New study attempts to characterize long-COVID in all its complexity

By Michael Greenwood, M.Sc.Jul 5 2021

Many studies have reported lingering effects from severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. These post-viral complications have come to be known as long-COVID. However, the rapid onset of the COVID-19 pandemic and rush to publish potentially helpful data has led to a lack of standardization of phenotypic reporting, making analysis of the data and the resulting discovery of trends difficult.

Study: Characterizing Long COVID: Deep Phenotype of a Complex Condition. Image Credit: Dragana Gordic / Shutterstock

This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources

In a new meta-study recently released on the medRxiv* preprint server, a large subset of such reports have been organized and categorized, mapping 287 unique clinical findings related to long-COVID.

Identifying phenotypic abnormalities

The group began by selecting articles relevant to long-COVID, excluding those with acute-COVID only time points or provided insufficient detail. Within the selected reports, 287 phenotypic abnormalities represented by specific terms were identified that were associated with long-COVID, with many being mapped to identical symptoms. For example, reporting of hepatic steatosis, steatosis, liver steatosis, fatty infiltration of the liver, and fatty liver could be grouped into a single term. Fatigue was the most commonly reported term, in 45.1% of cases, and nausea the least reported, in only 3.9%, though there was wide variation across all reported symptoms.

SARS-CoV-2 reportedly affects many organs throughout the body, and many symptoms associated with long-COVID are organ-specific. The group organized symptoms by organ affected, with respiratory problems in the lungs being the most frequently reported at 35.1%, with the specific symptom term being assigned dyspnea.

The symptoms of sleep impairment and decreased diffusing lung capacity for carbon monoxide (DLCO) were also highly ranked, followed by symptoms associated with gastrointestinal symptoms such as hepatic steatosis and diarrhea. Cells in the lungs and gastrointestinal system express high levels of the ACE2 receptor, explaining the more influential impact of SARS-CoV-2 in these organs. Some cells in the brain and nervous system also express the ACE2 receptor to high levels on their surface, implicated to be the cause of the loss of taste and smell experienced by many with COVID-19. This symptom (anosmia) was reported in 12.8% of those with long-COVID, and other brain- and nervous system-related symptoms were also highly reported. These symptoms included: anxiety (22.2%), cognitive impairment (18.6%), depression (21.1%), dysphagia (1%), and myalgia (13.8%).

Some of the reported symptoms were more or less common amongst those that had experienced severe COVID-19 compared with the only mildly or asymptomatic individuals, hepatic steatosis being reported more frequently amongst severe cases, for example.

Standardization of reporting

The Human Phenotype Ontology (HPO) is an international organization that sets a standardized vocabulary for the symptoms of diseases, phenotypic abnormalities. As the sometimes complex terms used to describe phenotypic abnormalities by the medical community are often unknown to the layperson, the group reflects on the potential advantages provided by implementing algorithms that would automatically categorize symptoms from survey-provided patient information, based on the system employed here. Further, data mining techniques could be employed that determine the frequency of long-COVID-related symptom complaints on social media, massively expanding the dataset available.

The standardization of the long-COVID reporting methodology utilized in this study could help to improve the treatment and diagnosis of long-COVID, and the application of machine learning could accelerate the acquisition of useful data using this process.

In order to determine which terms may be reported by an individual that could be better described by a HPO term, anosmia rather than loss of taste, for example, lists must be assembled by analysis of meta-data. Once refined to HPO terms, data can be better analyzed and classified by clinicians and researchers, allowing them to better understand the long-term effects of SARS-CoV-2 infection, and how they relate to disease severity.

This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources