SEARCH

SEARCH BY CITATION

Keywords:

  • bidi;
  • lung cancer;
  • incidence;
  • cohort study;
  • India

Abstract

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

The association of lung cancer incidence with bidi smoking was examined using a cohort study data in Karunagappally, Kerala, India. We sought interview of all the residents in Karunagappally with the population of 385,103 in 1991 census, and established a cohort of 359,619 (93% of the population in 1991) in the 1990s. There were 65,829 men aged 30–84 at interview after excluding those diagnosed as cancer or died of any cause before 1997. Among them, 212 newly diagnosed lung cancer cases were ascertained during the 8-year period between 1997 and 2004 through Karunagappally Cancer Registry. The relative risk (RR) of lung cancer was obtained from Poisson regression analysis of grouped data. Lung cancer incidence was relatively high among Moslem people and those with lower educational history. When taking into account attained age, religion and education, the RR between current bidi smokers and those who had never smoked bidis was 3.9 (95%CI = 2.6–6.0, p < 0.001). The lung cancer risk did not return to the level of non-smokers within 10 years after cessation. In further analyses using only those never smoked cigarettes to examine the effect of bidi smoking alone on lung cancer risk, current smokers of bidis had the RR of 4.6 (95%CI = 2.5–8.5, p < 0.001). Lung cancer incidence increased with larger amounts of bidi smoked a day (p < 0.001), with longer durations of smoking bidis (p < 0.001), and with younger ages starting smoking bidis (p < 0.001). Immediate measures should be taken to stop bidi smoking, which is common in south Asia. © 2008 Wiley-Liss, Inc.

In southern India, including Kerala State and Tamil Nadu, the most popular tobacco smoked is a locally-made bidi,1 which is made of 0.15–0.25 g of sun dried flaked tobacco rolled into a conical shape in a dried rectangular piece of Temburni leaf (Diospyros melanoxylon) and a thread securing the roll.2 Although the health consequences of bidi smoking has not been fully evaluated to date, its association with cancer risk was shown as early as 1955, when an Indian study reported an increased risk of oral cancer among bidi smokers.3 Recently, Gupta et al. reported the results of the cohort study of 99,570 individuals aged >35 years in Mumbai. Their study showed that bidi smoking increased all cause mortality by 64% and was associated with the increased mortality of all neoplasms, and oral and pharyngeal neoplasms.4 This is the first scientific international publication on the health effect of bidi smoking based on mortality.

Several studies have shown that bidi smoking also increases lung cancer risk.5–10 Gajalakshmi et al.10 reported the largest case-control study of lung cancer in India, which examined 778 lung cancer cases and 3,430 controls, including 1,503 cancer controls and 1,927 healthy controls in Trivandrum and Chennai, the capital of Kerala State and Tamil Nadu, respectively. Their study showed that the odds ratio of lung cancer was 6.45 for more than 30 years of bidi-only smokers.

In Karunagappally, Kerala, lung cancer is the leading cancer among men, according to our cancer registry data included in Cancer Incidence in Five Continents vol. VII,11 VIII12 and IX,13 published by International Agency for Research on Cancer (IARC). Krunagappally is a rural area, where only 2 major factories are operative. They are Indian Rare Earth factory, which processes monazite sands and the Kerala Metal and Minerals factory. However, since Karunagappally is known for high background radiation from thorium-containing monazite sand (Fig. 1),14–16 it has been concerned that lung cancer risk and other cancer risks are increased by exposure to high-level natural radiation and/or by synergistic effect between radiation and other factors, including bidi smoking. In the late 1980s, we started the project to establish a cohort of all residents of Karunagappally taluk mainly to study the potential health effects of high background radiation. However, our cohort study showed that the relatively high lung cancer incidence in this area is unlikely to be due to high-level natural radiation.17

thumbnail image

Figure 1. Location map of Karunagappally in Kerala.

Download figure to PowerPoint

In the present study, we analyzed the lung cancer risk among men in order to examine its association with bidi smoking, which is common in the study area. To our knowledge, this is the first cohort study to examine bidi-related lung cancer incidence, rather than mortality, among those who had never smoked cigarettes habitually.

Material and methods

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

Karunagappally taluk consisted of 12 panchayats. Taluk is an administrative unit, corresponding to a county, and panchayat is a sub unit of taluk. According to 1991 census, this taluk had a population of 385,103 (191,149 males and 193,954 females) residing in an area of 192 sq. km. In the late 1980s, we planned to establish a cohort of the entire residents in Karanagappally taluk in order to examine the risk of cancer in relation to natural radiation exposure, lifestyles and other factors.18 All the households (N = 71,674) in Karunagappally taluk were visited by 12 to 14 interviewers, starting from January 1, 1990 and ending on 31st December 1997. Using a 6-page standardized questionnaire, they collected information on socio-demographic factors, lifestyles and other factors of 359,619 residents, which correspond to 93% of Karunagappally population in 1991. Socio-demographic factors included religion, family income in Rupees, education and occupation.

