Abstract
Background
Increased risk of post-acute sequelae was found to occur in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2 reinfections). However, it is unclear whether these increases in post-acute risk are found in milder Omicron reinfections, or persist in a highly boosted population.
Methods
We utilised national COVID-19 databases and healthcare-claims records of Singapore to construct SARS-CoV-2 infected and uninfected cohorts over periods of Delta, Omicron BA.1/2, BA.4/5 and XBB predominance (1 July 2021–28 February 2023). The 300-day risk and excess burdens of pre-specified new-incident diagnoses across cardiovascular, neuropsychiatric, endocrine, auto-immune, renal, respiratory and gastrointestinal domains was compared across SARS-CoV-2 re-infected individuals (N = 57,222), SARS-CoV-2-infected individuals without documented re-infection (N = 1,239,119), and population-based un-infected controls (N = 3,409,170). Risk trajectories between groups were compared to examine whether differences in risk of post-acute sequelae persisted over follow-up time.
Results
There was an estimated 21% (hazards ratios (HR) = 1.21; 95% Confidence interval (CI) [1.15–1.28]) increase in risk of any post-acute sequelae, and increased post-acute risk of cardiovascular (HR = 1.20; 95% CI [1.09–1.32]), gastrointestinal (HR 1.26; 95% CI [1.16–1.38]), neurological (HR 1.32; 95% CI [1.23–1.42]), endocrine (HR 1.26; 95% CI [1.18–1.35]), respiratory (HR 1.63; 95% CI [1.44–1.83]) and renal (HR 1.28; 95% CI [1.12–1.47]) sequelae associated with SARS-CoV-2 reinfections. Associated risks of post-acute sequelae were greater in reinfections compared to first infections. Excess burdens per 1000 of any post-acute sequelae were also higher in reinfected individuals, and reinfected individuals also had higher outcome probabilities of post-acute sequelae over follow-up time. Risks of post-acute sequelae persisted in fully vaccinated and boosted individuals.
Conclusions
Reinfection with SARS-CoV-2 is associated with increased risk of post-acute sequelae. Reducing burden of post-acute complications due to SARS-CoV-2 necessitates strategies for preventing reinfection, such as updated vaccines with better effectiveness against infection.
Supplementary Information
The online version contains supplementary material available at 10.1186/s44263-025-00222-1.
Keywords: SARS-CoV-2, COVID-19, Multi-organ sequelae, Omicron, Delta, Reinfection
Background
Post acute sequelae following initial SARS-CoV-2 infection is extensively characterised; however, risk of chronic sequelae following subsequent re-infections is not well understood. While large population-based retrospective cohort studies in the pre-Omicron era found increased risks of post-acute sequelae and persistent symptoms across multiple organ systems following re-infection with SARS-CoV-2 [1, 2], other cohort studies that evaluated self-reported post-infection symptoms found reduced risk, or no increased risk, of symptom persistence following re-infection versus first infections [3–5]. Furthermore, the majority of studies that evaluated the risk of post-acute sequelae following re-infection were conducted in the pre-Omicron era [1–3], risk for re-infection substantially increased during Omicron predominance in the context of immunologic escape [6].
Furthermore, information on the impact of booster vaccination doses in attenuating risk of post-acute sequelae following SARS-CoV-2 re-infection is lacking. While risk of post-acute sequelae was elevated regardless of vaccination status in a cohort of USA veterans infected pre-Omicron, this preceded availability of booster vaccination [1]. Booster vaccination has been shown to reduce risk of post-acute sequelae following initial SARS-CoV-2 Omicron infection [7, 8] widespread availability of booster vaccination during COVID-19 endemicity might thus reduce the risk of post-acute sequelae following reinfection.
To address these gaps, we constructed a national population-based cohort of Singaporean adults infected with SARS-CoV-2 across Delta, Omicron BA.1/2, BA.4/5 and XBB periods of predominance (1 July 2021–28 February 2023). We estimated the 300-day risk and excess burdens of pre-specified new-incident diagnoses across multiple organ systems, comparing 57,222 individuals who were reinfected with SARS-CoV-2 (two infections) against 1,239,119 individuals who were infected once and 3,409,170 population-based controls.
Methods
Study setting and databases
To study the associated differences in the risk of post-acute sequelae following SARS-CoV-2 reinfection vs. infection, we utilised the national COVID-19 database from Singapore, a multi-ethnic Southeast Asian city-state, to construct cohorts of adult Singaporeans infected with SARS-CoV-2 during Delta and Omicron predominance (1 July 2021–28 February 2023). Omicron BA.1/2 displaced Delta as the predominant strain from January 2022 onwards, with subsequent BA.2, BA.4/5 and Omicron XBB waves [9, 10], up to end-2023 when the Omicron JN.1 subvariant emerged [11]. Vaccination status was defined by the number of mRNA vaccination doses recorded in the National Immunisation Registry, at the point of SARS-CoV-2 infection. Under the national vaccination programme that started in December 2020, monovalent BNT162b2 (Pfizer) and mRNA-1273 (Moderna) were originally approved for use in a two-dose primary series, with ≥ 90% of the population receiving mRNA vaccines [10]. Booster vaccinations were introduced from September 2021 onwards [10, 11]; during this period, high vaccination and boosting rates were encouraged via vaccination-differentiated public health measures [12]. Booster uptake was high; ≥ 90% of adults received at least one booster dose [13].
Risk of pre-specified new-incident individual and composite post-acute sequelae following SARS-CoV-2 infection was assessed using the national healthcare claims database. In Singapore, healthcare is delivered via a hybrid of public and private elements [14]. The national government-administered medical-savings scheme (Medisave) can be claimed against for inpatient care and also outpatient treatment of a wide range of conditions at both public and private healthcare providers; [14] this enabled comprehensive capture of new-incident diagnoses across different healthcare settings. This database was previously utilised to identify risk of post-acute sequelae following initial SARS-CoV-2 infection in the Singaporean population [15–18].
Cohort
A flowchart of cohort construction is provided (Fig. 1). Individuals were enrolled if they (1) were tested for SARS-CoV-2 (2) were Singaporean citizens/permanent residents. Individuals vaccinated with non-mRNA vaccines were excluded as they formed only a very small proportion of the population (< 5%) [10–12]. Individuals who died or had missing sociodemographic data were also excluded. Individuals with missing sociodemographic data were removed (> 0.01%).
Fig. 1.
Cohort construction flowchart. Individuals with a second SARS-CoV-2 infection are compared to population-based controls with no evidence of SARS-CoV-2 infection and individuals with only evidence of 1 SARS-CoV-2 infection. Individuals with one infection does not include individuals with second infection over follow-up period. Individuals with a second infection must report the second infection more than 90 days from the 1 st infection and more than 300 days must elapse before their 3rd infection. Individuals with more than 1 reinfection are dropped due to small sample size
Exposure
SARS-CoV-2 infection status was determined either by positive polymerase-chain-reaction (PCR) and/or rapid-antigen-test (RAT) and obtained from data recorded in the national registry maintained by the local Ministry of Health (MOH). Throughout the study period, subsidised SARS-CoV-2 testing was widely available at all hospitals and primary care clinics, including public primary-care-clinics (polyclinics) and Public-Health-Preparedness-Clinics (PHPCs), a nationwide network of more than 1000 private general-practitioner (GP) clinics activated to provide subsidised consultation and testing during pandemics [19]. Notification of all SARS-CoV-2 infections diagnosed at healthcare providers to our local MOH was mandatory under the Infectious Diseases Act [20], and public health messaging encouraged all individuals with symptoms of acute-respiratory-infection (ARI) to present to healthcare providers for diagnostic testing and treatment [19]. COVID-19 therapeutics were fully subsidised throughout the study period [21].
Exposure was taken as reinfection (2nd infection),with re-infections defined as a positive SARS-CoV-2 test (PCR and/or RAT) recorded in the national registry, occurring ≥ 90 days following the initial SARS-CoV-2 infection, with T1 considered the date of reinfection; a threshold of 90 days for re-infection post-index date is common in the literature [6]. Individuals with ≥ two infections were precluded from analysis due to low counts. The comparator groups comprised either individuals with only a single infection or population-based controls which had no evidence of infection. The date of first infection was taken as T0. The group comprising individuals with one infection excludes individuals with a second infection during the follow-up period.
Outcomes
We defined post-acute sequelae via pre-specified diagnoses of post-acute sequelae after SARS-CoV-2 reinfection contrasting against individuals with one infection or population-based controls with no evidence of infection. Population-based controls were assigned a T0 based on the distribution of T0 among individuals with at least one documented first-infection, whereas new-incident diagnoses among reinfected cases with two infections were compared against population-based controls or individuals with one infection. Comparator groups were assessed over a follow-up period spanning 31–300 days from T1. To ensure a similar distribution of follow-up time in contrasting individuals with no-reinfection vs. reinfection, we randomly assigned T1 to either individuals with one infection or population-based controls based on the distribution of T1 of reinfections. This procedure of assigning T1 was conducted among individuals with two infections who shared the same calendar month as the date of first infection or the index date of population-based controls. We also compared the risk of post-acute sequelae after reinfection with the risks of post-acute sequelae after one infection by assessing new-incident diagnoses over a follow-up period spanning 31–300 days from T0 for comparisons between individuals with one infection versus population-based controls (Fig. 2).
Fig. 2.
Schematic visualizing assignment of T0 (Date of first infection/pseudo-infection) or T1 (Date of reinfection/pseudo-reinfection) index dates for individuals with one infection and population-based controls
Diagnoses were based on International-Classification-of-Diseases, Tenth Revision (ICD-10) [22] codes recorded in the national healthcare claims database; referencing previous work on long-term sequelae post-SARS-CoV-2 infection in both the local population and the published literature, across auto-immune/cardiovascular/neurological/psychiatric/endocrine/renal/respiratory and gastrointestinal systems [1, 15–18]. We also defined the risk of all sequelae as any of the pre-specified diagnoses, and composite sequelae for a specific organ system as any of the pre-specified diagnoses in that specific organ system. Relevant definitions, and the list of ICD-10 codes, are detailed in the Supplementary material 2.
Covariates
We adjusted for differences in baseline characteristics by incorporating the following covariates: demographics (age, sex, ethnicity), comorbidity burden (Charlson’s comorbidity index [23]), presence of an immunocompromising condition, severity of acute illness (ambulatory vs. hospitalised), vaccination status, time since last vaccination before infection/re-infection as a continuous variable, preceding healthcare utilisation (any emergency-department [ED] visit or any hospitalisation in the year preceding COVID-19), and socioeconomic-status (SES). SES was classified by housing type, a key marker of SES in Singapore [19]; the majority of Singaporeans (≥ 90%) stay in owner-occupied public housing under a tiered subsidy scheme, with caps on household monthly income for eligibility to purchase more highly subsidised smaller-sized flats [19]. To control for potential waning vaccination efficacy, we also controlled for time-since vaccination and in comparisons between individuals with any infections, we controlled for the time-since infection prior to T0 or T1.
