Association of H2RAs exposure with 30-day mortality in patients with acute myocardial infarction in MIMIC-IV: a retrospective cohort study

Abstract

Histamine H2 Receptor Antagonist (H2RAs) administration reduces mortality in critically ill patients with HF, but the association between H2RAs exposure and acute myocardial infarction (AMI) remains unclear. This study aims to investigate the relationship between H2RAs administration and 30-day mortality in critically ill patients with AMI using the MIMIC-IV 3.0 database. A retrospective cohort study was conducted using data from the MIMIC-IV 3.0 database. This study enrolled adult AMI patients and set the primary endpoint as the mortality within 30 days following hospital admission. Multivariable Cox proportional hazards models were employed to adjust for potential confounding factors, including demographic variables, comorbid conditions, and illness severity. Our analysis comprised 4252 patients with AMI, among whom 1557 received H2RAs. The overall 30-day mortality rate observed in the cohort was 13.9%. Following adjustment for confounding factors, H2RAs administration was associated with an increased hazard ratio of 1.32 (95% confidence interval: 1.09-1.6). Further subgroup analyses and propensity score assessments have validated the reliability and consistency of these results. This study revealed a significant association between H2RA administration and an increased 30-day mortality rate in myocardial infarction patients. Our findings highlight the need for further investigation to understand the underlying mechanisms and assess the clinical implications of H2RA use in this susceptible patient population.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18032-8.

Introduction

Acute myocardial infarction (AMI) remains a leading cause of cardiovascular mortality worldwide, contributing to more than 8 million deaths annually¹. Despite advancements in reperfusion therapies and secondary prevention strategies, short-term prognosis within 30 days post-AMI remains suboptimal, particularly due to comorbidities and potential drug interactions that might exacerbate risks². Identifying critical factors accounting for short-term outcomes during this period can improve survival rates and guide therapeutic decisions in AMI patients.

Stress ulcer (SU), a common gastrointestinal complication in AMI patients, is associated with sympathetic activation, mucosal ischemia, and increased gastric acid secretion³. H2RAs and proton pump inhibitors (PPIs) are commonly prescribed medications for the prevention and treatment of SU, particularly in critically ill patients⁴. However, recent studies including the 2022 Chinese Society of Digestive Surgery clinical practice guideline⁵, have raised concerns that long-term use of acid-suppressive agents may increase the risk of pneumonia, osteoporosis-related fracture, C. difficile infection, micronutrient deficiencies, renal impairment, and cardiovascular events. Early randomized controlled trials (RCTs) showed comparable efficacy between H2RAs and PPIs in reducing SU risk (RR = 0.92, 95% CI 0.73-1.16)⁶, but mortality outcomes were not thoughtfully evaluated. Studies in the past decade suggested that PPIs elevated cardiovascular risks likely by inhibiting nitric oxide synthase inhibition or interfering with clopidogrel metabolism⁷. In contrast, H2RAs, due to their distinct pharmacodynamic profile, are considered a safer alternative⁸. Evidence directly linking H2RAs to mortality in AMI patients remains limited, with previous studies often constrained by small sample sizes or inadequate adjustment for confounders.

The Medical Information Mart for Intensive Care IV (MIMIC-IV) database, a publicly accessible electronic health record repository, provides granular data for analyzing drug-outcome associations in critical care settings⁹. This retrospective cohort study utilizes the MIMIC-IV 3.0 database to evaluate the association between H2RA exposure and 30-day mortality in AMI patients. The findings provide evidence for the selection of gastric mucosal protection strategies in AMI patients.

Methods

Data source

The MIMIC-IV v3.0 database was used as the data source in this study. MIMIC-IV includes electronic health record dataset from 364,627 patients and 546,028 hospitalized patients between 2008 and 2022⁹. It contains clinical data such as patient demographics, diagnoses, laboratory test results, medications, and vital signs. This was in accordance with the Reporting of studies conducted using observational routinely collected data for pharmacoepidemiology (RECORD-PE) reporting guidelines¹⁰. Songmei Guan passed the Protecting Human Research Participants exam and obtained access to the MIMIC database (certification number: 62437414). Informed consent was not required since all the data were deidentified. This study was approved by the Medical Research Ethics Committee of the Second Affiliated Hospital of Guangdong Medical University (No: YJKT2025-026–01).

