Graph-Based Fusion of Imaging and Non-Imaging Data for Disease Trajectory Prediction

This study proposes a graph convolutional neural networks (GCN) architecture for fusion of radiological imaging and non-imaging tabular electronic health records (EHR) for the purpose of clinical event prediction. We focused on a cohort of hospitalized patients with positive RT-PCR test for COVID-19 and developed GCN based models to predict three dependent clinical events (discharge from hospital, admission into ICU, and mortality) using demographics, billing codes for procedures and diagnoses and chest X-rays. We hypothesized that the two-fold learning opportunity provided by the GCN is ideal for fusion of imaging information and tabular data as node and edge features, respectively. Our experiments indicate the validity of our hypothesis where GCN based predictive models outperform single modality and traditional fusion models. We compared the proposed models against two variations of imaging-based models, including DenseNet-121 architecture with learnable classification layers and Random Forest classifiers using disease severity score estimated by pre-trained convolutional neural network. GCN based model outperforms both imaging-only methods. We also validated our models on an external dataset where GCN showed valuable generalization capabilities. We noticed that edge-formation function can be adapted even after training the GCN model without limiting application scope of the model. Our models take advantage of this fact for generalization to external data. © 2023 IEEE.

Virtual-Reality Simulation Can Teach VATS Lobectomy to Lobectomy-Naive Trainees Within 4 Weeks: A Cross-Over Study

Objective: The aim of the cross-over study was to evaluate skill acquisition in lobectomy-naive surgical trainees completing a 4-week program to learn VATS lobectomy on a virtual reality surgical simulator (LapSim). Method(s): Lobectomynaive surgical trainees (year 1 and 2 postgraduation) were enrolled during the COVID pandemic from March to June 2021 for a 4-week course on basic VATS skills and right upper lobectomy. They were divided into 2 groups. Both groups completed an initial assessment, Group 1 completed the course first, then both groups completed a second assessment. Then Group 2 completed the course, and both groups completed a final assessment. Skill acquisition was assessed using instrument movement, procedure time, and blood loss for both the trained operation and an untrained operation (left lower lobectomy). Result(s): 16 trainees were enrolled, 10 completed the training program. There was no difference in baseline assessment. After Group 1 completed the training, they outperformed Group 2 in all metrics but this did not reach statistical significance. After training Group 2 at week 8, there was no longer difference in performance from Group 1. After completing the training program, the entire cohort showed a significant improvement in basic VATS tasks as well as lobectomies. There was statistically significant improvement in both right upper lobectomy instrument movement (P=0.002) and time (P=0.009) and left lower lobectomy time (P=0.047). Conclusion(s): This study showed that VATS simulation training on LapSim allowed junior trainees to learn advanced VATS resection during a pandemic and within 4 weeks. The acquired skills is transferrable to untrained operations. (Table Presented).

Knowledge, attitude, and oral hygiene practices among law students in Davangere City: A cross-sectional study

Aim: The aim of this study was to assess dental health awareness and oral hygiene practice among law students of R L Law College, Davangere, Karnataka. Materials and Methods: A questionnaire containing 30 questions, regarding basic knowledge, attitude, and practice was distributed among 150 Law students of R L Law college of Davangere, Karnataka, India. The data obtained were tabulated in Microsoft Excel and statistically analyzed in the Statistical Package for the Social Sciences (IBM Corporation) software. Results: In the present study, only 39.4% of the participants were aware of routine dental check-ups. 8.5% of the participants visited the dental clinic twice a year and 57% of the participants had never visited a dental clinic. 37.3% of the participants were aware of oral hygiene aids. 62% of the participants used a tongue cleaner routinely and 55.5% brushed their teeth using only horizontal brushing technique. 48.6% were aware that consuming tobacco causes cancer. Only 9.2% consumed tobacco in various forms. 49% of the participants felt that all dental procedures were painful and expressed fear. 81% of the students were unaware of dental fluorosis. 65.5% of the participants agreed that dental health education is important and 21.8% thought that it might be important. Inferential statistics using the Pearson Chi-square test showed that there was a statistically significant correlation between the responses and the Year of law degree the participant was pursuing. Conclusion: In the present study, law students have satisfactory knowledge about good oral hygiene practices and oral health care. However, the knowledge acquired must be transferred into daily practice. This can be achieved by a change in their attitude toward oral hygiene maintenance. The inclusion of oral health-oriented education programs in their curriculum would improve their knowledge along with behavior, and they would be a good model to the community. This study will help us in devising an outline for the oral hygiene education of Law students of our country based on their current attitude, knowledge, and oral hygiene practice.