Statistical analysis
Baseline sociodemographic characteristics of infected and re-infected individuals along with standardised mean differences (SMDs) between groups were computed. For estimation of risks for each new-incident diagnosis, a sub-cohort of individuals without history of the diagnosis (in the past 5 years) was constructed. Within each sub-cohort, a propensity score of belonging to each exposure group was computed using logistic regression, using aforementioned covariates measured at baseline as explanatory variables. Overlap weights were computed as 1-propensity score for reinfected individuals; and equal to the propensity score for individuals with one infection. Covariate balance was evaluated by comparing SMDs between unexposed/exposed groups. SMD < 0.1 was taken as the threshold for good balance.
Hazard ratios (HRs) of new-incident post-acute diagnoses were then estimated using cause-specific hazard models, with overlap weights applied [24]. Cause-specific hazard models were used rather than competing risk regression models, as the incidence of competing risks such as death (~ 0.5–0.85%) was very low during the follow-up period. In this context, the bias introduced by censoring deaths as competing events is expected to be minimal, and cause-specific hazards provide a more straightforward interpretation of associations with the outcome of interest. Deaths occurring during follow-up were treated as censoring events. Bonferroni correction was employed to account for multiple-testing, where we took the significance levels of p < 0.05 for the combined primary outcome (all outcomes), p < 0.05/8 for combined organ-system outcomes, and p < 0.05/34 for individual outcomes.
Risk trajectories were estimated between vaccination groups by employing the Kaplan–Meier approach [25], with the same overlap weights applied, to compare between-group differences in outcome probabilities over the post-acute period. Outcome probabilities were taken as 1-survival probabilities per condition. Confidence intervals for risk-trajectories were obtained by estimating 1000 bootstrapped risk trajectories through simple bootstrapping. 95% intervals were obtained using the percentile method.
Weighted excess burdens (EB) per 1000 individuals for the T0 + 30 to T0 + 300 days or T1 + 30 to T1 + 300 days of follow-up of new incident diagnoses were computed based on differences in estimated incidence rates between exposure groups. Incidence rates in each group were computed by dividing the weighted incidence of that specific post-acute outcome by the weighted amount of person-time each person was observed for before censoring. HRs provide a measure of relative risk and EBs could be interpreted as a measure of absolute risk between groups. To explore potential effect modification, we conducted exploratory analyses in subgroups by age (18–65, 65 +), comorbidities (CCI = 0, CCI = 1 +), ethnicity (Chinese, non-Chinese), sex (male, female), SES (1–3 room flats, 4–5 room flats, private property). We also estimated the variant-specific risk of sequelae by periods of Omicron BA.1/2, BA.4/5 (6 January 2022–17 October 2022) and XBB pre-dominance (18 October 2022–28 February 2023).
We investigated the robustness of our results in multiple sensitivity analyses. We explored the use of stabilised inverse-probability weights or inverse-probability weights as alternative weighting schemes. We explored the use of the doubly-robust estimator, where the same explanatory variables used in the propensity score model were employed in the outcome regression, with inverse probability weights employed. To examine if risk estimates in 2nd infections were still higher compared to 1st infections in only boosted individuals, we subset our study population to boosted individuals. To examine whether our definition of reinfections (≥ 90 days following initial SARS-CoV-2 infection) skewed risk estimates, we reran our cohort defining reinfections being ≥ 60 days following initial SARS-CoV-2 infection. We employed the use of negative outcome controls to detect suspected/unsuspected sources of spurious bias, and may lessen concerns about unmeasured confounding and other latent biases [26]. Here, we explored the risk of new-incident malignancy post-reinfection as the negative outcome control; increased risk of malignancy was not expected following SARS-CoV-2 infection. Negative-outcome controls were tested against the main cohort and all subgroups, using the same analytical approaches. Analyses were conducted using R 4.3.1.
Results
After inclusion and exclusion criteria were met, 57,222 individuals with 2 SARS-CoV-2 infections were compared against 1,239,119 individuals with 1 SARS-CoV-2 infection and 3,409,170 population-based controls (Fig. 1). Baseline sociodemographic and clinical characteristics between groups, before and after overlap weighting, are presented in Table 1. After weighting, differences in demographic characteristics, socioeconomic status and comorbidity burden between the two groups were small, with all SMDs < 0.1 (Table 1). This finding was consistent across all sub-cohorts of individuals without history of the respective diagnoses.
Table 1.
Summary statistics of cohort before and after overlap weights (OW) were applied
| Covariate | 2nd infection (weighted) |
1st infection (weighted) |
2nd infection (weighted) |
No infection (weighted) |
1 st infection (weighted) |
No infection (weighted) |
|---|---|---|---|---|---|---|
| Age, years, mean (SD) | 45.6(17.4) | 45.7(17.5) | 45.6(17.4) | 45.7(17.6) | 47.4(17.6) | 47.5(17.6) |
| Time since vaccination, mean (SD), days | 163.8(219.2) | 163.8(226.0) | 162.4(219.9) | 162.4(214.2) | − 4.6(239.0) | − 4.6(230.2) |
| Vaccination | ||||||
| Unvaccinated | 1104(2.1%) | 1104(2.1%) | 1206(2.2%) | 1206(2.2%) | 16,173(2.1%) | 16,173(2.1%) |
| Fully vaccinated | 4394(8.3%) | 4394(8.3%) | 4656(8.3%) | 4656(8.3%) | 110,529(14.6%) | 110,529(14.6%) |
| Boosted (3rd dose) | 41,630(79.1%) | 41,630(79.1%) | 44,288(79.2%) | 44,288(79.2%) | 597,436(78.7%) | 597,436(78.7%) |
| Boosted (4th dose +) | 5513(10.5%) | 5513(10.5%) | 5773(10.3%) | 5773(10.3%) | 35,172(4.6%) | 35,172(4.6%) |
| Ethnicity | ||||||
| Chinese | 33,403(63.5%) | 33,403(63.5%) | 35,230(63.0%) | 35,230(63.0%) | 568,738(74.9%) | 568,738(74.9%) |
| Malay | 12,231(23.2%) | 12,231(23.2%) | 13,198(23.6%) | 13,198(23.6%) | 101,634(13.4%) | 101,634(13.4%) |
| Indian | 5409(10.3%) | 5409(10.3%) | 5792(10.4%) | 5792(10.4%) | 64,796(8.5%) | 64,796(8.5%) |
| Other | 1599(3.0%) | 1599(3.0%) | 1702(3.0%) | 1702(3.0%) | 24,142(3.2%) | 24,142(3.2%) |
| Sex | ||||||
| Male | 25,317(48.1%) | 25,317(48.1%) | 26,903(48.1%) | 26,903(48.1%) | 368,229(48.5%) | 368,229(48.5%) |
| Housing type | ||||||
| 1–2 rooms | 3365(6.4%) | 3365(6.4%) | 3617(6.5%) | 3617(6.5%) | 34,838(4.6%) | 34,838(4.6%) |
| 3 rooms | 9526(18.1%) | 9526(18.1%) | 10,199(18.2%) | 10,199(18.2%) | 114,899(15.1%) | 114,899(15.1%) |
| 4 rooms | 20,152(38.3%) | 20,152(38.3%) | 21,451(38.4%) | 21,451(38.4%) | 256,583(33.8%) | 256,583(33.8%) |
| 5 rooms/executive condominium | 17,501(33.2%) | 17,501(33.2%) | 18,427(33.0%) | 18,427(33.0%) | 303,661(40.0%) | 303,661(40.0%) |
| Private housing | 1481(2.8%) | 1481(2.8%) | 1534(2.7%) | 1534(2.7%) | 44,174(5.8%) | 44,174(5.8%) |
| Others | 615(1.2%) | 615(1.2%) | 695(1.2%) | 695(1.2%) | 5155(0.7%) | 5155(0.7%) |
| Charlson’s Comorbidity Index (CCI) | ||||||
| None | 44,709(84.9%) | 44,709(84.9%) | 47,456(84.9%) | 47,456(84.9%) | 660,827(87.0%) | 660,827(87.0%) |
| Mild: CCI Score 1–2 | 5894(11.2%) | 5894(11.2%) | 6290(11.2%) | 6290(11.2%) | 74,579(9.8%) | 74,579(9.8%) |
| Moderate: CCI Score 3–4 | 1334(2.5%) | 1334(2.5%) | 1425(2.5%) | 1425(2.5%) | 16,186(2.1%) | 16,186(2.1%) |
| Severe: CCI Score 5 + | 704(1.3%) | 704(1.3%) | 751(1.3%) | 751(1.3%) | 7717(1.0%) | 7717(1.0%) |
| Immunological conditions | ||||||
| > = 1 | 1628(3.1%) | 1628(3.1%) | 1732(3.1%) | 1732(3.1%) | 20,850(2.7%) | 20,850(2.7%) |
| Past hospital utilisation (up to 1 year prior) | ||||||
| > = 1 | 23(0.0%) | 76(0.1%) | 14,421(25.8%) | 14,421(25.8%) | 114,423(15.1%) | 114,423(15.1%) |
All standardised mean differences between comparator groups had values of less than 0.05, indicating good balance
Increased risk of post-acute sequelae across organ systems
We estimated the HRs and excess burden of pre-specified new-incident diagnoses over multiple organ systems 30 to 300 days post-T0 between individuals with one infection versus no infection. Overall, we estimated an 12% (HR:1.10; 95% confidence interval (CI) [1.04–1.14]) increase in the risk of any new-incident sequelae associated with individuals who had one infection versus population-based controls, with heightened risks across the cardiovascular, neurological, psychiatric, endocrine, respiratory, renal, and gastrointestinal domains (Table 2).
Table 2.
