International Journal of Infectious Diseases 134 (2023) 1–7

Contents lists available at ScienceDirect

International Journal of Infectious Diseases

journal homepage: www.elsevier.com/locate/ijid

Characterization of post-COVID syndromes by symptom cluster and

time period up to 12 months post-infection: A systematic review and
meta-analysis
Paul Kuodi 1,∗, Yanay Gorelik 1, Blessing Gausi 2, Tomer Bernstine 1, Michael Edelstein 1,3
1
Azrieli Faculty of Medicine, Bar-Ilan University, Safed, Israel
2
School of Public Health and Family Medicine, University of Cape Town, Cape Town, South Africa
3
Ziv Medical Centre, Safed, Israel

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: The objective of this study was to characterize post-COVID condition symptoms and symptom
Received 7 March 2023 clusters, their duration, and prevalence.
Revised 2 May 2023
Methods: We conducted a systematic review and random-effects meta-analysis of studies reporting post-
Accepted 2 May 2023
COVID-19 symptoms and clusters, respectively. We searched MEDLINE (via PubMed), Scopus, Web of Sci-
ence, Science Direct, Google Scholar, EBSCOhost, EMBASE, PsycINFO, Cochrane Library, and Mednar for
Keywords: literature reporting on the post-COVID condition up to August 2022.
SARS-CoV-2 Results: In the 76 included studies, we found that although most symptoms were reported less frequently
COVID-19 7-12 months after infection compared to earlier, over 20% of patients reported at least one post-COVID
Long COVID
condition-compatible symptom. In the seven studies reporting post-COVID symptom clusters, neurological
Post-COVID condition
clustering was consistently identified, followed by cardiorespiratory and systemic/inflammatory.
Conclusion: Post-COVID symptom clustering provides direction for research into the etiology, diagnosis,
and management of post-COVID conditions. Studies reporting post-COVID symptom clusters remain rare
due to the focus on individual symptom reporting. Studies on post-COVID symptom clusters should re-
place individual symptom reporting to accelerate our understanding of this emerging public health issue.
© 2023 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious
Diseases.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)

Introduction participants, other studies on non-hospitalized participants and yet

others report a mixture of hospitalized and non-hospitalized par-
By November 2022, the global cumulative confirmed cases of ticipants), and (iv) the different measurement tools used in each
SARS-CoV-2 infections and deaths from COVID-19 surpassed 600 study [5,6].
million and 6.5 million, respectively [1,2]. Among patients consid- Early definitions of post-COVID condition were sensitive rather
ered recovered from COVID-19, some patients continue to report than specific, owing to the recent emergence and lack of clarity
lingering symptoms with varying symptom patterns and character- around the condition. In October 2021, following a Delphi consen-
istics [3]. Post-COVID-19 condition is emerging as a major public sus, the World Health Organization (WHO) defined post-COVID-19
health problem following closely in the wake of the COVID-19 pan- condition as a “condition that occurs in individuals with a history
demic [4]. The reported prevalence of symptoms that characterize of probable or confirmed SARS-CoV-2 infection, usually 3 months
post-COVID condition differs across studies and is estimated at be- from the onset of COVID-19 with symptoms that last for at least
tween 10-30% [5]. Variation in post-COVID-19 prevalence is likely 2 months and cannot be explained by an alternative diagnosis.
due to: (i) a lack of uniformity in its case definition (ii) the differ- Symptoms may be new onset, following initial recovery from an
ences in reporting time by studies, (iii) the characteristics of partic- acute COVID-19 episode, or persist from the initial illness. Symp-
ipants included in each study (some studies report on hospitalized toms may also fluctuate or relapse over time” [7].
The definition remains broad, with a wide variety of symptoms
and timeframes potentially included. Despite the publication of
∗
Corresponding author. Tel: +972552481236. this case definition, there is still a lack of agreement among clini-
E-mail address: [email protected] (P. Kuodi). cians and researchers on what constitutes a post-COVID condition.