A one-stop shoulder clinic to save time and resources

There is now a need more than ever to streamline services. A one-stop shoulder clinic was introduced during the COVID-19 pandemic. A total of 861 patients were seen, saving 794 future appointments. 111 patients had an ultrasound scan and 285 patients had an ultrasound-guided procedure, saving an average waiting time of 134 days. 327 patients had physiotherapy, and the average Oxford Shoulder Score improved by 8.56 at 1 year.Copyright © Royal College of Physicians 2023. All rights reserved.

COVID-19 detection on chest x-ray image using yolo based architecture

The whole world has been facing the problem of novel Coronavirus (COVID-19) since 2020. Over 88 million cases are confirmed and around 5 lacks deaths are accounted. Using the Lung-Computed Tomography (CT) Lesion Segmentation dataset, deep learning techniques may be used to quickly identify COVID-19 and the exact region that is infected. Based on CT, it is easy to identify the problem and the infected area, then assisting treatment of COVID-19. In the literature survey, research study has considered many research papers worked done work on identification of COVID-19 using chest/lungs X-ray image, and with that identified what are the deep learning-based models or methodology they have used for detecting COVID-19 result. To overcome their result, Authors have proposed a latest methodology of deep learning with the YOLO variant 7x to get optimum result of COVID -19 detection from lungs X-ray image. To identify COVID-19, Authors have applied proposed methodology on publically avail X-ray image-based dataset of COVID-19, proposed methodology has achieved good performance to detect COVID infection from lungs. © 2023 IEEE.

Development and Use of Symptom Management Plans in Guiding Care of Children Referred to a Tertiary Palliative Care Service

Background The prevalence of children living with life limiting conditions is increasing. These children often require tertiary palliative care input for management of symptoms: specialist palliative care teams develop individual symptom management plans and/or syringe driver plans (SMP/SDP) to guide management of pain and distressing symptoms. There is no national guidance on how and when such plans should be used and their implementation varies across the country. Objectives Identify the development and use of SMP/SDP in management of symptoms in children referred to a tertiary palliative care service with particular attention medications used at the end of life (EoL). Methods Seven month (February-August 2020) prospective, comparative study of all children referred to a tertiary palliative care service including: recording of all referrals with number of SMP/SDP written type and number of medications prescribed in the SMP/SDP and used by the patient at the EoL. Results 102 patients were referred to the tertiary palliative care service. 87 SMPs were written for a total of 58 patients. 43 patients had a SMP/SDP at the EoL. Median medications included in a SMP was 13 (range 3-32) with 2 medications (range 0-14) used at EoL. Median number prescribed in SDPs was 6 (range 2-10) with 2 drugs used at EoL (range 0-7). The most common medications prescribed and used were opiates and midazolam. The most common symptoms were pain, breathlessness nausea vomiting and constipation. Conclusions SMPs/SDPs are written for a significant number of patients referred to palliative care. Often only a few drugs from the SMP/SDP were required at the end of life particularly opiates and midazolam. Further study is needed to understand the purpose use and impact of SMP/SDP on symptom control particularly at the EoL. Of note this project took place during the coronavirus pandemic, and hence bears repeating when circumstance change.

Patient characteristics, anaesthetic workload and techniques in the UK: an analysis from the 7th National Audit Project (NAP7) activity survey.

Detailed contemporary knowledge of the characteristics of the surgical population, national anaesthetic workload, anaesthetic techniques and behaviours are essential to monitor productivity, inform policy and direct research themes. Every 3-4 years, the Royal College of Anaesthetists, as part of its National Audit Projects (NAP), performs a snapshot activity survey in all UK hospitals delivering anaesthesia, collecting patient-level encounter data from all cases under the care of an anaesthetist. During November 2021, as part of NAP7, anaesthetists recorded details of all cases undertaken over 4 days at their site through an online survey capturing anonymous patient characteristics and anaesthetic details. Of 416 hospital sites invited to participate, 352 (85%) completed the activity survey. From these, 24,177 reports were returned, of which 24,172 (99%) were included in the final dataset. The work patterns by day of the week, time of day and surgical specialty were similar to previous NAP activity surveys. However, in non-obstetric patients, between NAP5 (2013) and NAP7 (2021) activity surveys, the estimated median age of patients increased by 2.3 years from median (IQR) of 50.5 (28.4-69.1) to 52.8 (32.1-69.2) years. The median (IQR) BMI increased from 24.9 (21.5-29.5) to 26.7 (22.3-31.7) kg.m-2 . The proportion of patients who scored as ASA physical status 1 decreased from 37% in NAP5 to 24% in NAP7. The use of total intravenous anaesthesia increased from 8% of general anaesthesia cases to 26% between NAP5 and NAP7. Some changes may reflect the impact of the COVID-19 pandemic on the anaesthetic population, though patients with confirmed COVID-19 accounted for only 149 (1%) cases. These data show a rising burden of age, obesity and comorbidity in patients requiring anaesthesia care, likely to impact UK peri-operative services significantly.