Hazard ratios (HR) and excess burdens (EB) per 1000 for composite and individual outcomes
| No infection v. one infection(A) | One infection v. two infections(B) | No infection v. two infections(C) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Condition | Hazard Ratio | HR | Excess burden | Hazard ratio | HR | Excess burden | Hazard Ratio | HR | Excess burden |
| (95% CI) | P-value | (95% CI) | (95% CI) | P-value | (95% CI) | (95% CI) | P-value | (95% CI) | |
| All outcomes | 1.12 (1.10–1.14)* | 0 | 2.96 (2.39–3.52) | 1.17 (1.11–1.24)* | 0 | 4.22 (2.02–6.41) | 1.21 (1.15–1.28)* | 0 | 5.16 (3.05–7.28) |
| All cardiovascular outcomes | 1.09 (1.06–1.12)^ | 0 | 0.52 (0.26–0.77) | 1.27 (1.15–1.41)^ | 0 | 1.68 (0.65–2.71) | 1.20 (1.09–1.32)^ | 0.0002 | 1.29 (0.27–2.30) |
| All neurological outcomes | 1.13 (1.11–1.16)^ | 0 | 1.25 (0.92–1.58) | 1.26 (1.17–1.36)^ | 0 | 2.86 (1.47–4.26) | 1.32 (1.23–1.42)^ | 0 | 3.34 (2.00–4.68) |
| All psychiatric outcomes | 1.08 (1.03–1.14)^ | 0.0032 | 0.15 (0.01–0.29) | 1.15 (0.97–1.36) | 0.1139 | 0.33 (− 0.27–0.94) | 1.15 (0.98–1.36) | 0.0877 | 0.35 (− 0.24–0.94) |
| All endocrine outcomes | 1.12 (1.10–1.15)^ | 0 | 1.97 (1.53–2.41) | 1.18 (1.11–1.27)^ | 0 | 2.88 (1.19–4.56) | 1.26 (1.18–1.35)^ | 0 | 3.86 (2.24–5.47) |
| All autoimmune outcomes | 1.06 (0.99–1.14) | 0.0876 | 0.06 (− 0.04–0.17) | 1.21 (0.95–1.53) | 0.1155 | 0.23 (− 0.20–0.66) | 1.27 (1.01–1.59) | 0.0373 | 0.29 (− 0.12–0.70) |
| All respiratory outcomes | 1.16 (1.11–1.21)^ | 0 | 0.37 (0.20–0.53) | 1.62 (1.43–1.84)^ | 0 | 1.93 (1.15–2.71) | 1.63 (1.44–1.83)^ | 0 | 1.94 (1.18–2.69) |
| All renal outcomes | 1.07 (1.02–1.13)^ | 0.004 | 0.15 (0.00–0.31) | 1.26 (1.09–1.46)^ | 0.0014 | 0.79 (0.08–1.50) | 1.28 (1.12–1.47)^ | 0.0003 | 0.84 (0.15–1.54) |
| All gastrointestinal outcomes | 1.19 (1.16–1.22)^ | 0 | 1.56 (1.25–1.87) | 1.19 (1.09–1.30)^ | 0.0001 | 1.65 (0.44–2.86) | 1.26 (1.16–1.38)^ | 0 | 2.15 (0.99–3.31) |
| Dysrhythmias | 1.12 (1.07–1.17)& | 0 | 0.27 (0.11–0.43) | 1.31 (1.13–1.52)& | 0.0003 | 0.84 (0.16–1.52) | 1.26 (1.09–1.45) | 0.0016 | 0.72 (0.05–1.38) |
| Inflammatory heart disease | 0.99 (0.59–1.65) | 0.966 | − 0.00 (− 0.01–0.01) | 2.67 (0.89–8.01) | 0.0793 | 0.04 (− 0.04–0.13) | 2.61 (0.95–7.22) | 0.0637 | 0.04 (− 0.04–0.12) |
| Ischemic heart disease | 1.10 (1.06–1.14)& | 0 | 0.32 (0.13–0.50) | 1.22 (1.06–1.40) | 0.0044 | 0.73 (− 0.01–1.47) | 1.14 (1.00–1.31) | 0.0474 | 0.50 (− 0.23–1.23) |
| Other heart disease | 1.03 (0.97–1.09) | 0.3946 | 0.04 (− 0.09–0.16) | 1.27 (1.05–1.53) | 0.0133 | 0.46 (− 0.08–1.00) | 1.12 (0.94–1.34) | 0.2015 | 0.24 (− 0.30–0.77) |
| Thrombotic disorders | 1.10 (1.01–1.21) | 0.0261 | 0.07 (− 0.02–0.15) | 1.18 (0.89–1.55) | 0.2563 | 0.15 (− 0.22–0.51) | 1.14 (0.88–1.49) | 0.3269 | 0.12 (− 0.24–0.48) |
| Cerebrovascular disease | 0.95 (0.91–1.00) | 0.0593 | − 0.10 (− 0.25–0.04) | 1.05 (0.87–1.26) | 0.5968 | 0.11 (− 0.47–0.68) | 0.96 (0.81–1.15) | 0.6895 | − 0.08 (− 0.65–0.48) |
| Cognitive disorders | 1.16 (1.10–1.23)& | 0 | 0.24 (0.12–0.37) | 1.36 (1.14–1.63)& | 0.0007 | 0.65 (0.09–1.20) | 1.34 (1.13–1.59)& | 0.0007 | 0.62 (0.08–1.17) |
| Peripheral neuropathies | 1.19 (1.10–1.29)& | 0 | 0.14 (0.05–0.23) | 1.34 (1.04–1.72) | 0.0218 | 0.31 (− 0.09–0.71) | 1.42 (1.12–1.80) | 0.0039 | 0.36 (− 0.02–0.75) |
| Episodic disorders | 1.20 (1.11–1.29)& | 0 | 0.17 (0.07–0.26) | 1.33 (1.07–1.65) | 0.0103 | 0.40 (− 0.06–0.86) | 1.39 (1.13–1.72) | 0.0018 | 0.46 (0.02–0.90) |
| Movement disorders | 1.04 (0.94–1.14) | 0.4702 | 0.02 (− 0.06–0.09) | 1.23 (0.88–1.73) | 0.2235 | 0.12 (− 0.17–0.42) | 1.18 (0.85–1.62) | 0.3256 | 0.10 (− 0.19–0.39) |
| Musculoskeletal disorders | 1.19 (1.13–1.25)& | 0 | 0.33 (0.19–0.47) | 1.34 (1.14–1.58)& | 0.0005 | 0.73 (0.12–1.34) | 1.45 (1.24–1.69)& | 0 | 0.89 (0.31–1.47) |
| Sensory disorders | 1.17 (1.11–1.23)& | 0 | 0.30 (0.16–0.44) | 1.28 (1.08–1.52) | 0.0053 | 0.55 (− 0.03–1.13) | 1.37 (1.16–1.62)& | 0.0002 | 0.69 (0.14–1.25) |
| Other neurological disorders | 1.22 (1.17–1.27)& | 0 | 0.57 (0.40–0.75) | 1.27 (1.12–1.45)& | 0.0003 | 0.96 (0.18–1.74) | 1.39 (1.23–1.58)& | 0 | 1.28 (0.54–2.02) |
| Mood disorders | 1.10 (1.02–1.19) | 0.0133 | 0.08 (− 0.01–0.18) | 1.04 (0.80–1.34) | 0.7904 | 0.04 (− 0.37–0.44) | 1.08 (0.84–1.38) | 0.5485 | 0.08 (− 0.31–0.48) |
| Stress/anxiety disorders | 1.17 (1.10–1.25)& | 0 | 0.18 (0.07–0.28) | 1.10 (0.88–1.39) | 0.4022 | 0.13 (− 0.32–0.59) | 1.12 (0.90–1.39) | 0.3116 | 0.15 (− 0.28–0.59) |
| Psychotic disorders | 0.81 (0.71–0.93) | 0.002 | − 0.06 (− 0.12–0.01) | 1.35 (0.89–2.06) | 0.1554 | 0.12 (− 0.12–0.36) | 1.17 (0.79–1.73) | 0.4296 | 0.07 (− 0.18–0.31) |
| Diabetes | 1.09 (1.07–1.12)& | 0 | 0.86 (0.53–1.18) | 1.25 (1.15–1.36)& | 0 | 2.18 (0.94–3.42) | 1.31 (1.21–1.42)& | 0 | 2.58 (1.39–3.78) |
| Dyslipidemia | 1.12 (1.10–1.14)& | 0 | 1.86 (1.41–2.30) | 1.18 (1.10–1.26)& | 0 | 2.79 (1.10–4.49) | 1.24 (1.16–1.32)& | 0 | 3.65 (2.02–5.28) |
| Thyroid | 1.05 (0.92–1.20) | 0.4386 | 0.02 (− 0.04–0.07) | 0.91 (0.54–1.53) | 0.7261 | − 0.03 (− 0.24–0.18) | 0.88 (0.54–1.45) | 0.6169 | − 0.04 (− 0.24–0.16) |
| Lupus | 1.00 (0.79–1.25) | 0.9713 | − 0.00 (− 0.03–0.03) | 1.89 (1.07–3.34) | 0.0282 | 0.11 (− 0.05–0.28) | 2.16 (1.26–3.70) | 0.0051 | 0.13 (− 0.03–0.29) |
| Connective tissue disorders | 1.09 (0.98–1.22) | 0.1018 | 0.04 (− 0.03–0.11) | 1.31 (0.90–1.91) | 0.1566 | 0.13 (− 0.14–0.40) | 1.27 (0.89–1.81) | 0.1939 | 0.11 (− 0.15–0.38) |
| Vasculitis | 0.69 (0.51–0.94) | 0.0181 | − 0.02 (− 0.05–0.00) | 0.82 (0.25–2.63) | 0.7337 | − 0.01 (− 0.11–0.08) | 0.74 (0.24–2.31) | 0.6032 | − 0.02 (− 0.11–0.07) |
| Spondyloarthropathies | 0.89 (0.71–1.12) | 0.3079 | − 0.01 (− 0.04–0.02) | 0.72 (0.29–1.79) | 0.4763 | − 0.04 (− 0.16–0.09) | 0.76 (0.31–1.84) | 0.5373 | − 0.03 (− 0.14–0.09) |
| Skin disorders | 1.15 (1.03–1.30) | 0.0159 | 0.05 (− 0.01–0.12) | 1.08 (0.73–1.61) | 0.6973 | 0.03 (− 0.22–0.29) | 1.17 (0.80–1.72) | 0.4149 | 0.07 (− 0.18–0.31) |
| COPD and bronchiectasis | 0.97 (0.89–1.05) | 0.443 | − 0.03 (− 0.12–0.06) | 1.90 (1.51–2.38)& | 0 | 0.76 (0.34–1.19) | 1.73 (1.41–2.14)& | 0 | 0.68 (0.26–1.10) |
| Asthma | 1.23 (1.17–1.30)& | 0 | 0.37 (0.23–0.50) | 1.61 (1.39–1.86)& | 0 | 1.45 (0.77–2.13) | 1.67 (1.46–1.92)& | 0 | 1.53 (0.88–2.19) |
| Pulmonary fibrosis | 1.13 (0.90–1.42) | 0.3101 | 0.01 (− 0.02–0.04) | 1.50 (0.77–2.94) | 0.2329 | 0.06 (− 0.09–0.21) | 1.68 (0.89–3.17) | 0.1072 | 0.07 (− 0.07–0.21) |
| Acute kidney disease | 1.06 (1.01–1.11) | 0.0307 | 0.11 (− 0.03–0.25) | 1.18 (1.02–1.38) | 0.0301 | 0.51 (− 0.16–1.18) | 1.21 (1.05–1.40) | 0.0091 | 0.58 (− 0.07–1.24) |
| End stage kidney disease | 1.19 (1.05–1.36) | 0.006 | 0.06 (− 0.00–0.11) | 1.76 (1.27–2.44)& | 0.0007 | 0.33 (0.03–0.62) | 1.58 (1.17–2.15) | 0.0031 | 0.28 (− 0.01–0.58) |
| Gastritis | 1.20 (1.17–1.23)& | 0 | 1.43 (1.14–1.72) | 1.22 (1.11–1.34)& | 0 | 1.67 (0.53–2.80) | 1.31 (1.20–1.43)& | 0 | 2.18 (1.10–3.26) |
| Irritable bowel syndrome | 1.28 (1.13–1.45)& | 0.0001 | 0.08 (0.02–0.13) | 1.42 (0.90–2.22) | 0.1276 | 0.11 (− 0.11–0.33) | 1.47 (0.95–2.27) | 0.0824 | 0.12 (− 0.09–0.32) |
