Initial symptoms and three months follow-up after acute COVID-19 in outpatients: An international prospective cohort study

Abstract Background Most studies on long-term follow-up of patients with COVID-19 focused on hospitalised patients. No prospective study with structured follow-up has been performed in non-hospitalised patients with COVID-19. Objectives To assess long-COVID and post-COVID (WHO definition: symptomatic at least 12 weeks), describe lingering symptoms, their impact on daily activities, and general practice visits and explore risk factors for symptom duration in outpatients. Methods A prospective study of adult outpatients with confirmed SARS-CoV-2 infection and symptoms consistent with COVID-19 in 11 European countries, recruited during 2020 and 2021 from primary care and the community. Structured follow-up by phone interviews (symptom rating, symptom impact on daily activities and general practice visits) was performed at weeks 2, 4, 8, and 12 by study personnel. Data was analysed descriptively by using correlation matrixes and Cox regression. Results Of 270 enrolled patients, 52% developed long-COVID and 32% post-COVID-syndrome. When only considering the presence of moderate or (very) severe symptoms at weeks 8 and 12, these percentages were 28% and 18%, respectively. Fatigue was the most often reported symptom during follow-up. The impact of lingering symptoms was most evident in sports and household activities. About half (53%) had at least one general practice contact during follow-up. Obese patients took twice as long to return to usual health (HR: 0.5, 95%CI: 0.3–0.8); no other risk profile could predict lingering symptoms. Conclusion Long-COVID and post-COVID are also common in outpatients. In 32%, it takes more than 12 weeks to return to usual health.


Introduction
The ongoing pandemic caused by SARS-CoV-2 remains an international public health emergency and has resulted in enormous global morbidity and mortality. Before the new Omicron variant spread by September 2021, more than 200 million COVID-19 cases and 4.3 million deaths were recorded globally (WHO dashboard) [1].
According to the WHO definition, a symptomatic COVID-19 case is a person who has developed signs and symptoms suggestive of COVID-19, including fever, cough, fatigue, anorexia, shortness of breath, myalgia and gastrointestinal manifestations [2]. Loss of smell and/or taste is also reported, as well as various mental and neurological symptoms [3]. Mild COVID-19 illness is defined as an uncomplicated respiratory tract infection with non-specific symptoms, including fever and cough and moderate illness as pneumonia without the need for supplementary oxygen [2]. About 80% of patients with COVID-19 have mild or moderate disease, and self-manage or are managed by their general practitioner (GP); up to 6% require hospital admission [4,5].
According to the WHO, long-COVID is defined as still experiencing symptoms 2 months after a SARS-CoV-2 infection, and post-COVID syndrome as symptoms for longer than 3 months which cannot be explained by an alternative diagnosis [6].
A systematic review of post-COVID-19 cases showed that more than half experienced symptoms 6 months after recovery, including functional mobility impairments, pulmonary abnormalities, and mental health disorders [7]. However, 80% of included patients were recruited in hospital and cases were retrospectively evaluated [7]. A recent study of patient data extracted from general practice records included nearly half a million patients with a COVID-19 diagnosis, with the most common complaints after 4 weeks being joint pain (2.5%) and anxiety (1.2%) [4].
To our knowledge, no prospective study with structured follow-up of the repertoire of COVID-19 symptoms at set times has been performed in outpatients. Therefore, we set out to determine the prevalence of long-COVID and post-COVID, to identify lingering symptoms and assess their impact on daily activities and number of general practice visits in adult outpatients with COVID-19. In addition, we explored risk factors related to a longer time to return to usual health.

Methods
We set up a European, prospective, exploratory study of adult outpatients recruited from primary care and the community with laboratory-confirmed SARS-CoV-2 infection and symptoms consistent with COVID-19. National research teams were invited through the General Practice Research on Infections Network (GRIN, https://www.grinweb.org/) and the VALUE-Dx Primary Care Network (https://www.value-dx.eu) [8]. Each research team included a minimum of 10 to a maximum of 30 patients.

Patient recruitment
Patients were informed by their GP, primary care nurse, or practice assistant when they contacted or consulted the practice for their symptoms via a leaflet in testing centres, through the research team's personal contacts and via snowballing (relatives, housemates and friends of recruited patients). Patients were asked to contact the research team. Patients were asked to provide verbal informed consent during the first phone call, which was confirmed by a signed informed consent form.
The inclusion criteria were: laboratory-confirmed SARS-CoV-2 infection (PCR), symptoms consistent with COVID-19, maximum of 1 week since a positive test, age 18 years or older, able and willing to provide informed consent and to comply with phone calls. Patients hospitalised due to COVID-19 were excluded.

