Despite decades of research and robust cancer preventive strategies, lung cancer remained the most common cancer since 1985 with a recorded 1.8 million new cases in 2012 (1). It has a poor 5-year survival rate and is the commonest cause of cancer related mortality worldwide accounting for 1.6 million deaths annually, which is an estimate of one-in-five deaths from all cancers (2). In the United Kingdom (UK), 43,463 individuals (13% of all cancer cases) were diagnosed with lung cancer in 2011 and 35,371 lung cancer deaths (22% of all cancer deaths) were reported in 2012 (3). Five percent of adult lung cancer patients (4% of men and 7% of women) diagnosed in 2010–2011 in England and Wales are predicted to survive ten or more years (3).

Although the risk factors for lung cancer are multifactorial, tobacco smoking is still primarily responsible for the development of the disease (4,5). Throughout history, socioeconomic status has been identified to contribute to health disparities. Health behaviors such as smoking, lack of exercise and poor diet partly explain socioeconomic disparity in health (6). Socioeconomic gradient in lung cancer reflects differences in health behavior and exposure to occupational and environmental carcinogens, and air pollution among the various socioeconomic groups (7). The patterns in lung cancer incidence are influenced by exposure many years before diagnosis, which is unfortunately higher in socioeconomic deprived geographical areas (8).

Rising trends in lung cancer incidence have been observed internationally and also within the UK (9). Conducting periodic trend analysis has been found to be invaluable because studying historical data provides insights into the disease pattern, which is expedient for projection and future resource planning. The aim of this study is to present an analysis of incidence of lung cancer trends in England over the period 2002–2011, as well as project rates up to year 2020. In addition, we compared trends in comparable countries in the world.

Lung cancer cases were defined, according to the tenth revision of the International Classification of Diseases (ICD-10), by code C33-34. In the UK, it is obligatory for the National Health Service (NHS) to provide agreed core cancer registration dataset to regional cancer registries. New cancer registrations from the National Cancer Registration Service are submitted to the Office of National Statistics (ONS) for validation and data processing (for England and Wales), and are later compiled into the National Cancer Data Repository (10).

Information on new lung cancer registration and age-specific incidence data for this study were obtained from the cancer registration series published yearly by ONS. The ONS database records incidence by age band, region, gender, and ICD codes. Lung cancer incidences are reported as age-standardised per 100,000 population, allowing for better comparisons between groups in the study. Mid-year population estimates were used to calculate annual incidence rates for each calendar year of study. All cases with ICD C33-34 in England between 2002 and 2011 were included.

We also obtained age-standardised incidence data for 10 other countries from Europe, North America, Australia and Asia (Japan) from their respective cancer repositories. The selection of European countries to this study was purely based on population size and accessibility to validated data. Lung cancer incidence data for Russia, Italy, Spain, Poland and France were obtained from EUREG registry data from European Cancer Observatory (ECO) and for Germany, from the German Centre for Cancer Registry Data (ZfKD). Incidence data outside Europe were obtained for Australia, from the Australian Institute of Health and Welfare; for Canada, from Public Health Agency of Canada (Canada); for Japan, from National Cancer Centre (Japan); and for USA, from National Cancer Institute (USA).

A total of 318,417 lung cancer cases were extracted from ONS during the study period and more males were diagnosed with lung cancer (57.4%) than females (42.6%). Lung cancer age-standardised rates were also higher in men than in women in all age groups and the overall rate ratio between male and female was 3:2, respectively (Table I).

As expected, rates increased directly with increasing age and the greatest percentage change was seen in the 80+ group (+2.6%). Although the bulk of the cases seen in the 10-year study were observed in the 70–79 age-group, a total of 35.3% of all cases, this age group showed a remarkable decrease in AAPC by -0.3%. Regionally, London was the only region that showed a decrease in overall trend in the last ten years (Table II), which mirrors England’s smoking prevalence data showing the region to have the lowest prevalence of smoking (15), as well as the largest percentage decrease in smoking prevalence between 2002 and 2011 (Table III).

