Current and recent smoking have been associated with a greater risk of prostate cancer recurrence and mortality, though the underlying mechanism is unknown.

To determine if telomere shortening, which has been associated with poor outcomes, may be a potential underlying mechanism, we prospectively evaluated the association between smoking status and telomere length in 567 participants in the Health Professionals Follow-up Study, who were surgically treated for prostate cancer. Using tissue microarrays (TMA), we measured telomere length in cancer and benign tissue, specifically stromal cells in the same TMA spot using a telomere-specific fluorescence in situ hybridization assay. Smoking status was collected via questionnaire 2-years before diagnosis. Adjusting for age, pathologic stage and grade, the median and standard deviation of the per-cell telomere signals were determined for each man for stromal cells and cancer cells by smoking categories. In sub-analyses, we restricted to men without major co-morbidities diagnosed before prostate cancer.

Overall, there were no associations between smoking status and telomere length or variability in stromal cells or cancer cells. However, among men without comorbidities, current smokers and former smokers who quit <10 years ago had the most variable telomere length in stromal cells (29.3% more variable than never smokers; P-trend = 0.0005) and in cancer cells (27.7% more variable than never smokers; P-trend = 0.05). Among men without comorbidities, mean telomere length did not differ by smoking status in stromal cells or cancer cells.

Telomere variability in prostate cells may be one mechanism through which smoking influences poor prostate cancer outcomes.