Trajectories Of Persisting Covid-19 Symptoms Up To 24 Months After Acute Infection: Findings From The Predi-Covid Cohort Study

Trajectories Of Persisting Covid-19 Symptoms Up To 24 Months After Acute Infection: Findings From The Predi-Covid Cohort Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Trajectories Of Persisting Covid-19 Symptoms Up To 24 Months After Acute Infection: Findings From The Predi-Covid Cohort Study Aurélie Fischer, Lu Zhang, Abir Elbéji, Paul Wilmes, Chantal J. Snoeck, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4456228/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 Apr, 2025 Read the published version in BMC Infectious Diseases → Version 1 posted 4 You are reading this latest preprint version Abstract Introduction : Long COVID is a multisystemic, fluctuating condition inducing a high burden on affected people. Despite the existence of some guidelines, its management remains complicated. We aimed to demonstrate that Long COVID evolution follows different trajectories from the initial infection until 24 months after and to identify the determinants of these trajectories. Methods Study participants from the Predi-COVID cohort included between May 2020 and September 2021 were digitally followed from their acute SARS-CoV-2 infection until a maximum of 24 months. Data from 10 common symptoms were collected at study inclusion, and months 12, 15, and 24 and used to create a total symptom score. Impact of symptoms on quality of life (sleep, respiratory quality of life, anxiety, stress, and fatigue) was assessed at month 24 using standardized questionnaires and ad-hoc questions. Latent classes mixed models were used to identify total score symptom trajectories and individual symptoms trajectories. Results We included 555 participants with at least 2 different time points available during follow-up. We identified 2 trajectories: T1 “Mild symptoms, fast resolution” (N = 376; 67.7%), and T2 “Elevated and persisting symptoms” (N = 179; 32.3%). Symptom severity was worse in T2 than in T1 at 24 months (high fatigue level: 64.8% vs 19.5%, altered respiratory quality of life: 42.6% vs 4.6%, anxiety: 24.1% vs 4.6%, stress: 57.4% vs 35.6%, and bad sleep: 75.9% vs 51.1%). Fatigue and pain-related symptom frequencies in T2 increased between acute infection and month 12, and remained elevated until 24 months. Women, elevated body mass index, diabetes, and chronic medications were associated with T2. Conclusion A third of our study population was in the T2 “Elevated and persisting symptoms” trajectory, presenting high symptom frequencies up to 24 months after initial infection, with a significant impact on quality of life. This work underlined the urgent need to better identify individuals most vulnerable to long-term complications to develop tailored interventions for them. COVID-19 SARS-COV-2 Long Covid symptoms trajectories latent class mixed models quality of life Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 BACKGROUND Four years since the pandemic started, it has been estimated that more than 65 millions of people are still suffering from long-term sequelae grouped under the term Long COVID or Post COVID which became a major public health issue worldwide[ 1 , 2 ]. It has been estimated that 10–20% of people infected by SARS-CoV-2 develop Long COVID. All age categories are concerned and people with mild acute illness represent a majority of them. The economic impact of Long COVID is also important with a varying number of people with Long COVID that had to stop working, reduce their working time, or retire earlier than foreseen, depending on the countries[ 3 , 4 ]. In the US, the annual total cost of Long COVID taking into account the cost of reduced earnings, of medical spendings and of reduced quality of life, has been estimated around $ 3.7 trillion, representing 17% of the GDP[ 5 ]. In the absence of medical treatment, the management of Long COVID primarily involves the incorporation of various strategies that encompass symptom-specific care such as neurocognitive issues, physical rehabilitation for senses like taste and smell, along with dietary and activity adjustments. Pacing stands out as the primary recommendation for managing activities, emphasizing the importance of balancing exertion with rest to prevent worsening of symptoms[ 6 ]. Vaccination has been consistently shown by studies to be an effective prevention measure with a decrease of 15 to 75% of Long COVID risk, with an average risk reduction of around 40%[ 7 – 9 ]. Despite progressing knowledge about biological mechanisms, epidemiology, clinical manifestation, and risk factors, Long COVID care still faces many challenges and unmet needs[ 9 ]. Long COVID has also been shown to be heterogeneous[ 10 ], with a wide variety of symptoms[ 2 ], and affected people could be classified into different sub-groups of various Long COVID severity[ 11 , 12 ]. Only a few studies reported long-term evolution (up to 24 months or more)[ 13 – 15 ] and it is crucial to better understand how and why some people with Long COVID evolve differently over time to help physicians to personalize the care of people with Long COVID. In this study, we hypothesized that COVID-19 persisting symptoms evolved following different trajectories with a differential impact on the quality of life of affected people. We thus aimed at 1) identifying symptom trajectories from the acute infection until 24 months after, among a cohort of initially SARS-CoV-2 positive adults, 2) describing individual characteristics and identifying the main determinants of the trajectories, and 3) assessing multi-dimensions of the quality of life of people in the different trajectories. METHODS Population and study design In this study we analyzed the data from participants in the Predi-COVID study, a prospective cohort study of persons with a PCR-confirmed SARS-CoV-2 infection in Luxembourg. The Predi-COVID study design and analysis plan has been published previously[ 16 ]. The study is registered in ClinicalTrials.gov (NCT04380987) and was approved by the National Research Ethics Committee of Luxembourg (study number 202003/07) in April 2020. All participants signed an informed consent before inclusion in the study. Data were collected longitudinally, from baseline to a maximum of 24 months. Baseline data were collected by phone by an experienced clinical research nurse at study inclusion, which was done in the 5 days after the PCR test result and consisted of individual characteristics and symptoms. Participants were then invited to complete detailed self-reported questionnaires on symptoms and quality of life at months 12, 15 and 24 after inclusion in the study. In addition to data collection, participants were invited to participate in an optional biological sample collection. For those who agreed, an inclusion visit was organized in the 5 days after the positive PCR test. Nasopharyngeal swabs were collected for viral load measurement. Study Design This study is a longitudinal analysis of participants' symptoms and health status from acute infection to a maximum of 24 months after. Participants included between May 1st, 2020 and September 30th, 2021, who provided the baseline data and completed at least one of the M12, M15 or M24 questionnaires were eligible for the present study (N = 555). Symptoms and quality of life We used a list of 10 symptoms (fatigue, cough, sore throat, diarrhea, chest pain, myalgia, shortness of breath, conjunctivitis, rash, and fever) at baseline, M12, M15, and M24. This list was elaborated based on the symptoms available at baseline, with the limited level of knowledge available at the pandemic’s start. The addition of symptoms reported by the participants at each time point corresponds to the “total symptom score” variable. Trajectories modeling We used latent class mixed modeling (LCMM)[ 17 ] to identify and describe distinct trajectories in the evolution of the total symptom score and of individual symptoms from baseline to M24. This method characterizes trajectories in repeated measurements, with the assumption that several underlying subpopulations or latent classes exist. The LCMM does not require the same number of measurements per participant or measurement time points. The time metric used was the time in days from baseline. We first tested different link functions, including linear and splines with different number of nodes and nodes location, to identify the best-fitting model with one class, which had the lowest Bayesian information criterion (BIC). We then estimate the model with selected link function with two to four classes to determine the optimal number of latent trajectories, appraising the entropy of the model. We applied a gridsearch to ensure the convergence of the model. We did not include covariates to predict latent class membership. Covariates The following covariates were used as potential determinants of belonging to a given trajectory: age, gender, body mass index (BMI), smoking status (never, former and current smoker), comorbidities (diabetes, asthma, cardiovascular diseases, and hypertension), regular treatments at time of study inclusion, and disease severity at inclusion proxied by the total number of symptoms. We fitted a generalized linear model on each imputed dataset and pooled the models for a single set of estimates following the Rubin’s rules to explore the association of a characteristic and the different trajectories. Each characteristic was explored with the adjustment of the other characteristics in the model. Regression coefficients (Beta) with 95% Confidence intervals were estimated. Descriptive statistics We described the continuous variables, when the skewness was between − 1 and 1, as mean ± SD, otherwise, as median[min,max], while the categorical variables as numbers (percentage). To determine the differences of distribution we used the student t-test for normally distributed continuous variables, the Wilcoxon test for non normally distributed continuous variables and the Fisher’s exact test for categorical variables. Missing values We did not need to impute missing values for the trajectories modeling as we only included participants who responded to the entire dataset of 10 symptoms. However, participants were included in this study if they completed at least 2 out of the 4 timepoints. We imputed the missing values in the covariates and generated 45 imputed datasets. We performed all the analysis with the R version 4.3.0[ 18 ]. We used lcmm R package for trajectory analysis, the mice R package for missing covariate values imputation, and the ggplot2 R package. Sensitivity analysis Impact of missing timepoints on total score trajectories To assess the impact of missing timepoints on the total score trajectories, we compared the trajectories obtained on data from the 555 participants who completed at least baseline data and one monthly questionnaire with trajectories obtained on 84 participants who completed the 4 timepoints. Quality of life