Lung

can

cer

and

Smok

ing

Rese

arch

Gend

er in

G e n d e rin Lung Cancer

and Smoking Research

Department of Gender, Women and HealthFamily and Community Health

© World Health Organization 2005

All rights reserved. Publications of the World Health Organization can be obtained fromWHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland(tel: +41 22 791 2476; fax: +41 22 791 4857; email: bookorders@who.int). Requestsfor permission to reproduce or translate WHO publications - whether for sale or for non-commercial distribution - should be addressed to WHO Press, at the above address(fax: +41 22 791 4806; email: permissions@who.int).

The designations employed and the presentation of the material in this publication donot imply the expression of any opinion whatsoever on the part of the World HealthOrganization concerning the legal status of any country, territory, city or area or of itsauthorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines onmaps represent approximate border lines for which there may not yet be full agreement.

The mention of specific companies or of certain manufacturers' products does notimply that they are endorsed or recommended by the World Health Organization inpreference to others of a similar nature that are not mentioned. Errors and omissionsexcepted, the names of proprietary products are distinguished by initial capital letters.

All reasonable precautions have been taken by WHO to verify the information con-tained in this publication. However, the published material is being distributed withoutwarranty of any kind, either express or implied. The responsibility for the interpreta-tion and use of the material lies with the reader. In no event shall the World HealthOrganization be liable for damages arising from its use.

The named authors alone are responsible for the views expressed in this publication.

Printed in Italy

Gender in lung cancer and smoking research

WHO Library Cataloguing-in-Publication Data

Payne, Sarah.Gender in lung cancer and smoking research / by Sarah Payne.

(Gender and health research series)

1.Lung neoplasms – epidemiology 2.Smoking 3.Health services accessibility4.Gender identity 5.Sex factors 6. Research I.Title II.Series.

ISBN 92 4 159252 4 (NLM classification: W 84.3)ISSN 1813-2812

ii

iii

Acknowledgements 1Preface 2Abstract 3List of abbreviations 4

1. Introduction 52. Gender issues in lung cancer risk 82.1 Differences between women and men in lung cancer

incidence, prevalence and mortality 8–Differentials by geographical region 11–Differentials in relation to social diversity 13

2.2 Sex, gender and lung cancer: explaining thedifferences between men and women 13–Gender-linked factors 14–Sex-linked factors 17

3. Gender issues in access to care 193.1 Smoking cessation 193.2 Screening and treatment 214. Lung cancer research: knowledge gaps and

recommendations for future research 234.1 Current challenges in lung cancer research 254.2 What is required to address the knowledge

gaps in lung cancer research? 28–Research methods and methodology 28–Smoking and tobacco control 29–Screening and treatment 30–Sex and lung cancer risk: biological factors 33

5. Conclusion 346. References 357. Additional resources 43

Contents

1

Acknowledgments

This document was prepared for theWHO Gender and Health ResearchSeries by Dr Sarah Payne, School forPolicy Studies, University of Bristol,Bristol, England.

The Gender and Health ResearchSeries was developed by theDepartment of Gender, Women andHealth (GWH), under the supervisionof Dr Claudia García-Moreno andwith support from Dr Salma Galal.

GWH gratefully acknowledges thevaluable comments received from:Dr A Sasco and Dr K Straif from theInternational Agency for Research onCancer, Lyon, France, Dr A Ullrichfrom the Programme on CancerControl (PCC) at WHO, and would liketo thank Ann Morgan for copy-editingthis series.

2

The WHO Gender and Health ResearchS e r i e s has been developed by theDepartment of Gender, Women andHealth (GWH), with assistance fromother WHO departments, in order toaddress some of the main issuesinvolved in integrating gender consider-ations into health research. This publi-cation on Gender in Lung Cancer andSmoking Research constitutes one ofthe booklets in this series.

Sex and gender are both importantdeterminants of health. Biological sexand socially-constructed gender inter-act to produce differential risks andvulnerability to ill health, and differ-ences in health-seeking behaviour andhealth outcomes for women and men.Despite widespread recognition ofthese differences, health research hashitherto, more often than not, failedto address both sex and gender ade-quately.

In applied health research, includ-ing the social sciences, the problemhas traditionally been viewed as oneof rendering and interpreting sex dif-ferentials in data analysis and explor-ing the implications for policies andprogrammes. However, examining thegender dimensions of a health issueinvolves much more than this; itrequires unravelling how gender rolesand norms, differences in access toresources and power, and gender-based discrimination influence maleand female health and well-being.

Integrating gender considerationsin health research contributes to bet-ter science and more focused

research, and, consequently, to moreeffective and efficient health policiesand programmes. With these ambi-tions in mind, the objectives of thegender and research series are to:

raise awareness of the need tointegrate gender in health research;

provide practical guidance on howto do this; and

identify policies and mechanismsthat can contribute to ‘engendering’health research.

The series is aimed at researchers,research coordinators, managers ofresearch institutions, and researchfunding agencies. It comprises book-lets covering both a general introduc-tion to engendering the researchprocess as well as topic-specificissues such as lung cancer, tuberculo-sis, and mental health. The researchseries will be extended to other healthtopics in time.

Each booklet will review the partic-ular health issue from a gender per-spective, identify best practices inaddressing gender in research and thegaps in gendered research, and makerecommendations to address thosegaps.

Preface

This paper reviews the need for a gen-der-sensitive approach to lung cancerresearch. Lung cancer is a majorcause of premature and avoidablemortality around the world, andalthough in more developed countriesmortality rates are beginning todecrease, especially in men, the num-ber of deaths from lung cancer in lessdeveloped countries is steadilyincreasing. While historically moremen than women have died from lungcancer as a result of higher levels ofsmoking, the male:female mortalityratio is now showing signs of narrow-ing. Both sex- and gender-linked fac-tors are important in the etiology oflung cancer. However, research intolung cancer needs to address gender

more specifically if we are to makeprogress in reducing both the majorrisk factor – tobacco use – and thenumber of deaths from this disease.This paper reviews what is currentlyknown about sex and gender influ-ences on lung cancer, identifies cur-rent gaps in gender research on lungcancer and smoking and suggestssome directions for the future.

3

Abstract

4

AMA American Medical Association

BMI Body mass index

CDC Centers for Disease Control and Prevention

COPD Chronic obstructive pulmonary disease

CT Computed Tomography

ETS Environmental tobacco smoke

HRT Hormone replacement therapy

IARC International Agency for Research on Cancer

RCT Randomized control trial

USDHHS United States Department of Health and Human Sciences

List of abreviations

5

The "epidemic" of lung cancer mortal-ity has been identified as a majorhealth issue confronting both devel-oped and developing countries. In2000, over one million people diedfrom lung cancer worldwide; 53% ofthese deaths occurred in the moredeveloped countries, the remaining47% in the less developed countries(GLOBOCAN, 2000). Overall, womenaccounted for just over a quarter of alllung cancer deaths. Estimates sug-gest that by 2030, all tobacco-relatedmortality, including lung cancer, willreach around 10 million deaths peryear, with the greatest increase com-ing from the less developed countries(Jha et al., 2002).

Although available data clearlydemonstrates differing trends in lungcancer mortality for men and women,it is only in the last few years thatthere has been an increased aware-ness of the differences in lung cancerrisk associated with both sex (i.e. bio-logical factors) and gender or socially-constructed factors (see Box 1, nextpage). Despite the growing body ofresearch which explores the ways inwhich different patterns of lung can-cer incidence, mortality and survivalmight be associated with sex andgender, uncertainties about a numberof aspects of the disease and how itdiffers for men and women remain. Atpresent, many of the research find-ings in this particular field are sugges-tive rather than established. Thisuncertainty reflects, at least in part,the complexity of the factors involved– we must think not only of the rolesplayed by sex and biology, and by

gender-related issues, such as accessto resources, sexual division of labourand health-seeking behaviour, butalso how these factors interact. Forexample, lung cancer is highly associ-ated with tobacco consumption, butalso occurs in those who have neversmoked. This implies that externalfactors, such as environmental tobac-co smoke (ETS), need consideration;in addition, research has suggestedthat exposure to domestic pollution(e.g. emissions from cooking fuels)and to environmental pollution mayalso have an impact on lung cancerincidence rates.

The objectives of this paper arethreefold: firstly, to review what iscurrently known about sex and genderinfluences on lung cancer risk; sec-ondly, to offer suggestions as to thekinds of research questions that needstill addressing; and thirdly, to identi-fy mechanisms that can contribute tothe engendering of lung cancer smok-ing research.

It is organized as follows: sectiontwo outlines the key differencesbetween women and men in terms oftheir patterns of lung cancer, high-lighting what is known about biologi-cal factors, sociocultural factors andthe interaction between sex and gen-der.

Section three discusses the role ofgender factors in access to healthcare, including screening, health pro-motion and smoking cessation sup-port, and treatment for the diseaseitself. Section four explores the cur-

5

1. Introduction

Sex and Gender

6

Sex is the term used to distinguish men and women on the basis of their bio-logical characteristics. Gender on the other hand refers to those distinguish-ing features that are socially constructed. Gender influences the control menand women have over the determinants of their health, for example, their eco-nomic position and social status, and their access to resources. Gender con-figures both the material and symbolic positions that men and women occu-py in the social hierarchy, and shapes the experiences that condition theirlives. Gender is a powerful social determinant of health that interacts withother variables such as age, family structure, income, education and social sup-port, and with a variety of behavioural factors.

What then do we mean by gender-sensitive research and why is it consid-ered to be so important? Research that fulfils this objective includes consider-ations of gender at all levels of the research process, from commissioning andstudy design through to dissemination of the results. Moreover, sex and gen-der must be identified as key variables, in all measures, reported separatelyand the differences discussed (Doyal, 2002).

Health research that is gender sensitive is necessary because sex and gen-der rank among the key factors, alongside socioeconomic status, ethnicityand age, that determine the health of women and men. Sex and gender affectbiological vulnerability, exposure to health risks, experiences of disease anddisability, and access to medical care and public health services. Researchwhich is gender in-sensitive may result in study design which is unable to dif-ferentiate between women and men in the identification of key findings andtheir policy implications. Gender-sensitive research, on the other hand, is morelikely to lead to improved outcomes in treatment and preventative interven-tions (Doyal, 2002).

