Radiol Oncol 2006; 40(4): 239-44.
Body mass index and lung cancer risk in never smokers
Katsunori Kagohashi, Hiroaki Satoh, Koichi Kurishima, Hiroichi Ishikawa, Morio Ohtsuka
Division of Respiratory Medicine, Institute of Clinical Medicine, University of Tsukuba, Japan
Background. A relationship between body mass index (BMI) and lung cancer risk in never smokers has
not been reported precisely. To evaluate the risk of lung cancer associated with BMI in never smokers, we
conducted a case-control study.
Methods. The relationship between BMI and the risk of lung cancer in never smokers was investigated in
a study of 204 lung cancer cases and 398 controls admitted between 1987 and 2005. Controls were selected
from hospitalized age-matched never-smoking patients with non-malignant respiratory disease.
Results. When compared with BMI of the leanest group (BMI<20.8) in men, no inverse association between
BMI and lung cancer was observed after the adjustment for age (the second BMI group: BMI? 20.8 to <
22.9; p=0.683, the third BMI group: BMI? 22.9 to < 24.9; p=0.745, and the highest BMI group: BMI? 25.0;
p=0.327). Similarly, no association in women was found between BMI and lung cancer in these three BMI
groups (the second group, p=0.639; the third group, p=0.667; the highest group, p=0.978) when compared
with that of the leanest BMI group.
Conclusions. Our present study indicated that the association between leanness and the risk of lung cancer
might be influenced by other factors such as smoking.
Key words: lung neoplasms – epidemiology; risk factors; adenocarcinoma; body mass index; smoking
Introduction
An elevated risk of lung cancer associated with lower levels of body mass index (BMI) has been reported in previous studies.1-9 However, the interpretation of the associa-Received 3 December 2006 Accepted 14 December 2006
Correspondence to: Hiroaki Satoh, MD, Division of Respiratory Medicine, Institute of Clinical Medicine, University of Tsukuba, Tsukuba-city, Ibaraki, 305-8575, Japan; Phone: + 81 29 853 3210; Fax: + 81 29 853 3320; E-mail: hirosato@md.tsukuba.ac.jp
tion between low BMI and lung cancer is complicated by the fact that weight loss may be a sign of smoking. In general, smokers tend to be lighter than non-smok-ers,1, 10-15 and it is believed that nicotine is responsible for the effect of smoking on body weight because nicotine appears to increase the metabolic rate.
We have recently performed a case-control study of subjects participating in a mass-screening program and found an increased risk of lung cancer for lower BMI in male patients.16 Most of them were current smokers.16  However, the inverse
240
Kagohashi K et al. / BMI and lung cancer in non-smokers
association between the risk of lung cancer and BMI was not found in female patients. This was inconsistent with the results of previous studies.1-9 The difference between the findings of previous studies and our own was most probably due to difference in study population. The subjects in previ-ous studies were symptomatic, but all the subjects in our recent study were asymp-tomatic preclinical patients.16 In addition, there were higher proportions of never-smoking women and lung adenocarcinoma in our study, and the different results might also be influenced by them.
Whereas, it remains an unsettled ques-tion whether there is an inverse association between the risk of lung cancer and BMI among never smokers, especially in those with lung adenocarcinoma. In order to evaluate the association between BMI and the risk of lung cancer in patients who never smoked, we conducted a hospital-based case-control study. In this study, we also estimated the association between BMI and the risk of lung adenocarcinoma in never-smokers.
Methods
Study Design
A hospital-based case-control study was conducted from January 1987 to September 2005 in Respiratory Division of Tsukuba University Hospital, to explore the risk of lung cancer associated with lower levels of BMI at the time of initial diagnosis.
Patients
All cases and controls were recruited at Tsukuba University Hospital and were identified from the medical record. All patients who were histopatologically con-firmed to have lung cancer were included in this analysis. Pathological diagnoses
Radiol Oncol 2006; 40(4): 239-44.
were based on the 1999 WHO classification of lung neoplasms. Only lung cancer pa-tients as well as control subjects, who had never smoked and without occupational or domestic exposure to other recognized car-cinogens, were included in this study. The hospital-based controls matched by age (±5 yr), gender, and time of hospitalization were recruited from our division, including inpatients with non-malignant respiratory diseases. Patients with other cancer at any site were excluded. This study was ap-proved by the institutional ethics commit-tee of University of Tsukuba.
Statistical analysis
BMI was calculated from body height and weight, which were measured and reported by nurses at the time of admission, using the formula for Quetelet’s index (expressed in kg/m2). BMI was categorized into four levels on the basis of the distribution in the total study population (BMI < 20.8, leanest; 20.8 ? BMI < 22.9, second; 22.9 ? BMI < 25.0, third; 25.0 ? BMI, highest), which was the same category as our previous study based on the results of community mass screening.16 We used the leanest category (BMI < 20.8) as the reference group for analyses.
