A new approach to defining cancer risks
To quote one of New York’s most famous adopted sons, Bob Dylan, and one of his most socially challenging lyrics – the times they are a changin’
Change, socioeconomics and indeed New York are relevant to this update on oral cancer risk. How we define oral cancer is changing. This is important not only for how the disease is managed and its prognosis but also in terms of the changing risk profile and factors associated with oral cancer, which in turn is essential for prevention. Oral cancer is increasingly falling into two distinct diseases: oral cavity cancer (OCC) – “mouth cancer” and oropharyngeal cancer (OPP) – “throat cancer”. Although the tumours do not always recognise such clear-cut boundaries and more often overlap the anatomical sites (particularly in the retromolar trigone – behind the wisdom teeth).
While oral cavity cancer rates are either stable or only marginally increasing, oropharyngeal cancer is the most rapidly increasing cancer in Scotland – with a threefold increase in incidence among men in the last decade, and a 2.5-fold increase among women. Oropharyngeal cancer increases are now greater than malignant melanoma, adeonocarcinoma of the oesophagus and cervical cancer 1. In 2013 there were 494 cases of oral cavity cancer and 343 cases of oropharyngeal cancer diagnosed in Scotland 2. This changing trend of flat-lining oral cavity cancer and increasing in oropharyngeal cancer is a global phenomenon and has been related to changing population risk factors – described as “controlling a tobacco epidemic while a human papillomavirus epidemic emerges” 3.
When we talk about risk we are talking about probability – the chance that an event/the disease/oral cancer diagnosis will occur. Patients and the public may view risk as completely random chance – the flip of a coin, 50/50, one in two, 50 per cent chance, it will either happen to me or it won’t.
A man’s lifetime risk of developing oral cavity cancer in Scotland is higher – about one in 42 men – but his risk of developing oral cavity cancer (at a younger age) by the age of 64 is about one in 135
However, risk estimates for cancer can be determined by undertaking studies on large groups of people, which identify the probability that an individual or group will develop the disease over a period of time. These studies also identify risk factors – characteristics or behaviours that are associated with increased risk. We generally define risk in two ways: absolute risk and relative risk.
This is the numeric chance or probability of developing oral cancer during a specified period of time. The Scottish Cancer Registry computes this over a whole lifetime. The absolute risk of developing oral cavity cancer in Scotland in a lifetime is estimated at 1.7 per cent, or to put it another way – about 1 in 59 persons will develop oral cavity cancer at some time in their lives. For comparison, about one in 12 persons in Scotland will develop lung cancer in their whole lives, while about one in 2.5 persons will develop any type of cancer. These lifetime risks have a lot to do with other factors such as gender and age. A man’s lifetime risk of developing oral cavity cancer in Scotland is higher – 2.4 per cent, or about one in 42 men – but his risk of developing oral cavity cancer (at a younger age) by the age of 64 is 0.7 per cent, or about one in 135 men 2.
This is a comparison or ratio rather than an absolute value. It provides an estimate of the relationship between a risk factor and outcome by comparing the number of cases in a group of people with a particular trait or behaviour with the number of cases in a (otherwise similar) group of people who don’t have that trait or behaviour. The risk of oral cavity cancer for people who smoke has been estimated at around 5.8 times higher than for those who don’t smoke – the relative risk is 5.8. Relative risk is also presented as a percentage. In the same example the risk of oral cavity cancer is 580 per cent higher than in those who don’t smoke 4.This percentage over 100 per cent and lack of an upper limit in relative risk estimates is counter-intuitive. Most people would think 100 per cent is the highest possible risk. But 100 per cent equates to a doubling of risk associated with a risk factor, while 200 per cent to a tripling of risk estimate.
Population attributable risk
This is another way of expressing relative risk at the population level. It is the difference in the rate of disease between a population exposed to a risk factor and a population not exposed to the risk factor. It is more commonly used in public health policy decisions, where the burden of the disease reduction can be calculated by “hypothetically” removing the risk factor in question.
The most comprehensive and up-to-date data on oral cancer risk can be found from research by the INHANCE (International Head And Neck Cancer Epidemiology) Consortium (www.inhance.utah.edu). It was established in 2004 as a collaboration of researchers from around the world-leading large epidemiology studies of head and neck cancer to improve the understanding of the causes, risks and mechanisms of head and neck cancer.
