Non-linear association between anion gap and 28-day mortality in critically ill patients with COVID-19: a cohort study from MIMIC IV database

Introduction

Global pandemics have profound effects on both economies and public health. The coronavirus disease 2019 (COVID-19) pandemic, which was caused by severe acute respiratory syndrome coronavirus 2 (SARS-COV-2), stands out as one of the most severe pandemics in recent history (1). The World Health Organization declared it a global pandemic on March 11, 2020, presenting an unparalleled global challenge with extensive ramifications for the global economy and society. As of October 13, 2024, the SARS-COV-2 pandemic has contributed to more than 776,618,091 cases and 7,071,324 deaths worldwide (2). Individuals infected with SARS-CoV-2 can exhibit a range of clinical manifestations, varying from asymptomatic infection to severe respiratory failure necessitating admission to the high-dependency unit or intensive care unit (ICU), and frequently resulting in fatalities (3). The primary cause for ICU admission of severe COVID-19 patients is severe respiratory failure. Nevertheless, critically ill COVID-19 patients frequently display metabolic irregularities that appear to impact their prognosis (4-6). Out of these metabolic abnormalities, it appears that metabolic acidosis is associated with a poorer prognosis (7,8).

The serum anion gap (AG) is a cost-effective and important parameter obtained from electrolyte measurements. It is extensively utilized in clinical practice to assess acid-base imbalances and to diagnose metabolic acidosis (9). It indicates the equilibrium of unmeasured anions and cations in the blood and is determined using the following formula (10):

Aniongap=sodium−(chloride+bicarbonate)[1]

Earlier studies have indicated that an increased AG may be associated with a worse outcome in critically ill patients (11,12). Many investigations have reported that an increased AG was linked to adverse outcomes in critically ill patients, including sepsis (13), influenza (14), heart failure (15), infective endocarditis (10), asthma (12), acute pancreatitis (16). However, its relationship with critically ill patients with COVID-19 has not been thoroughly explored.

This cohort study seeks to investigate the relationship between the serum AG and mortality in critically ill patients with COVID-19. We hypothesize that an increased AG is linked to a higher risk of mortality in this group. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-1964/rc).

Methods

Database

The data for this study were obtained from version 3.0 of the Medical Information Mart for Intensive Care IV, abbreviated as MIMIC IV database (https://mimic.mit.edu/) (17). The dataset includes information on 94,458 admissions of critically ill patients who were admitted to Beth Israel Deaconess Medical Center in Boston during the period from 2008 to 2022 (18). The database comprises a wide range of data elements, such as survival outcomes, vital signs, diagnostic information, laboratory analyses, and treatment methods. Due to the MIMIC database’s comprehensive and high-quality dataset as well as other advantages, it is increasingly being utilized by researchers for academic studies (19,20). One of the authors, Yingxiu Huang was granted access to the database after successfully completing an online training course and examination (Certificate ID: 56513391). As the dataset ensures patient privacy, consent for participation was not required. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

We performed a retrospective cohort study leveraging data from MIMIC IV version 3.0, encompassing all patients who were admitted to the ICU for the first time with a confirmed diagnosis of COVID-19 and whose sequence numbers were between 1 and 5. The gathered data comprised patient demographics, comorbidities, severity of illness scores, and laboratory metrics. The disease severity assessment was evaluated by Oxford acute severity of illness score (OASIS); Charlson Comorbidity Index (CCI). The comorbidities included sepsis, renal disease, diabetes, and sepsis. The treatment measures included antivirus drug use with 5 days of ICU admission, glucocorticoid use first day, ventilation first day of ICU stay. The primary endpoint of the study was mortality within 28 days.

Study population

Criteria

First, we included patients who had been diagnosed with COVID-19 with sequence numbers between 1 and 5 during their hospitalization from 2008 to 2022, and the criteria for diagnosis were according to the International Classification of Diseases (ICD-10), code = ‘U071’. Secondly, only the initial ICU admission record was chosen. Finally, we restricted our analysis to patients whose serum AG values were measured within 24 hours of their ICU admission.

