There has been wide variety in the definition of HNC in the literature, with some papers including cervical oesophagus or thyroid gland (6,7). However, our classification was based on the United Kingdom association of cancer registries (UKACR) (8) using International Classification of Diseases, 10th edition (ICD-10). HNC cases (C00-C14 and C30-C32); oral cavity (C00-06, C12-14); salivary gland; (C07-08); tonsil and oropharynx (C09-10); nasopharynx (C11); nasal cavity and middle ear (C30); accessory sinuses (C31); and larynx (C32).

Incidence was calculated using cancer registration data and gender- and age-specific population data for each region of England. A ten-year period was chosen in order to have sufficient data to allow examination in recent trends in the incidence of HNC. There has been a reported increase in the incidence of human papilloma virus-related HNC in recent years (9), therefore oropharyngeal (C10), base of tongue (C01) and tonsil cancers (C09) were analysed as a subgroup, as these sites have most frequently been associated with HPV-infection. This group will be referred to as ‘Potentially HPV-associated cancers’ (HPV-A); The HPV status of these cancers was not known but it has been found that 36–80% of cancers at these sites are HPV-associated (9).

Oral cavity cancer included ICD 10 codes C00, C02-06 and C12-14, according to the UKACR definition. Laryngeal cancers were also analysed separately (C32). Salivary gland (C07-08), nasopharynx (C11), nasal cavity and middle ear (C30) and accessory sinuses (C31) were not analysed as subgroups but are included in the overall HNC figures. Incident rates were reported as number of new cases per 100,000 person-years, and they were age-standardised according to the 2013 European Standard Population. T he study population was categorised according to age, region of residence, gender and cancer sub-types. For this analysis, all HNC patients were divided into five age groups, 0–29, 30–39, 40–49, 50–59, and 60+ years. Poisson regression models were used to examine time trends in the overall incidence of HNCs and time trends in five age categories, region of residence, gender and HNC specific subtypes in England between 2002 and 2011. All analyses were performed using Stata 13.1 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP, USA) and P-values <0.05 were considered statistically significant.

The incidence of HNC increased significantly in all regions. The North of England has the highest incidence per 100,000. Fig. 2a depicts the trend in incidence of HNC for all regions of England. The North East of England has the highest incidence of HNC in England; in 2011 the incidence was 27.6 and 10.2 per 100,000 for males and females, respectively, compared to the lowest incidence in London of 17.3 per 100,000 males and 6.8 per 100,000 females. The North East also recorded the highest incidence for the subtypes analysed, potentially HPV-A (Fig. 2b), oral cavity cancer (Fig. 2c) and laryngeal cancer (Fig. 2d).

The North West consistently recorded the second highest incidence of HNC from 2002–2011 (average 24.1 per 100,000 males and 9.9 per 100,000 females). The highest number of cases of HNC in England are recorded in the North West; an average of 813 male and 348 female cases of HNC were recorded each year from 2002–2011, compared to 325 and 132, respectively in the North East.

From 2002–2011, there was a significant increase in the incidence of HNC in all regions of England. The most consistent increases were found in the South West and Yorkshire and the Humber for males, with an average annual percentage increase of 3.7 and 3.1%, respectively. The East Midlands and North West reported the most consistent increases for females with average Annual Percentage Change (APC) 8.1% and 4.3%, respectively).

Potentially HPV-A cancer incidence increased significantly in males and females in all regions except for females in the East and London. The East Midlands and Yorkshire and the Humber display the most consistent increases in incidence, in males, between 2002 and 2011 with average APC of 14.3% (incidence increased from 1.9 to 5.9 per 100,000) and 8% (incidence increased from 2.6 to 5.1 per 100,000). In females, the most consistent increase was found in the South West, with average APC of 14.1%; incidence increased from 0.9 to 2.1 per 100,000 between 2002 and 2011.

The incidence of oral cavity cancer significantly increased in all regions except Yorkshire and the Humber and the North East, for males and females. There was no significant increase in the East of England for female oral cancers. The East Midlands and South West displayed the most consistent increases each year for males and females, respectively, with average APC of 5.1% (incidence increased from 6.2 to 9.2 per 100,000 from 2002–2011) and 5.7% (incidence increased from 3.5 to 5.6 per 100,000).

The North East showed a statistically significant decrease for Laryngeal cancer in males from 2002–2011 (incidence 9.2 per 100,000 to 7.6 per 100,000). However, the East Midlands report a statistically significant increase in incidence for females, from 0.8 to 1.5 per 100,000. All other regions display non-significant trends, indicating the incidence of laryngeal cancer is relatively stable.