The consortium includes data on 25,500 patients with head and neck cancer (including oral cavity cancer, oropharyngeal cancer, and larynx cancers), and 37,100 controls who did not have these cancers, from 35 studies from across the world. Overview papers have been published which detail INHANCE methods 5 and research findings 6.
It was a privilege to be invited to participate in the 12th Annual INHANCE consortium meeting in May this year. We convened at the Icahn School of Medicine at Mount Sinai New York City, on the weekend the new Freedom Tower was opened. During our meeting, we reflected on over 10 years of INHANCE research, while at the same time looking forward to taking on the ongoing challenges of the increasing and changing burden of head and neck cancer.
Oral Cancer Risk
It is well recognised that smoking tobacco and heavy alcohol consumption are the main risk factors for oral cancer. INHANCE provides us with an opportunity to understand this risk better, including providing precise estimates of risk, understanding the joint tobacco-alcohol effects and the dose-response, as well as investigating the risk associated with smokeless tobacco and the benefits of quitting both smoking and alcohol.
INHANCE provides sufficient numbers of people who never smoked or drank alcohol – thereby avoiding the problems of confounding – to identify true and precise risk estimates. Among “never” alcohol drinkers, cigarette smoking was associated with a two-fold increased risk of oral cavity and oropharnx cancers 7. And heavy alcohol drinking (three or more drinks per day vs never drinkers) among those who never used tobacco was also linked to increased risk but only among heavy alcohol consumers. However, it should be noted that those who reported never smoking and never drinking alcohol may differ in other ways from the wider population. The complexity of the relationship with smoking and alcohol was also unpicked in the estimates of the population attributable risk (PAR) for tobacco and alcohol of 64 per cent, made up of 0 per cent for alcohol alone, 24 per cent for tobacco alone, and 40 per cent for tobacco and alcohol combined 8 – but remember this risk description is mainly relevant at the population rather than individual risk level.
In terms of the dose-response relationship, risk for oral cancer increases with increased frequency and duration of both smoking and alcohol. However, fewer cigarettes per day over a longer period of time was worse (gave a greater risk for oral cancer) than more cigarettes per day over a shorter period of time. This contrasts with alcohol consumption which found that higher intake over a shorter period of time was worse (gave a higher risk for oral cancer) than a lower intake for a longer time 9. Moreover, there are no safe low-intake levels associated with negligible risk – so, the old sayings: “everything in moderation including moderation itself” or “a wee bit of what you fancy does you good” unfortunately do not hold up here.
Smokeless tobacco in the form of snuff (powdered tobacco) and tobacco chewing are not safe harm reduction alternatives some might want you to believe, with both associated with slight increased risk for oral cavity cancer. But good news does exist in the form of the benefits of quitting – with benefits appearing immediately (one to four years) after stopping smoking, and equating with those who never smoked after 20 years of quitting. The risk effects associated with heavy alcohol consumption last a bit longer with benefits of quitting taking 20 years to emerge 10.
The promise of the breakthrough in identifying genetic variants as strong markers of increased oral cancer risk has yet to fully materialise. Considerable research effort has found slight increased risks for oral cancer associated with the presence of genetic variants involved with alcohol metabolism, DNA repair pathways, and genes involved in the metabolism of nicotine 6. Similarly, there is limited evidence of dietary risk factors for oral cancer beyond confirming the protective effects associated with diets high in fresh fruit and vegetables (with approximately five or more fresh fruit and vegetable portions per day conferring a 50 per cent lower risk than those consuming low levels 11.
The INHANCE work which our team in Glasgow have led on is in relation to socioeconomic inequalities and determinants of oral cancer risk 12. You can see from the reference list with 75 co-authors the extent of the international collaborative effort involved. In our analysis we identified increased risks for oral cancer associated with low education and income relative to those in higher socioeconomic positions which were not explained by smoking or alcohol consumption, i.e. there were socioeconomic effects operating in two ways both in influencing risk behaviours (the causes of the causes) and also more direct or explained effects from low socioeconomic circumstances to oral cancer risk. Moreover these socioeconomic effects are of a similar magnitude (twofold increase) to risks associated with smoking and alcohol, and are strongest in countries where income inequalities are widest.