The role of gender in public health is now widely acknowledged and is acore component of many health programmes, both international and national.Sex and gender as determinants of health, and as components of a conceptu-al framework for health research, are discussed in more detail in the accom-panying booklet in this WHO Gender and Health Research Series.

rent knowledge gaps in lung cancerresearch and suggests directions forfuture research efforts. This discus-sion is illustrated with examples of"good practice" in terms of genderedhealth research; a number of studies,considered to be valuable contribu-tions to the objective of understand-ing differences between women andmen, have been selected from the lit-erature and summarized in a series ofboxes. In addition, a set of recom-mendations for engendering researchin relation to tobacco control is includ-ed, which can be usefully applied to

the wider field of lung cancerresearch. Section five concludes thepaper with a brief summary of thearguments in favour of gender-sensi-tive research and how this might bedeveloped. There is a degree ofurgency about the gender agenda inlung cancer research, which stemsfrom the growing global epidemic oftobacco use and the likelihood ofincreasing rates of lung cancer amongboth women and men around theworld in the next millennium.

7

Data show that more men thanwomen develop lung cancer, andmore men than women die from thedisease. For both women and men,the single most important risk factoris smoking. Despite the narrowing ofthe gap between men and women intobacco use in recent years, the fig-ures for lung cancer mortality stillshow higher rates for men thanwomen due to the time-lag betweenexposure to smoking and the develop-ment of cancer. However, as the gapin tobacco use continues to narrow,the male:female difference in lungcancer mortality is also expected todecrease further over time. Someresearch has suggested women maysuffer a greater risk of developinglung cancer than men for the samedegree of exposure to the various riskfactors; however, these findings haveyet to be confirmed.

2.1 Differences betweenwomen and men in lung cancerincidence, prevalence andmortality

As a disease, lung cancer is one of themost fatal forms of cancer, with verypoor prognosis once diagnosed. Thisis largely a consequence of the natu-ral history of lung cancer, which has avery rapid rate of growth comparedwith other cancers. Lung cancer isunlikely to be diagnosed opportunisti-cally during the course of other con-sultations. Screening of an "at risk"population is not currently part ofpublic health policy in any countryand despite much debate in the med-

ical literature over the potential ofnewer methods of screening(Henschke et al., 2001; Grannis,2002), evaluation of screening hasnot suggested it would be of signifi-cant value (Reich, 2002; Tyczynski,Bray & Parkin, 2003).

Due to the long time-lag betweenexposure to lung cancer risk factors,such as smoking, and the onset of thedisease itself, lung cancer incidenceand mortality for women and mentends to reflect prior and long-termexposures to risk. Broadly speaking,patterns of lung cancer incidence andmortality show higher rates of the dis-ease among men than women (seeTable 1, next page). In the UnitedStates of America (USA), for exam-ple, in 2000 the age-adjusted lungcancer incidence rate was 79.7 per100 000 population for males, com-pared with a rate of 49.7 per 100 000for females (SEER, 2003). Similarly, inthe United Kingdom, the age-stan-dardized lung cancer incidence rateamong males is approximately twicethat in women (70.4 per 100 000population in men and 34.9 per100 000 population in females in1999) (Cancer Research UK, 2003).

In many of the more developedcountries, the incidence of lung can-cer in men has reached a plateau andis now decreasing, whereas the num-ber of new cases in women continuesto increase (Bray et al., 2002; CDC,2002; Jemal et al., 2003). In con-trast, in less developed countries malelung cancer incidence is continuing toincrease. Although rates of lung can-

8

2. Gender issues in lung cancer risk

9

World 34.92 11.05 31.43 9.53

More developed countries 55.62 15.62 50.15 13.14

Less developed countries 24.79 8.44 22.02 7.40

Eastern Africa 3.08 2.13 2.84 1.95

Middle Africa 5.65 0.76 5.21 0.70

Northern Africa 15.41 2.76 14.22 2.54

Southern Africa 23.81 7.32 21.98 6.75

Western Africa 2.15 0.35 1.98 0.31

Carribbean 28.76 9.70 26.16 8.70

Central America 22.71 8.44 20.55 7.61

South America 25.28 8.34 22.60 7.41

Northern America 58.20 33.59 52.86 26.95

Eastern Asia 39.41 15.01 33.67 12.68

South-eastern Asia 27.83 9.07 25.68 8.36

South-central Asia 11.61 2.33 10.86 2.15

Western Asia 31.21 4.80 28.85 4.43

Eastern Europe 69.70 8.77 63.12 7.79

Northern Europe 44.32 18.85 45.12 18.07

Southern Europe 58.75 7.95 50.42 6.93

Western Europe 53.21 10.68 48.94 9.18

Australia/ New Zealand 42.10 18.18 36.70 14.80

Melanesia 4.73 2.86 4.37 2.64

Micronesia 51.87 18.6 47.93 17.13

Polynesia 38.36 14.24 35.44 2.10

Table 1Global lung cancer incidence and mortality rates, 2000

Cases per100 000 population a

Deaths per100 000 population b

a Age-standardized rate (world standardized rate).b Age-standardized rate.Source: GLOBOCAN 2000. Cancer incidence, mortality and prevalence worldwide. version 1.o I.

Male Female Male Female

cer incidence tend to be relatively lowamong women in most less developedcountries, rates are beginning toincrease in some countries (Ezzati &Lopez, 2003).

Due to the low survival rates forlung cancer for both men and women,mortality closely reflects incidence.Consequently, male lung cancer mor-tality is higher than that of females(Swerdlow et al., 1998; SEER, 2003),again a reflection of the differences inexposure to risk factors, particularlysmoking, over the last 50 years (seeTable 1, next page). In 2000, the age-adjusted male lung cancer mortalityrate in the USA was 76.9 per100 000 population, compared with afemale mortality rate of 41.2 per

100 000 population (SEER, 2003).Mirroring the patterns of incidence, inthe United Kingdom the age-standard-ized lung cancer mortality rate formen at 59.1 per 100 000 populationwas roughly double that for women(29.5 per 100 000 population) in2001 (Cancer Research UK, 2003).As in the case of incidence, whereasmen's lung cancer mortality rates aredecreasing in many developed coun-tries, women's are increasing(Pampel, 2003; SEER, 2003). In sev-eral countries, mortality rates havemore than doubled in women over a30-year period, 1968-1998, and eventripled in some, while rates in menover the same period have barelyincreased overall, and even declined insome countries (See Figure 1 below).

10

As the data in Table 2 (page 12)demonstrate, lung cancer prevalencefollows much the same pattern asincidence and mortality. Both 1-yearand 5-year prevalence is higher inmen than in women. Lung cancer isgenerally more prevalent in developedcountries than it is in the less devel-oped countries.

Further differences between menand women are revealed when lungcancer incidence is broken down byhistologic type. Women are more fre-quently diagnosed with adenocarcino-ma, whereas men are more likely tohave squamous cell carcinomas(Baldini & Strauss, 1997; de Perrot etal., 2000; Siegfried, 2001). These dif-ferences are of significance whenattempting to understand sex andgender factors in the etiology of lungcancer. Adenocarcinoma, for exam-ple, is associated with the major riskfactor for lung cancer, tobaccosmoke. However, this association isnot quite as strong as that for squa-mous cell cancer, and adenocarcino-ma is found more often than othertypes of cancer in non-smokers (Koyi,Hillerdal & Branden, 2002; Sy et al.,2003). In recent years, adenocarcino-ma has increased as a proportion of alllung cancers diagnosed (Blizzard &Dwyer, 2003; Sy et al., 2003). Thisincrease has been associated with theincreasing consumption of cigaretteswith lower nicotine and tar yieldswhich are also more often smoked bywomen (Fry, Menck & Winchester,1996; Levi et al., 1997; Shields,2002). These shifting patterns in theincidence of the various lung cancertypes suggest that different risk andprotective factors are operating formen and women (Axelsson &Rylander, 2002; Tewari & Disaia,2002).

Differentials by geographical region

Within the more developed countriesof the world, there are some markedregional variations in lung cancer pat-terns. The countries in easternEurope, for example, have the highestrates of male lung cancer mortality;women's lung cancer mortality on theother hand is greatest in northernEurope and in the USA (Tyczynski,Bray & Parkin, 2003). Generallyspeaking, lung cancer incidence andmortality are lower in developingcountries. However, while recent evi-dence indicates that rates are increas-ing annually in both men and women,the rate of this increase varies consid-erably between countries (Ezzati &Lopez, 2003).

The data in Table 1 (page 9) can beused to explore regional differences inthe ratio of male to female lung can-cer deaths. The widest gap exists inPolynesia, where male deaths out-number those of women by nearly 17to 1. A high male:female ratio is alsofound in eastern Europe, middleAfrica, southern Europe and westernAsia. The gap between men andwomen in terms of the number of lungcancer deaths is smallest in eastAfrica and Melanesia where themale:female ratios are 1.5 and 1.7,respectively. Both of these regionshave relatively low lung cancer mor-tality rates.

Richmond (2003) has distin-guished four different stages of the"tobacco epidemic" – the term usedto describe the worldwide rapidincrease in the use of tobacco andassociated mortality from lung cancer– and has characterized these in rela-tion to the level of economic develop-ment. Thus countries in the develop-

11

12

World 361 850 129 859 1 013 529 380 888

More developed countries 211 536 78 338 538 791 212 185

Less developed countries 150 314 51 521 474 738 168 703

Eastern Africa 505 352 1 279 873

Middle Africa 398 77 1 020 209

Northern Africa 2 475 505 6 257 1 288

Southern Africa 1 102 418 2 822 1 091

Western Africa 424 61 1 178 185

Carribbean 1275 523 3 123 1296

Central America 2 688 1 224 6 591 3 044

South America 10 003 4 151 24 792 10 492

Northern America 49 073 38 229 132 630 103 996

Eastern Asia 105 664 38 899 370 777 138 853

South-eastern Asia 14 276 5 786 35 563 14 503

South-central Asia 24 764 5 687 58 269 13 679

Western Asia 6 141 1 047 15 511 2 737

Eastern Europe 58 187 10 826 141 789 29 456

Northern Europe 12 531 6 574 29 294 16 030

Southern Europe 30 690 4 941 78 557 13 258

Western Europe 38 625 9 150 96 659 26 186

Australia/ New Zealand 2 988 1 386 7 305 3 658

Melanesia 16 12 50 35

Micronesia 21 11 55 19

Polynesia 4 <0.5 8 <0.5

Table 2Global lung cancer prevalence, all ages, 2000

1-year prevalence a 5-year prevalence b

a The 1-year prevalence at a fixed-point mid-year is calculated from the number of newcases in 2000, times the probability of surviving for 6 months or longer.

b The 5-year prevalence at a fixed-point mid-year is calculated from the number of newcases in 2000, times the probability of surviving for 4.5 years or longer.

