key: cord-0891523-txxylmrw
authors: frontera, Jennifer A.; Rahimian, Joseph O.; Yaghi, Shadi; Liu, Mengling; Lewis, Ariane; de Havenon, Adam; Mainali, Shraddha; Huang, Joshua; Scher, Erica; Wisniewski, Thomas; Troxel, Andrea B.; Meropol, Sharon; Balcer, Laura J.; Galetta, Steven L.
title: Treatment with Zinc is Associated with Reduced In-Hospital Mortality Among COVID-19 Patients: A Multi-Center Cohort Study
date: 2020-10-26
journal: Res Sq
DOI: 10.21203/rs.3.rs-94509/v1
sha: 6f9cf04546c8a4c38aeaf8c8db3072b5c2a01371
doc_id: 891523
cord_uid: txxylmrw

Background: Zinc impairs replication of RNA viruses such as SARS-CoV-1, and may be effective against SARS-CoV-2. However, to achieve adequate intracellular zinc levels, administration with an ionophore, which increases intracellular zinc levels, may be necessary. We evaluated the impact of zinc with an ionophore (Zn+ionophore) on COVID-19 in-hospital mortality rates. Methods: A multicenter cohort study was conducted of 3,473 adult hospitalized patients with reverse-transcriptase-polymerase-chain-reaction (RT-PCR) positive SARS-CoV-2 infection admitted to four New York City hospitals between March 10 through May 20, 2020. Exclusion criteria were: death or discharge within 24h, comfort-care status, clinical trial enrollment, treatment with an IL-6 inhibitor or remdesivir. Patients who received Zn+ionophore were compared to patients who did not using multivariable time-dependent cox proportional hazards models for time to in-hospital death adjusting for confounders including age, sex, race, BMI, diabetes, week of admission, hospital location, sequential organ failure assessment (SOFA) score, intubation, acute renal failure, neurological events, treatment with corticosteroids, azithromycin or lopinavir/ritonavir and the propensity score of receiving Zn+ionophore. A sensitivity analysis was performed using a propensity score-matched cohort of patients who did or did not receive Zn+ionophore matched by age, sex and ventilator status. Results: Among 3,473 patients (median age 64, 1947 [56%] male, 522 [15%] ventilated, 545[16%] died), 1,006 (29%) received Zn+ionophore. Zn+ionophore was associated with a 24% reduced risk of in-hospital mortality (12% of those who received Zn+ionophore died versus 17% who did not; adjusted Hazard Ratio [aHR] 0.76, 95% CI 0.60-0.96, P=0.023). More patients who received Zn+ionophore were discharged home (72% Zn+ionophore vs 67% no Zn+ionophore, P=0.003) Neither Zn nor the ionophore alone were associated with decreased mortality rates. Propensity score-matched sensitivity analysis (N=1356) validated these results (Zn+ionophore aHR for mortality 0.63, 95%CI 0.44-0.91, P=0.015). There were no significant interactions for Zn+ionophore with other COVID-19 specific medications. Conclusions: Zinc with an ionophore was associated with increased rates of discharge home and a 24% reduced risk of in-hospital mortality among COVID-19 patients, while neither zinc alone nor the ionophore alone reduced mortality. Further randomized trials are warranted.

In this study of 3,473 hospitalized COVID-19 patients, treatment with zinc and an ionophore was associated with a 24% reduced risk of in-hospital mortality in multivariable Cox regression analysis. A sensitivity analysis in a propensity score-matched cohort supported this nding.

Efforts to identify therapeutic interventions to ameliorate the impact of COVID-19 have resulted in over 200 ongoing clinical trials, evaluating a wide spectrum of medications including antivirals, antimalarials, immunomodulators, plasma-based therapies, stem cell therapies, anticoagulants and vitamin/mineral supplementation (1) . One potential therapy of interest is zinc, which has a range of immune-modulating effects including regulation of leukocyte function and tempering of in ammatory responses (2, 3) . Pre-clinical studies have shown that zinc acts to impair replication of RNA viruses, including SARS-CoV-1, through direct inhibition of RNA-dependent RNA polymerase (4) . It is postulated that zinc may similarly inhibit SARS-CoV-2 replication, though little data currently exists to support this hypothesis (5, 6) .

