Asia-Pacific Study Finds Most Long COVID Risks Subside After One Year

A large multi-regional cohort study examining healthcare databases across four Asia–Pacific countries and territories has revealed that while SARS-CoV-2 infection significantly elevates the short-term hazard of mortality and multi-organ clinical sequelae, the vast majority of these health risks substantially diminish or resolve entirely beyond the first three months. By analyzing longitudinal records from New Zealand, South Korea, Taiwan, and Hong Kong, the investigative network tracked individuals over short-term (0–90 days), medium-term (91–365 days), and long-term (>365 days) intervals. The findings demonstrate that although elevated vulnerabilities across cardiovascular, renal, metabolic, and respiratory systems occur during acute recovery, most complications gradually subside to levels comparable to uninfected controls within a year, offering critical baseline evidence for an underrepresented geographic population.

Published evidence prior to this research was largely derived from Western cohorts or single-database evaluations, which introduced potential limitations when generalizing outcomes to populations across the Asia–Pacific region. Variations in demographic composition, circulating viral variants, regional pandemic trajectories, and specific vaccination campaigns necessitated a targeted assessment across diverse healthcare delivery frameworks. To resolve these knowledge gaps, the researchers constructed a collaborative observational architecture pooling electronic medical records and health insurance data from the National Health Insurance System–National Health Insurance Database in South Korea, the Hospital Authority system in Hong Kong, the National Health Insurance Research Database in Taiwan, and the Medicines Intelligence Data Platform in New Zealand. Although data from Australia’s TREVI Coronavirus database was initially considered during the overall study design, it was ultimately excluded due to concerns regarding reporting disparities between sites.

The study cohort comprised 346,882 individuals aged 18 years or older with documented SARS-CoV-2 infection and 1,339,095 unexposed controls identified across the regional registries. Study periods varied according to jurisdictional data availability, spanning October 2020 through June 2021 in South Korea, January 2020 through August 2023 in Hong Kong, June 2021 through September 2021 in Taiwan, and January 2021 through December 2021 in New Zealand. Unexposed controls were assigned a pseudo-index date corresponding to the median inclusion date of positive cases within each respective database, synchronizing follow-up baselines. Each infected participant was matched to up to four uninfected comparators using propensity score matching within a caliper of 0.1 standard deviations of the propensity score logit, balancing variables including sex, age categories, vaccination doses administered within the preceding 180 days, and pre-existing comorbidities such as stroke, myocardial infarction, heart failure, hypertension, and diabetes.

Primary analyses evaluated incidence rates per 1,000 person-years alongside hazard ratios generated through Cox proportional-hazard regressions across prespecified post-infection intervals. Because statistical heterogeneity across regional datasets was considerable, with I-squared values exceeding 50 percent for the majority of monitored clinical outcomes, random-effects meta-analyses were deployed to synthesize site-specific effect estimates. The monitored clinical endpoints encompassed acute kidney injury, renal failure, interstitial lung disease, acute respiratory distress syndrome, major cardiovascular disease, deep vein thrombosis, atrial fibrillation, coronary artery disease, anxiety, dementia, liver injury, uncomplicated diabetes mellitus, and all-cause mortality. Participants who possessed a documented medical history of a given condition on or prior to the index date were strictly excluded from the corresponding disease-specific incidence evaluations.

During the short-term window of 0 to 90 days following acute infection, patients with COVID-19 experienced a markedly elevated hazard of all-cause mortality, reflected by a pooled random-effects hazard ratio of 7.57 with a 95 percent confidence interval spanning 6.30 to 9.11. Acute respiratory complications exhibited the highest relative elevations during this primary stage, with acute respiratory distress syndrome reaching a hazard ratio of 7.84 (95 percent confidence interval: 4.45 to 13.81) and interstitial lung disease demonstrating an estimated hazard ratio of 7.25 (95 percent confidence interval: 2.24 to 23.48). Significant multi-organ extrapulmonary risks were also documented during this initial window, including acute kidney disease with a hazard ratio of 2.96, deep vein thrombosis at 2.51, heart failure at 2.20, coronary artery disease at 2.18, anxiety at 1.91, and atrial fibrillation at 2.16.

Transitioning into the medium-term interval extending from 91 to 365 days, the comparative excess risks across most organ categories declined sharply. The pooled hazard ratio for all-cause mortality fell to 1.10 with a 95 percent confidence interval of 0.88 to 1.38, indicating that the initial mortality disparity observed during acute recovery did not persist uniformly across the wider cohort. However, persistent vulnerabilities remained identifiable in specific organ domains. The meta-analysis identified a continued elevated risk for acute respiratory distress syndrome between 91 and 365 days, retaining a statistically significant hazard ratio of 1.38 (95 percent confidence interval: 1.05 to 1.81). Furthermore, region-specific evaluations identified lingering medium-term risks for particular outcomes, including elevated hazard ratios for interstitial lung disease in South Korea and Taiwan, myocardial infarction in Hong Kong and Taiwan, and dementia in Hong Kong and Taiwan.

