Eosinophil Recovery in Hospitalized COVID-19 Patients is Associated with Lower Rates of ICU Admission and In-Hospital Mortality: An Observational Cohort Analysis

Background Admission eosinopenia (<100 cells/μL) is associated with poor clinical outcomes in hospitalized COVID-19 patients. However, the effects of eosinophil recovery (defined as reaching ≥50 eosinophils/μL) during hospitalization on COVID-19 outcomes have been inconsistent. Methods The study included 1,831 patients admitted to UCLA hospitals between February 2020 and February 2021 with PCR-confirmed COVID-19. Using competing risk regression and modeling eosinophil recovery as a time-dependent covariate, we evaluated the longitudinal relationship between eosinophil recovery and in-hospital outcomes including ICU admission, need for mechanical ventilation, and in-hospital mortality. All analyses were adjusted for covariates including age, BMI, tobacco smoke exposure, comorbidities known to be risk factors for COVID-19 mortality, and treatments including dexamethasone and remdesivir. Results Eosinophil recovery was evaluated in patients with <50 eosinophils/μL on admission (n=1282). These patients cumulatively amassed 11,633 hospital patient-days;3,985 of those days qualified as eosinophil recovery events, which were represented by 781 patients achieving at least one instance of eosinophil recovery during hospitalization. Despite no significant difference in the rate of mechanical ventilation, eosinophil recoverers had significantly lower rates of in-hospital mortality (aHR: 0.44 [0.29, 0.65], P=0.001) and ICU admission (aHR: 0.25 [0.11, 0.61], P=0.002). Conclusion Trending eosinophil counts during hospitalization is simple and can be performed in resource-limited healthcare settings to track the inflammatory status of a patient. Lack of eosinophil recovery events can identify those at risk for future progression to severe COVID.

Persistent memory despite rapid contraction of circulating T Cell responses to SARS-CoV-2 mRNA vaccination

Introduction While antibodies raised by SARS-CoV-2 mRNA vaccines have had compromised efficacy to prevent breakthrough infections due to both limited durability and spike sequence variation, the vaccines have remained highly protective against severe illness. This protection is mediated through cellular immunity, particularly CD8+ T cells, and lasts at least a few months. Although several studies have documented rapidly waning levels of vaccine-elicited antibodies, the kinetics of T cell responses have not been well defined. Methods Interferon (IFN)-γ enzyme-linked immunosorbent spot (ELISpot) assay and intracellular cytokine staining (ICS) were utilized to assess cellular immune responses (in isolated CD8+ T cells or whole peripheral blood mononuclear cells, PBMCs) to pooled peptides spanning spike. ELISA was performed to quantitate serum antibodies against the spike receptor binding domain (RBD). Results In two persons receiving primary vaccination, tightly serially evaluated frequencies of anti-spike CD8+ T cells using ELISpot assays revealed strikingly short-lived responses, peaking after about 10 days and becoming undetectable by about 20 days after each dose. This pattern was also observed in cross-sectional analyses of persons after the first and second doses during primary vaccination with mRNA vaccines. In contrast, cross-sectional analysis of COVID-19-recovered persons using the same assay showed persisting responses in most persons through 45 days after symptom onset. Cross-sectional analysis using IFN-γ ICS of PBMCs from persons 13 to 235 days after mRNA vaccination also demonstrated undetectable CD8+ T cells against spike soon after vaccination, and extended the observation to include CD4+ T cells. However, ICS analyses of the same PBMCs after culturing with the mRNA-1273 vaccine in vitro showed CD4+ and CD8+ T cell responses that were readily detectable in most persons out to 235 days after vaccination. Discussion Overall, we find that detection of spike-targeted responses from mRNA vaccines using typical IFN-γ assays is remarkably transient, which may be a function of the mRNA vaccine platform and an intrinsic property of the spike protein as an immune target. However, robust memory, as demonstrated by capacity for rapid expansion of T cells responding to spike, is maintained at least several months after vaccination. This is consistent with the clinical observation of vaccine protection from severe illness lasting months. The level of such memory responsiveness required for clinical protection remains to be defined.