Unlocking happiness: people's engagement for mental well-being during the lockdown of the COVID-19 outbreak

The COVID-19 lockdown has its own set of problems for the people. However, the lockdown opened many gateways for the people to make themselves happy and active in their liking. This exploratory research study has also attended to identify the constructs related to happiness resulting from lockdown. For this expressive exploration, the research was carried out on 506 respondents from the Gujarat state of India through a structured questionnaire, followed by information processing. The data was processed by applying the Kruskal Wallis test and Mann Whitney U test to derive the conclusion. When the questionnaire was administered, open-ended questions were asked to the respondents. They revealed that even in the tough time of the COVID-19 outbreak, they identified various activities that cater to happiness and chose to keep themselves engaged in motivating themselves to fight against a hard time as a ray of hope!

A Systematic Review of 'Fair' AI Model Development for Image Classification and Prediction

Purpose: The new challenge in Artificial Intelligence (AI) is to understand the limitations of models to reduce potential harm. Particularly, unknown disparities based on demographic factors could encrypt currently existing inequalities worsening patient care for some groups. Method(s): Following PRISMA guidelines, we present a systematic review of 'fair' deep learning modeling techniques for natural and medical image applications which were published between year 2011 to 2021. Our search used Covidence review management software and incorporates articles from PubMed, IEEE, and ACM search engines and three reviewers independently review the manuscripts. Result(s): Inter-rater agreement was 0.89 and conflicts were resolved by obtaining consensus between three reviewers. Our search initially retrieved 692 studies but after careful screening, our review included 22 manuscripts that carried four prevailing themes;'fair' training dataset generation (4/22), representation learning (10/22), model disparity across institutions (5/22) and model fairness with respect to patient demographics (3/22). We benchmark the current literature regarding fairness in AI-based image analysis and highlighted the existing challenges. We observe that often discussion regarding fairness are limited to analyzing existing bias without further establishing methodologies to overcome model disparities. Conclusion(s): Based on the current research trends, exploration of adversarial learning for demographic/camera/institution agnostic models is an important direction to minimize disparity gaps for imaging. Privacy preserving approaches also present encouraging performance for both natural and medical image domain. Copyright © 2022, Taiwanese Society of Biomedical Engineering.

SEVERE THROMBOCYTOPENIA DUE TO ELECTRON-BEAM POLYSULFONE DIALYZER MEMBRANE REACTION MASQUERADING AS CHRONIC IMMUNE THROMBOCYTOPENIC PURPURA

