Lung cancer correlates in Lebanese adults: A pilot case–control study
- DOI
- 10.1016/j.jegh.2013.06.005How to use a DOI?
- Keywords
- Lung cancer; Cigarette; Waterpipe; Indoor pollution; Outdoor pollution
- Abstract
Background: Lung cancer is one of the most prevalent types of cancers. However, there are no epidemiological studies concerning lung cancer and its risk factors in Lebanon. This study was carried out to determine the association between lung cancer and its most common risk factors in a sample of the Lebanese population.
Methods: A hospital-based case–control study was conducted. Patients were recruited in a tertiary health care center. A questionnaire in Arabic was designed to assess the possible risk factors for lung cancer.
Results: For females, cigarette smoking (ORa = 9.76) and using fuel for heating (ORa = 9.12) were found to be the main risk factors for lung cancer; for males, cigarette smoking (ORa = 156.98), living near an electricity generator (ORa = 13.26), consuming low quantities of fruits and vegetables (ORa = 10.54) and a family history of cancer (ORa = 8.75) were associated with lung cancer. Waterpipe smoking was significantly correlated with lung cancer in the bivariate analysis.
Conclusion: In this pilot study, it was found that in addition to smoking, outdoor and indoor pollution factors were potential risk factors of lung cancer. Additional studies would be necessary to confirm these findings.
- Copyright
- © 2013 Ministry of Health, Saudi Arabia. Published by Elsevier Ltd.
- Open Access
- This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Lung cancer is one of the most prevalent types of cancer, and it is one of the most commonly diagnosed cancers as well as the leading cause of cancer death, particularly in males [1,2]. Cigarette smoking is the main risk factor for the onset of lung cancer [3,4]. Although waterpipe tobacco contains the same chemicals found in cigarettes, few studies have assessed the association between waterpipe smoking and lung cancer: a review of the available evidence showed that waterpipe smoking may probably be associated with lung cancer [5], and a recent Indian study found a 6-times elevated risk for lung cancer in waterpipe smokers, in a region where waterpipe smoking is popular [6].
In addition, air pollution contributes to the occurrence of lung cancer [7–9]. Outdoor pollution such as urban dwelling [10,11], living near factories [12], and heavy traffic [13,9,14], and indoor pollutants such as environmental tobacco smoke [15], and the use of coal and fuel in cooking or heating increased the risk of lung cancer [16,17]. Other factors, such as a low consumption of fruits and vegetables [18–21], and occupational factors such as asbestos [22–24], metals and silica [25], diesel exhaust [26], pesticides [27,28] and organic dust [29] exposures also increase the risk of lung cancer. Furthermore, individuals with a positive family history of lung cancer [30,31] and a previous history of chronic obstructive pulmonary disease, emphysema, chronic bronchitis, pneumonia and tuberculosis have a higher risk for lung cancer [32].
In Lebanon, there are no epidemiological studies concerning lung cancer risk factors, although lung cancer is thought to be of primary importance in the region [33] and has a relatively high incidence in Lebanon, particularly in males [34]. This study was carried out to determine the association between the most common risk factors and lung cancer in a sample of the Lebanese population; more specifically, the present study aimed to investigate the association of lung cancer with waterpipe smoking and electricity generator exposure (a potential source of diesel exhaust to the community), two common exposures in the Lebanese population.
2. Materials and methods
2.1. Study design and population
The present study is a hospital-based, case–control study conducted between March and June 2012. Patients were recruited in a tertiary health care center on an ongoing prospective basis, when visiting for a medical consultation, or for receiving a treatment for lung cancer. Inclusion criteria were a medical diagnosis of any type of lung cancer; however, lung cancer patients who simultaneously had another type of cancer were excluded. All cases satisfying inclusion criteria were referred by physicians to the researchers; potential participants were approached by an interviewer to fill out the standardized questionnaire. The first 50 patients who agreed to participate in the study were enrolled.
Controls were selected from patients visiting the same hospital for other reasons and other visitors who had no medical problems. All patients admitted for urinary, kidney, orthopedic, endocrinology and gynecologic problems during the study period, in addition to their accompanying persons with no medical problems, were considered eligible; they were referred to the researchers by the physicians. Controls that had any type of cancer were excluded from the control group. For each case, two controls of the same sex were selected; the first 100 controls who agreed to participate in the study were enrolled.