Source: GLOBOCAN 2000. Cancer incidence, mortality and prevalence worldwide, version 1.o I.

Male Female Male Female

ing world in the early stages of thetobacco epidemic (stage 1) have lowlevels of lung cancer mortality butgrowing male lung cancer mortality(e.g. Malawi, Nigeria, Swaziland).Countries in stage 2 of the tobaccoepidemic (e.g. China, Mexico, ThePhilippines) have increasing rates ofboth male and female lung cancermortality. The more developed coun-tries, typified by Australia, The UnitedKingdom and the USA, are in stages 3and 4 of the tobacco epidemic: heresmoking is declining in both men andwomen but while lung cancer mortali-ty rates in men may be relatively stat-ic, or even falling, rates in women arecontinuing to rise as a consequence oftheir high level of tobacco use in ear-lier decades.

Although most of countries thatare currently in stages 1 and 2 of theepidemic, have a relatively widemale:female gap in lung cancer inci-dence and mortality, the propensityfor female lung cancer rates to risesubsequent to increased levels oftobacco use is a matter of growingpublic health concern.

Differentials in relation to socialdiversity

Patterns of lung cancer incidence andmortality in women and men vary sig-nificantly in relation to ethnic groupand also in relation to social class orincome. For example, work by Fry,Menck & Winchester (1996) found apoorer prognosis among blackAmericans of both sexes diagnosedwith lung cancer compared with awhite population. Evidence relating tothe cause of the poorer prognosis wasinconclusive, but the authors suggest-ed that socioeconomic factors, pover-ty in particular, were likely to be

important in this regard. Analysis oflung cancer incidence data for theUSA has revealed a bigger gap in mor-tality between black men and womenthan between white men and women;in the black population themale:female ratio of lung cancer inci-dence is 2:1, while in the white popu-lation it was only 1.5:1 (SEER, 2003).

Several studies in developed coun-tries have shown that lung cancermortality is higher among lowerincome groups (Steenland, Henley &Thun, 2002; Richmond, 2003), a find-ing that reflects differences in expo-sure to lung cancer risks, includingtobacco, and also the impact of depri-vation and poor socioeconomic condi-tions on lung health (Hart et al.,2001). Nevertheless, gender differ-ences remain, with higher lung cancermortality in men on lower incomesbeing a frequently reported outcome(Hart et al., 2001).

2.2 Sex, gender and lung can-cer: explaining the differencesbetween women and men

What then are the roles of sex andgender in determining the differencesbetween women and men in their riskof lung cancer morbidity and mortali-ty? Although there is a growing bodyof research which adds to our under-standing of the relative roles playedby sex and gender, and which alsohas begun to suggest ways in whichsex and gender may combine toincrease risks, the picture remainscomplex and not fully understood. Inparticular, there is a need to differen-tiate between established researchfindings and those that are more sug-gestive and also to identify where fur-ther research is necessary. Some

13

studies have suggested that womenare more vulnerable to lung cancerthan men at the same level of expo-sure – and it is this that has drivenmuch of the recent research into bothsex and gender-linked factors in theetiology of lung cancer. However, thisfinding is still speculative and contro-versial. The state of research into thekey factors that have been associatedwith differences in lung cancer pat-terns between women and men,beginning with those related to gen-der, is summarized in the followingsubsections.

Gender-linked factors

Gender-linked factors can be dividedinto those that affect exposure to riskfactors and those that affect accessto treatment. In the case of lung can-cer the most significant risk factor issmoking, and one of the most press-ing research questions, in terms of pri-mary prevention, is how do the deter-minants of smoking differ betweengirls and boys, and between womenand men? Other avenues of investiga-tion are concerned with differences insmoking behaviour between malesand females, and how these mightplay a role in determining lung cancerrisk.

Smoking

Worldwide, there are more male thanfemale current smokers. About 47%of all men and 11% of all womensmoke, with men accounting for four-fifths of all smokers (Jha et al.,2002). Table 3 (see next page) pres-ents data on smoking prevalence inselected countries for which data areavailable and illustrates the wide vari-ation that exits in the ratio ofmale:female smokers. In India, for

example, there are 12 times as manymen smokers as women smokers, butin Norway and Sweden the numbersare fairly even. The globally-averagedratio of male:female smokers is esti-mated to be 4.3:1 (Jha et al., 2002).

In a number of developed coun-tries, the proportion of women whosmoke has in recent years approachedand, in some cases overtaken, theproportion of men who smoke – par-ticularly among the younger agegroups (WHO, 2001). Taking the USAas a typical example, in 1998, 22% ofall women smoked compared with26% of all men – a narrowing of thesex ratio which reflects the dramaticdecrease in the number of male smok-ers over recent years (USDHHS,2001). Although rates of smoking arecurrently low among women in devel-oping countries, some parts of theworld, China for example (Tomlinson,1997), have seen an increase in thenumbers of women taking up smok-ing. According to recent estimates,60% of Chinese men and 8% ofChinese women smoke (WHO, 1997).The tobacco industry, in its search fornew markets, has adopted aggressivemarketing techniques in an effort todevelop smoking in this sector (Samet& Yoon, 2001; Richmond, 2003).

Current research indicates thatmale:female differences in smokingprevalence alone does not accountfor the observed patterns of lung can-cer incidence and mortality in womenand men, and thus, other factorsmust be playing a role. It has beensuggested that differences in smokingbehaviour, i.e. type of tobacco, depthof inhalation and speed of consump-tion of cigarettes, may provide part ofthe explanation. Differences betweenwomen and men that may be particu-

14

15

Male Female M:F ratio Period

India 29 2.4 12.1 : 1 1998-1999

Algeria 43.8 6.6 6.6 : 1 1997-1998

Japan 47.4 11.5 4.1 : 1 2000

Pakistan 36 9 4.0 : 1 1996

Uganda 52 17 3.1 : 1 unknown

Slovakia 41.1 14.7 2.3 :1 1998

Kenya 66.8 31.9 2.1 : 1 unknown

Cuba 48 26.3 1.8 : 1 1995

Spain 39.1 24.6 1.6 : 1 2001

USA 25.7 21 1.2 : 1 2000

Denmark 32 29 1.1 : 1 2000

Norway 31 32 1 : 1 1999-2000

Sweden 17.4 20.4 0.9 : 1 2000-2001

Nauru 49.8 59 0.8 : 1 1994

Table 3Smoking prevalencea in women and men in selected countries

a Percentage of the relevant population who smoke.

Source: Shafey O, Dolwick S, Guindon GE (eds). Tobacco control country profiles 2003. Atlanta,GA, American Cancer Society, 2003.

larly significant in terms of lung can-cer risk relate to the type of tobaccoused (i.e. composition of the ciga-rette, cigar or tobacco) and the way inwhich cigarettes are smoked (i.e.deep or superficial inhalation). Womensmokers are more likely to use ciga-rettes which are lower in tar and nico-tine, and which have been labelled as"mild" or "light" by the tobacco indus-try (Richmond, 2003). Indeed, suchlabels have been seen by the tobaccoindustry as having a greater appeal forwomen, in that they carry associa-tions with weight control. This associ-ation has been strengthened by thepractice of tobacco companies, notonly in their advertising, but alsothrough sponsorship of such activitiesas beauty pageants in some countries(Christofides, 2002). The WHOFramework Convention on TobaccoControl (WHO, 2003) has called for aglobal ban on the use of the terms“mild” and “light”, as well as strictercontrol over such sponsorship andadvertising.

Evidence that emerged in the1990s postulated that one effect ofcigarettes with a lower tar and nico-tine yield is that smokers tend to drawmore heavily on such cigarettes inorder to achieve the desired level ofnicotine delivery – thus increasingtheir exposure to tar and other car-cinogens (IARC, 2002a; Shields,2002). Were this to be the case,women smokers could conceivably beat greater risk from the gendered con-struction of "feminine" cigarettes thathave more severe health conse-quences. However, the interplay offactors here is particularly difficult tounravel. It is important to note thatthe effect of "low" tar cigarettes ondepth of inhalation may be strongerfor those who have switched brands;

Shields' (2002) research, forinstance, looks at people who haveattempted to reduce their exposure tolung cancer risk by switching brands,rather than at those who have onlyever smoked low tar/nicotine brands.

Although it is generally acceptedthat smokers may well compensatefor a lower tar and nicotine yield byadapting their smoking pattern, thequestion remains open as to whetherthis actually results in a higher risk oflung cancer for those who opt to uselow tar cigarettes. The Working Groupon Tobacco Carcinogenicity of theInternational Agency for Research onCancer (IARC) concluded that whilechanges in cigarette composition inthe last 50 years had contributed toreductions in lung cancer risk associ-ated with the number of cigarettessmoked, these same changes mayalso have affected patterns of ciga-rette consumption, particularly interms of how people smoke (IARC,2002a). A study by Woodward(2001) indicated that smokers whomaintain the same smoking patternand depth of inhalation may be less atrisk if they smoke cigarettes with lessthan 10 mg of tar, compared withthose who smoke cigarettes withhigher levels of tar.

Other risk factors

Other gender-linked risk factors rele-vant to the etiology of lung cancerinclude exposure to environmentaltobacco smoke (ETS), fumes andsmoke from certain cooking fuels andmethods, environmental pollution andemployment-related risks in certainindustrial settings. Some gender-linked factors may operate to reducethe risk of lung cancer, for example, adiet rich in fruit and vegetables.

16

However, at present evidence here islimited and inconclusive (IARC,2002b; 2003).

On the whole, research indicatesthat environmental factors may affectwomen more than men, particularlywith respect to the risks associatedwith cooking and with second-handsmoke or ETS. Studies in countrieswhere women are more likely to belifelong non-smokers have found highlevels of lung cancer among non-smoking women associated both withcooking vapours released from the oilused in cooking and with exposure tofumes from coal used in cooking inthe home (Gao et al., 1987; Granvilleet al., 2003; Keohavong et al., 2003;Kleinerman et al., 2002; Metayer etal., 2002; Wen Cheng & Lee, 2003).These are specific gender effects thatact to increase women's risks as aresult of the sexual division of labourin the home. Studies have also high-lighted a role for wider environmentalpollution as an additive factor in expo-sure to lung carcinogens (Hu et al.,2002; Wen Cheng & Lee, 2003).However, recent research wouldappear to indicate that although envi-ronmental risk factors by themselvesare significant, it is the combination ofthe exposure to such risks and genet-ic differences that add up to producehigher risks of lung cancer for womenin these circumstances (Siegfried,2001).