SESSION TITLE: Critical Thinking SESSION TYPE: Case Reports PRESENTED ON: 10/19/2022 09:15 am - 10:15 am INTRODUCTION: We describe a case of severe thrombocytopenia due to reaction with an electron-beam sterilized polysulfone (PS) membrane in a patient with a previous diagnosis of reported chronic immune thrombocytopenic purpura (ITP). This phenomenon has been previously described but is rarely reported. Electron-beam (e-beam) sterilized PS membranes are classically more biocompatible than cellulose-based membranes but adverse reactions may occur as demonstrated in our case. CASE PRESENTATION: An 84-year-old woman with ESRD on hemodialysis (HD) and reported chronic ITP refractory to glucocorticoids with severe thrombocytopenia at baseline presented for evaluation of chest pain. She was incidentally found to have severe thrombocytopenia and treated with high dose glucocorticoids with minimal improvement in her thrombocytopenia and transitioned to chronic glucocorticoid taper. She had a severe chronic thrombocytopenia despite glucocorticoids which was suspected to be chronic ITP and diagnosed after initiation of outpatient HD. HD was held the first few days of her admission. She was found to have multifocal pneumonia due to SARS-CoV-2 infection. She developed progressive hypoxemic respiratory failure requiring intubation with sepsis treated with vancomycin & piperacillin-tazobactam. BAL revealed ESBL Escherichia coli & transitioned to ertapenem. She developed recurrent thrombocytopenia following HD and her PLT would improve between HD sessions. Evaluation of usual culprits of thrombocytopenia was unrevealing. Reaction to the PS membrane was suspected and a cellulose-based dialyzer membrane was used instead for subsequent sessions of HD with recovery of the platelet counts to normal. The remainder of her course was significant for tracheostomy with ventilator dependence and surrogate pursued compassionate care. DISCUSSION: We describe an interesting case of severe thrombocytopenia due to PS membrane reaction which was previously labeled as chronic ITP. Usual culprits such as pseudothrombocytopenia, HIT, HIV, HCV, hypersplenism, alcohol use, nutritional deficiencies, and rheumatologic disease were excluded. Synthetic membranes like PS-membranes are traditionally regarded as more biocompatible but filter reactions are described [1]. It is hypothesized that e-beam radiation may affect dialyzer membrane integrity or structure, or produce intermediary products which may cause platelet activation, aggregation, and adsorption, and therefore thrombocytopenia [2]. There is a high prevalence of thrombocytopenia among critically ill patients undergoing HD [3]. CONCLUSIONS: Thrombocytopenia due to PS dialyzer membrane is a rarely reported phenomenon and may be underrecognized in critically ill patients. This entity should be considered in the differential diagnosis of patients undergoing HD who develop thrombocytopenia. Early recognition may reduce incidence of bleeding and need for blood products in these patients. Reference #1: Golli-Bennour EE, Kouidhi B, Dey M et al. Cytotoxic effects exerted by polyarylsulfone dialyser membranes depend on different sterilization processes. Int Urol Nephrol 2011;43: 483–490. Reference #2: Batalini F, Aleixo GF, Maoz A, Sarosiek S. Haemodialysis-associated thrombocytopenia: interactions among the immune system, membranes and sterilisation methods. BMJ Case Rep. 2019 Sep 4;12(9):e229594. doi: 10.1136/bcr-2019-229594. PMID: 31488440;PMCID: PMC6731774. Reference #3: Griffin BR, Jovanovich A, You Z, Palevsky P, Faubel S, Jalal D. Effects of Baseline Thrombocytopenia and Platelet Decrease Following Renal Replacement Therapy Initiation in Patients With Severe Acute Kidney Injury. Crit Care Med. 2019;47(4):e325-e331. doi:10.1097/CCM.0000000000003598 DISCLOSURES: No relevant relationships by Adefemi Adeyemo No relevant relationships by Zachary Chandler No relevant relationships by Bijal Patel No relevant relationships by Vandana Seeram

