Atopic status
Fifty-seven participants (10.7% of the study population) were atopic and showed a history of allergy to seasonal and/or perennial airborne allergens (either rhino-conjunctivitis or bronchial asthma). Of these, 32 (56.1%) were females (mean age 62±14, range 28-93 years). Atopic subjects were younger than non-atopic patients were (69±14 years, p = 0.01), and showed a lower prevalence of coronary heart disease (p <0,050). Notably, atopic subjects showed a much lower occurrence of severe or very severe COVID-19-related pulmonary involvement (33.3% vs 67.7% in non-allergic individuals, p <0.0001), and therefore a milder disease course. The protective effect of atopic status against severe lung disease was evident throughout all the age subset evaluated, particularly in patients below 60 years of age (p = 0.001) (Figure 1). Intriguingly, a reduced cumulative incidence of death was observed among atopic subjects (8.8% vs 18.9%; p = 0.059), albeit it did not reach statistical significance due to the small number of observations. The multiple logistic regression analysis was first performed including as potential confounders age, sex, smoking, and all the comorbidities. Since none of the comorbidities did contribute to the estimates, we excluded them from the final model, while we decided to still keep age. The adjusted estimate confirmed a significant association between atopic status and reduced severity of COVID-19: in fact, while adjusting for sex, age, and smoking status, non-atopic patients had a significantly higher risk of having severe Covid-19 (ORadj 3.0, 95% confidence interval 1.6-5.7, p = 0.001) (Table 1).
An unsupervised two-way hierarchical clustering analysis yielded several original patterns of associated clinical features in our population. The simultaneous evaluation of distinct comorbidities at the single patient level generated by the multiple logistic regression analysis identified a clear cluster between mild COVID-19-related pneumonia and atopy status, which were strictly associated (p <0.0001; ORa= 4.523, 95% CI= 2.221-9.221) (Figure 2).