key: cord-0256560-wn4kdip8
authors: Sahanic, S.; Tymoszuk, P.; Ausserhofer, D.; Rass, V.; Pizzini, A.; Nordmeyer, G.; Hampuumlfner, K.; Kurz, K.; Weber, P. M.; Sonnweber, T.; Boehm, A.; Aichner, M.; Cima, K.; Boeckle, B.; Holzner, B.; Rumpold, G.; Puelacher, C.; Kiechl, S.; Huber, A.; Wiedermann, C. J.; Sperner-Unterweger, B.; Tancevski, I.; Bellmann-Weiler, R.; Bachler, H.; Piccoliori, G.; Helbok, R.; Weiss, G.; Loeffler-Ragg, J.
title: Phenotyping of acute and persistent COVID-19 features in the outpatient setting: exploratory analysis of an international cross-sectional online survey
date: 2021-08-07
journal: nan
DOI: 10.1101/2021.08.05.21261677
sha: fdb05613fa7cf46eab32c28d0079843d0685a18f
doc_id: 256560
cord_uid: wn4kdip8

BACKGROUND. Long COVID, defined as presence of COVID-19 related symptoms 28 days or more after the onset of acute SARS-CoV-2 infection, is an emerging challenge to healthcare systems. The objective of this study was to phenotype recovery trajectories of non-hospitalized COVID-19 individuals. METHODS. We performed an international, multi-center, exploratory online survey study on demographics, comorbidities, COVID-19 symptoms and recovery status of non-hospitalized SARS-CoV-2 infected adults (Austria: n=1157), and Italy: n= 893). RESULTS. Working age subjects (Austria median: 43 yrs (IQR: 31 - 53), Italy: 45 yrs (IQR: 35 - 55)) and females (65.1% and 68.3%) predominated the study cohorts. Course of acute COVID-19 was characterized by a high symptom burden (median 13 (IQR: 9 - 18) and 13 (7 - 18) out of 44 features queried), a 47.6 - 49.3% rate of symptom persistence beyond 28 days and 20.9 - 31.9% relapse rate. By cluster analysis, two acute symptom phenotypes could be discerned: the non-specific infection phenotype and the multi-organ phenotype (MOP), the latter encompassing multiple neurological, cardiopulmonary, gastrointestinal and dermatological features. Clustering of long COVID subjects yielded three distinct subgroups, with a subset of 48.7 - 55 % long COVID individuals particularly affected by post-acute MOP symptoms. The number and presence of specific acute MOP symptoms and pre-existing multi-morbidity was linked to elevated risk of long COVID. CONCLUSION. The consistent findings of two independent cohorts further delineate patterns of acute and post-acute COVID-19 and emphasize the importance of symptom phenotyping of home-isolated COVID-19 patients to predict protracted convalescence and to allocate medical resources.

Coronavirus disease 2019 (COVID-19) displays a broad clinical spectrum ranging from asymptomatic to fatal courses of infection (1) and a large variability of symptom duration beyond acute illness.

The implications and consequences of such long-lasting, in many cases relapsing complaints for more than four weeks, often described as 'long COVID', are a growing health concern and a new burden to the health care systems. (2) For the sake of medical resource allocation, the terms 'ongoing symptomatic COVID-19' for persistence of post-infectious symptoms for 4 -12 weeks and 'post-COVID-19 syndrome' for symptoms lasting for 12 weeks or longer were introduced. (3) Long-lasting dyspnoea and fatigue are well characterized in hospitalized COVID-19 patients. Little is known about the clinical phenotype of outpatients, who constitute the majority of symptomatic COVID-19 cases. (4) (5) (6) Recent evidence suggests, that also this subset, commonly classified as mild COVID-19, experiences prolonged symptoms including chronic cough, shortness of breath, chest tightness, cognitive dysfunction, and fatigue. (7) (8) (9) (10) For this group, identification of subjects at risk of long COVID is urgently needed to effectively allocate healthcare resources.

With this international multi-center online survey of non-hospitalized COVID-19 patients, we sought to describe prevalence and patterns of acute and persistent manifestations and identify key factors impacting the presence and relapse of persistent symptoms and major physical impairment.

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint

The multi-center study 'Health after COVID-19 in Tyrol' (ClinicalTrials.gov: NCT04661462) was conducted in the neighboring European regions Tyrol (Austria) and South Tyrol (Italy) between the 30 th September 2020 and 5 th July 2021 as an anonymized online survey. (11) The participants were invited via public media calls (Tyrol and South Tyrol) and by general practitioners (South Tyrol).

The inclusion criteria encompassed confirmed SARS-CoV2 infection (PCR or seropositivity), Tyrol/South Tyrol residence and age ≥ 16 (Tyrol) or ≥18 years (South Tyrol). Exclusion criteria from further analysis were hospitalization because of COVID-19 and survey completion < 28 days after initial diagnosis of the infection (Figure 1 ).

This study was approved by the institutional review boards of the Medical University of Innsbruck (Austria) (approval number: 1257/2020) and the South Tyrol Province (Italy) (0150701). Each participant gave a digital informed consent at the survey start.

Data on demographics, socioeconomic status, comorbidities, smoking, daily medication relevant to SARS-CoV-2 infection, course and perception of acute SARS-CoV-2 infection, presence/duration of COVID-19-associated symptoms (44 items), symptom relapse and physical constitution during convalescence were recorded (Supplementary Table S1 ). The baseline cohort characteristics are shown in Table 1 and 2, the features of acute and post-acute COVID-19 in Table 3 .

Respondents retrospectively assigned their symptoms to pre-defined duration classes (absent, present for 1 -3 days, ≤ 1 week, ≤ 2 weeks, ≤ 4 weeks, ≤ 3 months, ≤ 6 months and > 6 months).

The symptoms were classified as (1) acute, when present in the first two weeks after clinical onset,

(2) sub-acute when present 2 -4 weeks after clinical onset and (3) persistent when lasting ≥ 4 weeks. For modeling, symptom numbers were stratified by quartiles (Supplementary Table S2 ).

Perception of acute infection, symptom relapse, subjective convalescence and rehabilitation need were surveyed as single questions each. Self-reported percent loss of physical performance 6 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint following COVID-19 was stratified as 0%, ≤ 25%, ≤ 50%, ≤ 75% and > 75%. The detailed variable stratification scheme is shown in the Supplementary Methods and Supplementary Table S1.

Statistical analysis was performed with R 4.0.5. The analysis pipeline is available at https://github.com/PiotrTymoszuk/health-after-COVID19-analysis-pipeline.

Frequency changes were assessed with χ2 test and differences in medians with Mann-Whitney U or is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint

Overall, 2065 individuals in Tyrol and 1075 in South Tyrol participated in the online survey. After exclusion of hospitalized respondents (n = 84 and n = 83, respectively) and questionnaires with an observation period < 28 days (n = 741 and n = 56, respectively), 1157 Tyrol and 893 South Tyrol surveys were eligible for further data analysis (Figure 1) . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint Characteristics of acute COVID-19 and recovery trajectories 8.3 -12.3% of participants experienced an asymptomatic SARS-CoV-2 infection, which is below the estimates of the Austrian (16.5% -26.9%) and Italian (up to 50%) health authority ( Table 3) . (22, 23) Almost half of the symptomatic participants described acute COVID-19 as a condition 'not experienced before' (47.8 -49.7%), followed by ' common cold-like' (25.4 -28.3%), 'influenza (flu)-like' (16.8 -20.7%) or 'gastroenteritis-like' (4.1 -7.1%) illness. In most participants (52.8 -60%), self-reported severe illness perception was limited to one week, for 18 -27.3% the feeling of severe sickness persisted for longer than 2 weeks, i. e. the duration of the official quarantine. In 29.5 -32.2% of the convalescents the queried symptoms resolved within two weeks. However, nearly half of the participants (Tyrol: 47.6%, South Tyrol: 49.3%) suffered from at least one persisting symptom for ≥ 28 days. Notably, 20.9 -31.9% experienced symptom relapse (Figure 2A , Table 3 ).