https://doi.org/10.1016/j.ijid.2023.05.003
1201-9712/© 2023 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Patients diagnosed with post-COVID condition present with a Cochrane Library, and Mednar. A hand-search of reference lists
wide variety of symptoms with the involvement of multiple body of included studies was conducted to source additional studies.
systems making the application of the proposed WHO definition The International Severe Acute Respiratory and Emerging Infection
in clinical settings challenging. In addition, the lack of an agreed Consortium (ISARIC) database was also searched for more studies.
pathophysiological mechanism that explains all post-COVID con-
dition symptoms further complicates discriminating symptoms Search strategy
attributable to post-COVID condition from other symptoms expe-
rienced for unrelated reasons. Thus far, it has been hypothesized A search strategy was developed in accordance with the
that dysregulation of the immune system, inflammatory reac- Cochrane Highly Sensitive Search guidelines [14]. The search strat-
tions, viral persistence, reactivation of pathogens together with egy was piloted on the PubMed database before conducting it on
host microbiome, autoimmunity, and coagulation activation could other databases. Full search strategies for all databases and regis-
contribute to the etiology of post-COVID condition [8,9]. It is ters are shown in the Supplementary Table 3.
increasingly suggested that the symptomatology of post-COVID
condition involves complex relationships between biological, Selection process
psychological, and social factors [10]. Consequently, to identify
and manage patients who suffer from this condition, the case Records retrieved from different databases were exported into
definition of post-COVID condition must be further refined. the Rayyan online platform for de-duplication. Two reviewers con-
Published studies report numerous and seemingly unrelated ducted title and abstract screening followed by full-text screening.
symptoms associated with post-COVID condition, from almost ev- Reviewers conducted screening independently and reconciled con-
ery system in the body with no clear picture of changes over flicts through discussion and consensus. The screening process is
time. Stratifying symptom reporting by time period and focusing as outlined in the PRISMA flow diagram, Figure 1.
on symptom clusters, rather than isolated individual symptoms of
post-COVID condition is likely to improve our estimation of the Data collection process
burden of disease over time and facilitate the identification of
the underlying pathophysiological mechanism(s) and the identifi- Two reviewers independently screened titles, abstracts, and full
cation of patients with more specificity and will open the way for texts of studies eligible for inclusion following predefined inclusion
targeted treatment. The aim of the systematic review and meta- and exclusion criteria. In addition, two reviewers independently
analysis was to describe the prevalence of post-COVID symptoms extracted data from included studies and assessed the quality of
up to 12 months and to identify post-COVID symptom clusters. included studies. At every stage of screening and quality assess-
ment, a third reviewer was consulted when there was a discrep-
Methods ancy in the data extracted between the first and the second re-
viewers. The review data was extracted into a predesigned data
Protocol extraction form, cleaned, and exported for analysis into R statis-
tical software, version 4.1.2, and STATA from Stata Corp, version 15.
Prior to commencing the search, a protocol was developed de-
tailing the research procedures including databases to be searched Data items
and study eligibility criteria.
The reporting of this systematic review and meta-analysis fol- The following data points were extracted: (i) study characteris-
lows the Preferred Reporting Items for Systematic Reviews and tics; author identification, contacts, and institutional affiliation(s),
Meta-analysis (PRISMA) guidelines [11]. The systematic review was date of publication, title of the study, study design, setting, study
registered on the International Prospective Register of Systematic objectives, study participants and characteristics, duration of follow
Reviews (PROSPERO) platform: Number: CRD4202126589 [12], and up and study outcomes, funding sources and disclosures by the
the protocol was published [13]. authors. (ii) review specific objectives; Physical Health Outcomes
of SARS-CoV-2 infection including; respiratory outcomes (cough,
Eligibility criteria shortness of breath), cardiovascular outcomes (palpitations, heart
attacks, stroke, blood clots in the vessels, arrhythmias, pulmonary
The inclusion and exclusion criteria of studies were defined emboli, easy bleeding), musculoskeletal outcomes (fatigue, mus-
a priori in the review protocol [13]. Observational studies pub- cle weakness, joint pain, myalgia), neurological outcomes (syncope,
lished in the peer-reviewed literature were eligible for inclusion. In dizziness, seizures, tremors, diplopia, insomnia, ageusia, anosmia,
terms of exposure, studies were eligible if they reported on partic- tinnitus, loss of balance, pins and needles, loss of sensation), gas-
ipants who tested for SARS-CoV-2 through serological or molecular trointestinal outcomes (abdominal pain, loss of appetite, consti-
tests with or without the inclusion of control groups. In terms of pation, diarrhea, nausea and vomiting) among other body sys-
outcomes, studies were included if they reported physical and/or tems. (iii) Mental Health Outcomes of SARS-CoV-2 infections data
mental health symptoms or symptom clusters or new onset symp- extracted for anxiety, depression, pain and discomfort, cognitive
toms, or ongoing symptoms following initial SARS-CoV-2 infec- function deterioration symptoms. Data were extracted for out-
tion. No restriction was applied with regards to the time elapsed comes at different time points including, 1-3 months, 3-6 months,
between SARS-CoV-2 infection and symptom reporting provided and 6-12 months. In addition to individual physical and mental
the symptoms were reported by patients previously infected with health symptoms, data were extracted for symptom clusters in
SARS-CoV-2. We restricted the literature search to studies pub- studies that reported them, including clustering techniques used
lished in English between January 2020 and August 2022. No re- and symptom clusters reported.
striction by geographical location was applied.
Risk of bias assessment
Information sources
Studies meeting the inclusion criteria were assessed for quality
We searched MEDLINE (via PubMed), Scopus, Web of Science, using the Effective Public Healthcare Panacea Project Quality As-
Science Direct, Google Scholar, EBSCOhost, EMBASE, PsycINFO, sessment tool [15]. The tool is appropriate for assessing the quality

