Background
Idiopathic pulmonary fibrosis (IPF) is a complex lung disease, characterized by progressive lung scarring. Severe COVID-19 is associated with substantial pneumonitis and has a number of shared major risk factors with IPF. This study aimed to determine the genetic correlation between IPF and severe COVID-19 and assess a potential causal role of genetically increased risk of IPF on COVID-19 severity.
Methods
The genetic correlation between IPF and COVID-19 severity was estimated with linkage disequilibrium (LD) score regression. We performed a Mendelian randomization (MR) study for IPF causality in COVID-19. Genetic variants associated with IPF susceptibility (
P
<5 × 10
−8
) in previous genome-wide association studies (GWAS) were used as instrumental variables (IVs). Effect estimates of those IVs on COVID-19 severity were gathered from the GWAS meta-analysis by the COVID-19 Host Genetics Initiative (4,336 cases & 623,902 controls).
Findings
We detected a positive genetic correlation of IPF with COVID-19 severity (rg=0·31 [95% CI 0·04–0·57],
P
= 0·023). The MR estimates for severe COVID-19 did not reveal any genetic association (OR 1·05, [95% CI 0·92–1·20],
P
= 0·43). However, outlier analysis revealed that the IPF risk allele rs35705950 at
MUC5B
had a different effect compared with the other variants. When rs35705950 was excluded, MR results provided evidence that genetically increased risk of IPF has a causal effect on COVID-19 severity (OR 1·21, [95% CI 1·06–1·38],
P
= 4·24 × 10
−3
). Furthermore, the IPF risk-allele at
MUC5B
showed an apparent protective effect against COVID-19 hospitalization only in older adults (OR 0·86, [95% CI 0·73–1·00],
P
= 2·99 × 10
−2
) .
Interpretation
The strongest genetic determinant of IPF, rs35705950 at
MUC5B
, seems to confer protection against COVID-19, whereas the combined effect of all other IPF risk loci seem to confer risk of COVID-19 severity. The observed effect of rs35705950 could either be due to protective effects of mucin over-production on the airways or a consequence of selection bias due to (1) a patient group that is heavily enriched for the rs35705950 T undertaking strict self-isolation and/or (2) due to survival bias of the rs35705950 non-IPF risk allele carriers. Due to the diverse impact of IPF causal variants on SARS-CoV-2 infection, with a possible selection bias as an explanation, further investigation is needed to address this apparent paradox between variance at
MUC5B
and other IPF genetic risk factors.
Funding
Novo Nordisk Foundation and Oak Foundation.