evaluation We described the impact of symptoms on quality of life in a subpopulation of 141 participants who completed the M24 questionnaire. Sleep quality was assessed using the PSQI questionnaire. A categorical variable was generated using the PSQI score. Poor sleep was defined as PSQI total score ≥ 5[ 19 ]. The respiratory quality of life was assessed using the VQ11 questionnaire, initially developed for COPD patients. One global score and 3 sub-scores (functional, psychological and relational) were calculated as described elsewhere and categorical variables were generated[ 20 ][ 21 ]. An altered respiratory quality of life was defined as VQ11 global score ≥ 22, an altered physical autonomy as functional component ≥ 8, an altered psychological quality of life as psychological component ≥ 10 and an altered relational quality of life as relational component ≥ 10. The stress level was assessed using the Perceived Stress Scale 4 (PSS 4) questionnaire. The final score ranged from 0 to 16, the highest score corresponding to a higher stress level. A PSS4 score of 6 and above was used to identify participants with high levels of stress[ 22 ]. The Fatigue Severity Scale (FSS9) which has recently been validated in COVID-19 population was used to measure the fatigue level[ 23 ]. The FSS9 score corresponded to the mean of the scores from the 9 items. A high level of fatigue was defined as a total score ≥ 36. The Generalized Anxiety Disorder 7-item (GAD7) has been used to grade the level of anxiety. A score above or equal to a cut-off of 10 was considered to identify generalized anxiety disorder[ 24 ]. SARS-CoV-2 viral load We described the viral load at inclusion in a subsample of participants who provided nasopharyngeal swabs at that time point. Briefly, SARS-CoV-2 viral RNA was extracted from 140 µL of swab supernatant and quantified by SARS-CoV-2 N gene RT-qPCR. The limit of detection (LOD) and limit of quantification (LOQ) of the assay were determined using 3-fold dilution series of the EDX SARS-CoV-2 Standard (BioRad) synthetic RNA transcripts containing five gene targets (E, N, ORF1ab, RdRP and S Genes of SARS-CoV-2 ) quantified by ddPCR by the manufacturer at 2x10 5 copies/mL) with 48 to 60 replicates of each dilution. The LOD was the lowest concentration with at least 95% detection rate and the LOQ as the lowest concentration quantifiable with a coefficient of variation below 35%. Both parameters were calculated using curve-fitting methods implemented in the R script developed by Merkes et al. 2019[ 25 ]. The LOD was 3.6 viral RNA copies/reaction and LOQ was 16.0 copies/reaction. Three replicates of 6 points 3-fold dilution series of the standard were included in each experiment to quantify SARS-CoV-2 viral RNA in clinical samples. Standard curves were analyzed and outliers were excluded when necessary to reach PCR efficiency ranging from 90–110% and R 2 above 0.98 in agreement with MIQE guidelines[ 26 ]. RNA extracts were tested in duplicates and average values were used for downstream analyses. Samples with Cq values below 40 were considered positive for SARS-CoV-2. When viral RNA concentration exceeded the upper range of the standard curves, RNA extracts were diluted in 10-fold series and retested in duplicates. When viral RNA concentrations were lower than the LOQ, samples were considered positive but no viral load was calculated. Viral loads were expressed in viral RNA copies/mL of swab supernatant. RESULTS Study population characteristics The study population was composed of 51.5% of women, mean age was 41.6 years (± 12.6), and mean BMI was 25.1kg/m 2 [16.7,55.1]. Thirty-two percent of the participants took at least one regular treatment and 6.3% had at least 2 comorbidities prior COVID-19 infection. The most frequent treatments were anti-hypertensive (10.4%), antibiotics (10.4%), and anti-cholesterol (7.4%). Total symptom score trajectories Based on the lowest BIC and the highest entropy, the optimal number of total score trajectories was identified as 2 (see Supplementary table 1 , additional file 1). The total score trajectories were named according to their characteristics: T1, mild symptoms, fast resolution, and T2, elevated and persisting symptoms. The trajectories are presented in Fig. 1 . Total symptom score evolution in T1 “Mild symptoms, fast resolution”, and T2 “Elevated and persisting symptoms”, from baseline up to 24 months after (in days). The grey areas show the 95% confidence intervals. The number of participants in each trajectory was 376/555 (67.7%) in T1 and 179/555 (32.3%) in T2. Participants in the T2 “Elevated and persisting symptom” trajectory were more frequently female (61.5% vs 46.8%), had a higher BMI (26.3 vs 24.7), were older (44 vs 40.5 years), had more frequently more than 2 comorbidities (10.6% vs 4.3%), and took more frequently at least 1 chronic medication (44.7% vs 26.3%) than participants in the T1 “Mild symptoms, fast resolution” trajectory. Participants characteristics in total study population and in each trajectory are summarized in table 1. The main determinants of experiencing a T2 “Elevated and persisting symptoms” trajectory were older age, being a female, higher BMI, multi comorbidities, diabetes, hypertension, the number and type of chronic medications (for pain, diabetes in particular) (see Fig. 2 ). When exploring symptom frequencies at each time point in the 2 trajectories we observed that fatigue, cough and fever were the most frequent symptoms at baseline in both trajectories. Symptom frequencies decreased in T1 from baseline until M24, at various speeds. In particular, fatigue decreased more slowly than couch or fever. In T2, fatigue, pain-related symptoms (chest pain, myalgia), shortness of breath, and conjunctivitis frequencies increased between baseline and M12 and remained elevated until M24. Cough frequency decreased between baseline and M12, and increased again between M15 and M24. Symptom frequencies in both trajectories are shown in Fig. 3 . Symptom frequencies are provided for each trajectory at baseline, M12, M15, and M24 (%). Individual symptom trajectories Individual symptom trajectories from baseline up to M24 were also identified and are summarized in Fig. 4 . Briefly, some symptoms evolved following 2 trajectories, one trajectory remaining at a low level and the other one increasing over time (chest pain, conjunctivitis, shortness of breath, myalgia, rash and cough). Diarrhea and sore throat evolved following 3 trajectories, one being low, one increasing and one decreasing. Fever and fatigue had particular patterns of evolution. Fever followed 2 trajectories, one including participants with low level and the other one with fever decreasing in a fast way after baseline. Fatigue was the most complex symptom in terms of individual trajectories as we identified 4 different trajectories: one with half of the participants experiencing low level of fatigue, but with a slight increase over time, the second trajectory with initial low level of fatigue but increasing and remaining at a high level until M24, the third one with initial high level of fatigue but decreasing rapidly over time, and the last one with fatigue being highly present from baseline until M24. Individual characteristics of participants in the 4 fatigue trajectories are provided in supplementary table 2 (see additional file 2). Individual symptom trajectories were modelled for the 555 participants from baseline until month 24 (in days) Sensitivity analysis The trajectories obtained on 84 participants with complete data at each timepoint were similar to those obtained on the population of 555 participants described above (See supplementary Fig. 1, additional file 3). We also described the quality of life of 138 participants who completed the month 24 questionnaire, in the total population and in the 2 trajectories. In brief, participants in the T2 “Elevated and persisting symptoms” trajectory had higher stress, fatigue and anxiety levels, and were more likely to experience poor sleep quality and poor respiratory quality of life than participants in the T1 “Mild symptoms, fast resolution” trajectory. They also less frequently recovered a similar life rhythm and professional activity as before SARS-CoV-2 infection, and they were more likely to experience a worsening of their relationships with their family or friends (see Table 2). The percentage of participants above the cut-off in each of the PSS4, FSS9, GAD7, PSQI and VQ11 scales is summarized in Fig. 5 and shows a degradation of these 5 indicators in participants from the T2 “Elevated and persisting symptoms” trajectory. Radar diagram showing the percentage of participants with high levels of fatigue, stress, anxiety and with poor sleep and respiratory quality of life in each trajectory using the specific cut-off score of each scale. The viral load was measured in nasopharyngeal swabs from 172 participants, collected during the study inclusion visit taking place within 5 days after the initial confirmation of infection. Among them, 145 (84.3%) still had detectable levels of viral RNA, and 129 (75%) had a measurable viral load. Viral RNA levels were below LoQ cut-off for 16 participants preventing viral load calculation. The median viral load at baseline was 1.2E6 [1.4E3,1.8E9] RNA copies/ml in the entire cohort, and was higher in T2 than in T1 (2.6E6 [1.5E3,1.8E9] and 9.3E5[1.4E3,1.3E9] RNA copies/ml respectively ; p = 0.139). DISCUSSION In this study we described the evolution of a score based on 10 COVID-19-related symptoms, from the initial infection up to 24 months after. We have observed two trajectories, with one third of our study participants experiencing a T2 “Elevated and persisting symptoms” trajectory, with some symptoms having increasing frequencies until month 24, and having their quality of life heavily impacted. Fatigue was the most frequent symptom in both total score trajectories and we identified 4 trajectories of fatigue taken individually. Comparison with literature Although an increasing number of studies describe Long Covid prevalence, subphenotypes and related symptoms at 12 or 24 months[ 11 , 27 – 29 ], few of them aimed at modeling the long-term trajectories of Long COVID evolution[ 13 , 15 ]. Our results are in coherence with these studies which showed also that a subpopulation of people with Long COVID experienced very long lasting symptoms with little recovery over time. Other studies focused on trajectories from specific symptoms like neurological or respiratory symptoms[ 30 – 32 ]. We found that fatigue was predominant in both trajectories. Its frequency increased over time in the T2 “Elevated and persisting symptoms”, whereas in the T1 “Mild symptoms, fast resolution” trajectory it remained on a higher level than other symptoms until M15 and decreased at M24. Looking at fatigue independently from other symptoms we identified 4 different trajectories, with 34% of our participants with either a high and persisting level of fatigue from the acute infection