Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection that causes coronavirus disease 2019 (COVID-19) has resulted in at least 764 million infections and 6.9 million deaths worldwide (1). Over 7 million cases and around three hundred and thirty thousand fatalities have occurred in Mexico to date (1, 2). Global collaborative efforts were rapidly mobilized to prevent or reduce the impact of COVID-19, including the development of effective vaccines within 1 year after the first case was recognized. However, while vaccination programs resulted in significant reductions in disease severity and mortality, reports of new or persistent symptoms after acute SARS-CoV-2 infection emerged in a subset of patients (3–7).

Several terms describe post-COVID-19 issues, including post-acute COVID syndrome, long COVID, post-acute sequelae of SARS-CoV-2 infection (PASC), persistent COVID, or post-COVID conditions (PCC), the term we will use in this manuscript (8–11). Current reports indicate that PCC occurs between 32 and 87% of patients with COVID-19, can affect multiple organs and systems, and can last more than a year in over 25% of affected patients (12–14). Reports suggest that the incidence and duration of PCC vary according to acute COVID-19 severity, vaccination status at the time of infection, and underlying comorbidities (15, 16). Additional studies have found delayed hospitalization and female sex to be associated with PCC (17, 18).

The World Health Organization (WHO) defines PCC as an illness in people who have a history of probable or confirmed SARS-CoV-2 infection, occurring within 90 days from the onset of COVID-19, with symptoms and effects that last for at least 2 months, and where the symptoms cannot be explained by an alternative diagnosis (9). Critical questions about the duration, severity, and co-occurrence of specific PCC symptoms (9, 10). In addition, studies of PCC have largely focused on the US, European, and Chinese populations, while prospective studies in patients from Latin America are lacking (8, 19).

Early during the pandemic (June 2020), we initiated the Longitudinal Instituto de la Nutrición-Stanford COVID-19 Collaborative Study (LINS) to establish an exploratory longitudinal cohort of Mexican patients admitted due to COVID-19. In this exploratory study, we aimed to provide detailed clinical phenotyping, including the onset, duration, and co-occurrence of persistent symptoms to better understand PCC in an under-studied population. In addition, we analysed a set of clinical covariates to determine their association with PCC.

Two-hundred and sixty-one patients were recruited in the study between July 2nd, 2020 and January 23rd, 2022. The median time between symptom onset and recruitment into the study was 11 days (interquartile range 9–13). For the examination of PCC, we excluded 45 patients who died during the index hospitalization, 14 patients who were followed for less than 90 days, and 10 patients who did not report persistent symptoms more than 90 days after acute COVID-19. Using the WHO statement as a guideline for PCC, the remaining 192 patients were classified as PCC patients and included in subsequent PCC analysis (Figure 1A) (9). The baseline demographic and clinical characteristics of included patients are described in Table 1. Characteristics of excluded patients are described in Supplementary Table S1. Patients with PCC were recruited during the predominance of either the B.1.1.519 variant (147, 76.6%) or the B.1.617.2 (Delta) variant (45, 23.4%, Figure 1B) (26). Among patients with PCC, the median age was 53 years; 67 (65.6%) were male, and 12 (6.2%) required mechanical ventilation. Only 35 (18.2%) patients had received a dose of SARS-CoV-2 vaccine 14 or more days before the onset of COVID-19 symptoms. The median follow-up was 405 days, with a range of 91 to 626. A total of 167 (87%) PCC patients had more than 180 days of follow-up, 105 (54.7%) were followed for more than 360 days, and 23 (12%) patients had follow-up for more than 540 days (Figure 1B). Six categories of symptoms (respiratory, mucocutaneous, neurological, functional impairment, gastrointestinal (GI), mood, sleep, and cognitive disorders (MSCD)), in total comprising 23 symptoms, were assessed at study inclusion and each follow-up phone call for each patient. Low oxygen saturation was the most common symptom among PCC patients at 90 days (112, 58.3%) (Figure 2A) and during acute COVID-19 (Supplementary Figure S1A). Among the 192 PCC patients examined, 108 (56.2%) had more than five symptoms at 90 days post-symptom onset (Figure 2B), vs. 166 (86.4%) during acute COVID-19 (Supplementary Figure S1B). The most commonly co-reported symptoms at 90 days post-symptom onset were low oxygen saturation and difficulty using the stairs (N = 73 patients), difficulties walking and using the stairs (N = 64), and anxiety and low oxygen saturation (N = 59). Hierarchical clustering of symptom co-occurrence revealed that MSCD, functional impairment, and respiratory symptoms tended to occur together. GI and some neurological symptoms (smell, hearing, and taste disturbances) generally did not co-occur with other symptoms (at most 24 patients reporting these symptoms also reported any other co-occurring symptom, Figure 2C). Similar analyses of symptom co-occurrence during acute COVID-19 showed that respiratory and functional impairment symptoms were most reported together (Supplementary Figure S1C).

