Analysis of clinical trial data to better understand Long Covid

Project Summary Long COVID is a new chronic illness that represents an emerging public health crisis about which the medical community needs more information. The COVID-OUT trial was a phase 3, randomized, placebo-controlled clinical trial of early outpatient treatment of COVID-19 using a 2x3 factorial design to efficiently test 3 district treatments: metformin, ivermectin, and fluvoxamine. COVID-OUT continued follow-up assessments through 10 months after randomization. This unique long-term follow-up included monthly surveys to assess whether participants had been diagnosed with long COVID, had persistent or new symptoms, had new diagnoses, new medications, and repeat infections and vaccinations. The COVID-OUT trial also collected and quantified viral load from nasal swab samples at baseline, Days 5 and 10. This proposal seeks funding to conduct secondary analyses of already-collected data from the COVID-OUT trial to improve knowledge and understanding about long COVID. The COVID-OUT dataset is uniquely comprehensive with viral load samples and information on chronic disease development and progression, and it has less than 2% missingness for clinical outcomes during acute infection, and less than 5% missing long-term data through 9 months. Dr. Bramante started the COVID-OUT trial as a KL2 scholar, and the proposed R03 would support Dr. Bramante's development into a fully independent translational researcher and conducting the proposed analyses will also give Dr. Bramante important insights into how to balance feasibility of clinical trial conduct and depth of data collected, which will inform future clinical trials. Aim 1 will replicate two newly emergent symptom-based definitions of long COVID, from ACTIV-6 and the RECOVER prospective cohort, in the COVID-Out monthly follow-up data. This will serve to understand whether the comparisons of medications versus placebo in the trial are supported across new definitions of long Covid. These are post-hoc, therefore hypothesis-generating analyses. Aim 1a will assess the specific symptom phenotypes identified in the RECOVER prospective cohort, and present descriptive analyses of symptoms at each month after randomization. Aim 2 will create a predictive model of Long COVID using this comprehensive dataset that includes: baseline demographic data, home medications and comorbidities, viral load data, outcomes and treatments received during initial acute COVID infection and subsequent infections, and vaccinations and boosters received before and after enrollment. This model can be repeated for the long COVID outcomes replicated in Aim 1, and will be important for adding information to the medical literature to better understand risks associated with developing long Covid. Aim 3 will assess whether changes in sleep, physical activity, and weight effect outcomes during acute COVID-19 infect or during long-term follow-up, as sleep, adiposity, and physical activity influence the immune system. The use of existing data for analyses proposed in this R03 would address communication and logistical roadblocks that exist when trying to define and efficiently research new diseases that arise in pandemic proportions.