Smoking and the Risk of Second Primary Lung Cancer Among Breast Cancer Survivors from the Population-Based UK Biobank Study

Abstract

Objective: Long-term breast cancer (BC) survivors are known to develop second malignancies, with second primary lung cancer (SPLC) one common type. Smoking was identified as a main risk factor for SPLC among BC survivors. These findings were limited to the U.S. and focused on smoking status, not incorporating cumulative smoking exposures (eg, pack-years). We examine SPLC incidence and evaluate the associations between SPLC risk and cumulative cigarette smoking exposures and other potential factors among BC survivors in a prospective European cohort.

Methods: Of 502,505 participants enrolled in the UK Biobank in 2006 to 2010, we identified 8429 patients diagnosed with BC between 2006 and 2016 and followed for second malignancies through 2016. Smoking information was collected at enrollment, and treatment data were collected using electronic health records. Multivariable cause-specific Cox regression (CSC) evaluated the association between each factor and SPLC risk.

Results: Of 8429 BC patients, 40 (0.47%) developed SPLC over 45,376 person-years. The 10-year cumulative SPLC incidence was 0.48% (95% CI = 0.33%-0.62%). The CSC analysis confirmed the association between SPLC and ever-smoking status (adjusted hazard-ratio (aHR) = 3.46 (P < .001). The analysis showed a 24% increment in SPLC risk per 10 smoking pack-years among BC survivors (aHR = 1.24 per-10 pack-years, P = .01). The associations between SPLC and other variables remained statistically insignificant. We applied the USPSTF lung cancer screening eligibility criteria and found that 80% of the 40 BC survivors who developed SPLC would have been ineligible for lung cancer screening.

Conclusion: In a large, European cohort, cumulative smoking exposure is significantly associated with SPLC risk among BC survivors.

Keywords: Breast Cancer; Breast cancer; Second Primary Lung Cancer; Smoking.

Figure 1

Cumulative incidence of SPLC in the presence of competing death among the entire breast cancer patients in (A) and by a smoking history in (B), smoking pack years in (C), and smoking duration (years) in (D) with a 5-year and 10-year point estimate. Note: The grey or orange colored areas represent the 95% confidence interval (CI) for the cumulative incidence of SPLC over time. All smoking variables are related to cigarette smoke exposure.