Association between metformin use and reduced short- and long-term all-cause mortality in patients undergoing cardiac surgery: a retrospective cohort study

Abstract

Aims

The use of metformin during the perioperative period remains a contentious issue. This study aimed to assess the impact of metformin use on overall mortality rates among patients in postoperative cardiac intensive care units.

Materials and methods

This study retrospectively analyzed data from 8675 cardiac surgery patients sourced from the Medical Information Marketplace in Intensive Care (MIMIC 3.0) database. The administration of metformin during the postoperative ICU stay was assessed. Utilizing Cox proportional hazards modeling, the study explored the correlation between metformin treatment and all-cause mortality (ACM). The impact of metformin on survival rates was examined using the Kaplan–Meier method. Subgroup analyses and interaction tests were conducted to ascertain the consistency of these associations across diverse demographic and clinical subgroups.

Results

The average age of participants in this study was 69.0 years, spanning from 61.0 to 76.0 years. A higher proportion of individuals who received metformin following surgery were male (75.6% vs 72.3%, P = 0.01), younger (67.19 ± 9.26 vs 67.95 ± 11.75, P = 0.006), and had lower Sequential Organ Failure Assessment Scores (4.92 ± 2.38 vs 5.31 ± 2.79, P < 0.001). When compared with patients who did not receive metformin after surgery, these individuals who received metformin following surgery had a lower prevalence of chronic kidney disease, congestive heart failure, and peripheral vascular disease and the higher prevalence of acute myocardial infarction, hypertension and diabetes. The metformin administration was correlated to decreased ACM at 28, 60, 90, and 365 days.

Conclusions

Utilizing metformin in the intensive care unit following surgery was correlated with the notable reduction in ACM among patients who have undergone cardiac procedures, particularly those with diabetes. However, further prospective research is necessary to validate these findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-025-02878-9.

Introduction

The data from the World Health Organization indicates that approximately 2 million cardiac surgeries were conducted worldwide annually. This procedure is considered as one of the most intricate and perilous procedures in medicine, with an in-hospital mortality rate ranging from 2 to 6% [1, 2]. Over the recent decades, cardiac surgery has experienced significant advancements, revolutionizing the treatment of heart diseases and improving patient outcomes. Continuous enhancements in surgical methods and postoperative care have notably enhanced patient outcomes, resulting in a substantial reduction in in-hospital mortality rates. However, the extended-term outcomes for individuals who have had cardiac surgery continue to be a significant concern. The complexity of the procedure and the multiple complications that patients may face postoperatively, such as arrhythmias, heart failure, new-onset stroke, and acute kidney injury (AKI) [3–6], pose a threat to patients’ long-term health. For example, over 30% of patients who have undergone cardiac surgery experience AKI, and the extent of this injury is closely linked to a higher mortality rate [3, 7]. Many factors, such as ischemia–reperfusion injury, surgical stress, and systemic inflammatory response, may have an impact on the occurrence of complications after cardiac surgery [8]. Several studies have shown that targeted pharmacologic interventions implemented in the early postoperative period can significantly reduce the incidence of postoperative complications, thereby significantly improving the prognosis of patients undergoing cardiac surgery [9]. Therefore, an in-depth exploration of the application of early pharmacologic interventions for cardiac surgery patients is highly significant both clinically and scientifically.

Metformin is not only a widely used oral hypoglycemic agent, but it is also gaining attention for its potential role in other areas, such as in cancer treatment, weight management and cognitive impairment improvement [10–13]. Studies have shown that metformin has significant anti-inflammatory effects. The research has demonstrated its efficacy in lowering C-reactive protein (CRP) levels, a primary biomarker of inflammation, across both human and animal studies [14, 15]. Esteghamati observed that the administration of metformin significantly decreased the levels of oxidative stress and inflammation among individuals with type 2 diabetes [16, 17]. The results imply the anti-inflammatory properties of metformin could potentially mitigate postoperative inflammation and decrease the incidence of postoperative complications, thereby optimizing patient prognosis. Consequently, it is imperative to pursue additional studies investigating the effect of metformin in patients undergoing cardiac surgery, as this could potentially enhance survival rates and provide valuable insights into postoperative management strategies.