Analysis showed significant decline in age-standardised rates of lung cancer in males in all age groups, from 2002 to 2011 with an AAPC of −1.0%. In contrast, the age-standardised rates of lung cancer in females increased steadily in all age groups over the same period with an overall AAPC of +1.9% (Table I and Fig. 1). The greatest change seen in women was an increase of +4.8% in the 80+ age group followed by those in 60–69 age group with an AAPC of +3.3% (Table I). Overall, lung cancer trend increased by +0.5% in England and our projections to 2020 revealed that incidence rates in men would continue to fall from the current 57.2 per 100,000 to 39.8 per 100,000 in 2020 whilst rates in women will increase from the current value of 38.7 to 40.3 per 100,000 using a very modest model (R²=0.69) (Fig. 2). Actually, the best fitting model (R²=0.91) predicted age-standardised rate of 50.5 per 100,000 in women by 2020.

While the decrease in lung cancer rates in men is consistent throughout this study, the lung cancer burden, defined as the total number of new registration per year, is nevertheless on the increase by an average of +0.3% per year in males; and by 3.0% per year in females. This higher rate of increase in females is responsible for narrowing the existing incidence gap between males and females, from a difference of 6,000 new registrations between men and women in 2002 to about 3,500 in 2011 (Fig. 1). Historically this gap was wider by about 18,000 back in 1972 (Fig. 2). Our projections, which incorporated the impact of expected population structure in England (model 2), showed that this existing gap between men and women would cease to exist by 2020 and women will then begin to have higher number of new lung cancer registrations per year subsequently (Fig. 3).

We also observed that although the North West recorded the largest number of new cases over the 10 years in both males and females (30,089 and 24,572 new cases, respectively), the North East region had the highest incidence rates in both males and females (87.1 and 64.2, respectively) (Table II). Rates varied differently in all the regions over the last 10 years, with a decrease in AAPC seen in men in all the regions except in East Midlands (+0.6%), South West (+0.5%) and North West (+0.4%). North East and London regions both have AAPC that is less than the England average (AAPC) for men. Regional trends in females varied between +1.1 and +4.1%, indicating an increase in female lung cancer incidence trend in all regions in England from 2002 to 2011. South West region had the highest rate of growth in female lung cancer incidence with an AAPC of +4.1 and London had the least (+1.1). The overall DASR (Directly Age-Standarised Rate) ratio between England worst (North East) and England best (South East) region is 1.7.

Further regional analyses showed London to have significantly lower than England average trend in both men and women (Fig. 4), which could be attributed to the low smoking prevalence in that region. London currently has the lowest prevalence of smoking (15) and their smoking data also showed the largest regional percentage decrease (−4.9%) in smoking prevalence between 2002 and 2011. London was the only region that demonstrated an overall trend (AAPC) decrease in lung cancer in the last ten years (Table II). Finally, analysis of lung cancer data from other countries revealed that Japan had the largest increase over the 10-year study in males (+1.6%) and France (+8.0%) in females. The largest decrease in men was seen in Italy (−4.6%) and Russian Federation in women (−1.3%) (Fig. 5).

This study showed that overall lung cancer incidence in England increased per year by an average of +0.4% between 2002 and 2011. Historically, men have always had higher lung cancer incidence rates than women across the country, with an incidence ratio as high as 6:1 documented between males and female, respectively in 1950s (16). However, the gender gap is closing due to decreasing rates in men and increasing rates in women, in spite of the decreasing prevalence of smoking. The current overall male to female incidence ratio demonstrated in this study was 3:2, but our projections showed the rate ratio will be 1:1 by 2020.

The overall trend in women increased by an average of +1.9, which seemed rather modest, but subgroup analysis revealed that some age-groups had AAPC values as high as +4.7 (80+) and +3.3 (60–69) during the period of study (Table I). If 9 out of 10 lung cancer cases occur at age of 60+ (13), understanding the increasing rate in women over 60 becomes very vital because it would lead to a significant lung cancer burden with an increasing ageing population in England. The projected models in Fig. 3 showed an impending lung cancer burden in England especially in women. For instance, if the current incidence rate is not halted, as describes by model 2, there would be 5,848 excess lung cancer cases in 2020 (when compared with the 2011 figures), with the female population accounting for 85.4% (4,996) of the excess cases. We also demonstrated that in spite of the projected decrease in incidence rates in men, there would still be an excess of 852 lung cancer cases by 2020, mainly caused by the projected increase in male population over 60+.

Implementing a comprehensive tobacco control that is based on the six policies identified by the World Bank has the potential of reducing incidence. An example is the California Tobacco Control programme which utilised these policies and saw a 61% decrease per capita cigarette consumption compared to just 41% across USA, during the same time period (1989–90 and 2006–07). Within 10 years (1988–1997), California recorded a decrease in lung cancer incidence by 1.9% per year (p<0.01). The overall incidence rate decreased by 14.0% in California and by 2.7% in non-California regions during the 10-year period (14). Our model 3 provides picture of lung cancer if a similar tobacco control programme was implemented across England, with the potential of reducing lung cancer incidence (new registrations) in males by 371 cases per year and in females by 340 cases per year. Effective tobacco control will reduce incidence by an overall 6,400 cases (3,340 in men and 3,060 in women) over the projected 9-year period.

Throughout this study, we saw a statistically significant trend increase in women partly explained by the steady rise in female smoking prevalence after the Second World War through to the early 1980s when smoking prevalence realistically started to decline (17,18). Smoking is the most significant risk factor in the development of lung cancer (19,20) and the presence of lag time between exposure to smoking and onset of lung cancer of ~20–35 years is why we are still experience a steady rate increase in women (21).

However, recent studies showed that smoking alone cannot account for the relatively high AAPC currently seen in females (22). These studies identified genetic factors that make women more susceptible to developing lung cancer, smoking patterns in women involving deeper inhalation and effects of oestrogen on tumor metabolism (23,24), all contribute to the increasing trend in women. Zang and Wynder described the relative risk of developing lung cancer in women to be ~1.5-fold higher than in men, despite men having greater number of cigarette pack-years (25).

We also established that regional inequalities still exists across England with higher incidence rates seen in the northern parts of the country and lower rates in the south creating a North-South divide (Table II and Fig. 6). For more than half a century, the northern part of England is known for its high deprivation, unemployment and higher smoking prevalence (26,27) and Quinn et al showed that in the most deprived region in England, lung cancer incidence was 2.5 and 3.0 times those in affluent areas in both males and females, respectively (20). Trends in males decreased across all the regions except in East-Midlands (+0.6), South-West (+0.5) and North-West (+0.4) regions. While trend increased in females in all the regions, South-West (+4.1), East of England (+3.3) and East-Midlands (+2.8) were the top 3 hotspots for female trend in England. The ratio between England worst region (North East) and England best region (South East) is 1.6 and 1.9 in both males and females, respectively (Table II).

Finally, benchmarking England data against 10 different countries was necessary to evaluate the impact of existing lung cancer preventive measures in England. Overall, trend seen in these countries is similar to that observed in England with only a few exceptions. For instance, all countries demonstrated a decreasing trend in men except in Japan where the trend in on the increase (+1.6). Similarly, the trend in women is on the increase in all the countries except in Russia and the USA where incidence decreased by −1.3 and −0.7, respectively. Although lung cancer incidence trend in males is decreasing in England but when compared with other countries, we concluded that there is room for improvement. Seven countries achieved better 10-year AAPC than England, whilst 5 countries had better AAPC in females (Fig. 5). Only 2 countries (USA and Russia) showed decreasing trend in both males and females and while the success seen in USA could be attributable to the wide-ranging smoking cessation programmes and lobbying in that country (28,29), that seen in Russia remains unclear. A possible explanation for the decreasing rates in Russia could be the low average life expectancy of 66 years recorded during our study period, whereas the average age of diagnosis for lung cancer is 70 years (30,31).

The strengths of this study include the large number of lung cancer cases and the use of ONS data, which minimises the chances of missing information on lung cancer incidence. However, the result of this study must be considered in the light of a number of limitations. First, it was practically impossible to stratify and depict trend by race, smoking and other important risk factors (as the data are not available). Secondly, we did not have access to stage or grade specific incident data, which would have allowed us to explore trend according to histological subtypes.

In conclusion, as the UK is already laden with the challenges of a growing and ageing population, it is necessary to re-strategise and develop high-quality preventive interventions that will complement existing national tobacco control measures, especially in women living in regions that have demonstrated very high AAPC values (32). Utilising this targeted approach will achieve a considerable decline in age-adjusted lung cancer incidence rate similar to those seen in comparable countries such as USA (33). Our models showed that, unless a targeted and comprehensive preventive strategy is implemented, lung cancer rates in women will be on par or higher than that expected in their male counterparts in England by the year 2020.