2
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) flow diagram for included studies, databases searched, registers, and other sources.

of all study designs. Two reviewers independently undertook the Results

quality of studies assessments. The discrepancy in the quality of a
study by reviewers was resolved through discussion before arriving Study selection process
at a consensus. The results of the quality assessment were summa-
rized in Supplementary Table 4. Publication bias assessment was From seven databases searched, and other literature sources
not feasible due to the small number of studies reporting the out- (the ISARIC document library and reference lists of included stud-
comes of interest. ies), 8503 and 105 records were retrieved, respectively. In total,
1081 duplicate records were excluded after merging the records
Effect measures from the seven databases. After duplicate title and abstract screen-
ing, 6037 records were excluded because they did not meet the
We reported the proportions of patients reporting the identified review’s inclusion criteria. 341 articles were retrieved for full-text
symptom clusters following SARS-CoV-2 infection. In addition, the screening. After full-text screening, 76 original research articles
proportions of individuals self-reporting each symptom were also met the inclusion criteria. Eight studies reported distinct clusters
reported by time period: up to 3 months postinfection, 4-6, and and 68 studies reported only individual symptoms. The study se-
7-12 months postinfection. lection process is summarized in a PRISMA flow diagram, Figure 1.

Synthesis methods Characteristics of included studies

Because most studies reported the proportion of patients re- Included studies comprised of cohort and cross-sectional stud-
porting individual symptoms but did not collect data in a way that ies with a global geographical representation. Twelve studies
enabled them to determine how symptoms clustered, the meta- (15.8%) were from Italy and nine (11.8%), were from the UK. The
analysis only included studies that reported symptom clusters. Re- remaining included studies (72.4%) were distributed among 27
viewers independently identified consistently reported symptom countries (Supplementary Table 1). Thirty-six studies (47.4%) re-
clusters from the eligible studies and reached a consensus through ported post-COVID outcomes at less than 3 months post-recovery
a discussion regarding how to pool the clusters in a meta-analysis from COVID-19 illness or SARS-CoV-2 testing. Twenty-eight stud-
according to dominant symptom profiles. For each agreed cluster ies (36.8%) reported post-COVID outcomes at 3-6 months and 12
category, data on the proportions of patients experiencing symp- studies (15.8%) reported the outcomes at 7-12 months post-SARS-
toms of a particular cluster were pooled using a random-effects CoV-2 infection or a COVID-19 illness.
model. Heterogeneity was assessed in the meta-analysis using Hig-
gins & Thompson’s I2 statistic and predictive intervals with signifi- Characterization of post-COVID symptoms
cance defined at 10% α -level [16]. I2 statistics of between 25% and
less than 50% were considered low heterogeneity. I2 statistics of Overall, the top 10 most frequently reported symptoms in the
between 50% and less than 75% were considered moderate hetero- included studies were fatigue (37.80%), post-exertional malaise
geneity and I2 statistics of greater than 75% were considered sub- (35.50%), sleep disorders (25.20%), shortness of breath (23.40%),
stantial heterogeneity. Results for the pooled estimates were pre- anxiety and depression (21.70%), brain fog (13.40%), loss of concen-
sented in forest plots while results of individual studies are shown tration (13.10%), altered smell (11.20%), persistent cough (10.60%),
in tables. and muscle pain (10.20%) (Table 1).

