Skip to main content

Longitudinal analyses reveal immunological misfiring in severe COVID-19

Our take —

This study tracked immune responses among 113 COVID-19 patients over time (up to 51 days after symptom onset), identifying distinct disease courses that were associated with inflammatory “signatures.” The magnitude of early inflammatory cytokine response correlated with disease outcomes, and those with severe disease had persistently high type 1 (antiviral) and type 3 (antifungal) inflammatory markers, and showed increases in type 2 (anti-helminth) markers. Levels of specific inflammatory markers early in disease may have prognostic value, though the study did not account for comorbidities that may also affect inflammation.

Study design

Prospective Cohort

Study population and setting

This study included 113 patients with moderate (n=80) or severe (n=33) COVID-19 disease who were admitted to a single hospital in New Haven, CT from March 18 to May 27, 2020. Patients were defined as moderate or severe depending on supplemental oxygen requirements and admission to ICU. As healthy controls, 108 volunteer healthcare workers were also included. Patients were sampled over multiple (1-7) time points over 3 to 51 days after symptom onset; measurements were taken of viral RNA using RT-qPCR of nasopharyngeal swabs, concentrations of plasma cytokines and chemokines, and profiles of leukocytes using flow cytometry on peripheral blood mononuclear cells.

Summary of Main Findings

COVID-19 patients had higher levels of T-cell activation and lower counts and frequencies of CD4+ and CD8+ cells than controls, as well as higher levels of monocytes, low-density neutrophils, and eosinophils that correlated with disease severity. The authors observed a core inflammatory response shared by COVID-19 patients defined by a cluster of cytokines with elevated concentrations: IL-1α, IL-1β, IL-17A, IL-12p70, and IFNα. After ten days following symptom onset, this core response persisted in those with severe disease but waned in those with moderate disease. Types 1, 2, and 3 immune response biomarkers continued to increase over time, broadly speaking, for patients with severe disease. Another cluster of elevated inflammatory biomarkers was observed in patients with severe disease: thrombopoietin (TPO), IL-33, IL-16, IL-21, IL-23, IFNλ, eotaxin, and eotaxin 3. Patients with moderate and severe disease had similar levels of SARS-CoV-2 RNA immediately after symptom onset, but those with moderate disease showed a decline over time, while those with severe disease did not. Concentrations of cytokines and interferons were correlated with SARS-CoV-2 viral loads. In an unsupervised clustering analysis of immune markers early in the disease course, three patient clusters correlated to disease outcomes were identified; these were related to four distinct immunologic signatures before 12 days after symptom onset. Early plasma concentrations of inflammatory markers such as IFNα, IFNλ, and IL-1Ra were associated with severe disease outcomes including death. In a clustering analysis using all longitudinal time points, severe disease was associated with immune signatures whose biomarkers overlapped considerably with the “core” cluster described above, and with those identified early in the disease course. Overall, those with moderate disease showed declines over time in type 1 and type 3 immune responses; those with severe disease showed sustained elevated type 1 and type 3 immune responses and increases in type 2 immune responses.

Study Strengths

This study longitudinally sampled laboratory-confirmed COVID-19 patients over multiple time points, and measured a large number of plasma immune biomarkers along with SARS-CoV-2 RNA. The authors were able to associate individual markers and clusters of markers with disease progression and outcomes over time. The definition of healthy controls required a negative result for both RT-PCR and antibody testing.


Comorbidities were not considered as possible confounding variables; these may have considerably influenced concentrations of inflammatory biomarkers. Patients were from a single institution and the sample size was not particularly large; it is unclear how generalizable these results are. Timing of sample collection and frequency of follow-up was not uniform.

Value added

Although many publications have examined immune responses to COVID-19, this is the most detailed study thus far to track immune responses longitudinally and link them to disease outcomes.

This review was posted on: 9 October 2020