Lung cancer (LC) is the leading cause of death from cancer worldwide,1 accounting for 1527000 deaths every year.2 In Spain alone, every year 20401 people die of LC, which is the third most common cause of death after ischaemic heart disease and cerebrovascular disease.3 In Spain, the incidence and mortality of LC continue to rise in women, while in men it reached its peak in the 1990s, and is currently on the decline.4

It is well known that the main causative factor of LC is tobacco use. Indeed, in countries such as the United States, the fall in smoking rates in the population has been accompanied by a reduction in the incidence and mortality of LC.5 However, around 10%–25% of cases of LC occur in never smokers. Even among never smokers, LC remains one of the principal causes of death from cancer6 and in non-smoking women it is the third most common cause of cancer death.7 In recent years this has led to increased interest in other LC risk factors, such as passive smoking, domestic exposure to radon, previous pulmonary comorbidities and genetic or dietary factors.8,9

Recent studies suggest that LC in never smokers (LCNS) is a distinct entity, to be distinguished both epidemiologically and biologically from LC in smokers.10 Furthermore, the rise in the non-smoking population throughout the world justifies the interest in determining how LC behaves in these patients. Accordingly, the aim of our study was to analyse the incidence, clinical characteristics and survival of patients with LCNS in comparison to that in patients who smoke.

This was an observational, retrospective study of a cohort of LC patients, diagnosed between 1 January 1999 and 31 December 2011 in the University Hospital Complex of Orense. All patients with a primary cytohistological diagnosis of LC were included. Case information was collected from the clinical documentation, bronchoscopy and anatomical pathology databases.

We designed a database for the collection of variables, including identification, age, sex, smoking habit, symptoms, comorbidities, fibreoptic bronchoscopy (FB) findings, diagnostic tests, stage, definitive diagnosis, treatment and date of death. The World Health Organisation (WHO) system was used for histological classification.11 Tumour involvement was classified according to the TNM staging system (7th edition) of the International Association for Study of Lung Cancer (IASLC).12 Date of death was obtained from the clinical records, or, if it was not recorded, a telephone call was made to the patient's home or the Death Registry of Galicia. Data were censored on 1 October 2012 (last study day). Never smokers were defined as subjects who had smoked less than 100 cigarettes during their lifetime, and ex-smokers were those who had stopped smoking at least 6 months earlier.13

Smoking is the main etiological factor for the development of LC. However, LC remains a serious health problem in the never smoker population, in whom the incidence is rising,14 with domestic radon and passive smoking emerging as the principal risk factors.15 In the majority of LC series, 15% of men and about 53% of women have no history of smoking.6 In our study, the percentage of patients with LCNS was 18.3%, with significant differences being observed in terms of sex: 7.7% were men, compared to 76.4% women. This difference coincides with the finding that never smoker females have a higher risk of developing LC than never smoker men,16 which could be explained by other factors, such as genetic predisposition, passive smoking or environmental exposure.

In our series, the most frequent histological strain in smokers was squamous cell carcinoma. In contrast, the never smokers were more often diagnosed with adenocarcinoma. This was found in 55.6% of cases, a rate similar to that of other publications.17 Adenocarcinomas were also more common in women, regardless of their smoking habit.15 Moreover, in the mostly female never smoker population, adenocarcinomas were more common than squamous cell carcinomas.17 Molecular and genetic factors were involved in these histological differences: epidermal growth factor receptor (EGFR) mutations occur almost exclusively in never smokers, and are more common in women and in adenocarcinomas.18,19 A different K-ras oncogene mutation pattern has also been observed in never smoker patients compared to ex-smokers or active smokers, with significantly more frequent guanine to adenosine transitions in codons 12 and 13 occurring in never smokers.20

It is not clear whether there is a difference between LCNS and LC in smokers with respect to TNM stage at the time of diagnosis. In our study, 51.3% of LCNS patients presented in advanced stage (IV), a significantly higher incidence than in smokers. This phenomenon has also been described in the literature, with frequent reports of patients being diagnosed in more advanced stages.21 One possible explanation for this finding is the more intensive medical follow-up of smokers, who frequently present comorbidities, such as COPD, that require closer clinical and radiological monitoring.

The late diagnosis of LC means that the number of patients who can be treated surgically is low: in our series, 14.4% of the never smokers and 15.2% of the smokers received surgical intervention. On the contrary, 48.5% received chemotherapy. Significant differences have been observed in the response rate to EGFR tyrosine kinase inhibitors, such as gefitinib, which is higher in never smokers than in smokers with advanced non-small cell lung cancer (36% vs 8%).22 Similar results were observed with erlotinib (24.7% vs 3.9%).23

LC survival is poor, ranging in most series between 6% and 16%.24–26 In our LCNS patients, 5-year survival was 9.37% and no significant differences were observed compared to that in the smoking population. The results in the literature in this area are varied: Nordquist et al.27 reported improved survival in non-smoking patients, while, like us, Toh et al.28 did not find any difference.

One of the most remarkable results of our study is that in never smoker patients, women had better survival than men; in addition, the multivariate analysis showed that the risk of death was higher in the male population. In this respect, the results of the studies are contradictory: some papers report that in LC, female sex is a factor for good prognosis,29 while others describe worse survival and treatment response in women.30 In a previous study carried out by our group,31 no differences were seen in survival between the sexes, although another recent Spanish publication32 describes improved survival in women with LC, with differences depending on the histological type.

Our study has some limitations. Firstly, its retrospective nature meant that no other possible etiological factors, such as radon exposure or passive smoking, thought to be the first and second leading causes of LCNS, respectively,8 were analysed. Our group is now participating in a multicentre study addressing this causative association, and we have seen that in the homes of patients the concentration of domestic radon was threefold higher than in the control population, according to already published preliminary results.9 Secondly, LC with EGFR mutations occurs almost exclusively in never smokers with adenocarcinoma.20 Since the determination of these mutations in our centre is recent, the impact of EGFR tyrosine kinase inhibitors on patient prognosis could not be examined.

To conclude, our study confirms that LCNS has clinical characteristics distinct from those of LC in smokers; it occurs in 18.3% of LC patients, more frequently in women, diagnosis is made in advanced stage and the histological type is predominantly adenocarcinoma. We did not find any significant differences in survival. In our opinion, therefore, this patient group merits further study, with regard to both causative factors and the response to tailored treatment in patients with certain mutation patterns.

The authors declare that they do not have any conflict of interests.