Our questionnaire had separate questions for bidi smoking, cigarette smoking and tobacco chewing. This design enabled us to examine single habits as well as combinations. Regarding bidi and cigarette smoking, we asked whether residents never smoked them, smoked them in the past or smoked them currently. To bidi smokers, further questions were asked to know the amount of bidis smoked a day, age starting smoking, and the duration of smoking. To ex-smokers, age stopped smoking bidis was also asked. The same types of questions were asked to cigarette smokers. Regarding tobacco chewing and pan using, we first asked residents whether they were former or current users of chewing tobacco or pan. When residents turned out to have those habits, we asked them to pick an answer from (i) pan only, (ii) pan + chewing tobacco, (iii) chewing tobacco only and (iv) others. For tobacco chewers, we asked questions on the frequency, starting age and duration of tobacco chewing. To those who answered to stopped tobacco chewing, age stopped chewing tobacco was also asked. To pan users, the same types of questions were asked. Pan is the mixture of Betel leaves, lime and areca nuts, which is obtained from the fruit of Areca catechu tree.2 In addition, indoor and outdoor radiation levels were measured at the time of house visit by using portable scintillometres.

There were 69,944 men who were 30–84 years old at the time of interview. We excluded those younger than 30 years of age since (i) lung cancer risk was low in that age range; and (ii) the majority of bidi smokers started bidi smoking in their 20s. Those aged 85 years or older were also excluded from analysis because the elderly were less likely to seek medical care for malignancy, and therefore the follow-up of this age group may result in lower completeness of cancer case ascertainment and lower accuracy of diagnosis. We also excluded Rare Earth workers, who might have been exposed to various occupational exposures (N = 1,428). In addition, 2,222 subjects, who had died or had been diagnosed as cancer before 1997, were excluded from analysis, leaving 66,294 men. Further, those who died within 3 years of interview were also excluded from analysis since their lifestyles might have been affected by their health conditions. Thus, there were 65,829 subjects available for statistical analysis.

In the present study, we analyzed cancer incidence during the period between January 1, 1997 and December 31, 2004. Cancer cases among the cohort were ascertained by the cancer registry in Karunagappally, which was officially initiated as of January 1, 1990. The registry reports have been presented in “Cancer Incidence in Five Continents” vol. VII,11 vol. VIII12 and vol IX.13 The major activities to identify cancer cases were (i) monthly routine visits to the Regional Cancer Centre (RCC) in Trivandrum, which is the comprehensive cancer centre in the state of Kerala, and more than half of cancer cases registered in Karunagappallly cancer cancer registry were those who sought medical treatment in RCC (unpublished data); (ii) annual visits to Trivandrum Medical College Hospital in Trivandrum; (iii) annual visits to major pathological laboratories in Karunagappally taluk and its neighboring areas, and in Trivandrum; (iv) annual visits to all the hospitals and medical practitioners in Karunagappally taluk; (v) 3–4 time visits to three primary health centers in the taluk, which have cancer screening facilities; (vi) our clinics to provide monthly follow-up care for local cancer patients, which became popular because it provides cancer patients with palliative care and palliative care home service as well and; (vii) our cancer screening camps conducted twice a year on average in all panchayats in the taluk. Our registry workers retrieved medical records and other relevant documents of cancer cases of Karunagappally residents diagnosed in the RCC and other medical facilities, and abstracted information on cancer cases diagnosed.

Death reports were obtained from the death registers kept in the vital statistics division of each panchayat. House visits of the deceased, to supplement information on cause of death, were started in 1997. The proportion of DCO cases in Karunagappaly cancer registry was 14% during 1991–1992,11 and was 10% during 1993–97.12 In 1997, when the active tracing back of all deaths was started, DCO% decreased substantially, and was 4% during 1998–2002.13 Therefore, we decided to analyze the cancer incidence data since 1997. The ratio between lung cancer incidence and lung cancer mortality was 36% during the period between 1997 and 2004 (unpublished data). The ratio between incidence and mortality (M/I%) for all cancer among men was 53% during the period between 1993 and 200119 and was similar to those in other major cancer registries in this country.20

The extent of migration among cohort members was assessed by conducting a door-to-door survey of all the households in the 6 panchayats (Chavara, Neendakara, Panmana, Alappad, Oachira and Thevalakkara) and in the remaining 6 panchayats in 2001 and 2003, respectively. The survey findings were linked to incident cases through name, address, age, house number etc. This survey showed that migration was negligible and people have lived here for centuries. Movement within a panchayat was 9.5% whereas migration to outside the taluk was 6% in the 13-year study period. Only 0.7% was lost to follow-up. The majority of migration took place for job opportunities in Gulf countries.

Statistical analysis was based on the data in cross-classifications by attained age (5-year category), and other covariates. Relative risk (RR) and 95% confidence intervals (95%CI) were obtained from Poisson regression analysis of grouped survival data,21 using the DATAB and AMFIT procedures of Epicure program.22 In the analysis examining the association of cancer risk with variables related to socio-demographic factors, which are summarized in Table II, the estimates of relative risks (RRs) were obtained by the following model:

  • equation image

where A2, A3, A4, A5, A6, A7, A8, A9, A10 and A11 are the indicator variables for attained age; R2 and R3 are indicator variables for religion; I2, I3, I4, and I5 are for the categories of family income; E2, E3, E4 and E5 are for the categories of educational history; O2, O3, and O4 are for the categories of occupation. Attained age at the time of the mid-point of one-year interval during the observational period (1997–2004) was calculated for each cohort members by the DATAB procedure of EPICURE program.21 The maximum likelihood estimates of β12 and β13, for example, are log relative risks (RRs) for the indicator variables R2 and R3, respectively, when compared to the reference category of R1, adjusting for other variables, i.e., family income, education and occupation, in addition to categorized attained age. Heterogeneity test was based on a global p-value for a set of indicator variables. In the analysis examining the association of cancer risk with bidi smoking, which has the three categories (those never smoked, ex-smokers and current smokers), the following model was used to estimate the RRs of ex-smokers (represented by S2) and current-smokers (represented by S3):

  • equation image

where H0 represents the baseline, or background, lung cancer incidence (e.g., incidence rate for never smokers) for cross-classified strata by attained age and socio-demographic variables. This model takes into account the potential confounding by attained age, religion and education, and corresponds to the model adjusting for attained age, religion and education. As far as adjustment for covariables (i.e., attained age, religion and education) is concerned, this model is even better than the model used for RR estimation for socio-demographic factors since this model taking into account the possible interactions among covariables. However, this model does not provide RR estimates for covariables, which were used for stratification. This kind of model is commonly used in the risk analysis of radiation-related risk.23, 24 The maximum likelihood estimates of β2 and β3 are log RRs for the indicator variables S2 and S3, representing ex-smokers and current-smokers, respectively, when compared to those never smoked, adjusting for attained age and socio-demographic factors. In the present study, among socio-demographic factors, only religion and education were considered as potential confounders and incorporated in the model for evaluating smoking (or tobacco chewing)-related lung cancer risk since they were found to be significantly related to lung cancer risk. The same types of statistical models were used for cigarette smoking and tobacco chewing. Trend test was conducted by assigning consecutive integer to exposure categories.

We tried to use 10-year intervals for the duration of smoking. However, we had to use combined categories because of the paucity of lung cancer cases. In the case of age starting smoking, we took into account that many started smoking around 20 years of age. When deciding the cutoff points of bidis and cigarettes smoked a day, the fact that people tend to give vague answers such as around 10 bidis or about 20 bidis was taken into account by using the intervals of 1–4, 5–14, 15–24 and 25 or more cigarettes. The cutoff points we used are those frequently used for the risk analysis of cigarette smoking.1

The entry into the cohort was January 1st, 1997 or the date of interview, which was started on January 1, 1990 and ended on December 31, 1997. A member of the cohort was considered to be censored when he/she was diagnosed as cancers other than lung cancer or died of causes other than lung cancer. Thus, the end of follow-up was the date of diagnosis for cancer cases, the date of death for those deceased, the end of follow-up (December 31, 2004) or the date attaining age of 85. In person-year calculation, we ignored the migration of cohort members since information on migration was available only for a part of our observation period (1997–2004). The use of available information on migration in person-year calculation caused only small changes in relative risk estimates.

Results

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

Among 65,829 men aged 30–84 at interview, 212 newly diagnosed lung cancer cases were ascertained during the 8-year period between January 1, 1997 and December 31, 2004. On average, study subjects were followed for 7.6 years. The frequency of various tobacco habits among those subjects are summarized in Table I. More than 60% of study subjects answered to have smoked bidis in the past or to smoke bidis currently. Among the subjects aged 70 or older at the time of interview, bidi smokers accounted for more than 70%. Those smoking only bidis had lower family income and lower levels of education when compared to those smoking only cigarettes (data not shown).

Table I. Distribution of Smoking Cigarettes and Bidis, and Tobacco Chewing Among Men According to Age, Religion and Education
 Education totalIlliteratePrimary/middleHigh school/CollegeUnknown
AgeAgeAgeAgeAge
30−70+30−70+30−70+30−70+30−70+
  • In some colums, total percentages do not add up to 100% due to rounding errors. The distinction between current and former smokers (users) were ignored in this table.

  • 1

    tc, tobacco chewing.

Hindu
Tobacco usage
TotalN44689319219363472255824161972438247147
 (%)100%100%100%100%100%100%100%100%100%100%
Bidi (bd) only29411930214040612728
Cigarette (cg) only10139121013101289
Tobacco chewing only114628316686
bd+cg5120416230
bd+tc1716181910182556
cg+tc4599652312
bd+cg+tc17122015211212774
None1141261448364
Unknown655764523740
Moslem
TotalN113096321423201603738937162913313
 (%)100%100%100%100%100%100%100%100%100%100%
Bidi (bd) only29392234244041592031
Cigarette (cg) only7117128127050
Tobacco chewing only1135194191490
bd+cg4020316040
bd+tc11021202612203777
cg+tc3465441300
bd+cg+tc2013261324141414100
None93711146320
Unknown655663504462
Christian
TotalN54565512453726124022550107495
 (%)100%100%100%100%100%100%100%100%100%100%
Bidi (bd) only37511832264849713540
Cigarette (cg) only611111671157140
Tobacco chewing only145701151781220
bd+cg3040203100
bd+tc151311227152500
cg+tc2385231100
bd+cg+tc20132319281513420
None821231115200
Unknown625362503740

Table II summarizes the results of lung cancer risk analysis with respect to socio-demographic factors. Lung cancer incidence was relatively high among Moslem people and those with lower educational history.