Out of 192 PCC patients, a total of 160 (83.3%) reported at least one symptom after 180 days, 93 (48.4%) after 360 days, and 22 (11.5%) after 540 days (Figure 3A). The results of KM analyses to examine the duration of PCC, symptom categories, and individual symptoms are shown in Figure 3. Persistence probabilities were interpolated from these KM curves to determine the probability of PCC, a symptom category (Supplementary Table S2), or an individual symptom persisting at 180-, 360-, and 540 days (Supplementary Table S3). The probability of PCC persisting for up to 360 days was 0.78 (Figure 3A, Supplementary Table S2). We found that MSCD symptoms were the most persistent at 180 days post-symptom onset (Figure 3B, Supplementary Table S2), of which anxiety, psychosocial difficulties, and depression were the most persistent (Figure 3C, Supplementary Table S3). MSCD symptoms were also the most persistent by 360 days (persistence probability 0.62), followed by neurological, functional impairment, respiratory, mucocutaneous, and GI symptom categories (Figure 3B, Supplementary Table S2). Anxiety, difficulty hearing, difficulty exercising, congestion, hair loss, and nausea were the most persistent in each symptom category 180 days post-symptom onset (Figures 3C–H, Supplementary Table S3). Mucocutaneous and GI symptom categories displayed the lowest persistence probability by 540 days post-symptom onset (0.08 and 0.26 respectively, Supplementary Table S2). In only 49 PCC patients (25.5%) all symptoms had resolved by the last phone call. Among those with more than 360 days of follow-up, 29 (27.6%) had resolved all symptoms. Symptom durations for each person are shown in Supplementary Figure S2.

The effects of baseline demographic and clinical characteristics (Table 1) on symptom duration were calculated with Cox regression models and reported with HRs (14, 28). Results for PCC and symptom categories were summarized using a heatmap (Figure 4A) that depicts the HR of the group of patients presenting each covariate, where a hazard ratio less than 1 indicates a longer time to resolution if that covariate is present. Significant associations (value of p <0.05) with symptom duration were extracted and subgroup KM with log-rank analyses were performed (Figures 4B–G). Among PCC patients, prior SARS-CoV-2 vaccination and acute COVID-19 symptom onset occurring during a period of Delta variant predominance were associated with a shorter time to PCC resolution (Figures 4B,C). Male sex was associated with a shorter time to resolution of functional impairment and respiratory symptoms (Figures 4D,E), and a clinical history of either hypertension or diabetes (a known risk factor for PCC), was associated with a longer time to resolution of functional impairment (Figures 4F,G) (7, 15). Similarly, multivariable analysis of associations of clinical covariates with symptom duration revealed that male sex was associated with a shorter duration of functional impairment and respiratory symptoms, and hypertension was associated with a longer duration of functional impairment (Supplementary Figure S3). The number of patients with each demographic and clinical covariate is shown for each symptom category (Supplementary Figure S4). Evaluation of the proportional hazard assumption of the Cox regressions is shown in Supplementary Figure S5 (28). HRs for the resolution of PCC or of each symptom category concerning a given covariate are shown in Supplementary Figure S6. Results for the effect of covariates on individual symptoms are shown as a heatmap (Supplementary Figure S7).

PCC is emerging as a considerable cause that may disproportionately burden the healthcare system of Latin American countries (9, 11). We performed detailed and longitudinal phenotyping of multiple PCC symptom dimensions in a cohort of Mexican patients with PCC followed clinically for a period of up to 2 years after hospitalization for COVID-19. Several major themes emerged from the analysis of individual PCC symptom prevalence, co-occurrence, and duration. (1) While most patients with PCC reported several persistent symptoms after acute COVID-19, the duration of individual PCC symptom categories varied substantially from patient to patient, (2) respiratory, neurological, MSCD, and functional impairments contributed the most to the persistence of PCC, and (3) prior SARS-CoV-2 vaccination and infection during the Delta variant wave were associated with shorter duration of PCC symptoms.

