Complications Post-COVID-19 and Risk Factors among Patients after Six Months of a SARS-CoV-2 Infection: A Population-Based Prospective Cohort Study

In October 2020, we conducted a population-based prospective cohort study to determine post-COVID-19 complications, recovery, return to usual health, and associated risk factors in 536 cases of COVID-19 outbreak in Borriana (Spain) by administering an epidemiological questionnaire via phone interviews. A total of 484 patients participated (90.3%), age mean 37.2 ± 17.1 years, and 301 females (62.2%). Mild illness was the most common COVID-19 manifestation. After six months, 160 patients (33.1%) suffered at least one complication post-COVID-19, and 47 (29.4%) of them sought medical assistance. The most frequent persistent symptoms were hair loss, fatigue, loss of smell or taste, and headache. Risk factors associated with a complication were female sex (adjusted relative risk, [aRR] = 1.93 95% confidence interval [CI] 1.41–2.65), age 35 years and above (aRR = 1.50 95% CI 1.14–1.99), B blood group (aRR = 1.51 95% CI 1.04–2.16), current smoker (RR = 1.61 95% CI 1.02–2.54), and at least a COVID-19 exposure (aRR = 2.13 95% CI 1.11–4.09). Male sex, age younger than 35 years, and low COVID-19 exposures were associated with better recovery and return to usual health. A third of patients presented persistent symptoms compatible with the long-COVID-19 syndrome. In conclusion, an active medical follow-up of post-COVID-19 patients must be implemented.


Introduction
After more than year of the COVID-19 pandemic, many aspects of this new and complicated disease are still poorly understood and characterized [1,2], such as the frequency

Explanation
In March 2020 a COVID-19 outbreak took place during the mass gathering events (MGEs) of the Falles festival in Borriana, a city of 34,000 inhabitants, in the Castellon province, Valencia Community (Spain), with 536 laboratory-confirmed cases of COVID-19 [32]. In October 2020 a population-based prospective cohort study of 536 COVID-19 patients was performed by the Public Health Center of Castelló de la Plana, and the Emergency and Microbiology and Clinical Analysis Services of Hospital de la Plana Vila-real [33].
We first performed a sero-survey and an epidemiological questionnaire survey in June 2020 that has been described by Domenech-Montoliu and co-authors [32]. A second sero-survey and questionnaire were run in October 2020. In summary, the sero-surveys included determinations of anti-SARS-CoV-2 N-antibodies in June and October 2020 with a flow chart of the study, and ABO blood groups and vitamin D status in October 2020. The sero-survey results were subject of two previous publications [33,34], and are no further discussed in the current manuscript. Health staff of Hospital La Plana Vila-real, and several health centers (Borriana, Vila-real, Onda, and Vall D'Uixo) performed the telephone interviews. In June 2020, we collected information about acute COVID-19 illness (symptoms, medical consultation, health status), demographic data (age, sex), lifestyle (smoking habit, alcohol consumption, physical exercise), body mass index (BMI = kg/m 2 , obesity ≥ 30.0), occupation as a proxy social class, COVID-19 exposure (measured as positive observation a person coughing at the MGEs, contact with a COVID-19 patient or a relative with COVID-19, attendance two or more MGEs). In the second (October 2020) telephone survey, we collected information on the evolution of the COVID-19 disease, complications, persistent symptoms, health recovery, return to the usual state of health, and medical consultation after the acute phase.