| Biliary tract disease | 1.09 (0.98–1.22) | 0.1154 | 0.04 (− 0.03–0.10) | 0.73 (0.47–1.14) | 0.1703 | − 0.13 (− 0.38–0.12) | 0.76 (0.50–1.18) | 0.2194 | − 0.11 (− 0.36–0.13) |
| Non− infectious hepatitis and cirrhosis | 1.19 (0.87–1.63) | 0.2795 | 0.01 (− 0.01–0.03) | 0.62 (0.15–2.60) | 0.5106 | − 0.02 (− 0.10–0.06) | 0.63 (0.16–2.56) | 0.5223 | − 0.02 (− 0.10–0.06) |
| Inflammatory bowel disease | 1.26 (1.16–1.38)& | 0 | 0.15 (0.07–0.23) | 1.59 (1.23–2.06)& | 0.0004 | 0.42 (0.05–0.79) | 1.54 (1.20–1.98)& | 0.0007 | 0.40 (0.04–0.76) |
Numbers in parentheses represent 95% Cis
(A)T0 + 30 to T0 + 300 contrasting population-based controls versus individuals with one infection
(B)T1 + 30 to T1 + 300 contrasting individuals who have 1 COVID-19 infection versus individuals with two infections
(C)T1 + 30 to T1 + 300 contrasting population-based controls versus individuals with two infections. A HR > 1 represents higher risk of that pre-specified sequelae in the follow-up period for reinfected individuals relative to infected individuals. HRs are estimated using Cox proportional hazards models with overlapping weights
*p.val < 0.05 (i.e. the Bonferroni corrected P-value, given that there is 1 tested outcome for all sequelae)
^p.val < 0.05/8 (i.e. the Bonferroni corrected P-value, given that there are 8 tested composite outcomes)
&p.val < 0.05/34 (i.e. the Bonferroni corrected P-value, given that there are 34 tested individual outcomes)
Risks of sequelae associated with reinfections were greater versus risks after the first infection. Individuals with reinfections had 17% (HR = 1.17; 95% CI [1.11–1.24]) and 21% (HR = 1.21; 95% CI [1.15–1.28]) increased risks of any sequelae when compared to individuals with one infection or no infection respectively. Associated risks of sequelae across different organ systems also increased for reinfections versus risks after the first infection, with increased risks in the cardiovascular, neurological, endocrine, respiratory, renal and gastrointestinal domains. Associated risks of outcomes cumulated over the follow-up time, with higher risks of new incident sequelae across organ systems in individuals with 2nd infection vs. 1 st infection (Fig. 3). Lastly, stratifying our analyses by days since infection for the group with two infections and comparing these to population-based controls, we found that associated risks of post-acute sequelae did not appear to attenuate with time since previous infection (Table 3).
Fig. 3.
Outcome probabilities (1-survival probabilities)*100 of composite diagnoses in individuals with one infection versus two infections from T0 + 30 to T0 + 300
Table 3.
Hazard ratios (HR) and excess burdens (EB) per 1000 for composite and individual outcomes from T1 + 30 to T1 + 300 contrasting population-based controls versus individuals with two infections
| Days since infection for 2nd infection group | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 0–120 days | 120–180 | 180–240 | 240–300 | > 300 | ||||||
| Condition | Hazard ratio | HR | Hazard ratio | HR | Hazard ratio | HR | Hazard ratio | HR | Hazard ratio | HR |
| (95% CI) | P-value | (95% CI) | P-value | (95% CI) | P-value | (95% CI) | P-value | (95% CI) | P-value | |
| All outcomes | 1.32 (1.01–1.73)* | 0.0389 | 1.18 (1.02–1.38)* | 0.0275 | 1.19 (1.07–1.32)* | 0.0019 | 1.12 (1.01–1.24)* | 0.0322 | 1.17 (1.05–1.30)* | 0.0037 |
| All cardiovascular outcomes | 1.54 (0.99–2.39) | 0.0534 | 1.35 (1.03–1.78)* | 0.0283 | 1.33 (1.09–1.61)* | 0.0043 | 1.21 (1.01–1.45)* | 0.0412 | 1.25 (1.03–1.51)* | 0.0251 |
| All neurological outcomes | 1.56 (1.12–2.18)* | 0.0089 | 1.51 (1.25–1.83)* | 0 | 1.22 (1.05–1.42)* | 0.0093 | 1.22 (1.06–1.40)* | 0.0053 | 1.22 (1.05–1.42)* | 0.0086 |
| All psychiatric outcomes | 0.64 (0.24–1.73) | 0.3767 | 0.77 (0.45–1.32) | 0.3421 | 1.07 (0.76–1.51) | 0.6857 | 1.34 (0.99–1.80) | 0.0575 | 1.35 (0.99–1.85) | 0.0549 |
| All endocrine outcomes | 1.31 (0.94–1.84) | 0.1135 | 1.30 (1.08–1.56)* | 0.0057 | 1.30 (1.14–1.48)* | 0.0001 | 1.14 (1.01–1.29)* | 0.0379 | 1.09 (0.96–1.24) | 0.1726 |
| All autoimmune outcomes | 0.81 (0.20–3.27) | 0.765 | 0.95 (0.47–1.94) | 0.8922 | 1.53 (1.00–2.35) | 0.0509 | 1.37 (0.92–2.04) | 0.1251 | 1.12 (0.71–1.77) | 0.6252 |
| All respiratory outcomes | 2.84 (1.78–4.52)* | 0 | 1.78 (1.31–2.41)* | 0.0002 | 2.04 (1.62–2.57)* | 0 | 1.44 (1.14–1.83)* | 0.0025 | 1.13 (0.85–1.51) | 0.3841 |
| All renal outcomes | 2.08 (1.24–3.48)* | 0.0057 | 1.18 (0.80–1.73) | 0.4086 | 1.25 (0.95–1.66) | 0.1134 | 1.24 (0.96–1.60) | 0.1033 | 1.31 (0.99–1.73) | 0.0608 |
| All gastrointestinal outcomes | 0.86 (0.52–1.41) | 0.5443 | 1.45 (1.17–1.80)* | 0.0006 | 1.18 (1.00–1.40) | 0.0538 | 1.11 (0.94–1.31) | 0.2053 | 1.18 (0.99–1.40) | 0.0574 |
| Dysrhythmias | 1.73 (0.97–3.10) | 0.0656 | 1.37 (0.92–2.04) | 0.12 | 1.26 (0.94–1.68) | 0.1161 | 1.15 (0.88–1.52) | 0.3008 | 1.48 (1.11–1.97)* | 0.0081 |
| Inflammatory heart disease | – | – | – | – | – | – | 7.09 (1.87–26.89)* | 0.004 | – | – |
| Ischemic heart disease | 1.15 (0.57–2.33) | 0.695 | 1.34 (0.94–1.93) | 0.1083 | 1.27 (0.97–1.67) | 0.0848 | 1.22 (0.95–1.56) | 0.1209 | 1.18 (0.91–1.54) | 0.2093 |
| Other heart disease | 1.28 (0.56–2.90) | 0.558 | 1.34 (0.82–2.18) | 0.2398 | 1.56 (1.12–2.19)* | 0.0092 | 1.24 (0.89–1.73) | 0.1985 | 0.94 (0.61–1.43) | 0.7575 |
| Thrombotic disorders | 4.84 (2.33–10.07)* | 0 | 1.00 (0.44–2.29) | 0.9962 | 1.45 (0.86–2.42) | 0.1607 | 0.55 (0.27–1.13) | 0.1031 | 1.29 (0.79–2.13) | 0.3094 |
| Cerebrovascular disease | 1.57 (0.77–3.21) | 0.2182 | 1.25 (0.80–1.94) | 0.3306 | 1.06 (0.73–1.53) | 0.7687 | 1.10 (0.80–1.52) | 0.5618 | 0.88 (0.61–1.27) | 0.4928 |
| Cognitive disorders | 0.96 (0.42–2.17) | 0.9161 | 1.73 (1.13–2.63)* | 0.0109 | 1.33 (0.94–1.90) | 0.1108 | 1.48 (1.08–2.03)* | 0.0141 | 1.29 (0.89–1.85) | 0.1754 |
| Peripheral neuropathies | 1.93 (0.72–5.23) | 0.1935 | 1.30 (0.66–2.56) | 0.4532 | 1.32 (0.80–2.16) | 0.2746 | 1.40 (0.91–2.13) | 0.1235 | 1.20 (0.72–2.01) | 0.4891 |
| Episodic disorders | 2.19 (1.01–4.74)* | 0.0472 | 1.74 (1.06–2.86)* | 0.0295 | 1.46 (0.97–2.19) | 0.0673 | 0.89 (0.56–1.40) | 0.6039 | 1.33 (0.86–2.07) | 0.2037 |
| Movement disorders | 3.00 (0.94–9.55) | 0.0635 | 1.45 (0.63–3.32) | 0.3814 | 1.58 (0.89–2.80) | 0.1197 | 0.92 (0.47–1.81) | 0.8082 | 1.23 (0.60–2.50) | 0.5718 |
| Musculoskeletal disorders | 1.87 (0.96–3.65) | 0.0661 | 2.02 (1.37–3.00)* | 0.0004 | 1.01 (0.71–1.42) | 0.9726 | 1.46 (1.09–1.94)* | 0.0104 | 1.27 (0.92–1.76) | 0.1402 |
| Sensory disorders | 1.31 (0.58–2.96) | 0.5113 | 0.88 (0.48–1.61) | 0.6807 | 1.21 (0.86–1.70) | 0.2655 | 1.40 (1.04–1.87)* | 0.0253 | 1.40 (1.01–1.93)* | 0.0437 |
| Other neurological disorders | 1.19 (0.61–2.31) | 0.6034 | 1.31 (0.93–1.83) | 0.1198 | 1.33 (1.04–1.71)* | 0.0251 | 1.24 (0.98–1.58) | 0.0773 | 1.27 (0.98–1.63) | 0.0693 |
| Mood disorders | 0.58 (0.14–2.37) | 0.4494 | 0.50 (0.21–1.22) | 0.1278 | 0.89 (0.53–1.51) | 0.6706 | 1.21 (0.76–1.93) | 0.4114 | 1.52 (0.98–2.36) | 0.0606 |
