Abstract
Background
In various disease contexts, magnesium abnormalities are associated with acute kidney injury (AKI) incidence. However, this association remains unclear and has not been systematically investigated in patients with cirrhosis. Hence, we aimed to elucidate the association between admission serum magnesium levels and AKI incidence in intensive care unit (ICU)-admitted cirrhotic patients.
Methods
A retrospective cohort study was conducted using MIMIC-IV2.2 data, focusing on critically ill patients with cirrhosis. We employed univariable and multivariable logistic regression and restricted cubic spline analyses to robustly address our research objectives. To further substantiate the findings, subgroup and sensitivity analyses were also conducted.
Results
Among the 3,228 enrolled ICU-admitted cirrhotic patients, 34.4% were female, and the overall AKI incidence was 68.6% (2,213/3,228). Multivariable logistic regression analysis revealed an independent relationship between elevated serum magnesium levels and increased AKI risk (OR = 1.55 [95% CI = 1.15–2.09], p = 0.004). Compared with individuals with serum magnesium levels < 1.6 mg/dL, individuals with serum magnesium levels in Q2 (1.6–2.6 mg/dL) and Q3 (≥2.6 mg/dL) had adjusted ORs for AKI of 1.89 (95% CI = 1.34–2.65, p < 0.001) and 2.19 (95% CI = 1.27–3.75, p = 0.005), respectively. The restricted cubic spline analysis revealed that AKI risk increased linearly with increasing serum magnesium levels. Subgroup analysis revealed that the association between serum magnesium levels and AKI incidence was remarkably stable in subgroup analysis (all Pinteraction >0.05).
Conclusions
High serum magnesium concentrations were significantly associated with an increased AKI risk in ICU-admitted patients with cirrhosis. Further randomized trials are needed to confirm this association.
Keywords: Cirrhosis, acute kidney injury, magnesium, intensive care unit
Introduction
Acute kidney injury (AKI) is a prevalent complication in patients with cirrhosis, affecting up to 50% of hospitalized individuals with cirrhosis and 58% of such patients in the intensive care unit (ICU) [1–5]. AKI is associated with increased morbidity and mortality, resulting in a very unfavorable prognosis and an increased likelihood of developing chronic kidney disease (CKD) following liver transplantation [3,5–7]. Therefore, a comprehensive understanding of AKI and its associated risk factors may contribute to the prevention and management of AKI, thereby enhancing the prognosis and quality of life of patients with cirrhosis.
Magnesium ranks as the fourth most abundant cation in the body and plays a pivotal role in a multitude of physiological and biochemical processes, including enzyme activation, nucleic acid stability, and protein synthesis [8–10]. It also plays a role in regulating nerve and cardiac function, supporting mitochondrial function, and maintaining cytoskeletal integrity [11–13]. Additionally, it acts as a catalyst for more than 300 intracellular reactions, including neurotransmitter release, energy generation and intracellular calcium regulation [8,14]. Dysmagnesemia not only interferes with various physiological activities but also results in the progression of diseases, including renal dysfunction. Moreover, recent studies have demonstrated that serum magnesium abnormalities are associated with the risk of developing AKI in different patient populations [10,15–17]. However, the relationship between serum magnesium concentrations and the risk of developing AKI in critically ill cirrhotic patients admitted to the ICU remains controversial and poorly understood. Therefore, the objective of the present study was to identify the association between serum magnesium levels at admission and the risk of developing AKI in critically ill cirrhotic patients admitted to the ICU.
Materials and methods
Data sources
In the present retrospective cohort study, we used the Medical Information Mart for Intensive Care-IV (MIMIC-IV, version 2.2) database, a substantial, publicly accessible repository encompassing deidentified health-related data from patients admitted to critical care units at the Beth Israel Deaconess Medical Center from 2008 to 2019 [18,19]. Bingwen Lin, as one of the authors, obtained authorization to access the database (Record ID: 51729969). The review boards of the Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center granted approval for the use of the MIMIC-IV database. As the data were anonymized, written informed consent was waived, and we conformed to the guidelines outlined in the ‘Strengthening the Reporting of Observational Studies in Epidemiology’ [20] for observational research.
Study participants
A total of 5871 patients who were diagnosed with liver cirrhosis during their initial ICU admission were included in the study. The diagnosis of liver cirrhosis was based on the ninth and tenth versions of the International Classification of Diseases. Patients who were not admitted to the ICU and/or who were less than 18 years of age were subsequently excluded. Additionally, participants without serum magnesium data on their first day of admission and with an ICU stay of less than 24 h were excluded. Ultimately, 3228 patients were included in this study. The patients were categorized into three groups based on their serum magnesium levels at ICU admission (Figure 1).
Figure 1.
The flow chart of participants.
Data collection
Structured Query Language was used to extract baseline characteristics, including age, sex, ethnicity, vital signs, laboratory results, comorbidities and disease severity score. The data were collected within 24 h of admission to the ICU and included the use of ventilators, renal replacement therapy (RRT), vasoactive drugs, and loop diuretics, as well as the length of ICU and hospital stay. The use of diuretics, vasoactive agents and magnesium supplements was defined as receiving any loop diuretic, vasoactive agent or magnesium supplement within 7 days after ICU admission, respectively. Additionally, the use of ventilators and undergoing RRT were recorded if they occurred within 24 h after ICU admission.
Exposure
The primary exposure variable of the current analysis was the total serum magnesium level, which was assessed as a continuous variable and was also classified into hypoMg (<1.6 mg/dL), normalMg (1.6–2.6 mg/dL) and hyperMg (>2.6 mg/dL)categories based on the admission value in the MIMIC-IV database. Serum magnesium concentrations were defined as the first measured value within 24 h after admission to the ICU.