Evidence relating to the impact ofETS on lung cancer risk has suggest-ed that women may be more at riskbecause women are more likely to livein a household where a partnersmokes than are men (Zang &Wynder, 1996). However, other stud-ies have pointed out that men aremore exposed to ETS in the work-

place due to: a) higher rates of smok-ing historically among men, and b) thesexual division of labour in the work-place is such that men are more likelyto work in all-male, high-smokingenvironments. As a result, men mayhave had a greater lifelong exposureto tobacco smoke (Siegfried, 2001).While some studies have suggestedthat at similar levels of exposure toETS, women are more susceptible tolung cancer (Kreuzer et al., 2002), therecent IARC study (2004) suggestsrisks are similar for women and men.

An increased risk of lung cancerhas also been documented amongthose exposed to occupational lungcarcinogens, such as insecticides,pesticides and diesel. Research of thistype has tended to focus on the tradi-tional male occupations (Nurminen &Karjalainen, 2001; Brown et al.,2002; Rachtan, 2002; Chan Yeung etal., 2003). Nevertheless, there issome evidence to suggest that theremay be differences between womenand men in the strength of the risk.For example, a large record-basedstudy of Swedish women and menworking in occupations where theywould be exposed to diesel emissionsreported an increased risk of lung can-cer for men (relative to the male pop-ulation), but not for women (Boffetaet al., 2001). However, this study didnot have access to information onsmoking.

Sex-linked factors

Sex-linked factors include both thosethat are related to biological differ-ences between women and men (inparticular, the role of reproductivehormones) and those that are relatedto genetic differences that may play apart in determining vulnerability in the

17

context of other factors. Sex-linkedfactors have been identified as beingthe most likely explanation ofwomen's increased vulnerability tolung cancer at the same level of expo-sure to risk factors, such as tobaccosmoke (Siegfried, 2001).

Research on hormones and lungcancer has indicated that later age atmenarche is associated with adecreased risk of lung cancer, as is alonger menstrual cycle length (Gao etal., 1987; Brenner et al., 2003). Onestudy reported differences in histolog-ic cancer type in relation to both thelength of the menstrual cycle andnumber of days of bleeding (Liao etal., 1996). However, Kreuzer et al.(2003), in a case-control study inGermany, found no clear associationbetween lung cancer incidence andreproductive events such as age atmenarche, length of menstrual cycle,number of live births or age atmenopause, although an associationwas found with use of exogenoushormones (i.e. oral contraceptives andhormone replacement therapy orHRT). However, other studies, forexample, the work of Blackman et al.(2002) failed to demonstrate a linkbetween the use of exogenous hor-mones and lung cancer.

One of the main problem areas inthe research looking at hormonal influ-ences on lung cancer risk is the differ-entiation of results by smoking status.For example, research into the role ofthe length of the menstrual cycle sug-gests that the nature of this impact onindividual risk might vary according tosmoking status. Seow et al. (2002)found that longer cycles and three ormore live births were associated withsignificantly reduced risks of lung can-cer, but only among lifelong non-

smokers. It should be noted that thisresearch was carried out in China andstudies have not assessed the applica-bility of these findings to non-Chinesewomen.

Other research focusing on geneticsusceptibility has demonstrated thatwomen are more likely than men tocarry genetic mutations that are asso-ciated with an increased risk of lungcancer (Dresler et al., 2000; Siegfried,2001; Stabile et al., 2002; Stabile &Siegfried, 2003; Sy et al., 2003).Studies involving gene expressionindicate that hormonal influences onlung disease may affect both theimpact of tobacco smoke on the lungs– increasing the risk of carcinogenesis –and the ability of the lungs to recoverfrom, or repair, damage (Haugen,2002; Stabile et al., 2002). Shields(2002) have suggested that geneticsusceptibility is an important elementof lung cancer risk and that suscepti-bility is associated with tobacco usethrough the metabolism of nicotineand cell repair (Taioli & Wynder,1994; Siegfried, 2001)

In sum, it is fairly clear then thatboth gender and sex play a part inshaping lung cancer risk. However,the indications are that, in addition tothe independent risks posed by each,there are important interactionsbetween the gender- and sex-linkedrisk factors that play a further part. Inother words, the impact of genderedfactors on lung cancer risk may wellbe mediated by sex-linked factors.

18

This section outlines the evidenceregarding gender differentials inaccess to health care and in the qual-ity of the care that is available to menand women in relation to lung cancer.Health care issues considered includehealth promotion and support forsmoking cessation, access to screen-ing, diagnosis, treatment for lung can-cer, and other factors relating to sur-vival.

3.1 Smoking cessation

Clearly, public health initiatives toreduce smoking are at the centre oflung cancer prevention strategies forboth men and women. Preliminaryresearch has suggested that smokingcessation has clearer advantages forwomen than for men in terms ofreduced cancer risk, although this is,at present, a matter of some contro-versy. Connett et al. (2003), forexample, found that women who quitsmoking achieved a more significantreduction in chronic obstructive pul-monary disease (COPD), a risk factorfor lung cancer, compared with men.However, the authors of this studydid not adjust for differences in otherrisk factors in the study population,such as occupational risk factors forCOPD. Research on smoking cessa-tion has also found differencesbetween women and men in their suc-cess in stopping smoking, and in thevalue of nicotine substitutes and othersupport systems in maintaining cessa-tion. Some studies on smoking cessa-tion have indicated that women giveup more often than men but alsorelapse more often (Borrelli et al.,

2001; Perkins, 2001); others suggestthere are few differences betweenmen and women in attempts to quitsmoking and relapse rates (USDHHS,2001).

Factors affecting success in quit-ting smoking are likely to be related toboth biology and gender. In terms ofthe biological or sex-linked factors,there is some evidence to suggestthat nicotine affects women and mendifferently. For example, nicotinewithdrawal appears to be moreintense for women (Perkins, 2001).Furthermore, whereas some studiesindicate that women are more readilyaddicted to nicotine, nicotine therapyalso appears to be less effective forwomen than men (Tewari & Disaia,2002). The other sex-linked factor insmoking cessation to emerge fromresearch relates to women of repro-ductive age, with some studies sug-gesting that women's success in theirattempts to quit smoking may berelated to the stage of menstrualcycle when the quit starts (Allen etal., 1999; Pomerleau et al., 2000).

Looking at the gender-related fac-tors in smoking cessation, Borrelli etal. (2001) in a study in the USA foundthat men who gained weight afterstopping smoking were more likelythan women who gained weight torelapse in the first three months of aquit. Other studies have shown thatwomen with a higher body massindex (BMI) are more likely to stayquit than women whose BMI is low,perhaps because women who areoverweight already are less concerned

19

3. Gender issues in access to care

about further increases in their weight(Osler et al., 1999). The level of socialsupport provided during attempts tostop smoking has been proposed as apossible factor in smoking cessation.While some research suggests socialsupport is more important for womenin cessation attempts (Gritz, Nielsen& Brooks, 1996; USDHHS, 2001),Westmaas, Wild & Ferrence (2002)concluded from their study of 93 menand 117 women that social supportmay be more closely associated withsmoking reduction for men than forwomen. During an attempt to stopsmoking, men's success was associ-ated with a higher self-reported influ-ence of partners, family and friends,whereas for women, increased self-reporting of partner, family andfriends' influence was associated withsmaller reductions in smoking(Westmaas, Wild & Ferrence, 2002).

Success in smoking cessation islargely governed by two factors: theunderlying reasons for tobacco take-up and the reasons for continued use,and again research suggests that boththese vary between women and men.Many of the studies of smoking initia-tion among young people that havebeen carried out to date indicate that,in developed countries at least, girlsmay be more likely than boys to con-tinue to smoke after experimentation(Best et al., 2001). Moreover, dietingmay be a factor influencing the transi-tion in girls from experimental smokerto more habitual tobacco use (Austin& Gortmaker, 2001).

Research into factors that are con-sidered to be important in continuedsmoking suggests that women aremore likely to remain smokers inresponse to depression, stress, thedesire to control weight and the need

for "time out" in caring responsibili-ties. Men on the other hand are morelikely to smoke as part of relaxation.Compared with men, women smokemore often in stressful and "higharousal" situations (Gritz, Nielsen &Brooks, 1996) and are more likely torelapse in the face of stressful lifeevents, particularly health and finan-cial stressors (McKee et al., 2003).Women are also reported as beingmore likely than men to use tobaccofor the sensory aspects of cigarettes(Perkins, 2001) and as a form of self-treatment for feelings of depressionand anxiety (Gritz, Nielsen & Brooks,1996).

Gendered differences betweenwomen and men in smoking triggersand associations (in the form oftobacco advertising) also affect thelikelihood of successful quits.Tobacco marketing targets womenand men smokers in different ways:messages to women tend to linksmoking with maturity, confidence,sexual attractiveness and beauty, par-ticularly in developing countries(WHO, 2001). The promotion of"mild" and "light" cigarettes has typi-cally involved associations withweight control (Christofides, 2002;Richmond, 2003). Despite calls for aban of the use of the terms "light"and "mild", and also for global con-trols on tobacco advertising, theseassociations remain important. It isessential that differences betweengirls/boys and women/men in the fac-tors underlying smoking initiation andcontinuation are fully understood ifprimary prevention is to be targetedeffectively at both sexes.

Pregnant women, as a group inregular contact with health careproviders, and because smoking dur-

20

ing pregnancy is known to adverselyaffect fetal and neonatal growth anddevelopment, are a key target groupfor cessation programmes. In terms ofthe relationship between pregnancyand smoking cessation, factors thatare likely to be of particular interestare both biological – related to repro-ductive factors which affect smokingbehaviour (e.g. nausea during theearly months) and the use of pharma-cological interventions – and gendered– related to social constructions ofmaternity. The safety of nicotinereplacement therapy in pregnancy hasnot been well studied (Dempsey &Benowitz, 2001). Investigation of thebenefits of cessation support fromhealth providers has produced mixedresults; nevertheless, it is generallyaccepted that cessation programmesneed to recognize a complex range offactors involved in smoking behaviourduring pregnancy, including, forexample, the role of stress (Ludman etal., 2000; Bullock et al., 2001).

