SEARCH

SEARCH BY CITATION

Keywords:

  • case-control studies;
  • family;
  • smoking;
  • lung neoplasms;
  • genetics

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Family history data from a case-control study of lung cancer conducted in the United Kingdom between 1999 and 2004 were analysed to estimate familial risks of the disease. Comparison of lung cancer prevalence in first-degree relatives of 1,482 female lung cancer cases and 1,079 female controls was undertaken using logistic regression adjusting for age and tobacco exposure. Overall, lung cancer in a first-degree relative was associated with a significant increase in the risk of lung cancer [odds ratio (OR) 1.49; 95% confidence interval (CI), 1.13–1.96]. For cases with early onset of the disease (< 60 years), the OR of lung cancer was 2.02 (95% CI, 1.22–3.34). Having 2 or more affected relatives was associated with an OR of 2.68 (95% CI, 1.29–5.55), with a significant trend in risk according to the number of relatives affected (p = 0.001). An increased risk of lung cancer associated with family history of the disease was observed when analysis was restricted to lifetime nonsmokers, although this did not reach significance (OR 1.23; 95% CI, 0.65–2.31). Results confirm previous findings and support the role of a familial predisposition to lung cancer. © 2005 Wiley-Liss, Inc.

Lung cancer is the most common cancer in the world and a major public health problem.1 In the United Kingdom it accounts for ∼14% of all cancers and ∼22% of all cancer deaths, representing the commonest cause of cancer death in both men and women.2 Tobacco smoking is undoubtedly the major aetiological risk factor, the risk being ∼10-fold higher in long-term smokers compared to nonsmokers.3, 4 Other recognised risk factors for lung cancer include exposure to radiation, asbestos, heavy metals (arsenic, chromium, nickel), polycyclic aromatic hydrocarbons, and chloromethyl ethers.5

Lung cancer is frequently cited as an example of a malignancy solely attributable to environmental exposure. However, it has long been postulated that individuals may differ in their susceptibility to environmental risk factors and there is increasing evidence for a familial risk. In the 1960s, Tokuhata and Lilienfeld6 were the first to demonstrate familial clustering of lung cancer. Since then several case-control epidemiological studies have shown an increase in lung cancer risk for relatives of lung cancer cases, with odds ratios ranging between 1.8 and 2.8.7, 8, 9, 10, 11 Cohort studies12, 13, 14 have found similar familial relative risks, with results ranging between 1.7 and 2.0. Laboratory support for an inherited predisposition to lung cancer has come from studies that suggest that the effect of exposure to carcinogens such as polycyclic aromatic hydrocarbons and aromatic amines, originating from tobacco smoke and other environmental sources, may be modified by host susceptibility factors, reviewed by Norppa15 and Goode et al.16

Several studies over the last 20 years have investigated gender differences in susceptibility to lung cancer.17, 18, 19, 20, 21, 22, 23, 24, 25 Case-control studies20, 22 have reported elevated odds ratios for all major lung cancer types in women compared to men, at every level of exposure to cigarette smoke. However, cohort studies23, 24, 25 have failed to replicate these observations and at present the role of modification of the tobacco effect by gender remains unclear (reviewed in IARC monographs, vol. 834).

To date, there have been few studies that investigate the association of family history to lung cancer risk in women, most of which have been based on small numbers.14, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 To further investigate the role of familial aggregation of lung cancer in women, we conducted a case-control study (GELCAPS, Genetic Lung Cancer Predisposition Study) in the United Kingdom. To our knowledge, this is the largest case-control study reported to date of familial risks of lung cancer in women, based on an analysis of family history data of 1,482 lung cancer cases and 1,079 controls.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Study population