Variable definitions and outcomes
The incidence of AKI within 7 days after admission was the main outcome of this study. AKI identification and classification followed the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines, defining the condition as follows: an sCr increase of ≥0.3 mg/dL (26.5 µmol/L) within 48 h, sCr elevation to ≥1.5 times baseline in the past 7 days and that is known or presumed to have occurred within the previous 7 days, or urine output <0.5 mL/kg/hour over a 6-h period [1,21]. If baseline sCr was not documented before ICU entry, the first recorded sCr postadmission was used as the baseline reference.
Statistical analysis
The baseline characteristics of the participants were presented based on the different tertiles of serum magnesium levels. Normally distributed continuous variables are expressed as the mean ± SD, while skewed continuous variables are described as the median (interquartile range [IQR]). Categorical variables are presented as frequencies (%). Comparisons of continuous variables among groups were conducted using the independent-samples Student’s t test or Mann–Whitney U test, depending on the normality of the distribution. Categorical data were compared using the chi-square test or Fisher’s exact test, as appropriate. For all analyses, as the percentages of missing data varied from 0.03 to 33.4%, we simply replaced the missing values with the mean or median value for continuous variables and set 9 as the cutoff for categorical variables.
We employed multivariable logistic regression models to assess the independent association between serum magnesium concentrations and the risk of developing AKI, adjusting for major covariates. The odds ratio (OR) and 95% confidence interval (CI) were used for this analysis. Serum magnesium concentrations were evaluated as both a categorical and continuous variable. The selection of confounders was based on judgment, previous scientific literature, significant covariates in the univariate analysis, their associations with the outcomes of interest, or a change in effect estimate of more than 10%.
We constructed four models: (1) Model 1 was unadjusted; (2) Model 2 was adjusted for sex, age, race, MAP, white blood cell count, and calcium, potassium and glucose concentrations; (3) Model 3 was additionally adjusted for CKD, liver disease, Charlson’s Comorbidity Index, SOFA score, SAPS II and MELD-Na score; and (4) Model 4 was additionally adjusted for RRT status, ventilation status, diuretic use, vasoactive drug use and magnesium supplementation status.
To investigate the potential linear relationship between serum magnesium concentrations and the risk of developing AKI, we used restricted cubic spline (RCS) regression with 4 knots at the 5th, 35th, 65th, and 95th percentiles of serum magnesium after adjusting for variables in Model 4. To assess the consistency of the serum magnesium-AKI association across different patient profiles, we conducted subgroup analyses via stratified logistic regressions. These analyses were categorized by sex, age, MELD-Na score, CKD presence, sepsis, vasoactive agent use, and magnesium supplementation. Interaction effects between subgroups were statistically evaluated using likelihood ratio tests, ensuring a robust examination of the relationship’s stability across various clinical contexts.
We conducted several sensitivity analyses to test the robustness of our findings. First, considering the severe renal impairment of CKD individuals, which might introduce confounding effects when examining the association between serum magnesium levels and the risk of developing AKI, we reevaluated the association between serum magnesium concentrations and the risk of developing AKI by excluding patients with CKD, including those on chronic dialysis. Second, we performed additional analyses using only the serum creatinine level to categorize AKI, considering that changes in urine output are less frequently used for AKI diagnosis. Third, we employed multiple imputation instead of simple substitution to address missing values and reduce bias introduced by missing data. Finally, to mitigate the effects of hypoproteinemia, we corrected magnesium levels using the initial albumin concentration and assessed the association between corrected magnesium levels and the risk of developing AKI [22].
All analyses were performed using R Statistical Software (Version 4.2.2, http://www.R-project.org, The R Foundation) and the Free Statistics analysis platform (Version 1.9, Beijing, China, http://www.clinicalscientists.cn/freestatistics). A two-sided p value <0.05 was considered to indicate statistical significance.
Results
Baseline characteristics
After strict screening according to the inclusion and exclusion criteria, a total of 3228 patients with cirrhosis who were admitted to the ICU were deemed eligible for this study.
The overall prevalence of AKI was 68.6%. The baseline characteristics of the study cohorts categorized by serum magnesium levels are outlined in Table 1. The overall mean age was 60.1 ± 12.3 years, and 1111 (34.4%) patients were female. Patients with higher serum magnesium levels exhibited a greater propensity for increased age, white blood cell count, calcium concentrations, potassium concentrations, serum creatinine concentrations, urea nitrogen concentrations, total bilirubin concentrations, INR, and APTT than did those in the lower serum magnesium group. Moreover, they also demonstrated a greater prevalence of comorbidities such as congestive heart failure, CKD, hepatorenal syndrome, ascites and liver disease. Furthermore, there was a discernible ascending pattern from the lowest to the highest categories of serum magnesium levels for SAPSII, the Charlson Comorbidity Index, the SOFA score, and the MELD-Na score. Additionally, the use of RRT and vasoactive agents was more prevalent in the highest serum magnesium group.
Table 1.
Baseline characteristics of patients.