Study population

This study included data from all adult patients (≥ 18 years) whose first three discharge diagnoses (primary, secondary, or tertiary) included AMI, as identified by ICD-9/10 codes. Only patients who were admitted to the ICU for the first time during hospitalization were included. The ICD-9/10 codes¹¹ are available in the online supplementary material, showing at Table S1.

Data extraction

The data used in the study were extracted from the MIMIC-IV database using Structured Query Language (SQL) in Navicat Premium (version 17.0.13). H2RAs (including famotidine, ranitidine, and cimetidine) were defined as their use within 24 h of ICU admission. Physical characteristics, vital signs, laboratory parameters, clinical parameters, co-morbidities, medications, and other information were collected. Variables with more than 40% missing values were excluded from the analysis. We imputed missing data of the covariates by using multiple imputations. For more details on specific extraction information and covariate screening, refer to Table S2.

Primary endpoint

The primary endpoint of this study was 30-day mortality, defined as death observed within 30 days of entry into the ICU.

Statistical analysis

Histogram distribution, Q-Q plots, and the Kolmogorov-Smirnov test were utilized to assess the normality of the variables. Continuous variables that exhibited a normal distribution are presented as the means ± standard deviations (SDs), whereas skewed continuous variables are presented as medians and interquartile ranges (IQRs). Categorical variables are reported as frequencies and percentages (%). Continuous variables across groups were compared using the independent samples Student’s t-test for normally distributed data and the Mann-Whitney U test for non-normally distributed data. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Hazard ratios (HRs) and 95% CIs for 30-day mortality were calculated via a Cox proportional hazards model. Multiple imputation was used to estimate the missing values of each variable. A multivariate adjustment model was used. The proportional hazards hypothesis was verified by examining the "log-log" plot and introducing the interaction with survival time. The outcome of 30-day mortality was analyzed using Kaplan-Meier survival curves based on H2RA usage and was compared with the log-rank test. The confounders were selected on the basis of clinical interest, prior scientific literature, all significant covariates in univariate analyses, or their association with an estimated change of more than 10% in the outcome or effect of interest. Multicollinearity was tested via the variance inflation factor (VIF) method, with a VIF ≥ 5 indicating the presence of multicollinearity. We constructed 3 models: Model 1 was adjusted for sex, age, acute myocardial infarction, BMI, revascularization, and the Charlson comorbidity index. Model 2 was additionally adjusted for PPI, RAASI, aspirin, beta-blockers, clopidogrel, congestive heart failure, and cardiogenic shock. Model 3 was additionally adjusted for heart rate, SBP, SPO₂, cardiac troponin T, hemoglobin, BUN, creatinine, partial thromboplastin time, lactate, PaO₂, and PaCO₂.

The subgroup analyses were conducted by subgroup variables. We used multiple imputation, which was based on 5 replications and a chained equation approach method in the R mice procedure, to maximize the statistical power and minimize bias that might account for missing data. We repeated all analyses with the complete data cohort for comparison. We conducted a series of sensitivity analyses to evaluate the robustness of the findings of the study and how our conclusions can be affected by applying various association inference models. In the sensitivity analysis, we applied several more association inference models: propensity score adjusted [PSA], propensity score matching [PSM]¹², inverse probability of treatment weighting [IPTW]¹³, and pairwise algorithm [PA]^14,15. The calculated effect sizes and p values from all these models were reported and compared.

All analyses were performed via R Statistical Software (Version 4.2.2, http://www.R-project.org, The R Foundation) and the Free Statistical analysis platform (Version 2.0, Beijing, China, http://www.clinicalscientists.cn/freestatistics). Free Statistics is a software package that provides intuitive interfaces for most common analyses and data visualization. It uses R as the underlying statistical engine, and the graphical user interface (GUI) is written in Python. Most analyses can be performed with just a few clicks. It is designed for reproducible analysis and interactive computing. A two-sided P value < 0.05 was considered statistically significant.

