The role of immune suppression in COVID-19 hospitalization: clinical and epidemiological trends over three years of SARS-CoV-2 epidemic

Abstract

Specific immune suppression types have been associated with a greater risk of severe COVID-19 disease and death. We analyzed data from patients >17 years that were hospitalized for COVID-19 at the "Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico" in Milan (Lombardy, Northern Italy). The study included 1727 SARS-CoV-2-positive patients (1,131 males, median age of 65 years) hospitalized between February 2020 and November 2022. Of these, 321 (18.6%, CI: 16.8-20.4%) had at least one condition defining immune suppression. Immune suppressed subjects were more likely to have other co-morbidities (80.4% vs. 69.8%, p < 0.001) and be vaccinated (37% vs. 12.7%, p < 0.001). We evaluated the contribution of immune suppression to hospitalization during the various stages of the epidemic and investigated whether immune suppression contributed to severe outcomes and death, also considering the vaccination status of the patients. The proportion of immune suppressed patients among all hospitalizations (initially stable at <20%) started to increase around December 2021, and remained high (30-50%). This change coincided with an increase in the proportions of older patients and patients with co-morbidities and with a decrease in the proportion of patients with severe outcomes. Vaccinated patients showed a lower proportion of severe outcomes; among non-vaccinated patients, severe outcomes were more common in immune suppressed individuals. Immune suppression was a significant predictor of severe outcomes, after adjusting for age, sex, co-morbidities, period of hospitalization, and vaccination status (OR: 1.64; 95% CI: 1.23-2.19), while vaccination was a protective factor (OR: 0.31; 95% IC: 0.20-0.47). However, after November 2021, differences in disease outcomes between vaccinated and non-vaccinated groups (for both immune suppressed and immune competent subjects) disappeared. Since December 2021, the spread of the less virulent Omicron variant and an overall higher level of induced and/or natural immunity likely contributed to the observed shift in hospitalized patient characteristics. Nonetheless, vaccination against SARS-CoV-2, likely in combination with naturally acquired immunity, effectively reduced severe outcomes in both immune competent (73.9% vs. 48.2%, p < 0.001) and immune suppressed (66.4% vs. 35.2%, p < 0.001) patients, confirming previous observations about the value of the vaccine in preventing serious disease.

Keywords: COVID-19; COVID-19 vaccination; SARS-CoV-2; disease outcome; hospitalization; immune suppression.

Figure 1

Temporal trends in patient characteristics throughout the study period. The graph in (A) represents the proportion of immune suppressed subjects among hospitalized patients. The bar graph (scale on the left) shows the number of hospitalized patients during each month while the dotted line (scale on the right) corresponds to the proportion of immune suppressed patients (for each timepoint the data of three months – the indicated timepoint ± 1 month – were used) with the confidence interval indicated by vertical lines. Timepoints corresponding to key events regarding vaccination or variant circulation are indicated by arrows. The graph in (B) shows the relative proportion of patients belonging to the three indicated age classes at the same three-month timepoints. The graph in (C) illustrates the proportions of patients hospitalized with no co-morbidities or co-morbidities in one considered category and of subjects with co-morbidities in more than one considered category at the same three-month timepoints; proportions calculated both considering and excluding immune suppression as a co-morbidity category are shown.

Figure 2

Potential predictors of severe outcomes. Odds ratios (dots) and 95% confidence intervals (lines) are indicated for each potential predictor of severe outcome when evaluated separately (simple logistic regression, dotted line) or after adjusting for the other variables (multiple logistic regression, continuous line). Results are labeled according to statistical significance as indicated in the legend. The vertical line indicates the cut-off for statistical significance.