The GELCAPS Consortium was established in March 1999 to investigate the role of genetic factors in the aetiology of lung cancer. Currently GELCAPS includes over 100 oncology departments throughout the United Kingdom specialising in the management of lung cancer. To date, over 5,000 cases with histologically or cytologically (only if not adenocarcinoma) confirmed primary lung cancer (accrual of ∼10% of all UK lung cancer cases in the last year) and 2,000 controls have been recruited. Spouses of cases with no personal history of smoking-related malignancies are being used as controls. A standardised questionnaire is used to collect basic demographic characteristics (sex, date of birth, ethnic group, country of birth and area of residence) in addition to details on active and past smoking history (including type of tobacco product, amount smoked, age at first cigarette and age at any major change of smoking habits), occupational history and personal past medical history. All questionnaires were self administered and no surrogate responders were used. For cigarette smokers, pack years are calculated as a cumulative dose indicator that is categorised in 3 groups (less than 20 pack years, 20–40 pack years and more than 40 pack years). Lifetime nonsmokers only include participants that reported to have never smoked, while smokers are defined as current smokers, participants having ever smoked cigarettes within 5 years of recruitment, and ex-smokers, participants that quit smoking at least 5 years prior to recruitment. An open question is used to obtain information on family history of cancer involving first-degree relatives. A positive history of lung cancer is only assigned when detailed information is provided that identified the family member affected by lung cancer. Ascertainment and collection of cases is solely systematic with no selection for family history or other known risk factors for lung cancer. Informed consent is obtained from all participants. The study is being undertaken with local ethical board approval in accordance with the tenets of the Helsinki Declaration.

For the purposes of this analysis all women of Caucasian British ancestry and current UK residents were selected from the GELCAPS cohort. Out of a total of 1,717 female lung cases and 1,276 female controls recruited in GELCAPS by May 2004, 1,482 cases and 1,079 controls satisfied the above eligibility criteria.

Statistical methods

All computations were undertaken using the statistical software package STATA-Version 7.0 (Stata Corporation, College Station, TX 77840, USA. http://www.stata.com). Differences between groups were assessed for independence using chi-square tests for categorical variables and t-tests for continuous variables. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using conditional logistic regression, adjusting for age and smoking history (pack years). Adjustment for asbestos exposure did not affect the primary analyses. A p-value of 0.05 was considered statistically significant in all analyses.

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Table I details the demographic characteristics of the female cases and controls included in the analysis. The mean age of participants was marginally higher in cases compared to controls (64.0 vs. 63.1 years, p = 0.04). Approximately one third of cases (503/1482) and controls (375/1079) were under 60 years of age (early onset group). The type of tobacco exposure was equally distributed between cases and controls with most participants using filtered cigarettes (83.4% vs. 85.3%, p = 0.48). Thirty-five (2.4%) of the cases reported a history of asbestos exposure compared to 8 (0.7%) of the controls (p = 0.002; OR of lung cancer by asbestos exposure 3.24; 95% CI, 1.50–7.01).

Table I. Demographic Characteristics of Female Lung Cancer Cases and Controls
 Cases N = 1482Controls N = 1079
  • 1

    p-value < 0.05;

  • 2

    CI: confidence interval;

  • 3

    p-value < 0.0001;

  • 4

    ever smoked cigarettes within 5 years of recruitment;

  • 5

    quit smoking more than 5 years prior to recruitment.

Age  
Mean age1 (95% CI2)64.0 (63.5–64.6)63.1 (62.5–63.7)
Less than 60 years (%)503 (33.9%)375 (34.8%)
60 years or older (%)979 (66.1%)704 (65.2%)
Smoking history  
Never-smokers3 (%)138 (9.3%)459 (42.5%)
All Smokers (%)1344 (90.7%)620 (57.5%)
 Mean years smoked3 (95% CI)40.1 (39.5–40.8)33.6 (32.6–34.7)
 Mean pack years3 (95% CI)40.5 (39.3–41.8)27.4 (25.9–28.9)
Current smokers34 (%)920 (62.1%)324 (30.0%)
 Mean years smoked3 (95% CI)43.9 (43.2–44.5)41.0 (39.9–42.0)
 Mean pack years3 (95% CI)44.4 (43.0–45.9)33.8 (31.9–35.7)
Ex-smokers35 (%)424 (28.6%)296 (27.4%)
 Mean years smoked3 (95% CI)32.0 (30.8–33.2)25.5 (24.0–27.0)
 Mean pack years3 (95% CI)31.9 (29.8–34.1)20.3 (18.1–22.4)
Type of tobacco exposure (%)  
Filtered cigarettes1121 (83.4%)529 (85.3%)
Unfiltered cigarettes57 (4.2%)26 (4.2%)
Both filtered and unfiltered137 (10.2%)50 (8.1%)
Cigarettes and cigar/pipe22 (1.6%)8 (1.3%)
Not specified7 (0.6%)7 (1.1%)
Reported asbestos exposure135 (2.4%)8 (0.7%)

As expected, there was a strong relationship between smoking and lung cancer risk. Cases were more likely to have ever smoked (90.7% vs. 57.7%, p < 0.0001) and had a significantly higher number of pack years (40.5 vs. 27.4, p < 0.0001) when compared to controls. The OR of lung cancer for every pack year exposure was 1.05 (95% CI, 1.04–1.05).