INTERSTITIAL LUNG DISEASE AFTER COVID-19 IN PATIENTS WITH CANCER AND WITHOUT: A PROSPECTIVE COHORT STUDY

SESSION TITLE: Long COVID: It Can Take Your Breath Away SESSION TYPE: Original Investigations PRESENTED ON: 10/16/2022 10:30 am - 11:30 am PURPOSE: Survivors of COVID-19 hospitaliaztion may be at high risk for interstitial lung disease (ILD). The incidence and natural history of post-COVID ILD may vary in cancer and non-cancer patients, particularly if survival is lower in cancer patients. We sought to determine the incidence of ILD at 3 and 6 months after hospital discharge in cancer and non-cancer patients METHODS: We analyzed a prospective cohort of patients discharged after COVID-19 hospitalization between March 2020 and March 2021. Cancer patients were referred to post-COVID clinics 3 months after discharge, while non-cancer patients self-referred to post-COVID clinics at a tertiary referral center. We classified patients into 4 groups: Group 1, asymptomatic without ILD;Group 2, symptomatic without ILD;Group 3, ILD with spontaneous improvement by 6 months;Group 4, persistent ILD at 6 months. Group 1 patients were not seen after initial visits, while all others returned at 6 and 12 months after discharge. We hypothesized that initial COVID-19 severity, measured by the Radiologic Severity Index (RSI) on admission CT, would be associated with irreversible ILD. RSI measures radiologic severity by measuring percentage of involvement (normal - 0, <25% involvement-1, 25-50% involvement-2, 50-75% involvement-3, >75% involvement–4) and multiplying by a score based on the predominant pattern of infiltrate (normal-1, ground glass-2, consolidation-3) in six zones (left, right;upper, middle, lower) to yield a score between 0-72. We used logistic regression to measure whether admission CT RSI was associated with ILD at 3 months. RESULTS: 609 cancer patients were hospitalized with COVID-19 during the study period, of whom 85 (14%) died in-hospital, and 31 (5%) were sent home to hospice. A further 63 (10%) patients died before post-COVID evaluation. Similar data were not available for non-cancer patients due to self-referral. 98 cancer patients and 75 non-cancer patients were seen in post-COVID clinics. Among cancer patients, 20 were Group 1 (20%);8, Group 2 (8%);42, Group 3 (43%);24, Group 4 (25%);5 (5%) lacked post-COVID imaging. ILD was seen in 68% of patients at 3 months, but only in 25% at 6 months. 6% of all hospitalized cancer patients developed ILD. Among non-cancer patients: Group 1, 2 patients (3%);Group 2, 3 patients (4%);Group 3, 26 patients (35%);Group 4, 14 patients (19%);30 (40%) had no post-COVID imaging. Higher RSI at COVID admission associated with ILD at 3 months in non-cancer patients (OR 1.03, 95% CI 1.00-1.07, p=0.054) but not cancer patients (OR 1.3, 95% CI 0.4-4.5). CONCLUSIONS: Initial COVID-19 severity is associated with ILD 3 months after discharge in non-cancer patients but not cancer patients. CLINICAL IMPLICATIONS: ILD is common in survivors of COVID-19 hospitalization, particularly in non-cancer patients hospitalized for severe infection. These data may guide patient selection for referral to post-COVID clinics. DISCLOSURES: No relevant relationships by Roberto Adachi No relevant relationships by Diwakar Balachandran No relevant relationships by Lara Bashoura No relevant relationships by Christopher Bertini No relevant relationships by Kodwo Dickson Owner/Founder relationship with Pulmotect, Inc Please note: 2010-present by Scott Evans, value=Royalty No relevant relationships by Saadia Faiz no disclosure on file for Bruno Granwehr;no disclosure submitted for Shannon Holloway;No relevant relationships by Maryam Kaous no disclosure on file for Fareed Khawaja;No relevant relationships by Lyndon Lee No relevant relationships by Joanna Manzano No relevant relationships by Isabel Mira-Avendano No relevant relationships by Alyssa Mohammed No relevant relationships by Mayoora Muthu No relevant relationships by Sungryong Noh Research relationship with United Therapeutics;PhaseBio Please note: $5001 - $20000 by Bela Patel, value=Grant/Research No relevant relationships by Vickie Shannon onsultant relationship with Psioxus Therapeutics Please note: 3/1/20-7/1/20 by Ajay Sheshadri, value=Consulting fee Consultant relationship with Enanta Pharmaceuticals Please note: 01/01/21-ongoing by Ajay Sheshadri, value=Consulting fee No relevant relationships by Hui Song

Life-threatening bradycardias in COVID-19 patients whilst supported on venous-venous ECMO

Background: Veno-venous extracorporeal membrane oxygenation (VV-ECMO) was increasingly used during the Coronavirus Disease 2019 (COVID-19) pandemic. COVID-19 has been associated with bradycardia however, causative mechanisms are unclear. We aimed to assess the incidence and impact of bradycardia in COVID-19 patients whilst on VV-ECMO. Method(s): Retrospective, single centre cohort study including 173 patients on VV-ECMO due to COVID 19 from Mach 2020 to March 2022. Patients with HR<60bpm were compared with non-bradycardic patients. Furthermore, bradycardic patients were divided into 3 subgroups. Groups 1 and 2 had a HR 40-60bpm, but only Group 2 received management for bradycardia;Group 3 had life-threatening bradycardia (HR<=40). Patient characteristics, clinical and laboratory findings, and patient outcomes were collected and analysed. Result(s): 88.5% of patients developed bradycardia during VV-ECMO, of which 43% developed life-threatening bradycardia. Bradycardic patients had significantly longer ECMO runs (29 (17-46) vs 12 (7-17) days, p<0.0001) and length of hospital stay (LOHS) (47 (30-67) vs 29 (22-36) days, p=0.0007). There was no difference in mortality (p=0.7973). Counterintuitively, the non-bradycardic population were more often treated with bradycardia-inducing medications, such as beta-blockers (p=0.0220). Patients in Group 3 had significantly longer ECMO duration (37 (23- 59) days, p=0.0004) and subsequently a longer LOHS (55 (37-73) days, p=0.0033) but showed no difference in mortality (p=0.3091) compared to other subgroups. Conclusion(s): Bradycardia was associated with increased ECMO duration and LOHS, especially in the case of life-threatening bradycardia. The underlying mechanisms behind the bradycardia remain unclear, however it seems unlikely bradycardia was precipitated by bradycardia-inducing medications.