| Stress/anxiety disorders | 0.58 (0.14–2.36) | 0.4465 | 1.07 (0.58–1.97) | 0.8301 | 0.99 (0.62–1.60) | 0.9822 | 1.28 (0.86–1.90) | 0.2303 | 1.14 (0.74–1.75) | 0.5484 |
| Psychotic disorders | 0.94 (0.12–7.35) | 0.95 | 1.46 (0.45–4.76) | 0.5279 | 1.34 (0.57–3.13) | 0.5018 | 1.16 (0.50–2.70) | 0.7237 | 1.58 (0.78–3.19) | 0.1993 |
| Diabetes | 1.00 (0.62–1.62) | 0.9991 | 1.56 (1.25–1.94)* | 0.0001 | 1.27 (1.07–1.50)* | 0.0052 | 1.25 (1.08–1.46)* | 0.0039 | 1.14 (0.96–1.34) | 0.1261 |
| Dyslipidemia | 1.06 (0.73–1.53) | 0.7749 | 1.12 (0.92–1.37) | 0.2519 | 1.25 (1.10–1.42)* | 0.0008 | 1.28 (1.14–1.44)* | 0 | 1.07 (0.94–1.22) | 0.3112 |
| Thyroid | 1.52 (0.21–11.14) | 0.6783 | 1.05 (0.32–3.43) | 0.9338 | 1.22 (0.44–3.37) | 0.6979 | 0.59 (0.18–1.88) | 0.3701 | 1.15 (0.48–2.77) | 0.7551 |
| Lupus | 5.20 (0.65–41.46) | 0.1196 | – | – | 2.21 (0.78–6.29) | 0.1374 | 2.22 (0.78–6.37) | 0.1369 | 1.96 (0.72–5.32) | 0.1876 |
| Connective tissue disorders | 1.27 (0.18–9.16) | 0.8094 | 1.55 (0.56–4.29) | 0.3964 | 1.41 (0.69–2.91) | 0.347 | 1.54 (0.80–2.97) | 0.1981 | 1.11 (0.54–2.26) | 0.7784 |
| Vasculitis | – | – | 2.89 (0.37–22.77) | 0.3143 | 2.90 (0.66–12.69) | 0.1573 | – | – | 1.17 (0.14–9.59) | 0.8844 |
| Spondyloarthropathies | – | – | – | – | 2.71 (0.94–7.79) | 0.0644 | 0.46 (0.06–3.45) | 0.4533 | – | – |
| Skin disorders | 0.79 (0.11–5.72) | 0.8173 | 1.20 (0.43–3.29) | 0.7289 | 0.80 (0.32–1.96) | 0.6243 | 1.37 (0.75–2.51) | 0.3069 | 1.14 (0.51–2.55) | 0.7516 |
| COPD and Bronchiectasis | 3.19 (1.37–7.41)* | 0.007 | 2.49 (1.40–4.45)* | 0.002 | 2.74 (1.84–4.09)* | 0 | 1.55 (1.01–2.40)* | 0.0467 | 1.31 (0.82–2.07) | 0.2556 |
| Asthma | 2.59 (1.48–4.52)* | 0.0009 | 1.93 (1.39–2.67)* | 0.0001 | 1.78 (1.35–2.35)* | 0 | 1.49 (1.14–1.95)* | 0.0034 | 1.19 (0.86–1.65) | 0.3044 |
| Pulmonary fibrosis | – | – | 3.29 (0.73–14.75) | 0.1205 | 2.06 (0.62–6.85) | 0.2366 | 1.96 (0.68–5.70) | 0.2156 | 0.40 (0.05–3.01) | 0.3722 |
| Acute kidney disease | 1.66 (0.92–3.01) | 0.0929 | 1.07 (0.70–1.62) | 0.7599 | 1.26 (0.94–1.70) | 0.118 | 1.10 (0.83–1.45) | 0.5057 | 1.27 (0.95–1.71) | 0.1119 |
| End stage kidney disease | 4.00 (1.56–10.24)* | 0.0039 | 3.36 (1.65–6.82)* | 0.0008 | 1.16 (0.59–2.31) | 0.666 | 1.68 (0.88–3.18) | 0.1132 | 1.60 (0.82–3.11) | 0.1659 |
| Gastritis | 0.90 (0.53–1.53) | 0.6984 | 1.51 (1.21–1.90)* | 0.0003 | 1.25 (1.05–1.49)* | 0.0131 | 1.14 (0.96–1.35) | 0.141 | 1.20 (1.00–1.43) | 0.0534 |
| Irritable bowel syndrome | – | – | 2.22 (0.88–5.58) | 0.0901 | 0.93 (0.34–2.52) | 0.8796 | 2.10 (1.08–4.10)* | 0.0295 | 0.88 (0.26–2.90) | 0.8278 |
| Biliary tract disease | – | – | 0.43 (0.11–1.76) | 0.2421 | 0.75 (0.33–1.70) | 0.4894 | 0.59 (0.24–1.46) | 0.2565 | 1.22 (0.57–2.59) | 0.6082 |
| Non-infectious hepatitis and cirrhosis | – | – | – | – | 1.98 (0.25–15.57) | 0.5156 | – | – | 1.46 (0.17–12.66) | 0.7311 |
| Inflammatory bowel disease | 1.28 (0.40–4.13) | 0.6749 | 1.55 (0.81–2.98) | 0.1858 | 1.36 (0.83–2.23) | 0.2274 | 1.66 (1.05–2.62)* | 0.0286 | 1.68 (0.94–3.00) | 0.0801 |
Analyses were stratified by days since infection for the group with two infections. A HR > 1 represents higher risk of that pre-specified sequelae in the follow-up period for reinfected individuals relative to infected individuals. HRs are estimated using Cox proportional hazards models with overlapping weights
Numbers in parentheses represent 95% CIs
*p.val < 0.05
Increased risk of individual complications after reinfection
Among individual outcomes, after Bonferroni correction, reinfections were associated with increased risks and excess burdens of dysrhythmias, cognitive disorders, musculoskeletal disorders, other neurological disorders, diabetes, dyslipidemia, chronic obstructive pulmonary disease (COPD), bronchiectasis, asthma and gastritis. Risks were replicated across comparisons between reinfections and individuals with one infection or population-based controls (Table 2). Risks persisted over the follow-up time since 2nd infection (Table 3) and also across different periods of BA.1/2, 4./5 and XBB predominance (Table 4).
Table 4.
Hazard ratios (HR) and excess burdens (EB) per 1000 for composite and individual outcomes
| No infection v. one infection(A) | One infection v. two infections(B) | No infection v. two infections(C) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Condition | BA.1/2 | BA.4/5 | XBB | BA.1/2 | BA.4/5 | XBB | BA.1/2 | BA.4/5 | XBB |
| HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | |
| All outcomes | 1.13 (1.10–1.15)* | 1.11 (1.08–1.14)* | 1.14 (1.09–1.21)* | 0.97 (0.64–1.48) | 1.28 (1.14–1.44)* | 1.20 (1.12–1.29)* | 0.89 (0.57–1.37) | 1.27 (1.12–1.43)* | 1.14 (1.06–1.24)* |
| All cardiovascular outcomes | 1.11 (1.06–1.16)^ | 1.08 (1.02–1.14) | 1.07 (0.97–1.18) | 1.15 (0.65–2.03) | 1.26 (1.02–1.55) | 1.24 (1.09–1.41)^ | 1.31 (0.72–2.39) | 1.37 (1.09–1.71)^ | 1.32 (1.15–1.51)^ |
| All neurological outcomes | 1.12 (1.08–1.16)^ | 1.15 (1.11–1.20)^ | 1.09 (1.01–1.17) | 1.37 (0.85–2.22) | 1.22 (1.03–1.45) | 1.32 (1.20–1.46)^ | 2.02 (1.20–3.41) | 1.21 (1.01–1.45) | 1.26 (1.14–1.40)^ |
| All psychiatric outcomes | 1.07 (0.99–1.16) | 1.12 (1.02–1.23) | 1.25 (1.04–1.49) | 1.28 (0.45–3.68) | 1.54 (1.14–2.08)^ | 1.01 (0.80–1.29) | 1.48 (0.48–4.50) | 1.72 (1.25–2.38)^ | 0.99 (0.77–1.27) |
| All endocrine outcomes | 1.14 (1.11–1.17)^ | 1.11 (1.07–1.15)^ | 1.14 (1.07–1.21)^ | 1.22 (0.77–1.94) | 1.35 (1.17–1.55)^ | 1.20 (1.10–1.32)^ | 1.25 (0.78–2.01) | 1.28 (1.10–1.48)^ | 1.13 (1.03–1.24) |
| All autoimmune outcomes | 1.13 (1.02–1.26) | 1.00 (0.89–1.14) | 1.11 (0.87–1.42) | 1.43 (0.32–6.44) | 1.20 (0.73–1.99) | 1.30 (0.95–1.77) | 0.84 (0.16–4.27) | 1.14 (0.67–1.93) | 1.23 (0.89–1.71) |
| All respiratory outcomes | 1.19 (1.11–1.28)^ | 1.12 (1.03–1.22) | 1.34 (1.16–1.56)^ | 2.52 (1.19–5.34) | 1.74 (1.35–2.23)^ | 1.65 (1.40–1.95)^ | 1.74 (0.80–3.79) | 1.86 (1.41–2.44)^ | 1.63 (1.37–1.94)^ |
| All renal outcomes | 1.12 (1.04–1.21)^ | 1.07 (0.98–1.17) | 1.02 (0.87–1.20) | 1.20 (0.59–2.44) | 1.31 (0.97–1.75) | 1.25 (1.03–1.52) | 1.63 (0.76–3.50) | 1.47 (1.07–2.02) | 1.21 (0.98–1.48) |
| All gastrointestinal outcomes | 1.15 (1.11–1.19)^ | 1.23 (1.17–1.28)^ | 1.25 (1.15–1.36)^ | 0.86 (0.44–1.66) | 1.26 (1.05–1.52) | 1.25 (1.11–1.40)^ | 0.70 (0.35–1.38) | 1.29 (1.05–1.57) | 1.17 (1.03–1.31) |
| Dysrhythmias | 1.14 (1.06–1.21)& | 1.08 (0.99–1.17) | 1.02 (0.88–1.19) | 1.04 (0.43–2.48) | 1.18 (0.85–1.62) | 1.34 (1.11–1.62) | 1.74 (0.67–4.50) | 1.34 (0.95–1.89) | 1.40 (1.15–1.71)& |
| Inflammatory heart disease | 1.92 (0.91–4.09) | 0.92 (0.35–2.40) | – | – | 6.13 (0.68–55.54) | 0.99 (0.14–7.18) | – | 1.85 (0.18–18.96) | 1.16 (0.15–9.00) |
| Ischemic heart disease | 1.10 (1.03–1.16) | 1.13 (1.05–1.21) | 1.09 (0.95–1.25) | 2.01 (1.09–3.72) | 1.20 (0.91–1.60) | 1.15 (0.96–1.38) | 1.98 (1.04–3.77) | 1.29 (0.95–1.74) | 1.22 (1.01–1.48) |
| Other heart disease | 1.06 (0.97–1.16) | 0.98 (0.88–1.09) | 0.98 (0.80–1.19) | 2.07 (0.93–4.58) | 1.37 (0.96–1.96) | 1.13 (0.88–1.45) | 2.18 (0.95–4.98) | 1.50 (1.03–2.21) | 1.34 (1.03–1.74) |