Table II details the distribution of histological subtypes of lung cancer observed in our study. Overall this compares favourably with previously published results from the UK,31 the female cases from GELCAPS exhibiting a relatively lower frequency of squamous cell carcinoma compared to that reported for male cases (27.9% vs. 52.0%, respectively) and a relatively higher frequency of adenocarcinoma (25.0% vs. 13.0%). Stratifying histology by age at diagnosis and smoking status, adenocarcinoma was the most common histological subtype observed in both the younger age group and the nonsmokers (p = 0.001 and p < 0.0001, respectively).

Table II. Distribution of Histological Subtype of Lung Cancer Among Study Cases
 SmokersLifetime non-smokers
All N = 1344<60 years N = 461≥ 60 years N = 883All N = 138<60 years N = 42≥ 60 years N = 96
Histology (%)      
 Small cell368 (27.4%)134 (29.1%)234 (26.5%)24 (17.4%)6 (14.3%)18 (18.8%)
 All nonsmall cell926 (68.9%)314 (68.1%)612 (69.3%)109 (79.0%)33 (78.6%)76 (79.2%)
  Squamous cell393 (29.2%)105 (22.8%)288 (32.6%)20 (14.5%)2 (4.8%)18 (18.8%)
  Adenocarcinoma300 (22.3%)124 (26.9%)176 (19.9%)70 (50.7%)26 (61.9%)44 (45.8%)
 Not specified50 (3.7%)13 (2.8%)37 (4.2%)5 (3.6%)3 (7.1%)2 (2.0%)

Table III shows data on the reported family history of lung cancer in first-degree relatives of the cases and controls. The prevalence of a history of lung cancer in at least 1 first-degree relative was 14.8% (220/1482) for the lung cancer cases and 9.8% (106/1079) for the control group. Altogether 177 fathers, 63 mothers and 127 siblings of the study population were affected by lung cancer. The adjusted OR for any family history of lung cancer was 1.49 (95% CI, 1.13–1.96). History of lung cancer in the father of study participants was reported by 9.5% of cases and 5.9% of controls, resulting in a significant increase of the risk of lung cancer (OR 1.70; 95% CI, 1.19–2.44). History of lung cancer in the mother or a sibling of a participant also resulted in an elevated, though not statistically significant risk of the disease (OR 1.53; 95% CI, 0.84–2.81 and OR 1.48; 95% CI, 0.96–2.29, respectively). Analysis of the data by the sex of the affected relative did not reveal a difference in the male/female distribution between cases and controls, both when all affected relatives were considered together (p = 0.15) or when only siblings were analysed (p = 0.06). The sex distribution of affected relatives did not differ by age (< 60 years and ≥60 years) or by histology. History of lung cancer in a male or female relative of a participant resulted in a significantly elevated risk of lung cancer for both groups (OR 1.44; 95% CI, 1.05–1.98 and OR 1.36; 95% CI, 1.08–1.71, respectively).

Table III. Odds Ratios of Lung Cancer Associated with Family History of Lung Cancer in a First-Degree Relative by Affected Family Member, Age Group, and Histology
 Cases Relative with lung cancerControls Relative with lung cancerCrude OR (95% CI)1Adjusted OR (95% CI)2
YesNoYesNo
  • 1

    OR: odds ratio; CI: confidence interval;

  • 2

    Adjusted for age and pack years tobacco exposure;

  • 3

    p-value < 0.0001;

  • 4

    p-value < 0.005;

  • 5

    p-value < 0.01;

  • 6

    p-value < 0.05.

Any first-degree relative with lung cancer2201,2621069731.603 (1.25–2.05)1.494 (1.13–1.96)
First-degree relative with lung cancer
 Father1201,262579731.624 (1.17–2.25)1.704 (1.19–2.44)
 Mother431,262209731.66 (0.97–2.83)1.53 (0.84–2.81)
 Sibling891,262389731.814 (1.22–2.66)1.48 (0.96–2.29)
Age
 <60 years75428263492.353 (1.47–3.75)2.025 (1.22–3.34)
 60+ years145834806241.366 (1.01–1.82)1.30 (0.93–1.80)
Histology
 Small cell543381069731.476 (1.03–2.08)1.34 (0.89–2.00)
 All non-small cell1538821069731.594 (1.22–2.07)1.475 (1.10–1.97)
  Squamous683451069731.813 (1.30–2.51)1.626 (1.11–2.37)
  Adenocarcinoma443261069731.24 (0.85–1.80)1.15 (0.77–1.70)

A family history of lung cancer was reported by 17.5% of young lung cancer cases (aged < 60 years at diagnosis) compared to 7.4% of controls, providing an OR of 2.02 (95% CI, 1.22–3.34). In contrast, family history of lung cancer was not significantly associated with lung cancer risk in the older age group (OR 1.30; 95% CI, 0.93–1.80).