| Variables | Total (n = 3228) | Magnesium(mg/dL) |
p | ||
|---|---|---|---|---|---|
| HypoMg (<1.6mg/dL) | Normal-Mg (1.6–2.6mg/dL) | HyperMg (>2.6mg/dL) | |||
| Participants (n) | 3228 | 213 | 2803 | 212 | |
| Age, years | 60.1 ± 12.3 | 56.6 ± 11.8 | 60.3 ± 12.4 | 61.7 ± 11.7 | <0.001 |
| Gender/Female, n (%) | 1111 (34.4) | 81 (38) | 950 (33.9) | 80 (37.7) | 0.272 |
| Marital status, n (%) | 0.016 | ||||
| Married | 1284 (39.8) | 65 (30.5) | 1124 (40.1) | 95 (44.8) | |
| Single | 1176 (36.4) | 91 (42.7) | 1022 (36.5) | 63 (29.7) | |
| Others | 768 (23.8) | 57 (26.8) | 657 (23.4) | 54 (25.5) | |
| Race, n (%) | 0.303 | ||||
| White | 2180 (67.5) | 145 (68.1) | 1902 (67.9) | 133 (62.7) | |
| Other | 1048 (32.5) | 68 (31.9) | 901 (32.1) | 79 (37.3) | |
| Cause of cirrhosis, n (%) | <0.001 | ||||
| Alcoholic cirrhosis of liver | 1527 (47.3) | 141 (66.2) | 1301 (46.4) | 85 (40.1) | |
| Biliary cirrhosis | 80 (2.5) | 4 (1.9) | 66 (2.4) | 10 (4.7) | |
| Other causes | 1621 (50.2) | 68 (31.9) | 1436 (51.2) | 117 (55.2) | |
| Ascites, n (%) | 1606(49.8) | 83(39) | 1383(49.3) | 140(66) | <0.001 |
| Heart rate(bpm) | 86.9 ± 16.3 | 92.2 ± 15.7 | 86.7 ± 16.4 | 84.2 ± 15.1 | <0.001 |
| Respiratory rate(bpm) | 18.8 ± 4.0 | 18.7 ± 4.0 | 18.8 ± 4.0 | 18.6 ± 3.8 | 0.764 |
| MAP(mmHg) | 77.2 ± 11.5 | 81.1 ± 11.9 | 77.3 ± 11.5 | 72.6 ± 10.2 | <0.001 |
| SPO2 (%) | 96.8 ± 2.1 | 96.7 ± 2.8 | 96.9 ± 2.0 | 96.9 ± 2.0 | 0.717 |
| White blood cells (109/L) | 10.6 (7.1, 15.9) | 8.6 (5.9, 12.6) | 10.6 (7.1, 15.9) | 12.9 (7.8, 19.1) | <0.001 |
| Platelets (109/L) | 98.0 (59.0, 151.2) | 81.0 (54.0, 125.0) | 100.0 (60.0, 153.0) | 99.5 (54.0, 151.2) | 0.018 |
| Glucose (mg/dL) | 148.4 ± 88.6 | 138.0 ± 76.0 | 149.9 ± 91.8 | 139.2 ± 45.3 | 0.052 |
| Magnesium (mg/dL) | 2.0 ± 0.4 | 1.4 ± 0.1 | 2.0 ± 0.2 | 3.0 ± 0.9 | <0.001 |
| Calcium (mmol/L) | 8.0 ± 0.9 | 7.7 ± 0.8 | 8.0 ± 0.9 | 8.5 ± 1.0 | <0.001 |
| Sodium(mEq/L) | 138.4 ± 5.7 | 139.1 ± 4.4 | 138.3 ± 5.7 | 138.4 ± 7.2 | 0.114 |
| Potassium(mEq/L) | 4.6 ± 0.9 | 4.3 ± 0.8 | 4.6 ± 0.9 | 4.9 ± 1.0 | <0.001 |
| Creatinine(mg/dL) | 1.2 (0.8, 2.1) | 0.9 (0.7, 1.4) | 1.2 (0.8, 2.0) | 2.2 (1.5, 3.8) | <0.001 |
| Urea nitrogen(mg/dL) | 26.0 (16.0, 44.0) | 17.0 (11.0, 28.0) | 25.0 (16.0, 42.0) | 60.0 (30.0, 93.2) | <0.001 |
| Total bilirubin(mg/dL) | 3.2 (1.3, 5.8) | 2.6 (1.2, 5.8) | 3.1 (1.2, 5.8) | 5.8 (2.2, 18.0) | <0.001 |
| ALT(IU/L) | 44.0 (24.0, 145.0) | 37.0 (21.0, 99.0) | 44.0 (24.0, 149.5) | 50.5 (24.0, 150.0) | 0.036 |
| AST(IU/L) | 87.5 (44.0, 285.0) | 87.0 (46.0, 217.0) | 87.0 (44.0, 297.5) | 105.0 (47.5, 262.5) | 0.701 |
| INR | 1.9 ± 1.1 | 1.8 ± 0.9 | 1.9 ± 1.1 | 2.2 ± 1.1 | <0.001 |
| APTT(S) | 48.4 ± 28.1 | 45.4 ± 26.4 | 48.1 ± 28.0 | 55.1 ± 29.2 | <0.001 |
| Albumin(g/dL) | 3.0 ± 0.6 | 2.9 ± 0.6 | 3.0 ± 0.6 | 3.2 ± 0.7 | <0.001 |
| Congestive heart failure, n (%) | 622 (19.3) | 22 (10.3) | 544 (19.4) | 56 (26.4) | <0.001 |
| Chronic kidney disease, n (%) | 660 (20.4) | 26 (12.2) | 557 (19.9) | 77 (36.3) | <0.001 |
| Hepatorenal syndrome, n (%) | 503 (15.6) | 19 (8.9) | 416 (14.8) | 68(32.1) | <0.001 |
| Liver disease, n (%) | 2883 (89.3) | 188 (88.3) | 2501 (89.2) | 194 (91.5) | 0.512 |
| Hypertension, n (%) | 1639 (50.8) | 115(54) | 1446(51.6) | 78(36.8) | <0.001 |
| Sepsis, n (%) | 2150 (66.6) | 147 (69) | 1850 (66) | 153 (72.2) | 0.138 |
| AKI 7 day, n (%) | 2213 (68.6) | 107 (50.2) | 1928 (68.8) | 178 (84) | <0.001 |
| AKI stage 7 day, n (%) | <0.001 | ||||
| 1 | 386 (12.0) | 15 (7) | 340 (12.1) | 31 (14.6) | |
| 2 | 924 (28.6) | 55 (25.8) | 822 (29.3) | 47 (22.2) | |
| 3 | 903 (28.0) | 37 (17.4) | 766 (27.3) | 100 (47.2) | |
| SAPSII | 37.8 ± 13.9 | 31.2 ± 12.0 | 37.6 ± 13.6 | 47.5 ± 14.6 | <0.001 |
| CCI | 6.7 ± 2.9 | 5.7 ± 2.6 | 6.7 ± 2.9 | 7.3 ± 2.6 | <0.001 |
| Sofa score | 4.8 ± 2.2 | 4.4 ± 1.8 | 4.8 ± 2.2 | 6.0 ± 2.8 | <0.001 |
| MELD-Na | 22.8 ± 11.4 | 18.9 ± 8.9 | 22.6 ± 11.3 | 30.2 ± 12.1 | <0.001 |
| Diuretic, n (%) | 1686 (52.2) | 97 (45.5) | 1475 (52.6) | 114 (53.8) | 0.157 |
| Vasoactive agent, n (%) | 1153 (35.7) | 65 (30.5) | 981 (35) | 107 (50.5) | <0.001 |
