Remdesivir and Tocilizumab for the Treatment of Severe COVID-19 in a Community Hospital: A Retrospective Cohort Study

Background. Growing evidence supports the use of remdesivir and tocilizumab for the treatment of hospitalized patients with severe COVID-19. The purpose of this study was to evaluate the use of remdesivir and tocilizumab for the treatment of severe COVID-19 in a community hospital setting. Methods. We used a de-identified dataset of hospitalized adults with severe COVID-19 according to the National Institutes of Health definition (SpO2 < 94% on room air, a PaO2/FiO2 < 300 mm Hg, respiratory frequency > 30/min, or lung infiltrates > 50%) admitted to our community hospital located in Evanston Illinois, between March 1, 2020, and March 1, 2021. We performed a Cox proportional hazards regression model to examine the relationship between the use of remdesivir and tocilizumab and inpatient mortality. To minimize confounders, we adjusted for age, qSOFA score, noninvasive positive-pressure ventilation, invasive mechanical ventilation, and steroids, forcing these variables into the model. We implemented a sensitivity analysis calculating the E-value (with the lower confidence limit) for the obtained point estimates to assess the potential effect of unmeasured confounding. Figure 1. Kaplan-Meier survival curves for in-hospital death among patients treated with and without steroids The hazard ratio was derived from a bivariable Cox regression model. The survival curves were compared with a log-rank test, where a two-sided P value of less than 0.05 was considered statistically significant. Figure 2. Kaplan-Meier survival curves for in-hospital death among patients treated with and without remdesivir The hazard ratio was derived from a bivariable Cox regression model. The survival curves were compared with a log-rank test, where a two-sided P value of less than 0.05 was considered statistically significant. Results. A total of 549 patients were included. The median age was 69 years (interquartile range, 59 - 80 years), 333 (59.6%) were male, 231 were White (41.3%), and 235 (42%) were admitted from long-term care facilities. 394 (70.5%) received steroids, 192 (34.3%) received remdesivir, and 49 (8.8%) received tocilizumab. By the cutoff date for data analysis, 389 (69.6%) patients survived, and 170 (30.4%) had died. The bivariable Cox regression models showed decreased hazard of in-hospital death associated with the administration of steroids (Figure 1), remdesivir (Figure 2), and tocilizumab (Figure 3). This association persisted in the multivariable Cox regression controlling for other predictors (Figure 4). The E value for the multivariable Cox regression point estimates and the lower confidence intervals are shown in Table 1. The hazard ratio was derived from a bivariable Cox regression model. The survival curves were compared with a log-rank test, where a two-sided P value of less than 0.05 was considered statistically significant. The hazard ratios were derived from a multivariable Cox regression model adjusting for age as a continuous variable, qSOFA score, noninvasive positive-pressure ventilation, and invasive mechanical ventilation. Table 1. Sensitivity analysis of unmeasured confounding using E-values CI, confidence interval. Point estimate from multivariable Cox regression model. The E value is defined as the minimum strength of association on the risk ratio scale that an unmeasured confounder would need to have with both the exposure and the outcome, conditional on the measured covariates, to explain away a specific exposure-outcome association fully: i.e., a confounder not included in the multivariable Cox regression model associated with remdesivir or tocilizumab use and in-hospital death in patients with severe COVID-19 by a hazard ratio of 1.64-fold or 1.54-fold each, respectively, could explain away the lower confidence limit, but weaker confounding could not. Conclusion. For patients with severe COVID-19 admitted to our community hospital, the use of steroids, remdesivir, and tocilizumab were significantly associated with a slower progression to in-hospital death while controlling for other predictors included in the models.

