAftSot. EASL clinical practice guidelines for the management of patients with decompensated cirrhosis. J Hepatol. 2018;69(2):406–60. [DOI] [PubMed] [Google Scholar]
- 3.Desai AP, Knapp SM, Orman ES, Ghabril MS, Nephew LD, Anderson M, Ginès P, Chalasani NP, Patidar KR. Changing epidemiology and outcomes of acute kidney injury in hospitalized patients with cirrhosis - a US population-based study. J Hepatol. 2020;73(5):1092–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Warner NS, Cuthbert JA, Bhore R, Rockey DC. Acute kidney injury and chronic kidney disease in hospitalized patients with cirrhosis. J Invest Medicine: Official Publication Am Federation Clin Res. 2011;59(8):1244–51. [DOI] [PubMed] [Google Scholar]
- 5.Gupta K, Bhurwal A, Law C, Ventre S, Minacapelli CD, Kabaria S, Li Y, Tait C, Catalano C, Rustgi VK. Acute kidney injury and hepatorenal syndrome in cirrhosis. World J Gastroenterol. 2021;27(26):3984–4003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Patidar KR, Ma AT, Juanola A, Barone A, Incicco S, Kulkarni AV, Hernández JLP, Wentworth B, Asrani SK, Alessandria C, et al. Global epidemiology of acute kidney injury in hospitalised patients with decompensated cirrhosis: the international club of Ascites GLOBAL AKI prospective, multicentre, cohort study. Lancet Gastroenterol Hepatol. 2025;10(5):418–30. [DOI] [PubMed] [Google Scholar]
- 7.Lee HA, Seo YS. Current knowledge about biomarkers of acute kidney injury in liver cirrhosis. Clin Mol Hepatol. 2022;28(1):31–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Yewale RV, Ramakrishna BS. Novel biomarkers of acute kidney injury in chronic liver disease: where do we stand after a decade of research? Hepatol Research: Official J Japan Soc Hepatol. 2023;53(1):3–17. [DOI] [PubMed] [Google Scholar]
- 9.Girish V, Maiwall R. Tracking the trajectory of kidney dysfunction in cirrhosis: the acute kidney injury: chronic kidney disease spectrum. Clin Mol Hepatol. 2025;31(3):730–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ruan GT, Xie HL, Zhang HY, Liu CA, Ge YZ, Zhang Q, Wang ZW, Zhang X, Tang M, Song MM, et al. A novel inflammation and insulin resistance related indicator to predict the survival of patients with cancer. Front Endocrinol. 2022;13:905266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gao A, Peng B, Gao Y, Yang Z, Li Z, Guo T, Qiu H, Gao R. Evaluation and comparison of inflammatory and insulin resistance indicators on recurrent cardiovascular events in patients undergoing percutaneous coronary intervention: a single center retrospective observational study. Diabetol Metab Syndr. 2025;17(1):157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Xu M, Zhang L, Xu D, Shi W, Zhang W. Usefulness of C-reactive protein-triglyceride glucose index in detecting prevalent coronary heart disease: findings from the National health and nutrition examination survey 1999–2018. Front Cardiovasc Med. 2024;11:1485538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Clària J, Stauber RE, Coenraad MJ, Moreau R, Jalan R, Pavesi M, Amorós À, Titos E, Alcaraz-Quiles J, Oettl K, et al. Systemic inflammation in decompensated cirrhosis: characterization and role in acute-on-chronic liver failure. Hepatology (Baltimore MD). 2016;64(4):1249–64. [DOI] [PubMed] [Google Scholar]
- 14.Arroyo V, Angeli P, Moreau R, Jalan R, Clària J, Trebicka J, Fernández J, Gustot T, Caraceni P, Bernardi M. The systemic inflammation hypothesis: towards a new paradigm of acute decompensation and multiorgan failure in cirrhosis. J Hepatol. 2021;74(3):670–85. [DOI] [PubMed] [Google Scholar]
- 15.MIMIC Online Documentation. https://mimic.mit.edu.