| Magnesium supplementation, n (%) | 2425 (75.1) | 206 (96.7) | 2112 (75.3) | 107 (50.5) | <0.001 |
| RRT, n (%) | 248 (7.7) | 10 (4.7) | 201 (7.2) | 37 (17.5) | <0.001 |
| Ventilator, n (%) | 1543 (47.8) | 101 (47.4) | 1338 (47.7) | 104 (49.1) | 0.927 |
| Length of hospital stay, day | 8.9 (5.1, 17.1) | 6.6 (3.8, 12.6) | 8.8 (5.1, 17.0) | 12.9 (6.1, 22.0) | <0.001 |
| Length of ICU stay, day | 2.3 (1.3, 4.8) | 1.9 (1.1, 3.6) | 2.3 (1.3, 4.8) | 2.7 (1.8, 5.9) | <0.001 |
The mean standard deviation (SD). Median (IQR) for skewed variables, and numbers (proportions) for categorical variables.
Abbreviations: bpm: beats per minute; MAP: mean arterial pressure; CCI: Charlson Comorbidity Index; SAPSII: Simplified Acute Physiology Score II; AKI: acute kidney injury; ALT: alanine transaminase; AST: aspartate transaminase; INR: International Normalized Ratio; APTT: activated partial thromboplastin time; MELD-Na: model for end-stage liver disease-sodium; SOFA: Sequential Organ Failure Assessment; RRT: renal replacement treatment.
Univariable logistic regression analysis of the incidence rate of AKI
Serum magnesium concentrations were associated with the risk of developing AKI (Table 2). Additionally, univariable analysis demonstrated that age, heart rate, respiratory rate, MAP, SPO2, white blood cells, potassium, total bilirubin, serum creatinine, urea nitrogen, ALT, AST, INR, APTT, congestive heart failure, CKD, hepatorenal syndrome, liver disease, sepsis, CCI, SAPSII, SOFA score, MELD-Na score, diuretic, RRT, vasoactive agent and ventilator use were associated with the risk of developing AKI in patients with cirrhosis (p < 0.05).
Table 2.
Univariable analysis for incidences of AKI.
| Variable | OR 95%CI | p Value |
|---|---|---|
| Female | 1.002 (0.857–1.172) | 0.9784 |
| Age | 1.009 (1.003–1.015) | 0.0026 |
| Marriage | ||
| Married | Ref | |
| Single | 0.843 (0.712–0.998) | 0.0477 |
| Other | 1.095 (0.899–1.332) | 0.3668 |
| Race/ethnicity | ||
| White | Ref | |
| Others | 1.259 (1.071–1.4795) | 0.0052 |
| Heart rate | 1.009 (1.004–1.014) | <0.001 |
| Respiratory rate | 1.042 (1.022–1.062) | <0.001 |
| MAP | 0.969 (0.963–0.975) | <0.001 |
| SPO2 | 1.039 (1.003–1.076) | 0.0336 |
| White blood cells | 1.07 (1.057–1.083) | <0.001 |
| Platelets | 0.999 (0.9979–0.9996) | 0.0041 |
| Glucose | 0.999 (0.9985–1) | <0.001 |
| Magnesium | 3.178 (2.495–4.049) | <0.001 |
| Calcium | 0.87 (0.798–0.947) | 0.0013 |
| Sodium | 1.004 (0.991–1.017) | 0.5716 |
| Potassium | 1.274 (1.169–1.388) | <0.001 |
| Total bilirubin | 1.064 (1.169–1.388) | <0.001 |
| Serum creatinine | 1.645 (1.514–1.787) | <0.001 |
| Urea nitrogen | 1.02 (1.016–1.023) | <0.001 |
| ALT | 1.0005 (1.0002–1.0008) | <0.001 |
| AST | 1.0003 (1.0001–1.0004) | <0.001 |
| INR | 1.495 (1.341–1.666) | <0.001 |
| APTT | 1.017 (1.013–1.021) | <0.001 |
| Albumin | 0.666 (0.585–0.758) | <0.001 |
| Congestive heart failure | 1.396 (1.146–1.7) | <0.001 |
| Chronic kidney disease | 1.763 (1.443–2.154) | <0.001 |
| Hepatorenal syndrome | 2.254 (1.776–2.862) | <0.001 |
| Liver disease | 1.443 (1.146–1.817) | 0.0018 |
| Hypertension | 0.819 (0.706–0.951) | 0.009 |
| Sepsis | 3.019 (2.583–3.528) | <0.001 |
| CCI | 1.102 (1.072–1.133) | <0.001 |
| SAPSII | 1.082 (1.074–1.091) | <0.001 |
| Sofa score | 1.187 (1.141–1.235) | <0.001 |
| MELD-Na | 1.06 (1.05–1.07) | <0.001 |
| Diuretic | 2.305 (1.979–2.684) | <0.001 |
| Vasoactive agent | 5.384 (4.417–6.563) | <0.001 |
| Magnesium supplementation | 1.021 (0.86–1.212) | 0.8131 |
| RRT | 6.842 (4.157–11.263) | <0.001 |
| Ventilator | 4.426 (3.747–5.229) | <0.001 |
Abbreviations: MAP: mean arterial pressure; CCI: Charlson Comorbidity Index; SAPSII: simplified acute physiology scoreII; AKI: acute kidney injury; ALT: alanine transaminase; AST: aspartate transaminase; INR: International Normalized Ratio; APTT: activated partial thromboplastin time; MELD-Na: model for end-stage liver disease-sodium; SOFA: Sequential Organ Failure Assessment; RRT: renal replacement treatment.