Exposure

The primary variable of the study was the serum AG, considered as a continuous variable. Additionally, AG was categorized into three tertile group derived from admission values in the MIMIC IV database: T1 (≤11 mEq/L), T2 (12–15 mEq/L), and T3 (≥16 mEq/L). The baseline serum AG was established from the greatest value obtained within 24 hours of ICU admission.

Covariates

A structured query language was used to retrieve pertinent data from the MIMIC IV dataset, which was subsequently saved in PostgreSQL. Patient demographics, including age, sex, race, vital signs, ratio of arterial partial pressure of oxygen to fractional inspired oxygen (PaO2/FiO2), and laboratory markers, including the AG, potential of hydrogen (PH), creatinine, white blood cells (WBC), lactate, and international normalized ratio (INR), on the day of admission to ICU, were among the extracted data. Among the comorbidities were sepsis, diabetes, hypertension, and renal failure. The illness severity was including OASIS and CCI. The use of glucocorticoids, antiviral medications, and ventilation were part of the therapy plan.

Statistical analysis

According to the tertile distribution of serum AG, patients were divided into three groups. Summary statistics were used to describe the characteristics. For data with a normal distribution, continuous variables were presented as mean and standard deviation (SD); for data with a skewed distribution, they were presented as median and interquartile range. Proportions (percentages) were used to display categorical variables. The Chi-square test was used for categorical variables, while the Student’s t-test or Mann-Whitney U test was applied for continuous variables to compare data in order to examine baseline characteristics. The K-Nearest Neighbors imputation approach was used to variables with missing data rates less than 50% (21).

We applied multivariable Cox regression models to determine the hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between serum AG and mortality risk, with the goal of evaluating the independent link between serum AG and 28-day mortality. We used an expanded Cox model to generate models that were modified for different factors. The reference group was the lowest tertile of serum AG. Model 1: not modified. Model 2: Take race, sex, and age into account. Model 3: Modify for race, sex, age, hypertension, diabetes, sepsis, ventilation, lactate, PH, PaO2/FiO2, INR, CCI, antivirus drug use, and glucocorticoid use. Trend tests were conducted by incorporating the median value of each tertile as a continuous variable in the models. Restricted cubic splines (RCS) were utilized to investigate the non-linear relationship between AG and mortality. Additionally, survival curves were created using Kaplan-Meier analysis and compared using log-rank analysis. To determine whether the effect of serum AG levels on mortality varied across different subgroups and stratifications, a subgroup analysis was performed.

R statistical software version 4.3.2 (http://www.R-project.org, The R Foundation) and Free Statistics software version 2.0 were used for all studies (22). Statistical significance was defined as a two-tailed test with a significance level of P<0.05.

Results

Participants characteristics

A total of 811 individuals were chosen from the first cohort of critically sick COVID-19 patients in the MIMIC IV database following a stringent selection process guided by the inclusion and exclusion criteria (Figure 1). The final analysis comprised a total of 811 patients, with males constituting 59.6% of the sample. The average age among the participants was 64.1 years. Among the patients, 222 experienced non-survival, resulting in a 28-day death rate of 27.4%. Table 1 lists each participant’s specific baseline characteristics.

Figure 1 Flowchart of the study. COVID-19, coronavirus disease 2019; ICU, intensive care unit; MIMIC IV, medical information mart for intensive care IV.

Non-linear association between mortality and serum AG

To explore the link between serum AG levels and 28-day outcome in patients with COVID-19, we used three models: Model 1 (unadjusted), Model 2 (adjusted for age, sex, and race), and Model 3 (adjusted for age, race, sex, hypertension, diabetes, sepsis, ventilation, lactate, PH, PaO2/FiO2, INR, CCI, antivirus drug use, glucocorticoid use). In the crude model, each unit increase in AG was associated with a 12% increase in mortality (HR =1.12, 95% CI: 1.09–1.15; Model 1) (see Table 2). In Model 3, a strong association between 28-day mortality and AG persisted after adjustment for potential confounders (HR =1.07, 95% CI: 1.04–1.11, P<0.001). Subsequently, The T3 group’s risk of 28-day mortality was noticeably higher than the T1 group’s when AG was divided into tertiles (HR =1.48, 95% CI: 1.01–2.17, P=0.042, Model 3). The p-values for trend test in all three models were less than 0.05.