3.2 Screening and treatment

In comparison with smoking cessa-tion, screening is at present lesseffective as a means of reducing lungcancer mortality, and for this reasonthere has been little research on gen-der differences in the value, or meth-ods, of screening programmes.Although several trials have been car-ried out (Henschke et al. 2001;Tyczynski, Bray & Parkin, 2003;Wisnivesky et al., 2003), screeningprogrammes for lung cancer have notbeen widely adopted. Trials havedemonstrated that computed tomog-raphy (CT) screening may be able todetect Stage I cancers and improvesurvival rates, but given the poorprognosis of lung cancers and the pro-hibitive financial and other costs asso-

ciated with such screening, large-scale screening programmes are notconsidered to be a cost-effectivemeans of reducing lung cancer mortal-ity. Until such time as studies usingrandomized-control trial (RCT)methodology become available toprove otherwise, neither mass screen-ing nor the screening of high-riskgroups (such as heavy smokers), islikely to be seen as a worthwhile pub-lic health strategy (Wisnivesky et al.,2003). Such RCTs would need toconsider gender dimensions and therelative value of screening for womenin comparison with men.

Similarly, there are few studiesthat have explored gender differencesin diagnosis. In 1991 a report fromthe American Medical Association(AMA) concluded that women andmen did not have equal access todiagnostic studies for lung cancer;research protocols often meant thatwomen were less eligible for inclusionin studies of this type. However, thestudy did not offer any suggestionsfor possible reasons for this finding. Alater study conducted by Lam et al.(1999) in the USA, which looked atsmoking-related changes in lung func-tion and in bronchial epithelial cells,identified differences in pre-invasivelesions and airflow obstructionbetween men and women. These dif-ferences could be of particular signifi-cance in the context of screening anddiagnostic tests. In Japan, wherescreening has been introduced, arecent study on the time lapsebetween screening and diagnosisreported no differences between menand women in the time betweenscreening and the first hospital visit(Kanashiki et al., 2003). These find-ings may have implications for bothdiagnosis and the point at which

21

treatment might start.

Research on differences betweenwomen and men in the kind of treat-ment offered once a diagnosis of lungcancer is made is also thin on theground, again probably because treat-ment has little impact on survival.Some research has reported thatwomen's survival rates are higherthan those for men, a finding thatmay, at least in part, be a reflection ofthe type of lung cancer diagnosed

(Alexiou et al., 2002; Radzikowska,Gaz & Roszkowski, 2002; Bremnes etal., 2003). For example, Radzikowska,Gaz & Roszkowski (2002) concluded,on the basis of their multivariateanalysis of data on 2875 women and17 686 men diagnosed with lung can-cer in Poland over a five-year period,that advanced stage, non-surgicaltreatment, age > 50 years at diagno-sis and male gender were all signifi-cant independent negative prognosticfactors.

22

There is now a large number of papersreporting primary research on sex andgender factors in lung cancer and sev-eral excellent reviews of the literaturein this subject area (see Box 2, nextpage). Although much of this stemsfrom clinical medicine, there are alsostudies from the field of health, psy-chology and from the social sciences.To date, research has to a largeextent focused on the biological fac-tors, namely, the reproductive andhormonal differences betweenwomen and men, and the part playedby genetics, in order to better under-stand the male:female differences inthe patterns of lung cancer mortality.Much less attention has been paid togender-linked factors that are alsolikely to play a key role in the etiologyof lung cancer. Where research hasexplored the issue of gender, it hastended to be relation to specificbehaviour, that is to say, smoking,although there has been in recentyears a growth in efforts to investi-gate the gender-linked risks associat-ed with exposure to emissions fromcooking and domestic fuels in someparts of the world. In addition, someresearch has been conducted on gen-der factors associated with exposureto environmental tobacco smoke andto occupational risk factors.

It is now widely acknowledgedthat gender-sensitive research isessential to the understanding of therole played by various risk factors and

their interactions in the etiology oflung cancer. One way of viewing theneed for gender-sensitive research isto ask what happens when research isgender in-sensitive. In terms of lungcancer research, empirical informationon sex differences in risk is lackingwhen research does not recognize thegendered nature of some of the risksidentified – for example, the fact thatwomen are responsible for domesticwork leads, in some countries, to anincreased exposure to the risks asso-ciated with cooking fuels. Whenresearch focuses solely on the biolog-ical factors that influence women'srisk of lung cancer, the opportunity tohighlight and challenge the risks stem-ming from gendered inequalities ismissed. Similarly, research on tobac-co use which is gender in-sensitivecannot question underlying motiva-tions for unhealthy behaviour.

In the field of lung cancer researchthere are now well-established gendereffects on lung cancer; there are alsoa number of findings which, as yet,can only be described as suggestive.Well-established effects include differ-ent patterns of smoking for womenand men over time, and different lev-els of exposure to occupational car-cinogens and to environmental fac-tors. Box 3 (see page 26) provides acritical review of one of the firstpapers to explore in depth the differ-ences in risk factors, especially smok-ing, experienced by women and men.

23

4. Lung cancer research: knowledgegaps and recommendations for

future research

24

Sex and gender differences in lung cancer

Summary

This paper is a review of epidemiological evidence relating to lung cancer and,in particular, examines the differences between women and men. The reviewlargely focuses on research papers from China, the USA and Europe. Thepaper includes useful data on lung cancer incidence and mortality, as well assmoking. The paper clearly outlines the meaning of sex and gender in the con-text of lung cancer risk and provides a valuable summary and discussion ofresearch in this field.

What gender issues does the paper cover?

The paper covers in detail the evidence relating to sex-linked factors in lungcancer and as such provides a useful summary of research in this field. Thesection on gender-linked factors is shorter and focuses on cigarette smoking.

Critique

This paper is highlighted because it distinguishes the different factors thatshape lung cancer incidence and mortality patterns with greater clarity thanmany other reviews of the subject. The summary of sex-linked factors in lungcancer is especially good. However, the paper does not reflect at length onthe role of gender in shaping lung cancer risk for women and men.

(Stabile & Siegfried, 2003)

Many of the gender-linked risk factorsare preventable, unlike some of theother risk factors associated with bio-logical differences between men andwomen where public health initiativeshave less to offer. For this reason,while gender-sensitive research is anapproach that needs incorporating inall aspects of work in this field, it ismost valuable in relation to increasingour understanding of the gender-spe-cific aspects of primary prevention,and of smoking cessation in particu-lar.

Considering the role it plays in lungcancer, the research focus on smok-ing is only natural. There is a largebody of literature on the psychologyof smoking and how this relates toself-image, other substance use andself-esteem, especially amongyounger people. Less attention hasbeen paid on the wider gender pic-ture, that is to say, much of the psy-chological literature on tobacco uselooks at individuals but not at thewider social context of smoking andthe structural factors that influencesmoking behaviours. Although thereare noteworthy exceptions to this(Graham & Blackburn, 1998; Graham& Der, 1999), the majority of studiesof this type concentrate on women.What this limited literature suggests isthat women and men's smokingbehaviour does differ in importantways but that such differences haveto be understood across the entire lifecourse and in the context of structur-al factors such deprivation and disad-vantage.

There is also very little research ongender differences in how tobacco isused – how men and women smokeand whether, for example, smokinglocation and circumstances affect

depth of inhalation, length of timetaken to smoke a cigarette and lengthof stub, all factors that may affectrisk. Again, these differences are lesslikely to be individual, and more likelyto be associated with structural fac-tors such as paid and unpaid work,caring responsibilities, stress, experi-ence of deprivation and so on. Thework of Prescott et al. (1998) is note-worthy in this respect as it one of thefew to attempt to explore the natureof gender differences in tobacco use(see Box 4, page 27).

4.1 Current challenges in lungcancer research

Although the volume of research onsex and gender differentials hasincreased in recent years, there arestill significant gaps in our under-standing of the ways in which sex-and gender-linked factors relate tolung cancer risk. There are a numberof reasons for this. One difficulty isthat sex and gender are terms that arenot used consistently across all disci-plines, or even within disciplines, andalso vary over time and place. Despitethe existence of a clear delineation inthe meaning of these concepts (WHO,1998), some researchers continue touse "gender" to refer to biological fac-tors and to men and women in astudy, rather than restricting the useof the term to reflect socially-con-structed differences between womenand men (see Box 4, page 27).

The second problem arisesbecause of the complex nature of theinteractions between sex and genderin the etiology of lung cancer. A num-ber of recent papers have highlightedthe epidemiological difficultiesencountered in trying to understandthe nature of the risks experienced by

25

26

Differences in lung cancer risk between men andwomen: examination of the evidence

Summary

This paper was one of the first to explore in detail the male:female differencesin lung cancer incidence in relation to the major risk factor, smoking. It fol-lows an earlier study by the same authors, but data are analysed and report-ed in more depth here. The analysis is based on a hospital-based case-controlstudy conducted in the USA, involving 1889 subjects with lung cancer (1108men, 781 women), and 2070 control subjects (1122 men, 948 women).Using pair-matched controls, adjusted odds ratios (ORs) of lung cancer formen and women were calculated. Results indicated that women with lungcancer were more likely to be never-smokers than men, that men with lungcancer started smoking earlier, reported inhaling more deeply and smokedmore cigarettes per day than women with lung cancer. However, dose-response ORs were higher for women than men, even after adjusting for fac-tors such as body mass index (BMI).

What gender issues does the paper cover?

The paper explores differences in lung cancer risk for women and men andanalyses the data by gendered factors – smoking history in particular. Thereis a recognition and attempt to deal with differences between women andmen in terms of smoking behaviour – age of smoking initiation, number of cig-arettes smoked, and depth of inhalation. The research included detailed and"precisely quantitated" smoking exposures as well as a careful analysis ofcancer incidence by histologic type. Differences related to sex and gendersuch as weight, height and BMI, were also considered.

Critique

This is a ground-breaking paper in the development of research on lung can-cer risk for women and men and has been highly influential in shaping subse-quent research. Discussion of this paper by other authors since its publicationhas focused on the way in which data were presented in relation to differ-ences between women and men smokers in the risk of lung cancer. The paperrepresents an important original contribution to the question of sex-linked dif-ferences between women and men in their risk of lung cancer.

Zang & Wynder, 1996

27

Mortality in women and men in relation to smoking

Summary

In this paper, pooled data from three prospective population studies inCopenhagen are used to compare total and cause-specific mortality in relationto smoking habits. A sample population, comprising more than 30 000 indi-viduals whose date and cause of death was recorded, was monitored between1964 and 1994. Information was collected from individuals via a self-admin-istered questionnaire on smoking behaviour in never-smokers, ex-smokers,those smoking fewer than 15 cigarettes a day, and those smoking more than15 cigarettes a day. Data were also collected for those using other forms oftobacco, and on inhalation.

