Infection with SARS-CoV-2 can lead to prolonged or new symptoms or conditions, often referred to as ‘Long COVID.’ Although definitions vary, the Centers for Disease Control and Prevention (CDC) refers to Long COVID as post-COVID conditions that emerge, recur, or persist beyond 4 weeks following the acute infection.¹ Long COVID encompasses a multitude of symptoms affecting individual’s neurocognitive, cardiovascular, pulmonary, musculoskeletal, and psychological health, and impacting people’s ability to work and function in their daily lives.

To date, several prior studies estimate the incidence of Long COVID following an acute COVID infection to be 10–30%;^2–8 however, characterization of the epidemiology of Long COVID has been challenging to study given the range of presentations of Long COVID, different sampling and follow-up methods, and the lack of a comparison group.⁹ Most prior studies did not systematically and prospectively evaluate participants from the time of SARS-CoV-2 infection through the post-acute period, which may reveal heterogeneity in symptom trajectory and severity. Furthermore, many of these studies were retrospective and relied on convenience sampling at the time of enrollment, thereby introducing risk of selection bias, recall bias, and confounding. Consequently, there is a need to better understand the post-COVID period by systematically investigating the persistence, resolution, and even emergence of new symptoms beyond the acute infectious period in a prospective manner. Moreover, there is a need to systematically evaluate long-term symptoms that may arise from SARS-CoV-2 infection and which may overlap with symptoms that are common among other post-infectious syndromes or medical conditions,¹⁰ or potentially even as a result of the physical and mental toll of a pandemic.

Accordingly, the Innovative Support for Patients with SARS-CoV-2 Infections Registry (INSPIRE) was designed to prospectively assess the longitudinal symptoms and outcomes of adults who test positive for SARS-CoV-2 compared with symptomatic adults who test negative by following them from time of testing for up to 18 months, and to determine whether symptoms and recovery differed by age and care setting when enrolled in the study. This initial interim analysis focuses on the early trajectories of participants at 3-month follow-up with the purpose of describing the prevalence and types of symptoms among persons who had acute SARS-CoV-2 infection compared to a group of persons with COVID-like symptoms who were SARS-CoV-2 negative.

In this prospective cohort study comparing symptomatic adults who test positive for SARS-CoV-2 versus symptomatic adults who test negative, we found that participants in the COVID+ group had more symptoms at baseline as compared to the COVID− group. At 3-month follow-up, most symptoms resolved; however, persistent or emergent symptoms were more common in the COVID+ group with approximately 1-in-2 persons (as compared with 1-in-4 persons in the COVID− group) having at least one symptom. The five most common persistent or emergent symptoms at 3 months in the COVID+ group in descending frequency were: tiredness, loss of smell, muscle aches, loss of taste, and headache. The five most common persistent or emergent symptoms in the COVID− group at 3 months in descending frequency were: tiredness, muscle aches, headache, runny nose, and joint pains.

We also found that the symptoms of post-infectious syndromes, including fatigue/ tiredness/exhaustion, muscle aches, joint pains, unrefreshing sleep, forgetfulness, difficulty thinking or concentrating, and dizziness or fainting were similarly common in both the COVID+ and COVID− participants at 3 months, with 1-in-6 individuals in both groups reporting severe post-infectious symptoms. The similarity of post-infectious symptoms in the COVID+ and COVID− groups support prior studies that these types of symptoms may follow a variety of viral and non-viral infectious insults.¹⁰ However, it should also be noted that individuals in the COVID− group were more likely to report having these symptoms prior to their acute illness, suggesting that other chronic conditions may explain the high rate of symptoms in the COVID− group. The COVID− group was older and had more comorbidities underscoring the toll of recovery for younger, healthier people with COVID. Additionally, these symptoms may not be a pure response to infection or underlying medical conditions but reflect a general societal effect of the pandemic (with social isolation, disruption of daily lives, financial stress, and grief);^20–22 presumably, this could affect both cases and controls leading to overestimation of the effects of SARS-CoV-2 infection and other acute illnesses. Taken together, the recovery from COVID can be marked by significant fatigue, sleep disturbances, muscle aches, and cognitive impairment, which concerningly, is not unlike other non-COVID illnesses that develop in older, more co-morbid cohorts.

Several prior studies awakened the public and medical community to the prevalence of Long COVID, and there have been calls for a more structured research initiative to better understand the epidemiology, characterization, and treatment of this condition.^{6, 7,8} However, some of these studies lacked an adequate control group and may have been subject to selection bias based on recruitment methods.⁹ More recent data demonstrate a higher prevalence of Long COVID symptoms among COVID+ as compared with COVID− individuals, although no difference in biomarkers and basic cardiopulmonary testing.²³ The CDC reports that 1-in-5 survivors of COVID-19 between the ages of 18–64 years and 1-in-4 aged ≥65 years experienced at least one condition that may be attributable to SARS-CoV-2 infection.²⁴ Our data builds on this early and important work, demonstrating not only that COVID+ individuals continue to have a range of symptoms at 3 months, but that COVID− individuals who had symptoms at the time of testing but were negative for SARS-CoV-2 infection, also have a high burden of symptoms at 3 month follow-up. This study and that of Wisk et al (submitted as companion paper; reference to follow) evaluating well-being at 3 months both highlight the importance of discerning the long-term effects of SARS-CoV-2 illness from other viral illnesses and conditions, and in relation to the general effects of a pandemic as well as to individuals’ underlying health status and social determinants of health. Future research could build upon this complex and intersecting framework.

There are several important limitations to consider. First, testing is imperfect – with false positive and false negative results – which may have led to misclassification. We did not query participants about the test type at time of enrollment or check antibody status, limiting our capacity to determine test performance and rates of misclassification. Additionally, we cannot determine whether symptoms in the COVID+ group are necessarily because of SARS-CoV-2 infection or due to another virus or condition. Relatedly, a minority of patients in the COVID− group had loss of taste and smell, which is common for COVID-19 illness; however, it can also be observed with other viral illnesses or may simply reflects the participant’s reporting of rhinorrhea as opposed to clinically defined anosmia. Several steps were taken to mitigate misclassification. We restricted the study cohort to those with symptoms that are most specific to COVID, which improves COVID-19 diagnostic test performance (i.e., the sensitivity and specificity of the testing for SAR-CoV-2 is higher compared to if we included an asymptomatic group).²⁵ Additionally, to be enrolled in the study, participants needed to provide verification of the COVID test result. Third, at each survey timepoint we asked whether participants tested positive for COVID in the interim from the last survey. Therefore, we believe the rate of misclassification to be low. Finally, any misclassification would bias the results to the null. Second, we observed a discordance between the report of symptoms at screening to confirm study eligibility versus upon completion of the baseline survey, particularly among COVID− participants, likely attributable to allowance of up to 42 days for the initial survey completion. COVID− participants completed baseline surveys on average 9 days following enrollment in comparison to COVID+ participants who completed baseline surveys on average 3 days after enrollment, thereby exacerbating known challenges in self-reported symptom recall. Third, this report is limited to three-month outcomes and may not reflect the longer-term impact. Future work will assess symptoms beyond the three-month time period. Fourth, focusing on interim data of the first 1,000 participants enrolled between December 2020 and September 2021 precluded analysis by different SARS-CoV-2 variants. Fifth, stratified analyses by age group were limited by small sample size. For this reason, we lowered the age limit of the older group to 50 and above; future analyses will focus on the 65 and older age group. Sixth, recruitment was based out of eight academic medical centers, which might reduce generalizability. However, sites were intentionally selected for diversity of geography and populations and were not limited to enrolling individuals from their health system; rather, individuals were broadly recruited and represent over 20 different states. Seventh, as the study required participants to have access to an internet-capable device, this may have led to selection bias. The response rate at 3 months was 70.3% with differences in response rates by demographic characteristics.

Adults with acute symptoms suggestive of SARS CoV-2 infection, both those COVID+ and those COVID−, experience a range of symptoms, most of which resolve by 3 months. Still, one-in-two persons in the COVID+ group and one-in-four persons in the COVID− group had at least one post-acute symptom at three months with the most common symptoms being fatigue, headache, loss of taste and/or smell, joint pains, and muscle aches. Post-infectious symptoms such as fatigue, sleep disturbance, and cognitive impairments, generally considered a hallmark feature of post-COVID conditions, occurred at similar rates in the COVID+ and COVID− groups. These findings suggest that recovery from COVID is at least as severe as other post-infectious syndromes occurring in an older, more co-morbid COVID- group.