Table II. Lung Cancer Risk with Respect to Socio-Demographic Factors
 Lung cancer casesNumber of subjectsPerson-yearsRR295%CI
  • 1

    One rupee is 2–3 cents in US dollars.

  • 2

    RR: Relative Risks were obtained by a multivariate analysis with the Poisson regression model of group data, incorporating the categorized variables of all the socio-demographic factors listed in this table and attained age.

Total21265,829501,508  
Religion  p for heterogeneity = 0.001
 Hindu13547,881364,6521Reference
 Moslem6111,94191,5791.81.3–2.5
 Christian166,00745,2770.90.5–1.5
Family income (Rs.1)  p for heterogeneity = 0.681
 <500154,28332,0411Reference
 501–12006819,307146,4551.00.6–1.8
 1201–25008325,047191,4940.90.5–1.6
 2501–35003510,99884,0851.00.5–1.8
 3500+116,19447,4330.60.3–1.4
Education  p for heterogeneity < 0.001
 Illiterate264,18930,5981Reference
 Primary School10017,046126,9551.30.8–1.9
 Middle School6217,368132,1651.10.7–1.7
 High School+college2326,508206,3260.40.2–0.7
Unknown17185,4630.40.1–2.6
Occupation  p for heterogeneity = 0.182
 Fishermen and Farmers6212,40493,7271Reference
 Unemployed4715,263116,1790.70.5–1.0
 Skilled workers63,89030,0160.60.3–1.5
 Others9734,272261,5860.70.5–1.0

The upper panel of Table III summarizes the risk analysis in relation to bidi smoking, cigarette smoking and tobacco chewing, using all the study subjects (N = 65,829). The regression analyses were stratified by attained age, religion and education. Those who answered to smoke bidis currently had a 3.9-fold increase of lung cancer incidence when compared to those never smoked bidis. When the data were stratified only by attained age and religion (this model does not take into account the effect of education), the RRs increased to 4.8, indicating the presence of confounding by education. On the other hand, when the data were stratified only by attained age and education (this model does not take into account the effect of religion), the RR changed only slightly. Lung cancer risk among former bidi smokers was higher than that of those who never smoked bidis but was lower than that of current bidi smokers. This cohort study failed to show any evident increase of lung cancer risk among cigarette smokers. It is of note, however, that the average amount of cigarettes smoked a day was only 6.6 cigarettes, and the mean duration of cigarette smoking was 23.1 years, and, on average, cigarette smokers started smoking at 22.5 years of age. Tobacco chewers had slightly lower risks of lung cancer than those who never had the habit of tobacco chewing.

Table III. Lung Cancer Risk with Respect to Bidi and Cigarette Smoking, and Tobacco Chewing
 Lung cancer casesNumber of subjectsPerson-yearsRR195%CI
  • 1

    Estimates of the relative risk (RR) for ex-smokers of bidi and current-smokers of bidis were obtained using Poisson regression, based on a relative risk model in which the relative risk was assumed to be the form of (β2S2+ β3S3), where S2 and S3 were indicator variables for ex-smokers and current-smokers, respectively. Those who never smoked bidis were used for the reference category. The regression analyses were stratified by attained age, religion and education. Similar models were used for cigarette smoking and tobacco chewing.

Total21265,829501,508  
Bidi smoking    p for heterogeneity < 0.001
 Never2731,324242,8791Reference
 Former295,97043,1462.61.5–4.5
 Current14325,446191,7993.92.6–6.0
 Unknown133,08923,6843.92.0–7.6
Cigarette smoking    p for heterogeneity = 0.130
 Never8329,309222,2371Reference
 Former235,69842,1771.20.7–1.9
 Current9727,880214,3611.41.0–1.9
 Unknown92,94222,7331.50.7–3.1
Tobacco chewing    p for heterogeneity = 0.025
 Never14742,390323,3151Reference
 Former184,48332,7150.70.4–1.2
 Current4718,956145,4780.70.5–0.9
In the following analysis, cigarette smokers were excluded
Bidi smoking    p for heterogeneity < 0.001
 Never1419,045147,5961Reference
 Former51,82112,8211.30.5–3.7
 Current607,59255,3564.62.5–8.5
 Unknown48516,4643.61.2–11.3
In the following analysis, bidi smokers were excluded
Cigarette smoking    p for heterogeneity = 0.688
 Never1419,045147,5961Reference
 Former31,85414,1081.60.5–5.9
 Current99,71875,6311.40.6–3.3
 Unknown17075,5442.70.4–21.2
In the following analysis, bidi smokers and cigarette smokers were excluded
Tobacco Chewing    p for heterogeneity = 0.827
 Never1014,35011,1451Reference
 Former14883,6201.50.2 – 12.4
 Current34,20732,5310.80.2 – 2.8

The lower panel of Table III summarizes the lung cancer risk of bidi smoking among those who never smoked cigarettes, the risk of cigarette smoking among those who never smoked bidis, and the risk of tobacco chewing among those who never smoked bidis or cigarettes. The RR of lung cancer for current bidi smokers slightly increased after restriction to those never smoked cigarettes. On the other hand, the lung cancer risk of cigarette smoking did not become evident after restriction to those who had not smoked bidis. Further, we made a comparison between those who smoked bidis and cigarettes currently and those who never smoked bidis or cigarettes. The RR for those smoking bidis and cigarettes was 4.2 (95%CI = 2.4–7.5), indicating that the lung cancer risk of bidi smoking is not magnified by cigarette smoking.