3
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Table 1
Frequency of reported symptoms and proportions of reported symptoms by the included studies.

Post-SARS-CoV-2 Symptoms or conditions n (studies Median proportion of patients Range

reporting) reporting symptoms

Fatigue 62 37.80% 0.2-98.0%

Shortness of breath 59 23.40% 1.2-87.6%
Muscle pain 48 10.20% 0.5-86.1%
Altered smell 46 11.20% 0.0-74.0%
Persistent cough 46 10.60% 0.0-81.2%
Chest pain 44 7.30% 0.8-75.6%
Altered taste 40 9.60% 0.0-64.0%
Sleep disorders 35 25.20% 0.2-71.4%
Anxiety 34 21.70% 0.1-48.6%
Brain fog 34 13.40% 0.7-82.0%
Diarrhea 34 3.80% 0.0-58.7%
Loss of concentration 30 13.10% 1.3-71.4%
Depression 29 21.70% 0.1-50%
Fever 29 1.70% 0.0-88.7%
Joint pain 26 9.50% 0.0-77.6%
Abdominal pain 24 1.40% 0.1-53.7%
Palpitations 20 9.70% 0.6-40.8%
Nausea or vomiting 20 2.10% 0.0-25.6%
Dizziness 17 6.70% 0.2-40%
Vision problems 16 3.50% 0.3-38.5%
Pins and needles 15 8.40% 0.1-60%
Hair loss 14 9.60% 0.9-24.8%
Problems with hearing 10 6.10% 1.5-24.0%
Muscle weakness 10 9.30% 3.1-62.7%
Post-exertional malaise 7 35.50% 14.5-50.8%
Constipation 6 2.6% 0.5-12.1%
Stroke 3 2.10% 0.7-13.3%
Arrhythmias 2 6.10% 2.1-10.0%
Blood clots in vessels 2 0.70% 0.2-1.1%
Menstrual pain 1 1.60% 1.6-1.6%
Heart attack 1 0.10% 0.1-0.1%
Transient ischemia 1 0.10% 0.1-0.1%
Any post-COVID symptoms 54 59.00% 14.0-100%

Fatigue, sleep disorders and shortness of breath, post-exertional Synthesis of symptom clusters
malaise, and anxiety were the most prevalent post-COVID symp-
toms reported between 1-3 months. Between 4-6 months post- Of the seven studies reporting clusters, the cardiorespiratory
COVID-19 illness, the most prevalent symptoms were fatigue, cluster was reported by four with a pooled estimate of 36%, con-
shortness of breath, post-exertional malaise, sleep disorders, and fidence interval (32-40%), Figure 3a. The pooled estimate for the
depression. Beyond 6 months, the most prevalent post-COVID presence of neurological cluster, based on three studies, was 72%,
symptoms were post-exertional malaise, muscle weakness, depres- confidence interval (45-92%), Figure 3b. Four studies reported on
sion, anxiety and fatigue (Figure 2). the systemic inflammatory cluster. The pooled estimate for studies
Compared with 4-6 months postinfection, the proportion of pa- reporting symptoms in the systemic inflammatory cluster was 46%
tients reporting symptoms 7-12 months postinfection decreased (17-77%), Figure 3c.
for all of the most common symptoms except post-exertional
malaise (Figure 2). However, over 20% of patients still re-
Discussion
ported post-exertional malaise (median, 49.1%), depression (me-
dian, 21.2%), and anxiety (median, 20.3%), Figure 2 and Supplemen-
In this systematic review and meta-analysis, we consistently
tary Table 2.
found a high proportion of patients reporting symptoms compat-
ible with post-COVID condition. Reported symptoms and patterns
Characterization of post-COVID symptom clusters were consistent across countries, demonstrating the emergence of
post-COVID condition as a global public health problem in parallel
Eight of the included studies reported on post-COVID symp- to the COVID-19 pandemic. Studies included in the review reported
tom clusters. Seven studies identified three distinct and consis- symptoms beginning at less than 3 months up to 1 year follow-
tent post-COVID symptom clusters while one study did not find ing the initial SARS-CoV-2 infection, confirming the frequently re-
any symptom clustering in the population studied. The consistently ported notion that the road to recovery from COVID-19 illness is a
reported clusters of symptoms centered around: (i) cardiorespi- protracted journey for some patients. The most frequently reported
ratory symptoms comprising fatigue, dyspnea, chest pain, muscle symptoms included fatigue, shortness of breath, muscle pain, al-
pain, headache, and palpitations, (ii) systemic inflammatory symp- tered smell, persistent cough, chest pain, altered taste, sleep dis-
toms including dizziness, gastrointestinal symptoms, muscle pain, orders, anxiety, brain fog, loss of concentration, anxiety, and de-
muscle weakness, hair loss, and sleep disorders, (iii) neurological pression similar to findings from other systematic reviews report-
symptoms including anosmia, paresthesia, headache, neuropathies, ing on post-COVID symptoms [21,22]. By summarizing the patterns
dizziness, vision and balance problems, memory problems, and of symptoms reported in different time periods, our review helps
poor concentration. Details of studies reporting post-COVID symp- characterize post-COVID-19 condition over time.
tom clusters, the clustering methods used, and the proportions of For most symptoms, the proportion of patients reporting them
patients reporting symptoms in a cluster are shown in Table 2. was substantially lower 7-12 months postinfection than earlier. Of