Persistent memory despite rapid contraction of circulating T Cell responses to SARS-CoV-2 mRNA vaccination.

Introduction: While antibodies raised by SARS-CoV-2 mRNA vaccines have had compromised efficacy to prevent breakthrough infections due to both limited durability and spike sequence variation, the vaccines have remained highly protective against severe illness. This protection is mediated through cellular immunity, particularly CD8+ T cells, and lasts at least a few months. Although several studies have documented rapidly waning levels of vaccine-elicited antibodies, the kinetics of T cell responses have not been well defined. Methods: Interferon (IFN)-γ enzyme-linked immunosorbent spot (ELISpot) assay and intracellular cytokine staining (ICS) were utilized to assess cellular immune responses (in isolated CD8+ T cells or whole peripheral blood mononuclear cells, PBMCs) to pooled peptides spanning spike. ELISA was performed to quantitate serum antibodies against the spike receptor binding domain (RBD). Results: In two persons receiving primary vaccination, tightly serially evaluated frequencies of anti-spike CD8+ T cells using ELISpot assays revealed strikingly short-lived responses, peaking after about 10 days and becoming undetectable by about 20 days after each dose. This pattern was also observed in cross-sectional analyses of persons after the first and second doses during primary vaccination with mRNA vaccines. In contrast, cross-sectional analysis of COVID-19-recovered persons using the same assay showed persisting responses in most persons through 45 days after symptom onset. Cross-sectional analysis using IFN-γ ICS of PBMCs from persons 13 to 235 days after mRNA vaccination also demonstrated undetectable CD8+ T cells against spike soon after vaccination, and extended the observation to include CD4+ T cells. However, ICS analyses of the same PBMCs after culturing with the mRNA-1273 vaccine in vitro showed CD4+ and CD8+ T cell responses that were readily detectable in most persons out to 235 days after vaccination. Discussion: Overall, we find that detection of spike-targeted responses from mRNA vaccines using typical IFN-γ assays is remarkably transient, which may be a function of the mRNA vaccine platform and an intrinsic property of the spike protein as an immune target. However, robust memory, as demonstrated by capacity for rapid expansion of T cells responding to spike, is maintained at least several months after vaccination. This is consistent with the clinical observation of vaccine protection from severe illness lasting months. The level of such memory responsiveness required for clinical protection remains to be defined.

Chronic Obstructive Pulmonary Disease is Not Associated with In-Hospital Mortality in COVID-19: An Observational Cohort Analysis.

Background: Chronic obstructive pulmonary disease (COPD) is associated with worsened outcomes in COVID-19 (coronavirus disease 2019). However, data remain fraught with heterogeneity and bias from comorbid conditions. Additionally, data on the impact of COPD-specific factors, such as pre-hospital medications and pulmonologist involvement, remain sparse. Objective: We report a single-center analysis of COPD patients hospitalized with COVID-19 compared to those without COPD. Primary outcomes include ICU admission, mechanical ventilation, and in-hospital mortality. Methods: We evaluated all patients ≥40 years admitted with PCR-confirmed COVID-19 between February 2020 and February 2021. COPD was defined by documented ICD-10 diagnosis of COPD, confirmed smoking history, and active bronchodilator use. We compared outcomes between COPD patients and the remainder of the COVID-19 cohort. Multivariable analyses were adjusted for age, sex, smoking status, and comorbid conditions. Results: Of 1537 hospitalized COVID-19 patients, 122 (7.9%) carried a diagnosis of COPD. The COPD cohort was older (74 ± 13 vs 66 ± 15 years, P < 0.001) and more often former smokers (P < 0.001). Comorbid conditions including diabetes, cardiovascular disease, and kidney disease were more prevalent in the COPD group (P < 0.001). After adjusting for comorbid conditions, the COPD cohort had higher severity scores and trended towards fewer hospital-free days. Among patients with COPD, pre-hospital use of aspirin was associated with decreased ICU admissions (aHR 0.56, P = 0.049) and mechanical ventilation (aHR 0.25, P = 0.008), while LAMAs (long-acting muscarinic antagonists) were associated with decreased in-hospital mortality (aHR 0.34, P = 0.047). Involvement of pulmonology in pre-hospital management of COPD was not found to significantly affect outcomes. Conclusion: When corrected for comorbid illnesses, COPD was associated with more severe disease but not with increased ICU admission, mechanical ventilation, or in-hospital mortality rates. Among COPD patients, prehospital treatment with aspirin and COPD-directed therapies were associated with improved outcomes.