THE FIRST AND SECOND WAVES OF THE COVID-19 PANDEMIC IN A COMMUNITY HOSPITAL ICU

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: Several countries have seen a two-wave pattern of the COVID-19 pandemic. However, clinical characteristics and outcomes between waves vary across regions. A study in England suggested a substantial improvement in survival amongst people admitted to critical care with COVID-19, with markedly higher survival rates in people admitted in the first wave compared with those admitted in the second wave, while a study in Africa, the second wave appeared to be much more aggressive. Therefore, regional-specific analyses are needed. METHODS: We retrospectively reviewed a de-identified dataset of patients with COVID-19 admitted to our community hospital ICU, from March 1, 2020, to February 28, 2021. Only molecularly confirmed COVID-19 cases defined by a positive result on an RT-PCR assay or NAAT of a specimen collected on a nasopharyngeal swab were included. We then identified patients from the first wave as those admitted during the initial peak of admissions observed at our hospital between March 1, 2020, and September 3, 2020. The second wave was defined as those admitted during the second peak of admissions observed between October 1, 2020, and February 28, 2021. Descriptive statistics were performed to summarize data. RESULTS: Between March 1, 2020, and February 28, 2021, a total of 190 patients were admitted to our community-hospital ICU. Of those, 132 (69.5%) were identified as patients from the first wave, and 58 (30.5%) were identified as patients from the second wave. The median age was not significantly different among patients from the first and second wave (69 years [IQR 59 – 78 years] vs. 69 years [IQR 61 – 77.25 years;p=.841]. Sex distribution was also not significantly different between the two waves (85/132 males [64.4%] vs. 40/58 males [69%];p=.541). A significantly higher rate of patients was admitted from long-term care facilities during the first wave compared to the second wave (77/132 [58.3%] vs. 7/58 [12.1%];p<.001). The distribution of comorbidities was similar between groups, except for neurocognitive disorders, which were mostly observed in the first wave (46/132 [34.8% vs. 7/58 [12.1%];p=.001). While the rates of invasive mechanical ventilation were similar between groups (75/132 [56.8%] vs. 36-58 [62.1%];p=.499, significant higher rates of patients received humidified high-flow nasal cannula (19/132 [14.4%] vs. 29/58 [50%];p<.001) and noninvasive ventilation (9/132 [6.8%] vs. 23/58 [39.7%];p<.001) during the second wave. Following the release of some pivotal clinical trials, more patients during the second wave received corticosteroids (87/132 [65.9%] vs. 56/58 [96.6%];p<.001) and remdesivir (19/132 [14.4%] vs. 48/58 [82.8%];p<.001). However, the in-hospital case-fatality rate was not significantly different between groups (68/132 [51.5%] vs. 32/58 [55.2%];p=.642). CONCLUSIONS: While epidemiological characteristics of patients with COVID-19 admitted to our ICU between the two waves were grossly similar, a significantly higher rate of patients was admitted from long-term care facilities during the first wave, and non-invasive ventilation and targeted therapies were used more during the second wave. The in-hospital case-fatality rate was not significantly different. CLINICAL IMPLICATIONS: In our community hospital in the Chicago North Shore area, the ICU case-fatality rate was not significantly different between two different waves of the COVID-19 pandemic. DISCLOSURES: No relevant relationships by Chul Won Chung, source=Web Response No relevant relationships by Goar Egoryan, source=Web Response No relevant relationships by Harvey Friedman, source=Web Response No relevant relationships by Emre Ozcekirdek, source=Web Response No relevant relationships by Ece Ozen, source=Web Response No relevant relationships by Bidhya Poudel, source=Web Response No relevant relationships by Guillermo Rodriguez-Nava, source=Web Response No relevant relationships by Daniela Trelles Garcia, source=Web Response No relevant relationships by Valer a Trelles Garcia, source=Web Response No relevant relationships by Maria Yanez-Bello, source=Web Response No relevant relationships by Qishuo Zhang, source=Web Response

