The relationship of airflow limitation with lung squamous cell carcinoma: evidence from mendelian randomization analysis.

BACKGROUND

Observational studies showed associations between smoking, and airflow limitation, with lung squamous cell carcinoma (LUSC). However, the causal association of airflow limitation with LUSC and the modification by smoking status for the association remains unclear.

METHODS

Genetic summary data were obtained from large genome-wide association studies (GWAS). One hundred two single nucleotide polymorphisms (SNPs) for airflow limitation (i.e., FEV1/FVC < 0.7) and 153 SNPs for smoking behavior were used as instrumental variables and the main MR analysis methods. The univariable and multivariable Mendelian Randomization (MR) in a two-sample setting were performed to assess the association of airflow limitation, and smoking behavior with LUSC.

RESULTS

In the univariable MR analysis, genetic predisposition towards airflow limitation [Inverse Variance-Weighted (IVW) method Odds Ratio (OR) = 4.83, 95% Confidence Interval (CI) 1.55 to 15.06, P = 0.006], age of smoking initiation (IVW method OR = 0.10, 95%CI 0.02 to 0.36, P < 0.001), cigarettes smoked per day (IVW method OR = 3.10, 95%CI 2.07 to 4.63, P < 0.001), ex-smoking (IVW method OR = 0.47, 95%CI 0.31 to 0.69, P < 0.001), current smoking status (IVW method OR = 13.08, 95%CI 2.53 to 67.84, P = 0.002), pack-years of smoking (Weighted median method OR = 11.49, 95%CI 3.71 to 35.63, P < 0.001) were associated with LUSC. In the multivariable MR analysis, the causal effect of airflow limitation was still observed on LUSC (IVW method OR = 2.97, 95% CI 1.09 to 8.04, P = 0.032 adjusted for age of smoking initiation and cigarettes smoked per day; IVW method OR = 3.24, 95% CI 1.09 to 9.58, P = 0.033 adjusted for ex-smoking, current smoking status, and pack years of smoking; IVW method OR = 2.91, 95% CI 1.01 to 8.41, P = 0.049 adjusted for 5 smoking behaviors mentioned above).

CONCLUSIONS

Our MR analysis demonstrated that airflow limitation is likely to be an independent predictor of LUSC.