Deep learning based COVID-19 detection using medical images: Is insufficient data handled well?

The deep learning is a prominent method for automatic detection of COVID-19 disease using medical dataset. This paper aims to give the perspective on the data insufficiency issue that exists in COVID-19 detection associated with deep learning. The extensive study on the available datasets comprising CT and X-ray images are presented in this paper, which can be very much useful in the context of deep learning framework for COVID-19 detection. Moreover, various data handling techniques that are very essential in deep learning models are discussed in detail. Advanced data handling techniques and approaches to modify deep learning models are suggested to handle the data insufficiency problem in deep learning based COVID-19 detection.

UTILITY OF RISK CALCULATORS AND INFLAMMATORY MARKERS IN THE ASSESSMENT OF OUTCOMES FOR PATIENTS WITH COVID-19 ON EXTRACORPOREAL MEMBRANE OXYGENATION

SESSION TITLE: ECMO and ARDS in COVID-19 Infections SESSION TYPE: Rapid Fire Original Inv PRESENTED ON: 10/17/2022 12:15 pm - 1:15 pm PURPOSE: Before the Coronavirus disease 2019 (COVID-19) pandemic, the use of extracorporeal membrane oxygenation (ECMO) specific scoring systems have been shown to predict survival better compared to general risk scores used in the intensive care unit (ICU). However, data is lacking on the utility of these scores in predicting mortality in COVID-19 patients managed with ECMO. Also, inflammatory markers have been reported to be predictors of mortality in patients with COVID-19 but have not been studied well in ECMO patients. Our study aims to assess the utility of standard ICU [Acute Physiology And Chronic Health Evaluation (APACHE-IV)] and ECMO specific [Respiratory Extracorporeal Membrane Oxygenation Survival Prediction (RESP)] risk calculators along with inflammatory markers in COVID-19 patients treated with ECMO. METHODS: This study is a retrospective analysis of all adult patients with COVID-related acute respiratory distress syndrome (ARDS) admitted to the ICUs of a quaternary care hospital between 03/01/2020 and 03/31/2022 who were managed with ECMO. Demographic and clinical characteristics, inflammatory markers at the time of hospital admission, and respiratory parameters at the time of ECMO cannulation of the patients who survived were compared and analyzed with patients who did not survive. RESULTS: Of the 36 COVID-19 patients managed with ECMO during the study period, 12 (33%) patients survived. Both groups were similar in terms of age, gender, and comorbidity burden (measured by Charlson Comorbidity Index). The severity of illness at the time of ICU admission (assessed by APACHE IV score) was not significantly different between two groups [Median (IQR) = 58 (50-67) vs. 60 (52-71), p= 0.627]. D-dimers at the time of ICU admission were significantly lower in the survivor group as compared to their counterpart [Median (IQR) ng/ml = 1238 (1021-1830) vs. 2990 (1298-16583), p= 0.031]. RESP score at the time of ECMO cannulation was significantly higher among the survivors as compared to non-survivors (Mean ± SD = 3.7 ± 2.2 vs. 1.3 ± 3.5, p= 0.042). CONCLUSIONS: Our study showed that higher D-dimers at the time of hospital admission and lower RESP score at the time of ECMO cannulation are associated with increased morality in patients with COVID-19 related ARDS placed on ECMO. Knowledge of these factors may assist with determining appropriate candidates for this limited resource as well as may enhance outcome predictions. Our study is limited by a relatively small sample size and therefore larger studies will be needed to validate our findings. CLINICAL IMPLICATIONS: This study shows that similar to pre-COVID studies, RESP score be useful in risk stratification of COVID-19 patients treated with ECMO. DISCLOSURES: No relevant relationships by ALEENA ARSHAD No relevant relationships by Dipak Chandy No relevant relationships by Oleg Epelbaum No relevant relationships by Daniel Greenberg No relevant relationships by Theresa Henson No relevant relationships by Areen Pitaktong No relevant relationships by Hamid Yaqoob

