RACIAL DISPARITIES IN HUMAN PAPILLOMAVIRUS PREVALENCE IN HEAD AND NECK CANCER PATIENTS: AN INTERNATIONAL POOLED AND META-ANALYSIS

Abstract

Head and neck cancer (HNC) is one of the top ten cancers in the world, and is caused by tobacco use, alcohol consumption and Human Papillomavirus (HPV). HPV associated HNC patients have improved survival rates compared to non -HPV associated HNC patients. This improved survival is due to HPV- positive tumors favorable response to chemotherapy and radiation. The literature has shown that there is a racial disparity in survival rates between Caucasians and African Americans, with African Americans having poorer survival rates. The aim of this study is to determine if the racial disparity among HNC patients is due to a difference in HPV prevalence between races. HPV prevalence in HNC was assessed by a meta-analysis of published articles (30/247) that reported race specific HPV prevalence. We also conducted a pooled analysis in which authors that assessed HPV in HNC were invited to submit their datasets. Meta-pooled prevalence estimates revealed that 20% of African American HNC patients had HPV-positive tumors, compared to 44% in Caucasians. However for both African American HNC patients and Caucasian HNC patients there was low to moderate heterogeneity between the studies (Q-test p-value = p < 0.001, I2 = 18.87%, and p= 0.008. I2 =65.47% respectively). The prevalence of HPV in African Americans was 60% and in Caucasians it was 39%. African Americans had a risk of oropharyngeal cancer that was no different from Caucasians (OR: 1.38, 95% CI: 0.53-3.62) but had an increased risk of death from oropharyngeal cancer (HR: 2.39, 95% CI 1.03-5.55) compared to Caucasians. The results of the pooled analysis does not support the concept that African Americans HNC patients have a lower prevalence of HPV, but substantiates the notion that African Americans have worse survival than Caucasians. However, these are preliminary results as the pooled analysis is still being conducted, the inclusion of more datasets in the analysis could alter these preliminary findings.