To examine the risk associated with smoking only bidis, further analysis was conducted using only who never smoked cigarettes. The regression analyses were stratified by attained age, religion and education. As shown in Table IV, lung cancer risk increased with the duration of bidi smoking (p < 0.001). Lung cancer risk associated with the number of bidis smoked a day and with the age starting smoking bidis were analyzed after excluding those who stopped smoking bidis by the time of interview. Increasing numbers of bidis smoked and younger ages of starting bidi smoking were related to an elevated lung cancer risk (p < 0.001). Table V shows the lung cancer risk after cessation of bidi smoking. At first, we conducted analysis using those who did not smoke cigarettes. After cessation, lung cancer risk decreased, and the lung cancer risk among those who stopped smoking bidis within 10years before interview had a lung cancer risk similar to those who never smoked bidis. However, the small number of lung cancer cases made it difficult to examine the change of lung cancer risk with the years since cessastion of bidi smoking. Therefore, we conducted the analysis using all the subjects including cigarette smokers. The results are shown in the lower panel of Table V. Lung cancer risk did not decrease within 10 year after the cessation of bidi smoking. However, those who stopped bidi smoking for 10 years or longer had a statistically significant decrease of lung cancer risk when compared to current bidi smokers. Furthermore, the RR of this category of ex-smokers was 0.3, and was the same as that of non-smokers. However, its wide confidence interval makes it difficult to tell whether this group has the same lung cancer incidence as non-smokers.

Table IV. Bidi Smoking and Lung Cancer Risk Among Men Who Never Smoked Cigarettes
Lung cancerNumber of casesPerson-years subjectsRR295%CI
  • 1

    Those who never smoked bidi or cigarettes.

  • 2

    Estimates of the relative risk (RR) for four categories of the duration of bidi-smoking were obtained using Poisson regression, based on a relative risk model in which the relative risk was assumed to be the form of (β2S2+ β3S3+ β4S4), where S2, S3 and S4 were indicator variables for the duration of 1–29 years, 30–44 years and 45 years or longer, respectively. Those who never smoked bidis were used for the reference category. The regression analyses were stratified by attained age, religion and education. Similar models were used for the variables of the amount of bidis smoked a day and age starting smoking bidis.

Duration of smoking bidis
 Never114190451475961Reference
 1–29154087312323.31.6–7.0
 30–44353411241204.12.1–7.8
 45+151916128334.11.8–9.0
   p for trend < 0.001  
Number of bidis/day (former bidi smokers were excluded from analysis)
 Never*14190451475961Reference
 1–4 (sticks)275655671.70.4–7.8
 5–14232806204054.72.3–9.4
 15–24202449178834.92.4–10.0
 25+151550112635.12.4–10.9
   p for trend < 0.001  
Age starting smoking bidis (former bidi smokers were excluded from analysis)
 1–17121549112364.11.8–9.1
 18–22313398247095.42.8–10.4
 23+172645194113.81.9–7.9
 Never114190451475961Reference
   p for trend < 0.001  
Table V. Cessation of Bidi Smoking and Lung Cancer Risk Among Men
 Lung cancer casesNumber of subjectsPerson-yearsRR*95%CI
  • *

    In the analysis excluding cigarette smokers, estimates of the relative risk (RR) were obtained using Poisson regression, based on a relative risk model in which the relative risk was assumed to be the form of (β2S23S34S4), where S2, S3 and S4 were indicator variables for two categories of years since quitting smoking bidis, and those never smoker bidis, respectively. In the analysis using all the subjects, three categories for years since quitting bidi smoking were used. Current bidi smokers were used for the reference category. The regression analysis was stratified by attained age, religion and education.

  • **

    p for trend < 0.001 even when those who never smoked bidi were excluded from analysis.

In the following analysis, cigarette smokers were excluded
Current bidi smokers607592553561Reference
Years since quitting     
Bidi smoking     
 0–9 years41,09777250.40.1–1.0
 10 years+168448120.20.02–1.2
 Never1419,0451475960.20.1–0.4
    p for trend < 0.001** 
In the following analysis, all the subjects were analyzed
Current bidi smokers143254461917991Reference
Years since quitting     
Bidi smoking     
 0–4 years142322168700.90.5–1.5
 5–9 years101517109660.90.5–1.7
 10 years+52031145930.30.1–0.8
 Never27313242428790.30.2–0.4
    p for trend < 0.001** 

To examine the possibility that bidi smoking masks the effect of cigarette smoking, further analyses of cigarette smoking were conducted using those who never smoked bidis. The RRs for those smoking 1–14 and 15 or more cigarettes a day did not show significantly increased lung cancer risks in this analysis (Table VI). Additionally, analysis using only current bidi smokers was conducted in order to examine the possibility that bidi smoking may magnify the risk of lung cancer associated with cigarette smoking. In this analysis, the RRs of lung cancer for those smoking 1–4, 5–14, 15–24 and 25 or more cigarettes a day ranged between 0.6 and 1.2, suggesting that bidi smoking does not magnify the lung cancer risk associated with cigarette smoking. When those who smoked 1–4 sticks of bidis were excluded from analysis, the results were similar.