Natural History and Trajectory of non-COVID-19 Acute Respiratory Distress Syndrome patients. An observational study for comparison to COVID-19 populations

Introduction: Previous studies on acute respiratory distress syndrome (ARDS) relate trajectories to initial degree of hypoxia1,2 Further work is required to deduce whether previous ARDS frameworks are applicable to COVID-19 ARDS patients. Objectives: How does hypoxia progression influence outcomes in non-COVID ARDS patients and does this differ from COVID-19 ARDS patients? Methods: Mechanically ventilated patients that met the Berlin ARDS Criteria1 were selected from the Medical Information Mart for Intensive Care (MIMIC) database.3 Daily blood gas and ventilatory settings were analysed, from the point of intubation to death or discharge, allowing longitudinal analysis with high granularity. Our primary outcome was how the trajectory of patients was dependent on their hypoxia progression. Secondary outcomes included how base characteristics and initial clinical parameters affect trajectory and outcomes. Comparative analysis was performed between the results of this study and a previous large COVID-19 ARDS study4 Results: 1,575 ICU admissions were included in the study. All results report this study first followed by the COVID-19 study.4 Overall survival rate was higher (70.2% vs 57.7%);less patients had initial moderate or severe hypoxia (54.5% vs 76.8%);less patients had worsening of hypoxia over the first 7 days (18.9% vs 31.8%);and more patients improved their hypoxia status (33.1% vs 23.5%). This study showed a smaller proportion of hypoxia nonresolvers compared to the COVID study (32.6% vs 57.9%). However, non-resolvers in the two studies had similar survival rates (58.6% vs 60.4%). Length of ICU stay (LOS) and duration of invasive mechanical ventilation (IMV) was significantly less in this study compared to the COVID-19 study regardless of hypoxia resolver status. Conclusions: Non-COVID ARDS patients have a more predictable natural history and trajectory compared to COVID-19 ARDs patients. Respiratory failure occurs less frequently and is quicker to resolve, resulting in a lower proportion of hypoxia non-resolvers. However hypoxia non-resolvers of both populations have similar survival outcomes. Despite this, COVID ARDS patient have much longer ICU length of stay and length of ventilation which has significant implications for provision of critical care resources. Further analysis of the impact of COVID-19 therapies on these outcomes is needed.

Massive pulmonary embolism post decannulation from Extra Corporeal Membrane Oxygenation: an unexpected complication

Introduction: Massive pulmonary embolism is a rare complication following Veno-Venous Extra Corporeal Membrane Oxygenation (VV-ECMO) decannulation. Management can be challenging. The authors present a case that required VV-ECMO re-cannulation and catheterdirected thrombolysis. Main body: 58-year-old gentleman, background of hypertension and asthma, admitted with severe respiratory failure secondary to COVID-19 pneumonitis. Due to lack of improvement with conventional ARDS treatment, he was referred and retrieved on VV-ECMO. After being off sweep gas for more than 24 hours he was decannulated on day 7. Five hours after decannulation the patient acutely deteriorated. He became tachycardic, hypotensive and hypoxic. A bedside TTE showed severely dilated and impaired right ventricle. The patient was started on milrinone and nitric oxide. Nevertheless, he deteriorated further and became profoundly hypoxic and hypercapnic, and a decision was made to start him on VV-ECMO. A TOE was done to guide cannulation and showed a thrombus in the RV and in the left pulmonary artery. Next day, a CT-pulmonary angiogram (CTPA) was done which showed saddle-shaped pulmonary embolism, with a large occlusive clot in the left main pulmonary artery causing complete non-perfusion of the left lung. After a multi-disciplinary team discussion, the patient had catheterdirected thrombolysis, with some haemodynamic improvement. Within 48 hours, TTE was repeated showing no significant improvement on RV function. CTPA showed very mild decrease of the clot burden. Decision was made to repeat catheter-directed thrombolysis and partial thrombectomy. Repeated imaging revealed decrease in the size of the left main pulmonary artery thrombus. It is thought that the massive pulmonary embolism could have been caused by showering of ECMO cannulas-related thrombi, which were dislodged during decannulation. Patient remained on VV-ECMO for 32 days and was decannulated successfully afterwards and was discharged home on apixaban and long-term pulmonary hypertension follow-up. Conclusion: ECMO cannulas related thrombi are not uncommon complications because of prolonged stay and coagulopathy related to ECMO circuit. However, massive embolism is rarely seen. The use of echocardiography was paramount on the differential diagnosis. In this TTE study, the right ventricle looks significantly dilated with severely impaired both longitudinal and radial functions. Additionally noted septal flattening in systole indicating RV pressure overload, diastolic notching of RVOT doppler trace consistent with significantly raised pulmonary artery pressure and mild to moderate tricuspid regurgitation. Otherwise, the left ventricle is small and has preserved function. (Figure Presented).