Statistical Methods
We performed the statistical analysis and considered complications, recovery, and return to the usual state of health as dependent variables; age, sex, ABO blood groups, oc-cupation, chronic disease, lifestyle, BMI, COVID-19 exposure during the MGEs, symptoms, and medical consultation of acute COVID-19 illness were independent variables.
In the univariate analysis, Chi2 and Fisher exact tests were used to compare qualitative variables and the Kruskal-Wallis test for quantitative variables. The incidence rate (IR) of a variable was obtained dividing the positive cases of this variable by the exposed population, and the relative risk (RR) of each independent variable, dividing the IR of the population exposed by the IR of the population no-exposed of this variable, with a 95% confidence interval (CI). After a revision of medical literature, we employed directed acyclic graphs (DAGs) [35], and the program DAGitty 3.0 version (Johannes Textor, Nijmegen, The Netherlands) [36,37] to define the potential confounders of each independent variable. We used inverse probability weighted regression [38] in the multivariable analysis to estimate adjusted incidence rates (aIR) and adjusted relative risks (aRR). We applied Stata ® 14.2 version (Stata Corp, College Station, TX, USA) or all calculations.
The study was part of public health surveillance activities of the COVID-19 outbreak control measures in the MGEs of the Falles festival [32], and it was exempt from the Ethics Review Board approval's protocol according to the Spanish legislation and regulations [39][40][41][42]. The study was approved by the director of the Public Health Center of Castelló de la Plana and the management of the Health Department of la Plana. All participants or the parents of minors provided the informed written consent to be included in the study.

Results
A total of 484 patients participated in the study with a response rate of 90.3% from the 536 COVID-19 laboratory-confirmed cases in MEGs. The mean ± standard deviation (SD) age of participants was 37.2 ± 17.1 (range 1-81) years, and 301 were females (62.2%). In the outbreak, a death attributable to COVID-19 was reported, 13 patients required hospitalization, and mild illness was the most frequent clinical presentation.
Patients' characteristics and potential risk factors of complications of the COVID-19 illness, health recovery, and return to the usual health before the acute illness in the six months of follow-up are shown in Table 1. A total of 160 participants reported at least one post-COVID-19 complication (33.1%), 47 (29.4%) of them sought medical care, but no hospitalizations took place. An 81.8% of the participants recovered, and 83.2% returned to their usual state of health before the COVID-19 episode. Complications were more frequent in females, elder age, B blood group, current smoking status, high COVID-19 exposure, and those who reported symptomatic acute COVID-19 episode, and sought medical care. Health recovery and return to the usual health were highest in males, young individuals, attendance of MEGs less than two events, asymptomatic acute illness, and fewer medical consultations. We show persistent post-COVID-19 symptoms by sex and four age groups in Table 2. The most frequent signs or symptoms were hair loss (24.8%), fatigue (17.4%) loss of smell or taste (16.1%), headache (15.1%) muscle pain (11.8%), insomnia (11.8%), anxiety (9.1%), weakness (8.5%), restlessness (8.1%), hands and foot's pain (7.4), and dyspnea (6.8%). Females reported suffering more symptoms than males, especially hair loss (35.9% versus 6.5% p = 0.000), fatigue (21.9% versus 9.8% p = 0.000), and loss of smell or taste (21.6% versus 7.1% p = 0.000). The mean (±SD) duration of persistent post-COVID-19 symptoms was 160.9 ± 45.5 days (range of 3-280 days). The mean (±SD) duration in females was 162.6 ± 44.5 (range 3-280) days, and in males 155.8 ± 48.8 (range 15-210) days without significant difference (p = 0.444).
Patients 65 years old and older reported more symptoms, with the exception of loss of smell or taste and the hair loss, which were more common in subjects in the range of 15-34 years and 35-64 years. In subjects 65 years and older, the most frequently reported symptoms were fatigue (41.2%), hand/foot pain (29.4%), and insomnia (29.4%).

