The changing aetiology of head and neck squamous cell cancer: A tale of three cancers?

Head and neck cancers are the sixth most common cancers worldwide and an important cause of ill health.(1) Each year about 9000 cases are reported in the UK.(2) Survival is poor and despite advances in treatment, has not improved until recently.(3) The majority of head and neck cancers are squamous cell carcinomas, affecting the larynx, oropharynx and oral cavity. The aetiology is changing and this has implications for prevention and treatment. Twenty years ago, tobacco and alcohol consumption were considered to be the established risk factors and it was estimated that these risk factors accounted for about 75% of cases.(4) More recently human papillomavirus (HPV) has emerged as a likely cause of oropharyngeal cancer.(4) Here we argue that each cancer site is predicted differently by smoking, alcohol and sexual behaviour. We also suggest that these risk factors do not fully explain the changing pattern of disease observed in the UK. This article is protected by copyright. All rights reserved.

NIHR funded UK-based clinical cohort study of head and neck cancer called Head and Neck 5000 (H&N5000). The cohort is described in detail elsewhere. 10 Briefly, we recruited people with a new diagnosis of head and neck cancer and obtained wide ranging consent, including agreement to collect clinical and longitudinal data, to store and analyse biological samples and for ongoing record linkage. Prior to treatment, participants completed questionnaires on health and lifestyle, quality of life and sexual history. The key variables used in this analysis were anatomical site (grouped according to ICD-10 code), age, gender and participant reported current smoking status, current drinking (units per week alcohol consumed) and history of giving oral sex. Alcohol consumption was categorised using the revised UK Department of Health guidelines 11

SAME?
In the UK, the incidence of laryngeal cancer declined by nearly 20% between 1990 and 2013, but this decline has levelled off since 2002. 2 People enrolled in H&N5000 who have laryngeal cancer are predominantly older, male smokers. Half of these people consume hazardous amounts of alcohol and a quarter report that they have never given oral sex. The established risk factors of smoking (90% had smoked) and alcohol (74% currently drank alcohol) are common behaviours amongst this recently diagnosed cohort of people with laryngeal cancer (see Figure 1). This pattern of behaviours is consistent with the risk factors reported in a large international consortium of case-control studies. 13 The consortium estimated that the joint effects of tobacco and alcohol use accounted for 89% of laryngeal cancers. 13 There was a substantial decline in the prevalence of smoking in most parts of the world between 1980 and 2012. 14

| OROPHARYNGEAL CANCER-JUST A CONSEQUEN CE OF THE SEXUAL REVOLUTION?
In the last 20-years, the incidence of oropharyngeal cancer in the UK has more than trebled. This increase is most marked in younger people. People with oropharyngeal cancer in the H&N5000 study are younger, less likely to smoke and more likely to have given oral sex than those with laryngeal cancer. Ninety three per cent of people with oropharyngeal cancer had given oral sex (93% of men and 90% of women). Men were more likely than women to have had more than 5 oral sex partners (44% of men and 13% of women).
The established risk factors of smoking and alcohol consumption are common (72% had smoked and 74% currently drink alcohol) but smoking is not as common as in those with laryngeal cancer.
While transmission of HPV to the upper aerodigestive tract is thought to be through oral sexual contact the epidemiologic evidence of the role of sexual behaviour in head and neck cancer aetiology is inconsistent. A recent meta-analysis examining the association between sexual behaviour and, head and neck cancer reported no association with oral sex practices (but 12 of 17 studies combined oral cavity and oropharynx cancers). 15 A case control study that just included people with oropharyngeal cancer showed that they were more likely than controls to have ever had oral sex and to have had multiple oral sex partners. 16 The profile of sexual behaviour for people with oropharyngeal tumours is markedly different from that for people with laryngeal cancer. It suggests a sexually transmitted aetiology but there are inconsistencies that make this association more complicated. The reported prevalence of oral sex with a partner of the opposite gender increased between the first UK National Survey of Sexual Attitudes and Life-  6 Thus, recent changes in sexual behaviour do not mirror the continued rise in the incidence of oropharyngeal cancer.
Sexual behaviour may explain some of the increase in incidence of oropharyngeal cancer in the UK, it does not seem to account for all of this change. A recent study in the UK found there was no change in the proportion of HPV-attributable cases over time (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011), however, during the same period, the incidence of oropharyngeal cancer doubled and similar findings have been made elsewhere in Europe. 17 HPV has an important role in the continuing increase in incidence of oropharyngeal cancer, but this may not be explained by changes in sexual behaviour and HPV infection alone.

AGAIN?
In the UK, the incidence of oral cavity cancer has risen by 50% since 1990 (Figure 1). In Figure 1, we show that people with oral cavity cancer in the H&N5000 study differed from those with laryngeal and oropharyngeal cancer. They were younger, more likely to be female and less likely to smoke than people with laryngeal cancer, but no more likely to have given oral sex. The prevalence of the established risk factors of smoking and alcohol was similar to that observed in people with oropharyngeal cancer but again less for giving oral sex. Oral cavity cancer is not typically associated with HPV infection 15  increased. 9 Furthermore, lifetime consumption appears to be important as the greater the number of years drinking varies by site, the risk of oral and pharyngeal cancer being higher than that for laryngeal cancer. 19 So longer term trends in alcohol consumption could explain some of the observed increase in incidence of oral cancer.
Interestingly, we observed a high proportion of women in H&N5000 with oral cavity cancer and a third of these women have never smoked and a third did not currently consume alcohol and 42% had never given oral sex (Table S1). These data suggest that this increase in the incidence in oral cancer may (in part) represent an emerging and distinct clinical entity of unknown aetiology.

Keypoints
• The incidence of laryngeal cancer is falling in the UK and is consistent with the decline in tobacco consumption • The incidence of oropharyngeal cancer is rising rapidly in the UK and is associated with sexual behaviour and Human Papilloma Virus

| HEAD AND NECK CANCER-A TALE OF THREE CANCERS?
We used 3 sources to explain recent trends in the incidence of head and neck squamous cell cancers in the UK, past risk factors, recent changes in the risk factors and current exposure in a large UK cohort. The data we have are imperfect but they represent the most contemporary UK data for these cancers. Changes in smoking, alcohol and HPV infection related to sexual behaviour explain some of the observed changes in incidence of head and

Change in cancer incidence
Cancer incidence: