Association of Preadmission Metformin Exposure and Outcomes in Acute Kidney Injury Patients with Type 2 Diabetes in Intensive Care Unit: a Cohort Study

Background: Acute kidney injury (AKI) occurred in more than half of ICU patients. The effective prevention and therapies strategy of AKI remains limited. This study is to assess AKI mortality among diabetes patients with or without preadmission prescriptions for metformin. Methods: We included AKI patients with type 2 diabetes in Medical Information Mart for Intensive Care (MIMIC)-III database. 30-day mortality, neutrophil-to-lymphocyte ratio, and length of stay (LOS) in the hospital were compared between those with and without preoperative prescriptions. The statistical approaches included multivariate regression, propensity score analysis and an inverse probability ‐ weighting model to ensure the robustness of our ndings. Results: In total, 4328 AKI patients with type 2 diabetes (998 in preadmission metformin usage group and 3330 in no preadmission metformin usage) were included in the analysis. The overall 30-day mortality was 14.2% (613/4328). There was 15.7% (523/3330) and 9.0% (90/998) of 30-day mortality in no preadmission metformin usage and preadmission metformin usage group, respectively. In the main analysis, preadmission metformin usage was associated with a 37% lower of 30-day mortality (HR=0.63,95% CI:0.50-0.80, p<0.0001) in inverse probability ‐ weighting model. Conclusions: This cohort study suggested that preadmission metformin usage may be associated with reduced risk-adjusted mortality in AKI patients with type 2 diabetes. Further randomized controlled trials are needed.


Background
Acute kidney injury (AKI) is now recognized as a major public health problem affecting millions of people worldwide and leading to decreased survival, increased progression of underlying chronic kidney disease (CKD), and sometimes to new onset of CKD [1]. The effective prevention and therapies strategy remains limited [2].
Metformin, the most common and rst-line biguanide antihyperglycemic agent, has anti-in ammatory properties and is associated with lower all-cause mortality compared with other hypoglycemics [3,4].
Several studies demonstrated that preadmission metformin usage was associated with reduced mortality among kidney disease patients [5,6]. Bell reported preadmission metformin was associated with a lower rate of 28-day mortality in AKI patients [7]. A recent multinational cross-sectional study demonstrated 57.3% intensive care unit (ICU) patients suffered AKI [8]. However, in critical care patients, the evidence of preadmission metformin decrease a risk of AKI mortality was still absented. Therefore we conducted a retrospective cohort study to determine the association of preadmission metformin and mortality in AKI patients with diabetes.

Methods
We enrolled a cohort of AKI patients with type 2 diabetes who were exposed or not to metformin prior to admission in the database of Medical Information Mart for Intensive Care (MIMIC)-III (version 1.4).
MIMIC-III is a real-world and publicly available clinical database contained more than 60,000 intensive care unit (ICU) admissions in Beth Israel Deaconess Medical Center between 2001 and 2012 [9]. We were approved to use the database. All reporting followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [10].
Study population AKI patients with type 2 diabetes were eligible in our study. AKI was de ned according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria on admission were considered eligible for study inclusion. KDIGO criteria are as follows [11]: increase in serum creatinine (SCr) to ≥1.5 times baseline must have occurred within the prior 7 days; or ≥0.3 mg/dl increase in SCr occurred within 48 h; or urine volume < 0.5 ml/kg/h for 6 h or more. The minimum of the SCr values available within the 7 days before admission was used as the baseline SCr [12,13]. When the preadmission SCr was not available, the rst SCr measured at admission was used as the baseline SCr [12,14]. The diagnosis of type 2 diabetes was based on International Classi cation of Disease, Ninth Revision (ICD-9). We only included adult patients (age>16 years). For patients admitted to the ICU more than once, only the rst ICU stay was considered.

Metformin Exposure
Preadmission metformin exposure was de ned as a record of using metformin in "Medications on admission" in MIMIC-III.

Covariates
We included the following variables: demographic characteristics, marital status, insurance, and service unit, heart rate, mean arterial pressure (MAP), respiratory rate, SPO2, white blood cell (WBC) count, hemoglobin, platelet, creatinine, lactate, glucose, SOFA score, simpli ed acute physiology score (SAPS) II score, ventilator use, vasopressor use, renal replace treatment (RRT) use, and comorbidity disease included cardiovascular disease, liver disease, malignancy, neurological disease, chronic pulmonary disease, hypertension. Vasopressor included norepinephrine, epinephrine, phenylephrine, vasopressin, dopamine, dobutamine, and Isuprel. We also included marital status and insurance. These variables included those representing the health habits of patients who received preadmission metformin that may capture a healthy user effect [15].