Symptoms
Return to the Usual Health Total (%) aIR 1  In the adjusted multivariable analysis (   Considering health recovery after the acute illness (Table 7), younger age than 35 years (aRR = 1.15 95% CI 1.06-1.25), and male sex (aRR = 1.09 95% CI 1.00-1.19) were associated with more recovery ( Table 7). The AB blood group subjects reported 100% recovery. Attendance with less than two MGEs was associated with more recovery (aRR = 1.11 95% CI 1.02-1.21).  The return to the usual health was associated with male sex (aRR = 1.11 95% CI 1.02-1.20), and age 0-34 years (aRR = 1.14 95% CI 1.05-1.24), no contact with a COVID-19 patient (aRR = 1.15 95% CI 1.06-1.25), and AB blood group (aRR = 1.15 95% CI 1.09-1.21) ( Table 8).   3 Adjusted for sex ABO blood group; 4 Adjusted for age sex; 5 Adjusted for age sex ABO blood goup; 6 Adjusted for age sex ABO blood group occupation lifestyle obesity; 7 Adjusted for age sex ABO blood group occupation lifestyle obesity; 8 Adjusted for age sex ABO blood group occupation lifestyle obesity; 9 Adjusted for age sex ABO blood group occupation lifestyle obesity information 14; 10 Adjusted age sex ABO blood group occupation lifestyle; 11 MGEs, mass gathering events; 12 Adjusted for age sex ABO blood group chronic disease occupation obesity other COVID-19 exposures; 13 Adjusted for age sex ABO blood group chronic disease occupation obesity other COVID-19 exposures; 14 Adjusted for age sex ABO blood group chronic disease occupation obesity other COVID-19 exposures; 15 Adjusted for age sex ABO blood group chronic disease occupation obesity other COVID-19 exposures; 16 Adjusted for age sex ABO blood group chronic disease occupation obesity.

