Tonsillectomy and Incidence of Oropharyngeal Cancers

Abstract

Background: Rising incidence of oropharyngeal cancers in numerous countries since the 1970s has been attributed to increased oral human papillomavirus (HPV) exposure. However, the contribution of coincidental declines in the surgical removal of the tonsils (tonsillectomy) is unknown. We quantified the association of tonsillectomy with risk of tonsillar, other oropharyngeal, and other head and neck cancers and the contribution of declines in tonsillectomies to cancer incidence trends.

Methods: We conducted a nation-wide cohort study in Sweden (1970-2009). Tonsillectomies (N = 225,718) were identified through national patient registers, which were linked with the cancer register. Cancer incidence in the tonsillectomy cohort was compared with Sweden's general population through standardized incidence ratios (SIR).

Results: Tonsillectomies were associated with reduced risk of tonsil cancers [SIRs 1+ years post-tonsillectomy = 0.31; 95% confidence interval (CI), 0.08-0.79 and 5+ years post-tonsillectomy = 0.17; 95% CI, 0.02-0.62], but unrelated to other oropharyngeal or other head and neck cancers (SIRs 1+ years post-tonsillectomy = 1.61; 95% CI, 0.77-2.95 and 0.92; 95% CI, 0.64-1.27, respectively). The cumulative incidence of tonsillectomy declined significantly (40%-50%) during 1970-2009. However, tonsil cancer incidence significantly increased during 1970-2009 both without and with corrections for declines in tonsillectomies (relative risks per 5-year periods = 1.23, P < 0.001 and 1.20, P < 0.001, respectively).

Conclusions: The reduced tonsil cancer risk with tonsillectomy reflects the removal of most of the relevant tissue. The absence of associations with other head and neck cancers indicates that tonsillectomy may not impact carcinogenesis at other sites.

Impact: The significant increases in oropharyngeal cancer incidence since the 1970s in Sweden appear independent of declines in tonsillectomies, reinforcing increased oral HPV exposure as the likely cause. Cancer Epidemiol Biomarkers Prev; 25(6); 944-50. ©2016 AACR.