Outcomes
The primary outcome was 30-day mortality. The secondary outcomes were neutrophil-to-lymphocyte ratio (NLR), and length of stay (LOS) in the hospital.

Statistical analysis
A descriptive analysis was performed for all participants. Categorical variables were expressed as numbers and percentages (%). Continuous variables were expressed as mean and standard deviation (SD) when normally distributed or median and interquartile range (IQR) when skewed. The chi-square tests (categorical variables) and One-Way ANOVA (normal distribution), Kruskal-Wallis (skewed distribution) test were used for comparison of categorical, normally, nonnormally distributed continuous variables, respectively.
To minimize the potential bias of treatment allocation and confounding, we generated a propensity score to estimate by logistic regression the likelihood that patients had preadmission metformin exposure [16]. A 1:1 nearest neighbor matching algorithm was applied using a caliper width of 0.01. The following variables were selected to generate the propensity score: age, sex, ethnicity, marital status, insurance, admission type, service unit, heart rate, MAP, respiratory rate, SPO2, WBC, creatinine, hemoglobin, platelet, ventilator use, vasopressor use, and SAPS II score. A standardized mean difference (SMD) was used to examine the PSM degree. A threshold of less than 0.1 was considered acceptable. On the PSM cohort, we used a 2-sided t test to compare preoperative NLR and LOS in hospital. We applied the Kaplan-Meier and log-rank analyses were used for 30-day survival curves.
Using the estimated propensity scores as weights, an inverse probabilities weighting (IPW) model was used to generate a weighted cohort [17]. A Cox proportional hazards regression was then performed to adjusted propensity score. We also used a univariable Cox proportional hazards regression model with the robust variance estimator to calculate the hazard ratio (HR) for mortality.
All analyses were performed with the statistical software packages R. version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria). EmpowerStats (X&Y Solutions, Inc., Boston, MA). The threshold of p <0.05 (two-sided) was considered statistically signi cant.

Sensitivity Analyses
To ascertain whether the results were sensitive to the matching method, we performed several additional sensitivity analyses using the full cohort. Multivariable logistic regression analyses were performed to assess the independent association after adjusted metformin prescription after hospital admission and adjusted for propensity score. Additionally, A previous study excludes patients with myocardial infarction during the previous month, who were contraindicated to metformin [18]. We exclude patients admitted to CCU for sensitivity analyses.

Population
We identi ed 4,328 individuals with type 2 diabetes who underwent AKI according to the KDIGO de nition. Of these patients, 998(23%) had preadmission metformin usage. Figure 1 presented the ow chart of the study patients. After propensity score matching, patient characteristics were balanced across groups ( Table 1). The extent of missing values was presented in eTables 1 in the Supplement.
On day 30, the mortality was 15.7% (523/3,330) and 9.0% (90/998) in no preadmission metformin usage and preadmission metformin usage group, respectively ( Table 2). The IPW demonstrated a signi cant bene cial effect of preadmission metformin usage in terms of 30-day mortality. The HR was 0.63 (95% CI, 0.50-0.80, P < 0.0001). The propensity score-matched mortality rates for the no preadmission metformin usage and preadmission metformin usage groups were 14.8% vs. 9%. In univariable Cox proportional hazards regression, The HR was 0.59 (95% CI, 0.45-0.77, P < 0.0001). Kaplan-Meier curve showed patients with preadmission metformin usage had lower mortality by day 30 (Log-rank test: p<0.0001, Figure 2). Secondary outcomes studies with propensity score matching We evaluated several secondary outcomes to investigate potential factors that might account for the bene cial effects of preadmission metformin usage after PSM. First, NLR had a numerical decrease among patients with preadmission metformin usage compared with those without preadmission metformin usage in admission