Cannabis consumption is associated with lower COVID-19 severity among hospitalized patients: a retrospective cohort analysis.

BACKGROUND: While cannabis is known to have immunomodulatory properties, the clinical consequences of its use on outcomes in COVID-19 have not been extensively evaluated. We aimed to assess whether cannabis users hospitalized for COVID-19 had improved outcomes compared to non-users. METHODS: We conducted a retrospective analysis of 1831 patients admitted to two medical centers in Southern California with a diagnosis of COVID-19. We evaluated outcomes including NIH COVID-19 Severity Score, need for supplemental oxygen, ICU (intensive care unit) admission, mechanical ventilation, length of hospitalization, and in-hospital death for cannabis users and non-users. Cannabis use was reported in the patient's social history. Propensity matching was used to account for differences in age, body-mass index, sex, race, tobacco smoking history, and comorbidities known to be risk factors for COVID-19 mortality between cannabis users and non-users. RESULTS: Of 1831 patients admitted with COVID-19, 69 patients reported active cannabis use (4% of the cohort). Active users were younger (44 years vs. 62 years, p < 0.001), less often diabetic (23.2% vs 37.2%, p < 0.021), and more frequently active tobacco smokers (20.3% vs. 4.1%, p < 0.001) compared to non-users. Notably, active users had lower levels of inflammatory markers upon admission than non-users-CRP (C-reactive protein) (3.7 mg/L vs 7.6 mg/L, p < 0.001), ferritin (282 µg/L vs 622 µg/L, p < 0.001), D-dimer (468 ng/mL vs 1140 ng/mL, p = 0.017), and procalcitonin (0.10 ng/mL vs 0.15 ng/mL, p = 0.001). Based on univariate analysis, cannabis users had significantly better outcomes compared to non-users as reflected in lower NIH scores (5.1 vs 6.0, p < 0.001), shorter hospitalization (4 days vs 6 days, p < 0.001), lower ICU admission rates (12% vs 31%, p < 0.001), and less need for mechanical ventilation (6% vs 17%, p = 0.027). Using propensity matching, differences in overall survival were not statistically significant between cannabis users and non-users, nevertheless ICU admission was 12 percentage points lower (p = 0.018) and intubation rates were 6 percentage points lower (p = 0.017) in cannabis users. CONCLUSIONS: This retrospective cohort study suggests that active cannabis users hospitalized with COVID-19 had better clinical outcomes compared with non-users, including decreased need for ICU admission or mechanical ventilation. However, our results need to be interpreted with caution given the limitations of a retrospective analysis. Prospective and observational studies will better elucidate the effects cannabis use in COVID-19 patients.

Dominant CD8+ T Cell Nucleocapsid Targeting in SARS-CoV-2 Infection and Broad Spike Targeting From Vaccination.