4
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Figure 2. Frequency of post-COVID symptoms reported by the included studies.

Table 2
Studies reporting symptom clusters, clustering method, and frequency of symptom clusters reported.

Study Clustering method Symptom clusters Symptoms cluster n (%)

Fernandez-de-las Unsupervised K-means clustering Fatigue, dyspnea, muscle pain, chest pain, Cardiorespiratory 696 (35.3)
penas et al. [5] breathlessness, dyspnea on exertion
Fatigue, muscle pain, hair loss Systemic/inflammatory 1273 (64.7)
Frontera et al. [17] Unsupervised hierarchical agglomerative Dyspnea, chest pain, fatigue Cardiorespiratory 106 (43.8)
cluster analysis
Grisanti et al. [18] Unsupervised K-means clustering Headache, memory problems, fatigue, vertigo, Neurological 76 (69.7)
anosmia/ageusia
Kenny et al. [19] Unsupervised hierarchical agglomerative Chest pain, palpitations, Fatigue, dyspnea Cardiorespiratory 87 (37.3)
cluster analysis Muscle pain, headache, dizziness, joint pain, Systemic inflammatory 145 (62.2)
cough, GI symptoms
Fatigue, dyspnea, cough
Peluso et al. [20] Unsupervised K-means clustering Anosmia, Paresthesia, headache neuropathies, Neurological 82 (51.2)
dizziness, vision and balance problems,
memory problems, poor concentration
Reese et al. [25] Unsupervised K-means clustering Palpitations, fatigue, chest pain, headache, Cardiorespiratory 391 (31.7)
muscle pain, cough, tachycardia
Cough, fever, dizziness, muscle pain, GI Systemic inflammatory 644 (52.2)
symptoms, sleep problems, fatigue, anxiety
and depression, weight loss, headache
Ziauddeen et al. Headache, fatigue, dizziness, memory, and Neurological 2243 (88.8)
[27] concentration problems
Dizziness, headache, muscle pain, joint pain Systemic inflammatory 283 (11.4)
Sivan et al. [31] Unsupervised K-means clustering and No clusters were found NA NA
hierarchical agglomerative cluster analysis
GI, Gastrointestinal
n: number of participants reporting post-COVID symptoms out of the study population

the nine most commonly reported symptoms, only one was re- included studies. Identification of these clusters adds to the emerg-
ported more frequently at 7-12 months than earlier. Although our ing body of evidence that post-COVID condition occurs in clinical
study design does not allow us to formally analyze trends over sub-types. Describing post-COVID symptom clustering is an impor-
time, this suggests that for a high proportion of patients, most tant step in better understanding post-COVID conditions and their
symptoms disappear after 6 months while a minority go on ex- underlying mechanisms. Describing symptom clustering of symp-
periencing symptoms beyond that time period. These findings are toms can help suggest one or several etiologies and pathologi-
compatible with large longitudinal studies that found that the fre- cal mechanisms for post-COVID-19 condition, paving the way for
quency of most symptoms decreased over time and increased for evidence-informed treatment options. Our meta-analysis, the first
a minority of symptoms [23,24]. of its kind to attempt the pooling of results from studies report-
Clustering of post-COVID symptoms was reported by few of the ing post-COVID symptom clusters, suggests the existence of dis-
included studies [17–20,25–27]. Nevertheless, distinct post-COVID tinct cardiorespiratory, nervous, and immune system clusters, and
symptom clusters were consistently identified for cardiorespira- potentially different pathophysiological mechanisms contributing
tory, systemic inflammatory, and neurological systems across the to the burden of post-COVID-19 condition. Acute SARS-CoV-2 in-

