Our cohort investigation employed a retrospective design that matched subjects using propensity scores, utilizing the MIMIC-IV electronic database (version 3.1) as its foundation. A comprehensive list of patients from MIMIC-IV was assembled, encompassing all medical record numbers linked to individuals admitted to either the intensive care unit (ICU) or the emergency department from 2008 to 2022. This included a total of 546,028 hospitalizations and 94,458 distinct stays in the ICU at Beth Israel Deaconess Medical Center located in Boston, MA, USA. The corresponding author (Xin Xue) successfully completed the exam for the Collaborative Institutional Training Initiative (CITI) program and received a certificate (Record ID: 53098597). The research adhered to the ethical principles set forth in the Declaration of Helsinki and complied with the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [16]. Given that the data utilized in this clinical research were sourced from public databases, the Ethics Review Committee at the Second Hospital of Jilin University classified it as exempt from an ethical review.

Acute heart failure was defined by pulmonary congestion, systemic congestion, and inadequate perfusion of tissues and organs. It was categorized into four primary types: acute decompensated heart failure, acute pulmonary edema, cardiogenic shock, and isolated right ventricular failure.

In the MIMIC-IV database, patients meeting inclusion criteria for AHF were identified based on the International Classification of Diseases 9th edition (ICD-9) codes (42833, 42823, 42843, 42831, 42821, 42841) and the ICD-10 codes (I5033, I5023, I5021, I5031, I5043, I5041). Patients under 18 years of age, those who had been given magnesium sulfate prior to ICU admission, or instances where blood magnesium data was unavailable were excluded from the study. Furthermore, we limited our analysis to the initial admission records of patients who had multiple ICU admissions.

The use of intravenous magnesium sulfate during the ICU stay was the exposure being examined, without any restrictions. Data concerning the administration of magnesium sulfate were gathered from the prescription table. The main outcome measured was mortality from any cause within 28 days. Additionally, secondary outcomes encompassed mortality from any cause at 90 days and at 365 days.

Data was extracted utilizing Structured Query Language (SQL) across five distinct categories: (1) Demographics, which encompassed age, gender, race, and types of intensive care units (ICUs); (2) Physical assessments and laboratory tests, detailing heart rate, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), respiratory rate, body temperature, and magnesium status at the onset of acute heart failure (including hypomagnesemia, hypermagnesemia, and normomagnesemia), along with white blood cell count (WBC), platelet count, glucose levels, creatinine levels, magnesium levels, potassium levels, sodium levels, calcium levels, chloride levels, phosphate levels, albumin, aspartate aminotransferase, and alanine aminotransferase levels; (3) Comorbidities, which included QT prolongation syndromes, hypertension, diabetes, myocardial infarction, cerebrovascular diseases, peripheral vascular diseases, dementia, chronic pulmonary diseases, renal disease, severe liver disease, rheumatic diseases and malignant cancer; (4) Special treatments, consisting of mechanical ventilation, intravenous diuretics (furosemide), vasodilator (sodium nitroprusside or nitroglycerin), vasopressor (norepinephrine, dopamine or dobutamine), digoxin, statins, angiotensin converting enzyme inhibitor (ACEI), angiotensin II receptor blockers (ARB), aspirin, $\beta$-blockers, and adenosine diphosphate (ADP) receptor antagonists; (5) Clinical scores, including Charlson comorbidity index, Acute Physiology and Chronic Health Evaluation Ⅲ (APACHE Ⅲ), Simplified Acute Physiology Score Ⅱ (SAPS Ⅱ) and Oxford acute severity of illness score (OASIS).

This research utilized a retrospective approach, dividing participants into two categories: the group that received magnesium sulfate (designated as the magnesium sulfate group) and the group that did not receive it (referred to as the no-magnesium sulfate group). The missing values for each variable were estimated through multiple imputation methods. The presence of multicollinearity among covariates was evaluated by examining the variance inflation factor ($>$10) and tolerance values ($>$0.1).