The Medical Information Mart for Intensive Care IV (MIMIC-IV) constitutes the comprehensive, multiple-center database dedicated to intensive care medicine, encompassing data on individuals admitted to ICUs from 2008 through 2022 [18]. This study investigated the correlation between early metformin use after cardiac surgery and clinical prognosis. This was accomplished by examining the clinical records of postoperative cardiac patients within the MIMIC-IV database. Our objective was to establish a robust scientific foundation to facilitate timely prognostic assessments and to enhance the planning of therapeutic interventions in clinical settings.

Methods

Study design and population

This study examines a retrospective observational cohort study, using data sourced from the MIMIC-IV database (version 3.0) [18]. The database, which is openly accessible, comprises 94,458 ICU admission records gathered. The author, B.C., who holds a certification from the Collaborative Institutional Training Initiative (CITI) with the identification number (14008404), was responsible for gathering the data used in this study. Given that anonymized data were used for this secondary analysis, informed consent was not sought. The dataset was accessible to researchers who met the criteria for data utilization. Adhering to the ethical standards set forth by the Helsinki Declaration, this study also complied with the STROBE guidelines, which are designed to enhance the quality of reporting for observational research in the field of epidemiology [19].

From the MIMIC-IV database, the research encompassed 8675 individuals who were hospitalized in the ICU after cardiac surgery. The exclusion criteria were defined as follows: (1) individuals under 18 years of age (n = 0); (2) ICU stays shorter than 24 h (n = 11,766); (3) nonprimarily admitted patients and those not admitted for the first time to the ICU (n = 1603); (4) patients who were not undergoing cardiac surgery (n = 43,322). The final sample consisted of 8675 patients (Fig. 1).

Fig. 1.

The flowchart of this study

Data extraction

Patient data, encompassing both clinical and demographic aspects, were retrieved through the execution of SQL queries utilizing the PostgreSQL database management system (version 13.7.2), complemented by the Navicat Premium client tool (version 16). The initial assessment captured a range of attributes such as age, race and gender, alongside the mean values of critical physiological parameters, including respiratory rate, heart rate, mean noninvasive arterial blood pressure and temperature, all recorded for the first time in ICU admission. Additionally, a comprehensive set of laboratory indicators was examined, consisting of white blood cell (WBC) counts, serum chloride, sodium, potassium levels and red blood cell (RBC) counts. Comorbid conditions included chronic kidney disease (CKD), diabetes, hypertension, heart failure, cerebrovascular disease, peripheral vascular disease, myocardial infarction, liver diseases and stroke among the patient cohort. Therapeutic measures include the use of vasopressors. The assessment tools utilized in this study comprised the Acute Physiology Score III (APS III) and the Sequential Organ Failure Assessment (SOFA).

Outcomes

The study’s primary outcomes were the rates of mortality from all causes mortality (ACM) at 28 days, 60 days, 90 days, and 365 days post-baseline (Supplementary table).