until 24 months after, or an initial low level of fatigue importantly increasing until month 12 and reaching a maximum between month12 and month 24. This tendency of fatigue persistence has been recently described in a recent meta-analysis on the neurological symptoms of Long COVID at 12 months[ 31 ] and another study also described a worsening of fatigue over time[ 33 ]. Being a woman and of higher age were risk factors to experience the T2 persisting Long COVID trajectory. We also showed that preexisting comorbidities like diabetes, obesity and hypertension, and associated treatments, but also treatments for pain, inflammation and anxiolytics, were associated with a higher risk of developing a severe form of Long COVID. These findings are in line with results previously described[ 15 , 34 ]. There are few studies describing the quality of life of people with Long COVID, and they generally focus on overall quality of life using questionnaires like SF12, EQ-5D-3L, or EQ-5D-5L[ 35 , 36 ] or on only one specific aspect like fatigue[ 37 ]. A recent study described the quality of life of people with Long COVID at a median time of 197.5 days after initial infection using various scales (including GAD7, PHQ9, MOCA) and showed subpopulations with a higher impact on quality of life[ 11 ]. Our study is providing additional information on the multiple aspects of quality of life that are impacted by Long COVID 24 months after acute infection. We showed that being in the T2 “Elevated and persisting symptoms” was associated with a multidimensional alteration of quality of life (altered sleep and respiratory quality of life, increase of fatigue, stress and anxiety). The impaired respiratory quality of life observed at month 24 in people belonging to the T2 highly persisting trajectory could be explained by a limited recovery in lung function 2 years after initial infection[ 30 ]. Participants in the T2 persisting trajectory had a higher SARS-CoV-2 viral load during acute infection, even though this result was not statistically significant due to the low number of data available. Previously, some studies found no relation between viral load and early COVID-19 clinical outcomes [ 38 , 39 ], however another study suggested a correlation between higher viral load during acute infection and Long COVID[ 40 ]. It would be of interest to deeper investigate this finding as it may provide new insight on Long COVID determinants and biological mechanisms. Strengths and limitations Our study has several strengths. First, all study participants had an initial PCR-confirmed SARS-CoV-2 infection and were prospectively followed up to 24 months after. Trajectories have been modeled based on 10 symptoms collected systematically at each timepoint from day 0 to month 24. Finally, study participants were in majority non hospitalized individuals, enhancing the result’s generalizability since the majority of people with Long COVID undergo mild infections. This study also has some limitations. The high number of participants who did not complete the questionnaire at months 15 and 24 might have led to an overestimation of Long COVID symptoms at 24 months, as people who completed the questionnaire were experiencing more symptoms than participants who completed only the questionnaire at 12 months. However, our sensitivity analysis on participants who completed the full set of questionnaires showed similar trajectories, confirming the reliability of our results. In addition, symptoms were self-reported, and we could not fully assert that reported symptoms were linked to Long COVID and we could not exclude that other concomitant health issues could have interfered. CONCLUSIONS Our findings demonstrated a high diversity in the long-term evolution of Long COVID. One-third of study participants are still suffering from symptoms 24 months after the acute illness with a significant impact on various dimensions of their quality of life. This work underlined the need to identify the individuals most vulnerable to long-term sequelae to develop tailored care interventions. Declarations Ethics approval and consent to participate: The study is registered in ClinicalTrials.gov (NCT04380987) and was approved by the National Research Ethics Committee of Luxembourg (study number 202003/07) in April 2020. All participants signed an informed consent before inclusion in the study. Consent for publication: Not applicable Availability of data and materials: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Competing interest : The authors declare that they have no competing interests. Funding: This work was supported by the Luxembourg Government through the CoVaLux Programme and the Luxembourg Institute of Health (Grant number 16954531). The Predi-COVID study was supported by the Luxembourg National Research Fund (FNR) (Predi-COVID, grant number 14716273), the André Losch Foundation and by European Regional Development Fund (FEDER, convention 2018-04-026-21). Author Contributions: A.F. and G.F. had full access to study data and took responsibility for the integrity of the data and the accuracy of the data analysis. L.Z., A.E. and A.F. performed the statistical analysis. A.F., G.F., and L.Z. designed the study and drafted the manuscript. C.S. took responsibility of the viral load determinations. J.L and P.O were involved in the study design and results interpretation. G.F., M.O., and P.W. obtained the funding. A.F. provided administrative, technical, or material support. All authors read and approved the final manuscript. Acknowledgments: We are thankful to all the participants of the Predi-COVID study. We also acknowledge the involvement of the interdisciplinary and inter-institutional study team that contributed to Predi-COVID. The full list of the Predi-COVID team can be found here: https://www.lih.lu/en/predi-covid-project-team/. References The Lancet. Long COVID: 3 years in. Lancet. 2023;401: 795. Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol. 2023;21: 133–146. Impact of long COVID-19 on work: a co-produced survey. Lancet. 2023;402: S98. Salmon D, Slama D, Linard F, Dumesges N, Le Baut V, Hakim F, et al. Patients with Long COVID continue to experience significant symptoms at 12 months and factors associated with improvement: A prospective cohort study in France (PERSICOR). Int J Infect Dis. 2024;140: 9–16. [No title]. [cited 19 Apr 2024]. Available: https://scholar.harvard.edu/files/cutler/files/long_covid_update_7-22.pdf Website. Available: https://world.physio/sites/default/files/2021-07/Briefing-Paper-9-Long-Covid-FINAL-English-2021_0 COVID-19 vaccination for the prevention and treatment of long COVID: A systematic review and meta-analysis. Brain Behav Immun. 2023;111: 211–229. The impact of COVID-19 vaccination prior to SARS-CoV-2 infection on prevalence of long COVID among a population-based probability sample of Michiganders, 2020-2022. Ann Epidemiol. 2024;92: 17–24. Al-Aly Z, Topol E. Solving the puzzle of Long Covid. Science. 2024;383: 830–832. Zhang H, Zang C, Xu Z, Zhang Y, Xu J, Bian J, et al. Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes. Nat Med. 2023;29: 226–235. Kitsios GD, Blacka S, Jacobs JJ, Mirza T, Naqvi A, Gentry H, et al. Subphenotypes of self-reported symptoms and outcomes in long COVID: a prospective cohort study with latent class analysis. BMJ Open. 2024;14: e077869. Fischer A, Badier N, Zhang L, Elbéji A, Wilmes P, Oustric P, et al. Long COVID Classification: Findings from a Clustering Analysis in the Predi-COVID Cohort Study. Int J Environ Res Public Health. 2022;19. doi:10.3390/ijerph192316018 Wynberg E, Verveen A, van Willigen HDG, Nieuwkerk P, Davidovich U, Lok A, et al. Two-year trajectories of COVID-19 symptoms and their association with illness perception: A prospective cohort study in Amsterdam, the Netherlands. Influenza Other Respi Viruses. 2023;17: e13190. Ballouz T, Menges D, Anagnostopoulos A, Domenghino A, Aschmann HE, Frei A, et al. Recovery and symptom trajectories up to two years after SARS-CoV-2 infection: population based, longitudinal cohort study. BMJ. 2023;381: e074425. Servier C, Porcher R, Pane I, Ravaud P, Tran V-T. Trajectories of the evolution of post-COVID-19 condition, up to two years after symptoms onset. Int J Infect Dis. 2023;133: 67–74. Fagherazzi G, Fischer A, Betsou F, Vaillant M, Ernens I, Masi S, et al. Protocol for a prospective, longitudinal cohort of people with COVID-19 and their household members to study factors associated with disease severity: the Predi-COVID study. BMJ Open. 2020;10: e041834. Proust-Lima C, Philipps V, Liquet B. Estimation of extended mixed models using latent classes and latent processes: The R package lcmm. J Stat Softw. 2017;78. doi:10.18637/jss.v078.i02 Website. Available: http://www.r-project.org/index.htm Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28: 193–213. Ninot G, Soyez F, Préfaut C. A short questionnaire for the assessment of quality of life in patients with chronic obstructive pulmonary disease: psychometric properties of VQ11. Health Qual Life Outcomes. 2013;11: 179. [No title]. [cited 4 Apr 2022]. Available: https://www.has-sante.fr/upload/docs/application/pdf/2021-07/iqss_guide_proms_specifiques_bpco_2021.pdf Malik AO, Peri-Okonny P, Gosch K, Thomas M, Mena C, Hiatt WR, et al. Association of Perceived Stress Levels With Long-term Mortality in Patients With Peripheral Artery Disease. JAMA Netw Open. 2020;3: e208741. Naik H, Shao S, Tran KC, Wong AW, Russell JA, Khor E, et al. Evaluating fatigue in patients recovering from COVID-19: validation of the fatigue severity scale and single item screening questions. Health Qual Life Outcomes. 2022;20: 170. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166: 1092–1097. Merkes CM. Generic qPCR limit of detection (LOD) / limit of quantification (LOQ) calculator. U.S. Geological Survey; 2019. doi:10.5066/P9GT00GB Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55: 611–622. Bowe B, Xie Y, Al-Aly Z. Postacute sequelae of COVID-19 at 2 years. Nat Med. 2023;29: 2347–2357. Prevalence of post-acute coronavirus disease 2019 symptoms twelve months after hospitalization in participants retained in follow-up: analyses stratified by gender from a large prospective cohort. Clin Microbiol Infect. 2023;29: 254.e7–254.e13. Harris E. Some People Still Have Long COVID Symptoms After 2 Years. JAMA. 2023;330: 1127–1127. Iversen KK, Ronit A, Ahlström MG, Nordestgaard BG, Afzal S, Benfield T. Lung Function Trajectories in Mild COVID-19 With 2-year Follow-up. J Infect Dis. 2024; jiae037. Giussani G, Westenberg E, Garcia-Azorin D, Bianchi E, Khan Y, Khan AH, et al. Prevalence and Trajectories of Post-COVID-19 Neurological Manifestations: A Systematic Review and Meta-Analysis. Neuroepidemiology. 2024;58: 120–133. Taquet M, Sillett R, Zhu L, Mendel J, Camplisson I, Dercon Q, et al. Neurological and psychiatric risk trajectories after SARS-CoV-2 infection: an analysis of 2-year retrospective cohort studies including 1 284 437 patients. Lancet Psychiatry. 