Data that demonstrated a normal distribution, as verified by the Kolmogorov-Smirnov test, were reported as mean $\pm$ standard deviation; a t-test was conducted to assess differences between the two groups. For data exhibiting skewed distributions, results were shown as median (interquartile range), with group comparisons performed using a rank sum test. Count data were indicated by the number of cases (n) and their corresponding percentages (%), analyzed through the chi-square test. A multivariate cox proportional hazards model was constructed to evaluate the hazard ratio (HR) and its associated 95% confidence interval (CI) concerning magnesium sulfate treatment and the risk of all-cause mortality in patients with acute heart failure. The Kaplan–Meier approach was employed to study the cumulative incidence of 28-day all-cause mortality, with comparisons made through the log-rank test. For secondary outcomes, HR and 95% CI were determined using multivariate cox regression models.

Fig. 1 depicts the procedure for patient selection. The MIMIC IV(3.1) database contains 94,458 ICU hospitalization records. Among them, 63,419 participants were excluded due to non-acute heart failure, 9248 due to missing blood magnesium data, 10,696 due to repeated hospitalization, and 1064 due to magnesium sulfate administration before admission to the ICU. After excluding records that did not meet the inclusion criteria, 10,031 patients were included in the matching cohort, with 6794 (67.7%) receiving intravenous magnesium sulfate while in the ICU. The matched cohort consisted of 6240 patients, 3120 in each group.

Fig. 1.

The flow chart of participant selection. ICU, intensive care unit; MIMIC-IV, Medical Information Mart for Intensive Care IV; AHF, acute heart failure.

Table 1 presents the baseline characteristics prior to and following matching. Within the full cohort, serum magnesium concentrations were markedly reduced in the intravenous magnesium sulfate group when compared to the non-intravenous magnesium sulfate group (p 0.001). The occurrence of hypomagnesemia was notably more prevalent in the magnesium sulfate group (1.32%, 90/6794) in contrast to the non-intravenous group (0.53%, 17/3237) (p 0.001). Patients who received magnesium sulfate were younger and had more hospitalizations in the cardiovascular care unit, had a lower Charlson comorbidity index, and were more likely to receive diuretics via intravenous infusion, vasodilator, or vasopressors. The process of matching enhanced the balance of variables, achieving an absolute standardized mean difference (SMD) of less than 0.10, though some imbalances persisted in variables that were not included in the propensity score matching (Table 1).

The rate of all-cause mortality over a 28-day period was found to be 13.2% (413/3120) in the group that received magnesium sulfate, compared to 15.8% (493/3120) in the group that did not receive the treatment. In the matched cohort dataset, the variables that demonstrated statistical significance (p 0.05) during the univariate analysis were considered for inclusion in the multivariate analysis to ensure thorough adjustment (Supplementary Table 1). Fig. 2 presents the Kaplan–Meier curve, which reflects the all-cause mortality rates over 28 days related to the administration of magnesium sulfate in the matched cohort. Findings from the multivariable analysis indicated an association between the use of magnesium sulfate and a decreased all-cause mortality rate throughout the 28 days (HR, 0.81; 95% CI, 0.71–0.93; p = 0.004), corroborated by the results from the univariable analysis (HR, 0.81; 95% CI, 0.71–0.93; p = 0.002).

Fig. 2.

Kaplan–Meier curve for 28-day all-cause mortality according to magnesium sulfate use in the matched cohort. CI, confidence interval.

The findings of subgroup analyses regarding 28-day all-cause mortality within the matched cohort are illustrated in Fig. 3. The analysis revealed a notable interaction between sex and the administration of magnesium sulfate among patient subgroups categorized by gender (p for interaction = 0.025). In female patients, the relationship between magnesium sulfate administration and mortality was significantly more pronounced when compared to males (HR, 0.70; 95% CI, 0.57–0.86; p = 0.001). The association between magnesium sulfate use and mortality was similar among patient subgroups stratified by age, Charlson score, and intensive care unit type (p for interaction $>$0.05). In all subgroups, patients who received magnesium sulfate had a lower risk of all-cause death than those who did not receive magnesium sulfate. In all subgroups, patients who received intravenous magnesium sulfate exhibited a lower risk of all-cause mortality compared to those who did not receive intravenous magnesium sulfate (HR 1).

Fig. 3.

Subgroup analyses for 28-day all-cause mortality in the matched cohort. yr, year; HR, hazard ratio.