| Thrombotic disorders | 1.10 (0.97–1.25) | 1.08 (0.92–1.27) | 1.34 (1.00–1.79) | 0.50 (0.06–3.88) | 1.13 (0.63–2.02) | 1.28 (0.89–1.83) | 0.69 (0.08–5.61) | 1.48 (0.79–2.75) | 1.22 (0.83–1.78) |
| Cerebrovascular disease | 0.99 (0.93–1.07) | 0.94 (0.85–1.03) | 0.92 (0.77–1.09) | 1.72 (0.84–3.54) | 0.67 (0.42–1.07) | 1.04 (0.82–1.31) | 2.85 (1.26–6.44) | 0.66 (0.41–1.07) | 1.13 (0.88–1.45) |
| Cognitive disorders | 1.15 (1.06–1.25)& | 1.33 (1.20–1.47)& | 1.04 (0.86–1.25) | 0.63 (0.19–2.11) | 1.73 (1.27–2.37)& | 1.25 (0.98–1.60) | 0.88 (0.24–3.23) | 1.95 (1.37–2.78)& | 1.25 (0.96–1.62) |
| Peripheral neuropathies | 1.14 (1.01–1.29) | 1.24 (1.08–1.43) | 1.13 (0.87–1.48) | 0.96 (0.12–7.91) | 1.33 (0.78–2.28) | 1.38 (0.99–1.92) | 1.17 (0.13–10.25) | 1.22 (0.70–2.14) | 1.32 (0.93–1.86) |
| Episodic disorders | 1.10 (0.98–1.23) | 1.33 (1.16–1.52)& | 1.11 (0.86–1.42) | – | 0.92 (0.53–1.61) | 1.59 (1.22–2.08)& | – | 1.00 (0.56–1.78) | 1.52 (1.15–2.02) |
| Movement disorders | 1.01 (0.87–1.17) | 1.10 (0.92–1.31) | 0.99 (0.72–1.35) | 1.26 (0.15–10.26) | 0.76 (0.31–1.84) | 1.31 (0.86–2.01) | 2.16 (0.22–21.10) | 0.81 (0.32–2.04) | 1.51 (0.97–2.35) |
| Musculoskeletal disorders | 1.18 (1.09–1.28)& | 1.20 (1.10–1.32)& | 1.11 (0.94–1.32) | 4.01 (1.71–9.39)& | 1.44 (1.01–2.04) | 1.42 (1.15–1.76)& | 2.98 (1.15–7.77) | 1.42 (0.97–2.07) | 1.31 (1.05–1.64) |
| Sensory disorders | 1.18 (1.09–1.27)& | 1.12 (1.02–1.23) | 1.24 (1.06–1.46) | 2.49 (0.85–7.32) | 1.16 (0.78–1.73) | 1.50 (1.20–1.86)& | 2.60 (0.84–8.04) | 1.20 (0.79–1.83) | 1.37 (1.10–1.72) |
| Other neurological disorders | 1.18 (1.11–1.26)& | 1.26 (1.16–1.36)& | 1.22 (1.07–1.40) | 0.57 (0.18–1.86) | 1.44 (1.09–1.89) | 1.43 (1.21–1.69)& | 1.57 (0.41–5.96) | 1.38 (1.03–1.86) | 1.31 (1.10–1.56) |
| Mood disorders | 1.02 (0.91–1.14) | 1.26 (1.10–1.46)& | 1.36 (1.04–1.77) | 1.82 (0.52–6.38) | 1.32 (0.83–2.09) | 0.87 (0.59–1.27) | 2.47 (0.66–9.31) | 1.44 (0.87–2.36) | 0.80 (0.54–1.19) |
| Stress/anxiety disorders | 1.16 (1.05–1.28) | 1.18 (1.04–1.34) | 1.30 (1.03–1.64) | 0.90 (0.12–6.86) | 1.79 (1.22–2.62) | 1.05 (0.77–1.44) | 0.68 (0.08–5.41) | 2.04 (1.35–3.11)& | 1.02 (0.74–1.41) |
| Psychotic disorders | 0.97 (0.80–1.17) | 0.76 (0.59–0.97) | 0.72 (0.45–1.17) | 2.06 (0.43–9.85) | 1.25 (0.58–2.71) | 0.96 (0.51–1.80) | 2.69 (0.43–16.81) | 1.46 (0.64–3.34) | 1.08 (0.56–2.09) |
| Diabetes | 1.11 (1.07–1.14)& | 1.07 (1.03–1.12) | 1.04 (0.95–1.13) | 1.18 (0.71–1.96) | 1.48 (1.24–1.76)& | 1.21 (1.08–1.36)& | 1.21 (0.72–2.02) | 1.39 (1.15–1.68)& | 1.17 (1.03–1.32) |
| Dyslipidemia | 1.13 (1.10–1.16)& | 1.10 (1.06–1.14)& | 1.13 (1.06–1.21)& | 1.09 (0.69–1.74) | 1.28 (1.11–1.47)& | 1.21 (1.11–1.33)& | 1.21 (0.75–1.96) | 1.21 (1.05–1.40) | 1.16 (1.05–1.27) |
| Thyroid | 0.93 (0.75–1.14) | 1.08 (0.86–1.35) | 1.23 (0.82–1.87) | – | 0.46 (0.11–1.87) | 0.90 (0.45–1.81) | – | 0.45 (0.11–1.90) | 0.92 (0.45–1.91) |
| Lupus | 0.99 (0.68–1.45) | 0.93 (0.66–1.33) | 1.28 (0.54–3.01) | – | 3.46 (1.25–9.60) | 2.45 (1.25–4.80) | – | 2.55 (0.81–8.10) | 2.39 (1.17–4.88) |
| Connective tissue disorders | 1.24 (1.05–1.47) | 0.99 (0.81–1.19) | 1.01 (0.70–1.45) | 2.13 (0.24–18.80) | 1.05 (0.43–2.54) | 1.24 (0.75–2.03) | 0.61 (0.06–5.87) | 1.06 (0.42–2.65) | 1.32 (0.79–2.23) |
| Vasculitis | 0.61 (0.39–0.98) | 0.89 (0.52–1.53) | 0.62 (0.19–1.98) | – | 0.98 (0.14–7.10) | 0.98 (0.24–4.00) | – | 1.54 (0.18–12.86) | 0.98 (0.23–4.10) |
| Spondyloarthropathies | 0.93 (0.66–1.31) | 0.93 (0.62–1.41) | 1.23 (0.56–2.71) | – | 1.37 (0.32–5.79) | 0.61 (0.15–2.47) | 0.00 (0.00–0.00)& | 1.55 (0.34–6.98) | 0.54 (0.13–2.26) |
| Skin disorders | 1.18 (1.00–1.40) | 1.09 (0.88–1.35) | 1.24 (0.82–1.87) | 1.40 (0.16–12.26) | 1.11 (0.49–2.52) | 1.05 (0.59–1.87) | 2.16 (0.22–21.28) | 1.04 (0.44–2.46) | 0.93 (0.52–1.68) |
| COPD and Bronchiectasis | 1.03 (0.91–1.16) | 0.91 (0.78–1.06) | 0.99 (0.76–1.28) | 2.96 (0.85–10.31) | 1.23 (0.72–2.10) | 1.98 (1.51–2.61)& | 3.86 (0.80–18.62) | 1.75 (0.99–3.08) | 2.07 (1.54–2.79)& |
| Asthma | 1.23 (1.14–1.34)& | 1.19 (1.08–1.31)& | 1.57 (1.31–1.87)& | 2.37 (0.95–5.91) | 2.01 (1.53–2.64)& | 1.65 (1.37–2.00)& | 1.30 (0.53–3.17) | 2.02 (1.50–2.72)& | 1.57 (1.29–1.93)& |
| Pulmonary fibrosis | 1.40 (0.99–1.97) | 1.00 (0.65–1.52) | 1.15 (0.54–2.48) | – | 2.52 (0.91–7.01) | 1.11 (0.35–3.48) | – | 2.87 (0.87–9.46) | 0.98 (0.30–3.18) |
| Acute kidney disease | 1.10 (1.02–1.18) | 1.07 (0.98–1.18) | 1.03 (0.87–1.21) | 0.90 (0.41–2.00) | 1.23 (0.90–1.69) | 1.17 (0.95–1.44) | 1.49 (0.64–3.50) | 1.39 (0.99–1.95) | 1.11 (0.89–1.38) |
| End stage kidney disease | 1.30 (1.08–1.56) | 0.92 (0.72–1.17) | 1.37 (0.89–2.09) | 2.41 (0.50–11.63) | 1.92 (1.06–3.48) | 1.60 (1.04–2.45) | 2.29 (0.40–13.16) | 1.94 (1.00–3.73) | 1.85 (1.17–2.92) |
| Gastritis | 1.15 (1.10–1.19)& | 1.24 (1.18–1.30)& | 1.27 (1.16–1.38)& | 1.22 (1.11–1.34)& | 1.37 (1.13–1.67)& | 1.28 (1.13–1.44)& | 0.86 (0.40–1.82) | 1.37 (1.11–1.68) | 1.17 (1.03–1.33) |
| Irritable bowel syndrome | 1.25 (1.03–1.50) | 1.27 (1.00–1.60) | 1.78 (1.14–2.79) | 1.42 (0.90–2.22) | 2.09 (0.92–4.75) | 1.07 (0.53–2.16) | 3.53 (0.32–39.39) | 2.33 (0.95–5.72) | 0.95 (0.46–1.94) |
| Biliary tract disease | 1.16 (0.99–1.37) | 1.00 (0.81–1.24) | 1.28 (0.89–1.84) | 0.73 (0.47–1.14) | 0.60 (0.19–1.87) | 0.76 (0.42–1.38) | 0.29 (0.04–2.29) | 0.47 (0.15–1.53) | 0.78 (0.42–1.44) |
| Non–infectious hepatitis and cirrhosis | 1.60 (0.97–2.64) | 1.37 (0.81–2.30) | 1.07 (0.38–2.98) | 0.62 (0.15–2.60) | – | 0.71 (0.10–5.10) | – | – | 0.46 (0.06–3.39) |
| Inflammatory bowel disease | 1.30 (1.14–1.50)& | 1.27 (1.09–1.48) | 1.31 (0.99–1.75) | 1.59 (1.23–2.06)& | 1.83 (1.08–3.10) | 1.53 (1.08–2.17) | 0.25 (0.03–2.02) | 2.33 (1.33–4.11) | 1.72 (1.20–2.45) |
Numbers in parentheses represent 95% Cis
(A)T0 + 30 to T0 + 300 contrasting population-based controls versus individuals with one infection
(B)T1 + 30 to T1 + 300 contrasting individuals who have 1 COVID-19 infection versus individuals with two infections
(C)T1 + 30 to T1 + 300 contrasting population-based controls versus individuals with two infections for periods of BA.1/2, 4./5 or XBB predominance. A HR > 1 represents higher risk of that pre-specified sequelae in the follow-up period for reinfected individuals relative to infected individuals. HRs are estimated using Cox proportional hazards models with overlapping weights
*p.val < 0.05 (i.e. the Bonferroni corrected P-value, given that there is 1 tested outcome for all sequelae)
^p.val < 0.05/8 (i.e. the Bonferroni corrected P-value, given that there are 8 tested composite outcomes)
&p.val < 0.05/34 (i.e. the Bonferroni corrected P-value, given that there are 34 tested individual outcomes)
Subgroup analyses
Younger individuals aged less than 65 had generally higher risk of composite and individual sequelae after reinfection vs. older individuals. Females and non-Chinese also had significantly higher risks of sequelae compared to males and Chinese respectively. Individuals of low/middle socioeconomic status had higher risks of sequelae after reinfection versus individuals of high socioeconomic status.