ENABLING REMOTE RESEARCH THROUGH DECENTRALIZED RANDOMIZED CONTROLLED CLINICAL TRIALS

BACKGROUND: Traditionally, randomized clinical trials have relied on physical research centers to support subject recruitment and participation. The COVID-19 pandemic has highlighted the need to interact with subjects who are unable to physically visit research centers. By leveraging remote technology, such clinical trials may reach subjects in isolation and broaden geographical reach. We describe a fully remote, multisite randomized controlled clinical trial of outpatient COVID-19 treatments using a technologyenabled, decentralized approach. METHODS: We conducted a remote double-blind, randomized placebocontrolled trial (COVID-OUT). We identified subjects through medical records, patient advocacy groups, testing facilities, and multiformat advertising. They were recruited via brochure, electronic message, telephone outreach, and self-referral. Research staff across sites used the Research Electronic Data Capture (REDCap) system to manage local and central enrollment and were reallocated dynamically based on trial needs. Subjects were screened by phone, consented and randomized electronically and delivered study medication by courier or same-day mail. They were followed via their preferred communication method (phone, video, text, or email) to determine the study endpoints. RESULTS: 1195 non-hospitalized adults aged 30-85 years with laboratory confirmed infection with SARS-CoV-2 were enrolled into the COVID-OUT trial through January 6, 2022 over a span of 7 months. Initially starting as a 2- arm trial with 7 sites, the study expanded to a 6-arm trial with recruitment at 8 sites. To date, 9600 subjects have been screened with an enrollment rate of 12% from 822 zip codes. 25 research coordinators are involved across 8 sites, and the rate of study completion is 90%. CONCLUSIONS: Decentralized remote studies offer an efficient, low-touch way of performing research in the COVID-19 era. Our decentralized study design enables research with infectious, isolated subjects in widespread geographies, while maintaining safety of subjects, the research team, and public atlarge. Coordination across sites via RED Cap enabled programmatic efficiencies, including the ability to redistribute staffing support across enrollment sites for study drug distribution, follow-up calls, recruitment, and event reporting. We decreased overall costs by less need for physical research space. The decentralized infrastructure enabled nimble adaptations of the protocol, including increasing follow-up periods to assess long-COVID symptoms and adding study arms for additional outpatient treatments. Given the widespread availability of mobile phones and remote communication, decentralized trials show promise for improving reach and efficiency in both pandemic and nonpandemic times.