IN-HOSPITAL MORTALITY IN MECHANICALLY VENTILATED PATIENTS WITH COVID-19 WITH DIFFERENT ARDS PHENOTYPES

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: The purpose of the study is to investigate the in-hospital mortality of mechanically ventilated, COVID-19 (i.e., severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)) patients with high lung compliance (i.e., atypical acute respiratory distress syndrome (ARDS)) compared to those with low lung compliance (i.e., classic ARDS). METHODS: It is a retrospective cohort study of patients older than 18 years diagnosed with COVID-19 infection that required mechanical ventilation (MV) for at least 24 hours between January 20, 2020, and April 30, 2020. Atypical ARDS was defined as driving pressure is less than 15 cm H2O throughout the period of MV, suggesting compliant lung based on the currently available evidence (Amato, Marcelo BP, et al., 2015). If it was impossible to maintain driving pressure less than 15 cm H2O for more than two days, the case was defined as classic ARDS with low compliance. Patients who required mechanical ventilation not more than 24 hours or expired within 24 hours since intubated and those transferred to another hospital were excluded. Patients who received remdesivir were also excluded because 95% of the patients did not receive it during their index hospitalization. The outcome was adjusted by age, sex, days of onset to ICU, the severity of illness estimated by APACHE score, and severity of ARDS based on PaO2/FiO2 ratio. RESULTS: A total of 60 patients that required mechanical ventilation for COVID-19 induced ARDS during the study period were reviewed per inclusion and exclusion criteria. In-hospital mortality of 30 patients of the atypical ARDS group was 50% during the index hospitalization whereas it was 53 % for 30 patients of the classic ARDS group (p=0.80) when both were treated with the same ARDS protocol, including low tidal volume and higher PEEP. The average duration of mechanical ventilation, length of ICU and hospital stay was 10.23, 12.33, and 12.93 days for the atypical ARDS group, respectively, compared to 16.57, 18.33, and 19.33 for the classic ARDS group (p=0.003, 0.011, and 0.004, respectively). The classic ARDS group required prone positioning (67% vs. 37%;p=0.02) and use of paralytics (73% vs. 43%;p=0.018) more frequently compared to the atypical ARDS group. CONCLUSIONS: In this retrospective cohort study of 60 patients that required mechanical ventilation for COVID-19 induced ARDS between January 20, 2020, and April 30, 2020, in-hospital mortality was not significantly different between the atypical ARDS group and the classic ARDS group. However, the duration of mechanical ventilation, length of ICU and hospital stay was significantly shorter in the atypical ARDS group compared to the classic ARDS group. CLINICAL IMPLICATIONS: The difference in the duration of mechanical ventilation between the two groups may suggest a different pathophysiologic process and a need for a different approach to COVID-induced ARDS depending on lung compliance. DISCLOSURES: No relevant relationships by Mariam Charkviani, source=Web Response No relevant relationships by Chul Won Chung, source=Web Response No relevant relationships by Harvey Friedman, source=Web Response No relevant relationships by Jooseob Lee, source=Web Response No relevant relationships by Guillermo Rodriguez-Nava, source=Web Response No relevant relationships by Daniela Trelles Garcia, source=Web Response No relevant relationships by Maria Yanez-Bello, source=Web Response