After adjusting for potential confounders, the RCS analysis demonstrated a nonlinear relationship between AG and mortality (P for non-linearity =0.006) (Figure 2). We utilized a two-piecewise model to examine the relationship between 28-day mortality and AG. Our findings identified an inflection point at roughly 18 mEq/L (Table 3). In the threshold analysis, the HR for mortality in subjects with serum AG less than 18 was 1.009 (95% CI: 0.954–1.067, P=0.76) (Table 3), indicating that AG was not significantly associated with mortality risk at this level. Conversely, a positive association of AG with mortality was observed when AG was 18 or higher, with a HR of 1.191 (95% CI: 1.098–1.293, P<0.001) (Table 3). This suggests that each 1 mEq/L an elevation in serum AG is linked to a 19.1% increase in mortality risk when AG ≥18 mEq/L.

Figure 2 Relationship between serum AG and 28-day mortality. Solid and dashed red lines represent the predicted value and 95% confidence intervals, and the horizontal dashed line marks the null line (HR =1). The blue area density plot shows the distribution of patient proportions. Adjusted for age, race, sex, ventilation, PO2, MBP, diabetes, hypertension, sepsis, INR, CCI, OASIS, antivirus drug use, glucocorticoid use. Only 99.5% of data was displayed. AG, anion gap; CCI, Charlson Comorbidity Index; HR, hazard ratio; INR, international normalized ratio; MBP, mean blood pressure; OASIS, Oxford acute severity of illness score; PO2, partial pressure of oxygen.

Kaplan-Meier survival curve analysis

The Kaplan-Meier curve illustrated that the 28-day cumulative survival rates were significantly lower in the T3 group than in both the T1 and T2 groups, suggesting a clear disparity in survival outcomes among the tertiles (P<0.001) (Figure 3).

Figure 3 Kaplan-Meier survival curves for critically ill patients with COVID-19 based on serum AG tertile. AG, anion gap; COVID-19, coronavirus disease 2019.

Subgroup analyses

We conducted an subgroup analysis to examine the strength of the correlation between serum AG levels and mortality, which was stratified by age (<65, and ≥65 years), gender, antiviral treatment, renal disease, and diabetes. The results of this analysis are presented within Figure 4’s forest plot. The findings showed that there was a consistent correlation between blood AG levels and death in a range of individuals of different ages, antiviral treatments, renal disease status, and diabetes. According to the result of interaction analysis, no interaction existed between serum AG and subgroups (P>0.05).

Figure 4 Subgroup analyses for the association of serum AG with 28-day mortality in the critically ill patients with COVID-19. Adjusted for age, race, sex, ventilation, PO2, MBP, diabetes, hypertension, sepsis, INR, CCI, OASIS, antivirus drug use, glucocorticoid use. AG, anion gap; CCI, Charlson Comorbidity Index; CI, confidence interval; COVID-19, coronavirus disease 2019; HR, hazard ratio; INR, international normalized ratio; MBP, mean blood pressure; OASIS, Oxford acute severity of illness score; PO2, partial pressure of oxygen.

Discussion

In severely ill COVID-19 patients admitted to the ICU, this retrospective cohort analysis found a strong link between great serum AG levels and an increased likelihood of mortality. After controlling for confounding variables, the findings from the multivariate Cox proportional hazards regression analysis indicate that elevated serum AG levels remain independently associated with an increased risk of mortality with 28 days in critically sick COVID-19 patients. Furthermore, compared to the lowest AG group, the highest AG group showed a 1.51-fold higher risk of all-cause mortality. Notably, this is the first study to look into the connection between critically ill patients with COVID-19 and serum AG levels.