Results

Baseline characteristics of the study population

A total of 13,309 patients with a diagnosis of AMI were identified in the database. After excluding patients who spent less than 24 h in the ICU or lacked explicit records regarding H2RA use, 4252 patients met the inclusion and exclusion criteria (Fig. 1). Of these, 1557 cases were H2RA users and 2695 were non-H2RA users. The age range of the cohort was 21-99 years. Among these patients, 614 were diagnosed with ST-segment elevation myocardial infarction.

Fig. 1.

Selection of the study population from the MIMIC-Ⅳ database.

Baseline characteristics before matching are summarized in Table 1. Compared to the non-H2RA group, H2RA users were younger (P < 0.001), included fewer women (P < 0.001), had lower Charlson comorbidity index scores, fewer ST-segment elevation myocardial infarction (STEMI) cases (P < 0.001), and a higher proportion of patients who underwent revascularization (P < 0.001). There was no significant difference in body mass index between the two groups (P = 0.104), whereas ethnicity differed significantly (P = 0.002).

Table 1.

Baseline characteristics of the study participants.

Characteristics	All participant (n = 4252)	Patient without H2RAs (n = 2695)	Patient with H2RAs (n = 1557)	P value
STEMI, n (%)	614 (14.4)	452 (16.8)	162 (10.4)	< 0.001
Gender, male, n (%)	2,802 (65.9)	1,705 (63.3)	1,097 (70.5)	< 0.001
Age, years	70.6 ± 12.6	71.4 ± 13.1	69.1 ± 11.5	< 0.001
BMI, kg/m2	28.9 ± 6.3	28.7 ± 6.3	29.1 ± 6.3	0.104
Ethnicity, White n (%)	2,656 (62.5)	1,731 (64.2)	925 (59.4)	0.002
Revascularization, n (%)	1,212 (28.5)	508 (18.8)	704 (45.2)	< 0.001
Charlson comorbidity index	6.9 ± 2.6	7.1 ± 2.6	6.4 ± 2.4	< 0.001
Vital signs
Troponin, ng/mL	1.3 (0.4, 3.7)	1.3 (0.4, 3.8)	1.1 (0.4, 3.5)	0.029
Heart rate, beats per minute	99.8 ± 18.7	100.0 ± 19.8	99.5 ± 16.6	0.396
SBP, mmHg	87.8 ± 14.8	88.9 ± 15.6	85.8 ± 13.2	< 0.001
DBP, mmHg	44.9 ± 10.5	45.3 ± 11.3	44.4 ± 9.1	0.009
Respiratory rate, breaths per minute	28.0 ± 5.9	28.1 ± 5.8	27.8 ± 6.0	0.168
SPO2, %	91.7 ± 5.6	91.5 ± 5.6	92.1 ± 5.6	< 0.001
Hemoglobin, g/dL	10.2 ± 2.3	10.5 ± 2.4	9.7 ± 2.0	< 0.001
BUN, mg/dL	22.0 (16.0, 35.0)	24.0 (17.0, 39.0)	19.0 (14.0, 28.0)	< 0.001
Creatinine, mg/dL	1.1 (0.9, 1.7)	1.2 (0.9, 1.8)	1.1 (0.8, 1.4)	< 0.001
Sodium, mmol/L	139.1 ± 4.0	139.2 ± 4.2	139.0 ± 3.8	0.043
Potassium, mmol/L	4.7 ± 0.8	4.7 ± 0.8	4.7 ± 0.7	0.44
PTT, s	47.0 (31.2, 81.7)	53.7 (31.6, 91.5)	39.6 (30.7, 67.3)	< 0.001
Lactate, mmol/L	2.4 (1.7, 3.4)	2.3 (1.5, 3.3)	2.5 (1.9, 3.5)	< 0.001
PaO2, mmHg	87.0 (70.0, 114.0)	84.0 (66.0, 112.0)	91.0 (74.0, 117.0)	< 0.001
PaCO2, mmHg	46.4 ± 10.5	44.8 ± 11.4	48.1 ± 9.2	< 0.001
Comorbidity
Congestive heart failure, n (%)	2,135 (50.2)	1,482 (55)	653 (41.9)	< 0.001
Peptic ulcer disease, n (%)	83 (2.0)	67 (2.5)	16 (1)	< 0.001
Diabetes, n (%)	1,775 (41.7)	1,111 (41.2)	664 (42.6)	0.365
Hypertension, n (%)	3,329 (78.3)	2,094 (77.7)	1,235 (79.3)	0.217
Cardiogenic shock, n (%)	711 (16.7)	481 (17.8)	230 (14.8)	0.01
Medications
PPI, n (%)	1,209 (28.4)	1,050 (39)	159 (10.2)	< 0.001
RAASI, n (%)	1,323 (31.1)	1,124 (41.7)	199 (12.8)	< 0.001
Aspirin, n (%)	2,770 (65.1)	1,616 (60)	1,154 (74.1)	< 0.001
CCB, n (%)	212 (5.0)	123 (4.6)	89 (5.7)	0.096
Diuretic, n (%)	1,857 (43.7)	1,070 (39.7)	787 (50.5)	< 0.001
β-blockers, n (%)	2,166 (50.9)	1,384 (51.4)	782 (50.2)	0.478
Clopidogrel, n (%)	933 (21.9)	718 (26.6)	215 (13.8)	< 0.001
Statin, n (%)	2,038 (47.9)	1,257 (46.6)	781 (50.2)	0.027
Outcomes
In-hospital mortality, n (%)	464 (10.9)	314 (11.7)	150 (9.6)	0.042
Los ICU, Median (IQR)	2.4 (1.5, 4.2)	2.4 (1.6, 4.2)	2.3 (1.4, 4.3)	0.462
Mor 30d, n (%)	590 (13.9)	409 (15.2)	181 (11.6)	0.001