More than two-thirds of all HNC patients were in the 60+ year age group from 2002–2011. Dobaree et al reported that 60% of HNC patients were aged 40–69 years in South East England between 2000 and 2004. A study conducted in Germany reported a similar age at diagnosis, with 50% of HNC patients being aged 60–79 years (10,11) and a study in the United States reported mean age at diagnosis of 62 years (19) (www.seer.cancer.gov/faststats/selections.php). Over the ten year period we studied, the most marked increases in HNC incidence were found in the 55–59 year age category for males and 65–69 year age category for females, which is mostly due to the increase in oropharyngeal, base of tongue and tonsil cancer as discussed below.

The highest incidence of oral cavity cancer was found in the over 80′s, for both males and females. The most significant increases in incidence of oral cavity cancer were found in males aged 60–64 years and females aged 65–69 years. Results from Portugal show that the highest incidence of oral cancer in females is in the over 75 year age group, whereas for males, the 60–64 year age group has highest incidence, further supporting the concept that females with oral cancer tend to be older than males (13). Oral cancer incidence is increasing in the Netherlands in females; however, rates are stable in males. This observation is similar to those reported from France (18,20). Other studies have reported reduced incidence of oral cancer; a reduction of 1.5% per year was found between 1995 and 2004 in the United States (20). However, in the time period 2003–2010, there was an annual increase of 0.2% in oral cancers in males (19). In fact, cancers of the tongue increased in males by 2.4% annually between 1999 and 2010 and in females by 0.6% annually between 1992 and 2010.

Information on HPV-status was not available from the ONS, therefore our classification of potentially HPV-A as cancers of the tonsil, oropharynx and base of tongue is based on reports in the literature that have found between 36% and >80% of cancers at these sites are infected by HPV, implying, but not proving, causation (9). A recent study confirmed 55% of oropharyngeal cancers were HPV-infected and HPV infection poses an increased risk of developing oropharyngeal cancer: OR 3.5 (95% CI 2.1–5.9) (21,22).

In our study, potentially HPV-A cancer seems to affect younger individuals, with the highest incidence found in those aged 60 years and above; it also affects proportionally more people aged 40–49 years than other HNC. Other HNC have higher incidence in older age categories, however, incidence of potentially HPV-A cancer decreases in persons aged over 65 years. The incidence of potentially HPV-A cancer doubled in males aged 55–59 and 65–69 years and almost doubled in females aged 65–69 years between 2002 and 2011. This finding is in agreement with National Cancer Information Service data, which show that the incidence of oropharyngeal cancer doubled in England between 1990 and 2006 (3); this rise is thought to be due to HPV-infection. However, the incidence increased sharply between 2006 and 2011 perhaps also because of increased HPV-infection (Fig. 1b).

Researchers from Canada have also reported a marked increase in tonsil and base of tongue cancer, again, believed to be due to HPV infection (23). They reported the largest increase in persons aged 50–59 years (incidence 4.4 to 8.9 per 100,000 between 1997 and 2009), which is similar to our findings in England. Researchers from Portugal reported that oropharyngeal cancer in men increased by 3.5% per year and in females by 2% per year between 1998 and 2007 (13) and researchers from the United States reported 2.3% annual increase in the incidence of oropharyngeal and tonsillar cancer for males between 1992 and 2010 (19,24). Of note, a 0.4% annual reduction in female oropharyngeal and tonsil cancers is reported in the United States in the period 1992–2010 (19) but the reason for this increase is unclear.

The ratio of males: females HNC remained relatively stable over the study period (approximately 2:1). The incidence of HNC in males is higher than females throughout all age groups; however, the relative proportion of females affected increases in the older age categories, likely due to a larger female population. The male: female ratio is most even for cancer of the oral cavity; subgroup analysis showed that there are more female oral cavity cancer patients over 80 years of age than male patients (58 and 42%, respectively). Females have longer life-expectancy than males and the population of females is much larger in this age group (1,564,400 and 920,700, respectively in 2011).

Laryngeal cancer had the highest ratio of male:female cancers (4.6:1) and this ratio is very similar throughout all age groups, even the most elderly, which is different to the trend observed with oral cavity cancer: Coupland et al reports a male:female ratio of 4.8:1 for laryngeal cancer in the South East of England between 1985 and 2004 (17). In other countries, the ratio is much higher: Lithuania and Portugal have a male:female ratio of 25:1 and 36:1, respectively (2). The incidence in males varies widely from country to country from 11.9 per 100,000 in Hungary to 1.8 per 100,000 in Sweden (2). Smoking and alcohol consumption are the major risk factors for all HNC, however, for laryngeal cancer the Population Attributable Risk (PAR) has been found to be 89% compared to 64% for oral cavity cancer (25). Males also have a higher PAR for smoking and alcohol compared to women (74% compared to 57%, respectively) (25); these findings may help to explain the much higher incidence in males than females.

