Evidence from clinical and epidemiological studies indicates that asthma is associated with allergic diseases including hay fever, allergic rhinitis, and eczema (Zhu et al., 2018). In 2018, Zhu et al. (2018) used the asthma and allergic disease genome-wide association study (GWAS) datasets including 33,593 cases and 76,768 controls of European ancestry from the UK Biobank and conducted a genetic correlation analysis using the linkage disequilibrium score regression (LDSC) method. They found that asthma had a positive genetic correlation with allergic diseases, which indicated shared genetic etiology (Zhu et al., 2018). In 2017, Ferreira et al. conducted a GWAS analysis of 360,838 samples with kinds of allergic disease phenotypes (Ferreira et al., 2017). They demonstrated that asthma, hay fever, and eczema share many common genetic variants (Ferreira et al., 2017). Importantly, these risk variants could regulate and dysregulate the expression of immune-related genes (Ferreira et al., 2017).

Many observational studies have been conducted to evaluate the association between smoking and asthma risk, but these studies have produced inconsistent findings for this association (Skaaby et al., 2017). Mendelian randomization (MR) analysis methods have been used widely to infer causal associations (Liu et al., 2018; Luo et al., 2019; Zhao and Schooling, 2019; Zhao et al., 2019; Zhuang et al., 2019a, b). Skaaby et al. (2017) used the rs16969968 and rs1051730 single-nucleotide variants, which were in linkage disequilibrium with each other, to genotype 231,020 individuals by observational and MR analyses. The observational analysis revealed similar asthma risk in current smokers and never smokers, and the MR analysis showed that current smokers had higher risk of asthma than never smokers (Skaaby et al., 2017).

Other MR studies using rs16969968 as an instrumental variable for smoking heaviness have reported its association with other human habits, such as coffee consumption (Bjorngaard et al., 2017) and alcohol use (Taylor et al., 2018), and factors such as depression and anxiety (Taylor et al., 2014), and blood pressure and resting heart rate (Linneberg et al., 2015). MR studies that use a single genetic variant as the instrumental variable have some limitations (Rosa et al., 2019). Therefore, an MR analysis using multiple significant and independent genetic variants as potential instrumental variables may increase the statistical power and precision, as was found in some recent studies (Relton and Davey Smith, 2015; Davies et al., 2018; Rosa et al., 2019).

Wootton et al. (2019) performed a GWAS of lifetime smoking behavior (including smoking duration, heaviness, and cessation) by analyzing a total of 462,690 samples from the UK Biobank. They identified 124 genetic variants associated with lifetime smoking that met a genome-wide level of significance of P < 5 × 10^–8 (Wootton et al., 2019). Zhu et al. (2018) conducted a GWAS of asthma with 14,085 asthma subjects and 76,768 controls of European ancestry. We used these lifetime smoking and asthma GWAS datasets to perform an updated MR analysis to evaluate the causal association between lifetime smoking and asthma risk.

Epidemiological studies indicate that asthma and allergic disease co-occur and share a similar and very close course especially in the children (Ober and Yao, 2011; Guibas et al., 2017; Schoettler et al., 2019). It is reported that allergic asthma is most common in kinds of diseases in childhood (Cookson and Moffatt, 2000). It is known that both asthma and allergic disease are complex human diseases, which are caused by both genetic and environmental contributions (Ober and Yao, 2011; Guibas et al., 2017).

Until recently, large-scale GWAS and meta-analyses of GWAS have been conducted and identified common genetic variant, risk genes, or pathways that contribute to both diseases (Ferreira et al., 2017; Zhu et al., 2018). In a recent GWAS, Ferreira et al. analyzed 360,838 samples and demonstrated common genetic variants to be shared by asthma, hay fever, and eczema (Ferreira et al., 2017). Importantly, these risk variants could regulate and dysregulate the expression of immune-related genes (Ferreira et al., 2017). In 2018, Zhu et al. (2018) analyzed 33,593 cases and 76,768 controls and identified a positive genetic correlation between asthma and allergic diseases.

However, Skaaby et al. (2017) found that smoking was significantly associated with increased risk of asthma, reduced hay fever, and reduced allergic sensitization risk by an MR analysis. These findings are inconsistent, and currently, no consensus exists. In their MR analysis, Skaaby et al. (2017) selected only one single genetic variant rs16969968 as the instrumental variable. It is reported that MR studies that use a single genetic variant as the instrumental variable have some limitations (Relton and Davey Smith, 2015; Davies et al., 2018; Rosa et al., 2019). Hence, an MR analysis using multiple significant and independent genetic variants as potential instrumental variables may increase the statistical power and precision (Relton and Davey Smith, 2015; Davies et al., 2018; Rosa et al., 2019).

Here, we conducted an updated MR analysis to investigate the causal association between lifetime smoking and asthma risk using the 124 genetic variants from the lifetime smoking GWAS dataset as the instrumental variables (Wootton et al., 2019). We first performed a pleiotropy analysis using the MR–Egger intercept test, and we found no significant pleiotropy. We further conducted a pleiotropy analysis using MR-PRESSO global test, and we found significant pleiotropy (Verbanck et al., 2018). In the MR analysis, we first selected three methods including IVW, weighted median regression, and MR–Egger methods, and we clearly demonstrated that increased lifetime smoking was significantly associated with decreased asthma risk. We further conducted an MR analysis using MR-PRESSO method and verified the significant association between increased lifetime smoking and decreased asthma risk. MR-PRESSO outlier test further supported the significant association after removing the outlier. Importantly, no significant difference was observed in the causal estimates before and after correction for outliers. Finally, the leave-one-out permutation showed that only one single genetic variant rs1050847 could significantly affect the estimates between increased lifetime smoking and decreased asthma risk.

Our results are consistent with the results from a large-scale population-based international cohort study by Cerveri et al. (2012) who analyzed 9,092 non-asthma and 1,045 asthma samples. They found that the proportion (26%) of smokers in the asthma group was significantly less than the proportion (31%) of smokers in the non-asthma group (P < 0.001; Cerveri et al., 2012). In the asthma group, smoking increased the mean asthma score (Cerveri et al., 2012). A study using mouse mast cells showed that cigarette smoke prevented allergies by decreasing the reaction of immune cells to allergens (Mortaz et al., 2009).

All datasets presented in this study are included in the article/supplementary material.

HZ designed the project. MS analyzed the data. All authors contributed to the first draft of the manuscript and read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.