2022;9: 815–827. Mazza MG, Palladini M, Villa G, De Lorenzo R, Rovere Querini P, Benedetti F. Prevalence, trajectory over time, and risk factor of post-COVID-19 fatigue. J Psychiatr Res. 2022;155: 112–119. Mateu L, Tebe C, Loste C, Santos JR, Lladós G, López C, et al. Determinants of the onset and prognosis of the post-COVID-19 condition: a 2-year prospective observational cohort study. Lancet Reg Health Eur. 2023;33: 100724. Kim Y, Bae S, Chang H-H, Kim S-W. Long COVID prevalence and impact on quality of life 2 years after acute COVID-19. Sci Rep. 2023;13: 11207. Smith P, De Pauw R, Van Cauteren D, Demarest S, Drieskens S, Cornelissen L, et al. Post COVID-19 condition and health-related quality of life: a longitudinal cohort study in the Belgian adult population. BMC Public Health. 2023;23: 1433. Walker S, Goodfellow H, Pookarnjanamorakot P, Murray E, Bindman J, Blandford A, et al. Impact of fatigue as the primary determinant of functional limitations among patients with post-COVID-19 syndrome: a cross-sectional observational study. BMJ Open. 2023;13: e069217. Kuri-Ayache M, Rivera-Cavazos A, Pérez-Castillo MF, Santos-Macías JE, González-Cantú A, Luviano-García JA, et al. Viral load and its relationship with the inflammatory response and clinical outcomes in hospitalization of patients with COVID-19. Front Immunol. 2022;13: 1060840. Abdulrahman A, Mallah SI, Alqahtani M. COVID-19 viral load not associated with disease severity: findings from a retrospective cohort study. BMC Infect Dis. 2021;21: 688. Girón Pérez DA, Fonseca-Agüero A, Toledo-Ibarra GA, Gomez-Valdivia J de J, Díaz-Resendiz KJG, Benitez-Trinidad AB, et al. Post-COVID-19 Syndrome in Outpatients and Its Association with Viral Load. Int J Environ Res Public Health. 2022;19. doi:10.3390/ijerph192215145 Tables Tables are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Additionalfile1Supplementarytable1.pdf Additional file 1: Additional file 1 - Supplementary Table 1.pdf Supplementary table 1: Determination of the optimal class number. The optimal number of classes is determined by the lowest BIC and the highest entropy. Additionalfile2SupplementaryTable2.xlsx Additional file 2: Additional file 2 - Supplementary Table 2.xlsx Supplementary table 2: Participant's individual characteristics in fatigue trajectories. Additionalfile3Supplementaryfigure1.pdf Additional file 3 : Additional file 3 -Supplementary Figure 1.pdf Supplementary figure 1: Complete case analysis: total symptom score evolution in T1 and T2 from baseline up to 24 months after (in days) for 84 participants who completed the 4 timepoints. Table1Participantsindividualcharacteristicsfinal.xlsx Table2ParticipantsQOL.xlsx Cite Share Download PDF Status: Published Journal Publication published 25 Apr, 2025 Read the published version in BMC Infectious Diseases → Version 1 posted Editorial decision: Revision requested 23 May, 2024 Submission checks completed at journal 23 May, 2024 Editor assigned by journal 23 May, 2024 First submitted to journal 21 May, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4456228","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":305853829,"identity":"901ac5b8-8a33-4026-860c-3e9d33969cdd","order_by":0,"name":"Aurélie Fischer","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7klEQVRIie3QvWrDMBDA8QsH0iLba0MhfYULhkwhfpUrBXsKGAqlQ4dMfoY+jsGQLCZzpmAvmQ1ZOoQ0p5gOHSqtHfQfhGz0Qx8AodA/TNmhI4CsAUCZz2I08sdBEjuwEMKRpMoSdpDpBsYFou7kuQIPocO677hcAmndn8v3Y1HpqLbQQYqUmHLI0KSPn+3rusKYPSRXD0yN3MUojCoWYshJskOuv5i+hegTRlculI/YXeTFavtiC4w2zH7SnlAO9mLI3sVseW4PVvPeQXb5ZBguqxklu/5sPvgpSdp5N7z9TX4yv75qPwiFQqGQqxvh8UHG/62zGwAAAABJRU5ErkJggg==","orcid":"","institution":"Deep Digital Phenotyping Research Unit. Department of Precision Health. Luxembourg Institute of Health","correspondingAuthor":true,"prefix":"","firstName":"Aurélie","middleName":"","lastName":"Fischer","suffix":""},{"id":305853832,"identity":"597d6846-5503-42ac-a20d-8a9dbcbd6f5a","order_by":1,"name":"Lu Zhang","email":"","orcid":"","institution":"Bioinformatics Platform. Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Lu","middleName":"","lastName":"Zhang","suffix":""},{"id":305853835,"identity":"92701048-d647-48d8-87ec-a6b60f60418b","order_by":2,"name":"Abir Elbéji","email":"","orcid":"","institution":"Deep Digital Phenotyping Research Unit. Department of Precision Health. Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Abir","middleName":"","lastName":"Elbéji","suffix":""},{"id":305853836,"identity":"efc0d65a-4faf-4fc7-a283-397679d2496d","order_by":3,"name":"Paul Wilmes","email":"","orcid":"","institution":"Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Wilmes","suffix":""},{"id":305853837,"identity":"3f115f89-93bb-4b4c-a9e3-aef243af7994","order_by":4,"name":"Chantal J. Snoeck","email":"","orcid":"","institution":"Clinical and Applied Virology group, Department of Infection and Immunity. Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Chantal","middleName":"J.","lastName":"Snoeck","suffix":""},{"id":305853838,"identity":"5dca5635-9a54-4b51-bb7d-2ba9bc8a5df6","order_by":5,"name":"Jérôme Larché","email":"","orcid":"","institution":"Long Covid Center, Clinique du Parc","correspondingAuthor":false,"prefix":"","firstName":"Jérôme","middleName":"","lastName":"Larché","suffix":""},{"id":305853841,"identity":"5411af8e-4753-4fdf-87f5-a918eee4f521","order_by":6,"name":"Pauline Oustric","email":"","orcid":"","institution":"Association #ApresJ20 Covid Long France","correspondingAuthor":false,"prefix":"","firstName":"Pauline","middleName":"","lastName":"Oustric","suffix":""},{"id":305853842,"identity":"0ec2b77b-2da4-4a3e-9262-9e695b64207a","order_by":7,"name":"Markus Ollert","email":"","orcid":"","institution":"Department of Infection and Immunity. Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"","lastName":"Ollert","suffix":""},{"id":305853843,"identity":"856f4b7c-9a29-4582-b5f6-6bc5fdeaf3c5","order_by":8,"name":"Guy Fagherazzi","email":"","orcid":"","institution":"Deep Digital Phenotyping Research Unit. Department of Precision Health. Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Guy","middleName":"","lastName":"Fagherazzi","suffix":""}],"badges":[],"createdAt":"2024-05-21 16:41:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4456228/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4456228/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12879-025-11023-0","type":"published","date":"2025-04-25T15:58:28+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":57953889,"identity":"368994e9-6789-4387-b130-cbbbe1cca41f","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":157545,"visible":true,"origin":"","legend":"\u003cp\u003eTotal symptom score trajectories\u003c/p\u003e\n\u003cp\u003eTotal symptom score evolution in T1 “Mild symptoms, fast resolution”, and T2 “Elevated and persisting symptoms”, from baseline up to 24 months after (in days).\u003c/p\u003e\n\u003cp\u003eThe grey areas show the 95% confidence intervals.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/591b95c1067b58e5f65886e3.png"},{"id":57953890,"identity":"99e0695f-47a0-4a43-aceb-5243783a4146","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":357543,"visible":true,"origin":"","legend":"\u003cp\u003eDeterminants of being in T2 (Elevated and persisting symptoms) vs T1 (Mild symptoms, fast resolution).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/25a958e471faad0b0505456c.png"},{"id":57954796,"identity":"f2e87e78-c664-4e0b-b6c1-775124537eb0","added_by":"auto","created_at":"2024-06-07 23:20:15","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":278749,"visible":true,"origin":"","legend":"\u003cp\u003eSymptom frequencies in T1 and T2 trajectories.\u003c/p\u003e\n\u003cp\u003eSymptom frequencies are provided for each trajectory at baseline, M12, M15, and M24 (%).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/8bab29ba8016fc150c8545b4.png"},{"id":57953892,"identity":"b28f8a68-f2aa-447b-838e-96d450b0bcec","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":508502,"visible":true,"origin":"","legend":"\u003cp\u003eIndividual symptoms trajectories\u003c/p\u003e\n\u003cp\u003eIndividual symptom trajectories were modelled for the 555 participants from baseline until month 24 (in days)\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/24a57b8288aa94bda1240b08.png"},{"id":57953894,"identity":"d7ad3fbc-5d73-467f-b4fc-4df9490919a9","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":492336,"visible":true,"origin":"","legend":"\u003cp\u003eParticipants with altered quality of life at M24\u003c/p\u003e\n\u003cp\u003eRadar diagram showing the percentage of participants with high levels of fatigue, stress, anxiety and with poor sleep and respiratory quality of life in each trajectory using the specific cut-off score of each scale.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/0d43a801263e2f3720c76424.png"},{"id":81569792,"identity":"21f4d0e3-f1ae-4475-8ca1-b02f07771c03","added_by":"auto","created_at":"2025-04-28 16:11:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2669038,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/edeea248-4dd1-4472-a87c-49776bb05365.pdf"},{"id":57954795,"identity":"bcf5beaf-67fd-4ca1-b825-9eb976904b91","added_by":"auto","created_at":"2024-06-07 23:20:15","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":50841,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 1: Additional file 1 - Supplementary Table 1.pdf\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary table 1: \u003c/strong\u003eDetermination of the optimal class number. The optimal number of classes is determined by the lowest BIC and the highest entropy.\u003c/p\u003e","description":"","filename":"Additionalfile1Supplementarytable1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/6d19370e8b5f286ca21d9772.pdf"},{"id":57954794,"identity":"189a71f0-15af-430b-9fad-f3b6b7038109","added_by":"auto","created_at":"2024-06-07 23:20:15","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":33165,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 2: Additional file 2 - Supplementary Table 2.xlsx\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary table 2:\u003c/strong\u003e Participant's individual characteristics in fatigue trajectories.\u003c/p\u003e","description":"","filename":"Additionalfile2SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/00f7e53ae3c55050e1360a7f.xlsx"},{"id":57953897,"identity":"1b42fe56-790c-4887-8c75-7fd457690736","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":145137,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 3 : Additional file 3 -Supplementary Figure 1.pdf\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary figure 1: \u003c/strong\u003eComplete case analysis: total symptom score evolution in T1 and T2 from baseline up to 24 months after (in days) for 84 participants who completed the 4 timepoints.