Asthma is one of the most common chronic respiratory disorders (1), affecting affects over 300 million people worldwide and bringing a huge economic and social burden (2). Accumulating evidence has shown that inflammation might be involved in the pathogenesis of asthma (3) and individuals with brain inflammation have a likelihood of being predisposed to epileptogenesis (4, 5). These findings have drawn much attention to exploring the association of asthma with epilepsy. Indeed, two previously published population-based studies of adults revealed that patients with epilepsy were often accompanied by physical comorbidities such as asthma (6, 7). In addition, numerous case-control studies have announced that the prevalence of asthma was related to higher odds of epilepsy either in children (8) or in adults (9). These data suggest that asthma might be associated with high susceptibility to epilepsy. However, data from other case-control studies have displayed discordant findings, with a retrospective study among children suggesting that idiopathic epilepsy is not etiologically connected with asthma (10). Furthermore, observational studies cannot prove the causal inference due to their sensitivities to residual confounding and reverse causality.

Mendelian randomization (MR), using genetic connections to inquire about the causal impact of a risk factor on an outcome (11), is an effective method for gaging causal inference. This approach can not only limit reverse causality but also greatly reduce the likelihood of residual confounding (12). Based on the inconsistent findings of the aforementioned retrospective cohort studies, we undertook a 2-sample MR approach to assess whether asthma causally contributed to an increased risk of epilepsy.

Using the IVW method, genetically predicted asthma was associated with an increased risk of epilepsy in the discovery stage (ILAEC: OR = 1.112, 95% CI: 1.023-1.209, P = 0.012). Directional consistent results were obtained in sensitivity analyses using simple median, weighted median, maximum likelihood, and MR-PRESSO approaches ( Figure 2A ). In the replication stage, estimates of the FinnGen dataset showed the same trend direction as the results of ILAEC ( Figure 2B ). No obvious causal effects of childhood-onset asthma and adult-onset asthma on epilepsy were found in both the discovery stage and replication stage ( Figure 2 ). There was no obvious pleiotropy observed in the MR-Egger intercept test, but potential pleiotropy of childhood-onset asthma on epilepsy (P=0.037) in the discovery stage was observed in the MR-PRESSO test ( Table 2 ). Cochran-Q test also showed heterogeneity in evaluating the causal association between childhood-onset asthma and epilepsy in the discovery stage ( Table 2 ). The corrected estimate after removing the outlier (rs1893380) identified by the MR-PRESSO test showed a similar result, suggesting good stability. All the F-statistic values were larger than 10 across the MR study, indicating good instrumental strength.

A further meta-analysis of ILAEC and FinnGen also showed a causal effect of asthma on epilepsy (OR = 1.085, 95% CI: 1.012-1.164, P = 0.022), which was validated in a sensitivity analysis using other approaches ( Figure 3 ; Supplementary Figure S1 ). The meta-analysis results from both the fixed-effect model and the random-effect were largely consistent across different statistical methods ( Figure 3 , Supplementary Figure S1 ).

In this study, we took advantage of the 2-sample MR method to analyze the causal relationship between asthma and epilepsy. The main results consistently suggested that asthma was associated with a higher risk of epilepsy. Furthermore, several sensitivity analyses were used based on their different underlying assumptions and similar results were observed, which further strengthened the credibility of the results.

Previous reports have investigated the relationship between asthma and epilepsy, but the results were inconsistent. A population-based study found that most adult patients with epilepsy presently have symptomatic asthma (6). Meanwhile, a U.S. National Health Interview Survey found that adult patients with epilepsy were more often to record physical comorbidities like asthma (7). Previous studies among US children aged 0-17 years reported that the lifetime prevalence of asthma was related to a higher risk of epilepsy (2.30 [1.50-3.52]) (8). Similarly, a recent cohort study including 150,827 asthma patients showed that the asthma patients had an increased risk of epilepsy than health controls (hazard ratio=1.39) (9). All these findings indicated that asthma was associated with the risk of epilepsy, which was consistent with the results of our MR study based on data from the ILAEC and FinnGen Consortium. Although an early study among children suggested that there was no etiological relationship between asthma and epilepsy, the result may be attributed to small samples (10).