Statistical analysis

Data exhibiting a normal distribution were expressed as means ± standard deviation. Non-normally distributed data were described using medians and interquartile ranges. Categorical data were represented through counts and percentage values. In the analysis of continuous variables, parametric data were compared using Student’s t test, whereas nonparametric data were assessed with the Mann–Whitney U test. For categorical data, the chi-square test was utilized, with Fisher’s exact test employed when necessary. Subjects were categorized into two distinct groups based on their utilization of metformin. Assessment of the estimated mortality rates and related disparities was conducted using Kaplan–Meier survival analysis. Subsequently, the Log-Rank test was applied to evaluate the statistical significance across groups. Furthermore, both univariate and multivariate analyses were performed employing Cox proportional hazards regression to explore the relationship between metformin treatment and key outcomes of interest. For the multivariate Cox models, a selection of covariates was made based on established literature and their clinical significance. These included: Model 1 which was unadjusted, Model 2, adjusted for age, gender, and race, and a comprehensive model (Model 3) that further controlled for age, race, gender, vital signs, laboratory parameters, comorbidities, and treatment measures and scores. Subgroup analyses were performed to assess the impact of metformin use on various patient characteristics, such as gender, age, hypertension, heart failure, CKD, cerebrovascular disease and SOFA score. Statistical analyses were performed utilizing SPSS 25.0 software (developed by IBM, based in the USA) alongside R 4.1.2 (provided by the R Foundation). Statistical significance was established with a p value threshold of less than 0.05.

Results

Baseline characteristics

The compilation of baseline characteristics and clinical outcomes, stratified by metformin use, is presented in Table 1. A total of 8675 postoperative cardiac patients (2357 females and 6318 males) who fulfilled the inclusion criteria. The average age of participants in this study was 69.0 years, spanning from 61.0 to 76.0 years. Of these, 16.95% (1470/8675) of patients were treated with metformin. The 28-day mortality rate was 0.3% (5/1470) in metformin users and 1.5% (109/7205) in nonusers (P < 0.001, Table 1). The 60-day mortality rate was 0.6% (9/1470) in metformin users and 2.3% (168/7205) in nonusers (P < 0.001, Table 1). The 90-day mortality rate was 0.9% (13/1470) in metformin users and 2.7% (198/7205) in non-users (P < 0.001, Table 1). The 365-day mortality rate was 2.2% (32/1470) in metformin users and 5.1% (367/7205) in nonusers (P < 0.001, Table 1). The findings indicated that a larger share of individuals administered metformin following surgery were male, exhibited younger ages, and possessed lower SOFA scores. These patients had fewer comorbidities such as peripheral vascular disease, congestive heart failure, and CKD. However, these individuals displayed a higher incidence of acute myocardial infarction, diabetes and hypertension.

Table 1.

Baseline characteristics

Name	Nonmetformin (N = 7205)	Metformin (N = 1470)	Total (N = 8675)	P
Gender (male)	5207 (72.3%)	1111 (75.6%)	6318 (72.8%)	0.01
Age	69.0 (61.0 to 76.0)	68.0 (61.0 to 74.0)	69.0 (61.0 to 76.0)	< 0.001
Serum sodium	139.0 (137.0 to 141.0)	139.0 (137.0 to 140.0)	139.0 (137.0 to 141.0)	0.225
Serum potassium	4.3 (4.0 to 4.6)	4.3 (4.0 to 4.6)	4.3 (4.0 to 4.6)	0.91
Serum Chloride	109.0 (106.0 to 111.0)	108.0 (106.0 to 110.0)	109.0 (106.0 to 111.0)	< 0.001
White blood cell	12.1 (9.1 to 15.6)	12.5 (9.4 to 16.1)	12.2 (9.1 to 15.7)	0.008
Red blood cell	3.1 (2.7 to 3.5)	3.2 (2.8 to 3.6)	3.1 (2.8 to 3.6)	< 0.001
Heart rate	80.0 (74.0 to 86.0)	80.0 (75.0 to 86.0)	80.0 (74.0 to 86.0)	0.42
Mean arterial pressure	72.0 (65.0 to 80.0)	73.0 (67.0 to 80.0)	73.0 (66.0 to 80.0)	0.017
Respiratory rate	15.0 (12.0 to 17.0)	15.0 (12.0 to 17.0)	15.0 (12.0 to 17.0)	0.578
Temperature	36.6 (36.4 to 36.9)	36.6 (36.4 to 36.9)	36.6 (36.4 to 36.9)	0.612
APS III	33.0 (26.0 to 44.0)	33.0 (26.0 to 41.0)	33.0 (26.0 to 44.0)	0.306
SOFA	5.0 (3.0 to 7.0)	5.0 (3.0 to 6.0)	5.0 (3.0 to 7.0)	< 0.001
Hypertension	4031 (55.9%)	1032 (70.2%)	5063 (58.4%)	< 0.001
Myocardial infarction	943 (13.1%)	250 (17%)	1193 (13.8%)	< 0.001
Heart failure	1835 (25.5%)	313 (21.3%)	2148 (24.8%)	< 0.001
Peripheral vascular disease	387 (5.4%)	50 (3.4%)	437 (5%)	0.002
Cerebrovascular disease	728 (10.1%)	150 (10.2%)	878 (10.1%)	0.945
CKD	1172 (16.3%)	161 (11%)	1333 (15.4%)	< 0.001
Diabetes	1522 (21.1%)	1423 (96.8%)	2945 (33.9%)	< 0.001
Liver disease	259 (3.6%)	57 (3.9%)	316 (3.6%)	0.652
COPD	365 (5.1%)	70 (4.8%)	435 (5%)	0.674
Stroke	541 (7.5%)	109 (7.4%)	650 (7.5%)	0.944
Vasopressors	7107 (98.6%)	1455 (99%)	8562 (98.7%)	0.357