The E-value suggests that the link between the use of magnesium sulfate and all-cause mortality is strong (E-value: 1.77; upper confidence interval: 1.36). This implies that for the observed relationship between magnesium sulfate administration and overall mortality to be impacted, unmeasured confounders would need to have a risk ratio of at least 1.77. Specifically, only those confounders that exceed a risk ratio of 1.77 would have the potential to alter the findings of the correlation analysis (E-value for point estimate: 1.77; E-value for confidence interval: 1.36) (Supplementary Fig. 1). Additionally, we performed sensitivity analyses on the full dataset to assess the robustness of the results obtained from the matched cohort. In the subset that received magnesium sulfate, the mortality rate over 28 days was 13.1% (892/6794), compared to 15.9% (514/3237) in the non-treatment group. The administration of magnesium sulfate was associated with a reduced rate of all-cause mortality at 28 days, as demonstrated by both univariable analysis (HR, 0.80; 95% CI, 0.72–0.90; p 0.001) and multivariable analysis (HR, 0.77; 95% CI, 0.69–0.87; p 0.001) (Supplementary Table 2).

The 90-day mortality rate from all causes in the group receiving magnesium sulfate was noted to be 21.4% (667/3120), whereas the group that did not receive magnesium sulfate recorded a rate of 23.5% (733/3120). In multivariable analysis, the administration of magnesium sulfate was associated with a decreased risk of all-cause mortality at 90 days (HR, 0.88; 95% CI, 0.79–0.99; p = 0.029), a result that was further corroborated by univariable analysis (HR, 0.88; 95% CI, 0.80–0.98; p = 0.021). Regarding the 365-day all-cause mortality rate, the magnesium sulfate group showed a rate of 32.4% (1010/3120), in contrast to a 36.1% (1125/3120) rate for the group not receiving it. Once more, the use of magnesium sulfate was associated with a reduced 90-day all-cause mortality risk in multivariable analysis (HR, 0.88; 95% CI, 0.80–0.96; p = 0.006) and was also validated in univariable analysis (HR, 0.87; 95% CI, 0.80–0.95; p = 0.001) (Table 2).

Research has previously examined the link between magnesium supplementation and the likelihood of all-cause mortality across different populations. A retrospective cohort analysis conducted by Gu et al. [17] indicated that the use of intravenous magnesium sulfate significantly decreases mortality rates in critically ill patients suffering from sepsis. In a similar vein, Barbosa et al. [19] reported that intravenous magnesium sulfate also diminishes the risk of death among critically ill individuals experiencing acute kidney injury. Although Barbosa et al.’s investigation [19] was primarily restricted to a population with low magnesium levels, the findings from Gu et al. [17] illustrated that the positive impact of intravenous magnesium sulfate on mortality was not reliant on serum magnesium concentrations. In cases of heart failure, there is a strong correlation between magnesium and the advancement of the disease [20]. Nonetheless, the exploration of the connection between magnesium supplementation and the all-cause mortality risk remains under-researched. A propensity score-matched cohort study by Adamopoulos et al. [21] disclosed that in patients diagnosed with chronic heart failure, serum magnesium levels falling below 2 mEq/L were linked to higher cardiovascular mortality. This study, however, did not explore the impact of intravenous magnesium sulfate supplementation on mortality risk and was limited to patients with chronic heart failure. Research by Zhao et al. [22] indicated that individuals with higher dietary magnesium intake exhibited a reduced likelihood of developing congestive heart failure. The focus of their investigation was on dietary magnesium supplementation, leaving the relationship between intravenous magnesium supplementation and mortality unaddressed. A systematic review that analyzed serum magnesium levels and heart failure outcomes included 13,539 patients diagnosed with heart failure with reduced ejection fraction (HFrEF) and assessed how serum magnesium levels influenced cardiovascular mortality, all-cause mortality, and cardiovascular morbidity. Among the studies reviewed, hypomagnesemia was recognized as an independent risk factor for cardiovascular death, including instances of sudden cardiac death [23]. Nevertheless, this review failed to assess whether intravenous magnesium sulfate enhances the outcomes for patients with heart failure.