Silent Hypoxemia in Covid-19 and Non-Covid-19 Pneumonias

Rationale Multiple anecdotal reports describe patients with COVID-19 presenting with what is referred to as silent hypoxemia (without dyspnea, tachypnea, or respiratory distress). However, there is a paucity of literature on this problem. We conducted a study to determine the prevalence of silent hypoxemia in patients with COVID-19 infection at presentation in the emergency department (ED) or on admission, and compared it with that of patients with pneumonia by any other etiology at Memorial Hermann Health System. In addition, we analyzed the differences between clinical, laboratory, hospital utilization, and mortality. Methods From electronic medical records, we identified pneumonia patients with a positive result for coronavirus on reverse transcriptasepolymerase chain reaction (RT-PCR) nasopharyngeal swab test, or diagnosis of pneumonia by influenza or any other etiology for one year, presenting to the ED or admitted to any MHHS hospital. We extracted age, sex, race/ethnicity, vital signs, oxygen saturation, intensive care unit (ICU) admission, and hospital stay for all patients. Results A total of 17,475 COVID-19 patients were compared with 2,343 patients with influenza/pneumonia of any etiology. The two groups were similar with respect to age, sex, and ethnicity. The frequency of silent hypoxemia was greater in the COVID-19 (n=232, 1.3%) group compared to influenza/pneumonia (n= 3, 0.13%) patients (p<0.001). For patients with an oxygen saturation < 90, the proportion who had silent hypoxemia was 11.7% (232/1981) in the COVID-19 patients and 2.6% (3/117) in the influenza and other viral pneumonia patients. Between COVID-19 patients admitted to the hospital, patients with symptomatic hypoxemia had a longer hospital stays (median: 11 days, IQR: 6-22) compared to those with silent hypoxemia (median: 8 days, IQR: 4-18) (p=0.001). The symptomatic hypoxemia group was more likely to be admitted to the ICU (n=457, 26.1%) compared to the silent hypoxemic patients (n=33, 15.5%) (p<0.001) and the mortality rate was higher (n=454, 26.0%) vs (36, 15.5%) (p=0.001). Conclusion The prevalence of silent hypoxemia for adults with viral pneumonia who were seen in the ED or hospitalized in a 17- hospital system in Houston, was higher in the COVID-19 infection patient group compared to patients with other pneumonias or any other etiology. Among the COVID-19 subjects, those with symptomatic hypoxemia had worse outcomes compared to those with silent hypoxemia. More studies are needed to confirm these findings.

Unsuspecting Adult Case of Rhinovirus Pleuropericarditis

Rhinoviridae are the most common cause of upper respiratory tract infections, especially in children, and often referred to as “the common cold”. Symptoms are usually mild, nasopharyngeal in nature;they have, however, been implicated in cases of infantile viral pericarditis. Its role in the presentation of adult viral pericarditis remains unclear. We present the case of a 45-year-old male with a past medical history of pre-diabetes, hyperlipidemia and hypertension with complaints of severe left-sided chest pain that worsened with movement and coughing but improved when lying supine. Two weeks prior to presentation, he had developed an intermittent cough, treated with antibiotics and steroids. On presentation to the ED, the patient was afebrile but hypotensive to 80/52 mmHg, tachycardic to 116 BPM, hypoxic to 88% on room air, improving to 91% with 3L nasal cannula. Physical examination was notable for wheezing and egophony. Laboratory findings were concerning for WBC 19.97x10-3/uL, Hgb 13.4 g/dL, CRP 176 mg/L, Ferritin 772 ug/L, D-dimer 3.70 ug/mL FEU;procalcitonin 0.2 ng/mL and troponin <0.015 ng/mL. Respiratory viral panel revealed negative COVID-19 test but positive for rhinovirus/enterovirus. Electrocardiogram showed sinus tachycardia. Chest computed tomography demonstrated moderate pericardial effusion, ground glass attenuation of the lungs bilaterally with moderate left pleural effusion and reflux of contrast into the hepatic veins, suggestive of right heart failure. Echocardiogram demonstrated small to moderate pericardial effusion. The patient was admitted with the diagnosis of acute rhino/enteroviral-associated pleuropericarditis. Broad-spectrum antibiotics, prednisone, colchicine and indomethacin were commenced. Upon clinical stabilization of his condition, steroids were discontinued and he was discharged home with close follow-up. While rhinovirus has been associated with infantile viral pericarditis, it is implicated in pneumonia and COPD exacerbations in adults but rarely reported as a cause of adult pericarditis. A case-control study of adults diagnosed with acute idiopathic pericarditis had an independent association with an upper respiratory tract infection or gastroenteritis in the month preceding pericarditis diagnosis but did not delineate causative viruses. Therefore in cases of unknown causes of viral pericarditis, thorough history is vital. Steroids as part of the treatment algorithm for pleuropericarditis management has long been debated. Older literature has not favored the use of steroids due to high recurrence rate. However, Perrone et al refuted this point, noting that low-dose steroids with gradual tapers have equal efficacy and recurrence rates as compared with NSAIDs/colchicine. Therefore, steroids may be a reasonable option for patients with contraindications to NSAIDs/colchicine.