\u003c/p\u003e","description":"","filename":"Additionalfile3Supplementaryfigure1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/4843859d7c0ecbf1d1a71603.pdf"},{"id":57953898,"identity":"b4749443-9210-45a5-baba-9a945be5a1f4","added_by":"auto","created_at":"2024-06-07 23:12:16","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":28627,"visible":true,"origin":"","legend":"","description":"","filename":"Table1Participantsindividualcharacteristicsfinal.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/e054f95558b6bf9cbad2464d.xlsx"},{"id":57953895,"identity":"73c0d2be-a9ac-4c4d-960d-ad27baae1f3a","added_by":"auto","created_at":"2024-06-07 23:12:15","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":32678,"visible":true,"origin":"","legend":"","description":"","filename":"Table2ParticipantsQOL.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4456228/v1/219c740d667c26d4d69e5729.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Trajectories Of Persisting Covid-19 Symptoms Up To 24 Months After Acute Infection: Findings From The Predi-Covid Cohort Study","fulltext":[{"header":"BACKGROUND","content":"\u003cp\u003eFour years since the pandemic started, it has been estimated that more than 65 millions of people are still suffering from long-term sequelae grouped under the term Long COVID or Post COVID which became a major public health issue worldwide[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. It has been estimated that 10\u0026ndash;20% of people infected by SARS-CoV-2 develop Long COVID. All age categories are concerned and people with mild acute illness represent a majority of them. The economic impact of Long COVID is also important with a varying number of people with Long COVID that had to stop working, reduce their working time, or retire earlier than foreseen, depending on the countries[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. In the US, the annual total cost of Long COVID taking into account the cost of reduced earnings, of medical spendings and of reduced quality of life, has been estimated around \u003cspan\u003e$\u003c/span\u003e3.7 trillion, representing 17% of the GDP[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn the absence of medical treatment, the management of Long COVID primarily involves the incorporation of various strategies that encompass symptom-specific care such as neurocognitive issues, physical rehabilitation for senses like taste and smell, along with dietary and activity adjustments. Pacing stands out as the primary recommendation for managing activities, emphasizing the importance of balancing exertion with rest to prevent worsening of symptoms[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Vaccination has been consistently shown by studies to be an effective prevention measure with a decrease of 15 to 75% of Long COVID risk, with an average risk reduction of around 40%[\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eDespite progressing knowledge about biological mechanisms, epidemiology, clinical manifestation, and risk factors, Long COVID care still faces many challenges and unmet needs[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eLong COVID has also been shown to be heterogeneous[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], with a wide variety of symptoms[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], and affected people could be classified into different sub-groups of various Long COVID severity[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Only a few studies reported long-term evolution (up to 24 months or more)[\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] and it is crucial to better understand how and why some people with Long COVID evolve differently over time to help physicians to personalize the care of people with Long COVID.\u003c/p\u003e \u003cp\u003eIn this study, we hypothesized that COVID-19 persisting symptoms evolved following different trajectories with a differential impact on the quality of life of affected people.\u003c/p\u003e \u003cp\u003eWe thus aimed at 1) identifying symptom trajectories from the acute infection until 24 months after, among a cohort of initially SARS-CoV-2 positive adults, 2) describing individual characteristics and identifying the main determinants of the trajectories, and 3) assessing multi-dimensions of the quality of life of people in the different trajectories.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePopulation and study design\u003c/h2\u003e \u003cp\u003eIn this study we analyzed the data from participants in the Predi-COVID study, a prospective cohort study of persons with a PCR-confirmed SARS-CoV-2 infection in Luxembourg. The Predi-COVID study design and analysis plan has been published previously[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. The study is registered in ClinicalTrials.gov (NCT04380987) and was approved by the National Research Ethics Committee of Luxembourg (study number 202003/07) in April 2020. All participants signed an informed consent before inclusion in the study.\u003c/p\u003e \u003cp\u003eData were collected longitudinally, from baseline to a maximum of 24 months. Baseline data were collected by phone by an experienced clinical research nurse at study inclusion, which was done in the 5 days after the PCR test result and consisted of individual characteristics and symptoms. Participants were then invited to complete detailed self-reported questionnaires on symptoms and quality of life at months 12, 15 and 24 after inclusion in the study.\u003c/p\u003e \u003cp\u003eIn addition to data collection, participants were invited to participate in an optional biological sample collection. For those who agreed, an inclusion visit was organized in the 5 days after the positive PCR test. Nasopharyngeal swabs were collected for viral load measurement.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design\u003c/h2\u003e \u003cp\u003eThis study is a longitudinal analysis of participants' symptoms and health status from acute infection to a maximum of 24 months after. Participants included between May 1st, 2020 and September 30th, 2021, who provided the baseline data and completed at least one of the M12, M15 or M24 questionnaires were eligible for the present study (N\u0026thinsp;=\u0026thinsp;555).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eSymptoms and quality of life\u003c/h2\u003e \u003cp\u003eWe used a list of 10 symptoms (fatigue, cough, sore throat, diarrhea, chest pain, myalgia, shortness of breath, conjunctivitis, rash, and fever) at baseline, M12, M15, and M24. This list was elaborated based on the symptoms available at baseline, with the limited level of knowledge available at the pandemic\u0026rsquo;s start.\u003c/p\u003e \u003cp\u003eThe addition of symptoms reported by the participants at each time point corresponds to the \u0026ldquo;total symptom score\u0026rdquo; variable.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eTrajectories modeling\u003c/h2\u003e \u003cp\u003eWe used latent class mixed modeling (LCMM)[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] to identify and describe distinct trajectories in the evolution of the total symptom score and of individual symptoms from baseline to M24. This method characterizes trajectories in repeated measurements, with the assumption that several underlying subpopulations or latent classes exist. The LCMM does not require the same number of measurements per participant or measurement time points. The time metric used was the time in days from baseline. We first tested different link functions, including linear and splines with different number of nodes and nodes location, to identify the best-fitting model with one class, which had the lowest Bayesian information criterion (BIC). We then estimate the model with selected link function with two to four classes to determine the optimal number of latent trajectories, appraising the entropy of the model. We applied a gridsearch to ensure the convergence of the model. We did not include covariates to predict latent class membership.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eCovariates\u003c/h2\u003e \u003cp\u003eThe following covariates were used as potential determinants of belonging to a given trajectory: age, gender, body mass index (BMI), smoking status (never, former and current smoker), comorbidities (diabetes, asthma, cardiovascular diseases, and hypertension), regular treatments at time of study inclusion, and disease severity at inclusion proxied by the total number of symptoms.\u003c/p\u003e \u003cp\u003eWe fitted a generalized linear model on each imputed dataset and pooled the models for a single set of estimates following the Rubin\u0026rsquo;s rules to explore the association of a characteristic and the different trajectories. Each characteristic was explored with the adjustment of the other characteristics in the model. Regression coefficients (Beta) with 95% Confidence intervals were estimated.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eDescriptive statistics\u003c/h2\u003e \u003cp\u003eWe described the continuous variables, when the skewness was between \u0026minus;\u0026thinsp;1 and 1, as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD, otherwise, as median[min,max], while the categorical variables as numbers (percentage). To determine the differences of distribution we used the student t-test for normally distributed continuous variables, the Wilcoxon test for non normally distributed continuous variables and the Fisher\u0026rsquo;s exact test for categorical variables.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eMissing values\u003c/h2\u003e \u003cp\u003eWe did not need to impute missing values for the trajectories modeling as we only included participants who responded to the entire dataset of 10 symptoms. However, participants were included in this study if they completed at least 2 out of the 4 timepoints.\u003c/p\u003e \u003cp\u003eWe imputed the missing values in the covariates and generated 45 imputed datasets.\u003c/p\u003e \u003cp\u003eWe performed all the analysis with the R version 4.3.0[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. We used lcmm R package for trajectory analysis, the mice R package for missing covariate values imputation, and the ggplot2 R package.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003eImpact of missing timepoints on total score trajectories\u003c/h2\u003e \u003cp\u003eTo assess the impact of missing timepoints on the total score trajectories, we compared the trajectories obtained on data from the 555 participants who completed at least baseline data and one monthly questionnaire with trajectories obtained on 84 participants who completed the 4 timepoints.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eQuality of life evaluation\u003c/h2\u003e \u003cp\u003eWe described the impact of symptoms on quality of life in a subpopulation of 141 participants who completed the M24 questionnaire.