4 Discussion
This is the first comprehensive bidirectional and multivariable MR analysis to investigate the potential causal link between allergic disease and COVID-19. Our results revealed that allergic diseases have a protective causal association with severe COVID-19, including self-reported asthma and/or hay fever (or AR) and/or eczema (or AD). Though the heterogeneity test indicated a minor problem, heterogeneity is pervasive across MR analyses25 and the existence of heterogeneity does not render an MR study inadmissible when horizontal pleiotropy is absent21. The expression level of angiotensin-converting enzyme 2 (ACE2), the receptor for SARS-CoV-2, decreases with the increase of environmental allergens, which may be one of the reasons for protecting allergic patients from severe COVID-1926.
However, SARS-CoV-2 infection was not found to be causally associated with allergic diseases, and severe COVID-19 did not in turn lead to an increase or decrease in the risk of the allergic disease. The Korea National Health and Nutrition Examination Survey also did not show a significant reduction in the incidence of each allergic disease (asthma, AD, AR), whether self-reported or physician-diagnosed, in 2020 compared with 201927. Although COVID-19 does not cause allergic diseases, viral infections (not just SARS-CoV-2) may exacerbate the symptoms of allergic diseases, so telemedicine is still advised during the COVID-19 pandemic28.
Further analysis of various types of allergic diseases and asthma showed that AD and asthma (especially mixed asthma and childhood asthma) were causal protective factors for severe COVID-19. AD, a form of eczema, is an allergic skin disease often related to asthma, food allergy, allergic conjunctivitis, and AR29. In observational studies, the relationship between AD and COVID-19 remained murkier: some studies30-32 suggested that AD was associated with increased risk for COVID-19, while others suggested that it was associated with reduced risk33-35 or had no effect36. Cohort studies have shown a reduced risk of SARS-CoV-2 infection in patients with AD treated with Dupilumab35. This may be attributable to the fact that Dupilumab lowered the incidence of severe infections, such as herpes simplex and skin infections37, and thus SARS-CoV-2. In addition, Dupilumab was also found to reduce the risk of SARS-CoV-2 infection in patients with asthma38.
A meta-analysis concluded that asthma was not associated with higher SARS-CoV-2 infection or a worse prognosis and that patients with asthma had lower mortality than those without asthma39. Previous MR analysis studies have also suggested that asthma was a protective factor for SARS-CoV-2 infection and severe COVID-197,8. Although similar results were obtained only in the multivariate MR analysis with AR and AD as references, the two-sample MR analysis of asthma-severe COVID-19 in this study was negative. After comparing the differences between this study and these two previous studies, three possible reasons for the different results were found. First, limited by research time, the study that suggested asthma was a protective factor for SARS-CoV-2 infection8 used a relatively old COVID-19 database (the COVID-19 Host Genetic Initiative GWAS meta-analyses round 4, released on October 20, 2020), when the COVID-19 pandemic was just beginning, and the amount of GWAS data for COVID-19 patients was far from enough. Second, ”COVID-19 infection” and ” COVID-19 hospitalization” were chosen as the outcomes in these two MR studies, instead of ”very severe respiratory confirmed COVID-19” as in our study. Differences in the selection of outcomes may have contributed to the differences in results. Thirdly, the choices of LD value in these two studies were more liberal than that in our study, and both of them chose r2<0.01. Our choice (r2<0.001) was 10 times more stringent. These factors are all optional and acceptable. Given the positive result of multivariate MR analysis after adjusting for AD and AR and the further analysis of asthma subtypes, we still consider asthma a protective factor for severe COVID-19.
Both mixed asthma and childhood asthma have causal protective effects against severe COVID-19, with childhood asthma playing a major role. To our knowledge, this study is the first to suggest that mixed asthma is a protective factor against severe COVID-19. Mixed asthma was described in the FinnGen database as the “combination of conditions listed in predominantly allergic asthma and nonallergic asthma”. Conditions listed in “predominantly allergic asthma” included allergic (bronchitis, rhinitis with asthma), atopic asthma, extrinsic allergic asthma, and hay fever with asthma. Symptoms of “nonallergic asthma” included idiosyncratic asthma and intrinsic nonallergic asthma. We hypothesized that the development of mixed asthma is involved in the interaction of exogenous anaphylaxis and endogenous infection, leading to a more active asthma-related immune response. However, the specific pathophysiological reasons still need to be further explored.
Childhood asthma was considered a major protective factor against severe COVID-19 in our study, and there have been many reports on childhood asthma and severe COVID-19. In Spain, children were thought to usually develop mild COVID-1940. Clinical observations indicated that allergies or asthma were not hazardous conditions in pediatric patients with COVID-1941. A meta-analysis of COVID-19 patients in children and young people found that patients with asthma were less likely to be admitted to critical care and less likely to die42. Data from a pediatric referral hospital indicated that the prevalence of asthma in pediatric patients with COVID-19 was low, with varied clinical manifestations and laboratory findings40. New confirmed cases of childhood asthma in Japan dropped significantly after the COVID-19 pandemic began, and 15 months later they have not recovered43. The impact was particularly strong for younger children. New diagnoses of atopic dermatitis also fell slightly43. As mentioned above, no significant decrease was shown in the incidence rates of allergic diseases (asthma, AD, AR) in Korea27. However, data from the nationally representative Korean Adolescent Risk Behavior Survey presented that the prevalence of allergy among Korean adolescents increased before 2019, but decreased significantly in 202044. Regardless of prior trends, the prevalence of three allergic diseases, asthma, AD, and AR, all decreased in adolescents in 202044. All of these studies showed an association between COVID-19 and asthma in children.
Our study has several strengths. Firstly, large datasets covering multidimensional phenotypes were used, and the F-statistics were also large enough to prevent any weak instrumental bias. Second, compared with traditional observational studies, MR analysis is usually less affected by confounding factors and reverse causality, leading to a higher level of evidence. Third, the sample used was largely derived from populations of European ancestry, which minimized stratification bias. In addition, the pleiotropy that IVs do not have also illustrates the robustness of this study. Our study also has certain limitations, such as uneven weighting among phenotypes and a relatively small number of cases in some phenotypes. Moreover, most of the included populations were of European ancestry, which does not represent the general population, and verification of the results in populations with different ancestries is required.
In summary, this study, based on the population genetic variation model, pointed out the protective effect of allergic diseases against severe COVID-19. More specifically, AD and asthma (both mixed and childhood asthma) were protective, but childhood asthma played a major role. This protective effect may come from the persistent inflammatory response and ACE2. More research is still needed to figure out why the protective effect is stronger in children than in adults.