key: cord-0909877-rbetn5x5
authors: Saeed, Omar; Castagna, Francesco; Agalliu, Ilir; Xue, Xiaonan; Patel, Snehal R.; Rochlani, Yogita; Kataria, Rachna; Vukelic, Sasa; Sims, Daniel B.; Alvarez, Chikezie; Rivas‐Lasarte, Mercedes; Garcia, Mario J.; Jorde, Ulrich P.
title: Statin Use and In‐Hospital Mortality in Patients With Diabetes Mellitus and COVID‐19
date: 2020-12-11
journal: J Am Heart Assoc
DOI: 10.1161/jaha.120.018475
sha: 7bd7464a9d54d0e5f4202aad6c6e73ec3b36ee65
doc_id: 909877
cord_uid: rbetn5x5

BACKGROUND: Severe coronavirus disease 2019 (COVID‐19) is characterized by a proinflammatory state with high mortality. Statins have anti‐inflammatory effects and may attenuate the severity of COVID‐19. METHODS AND RESULTS: An observational study of all consecutive adult patients with COVID‐19 admitted to a single center located in Bronx, New York, was conducted from March 1, 2020, to May 2, 2020. Patients were grouped as those who did and those who did not receive a statin, and in‐hospital mortality was compared by competing events regression. In addition, propensity score matching and inverse probability treatment weighting were used in survival models to examine the association between statin use and death during hospitalization. A total of 4252 patients were admitted with COVID‐19. Diabetes mellitus modified the association between statin use and in‐hospital mortality. Patients with diabetes mellitus on a statin (n=983) were older (69±11 versus 67±14 years; P<0.01), had lower inflammatory markers (C‐reactive protein, 10.2; interquartile range, 4.5–18.4 versus 12.9; interquartile range, 5.9–21.4 mg/dL; P<0.01) and reduced cumulative in‐hospital mortality (24% versus 39%; P<0.01) than those not on a statin (n=1283). No difference in hospital mortality was noted in patients without diabetes mellitus on or off statin (20% versus 21%; P=0.82). Propensity score matching (hazard ratio, 0.88; 95% CI, 0.83–0.94; P<0.01) and inverse probability treatment weighting (HR, 0.88; 95% CI, 0.84–0.92; P<0.01) showed a 12% lower risk of death during hospitalization for statin users than for nonusers. CONCLUSIONS: Statin use was associated with reduced in‐hospital mortality from COVID‐19 in patients with diabetes mellitus. These findings, if validated, may further reemphasize administration of statins to patients with diabetes mellitus during the COVID‐19 era.

C oronavirus disease 2019 , caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has rapidly spread globally. 1, 2 Although nearly 80% of cases recover, COVID-19 still appears to retain an elevated mortality with the number of deaths escalating to over 250 000 worldwide through April 2020. 2 The severity of COVID-19 is characterized by a proinflammatory state evident by elevated serum biomarkers of inflammation including C-reactive protein, interleukin-6, and ferritin. 3, 4 These biomarkers are predictive of death during SARS-CoV-2 infection. 4 Statins have known anti-inflammatory properties 5 and may represent an agent for modulating the host response during COVID-19. 6 Pleotropic effects of statins also reduce reactive oxygen species and platelet reactivity, and acute administration improves survival in proinflammatory states such as myocardial infarction. 7, 8 In addition, statins can limit viral endotheliitis, 9 which has been implicated during SARS-CoV-2 infection. 10 On the contrary, animal studies indicate that statins increase membrane-bound angiotensin-converting enzyme 2 receptor, 11 which is a port of entry for SARS-CoV-2, thereby raising concern for statin usage during the COVID-19 pandemic. 12 As statins are one of the most commonly prescribed to patients with cardiovascular disease and diabetes mellitus, we sought to determine if statin use is associated with survival during COVID-19 by evaluating the association of in-hospital mortality and statin therapy.

The data that form the basis of the presented findings are available from the corresponding author upon reasonable request.

A retrospective single-center review of all patients who were admitted to Montefiore Medical Center in the Bronx, New York, with a confirmed COVID-19 diagnosis from March 1, 2020 to May 2, 2020, was conducted. COVID-19 was confirmed by a positive real-time reverse transcriptase polymerase chain reaction assay. A sample for positive real-time reverse transcriptase polymerase chain reaction was obtained by either nasopharyngeal or oropharyngeal swab. The follow-up period was from March 1, 2020 to May 4, 2020. The study was approved by the Institutional Review Board (2020-11308), and informed consent was waived.

Baseline demographic, clinical, and laboratory variables were retrieved from the electronic medical record system. The exposure was statin administration. Patients were grouped as those who received and those who did not receive a statin during the hospitalization. The primary outcome was in-hospital mortality during the follow-up period.