Serum AG serves as an economical and effective tool for distinguishing among various acid-base disorders, particularly metabolic acidosis. The advancement of automated analyzers has facilitated easier access to serum AG, facilitating electrolyte testing in sizable patient groups, especially those with severe illnesses (23). Therefore, an increased AG may have predictive value for numerous disorders in subjects who were in critical condition. Some prior studies have indicated that serum AG showed a positive association with outcomes in other critical conditions, such as sepsis (13), chronic obstructive pulmonary disease (COPD) (24), influenza (14), asthma (25), infective endocarditis (10), acute pancreatitis (26), and cirrhosis (27). Serum AG serves as a more general and reliable risk factor for critically ill condition (9), according to the data, including the results of the current investigation. This underscores the clinical relevance of this readily available prognostic indicator. Elevated AG was found to be significantly associated with higher fatality rates and longer hospital stays in a cohort study that included 500 severely ill patients in ICU (28). Likewise, in a multicenter research, Li et al. (11) reported that among critically ill patients, serum AG of ≥16 mmol/L following hospitalization in the ICU was linked to a higher fatality rate with a large sample. Our research revealed that in severely ill patients with COVID-19, an elevated AG was associated with an increased risk of 28-day mortality, consistent with previous research. This highlights the need for clinical interventions to reduce mortality in this group. Additionally, we observed significant differences in mortality rates among the AG groups, reinforcing the resilience of our results.

PH and lactate levels are critical markers of metabolic and respiratory acid-base balance in critically ill patients (29,30). Their inclusion in the model helps account for the physiological stress and tissue hypoperfusion associated with COVID-19, which may confound the relationship between the AG and mortality (31). Despite the inclusion of additional covariates, the association between the AG and 28-day mortality remained significant, indicating that the relationship is robust. Future studies could explore the combined effects of the AG, PH, and lactate levels in predicting clinical outcomes in COVID-19 patients.

Increased AG levels are frequently seen in severe patients with COVID-19, and this can happen for certain causes. First, an increased serum AG may be the consequence of metabolic acidosis. When metabolic acidosis occurs, the body produces too much lactate, resulting in a rise of anions in the bloodstream, consequently, an elevated AG (32). Second, elevated AG levels may also result from the buildup of other acids in the extracellular fluid, including acetoacetic acid and beta-hydroxybutyric acid. Third, severe individuals with COVID-19 are susceptible to acute renal damage, which can hinder the function of the kidneys and decrease the excretion of unmeasured anions, contributing to elevated AG levels (33). Thus, renal insufficiency may possibly be the cause of the correlation between AG and death (34).

There are various advantages in the current study. First, using data from the large-scale, real-world dataset of the MIMIC IV database, renowned for its high-quality data, the research first benefits from a comparatively large sample size. Second, subgroup and sensitivity analyses were also included in the study, which adds to the findings’ resilience. Nevertheless, our study is subject to several limitations. First, the selection bias may be present due to the retrospective nature of the analysis. Second, we extracted AG data solely for patients at the time of their admission to the ICU. Consequently, we failed to evaluate patterns in AG changes, which might have impacted the accuracy of our results. Third, since this was an observational study, we were unable to verify the underlying mechanism that was thought to relate a higher AG in relation to mortality of COVID-19.

Conclusions

According to this study, among severely sick patients with COVID-19, a higher serum AG at admission is independently associated with elevated 28-day mortality. To confirm these results and look into the mechanisms that may explain them, more research is required.

