Identifying subtypes of Long COVID: a systematic review
Bibliographic record
Abstract
Background: Long COVID, a persistent condition following SARS-CoV-2 infection, exhibits diverse symptoms across multiple organ systems. This study aims to summarize the existing clustering and classification approaches to support the management of Long COVID. Methods: Following PRISMA guidelines, we systematically searched PubMed, Embase, Web of Science, and Google Scholar from their inception to January 21, 2025, and updated the search on October 1, 2025, to identify studies that presented a way to categorize Long COVID patients or symptoms. Data extraction and quality assessment were conducted for eligible studies. We presented symptom co-occurrence networks, and performed meta-analysis to estimate the percentage of different organ system-based symptom clusters. In addition, we conducted an exploratory analysis of the determinants of different symptom clusters. The protocol was registered in OSF (https://doi.org/10.17605/OSF.IO/J483F). Findings: Forty-seven cohort studies and 17 cross-sectional studies categorizing Long COVID subtypes or symptoms were included, encompassing 2.43 million participants across 20 countries. The methodological quality of the cohort studies was on average high (mean Newcastle-Ottawa scale score: 7.5/9), and of the 17 cross-sectional studies moderate (mean Joanna Briggs Institute tool score: 0.61/1.00). Patients or symptoms were categorized either according to the co-occurrence of symptoms (n = 30 studies, 46.9%); by the affected organ system (n = 16, 25.0%); by severity stratification (n = 9, 14.1%); by clinical indicators (n = 3, 4.7%); or by using other ways of classification (n = 6, 9.4%). Among the 30 studies defining patient clusters by the co-occurrence of symptoms, fatigue was the most frequently used descriptor for a cluster, either alone or together with other symptoms (n = 15 studies). Pairwise co-occurrence analysis revealed some commonly used symptom dyads, including olfactory-gustatory dysfunction (n = 10 times), anxiety-depression (n = 10) and joint pain/swelling-muscle pain (n = 9). Fatigue was a recurrent core symptom, frequently co-occurring with joint pain/swelling (n = 9 times) or muscle pain (n = 7), cognitive symptoms (n = 7), and dyspnea (n = 7). Meta-analysis of the organ system-based subtypes showed that respiratory symptom cluster had the highest pooled percentage (47% [95% CI: 29%-65%]), followed by neurological (31% [95% CI: 3%-60%]) and gastrointestinal clusters (28% [95% CI: 0%-57%]). These percentages represent the proportion of Long COVID patients with each symptom cluster within the 16 included organ system-based subtyping studies, not population-level prevalence of Long COVID. Exploratory analysis indicated that symptom subtypes were influenced by factors such as sex, age, virus variant, and comorbidities. Interpretation: This review identified four major approaches for categorizing Long COVID patients and their symptoms. Symptom co-occurrence and organ system were the most commonly used subtypes used in categorization. Fatigue and olfactory-gustatory dysfunction emerged as recurrent core symptoms across multiple subtypes of Long COVID. Funding: This work was supported by the K. C. Wong Education Foundation, Hong Kong, the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (2024-I2M-ZD-011), the Beijing Nova Program (20240484523), the Elite Medical Professionals Project of China-Japan Friendship Hospital (NO. ZRJY2024-GG03), and the National High Level Hospital Clinical Research Funding.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.035 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".