In our cohort of patients hospitalized with COVID-19, PCC was surprisingly common, with 95% of enrolled surviving patients affected. Given we only evaluated patients hospitalized with severe COVID-19, the high incidence of PCC observed is likely to represent an overestimation of the overall incidence of PCC in people who had COVID-19. For example, a study performed in Mexico using data from the national health survey (ENSANUT) found a prevalence of 4.5% of symptoms beyond 3 months (33). However, the cross-sectional studies are limited by the potential for recall bias, which is diminished in our study given data was prospectively collected (33, 34). However, a high frequency of PCC incidence was reported in a landmark study by Carfi et al. where almost 90% of patients from an Italian post-COVID clinic had persistent symptoms 60 days after acute symptom onset (5). Among studies with long-term follow-up ≥6 months, a high proportion of persistent symptoms is uniformly reported (13, 18). The PHOSP-COVID study of discharged COVID-19 patients in the United Kingdom and an additional study by Seeßle et al. found a low proportion of recovery at 1 year (28.9 and 22.9%, respectively), while a study by Huang et al. reported 51% recovery at 1 year (13, 17, 18). These studies found fatigue as the most common symptom at 1 year, which contrasts with the high burden of low-oxygen saturation, anxiety, and difficulty using the stairs in our cohort (13, 14, 17, 18). A study derived from chart data from more than 1.2 million COVID-19 survivors showed differential recovery rates for specific PCC symptoms, where cognitive disorders, psychotic disorders, and epilepsy or seizures persisted longer than mood and anxiety disorders (35). Thus, the long-term symptoms appear to vary according to the studied population. Unlike other studies in non-Hispanic populations, we did not observe a correlation between age and an increased risk of developing PCC (7). Interestingly, a study by Jia et al. which had a high proportion of Hispanic patients (around 40%), also did not find a higher risk of PCC according to age (14). Our study, however, had relatively few people of younger age, which is concordant with the admission of sicker individuals who were usually older (36). Thus, a larger number of people of lower age may be needed to detect a correlation between age and PCC in our population.

To date, there are no targeted treatments for PCC, and interventions remain symptom-based. Our results indicate that in the cohort examined, having received prior SARS-CoV-2 vaccination may accelerate the resolution of PCC, as observed in other studies (37–39). Taken together, these results suggest vaccination may not only protect against severe acute disease but also reduce the length of PCC.

Currently, the precise causes of PCC are unknown (40). Recent studies have begun to investigate the role of genetic diversity as a possible causal factor (41). This emphasizes the importance of studies that examine PCC in different ethnic populations.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Comité de Investigación en Humanos, INCMNSZ. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

BG and SV-F: conceptualization. IN, JG, SF-S, DF, JM, JK, HC, EG, CB, KN, GN, JC, DM, BG, and SV-F: methodology. SF-S, DF, JM, JK, HC, EG, and AB: investigation. LI-W, AG-G, UV-G, JF-G, JH-G, and GR-G: data collection. IN, JG, and LI-W: data curation. IN and JG: formal analysis and writing – original draft. LI-W, CB, KN, DM, BG, and SV-F: project administration. GN, BG, and SV-F: funding acquisition. IN, JG, SF-S, DF, LI-W, AG-G, UV-G, JM, JK, HC, EG, AB, JF-G, PD-B, JH-G, GR-G, CB, KN, GN, JC, DM, BG, and SV-F: writing – review and editing. IN, JG, and LI-W directly accessed and verified the underlying data reported in the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the Bill and Melinda Gates Foundation COVID-19 (Pilot Award OPP1113682 to GN, DM, HC, and INV-002704 to GN); Consejo Nacional de Ciencia y Tecnología, Gobierno de México (grant A1-S-18342 to SV-F); Fundación Gonzalo Río Arronte (Grant SL00206 to SV-F); The Stanford Centre for Human Systems Immunology (to BG); CEND COVID Catalyst Award (to BG); the Doris Duke Charitable Foundation (to BG); German Academic Exchange Service with funds from the German Federal Ministry of Education and Research (to AB); the Centers for Disease Prevention and Control (contract number 75D30120C08009 to JK); The Sunshine Foundation (to KN), David A. Crown Foundation (to KN), and Sean N. Parker Foundation (to KN). National Institutes of Health (5U19AI100627-09 Systems Approach to Immunity and Inflammation COVID supplement to GN, DM, and HC); US Food and Drug Administration Medical Countermeasures Initiative (contracts 75F40120C00176 and HHSF223201610018C to GN, DM, and HC). This article reflects the views of the authors and should not be construed as representing the views or policies of the US FDA, CDC, NIH, or other funding sources listed here.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.