The number of acute COVID-19 symptoms present in the first two weeks (Tyrol, median: 13 (IQR: 9 -18), South Tyrol: 13 (IQR: 7 -18) out of 44 symptoms queried) and the weekly symptom count resolution rate of 35% were comparable in the study collectives ( Figure 2B , Figure   2C ).

Besides highly frequent non-specific infection symptoms (fatigue, headache, joint pain, myalgia, is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint Furthermore, 13.7 -14% of participants reported loss of hair and 4.38 -5.21% a weight reduction > 5 kg during convalescence. 23.8 -25.3% of respondents reported a physical performance loss exceeding 25% and for 7.3 -7.7% it was even more than 50%. Over one third of the study populations (36.7 -46%) reported on an incomplete recovery and 13.2 -17% declared a need for rehabilitation ( Table 3) . Table S4 ). Of note, neither the NIP nor MOP symptom count differed between the study cohorts ( Table 3) .

By an analogical procedure, we found three phenotypes of persistent COVID-19 manifestations: (1) 'hyposmia/anosmia phenotype' (HAP) encompassing closely co-occurring smell and taste disorder, 

By means of univariate, age-, sex and observation time-adjusted serial modeling we found 15 factors significantly correlating with the acute symptom count in both study collectives 10 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint (Supplementary Figure S7A) . Among them, hallmarks of subjective infection severity: contact with a physician, need of symptomatic treatment with anti-pyretic or antibiotic drugs and duration of the home isolation > 14 days were associated with a 30 -46% increase of the symptom number in the study populations. Additionally, measures of overall health status before infection such as ≥3 comorbidities, obesity, pulmonary disease and > 2 respiratory infections per year, depression or anxiety and sleep disorders were linked to > 10% higher acute COVID-19 symptom burden. In turn, a 23 -29% lower symptom count was observed in males and in participants not having received any symptomatic therapy during acute SARS-CoV2 infection (Supplementary Figure   S7B , Supplementary Tables S5 -S7, for pooled β estimates, see: Supplementary Table S8 ).

By multi-parameter LASSO modeling, the readouts for subjective severe illness perception factors linked to a higher acute symptom burden were: need to consult a physician (β = 1.18 -1.24) , antipyretic treatment (1.02 -1.06) and ≥ 3 pre-existing co-morbidities (1.05 -1.12). Opposite, no symptomatic therapy (β = 0.93 -0.96) and male sex (0.89 -0.93) could be corroborated as independent co-variates linked to low risk for perception of severe illness (Figure 6 , Table S8 ). (18) Factors linked to symptom persistence or relapse and major physical loss of performance By uni-variate modeling, 30 factors correlating with the risk of developing persistent symptoms, relapse and more than 50% physical performance loss were identified (Supplementary Figure S8) .

Especially, the overall number of symptoms, count of MOP manifestations and fatigue during acute SARS-CoV2 infection were associated with > 6-fold risk of symptom persistence, relapse and physical performance loss following COVID-19. Neuro-cognitive symptoms present in acute COVID-19 such as forgetfulness, concentration deficits, confusion and dizziness constituted another group of shared unfavorable risk factors linked to linked to an > 3-fold risk of the investigated long-COVID features. Furthermore, acute cardiopulmonary manifestations such as tachycardia, heart palpitations, shortness of breath and dyspnea were associated with an > 2.5-fold increase in risk of developing the investigated long COVID-19 features. Interestingly, male sex and no symptom-specific therapy during acute SARS-CoV2 infection were associated with a 34 -45% reduced risk of development of persistent symptoms and relapse but not with significantly physical impairment (Supplementary Figures S9 -S11, Table S7 , for pooled β estimates, see:

