Inflammatory phenotypes may be more important than age and comorbidities in predicting clinical outcomes in hospitalised patients with COVID-19

Abstract

Background

In critically ill patients with coronavirus disease 2019 (COVID-19), distinct hyperinflammatory and hypoinflammatory phenotypes have been described, with different outcomes and responses to therapy. We investigated if similar phenotypes exist in non-severe illness.

Methods

Consecutive patients with polymerase chain reaction (PCR) confirmed SARS-CoV-2 were examined. Baseline demographics and laboratory investigations were tabulated, including serum C-reactive protein (CRP). Patients were divided into those who were hyperinflammatory (defined as CRP>17mg/L) or hypoinflammatory. Adverse outcomes, defined as requiring oxygenation, intensive care, or death, were recorded during the hospital stay. Clinical characteristics and outcomes were compared.

Results

Of the 1781 patients examined, 276 (15.5%) had a hyperinflammatory phenotype. They were older (51.8±17.2 versus 40.3±13.8 years, p<0.001), had a lower PCR cycle threshold (PCR cycle threshold value 19.3±6.3 versus 22.7±15.4, p=0.025) at presentation, and more medical comorbidities. The hyperinflammatory phenotype was independently associated with adverse clinical outcomes, even after adjusting for age, past medical history and viral load on multivariable analyses (adjusted odds ratio 5.78, 95% confidence interval 2.86 – 11.63).

Conclusions

Even in non-severe COVID-19, there are distinct hyper- and hypoinflammatory phenotypes, with the hyperinflammatory phenotype strongly associated with adverse clinical outcomes, that could be distinguished with a simple biomarker.