\u003c/p\u003e \u003cp\u003eSleep quality was assessed using the PSQI questionnaire. A categorical variable was generated using the PSQI score. Poor sleep was defined as PSQI total score\u0026thinsp;\u0026ge;\u0026thinsp;5[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe respiratory quality of life was assessed using the VQ11 questionnaire, initially developed for COPD patients. One global score and 3 sub-scores (functional, psychological and relational) were calculated as described elsewhere and categorical variables were generated[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e][\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. An altered respiratory quality of life was defined as VQ11 global score\u0026thinsp;\u0026ge;\u0026thinsp;22, an altered physical autonomy as functional component\u0026thinsp;\u0026ge;\u0026thinsp;8, an altered psychological quality of life as psychological component\u0026thinsp;\u0026ge;\u0026thinsp;10 and an altered relational quality of life as relational component\u0026thinsp;\u0026ge;\u0026thinsp;10.\u003c/p\u003e \u003cp\u003eThe stress level was assessed using the Perceived Stress Scale 4 (PSS 4) questionnaire. The final score ranged from 0 to 16, the highest score corresponding to a higher stress level. A PSS4 score of 6 and above was used to identify participants with high levels of stress[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe Fatigue Severity Scale (FSS9) which has recently been validated in COVID-19 population was used to measure the fatigue level[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. The FSS9 score corresponded to the mean of the scores from the 9 items. A high level of fatigue was defined as a total score\u0026thinsp;\u0026ge;\u0026thinsp;36.\u003c/p\u003e \u003cp\u003eThe Generalized Anxiety Disorder 7-item (GAD7) has been used to grade the level of anxiety. A score above or equal to a cut-off of 10 was considered to identify generalized anxiety disorder[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eSARS-CoV-2 viral load\u003c/h2\u003e \u003cp\u003e We described the viral load at inclusion in a subsample of participants who provided nasopharyngeal swabs at that time point. Briefly, SARS-CoV-2 viral RNA was extracted from 140 \u0026micro;L of swab supernatant and quantified by SARS-CoV-2 N gene RT-qPCR. The limit of detection (LOD) and limit of quantification (LOQ) of the assay were determined using 3-fold dilution series of the EDX SARS-CoV-2 Standard (BioRad) synthetic RNA transcripts containing five gene targets (E, N, ORF1ab, RdRP and S Genes of SARS-CoV-2 ) quantified by ddPCR by the manufacturer at 2x10\u003csup\u003e5\u003c/sup\u003e copies/mL) with 48 to 60 replicates of each dilution. The LOD was the lowest concentration with at least 95% detection rate and the LOQ as the lowest concentration quantifiable with a coefficient of variation below 35%. Both parameters were calculated using curve-fitting methods implemented in the R script developed by Merkes et al. 2019[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The LOD was 3.6 viral RNA copies/reaction and LOQ was 16.0 copies/reaction. Three replicates of 6 points 3-fold dilution series of the standard were included in each experiment to quantify SARS-CoV-2 viral RNA in clinical samples. Standard curves were analyzed and outliers were excluded when necessary to reach PCR efficiency ranging from 90\u0026ndash;110% and R\u003csup\u003e2\u003c/sup\u003e above 0.98 in agreement with MIQE guidelines[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. RNA extracts were tested in duplicates and average values were used for downstream analyses. Samples with Cq values below 40 were considered positive for SARS-CoV-2. When viral RNA concentration exceeded the upper range of the standard curves, RNA extracts were diluted in 10-fold series and retested in duplicates. When viral RNA concentrations were lower than the LOQ, samples were considered positive but no viral load was calculated. Viral loads were expressed in viral RNA copies/mL of swab supernatant.\u003c/p\u003e \u003c/div\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eStudy population characteristics\u003c/h2\u003e \u003cp\u003eThe study population was composed of 51.5% of women, mean age was 41.6 years (\u0026plusmn;\u0026thinsp;12.6), and mean BMI was 25.1kg/m\u003csup\u003e2\u003c/sup\u003e [16.7,55.1]. Thirty-two percent of the participants took at least one regular treatment and 6.3% had at least 2 comorbidities prior COVID-19 infection.\u003c/p\u003e \u003cp\u003eThe most frequent treatments were anti-hypertensive (10.4%), antibiotics (10.4%), and anti-cholesterol (7.4%).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eTotal symptom score trajectories\u003c/h2\u003e \u003cp\u003eBased on the lowest BIC and the highest entropy, the optimal number of total score trajectories was identified as 2 (see Supplementary table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, additional file 1).\u003c/p\u003e \u003cp\u003eThe total score trajectories were named according to their characteristics: T1, mild symptoms, fast resolution, and T2, elevated and persisting symptoms. The trajectories are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTotal symptom score evolution in T1 \u0026ldquo;Mild symptoms, fast resolution\u0026rdquo;, and T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo;, from baseline up to 24 months after (in days).\u003c/p\u003e \u003cp\u003eThe grey areas show the 95% confidence intervals.\u003c/p\u003e \u003cp\u003eThe number of participants in each trajectory was 376/555 (67.7%) in T1 and 179/555 (32.3%) in T2. Participants in the T2 \u0026ldquo;Elevated and persisting symptom\u0026rdquo; trajectory were more frequently female (61.5% vs 46.8%), had a higher BMI (26.3 vs 24.7), were older (44 vs 40.5 years), had more frequently more than 2 comorbidities (10.6% vs 4.3%), and took more frequently at least 1 chronic medication (44.7% vs 26.3%) than participants in the T1 \u0026ldquo;Mild symptoms, fast resolution\u0026rdquo; trajectory.\u003c/p\u003e \u003cp\u003eParticipants characteristics in total study population and in each trajectory are summarized in table 1.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe main determinants of experiencing a T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; trajectory were older age, being a female, higher BMI, multi comorbidities, diabetes, hypertension, the number and type of chronic medications (for pain, diabetes in particular) (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhen exploring symptom frequencies at each time point in the 2 trajectories we observed that fatigue, cough and fever were the most frequent symptoms at baseline in both trajectories. Symptom frequencies decreased in T1 from baseline until M24, at various speeds. In particular, fatigue decreased more slowly than couch or fever. In T2, fatigue, pain-related symptoms (chest pain, myalgia), shortness of breath, and conjunctivitis frequencies increased between baseline and M12 and remained elevated until M24. Cough frequency decreased between baseline and M12, and increased again between M15 and M24. Symptom frequencies in both trajectories are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSymptom frequencies are provided for each trajectory at baseline, M12, M15, and M24 (%).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eIndividual symptom trajectories\u003c/h2\u003e \u003cp\u003eIndividual symptom trajectories from baseline up to M24 were also identified and are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Briefly, some symptoms evolved following 2 trajectories, one trajectory remaining at a low level and the other one increasing over time (chest pain, conjunctivitis, shortness of breath, myalgia, rash and cough). Diarrhea and sore throat evolved following 3 trajectories, one being low, one increasing and one decreasing. Fever and fatigue had particular patterns of evolution. Fever followed 2 trajectories, one including participants with low level and the other one with fever decreasing in a fast way after baseline.\u003c/p\u003e \u003cp\u003eFatigue was the most complex symptom in terms of individual trajectories as we identified 4 different trajectories: one with half of the participants experiencing low level of fatigue, but with a slight increase over time, the second trajectory with initial low level of fatigue but increasing and remaining at a high level until M24, the third one with initial high level of fatigue but decreasing rapidly over time, and the last one with fatigue being highly present from baseline until M24. Individual characteristics of participants in the 4 fatigue trajectories are provided in supplementary table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003e2\u003c/span\u003e (see additional file 2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIndividual symptom trajectories were modelled for the 555 participants from baseline until month 24 (in days)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eThe trajectories obtained on 84 participants with complete data at each timepoint were similar to those obtained on the population of 555 participants described above (See supplementary Fig.\u0026nbsp;1, additional file 3).\u003c/p\u003e \u003cp\u003eWe also described the quality of life of 138 participants who completed the month 24 questionnaire, in the total population and in the 2 trajectories. In brief, participants in the T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; trajectory had higher stress, fatigue and anxiety levels, and were more likely to experience poor sleep quality and poor respiratory quality of life than participants in the T1 \u0026ldquo;Mild symptoms, fast resolution\u0026rdquo; trajectory. They also less frequently recovered a similar life rhythm and professional activity as before SARS-CoV-2 infection, and they were more likely to experience a worsening of their relationships with their family or friends (see Table\u0026nbsp;2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe percentage of participants above the cut-off in each of the PSS4, FSS9, GAD7, PSQI and VQ11 scales is summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and shows a degradation of these 5 indicators in participants from the T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; trajectory.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eRadar diagram showing the percentage of participants with high levels of fatigue, stress, anxiety and with poor sleep and respiratory quality of life in each trajectory using the specific cut-off score of each scale.\u003c/p\u003e \u003cp\u003eThe viral load was measured in nasopharyngeal swabs from 172 participants, collected during the study inclusion visit taking place within 5 days after the initial confirmation of infection. Among them, 145 (84.3%) still had detectable levels of viral RNA, and 129 (75%) had a measurable viral load. Viral RNA levels were below LoQ cut-off for 16 participants preventing viral load calculation.