CD8+ T cells have key protective roles in many viral infections. While an overall Th1-biased cellular immune response against SARS-CoV-2 has been demonstrated, most reports of anti-SARS-CoV-2 cellular immunity have evaluated bulk T cells using pools of predicted epitopes, without clear delineation of the CD8+ subset and its magnitude and targeting. In recently infected persons (mean 29.8 days after COVID-19 symptom onset), we confirm a Th1 bias (and a novel IL-4-producing population of unclear significance) by flow cytometry, which does not correlate to antibody responses against the receptor binding domain. Evaluating isolated CD8+ T cells in more detail by IFN-γ ELISpot assays, responses against spike, nucleocapsid, matrix, and envelope proteins average 396, 901, 296, and 0 spot-forming cells (SFC) per million, targeting 1.4, 1.5, 0.59, and 0.0 epitope regions respectively. Nucleocapsid targeting is dominant in terms of magnitude, breadth, and density of targeting. The magnitude of responses drops rapidly post-infection; nucleocapsid targeting is most sustained, and vaccination selectively boosts spike targeting. In SARS-CoV-2-naïve persons, evaluation of the anti-spike CD8+ T cell response soon after vaccination (mean 11.3 days) yields anti-spike CD8+ T cell responses averaging 2,463 SFC/million against 4.2 epitope regions, and targeting mirrors that seen in infected persons. These findings provide greater clarity on CD8+ T cell anti-SARS-CoV-2 targeting, breadth, and persistence, suggesting that nucleocapsid inclusion in vaccines could broaden coverage and durability.

Clinical Characteristics and Outcomes of Coronavirus Disease 2019 Patients Who Received Compassionate-Use Leronlimab.

BACKGROUND: Leronlimab, a monoclonal antibody blocker of C-C chemokine receptor type 5 originally developed to treat human immunodeficiency virus infection, was administered as an open-label compassionate-use therapeutic for coronavirus disease 2019 (COVID-19). METHODS: Twenty-three hospitalized severe/critical COVID-19 patients received 700 mg leronlimab subcutaneously, repeated after 7 days in 17 of 23 patients still hospitalized. Eighteen of 23 received other experimental treatments, including convalescent plasma, hydroxychloroquine, steroids, and/or tocilizumab. Five of 23 received leronlimab after blinded, placebo-controlled trials of remdesivir, sarilumab, selinexor, or tocilizumab. Outcomes and results were extracted from medical records. RESULTS: Mean age was 69.5 ±â€…14.9 years; 20 had significant comorbidities. At baseline, 22 were receiving supplemental oxygen (3 high flow, 7 mechanical ventilation). Blood showed markedly elevated inflammatory markers (ferritin, D-dimer, C-reactive protein) and an elevated neutrophil-to-lymphocyte ratio. By day 30 after initial dosing, 17 were recovered, 2 were still hospitalized, and 4 had died. Of the 7 intubated at baseline, 4 were fully recovered off oxygen, 2 were still hospitalized, and 1 had died. CONCLUSIONS: Leronlimab appeared safe and well tolerated. The high recovery rate suggested benefit, and those with lower inflammatory markers had better outcomes. Some, but not all, patients appeared to have dramatic clinical responses, indicating that unknown factors may determine responsiveness to leronlimab. Routine inflammatory and cell prognostic markers did not markedly change immediately after treatment, although interleukin-6 tended to fall. In some persons, C-reactive protein clearly dropped only after the second leronlimab dose, suggesting that a higher loading dose might be more effective. Future controlled trials will be informative.

Humoral responses to SARS-CoV-2 mRNA vaccines: Role of past infection.