Finally, no discussion about oral cancer, and certainly not oropharyngeal cancer, can be complete without talking about sex and oral HPV (human papillomavirus) infection. Oral HPV is mainly associated with oropharyngeal cancer risk, with up to 80 per cent of cases having HPV identified. There are over 200 HPV types, but as for cervical cancer, HPV16 and 18 subtypes are the main high-risk oncogenic types. Increased risk for oropharyngeal cancer has been estimated as high as 15 times greater in those with oral HPV16 infection in the ground breaking New England Journal of Medicine paper 13. However, the natural history of oral HPV infection is not well understood (in terms of prevalence, persistence, and determinants). The only large epidemiological study of oral HPV prevalence has been undertaken in US 14. They found a prevalence of 7 per cent, with slightly greater peaks (around 10 per cent) among 25-30 and 50-55-year-olds, and among men. Risk factors identified included smoking and alcohol, number of sexual partners/oral sex partners, but also open mouth kissing. We are currently completing a feasibility study to undertake a similar study in dental practices in Scotland – HOPSCOTCH (HPV Oral Prevalence in Scotland) study ( http://www.sohrc.org/projects/hopscotch/ ). We are grateful for the outstanding support that we have received from dental practices and teams across Scotland in stepping up to this important research area, and we look forward to disseminating our feasibility study findings and taking forward a full population study in dental practices in due course.
INHANCE studies also point to a slight increased risk for oral cancer associated with six or more lifetime sexual partners, four or more lifetime oral sex partners, and early age (≤18) of sexual debut 15. However, it is worth noting again that this research and our understanding is at a far earlier stage, perhaps several decades behind our knowledge of the role of HPV in cervical cancer.
Communicating risks for oral cancer is not straightforward. I had a go at trying to explain the oral cancer risks associated with alcohol drinking on the BBC Radio 4 statistics programme More or Less [ http://www.bbc.co.uk/programmes/b03qfzgx ]. I am not convinced I helped communicate this complex risk issue particularly well. This reflection is not helped by my students who let me know I sounded drunk on the interview!
This is a key challenge for clinicians and public health practitioners. Effective risk communication can stimulate health behaviour/belief change and reduce risk levels 16. One of the major barriers to communicating risk effectively is the difficulty both patients and clinicians have in understanding statistics and numbers, e.g. even among highly educated adults in a US survey only 21 per cent correctly identified that one in 1,000 was the same as 0.1 per cent 17.
Communicating risk more generally, in relation to treatment options and associated risks, has similar and perhaps even greater complexity for clinicians and patients, and is beyond the scope of this article.
INHANCE researchers are currently developing and validating a risk-prediction model that could be used to identify those at highest risk of oral cavity and oropharyngeal cancer which could potentially guide opportunistic screening, and risk factor counselling interventions. Such personalised risk information can be presented to individuals based on their characteristics and behaviours and can improve decision making in relation to screening 18. Such risk tools already exist for presenting breast cancer risk [ http://www.cancer.gov/bcrisktool/ ] and are widely available for cardiovascular risk [ http://www.qrisk.org/ ]. The major risk factors identified (above) and going forward in this model for oral cancer risk profiling are smoking, alcohol, and socioeconomic status, alongside age and gender determinants.
To paraphrase Johannes Clemmesen (the founder of the Danish Cancer Registry), the purpose of all cancer epidemiology studies is to prevent it.
Understanding risk is the key first step in the pathway to prevention. Prevention approaches therefore depend on whether the cancer is HPV-driven or non-HPV driven.
The primary prevention for HPV-driven oral cancer is likely to be via the HPV vaccination. There is one proof of principle study which demonstrates that the HPV vaccine (designed for cervical cancer prevention) prevents oral HPV infection. However, more evidence is needed to fully inform policy in relation to extending the vaccination to males. Although the case could be (and has been in other countries) argued on equity grounds, where men who have sex with men or men who have sex with women outside of the vaccinated population will not benefit from the hypothetical and assumed “herd immunity” to the population now that girls have been widely vaccinated (~90 per cent of 12 year olds from 2008 in Scotland). In addition to HPV vaccination, in theory HPV-driven oral cancer could be prevented via behavioural modification/safer sexual practices (i.e. condom use or dental (rubber) dam use for oral sex – for more information see http://www.nhs.uk/chq/Pages/970.aspx ).