Positive associations were confirmed for both men and women for smok-ing and lung cancer (together with other causes of death). The authors notedthat while relative risks associated with smoking were higher for women inrelation to respiratory and vascular disease, there were no differencesbetween women and men in the relative risk of smoking-related cancers. Theauthors cautiously concluded that although women may be more sensitivethan men in terms of some causes of death, lung cancer in not among them.

What gender issues does the paper cover?

The study uses a sufficiently large sample size, in terms of numbers of bothmen and women, to be able to address questions of difference in relative riskfor women and men. The population also includes adequate numbers offemale heavy smokers, which is important in the calculation of relative risk forthat specific category. It also considers factors such as age at smoking debutand reported inhalation, for women and men, and due to the longitudinalnature of the study, differences in tobacco use over time were also observed.In this respect, the paper recognizes the importance of gender differences intobacco use.

Critique

The authors use "gender" to refer to women and men, to social differencessuch as age at smoking debut, and to differences between women and menin parameters that reflect sex as well as gender (all-cause mortality for exam-ple). However, the broad use of the word, gender, may be the result of edito-rial policy of the journal. Like many studies of its type, it suffers from the usualdifficulties associated with collecting accurate data from individuals via themeans of a self-administered questionnaire. Possible sources of error includememory error and unwillingness to report actual consumption. This paper isbased on large-scale survey data with a relatively high response rate (77% forthe first examination).

Prescott et al., 1998

women and men, particularly the addi-tive as opposed to the interactiveproperties of some risks (Perneger,2001; Perneger, 2002; Risch & Miller,2002; Zang, 2002). Thirdly, whereasthe type of carcinoma, length of sur-vival and genetic mutation are all vari-ables that may be measured with rea-sonable scientific accuracy, smokingbehaviour, a particularly importantfactor, is much more difficult toquantify. Many researchers discusscigarette smoking in terms of "packyears", which is calculated as thenumber of packs smoked per daytimes the number of years of smok-ing. However, we do not know howaccurately this measure reflects actu-al smoking practice, and how theaccuracy of the measure might varyfor women and men.

We do not fully understand the rel-ative risks for women and men whosmoke, or how these risks are modi-fied by other factors (USDHHS,2001). This is due, in part, to theproblems associated with monitoringany self-reported behaviour that isknown to have adverse health conse-quences and, partly to the inherentcomplexity of the factors that influ-ence smoking behaviour.

4.2 What is required to addressthe knowledge gaps in lung cancerresearch?

Much of the research carried out inrecent years has focused on explana-tions of the apparently greater vulner-ability of women compared to men atthe same level of exposure. There hasbeen less exploration of men's vulner-ability to lung cancer – how thisvaries in response to different risks,how it varies across the life course,

and what such variations might tell usabout the risks for both men andwomen. Although there is a continu-ing need to develop the scientificbasis of our understanding of lungcancer risk for women and men, thereis also a need to further our under-standing of the ways in which suchrisks are mediated by other forms ofdiversity, in particular ethnicity andincome. Nevertheless, given that themost significant risk factor in lungcancer is known and preventable, pri-ority should be given to interventionsto promote and support smoking ces-sation.

Research methods and methodology

In exploring the impact of differentfactors on lung cancer risk and onprognosis and experience of the dis-ease, researchers need to identifysex-linked and gender-linked factorsconsistently. Gender-sensitive researchthus implies a requirement for aframework in which sex and genderare clearly conceptualized. In addition,the conceptual framework wouldneed to include explicit recognition ofthe ways in which sex and gendermight interact.

To ensure gender-sensitivity, suffi-cient numbers of both women andmen must be included in researchstudies so as to allow meaningfulanalysis, not just from the point ofview of generating disaggregated datafor women and men but also for spe-cific subgroups of each. The role ofhormonal factors in lung cancermeans that age and stage in reproduc-tive life course are important consider-ations for women which need to betaken into account when planningsample size.

28

Lung cancer research needs toensure the active involvement ofwomen and men at every stage of thecycle, beginning with the commission-ing process, through the developmentstages, the actual experimental work,and right up to the data analysis andreporting stage (Christofides, 2002).In addition, findings must be analysedand reported in such a way that it isclear who has participated in theresearch – both as "researched" andas researchers. The way in which par-ticipation in the research is arranged,not just the sample design and recruit-ment, but also the handling of theproject itself, should also be transpar-ent.

A useful resource in this context issome work produced by a workinggroup charged with the task of identi-fying research needs for effectivetobacco control (Samet et al., 1998).The recommendations of the workinggroup, which are set out in Table 4(see page 31), offer a helpful guid-ance on the principal methods thatwould be of use in such research, andmost significantly, recognize the needto take gender into account in bothobservational studies and interventiontrials.

Smoking and tobacco control

Although the relationship betweenrisk factors and vulnerability to lungcancer may differ for women andmen, there is a need to prioritizeresearch that addresses the most sig-nificant risk factor – smoking. Keyquestions are:

In what ways does smoking behav-iour differ for women and men?

How does smoking behaviour vary

during the course of an individual'slifetime and in the context of otherresponsibilities, such as caring work?(For example, does smoking behav-iour, including depth of inhalation andnumber of puffs taken, alter in relationto responsibilities such as the care ofyoung children?)

How does smoking behaviour varyby income and ethnicity?

Research in some countries sug-gests women in particular are likely tounderreport smoking due to culturalpressures (Christofides, 2002); insuch cases the need for research onsmoking behaviour which is aware ofthese gendered cultural factors, andaddresses women appropriately inorder to collect accurate information,is especially acute.

The relationship between women'ssmoking and changes in women'seconomic and social position is ofprime importance, particularly in lessdeveloped countries where questionsrelating to the likely impact of devel-opment on gendered differences insmoking behaviour need to be asked.Such research also needs to considerhow these influences can beaddressed. Research in Europe onwomen's smoking suggests that asso-ciations between social disadvantage,life course influences and smoking areimportant; here the aim of gender-sensitive research is to guide thedevelopment of suitable policies withthese influences in mind (Bostock,2003).

At present, evidence regarding dif-ferences in the ways in which tobac-co advertising affects women andmen is mostly only suggestive, yetthe nature of the tobacco industry's

29

gender-specific approach to sellingtheir goods highlights the need formore work in this field if strategies areto be effective in the future.Questions that should be added tothis agenda include those relating tothe impact of product placement infilms and television.

In terms of the solution to the lungcancer epidemic, primary preventionin relation to smoking is critical. Thus,there is a very definite need forresearch to evaluate the effectivenessof tobacco control initiatives as theyaffect men and women, as well assubgroups such as young women andyoung men, and women and menfrom different ethnic groups withincountries (Greaves & Barr, 2000).Available evidence, albeit rather limit-ed at this point, relating to ethnic dif-ferences in smoking initiation and ces-sation indicates the need for tobaccocontrols to take account of differentpatterns of smoking and tobacco use.Tobacco control targets that areappropriate to the stage a country hasreached in the tobacco epidemic arealso needed, and in this context,research needs to be sensitive to localconditions and knowledge(Christofides, 2002).

Given that smoking cessation is aglobal public health issue, research onfactors that influence smoking cessa-tion is high on the list of priorities. Astudy by Osler et al. (1999), whichlooks at unaided smoking cessation ina large sample of smokers in the USA,provides some valuable insights intobehavioural and social determinantsof smoking cessation (see Box 5,page 32), but more research in thisarea is still required. More research isalso urgently needed to add to currentunderstanding of the biological differ-

ences in the effectiveness and valueof pharmacological interventions inattempts to quit smoking. Thisresearch must also include genderquestions about the effectiveness ofother forms of support in smokingcessation, and needs to be carried outso as to address questions of differ-ences between groups of women andmen, again related to age, income orclass and ethnicity.

Finally, there are outstanding ques-tions relating to the impact of employ-ment in the tobacco industry on menand women's use of tobacco. Heretoo gender factors are likely to be sig-nificant, with many women depend-ent on tobacco cultivation for a liveli-hood, despite the low wages paid inthis sector (Mosoba Mosoba, 2003).The concern here is not confined tothe increased risk of smoking initia-tion, but also the fact that womenworking in this sector are exposed tooccupational health risks arising fromthe processing of the tobacco. Therisks of miscarriages, chest infections,poisoning from chemicals and fumi-gants, and cancer, for example, are allincreased in female tobacco workers(Mosoba Mosoba, 2003).

Screening and treatment

Research into the treatment andscreening of lung cancer is generallyaccorded lower priority than researchthat adds to the value of primary pre-vention. Nevertheless, there are someimportant gaps in our knowledge herethat need to be addressed. In terms ofdiagnosis, more information is neededif we are to be able to establish con-clusively the relative values of differ-ent diagnostic tests for women andmen. With treatment, little is knownabout the differences that may exist

30

31

Surveytype

Recommendations Gender dimension

Observationalstudies/generalsurveys

General population studies ontobacco use need to be:

standardized;replicable;of sufficient size to providestable age and sex-specificprevalence estimates;conducted every 5 - 10 years;flexible enough for morefrequent repeats if interventionsneed evaluation.

Surveys of health-care providersand others concerned with publichealth need to cover:

smoking among health-care workers;evaluation of interventions toreduce smoking in such workers.

Children's surveys to cover:school-based surveys;decisions to smoke;impact of tobacco advertisingand other forms of promotion.

Sample size critical, as isrecruitment and, wherecohorts are followed, a meansof maintaining contact; maydiffer for women and men.

Gender-specific questionsneed to be addressed regard-ing the context of tobaccouse, smoking "career" etc.

Need to be aware of culturaldifferences affecting reportingof smoking by women andmen.

Health-care workers differenti-ated by gender and profes-sional status – associatedlinks with smoking and cessa-tion need to be considered.

Girls and boys' motivationsregarding smoking need to beaddressed separately.

Observationalstudies/sur-veys ofadverse healtheffects

Repeat studies of earlier work toupdate findings, reinforce messageregarding adverse effects.

Case-control studies of tobaccorelated diseases.

Where earlier work hassex/gender gaps, need to buildon original methods to enablegeneration of appropriate data.

Case-control but also need toreport results separately forwomen and men.

Interventionstudies

All interventions should includean evaluation.

Studies should include modificationplans to follow ongoing evaluation.

Consider relevance of studiesacross different countries.

Differences in nature of nicotineaddiction and impact on cessationand other interventions need to beaddressed explicitly.

Evaluation needs to considerneeds of men and womenexplicitly and report separately.

Studies in other countries maynot have separate data onwomen and men.

Gender differences in nicotineimpact and addiction need tobe considered.