2.2. Data collection
After informed consent, a questionnaire in Arabic was administered by a face-to-face interview for cases and controls to assess the possible risk factors for lung cancer. This questionnaire had been pre-tested with 15 persons to make sure that the questions were well understood and could be adequately answered. Data was collected between March and June 2012. Information gathered comprised: socio-demographic characteristics (age, sex, education level, marital status, income, height and weight); current and past smoking habits (cigarette and waterpipe smoking, age of starting, number of cigarettes or waterpipes, age of quitting); current and past occupations; exposure to gases or toxic substances; living area and pollution (urban or rural, near the road, near an electricity generator); indoor pollution and passive smoking (heating, cooking, smokers at the house, smokers at work); fruits and vegetables consumption; nutritional supplements intake; past pulmonary diseases or affections (chronic respiratory symptom or disease, and childhood respiratory infection or problem); and cancer and lung cancer history in the family. For lung cancer patients, additional information was collected concerning their affection (age at diagnosis, cancer type, and treatment).
2.3. Sample size
This pilot study was conducted to ascertain the relationship between waterpipe smoking and electricity generator exposure (a potential source of diesel exhaust to the community) with lung cancer; however, there was no reference value to base the sample size calculations on; therefore it was decided to take the first 50 cancer patients as cases and the first 100 controls who would agree to participate in the study.
2.4. Statistical analysis
Data was entered and analyzed using SPSS, version 18.0. Descriptive analysis was performed to observe the distribution of the participants. Bivariate analysis was computed to observe the distribution of various socio-demographic and other independent variables among the main outcome variable which is “having lung cancer”. Chi-square test and Fisher’s exact test (when expected values were lower than 5) were used for dichotomous variables. The odds ratio was calculated for each of the variables with its 95% confidence interval. Moreover, stratification analysis by gender was also performed to detect any interaction or confounding factors.
Significant associations were subsequently entered to perform multivariate analysis for each gender separately. A stepwise forward logistic regression analysis was chosen to determine the predictors of having lung cancer, therefore the exponential β coefficient was used in order to estimate the adjusted odds ratio for each independent factor, with its 95% confidence interval; “having lung cancer” was the dependent variable. Hosmer–Lemeshow test was carried out to measure goodness of fit; a Hosmer–Lemeshow test >0.05 indicated that the model is adequate. For all analyses, p-values <0.05 were considered to be statistically significant.
3. Results
3.1. Description of the sample
There was a very low refusal rate (<5%). The final sample consisted of 150 participants of whom 50 were lung cancer patients [adenocarcinoma (32%), non-small cell carcinoma (10%), small cell lung carcinoma (28%), squamous-cell cancers (16%) and non-specified (14%)]; no significant difference was found between males and females regarding cancer histological types (p = 0.893).
Controls were 100 individuals [59% of the controls were patients visiting for other medical problems (urinary, kidney, orthopedic, endocrinologic and gynecologic problems) and 41% were visitors that had no medical problems; they were accompanying other patients]. Moreover, 62% of the participants were males and 38% were females. The average age for cases and controls was 59.58 years (±6.03) and 59.82 years (±6.31), respectively. Socio-demographic characteristics for cases and controls are shown in Table 1. No significant difference was found between cases and controls for any of the socio-demographic variables: age (p-value = 0.663), gender (p-value = 1.000), region (p-value = 0.769), marital status (p-value = 0.430), educational level (p-value = 0.511), monthly income (p-value = 0.129), and body mass index (p-value = 0.098) (Table 1).