APS III Acute Physiology and Chronic Health Evaluation III, SOFA sequential organ failure assessment, CKD chronic kidney disease, COPD chronic obstructive pulmonary disease

Survival correlation

Both the log-rank test and the Kaplan–Meier curve showed that metformin users had a higher probability of survival as compared to nonusers in Fig. 2. Multivariate analysis indicated the utilization of metformin was correlated to lower 28-day ACM (HR = 0.34, 95% CI 0.13–0.88, P = 0.027), 60-day ACM (HR = 0.32, 95% CI 0.15–0.65, P = 0.002), 90-day ACM (HR = 0.43, 95% CI 0.23–0.79, P = 0.007) and 365-day all-cause mortality (HR = 0.50, 95% CI 0.34–0.75, P = 0.001) in postcardiac surgery patients.

Fig. 2.

Association of metformin use with the probability of short- and long-term survival in patients after cardiac surgery. A Metformin use with the probability of 28-day survival after cardiac surgery. B 60 days. C 90 days. D 365 days. HR hazard ratio

Subgroup analysis

Stratified analyses were conducted to evaluate the correlation between metformin treatment and all-cause mortality among cardiac surgery patients across various subgroups (Table 2). The analyses were stratified by gender, age, hypertension, heart failure, CKD, cerebrovascular disease and SOFA score. The findings indicated no notable interaction between metformin treatment and the majority of subgroups (P for interaction > 0.05). However, a significant interaction (P for interaction = 0.041) was noted specifically within the category of patients with diabetes (Fig. 3).

Table 2.

Univariate and multivariate Cox regression modeling elucidating the association between metformin and all-cause mortality in postcardiac surgery patients

	HR (95% CI)
	Model 1	P value	Model 2	P value	Model 3	P value
28 days ACM	0.22 (0.09–0.55)	0.001	0.23 (0.09–0.57)	0.001	0.34 (0.13–0.88)	0.027
60 days ACM	0.26 (0.13–0.51)	< 0.001	0.27 (0.14–0.54)	< 0.001	0.32 (0.15–0.65)	0.002
90 days ACM	0.32 (0.18–0.56)	< 0.001	0.33 (0.19–0.58)	< 0.001	0.43 (0.23–0.79)	0.007
365 days ACM	0.42 (0.29–0.60)	< 0.001	0.44 (0.31–0.64)	< 0.001	0.50 (0.34–0.75)	0.001

HR Hazard Ratio, 95% CI 95% confidence interval, ACM all causes mortality

Model 1 was adjusted for none

Model 2 was adjusted for age, gender, ethnicity

Model 3 was adjusted for age, sex, race, hypertension, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, renal disease, diabetes mellitus, hepatic disease, chronic lung disease, stroke, use of vasopressor medications, sodium, potassium, chloride, leukocytes, erythrocytes, mean heart rate, mean arterial blood pressure, mean respiratory rate, mean body temperature, Acute Physiology and Chronic Health Evaluation III, Sequential Organ Failure Assessment

Fig. 3.

Association between metformin use and all-cause mortality in patients undergoing cardiac surgery between different subgroups. HR hazard ratio, 95% CI 95% confidence interval, SOFA sequential organ failure assessment, CKD chronic kidney disease, 0 absence of the disease state, 1 presence of the disease state. Along the X axis, the left side (HR < 1) indicates “Favors metformin,” while the right side (HR > 1) indicates “Favors nonmetformin”

Discussion

Cardiac surgery is essential for the treating of heart disease. Nonetheless, patients in the postsurgical period are at an elevated risk of mortality, particularly during the initial postoperative phase. The aim of this research was to assess the influence of metformin treatment during the initial postoperative period on the survival rate of patients who have had cardiac surgery, with the intention of developing novel clinical management strategies. Previous research has established that metformin offers notable cardiovascular advantages for patients suffering from cardiovascular conditions, such as decreased rates of ACM and cardiovascular events [20, 21]. In addition, it has been demonstrated that metformin is effective in lowering death rates among individuals with diabetes who also have heart failure [22]. These findings provide a solid theoretical basis for the present study, suggesting that metformin may have a positive contribution to postoperative recovery in cardiac surgery patients.

In this study, the annual mortality rate for individuals who had cardiac surgery was found to be 4%, aligning with previously documented mortality rates [23]. We investigated the correlation between metformin treatment and mortality rates among cardiac surgery patients, and the results from multivariable-adjusted Cox regression analyses indicated that metformin use was linked to a decline in ACM post cardiac surgery. Subgroup analyses, which accounted for variables such as gender, age, hypertension, heart failure, CKD, cerebrovascular disease, and SOFA score, produced similar findings. Furthermore, an interactive effect was noted within the subgroup of patients with diabetes.

The impact of metformin administered during the perioperative period on patient outcomes following cardiac surgery remains a subject of debate. Post cardiac surgical procedures, individuals commonly undergo an inflammatory reaction and oxidative stress, both of which are significantly linked to the emergence of complications in the postoperative phase. Metformin exhibits anti-inflammatory properties and mitigates oxidative stress, which may contribute to the decreased likelihood of developing postoperative atrial fibrillation after cardiac procedures [24, 25]. Prior research has indicated that the incidence of postoperative atrial fibrillation after cardiac procedures did not significantly diminish in diabetic individuals who had been treated with metformin before the surgery. However, this did not prevent it from positively affecting the overall prognosis after cardiac surgery through anti-inflammatory and antioxidative stress mechanisms. For example, metformin triggers AMPK activation by suppressing mitochondrial respiratory chain complex I. This inhibition results in the diminished electron transport across the mitochondrial membrane and decrease in mitochondrial membrane potential. Consequently, this leads to a reduction in mitochondrial oxygen consumption and ATP production, ultimately increasing the AMP/ATP ratio [26, 27]. The phosphorylation and consequent activation of AMPK encourages the shift of macrophages towards the M2 phenotype and concurrently inhibits the secretion of pro-inflammatory cytokines [28]. In addition, AMPK activation promotes nuclear translocation of the transcription factor Nrf2, thereby augmenting the expression of the detoxifying enzyme heme oxygenase-1 (HO-1), and mitigates cellular damage induced by lipopolysaccharide (LPS) [29, 30].