Two mRNA vaccines (BNT162b2 and mRNA-1273) against severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) are globally authorized as a two-dose regimen. Understanding the magnitude and duration of protective immune responses is vital to curbing the pandemic. We enrolled 461 high-risk health services workers at the University of California, Los Angeles (UCLA) and first responders in the Los Angeles County Fire Department (LACoFD) to assess the humoral responses in previously infected (PI) and infection naïve (NPI) individuals to mRNA-based vaccines (BNT162b2/Pfizer- BioNTech or mRNA-1273/Moderna). A chemiluminescent microparticle immunoassay was used to detect antibodies against SARS-CoV-2 Spike in vaccinees prior to (n = 21) and following each vaccine dose (n = 246 following dose 1 and n = 315 following dose 2), and at days 31-60 (n = 110) and 61-90 (n = 190) following completion of the 2-dose series. Both vaccines induced robust antibody responses in all immunocompetent individuals. Previously infected individuals achieved higher median peak titers (p = 0.002) and had a slower rate of decay (p = 0.047) than infection-naïve individuals. mRNA-1273 vaccinated infection-naïve individuals demonstrated modestly higher titers following each dose (p = 0.005 and p = 0.029, respectively) and slower rates of antibody decay (p = 0.003) than those who received BNT162b2. A subset of previously infected individuals (25%) required both doses in order to reach peak antibody titers. The biologic significance of the differences between previously infected individuals and between the mRNA-1273 and BNT162b2 vaccines remains uncertain, but may have important implications for booster strategies.

Using Serologic Testing to Assess the Effectiveness of Outbreak Control Efforts, Serial Polymerase Chain Reaction Testing, and Cohorting of Positive Severe Acute Respiratory Syndrome Coronavirus 2 Patients in a Skilled Nursing Facility.

We characterized serology following a nursing home outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) where residents were serially tested by reverse-transcription polymerase chain reaction (RT-PCR) and positive residents were cohorted. When tested 46-76 days later, 24 of 26 RT-PCR-positive residents were seropositive; none of the 124 RT-PCR-negative residents had confirmed seropositivity, supporting serial SARS-CoV-2 RT-PCR testing and cohorting in nursing homes.

Primary, Recall, and Decay Kinetics of SARS-CoV-2 Vaccine Antibody Responses.

Studies of two SARS-CoV-2 mRNA vaccines suggested that they yield ∼95% protection from symptomatic infection at least short-term, but important clinical questions remain. It is unclear how vaccine-induced antibody levels quantitatively compare to the wide spectrum induced by natural SARS-CoV-2 infection. Vaccine response kinetics and magnitudes in persons with prior COVID-19 compared to virus-naïve persons are not well-defined. The relative stability of vaccine-induced versus infection-induced antibody levels is unclear. We addressed these issues with longitudinal assessments of vaccinees with and without prior SARS-CoV-2 infection using quantitative enzyme-linked immunosorbent assay (ELISA) of anti-RBD antibodies. SARS-CoV-2-naïve individuals achieved levels similar to mild natural infection after the first vaccination; a second dose generated levels approaching severe natural infection. In persons with prior COVID-19, one dose boosted levels to the high end of severe natural infection even in those who never had robust responses from infection, increasing no further after the second dose. Antiviral neutralizing assessments using a spike-pseudovirus assay revealed that virus-naïve vaccinees did not develop physiologic neutralizing potency until the second dose, while previously infected persons exhibited maximal neutralization after one dose. Finally, antibodies from vaccination waned similarly to natural infection, resulting in an average of ∼90% loss within 90 days. In summary, our findings suggest that two doses are important for quantity and quality of humoral immunity in SARS-CoV-2-naïve persons, while a single dose has maximal effects in those with past infection. Antibodies from vaccination wane with kinetics very similar to that seen after mild natural infection; booster vaccinations will likely be required.

Discharge Clinical Characteristics and Post-Discharge Events in Patients with Severe COVID-19: A Descriptive Case Series.