The North East and North West have the highest incidence of HNC and all subtypes analysed, with incidence rates that are consistently above the national average. For HNC overall (C00-14, C30-32), all Regions of England showed a significant increase in incidence in the period 2002–2011. Not all regions had a statistically significant increase in oral cavity cancer but none displayed a decrease in incidence. Incidence of potentially HPV-A cancers increased in all regions for males, particularly the East Midlands and Yorkshire and the Humber; interestingly Yorkshire and the Humber was one of the regions with no statistically significant increase in oral cancer incidence. For females, it was the South West with the most marked increase in potentially HPV-A cancers. Laryngeal cancer incidence significantly decreased in the North East, although even with the decrease, the incidence is still the highest in the country. A previous study has found a decreasing incidence in laryngeal cancer from the North to the South of England, thought to be due to the changes in the industrial landscape and supporting the concept of the ‘North-South divide’, which is a term used to describe gross differences in socio-economic status for individuals living in the North and South of England (3).

It has previously been reported that there is increased incidence of HNC in lower socio-economic groups and a link with deprivation has been established (26). Since smoking and alcohol consumption are the two most significant risk factors for HNC, differences in the rates of smoking and alcohol consumption could help to explain the regional variations in incidence of the disease. Smoking rates in the UK reduced dramatically between 1974 and 2011 (51-20% of men and 41–18% of women) but regional variations still exist: in 2011 smoking rates were higher in the North West (22%) than the rest of the England (20%), although this is not statistically significant (27) Residential area deprivation is a strong independent predictor for smoking (OR 1.85 CI 1.57–2.13) and the North West contains over half of the 1% most deprived areas of England (28,29). Smoking is more prevalent in the routine and manual occupations than managerial and professional (29 vs. 14% for males; 26 vs. 12% for females) and the percentage of never smokers is lower. It is known that people in lower socio-economic groups are not only more likely to smoke but generally start smoking at younger age, smoke more heavily and are less likely to quit smoking, each of which increases their risk of HNC (27).

Heavy drinking on at least one day in the week (>8 units for men, >6 units for women) was most common in the North West and Yorkshire and the Humber in 2011 (23 vs. 18%). In contrast to smoking, regular alcohol consumption (at least 5 days per week) is more prevalent amongst the managerial/professional group than the routine/manual group (19 vs. 13% for males) and in those who are in employment compared to the unemployed (15 vs. 6% for males) (27). Smoking and drinking are also closely linked: according to ONS data (2011), 14% of male smokers consumed greater than 12 units of alcohol on one day in the preceding week compared to 8% of non-smokers. It has also been found that, amongst young people aged 11–15 years, occasional and regular smokers are much more likely to drink alcohol (OR 2.85 and 3.65, respectively) than non-smokers (30).

Ethnicity has also been considered as a risk factor for HNC, due to the marked differences in incidence of HNC around the world. However, some of the differences are cultural rather than ethnic e.g. chewing betel or tobacco consumption. These may not be transferred along ethnic lines due to availability in England. Incidence of HNC is highest in South and South East Asia and the incidence of HNC amongst South East Asians living in the UK is higher than in other ethnic groups (6,31): Csikar et al reported an incidence of 7.2 per 100,000 and 6.0 per 100,000, respectively. They concluded that areas in which many South Asian women live, there may be a higher incidence of head and neck (particularly oral) cancer. However, London has the highest percentage of Indian, Bangladeshi and Pakistani persons in England (1.8% of the population) and our study shows the incidence of HNC was the lowest of all the Regions (16.4 per 100,000 males and 6.8 per 100,000 females). Smokeless tobacco use is highest in Bangladeshi women (16% of the Bangladeshi population) and this is thought to account for the higher incidence of oral cancer in this group (6). Noteworthy, there appears to be no link between oral and pharyngeal cancer with deprivation, in South Asian males and females living in the UK. Presumably, we cannot compare deprivation in South Asia (poor sanitation and poor diet) with deprivation in the UK - smoking, drinking and poor diet. This supports the concept that ethnic-specific risk factors account for the higher incidence of oral/HNC in this group of the population (31).

In conclusion, our study has confirmed that the incidence of head and neck cancer continues to rise in England. Between 2002 and 2011, incidence increased from 12.2 to 15.9 per 100,000. Oral cavity cancer incidence is also increasing, in males and females, whereas incidence of laryngeal cancer is stable. The incidence of cancers at sites strongly associated with HPV infection doubled, in males and females in high-risk age-groups in this ten-year period. Regional variation exists and further work is needed to establish the role of deprivation and socioeconomic status on HNC incidence. Cancer statistics form an important part of health care planning and this information may be used to inform researchers when planning studies, screening and health awareness programmes in different Regions of England.