Multivariable logistic regression analysis of the association between serum magnesium concentrations and the risk of developing AKI
After adjusting for potential confounding variables in the sample-weighted multivariable analyses (Table 3), a positive association was observed between the serum magnesium concentration and the risk of developing AKI, regardless of whether serum magnesium concentrations were analyzed as a continuous variable or categorical variable. As a continuous variable, higher serum magnesium concentrations were associated with an increased risk of developing AKI (unadjusted OR, 3.18; 95% CI, 2.49–4.05; p < 0.001; Table 3). According to the multivariable logistic regression model, this relationship remained significant, as a 1 mg/dL increase in the serum magnesium concentration was associated with a 55% greater risk of developing AKI (adjusted OR, 1.55; 95% CI, 1.15–2.09; p = 0.004; Table 3, Model 4). When considered a categorical variable, there was a clear upward trend toward a greater incidence of AKI among patients with increasing serum magnesium concentration tertiles (p for trend = 0.001).
Table 3.
Association between magnesium concentrations and acute kidney injury among liver cirrhotic patients.
| Variable | Model 1 | Model 2 | Model 3 | Model 4 | ||||
|---|---|---|---|---|---|---|---|---|
| OR (95%CI) | p Value | OR (95%CI) | p Value | OR (95%CI) | p Value | OR (95%CI) | s Value | |
| Magnesium (mg/dL) | 3.18 (2.49–4.05) | <0.001 | 2.51 (1.94–3.25) | <0.001 | 1.55 (1.17–2.04) | 0.002 | 1.55 (1.15–2.09) | 0.004 |
| Magnesium tertiles, mg/dL | ||||||||
| Q1(<1.6) | 1 (Ref) | 1 (Ref) | 1 (Ref) | 1 (Ref) | ||||
| Q2 (1.6 ≤ Mg < 2.6) | 2.18 (1.65–2.89) | <0.001 | 1.83 (1.36–2.46) | <0.001 | 1.63 (1.19–2.23) | 0.002 | 1.89 (1.34–2.65) | <0.001 |
| Q3(Mg ≥ 2.6) | 5.19 (3.29–8.17) | <0.001 | 3.6 (2.23–5.81) | 0.011 | 1.84 (1.1–3.07) | 0.02 | 2.19 (1.27–3.75) | 0.005 |
| p For trend | <0.001 | 0.002 | 0.005 | 0.001 | ||||
Abbreviations: OR: odds ratio; C: confidence interval; CCI: Charlson Comorbidity Index; SAPSII: Simplified Acute Physiology Score II; MELD-Na: model for end-stage liver disease-sodium; SOFA: Sequential Organ Failure Assessment; RRT: renal replacement treatment.
Model 1: No adjusted.
Model 2: Adjusted for gender, age, race, MAP, white blood cell, calcium, potassium, and glucose.
Model 3: Adjusted for model 2 plus chronic kidney disease, liver disease, Charlson’s Comorbidity Index, SOFA Score, SAPS II, and MELD-Na.
Model 4: Adjusted for model 3 plus RRT, ventilation, diuretic, vasoactive drug and magnesium supplementation.
Multivariable-adjusted regression analysis revealed that the odds ratios (95% CIs) for the risk of developing AKI were 1.89 (95% CI = 1.34–2.65; p < 0.001) for Q2 and 2.19 (95% CI = 1.27–3.75; p = 0.005) for Q3 of serum magnesium concentrations (Q1 as the reference; Table 3, Model 4).
RCS regression model
Using a RCS model, we showed that the risk of developing AKI increased linearly with increasing serum magnesium levels (Supplementary Figure 1).
Subgroup, sensitivity and additional analyses
Several sensitivity analyses were conducted to test the robustness of our main findings regarding the association between serum magnesium concentrations and the risk of developing AKI. The results presented in Supplementary Tables 1–4 demonstrated that the significant association between elevated serum magnesium concentrations and the risk of developing AKI persisted across all four sensitivity analyses.
We also performed stratified analyses according to sex, age, MELD-Na score, chronic renal disease status, sepsis status, vasoactive agent use and magnesium supplementation status (Figure 2). We found that the risk estimates were generally similar for incident AKI across subgroups (all p values for interactions >0.05).
Figure 2.
Subgroup analysis for the associations between serum magnesium concentrations and acute kidney injury. Each stratification adjusted for all confounders (Table 3, Model 4), except for the stratification factor itself.
Discussion
In this large population-based cohort study, we consistently observed a significant positive association between serum magnesium levels and the risk of developing AKI in a multivariable logistic regression analysis after adjustment for confounders. Similar patterns of association were observed for subsequent subgroup analysis.
The relationship between high serum magnesium concentrations and the risk of developing AKI has been a subject of discussion in the preceding academic literature. In a multicenter retrospective cohort investigation involving 6,124 adult patients subjected to cardiac surgery, higher early postoperative serum magnesium levels were shown to correlate with a heightened risk of postoperative AKI [23]. However, in pairs, another retrospective observational study delved into a cohort of 1,685 hospitalized patients diagnosed with COVID-19, revealing that patients exhibiting hypermagnesemia experienced a significantly greater prevalence of AKI (65 versus 50%; p < 0.001) and AKI necessitating continuous RRT (CRRT) (18 versus 5%; p < 0.001) relative to their nonhypermagnesemic counterparts [24]. Additionally, a retrospective observational study encompassing 3,669 critically ill children suggested a significant association between hypermagnesemia and increased AKI (OR = 1.52, 95% CI = 1.27–1.82, p < 0.001) [25]. Even so, studies on patients with cirrhosis are limited. In our current study, we observed that higher serum magnesium concentrations were associated with a significantly greater risk of developing AKI in patients with cirrhosis, which was consistent with previous studies.
In contrast, several alternative investigations with diverse sample populations have produced contrasting outcomes. For instance, a prospective cohort study conducted in three ICUs in Brazil, which included 7,042 critically ill patients, revealed that hypomagnesemia, but not hypermagnesemia, at ICU admission was associated with the risk of developing AKI (OR = 1.25; 95% CI = 1.08–1.46) [12]. Similarly, hypomagnesemia, but not hypermagnesemia, was associated with the risk of developing AKI in other kinds of patients, including those with acute pancreatitis, traumatic brain injury, or malignancy [10,26,27]. Interestingly, one study with a large sample size verified that both hypermagnesemia and hypomagnesemia could increase the risk of developing AKI in hospitalized patients in the general population [14]. We hypothesize that the inconsistency across the findings of these studies might partly stem from variations in study demographics, sample sizes, and the consideration or lack thereof of potentially confounding variables such as patient heterogeneity, necessitating further research to validate these assumptions.
In our research, we focused on individuals afflicted with cirrhosis, a condition predisposing them to a greater likelihood of developing electrolyte imbalances, which are typically characterized by reduced serum magnesium levels (i.e., hypomagnesemia). This decline can arise due to diminished magnesium intake, compromised fat absorption, the use of diuretics, renal tubular acidosis, and elevated serum concentrations of growth hormone and glucagon [28,29]. Nonetheless, in our present study cohort, elevated serum magnesium levels were prevalent, and we indeed observed a corresponding increase in the incidence of AKI concurrent with heightened serum magnesium concentrations. The definitive mechanism linking hypermagnesemia to an increased risk of developing AKI remains unclear. However, magnesium plays a pivotal role in maintaining cardiovascular homeostasis, serving to mitigate vasoconstriction induced by endogenous catecholamines and augmenting the efficacy of naturally occurring vasodilatory agents [30–32]. Abnormally high serum magnesium levels could plausibly disrupt renal hemodynamics and function and may lead to an imbalance in vascular tone homeostasis, culminating in an increased risk of developing AKI through overstimulated vasoconstriction and vasodilation responses [14]. Nonetheless, additional scientific inquiry is essential to fully expound upon the complex pathophysiological processes implicated in the relationship between serum magnesium levels and the occurrence of AKI. Further mechanistic studies, including animal models and controlled clinical trials, are crucial to decipher the exact pathways by which hypermagnesemia might contribute to AKI development.
Considering patients afflicted with acute liver failure or acute-on-chronic liver failure, particularly those presenting with concurrent hypermagnesemia, the incorporation of novel extracorporeal detoxification therapies, such as the Coupled Plasma Filtration and Adsorption (CPFA) system [33] or the Molecular Adsorbent Recirculating System (MARS) [34], may prove beneficial for hepatotoxin elimination and amelioration of hypermagnesemia, potentially preventing the development of AKI.
Our study had several strengths. First, it incorporated a relatively large, population-based cohort, providing ample statistical power to investigate the relationship between serum magnesium levels and the risk of developing AKI. Second, we conducted several sensitivity analyses to bolster the robustness of our findings. (1) By excluding participants with preexisting CKD, which could bias our results, we confirmed that the association between serum magnesium concentrations and the risk of developing AKI risk remained stable. (2) Our cohort study revealed an AKI incidence rate of 68.6%, a notably high figure possibly stemming from the combined application of both serum creatinine and urine output criteria for diagnosing AKI. According to a supplementary sensitivity analysis in which AKI was solely diagnosed based on serum creatinine levels, the incidence decreased significantly to 40.3%, which is consistent with prior literature [3,5]. Despite this adjustment, the results consistently supported our initial findings. (3) To minimize potential biases due to missing data, we employed an alternative method of managing missing values, and the results were consistent with our primary conclusions. (4) Considering that hypoproteinemia is prevalent among critically ill patients with cirrhosis, we further validated our results using albumin-calibrated serum magnesium measurements, ensuring that the impact of low protein levels on magnesium readings did not distort our findings. The results held strong even after this adjustment. Third, the findings have good potential for real-world applications. The current study, therefore, offers robust empirical evidence on the relationship between serum magnesium concentrations and the risk of developing AKI in patients with cirrhosis, taking meticulous account of potential confounding factors and biases. Hence, these results imply that monitoring serum magnesium levels may be a useful tool for identifying individuals at greater risk of developing AKI. Finally, our study demonstrated that the strength and consistency of the significant associations between serum magnesium levels and the incidence of AKI were noteworthy, persisting across various subgroups stratified by age, sex, MELD-Na score, the presence of CKD, sepsis, the use of vasoactive agents, and magnesium supplementation.