| Variable | Total | Cases | Controls | p-Value |
|---|---|---|---|---|
| Age | ||||
| <50 years | 1 (0.7%) | 1 (2%) | ||
| 50–60 years | 85 (56.7%) | 27 (54%) | 58 (58%) | |
| 60–70 years | 52 (34.7%) | 18 (36%) | 34 (34%) | |
| >70 years | 12 (8%) | 4 (8%) | 8 (8%) | 0.663b |
| Gender | ||||
| Males | 93 (62%) | 31 (62%) | 62 (62%) | |
| Females | 57 (38%) | 19 (38%) | 38 (38%) | 1.000a |
| Region | ||||
| Beirut | 61 (40.7%) | 19 (38%) | 42 (42%) | |
| Mount Lebanon | 89 (59.3%) | 31 (62%) | 58 (58%) | 0.769a |
| Marital status | ||||
| Single | 17 (11.3%) | 8 (16%) | 9 (9%) | |
| Married | 119 (79.3%) | 38 (76%) | 81 (81%) | |
| Widowed/divorced | 14 (9.3%) | 4 (8%) | 10 (10%) | 0.430a |
| Educational level | ||||
| No formal education | 3 (2%) | 2 (4%) | 1 (1%) | |
| Primary | 70 (46.7%) | 25 (50%) | 45 (45%) | |
| Secondary | 54 (36%) | 17 (34%) | 37 (37%) | |
| University | 23 (15.3%) | 6 (12%) | 17 (17%) | 0.511a |
| Monthly income | ||||
| <1,000,000 LL | 26 (17.3%) | 10 (20%) | 16 (16%) | |
| 1,000,000–2,000,000 LL | 59 (39.3%) | 25 (50%) | 34 (34%) | |
| 2,000,000–4,000,000 LL | 58 (38.7%) | 13 (26%) | 45 (45%) | |
| >4,000,000 LL | 7 (4.7%) | 2 (4%) | 5 (5%) | 0.129a |
| BMI | ||||
| Normal | 100 (66.7%) | 32 (64%) | 68 (68%) | |
| Overweight | 36 (24%) | 16 (32%) | 20 (20%) | |
| Obese | 14 (9.3%) | 2 (4%) | 12 (12%) | 0.098a |
BMI, body mass index.
Chi-square test.
Fisher’s exact test.
Socio-demographic characteristics of the study participants.
3.2. Lung cancer and smoking
Table 2 shows the bivariate analysis results. It can be seen that former cigarette smokers had a 17-times higher risk of lung cancer compared with non-smokers (OR = 16.55/95% CI: 5.53–49.55). Also, former waterpipe smokers had a significant 6-times higher risk compared with non-smokers (OR = 6.0/95% CI: 1.78–20.26).
| Variable | Cases | Controls | OR (95% CI)a | p-Value |
|---|---|---|---|---|
| Former waterpipe smoker | ||||
| No | 40 (80%) | 96 (96%) | 1.0 (Ref.) | |
| Yes | 10 (20%) | 4 (4%) | 6.0 (1.78–20.26) | 0.004b |
| Ever lived near a generator | ||||
| No | 29 (58%) | 78 (78%) | 1.0 (Ref.) | |
| Yes | 21 (42%) | 22 (22%) | 2.57 (1.23–5.35) | 0.018b |
| Former cigarette smoker | ||||
| No | 4 (8%) | 59 (59%) | 1.0 (Ref.) | |
| Yes | 46 (92%) | 41 (41%) | 16.55 (5.53–49.55) | <0.001b |
| Daily cigarette consumption | ||||
| <25 cig. | 4 (8.7%) | 22 (53.7%) | 1.0 (Ref.) | |
| 25–35 cig. | 16 (34.8%) | 12 (29.3%) | 7.33 (1.99–26.95) | |
| >35 cig. | 26 (56.5%) | 7 (17.1%) | 20.42 (5.28–79.03) | <0.001b |
| Residence | ||||
| Rural | 3 (6%) | 20 (20%) | 1.0 (Ref.) | |
| Urban | 47 (94%) | 80 (80%) | 3.92 (1.10–13.89) | 0.045b |
| Ever lived near road traffic | ||||
| No | 12 (24%) | 45 (45%) | 1.0 (Ref.) | |
| Yes | 38 (76%) | 55 (55%) | 2.59 (1.21–5.54) | 0.020b |
| Use of fuel for heating | ||||
| No | 22 (44%) | 69 (69%) | 1.0 (Ref.) | |
| Yes | 28 (56%) | 31 (31%) | 2.83 (1.41–5.71) | 0.005b |
| Exposure to passive smoking at home | ||||
| No | 19 (38%) | 58 (58%) | 1.0 (Ref.) | |
| Yes | 31 (62%) | 42 (42%) | 2.25 (1.12–4.52) | 0.033b |
| Exposure to passive smoking at work | ||||
| No | 23 (46%) | 66 (66%) | 1.0 (Ref.) | |
| Yes | 27 (54%) | 34 (34%) | 2.28 (1.14–4.56) | 0.030b |
| Fruit and vegetable consumption | ||||