ASSOCIATION BETWEEN ETHNICITY AND COVID-19 PATIENTS TREATED WITH EXTRACORPOREAL MEMBRANE OXYGENATION

SESSION TITLE: Impact of Health Disparities and Differences SESSION TYPE: Rapid Fire Original Inv PRESENTED ON: 10/19/2022 11:15 am - 12:15 pm PURPOSE: As of March 25, 2022, age-standardized data reported by the Centers for Disease Control and Prevention showed that Hispanic, Black and American Indian or Alaska Native are about twice as likely to die from coronavirus disease 2019 (COVID-19) as their White and Asian counterparts. However, there is paucity of data regarding the effect of race on outcomes in COVID-19 related acute respiratory distress syndrome (ARDS) patients managed with extracorporeal membrane oxygenation (ECMO). Our study aims to understand the differences in the outcome between White/Asian and other ethnically minority COVID-19 patients treated with ECMO in our intensive care unit (ICU). METHODS: Retrospective analysis of adult patients with COVID-19 related ARDS treated with ECMO in the ICUs of a quaternary care hospital between 03/01/2020 and 03/31/2022. Patients were divided into two groups: White/Asian (WA) and Other Minorities (OM). Demographics, clinical characteristics, and outcomes of the two groups were compared. RESULTS: Of the 36 COVID-19 patients managed with ECMO during the study period, 18 (50%) patients belonged to the WA group while 18 (50%) patients belonged to the OM group. In the WA group, 16 (89%) were white and 2 (11%) were Asians whereas in the OM group, 16 (89%) patients were Hispanics and 2 (11%) patients were African-American. Both groups were similar in terms of age, gender, comorbidity burden (measured by Charlson Comorbidity Index), and severity of illness at the time of ICU admission (assessed by APACHE-IV score). Mean RESP score was lower in the OM group but was not statistically significant (1.3 ± 3.9 vs 2.9 ± 2.3, p= 0.157). This was reflected in the higher hospital mortality in the OM group compared to the WA group [n= 9 (50%) vs. 15 (83%), p=0.075]. There was no significant difference between the groups in the rate of ECMO-related complications, including major bleeding requiring transfusion, transaminitis (alanine transaminase greater than 5 times of upper normal limit), stroke, myocardial dysfunction (defined as an ejection fraction < 30%), acute kidney injury requiring dialysis and positive sterile fluid cultures. CONCLUSIONS: Our study showed higher mortality in ethnically minority patients compared to the white and Asian population but the difference was not statistically significant. It is possible that the relatively small number of patients in our study led to a beta error. Higher mortality rates among people of color have been attributed to low socio-economic status, structural inequities in health care and differences in vaccination rates. CLINICAL IMPLICATIONS: Larger studies are needed to further explore differences in clinical characteristics and outcomes of COVID-19 patients of different races and ethnicities treated with ECMO. DISCLOSURES: No relevant relationships by ALEENA ARSHAD No relevant relationships by Dipak Chandy No relevant relationships by Subo Dey No relevant relationships by Oleg Epelbaum No relevant relationships by Daniel Greenberg No relevant relationships by Theresa Henson No relevant relationships by Lawrence Huang No relevant relationships by Daniel Peneyra No relevant relationships by Areen Pitaktong No relevant relationships by Hamid Yaqoob