In multi-parameter LASSO modeling, acute forgetfulness (OR = 1.52 -1.88), dizziness (1.01 -1.09) and tachypnea (1.38 -1.5) were corroborated as independent unfavorable risk factors of 11 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint (Figure 7, Supplementary Table S10 ). Applying the LASSO approach to the relapse and major physical impairment risk, the overall number of acute symptoms (> 75 th percentile, relapse: OR = 1.14 -1.56, performance loss: 1.23 -1.26) and chest pain during acute infection (relapse: OR = 1.14 -1.56, OR = 1.02) were discerned as independent is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint Furthermore, the overall number of acute infection symptoms, fatigue, smell and taste disorders, acute MOP symptom count and particular MOP features including neuro-cognitive deficits, sleeplessness and cardiopulmonary abnormalities could be proposed as strong correlates of elevated symptom persistence, relapse and major physical impairment risk. Analogically to the previous reports we found lower symptom persistence and recurrence rates in males and subjects having is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint COVID such as symptom persistence, relapse and physical impairment. These findings may help to develop predictive tools to identify individuals at risk for long COVID and establish concepts of therapy and early rehabilitation. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ;

Simple matching distances between the symptoms of COVID-19 present in the first two weeks after symptom onset were subjected to clustering with PAM (partitioning around medoids) algorithm. Consensus clusters of the same symptoms in both cohorts were termed phenotypes. Values of simple matching coefficient between each two symptoms assigned to the consensus non-specific 28 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint (A) Numbers of significant independent factors associated with the risk of developing persistent symptoms in each of the study populations presented in a Venn plot. Common significant factors are listed next to the plot. (B) Values of non-zero regression coefficients of the significant factors correlating with the risk of developing persistent symptoms presented in the plot. Numbers of complete observations and pseudo-R^2^ statistics for each LASSO model are presented next to the plot. Obs.time: observation time (test to participation), Pre-CoV sleep disord.: sleep disorders before COVID-19, Surv. completion: survey completion, Acute imp. concentration: acute impaired concentration, Subj. CoV percept.: subjective perception of acute COVID-19. # acute MOP symptoms: sum of acute multi-organ phenotype symptoms, # acute symptoms: sum of acute COVID-19 symptoms, Daily ACE: daily intake of ACE inhibitors, CoV outbreak: SARS-CoV2 outbreak, Acute bl. marm. skin.: acute blue marmorate skin, 2Q: second quartile, 4Q fourth quartile.

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.05.21261677 doi: medRxiv preprint

Chronic COVID syndrome: Need for an appropriate medical terminology for long-COVID and COVID long-haulers

Overview | COVID-19 rapid guideline: managing the long-term effects of COVID-19 | Guidance | NICE

Cardiopulmonary recovery after COVID-19 -an observational prospective multicenter trial

6-month consequences of COVID-19 in patients discharged from hospital: a cohort study

Investigating phenotypes of pulmonary COVID-19 recovery -a longitudinal observational prospective multicenter trial

Persistent symptoms 3 months after a SARS-CoV-2 infection: the post-COVID-19 syndrome?

Attributes and predictors of long COVID

Long COVID in a prospective cohort of home-isolated patients

Characterizing long COVID in an international cohort: 7 months of symptoms and their impact

The Computer-based Health Evaluation Software (CHES): A software for electronic patient-reported outcome monitoring

Fitting linear mixed-effects models using lme4

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics

Similarity measures for categorical data: A comparative evaluation

A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise

A basic introduction to fixed-effect and randomeffects models for meta-analysis

Regression Shrinkage and Selection via the Lasso

Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing

Austrian Agency for Health and Food Safety (AGES)

Integrated surveillance of COVID-19 in Italy

Austrian Agency for Health and Food Safety (AGES)

Symptom clusters in COVID-19: A potential clinical prediction tool from the COVID Symptom Study app

Post-acute COVID-19 syndrome

Symptom profiles and progression in hospitalized and nonhospitalized patients with coronavirus disease

Neurological outcome and quality of life 3 months after COVID-19: A prospective observational cohort study