- 16.Rajakumar A, Appuswamy E, Kaliamoorthy I, Rela M. Renal dysfunction in cirrhosis: critical care management. Indian J Crit Care Medicine: peer-reviewed Official Publication Indian Soc Crit Care Med. 2021;25(2):207–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Khemichian S, Francoz C, Nadim MK. Advances in management of hepatorenal syndrome. Curr Opin Nephrol Hypertens. 2021;30(5):501–6. [DOI] [PubMed] [Google Scholar]
- 18.Nadim MK, Durand F, Kellum JA, Levitsky J, O’Leary JG, Karvellas CJ, Bajaj JS, Davenport A, Jalan R, Angeli P, et al. Management of the critically ill patient with cirrhosis: A multidisciplinary perspective. J Hepatol. 2016;64(3):717–35. [DOI] [PubMed] [Google Scholar]
- 19.Zhang Z. Multiple imputation with multivariate imputation by chained equation (MICE) package. Annals Translational Med. 2016;4(2):30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kellum JA, Lameire N. Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care. 2013;17(1):204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Huo G, Tang Y, Liu Z, Cao J, Yao Z, Zhou D. Association between C-reactive protein-triglyceride glucose index and stroke risk in different glycemic status: insights from the China health and retirement longitudinal study (CHARLS). Cardiovasc Diabetol. 2025;24(1):142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med. 1989;8(5):551–61. [DOI] [PubMed] [Google Scholar]
- 23.Trebicka J, Fernandez J, Papp M, Caraceni P, Laleman W, Gambino C, Giovo I, Uschner FE, Jimenez C, Mookerjee R, et al. The PREDICT study uncovers three clinical courses of acutely decompensated cirrhosis that have distinct pathophysiology. J Hepatol. 2020;73(4):842–54. [DOI] [PubMed] [Google Scholar]
- 24.Bernardi M, Caraceni P. Novel perspectives in the management of decompensated cirrhosis. Nat Reviews Gastroenterol Hepatol. 2018;15(12):753–64. [DOI] [PubMed] [Google Scholar]
- 25.Parvathareddy VP, Wu J, Thomas SS. Insulin resistance and insulin handling in chronic kidney disease. Compr Physiol. 2023;13(4):5069–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gomez H, Ince C, De Backer D, Pickkers P, Payen D, Hotchkiss J, Kellum JA. A unified theory of sepsis-induced acute kidney injury: inflammation, microcirculatory dysfunction, bioenergetics, and the tubular cell adaptation to injury. Shock (Augusta Ga). 2014;41(1):3–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Engelmann C, Clària J, Szabo G, Bosch J, Bernardi M. Pathophysiology of decompensated cirrhosis: portal hypertension, circulatory dysfunction, inflammation, metabolism and mitochondrial dysfunction. J Hepatol. 2021;75(Suppl 1Suppl 1):S49–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ye Z, An S, Gao Y, Xie E, Zhao X, Guo Z, Li Y, Shen N, Zheng J. Association between the triglyceride glucose index and in-hospital and 1-year mortality in patients with chronic kidney disease and coronary artery disease in the intensive care unit. Cardiovasc Diabetol. 2023;22(1):110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Fritz J, Brozek W, Concin H, Nagel G, Kerschbaum J, Lhotta K, Ulmer H, Zitt E. The association of excess body weight with risk of ESKD is mediated through insulin Resistance, Hypertension, and hyperuricemia. J Am Soc Nephrology: JASN. 2022;33(7):1377–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jin Q, Liu T, Qiao Y, Liu D, Yang L, Mao H, Ma F, Wang Y, Peng L, Zhan Y. Oxidative stress and inflammation in diabetic nephropathy: role of polyphenols. Front Immunol. 2023;14:1185317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Yang Z, Gong H, Kan F, Ji N. Association between the triglyceride glucose (TyG) index and the risk of acute kidney injury in critically ill patients with heart failure: analysis of the MIMIC-IV database. Cardiovasc Diabetol. 2023;22(1):232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cullaro G, Kanduri SR, Velez JCQ. Acute kidney injury in patients with liver disease. Clin J Am Soc Nephrology: CJASN. 2022;17(11):1674–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ruan GT, Shi JY, Xie HL, Zhang HY, Zhao H, Liu XY, Ge YZ, Zhang XW, Yang M, Zhu LC, et al. Prognostic importance of an indicator related to systemic inflammation and insulin resistance in patients with Gastrointestinal cancer: a prospective study. Front Oncol. 2024;14:1394892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zhao DF. Value of C-Reactive Protein-Triglyceride glucose index in predicting cancer mortality in the general population: results from National health and nutrition examination survey. Nutr Cancer. 2023;75(10):1934–44. [DOI] [PubMed] [Google Scholar]
- 35.Tang S, Wang H, Li K, Chen Y, Zheng Q, Meng J, Chen X. C-reactive protein-triglyceride glucose index predicts stroke incidence in a hypertensive population: a National cohort study. Diabetol Metab Syndr. 2024;16(1):277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Huelin P, Solà E, Elia C, Solé C, Risso A, Moreira R, Carol M, Fabrellas N, Bassegoda O, Juanola A, et al. Neutrophil Gelatinase-Associated Lipocalin for assessment of acute kidney injury in cirrhosis: A prospective study. Hepatology. 2019;70(1):319–33. [DOI] [PubMed] [Google Scholar]
- 37.Brilland B, Boud’hors C, Wacrenier S, Blanchard S, Cayon J, Blanchet O, Piccoli GB, Henry N, Djema A, Coindre JP, et al. Kidney injury molecule 1 (KIM-1): a potential biomarker of acute kidney injury and tubulointerstitial injury in patients with ANCA-glomerulonephritis. Clin Kidney J. 2023;16(9):1521–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Qin Z, Li H, Jiao P, Jiang L, Geng J, Yang Q, Liao R, Su B. The value of urinary interleukin-18 in predicting acute kidney injury: a systematic review and meta-analysis. Ren Fail. 2022;44(1):1717–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bell M, Larsson A, Venge P, Bellomo R, Mårtensson J. Assessment of cell-cycle arrest biomarkers to predict early and delayed acute kidney injury. Dis Markers. 2015;2015:158658. [DOI] [PMC free article] [PubMed]
- 40.Nadim MK, Garcia-Tsao G. Acute kidney injury in patients with cirrhosis. N Engl J Med. 2023;388(8):733–45. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data analyzed during the current study are available from the corresponding author on reasonable request.