Robustness checks
We explored the use of alternative weighting schemes. These methods reproduced the elevated risk sequelae in individuals with reinfections, but were not used for the reporting of main or subgroup analyses due to poorer balance in sociodemographic characteristics between groups. Analysis of negative-exposure controls demonstrated that our analytical plan did not generate spurious estimates of increased risk of lymphomas. Reconducting our analysis to only boosted individuals reproduced heightened risk estimates in reinfections, with near-identical risk estimates. Redefining reinfections as those which occurred at least 60 days after their first infection did not alter risks estimates. All subgroup analyses and robustness checks are reported in the Supplementary materials 1, 3, and 4.
Discussion
In this population-wide study during Delta and Omicron-predominant transmission, reinfection was associated with a significantly increased risk (21%) of post-acute sequelae across multiple organ systems when compared against individuals with no infections. Increased risk was observed even after vaccine breakthrough re-infection in boosted individuals. An excess number of around 5 diagnoses of new-incident conditions per 1000 is expected in reinfections and 3 diagnoses of new-incident conditions per 1000 is expected in 1st infections. This translates to around 295 and 3668 new diagnoses attributable to reinfections and 1st infections, respectively. These findings highlight the prolonged burden that SARS-CoV-2 infection may place on health systems and the importance of preventing reinfection by SARS-CoV-2.
These findings are congruent with prior population-based cohort studies in the pre-Omicron era that found elevated risks of post-acute sequelae following SARS-CoV-2 re-infection [1, 2]. In a cohort of US veterans infected in the pre-Omicron era, re-infected individuals had up to 110% higher risk of any sequelae up to 6 months following SARS-CoV-2 infection, compared to those with no re-infection [1]. Similarly, a cohort study from Spain and the UK that compared first COVID-19 cases and re-infections showed that the risk of post-acute COVID-19 symptoms was increased by ~ 50% after re-infection [2]. More modest estimates of post-acute sequelae following re-infection during Omicron-predominant infection may reflect the milder nature of initial infection attributed to Omicron [27]. Across organ systems, highest risk of post-acute sequelae following re-infection was reported in the respiratory system, with 51% increase in the risk of any post-acute respiratory sequelae following re-infection. Previous studies in the local population found much lower risks of post-acute sequelae outside of the respiratory system following initial Omicron infection [17, 18]. Yet, despite prior reports of increased risk of auto-immune disorders following SARS-CoV-2 infection [28], we found no significant association between either 1st or 2nd infections on subsequent risk of new-incident autoimmune disorders. Differences in effect estimates may reflect variant-specific differences in risks, vaccination profile, or the relatively smaller study sample available for analysis locally to capture new-incident autoimmune conditions. Strategies to minimise re-infection risk may still be necessary in these at-risk groups during COVID-19 endemicity [29].
Significant heterogeneity has been reported in literature on risk of post-acute sequelae following COVID-19 re-infection. While a previous meta-analysis demonstrated presence of post-COVID symptoms in 30% of patients 2 years after COVID-19, there were high heterogeneities across study populations [30]. In a German cohort that spanned pre-Omicron and Omicron eras, individuals who did not self-report symptoms congruent with post-COVID-19-condition after their first infection had 50% reduced risk of developing post-COVID-19-condition after re-infection [3]. Similarly, an Omicron UK cohort found an adjusted-odds-ratio of 0.72 for self-reported post-acute sequelae of COVID-19 after re-infection compared to a first infection [5]. In an electronic-health-record (EHR)-based US study that spanned pre-Omicron and Omicron eras, provider-diagnosed post-acute sequelae of COVID-19 occurred more frequently following initial infection than reinfection, across all variant epochs [6]. This discordance could stem from differences in post-acute sequelae definitions [1, 2].
Elevated risk of post-acute sequelae following re-infection was observed even when restricted to boosted individuals. In the US Veterans’ Affairs cohort, reinfection with pre-Omicron variants was associated with higher risk of any sequelae in different organ systems regardless of vaccination status [1]. Similarly, in a large UK primary care cohort, no evidence was observed between reinfection with the Omicron variant and risk of new-onset post-acute sequelae by vaccination status [5]. Our results demonstrate that increased risk of post-acute sequelae following re-infection (versus no re-infection) was evident even in fully boosted individuals, suggesting that hybrid immunity from prior infection and vaccine-induced immunity does not fully attenuate the risk of post-acute sequelae following SARS-CoV-2 re-infection. Tailored strategies to prevent re-infection, regardless of prior infection history and vaccination status, such as updated vaccines, may thus remain relevant during COVID-19 endemicity notwithstanding the widespread availability of booster vaccination, especially in at-risk groups.
The strengths of our study include: usage of comprehensive nationwide registries to classify SARS-CoV-2 infection and vaccination status, reducing potential misclassification. New-incident diagnoses were identified via comprehensive healthcare claims database with national-level coverage, minimizing selection bias caused by loss-to-follow-up. There are several limitations. (1) During the Omicron era, although ≥ 90% of COVID-19 cases presented in a timely manner to primary care (within 48 h of symptom onset) for diagnosis and treatment [19], given the context of widely available free SARS-CoV-2 testing, mild or asymptomatic infections might not have been picked up if testing was not performed [31]; underestimation of re-infection rates would have biased estimates conservatively. (2) Results from home-based testing were not included in our test registry, which may have resulted in potential misclassification of re-infection; however, public health messaging during the period encouraged individuals to present to healthcare providers for confirmatory diagnosis [19]. (3) Administrative claims data were utilised to capture diagnoses; however, healthcare claims data might under-report estimates of milder sequelae not affecting reimbursement, particularly in individuals with mild SARS-CoV-2 infection diagnosed in primary care [32]. (4) Vaccination/boosting may drive differences in health-seeking behaviour, which impact post-acute sequelae diagnosis, although prior healthcare utilisation was controlled for. (5) We did not account for the usage of COVID-19 treatments, which can potentially affect severity of initial disease and post-acute sequelae risk; however, uptake of oral antivirals was low (< 1%) [21], and receipt of early nirmatrelvir/ritonavir did not reduce risk of post-acute sequelae following SARS-CoV-2 infection in our local population [33]. (6) Individuals with a greater predisposition to post-acute sequelae may already have experienced sequelae following a first infection, and therefore were not in the sample eligible to experience new-onset sequelae following a second infection. This would bias risk estimates of post-acute sequelae in a conservative direction. (7) Subgroup analyses were exploratory in nature and relatively lower number of samples available for analysis in some subgroups may have led to differential estimates of risk by chance.
Conclusions
Reinfection was associated with increased risk of diagnoses of any post-acute sequelae and sequelae across multiple organ systems, even in boosted individuals in Singapore. The findings highlight the continued significance of re-infection in influencing the risk of post-acute sequelae. Our results emphasise the need for preventative strategies against SARS-CoV-2 infection, such as updated vaccines that have better effectiveness against infection.
Supplementary Information
Supplementary material 1: Results for subgroup analyses across age, sex, vaccination, comorbidity status and ethnicities.
Supplementary material 2: Description and list of ICD codes used for analysis.
Supplementary material 3: Results for sensitivity analyses across different weighting schemes.
Supplementary material 4: Results for analyses across different periods of COVID-19 variant predominance.
Acknowledgements
Not applicable.
Authors’ contributions
LEW, JTL, LJP contributed to literature search and writing of the manuscript. JTL, JYJT, CC, BO, DCL and KBT contributed to critical review and editing of the manuscript. DCL and KBT provided supervision. KBT, LEW, JTL contributed to study design. JYJT contributed to data collection, and LEW, JTL and JYJT contributed to data analysis. LEW, JTL and JYJT directly accessed and verified the underlying data reported in the manuscript. All authors read and approved the final manuscript.
Funding
LEW is supported by the National Medical Research Council, Singapore, through the Clinician Scientist New Investigator Grant. JTL is supported by the Ministry of Education (MOE), Singapore Start-up Grant.
Data availability
The databases with individual-level information used for this study are not publicly available due to personal data protection. Deidentified data can be made available for research, subject to approval by the Ministry of Health of Singapore. All inquiries should be sent to the corresponding author.