Under normal physiological conditions, intracellular levels of zinc are maintained at low and possibly subtherapeutic levels by metallothioneins (7) . In order to effectively increase intracellular zinc levels, co-administration with an ionophore is necessary. Ionophores are lipid soluble compounds that act by promoting intracellular transport of non-lipid soluble zinc, independent of zinc binding proteins present in the plasma membrane. Several compounds have been identi ed as zinc ionophores including pyrithione(4), pyrrolidine dithiocarbamate(4), hinokitiol (8) , resveratrol (9) , plant polyphenols such as quercetin and epigallocatechin-gallate (10) , chloroquine and hydroxychloroquine (11) . One randomized trial of hydroxychloroquine in non-hospitalized adults with early COVID-19 examined a subgroup of 63 patients (15% of the analyzed cohort of 423) who received zinc (12) . Though there was a 3% reduction in disease severity among those who received both zinc and hydroxychloroquine, this difference did not meet statistical signi cance in this small subgroup.

Based on the putative antiviral effect of zinc in preclinical studies of SARS-CoV-1(4), we hypothesized a priori that administration of zinc along with an ionophore would be e cacious in the treatment of patients with SARS-CoV-2 infection. Therefore, zinc administration was suggested by hospital pharmacists among patients who were already receiving the ionophore hydroxychloroquine as part of a hospital-wide guideline for the management of patients with COVID-19.

The primary aim of this study was to compare rates of in-hospital mortality among adult hospitalized COVID-19 patients who received zinc plus an ionophore (Zn + ionophore) to those who did not.

A multi-center cohort study of consecutive hospitalized patients admitted between March 10, 2020 and May 20, 2020 was conducted. Inclusion criteria were: aged ≥ 18 years, hospital admission, and reverse-transcriptase-polymerase-chain-reaction (RT-PCR) con rmed SARS-CoV-2 infection. Exclusion criteria were: death or discharge < 24 hours from admission, comfort care status, participation in a clinical trial (drug or convalescent plasma), treatment with an interleukin-6 inhibitor/modulator (including clazakizumab, tocilizumab, or sarilumab), treatment with Remdesivir, treatment with zinc or hydroxychloroquine for < 24 hours, or treatment in an emergency department or outpatient setting only.

A sensitivity analysis was performed in a propensity score-matched cohort of patients who did or did not receive Zn + ionophore, matched by age, sex, and invasive mechanical ventilation status.

This study included patients admitted to four NYU Langone Hospitals located in Manhattan, Brooklyn and Mineola, New York. All four hospitals utilize the same electronic medical record ([EMR], Epic) and information technology center, and have integrated clinical protocols for patient management. This study was approved with a waiver of authorization and informed consent by the NYU Grossman School of Medicine Institutional Review Board based on the low risk to patients.

Demographics, past medical history, medication utilization and in-hospital outcomes (in-hospital mortality, discharge disposition, ventilator days and hospital length of stay) were extracted from the EMR. The maximum recorded Sequential Organ Failure Assessment (SOFA) Score was used to assess severity of illness and has been shown to be predictive of organ failure and in-hospital mortality (13) (14) (15) . Speci c in-hospital complications including neurological events (e.g. stroke, toxic/metabolic encephalopathy, hypoxic/ischemic brain injury, seizures), and acute respiratory failure requiring invasive mechanical ventilation were ascertained via manual chart review following the Global Consortium Study-NeuroCOVID study protocol and data dictionary (16) . The in-hospital complication of acute renal failure was assessed by EMR data query based on ICD-10 diagnoses.

Patients were coded as having received Zn + ionophore if the combination of zinc and hydroxychloroquine were administered together at any time during hospitalization for a minimum duration of one day. Data on the dose, route, duration of administration and time from admission to rst dose were extracted. A time-dependent Zn + ionophore covariate was used to account for "immortal time bias", which can occur when an event is observed more frequently in patients who survive long enough to be diagnosed or treated with a particular medication (17) . This methodology also allowed us to account for the time from admission to exposure to medication when predicting the hazard of in-hospital death.

A health system-wide inpatient COVID-19 treatment algorithm was developed by the hospital infection control and infectious diseases departments and was iteratively revised over the study time frame (see Supplemental Methods). This guideline broadly followed Infectious Diseases Society of America Guidelines for the treatment and management of COVID-19 (18) . Hospital guidelines for the treatment for COVID-19 patients at the time (instituted in March, 2020) recommended administration of hydroxychloroquine (400 mg BID for one day then 200 mg BID for four days) in patients whose oxygen saturation was < 94% on room air and whose QTc interval was < 500 ms. Zinc sulfate (220 mg [50 mg Zinc] PO once or twice daily) was suggested in combination with the ionophore hydroxychloroquine(11) beginning March 26, 2020, because achieving elevated intracellular levels with zinc alone is di cult (19) . Zinc was therefore administered at the same time as hydroxychloroquine, typically for four days or until discharge. Zinc could be prescribed for longer courses at the discretion of the treating physician. During the timeframe of this study (prior to publication of the RECOVERY trial(20)), corticosteroids could be considered in patients with bilateral opacities on chest imaging and a PaO2/FiO2 < 250 mmHg, as long as the patient did not have an active bacterial infection and was not immunosuppressed. Corticosteroids could also be used for other indications, such as COPD exacerbation, refractory septic shock or suspected adrenal insu ciency. A protocol of early proning was promoted for patients prior to intubation.

The primary outcome was in-hospital death. To avoid time to event bias among patients who were discharged, a dummy variable of 75 days was used as the event time for right censored patients who were not dead or discharged to hospice. Seventy-ve days was selected based on the prolonged length of stay (LOS) observed in this cohort (maximum LOS 71.4 days).

Demographic variables, past medical history, clinical features, concomitant medication usage, hospital complications, and in-hospital outcomes were compared between COVID-19 patients who did or did not receive Zn + ionophore using the Mann-Whitney-U (Wilcoxon rank-sum) test for continuous variables and Chi-square test or Fisher's exact test for categorical values, as appropriate.

A multivariable Cox proportional hazards model was t for the time to in-hospital death using a time-dependent Zn + ionophore covariate to account for immortal time bias and the fact that medication usage with variable initiation times violated the proportional hazards assumption over the hospitalization period (17) . This model was adjusted for confounders including age, sex, race, body mass index (BMI), diabetes, week of admission, hospital location, maximum sequential organ failure assessment (SOFA) score recorded during hospitalization, intubation/invasive mechanical ventilation, acute renal failure, neurological complications, treatment with corticosteroids, azithromycin or lopinavir/ritonavir and the propensity score for receiving Zn + ionophore. The propensity score was generated using logistic regression to model the predicted probability of receipt of Zn + ionophore as a function of age, sex, and ventilator status (as a marker of disease severity). The other covariates were selected based on known predictors of inhospital death, biological plausibility and bivariate associations within our own data. Similar Cox regression models were t for Zn alone, hydroxychloroquine alone, corticosteroids, azithromycin and lopinavir/ritonavir administration using the propensity score predicted probability speci c to each medication and adjusting for the same covariates as the primary analysis. Subgroup analyses were conducted in dichotomized categories of age (dichotomized at the median), sex, ventilator status, corticosteroid, azithromycin and lopinavir/ritonavir use and forest plots were constructed. To determine if there was a difference in treatment effect for once versus twice daily zinc in addition to the ionophore, a Cox regression model was t with a time-dependent covariate for twice daily dosing compared to a reference of once daily dosing, adjusting for the same covariates as the primary analysis, as well as the propensity score for once versus twice daily dosing (generated by logistic regression analysis adjusting for age, sex and ventilator status).

In a sensitivity analysis, a propensity score matched cohort was generated (matching for age, sex and ventilator status, as above) to reduce the imbalance in baseline characteristics between treatment groups when evaluating the treatment effects on the outcome of in-hospital mortality. In each propensity score-matched cohort, patients who did or did not receive Zn + ionophore were matched in a 1:1 ratio (random case order when drawing matches without replacement, with priority given to exact matches) with a 0.00003 propensity score matching radius. Similar Cox proportional hazard models were t for the time to in-hospital death using a time-dependent Zn + ionophore covariate and adjusting for the same confounder covariates as the primary analysis. All analyses were conducted using IBM SPSS Statistics for Mac version 25 (IBM Corp., Armonk, NY).

Between March 10 and May 20, 2020, 4,491 adult SARS-CoV-2 RT-PCR positive patients were hospitalized at our study sites. After excluding patients who died or were discharged within 24 hours, patients who received comfort measures only, those enrolled in clinical trials and those who received an IL-6 inhibitor or Remdesivir, 3,473 were included in analysis (Fig. 1 ). Of these 3,473 patients, 1,066 (29%) patients received Zn + ionophore beginning at a median of 0.5 days from admission. The median duration of treatment was three days (interquartile range interquartile range [IQR] 1.5-4.3 days). The median age was 64 years (IQR 50-76), 56% were male and 15% required invasive mechanical ventilation (Table 1) . Patients who received Zn + ionophore were more often male, black, diabetic, had a higher BMI, were more often treated with corticosteroids and azithromycin, and less often treated with lopinavir/ritonavir compared to patients who did not receive Zn + ionophore ( Table 2 ). In univariate analyses, rates of in-hospital mortality were signi cantly lower among patients who received Zn + ionophore compared to those who did not (12% died versus 17%, P < 0.001). Similarly, rates of discharge home were signi cantly higher among patients who received Zn + ionophore (72% versus 67% of patients who did not receive Zn + ionophore, P < 0.001). Table 1 Cohort characteristics of patients who received zinc with an ionophore (N = 1,097) did not affect mortality rates (aHR 1.14, 95% CI 0.89-1.44, P = 0.296) and treatment with the ionophore hydroxychloroquine alone (N = 2,299) appeared to be harmful (aHR 1.60, 95% CI 1.22-2.11, P = 0.001). In subgroup analysis, there were no signi cant interactions by age, intubation status, or use of other COVID-19 speci c medications (corticosteroids, azithromycin, lopinavir/ritonavir), however, there was a suggestion of greater bene t of Zn + ionophore among males (Fig. 2) . Among patients who received Zn + ionophore treatment, we did not identify a difference in treatment effect between once daily or twice daily dosing of zinc sulfate after adjusting for the same covariates used in the primary Cox regression analysis (aHR 0.90, 95% CI 0.56-1.45, P = 0.655). In similar adjusted Cox regression analyses, corticosteroid use (N = 478) was associated with higher rates of in-hospital mortality (aHR 2.2, 95% CI 1.74-2.69, P < 0.001). Among patient who received corticosteroids, 272/478 (57%) received methylprednisolone, 170/478 (36%) received prednisone and 78/478 (16%) received dexamethasone. The median time from admission to rst dose of corticosteroids was 3 days (IQR 0.8-8.7 days) and the median duration of treatment was 4 days (IQR 1-7 days). There was no impact on in-hospital mortality for azithromycin (N = 2,014; aHR 1.22, 95% CI 0.99-1.51, P = 0.066) or lopinavir/ritonavir (N = 226; aHR 1.20, 95% CI 0.91-1.60, P = 0.204).

A sensitivity analysis in a propensity score-matched cohort produced consistent results. A total of 1,356 patients were included in this analysis (N = 678 Zn + ionophore, N = 678 No Zn + ionophore). The Zn + ionophore versus no Zn + ionophore groups were well matched for age, sex and intubation/invasive mechanical ventilation status (Supplemental Table 1 ). In this propensity score-matched cohort analysis, the median time from admission to Zn + ionophore administration was 0.6 days (IQR 0.3-0.9) and the median duration of treatment was 3 days (IQR 1. 6-4.4) . Similar to the primary analysis, Zn + ionophore was associated with a 37% reduced risk of in-hospital death (aHR 0.63, 95% CI 0.44-0.92, P = 0.015), while there was no signi cant association of in-hospital mortality with Zn alone (aHR 1.11, 95% CI 0.77-1.60, P = 0.587) or the ionophore hydroxychloroquine alone (aHR 0.97, 95% CI 0.59-1.60, P = 0.897).

In this large cohort study with propensity score-matched sensitivity analysis, we evaluated the a priori hypothesis based on biologic plausibility, that treatment with zinc along with an ionophore (which increases intracellular zinc levels) would have a bene cial impact on patients with COVID-19. We found that Zn + ionophore was associated with a 24% relative risk reduction in rates of in-hospital death compared to patients who did not receive this medication combination. Neither zinc alone nor the ionophore alone (hydroxychloroquine) was associated with a reduction in mortality rates. In fact, hydroxychloroquine appeared to be harmful, though our primary analysis was not designed to evaluate the impact of hydroxychloroquine. The fact that a protective effect of Zn + ionophore was observed with fewer patients (Zn + ionophore N = 1,006) than the zinc alone group (N = 1,097) supports the biological hypothesis that an ionophore may be required to increase intracellular zinc levels to achieve therapeutic e cacy. Patients who received Zn + ionophore were also signi cantly more likely to be discharged home than those who did not. Hospital length of stay and ventilator days were signi cantly longer among patients who received Zn + ionophore, perhaps because in-hospital mortality rates were lower. The low cost of zinc sulfate (21)) makes it an attractive therapeutic candidate compared to other drugs being evaluated for COVID-19 management such as remdesivir ($520.00 /dose(22)) or IL-6 blockers ($277.00 /600 mg dose of tocilizumab (23)). The cost of ionophores that are co-administered with zinc is similarly low (e.g. hydroxychloroquine: $4.36 /dose (24)).

Strengths of this study include the large sample size, and exclusion of comfort care patients or those who were dead or discharged within 24 hours, because these groups are less likely to receive COVID-19-speci c medication and their inclusion could confound mortality analyses. Additionally, we excluded patients who received IL-6 inhibitors (tocilizumab, sarilumab, clazikuzumab) or remdesivir to avoid inadvertent observation of bene cial effect related to these agents. Similarly, patients enrolled in clinical trials were excluded because even patients randomized to a placebo group are treated differently than patients not participating in a trial. We did include patients who received corticosteroids, however, a variety of different steroid-class medications, doses and durations were utilized. Furthermore, this study was conducted prior to the RECOVERY trial (20) , and use of corticosteroids was limited to severely hypoxic patients, which may explain the increased in-hospital mortality rates observed with corticosteroid use in this cohort. To limit the confounding impact of corticosteroids, we adjusted for their use in the multivariable Cox regression analyses. We also adjusted for the date of admission and hospital location to temper any effect that access to resources may have played in determining outcomes (25) . Patients who received Zn + ionophore were actually enrolled earlier in the study window than those who did not, which could have biased this group toward worse outcomes due to resource scarcity, however, we still observed a bene cial effect. In addition to the above strengths, we also con rmed our results in a propensity score matched analysis, which controlled for baseline predictors of zinc use and in-hospital mortality.

There are limitations to this study. First, because this was an observational study, we are unable to fully account for all differences between patients who did or did not receive Zn + ionophore, despite adjusting for a variety of confounders in multivariable and propensity score-matched analyses. A number of randomized trials of zinc among COVID-19 patients are currently underway and should provide more de nitive data. Second, we did not measure serum zinc levels prior to or after treatment, nor do we have any data on changes in intracellular zinc levels. It should be noted, however, that plasma zinc levels do not correlate well with tissue levels and active in ammation can acutely lower plasma zinc levels (26) . Erythrocyte zinc concentrations may be a more accurate measure, but may not be readily available at many centers in a timely fashion. Additionally, the optimal duration of treatment with zinc is unknown. In our study, even a brief median duration of 3 days of Zn + ionophore use was associated with improved outcomes, though whether longer treatment would further improve outcomes or increase risk of zinc toxicity is unknown. We also did not have data on whether patients were taking zinc and/or hydroxychloroquine prior to admission. Zinc toxicity generally includes non-speci c gastrointestinal symptoms, but prolonged exposure to high doses of zinc can lead to copper de ciency, which is associated with neurological abnormalities including myeloneuropathy and cognitive de cits (27), as well as hepatosplenomegaly and osteoporosis. Third, since only adults were included in this study, our results cannot be generalized to children. Fourth, hydroxychloroquine was used as the zinc ionophore out of expediency, since at the time of the study, hydroxychloroquine had hypothetical bene t in COVID-19 (28, 29) , in addition to functioning as a supporting agent to promote increased intracellular zinc levels. The use of hydroxychloroquine in Figure 2 Forest plot of subgroups evaluating the impact of treatment with zinc plus an ionophore on in-hospital mortality

This is a list of supplementary les associated with this preprint. Click to download.