Analysis of the data by histological subtype of lung cancer revealed a significant association between lung cancer risk and family history in patients with nonsmall cell disease (OR 1.47; 95% CI, 1.10–1.97). This association remained statistically significant when the analysis was restricted to cases with squamous cell carcinoma (OR 1.62; 95% CI, 1.11–2.37) but failed to reach significance in those with adenocarcinoma (OR 1.15; 95% CI, 0.77–1.70). The OR of lung cancer associated with positive family history of the disease was also elevated for cases with a diagnosis of small cell lung cancer, though not statistically significant (OR 1.39; 95% CI, 0.89–2.00).

In total, 138 cases (9.3%) and 459 controls (42.5%) were lifetime nonsmokers. The mean age of these cases and controls did not differ significantly (65.2 and 64.2 years, respectively, p = 0.13). Table IV shows the odds ratios for lung cancer by family history and smoking status, nonsmokers with no family history of lung cancer representing the reference group. A family history of lung cancer in a first-degree relative was reported by 10.9% of lifetime nonsmoker cases compared to 8.7% of controls. The calculated risk of lung cancer associated with family history in lifetime nonsmokers was increased but did not reach statistical significance (OR 1.23; 95% CI, 0.65–2.31). Restriction of the analysis to smokers (1,344 cases and 620 controls; mean age 62.3 and 63.9, respectively, p = 0.001) found a further increased risk of lung cancer associated with positive family history of the disease (OR 10.65; 95% CI, 7.55–15.01) compared to the risk observed for those smokers with no family history (OR 7.15; 95% CI, 5.70–8.96). There was no evidence of interaction between smoking status//pack years and family history of lung cancer.

Table IV. Odds Ratios for Lung Cancer by Smoking Status and Family History of Lung Cancer (Reference Category is Nonsmokers with No Positive History of Cancer)
Smoking statusNo positive family history of lung cancerPositive family history of lung cancer
CaseControlOR (95% CI)1CaseControlOR (95% CI)1
  • 1

    OR: Odds ratio, CI: confidence intervals, adjusted for age;

  • 2

    p-value <0.0001.

Nonsmokers1234191.015401.23 (0.65–2.31)
Ever-smokers1,1395547.15 (5.70–8.96)22056610.65 (7.55–15.01)2

Table V shows the risks of lung cancer by number of relatives affected by lung cancer. While there are few families with multiple occurrences of lung cancer in first-degree relatives, families of cases are disproportionately more often affected. In total, 42 of the cases (2.8%) reported a history of lung cancer in 2 or more of their first-degree relatives compared to 11 of the controls (1.0%). Having multiply affected relatives was associated with an increased OR of lung cancer of 2.68 (95% CI, 1.29–5.55), compared to an OR of 1.35 (95% CI, 1.01–1.81) associated with having a single affected relative. There was a significant linear trend in risk according to the number of relatives affected (p = 0.001).

Table V. Odds Ratios for Lung Cancer According to the Number of First-Degree Relatives (Parents, Siblings) Affected by Lung Cancer, in a Population of UK Females
Number of relatives affectedCasesControlsCrude OR (95% CI)1Adjusted OR (95% CI)2
No.%No.%
  • 1

    OR: Odds ratio, CI: confidence intervals;

  • 2

    adjusted for age and pack years tobacco exposure;

  • 3

    p-value for trend <0.001.

01,26285.297390.2  
117812.0958.81.44 (1.11–1.88)1.35 (1.01–1.81)
2332.2100.9  
390.6002.94 (1.51–5.75)2.683 (1.29–5.55)
40010.1  
Total1,4821001,079100  

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The results of our study show evidence of familial aggregation of lung cancer. We detected an elevated risk of lung cancer for women with an affected first-degree family member, which remains significant after adjustment for age and tobacco exposure. This familial lung cancer risk is further increased in the younger onset age group of cases and with each additional affected family member.