Declarations
Ethics approval and consent to participate
This study was performed as part of national public health research under the Infectious Diseases Act, Singapore [34]; separate ethics review by an Institutional Review Board was not required. The research conformed to the principles of the Helsinki Declaration.
Consent to publication
Not applicable.
Competing interests
JTL is an editorial board member of BMC Global and Public Health. The remaining authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Jue Tao Lim, Liang En Wee and Janice Yu Jin Tan contributed equally to this work.
Contributor Information
Jue Tao Lim, Email: juetao.lim@ntu.edu.sg.
Liang En Wee, Email: ian.wee.l.e@singhealth.com.sg.
References
- 1.Bowe B, Xie Y, Al-Aly Z. Acute and postacute sequelae associated with SARS-CoV-2 reinfection. Nat Med. 2022;28(11):2398–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kostka K, Roel E, Trinh NTH, Mercadé-Besora N, Delmestri A, Mateu L, et al. The burden of post-acute COVID-19 symptoms in a multinational network cohort analysis. Nat Commun. 2023;14(1):7449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mikolajczyk R, Diexer S, Klee B, Pfrommer L, Purschke O, Fricke J, et al. Likelihood of post-COVID condition in people with hybrid immunity; data from the German National Cohort (NAKO). J Infect. 2024;89(2). Available from: https://www.journalofinfection.com/article/S0163-4453(24)00140-3/fulltext. Cited 18 Sept 2025. [DOI] [PubMed]
- 4.Peghin M, Martino MD, Palese A, Chiappinotto S, Fonda F, Gerussi V, et al. Post-COVID-19 syndrome 2 years after the first wave: the role of humoral response, vaccination and reinfection. Open Forum Infect Dis. 2023;10(7):ofad364. [DOI] [PMC free article] [PubMed]
- 5.Bosworth ML, Shenhuy B, Walker AS, Nafilyan V, Alwan NA, O’Hara ME, et al. Risk of new-onset long COVID following reinfection with severe acute respiratory syndrome coronavirus 2: a community-based cohort study. Open Forum Infect Dis. 2023;10(11):ofad493. [DOI] [PMC free article] [PubMed]
- 6.Hadley E, Yoo YJ, Patel S, Zhou A, Laraway B, Wong R, et al. Insights from an N3C RECOVER EHR-based cohort study characterizing SARS-CoV-2 reinfections and Long COVID. Commun Med. 2024;4(1):129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Huh K, Kim YE, Bae GH, Moon JY, Kang JM, Lee J, et al. Vaccination and the risk of post-acute sequelae after COVID-19 in the Omicron-predominant period. Clin Microbiol Infect. 2024;30(5):666–73. [DOI] [PubMed] [Google Scholar]
- 8.Lundberg-Morris L, Leach S, Xu Y, Martikainen J, Santosa A, Gisslén M, et al. Covid-19 vaccine effectiveness against post-covid-19 condition among 589 722 individuals in Sweden: population based cohort study. BMJ. 2023;22(383):e076990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tan CY, Chiew CJ, Pang D, Lee VJ, Ong B, Lye DC, et al. Vaccine effectiveness against Delta, Omicron BA.1, and BA.2 in a highly vaccinated Asian setting: a test-negative design study. Clin Microbiol Infect. 2023;29(1):101–6. [DOI] [PMC free article] [PubMed]
- 10.Wee LE, Pang D, Chiew C, Tan J, Lee V, Ong B, et al. Long-term real-world protection afforded by third mRNA doses against symptomatic severe acute respiratory syndrome coronavirus 2 infections, coronavirus disease 19-related emergency attendances and hospitalizations amongst older singaporeans during an omicron XBB wave. Clin Infect Dis. 2023;77(8):1111–9. [DOI] [PubMed] [Google Scholar]
- 11.Chong C, Wee LE, Jin X, Zhang M, Malek MIA, Ong B, et al. Risks of SARS-CoV-2 JN.1 infection and COVID-19–associated emergency department visits/hospitalizations following updated boosters and prior infection: a population-based cohort study. Clin Infect Dis. 2024;79(5):1190–6. [DOI] [PubMed]
- 12.Wee LE, Yap AJW, Dickens B, Tan S, Ong B, Lye DC, et al. Access to COVID-19 vaccination by socio-economic status in older Singaporean adults: a population-based cohort study. Public Health. 2024;1(233):38–44. [DOI] [PubMed] [Google Scholar]
- 13.COVID-19 Vaccinations Recommendations and Uptake. Ministry of Health. Available from: https://www.moh.gov.sg/newsroom/covid-19-vaccinations-recommendations-and-uptake/. Cited 6 May 2025
- 14.Tan CC, Lam CSP, Matchar DB, Zee YK, Wong JEL. Singapore’s health-care system: key features, challenges, and shifts. The Lancet. 2021;398(10305):1091–104. [DOI] [PubMed]
- 15.Lim JT, Liang En W, Tay AT, Pang D, Chiew CJ, Ong B, et al. Long-term cardiovascular, cerebrovascular, and other thrombotic complications in COVID-19 survivors: a retrospective cohort study. Clin Infect Dis. 2024;78(1):70–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wee LE, Lim JT, Tay AT, Pang D, Dickens B, Chiew CJ, et al. Long-term neuropsychiatric sequelae of Delta versus Omicron SARS-CoV-2 infection. Clin Microbiol Infect. 2024;30(4):531–9. [DOI] [PubMed] [Google Scholar]
- 17.Wee LE, Lim JT, Tay AT, Pang D, Chiew CJ, Chia YW, et al. Long-term cardiovascular, cerebrovascular, and thrombotic complications after SARS-CoV-2-Omicron infection: a retrospective cohort study. Clin Microbiol Infect. 2024;30(10):1319–26. [DOI] [PubMed] [Google Scholar]
- 18.Wee LE, Lim JT, Tay AT, Chiew CJ, Ong B, Lye DCB, et al. Autoimmune sequelae after delta or omicron variant SARS-CoV-2 infection in a highly vaccinated cohort. JAMA Netw Open. 2024;7(8):e2430983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tan RY, Wong B, Lim R, Lee CL, Tan J, Tan KB, et al. Factors associated with delayed diagnosis of symptomatic adult COVID-19 cases presenting to primary care: a population-wide study during transition from Delta to Omicron BA.1 in Singapore. The Lancet Regional Health - Western Pacific. 2023;41:100919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Infectious Diseases Act 1976 - Singapore Statutes Online. Available from: https://sso.agc.gov.sg/Act/IDA1976. Cited 30 Oct 2025.
- 21.Tan RY, Dickens B, Tan J, Lee CL, Wong B, Lim R, et al. Oral antiviral utilisation among older adults with COVID-19 in primary care: A population-wide study during successive Omicron waves in Singapore. Ann Acad Med Singap. 2024;53(2):117–20. [DOI] [PubMed] [Google Scholar]
- 22.ICD-10 Version:2019. Available from: https://icd.who.int/browse10/2019/en. Cited 30 Oct 2025.
- 23.Glasheen WP, Cordier T, Gumpina R, Haugh G, Davis J, Renda A. Charlson comorbidity index: ICD-9 update and ICD-10 translation. Am Health Drug Benefits. 2019;12(4):188–97. [PMC free article] [PubMed] [Google Scholar]
- 24.Lin DY, Wei LJ. The Robust Inference for the Cox Proportional Hazards Model. J Am Stat Assoc. 1989;84(408):1074–8. [Google Scholar]
- 25.Bland JM, Altman DG. Survival probabilities (the Kaplan-Meier method). BMJ. 1998;317(7172):1572–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Al-Aly Z, Xie Y, Bowe B. High-dimensional characterization of post-acute sequelae of COVID-19. Nature. 2021;594(7862):259–64. [DOI] [PubMed] [Google Scholar]
- 27.Yu W, Guo Y, Zhang S, Kong Y, Shen Z, Zhang J. Proportion of asymptomatic infection and nonsevere disease caused by SARS-CoV-2 Omicron variant: a systematic review and analysis. J Med Virol. 2022;94(12):5790–801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chang R, Chen TYT, Wang SI, Hung YM, Chen HY, Wei CCJ. Risk of autoimmune diseases in patients with COVID-19: a retrospective cohort study. eClinicalMedicine. 2023;56. Available from: https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(22)00512-0/fulltext. Cited 18 Sept 2025 [DOI] [PMC free article] [PubMed]
- 29.Wee LE, Tan JYJ, Chiew CJ, Abisheganaden JA, Chotirmall SH, Lye DCB, et al. A nationwide cohort study of delta and omicron SARS-CoV-2 outcomes in vaccinated individuals with chronic lung disease. Chest. 2024;166(4):685–96. [DOI] [PubMed] [Google Scholar]
- 30.Fernandez-de-las-Peñas C, Notarte KI, Macasaet R, Velasco JV, Catahay JA, Ver AT, et al. Persistence of post-COVID symptoms in the general population two years after SARS-CoV-2 infection: a systematic review and meta-analysis. J Infect. 2024;88(2):77–88. [DOI] [PubMed] [Google Scholar]
- 31.Hoeggerl AD, Nunhofer V, Weidner L, Lauth W, Zimmermann G, Badstuber N, et al. Dissecting the dynamics of SARS-CoV-2 reinfections in blood donors with pauci- or asymptomatic COVID-19 disease course at initial infection. Infect Dis. 2024;56(11):954–64. [DOI] [PubMed] [Google Scholar]
- 32.Jeffrey K, Woolford L, Maini R, Basetti S, Batchelor A, Weatherill D, et al. Prevalence and risk factors for long COVID among adults in Scotland using electronic health records: a national, retrospective, observational cohort study. eClinMed. 2024;71:102590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wee LE, Lim JT, Tay AT, Chiew CJ, Young BE, Wong B, et al. Nirmatrelvir/ritonavir treatment and risk for postacute sequelae of COVID-19 in older Singaporeans. Clin Microbiol Infect. 2025;31(1):93–100. [DOI] [PubMed] [Google Scholar]
- 34.Infectious Diseases Act 1976 - Singapore Statutes Online. Available from: https://sso.agc.gov.sg/Act/IDA1976. Cited 18 Sept 2025.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material 1: Results for subgroup analyses across age, sex, vaccination, comorbidity status and ethnicities.
Supplementary material 2: Description and list of ICD codes used for analysis.
Supplementary material 3: Results for sensitivity analyses across different weighting schemes.
Supplementary material 4: Results for analyses across different periods of COVID-19 variant predominance.
Data Availability Statement
The databases with individual-level information used for this study are not publicly available due to personal data protection. Deidentified data can be made available for research, subject to approval by the Ministry of Health of Singapore. All inquiries should be sent to the corresponding author.