H2RAs, histamine H2 receptor antagonist; STEMI,ST-segment elevation myocardial infarction; BMI, body mass index; Revascularization, Coronary artery bypass grafting or percutaneous coronary intervention; SBP, systolic blood pressure; DBP, diastolic blood pressure; SPO2, Percutaneous arterial oxygen saturation; BUN, Serum urea nitrogen ;PTT, Partial thromboplastin time ;PaO2,Arterial partial pressure of oxygen;PaCO₂,Arterial blood carbon dioxide partial pressure; PPI, Proton pump inhibitor; RAASI, Renin angiotensin aldosterone system inhibitor; CCB, Calcium channel antagonists; LOS, length of stay; ICU, indicates intensive care unit.

Regarding vital signs and laboratory values. H2RA users showed lower systolic BP and hemoglobin levels, higher SpO₂, and better renal function indices (all P < 0.05). They also had slightly lower STEMI-presentation troponin levels (P = 0.029), shorter partial thromboplastin time (PTT), and marginally higher lactate values (both P < 0.001). The prevalence of congestive heart failure, peptic ulcer disease, and cardiogenic shock was lower in H2RA users (all P ≤ 0.01), while the rates of diabetes and hypertension were comparable between groups.

In terms of medications, H2RA users received less PPI, RAAS inhibitor, or clopidogrel, but more aspirin, diuretics, and statins (all P < 0.05), whereas β-blocker and CCB use did not differ. For the outcomes, both in-hospital and 30-day mortality were modestly lower among H2RA users (9.6% vs. 11.7% and 11.6% vs. 15.2%; P = 0.042 and 0.001, respectively); ICU length of stay was similar between groups (P = 0.46).

Associations between H2RA exposure and primary event

First, we evaluated the 30-day mortality rate of 4252 patients with followed up. The overall incidence was 13.9%. In the meantime, H2RAs exposure was significantly associated with increased 30-day mortality, as shown by Kaplan-Meier curves (Supplementary material, Fig. S1). In univariate Cox regression, 30-day mortality was associated with older age, STEMI presentation, higher troponin, lower systolic blood pressure, heart failure, cardiogenic shock, and the use of PPIs, whereas H2RA exposure showed a protective point estimate (Supplementary Table S3).

Multivariate Cox regression analyses were then performed to evaluate the relationship between H2RA use and 30-day mortality (Table 2). In the unadjusted model, H2RAs administration after ICU admission was linked to a 32% higher risk of 30-day mortality compared to non-H2RAs users (P = 0.005). In the minimally adjusted model (Model I), the association did not reach statistical significance (HR 1.18, 95% CI 0.98–1.41; P = 0.077). However, after adjusting for demographic factors, cardiovascular risk factors, comorbidities, and in-hospital treatments (Model II), H2RA use was associated with a significantly higher risk of 30-day mortality (hazard ratio [HR] 1.36, 95% confidence interval [CI] 1.13–1.65; P = 0.001). This association remained consistent after further adjustment for vital signs and laboratory parameters (Model III: HR 1.32, 95% CI 1.09–1.60; P = 0.005).