EXTRACORPOREAL MEMBRANE OXYGENATION: A HOPE FOR THE SICKEST PATIENTS WITH COVID-19

SESSION TITLE: ECMO and ARDS in COVID-19 Infections SESSION TYPE: Rapid Fire Original Inv PRESENTED ON: 10/17/2022 12:15 pm - 1:15 pm PURPOSE: The role of extracorporeal membrane oxygenation (ECMO) for patients with coronavirus disease 2019 (COVID-19) related acute respiratory distress syndrome (ARDS) is evolving. Data from different waves of the pandemic has shown that mortality of COVID-19 patients treated with ECMO ranges from 40-94%. Pre-COVID studies have shown that ECMO is associated with bleeding in 30-50% of patients, thrombotic complications in about 10% and secondary infections in 40% of patients. However, there is a paucity of data regarding complications with the use of ECMO in COVID-19 patients. In this study, we describe the hospital course and complications seen in the COVID-19 patients admitted to our Intensive Care Unit (ICU) treated with ECMO. METHODS: Retrospective cohort analysis of adult patients with COVID-19 related ARDS admitted to the ICUs of a quaternary care hospital between 03/01/2020 and 03/31/2022 who were managed with ECMO. RESULTS: Of the 36 COVID-19 patients treated with ECMO, 23 (64%) patients were male. Median age was 48 years (IQR 36-59). Patients had a mean BMI of 36 ± 12. Median Charlson Comorbidity Index (assesses comorbidities) was 1 (0-2) and median APACHE-IV score (assesses severity of illness at the time of ICU admission) was 60 (51-72). Prior to initiation of ECMO, 14 (39%) patients were proned and 29 (81%) patients received a trial of neuromuscular blockade. Patients had high plateau pressures (mean 31 ± 8 cm H20) with pO2/FiO2 ratios consistent with severe ARDS (mean 63 ± 17) at the time of ECMO cannulation. Mean Respiratory ECMO Survival Prediction (RESP) score was 2.1 ± 3.3. The most common complications were bleeding requiring transfusion seen in 94% of patients and positive sterile fluid cultures (53% patients). Hemorrhagic stroke was seen in 3 patients (8%). None of the patients had limb ischemia or clotting of the cannula requiring catheter exchange. Withdrawal of care occurred in 3 patients (8%). 13 (35%) patients were successfully decannulated from ECMO;however only 12 (33%) patients were discharged alive. CONCLUSIONS: Our study shows a survival rate in COVID-19 patients treated with ECMO that is comparable to previously reported studies. High bleeding and infection rates can possibly be explained by steroid use and COVID-19 disease specific characteristics. CLINICAL IMPLICATIONS: Our study describes the hospital course of the COVID-19 patients treated with ECMO and can be used to evaluate it's role in the management of severe COVID-19 patients refractory to conventional ventilatory management. DISCLOSURES: No relevant relationships by ALEENA ARSHAD No relevant relationships by Dipak Chandy No relevant relationships by Oleg Epelbaum No relevant relationships by Daniel Greenberg No relevant relationships by Theresa Henson No relevant relationships by Lawrence Huang No relevant relationships by Daniel Peneyra No relevant relationships by Areen Pitaktong No relevant relationships by Hamid Yaqoob

Emergency Department Management Using Regression Models

COVID-19, a global pandemic has been ravaging the world. The Emergency departments are flooded because of this global pandemic. To provide a good service in the Emergency Departments (ED) in hospitals as a part of smart healthcare, tools that analyze, program, plan or prioritize is required to use the available resources (staff and treatment equipment) in a fine possible way. In this paper, the various queuing methods that are implemented to tackle the patient flow in EDs as well as outpatient departments in the already existing systems are surveyed, and a method is suggested. The previous papers have taken queuing theory into account. Here three different regression techniques namely Linear regression, polynomial regression, and support vector regression are considered for the prediction of the patient flow in Emergency Departments. The hazardous COVID-19 pandemic and its impact on the mounting crisis in the EDs is also discussed. The challenges and suggestive methods are also discussed here. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Comparison of Extracorporeal Membrane Oxygenation Survivors with Non-Survivors;A Retrospective Analysis at a Quaternary Center

Rationale: International organizations, including the World Health Organization and Extracorporeal Life Support Organization have recommended the use of extracorporeal membrane oxygenation (ECMO) in the management of patients with Coronavirus Disease-19 (COVID-19) related acute respiratory distress syndrome (ARDS) based on favorable outcomes shown by some studies from earlier parts of the pandemic. Mortality rates of COVID-19 patients with the use of ECMO has ranged from 40-94%. Some reports suggest early initiation of ECMO leads to better outcomes before refractory hypoxemia leads to multi-organ failure. However, the predictors of mortality among COVID-19 patients treated with ECMO remain unclear. Also, ECMO has been associated with potentially life-threatening bleeding and thromboembolic complications. This study aims to identify the risk factors associated with the mortality in COVID-19 patients managed with ECMO and to assess the effect of ECMO related complications on mortality. Methods: Retrospective analysis of adult patients with COVID-related ARDS treated with ECMO at the ICUs of a quaternary care hospital between 03/01/2020 and 08/31/2021. Demographics, clinical characteristics, and outcomes of the patients who survived were compared with those who did not survive. Hemorrhagic complications were defined as bleeding requiring transfusion, hemorrhagic stroke and coagulopathy with International Normalized Ratio (INR) > 3. Thrombotic complications were defined as limb ischemia and ischemic stroke. Myocardial dysfunction was defined as a drop in ejection fraction to less than 30%, liver dysfunction as alanine transaminase (ALT) greater than 5 times of upper normal limit (ULN), and kidney involvement as acute kidney injury (AKI) requiring dialysis. Results: Of the 31 COVID-19 patients managed with ECMO during the study period, 11 (36%) patients survived. Both groups were similar in terms of age, gender, comorbidity burden (measured by Charlson Comorbidity Index), and severity of illness at the time of ICU admission (assessed by APACHE-IV score). Days spent on mechanical ventilation (MV) before ECMO cannulation were lower in survivors as compared to non-survivors but the difference was not statistically different. The incidence of complications was not statistically different between two groups. Conclusion: Our study shows a survival rate in COVID-19 patients treated with ECMO that is similar to previously reported studies. Our study did not reveal any significant predictive differences between survivors and nonsurvivors, thereby continuing to make the process of patient selection for ECMO challenging during this pandemic. Our study is limited by a relatively small sample size and therefore larger studies will be needed to confirm our findings.