Continuous data are displayed as mean±SD or median 25% to 75% interquartile range (IQR), and characteristics were compared between statin users and nonusers using the Student t test, or Wilcoxon ranksum test. Categorical data are shown as percent and were compared by the chi-squared test. As hospital discharge is a competing event with in-hospital mortality, instead of Kaplan-Meier estimates, we used competing events analysis to estimate the cumulative incidence of in-hospital mortality since admission. 13 Patients who remained in the hospital at the end of the follow-up period on May 4, 2020, were censored. The difference in the cumulative incidence of in-hospital mortality was compared between statin users and nonusers separately among patients with and without diabetes mellitus by Fine and Gray's method. 14 

A multivariable regression model with subdistribution hazard ratios (HRs) 13 What Are the Clinical Implications?

• These findings may further strengthen and refocus administration of statins to patients with diabetes mellitus in the coronavirus disease 2019 era.

COVID-19 coronavirus disease 2019 IPTW inverse probability treatment weighting PS propensity score SARS-CoV-2 severe acute respiratory syndrome coronavirus 2 oximetry measurement, serum glucose, serum lactic acid, serum creatinine, and intravenous antibiotic use during hospitalization.

To limit potential residual confounding by indication, we also used propensity score (PS) matching and inverse probability treatment weighting (IPTW) to examine the association between statin use and in-hospital mortality in a Cox proportional hazards model among patients with diabetes mellitus. 15 Covariate selection for PS matching and IPTW was done in accordance with criteria described previously. 16, 17 Variables included for PS matching were true confounders, or those that were associated with the outcome of interest (ie, in-hospital death). We did not include variables that were associated only with exposure (ie, statin use) and not the outcome (ie, in-hospital death), since it has been demonstrated that these covariates increase the variance of the estimated exposure effect without decreasing bias. 16 The covariates in the propensity analyses included age, sex, body mass index, days of symptoms before admission, history of atherosclerotic heart disease, Charlson comorbidity index, presenting diastolic blood pressure, respiratory rate, pulse oximetry measurement, heart rate, serum glucose, lactic acid, serum creatinine, serum troponin levels, usage of angiotensin-converting enzymes inhibitors, usage of angiotensin receptor blockers, serum troponin levels, and intravenous antibiotics during hospitalization. PS matching was carried out through a 1:1 greedy matching algorithm, with a caliper width of 0.1 SD, and through IPTW. The PS and IPTW analyses were carried out using treatment effects in Stata (StataCorp, College Station, TX).

The proportion of patients with diabetes mellitus and missing data for a covariate that was included in multivariable competing risk regression and propensity score models was <5%, with the exception of glucose (9.2%), lactic acid (9.1%), and serum troponin (15.4%).

Since troponin was missing in >10% of the patients with diabetes mellitus, we created a categorical variable using quartile cutoff point with an additional category for missing values and added this variable in the PS models. Given a skewed distribution, quartiles 1 and 2 were collapsed as a single category and served as a reference for quartile 3 (intermediate-level troponin), quartile 4 (high-level troponin), and missing troponin categories. The final data set for the multivariable competing risks regression model included 2039 (90%) of the 2266 patients with diabetes mellitus, while the models using PS matching and IPTW included 1902 (84%) of patients with diabetes mellitus.

A P<0.05 was considered statistically significant. Stata 16 (StataCorp) and R (R Foundation for Statistical Computing, Vienna, Austria) software was used for all statistical analysis.

Overall, 4252 patients (65±16 years old; 47% female) were admitted with COVID-19. Thirty-seven percent (n=1570) were Hispanic. Diabetes mellitus (53%), hypertension (72%), and atherosclerotic heart disease (26%) were highly prevalent comorbidities (Table 1) . On average, patients presented 3 (IQR, 0-7) days after onset of symptoms and had elevated inflammatory markers including C-reactive protein (10.6; IQR, 4.5-18.9 mg/dL) and ferritin (785; IQR, 369-1607 ng/mL).

Within all admitted patients, 32% (n=1355) received a statin. Those on statin therapy were older (69±12 versus 63±17 years; P<0.01) and had a higher Charlson comorbidity index (5; IQR, 3-8 versus 3; IQR, 1-5; P<0.01). Despite a higher comorbidity burden, we observed a lower cumulative in-hospital mortality in those receiving a statin (23% versus 27%; P<0.01). Seventysix percent of the patients received atorvastatin. Further stratification analysis revealed that this effect was modified by diabetes mellitus. Patient with diabetes mellitus (n=2266) on a statin had a lower cumulative in-hospital mortality (24% versus 39%; P<0.01; Figure 

Patients with diabetes mellitus admitted for COVID-19 were 68±13 years old, 48% were female, 36% classified as Hispanic, 87% had a history of hypertension, and 36% had atherosclerotic heart disease. Their blood glucose at presentation was 180 (IQR, 120-274 mg/dL), and the glycosylated hemoglobin if available within 3 months before admission was 7.8 (IQR, 6.5%-9.7%; n=507).