| High | 13 (26%) | 73 (73%) | 1.0 (Ref.) | |
| Low | 37 (74%) | 27 (27%) | 7.70 (3.56–16.63) | <0.001b |
| Past multivitamin intake | ||||
| No | 43 (87.8%) | 94 (95.9%) | 1.0 (Ref.) | |
| Yes | 6 (12.2 %) | 4 (4.1%) | 3.28 (0.88–12.22) | 0.132b |
| Past occupational exposure to gases and toxins | ||||
| No | 28 (56%) | 77 (77%) | 1.0 (Ref.) | |
| Yes | 22 (44%) | 23 (23%) | 2.63 (1.27–5.44) | 0.014b |
| Chronic respiratory disease | ||||
| No | 38 (76%) | 91 (91%) | 1.0 (Ref.) | |
| Yes | 12 (24%) | 9 (9%) | 3.19 (1.24–8.20) | 0.025b |
| Chronic respiratory symptoms | ||||
| No | 35 (70%) | 90 (90%) | 1.0 (Ref.) | |
| Yes | 15 (30%) | 10 (10%) | 3.86 (1.58–9.40) | 0.004b |
| Cardiovascular problems | ||||
| No | 37 (74%) | 78 (78%) | 1.0 (Ref.) | |
| Yes | 13 (26%) | 22 (225) | 1.25 (0.57–2.74) | 0.733b |
| Respiratory problem during childhood | ||||
| No | 38 (82.6%) | 88 (94.6%) | 1.0 (Ref.) | |
| Yes | 8 (17.4%) | 5 (5.4%) | 3.71 (1.14–12.06) | 0.048b |
| Respiratory infection during childhood | ||||
| No | 35 (72.9%) | 90 (93.7%) | 1.0 (Ref.) | |
| Yes | 13 (27.1%) | 6 (6.3%) | 5.57 (1.96–15.81) | 0.001b |
| Family history of respiratory diseases | ||||
| No | 17 (34%) | 15 (15%) | 1.0 (Ref.) | |
| Yes | 33 (66%) | 85 (85%) | 2.92 (1.31–6.51) | 0.014b |
| Family history of cancer | ||||
| No | 24 (48%) | 21 (21%) | 1.0 (Ref.) | |
| Yes | 26 (52%) | 79 (79%) | 3.47 (1.67–7.24) | 0.001b |
| Family history of lung cancer | ||||
| No | 11 (22%) | 6 (6.3%) | 1.0 (Ref.) | |
| Yes | 39 (78%) | 90 (93.7%) | 4.42 (1.53–12.79) | 0.008b |
CI, confidence interval; OR, odds ratio.
Crude odds ratio.
Chi-square test.
Comparison of potential and known risk factors between cases and controls.
In smokers, those who smoked 25–35 cigarettes daily had a 7-times higher risk than those who smoked less than 25 cigarettes daily (OR = 7.33/95% CI: 1.99–26.95), while smokers who smoked more than 35 cigarettes daily had a much higher risk with an OR of 20.42 (95% CI: 5.28–79.03) compared with the reference group. No significant differences were found between different histological types and smoking (p = 0.632).
3.3. Lung cancer and pollutants
Bivariate analysis also shows an association between outdoor and indoor pollution and lung cancer. People living in urban areas had a higher risk of lung cancer compared with those living in rural areas (OR = 3.92/95% CI: 1.10–13.89). People who have lived near road traffic and near a generator had a higher risk than those who never did, with an OR of 2.59 (95% CI: 1.21–5.54) and 2.57 (95% CI: 1.23–5.35), respectively. People who used fuel for heating also had a higher risk of lung cancer with an OR of 2.83 (95% CI: 1.41–5.71) compared with those who did not.
Exposure to passive smoking at home and at work is also associated with a higher risk of lung cancer: people exposed to passive smoking at home had an OR of 2.25 (95% CI: 1.12–4.52), and those exposed at work had an OR of 2.28 (95% CI: 1.14–4.56) compared with non-exposed. Analysis concerning the use of fuel in cooking was not computed because all of the 150 participants used butane gas for cooking. Moreover, no significant differences were found between different histological types and other factors, particularly for those living close to electricity generators (p = 0.410).