\u003c/p\u003e \u003cp\u003eThe median viral load at baseline was 1.2E6 [1.4E3,1.8E9] RNA copies/ml in the entire cohort, and was higher in T2 than in T1 (2.6E6 [1.5E3,1.8E9] and 9.3E5[1.4E3,1.3E9] RNA copies/ml respectively ; p\u0026thinsp;=\u0026thinsp;0.139).\u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this study we described the evolution of a score based on 10 COVID-19-related symptoms, from the initial infection up to 24 months after. We have observed two trajectories, with one third of our study participants experiencing a T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; trajectory, with some symptoms having increasing frequencies until month 24, and having their quality of life heavily impacted. Fatigue was the most frequent symptom in both total score trajectories and we identified 4 trajectories of fatigue taken individually.\u003c/p\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eComparison with literature\u003c/h2\u003e \u003cp\u003eAlthough an increasing number of studies describe Long Covid prevalence, subphenotypes and related symptoms at 12 or 24 months[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], few of them aimed at modeling the long-term trajectories of Long COVID evolution[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Our results are in coherence with these studies which showed also that a subpopulation of people with Long COVID experienced very long lasting symptoms with little recovery over time. Other studies focused on trajectories from specific symptoms like neurological or respiratory symptoms[\u003cspan additionalcitationids=\"CR31\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eWe found that fatigue was predominant in both trajectories. Its frequency increased over time in the T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo;, whereas in the T1 \u0026ldquo;Mild symptoms, fast resolution\u0026rdquo; trajectory it remained on a higher level than other symptoms until M15 and decreased at M24. Looking at fatigue independently from other symptoms we identified 4 different trajectories, with 34% of our participants with either a high and persisting level of fatigue from the acute infection until 24 months after, or an initial low level of fatigue importantly increasing until month 12 and reaching a maximum between month12 and month 24. This tendency of fatigue persistence has been recently described in a recent meta-analysis on the neurological symptoms of Long COVID at 12 months[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] and another study also described a worsening of fatigue over time[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBeing a woman and of higher age were risk factors to experience the T2 persisting Long COVID trajectory. We also showed that preexisting comorbidities like diabetes, obesity and hypertension, and associated treatments, but also treatments for pain, inflammation and anxiolytics, were associated with a higher risk of developing a severe form of Long COVID. These findings are in line with results previously described[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThere are few studies describing the quality of life of people with Long COVID, and they generally focus on overall quality of life using questionnaires like SF12, EQ-5D-3L, or EQ-5D-5L[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] or on only one specific aspect like fatigue[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. A recent study described the quality of life of people with Long COVID at a median time of 197.5 days after initial infection using various scales (including GAD7, PHQ9, MOCA) and showed subpopulations with a higher impact on quality of life[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Our study is providing additional information on the multiple aspects of quality of life that are impacted by Long COVID 24 months after acute infection. We showed that being in the T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; was associated with a multidimensional alteration of quality of life (altered sleep and respiratory quality of life, increase of fatigue, stress and anxiety).\u003c/p\u003e \u003cp\u003eThe impaired respiratory quality of life observed at month 24 in people belonging to the T2 highly persisting trajectory could be explained by a limited recovery in lung function 2 years after initial infection[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eParticipants in the T2 persisting trajectory had a higher SARS-CoV-2 viral load during acute infection, even though this result was not statistically significant due to the low number of data available. Previously, some studies found no relation between viral load and early COVID-19 clinical outcomes [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e], however another study suggested a correlation between higher viral load during acute infection and Long COVID[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. It would be of interest to deeper investigate this finding as it may provide new insight on Long COVID determinants and biological mechanisms.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eStrengths and limitations\u003c/h2\u003e \u003cp\u003eOur study has several strengths. First, all study participants had an initial PCR-confirmed SARS-CoV-2 infection and were prospectively followed up to 24 months after. Trajectories have been modeled based on 10 symptoms collected systematically at each timepoint from day 0 to month 24. Finally, study participants were in majority non hospitalized individuals, enhancing the result\u0026rsquo;s generalizability since the majority of people with Long COVID undergo mild infections.\u003c/p\u003e \u003cp\u003eThis study also has some limitations. The high number of participants who did not complete the questionnaire at months 15 and 24 might have led to an overestimation of Long COVID symptoms at 24 months, as people who completed the questionnaire were experiencing more symptoms than participants who completed only the questionnaire at 12 months. However, our sensitivity analysis on participants who completed the full set of questionnaires showed similar trajectories, confirming the reliability of our results.\u003c/p\u003e \u003cp\u003eIn addition, symptoms were self-reported, and we could not fully assert that reported symptoms were linked to Long COVID and we could not exclude that other concomitant health issues could have interfered.\u003c/p\u003e \u003c/div\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eOur findings demonstrated a high diversity in the long-term evolution of Long COVID. One-third of study participants are still suffering from symptoms 24 months after the acute illness with a significant impact on various dimensions of their quality of life. This work underlined the need to identify the individuals most vulnerable to long-term sequelae to develop tailored care interventions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003eThe study is registered in ClinicalTrials.gov (NCT04380987) and was approved by the National Research Ethics Committee of Luxembourg (study number 202003/07) in April 2020. All participants signed an informed consent before inclusion in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e Not applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials:\u003c/strong\u003e The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eCompeting interest\u003c/strong\u003e: The authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThis work was supported by the Luxembourg Government through the CoVaLux Programme and the Luxembourg Institute of Health (Grant number 16954531). The Predi-COVID study was supported by the Luxembourg National Research Fund (FNR) (Predi-COVID, grant number 14716273), the Andr\u0026eacute; Losch Foundation and by European Regional Development Fund (FEDER, convention 2018-04-026-21).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eA.F. and G.F. had full access to study data and took responsibility for the integrity of the data and the accuracy of the data analysis. L.Z., A.E. and A.F. performed the statistical analysis. A.F., G.F., and L.Z. designed the study and drafted the manuscript. C.S. took responsibility of the viral load determinations. J.L and P.O were involved in the study design and results interpretation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eG.F., M.O., and P.W. obtained the funding. A.F. provided administrative, technical, or material support. All authors\u0026nbsp;read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e We are thankful to all the participants of the Predi-COVID study. We also acknowledge the involvement of the interdisciplinary and inter-institutional study team that contributed to Predi-COVID. The full list of the Predi-COVID team can be found here: https://www.lih.lu/en/predi-covid-project-team/.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eThe Lancet. Long COVID: 3 years in. Lancet. 2023;401: 795.\u003c/li\u003e\n\u003cli\u003eDavis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol. 2023;21: 133\u0026ndash;146.\u003c/li\u003e\n\u003cli\u003eImpact of long COVID-19 on work: a co-produced survey. Lancet. 2023;402: S98.\u003c/li\u003e\n\u003cli\u003eSalmon D, Slama D, Linard F, Dumesges N, Le Baut V, Hakim F, et al. Patients with Long COVID continue to experience significant symptoms at 12 months and factors associated with improvement: A prospective cohort study in France (PERSICOR). Int J Infect Dis. 2024;140: 9\u0026ndash;16.\u003c/li\u003e\n\u003cli\u003e[No title]. [cited 19 Apr 2024]. Available: https://scholar.harvard.edu/files/cutler/files/long_covid_update_7-22.pdf\u003c/li\u003e\n\u003cli\u003eWebsite. Available: https://world.physio/sites/default/files/2021-07/Briefing-Paper-9-Long-Covid-FINAL-English-2021_0\u003c/li\u003e\n\u003cli\u003eCOVID-19 vaccination for the prevention and treatment of long COVID: A systematic review and meta-analysis. Brain Behav Immun. 2023;111: 211\u0026ndash;229.\u003c/li\u003e\n\u003cli\u003eThe impact of COVID-19 vaccination prior to SARS-CoV-2 infection on prevalence of long COVID among a population-based probability sample of Michiganders, 2020-2022. Ann Epidemiol. 2024;92: 17\u0026ndash;24.\u003c/li\u003e\n\u003cli\u003eAl-Aly Z, Topol E. Solving the puzzle of Long Covid. Science. 2024;383: 830\u0026ndash;832.\u003c/li\u003e\n\u003cli\u003eZhang H, Zang C, Xu Z, Zhang Y, Xu J, Bian J, et al. Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes. Nat Med. 2023;29: 226\u0026ndash;235.\u003c/li\u003e\n\u003cli\u003eKitsios GD, Blacka S, Jacobs JJ, Mirza T, Naqvi A, Gentry H, et al. Subphenotypes of self-reported symptoms and outcomes in long COVID: a prospective cohort study with latent class analysis. BMJ Open. 2024;14: e077869.\u003c/li\u003e\n\u003cli\u003eFischer A, Badier N, Zhang L, Elb\u0026eacute;ji A, Wilmes P, Oustric P, et al. Long COVID Classification: Findings from a Clustering Analysis in the Predi-COVID Cohort Study. Int J Environ Res Public Health. 