A study that involved the Translational Immunology research group of Vall d'Hebron Research Institute (VHIR) and the Immunology Service of Vall d'Hebron University Hospital has classified patients with COVID-19 into three groups according to their immunological response to infection and has demonstrated its ability to predict the clinical progression of the disease. Stratifying patients at the time of hospital admission according to the risk of developing severe disease may facilitate the administration of appropriate treatments for each case. The results were published in Nature Communications.

The work, led by the Erasmus University Medical Center Rotterdam (The Netherlands), was born with the aim of identifying mechanisms related to the severity of COVID-19. In this sense, a machine learning system was used to find a possible relationship between the immune response and the progression of the disease. Specifically, patterns of proinflammatory, anti-inflammatory and antiviral cytokine levels in blood and the generation of antibodies against SARS-CoV-2 were studied. The trial was initially conducted with 50 patients from the Netherlands and the results were validated in a sample of 88 patients from Vall d'Hebron University Hospital.

The researchers identified three types of profiles according to the acute response of the patients' immune system to SARS-CoV-2: an excessive inflammation immunotype (EXI), a low antibody immunotype (LAI) and a balanced response immunotype (BRI).

Compared to uninfected patients, all three profiles had high levels of proinflammatory cytokines, but there were relevant differences between them. For example, the BRI types were characterized by lower levels of proinflammatory cytokines and a robust antibody response against SARS-CoV-2. EXI types had a much more inflammatory profile and high antibody levels. Finally, LAI types had high levels of the cytokine IFN-alpha but low antibody activity, indicative of a delayed response over time.

"We found that hospitalized patients with EXI and LAI profiles tended to have a more severe progression and required longer hospitalization in the hospital and ICU than in the case of BRI", explains Dr. Ricardo Pujol, head of the Translational Immunology group at VHIR during the course of the study and current advisor in immunology at the Vall d'Hebron Institute of Oncology (VHIO). It is also remarkable that no patient with BRI profile died.

Therefore, it is demonstrated that the identification of patients with EXI or LAI profiles allows predicting, at the time of hospital admission, the severity of the disease. In the case of patients with EXI profile, therapies against some cytokines could be used to regulate the immune response. On the other hand, patients with LAI could benefit from the administration of antibodies against SARS-CoV-2.

However, an easier algorithm to enable rapid classification of patients on admission has not yet been generated, and work is underway to generate one. A prospective study is also being completed - the one published so far is retrospective - with the Germans Trais and Bellvitge Hospitals to validate and extend the findings. This second study is led by ICS hospitals with the support of the Erasmus Hospital group (Prof. Peter Katsikis) for the generation of the classifier algorithm, a task in which the Statistics and Bioinformatics Unit of VHIR is involved.