Overall, patients with diabetes mellitus had an in-hospital mortality of 32% over a follow-up period of 5.9 (IQR, 3.4-10.8) days. Those who died in the hospital were older (74±12 versus 67±13 years; P<0.01), were more frequently men (57% versus 50%, P<0.01), and had a greater proportion of known atherosclerotic heart disease (44% versus 32%; P<0.01). Nonsurvivors had lower blood pressure, reduced oxygen saturation at presentation, and higher levels of inflammatory biomarkers (Table 2) . (1) Simvastatin 247 (18) The following were the available samples sizes for variables with >5% missing data: alanine transaminase (n=3901), ferritin (n=1663), lactate dehydrogenase (n=2635), C-reactive protein (n=2083), d-dimer (n=1857), procalcitonin (n=1537), lactic acid (n=3705), creatinine (n=4108), glucose (n=3703), and troponin (n=3384). ACE indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; ASHD, atherosclerotic heart disease; BNP, B-type natriuretic peptide; and COVID-19, coronavirus disease 2019.

*Other includes Asian, Pacific Islander, American Indian, Alaskan Native, or Other. 16) in comparison with those who did not receive a statin (Table 3) .

There was a 49% reduction in hospital mortality (HR, 0.51; 95% CI, 0.43-0.61) associated with stain use among patients with diabetes mellitus admitted with COVID-19 in the multivariable competing risk regression model (Table 4) . Through PS caliper matching, statin recipients and nonrecipients had similar distributions of PS ( Figure S1 ), and standardized average differences among covariates were greatly reduced ( Figure S2 ). PS analyses with either caliper matching (average treatment effect on the treated: HR, 0.88; 95% CI, 0.83-0.94; P<0.001) or IPTW (average treatment effect on the treated: HR, 0.88; 95% CI, 0.84-0.92; P<0.001) showed a statistically significant 12% reduced risk of in-hospital mortality in patients with diabetes mellitus receiving statin therapy (Table 4 ).

In this analysis involving a large cohort of hospitalized patients with COVID-19, statin use was associated with reduced in-hospital mortality in patients with diabetes mellitus. This observation was made despite older age and higher prevalence of hypertension and atherosclerotic heart disease in statin users with diabetes mellitus. Patients with diabetes mellitus either on or off statin therapy presented to the hospital in a similar time frame from the onset of symptoms; however, those on a statin had lower inflammatory markers. After matching and weighing confounding clinical covariables and indication, there remained an association between statin therapy and reduced in-hospital mortality.

There is a paucity of studies that have assessed the association between statin use and in-hospital mortality during COVID-19. An observational analysis from China reported an association between statin use and improved survival among hospitalized patients with COVID-19. 18 Unlike this study, our investigation focuses on a cohort of sicker hospitalized patients within the United States, which had notably greater overall in-hospital mortality. Moreover, in our cohort, we observed higher survival within patients with diabetes mellitus, who may carry mitigating characteristics for statin therapy.

Type 1 diabetes mellitus is as an autoimmune disease with destruction of pancreatic beta cells. During type 1 diabetes mellitus, there is persistent inflammation mediated through interferon gamma, tumor necrosis factor alpha, and interleukin-1 beta. 19 Type 2 diabetes mellitus is characterized by a chronic lowgrade inflammatory state promoted by a variety of pathways including immune dysregulation, alterations in gut microorganisms, and metabolic syndrome. 20 Patients with diabetes mellitus can have an initial delay in the adaptive immune response to viral infection and may subsequently experience an exaggerated hyperinflammatory response to SARS-CoV-2 infection. 21 Indeed, mortality during COVID-19 is disproportionally elevated in patients with diabetes mellitus, who incur a 2-to 3-fold greater adjusted risk of death, 22, 23 and early studies have noted higher levels of interleukin-6 and C-reactive protein with COVID-19. 24 Statins are known to have anti-inflammatory effects that are clinically evident through a reduction in inflammatory biomarkers such as C-reactive protein, 5 and these effects are also apparent in patients with diabetes mellitus. 25 In our cohort, patients with diabetes mellitus in the statin group presented with lower inflammatory markers and had reduced mortality. In contrast, there was no difference in inflammatory markers at presentation in patients without diabetes mellitus either on or off statins ( Table 3 ). Presuming that in-hospital statin administration was indicative of outpatient usage and adherence, one might speculate that statin therapy blunted the COVID-19 inflammatory response before admission as well as during hospitalization, and improved survival was related to ongoing chronic statin use. This hypothesis would have to be prospectively tested. Guidelines recommend statin therapy for adults 40 to 75 years of age with diabetes mellitus to reduce cardiovascular disease, 26 but these agents remain underused. 27 Usage of statins in patients with diabetes mellitus is reported to be low, in the 40% to 50% range, with further underusage in minorities, 27 which was consistent within our cohort. Our findings, if validated, may further strengthen and refocus administration of statins to patients with diabetes mellitus during the COVID-19 era.