2022;19. doi:10.3390/ijerph192316018\u003c/li\u003e\n\u003cli\u003eWynberg E, Verveen A, van Willigen HDG, Nieuwkerk P, Davidovich U, Lok A, et al. Two-year trajectories of COVID-19 symptoms and their association with illness perception: A prospective cohort study in Amsterdam, the Netherlands. Influenza Other Respi Viruses. 2023;17: e13190.\u003c/li\u003e\n\u003cli\u003eBallouz T, Menges D, Anagnostopoulos A, Domenghino A, Aschmann HE, Frei A, et al. Recovery and symptom trajectories up to two years after SARS-CoV-2 infection: population based, longitudinal cohort study. BMJ. 2023;381: e074425.\u003c/li\u003e\n\u003cli\u003eServier C, Porcher R, Pane I, Ravaud P, Tran V-T. Trajectories of the evolution of post-COVID-19 condition, up to two years after symptoms onset. Int J Infect Dis. 2023;133: 67\u0026ndash;74.\u003c/li\u003e\n\u003cli\u003eFagherazzi G, Fischer A, Betsou F, Vaillant M, Ernens I, Masi S, et al. Protocol for a prospective, longitudinal cohort of people with COVID-19 and their household members to study factors associated with disease severity: the Predi-COVID study. BMJ Open. 2020;10: e041834.\u003c/li\u003e\n\u003cli\u003eProust-Lima C, Philipps V, Liquet B. Estimation of extended mixed models using latent classes and latent processes: The R package lcmm. J Stat Softw. 2017;78. doi:10.18637/jss.v078.i02\u003c/li\u003e\n\u003cli\u003eWebsite. Available: http://www.r-project.org/index.htm\u003c/li\u003e\n\u003cli\u003eBuysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28: 193\u0026ndash;213.\u003c/li\u003e\n\u003cli\u003eNinot G, Soyez F, Pr\u0026eacute;faut C. A short questionnaire for the assessment of quality of life in patients with chronic obstructive pulmonary disease: psychometric properties of VQ11. Health Qual Life Outcomes. 2013;11: 179.\u003c/li\u003e\n\u003cli\u003e[No title]. [cited 4 Apr 2022]. Available: https://www.has-sante.fr/upload/docs/application/pdf/2021-07/iqss_guide_proms_specifiques_bpco_2021.pdf\u003c/li\u003e\n\u003cli\u003eMalik AO, Peri-Okonny P, Gosch K, Thomas M, Mena C, Hiatt WR, et al. Association of Perceived Stress Levels With Long-term Mortality in Patients With Peripheral Artery Disease. JAMA Netw Open. 2020;3: e208741.\u003c/li\u003e\n\u003cli\u003eNaik H, Shao S, Tran KC, Wong AW, Russell JA, Khor E, et al. Evaluating fatigue in patients recovering from COVID-19: validation of the fatigue severity scale and single item screening questions. Health Qual Life Outcomes. 2022;20: 170.\u003c/li\u003e\n\u003cli\u003eSpitzer RL, Kroenke K, Williams JBW, L\u0026ouml;we B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166: 1092\u0026ndash;1097.\u003c/li\u003e\n\u003cli\u003eMerkes CM. Generic qPCR limit of detection (LOD) / limit of quantification (LOQ) calculator. U.S. Geological Survey; 2019. doi:10.5066/P9GT00GB\u003c/li\u003e\n\u003cli\u003eBustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55: 611\u0026ndash;622.\u003c/li\u003e\n\u003cli\u003eBowe B, Xie Y, Al-Aly Z. Postacute sequelae of COVID-19 at 2 years. Nat Med. 2023;29: 2347\u0026ndash;2357.\u003c/li\u003e\n\u003cli\u003ePrevalence of post-acute coronavirus disease 2019 symptoms twelve months after hospitalization in participants retained in follow-up: analyses stratified by gender from a large prospective cohort. Clin Microbiol Infect. 2023;29: 254.e7\u0026ndash;254.e13.\u003c/li\u003e\n\u003cli\u003eHarris E. Some People Still Have Long COVID Symptoms After 2 Years. JAMA. 2023;330: 1127\u0026ndash;1127.\u003c/li\u003e\n\u003cli\u003eIversen KK, Ronit A, Ahlstr\u0026ouml;m MG, Nordestgaard BG, Afzal S, Benfield T. Lung Function Trajectories in Mild COVID-19 With 2-year Follow-up. J Infect Dis. 2024; jiae037.\u003c/li\u003e\n\u003cli\u003eGiussani G, Westenberg E, Garcia-Azorin D, Bianchi E, Khan Y, Khan AH, et al. Prevalence and Trajectories of Post-COVID-19 Neurological Manifestations: A Systematic Review and Meta-Analysis. Neuroepidemiology. 2024;58: 120\u0026ndash;133.\u003c/li\u003e\n\u003cli\u003eTaquet M, Sillett R, Zhu L, Mendel J, Camplisson I, Dercon Q, et al. Neurological and psychiatric risk trajectories after SARS-CoV-2 infection: an analysis of 2-year retrospective cohort studies including 1 284 437 patients. Lancet Psychiatry. 2022;9: 815\u0026ndash;827.\u003c/li\u003e\n\u003cli\u003eMazza MG, Palladini M, Villa G, De Lorenzo R, Rovere Querini P, Benedetti F. Prevalence, trajectory over time, and risk factor of post-COVID-19 fatigue. J Psychiatr Res. 2022;155: 112\u0026ndash;119.\u003c/li\u003e\n\u003cli\u003eMateu L, Tebe C, Loste C, Santos JR, Llad\u0026oacute;s G, L\u0026oacute;pez C, et al. Determinants of the onset and prognosis of the post-COVID-19 condition: a 2-year prospective observational cohort study. Lancet Reg Health Eur. 2023;33: 100724.\u003c/li\u003e\n\u003cli\u003eKim Y, Bae S, Chang H-H, Kim S-W. Long COVID prevalence and impact on quality of life 2 years after acute COVID-19. Sci Rep. 2023;13: 11207.\u003c/li\u003e\n\u003cli\u003eSmith P, De Pauw R, Van Cauteren D, Demarest S, Drieskens S, Cornelissen L, et al. Post COVID-19 condition and health-related quality of life: a longitudinal cohort study in the Belgian adult population. BMC Public Health. 2023;23: 1433.\u003c/li\u003e\n\u003cli\u003eWalker S, Goodfellow H, Pookarnjanamorakot P, Murray E, Bindman J, Blandford A, et al. Impact of fatigue as the primary determinant of functional limitations among patients with post-COVID-19 syndrome: a cross-sectional observational study. BMJ Open. 2023;13: e069217.\u003c/li\u003e\n\u003cli\u003eKuri-Ayache M, Rivera-Cavazos A, P\u0026eacute;rez-Castillo MF, Santos-Mac\u0026iacute;as JE, Gonz\u0026aacute;lez-Cant\u0026uacute; A, Luviano-Garc\u0026iacute;a JA, et al. Viral load and its relationship with the inflammatory response and clinical outcomes in hospitalization of patients with COVID-19. Front Immunol. 2022;13: 1060840.\u003c/li\u003e\n\u003cli\u003eAbdulrahman A, Mallah SI, Alqahtani M. COVID-19 viral load not associated with disease severity: findings from a retrospective cohort study. BMC Infect Dis. 2021;21: 688.\u003c/li\u003e\n\u003cli\u003eGir\u0026oacute;n P\u0026eacute;rez DA, Fonseca-Ag\u0026uuml;ero A, Toledo-Ibarra GA, Gomez-Valdivia J de J, D\u0026iacute;az-Resendiz KJG, Benitez-Trinidad AB, et al. Post-COVID-19 Syndrome in Outpatients and Its Association with Viral Load. Int J Environ Res Public Health. 2022;19. doi:10.3390/ijerph192215145\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-infectious-diseases","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"infd","sideBox":"Learn more about [BMC Infectious Diseases](http://bmcinfectdis.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/infd","title":"BMC Infectious Diseases","twitterHandle":"#bmcinfectdis","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"COVID-19, SARS-COV-2, Long Covid symptoms, trajectories, latent class mixed models, quality of life","lastPublishedDoi":"10.21203/rs.3.rs-4456228/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4456228/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eIntroduction\u003c/h2\u003e \u003cp\u003e: Long COVID is a multisystemic, fluctuating condition inducing a high burden on affected people. Despite the existence of some guidelines, its management remains complicated. We aimed to demonstrate that Long COVID evolution follows different trajectories from the initial infection until 24 months after and to identify the determinants of these trajectories.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eStudy participants from the Predi-COVID cohort included between May 2020 and September 2021 were digitally followed from their acute SARS-CoV-2 infection until a maximum of 24 months. Data from 10 common symptoms were collected at study inclusion, and months 12, 15, and 24 and used to create a total symptom score. Impact of symptoms on quality of life (sleep, respiratory quality of life, anxiety, stress, and fatigue) was assessed at month 24 using standardized questionnaires and ad-hoc questions. Latent classes mixed models were used to identify total score symptom trajectories and individual symptoms trajectories.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003e We included 555 participants with at least 2 different time points available during follow-up. We identified 2 trajectories: T1 \u0026ldquo;Mild symptoms, fast resolution\u0026rdquo; (N\u0026thinsp;=\u0026thinsp;376; 67.7%), and T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; (N\u0026thinsp;=\u0026thinsp;179; 32.3%). Symptom severity was worse in T2 than in T1 at 24 months (high fatigue level: 64.8% vs 19.5%, altered respiratory quality of life: 42.6% vs 4.6%, anxiety: 24.1% vs 4.6%, stress: 57.4% vs 35.6%, and bad sleep: 75.9% vs 51.1%). Fatigue and pain-related symptom frequencies in T2 increased between acute infection and month 12, and remained elevated until 24 months. Women, elevated body mass index, diabetes, and chronic medications were associated with T2.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eA third of our study population was in the T2 \u0026ldquo;Elevated and persisting symptoms\u0026rdquo; trajectory, presenting high symptom frequencies up to 24 months after initial infection, with a significant impact on quality of life. This work underlined the urgent need to better identify individuals most vulnerable to long-term complications to develop tailored interventions for them.\u003c/p\u003e","manuscriptTitle":"Trajectories Of Persisting Covid-19 Symptoms Up To 24 Months After Acute Infection: Findings From The Predi-Covid Cohort Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-07 23:12:10","doi":"10.21203/rs.3.rs-4456228/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-05-23T09:08:40+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-05-23T08:27:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-05-23T08:27:11+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Infectious Diseases","date":"2024-05-21T16:40:18+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-infectious-diseases","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"infd","sideBox":"Learn more about [BMC Infectious Diseases](http://bmcinfectdis.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/infd","title":"BMC Infectious Diseases","twitterHandle":"#bmcinfectdis","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c0d0de86-92d9-454e-aea3-146b5f0ad1a2","owner":[],"postedDate":"June 7th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-04-28T16:04:30+00:00","versionOfRecord":{"articleIdentity":"rs-4456228","link":"https://doi.org/10.1186/s12879-025-11023-0","journal":{"identity":"bmc-infectious-diseases","isVorOnly":false,"title":"BMC Infectious Diseases"},"publishedOn":"2025-04-25 15:58:28","publishedOnDateReadable":"April 25th, 2025"},"versionCreatedAt":"2024-06-07 23:12:10","video":"","vorDoi":"10.1186/s12879-025-11023-0","vorDoiUrl":"https://doi.org/10.1186/s12879-025-11023-0","workflowStages":[]},"version":"v1","identity":"rs-4456228","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4456228","identity":"rs-4456228","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}