Table 2.

Associations between H2RA use and outcomes in myocardial infarction patients.

Outcome	Model Ⅰ		Model Ⅱ		Model Ⅲ
	HR (95% CL)	P value	HR (95% CL)	P value	HR (95% CL)	P value
Patients without H2RAs	1(Ref)		1(Ref)		1(Ref)
Patients with H2RAs	1.18(0.98–1.41)	0.077	1.36(1.13–1.65)	0.001	1.32(1.09–1.6)	0.005

Model I: Adjusted for sex, age, STEMI, BMI, revascularization and the Charlson comorbidity index.

Model II: Adjusted for variables in Model I plus PPIs, the RAASI, aspirin, β-blockers, clopidogrel, congestive heart failure and cardiogenic shock.

Model III: Adjusted for variables in Model II plus heart rate, SBP, SPO2, cardiac troponin T, hemoglobin, BUN, creatinine, PTT, lactate, PaO2 and PaCO2.

Propensity score matching analysis

To strengthen the validity of our findings, propensity score matching (PSM)¹⁶ was performed, adopting a 1:2 nearest neighbor matching algorithm with a caliper width of 0.2. This resulted in 1512 patients exposed to H2RAs being matched to an equal number of non-H2RA-exposed patients. After matching, the baseline characteristics of the two groups were well balanced, with all standardized mean differences (SMDs) for the variables less than 0.1, indicating comparability between groups (Table S4). In the matched cohort of 3024 patients, the 30-day mortality rate was 15.67%.

For the PSM cohort, Cox proportional-hazards regression was applied to estimate the association between H2RA use and 30-day mortality. Incorporating the estimated propensity scores were set as weights. A doubly robust estimation combines a multivariate Cox regression model with a propensity score model that was also applied to estimate the independent associations in the full cohort¹⁷. To evaluate the robustness of our findings, inverse probability of treatment weighting (IPTW), pairwise algorithm (PA)^14,15, and overlap weight (OW)¹⁸ models were used. The choice of variables to generate the propensity score was consistent with the variables for multivariate analysis Model III. To evaluate the discriminative ability of the final multivariable model, we constructed a receiver operating characteristic (ROC) curve for 30-day mortality (Supplementary Fig. S2). The area under the curve (AUC) was 0.814, indicating good predictive performance Propensity score analyses revealed that an increased risk of death within 30 days was significantly associated with H2RAs exposure. The results of the propensity score analysis are shown in Table 3.

Table 3.

Associations between H2RA use and outcomes in the multivariable analysis and propensity score analyses.

Analysis	30-day mortality (%)	P
Multivariable analysis-hazard ratio(95%Cl)	1.32 (1.09–1.6)	0.005
Adjusted for propensity score	1.34 (1.09–1.63)	0.004
Adjusted for propensity score matching	1.46 (1.22–1.75)	< 0.001
With inverse probability weighting	1.63 (1.42–1.87)	0.006
With pairwise algorithmic	1.28 (1.01–1.62)	0.017
With overlap weight	1.29 (0.98–1.7)	0.013

Subgroup analyses

We also performed stratified analyses according to sex, age, BMI, STEMI and so on. Across all strata, the association between H2RA exposure and 30-day mortality remained directionally consistent. We found that the relationship between H2RAs and 30-day mortality was statistically significant across all subgroups. However, for 30-day mortality, we observed a significant difference in the STEMI subgroup (interaction P < 0.001). This suggests that the impact of H2RA exposure on mortality may vary depending on whether patients have STEMI. Furthermore, to evaluate the robustness of our findings, we conducted propensity score matching analysis, adjusting for primary confounding covariates between the H2RA groups. Consistent results were observed even after adjusting for multiple factors, including the use of dual antiplatelet therapy (DAPT), which is particularly relevant given that acid-suppressive therapy is typically prescribed in STEMI patients to reduce the risk of major gastrointestinal bleeding while on dual DAPT. The subgroup analysis results before and after propensity score matching were shown in Fig. 2.