This investigation has several limitations. First, given the retrospective nature of our investigation and admission of patients not necessarily followed chronically at our institution, we were unable to firmly differentiate patients who received a statin into subcategories of prehospital user or in-hospital user and those with cessation of chronic statin use. However, it has not been our institutional practice to initiate new statin therapy in patients with acute COVID-19 illness. In addition, typical clinical parameters that might lead to discontinuation of statin therapy, such as a significant elevation in liver enzymes including alanine aminotransferase, was neither apparent nor clinically meaningfully different between those who did and did not receive statins. Patients on statins may have unmeasured confounders that could not be assessed through matching and weighting methods. However, such unmeasured confounders would also be present in statin users without diabetes mellitus, who did not show any difference in hospital mortality in comparison to nonusers. The fidelity of the data may be limited by the accuracy and level of documentation within the electronic medical records.

Because of the single-center design, generalizability may be limited. These observational findings do not assess the efficacy of statins as a therapeutic agent for COVID-19 in patients with diabetes mellitus, which can be accomplished only through a randomized control trial.

In summary, in this observational analysis, statin administration to patients with diabetes mellitus was associated with a reduced risk of in-hospital mortality during COVID-19. Further study to validate our findings and investigate underlying mechanisms is needed.

Received July 13, 2020; accepted October 13, 2020. 

None.

Figures S1-S2 ATE indicates average treatment effect of statins in patients with diabetes mellitus admitted with COVID-19; ATT, average treatment effect in patients with diabetes mellitus treated with a statin; and IPTW, inverse probability treatment weighing.

*The multivariable competing risk regression model was adjusted for age, sex, history of atherosclerotic heart disease, Charlson comorbidity index, presenting diastolic blood pressure, respiratory rate, pulse oximetry measurement, serum glucose, serum lactic acid, serum creatinine, and intravenous antibiotic use during hospitalization. † Covariates in the propensity analyses (caliper matching and IPTW) included: age, sex, body mass index, days of symptoms before admission, history of atherosclerotic heart disease, Charlson comorbidity index, presenting diastolic blood pressure, respiratory rate, pulse oximetry measurement, heart rate, serum glucose, lactic acid, serum creatinine, serum troponin levels, usage of angiotensin converting enzymes inhibitors, usage of angiotensin receptor blockers, serum troponin level, and intravenous antibiotics during hospitalization.

Covid-19-navigating the uncharted

Mild versus severe COVID-19: laboratory markers

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Reduction in C-reactive protein and LDL cholesterol and cardiovascular event rates after initiation of rosuvastatin: a prospective study of the JUPITER trial

Hiding in plain sight: an approach to treating patients with severe COVID-19 infection

Pleiotropic effects of statins on the cardiovascular system

Early statin treatment following acute myocardial infarction and 1-year survival

Hydroxymethylglutaryl coenzyme a reductase inhibition limits cytomegalovirus infection in human endothelial cells

Endothelial cell infection and endotheliitis in COVID-19

Tissue specific up regulation of ACE2 in rabbit model of atherosclerosis by atorvastatin: role of epigenetic histone modifications

COVID-19, ACE2, and the cardiovascular consequences

A note on competing risks in survival data analysis

A class of K-sample tests for comparing the cumulative incidence of a competing risk

Propensity-score matching with competing risks in survival analysis

Variable selection for propensity score models

Methods for constructing and assessing propensity scores

In-hospital use of statins is associated with a reduced risk of mortality among individuals with COVID-19

The role of inflammation in diabetes: current concepts and future perspectives

Cells of innate and adaptive immunity in type 2 diabetes and obesity

COVID-19 pandemic, coronaviruses, and diabetes mellitus

Diabetes mellitus is associated with increased mortality and severity of disease in COVID-19 pneumonia-a systematic review, meta-analysis, and meta-regression

Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan

Diabetes is a risk factor for the progression and prognosis of COVID-19

Ten Cate H. Anti-inflammatory and anticoagulant effects of pravastatin in patients with type 2 diabetes

ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines

Trends and disparities in statin use and low-density lipoprotein cholesterol levels among US patients with diabetes