3.4. Fruits and vegetables intake
A low fruit and vegetable consumption is associated with an elevated risk of lung cancer with an OR of 7.70 (95% CI: 3.56–16.63/p-value <0.001) for people who had a low consumption of fruits and vegetables, in comparison with people who had a high consumption. But concerning the past intake of multivitamins, no significant difference was found between people who took multivitamins and those who did not (p-value = 0.132 > 0.05).
3.5. Occupational exposure
Concerning past occupational exposure to gases and toxins and the risk for lung cancer, people who had a past occupational exposure had a 2.6 times excess risk for developing lung cancer, compared with those who had never been exposed (OR = 2.63/95% CI: 1.27–5.44/p-value = 0.014).
3.6. Family and personal history
Having a chronic respiratory disease (OR = 3.19/95% CI: 1.24–8.20), chronic respiratory symptoms (OR = 3.86/95% CI: 1.58–9.40), a respiratory problem during childhood (OR = 3.71/95% CI: 1.14–12.06) or a respiratory infection during childhood (OR = 5.57/95% CI: 1.96–15.81) were all significantly associated with an elevated risk of lung cancer (p-values < 0.05). On the other hand, having cardio-vascular problems was not associated with a higher risk of lung cancer (p-value = 0.733). In addition, having a positive family history of respiratory diseases (OR = 2.92/95% CI: 1.31–6.51), a positive family history of cancer (OR = 3.47/95% CI: 1.67–7.24), or a positive family history of lung cancer (OR = 4.42/95% CI: 1.53–12.79) were also significantly associated with a higher risk of lung cancer (p-values < 0.05).
3.7. Stratification over gender
Stratification analysis by gender detected multiple qualitative interactions. In fact, the use of fuel for heating was significantly associated with a higher risk of lung cancer in females (p-value = 0.002) but not in males (p-value = 0.396). On the other hand, waterpipe smoking was associated with a higher risk of lung cancer in males (p-value = 0.009) but not in females (p-value = 0.321). Moreover, a positive significant association was found in males who lived near a generator (p-value = 0.013), who were exposed to passive smoking at work (p-value = 0.007) and who had an occupational exposure to gases and toxins (p-value = 0.028), but no association was found in females (p-values >0.05).
3.8. Multivariate analysis
Multivariate analysis results for females and males are shown in Table 3. For females, 2 variables were retained in the model: use of fuel for heating (ORa = 9.76/95% CI: 2.15–44.30/p-value = 0.003) and cigarette smoking (ORa = 9.12/95% CI: 1.81–46.00/p-value = 0.007). The model is adequate with a non-significant Hosmer–Lemeshow test (p-value = 0.457). 40.2% of the variability of the dependent variable is explained by the two variables retained in the model.
| Variables retained in the model for femalesa | ORa | 95% CI | p-Value | Nagelkerke R square |
|---|---|---|---|---|
| Use of fuel for heating | 9.76 | 2.15–44.30 | 0.003 | 0.402 |
| Cigarette smoking | 9.12 | 1.81–46.00 | 0.007 | |
| Variables retained in the model for malesb | ||||
|
|
||||
| Cigarette smoking | 156.98 | 5.90–417.4 | 0.003 | 0.684 |
| Living near a generator | 13.26 | 1.85–95.04 | 0.010 | |
| Low fruit and vegetable consumption | 10.54 | 1.76–63.24 | 0.010 | |
| Family history of cancer | 8.75 | 1.22–63.03 | 0.031 | |
Variables included but not retained in the model: age, waterpipe smoking, residence, occupational exposure, living near a road, living near a generator, fruits and vegetables consumption, exposure to passive smoking at home, exposure to passive smoking at work, having respiratory problems, having a respiratory problem during childhood, family history of respiratory disease, family history of cancer and family history of lung cancer.
Variables included but not retained in the model: age, waterpipe smoking, residence, occupational exposure, living near a road, use of fuel for heating, exposure to passive smoking at home, exposure to passive smoking at work, having respiratory problems, having a respiratory problem during childhood, family history of respiratory disease and family history of lung cancer.
Multivariate analysis: risk factors for lung cancer by stepwise forward logistic regression.
