Genetic causal correlations between allergic diseases and
COVID-19: A systematic two-sample and bidirectional MR study
Word count: 597
To the Editor,
Allergic diseases (ADs) such as asthma are presumed risk factors for
COVID-19 infection with greater severity.1 However,
recent observational studies suggest that the assumed correlation
remains controversial.2 We aimed to answer the
following questions: (1) Are there causal correlations between various
ADs and COVID-19 infection/severity? (2) If so, how do they affect each
other?
We employed a two-sample, bidirectional Mendelian randomization (MR)
analysis to systematically explore genetic correlations between ADs and
COVID-19 infection/severity.3 The primary MR analyses
contained three successive parts. First, we treated each COVID-19
phenotype (18,152 critically ill cases, A2; 44,986 hospitalized cases,
B2; 159,840 reported infection cases, C2) as exposure and each AD (broad
allergic disease (BAD), asthma, allergic dermatitis (ADE), shrimp
allergy (SA) and peach allergy (PA)) as outcome to evaluate whether the
COVID-19 outbreak would genetically affect the incidence rates of ADs
(Fig. S1A and S1C ). Conversely, we considered each
type of AD as an exposure and COVID-19 phenotypes as outcomes to
identify genetic risk factors for COVID-19 infection/severity
(Fig. S1B and S1C ). Secondly, we employed two
available independent GWAS datasets (Asthma2018 and ADE2021) to further
validate those significant correlations (Fig. S1D ). Finally, we
investigated the mechanisms underlying the significant correlations
based on well-known theories of pathogenesis for COVID-19 or ADs
(Fig. S1E ). After multiple pre-processing steps, we applied the
most suitable MR approach upon different scenarios, and thus a p-value
<0.05 was considered statistically significant. Details are
presented in supporting information (SI ) and Table S1
(TS1) .
Our MR results consistently indicated that all COVID-19 phenotypes were
causally associated with an increased asthma prevalence
(OR>4.21, p <1.30×10-78)
(Table 1 ). Validation results from Asthma2018 further confirmed
the causal correlation of critically ill with asthma using two different
MR methods (both OR=1.05,p <4.19×10-2) (SI-TS2 ). Both
critically ill and hospitalized cases were causally associated with a
decreased prevalence of ADE (OR<0.93,p <2.65×10-2) (Table 1 ),
although it did not obtain further validation. Interestingly,
hospitalized cases showed a significant association with an increased
rate of PA (OR=2.95, p =3.86×10-2)
(Table 1 ).
MR analyses found that asthma was a causally protective factor for
hospitalized cases (OR=0.9975, p =4.28×10-2;Table 2 ), which was confirmed by another GWAS dataset
(Asthma2018) upon adopting multiple MR methods (OR=0.87-0.98,p <5.61×10-3) (SI-TS2 ).
Besides, the MR analyses implied that SA was also a risk factor for the
COVID-19 infection/severity (OR>1.04,p <6.73×10-3) (Table 2 ).
In the investigation of underlying molecular mechanisms, MR analyses
indicated that COVID-19 phenotypes, especially severe symptoms, were
causally correlated to hematological traits (HT and MCH) and
immune-related cell counts (including LYMPH and CD3+ T, CD56+ NK, and
CD8+ T cell counts), implying that a cytokine storm induced by COVID-19
infection/severity may cause a substantial damage to host immune
response and further induce various ADs
(SI-TS3 ).4 In turn, MR analyses for causal
effects of ADs on ACE2 protein expression in peripheral blood suggested
that asthma was causally correlated with a decreased ACE2 protein
expression with two different methods (SI-TS4, both OR=0.9997,p =4.30×10-2), while SA with a marginally
increased protein expression (OR>1.01,p <8.90×10-2) (SI-TS4 ). These
findings partially explained that ADs such as asthma and SA may have
causal effects on COVID-19 infection/severity by affecting ACE2 protein
expression, which directly exerts a biological function in peripheral
blood tissue.5 Several previous studies also suggested
that allergic sensitization in asthma patients was associated with a low
ACE2 expression in respiratory tracts.6
In conclusion, our MR analyses suggest a bidirectional causal effect
between COVID-19 phenotypes and ADs, especially asthma. The underlying
molecular mechanisms of the causal effects may be beneficial in
developing effective therapeutic strategies for allergic patients with
COVID-19 infection and for long-term COVID-19 symptoms physical
characteristics should be prioritized.