Determinants of the onset and prognosis of the post-COVID-19 condition: a 2-year prospective observational cohort study

Abstract

Background: At least 5-10% of subjects surviving COVID-19 develop the post-COVID-19 condition (PCC) or "Long COVID". The clinical presentation of PCC is heterogeneous, its pathogenesis is being deciphered, and objective, validated biomarkers are lacking. It is unknown if PCC is a single entity or a heterogeneous syndrome with overlapping pathophysiological basis. The large US RECOVER study identified four clusters of subjects with PCC according to their presenting symptoms. However, the long-term clinical implications of PCC remain unknown.

Methods: We conducted a 2-year prospective cohort study of subjects surviving COVID-19, including individuals fulfilling the WHO PCC definition and subjects with full clinical recovery. We systematically collected post-COVID-19 symptoms using prespecified questionnaires and performed additional diagnostic imaging tests when needed. Factors associated with PCC were identified and modelled using logistic regression. Unsupervised clustering analysis was used to group subjects with PCC according to their presenting symptoms. Factors associated with PCC recovery were modelled using a direct acyclic graph approach.

Findings: The study included 548 individuals, 341 with PCC, followed for a median of 23 months (IQR 16.5-23.5), and 207 subjects fully recovered. In the model with the best fit, subjects who were male and had tertiary studies were less likely to develop PCC, whereas a history of headache, or presence of tachycardia, fatigue, neurocognitive and neurosensitive complaints and dyspnea at COVID-19 diagnosis predicted the development of PCC. The cluster analysis revealed the presence of three symptom clusters with an additive number of symptoms. Only 26 subjects (7.6%) recovered from PCC during follow-up; almost all of them (n = 24) belonged to the less symptomatic cluster A, dominated mainly by fatigue. Recovery from PCC was more likely in subjects who were male, required ICU admission, or had cardiovascular comorbidities, hyporexia and/or smell/taste alterations during acute COVID-19. Subjects presenting with muscle pain, impaired attention, dyspnea, or tachycardia, conversely, were less likely to recover from PCC.

Interpretation: Preexisting medical and socioeconomic factors, as well as acute COVID-19 symptoms, are associated with the development of and recovery from the PCC. Recovery is extremely rare during the first 2 years, posing a major challenge to healthcare systems.

Funding: Fundació Lluita contra les Infeccions.

Keywords: COVID-19; Long COVID-19; Post-COVID-19 condition.

Fig. 1

Factors associated with the post-COVID-19 condition. (a) Pre-existing conditions (odds ratio with 95% confidence interval; univariate logistic regression model). (b) Symptoms at COVID-19 diagnosis (odds ratio with 95% confidence interval; univariate logistic regression model).

Fig. 2

Post-COVID-19 condition clusters. (a) Silhouette plot suggesting a cut-off point on 3 clusters. (b) Dendogram of hierarchical cluster analysis on a set of dissimilarities using the Daisy method for mixed types of variables and the Gower metric. (c) 2-dimensional clustering plot, where each subject is represented by the principal components, and an ellipse is drawn around each cluster. (d) Heat map of subjects according to cluster membership and symptom prevalence.

Fig. 3

Model of recovery from the post-COVID-19 condition. Directed acyclic graph of causal relationships between factors present during the acute COVID-19 episode and their influence on the chance of recovery from the PCC. Green arrows indicate positive associations; red arrows show negative associations. The statistical model metrics are shown below the graph. ICU indicates intensive care unit. Symp p (protective) indicates symptoms developed during acute COVID-19 which are associated with recovery from PCC, i.e., hyporexia and smell and taste alterations. Symp r (risk) indicates symptoms developed during acute COVID-19 which are associated with lack of recovery from PCC, i.e., myalgia, dyspnea, tachycardia, neurocognitive disorder. Educational level indicates tertiary education or higher. Vascular comorbidities include hypertension, diabetes and dyslipidemia.

Fig. 4

Post-COVID-19 condition symptoms and SARS-CoV-2 vaccination. Stacked bar plot of overall symptom status reported by subjects with PCC one month after each SARS-CoV-2 vaccine dose administration, relative to their symptom status during the month prior to receiving the corresponding vaccine dose. Subjects’ reports belonged to 6 qualitative categories: (i) Equal: no changes in symptoms (green); (ii) Better, sustained: sustained improvement during the month following the vaccine dose (light orange); (iii) Better, transient: transient improvement following the vaccine dose with return to the pre-dose status by the end of the month (yellow); (iv) Worse, sustained: sustained worsening during the month following the vaccine dose (red); (v) Worse, transient: transient improvement following the vaccine dose with return to the pre-dose status by the end of the month (dark orange). Subjects with remission of all PCC symptoms (Cured) are shown in blue, unknown values in dark grey and no vaccine doses in light grey.