PREDICTORS AND OUTCOMES OF BACTERIAL AND FUNGAL SUPERINFECTIONS IN CRITICALLY ILL PATIENTS WITH COVID-19

TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: Secondary infections are well-described complications in patients with viral pneumonia and are associated with increased mortality and morbidity (1). A recent meta-analysis reported that secondary infections in critically ill Coronavirus Disease 19 (COVID-19) patients in intensive care unit (ICU) are around 14% (2). Most of the studies are limited by false-positive cultures in non-sterile body fluids (2). Our study aims to assess the burden of secondary infections by focusing on the positive rate of sterile body fluid cultures in ICU patients, to examine associated risk factors and to assess their impact on outcome. METHODS: Retrospective analysis of all adult patients with COVID-related acute respiratory distress syndrome (ARDS) admitted to the ICUs of a quaternary care hospital between 03/01/2020 and 05/31/2020. Patients who had positive sterile body fluid (blood, peritoneal, pleural or cerebrospinal fluid) cultures for bacteria or fungi, were compared with those without positive cultures. RESULTS: Of the 210 patients admitted to our ICU with COVID-19 related ARDS, 55 patients (26%) had positive sterile body fluid cultures, of which 37 grew bacteria, 7 fungi and 11 both. Patients with positive cultures were similar in terms of baseline demographics, comorbidities and severity of illness at ICU admission to those patients without positive cultures. Peak inflammatory markers were significantly higher in the positive culture group. In the positive culture group, the rates of receipt of steroids (78% vs. 61%, p=0.02) and convalescent plasma (64% vs. 45%, p=0.03) were significantly higher. The prevalence of acute kidney injury (AKI) requiring dialysis was significantly greater in the positive culture group (55% vs. 28%, p<0.001). There was no difference in terms of mortality between the two groups. However, mechanical ventilation (MV) free days at day 28 were significantly lower in the positive culture group (Mean ± S.D: 2.7 ± 5.7 vs. 6 ± 9.9, p=0.004). CONCLUSIONS: Our study shows that the incidence of superinfections in COVID-19 patients admitted to ICUs is higher than previously reported. The higher incidence is comparable to the rate of superinfections in the H1N1 influenza pandemic of 2009 (2). However, the high rates of steroid and plasma administration in our positive culture group may also have contributed to the high incidence of secondary infections. Our results have also shown that these infections may prolong ventilatory support as well as increase the likelihood of developing AKI requiring dialysis. Therefore, intensivists need to be vigilant about not missing superinfections in patients with COVID-19 which can negatively impact patient outcomes. CLINICAL IMPLICATIONS: This study will help to identify the risk factors associated with higher incidence of secondary infections in patients with COVID-19, and will assist physicians to identify and treat them early in the course of disease. 1. I. Martín-Loeches, A. Sanchez-Corral, E. Diaz, R.M. Granada, et al., H1N1 SEMICYUC Working Group Community-acquired respiratory coinfection in critically ill patients with pandemic 2009 influenza A(H1N1) virus. Chest, 139 (2011), pp. 555-562. 2. Lansbury L, Lim B, Baskaran V, Lim WS. Co-infections in people with COVID-19: a systematic review and meta-analysis. J Infect. 2020;81(2):266-275. DISCLOSURES: No relevant relationships by ALEENA ARSHAD, source=Web Response No relevant relationships by Dipak Chandy, source=Web Response No relevant relationships by Oleg Epelbaum, source=Web Response No relevant relationships by Daniel Greenberg, source=Web Response No relevant relationships by Muhammad Rizwan, source=Web Response No relevant relationships by Hamid Yaqoob, source=Web Response

The effect of chronic immunosuppression on the severity and outcomes of COVID-19 icu patients

Rationale: There is currently limited and conflicting data regarding the effect of immunosuppression on severity and outcomes of Coronavirus Disease 2019 (COVID-19). Mortality rates of solid organ transplant recipients hospitalized with COVID-19 range as widely as 5-67%. Some of these reports therefore seem to suggest that immunosuppressed patients have a similar mortality when compared to non-immunosuppressed patients. Also, there is limited data on the incidence of bacterial and fungal superinfections in these critically ill COVID-19 patients who are immunosuppressed. Our study aims to understand the impact of immunosuppression on the clinical course and outcomes of COVID-19 patients admitted to our intensive care units (ICUs). Methods: This study is a retrospective analysis of all adult patients with COVID-related acute respiratory distress syndrome (ARDS) admitted to the ICUs of a quaternary care hospital between 03/01/2020 and 05/31/2020. Patients on chronic immunosuppressive medications were identified. Demographic and clinical characteristics, inflammatory markers at the time of ICU admission and clinical outcomes of these patients were compared and analyzed with patients who were not immunosuppressed. Means, medians and categorical variables were compared using t-test, Mann-Whitney U-test and Fisher's exact test, respectively. P-value of <0.05 was taken to be statistically significant. Results:Of the 210 patients admitted to our ICUs with COVID-related Acute Respiratory Distress Syndrome, 23 (11%) were taking immunosuppressant medications before they were admitted to our unit. 11 patients had a history of organ transplantation (Liver: 4, Kidney: 4, Heart: 2 and Stem cell transplantation: 1). There was no statistically significant difference between the two groups in terms of demographics, comorbidities, severity as indicated by inflammatory markers and outcomes such as death, acute kidney injury (AKI) requiring dialysis, and bacterial or fungal superinfection. Conclusion: Our study seems to imply that there is no significant difference in the severity and outcomes of the immunosuppressed patients who were admitted to our ICUs. Our study did show an increased incidence of mortality (52.17% vs. 44.02%) and an increased rate of positive sterile fluid cultures (34.78% vs. 25.13%) in these immunosuppressed patients but the difference was not statistically significant. It is possible that an increased sample size may have revealed statistically significant differences. Therefore, larger studies are needed to determine if immunosuppression impacts the outcome of critically ill COVID-19 patients.

Impact of elevated troponins on mortality of ICU patients

Introduction: Myocardial injury as evidenced by elevated cardiac troponin levels may occur in 7-36% of hospitalized patients with coronavirus 2019 (COVID-19). Studies have shown that COVID-19 patients with elevated cardiac troponin are more likely to require the intensive care unit (ICU) and mechanical ventilation and have a higher risk of death. However, studies on the impact of elevated troponins at the time of hospital admission on mortality of ICU patients are limited. Here we aim to characterize, compare, and analyze COVID-19 patients in our institution's ICUs who had elevated troponins. Methods:This study is a retrospective analysis of adult patients with COVID-19 admitted to the ICUs of a tertiary care hospital between 03/01/2020 and 05/31/2020. Myocardial injury was defined as troponin ≥ 0.04 ng/mL. Only patients who had troponins measured within 24 hours of hospital admission were included. We excluded patients with a past medical history of chronic kidney disease (CKD) or end-stage renal disease (ESRD). We also excluded those patients who were transferred to our ICU from another hospital and where admission labs from the transferring institution were not available for our review. We compared the clinical characteristics and outcomes of critically ill COVID-19 patients who had elevated troponins with those who did not.Results: Of the 210 patients admitted with COVID-related Acute Respiratory Distress Syndrome, 178 had troponin measured within 24 hours of hospital admission. 25 patients had a past medical history of CKD or ESRD and 11 patients were transferred from another hospital with no labs available from their day of admission. Of the 142 remaining patients, 72 (50.7%) had an elevated troponin while 70 (49.9%) had a normal troponin at presentation. Patients with elevated troponins were older (67.5 ± 15.1 vs 60.2 ± 14.2), and predominantly male (72.2% vs. 54.3%). Inflammatory markers were significantly elevated in both groups but not statistically different. Therapies provided were similar in both groups. Most importantly, patients with an elevated troponin had significantly increased risk of mortality compared to those with a normal troponin at admission (55.6% vs. 34.3%).Conclusion: Our results suggest that myocardial injury is a common occurrence in COVID-19 patients requiring ICU admission. Additionally, mortality in this population is significantly higher further indicating myocardial injury as a predictor of mortality. To determine if elevated troponin levels are independent of other risk factors for mortality in COVID-19 such as age, and gender will need to be determined in future studies.