5
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Figure 3. Pooled estimate of proportions of patients reporting symptoms belonging to, (a) cardiorespiratory, (b) neurological, and (c) systemic inflammatory clusters.
CI, confidence interval.

fection has been reported to cause long-term damage to multiple tistical clustering with clinical consensus building, and a better
body organs including the lungs, heart, liver, pancreas, kidneys, and understanding of the underlying pathophysiological mechanisms
nervous system among others [28–30]. associated with long COVID outcomes. Secondly, most included
This review faced several limitations. First, studies reporting studies reported symptoms within the first 3 months following
post-COVID symptom clustering were scarce in the literature mak- infection, putting them outside of the WHO case definition of post-
ing a meta-analysis feasible for only the three clusters that were COVID condition. The stratification by time enables us to over-
reported by the seven included studies. This means that the meta- come this limitation and to generate a hypothesis with regard
analysis may have missed less common symptom clusters and that to the evolution of post-COVID condition over time. Longitudinal
the accuracy and precision of the pooled estimates for the three in- studies with longer follow-up times are therefore required to bet-
cluded studies are limited. To advance our understanding of post- ter understand the duration and severity of post-COVID condition
COVID condition, we encourage researchers in the field to not just over time.
report individual symptoms but clusters. Overlapping symptoms
found in this review however may demonstrate the limits of the Conclusion
clustering techniques employed by primary studies to completely
classify symptoms into specific clusters. For example, shortness of A wide range of symptoms compatible with post-COVID condi-
breath was found in the neurological and cardiorespiratory clus- tion are frequently reported among patients infected with SARS-
ters. Since our methodology is dependent on data reported by in- CoV-2 with different clinical manifestations among affected pa-
cluded studies, we were only able to synthesize outcomes from tients. Even though the reporting seems to decrease over time
the studies as reported. It is important to note that the cluster- for most symptoms, a subset continues to report symptoms be-
ing reported in the included study is based on the statistical like- yond a year after infection. The clustering of symptoms suggests
lihood of a particular symptom (or group of symptoms) being post-COVID condition occurs in sub-types and cardiorespiratory,
reported given the status of another symptom or groups of neurologic, and systemic/inflammatory clusters seem to be consis-
symptoms, rather than clinical clustering based on body systems tently identified. This clustering provides direction for research into
obtained from expert consensus. This explains why seemingly the etiology, diagnosis, and management of post-COVID condition.
unrelated clinical symptoms can cluster mathematically even if Studies on clusters remain rare and we encourage research teams
clinically they do not form a homogenous group of symptoms. working on post-COVID condition to switch from reporting individ-
This limitation necessitates the development of better cluster- ual symptoms to clusters to accelerate our understanding of this
ing techniques, the adoption of approaches that combine sta- emerging public health issue.