SupplementaryAppendixZincBMCID.docx

Alternative ionophores such as hinokitiol(8), resveratrol(9) (found in red wine), quercetin (a avonoid found in red wine, kale, onions, green tea) or epigallocatechin-gallate(10) (found in green tea) might prove useful, but little data currently exists supporting their use in COVID-19 patients. Conclusions This is the rst large cohort study to evaluate the impact of zinc on outcome in hospitalized COVID-19 patients. Our results demonstrate a signi cant reduction in hospital morality rates when zinc is combined with an ionophore to maximize intracellular zinc levels. The low cost of zinc, compared to other treatments such as remdesivir or IL-6 blockers

Consent for publication: Not applicable Availability of Data: The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request. Competing Interests: NIH/NIA grant 3P30AG066512-01S1 to TW, JAF and LB and NIH/NINDS grant 3U24NS11384401S1 to AT

drafted the manuscript, provided nal approval and agrees to be accountable for all aspects of the work JOR-study concept, interpretation of data, critical revisions, nal approval, agrees to be accountable SY-study design and concept, interpretation of data, critical revisions, nal approval, agrees to be accountable ML-analyzed and interpreted the data, critical revisions, nal approval, agrees to be accountable AL-interpretation of data

AdH-interpretation of data, critical revisions, nal approval, agrees to be accountable SM-interpretation of data, critical revisions

Ongoing Clinical Trials for the Management of the COVID-19 Pandemic

Zinc as a Gatekeeper of Immune Function

Zinc Signals and Immunity

Zn(2+) inhibits coronavirus and arterivirus RNA polymerase activity in vitro and zinc ionophores block the replication of these viruses in cell culture

Zinc and respiratory tract infections: Perspectives for COVID19 (Review)

Can Zn Be a Critical Element in COVID-19 Treatment?

The Role of Zinc in Antiviral Immunity

Antiviral activity of the zinc ionophores pyrithione and hinokitiol against picornavirus infections

Effect of resveratrol and zinc on intracellular zinc status in normal human prostate epithelial cells

Zinc ionophore activity of quercetin and epigallocatechin-gallate: from Hepa 1-6 cells to a liposome model

Chloroquine is a zinc ionophore

Hydroxychloroquine in Nonhospitalized Adults With Early COVID-19: A Randomized Trial

Serial evaluation of the SOFA score to predict outcome in critically ill patients

The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine

Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on "sepsis-related problems" of the European Society of Intensive Care Medicine

Global Consortium Study of Neurological Dysfunction in COVID-19 (GCS-NeuroCOVID): Study Design and Rationale. Neurocrit Care

Problem of immortal time bias in cohort studies: example using statins for preventing progression of diabetes

Infectious Diseases Society of America Guidelines on the Treatment and Management of Patients with COVID-19

Zinc in Cellular Regulation: The Nature and Signi cance of "Zinc Signals

Dexamethasone in Hospitalized Patients with Covid-19 -Preliminary Report

Drug Information Lexicomp 2020

Drug Information Lexicomp 2020

Drug Information Lexicomp 2020

Why Surviving the Virus Might Come Down to Which Hospital Admits You

Acute in ammatory response does not affect erythrocyte concentrations of copper, zinc and selenium

Copper de ciency myelopathy produces a clinical picture like subacute combined degeneration

Hydroxychloroquine, a less toxic derivative of chloroquine, is effective in inhibiting SARS-CoV-2 infection in vitro

Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial

Hydroxychloroquine with or without Azithromycin in Mild-to-Moderate Covid-19

JH-acquisition of data, critical revisions, nal approval, agrees to be accountable ES-acquisition of data, critical revisions, nal approval, agrees to be accountable SM-interpretation of data, critical revisions, nal approval, agrees to be accountable AT-interpretation of data, critical revisions, nal approval, agrees to be accountable TW-interpretation of data, critical revisions, nal approval, agrees to be accountable LB-interpretation of data, critical revisions, nal approval, agrees to be accountable SG-interpretation of data, critical revisions, nal approval, agrees to be accountable Acknowledgments: Not applicable