Our research involving individuals experiencing acute heart failure revealed that supplementation with intravenous magnesium sulfate lowered the risk of mortality from all causes in both a matched cohort and the general population. Furthermore, the use of intravenous magnesium sulfate correlated with a decrease in mid-term (90 days) and long-term (365 days) mortality rates. Taken together, these findings reinforce the link between magnesium treatment and improved outcomes for patients suffering from heart failure.

Moreover, our research revealed variations based on gender in how magnesium sulfate supplementation relates to the risk of overall mortality among patients experiencing acute heart failure. In particular, the link between intravenous magnesium sulfate supplementation and all-cause mortality risk seemed to be more significant in female patients. These results are consistent with the findings of Zhang et al. [24], who likewise noted that magnesium intake from diet lowers the mortality risk associated with congestive heart failure in women, whereas such a connection was not found in men.

Magnesium, the second most commonly found intracellular cation after potassium, is involved in over 600 enzymatic processes, such as energy metabolism and protein synthesis [25]. The role of magnesium ions is significant in myocardial excitation-contraction coupling [26]. Often recognized as a natural opponent to calcium, Mg²⁺ competes with Ca²⁺ for attachment sites on proteins and calcium transporters [7, 27]. The influence of Mg²⁺ on cardiomyocytes largely stems from its effect on calcium mobilization. It binds to calmodulin, troponin C, and parvalbumin, where a reduction in Mg²⁺ concentration can result in alterations to the free Ca²⁺ fraction [28]. Furthermore, Mg²⁺ has the ability to impact the key calcium-transporting proteins within cardiomyocytes, acting as a substrate alongside Adenosine Triphosphate (ATP) for cardiac Ca²⁺-ATPases and altering the affinity of the Na⁺-Ca²⁺ exchanger type 1 (NCX1) for calcium. However, empirical studies investigating the effects of Mg²⁺ on NCX1 and Sarco/Endoplasmic Reticulum Calcium ATPase (SERCA) activity are limited, with most existing data being derived from modeling and in vitro assessments [29, 30]. Regardless, maintaining ideal levels of [Mg²⁺] in cardiac cells is crucial for optimal heart function. Research indicates that magnesium supplementation can mitigate oxidative stress damage and provide protective benefits [31, 32]. Both the disruption of cardiac excitation-contraction coupling and oxidative stress are significant contributors to heart failure. The administration of intravenous magnesium sulfate may potentially affect these physiological mechanisms, thereby offering a positive effect on heart failure patient outcomes.

Magnesium sulfate primarily serves in the treatment of moderate to severe gestational hypertension, eclampsia, hypomagnesemia, torsade de pointes, and pediatric convulsions. Currently, it is not regarded as a standard intervention for acute heart failure. Despite the absence of randomized controlled trials (RCTs) assessing the use of magnesium sulfate specifically for acute heart failure, it was identified last century that a sudden increase in serum magnesium levels after intravenous magnesium therapy can reduce the incidence of ventricular arrhythmias in patients experiencing congestive heart failure [33]. These ventricular arrhythmias significantly contribute to mortality rates in heart failure patients. Findings from the second Leicester Intravenous Magnesium Intervention Trial (LIMIT-2) indicate that intravenous magnesium sulfate is a straightforward, safe, and broadly applicable treatment option. Its effectiveness in lowering early mortality associated with myocardial infarction is similar to that achieved with thrombolytic or antiplatelet therapies, yet it operates independently of these treatments. Myocardial infarction stands as a leading factor in the onset of acute heart failure [34]. This research demonstrates that intravenous magnesium sulfate can reduce short-term, mid-term, and long-term all-cause mortality for patients suffering from acute heart failure, highlighting the necessity for additional prospective RCTs to explore this further.

The current research possesses several notable limitations. Firstly, due to its design being retrospective and observational, the outcomes may be subject to residual bias and unmeasured confounding variables, even with the use of propensity score matching and multivariable analyses. Nevertheless, the E-value indicates that the findings from the association analysis are substantial. Secondly, establishing a causal connection was not feasible, necessitating additional RCTs. Thirdly, as the severity of patients’ clinical conditions may change over time, this can potentially affect the results. Fourthly, the guidelines for starting and stopping magnesium sulfate treatment were not recorded. Lastly, while the safety profile of magnesium sulfate was not evaluated in this research, it is important to note that adverse effects from magnesium supplementation are infrequent outside of critical care environments.