6
P. Kuodi, Y. Gorelik, B. Gausi et al. International Journal of Infectious Diseases 134 (2023) 1–7

Declarations of competing Interest [10] Saunders C, Sperling S, Bendstrup E. A new paradigm is needed to explain
long COVID. Lancet Respir Med 2023;11:e12–13. doi:10.1016/S2213-2600(22)
00501-X.
The authors have no competing interests to declare. [11] Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for system-
atic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev
Funding 2015;4:1. doi:10.1186/2046- 4053- 4- 1.
[12] Booth A, Clarke M, Dooley G, et al. The nuts and bolts of PROSPERO: an
international prospective register of systematic reviews. Syst Rev 2012;1:2.
This research did not receive any specific grant from funding doi:10.1186/2046- 4053- 1- 2.
agencies in the public, commercial, or not-for-profit sectors. [13] Kuodi P, Gorelik Y, Edelstein M. Characterisation of the long-term physical and
mental health consequences of SARS-CoV-2 infection: a systematic review and
meta-analysis protocol. PLoS One 2022;17:e0266232. doi:10.1371/journal.pone.
Ethical approval 0266232.
[14] Lefebvre C, Manheimer E, Glanville J. Searching for studies. Cochrane handbook
for systematic reviews of interventions: Cochrane book series. Chichester: John
This study did not require ethical approval since we utilized
Wiley & Sons; 2008. p. 95–150. doi:10.1002/9780470712184.ch6.
publicly available data from peer-reviewed published research ar- [15] Thomas H, Ciliska D, Dobbins M. Quality assessment tool for quanti-
ticles. tative studies. Effective Public Health Practice Project McMaster Univer-
sity; 2003. https://www.ephpp.ca/PDF/Quality%20Assessment%20Tool_2010_2.
pdf[accessed 12 December 2022]
Author contributions [16] Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, et al. Cochrane
Handbook for Systematic Reviews of Interventions version 6.3. Chichester: John
ME: conceptualized the original research idea, designed the Wiley & Sons; 2022.
[17] Frontera JA, Thorpe LE, Simon NM, de Havenon A, Yaghi S, Sabadia SB, et al.
methodology, and supervised the analysis and interpretation of the Post-acute sequelae of COVID-19 symptom phenotypes and therapeutic strate-
results. PK: formally analyzed the data, interpreted the findings, gies: a prospective, observational study. PLoS One 2022;17:e0275274. doi:10.
and wrote the first version of the manuscript. YG, BG, PK and 1371/journal.pone.0275274.
[18] Grisanti SG, Garbarino S, Barisione E, Aloè T, Grosso M, Schenone C, et al.
TB: conducted article screening and assessment of quality of in- Neurological long-COVID in the outpatient clinic: two subtypes, two courses.
cluded studies, and carried out data extraction from included stud- J Neurol Sci 2022;439:120315. doi:10.1016/j.jns.2022.120315.
ies. All authors critically reviewed and edited the manuscript and [19] Kenny G, McCann K, O’Brien C, Savinelli S, Tinago W, Yousif O, et al. Identi-
fication of distinct post-COVID-19 condition clinical phenotypes through clus-
approved the final version.
ter analysis of self-reported symptoms. Open Forum Infect Dis 2022;9 ofac060).
doi:10.1093/ofid/ofac060.
Data availability [20] Peluso MJ, Kelly JD, Lu S, Goldberg SA, Davidson MC, Mathur S, et al. Persis-
tence, magnitude, and patterns of postacute symptoms and quality of life fol-
lowing onset of SARS-CoV-2 infection: cohort description and approaches for
Data generated from the review are included in this article, measurement. Open Forum Infect Dis 2022;9:ofab640. doi:10.1093/ofid/ofab640.
whereas original data are publicly included articles. Processed and [21] Han Q, Zheng B, Daines L, Sheikh A. Long-term sequelae of COVID-19: a sys-
unprocessed data extracted from included studies and used for the tematic review and meta-analysis of one-year follow-up studies on post-COVID
symptoms. Pathogens 2022;11:269. doi:10.3390/pathogens11020269.
analyses are available upon request. Original articles are publicly [22] van Kessel SAM, Olde Hartman TC, Lucassen PLBJ, van Jaarsveld CHM. Post-
accessible through the databases we searched. acute and long-COVID-19 symptoms in patients with mild diseases: a system-
atic review. Fam Pract 2022;39:159–67. doi:10.1093/fampra/cmab076.
[23] Tran VT, Porcher R, Pane I, Ravaud P. Course of post COVID-19 disease symp-