Table 4Conducting effective research for global tobacco control:recommended strategies (Source: adapted from Samet et al., 1998)

32

Gender and determinants of smoking cessation:a longitudinal study

Summary

This paper analyses spontaneous unaided smoking cessation in a longitudinalsample population in the USA. The paper claims to be the first to analysesocial, behavioural and health-related factors on spontaneous smoking cessa-tion in a study population which includes large numbers of both men andwomen.

The results reported are based on a sample of 4535 men and 4550 womenwho were smokers at baseline and attended the first re-examination and 2942men and 3111 women who also attended the second follow-up examination.Logistic regression was carried out to investigate the determinants of smok-ing cessation at each interval.

What gender issues does the paper cover?

The analysis presented in the paper revealed that sex and education, and sexand BMI, were key determinants of smoking cessation. The study confirmsfindings elsewhere that men are more likely to quit smoking after controllingfor factors such as age, education and amount smoked, and also that womenare more likely to relapse after quitting.

Critique

The strengths of the study lie in its large population size for both men andwomen. The focus on unaided cessation is also important, in light of otherresearch that suggests there may be differences between women and men intheir use of support in smoking cessation. Most other studies report on differ-ences among women and men in aided cessation, largely because this groupis more accessible to researchers by virtue of being in regular contact withhealth care professionals during their attempt to stop smoking. This papertakes a longitudinal approach, which allows the exploration of a range ofpotential determinants in the absence of health care intervention. The majorweakness is the interchangeable use of the terms gender and sex, and thelack of clarity in what these terms mean. However, it is possible that this maybe the result of editorial policy of the journal, rather than the fault of theauthors.

Osler et al., 1999

between women and men in the ben-efits of different interventions for lungcancer. This is of course a difficultarea of research given poor survivalrates, and ethical issues would be crit-ical. Similarly, there is little researchwhich evaluates the potential of sepa-rate screening for women and men.Screening trials need to considerways of reporting data that allow dis-aggregated analysis. Questions thatneed addressing include whether thepotential benefits of screening differfor women and men and whether dif-ferent criteria for selecting popula-tions for screening exist.

The consequences of lung cancerfor those receiving a positive diagno-sis is another area where not muchwork has been done, at least, not interms of gender. Research whichexplores gender issues not only in thecontext of treatment itself (i.e. thespeed of diagnosis and appropriatetherapies) but also in the context ofthe delivery of care at all stages of thedisease and support for others affect-ed by the diagnosis (i.e. families) isurgently required.

Sex and lung cancer risk: biologicalfactors

A good deal of research has beendone already on the role of geneticand reproductive factors in the etiolo-

gy of lung cancer, but less on the roleof hormones. What is required in thisarea is more research that leads to aposition of established findings, asopposed to those that are merely sug-gestive, especially in terms of the roleof hormones in the etiology of lungcancer. Future research also needs toaddress the additive and/or interactiveeffects between sex and gender.

More research is also needed ondifferences between women and menwith regard to the impact of protec-tive factors such as diet and exercise.In particular, we need to knowwhether these are really protective(ongoing research includes theEuropean Prospective Investigationinto Cancer and Nutrition) andwhether there are sex and/or genderdifferences in their effects in lungcancer risk. There is some indicationthat diets rich in vegetables, forexample, confer positive benefits andthat these may differ for women andmen (Axelsson & Rylander, 2002;Chan Yeung et al., 2003).Male:female differences in impact ofdiet is likely to be related to sex-linkedfactors, but access to different dietsis mediated by gender factors.

33

This paper has reviewed the availableevidence as it relates to sex and gen-der issues in lung cancer and offeredsome suggestions on how research inthis field might be developed towardsa more gender-sensitive model.Although there is a considerable bodyof research that explores differencesbetween women and men in their riskof lung cancer and, to a lesser extent,which looks at survival and differ-ences in smoking behaviour, it is farfrom complete. Concepts of sex andgender are not consistently applied,and there is often a focus on biologi-cal risk factors while truly gender-sen-sitive research which adds to ourunderstanding of differences between

women and men and our understand-ing of appropriate public health meas-ures remains limited. What we need,therefore, is research which explicitlyaddresses risk factors, in particulartobacco use, from a gender-sensitiveperspective; which devises researchprotocols, analyses data and reportson findings for women and men sepa-rately; and which considers evidencein relation to the potential interactionbetween sex, or biology, and genderor social structures.

34

5. Conclusion

Allen SS et al. (1999). Withdrawal and pre-menstrual symptomatology duringthe menstrual cycle in short-term smoking abstinence: effects ofmenstrual cycle on smoking abstinence. Nicotine and TobaccoResearch, 1:129-142.

Alexiou C et al. (2002). Do women live longer following lung resection forcarcinoma? European Journal of Cardio-thoracic Surgery,21:319-325.

American Medical Association, Council on Ethical and Judicial Affairs (1991).Gender disparities in clinical decision making. Journal of the AmericanMedical Association, 266:559-562.

Austin S, Gortmaker S (2001). Dieting and smoking initiation in earlyadolescent girls and boys: a prospective study. American Journalof Public Health, 91:446-450.

Axelsson G, Rylander R (2002). Diet as risk for lung cancer: a Swedishcase-control study. Nutrition and Cancer, 44:145-151.

Baldini EH, Strauss GM (1997). Women and lung cancer: waiting to exhale.Chest, 112 (Suppl.):S229-S234.

Best D et al. (2001). Ethnic and gender differences in drinking and smokingamong London adolescents. Ethnicity and Health, 6:51-57.

Blackman JA et al. (2002). Estrogen replacement therapy and risk of lungcancer. Pharmacoepidemiology and Drug Safety, 11:561-567.

Blizzard L, Dwyer T (2003). Case-control study of lung cancer during1994-1997 in the birth cohort in Tasmania, Australia, with an excessof female cases during 1983-1992. Cancer Causes and Control,14:123-129.

Boffetta P et al. (2001). Occupational exposure to diesel engine emissionsand risk of cancer in Swedish men and women. Cancer Causesand Control, 12:365-374.

Borrelli B et al. (2001). Influences of gender and weight gain on short-termrelapse to smoking in a cessation trial. Journal of Consultingand Clinical Psychology, 69:511-515.

Bostock Y (2003). Searching for the solution: women, smoking and

35

6. References

inequalities in Europe. London, NHS Health Development Agency.

Bray F et al. (2002). Estimates of cancer incidence and mortality in Europein 1995. European Journal of Cancer, 38:99-166.

Bremnes RM et al. (2003). The value of prognostic factors in small cell lungcancer: results from a randomised multicenter study with minimum5-year follow-up. Lung Cancer, 39:303-313.

Brenner AV et al. (2003). Menstrual and reproductive factors and risk oflung cancer among Chinese women, Eastern Gansu Province,1994-1998. Journal of Epidemiology, 13:22-28.

Brown LM et al. (2002). Exposures in the painting trades and paintmanufacturing industry and risk of cancer among men and womenin Sweden. Journal of Occupational and Environmental Medicine,44:258-264.

Bullock LF et al. (2001). Retrospective study of the association of stressand smoking during pregnancy in rural women. Addictive Behaviors,26:405-413.

Cancer Research UK (2003). Malignant neoplasm of trachea, bronchusand lung. Statistics: ICD-9 162, ICD-10 C33-C34 [online database].(http://www.cancerresearchuk.org, accessed 8 September 2003).

CDC (2002). Recent trends in mortality rates for four major cancers, by sexand race/ethnicity – United States, 1990-1998. Morbidity andMortality Weekly Report, 51:49-53.

Christofides N (2002). Why gender issues are an essential component in anyresearch strategy relating to tobacco use in developing countries[presented at the Global Health Forum 6, Arusha, Tanzanzia,12-15 November 2002, available online] (http://www.globalfo-rumhealth.org/forum_6/sessions/2wednesday/4GNCDChristofidesWhyGender.pdf, accessed 6 August 2004).

Chan Yeung M et al. (2003). Risk factors associated with lung cancer inHong Kong. Lung Cancer (Amsterdam), 40:131-140.

Connett J et al. (2003). Changes in smoking status affect women more thanmen: results of the lung health study. American Journal ofEpidemiology, 157:973-979.

de Perrot M et al. (2000). Sex differences in presentation, managementand prognosis of patients with non-small cell lung carcinoma. Journalof Thoracic and Cardiovascular Surgery, 119:21-26.

36

Dempsey, D. A. and N. L. Benowitz (2001). Risks and benefits of nicotineto aid smoking cessation in pregnancy. Drug Safety: AnInternational Journal of Medical Toxicology and Drug Experience24(4): 277-322

Doyal L (2002). Sex, gender and the 10/90 gap in health research: a briefingdocument and resource guide. Geneva, Global Forum for HealthResearch (http://www.globalforumhealth.org/pages/index.asp,accessed 12 December 2003).

Dresler C et al. (2000). Gender differences in genetic susceptibility for lungcancer. Lung Cancer, 30:153-160.

Ezzati M, Lopez AD (2003). Measuring the accumulated hazards of smoking:global and regional estimates for 2000. Tobacco Control, 12:79-85.

Fry, W, Menck HR, Winchester DP (1996). The National Data Base Reporton Lung Cancer. Cancer, 77:1947-1955.

Gao YT et al. (1987). Lung cancer among Chinese women. InternationalJournal of Cancer, 40:604-609.

GLOBOCAN 2000. Cancer incidence, mortality and prevalence worldwide.IARC Cancer Base No. 5. Version 1.o.1 [online database]. Lyon,International Agency for Research on Cancer, 2001.

Graham H, Blackburn C (1998). The socio-economic patterning of healthand smoking behaviour among mothers with young children onincome support. Social Science and Medicine, 20:215-240.

Graham H, Der G (1999). Patterns and predictors of tobacco consumptionamongst women. Health Education Research, 14:611-618.

Grannis FW (2002). Lung cancer screening: conundrum or contumacy?Chest, 122:1-2.

Granville CA et al. (2003). Mutation spectra of smoky coal combustionemissions in Salmonella reflect the TP53 and KRAS mutations in lungtumors from smoky coal-exposed individuals. Mutation Research,525:77-83.

Greaves L, Barr V (2000). Filtered policy: women and tobacco in Canada.Vancouver, BC, British Columbia Centre of Excellence for Women'sHealth (http://www.bccewh.bc.ca/PDFs/filteredpolicy.pdf, accessed20 December 2003).

Gritz ER, Nielsen IR, Brooks LA (1996). Smoking cessation and gender:the influence of physiological, psychological, and behavioural factors.