For males, four variables were retained in the model: cigarette smoking (ORa = 156.98/95% CI: 5.90–417.4/p-value = 0.003), living near a generator (ORa = 13.26/95% CI: 1.85–95.04/p-value = 0.010), low fruit and vegetable consumption (ORa = 10.54/95% CI: 1.76–63.24/p-value = 0.010), and a positive family history of cancer (ORa = 8.75/95% CI: 1.22–63.03/p-value = 0.031). The model is adequate with a non-significant Hosmer–Lemshow test (p-value = 0.968). 68.4% of the variability of the dependent variable is explained by the four variables retained in the model.
4. Discussion
This study revealed a number of associations between probable risk factors and lung cancer in the Lebanese population. First, it was noted that the number of males with lung cancer in the sample was more than females. The higher number of males probably reflects the fact that males are at a higher risk of developing lung cancer compared with females [2]. However, the selection method of the cases and controls did not allow for excess risk assessment of lung cancer in males compared with females.
The results of this case–control study revealed that cigarette smoking is the main risk factor for developing lung cancer in the Lebanese population. Former cigarette smokers had an important elevated risk of lung cancer compared with never smokers. This result was expected because cigarette smoking has been established as the main risk factor for developing lung cancer in multiple studies in different countries [3,4]. In addition, there is biological evidence that the smoke itself contains known carcinogenic chemicals that promote the initiation and growth of tumors [35,36].
Female smokers had a 9-times elevated risk compared with female non-smokers, and male smokers a 157-times elevated risk compared with male non-smokers; this number is definitely higher than expected owing to the small number of participants. The relatively low risk related to cigarette smoking in females compared with males could be explained by the fact that the prevalence of heavy smoking in females in the general population is less than the prevalence of heavy smoking in males of the Lebanese population [37].
In addition, a dose–effect relationship was found in this study. People who smoked 25–35 cigarettes daily had more risk of having lung cancer compared with those who smoked less than 25 cigarettes daily, considered as the reference group. And a very high risk was found in those who smoked more than 35 cigarettes compared with the reference group. This is concordant with the results of another study that found that the risk of developing lung cancer is directly related to the number of cigarettes smoked [38].
Although waterpipe smoking has not been retained in the multivariate analysis, it showed an important association with lung cancer in the bivariate analysis. An excess of lung cancer risk was found in waterpipe smokers compared with non-smokers, similar to the one found in a recent Indian study [6]. Waterpipe smoking could have been excluded from the multivariate model because of the small number of waterpipe smokers in the sample or because waterpipe smoking was not popular in the past decades as it is in young adults nowadays; moreover, mixed smoking of waterpipe and cigarettes could also explain these results. Additional studies are necessary to clarify this issue.
In the multivariate analysis for females, one variable was retained in the model, in addition to cigarette smoking, which is the use of fuel for heating. Females who lived in houses where fuel was used for heating had an elevated risk of having lung cancer compared with those who did not. This association between the use of fuel and especially solid fuel for heating and cooking and an elevated risk of lung cancer is found in multiple studies [16,17,39]. But this risk factor was not retained in the multivariate analysis model for males despite its significant association with lung cancer in the bivariate analysis. When the stratification by gender was performed, some variables appeared to be not significantly associated with the risk of lung cancer in females. There were only 57 female participants in the study, of which only 19 were lung cancer patients; the small number of female participants could be a possible explanation for the small number of variables retained in the multivariate analysis model for females.
For males, in addition to cigarette smoking, three variables were also retained in the model as risk factors for lung cancer. Males living in a residence near a generator had a higher risk of having lung cancer than those who never did. The presence of generators near the residence, a potential source of diesel exhausts, is frequent in Lebanon. The risk of lung cancer related to this factor has not been assessed in other studies, but living near a generator could be viewed as an indicator of the air pollution near the residence, and air pollution has been identified as an important risk factor of developing lung cancer in different studies [7–9], and exposure to diesel exhaust has been shown to increase the risk of lung cancer [26]. A low consumption of fruits and vegetables was also associated with an excess risk of lung cancer in males. This result does not show the association between specific types of fruits or vegetables and lung cancer risk, but it is concordant with the results of different studies that showed that a high consumption of fruits and vegetables protects against lung cancer [18–20]. Finally males who had a positive family history of cancer had more risk of developing lung cancer compared with those whose family history of cancer was negative. A Turkish study found that the presence of different types of cancer in the history of the family increased the risk of lung cancer [31]. Another American study found an increased risk of lung cancer in association with a positive family history of different types of cancer in non-smokers [40].