BACKGROUND: As the SARS-CoV-2 pandemic continues, little guidance is available on clinical indicators for safely discharging patients with severe COVID-19. OBJECTIVE: To describe the clinical courses of adult patients admitted for COVID-19 and identify associations between inpatient clinical features and post-discharge need for acute care. DESIGN: Retrospective chart reviews were performed to record laboratory values, temperature, and oxygen requirements of 99 adult inpatients with COVID-19. Those variables were used to predict emergency department (ED) visit or readmission within 30 days post-discharge. PATIENTS (OR PARTICIPANTS): Age ≥ 18 years, first hospitalization for COVID-19, admitted between March 1 and May 2, 2020, at University of California, Los Angeles (UCLA) Medical Center, managed by an inpatient medicine service. MAIN MEASURES: Ferritin, C-reactive protein, lactate dehydrogenase, D-dimer, procalcitonin, white blood cell count, absolute lymphocyte count, temperature, and oxygen requirement were noted. KEY RESULTS: Of 99 patients, five required ED admission within 30 days, and another five required readmission. Fever within 24 h of discharge, oxygen requirement, and laboratory abnormalities were not associated with need for ED visit or readmission within 30 days of discharge after admission for COVID-19. CONCLUSION: Our data suggest that neither persistent fever, oxygen requirement, nor laboratory marker derangement was associated with need for acute care in the 30-day period after discharge for severe COVID-19. These findings suggest that physicians need not await the normalization of laboratory markers, resolution of fever, or discontinuation of oxygen prior to discharging a stable or improving patient with COVID-19.

Obesity and smoking as risk factors for invasive mechanical ventilation in COVID-19: A retrospective, observational cohort study.

PURPOSE: To describe the trajectory of respiratory failure in COVID-19 and explore factors associated with risk of invasive mechanical ventilation (IMV). MATERIALS AND METHODS: A retrospective, observational cohort study of 112 inpatient adults diagnosed with COVID-19 between March 12 and April 16, 2020. Data were manually extracted from electronic medical records. Multivariable and Univariable regression were used to evaluate association between baseline characteristics, initial serum markers and the outcome of IMV. RESULTS: Our cohort had median age of 61 (IQR 45-74) and was 66% male. In-hospital mortality was 6% (7/112). ICU mortality was 12.8% (6/47), and 18% (5/28) for those requiring IMV. Obesity (OR 5.82, CI 1.74-19.48), former (OR 8.06, CI 1.51-43.06) and current smoking status (OR 10.33, CI 1.43-74.67) were associated with IMV after adjusting for age, sex, and high prevalence comorbidities by multivariable analysis. Initial absolute lymphocyte count (OR 0.33, CI 0.11-0.96), procalcitonin (OR 1.27, CI 1.02-1.57), IL-6 (OR 1.17, CI 1.03-1.33), ferritin (OR 1.05, CI 1.005-1.11), LDH (OR 1.57, 95% CI 1.13-2.17) and CRP (OR 1.13, CI 1.06-1.21), were associated with IMV by univariate analysis. CONCLUSIONS: Obesity, smoking history, and elevated inflammatory markers were associated with increased need for IMV in patients with COVID-19.

A machine learning algorithm to increase COVID-19 inpatient diagnostic capacity.

Worldwide, testing capacity for SARS-CoV-2 is limited and bottlenecks in the scale up of polymerase chain reaction (PCR-based testing exist. Our aim was to develop and evaluate a machine learning algorithm to diagnose COVID-19 in the inpatient setting. The algorithm was based on basic demographic and laboratory features to serve as a screening tool at hospitals where testing is scarce or unavailable. We used retrospectively collected data from the UCLA Health System in Los Angeles, California. We included all emergency room or inpatient cases receiving SARS-CoV-2 PCR testing who also had a set of ancillary laboratory features (n = 1,455) between 1 March 2020 and 24 May 2020. We tested seven machine learning models and used a combination of those models for the final diagnostic classification. In the test set (n = 392), our combined model had an area under the receiver operator curve of 0.91 (95% confidence interval 0.87-0.96). The model achieved a sensitivity of 0.93 (95% CI 0.85-0.98), specificity of 0.64 (95% CI 0.58-0.69). We found that our machine learning algorithm had excellent diagnostic metrics compared to SARS-CoV-2 PCR. This ensemble machine learning algorithm to diagnose COVID-19 has the potential to be used as a screening tool in hospital settings where PCR testing is scarce or unavailable.