Statin Therapy May Be Associated with Reduced Mortality Risk of Sepsis Patients: A Retrospective Study from the MIMIC-IV Database


 BackgroundIt is controversial whether statin therapy is beneficial for sepsis patients. A large retrospective cohort study was conducted to evaluate the association between statin therapy and mortality in sepsis patients.MethodsAdult (≥18 years) sepsis patients were enrolled and divided into two groups: the statin group and the no-statin group. Data including demographic features, vital signs, laboratory tests, and comorbidities from MIMIC-IV v1.0 were extracted. Delirium was assessed via the Confusion Assessment Method for the ICU (CAM-ICU). Ninety-day mortality, 28-day mortality and the incidence of delirium after statin therapy were evaluated using multivariable logistic analysis, the PSM model and subgroup analysis based on univariate analysis. ResultsIn univariate analysis and multivariable logistic analysis, statin therapy showed a significant association with both reduced 28-day and 90-day mortality (90-day mortality: OR 0.58, 95% CI: 0.46-0.72, p<0.001; 28-day mortality: OR 0.47, 95% CI: 0.37-0.60, p<0.001), while no relationship was found between statin therapy and delirium (OR 0.92, 95% CI: 0.49-1.72, p=0.787). In further PSM model and subgroup analyses or sensitivity analysis, consistent results were shown. ConclusionStatin therapy is significantly associated with 28-day and 90-day mortality without decrease in the incidence of delirium.


Abstract Background
It is controversial whether statin therapy is bene cial for sepsis patients. A large retrospective cohort study was conducted to evaluate the association between statin therapy and mortality in sepsis patients.

Methods
Adult (≥18 years) sepsis patients were enrolled and divided into two groups: the statin group and the nostatin group. Data including demographic features, vital signs, laboratory tests, and comorbidities from MIMIC-IV v1.0 were extracted. Delirium was assessed via the Confusion Assessment Method for the ICU (CAM-ICU). Ninety-day mortality, 28-day mortality and the incidence of delirium after statin therapy were evaluated using multivariable logistic analysis, the PSM model and subgroup analysis based on univariate analysis.

Conclusion
Statin therapy is signi cantly associated with 28-day and 90-day mortality without decrease in the incidence of delirium. Background Sepsis is de ned as life-threatening organ dysfunction caused by a dysregulated host response to infection (Sepsis-3) [1]. It is a major challenge worldwide due to its high morbidity and mortality. The incidence of sepsis continued to increase, although the total in-hospital mortality fell in the United States from 1979 through 2000 [2]. Epidemiological studies of Spain showed a similar tendency from 2008 to 2017 [3]. In China, a recent clinical study of secondary analysis of a population-based database revealed that the standardized mortality rate was 67 cases per 100,000 population per year [4].
It is prevalently recognized that the in ammatory response plays a vital role in the progression of sepsis [5][6][7]. However, agents used to regulate the in ammatory cascade are rare, and most of them lack convincing evidence [8][9][10]. Since sepsis may cause an in ammatory and hypercoagulable state of endothelial cells [5], statins, well-known hydroxymethyl glutaryl coenzyme A (HMG-CoA) reductase inhibitors [11], were studied in the sepsis population to explore their potential role in the prevention and treatment of sepsis [12,13].
Over 20 million Americans, and millions more worldwide, take statins, such as atorvastatin, rosuvastatin, simvastatin, pitavastatin, pravastatin, uvastatin, and lovastatin, to lower cholesterol synthesis [14,15]. Interestingly, statins also have complex pleiotropic effects [12], which include multiple anti-in ammatory actions by inhibiting macrophages and neutrophils as well as endothelial cell activation via changes in the in ammatory signaling pathway [16]. Therefore, several studies have focused on whether statin therapy may affect sepsis mortality. In 2009, two observational studies reported that statin therapy was associated with reduced mortality in patients with infection or sepsis [17,18]. Some meta-analyses also revealed a protective role of statins in sepsis patients [19,20]. However, a small randomized clinical study failed to con rm a signi cant association between statin therapy and decreased sepsis mortality [21].
Of note, statins were reported to prevent delirium in critically ill patients [23], and the possible mechanism by which satins protect against delirium includes suppressing upregulated toll-like receptors (TLRs), reducing the expression of tumor necrosis factor-α (TNF-α) and interleukin 1β (IL-1β), and decreasing leukocyte adhesion [22]. In contrast, the MoDUS study established that early administration of statins did not attenuate the duration of delirium or coma in critically ill patients on mechanical ventilation [24]. Therefore, it remains controversial whether statin therapy is bene cial for sepsis patients. Here, we conducted a single-center, retrospective cohort study from the MIMIC-IV database to mainly evaluate the impact of statin therapy on mortality and delirium in patients with sepsis.