Data collection
After patient contact was established, the first phone interview was performed. Patients were included from October 2020 until June 2021 in Belgium (Antwerp and Leuven), Denmark, France, Germany, Georgia, Ireland, Italy, the Netherlands, Poland, Romania, and Sweden. Upon inclusion, information on age, sex, weight/height, smoking and comorbidity defined as chronic respiratory condition (e.g. asthma, COPD, CF), diabetes, cardiovascular disease or 'other' was obtained. Furthermore, the presence of COVID-19 symptoms (fever, cough, shortness of breath, sore throat, fatigue/tiredness, headache, loss of appetite, gastrointestinal manifestation, myalgia, impaired taste, impaired smell, mental symptoms, emotional sensitivity), was captured.
All patients received follow-up calls at weeks 2, 4, 8, and 12 from study personnel. Outside the scope of this manuscript, after 12 weeks, monthly follow-up was done until symptom resolution up to a maximum of 1 year. Patients were asked to rate their symptoms as: absent, very mild, mild, moderate, severe, or very severe. Impact of symptoms on daily activities (work/education, income, caring for (grand-) children, household activities, sports, social life, non-sport hobbies) was rated as: no, slight, moderate, quite a bit, or severe.
Data was collected in a secured web-based tool (SurveyXact) hosted by Aalborg University Denmark, only including the patient study ID. Each research team kept patient's name, phone number and study ID separately and secured until follow-up was completed.
A convenience sample of 250 patients was arbitrarily considered appropriate to answer the research question. This would allow for loss of follow-up of 10% and result in 75 patients (one-third expected) experiencing lingering symptoms at week 12 for analysis.

Study endpoints
The primary endpoints of the study were the prevalence of long-COVID (defined as still symptomatic 8 weeks after a SARS-CoV-2 infection) and post-COVID (defined as symptomatic for 12 weeks, or longer, which an alternative diagnosis cannot explain) in outpatients. Secondary endpoints were duration of COVID-19-related symptoms (measured at 2, 4, 8 and 12 weeks), how these impacted patients 'daily activities' and general practice visits, and risk factors (e.g. age, sex, body mass index (BMI), and comorbidity) related to time to return to usual health.

Statistical methods
Categorical data are presented as numbers and proportions and continuous data as medians with lowest and highest values. To visualise correlations between symptoms present at weeks 2 and 12, and correlations between individual symptoms and daily activity impact present at week 12, correlation matrices were generated (including correlation coefficients) in Tableau Professional Desktop Edition (2021.4.3).
A Cox regression was performed for the composite outcome 'health returned to usual.' Patients were considered to reach this outcome at the moment they did not report any more symptoms or the week they confirmed a return to normal health. Patients who did not fall in either category at week 12 were censored. A backward regression using presence of each symptom at baseline, sex, age, any comorbidity, as well as separately each individual comorbidity, overweight and obesity, was performed. Sex and age were kept in the model, but other variables remained if significant. p-Values < 0.05 were considered significant. Results are presented as hazard ratios (HR) with 95% confidence intervals (95%CI). Data analysis was performed using SPSS v27.0 software (IBM, Armonk, NY).

Specification of COVID-19-related symptoms during the first 12 weeks
At inclusion, fatigue/tiredness was the most often reported symptom (79%, 213/270), followed by headache (73%, 196/270) and cough (67%, 180/270). Figure 2 shows the trend of reported symptoms and their severities from week 2 to week 12. At week 2, fatigue, cough, loss of/impaired taste and/or smell and headache were most often reported and with highest severity. At week 12, fatigue (20%, 49/248) and loss of smell (17%, 30/248) remained the most often reported symptoms, of whom, respectively, 20% and 27% rated these symptoms as (very) severe. Additionally, myalgia, headache and shortness of breath persisted until week 12 in about 9% of patients.  Specification of impact of symptoms on daily activities Table 2 shows that lingering symptoms of COVID-19 negatively impacted most daily activities. Initially, education/work, household activities and social life were most often reported to be affected. Continued impact of symptoms was most evidents for sports, followed by education/work and household activities.    Figure 4 shows the correlation between individual symptoms at week 2 and daily activities' impact at week 12. At week 12, most severe impact was on household activities, social life and sports. Fatigue, mental symptoms, emotional sensitivity and shortness of breath at week 2 showed the strongest correlations with daily activities impact at week 12.

Patient factors and symptoms related to time to return to usual health
The Cox regression with time to return to usual health showed a significant HR for obesity (HR: 0.5, 95%CI: 0.3-0.8), sore throat (HR: 0.7, 95%CI: 0.5-0.97) and gastrointestinal manifestations (HR: 0.7, 95%CI: 0.5-0.9). The HR for obesity can be interpreted as doubling the risk of post-COVID when obese ( Figure 5). No other patient characteristics were significantly associated with longer return to usual health.

General practice visits related to COVID-19
Over half of all included patients (53%) had at least one contact with a GP or nurse for COVID-19-related symptoms during the 12 weeks of follow-up. Overall, 16% of patients consulted their general practice three or more times. The median value of contacts was 1 (min: 0-max: 10). Of those patients still symptomatic at week 12 (n ¼ 79), 20% consulted their general practice three or more times.  Univariate correlations between presence and severity of symptoms week 2 (column) and impact on daily life at week 12 (row). Darker tones correspond to stronger correlation.