The familial risk of lung cancer documented in this and previously published studies6, 7, 8, 9, 10, 11, 26, 29 could be due to shared environmental as well as genetic factors. As most studies of familial lung cancer have examined cases who smoked cigarettes, familial aggregation of smoking habits could potentially explain the excess number of lung cancers in case families relative to control families. Some investigators have attempted to address this issue by taking into account smoking habits of the family members. Tokuhata and Lilienfeld6 found an excess risk of lung cancer in case relatives compared to control relatives irrespective of the relative's smoking history. To minimise the impact of shared tobacco habits in families, a number of studies have estimated familial risks associated with nonsmoker status. The largest study is that of Wu et al.30, a multicentre study of lung cancer in 646 female lung cancer patients in the United States (lifetime nonsmokers) and 1,252 population controls (also lifetime nonsmokers). Lung cancer occurred nonsignificantly more frequently in first-degree relatives of lung cancer patients than in comparative relatives of population controls (adjusted OR 1.29; 95% CI, 0.9–1.90). Schwartz et al.32 also conducted a population-based study of familial lung cancer risk in 257 nonsmoking lung cancer cases and 277 nonsmoking controls from Detroit. Nonsmoking cases were more likely than nonsmoking controls to have a first-degree relative with lung cancer (OR 1.4; 95% CI, 0.8–2.5). Wang et al.33 reported on the family history of any cancer in 135 lifetime nonsmoking female lung cancer cases and an equal number of matched controls in China. Cases were significantly more likely than controls to report a positive family history of any cancer, including lung cancer (OR 2.29; 95% CI, 1.01–5.17). Brownson et al.34 examined the association between family history of cancer and risk of lung cancer in 618 lung cancer cases (432 lifetime nonsmokers and 186 exsmokers) and 1,402 controls from Missouri. They reported a significantly increased risk for developing lung cancer associated with a history of lung cancer in a first-degree relative (OR 1.3; 95% CI, 1.0–1.8). However, when the analysis was limited to the lifetime nonsmokers the risk decreased to 1.1 (95% CI, 0.8–1.5). Similarly, Mayne et al.35 in 1999, conducted a population-based study of familial lung cancer risk in 437 never and former smokers (197 lifetime nonsmokers) and an equal number of matched controls from New York State. They reported a significantly increased risk of lung cancer associated with a positive history of the disease in a first-degree relative. Data on the lifetime nonsmokers alone was not reported. Finally, Wu et al.36 studied 108 female nonsmoking lung cancer patients from Taiwan and their 108 spouses as controls. They found family history of lung cancer (in a first-degree relative) to be a significant predictor of lung cancer risk (OR 5.7; 95% CI, 1.9–16.9). Examination of a life-long nonsmoking subgroup of female cases and controls in our study revealed an elevated risk of lung cancer associated with having a first-degree relative with the disease, though this did not reach statistical significance (OR 1.23; 95% CI, 0.65–2.31). However, the numbers of nonsmokers involved are small and preclude detection of a small increase in risk associated with family history.

Further investigation of the data in our study revealed a significant elevation in the risk of lung cancer in subjects exposed to tobacco, who reported a family history of the disease. Formal testing did not detect a significant interaction between tobacco exposure and family history. The relationship between smoking and family history of lung cancer has been investigated by some previous studies6, 7, 28, 29 with only Horwitz et al.28 detecting a significant interaction.

Risks of lung cancer associated with family history of the disease were strongly related to early age of diagnosis and number of relatives affected. This is concordant with previous reports7, 32, 37 and is supportive evidence of a genetic susceptibility. Sellers et al.,37, 38 Bailey-Wilson et al.39 and Gauderman et al.40 have carried out segregation analyses of lung cancer pedigrees, allowing for variable age of onset of disease and smoking history. Findings are compatible with a codominant model of inheritance of a rare gene that produces earlier age of onset of cancer. Such fitted models, however, inevitably depend heavily on the specific and speculative assumptions about the age distribution of incidence in susceptible individuals.41 An apparently superior goodness of fit could be a consequence of the assumed age distribution. It is therefore equally plausible that at least part of any familial predisposition to lung cancer is attributable to the action of low penetrance alleles, and a polygenic model cannot be excluded on the basis of existing data.

