A recent published in Nature Medicine, "Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes" by Hao Zhang, PhD, et al. used a data-driven framework to stratify patients suffering with post-acute SARS-CoV-2 (PASC), also referred to as long COVID, into different subcategories of sequelae.
The bulk of the patients in the study were categorized as "subphenotype 1" -- these were the sickest patients, and had cardiac and renal involvement. A close second was "subphenotype 2," which included patients with respiratory and concomitant sleep dysfunction and anxiety symptoms. "Subphenotype 3" patients had predominant neurological and musculoskeletal presentation, and the final, and smallest category was "subphenotype 4," including patients with significant digestive symptoms.
The researchers' perspectives and findings are important, and may better serve the PASC population and physicians, like me, who care for them.
There has been much dialogue over the past 2 years to understand what constitutes PASC. Indeed, the diagnosis remains controversial in some institutions, which can create barriers for patients struggling and trying to seek care. At the core of the recent study is a clear consensus that PASC does exist, albeit not as standardized as physicians would like for the clarity required to help patients effectively and confidently. The data mining from the INSIGHT cohort and the OneFlorida+ cohort clearly presents the wide array of symptoms experienced by patients at some point following acute infection with SARS-CoV-2.
Earlier studies that have identified PASC as an authentic post-viral entity have focused on the experienced by that encompass almost every organ system of the body. This has contributed to confusion about exactly what PASC is and how it should be classified, and further contributed to the lack of a cohesive paradigm within which physicians can work to treat PASC patients.
However, the study in Nature Medicine, with its unique database analytics approach, helped categorize different subphenotypes of PASC that focus on symptoms at presentation, along with history of previous organ dysfunction and demographics, to better enlighten physicians working with this patient population.
The categories presented by the Nature Medicine authors are consistent with the more prominent patient-reported symptoms identified in other studies; namely, those that involve the cardiovascular, respiratory, neurological, and gastrointestinal systems. Perhaps more importantly, it provides algorithmic fodder to prevent the risk of misdiagnosis, which is always a concern when a patient presents with many of these symptoms. Furthermore, it helps to decrease the notion that psychosomatism underlies this presentation.
In our clinic, while we have an admittedly skewed population of neurological symptoms, a full history reveals accompanying symptoms of gastrointestinal dysfunction and dysautonomia, consistent with the overlap of symptoms of the subphenotypes. This is highlighted by the authors and magnifies the importance of further dissecting the subphenotypes based on demographics and co-morbid conditions. It is interesting to note that because of the large population of connective tissue disorders seen in our clinic, our patient population is very much in line with what the authors deem "subphenotype 3" (musculoskeletal and nervous).
Among our PASC patients, along with neurological symptoms of headaches, cognitive impairment, fatigue, paresthesias, and sleep changes, many develop connective tissue compromise following the acute infection stage and are found to have co-existing mast cell activation syndrome (MCAS). This may explain the higher incidence of prescription rates of anti-allergy and anti-inflammatory medications as well as skin problems and rash noted of this subphenotype.
Stratifying patients in this manner can help inform direction for treatment regimens based on subphenotype. As it currently stands, patients are either offered symptomatic treatment or various regimens that may have a less than favorable risk profile (i.e., triple anticoagulant therapy) for proposed unifying mechanisms of PASC. Thus far, have led to significant improvements overall for PASC patients.
I can only hope the subphenotypes described by Zhang et al. will direct additional research into perhaps more targeted treatment options for prominent symptoms of specific PASC subphenotypes. An additional benefit may also be improved diagnostic coding for more efficient insurance reimbursements and disability approvals, as cost of care has been a burden for many patients with PASC.
Finally, this study offered insight into which population may be at greater risk for long COVID after acute infection by SARS-CoV-2 -- a question often asked by patients. A formula based on severity of acute infection, co-morbid conditions, medications, gender, and age may provide the context with which physicians and researchers can engage in conversations about who is most vulnerable. Importantly, this insight may prompt further research into potential methods to either reduce the risk of PASC or prevent it completely.
Overall, the complexity of post-viral sequelae pathophysiology in general, and PASC in particular, requires strong data compilation and data-driven sensitivity analysis to guide identification, assessment, diagnosis, and treatment. The pandemic has given us an opportunity to study what may be the early stages of a chronic, progressive disorder that results from an acute viral illness.
If we are able to learn who is most at risk and why, we can better strategize about how best to protect them. At the very least, this study will help physicians query patients with a history of a positive COVID infection and their symptoms to evaluate if they fall into a subphenotype of PASC that will guide next best steps in treatment and management options.
is a neurologist and medical director for the EDS/Chiari Center at Mount Sinai South Nassau in New York, and is also in private practice in Seattle.