Fig. 2.

Subgroup analysis before and after propensity score matching.

Exposure to H2RAs versus exposure to PPIs in relation to 30-day mortality in patients with AMI

In addition to the analysis of H2RA exposure, we analyzed the 30-day mortality rate of AMI patients exposed to PPIs. In the overall population, the hazard ratios for the use of PPIs after multivariate analysis adjusted for confounding factors and PSM analysis were 1.8, 1.26 and 1.39, respectively (Supplementary Table S3, S5 ~ S8and Fig. S2). Subgroup analyses for PPI exposure were also performed and are presented in Supplementary Tables S9, S10. To further explore 30-day mortality rate of both medications, we cross-grouped the drug-using population. The basic population characteristics of the groups are shown in the Supplementary materials (Table S11). The baseline characteristics of the four groups of people varied greatly at all levels. In this regard, we screened out patients who only used H2RAs or PPIs, matched the screened patients for propensity scores, and compared the matched data for survival curves. Moreover, to ensure the stability of the analysis, we also deleted all missing values and conducted the analysis. Survival analysis revealed that after PSM adjustment, patients with AMI who received H2RAs had a higher 30-day mortality rate than those used PPIs. The results of the data analysis of multiple imputation of the screened covariates with the use of PPI as the dependent variable and 30-day mortality as the outcome variable (P = 0.012) and the results of the data analysis of the deletion of missing covariates (P = 0.0065) were statistically significant. The results are shown in Supplementary Tables S12, S13 and Figs. S2 ~ 3.

Discussion

Summary of findings

In this study, we used the MIMIC-IV database to explore the relationship between H2RAs medication and 30-day mortality in critically ill AMI patients. Our analysis included 4252 patients, with 1557 exposed to H2RAs. After adjusting for multiple confounders, H2RAs administration was significantly associated with a 32% increased risk of 30-day mortality (HR = 1.32, 95% CI: 1.09-1.6, p = 0.005). This association remained consistent across various sensitivity analyses, including PSM and IPTW. Notably, the increased mortality risk was more pronounced in STEMI patients, indicating a potentially greater adverse effect of H2RAs in this subgroup. Moreover, compared with PPIs, H2RAs medication exihibited a greater risk in AMI, as survival analysis revealed statistically significant results for H2RAs versus PPIs after PSM. However, it should be noted that H2RAs users in the baseline data had fewer severe illnesses and lower mortality rates, possibly due to clinicians’ empirical choices. However, even after adjusting for these potential biases, H2RAs administration was still significantly associated with increased 30-day mortality, suggesting a potentially adverse impact on AMI patient prognosis.

Comparison with previous studies

A cohort study by Huang et al.²⁰ demonstrated that H2RAs administration were associated with reduced all-cause mortality in critically ill heart failure patients, comparable to β-blockers (HR = 0.72, 95% CI 0.58-0.89)¹⁹. In contrast, our observation showed an increased mortality in AMI patients, suggesting context-dependent effects of H2RAs. This discrepancy may reflect differences in underlying pathophysiology, as heart failure patients often exhibit chronic sympathetic activation where H2RAs-mediated histamine modulation could confer benefits²⁰, Conversely, as gastric acid secretion inhibitors (SGAS), H2RAs raise gastric pH, blunt oral nitrite-induced hypotension, decrease gastric NO formation, and lower circulating nitrosylated substances (RXNO), potentially promoting acute ischemic events²¹.

The association between acid suppressants and cardiovascular risk has been extensively debated. A nationwide study conducted by Sehested et al. (2018) revealed that prolonged use of proton pump inhibitors (PPIs) is linked to an elevated risk of myocardial infarction (MI), with the risk increasing in a dose-dependent manner. Specifically, individuals using high doses of PPIs exhibited a 43% greater risk of MI, indicated by a hazard ratio (HR) of 1.43 (95% confidence interval: 1.30-1.57). In contrast, low-dose PPI usage did not show a statistically significant increase in risk. These findings highlight the critical role of PPI dosage in relation to cardiovascular risk²². Whereas Hsu et al. (2020) reported no increased AMI risk with H2RAs compared with PPIs in a large insurance claims analysis (OR = 0.98, 95% CI 0.92-1.05)²³. Our results extend these observations by demonstrating a 32% mortality risk increase, specifically with H2RAs, in AMI patients, even after strick adjustment for confounding factors.

Moreover, our subgroup analysis revealed increased mortality in STEMI patients receiving H2RAs (HR = 2.68, 95% CI 1.7-4.21), which was consistent with a retrospective nationwide cohort study by Jimin Jeon et al. (2022): they found that patients with MI treated with H2RAs increased the risk in developing opneumonia (HR=1.50, 95%CI 1.16–1.93), which increases the mortality in MI patients ²⁴. The relatively neutral association of H2RAs with PPIs in our study (HR = 1.26, 95% CI 1.00-1.58) supported the hypothesis that H2RAs and PPIs exert distinct cardiovascular effects through different pathways.

Potential underlying mechanisms

The mechanisms underlying H2RAs-associated mortality in AMI patients may involve multiple pathways. First, The influence of H2 receptors on heart function is dependent on the specific context, with their effects differing based on the type of stress applied. Under hypoxic conditions, an increased expression of H2 receptors seems to provide a protective effect on atrial function, likely by augmenting the positive inotropic response associated with histamine. However, in scenarios involving ischemia–reperfusion injury and LPS-induced dysfunction of myocardial contractility, heightened H2 receptor expression may yield negative outcomes. These adverse effects could be related to the desensitization of histamine receptors, which might contribute to a further decline in cardiac performance²⁵. Second, acid suppressants may alter drug metabolism, which is critical in AMI management. H2RAs such as cimetidine inhibit cytochrome P450 enzymes, potentially reducing clopidogrel activation⁷. While our analysis adjusted for antiplatelet use, residual pharmacokinetic interactions persisted. This is particularly relevant given the observed interaction between H2RAs and STEMI outcomes, where effective antiplatelet therapy is paramount. Third, gastric pH modulation may impair nitrate bioconversion. Sanches-Lopes et al. (2020) demonstrated that acid suppression reduces systemic nitrite bioavailability by 40-60%, diminishing the antihypertensive effects of nitrates (p < 0.001)²¹. In AMI patients requiring vasodilation, this might exacerbate myocardial oxygen supply–demand imbalance. Furthermore, chronic acid suppression has been associated with endothelial dysfunction through asymmetric dimethylarginine accumulation, potentially accelerating atherosclerotic processes²⁶. Finally, confounding by indications remains a critical consideration. As noted by Blackburn et al. (2010), acid-suppressing drugs are frequently prescribed preemptively before ischemic events, creating temporal confounding in observational studies²⁷. Our sensitivity analyses using propensity score matching and inverse probability weighting mitigated this concern, yet residual confounding from unmeasured severity indicators (e.g., occult bleeding risk) may persist.

Strengths and limitations

Our study has notable methodological strengths that increase the validity of the findings. First, the MIMIC-IV database has a comprehensive repository of granular clinical data from over 40,000 ICU admissions, allowing this study to adjust for a wide array of potential confounders, including demographic factors, laboratory parameters, comorbidities, and pharmacotherapies⁹. It also enables robust adjustment for multiple potential confounders spanning demographics, laboratory parameters, comorbidities, and pharmacotherapies. Second, the application of advanced causal inference techniques, including propensity score matching, inverse probability weighting, and doubly robust estimation, increased the reliability of our conclusions by addressing indication bias and treatment heterogeneity¹². Third, the consistency of results across multiple sensitivity analyses (e.g., complete-case analysis, subgroup stratification, and machine learning-based imputation) significantly support the observational results regarding the association between H2RAs and mortality²⁸.

Nevertheless, several limitations warrant consideration. First, as an observational study, residual confounding from unmeasured variables (e.g., socioeconomic status, dietary habits, or undocumented medication adherence) may persist despite comprehensive adjustments²⁹. Second, the single center design and exclusive focus on U.S. ICUs limit generalizability to non-Western populations, where AMI management protocols and H2RAs prescribing patterns may differ substantially^30,31. Third, while we applied strict time criteria for drug exposure assessment, our analyses did not consider changes in H2RAs dosing regimens, especially drug doses, while analyses lacking dose data lacked assessment of potential threshold effects, and the lack of dosage data precludes evaluation of a potential threshold effect³². Fourth, the MIMIC-IV database documents only all-cause in-hospital and 30-day mortality; cause-specific deaths (e.g., cardiovascular events, gastrointestinal bleeding, sepsis) could not be ascertained. Moreover, no data on long-term survival or non-fatal outcomes beyond 30 days are available.

Clinical Implications

Our study revealed that the use of H2RA in critically ill AMI patients was associated with increased 30-day mortality. Given that both H2RAs and PPIs medication may increase the risk of AMI, clinicians should carefully evaluate the risk–benefit profile of these drugs, especially in patients with STEMI. While our findings highlight the potential risk of H2RAs medication in AMI patients, it is critical to explore alternative strategies to prevent stress ulcers in this high risk population. Considering nonpharmacological strategies, such as careful monitoring of these patients and the implementation of preventive measures such as early enteral nutrition, may be viable options. For pharmacological alternatives, sucralfate is considered a safer option because it does not elevate gastric pH and carries a significantly lower risk of ICU-acquired pneumonia compared with PPIs (OR 1.65, 95% CI 1.20-.27) and H2RAs (OR 1.30, 95% CI 1.08-1.58), while showing no increase in all-cause mortality^33–35.

Conclusions

This study demonstrates that H2RAs administration in critically ill patients with AMI is associated with an increased risk of 30-day mortality. Our findings highlight the need for further research to uncover the potential risks of H2RAs in this vulnerable patient population and to guide clinical decision-making. Future studies should focus on elucidating the mechanisms underlying this association and evaluating the safety and efficacy of alternative acid-suppressive therapies in AMI patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

H2RAs

Histamine H2-receptor antagonist

AMI

Acute myocardial infarction

STEMI

ST-segment elevation myocardial infarction

BMI

Body mass index

SOFA

Sequential organ failure assessment score

SBP

Systolic blood pressure

DBP

Diastolic blood pressure

BUN

Urea nitrogen

INR

International normalized ratio

PT

Prothrombin time

WBC

White blood cell

PTT

Partial thromboplastin time

ALT

Alanine aminotransferase

ALP

Alkaline phosphatase

AST

Aspartate aminotransferase

GGT

Glutamyl transpeptidase

Ph

Blood potential of hydrogen

SO2

Blood oxygen saturation

PO2

Partial pressure of oxygen

PCO2

Partial pressure of carbon dioxide

P/F

Ratio of the partial pressure of oxygen in arterial blood to the fraction of inspired oxygen

CCB

Calcium channel blocker

RAAS

Renin angiotensin aldosterone system

PPIs

Proton pump inhibitors

HMG-CoA

Hydroxymethylglutaryl-CoA reductase inhibitor

ICU

Intensive care unit

LOS

Length of stay

SGAS

Gastric acid secretion inhibitors

RXNO

Nitrosylated substances

SMD

Standardized mean difference

Author contributions

All the authors contributed significantly to the work reported in this study. G.S. is responsible for the research design, drafting and writing of the article. C.W. is responsible for conceiving and revising the manuscript. W.T is responsible for data analysis and interpretation. S.D. and Q.J. is responsible for data extraction, collection and analysis, and each author participates in the revision or critical comment of the article; provides final approval for the release to be released; agrees with the journal in which the article is submitted; and accepts responsibility for all aspects of the work

Funding

This work was supported by the Zhanjiang Science and Technology Development Special Projects (No. 2021A05086, 2021A05101, and 2022A01147), the High-level Talent Startup Fund from the Second Affiliated Hospital of Guangdong Medical University (No. 21H03, 23H02 and 23H03), and the Special Project for Clinical and Basic Sci & Tech Innovation of Guangdong Medical University (GDMULCJC2024048). The funding organization had no involvement in the study’s design; data collection, analysis, or interpretation; or the decision to publish the findings.

Data availability

The data supporting the findings of this study are available from the supplementary files.

Declarations

Competing interests

The authors declare that there are no conflicts of interest.

Ethical approval

This study was approved by the Medical Research Ethics Committee of the Second Affiliated Hospital of Guangdong Medical University (No. YJKT2025-026-01). We confirm that all methods were conducted in accordance with relevant guidelines and regulations.