Discussion
Our results suggest that a third of COVID-19 patients with mild illness experienced post-COVID-19 symptoms six months after the acute episode. The most frequent symptoms were loss of hair, fatigue, and loss of smell or taste. Acute symptoms of illness, female sex and elder age were associated with complications, and a less frequent recovery and return to usual health. As highlighted findings of our study, several exposures to COVID-19, including contact with a COVID-19 patient, attendance to two or more MGEs, and observing a person with a cough at MEGs, were risk factors of complications. On the other hand, lower COVID-19 exposures were associated with better recovery and return to the usual health. These results are in line with other studies that found COVID-19 exposures were associated with post-traumatic stress symptoms of COVID-19 patients, compatible with LCS [43][44][45].
Complications of COVID-19 have been observed in many prospective cohort studies indicating the importance of a follow-up of COVID-19 patients [19,[46][47][48]. In general, the frequency and severity of these complications or "sequelae" are associated with acute COVID-19 illness [4,31,49], and our results are consistent with these studies, these being hair loss, fatigue, and loss of smell or taste such as more frequent reported persistent symptoms in a multi-system affectation [21,46,[50][51][52][53]. Considering that our cohort had few hospitalized patients, the incidence of complications was lower in comparison with other cohort studies [49,51,52]. Recovery and return to the usual health were higher than those found by studies in hospitalized and non-hospitalized patients, and medical consultation was lower [54,55]. Soraas and co-authors 2021 [56] analyzed a cohort of non-hospitalized COVID-19 patients, and found 36% reported worse health status one year after illness. Persistent symptoms in cohorts of non-hospitalized patients presented large variations in this percentage, ranging from 20% to 61%, possible due to differences in the duration of follow-up, age and sex distributions, and sample sizes [57,58].
Several symptoms during acute illness were predictive of the post-COVID-19 persistence of symptoms, including weakness, skin lesions, loss of smell or taste, headache, and more of five acute symptoms, in line with cohort studies of non-hospitalized patients [31,51,56,59], but not others [60]. In a study in Norway, fatigue was present in 46% of non-hospitalized COVID-19 patients after a mean of four months; persistence of fatigue was associated with female sex, high numbers of symptoms in the acute illness, and confusion [61]. In addition, dyspnea increased the possibility of post-COVID-19 persistent symptoms and poor health recovery.
We found that females and elder patients experienced a higher frequency of post-COVID-19 symptoms compared with males and young individuals, as has been found in several studies [19,22,31,49,51,61]. The highest incidence of symptoms in females is not well understood and could be associated with the immune viral response [53,56]. Older age can worsen the outcome considering the inflammatory and immunodeficiency state of aging [61][62][63]. This poor recovery could also be associated with a more frequent severe course of the disease, probably more COVID-19 exposures, and a post-viral disease syndrome which could increase with age [64].
Another associated factor with persistent symptoms was the B blood group compared to the O blood group. A more severe course of COVID-19 disease in the B-group has been described including thrombotic complications [65][66][67]. The causes of the B-group having more complications compared to the O-group are unknown. Some studies have found that the O-group presents more resistance to some infectious diseases [68], higher physiologic capacity [69], low levels of SARS-CoV-2 IgG compared to non-O groups [70], and some protection from the anti-A/anti-B presence and furin cleavage [71]. In our study [34] the persistence of SARS-CoV-2 antibodies was lower in the O-group, and it has been found that B-group presented higher SARS-CoV-2 neutralizing antibody titers than the other ABO [72].
Current smoking status was associated with more complications, but it was not associated with heath recovery and return to the usual health. In some studies, current smoking is a risk factor of illness severity, but there are different results with respect to complications [56,59,[73][74][75][76]. However, the incidence of COVID-19 in current smokers is lower than in non-smokers [77].
We have not observed an association between BMI, obesity, regular exercise, or alcohol consumption and post-COVID-19 symptoms, in contrast with other studies [31,53]. In addition, high levels of anti-SARS-CoV-2 antibodies or their increase between June 2020 and October 2020 were not associated with complications, recovery, and return to the usual health [33], contrary to other studies [25,28,57]. However, attendance of MGEs, observing a person with a cough at MGEs, and contact with a COVID-19 patient were associated with complications, inferior recovery, and less frequent return to usual health. All of the above highlights the importance of intensity of COVID-19 exposure, duration, viral load, and the place where the exposure occurred [78][79][80][81]. In experimental models, the differences between the persistence of virus infection and the quantity of inoculum have been found [82].
Our study has some strengths such as its prospective design, a population-based approach, a high participation rate, a method to explore potential confounding variables [37], and the use of multivariate analysis in order to estimate adjusted risk factors for different variables. However, these results could be more accurate. In addition, we studied not only complications, but also health recovery and return to a usual state of health.
The study limitations include the use of a questionnaire to ascertain reported complications and symptoms of post-COVID-19 disease, with no medical examination of patients. As a consequence, we cannot exclude information bias even if the questionnaires were administrated by physicians and nurses. Most of the patients suffered a mild COVID-19 illness, and the participant population is not representative of severe COVID-19 disease. We cannot discard a residual confounding despite the analysis. COVID-19 is a new disease and some unknown variables not-collected in our study could have an impact on our observations. The causes of LCS are unknown, but some aspects could be considered, including viral persistence, post-infectious myalgic encephalomyelitis/chronic fatigue syndrome, metabolism alteration, disproportionate autoimmunity, pathological inflammation, disruption of the autonomic nervous system, post-traumatic stress, underlying chronic disease, damage to the lungs, brain, heart, kidney, and other organs [5,17,18,26,[83][84][85][86].
Considering the high incidence of persistence of COVID-19 related symptomatology found after COVID-19 infections, some points should be addressed in order to improve the prevention and management of the LCS, such as the medical follow-up of LCS patients to establish their natural history, determine laboratory tests needed, validate functional scales, accrue more information of potential risk factors, and apply appropriate therapies [22,59,84,87,88]. Our prospective design and the population-based approach have advantages in obtaining results with less bias than other designs.

Conclusions
Despite a majority of mild illness, a third of patients presented persistent symptoms compatible with the LCS, and several risk factors were found. An active medical follow-up of post-COVID-19 patients must be implemented. Funding: This research received no external funding.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki. The study was part of the public health surveillance as a prolongation of the COVID-19 outbreak in the Falles festival in Borriana control measures, which was exempted from Ethics Review Board approval's protocol according to the Spanish legislation and regulations, including (33) the General Law of Health, (34) the Law of Cohesion and Quality of the National System of Health, and (35) the Law General of Public Health. The study was approved by the director of the Public Health Center of Castellon and the management of the Health Department of La Plana. In addition, (36) the cohort was following to respond to a new disease, the COVID-19 pandemic.
Informed Consent Statement: All participants or the parents of minors provided the informed written consent to be included in the study.
Data Availability Statement: Data of the study can be consulted if the authors are requested. Dataset: borrianacohort.dta.