Secondary prevention – whereby we interrupt disease progression via early detection (i.e. “screening”) and early treatment – does have potential to prevent oral cancer. It will remain necessary for oropharyngeal cancer for decades to come (even with the prevention prospects of the HPV vaccine) as there will be a substantial unvaccinated cohort who will suffer the future oropharyngeal cancers. Unfortunately, there is no validated “screening” method for oropharyngeal cancer (although HPV16 E6 antibody serology blood test could be promising). Problems and research evidence gaps remain in relation to oropharyngeal cancer screening include: there is no identifiable precancerous lesion (like the Pap smear detected cervical intraepithelial neoplasm (CIN)), uncertainty about effectiveness of early intervention and treatment, and demonstrable reductions in cancer mortality.
The same difficulties for direct visual inspection of the oropharynx do not exist for comprehensive visual inspection of the oral cavity, and there is some limited evidence of effectiveness and cost-effectiveness of opportunistic screening 19 20. Research questions remain (and we have three PhD students working on some of them) including: what constitutes best practice for oral examination/screening? Can risk assessment and profiling assist in focusing on groups or on recall interval? Given low volume of disease is early detection a realistic proposition? Are there inequalities in access and uptake of the opportunity to screening? And what are the barriers and facilitators to delivering screening?
Non-HPV driven (oral cavity cancer) prevention in or via dental practice is important. Again we have research in our group ongoing in this area. The principles of prevention for oral cancer should reflect the evidence from our understanding of risk. These principles include: i) age is not an issue in terms of risk factors – work led by Tatiana Macfarlane, Aberdeen Dental School has shown that even among young adults with oral cancer the same risk factors smoking and alcohol dominate 21; ii) risk can reduce when behaviours stop; iii) oral health assessment is an important first step in any prevention intervention (we must ask the questions); iv) signposting and/or referring for more intensive preventive intervention services; and v) the role of tailoring advice and support to individual patients needs – recognising the dominant role of socioeconomic circumstances.
Public health and policy response needs to focus on the upstream structural causes of the causes; on what has been defined as the “common risk factor” approach 22 (Sheiham and Watt 2000) – risk factors for oral cancer overlap with periodontal disease, with other cancers, with cardiovascular disease and so on…; and on multiple risk factors – our research has shown that risk factors do not exist in isolation – they cluster: people who smoke also drink heavily and have poor diets, this clustering is even more socioeconomically determined 23. Policy developments also need to extend to the increasing preventive focus and wider healthcare role of dental practitioners and teams.
And as the times they are a changin’ – we would do well as a profession to more proactively contribute to the political and policy discourse; to advocate for societal change for tackling health inequalities; to prioritise research and development to tackle the burden of oral cancer – a burden on health services and society, but an even greater burden of suffering on communities, families, and patients.
About the Author
David Conway, Clinical Senior Lecturer /Honorary Consultant in Dental Public Health, Community Oral Health Group. University of Glasgow Dental School / NHS National Services Scotland. Tel. 0141 211 9750
- Junor EJ et al (2010) Oropharyngeal cancer. Fastest increasing cancer in Scotland, especially in men. BMJ 340:c2512.
- Scottish Cancer Registry (2015) Cancer Statistics. NHS National Services Scotland. Information Services Division. http://www.isdscotland.org/Health-Topics/Cancer/Cancer-Statistics/
- Hashibe M, Sturgis EM (2013) Epidemiology of oral-cavity and oropharyngeal carcinomas: controlling a tobacco epidemic while a human papilloma virus epidemic emerges. Otolaryngol Clin North Am 46(4):507-20.
- Lee YC, et al. (2009) Active and involuntary tobacco smoking and upper aerodigestive tract cancer risks in a multicenter case-control study. Cancer Epidemiol Biomarkers Prev 12: 3353-61.
- Conway DI, et al. (2009) Enhancing epidemiologic research on head and neck cancer: INHANCE – The INternational Head And Neck Cancer Epidemiology consortium. Oral Oncol 45(9):743-6.