Nevertheless, several limitations merit consideration. First, as a retrospective study, our investigation is inherently limited in establishing causal relationships. Therefore, while we observed an association between serum magnesium levels and the risk of developing AKI, we cannot conclusively determine the direction of this relationship, including the possibility of reverse causality, where developing AKI might influence serum magnesium concentrations. Second, categorizing AKI into three main subdivisions based on its direct etiology—(a) prerenal or functional AKI resulting from a reduction in effective circulating volume, (b) AKI attributed to organic causes within the kidney, and (c) obstructive AKI due to urinary tract obstruction—facilitates a deeper understanding of its pathogenic mechanisms and enables more targeted guidance for clinical management and treatment. However, data about the causes of AKI and the parameters of circulating volume or renal blood flow were limited in the MIMIC-IV database. We suggest that future research incorporate these additional clinical variables to further elucidate the multifactorial nature of AKI, especially in the context of hypermagnesemia. Third, the findings are based on a dataset derived from adult patients in the United States within a single ICU; hence, generalizability to other populations worldwide necessitates corroborating evidence from future multicenter, prospective studies. Last, despite making adjustments for numerous potential confounders through maximizing sample size, unmeasured or residual confounders may still be present, inherent to retrospective designs. Notwithstanding these limitations, our data offer a substantial exploration of the association between serum magnesium levels and the risk of developing AKI, contributing valuable insights to the existing body of knowledge.
Conclusion
In patients with cirrhosis admitted to the ICU, the occurrence of AKI increased with increasing serum magnesium. Therefore, the measurement of serum magnesium concentrations may be helpful for risk stratification and early prediction of AKI in ICU cirrhotic patients. Further investigations are warranted to validate our findings and explore the detailed relationships and potential underlying mechanisms involved.
Supplementary Material
Acknowledgments
We are grateful to all of the participants for their valuable contributions. Additionally, we thank the Free Statistics team (Beijing, China) for providing technical assistance and practical data analysis and visualization tools.
Funding Statement
This work was supported by a grant from the Startup Fund for Scientific Research, Fujian Medical University [Grant number: 2021QH1088].
Ethical approval
Not applicable. This study was an analysis of de-identified datasets from a publicly available database.
Author contributions
BL and XX designed the research; BL extracted and analyzed the data; BL wrote the manuscript; PH and WX collated and interpreted the data; and WX and JL modified the manuscript and interpreted the analysis. XX reviewed the manuscript. All authors contributed to the article and approved the final submission.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The datasets generated and analyzed in this study can be obtained from the authors upon reasonable request.
References
- 1.Nadim MK, Garcia-Tsao G.. Acute kidney injury in patients with cirrhosis. N Engl J Med. 2023;388(8):733–745. doi: 10.1056/NEJMra2215289. [DOI] [PubMed] [Google Scholar]
- 2.Tandon P, James MT, Abraldes JG, et al. Relevance of new definitions to incidence and prognosis of acute kidney injury in hospitalized patients with cirrhosis: a retrospective population-based cohort study. PLOS One. 2016;11(8):e0160394. doi: 10.1371/journal.pone.0160394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Belcher JM, Garcia-Tsao G, Sanyal AJ, et al. Association of AKI with mortality and complications in hospitalized patients with cirrhosis. Hepatology. 2013;57(2):753–762. doi: 10.1002/hep.25735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Flamm SL, Wong F, Ahn J, et al. AGA clinical practice update on the evaluation and management of acute kidney injury in patients with cirrhosis: expert review. Clin Gastroenterol Hepatol. 2022;20(12):2707–2716. doi: 10.1016/j.cgh.2022.08.033. [DOI] [PubMed] [Google Scholar]
- 5.Desai AP, Knapp SM, Orman ES, et al. Changing epidemiology and outcomes of acute kidney injury in hospitalized patients with cirrhosis – a US population-based study. J Hepatol. 2020;73(5):1092–1099. doi: 10.1016/j.jhep.2020.04.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Amathieu R, Al-Khafaji A, Sileanu FE, et al. Significance of oliguria in critically ill patients with chronic liver disease. Hepatology. 2017;66(5):1592–1600. doi: 10.1002/hep.29303. [DOI] [PubMed] [Google Scholar]
- 7.Nadim MK, Genyk YS, Tokin C, et al. Impact of the etiology of acute kidney injury on outcomes following liver transplantation: acute tubular necrosis versus hepatorenal syndrome. Liver Transpl. 2012;18(5):539–548. doi: 10.1002/lt.23384. [DOI] [PubMed] [Google Scholar]
- 8.Altura BM. Basic biochemistry and physiology of magnesium: a brief review. Magnes Trace Elem. 1991;10(2–4):167–171. [PubMed] [Google Scholar]
- 9.Volpe SL. Magnesium in disease prevention and overall health. Adv Nutr. 2013;4(3):378S–383S. doi: 10.3945/an.112.003483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Liu Z, Wang R, He M, et al. Hypomagnesemia is associated with the acute kidney injury in traumatic brain injury patients: a pilot study. Brain Sci. 2023;13(4):593. doi: 10.3390/brainsci13040593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Khalili H, Rahmani H, Mohammadi M, et al. Intravenous magnesium sulfate for prevention of vancomycin plus piperacillin-tazobactam induced acute kidney injury in critically ill patients: an open-label, placebo-controlled, randomized clinical trial. Daru. 2021;29(2):341–351. doi: 10.1007/s40199-021-00411-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ribeiro HS, Burdmann EA, Vieira EA, et al. Association of magnesium abnormalities at intensive care unit admission with kidney outcomes and mortality: a prospective cohort study. Clin Exp Nephrol. 2022;26(10):997–1004. doi: 10.1007/s10157-022-02245-6. [DOI] [PubMed] [Google Scholar]
- 13.Saris NE, Mervaala E, Karppanen H, et al. Magnesium. An update on physiological, clinical and analytical aspects. Clin Chim Acta. 2000;294(1–2):1–26. doi: 10.1016/s0009-8981(99)00258-2. [DOI] [PubMed] [Google Scholar]
- 14.Cheungpasitporn W, Thongprayoon C, Erickson SB.. Admission hypomagnesemia and hypermagnesemia increase the risk of acute kidney injury. Ren Fail. 2015;37(7):1175–1179. doi: 10.3109/0886022x.2015.1057471. [DOI] [PubMed] [Google Scholar]
- 15.Santos MS, Seguro AC, Andrade L.. Hypomagnesemia is a risk factor for nonrecovery of renal function and mortality in AIDS patients with acute kidney injury. Braz J Med Biol Res. 2010;43(3):316–323. doi: 10.1590/s0100-879x2010007500002. [DOI] [PubMed] [Google Scholar]
- 16.Alves SC, Tomasi CD, Constantino L, et al. Hypomagnesemia as a risk factor for the non-recovery of the renal function in critically ill patients with acute kidney injury. Nephrol Dial Transplant. 2013;28(4):910–916. doi: 10.1093/ndt/gfs268. [DOI] [PubMed] [Google Scholar]
- 17.Naksuk N, Hu T, Krittanawong C, et al. Association of serum magnesium on mortality in patients admitted to the intensive cardiac care unit. Am J Med. 2017;130(2):229.e225–229.e213. doi: 10.1016/j.amjmed.2016.08.033. [DOI] [PubMed] [Google Scholar]
- 18.Johnson A, Bulgarelli L, Pollard T, et al. Mimic-IV (version 2.2). PhysioNet; 2020. [cited 2021 Aug 23]. Available from: https://physionet.org/content/mimiciv/1.0/.