Acknowledgments

The authors thanks Dr. Qilin Yang of Department of Critical Care, The Second Affiliated Hospital of Guangzhou Medical University and the Physician Scientist Team for guidance on data extraction and analysis.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-1964/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-1964/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-1964/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. One of the authors (Yingxiu Huang) has passed the “Protecting Human Research Participants” examination, could access the database and was responsible for data extraction (Certificate ID: 56513391). MIMIC IV database used in the present study was approved by both Institutional Review Boards (IRB) of Beth Israel Deaconess Medical Center (2001-P-001699/14) and the Massachusetts Institute of Technology (No. 0403000206). The individual information of the patients included in this database was anonymous, and ethical review and informed consent were waived. The authors have also complied with all relevant ethical regulations regarding the use of the data for the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Li X, Wang W, Zhao X, et al. Transmission dynamics and evolutionary history of 2019‐nCoV. J Med Virol 2020;92:501-11. [Crossref] [PubMed]
2. COVID-19 deaths | WHO COVID-19 dashboard. datadot. (cited 2024 Oct 28). Available online: https://data.who.int/dashboards/covid19/cases
3. Chung YS, Lam CY, Tan PH, et al. Comprehensive Review of COVID-19: Epidemiology, Pathogenesis, Advancement in Diagnostic and Detection Techniques, and Post-Pandemic Treatment Strategies. Int J Mol Sci 2024;25:8155. [Crossref] [PubMed]
4. Sanghani H, Bansal S, Parmar V, et al. Study of Arterial Blood Gas Analysis in Moderate-to-Severe COVID-19 Patients. Cureus 2022;14:e26715. [Crossref] [PubMed]
5. Sotiropoulou Z, Antonogiannaki EM, Koukaki E, et al. Evaluation of the Acid-Base Status in Patients Admitted to the ICU Due to Severe COVID-19: Physicochemical versus Traditional Approaches. J Pers Med 2023;13:1700. [Crossref] [PubMed]
6. Zemlin AE, Sigwadhi LN, Wiese OJ, et al. The association between acid-base status and clinical outcome in critically ill COVID-19 patients admitted to intensive care unit with an emphasis on high anion gap metabolic acidosis. Ann Clin Biochem 2023;60:86-91. [Crossref] [PubMed]
7. Melo Mendes IC, Martins de Oliveira AL, Pinheiro Trindade PM, et al. Metabolic acidosis with elevated anion gap and euglycemic ketoacidosis in pregnant and postpartum women with severe Covid-19. Ann Med 2025;57:2445189. [Crossref] [PubMed]
8. Yasseen BA, Elkhodiry AA, El-Sayed H, et al. The role of neutrophilia in hyperlactatemia, blood acidosis, impaired oxygen transport, and mortality outcome in critically ill COVID-19 patients. Front Mol Biosci 2024;11:1510592. [Crossref] [PubMed]
9. Achanti A, Szerlip HM. Acid-Base Disorders in the Critically Ill Patient. Clin J Am Soc Nephrol 2023;18:102-12. [Crossref] [PubMed]
10. Huang Y, Ao T, Zhen P, et al. Association between serum anion gap and 28-day mortality in critically ill patients with infective endocarditis: a retrospective cohort study from MIMIC IV database. BMC Cardiovasc Disord 2024;24:585. [Crossref] [PubMed]
11. Li R, Jin X, Ren J, et al. Relationship of Admission Serum Anion Gap and Prognosis of Critically Ill Patients: A Large Multicenter Cohort Study. Dis Markers 2022;2022:5926049. [Crossref] [PubMed]
12. Xu C, Sun L, Dong M, et al. Serum Anion Gap is Associated with Risk of All-Cause Mortality in Critically Ill Patients with Acute Myocardial Infarction. Int J Gen Med 2022;15:223-31. [Crossref] [PubMed]
13. Lou Z, Zeng F, Huang W, et al. Association between the anion-gap and 28-day mortality in critically ill adult patients with sepsis: A retrospective cohort study. Medicine (Baltimore) 2024;103:e39029. [Crossref] [PubMed]
14. Huang Y, Ao T, Zhen P, et al. Association between the anion gap and mortality in critically ill patients with influenza: A cohort study. Heliyon 2024;10:e35199. [Crossref] [PubMed]