EXPERIENCE WITH REMDESIVIR FOR THE TREATMENT OF CRITICAL COVID-19 IN A COMMUNITY-BASED HOSPITAL IN EVANSTON, ILLINOIS

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: In late December 2019, a novel coronavirus named SARS-CoV-2 was discovered in Wuhan, China using deep unbiased sequencing in samples from patients with pneumonia. From its discovery, SARS-CoV-2 has caused global public health emergencies, economic crises, and innumerable deaths. To date, only corticosteroids have been proven to be effective in reducing mortality from COVID-19. From antiviral agents, remdesivir has been recently recognized as a promising therapy against COVID-19, but its mortality benefit is still a matter of controversy. In this study, we analyzed the effect of remdesivir on in-hospital death in our community hospital in the Chicago North Shore. METHODS: We retrospectively reviewed a de-identified dataset of 190 patients with COVID-19 admitted to a community hospital Intensive Care Unit (ICU) in Evanston, Illinois, from March 2020 to December 2020. Only molecularly confirmed COVID-19 cases defined by a positive result on a reverse-transcriptase-polymerase-chain-reaction (RT-PCR) assay or nucleic acid amplification test (NAAT) of a specimen collected on a nasopharyngeal swab were included. We performed a Cox proportional hazards model to analyze the effect of remdesivir on the hazard of in-hospital death in our patient population. To minimize confounders, age, qSOFA score, invasive mechanical ventilation, and other targeted COVID-19 therapies used at any given time (including corticosteroids, tocilizumab, hydroxychloroquine, colchicine, azithromycin, and atorvastatin) were forced as covariables into the model. For sensitivity analysis, we calculated the E value (with the lower confidence limit) for the obtained point estimate. The E value is defined as the minimum strength of association on the risk ratio scale that an unmeasured confounder would need to have with both the exposure and the outcome, conditional on the measured covariates, to explain away a specific exposure-outcome association fully. RESULTS: Between 190 patients admitted to the ICU, the median age was 69 years (IQR, 59 – 78 years), 125 (65.8%) were male, 62 (23.6 %) were White, and 84 (44.2%) were admitted from a long-term care facility. Of those patients, 143 (75.3) received corticosteroids, 67 (35.3%) received remdesivir, and 66 (34.7%) received both. Among survivors, 34/90 (37.8%) received remdesivir compared to 33/100 (33%) nonsurvivors. The Cox regression model showed decreased hazard of in-hospital death associated with the administration of remdesivir (Hazard Ratio [HR] 0.55;95% CI 0.29 – 0.94, p=.028). The E value for the point estimate was 3.04 and the E value for the lower confidence interval was 1.32, meaning that a confounder not included in the multivariable Cox regression model associated with remdesivir use and in-hospital mortality in patients with critical COVID-19 by a hazard ratio of 1.32-fold each could explain away the lower confidence limit, but weaker confounding could not. CONCLUSIONS: According to the data presented above, we concluded that in our patient population, the patients who did not receive remdesivir had a 65% chance of dying sooner compared to the ones who did receive remdesivir (when probability = HR/HR + 1). This could indicate a potential mortality benefit of remdesivir in critically ill patients. CLINICAL IMPLICATIONS: In our patient population, the use of remdesivir was associated with a slower progression to death in critically ill patients with COVID-19. DISCLOSURES: No relevant relationships by Chul Won Chung, source=Web Response No relevant relationships by Goar Egoryan, source=Web Response No relevant relationships by Harvey Friedman, source=Web Response No relevant relationships by Emre Ozcekirdek, source=Web Response No relevant relationships by Ece Ozen, source=Web Response No relevant relationships by Bidhya Poudel, source=Web Response No relevant relationships by Guillermo Rodriguez-Nava, source=Web Response No relevant relationships by Daniela Trelles Garcia, source=Web Response No relevant relationships by Maria Y nez-Bello, source=Web Response No relevant relationships by Qishuo Zhang, source=Web Response

Using machine learning for prediction of poor clinical outcomes in adult patients hospitalized with COVID-19

Background: As the ongoing COVID-19 pandemic develops, there is a need for prediction rules to guide clinical decisions. Previous reports have identified risk factors using statistical inference model. The primary goal of these models is to characterize the relationship between variables and outcomes, not to make predictions. In contrast, the primary purpose of machine learning is obtaining a model that can make repeatable predictions. The objective of this study is to develop decision rules tailored to our patient population to predict ICU admissions and death in patients with COVID-19. Methods: We used a de-identified dataset of hospitalized adults with COVID- 19 admitted to our community hospital between March 2020 and June 2020. We used a Random Forest algorithm to build the prediction models for ICU admissions and death. Random Forest is one of the most powerful machine learning algorithms;it leverages the power of multiple decision trees, randomly created, for making decisions. Results: 313 patients were included;237 patients were used to train each model, 26 were used for testing, and 50 for validation. A total of 16 variables, selected according to their availability in the Emergency Department, were fit into the models. For the survival model, the combination of age >57 years, the presence of altered mental status, procalcitonin ≥3.0 ng/mL, a respiratory rate >22, and a blood urea nitrogen >32 mg/dL resulted in a decision rule with an accuracy of 98.7% in the training model, 73.1% in the testing model, and 70% in the validation model (Table 1, Figure 1). For the ICU admission model, the combination of age < 82 years, a systolic blood pressure of ≤94 mm Hg, oxygen saturation of ≤93%, a lactate dehydrogenase >591 IU/L, and a lactic acid >1.5 mmol/L resulted in a decision rule with an accuracy of 99.6% in the training model, 80.8% in the testing model, and 82% in the validation model (Table 2, Figure 2). Conclusion: We created decision rules using machine learning to predict ICU admission or death in patients with COVID-19. Although there are variables previously described with statistical inference, these decision rules are customized to our patient population;furthermore, we can continue to train the models fitting more data with new patients to create even more accurate prediction rules. (Table Presented).

