In a large cohort of PCR-confirmed COVID-19 outpatients, we evaluated the prevalence of persistent symptoms 3 to 10 months after diagnosis. We compared COVID-positive to a cohort of patients with negative SARS-CoV-2 PCR with symptoms suggestive of COVID during the same period in the same study sites. In these populations consisting mostly of young and healthy participants, most working as HCW, we showed that 53% of patients with PCR-confirmed COVID-19 report persistent symptoms, while 37% with negative SARS-CoV-2 PCR do. Interestingly, only 4 out of the 14 surveyed persistent symptoms were independently associated with COVID status, i.e., fatigue, smell or taste disorder, dyspnea, and memory impairment, irrespective of the timing of the survey.
The type and prevalence of long-lasting symptoms in COVID-positive are consistent with the existing literature. Two studies, evaluating more than 100 laboratory-confirmed COVID-19 outpatients, reported the prevalence of symptoms lasting more than 6 months after diagnosis: 33–46% participants with at least one symptom, 14–22% with fatigue, 8% with dyspnea, and 14–15% with change in sense of smell or taste 9,10. The differences between studies may be explained by different selection criteria, data collection methods, and definitions of symptoms.
Strikingly, in our study, the proportion of COVID-positive patients with long-term symptoms was similar between the three surveyed periods. It suggests that symptoms which are present 3 months after a COVID-19 diagnosis may last for several months. Our findings are supported by a couple of longitudinal studies, which analyzed the evolution of residual symptoms and showed a stable proportion after the acute phase up to 7 months 17,18.
In our study, 23% of COVID-positive required at least one additional medical visit for COVID-19-related symptoms. These data show the burden on the ambulatory healthcare facilities that may be related to the post-COVID-19 syndrome. In contrast with previous findings 19, we did not find a statistical difference in the proportion of patients requiring an additional medical consultation between COVID-positive and COVID-negative, which may be related to the timing of the survey, COVID-negative being surveyed at a later time point.
Our results shed more light on the association between SARS-CoV-2 infection and long-term symptoms. While some symptoms were more frequent in PCR-confirmed COVID-19 outpatients than in COVID-negative, the prevalence of other symptoms was not different between both groups. Specifically, the prevalence of headache and sleep disorders, two predominant persistent symptoms in COVID-negative patients, was similar between COVID-negative and COVID-positive. Both symptoms may originate from mixed etiologies including psychosocial factors and/or non-specific post-infectious consequences. The COVID-19 pandemic has greatly impacted our psychosocial behaviors due to confinement, social distancing, masks, and changes in working conditions, especially for HCW. Some symptoms may be a consequence of these unprecedented changes 20.
One study evaluating the sequelae in outpatients 6 months after COVID-19 included a small control group of healthy volunteers and showed that 33% of COVID-positive outpatients and 5% of volunteers presented symptoms 9. However, the limited sample size (21 healthy volunteers) and normal health status prevent comparison with our results. Another report also showed that 26% of seropositive HCW present long-term symptoms 8 months after mild COVID, while only 9% of seronegative do 21. Inclusion of patients irrespective of the presence of symptoms (no or mild prior symptoms) might explain the lower prevalence compared to our results.
Overweight and obesity are increasingly associated with poor outcomes in COVID-19 22,23 and we found that it was an independent predictor for the presence of any long-term symptom, fatigue, and dyspnea. One study reported an association between body mass index and persistent symptoms after more than 28 days 24. The mechanisms responsible for the post-COVID-19 syndrome are not understood yet and may be multiple 25,26. Long-term symptoms may be due to higher initial organ damages related to a higher infectious burden or a dysregulated inflammatory response. This may be true in people with overweight and obesity 27, since the adipose tissue has a high expression of angiotensin-converting enzyme 2 receptors and secretes pro-inflammatory cytokines.
We identified female gender as a predictor for the post-COVID-19 syndrome, in particular for persistent fatigue and smell or taste disorder. Although the results are inconsistent, some studies have described an association between female gender and long-COVID 24, persistent fatigue 7,8,28, post-exertional polypnea 28, and anxiety or depression 8. There may be a difference in the immune response according to the gender, as illustrated by the higher representation of women in autoimmune diseases 29, that may explain divergent findings with COVID-19.
Our study has some limitations. First, most symptoms are subjective and prone to observer bias. Symptoms may also come from an intercurrent condition at the time of COVID-19 diagnosis and not always represent sequelae of SARS-CoV-2 infection. However, inclusion of a symptomatic COVID-negative group supports the association between some persistent symptoms and SARS-CoV-2 infection. Second, inclusion criteria were slightly different between COVID-positive and COVID-negative. COVID-positive presented with any symptom suggestive of COVID-19 (cough, dyspnea, sore throat, history of fever, myalgia, or smell or taste disorders), while the inclusion criteria of COVID-negative patients were the presence of cough, dyspnea, or history of fever. However, we do not think that the symptom difference at first consultation modified our results. Indeed, among COVID-positive patients, 361/418 (86%) presented with cough, dyspnea, or history of fever, which were the inclusion criteria of the COVID-negative group. Another limitation is related to the timing of the survey, which was performed later in COVID-negative. This difference could lead to an underestimation of the prevalence of persisting symptoms in COVID-negative. However, we added a subgroup analysis restricted to COVID-positive and COVID-negative surveyed during the same period (the third one) and found the same results than in the whole cohort. We also observed a difference in the characteristics of COVID-positive surveyed during the first period compared to COVID-positive surveyed during the other two periods. This difference is the result of chance as patients as clinicians called patients randomly. However, we adjusted the multivariate analyses for these characteristics and for the timing of the survey.
Our study suffers from a selection bias due to SARS-CoV-2 test criteria (HCW or presence of a risk factor of adverse outcome) at the time of the study, which may prevent generalization of our findings to a broader population. We added a subgroup analysis restricted to HCW and found the same results than in the complete population. This information suggests that our results are generalizable to HCW. Furthermore, we cannot formally exclude that some COVID-negative had an undiagnosed SARS-CoV-2 infection due to a false negative PCR result. However, we used a validated SARS-CoV-2 RT-PCR test on a nasopharyngeal swab performed by a dedicated trained medical team to minimize technical and sample collection bias 12,30.
In conclusion, our study shows that more than half of outpatients with mild-to-moderate COVID-19 report long-term symptoms 3 to 10 months after diagnosis and that 21% seek medical care for this reason. These data suggest that the post-COVID syndrome places a significant burden on society and especially healthcare systems. There is an urgent need to inform physicians and political authorities about the natural long-term course of COVID-19 in order to plan an appropriate and dedicated management of those with disabling persistent symptoms.