Long COVID Clinical Phenotypes up to 6 Months After Infection Identified by Latent Class Analysis of Self-Reported Symptoms

Abstract

Background The prevalence, incidence, and interrelationships of persistent symptoms after SARS-CoV-2 infection (Long COVID) vary. There are limited data on specific phenotypes of persistent symptoms. Using latent class analysis (LCA) modeling, we sought to identify whether specific phenotypes of COVID-19 were present three months and six months after acute infection. Methods This was a multicenter, prospective study of symptomatic adults tested for SARS-CoV-2 with prospectively collected data on general symptoms and fatigue-related symptoms up to six-months post-diagnosis. Using LCA, we identified symptomatically homogenous groups among participants with COVID-19 (COVID-positive) and among others without COVID-19 (COVID-negative) at each time period for both general and fatigue-related symptoms. Results Among 5,963 baseline participants (4,504 COVID-positive and 1,459 COVID-negative), 4,056 had three-month and 2,856 had six-month data at the time of analysis. We identified four distinct phenotypes of post-COVID conditions at three- and six-months for both general and fatigue-related symptoms; minimal symptom groups represented 70% of participants at three and six months. When compared with the COVID-negative cohort, COVID-positive participants had higher occurrence of loss of taste and smell, as well cognition problems. There was substantial class-switching over time; those in one symptom class at three months were equally likely to remain or enter a new phenotype at six months. Conclusions We identified distinct classes of post-COVID phenotypes for general and fatigue-related symptoms. Most participants had minimal or no symptoms at three and six months follow-up. Significant proportions of participants changed symptom groups over time, suggesting that symptoms present during the acute illness may differ from prolonged symptoms and that post-COVID conditions may have a more dynamic nature than previously recognized. Trial Registration NCT04610515

Long COVID-associated symptoms prevalent in both SARS-CoV-2 positive and negative individuals: A prospective follow-up study

“…More than two thirds of our CoV(+) outpatients reported at least one symptom, which accords with numerous other investigations [ [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [14] , [15] , [16] , [17] , [18] ]. The frequency of the most common symptoms among CoV(+) patients was slightly higher or in the same range as in previous outpatient studies: headache (27–36 ​% versus 0.4–14 ​% at various time points), fatigue (26–32 ​%/0.2–31 ​%), insomnia (18–31 ​%/2–10 ​%), weakness or tiredness (21–26 ​%/24–31 ​%), and joint symptoms (18–26 ​%/1–16 ​%) [ [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [14] , [15] , [16] , [17] , [18] ].…”

“…At 3–4 months the rates of impaired olfaction were 19 ​% in the CoV(+) and 5 ​% in the CoV(−) group and of impaired taste 15 ​% and 5 ​%, respectively; the differences were found throughout the follow-up. Other studies applying control groups also report respective rates [ 5 , 7 , 9 , 10 , 12 , 14 , 17 , 18 , 21 ]. Although seemingly trivial, these symptoms are associated with a lower quality of life [ 22 ].…”

“…However, CoV(−) control groups are missing in most reports, nor have they been set as inclusion criterion in the numerous reviews and meta-analyses [ [2] , [3] , [4] ]. Indeed, a recent cohort study involving CoV(+) outpatients and CoV(−) controls suggests for outpatients lower long COVID symptom frequencies [ 17 ] than those previously reported for inpatients. While restricting our prospective outpatient study to unvaccinated volunteers, our design also covered, so as to enable comparisons, three CoV(+) groups (outpatients and inpatients at wards or ICU) and CoV(−) controls.…”

“…CoV(+) and SARS-CoV-2-negative i.e. CoV(−) outpatients remains a minority [ [15] , [16] , [17] , [18] ]. Vaccines may provide protection and thus appear as a confounding factor [ 18 , 19 ].…”

Clusters of post-acute COVID-19 symptoms: a latent class analysis across 9 databases and 7 countries

“…Machine learning methods including clustering techniques have been used to identify patterns of associated symptoms in the approach to identify clinical subgroups of PCC [10][11][12][13] , which can subsequently facilitate targeted research into potential treatments and optimized care for affected individuals. However, previous studies using clustering approaches utilized small, prospectively collected datasets such as self-reported surveys and are therefore often not representative of people with PCC [10][11][12][13] . Moreover, the methodological heterogeneity of these studies in terms of setting, design, symptom de nition, and statistical approach prevents general comparison of clusters across studies.…”