- Winn DM, et al. (2015) The INHANCE consortium: Towards a better understanding of the causes and mechanisms of head and neck cancer. Oral Dis doi: 10.1111/odi.12342
- Hashibe M, et al. (2007) Alcohol drinking in never users of tobacco, cigarette smoking in never drinkers, and the risk of head and neck cancer: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. J Natl Cancer Inst 99(10): 777-89.
- Hashibe M, et al. (2009) Interaction between tobacco and alcohol use and the risk of head and neck cancer: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. Cancer Epidemiol Biomarkers Prev 18(2):541-50.
- Lubin JH, et al. (2009) Total exposure and exposure rate effects for alcohol and smoking and risk of head and neck cancer: a pooled analysis of case-control studies. Am J Epidemiol 170(8):937-47.
- Marron M, et al. (2010) Cessation of alcohol drinking, tobacco smoking and the reversal of head and neck cancer risk. Int J Epidemiol 39(1):182-96.
- Chuang SC, et al. (2012) Diet and the risk of head and neck cancer: a pooled analysis in the INHANCE consortium. Cancer Causes Control 23(1):69-88
- Conway DI, Brenner DR, McMahon AD, Macpherson LM, Agudo A, Ahrens W, Bosetti C, Brenner H, Castellsague X, Chen C, Curado MP, Curioni OA, Dal Maso L, Daudt AW, de Gois Filho JF, D’Souza G, Edefonti V, Fabianova E, Fernandez L, Franceschi S, Gillison M, Hayes RB, Healy CM, Herrero R, Holcatova I, Jayaprakash V, Kelsey K, Kjaerheim K, Koifman S, La Vecchia C, Lagiou P, Lazarus P, Levi F, Lissowska J, Luce D, Macfarlane TV, Mates D, Matos E, McClean M, Menezes AM, Menvielle G, Merletti F, Morgenstern H, Moysich K, Müller H, Muscat J, Olshan AF, Purdue MP, Ramroth H, Richiardi L, Rudnai P, Schantz S, Schwartz SM, Shangina O, Simonato L, Smith E, Stucker I, Sturgis EM, Szeszenia-Dabrowska N, Talamini R, Thomson P, Vaughan TL, Wei Q, Winn DM, Wunsch-Filho V, Yu GP, Zhang ZF, Zheng T, Znaor A, Boffetta P, Chuang SC, Ghodrat M, Amy Lee YC, Hashibe M, Brennan P (2015). Estimating and explaining the effect of education and income on head and neck cancer risk: INHANCE consortium pooled analysis of 31 case-control studies from 27 countries. Int J Cancer 136(5):1125-39.
- D’Souza G, et al. (2007) Case-control study of human papillomavirus and oropharyngeal cancer. N Engl J Med 356(19):1944-56
- Gillison ML, et al. (2012) Prevalence of oral HPV infection in the United States, 2009-2010. JAMA 307(7):693-703.
- Heck JE, et al. (2010) Sexual behaviours and the risk of head and neck cancers: a pooled analysis in the International Head and Neck Cancer Epidemiology (INHANCE) consortium. Int J Epidemiol 39(1):166-81.
- Ahmed H, et al. (2012) Communicating risk. BMJ 344.
- Gigerenzer G (2002) Reckoning with risk – learning to live with uncertainty. 1st ed. Penguin Press. London.
- Edwards AG, et al. (2006) Personalised risk communication for informed decision making about taking screening tests. Cochrane Database Syst Rev 4:CD001865.
- SIGN (2006) Diagnosis and management of head and neck cancer. NHS Quality Improvement Scotland.
20 Speight PM et al. (2006) The cost-effectiveness of screening for oral cancer in primary care. Health Technol Assess 10(14):1-144, iii-iv. Review.
- Macfarlane TV, et al (2010) The aetiology of upper aerodigestive tract cancers among young adults in Europe: the ARCAGE study. Cancer Causes Control;21(12):2213-21.
- Sheiham A, Watt RG(2000) The common risk factor approach: a rational basis for promoting oral health. Community Dent Oral Epidemiol 28: 399-406
- Lawder R, Harding O, Stockton D, Fischbacher C, Brewster DH, Chalmers J, Finlayson A, Conway DI. BMC Public Health. 2010 Jun 11;10:330.
CPD responses closedThe CPD quiz for this article is now closed. Please check the listings for the current quizzes