Data source
We performed a retrospective study using records from the Multiparameter Intelligent Motoring in Content was obtained for the original data collection. Therefore, the ethical approval statement and the need for informed consent were waived. One of the authors (SY) obtained access (Certi cation number 35931520) and extracted the data.

Inclusion and exclusion criteria
Adult (age≥18 years old) patients diagnosed with sepsis using International Classi cation of Disease, 10th Revision (ICD-10) according to the de nition of sepsis [1] and receiving statin therapy for more than 7 days were considered to be eligible for the study cohort. The exclusion criteria were as follows: (1) repeated admission to the ICU, (2) receiving statin therapy before ICU admission, and (3) patients whose baseline creatine kinase level was 10 times or ALT level was 8 times higher than the upper normal range.

Sample size
The sample size was constrained by the number of patients meeting the inclusion criteria and receiving one of the statins for more than 7 days.

De nitions
Sepsis is de ned as life-threatening organ dysfunction caused by a dysregulated host response to infection (Sepsis 3.0). That is, infectious patients with a SOFA score of 2 points or more are diagnosed with sepsis [1]. Statins include atorvastatin, rosuvastatin, simvastatin, pitavastatin, pravastatin, uvastatin, and lovastatin [15].

Data collection
Demographics, vital signs, disease severity, laboratory tests, comorbidities, medicines, and clinical outcomes, such as follow-up records were extracted from the database. The severity of sepsis was represented by the Simpli ed Acute Physiology Score II (SAPS II) [26] and Sequential Organ Failure Assessment (SOFA) score [27]. Comorbidities were de ned using ICD-10 codes [28]. Medicines mainly collected included some cardiovascular system related drugs, such as calcium channel blockers (CCBs), beta blockers, angiotensin converting enzyme inhibitors (ACEIs), angiotensin II receptor blockers (ARBs), thiazides, spironolactone, clopidogrel and aspirin.

Outcomes
The primary outcome was the association between statin therapy and 28-day mortality for sepsis patients. The secondary outcomes were 90-day mortality and the incidence of delirium.

Statistical analysis
Baseline characteristics for continuous variables are described as the median and interquartile range (IQR). Categorical variables are presented as the count and percentage (%). The Mann-Whitney U test, chisquare test, or Fisher's exact test was used for comparisons between the statin and no-statin groups.
Univariate and multivariable logistic analyses were conducted to explore the association between statin use and outcomes with odd ratios (ORs) and 95% con dence intervals (CIs). Adjustment confounders included the baseline characteristics listed in Table 1. In addition, we applied propensity score matching (PSM) to balance the baseline variables between groups. The propensity score revealed the likelihood of receiving statin therapy with a condition of prespeci ed confounders (Table 1). We matched patients based on their propensity score with a caliper less than 0.02, and the matching ratio was 1:1. After matching, the standard mean difference (SMD) for each covariate was calculated, and an SMD less than 0.1 was an indicator for balance.
In subgroup analyses, we examined the association of statin use and 90-day mortality strati ed by age, use of cardiovascular-related drugs, SOFA scores, and mechanical ventilation. A prespeci ed sensitivity analysis was performed under the following conditions: patients who died in the rst 24 h after ICU admission.
All p values (two-sided) less than 0.05 indicated statistical signi cance. All analyses were performed using SAS 9.4 and R Version 4.1.0.

Patient characteristics
An initial cohort of 35010 patients was eligible for the study. Among them, 7871 patients were repeatedly admitted to the ICU, 4694 had received statin therapy before ICU admission, and 1125 had baseline creatine kinase levels 10 times or ALT levels 8 times higher than the upper normal range. Finally, 21320 patients were included in our study. A total of 2158 (10.2%) patients received statin therapy for more than 7 days (statin group), and 19162 (89.8%) patients did not receive statins or had received statins for less than 7 days (no-statin group). Figure 1 is a ow chart describing the procedure for the selection of subjects.
For patients receiving statins therapy, age, weight, insurance, admitted type, SOFA score, SAPS II, mean arterial pressure (MAP), lactate level, platelets, urine output, blood urea nitrogen (BUN), creatine level, heart rate, uid input within 24 h, AST, ALT, total bilirubin, acute kidney injury (AKI), and comorbidities such as malignancy, diabetes, renal disease, cerebrovascular disease, liver disease, cardiovascular disease and cardiovascular-related drugs, except for spironolactone differed signi cantly compared with patients in the no-statin group. However, ethnicity, sex, rst GCS assessment, hemoglobin (Hb), white blood cells (WBCs), respiratory rate, temperature, SpO 2 , albumin level, positive blood culture, vasopressor use, mechanical ventilation, CRRT, RASS evaluation, and comorbidities such as hypertension showed no signi cant difference between the two groups (Table 1).     Figure 2. The SMD of each baseline variable is shown in Table 3.  When specially excluding patients who died in the rst 24 h after ICU admission (379 patients excluded), we determined that statin therapy exhibited signi cant association with both decreased 28-day and 90day mortality and no association with the incidence of delirium (Table 5).

Discussion
In this retrospective cohort study, we found that statin therapy more than 7 days was signi cantly associated with decreased 28-and 90-day mortality compared with patients who did not receive statins or had received statins for less than 7 days.
Statins are widely used in elderly people to reduce cholesterol, and drugs may soon become even more popular [14]. However, statins are not common prescriptions for patients with critically ill conditions, such as sepsis. In an early observational study, the authors found that 33% of sepsis patients were treated with statins before ICU admission [29], while only 26% of sepsis patients received statins after ICU admission [30]. In our study, we reported that only 2158 patients, which accounted for 10.1% of all eligible patients, received statin treatment more than 7 days, which indicates that most sepsis patients in the ICU did not or could not receive statins due to various subjective and objective reasons. For example, sepsis patients with unstable hemodynamics or undergoing abdominal surgery might have to prohibit eating and drinking, and clinicians may neglect or disregard prescribing statins for patients due to their undetermined effects.
In addition to inhibiting cholesterol and isoprenoids, statins have also been proven to have antioxidant effects via inhibition of isoprenoid formation and nicotinamide adenine dinucleotide phosphate oxidase [31] and anti-in ammatory properties with reduction of C-reactive protein and nuclear factorkappa B [32], which might be bene cial to sepsis prognosis. Other pleiotropic effects include immunomodulation, normalization of sympathetic out ow, plaque stabilization and suppression of the coagulation cascade or platelet aggregation [12]. The powerful effects of statins independent of lowering lipids and lipoproteins have piqued the interest of clinicians. In our study, we found that statin therapy for more than 7 days was greatly associated with reduced 28-day and 90-day mortality, which was consistent with several prospective observational studies [33,34] and a meta-analysis [19]. In our subgroup analysis of this study, we also reported that sepsis patients who received statins for more than 7 days had an improved prognosis regardless of age, severity of the disease (SOFA), whether taking cardiovascular-related drugs or receiving mechanical ventilation at the same time. The sensitivity analysis showed a consistent tendency, which indicated that our results were relatively stable and reliable. However, there are also some meta-analysis reports that included randomized clinical trials and concluded that statin therapy did not improve the survival of sepsis patients [35][36][37]. In addition, researchers demonstrated that there was no difference in changes in interleukin-6 (IL-6) with statin therapy in a randomized trial [38]. Previous studies included fewer subjects or populations without sepsis [19,33,34]. Some studies concentrated on the impact of prior statin therapy before ICU admission [39][40][41], which constrained the clinical guiding role of stains, since we could not predict whether and when one person would be hospitalized in the ICU. Unfortunately, there are no records on some in ammatory mediators, such as C-reactive protein (CRP), procalcitonin (PCT), IL-6 and IL-1 in the MIMIC-IV database, and we could not further analyze the relationship between statins and in ammatory cytokines to explore the potential anti-in ammatory effects of statins on mortality reduction. Speci cally, we included patients taking statins for more than 7 days after ICU admission to avoid accidental results.
Further large randomized controlled trial and mechanistic experiments are needed to unveil the mystery of statins in the sepsis population.
In this study, we did not nd any signi cant association between statin therapy and delirium. Delirium occurs early in critically ill patients, and there are multiple etiologies of delirium, including sepsis, electrolyte disturbance, kidney injury, exposure to sedation and analgesia [24]. Mechanistically, sepsisassociated delirium (SAD) occurs when a combination of neuroin ammation and disturbances in cerebral perfusion, blood brain barrier (BBB) and neurotransmission exist [42]. Consistent with our research, the MoDUS study, a prospective, randomized, double-blind and placebo-controlled trial, also found no evidence that statins attenuate delirium and coma in critically ill patients [24]. By analyzing serum CRP concentrations, the authors revealed that no difference was found between the statin therapy group and the placebo group. In addition, it is possible that the pleiotropic effects of statins were overwhelmed by other potential confounding factors [43]. The timing, duration, and dosage of statin administration in sepsis patients remain unknown, and require further research.
Our study has several limitations. First, we did not focus on the speci c kinds or dosages of statins. Different kinds of statins may have different effects in patients with sepsis. Some previous studies have shown that simvastatin at doses as high as 80 mg daily maintained a high therapeutic range, and the median dose was 40 mg [23]. Second, limited by retrospective study of a database, we had no access to learn about the prior statin therapy of our included subjects. Therefore, the survival advantage of continued statin therapy in prior statin users may cause type I error. The in uence of the duration of previous statin use or cessation time is uncertain in sepsis conditions.

Conclusions
Statin therapy is signi cantly associated with 28-and 90-day mortality, without reducing the incidence of delirium. More high-quality laboratory experiments and multicenter clinical trials are needed to con rm the bene ts of statin administration for sepsis treatment. Availability of data and materials

Abbreviations
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.   Clinical outcomes adjusted for confounders or after PSM.