Supplementary materials toms over time in the ComPaRe long COVID prospective e-cohort. Nat Commun
2022;13:1812. doi:10.1038/s41467- 022- 29513- z.
Supplementary material associated with this article can be [24] Johnson SU, Amundsen OM, Johnson MS, Hoffart A, Halsøy Ø, Skjerdingstad N,
et al. Psychiatric symptoms in COVID-19-positive individuals in the general
found, in the online version, at doi:10.1016/j.ijid.2023.05.003. population: trajectories of depression, anxiety, and insomnia. Psychiatry Res
2022;317:114806. doi:10.1016/j.psychres.2022.114806.
References [25] Reese JT, Blau H, Casiraghi E, Bergquist T, Loomba JJ, Callahan TJ, et al. Gener-
alisable post-COVID-19 condition subtypes: findings from the NIH N3C and RE-
[1] World Health Organization COVID-19 Dashboard; 2023 https:// covid19.who.int/ COVER programmes. EBiomedicine 2023;87:104413. doi:10.1016/j.ebiom.2022.
[accessed 26 February 2023]. 104413.
[2] Abdelrahman MM, Abd-Elrahman NM, Bakheet TM. Persistence of symptoms [26] Fernández-de-las-Peñas C, Martín-Guerrero JD, Florencio LL, et al. Clus-
after improvement of acute COVID-19 infection, a longitudinal study. J Med Vi- tering analysis reveals different profiles associating long-term post-COVID
rol 2021;93:5942–6. doi:10.1002/jmv.27156. symptoms, COVID-19 symptoms at hospital admission and previous med-
[3] Carfì A, Bernabei R, Landi F. Gemelli Against COVID-19 Post-Acute Care ical co-morbidities in previously hospitalized COVID-19 survivors. Infection
Study Group. Persistent Symptoms in Patients After Acute COVID-19. JAMA 2023;51:61–9. doi:10.1007/s15010- 022- 01822- x.
2020;324:603–5. doi:10.1001/jama.2020.12603. [27] Ziauddeen N, Gurdasani D, O’Hara ME, Hastie C, Roderick P, Yao G, Alwan NA.
[4] Reuschke D, Houston D. The impact of Long COVID on the UK workforce. Appl Characteristics and impact of Long Covid: findings from an online survey. PLoS
Econ Lett 2022. doi:10.1080/13504851.2022.2098239. One 2022;17:e0264331. doi:10.1371/journal.pone.0264331.
[5] Fernández-de-Las-Peñas C, Palacios-Ceña D, Gómez-Mayordomo V, Floren- [28] Dennis A, Wamil M, Alberts J, Oben J, Cuthbertson DJ, Wootton D, et al. Mul-
cio LL, Cuadrado ML, Plaza-Manzano G, Navarro-Santana M. Prevalence of tiorgan impairment in low-risk individuals with post-COVID-19 syndrome: a
post-COVID-19 symptoms in hospitalized and non-hospitalized COVID-19 sur- prospective, community-based study. BMJ Open 2021;11:e048391. doi:10.1136/
vivors: a systematic review and meta-analysis. Eur J Intern Med 2021;92:55–70. bmjopen- 2020- 048391.
doi:10.1016/j.ejim.2021.06.009. [29] Froidure A, Mahsouli A, Liistro G, De Greef J, Belkhir L, Gérard L, et al. Integra-
[6] Seeßle J, Waterboer T, Hippchen T, Simon J, Kirchner M, Lim A, Müller B, tive respiratory follow-up of severe COVID-19 reveals common functional and
Merle U. Persistent symptoms in adult patients 1 year after coronavirus dis- lung imaging sequelae. Respir Med 2021;181:106383. doi:10.1016/j.rmed.2021.
ease 2019 (COVID-19): a prospective cohort study. Clin Infect Dis 2022;74:1191– 106383.
8. doi:10.1093/cid/ciab611. [30] Rass V, Beer R, Schiefecker AJ, Kofler M, Lindner A, Mahlknecht P, et al. Neuro-
[7] Soriano JB, Murthy S, Marshall JC, Relan P, Diaz JV. A clinical case defi- logical outcome and quality of life 3 months after COVID-19: a prospective ob-
nition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis servational cohort study. Eur J Neurol 2021;28:3348–59. doi:10.1111/ene.14803.
2022;22:e102–7. doi:10.1016/S1473-3099(21)00703-9. [31] Sivan M, Parkin A, Makower S, Greenwood DC. Post-COVID syndrome symp-
[8] Castanares-Zapatero D, Chalon P, Kohn L, Dauvrin M, Detollenaere J, Maertens toms, functional disability, and clinical severity phenotypes in hospitalized
de Noordhout C, Primus-de Jong C, Cleemput I. Van den Heede K. Patho- and nonhospitalized individuals: a cross-sectional evaluation from a commu-
physiology and mechanism of long COVID: a comprehensive review. Ann Med nity COVID rehabilitation service. J Med Virol 2022;94:1419–27. doi:10.1002/
2022;54:1473–87. doi:10.1080/07853890.2022.2076901. jmv.27456.
[9] Batiha GE, Al-Kuraishy HM, Al-Gareeb AI, Welson NN. Pathophysiology of
Post-COVID syndromes: a new perspective. Virol J 2022;19:158. doi:10.1186/
s12985- 022- 01891- 2.