Discussion
In this retrospective, propensity score-matched cohort study, preadmission metformin usage provided to AKI patients with diabetes were associated with a lower risk-adjusted 30-day mortality, compared with no preadmission metformin usage. This association was reliable in additional models which control for the indicated bias.
In our study, 23.1% (998/4,328) AKI patients with type 2 diabetes received preadmission metformin exposure, which is lower compared with previous studies. Katherine et al. reported 59% of type 2 diabetes invidious used metformin before surgical procedure [19] and Samira et al. reported 40% AKI patients type 2 diabetes used metformin [7]. The de nition of metformin exposure may lead to this diversity. In MIMIC database metformin exposure was de ned as the recent history of metformin use and in Katherine and Samira's cohort, metformin exposure was de ned as prescribed metformin in the 180 days before surgical procedure or hospital admissions [7,19].
In a multicenter cohort, CKD (eGFR 30-60 ml/min/1.732) patients treated with metformin may decrease mortality rates compared to those not on metformin [20]. A recent study also showed among patients with diabetes and reduced kidney function persisting with monotherapy, treatment with metformin, compared with a sulfonylurea, was associated with a lower risk of major adverse cardiovascular events [21]. These studies indicated metformin may bene t diabetes patients with kidney disease and U.S. Food and Drug Administration (FDA) has also revised their recommendation stating that it can be used in mild renal impairment and some with moderate renal impairment [22]. Similar to our ndings, Bell et al. conducted an observational cohort study in Scotland (n=4944) and described preadmission metformin was associated with a lower rate of 28-day mortality (HR 0.82;95% CI,0.70-0.95, P= 0.01) [7]. Compared with our study, Bell et al. lacked data of critical care patients. In our cohort, we focused only on AKI patients in ICU and considered some critical risk factors such as SAPS II score, vasopressor usage, and ventilator usage. Thus, our study extended these ndings by demonstrating that preadmission metformin usage was also associated with a lower risk of all-cause motility in AKI patients from ICU.
Alike ndings in previous research, metformin was associated with lower NLR, which is a marker of systemic in ammation [23]. In other cohorts, a lower level of NLR is associated with a decreased risk of mortality in AKI patients [24,25]. However, it remains unclear the potential mechanism of preadmission metformin usage associated with lower mortality in AKI patients with diabetes. Metformin could improve autophagy and mitochondrial function in diabetes [26]and decrease in ammation by down-regulate proin ammatory cytokines, such as IL-6 and TNF-α [27,28]. Besides, metformin may have a potential role in antimicrobial therapy. Laboratory tests showed the effectiveness of metformin on multiple pathogens, including Trichinella spiralis, Staphylococcus aureus, Pseudomonas aeruginosa, hepatitis B virus, hepatitis C virus, and human immunode ciency virus [29]. The antimicrobial effect may be involved in the bene cial impact of sepsis. Furthermore, some studies show metformin could affect gut microbiota [30]and may bene t the AKI patients.
Several limitations are noteworthy. First, the potential for residual confoundings may exist, as with all retrospective analyses. We adjusted as many possible confoundings as we can and did a good balance in the PSM cohorts in this study. Second, as the study population contains only AKI patients with type 2 diabetes, it may not be generalizable to AKI patients with type 1 diabetes. Third, we were unable to exclude patients with myocardial infarction during the previous month, who were contraindicated to metformin [18]. We exclude patients admitted to CCU for instead. The result was still robust and reliable. Fourth, the record of metformin in "Medications on admission" in this study are likely to be more prone to unrecorded. The preadmission metformin usage in AKI patients with diabetes is lower than previously reported. However, it is noteworthy that the potential exposed misclassi cation resulting from such errors would bias toward the null, thus result in an underestimation of the association between preadmission metformin usage and 30-day mortality.

Conclusions
This cohort study suggested that preadmission metformin usage was associated with reduced riskadjusted mortality in AKI patients with type 2 diabetes. This association warrants further investigation. The establishment of this database was approved by the Massachusetts Institute of Technology (Cambridge, MA) and Beth Israel Deaconess Medical Center (Boston, MA), and consent was obtained for the original data collection. Therefore, the ethical approval statement and the need for informed consent were waived for this manuscript.

Consent for publication
Not applicable Availability of data and materials Data in the article can be obtained from MIMIC-III database (https://mimic.physionet.org/)

Competing interests
The authors declare that they have no competing interests.

Funding
There was no external for this work.

Authors' contributions
Qilin Yang conducted data analysis and wrote the manuscript. Jiezhao Zheng conducted data analysis.
Deliang Wen conducted data collection and data analysis. Xiaohua Chen conducted the data collection. Weiyan Chen conducted data collection and data interpretation. Weixiao chen drew the gure. Xuming Xiong designed the study and reviewed the manuscript. Zhenhui Zhang designed the study and reviewed the manuscript.  * Shown is the hazard ratio from the multivariable Cox proportional-hazards model, with adjusted for all covariates in table 1. † Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with the same strata and covariates with inverse probability weighting according to the propensity score. ‡ Shown is the hazard ratio from a multivariable Cox proportional-hazards model with the same strata and covariates with matching according to the propensity score. The analysis included 982 patients (982 who received preadmission metformin prescription and 982 who did not). § Shown is the hazard ratio from a multivariable Cox proportional-hazards model with the same strata and covariates, with additional adjustment for the propensity score.  The ow chart of the study Figure 2 Kaplan-Meier Survival Curves for day 30 of AKI patients with type 2 diabetes

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. etable.docx