37

Journal of American Medical Women's Association, 51:35-42.

Hart CL et al. (2001). Social class differences in lung cancer mortality: riskfactor explanations using two Scottish cohort studies. InternationalJournal of Epidemiology, 30:268-274.

Haugen A (2002). Women who smoke: are women more susceptible totobacco-induced lung cancer? Carcinogenesis, 23:227-229.

Henschke CI et al. (2001). Early lung cancer action project: a summaryof the findings on baseline screening. The Oncologist, 6:147-152.

Hu J et al. (2002). Risk factors for lung cancer among Canadian women whohave never smoked. Cancer Detection and Prevention, 26:129-138.

IARC (2004) Tobacco smoke and involuntary smoking: summary of datareported and evaluation. Lyon, International Agency for Researchon Cancer. Monograph Vol. 83 (http://monographs.iarc.fr/htdocs/indexes/vol83index.html, accessed 20 December 2003).

IARC (2002b) Handbooks of cancer prevention. Vol. 6. Weight controland physical activity. Lyon, International Agency for Researchon Cancer.

IARC (2003) Handbooks of cancer prevention. Vol. 8. Fruit and vegetables.Lyon, International Agency for Research on Cancer.

Jemal A et al. (2003). Lung cancer rates convergence in young menand women in the United States: analysis by birth cohortand histologic type. International Journal of Cancer, 105:101-107.

Jha P et al. (2002). Estimates of global and regional smoking prevalencein 1995 by age and sex. American Journal of Public Health,92:1002-1006.

Kanashiki M et al. (2003). Time from finding abnormality on mass-screeningto final diagnosis of lung cancer. Oncology Report, 10:649-652.

Keohavong P et al (2003). K-ras mutations in lung carcinomas fromnon-smoking women exposed to unvented coal smoke in China.Lung Cancer, 41:21-27.

Kleinerman RA et al. (2002). Lung cancer and indoor exposure to coal andbiomass in rural China. Journal of Occupational and EnvironmentalHealth, 44:338-344.

Koyi H, Hillerdal G, Branden E (2002). A prospective study of a total materialof lung cancer from a county in Sweden 1997-1999: gender,

38

symptoms, type, stage, and smoking habits. Lung Cancer, 36:9-14.

Kreuzer M et al. (2002). Risk factors for lung cancer among non-smokingwomen. International Journal of Cancer, 100:706-713.

Kreuzer M et al. (2003). Hormonal factors and risk of lung cancer amongwomen? International Journal of Epidemiology, 32:263-271.

Lam S et al. (1999). Sex-related differences in bronchial epithelial changesassociated with tobacco smoking. Journal of National CancerInstitute, 91:691-696.

Levi F et al. (1997). Lung carcinoma trends by histologic type in Vaudand Neuchatel, Switzerland, 1974-1994. Cancer, 79: 906-914.

Liao M et al. (1996). A study of the association between squamouscell carcinoma and adenocarcinoma in the lung, and history ofmenstruation in Shanghai women, China. Lung Cancer,14 (Suppl. 1):S215-221.

Ludman EJ et al. (2000). Stress, depressive symptoms, and smokingcessation among pregnant women. Health Psychology (OfficialJournal of the Division of Health Psychology, American PsychologicalAssociation), 19:21-27.

McKee S et al. (2003). Sex differences in the effects of stressful life eventson changes in smoking status. Addiction, 98:847-855.

Metayer C et al. (2002). Cooking oil fumes and risk of lung cancer in womenin rural Gansu, China. Lung Cancer, 35:111-117.

Mosoba Mosoba T (2003). Gender and health implications in tobaccofarming: the case of Kuria District. In: Proceedings of the AmanitareAfrican Reproductive and Sexual Rights Conference, Johannesburg,South Africa, 4-7 February 2003 (http://www.amanitare.org/linked%20documents/conference%20papers/mosoba%20final%20Ama%20Paper.rtf, accessed 20 December 2003).

Nurminen M, Karjalainen A (2001). Epidemiologic estimate of the proportionof fatalities related to occupational factors in Finland. ScandinavianJournal of Work, Environment and Health, 27:161-213.

Osler M et al. (1999). Gender and determinants of smoking cessation:a longitudinal study. Preventive Medicine, 29: 57-62.

Pampel FC (2003). Declining sex differences in mortality from lung cancerin high-income nations. Demography, 40:45-65.

39

Perkins K (2001). Smoking cessation in women: special considerations.CNS Drugs, 15:391-411.

Perneger TV (2001). Sex, smoking, and cancer: a reappraisal. Journal ofthe National Cancer Institute, 93:1601-1602.

Perneger TV (2002). Re: sex, smoking, and cancer: a reappraisal. Discussion.Journal of the National Cancer Institute, 94:309-310.

Pomerleau CS et al. (2000). Short-term weight gain in abstaining womensmokers. Journal of Substance Abuse Treatment, 18:339-342.

Prescott E et al. (1998). Mortality in women and men in relation to smoking.International Journal of Epidemiology, 27:27-32.

Rachtan J (2002). A case-control study of lung cancer in Polish women.Neoplasma, 49:75-80.

Radzikowska E, Gaz P, Roszkowski K (2002). Lung cancer in women: age,smoking, histology, performance status, stage, initial treatmentand survival. Population-based study of 20 561 cases. Annals ofOncology (Official Journal of the European Society For MedicalOncology), 13:1087-1093.

Reich JM (2002). Improved survival and higher mortality: the conundrumof lung cancer screening. Chest, 122:329-337.

Richmond R. (2003). You've come a long way baby: women and thetobacco epidemic. Addiction, 98:553-557.

Risch HA, Miller AB (2002). Re: sex, smoking, and cancer: a reappraisal.Journal of the National Cancer Institute, 94:308-309.

Samet JM, Yoon S (2001). Women and the tobacco epidemic: challengesfor the 21st century. Geneva, World Health Organization(WHO/NMH/TFI/01.1; http://www.who.int/tobacco/media/en/WomenMonograph.pdf, accessed 8 September 2003).

Samet JM et al. (1998). Research for effective global tobacco controlin the 21st century: report of a working group convened during the10th World Conference on Tobacco or Health. Tobacco Control,7:72-77.

SEER (2003). SEER incidence and mortality age-adjusted rates for lungcancer. [online database]. Bethesda, MD, National Cancer Institute(http://canques.seer.cancer.gov/cgi-bin/, accessed 8 September2003).

40

Seow A et al. (2002). Diet, reproductive factors and lung cancer risk amongChinese women in Singapore: evidence for a protective effect of soyin non-smokers. International Journal of Cancer, 97:365-371.

Shields PG (2002). Molecular epidemiology of smoking and lung cancer.Oncogene, 21:6870-6876.

Siegfried JM (2001) Women and lung cancer: does oestrogen play a role?Lancet Oncolology, 2:506-513.

Stabile LP, Siegfried JM (2003). Sex and gender differences in lung cancer.Journal of Gender Specific Medicine, 6:37-48.

Stabile LP et al. (2002). Human non-small cell lung tumors and cells derivedfrom normal lung express both estrogen receptor alpha and betaand show biological responses to estrogen. Cancer Research,62:2141-2150.

Steenland K, Henley J, Thun M (2002). All-cause and cause-specific deathrates by educational status for two million people in two AmericanCancer Society cohorts, 1959-1996. American Journal ofEpidemiology, 156:11-21.

Swerdlow A et al. (1998). Trends in cancer incidence and mortality inScotland: description and possible explanations. British Journalof Cancer, 77(Suppl. 3):1-16.

Sy SM et al. (2003). Genetic alterations of lung adenocarcinoma in relationto smoking and ethnicity. Lung Cancer, 41:91-99.

Taioli E, Wynder E (1994). Re: endocrine factors and adenocarcinomaof the lung in women. Journal of the National Cancer Institute,86:869-870.

Tewari KS, Disaia PJ (2002). Lung and bronchus cancer in women: a 21st-century epidemic. Clinical Obstetrics and Gynecology, 45:784-811.

Tomlinson, R (1997). China's smoking epidemic grows. British MedicalJournal, 315:501-504.

Tyczynski JE, Bray F, Parkin DM (2003). Lung cancer in Europe in 2000:epidemiology, prevention, and early detection. Lancet Oncology,4:45-55.

USDHHS (2001). Women and smoking: a report of the Surgeon General– 2001. Rockville, MD, United States Department of Health andHuman Services.

41

Wen Cheng Y, Lee H (2003). Environmental exposure and lung canceramong non smokers: an example of Taiwanese female lung cancer.Journal of Environmental Science and Health, 21: 1-28.

Westmaas JL, Wild TC, Ferrence R (2002). Effects of gender in social controlof smoking cessation. Health Psychology (Official Journal of theDivision of Health Psychology, American Psychological Association),21:368-376.

WHO (1997). The smoking epidemic: a fire in the global village [pressrelease]. Geneva, World Health Organization (WHO/6125 August 1997).

WHO (1998). Gender and health. Geneva, World Health Organization(document number).

WHO (2001). WHO urges countries to prevent a tobacco epidemic amongwomen and girls [press release]. Geneva, World Health Organization(WHO/27 30 May 2001).

WHO (2003). Resolution WHA56.1.WHO Framework Convention on TobaccoControl. In: Fifty-sixth World Health Assembly, Geneva, 19-28 May2003. Resolutions and decisions, Annexes. Geneva, World HealthOrganization: 1-25 (WHA56/2003/REC/1).

Wisnivesky JP et al. (2003). The cost-effectiveness of low-dose CTscreening for lung cancer: preliminary results of baseline screening.Chest, 124:614-621.

Woodward M (2001). Is compulsory restriction of tar yield of cigarettes aworthwhile public health policy? American Journal of PreventiveMedicine, 21:284-290.

Zang EA (2002). Re: sex, smoking, and cancer: a reappraisal. Journal ofthe National Cancer Institute, 94:308-309.

Zang E, Wynder EL (1996). Differences in lung cancer risk between men andwomen: examination of the evidence. Journal of the National CancerInstitute, 88:183-192.

42

43

7. Additional resources

The following is a list of web sites that provide information about genderissues in lung cancer and related topics that may be of interest to readers.

http://www.tobaccofreekids.org/reports/women/

http://www.who.int/gender/documents/en/Gender_Tobacco_2.pdf

http://www.cdc.gov/tobacco/global/GYTS/globaluse01.htm

http://tc.bmjjournals.com/

http://www.inwat.org/ppp/gendersensitivepolicy.pdf