The underlying mechanism mediating the association between asthma and epilepsy remains largely unknown. The potential reasons connecting asthma and epilepsy are anoxia and hypocapnia owing to repeated asthma attacks. In addition, chronic inflammation is a common pathological feature shared by asthma and epilepsy (3, 32, 33). Previous studies demonstrated that circulating cytokines might penetrate through the blood-brain barrier and then result in chronic neuroinflammation and neuronal damage, eventually increasing the susceptibility to epileptogenesis (4, 5, 34). Moreover, emerging evidence shows that the respiratory system has a tight relationship with the central nervous system, which goes beyond the classically known connections such as blood supply and oxygen saturation. Studies showed that respiratory system diseases such as asthma (35) and chronic obstructive pulmonary disease (36) might increase the risk of stroke, which was a risk factor for epilepsy. In addition, clinical data suggested that chronic obstructive pulmonary disease was associated with an increased risk for the development of seizures in patients with stroke (37). Although oxygen desaturation may be one of the risk factors for epilepsy in asthma patients (38), further work is needed to explore the exact mechanisms by which asthma causes an increased risk of epilepsy.

Asthma can be divided into childhood-onset asthma and adult-onset asthma based on the age of onset. Childhood-onset asthma may be related to genetic factors (39, 40), perinatal factors (41), or respiratory infections (42), while adult-onset asthma may be related to environmental and occupational factors such as obesity and smoking (43). Even though the mechanisms contributing to childhood-onset and adult-onset asthma might be different, our MR study found no causal associations between the age onset of asthma and epilepsy. These data suggested that asthma causally increased the risk of epilepsy independent of the age onset of asthma. The potential reason for these unexpected results might be due to the small sample size of childhood-onset and adult-onset asthma, which might lead to lower statistical power. It is worth noting that the proportion of cases with asthma was 13.8%, while in childhood-onset asthmatic and adult-onset asthma were 4.4% and 8.1%, respectively. In addition, although the causal relationship was not significant for childhood-onset asthma and adult-onset asthma on epilepsy, most of the OR values were larger than 1, suggesting a potential risk effect of childhood-onset asthma and adult-onset asthma on epilepsy.

This study has some limitations: first, the nonlinear connection between asthma and the risk of epilepsy cannot be eliminated due to the linear effect assumption in MR analysis. Second, although no evidence of pleiotropy was detected in the MR-Egger intercept test, potential pleiotropy was observed between childhood-onset asthma and epilepsy (P=0.037) in the MR-PRESSO test. Third, there was obvious heterogeneity between childhood-onset asthma and epilepsy from ILAEC datasets, which might be due to the mixed population of ILAEC (531 and 147 individuals of Asian and African descent, respectively). Fourth, our study is mainly based on Europeans, thus generalization of the findings to other ethnic groups needs to be cautious. Fifth, to better fulfill the independence assumption for the MR study, we used a relatively stringent way to exclude the SNPs associated with potential confounders of epilepsy from the IVs, which might weaken the statistical power of the MR study. Sixth, due to individual data not being publically available, we were unable to properly account for the potential sample overlap between the GWAS datasets of asthma and epilepsy, which might lead to bias in the overall estimates. Finally, there are other possible unmeasured and residual confounding factors like many other epidemiological studies, which might drive the bias of the overall estimates. For example, as asthma was caused due to an overactive immune response (44), many instrumental variables for asthma were associated with peripheral blood cells ( Supplementary Data Sheets , 7 ). Although previous studies suggested that inflammatory factors were also implicated in epilepsy (33), however, asthma, characterized by chronic inflammation and bronchial hyperresponsiveness, is a disease strongly related to the inflammatory response (45). If all instrumental variables related to peripheral blood cells were excluded, the number of instrumental variables would be dramatically reduced. Thus, like other previously published MR studies on asthma (46, 47), the SNPs related to peripheral blood cells were not removed from the instrumental variables in our MR study, which could not rule out the potential influence of inflammatory factors on the causal relationship between asthma and epilepsy.

In conclusion, the present MR study suggests that asthma is associated with an increased risk of epilepsy independent of the age onset of asthma. Further studies are warranted to investigate the potential mechanism mediating the causal effect of asthma on epilepsy.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

PT, XG, and RL conceived and designed the project. PT, XG, and LC collected and analyzed the data. XG and PT drafted the manuscript. RL revised the manuscript. All authors approved the final version of the manuscript.