The population in the present study mostly consisted of women, which differs from the results observed in São Paulo (Brazil) and England where males were more frequent among those diagnosed with COVID-19 [15, 16]. A higher COVID-19 mortality rate in males has also been described in 37 of the 38 countries that provide data by sex . However, the effect of sex on SARS-CoV-2 infection still requires further investigation , and genetic, immunological, hormonal, sociobehavioral, economic, and lifestyle factors must be considered to identify such differences [19, 20].
More than 95% of the study population consisted of symptomatic individuals, most likely because individuals with symptoms suggestive of COVID-19 more frequently sought diagnostic laboratory services.
The humoral immune response has been described and suggested as fundamental for protection against SARS-CoV-2 [21, 22]. In addition, the kinetics of the emergence of anti-SARS-CoV-2 antibodies revealed the absence of seroconversion or even a loss of antibodies after infection [23, 24]. Patel et al.  reported the 42% and 58% of health professionals from Tennessee (USA) exhibited a persistent IgG response and a nonpersistent IgG response, respectively, 60 days after diagnosis. Robbiani et al.  concluded in their study with convalescent COVID-19 patients that many did not have high titers of neutralizing antibodies 39 days after symptom onset. Demonbreun et al.  observed loss of anti-SARS-CoV-2 antibody seropositivity after 120 days in 25% of a cohort in Chicago (USA).
The impact of this high percentage of individuals who lost detectable levels of circulating antibodies on the occurrence of SARS-CoV-2 reinfection remains unknown, as has been reported in some countries . Considering that the response time of anti-SARS-CoV-2 antibodies seems to be short in a large portion of the population and that cellular immunity mediated by T lymphocytes is important and lasting , Altmann & Boyton  suggest that conducting population screening tests for anti-SARS-CoV-2 antibodies and T cells would be very useful.
Most infected individuals generate antibody and T cell responses with magnitudes correlated with the time of infection and disease severity . In this study, more than 95% of the individuals evaluated were symptomatic, with greater IgG persistence among those with symptoms lasting ≥ 21 days and lower persistence among those with symptoms lasting ≤ 7 days. The length of hospital stay, supplemental oxygen use, the number of reported symptoms, and moderate and severe pulmonary involvement were also more frequent in individuals with a persistent IgG response, indicating a close relationship between COVID-19 severity and the magnitude of the immune response, as previously suggested [29, 31]. Length of hospital stay and the quality and quantity of the antibody response may be associated with clinical manifestations and disease course, thus justifying further investigations .
The proinflammatory response plays an important role in the pathogenesis of COVID-19, especially in the development of more severe forms of the disease. At a more advanced stage of infection, cells secrete high levels of IL-1β, IL-6, and tumor necrosis factor (TNF) . In the present study, we evaluated whether individuals with a persistent IgG response had a sustained inflammatory response mediated by IL-1β and IL-6 90 days after diagnosis. The results showed that after 90 days, the individuals presented basal levels of cytokines, including those (n = 44) who reported experiencing post-COVID-19 symptoms (shortness of breath, anosmia, fatigue on exertion, hair loss, fatigue, dizziness, back pain, and muscle and joint pain). These findings demonstrate that although inflammatory cytokines are present at high levels in patients with COVID-19 admitted to hospitals , cytokine levels in individuals recovering from the disease do not persist 90 days after infection, even in patients who maintain symptoms related to COVID-19.
The percentage of individuals aged ≥ 60 years was higher in the group with a persistent IgG response. Klein et al. (2020) showed that older age was associated with increased antibody responses to COVID-19, which emerged as a factor that can be used to identify individuals with a high probability of presenting strong antiviral antibody responses  in contrast to other studies pointing to natural impairment of the immune response in the elderly population due to senescence [36, 37]. However, in addition to senescence, aging is accompanied by changes in the immune profile characterized by chronic subclinical systemic inflammation (inflamm-aging) that may contribute to the disproportionate SARS-CoV-2 mortality rate among elderly patients .