Logistic regression was used to exam-ine the effect of BMI on lung cancer risk. Results with a p value less than 0.05 were regarded as significant. The software pack-age SSPE (SSPE Inc., Chicago, IL, USA) and Microsoft Excel (Microsoft, Redmond, WA, USA) were used to perform the analyses stated above.
Results
During the study period, 919 patients with primary lung cancer were diagnosed in our division. Among them, a total of 204
Kagohashi K et al. / BMI and lung cancer in non-smokers
241
(22.2 %) documented never-smokers with primary lung cancer were registered.
Table 1 shows the characteristics of 204 lung cancer cases. Three-fourths of all cas-es were women. Although the age range for this study was between 22 and 85 years, median age of the male and female patients was 64 and 67 years, respectively. Twenty-four (64.9 %) of 37 male patients, and 151 (90.4 %) of 167 female patients had lung ad-enocarcinoma. There was no patient with small cell lung cancer in both genders (Table 1). Among the 204 patients, 81 had stage IA-IIIA, 35 had stage IIIB, and 88 had stage IV disease.
Table 2 shows odd ratios and 95 percent confidence intervals for the association between the risk of lung cancer and lower levels of BMI at the time of diagnosis with lung cancer. When compared with BMI of the leanest group (BMI<20.8) in men, no inverse association between BMI and lung cancer was observed after the adjustment for age (the second BMI group: BMI? 20.8 to < 22.9; p = 0.683, the third BMI group: BMI? 22.9 to < 24.9; p = 0.745, and the high-est BMI group: BMI? 25.0; p = 0.327). In women, no association was found between
Table 1. Characteristics of 204 lung cancer patients
Characteristics	Men	Women
Number of patients	37	167
Age (median, range), Yr	64, 22 – 80	67, 35 – 85
Histologic types		
Adenocarcinoma	24	151
Squamous cell carcinoma	10	9
Large cell carcinoma	1	4
Others	2	3
Stage		
IA – IIIA	11	70
IIIB	11	24
IV	15	73
BMI and lung cancer in these three BMI group (the second BMI group, p = 0.639; the third BMI group, p = 0.667; the high-est BMI group, p = 0.978) when compared with that of the leanest BMI group. In male adenocarcinoma, the second BMI group (p=0.967), the third BMI group (p=0.310), and the highest BMI group (p = 0.378) did not exhibit higher odd ratio than that of the leanest BMI group after the adjustment for age.
As above mentioned, 175 of 204 patients were adenocarcinomas. There were too-few
Table 2. Association of BMI with lung cancer
	Case	Control			
BMI	Patients, No.	Subjects No.	OR	95% CI	p-value
Men					
< 20.8	13	32	1.0		
20.8 – 22.8	10	20	0.8	0.3 – 2.2	0.683
22.9 – 24.9	5	10	0.8	0.2 – 2.8	0.745
? 25.0	9	13	0.6	0.2 – 1.7	0.327
Women					
< 20.8	67	131	1.0		
20.8 – 22.8	43	75	0.9	0.6 – 1.4	0.639
22.9 – 24.9	23	51	1.1	0.6 – 2.0	0.667
? 25.0	34	66	1.0	0.6 – 1.7	0.978
OR: odd ratio; CI: confidence interval
Radiol Oncol 2006; 40(4): 239-44.
242
Kagohashi K et al. / BMI and lung cancer in non-smokers
Table 3. Association of BMI with lung adenocarcinoma
	Case	Control			
BMI	Patients, No.	Subjects No.	OR	95%CI	p-value
Men					
< 20.8	8	26	1.0		
20.8 – 22.8	6	19	1.0	0.3 – 3.5	0.967
22.9 – 24.9	4	6	2.2	0.5 – 9.6	0.310
? 25.0	6	11	1.7	0.5 – 6.3	0.378
Women					
< 20.8	61	120	1.0		
20.8 – 22.8	36	67	1.1	0.6 – 1.8	0.831
22.9 – 24.9	22	46	0.9	0.5 – 1.7	0.841
? 25.0	32	58	1.1	0.6 – 1.8	0.762
OR: odd ratio; CI: confidence interval
patients to stratify the histologic subtypes in both men and women, therefore, we examined the association only in adenoca-cinoma cases (Table 3). In men, the second BMI group (p = 0.967), the third BMI group (p = 0.310), and the highest BMI group (p = 0.378) did not exhibit higher odd ratio than that of the leanest BMI group after the ad-justment for age. In women, no association was found between BMI and lung cancer in these three BMI groups (the second BMI group, p = 0.831; the third BMI group, p = 0.841; the highest BMI group, p = 0.762) when compared with that of leanest BMI group. For adenocarcinoma, therefore, the inverse association was not observed in both genders.