- 19.Goldberger AL, Amaral LA, Glass L, et al. Physiobank, physiotoolkit, and physionet: components of a new research resource for complex physiologic signals. Circulation. 2000;101(23):E215–E220. doi: 10.1161/01.cir.101.23.e215. [DOI] [PubMed] [Google Scholar]
- 20.von Elm E, Altman DG, Egger M, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335(7624):806–808. doi: 10.1136/bmj.39335.541782.AD. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179–c184. doi: 10.1159/000339789. [DOI] [PubMed] [Google Scholar]
- 22.Kroll MH, Elin RJ.. Relationships between magnesium and protein concentrations in serum. Clin Chem. 1985;31(2):244–246. doi: 10.1093/clinchem/31.2.244. [DOI] [PubMed] [Google Scholar]
- 23.Xiong C, Shi S, Cao L, et al. Association of early postoperative serum magnesium with acute kidney injury after cardiac surgery. Ren Fail. 2023;45(1):2170244. doi: 10.1080/0886022x.2023.2170244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Stevens JS, Moses AA, Nickolas TL, et al. Increased mortality associated with hypermagnesemia in severe COVID-19 illness. Kidney360. 2021;2(7):1087–1094. doi: 10.34067/kid.0002592021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Morooka H, Tanaka A, Kasugai D, et al. Abnormal magnesium levels and their impact on death and acute kidney injury in critically ill children. Pediatr Nephrol. 2022;37(5):1157–1165. doi: 10.1007/s00467-021-05331-1. [DOI] [PubMed] [Google Scholar]
- 26.Yu XQ, Deng HB, Liu Y, et al. Serum magnesium level as a predictor of acute kidney injury in patients with acute pancreatitis. World J Clin Cases. 2021;9(35):10899–10908. doi: 10.12998/wjcc.v9.i35.10899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Shen D, Wang Y, Xu J, et al. The effect of admission serum magnesium on the acute kidney injury among patients with malignancy. Cancer Manag Res. 2020;12:7199–7207. doi: 10.2147/cmar.s262674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Cohen-Hagai K, Feldman D, Turani-Feldman T, et al. Magnesium deficiency and minimal hepatic encephalopathy among patients with compensated liver cirrhosis. Isr Med Assoc J. 2018;20(9):533–538. [PubMed] [Google Scholar]
- 29.Rocchi E, Borella P, Borghi A, et al. Zinc and magnesium in liver cirrhosis. Eur J Clin Invest. 1994;24(3):149–155. doi: 10.1111/j.1365-2362.1994.tb00980.x. [DOI] [PubMed] [Google Scholar]
- 30.Bäcklund M, Pere P, Lepäntalo M, et al. Effect of intra-aortic magnesium on renal function during and after abdominal aortic surgery: a pilot study. Acta Anaesthesiol Scand. 2000;44(5):605–611. doi: 10.1034/j.1399-6576.2000.00520.x. [DOI] [PubMed] [Google Scholar]
- 31.Shechter M, Sharir M, Labrador MJ, et al. Oral magnesium therapy improves endothelial function in patients with coronary artery disease. Circulation. 2000;102(19):2353–2358. doi: 10.1161/01.cir.102.19.2353. [DOI] [PubMed] [Google Scholar]
- 32.Altura BM, Altura BT.. New perspectives on the role of magnesium in the pathophysiology of the cardiovascular system. II. Experimental aspects. Magnesium. 1985;4(5–6):245–271. [PubMed] [Google Scholar]
- 33.Donati G, Angeletti A, Gasperoni L, et al. Detoxification of bilirubin and bile acids with intermittent coupled plasmafiltration and adsorption in liver failure (HERCOLE study). J Nephrol. 2021;34(1):77–88. doi: 10.1007/s40620-020-00799-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bañares R, Catalina MV, Vaquero J.. Molecular adsorbent recirculating system and bioartificial devices for liver failure. Clin Liver Dis. 2014;18(4):945–956. doi: 10.1016/j.cld.2014.07.011. [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 datasets generated and analyzed in this study can be obtained from the authors upon reasonable request.