In the long-term observation period extending past 365 days, evaluated primarily within cohorts possessing long-term tracking data, overall hazards for most systemic sequelae and all-cause mortality attenuated further toward control levels. The pooled hazard ratio for all-cause mortality across long-term follow-up settled at 0.91 with a 95 percent confidence interval of 0.41 to 2.02. Deep vein thrombosis yielded a long-term hazard ratio of 1.14 (95 percent confidence interval: 0.81 to 1.60), coronary artery disease recorded 1.05 (95 percent confidence interval: 0.62 to 1.77), and major cardiovascular disease demonstrated an overall pooled hazard ratio of 1.07 (95 percent confidence interval: 0.60 to 1.89). Despite this general stabilization, site-specific analyses revealed enduring long-term elevations in select populations; notably, Hospital Authority records in Hong Kong demonstrated persistently elevated risks beyond one year for major cardiovascular disease at 1.41, acute respiratory distress syndrome at 1.34, and all-cause mortality at 1.36.

Stratified subgroup evaluations identified meaningful differences in disease vulnerability based on demographic baseline profiles. Participants aged older than 65 years exhibited considerably greater susceptibility to post-acute sequelae than patients aged 65 years or younger. Over the short-term window, all-cause mortality risks were markedly amplified among older adults across South Korea (hazard ratio 9.69 versus 6.09 in younger cohorts), Hong Kong (hazard ratio 7.93 versus 3.60), and Taiwan (hazard ratio 10.18 versus 6.77). Conversely, comparative analyses between male and female participants demonstrated broadly similar risk patterns. Sensitivity analyses further substantiated the core findings: censoring participants at the time of reinfection produced lower hazard ratios across sequelae compared to the primary analytical model, while adjustments for multiple testing using Bonferroni correction with an adjusted significance threshold of 0.003 maintained consistency with the unadjusted results.

Pathophysiological mechanisms described in the source literature contextualize these empirical trajectories. The pronounced early risk of acute respiratory distress syndrome reflects viral targeting of respiratory epithelium, inciting alveolar damage, localized thromboinflammation, and severe systemic immune dysregulation. Persistent medium-term pulmonary sequelae likely stem from post-inflammatory pulmonary fibrosis, characterized by documented vascular remodeling and ongoing alveolar epithelial injury that can take months to resolve. Furthermore, the exacerbation of cardiovascular complications is linked to direct endothelial cell involvement and sustained cytokine activation. The reduction in excess risk observed after six months suggests ongoing physiological recovery, gradual dampening of systemic hyperinflammation, and tissue repair in surviving patients, though lingering damage may sustain risks in frail or multimorbid individuals.

The study authors documented several analytical limitations inherent to multi-database observational research. Surveillance bias could arise if individuals with known COVID-19 diagnoses engaged in more frequent medical consultations, though the severity of the primary endpoints mitigated the risk of detection artifacts. Conversely, healthy survivor bias may have underestimated certain risks, and milder clinical manifestations of long COVID such as fatigue, localized pain, and brain fog could not be systematically captured via administrative coding systems. Variations in diagnostic practices, claims billing conventions such as those within the South Korean NHIS-NHID, differing circulating viral strains across recruitment windows, unmeasured socioeconomic factors, and the absence of follow-up exceeding 365 days in Taiwan and New Zealand bounded the scope of long-term inferences, underscoring the need for ongoing surveillance among high-risk individuals.

Subject of Research: Post-acute clinical sequelae and mortality following SARS-CoV-2 infection across Asia-Pacific populations

Article Title: Short-, medium- and long-term sequelae following COVID-19 infection: a network cohort study across multiple Asia–Pacific regions

Article References: Zhang, R., Lam, I. C. H., Chai, Y., Liu, X., Tan, G. S. Q., Ilomäki, J., Bell, J. S., Tsai, D. H.-T., Lai, E. C.-C., Shen, C.-Y., Cheng, M. C.-Y., Harrison, J., Chan, A. H. Y., Hong, B., Shin, J.-Y., Abtahi, S., Klungel, O. H., Chui, C. S. L., Li, X., … Wan, E. Y. F. (2026). Short-, medium- and long-term sequelae following COVID-19 infection: a network cohort study across multiple Asia–Pacific regions. The Lancet Regional Health – Western Pacific, 76, Article 102014. https://doi.org/10.1016/j.lanwpc.2026.102014

Image Credits: AI Generated

DOI: 10.1016/j.lanwpc.2026.102014

Keywords: COVID-19, SARS-CoV-2, Post-Acute Sequelae of SARS-CoV-2, Long COVID, Cohort Study, Asia-Pacific Healthcare, All-Cause Mortality, Acute Respiratory Distress Syndrome, Interstitial Lung Disease, Cardiovascular Disease, Propensity Score Matching, Epidemiology, Multi-Organ Complications, Reinfection Risk, Electronic Health Records, Public Health Surveillance