15. Huang Z, Wang S, Yang S. Association between serum anion gap and risk of in-hospital mortality in patients with acute heart failure. Sci Rep 2024;14:4858. [Crossref] [PubMed]
16. Cheng B, Li D, Gong Y, et al. Serum Anion Gap Predicts All-Cause Mortality in Critically Ill Patients with Acute Kidney Injury: Analysis of the MIMIC-III Database. Dis Markers 2020;2020:6501272. [Crossref] [PubMed]
17. Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data 2023;10:1. [Crossref] [PubMed]
18. Johnson A, Bulgarelli L, Pollard T, et al. Mimic-iv. PhysioNet; (cited 2024 Oct 28). Available online: https://physionet.org/content/mimiciv/3.0/
19. Wang H, Gao C, Dantona C, et al. DRG-LLaMA: tuning LLaMA model to predict diagnosis-related group for hospitalized patients. NPJ Digit Med 2024;7:16. [Crossref] [PubMed]
20. Liu K, Ma XY, Xiao H, et al. Association between the ROX index and mortality in patients with acute hypoxemic respiratory failure: a retrospective cohort study. Respir Res 2024;25:143. [Crossref] [PubMed]
21. El-Menyar A, Naduvilekandy M, Asim M, et al. Machine learning models predict triage levels, massive transfusion protocol activation, and mortality in trauma utilizing patients hemodynamics on admission. Comput Biol Med 2024;179:108880. [Crossref] [PubMed]
22. Huang Y, Ao T, Zhen P, et al. Association between serum bicarbonate and 28-day mortality in critically ill patients with infective endocarditis: a cohort study from MIMIC-IV. Sci Rep 2025;15:312. [Crossref] [PubMed]
23. Glasmacher SA, Stones W. Anion gap as a prognostic tool for risk stratification in critically ill patients - a systematic review and meta-analysis. BMC Anesthesiol 2016;16:68. [Crossref] [PubMed]
24. Chen X, Yang Q, Gao L, et al. Association Between Serum Anion Gap and Mortality in Critically Ill Patients with COPD in ICU: Data from the MIMIC IV Database. Int J Chron Obstruct Pulmon Dis 2024;19:579-87. [Crossref] [PubMed]
25. Wang S, Teng H, Han H, et al. The relationship between albumin corrected anion gap levels and mortality in patients with asthma in the ICU. Sci Rep 2023;13:16903. [Crossref] [PubMed]
26. Zhang Y, Ma H, Wang R, et al. Correlation between baseline anion gap and early acute kidney injury in patients with acute pancreatitis in the intensive care unit: A single-center retrospective cohort study. PLoS One 2025;20:e0315386. [Crossref] [PubMed]
27. Kou Y, Du S, Zhang M, et al. Serum Anion Gap at Admission Predicts All-Cause Mortality in Critically Ill Patients With Cirrhosis: A Retrospective Cohort Study. Clin Transl Gastroenterol 2024;15:e1. [Crossref] [PubMed]
28. Dafal A, Kumar S, Agrawal S, et al. Admission Anion Gap Metabolic Acidosis and Its Impact on Patients in Medical Intensive Care Unit. J Lab Physicians 2021;13:107-11. [Crossref] [PubMed]
29. Kieninger M, Sinning A, Vadász T, et al. Lower blood pH as a strong prognostic factor for fatal outcomes in critically ill COVID-19 patients at an intensive care unit: A multivariable analysis. PLoS One 2021;16:e0258018. [Crossref] [PubMed]
30. Trebuian CI, Brici OM, Sutoi D, et al. Lactate Levels and Clearance: Key Predictors of Prognosis for COVID-19 and Non-COVID-19 Septic Shock Patients in the Emergency Department. Clin Pract 2024;14:834-45. [Crossref] [PubMed]
31. Yu H, Chen S. Association between anion gap and the 30-day mortality of patients with ventilator-associated pneumonia: a study of the MIMIC-III database. J Thorac Dis 2024;16:2994-3006. [Crossref] [PubMed]
32. Al-Azzam N, Khassawneh B, Al-Azzam S, et al. Acid-base imbalance as a risk factor for mortality among COVID-19 hospitalized patients. Biosci Rep 2023;43:BSR20222362. [Crossref] [PubMed]
33. Gao P, Min J, Zhong L, et al. Association between albumin corrected anion gap and all-cause mortality in critically ill patients with acute kidney injury: a retrospective study based on MIMIC-IV database. Ren Fail 2023;45:2282708. [Crossref] [PubMed]
34. Zhong L, Xie B, Ji XW, et al. The association between albumin corrected anion gap and ICU mortality in acute kidney injury patients requiring continuous renal replacement therapy. Intern Emerg Med 2022;17:2315-22. [Crossref] [PubMed]