The occurrence of AKI represents a prevalent and severe issue following coronary artery bypass grafting (CABG) procedures [31–34]. The research indicates that the administration of metformin correlates with a diminished incidence of AKI post CABG in individuals diagnosed with type 2 diabetes [35, 36]. This association might stem from the observation that metformin is linked to an enhanced myocardial perfusion reserve and improved survival rates. In a retrospective, multicenter, longitudinal cohort study, individuals afflicted by type 2 diabetes using metformin had an elevated myocardial perfusion reserve and diminished occurrence of ACM and major adverse cardiovascular events (MACE) compared to patients without metformin [37]. This suggests that metformin may enhance cardiac function by improving myocardial perfusion, which in turn improves postoperative cardiac prognosis [38]. Myocardial ischemia–reperfusion injury (MIRI) is a common pathological injury in cardiac surgery, and metformin may regulate multiple signaling pathways by activating the AMPK signaling pathway. Activation of AMPK can regulate energy metabolism, antioxidative stress, attenuate inflammatory response, regulate autophagy, and inhibit apoptosis, thereby attenuating myocardial ischemia–reperfusion injury. This mechanism is important for myocardial protection and functional recovery after cardiac surgery, and helps to improve patients'postoperative prognosis [39, 40]. Although short-term metformin treatment is considered safe, it does not seem to serve as an effective approach for mitigating myocardial damage during surgery in nondiabetic individuals who are undergoing for CABG [41]. While lactic acidosis associated with metformin is an uncommon and severe adverse event, ongoing metformin therapy throughout the perioperative period does not correlate with a heightened risk of lactic acidosis in CABG patients [42].

There are several limitations associated with this study. To begin with, our findings indicate a correlation between metformin administration and enhanced clinical outcomes in cardiac patients, but we could not thoroughly clarify the precise mechanisms and pathways through which metformin enhances outcomes. Secondly, this study was conducted as a retrospective observational analysis at a single center with a limited sample size, utilizing data from the MIMIC-IV database covering the period from 2008 to 2022, included data on critically ill patients, which may introduce variability due to potential discrepancies in diagnostic criteria. Then, MIMIC database lacks dedicated cardiac surgery risk scores and many of their key components, such as ejection fraction and surgical urgency. The absence of this data may have affected our comprehensive assessment of patient risks. Second, we did not differentiate between types of cardiac surgery. Different surgeries can have varying risks and prognoses, so this lack of classification may have limited our ability to more precisely analyze the impact of metformin. We acknowledge these limitations. Future studies with more detailed cardiac surgery data could further explore these relationships. In subsequent studies, we intend to place a greater emphasis on examining the interaction between the categorization of surgical interventions and metformin treatment. Additionally, conducting an extensive randomized controlled trial is necessary to further substantiate the findings of our study. The multifaceted impact of metformin merits extensive further investigation.

Conclusion

The utilization of metformin in the ICU following surgery has been linked to a marked reduction in overall mortality rates among individuals who have undergone cardiac procedures, especially those with diabetes. Nonetheless, additional prospective research is required to substantiate these observations.

Author contributions

BC participated in data collection and analysis and drafted the manuscript. YHT participated in data merging and analysis and drafted the manuscript. YMK participated in data interpretation and drafted the manuscript. CYW contributed significantly to the study design and manuscript writing. All other authors contributed to the interpretation, discussion and editing of the manuscript. All authors consented to publication in this journal.

Funding

This research is financially supported by the project from The Science and Education Cultivation Fund of the National Cancer and Regional Medical Center of Shanxi Provincial Cancer Hospital (SD202309) and Shanxi Provincial Health Commission scientific research project (2023XG046).

Data availability

The data that support the findings of this study are openly available in MIMIC IV database at https://mimic.mit.edu.

Declarations

Ethics approval and consent to participate

Beth Israel Deaconess Medical Center (BIDMC) managed study approvals with human participants, anonymizing patient data for privacy. Ethical Committee of the Beth Israel Deaconess Medical Center exempted informed consent requirements.

Competing interests

The authors declare no competing interests.