In fact, other factors have also shown an association with lung cancer risk in bivariate analysis, like living in an urban area and near road traffic, the exposure to passive smoking at home and at work, and past occupational exposure to gases and toxic substances. Having a personal respiratory disease history, a family history of respiratory disease, and a family history of lung cancer were all associated with a higher risk of lung cancer in bivariate analyses, but none of these factors was retained in the multivariate analysis model; this could be owing to confounding by other factors or to a low power of the study. The small sample size might also explain high adjusted OR for some factors (due to sparse data), very wide confidence intervals obtained in the multivariate analysis and non-significant differences among cancer subtypes. More research, using a larger number of participants, should be done to verify these associations.
A selection bias is present in this study; this bias is frequently found in the case–control design. Since the participants were recruited at the same hospital and were all from Beirut and Mount Lebanon, this sample may not geographically represent the general Lebanese population. However, there is no reason to believe that controls do not represent the Lebanese population from the exposure point of view for two reasons: there were very few refusals (<5%) and percentages of exposure in controls are similar to those in the general population, according to a previous study conducted in Lebanon [41]. An admission bias is also possible because the patients and controls were recruited at the hospital and might have the same risk factors, but for another health problem; this may underestimate the association of some factors with the disease.
There are also different information biases present in the study. The questionnaire comprised an important number of questions that needed information recall, which may have generated a recall bias. There is a risk of bias on the part of the interviewer; being aware of the different risk factors, the interviewer might be more insistent when interviewing cases than controls. Nevertheless, the interviewer was trained to pass the questionnaires in a standardized manner, which decreases the likelihood of this bias. The number of cigarettes was also calculated based on the information reported by the participants, which also affects the classification of the participants.
Owing to the multiple qualitative interactions detected in the stratification analysis, multivariate analysis was done, for each sex separately, to adjust variables to possible confounding factors. Nevertheless, cases did not differ from controls in their socio-demographic characteristics, which reduce the risk of confounding bias. However, some variables may have been left not studied which may cause some residual confounding.
In addition, the nature of the study does not allow the researchers to prove any cause effect relationship, and the small number of participants does not allow for generalization of these results to the general Lebanese population or to study the subtypes of lung cancer, but it does highlight associations that may probably be risk factors for lung cancer in the Lebanese population and that would invite further research and investigations. Larger scale studies with participants from all the Lebanese regions should be done to confirm these results; using biomarkers of exposure and of effect will also be useful in such studies.
Meanwhile, adapting public health education efforts to control the tobacco epidemic should be culturally adapted to the Lebanese population, using modified versions of the WHO’s FCTC recommendations and introducing elements of smoking behavior and of pollution that are specific to Lebanon: exposure to waterpipe smoking and to electricity generators, which have been shown to be related to several chronic respiratory diseases, such as chronic bronchitis and chronic obstructive pulmonary diseases [42–45]. Including these cultural elements in public health policies seems of utmost importance to affect people’s behavior [46].
5. Conclusion
This study is the first Lebanese study that aimed to assess the risk factors for developing lung cancer in the Lebanese population. Its results showed that cigarette smoking can be seen as the major risk factor for developing the disease. More awareness campaigns should be carried out to remind people of the harm caused by tobacco smoking for human health. Other factors, related to indoor and outdoor pollutants, also showed an association with the risk of lung cancer and can also be viewed as possible risk factors. Additional studies are necessary to confirm these results.
Conflict of interest
None declared.
References
Cite this article
TY - JOUR AU - Joseph Aoun AU - Nadine Saleh AU - Mirna Waked AU - Joseph Salamé AU - Pascale Salameh PY - 2013 DA - 2013/07/23 TI - Lung cancer correlates in Lebanese adults: A pilot case–control study JO - Journal of Epidemiology and Global Health SP - 235 EP - 244 VL - 3 IS - 4 SN - 2210-6014 UR - https://doi.org/10.1016/j.jegh.2013.06.005 DO - 10.1016/j.jegh.2013.06.005 ID - Aoun2013 ER -