Main findings
This prospective, exploratory study of adult outpatients with SARS-CoV-2 infection performed in eleven European countries shows that 52% can be considered to have long-COVID (symptomatic at week 8) and 32% post-COVID (WHO definition: symptomatic at week 12). When only considering the presence of any symptom rated as moderate or (very) severe at weeks 8 and 12, these percentages decrease to 28% and 18%, respectively. During the first 12 weeks after a positive test, fatigue was the most often reported symptom, and sports, education/ work and household activities were most often affected. Obese patients appeared to take twice as long to return to usual health. Apart from sore throat and gastrointestinal manifestations at baseline (weak association), no other relevant risk factors could be identified.

Comparison with existing literature: Return to usual health
Compared to a US outpatient cohort, which reported 35% not returned to usual health 3 weeks after their positive test, our cohort showed that 74% had not returned to usual health at week 4 [9]. Data obtained through the COVID Symptom Study app, in which individuals self-reported their symptoms prospectively, showed lower estimates of 13% with symptoms at 4 weeks, 4.5% at 8 weeks and 2.3% at 12 weeks [10]. However, inclusion characteristics, setting, number and items questioned, and the definition of 'return to usual health' differed between studies.
Comparison with existing literature: Lingering symptoms We found fatigue/tiredness the most common symptom at weeks 8 and 12, followed by shortness of breath and slowly recovering loss of smell. A study of long-COVID patients also identified these symptoms with fatigue, post-exertional malaise and cognitive dysfunction, the most frequently reported symptoms after 6 months [11]. Ongoing symptoms, particularly fatigue and shortness of breath, impacted daily activities in 15% of our cohort at 12 weeks. Even though lingering symptoms may be mild, ongoing impact and lack of a total return to usual health is relatively high. This impact can also be felt economically, for instance in work absenteeism and health care costs; three or more general practice visits for symptoms related to COVID-19 were reported by one in six patients (16%).  manifestations and sore throat as initial symptoms were weakly associated. However, one consistent finding is that obesity is a strong predictor, appearing to double the chances of post-COVID, similar to other studies [9]. Our study did not identify other (chronic) conditions but other studies have specified depression and asthma to impact symptom duration [10]. Age was not a significant predictor in our study, which may be due to the wide range of ages. Other studies may have introduced bias in recruitment of certain age groups when using digital modes of recording (COVID Symptom Study app) or due to recruitment of more severe and often older patients. Similar to our study, the COVID Symptom Study app also found the co-occurrence of anosmia (loss of smell) and ageusia (loss of taste) associated with duration of symptoms, which is most likely just due to slow recovery and not necessarily a risk factor for prolonged symptom duration [12].

Strengths and limitations
The strength of our study is the prospective data capture using a structured questionnaire, with regular long-term follow-up and minimal loss to follow-up. We included outpatients from many countries, healthcare systems, ages, and cultural backgrounds, with representative comorbidity prevalences [8]. At the time of patient inclusion, routine PCR-based testing was implemented in all participating countries. Patients were recruited after they contacted or consulted the GP via testing centres and from the community to capture a broad case-mix, independent of thresholds for consulting the GP.
However, this is a convenience sample of COVID-19 outpatients recruited via different pathways. One can imagine that patients seeking help from their GP were more severely ill at baseline and, therefore, potentially prone to longer symptom duration. Also, our study's total number of patients was relatively small (n ¼ 270). Studies with higher numbers have found more correlations in their data but still did not identify a clear risk profile with underlying conditions beyond obesity. In addition, we did not include information on the patients' physical and mental health before COVID-19.
For generalisability, our study included patients before the emergence of the omicron variant, assuming that mainly delta dominated our cohort [13]. The omicron variant is reported to be less severe and, therefore, expected to result in less long-COVID [14]. Additionally, we can assume that the vast majority of patients in our cohort were unvaccinated, as vaccination only became available at the end of the study's inclusion period. Our findings, therefore, mostly apply to unvaccinated patients. With vaccinated people less likely to report symptoms of long-COVID, some studies estimate that vaccination halves the risk of long-COVID; our estimates of long-COVID may be higher than prevalent in a vaccinated population [12,15].
Nevertheless, this cohort will address a gap in knowledge of outpatients [16]. In some countries, long-COVID clinics have been set up. Still, extensive prospective research with follow-up of a year into long-COVID and progress in identifying treatments as part of its COVID-19 therapeutics plan has not been reported [16].

Conclusion
Long-COVID and post-COVID are common in predominantly unvaccinated outpatients. One in four reports moderate to (very) severe symptoms 8 weeks after a positive SARS-CoV-2 test, one in 6 after 12 weeks. Predicting a longer return to usual health remains difficult as, apart from obesity, no apparent risk profile appeared. The strong association of obesity with a longer return to usual health may be an indication for targeted vaccination and requires further study.