Table VI. Cigarette Smoking and Lung Cancer Risk Among Men
Number of cigarettes smoked a dayLung cancerNumber of subjectsPerson-yearsRR195%CI
  • **

    Trend test was conducted excluding former smokers.

  • 1

    Estimates of the relative risk (RR) for four categories of the number of cigarettes smoked a day were obtained using Poisson regression, based on a relative risk model in which the relative risk was assumed to be the form of β2S2+ β3S3+ β4S4+ β5S5+ β6S6, where S2, S3, S4, S5 and S6 were indicator variables for categories for ex-smokers, 1–4, 5–14, 15–24 and 25 or more cigarettes a day, respectively. The regression analysis was stratified by attained age, religion and education.

  • 2

    Categories of 1–4 and 5–14 were combined for calculation of RR and 95%CI.

  • 3

    Categories of 15-24 and 25+ were combined for calculation of RR and 95%CI.

The following analysis was restricted to those who never smoked bidis only
 Never14190451475961Reference
 Ex-smokers31854141081.60.5–5.9
 1–40358827987  
 5–1474715367451.320.5–3.3
 15–242106381972.330.5–10.5
 25+03292520  
    p for trend = 0.142** 
The following analysis was restricted to current bidi smokers only
 Never60759253561Reference
 Ex-smokers042321not determined 
 1–4408752662890.80.5–1.1
 5–14397380568481.20.8–1.8
 15–243127498570.60.2–1.9
 25+125319501.10.1–7.6
    p for trend > 0.5** 

Discussion and conclusions

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

In the present study, the RR of lung cancer among bidi smokers was 3.9. Cigarette smoking is probably the most important potential confounder when examining the risk of bidi smoking. However, the exclusion of cigarette smokers made the RR of lung cancer risk among current bidi smokers to be 4.6. Therefore, the excess lung cancer risk among bidi smokers is unlikely to be explained by confounding by cigarette smoking. Lung cancer risk was related to larger amounts of bidis smoked a day, longer durations of bidi smoking, and younger ages at starting smoking bidis. The observations also make it difficult to attribute the observed relationship between lung cancer risk and bidi smoking to confounding by other risk factors such as cigarette smoking and radiation. Although a large scale case-control study by Gajalakshmi et al. clearly showed an increased lung cancer risk among bidi smokers,10 the elevated lung cancer incidence among bidi smokers shown by the present cohort study strengthened the evidence for the association of lung cancer risk with bidi smoking.

Cessation of cigarette smoking is known to reduce lung cancer risk.1, 25 However, it takes about 20 years for the lung cancer risk among ex-smokers to return to the level of non-smokers after smoking cessation.1, 25 Although lung cancer risk is reported to decrease after cessation of bidi smoking,4–10 there has been few scientific reports that described how many years are necessary for lung cancer risk of ex-smokers to return to the level of non-smokers. In the present study, the excess lung cancer risk among those who answered to have stopped smoking within 10 years before interview had the cancer incidence similar to that among non-smokers. When the analysis was restricted to those who never smoked cigarettes, lung cancer incidence decreased even among those who stopped smoking during 0–9 years before interview. Note here, however, that the number of former bidi smokes is small in the analysis limited to those never smoked cigarettes, and therefore the RRs had wide confidence intervals. Although the analysis including cigarette smokers might have been affected by the presence of cigarette smokers, the results obtained from the analysis of data including cigarette smokers seem more reliable since the number of lung cancer cases was much larger. Note also that the effect of cigarette smoking was not strong in the present study, and therefore confounding by cigarette smoking is expected to be small in the analysis of lung cancer risk associated with bidi smoking.

In the present study, the relative risk of lung cancer associated with smoking 15–24 cigarettes a day, when compared to non-smokers, was 1.4, and was much smaller than what was observed in the US and Europe.1 However, it should be noted that the duration of cigarette smoking in our study population was relatively short on average (on average 23.1 years) partially because the average age starting smoking was relatively old. Doll and Peto has shown that the relative risk of cigarette smoking is more strongly related to duration of the habit than daily cigarette consumption.26 In addition, average amount of cigarettes smoked a day was only 6.6 cigarettes. Therefore, relatively low cumulative amount of cigarettes smoked by this cohort is another factor that made it difficult to detect lung cancer risk related to cigarette smoking. Taking those points into account, the hazard of cigarette is considered not to be full-blown in this population, yet. Therefore, lung cancer risk associated with cigarette smoking in this population will become much larger when the duration of smoking becomes longer.

Smoking both bidis and cigarettes did not appear to increase lung cancer risk when compared to smoking bidis alone or smoking cigarettes alone. However, we cannot deny the possibility that bidi smoking magnifies the lung cancer risk associated with heavy cigarette smoking or that heavy cigarette smoking may magnify the association of lung cancer risk with bidi smoking.