Discussion
An inverse gradient between BMI and the incidence of lung cancer has been reported in several case control and cohort studies.1-9 However, the interpretation of the asso-ciation between low BMI and lung cancer is complicated by the fact that low BMI may be influenced by other factors such
Radiol Oncol 2006; 40(4): 239-44.
as smoking. It is possible that smokers tend to have lighter body weight, possibly a consequence of the metabolic effects of nicotine.
In order to investigate the association between lower levels of BMI and the risk of lung cancer in never-smoking patients, we, therefore, conducted a hospital-based case-control study.
The results of the present study indicate two important points. The first point of importance is that we found the absence of an inverse gradient between BMI and the risk of lung cancer in never-smoking female patients, which was consistent with our recent study of subjects participating in a mass screening program.16 Interestingly, Rauscher et al reported an elevated risk of lung cancer associated with not “low” but “high” levels of BMI in non-smoking female patients.17 The difference between the find-ings of Rauscher and our own was prob-ably due to different study populations. All patients and controls in our study were never-smokers. On the other hand, howev-er, not all subjects in the study by Rauscher et al were never-smokers.17 They included 188 patients who haven’t smoked more
Kagohashi K et al. / BMI and lung cancer in non-smokers
243
than 100 cigarettes in their lifetime and 224 patients who haven’t smoked more than 100 cigarettes during the last 10 years.17 Therefore, there was a possibility that the results were influenced by the residual ef-fects of smoking.
The majority of published results inves-tigating the association between lower BMI and lung cancer risk were based on stud-ies conducted in Western countries where the prevalence of obesity is high.2,8,17 The difference in categorized BMI levels might also influence the difference between the findings of previous authors and our own. The second important point is that both thin male and female never smokers did not have an increased risk of lung adeno-carcinoma. The results of the present study indicate that lower BMI is not significantly associated with the risk of lung adenocarci-noma. Some lung adenocarcinomas in never smokers can arise without the growth pro-moting effects of the carcinogens present in cigarette smoke,18 one can postulate that the mechanism of carcinogenesis of adenocarcinomas arising in smokers and never-smokers may be different, and these distinct tumorigenic mechanisms can imply differences in tumor biology, demographic characteristics as Brownson and colleague
suggested.19
Although we showed the above-men-tioned two findings of importance, we must acknowledge the limitation of this study. First, it was a hospital-based case-control study. As with any hospital-based case-control study, it has been suggested that such case-control studies may reflect the presence of disease other than lung cancer in control subjects.2 The second limitation of the present study is that it included only small number of patients and controls in a single institute. The third, 123 (60.3%) of 204 patients had locally advanced or meta-static lung cancer and stage of the disease would be important in that more advanced
disease might itself be associated with weight loss. It is interesting to know the as-sociation between BMI and the risk of lung cancer among patients with early disease, but we could not evaluate the association because of small number of study population. The last, we could not examine the im-pact of environmental tobacco smoke (ETS) exposure on this association. An accurate assessment of the amount of ETS exposure in never smokers is necessary for determin-ing the lung cancer risks associated with ETS exposure.
It is well known that smokers tend to be leaner than non-smokers.1,10-15 Several previous studies have reported an asso-ciation between leanness and risk of lung cancer, mainly among smokers,12,14 and among men with smoking-related disease.14 Inconsistence with previous studies,1-9 our results indicated the possibility that the previously reported association between leanness and the risk of lung cancer might be influenced by other factors such as smoking. A large cohort study will be needed to confirm the current results.
References
1.   Nomura A, Heilbrun LK, Stemmermann GN. Body mass index as a predictor of cancer in men. J Natl Cancer Inst 1985; 74 : 319-23.
2.   Kabat GC, Wynder EL. Body mass index and lung cancer risk. Am J Epidemiol 1992; 135: 769-74.
3.   Knekt P, Heliovaara M, Rissanen A, Aromaa A, Seppanen R, Teppo L, et al. Leanness and lung-cancer risk. Int J Cancer 1991; 49: 208-13.
4.   Chyou PH, Nomura A, Stemmermann GN. A prospective study of weight, body mass index and other anthropometric measurements in relation to site-specific cancers. Int J Cancer 1994; 57: 313-7.
5.   Drinkard CR, Sellers TA, Potter JD, Zheng W, Bostick RM, Nelson CL, et al. Association of body mass index and body fat distribution with risk of lung cancer in older women. Am J Epidemiol 1995; 14 2 : 600-7.
Radiol Oncol 2006; 40(4): 239-44.