THE ANTI-INFECTIVE AND IMMUNOMODULATOR PARADOX: INCREASED RISK OF COINFECTION WITH HYDROXYCHLOROQUINE IN CRITICAL PATIENTS WITH COVID-19

SESSION TITLE: Late-breaking Abstract Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: October 18-21, 2020 PURPOSE: Coronavirus disease 2019 (COVID-19) is a new entity that has rapidly spread globally, claiming thousands of lives. Hydroxychloroquine, an agent used to prevent malaria and to treat autoimmune disorders, was being administered to COVID-19 cases to slow or prevent the disease. However, its use was rushed without sufficient evidence on efficacy and safety. METHODS: We retrospectively reviewed a de-identified dataset of 98 patients with COVID-19 admitted to a community hospital Intensive Care Unit (ICU) in Cook County, Illinois, from March 2020 to May 2020. Only confirmed COVID-19 cases, defined by a positive result on a reverse-transcriptase-polymerase-chain-reaction (RT-PCR) assay of a specimen collected on a nasopharyngeal swab were included. Co-infections were identified as the presence of positive blood cultures, sputum cultures, Legionella or pneumococcus urine antigen test, or respiratory viral panel. We performed a multivariable logistic regression analysis forcing variables that could be associated with increased risk of infection into the model, including central line placement, intubation, tocilizumab, intravenous steroids, colchicine, and hydroxychloroquine. RESULTS: Of 98 patients, the median age was 67 years (interquartile range, 57.75 – 74.25 years), 66 (67.3%) were males, 32 (32.7%) were Caucasian, and 56 (57.1%) were admitted from a Long-Term Care Facility (LTCF). 83.7% of the individuals had two or more comorbidities;the most frequent were hypertension (68.4%) and diabetes (51%). The most common targeted interventions included intravenous steroids (64.6%), azithromycin (42.9%), and hydroxychloroquine (34.7%). Among the group treated with hydroxychloroquine, 16 (47.1%) patients were found to have co-infections compared to 13 (20.3%) patients not treated with hydroxychloroquine (p=.006). The multivariable logistic regression showed increased odds of co-infection associated with the administration of hydroxychloroquine (odds ratio [OR] 4.04;95% CI 1.37 – 11.98, p=.012;Hosmer and Lemeshow goodness-of-fit test p=.724) and central line placement (OR 7.27;95% CI 1.93 – 27.31;p=.003). CONCLUSIONS: In this retrospective analysis of 98 adults with COVID-19 hospitalized in a community ICU, the patients who received hydroxychloroquine were found to have increased risk of co-infections. CLINICAL IMPLICATIONS: Hydroxychloroquine may increase the risk of co-infections in critical COVID-19 patients DISCLOSURES: No relevant relationships by Daniel Bustamante-Soliz, source=Web Response No relevant relationships by Chul Won Chung, source=Web Response No relevant relationships by Harvey Friedman, source=Web Response No relevant relationships by Elizabeth Patino, source=Web Response No relevant relationships by Guillermo Rodriguez-Nava, source=Web Response No relevant relationships by Daniela Trelles Garcia, source=Web Response No relevant relationships by Valeria Trelles Garcia, source=Web Response No relevant relationships by Maria Yanez-Bello, source=Web Response