TY - JOUR
T1 - Tobacco smoking and the risk of upper aero-digestive tract cancers
T2 - A reanalysis of case-control studies using spline models
AU - Polesel, Jerry
AU - Talamini, Renato
AU - La Vecchia, Carlo
AU - Levi, Fabio
AU - Barzan, Luigi
AU - Serraino, Diego
AU - Franceschi, Silvia
AU - Dal Maso, Luigino
PY - 2008/5/15
Y1 - 2008/5/15
N2 - Although tobacco smoking has long been recognized as a major risk factor for cancer of the upper aero-digestive tract (UADT, i.e., oral cavity, pharynx, larynx, and oesophagus), very few studies have provided estimates of the effect of very low tobacco consumption. Step-functions have been the common statistical methods for risk estimates, but the choice of reference category and of interval cutpoints influence the results, especially when data are sparse. In the present analysis, the dose-response relationship between UADT cancers and tobacco smoking was evaluated through logistic regression spline models. We included 1,241 UADT male cases and 2,835 male controls pooled from a large series of case-control studies conducted in northern Italy and in the Swiss Canton of Vaud during the last 2 decades. For cancers of the pharynx, larynx and oesophagus, the risk steadily increased with number of cigarettes/day. The risk of oral, pharyngeal and oesophageal cancers was significantly higher in smokers than in nonsmokers beginning with as low as 2 cigarettes/day. The effect of tobacco smoking at low levels seemed less evident for laryngeal cancer since the raise in risk begun with 6 cigarettes/day. In conclusion, for all the examined UADT sites, a monotonic dose-response relationship between cancer risk and cigarette smoking emerged. The excess of risk among people smoking 2 cigarettes/day highlights the absence of any harmless level for cigarette smoking, and it further supports the need of public health programs against tobacco smoking.
AB - Although tobacco smoking has long been recognized as a major risk factor for cancer of the upper aero-digestive tract (UADT, i.e., oral cavity, pharynx, larynx, and oesophagus), very few studies have provided estimates of the effect of very low tobacco consumption. Step-functions have been the common statistical methods for risk estimates, but the choice of reference category and of interval cutpoints influence the results, especially when data are sparse. In the present analysis, the dose-response relationship between UADT cancers and tobacco smoking was evaluated through logistic regression spline models. We included 1,241 UADT male cases and 2,835 male controls pooled from a large series of case-control studies conducted in northern Italy and in the Swiss Canton of Vaud during the last 2 decades. For cancers of the pharynx, larynx and oesophagus, the risk steadily increased with number of cigarettes/day. The risk of oral, pharyngeal and oesophageal cancers was significantly higher in smokers than in nonsmokers beginning with as low as 2 cigarettes/day. The effect of tobacco smoking at low levels seemed less evident for laryngeal cancer since the raise in risk begun with 6 cigarettes/day. In conclusion, for all the examined UADT sites, a monotonic dose-response relationship between cancer risk and cigarette smoking emerged. The excess of risk among people smoking 2 cigarettes/day highlights the absence of any harmless level for cigarette smoking, and it further supports the need of public health programs against tobacco smoking.
KW - Dose-response
KW - Spline models
KW - Tobacco smoking
KW - Upper aero-digestive tract cancer
UR - http://www.scopus.com/inward/record.url?scp=42149153209&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=42149153209&partnerID=8YFLogxK
U2 - 10.1002/ijc.23385
DO - 10.1002/ijc.23385
M3 - Article
C2 - 18224689
AN - SCOPUS:42149153209
VL - 122
SP - 2398
EP - 2402
JO - International Journal of Cancer
JF - International Journal of Cancer
SN - 0020-7136
IS - 10
ER -