Results from epidemiological studies on familial aggregation of lung cancer require careful interpretation. The first problem lies in eliminating potential bias due to the association between exposure to environmental factors among relatives. Smoking is the most important environmental risk factor of lung cancer, and the association between a person's smoking habits and that of his parents or siblings has been well described.42 Unless adjustment is made for smoking habits, an above-expected incidence of lung cancer in relatives of lung cancer patients would be observed, even in the absence of any genetic effect. In our analysis, we were not able to adjust for the smoking habits of first-degree relatives as this information was not available. However, it has been proposed by Lee43 that the extent of bias from this source is perhaps not as large as one might anticipate and he suggested that an unbiased estimate of the risk of lung cancer in a relative is obtained if accurate adjustment is made for the smoking habits of the index case.

In our study, we did not have the opportunity to account for other known environmental factors that have been incriminated in lung cancer risk, such as specific types of occupational exposures, environmental tobacco smoke (ETS) and urban/suburban residence. The restriction of our analysis to females limits the impact of occupational factors, such as asbestos exposure, on the familial aggregation observed. However, ETS and area of residence might account for some of the familial risks observed, especially in nonsmokers.

A further limitation of our study is the lack of adjustment for family size and age of relatives. The lack of this information raises the possibility of bias where familial risks in all first-degree relatives or siblings are calculated. However, this is unlikely to be a strong influence on our study as familial aggregation of lung cancer was observed when history of parental lung cancer alone was considered. Furthermore, adjustment of family size in previously published studies6, 11, 18, 26, 44 have not shown this to be a significant contributor to lung cancer risk.

In our study, participants themselves provided information on diagnosis, without verification of the diagnosis through medical sources or health records. Bondy et al.45 examined the validity of patient reports of a family history of cancer by validating them using medical records. In this study, subjects correctly identified the primary site of cancer for 88% of cases in first-degree relatives and for lung cancer in 85%. It is thus unlikely that lung cancer misclassification will represent a significant source of bias.

There is the issue inherent to all case-control studies of recall bias. In theory, diagnosis of lung cancer in an individual might result in an increased knowledge or awareness of lung cancer in relatives. We do not believe that recall bias has severely distorted our findings for 2 reasons. First, lung cancer in a first-degree relative is a disease severe enough to be equally remembered by both cases and controls. Second, the controls used in our study were spouses of lung cancer cases, so one would expect the recent diagnosis of lung cancer in their spouse would have increased their awareness of the disease in their family by a similar degree, reducing this as a source of potential bias. There is 1 study in the literature that has completely excluded the possibility of recall bias. Osann et al.29 conducted a case-control study to investigate familial aggregation of lung cancer, where the epidemiological information was prospectively collected. A relative risk of 1.9 was observed in this study, which is similar to the findings of ours and other previously published studies.7, 8, 9, 10, 11, 12, 13, 14

The use of spouse controls in case-control studies has been criticised46, 47 as not representative of a true sample of the population from which the cases arose. It may, thus, introduce selection bias in the study through assortative mating and overmatching for a series of environmental and socioeconomic factors, resulting in a distorted estimate of familial aggregation of disease. Some epidemiological studies have found increased concordance of smoking habits between spouse pairs.48, 49, 50 The use of spouse controls may thus lead to an underestimation of the familial aggregation detected in our study; however we are unable to exclude the introduction of positive bias through the use of this control group.

In summary, our study showed a significant increase in the risk of lung cancer associated with a family history of the disease in a first-degree relative in a population of British women. This is, to our knowledge, the largest case-control study that investigated the familial risk of lung cancer in women, with all 1,482 lung cancer cases included in the study histologically or cytologically verified by the recruiting centre. Accepting the caveats discussed, our data is consistent with a genetically determined risk, as we have found further elevation of the lung cancer risk with earlier age of onset of the disease and with multiple affected family members. These findings support the hypothesis that genetic susceptibility to lung cancer might act as both an independent risk factor and an effect modifier of environmental risk factors. Recently, Bailey-Wilson et al.51 provided direct evidence for an inherited predisposition to lung cancer. This followed a genome-wide linkage search of 52 pedigrees segregating lung and other tobacco related cancers. Analysis under an autosomal dominant model provided significant evidence for a major susceptibility locus influencing the risk of lung cancer on chromosome 6q23–25.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Athena Matakidou was in receipt of a clinical research fellowship from the Allan J. Lerner Fund. Tim Eisen and Richard Houlston are supported by grants from Cancer Research UK. The authors thank all the study participants and their clinicians.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References