Bidi tobacco contains higher levels of hazardous chemicals when compared to cigarettes.27–29 A study has shown that the mainstream smoke of bidi contains a much higher concentration of carcinogenic hydrocarbons such as benz(a)anthracene and benzopyrene than US cigarettes.1, 28 Further, bidi smokers were found to take almost five puffs per minute compared to the cigarette smokers who smoked two puffs per minitue.27 A couple of studies reported the possibility that bidi smoking is more strongly associated with lung cancer than cigarette smoking is.8, 10 However, such a comparison seems to depend on the duration of smoking, daily consumption of bidis or cigarettes and age starting smoking. Other factors such as passive smoking and dietary habits may also need to be taken into account. In the present study, we could not address this question since the lung cancer risk associated with cigarette smoking was low probably because of a relatively small number of medium to heavy smokers and because of relatively short duration of cigarette smoking. However, from the view point of preventing cancer associated with smoking, whether smoking bidis or cigarettes may not be important since bidi smoking is at least as hazardous as cigarette smoking.

In India, official mortality statistics may be underestimates and the accuracy of information on cause of death is sometimes questionable, particularly, among poor people in rural areas. Therefore, an important feature of the present study is its use of cancer incidence data, rather than mortality data. Even though lung cancer is mostly fatal, the use of cancer incidence data increased the accuracy of diagnosis and completeness in case ascertainment. However, although we made almost every effort to ascertain cancer cases in Karunagappally, we might have failed to identify at least a small percentage of cancer cases since bidi smoking is more common among poor people. In the present study, however, the results did not change sizably when the lowest income group was excluded from analysis (data not shown).

Migration of people is another concern, particularly, for cancer case ascertainment. However, the migration of people in Karunagapally, in its neighboring areas and in Trivandrum is unlikely to have caused serious incompleteness of cancer case ascertainment in the present study since various activities to ascertain cancer cases were conducted not only in Karunagappaly but also in its neighboring areas and in Trivandrum. We have to admit though that cancer cases among those who migrated out of the Karunagapally and Trivandrum areas were unlikely to be identified by us. Since the majority of migration took place for job opportunities in Gulf countries, it was difficult for us to identify cancer cases that were diagnosed and treated in those countries unless they came back to their home town to undergo medical treatment.

A disadvantage of a cohort study is the fact that the lifestyle of cohort members, examined at the start of its following-up, may change during follow-up. In the present study, no attempt was made to re-interview the cohort members. Generally speaking, however, the number of subjects who start smoking after age 30 is considered to be limited. Indeed, most of the bidi smokers examined in the present study took up smoking before age 30. On the other hand, it is difficult to estimate the number of subjects who quit smoking during follow-up. Because of such problems, the RRs for bidi smoking presented in the present study might have been underestimated. In addition, years after cessation of smoking is probably underestimated since the information used in the present study is the years after cessation until the time of interview.

In conclusion, lung cancer incidence was strongly related to the amount of bidis smoked a day, duration of smoking, and age at starting smoking. After quitting smoking bidis, the excess lung cancer risk did not return to the cancer incidence of non-smokers within 10 years after quitting bidi smoking. To our knowledge, this is the first cohort study that compared lung cancer incidence, rather than mortality, among those smoking only bidis with that among non-smokers. Immediate measures should be taken to stop bidi smoking, which is common in south Asia.

