Long COVID symptom clusters that could reshape patient care

By Hugo Francisco de SouzaReviewed by Lauren HardakerJan 12 2026

By analyzing millions of patients worldwide, researchers reveal consistent patterns in Long COVID symptoms, offering clinicians a clearer roadmap for diagnosis, referral, and tailored treatment.

Study: Identifying subtypes of Long COVID: a systematic review. Image credit: Dragana Gordic/Shutterstock.com

In a recent systematic review published in eClinicalMedicine, researchers synthesized data from 64 studies encompassing 2.43 million participants across 20 countries to synthesize existing approaches to Long COVID subtyping and propose a structured, subtype-oriented management framework.

The review summarized how included studies grouped patients or symptoms, leveraging study-level meta-analyses focused primarily on organ systems, based on symptom groupings to examine patterns in symptom clustering.

Review analyses identified four primary methods for categorizing patients: by symptom co-occurrence, affected organ systems, disease severity, and clinical indicators. The analysis highlights fatigue as a central, recurring symptom, affecting 37 % of patients and frequently overlapping with neurological and respiratory issues, alongside recurrent olfactory and gustatory dysfunction.

The study proposes a structured management framework to guide clinicians toward more personalized, precision-based treatments for this heterogeneous condition.

Extreme symptom diversity complicates Long COVID diagnosis

Since the onset of the COVID-19 pandemic, Long COVID, defined by the World Health Organization (WHO) as COVID-19 symptoms persisting for at least two months after initial infection, has challenged medical professionals due to its extreme heterogeneity. A growing body of evidence reports that patients present with a diverse spectrum of symptoms, ranging from brain fog and crushing fatigue to heart palpitations and gastrointestinal distress. While individual studies have attempted to group these symptoms, a unified “map” of Long COVID remains lacking.

Without a clear explanation of COVID-19 subtypes, patients are often subjected to a "one-size-fits-all" approach that fails to address their specific pathology. The identification and description of shared biological mechanisms that support targeted therapy is hence imperative, especially given the gradual rise in Long COVID incidence worldwide.

PRISMA-guided review compiles global Long COVID clustering studies

The present systematic review aims to inform future clinical policy by consolidating global data on Long COVID incidence, specific demographics, and symptomatic frequency. The study adheres to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines.

Study data were obtained via a custom keyword search of several major online scientific repositories, including PubMed, Embase, Web of Science, and Google Scholar, for literature published up to October 1, 2025.

Identified publications were subjected to a strict title, abstract, and full-text screening process, ensuring that only those that explicitly categorized Long COVID patients or symptoms, using clustering algorithms, descriptive grouping, or expert-driven classification, were included in subsequent analyses.

The review employed a meta-analytical approach to estimate the prevalence of symptom clusters based on specific organ systems. Given the high inter-publication heterogeneity observed across sample size and methodologies employed, a random-effects model, which accounts for high variability between different study populations, was used to pool cluster frequencies reported across studies. Additionally, "pairwise co-occurrence analysis" was employed to visualize which symptoms tend to be grouped together at the study level, thereby creating a network map of the disease.

Fatigue-centered and organ-based clusters dominate Long COVID

Publication screening identified 64 high-quality studies, 47 cohort and 17 cross-sectional, that met the study inclusion criteria, comprising a total sample size of 2,430,177 participants across 20 countries. Notably, most of the sample population was comprised of participants from the United States and Europe.

Random-effects analyses identified four major approaches to classifying Long COVID, with symptom co-occurrence, 46.9 %, and organ system, 25.0 %, being the most prevalent. Specifically, prevalence analyses established the following distinct organ systems, based on symptom clusters among Long COVID patients:

1. Respiratory Cluster

This was the most dominant cluster, found in 47 % of patients (95 % CI: 29 %–65 %).

2. Neurological Cluster

Including cognitive impairment and sensory issues. This cluster was prevalent in 31 % of patients (95 % CI: 3 %–60 %).

3. Gastrointestinal Cluster

Including the entire spectrum of gastrointestinal afflictions from chronic indigestion to Irritable Bowel Syndrome (IBS). Found in 28 % of patients (95 % CI: 0 %–57 %).

4. Fatigue

Identified as a "core" symptom, fatigue formed its own cluster in 37 % of cases (95 % CI: 19 %–55 %).

Importantly, these clusters were not mutually exclusive patient subtypes but reflect symptom groupings reported within individual studies.

Co-occurrence analysis identified "fatigue" as a “central hub”, frequently manifesting alongside joint pain, cognitive dysfunction, and dyspnea, shortness of breath. These co-occurrence patterns reflect the frequency with which symptoms were clustered together across studies, rather than the prevalence of patient-level co-occurrence. Other strong "dyads", pairs, included anxiety appearing with depression (n = 10 studies) and loss of smell pairing with loss of taste (n = 10 studies).

An exploratory analysis investigating potential demographic influences suggested that sex, age, and factors such as socioeconomic background and comorbidities may influence symptom clustering patterns. For example, females demonstrated a significantly higher risk for neuropsychiatric symptoms and fatigue, whereas males were found to be more prone to respiratory symptoms.

Symptom-oriented framework offers clearer path to personalized care

The present systematic review and meta-analysis demonstrate, through comprehensive synthesis, that Long COVID is a multisystem condition that can be systematically described using recurring symptom clustering patterns across studies. Consequently, the study proposes a "symptom subtype-oriented management framework," where general clinicians first classify patients by symptom co-occurrence, map them to relevant organ systems, incorporate severity-based stratification when needed, and then refer them to specialists, for example, neurologists or pulmonologists, for further treatment.

While the heterogeneity of the original data limits the review, it provides a vital foundation for future precision medicine and personalized interventions. Future research integrating these clinical clusters with biological data, multi-omics, and other mechanistic approaches may help uncover the specific mechanisms driving each subtype, moving us closer to targeted cures.