244
Kagohashi K et al. / BMI and lung cancer in non-smokers
6.   Kark JD, Yaari S, Rasooly I, Goldbourt U. Are lean smokers at increased risk of lung cancer? The Israel Civil Servant Cancer Study. Arch Intern Med 1995; 155: 2409-16.
7.   Goodman MT, Wilkens LR. Relation of body size and risk of lung cancer. Nutr Cancer 1993; 20: 179-86.
8.   Olson JE, Yang P, Schmitz K, Vierkant RA, Cerhan JR, Sellers TA. Differential association of body mass index and fat distribution with three major histologic types of lung cancer: evidence from a cohort of older women. Am J Epidemiol 2002; 156: 606-15.
9.   Albanes D, Jones DY, Schatzkin A, Micozzi MS, Taylor PR. Adult stature and risk of cancer. Cancer Res 1988; 48: 1658-62.
10. Garrison RJ, Feinleib M, Castelli WP, McNamara PM. Cigarette smoking as a confounder of the re-lationship between relative weight and long-term mortality: the Fremingham Heart Study. JAMA 1983; 249: 2199-203.
11. Vandenbroucke JP, Mauritz BJ, de Bruin A, Verheesen JH, van der Heide-Wessel C, van der Heide RM. Weight, smoking, and mortality. JAMA 1984; 252: 2859-60.
12. Sidney S, Friedman GD, Siegelaub AB. Thinness and mortality. Am J Public Health 1987; 77: 317-22.
13. Marti B, Tuomilehto J, Korhonen HJ, Kartovaara L, Vartiainen E, Pietinen P, et al. Smoking and lean-ness: evidence for change in Finland. BMJ 1989; 298: 1287-90.
14. Wannamethee G, Shaper AG. Body weight and mortality in middle aged British men: impact of smoking. BMJ 1989; 299: 1497-502.
15. Henley SJ, Flanders WD, Manatunga A, Thun MJ. Leanness and lung cancer risk: fact or artifact? Epidemiology 2002; 13: 268-76.
16. Kanashiki M, Sairenchi T, Saito Y, Ishikawa H, Satoh H, Sekizawa K. Body mass index and lung cancer. A case-control study of subjects participat-ing in a mass-screening program. Chest 2005; 128: 1490-6.
17.  Rauscher GH, Mayne ST, Janerich DT. Relation between body mass index and lung cancer risk in men and women never and former smokers. Am J Epidemiol 2000; 152: 506-13.
18. Nordquist LT, Simon GR, Cantor A, Alberts WM, Bepler G. Improved survival in never smokers vs current smokers with primary adenocarcinoma of the lung. Chest 2004; 126: 347-51.
19. Brownson RC, Loy TS, Ingram E, Myers JL, Alavanja MC, Sharp DJ, et al. Lung cancer in nonsmoking women. Histology and survival patterns. Cancer 1995; 75: 29-33.
Radiol Oncol 2006; 40(4): 239-44.
Slovenian abstracts
277
Radiol Oncol 2006; 40(4): 239-44.
Indeks telesne mase in tveganje za pljučnega raka pri nekadilcih
Kagohashi K, Satoh H, Kurishima K, Ishikawa H, Ohtsuka M
Izhodišča. Razmerje med indeksom telesne mase (BMI) in tveganjem za pljučnega raka pri nekadilcih do sedaj še ni bilo natančno opredeljeno. Da bi to tveganje ocenili, smo naredili študijo primerov.
Metode. V letih 1987 do 2005 smo v raziskavi obravnavali 204 bolnikov nekadilcev s pljučnim rakom in 398 bolnikov v kontrolni skupini. Starostno primerljive bolnike v kontrolni skupini smo izbrali med nekadilci, ki so se zdravili v naši bolnišnici zaradi nemaligne pljučne bolezni.
Rezultati. Ko smo primerjali starostno primerljivo skupino bolnikov z najmanjšim BMI (BMI < 20.8), nismo ugotovili manjše povezave s pljučnim rakom kot v skupinah z višjim BMI (v drugi skupini z BMI ? 20,8 do < 22,9, p = 0,683; v tretji skupini z BMI ? 22,9 do < 24,9, p = 0,745; in v skupini z najvišjim BMI, BMI ? 25,0, p = 0,327).
Prav tako primerjava omenjenih skupin pri ženskah ni pokazala povezave med BMI in pojavnostjo pljučnega raka (pri drugi skupini je bil p = 0,639; pri tretji p = 0,667; in pri skupini z najvišjim BMI p = 0,978).
Zaključki. Naša raziskava ni pokazala, da bi pri nekadilcih BMI bil povezan s stopnjo tveganja za pljučnega raka. Na obolevanje za pljučni rak še vedno najbolj vplivajo drugi dejavniki, največ kajenje.
Radiol Oncol 2006; 40(4): 273-8.