Acknowledgements

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

We acknowledge the research and technical staff of Natural Background Radiation Cancer Registry, Karunagappally, and Regional Cancer Center (RCC), Trivandrum. We also acknowledge medical officers of RCC, who immensely helped us detecting cancer cases and providing support services to cancer patients.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion and conclusions
  6. Acknowledgements
  7. References
  • 1
    IARC.Tobacco smoke and involuntary smoking. IARC monographs on the evaluation of carcinogenic risks to humans, vol. 83. Lyon: IARC Publication, 2004.
  • 2
    ReddyS,GuptaPC, eds. Report on Tobacco control in India, Ministry of Health and Family Welfare, New Delhi: Government of India, 2004.
  • 3
    Sanghvi LD,Rao KCM,Khanolkar VR. Smoking and Chewing of Tobacco in relation to Cancers of the upper alimentary tract. Brit Med J 1955; 1: 11114.
  • 4
    Gupta PC,Pednekar MS,Parkin DM,Sankaranarayanan R. Tobacco associated mortality in Mumbai (Bombay) India. Results of the Bombay Cohort Study. Int J Epidemiol 2005; 34: 1395402.
  • 5
    Notani PN,Jayant K,Sanghvi LD. Assessment of morbidity and mortality due to tobacco usage in India. In: SanghviLD,NotaniPN, eds. Tobacco and health: the Indian scene, Bombay: Tata Memorial Centre, 1987.
  • 6
    Notani PN,Rao DN,Sirsat MV,Sanghvi LD. A study of lung cancer in relation to bidi smoking in different religious communities in Bombay. Indian J Cancer 1977; 2: 11521.
  • 7
    Jussawalla DJ,Jain DK. Lung cancer in Greater Bombay: correlations with religion and smoking habits. Br J Cancer 1979; 40: 43748.
  • 8
    Dikshit RP,Kanhere S. Tobacco habits and risk of lung, oropharyngeal and oral cavity cancer: a population-based case-control study in Bhopal. India. Int J Epidemiol 2000; 29: 60914.
  • 9
    Gupta D,Boffetta P,Gaborieau V,Jindal SK. Risk factors of lung cancer in Chandigarh. India. Indian J Med Res 2001; 113: 14250.
  • 10
    Gajalakshmi V,Hung RJ,Mathew A,Varghese C,Brennan P,Boffetta P. Tobacco smoking and chewing, alcohol drinking and lung cancer risk among men in southern India. Int J Cancer 2003; 107: 4417.
  • 11
    Nair MK,Amma S,Gangadharan P. Cancer incidence in Karunagapally 1990–1992, Kerala, India. In: ParkinDM,WhelanSL,FerlayJ,RaymondL,YoungJ, eds.Cancer Incidence in five continents, vol. VII. IARC Publication No. 143: Lyon: International Agency for Research on Cancer 1997: 3503.
  • 12
    Nair MK,Gangadharan P,Jayalakshmi P,Mani KS. Cancer incidence in Karunagapally 1993–1997, Kerala, India. In: ParkinDM,WhelanSLFerlayJ,TeppoL,ThomasDB, eds. Cancer incidence in five continents, vol. VIII.; IARC Publication No. 155. Lyon: International Agency for Research on Cancer 2002: 2401.
  • 13
    Jayalekshmi P,Rajan B. Cancer incidence in Karunagapally 1998–2002, Kerala, India. In: CuardoMP,EdwardsB,ShinHR,StormH,FerlayJ,HeanueM,BoyleP, eds.Cancer incidence in five continents, vol. IX. IARC Publication No. 160. Lyon: International Agency for Research on Cancer, 2007.
  • 14
    Gopal Iyengar AR,Sundaram R,Mistry KB,Sunta CM,Nambi KSV,Kathuria SP et al. Evaluation of the long-term effect of High Background Radiation on selected population groups of the Kerala Coast. Proceedings of the. IV International conference on peaceful uses of Atomic Energy. Geneva: United Nations 1972: 113151
  • 15
    WHO. Effect of Radiation on human heredity. WHO Technical Report series No 166. Investigations of areas of High Natural Radiation. Geneva: World Health Organization 1959.
  • 16
    Nair MK,Nambi KSV,Amma NS,Gangadharan P,Jayalekshmi P,Jayadevan S,Cherian V,Reghuram KN Population study in the high natural background radiation area in Kerala. India Radiat Res 1999; 152( 6 Suppl): S145S148.
  • 17
    Jayalekshmi P,Gangadharan P,Binu VS,Reghuram KN,Nair MK,Rajan B,Akiba S. What did we learn from epidemiological studies in High background radiation areas in India? In: SugaharaT,SasakiY,MorishimaH,HayataI,SohrabiM,AkibaS, eds. Radiation dose and health effects. Proceedings of 6th high levels of natural radiation and radon areas. International conference series No. 1276. Amsterdam: Elsevier, 2004: 1025.
  • 18
    Nair MK,Gangadharan P,Jayalakshmi P,Reghuram KN,Mani KS. Natural background radiation cancer registry, technical report 1990–1999. Trivandrum: Regional Cancer Center, 2004.
  • 19
    Gangadharan P,Jayalekshmi O,Binu VS,Sreekumar MR. Cancer morbidity and mortality in Karunagappally, 1993–2001. Natural background radiation cancer registry. Trivandrum: Regional Cancer Center, 2004.
  • 20
    NandakumarA,RamnathT,RoselindFS,ShobanaB,PrabhuK,ShahB,ChaundhryK, eds. Two-year report of the population based cancer registries. Bangalore: National Cancer Registry Programme, 2005.
  • 21
    Breslow NE,Day NE.Statistical methods in cancer research. Volume 2 - The design and analysis of cohort studies, IARC Scientific Publications, No. 82. Lyon: International Agency for Research on Cancer, 1987.
  • 22
    Preston D,Lubin J,Pierce D,McConney M. Epicure users manual guide. Seattle: Hirosoft International Corporation, 1993.
  • 23
    Preston DL,Shimizu Y,Pierce DA,Suyama A,Mabuchi K. Studies of mortality of atomic bomb survivors. Report 13: Solid cancer and noncancer disease mortality: 1950–1997. Radiat Res 2003; 160: 381407.
  • 24
    Cardis E,Vrijheid M,Blettner M,Gilbert E,Hakama M,Hill C,Howe G,Kaldor J,Muirhead CR,Schubauer-Berigan M,Yoshimura T,Bermann F, et al. The 15-Country Collaborative Study of Cancer Risk among Radiation Workers in the Nuclear Industry: estimates of radiation-related cancer risks. Radiat Res 2007; 167: 396416.
  • 25
    US Department of Health and Human Services. The health benefits of smoking cessation. A report of the surgeon general. DHHS publication No (CDC) 90-8416. Rockville, MD: US Government printing office, 1990.
  • 26
    Doll R,Peto R. Cigarette smoking and bronchial carcinoma: dose and time relationships among regular smokers and lifelong non-smokers. J Epidemiol Community Health 1978; 32: 30313.
  • 27
    Hoffman D,Sanghvi LD,Wynder EL. Comparative chemical analysis of Indian bidi and American cigarette smoke. Int J Cancer 1974; 14: 4953.
  • 28
    Gupta PC,Murti PR,Bhonsle RB. Epidemiology of cancer by tobacco products and the significance of TSNA. Crit Rev Toxicol 1996; 26: 183198.
  • 29
    Jayant K,Pakhale SS. Toxic constituents in bidi smoke. In: SanghaviLD,NotaniPN, eds. Tobacco and health: indian scene. Bombay: Tata Memorial Center 1989: 10110.