Obesity and severe COVID-19: A retrospective analysis of intensive care unit patients at a quaternary medical center

Introduction In obese patients, it is thought that the excess of macronutrients stimulates adipocytes to generate cytokines such as interleukin-6 creating a pro-inflammatory and oxidative state leading to defective innate immunity. This environment potentially creates a conducive ground for the hyperinflammatory response mediated by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Therefore, in addition to accepted risk factors such as advanced age, pregnancy and diabetes, obesity is thought to be an important entity predisposing to severe Coronavirus Disease 2019 (COVID-19) and a potentially worse prognosis. The aim of this study is to understand the impact of obesity on the clinical course and outcome of intensive care unit (ICU) patients with COVID-19. Methods All COVID-19 ICU admissions between March 1st and May 31st, 2020 were obtained from our hospital's data base. Patients age >18 years with a positive SARS-CoV-2 polymerase chain reaction from a nasopharyngeal swab were included. ICU patients who were admitted for non-respiratory reasons with an incidental positive test result were excluded. Patients were divided into two groups obese: (BMI >/=30) and non-obese (BMI <30). All categorical variables are expressed as frequencies and percentages. Comparisons of categorical variables were performed using Fisher's Exact Test. Continuous variables are expressed as median (IQR). Medians were compared using the Mann-Whitney U-Test. P-value of <0.05 was taken to be statistically significant. The institutional review board of New York Medical College approved this study. Results Of the 210 patients who met our inclusion criteria, 95 patients were obese and 115 were nonobese. Obese patients were significantly younger than their non-obese counterparts (p=0.005). Aside from this, there was no significant difference between the groups in terms of gender, race, baseline comorbidities and severity of illness at the time of ICU admission. Outcomes such as mechanical ventilation, renal replacement therapy, stroke and death were not different between the two groups (Table 1). Conclusion Unlike other studies, ours did not find significant differences in the outcomes of critically ill obese patients with COVID-19. This might be because other authors used a BMI >25 as their cutoff whereas we used BMI >/= 30. Another potential explanation is that in our study, obese patients were younger, a factor which may have neutralized the negative impact of obesity. Our study is limited by sample size and the single center factor but suggests that obesity alone may not significantly affect prognosis although this will need to be confirmed by larger studies.

The incidence of stroke in critically ill COVID-19 patients and its association with elevated D-Dimers

Introduction:Ischemic and hemorrhagic strokes appear to be complications of Coronavirus Disease 2019 (COVID-19). Implicated mechanisms include severity of inflammation in these patients that leads to endothelial injury and a more severe prothrombotic state. Patients with a more severe presentation appear to be at higher risk. Based on observational data, intracranial hemorrhage occurs in 0.2-0.9% of hospitalized COVID-19 patients, while ischemic stroke occurs in 0.9-2.5%. Small reports have indicated that d-dimer levels preceding the stroke are higher than in COVID-19 patients without strokes. In our study we aimed to determine the incidence of stroke in our intensive care unit (ICU) patients with COVID-19, and further assess the association between ddimer levels at the time of ICU admission and ischemic/hemorrhagic stroke.Methods: This study is retrospective analysis of all adult patients with COVID-related acute respiratory distress syndrome (ARDS) admitted to the ICUs of a tertiary care hospital between 03/01/2020 and 05/31/2020. We identified patients who had a stroke based on CT head findings. Demographic and clinical characteristics, inflammatory markers at the time of ICU admission and clinical outcomes of patients with stroke were compared and analyzed with patients who did not have stroke. Means, Medians and categorical variables were compared using t-test, Mann-Whitney U-test and Fisher's exact test, respectively. P-value of <0.05 was taken to be statistically significant.Results:Of the 210 patients admitted to our ICUs with COVID related ARDS, 20 (9.5%) had strokes (18 ischemic, 2 hemorrhagic) at some point in their hospitalization. There was no statistically significant difference between the groups in terms of demographics, comorbidities, and outcomes such as death. Therapies provided, including steroids and therapeutic anticoagulation, were similar in both groups. D-dimer levels at the time of ICU admission in the patients with stroke were significantly higher than in patients without stroke (p= 0.025). Conclusion: Our study shows that the incidence of stroke in COVID-19 patients admitted to ICUs is higher than previously reported, despite most being on anticoagulation. Our study also demonstrated that COVID-19 ICU patients who developed stroke during their hospitalization had significantly higher d-dimer levels at the time of ICU admission. Our findings suggest that intensivists need to be constantly vigilant about not missing strokes in critically ill COVID-19 patients. The findings also suggest that the magnitude of d-dimer elevation at the time of ICU admission may be a useful predictor of developing strokes during the hospitalization. Larger studies are needed to confirm our findings.