Researchers may have uncovered why lung cancer afflicts some smokers and not others, according to data presented at the American Association for Cancer Research 100th Annual Meeting 2009, Denver.

"A history of smoking has always been thought of as a predictor of lung cancer, but it is actually not very accurate," said Dr. Jian-Min Yuan, associate professor of public health at the University of Minnesota. "Smoking absolutely increases your risk, but why it does so in some people but not others is a big question."

Yuan and colleagues hypothesised that the presence of the metabolite NNAL in a patient's urine might predict risk of lung cancer. This metabolite has been shown to induce lung cancer in laboratory animals, but the effect in humans had not yet been studied.

Researchers collected data from 18,244 men enrolled in the Shanghai Cohort Study and 63,257 men and women from the Singapore Chinese Health Study. In addition to in-person interviews to assess levels of cigarette smoking, dietary and other lifestyle factors, researchers collected blood and urine samples from more than 50,000 patients.

To evaluate the impact of NNAL, researchers identified 246 current smokers who later developed lung cancer and 245 smokers who did not develop lung cancer during the 10-year period following initial interview and collection of urine samples.

Levels of NNAL in the urine were divided into three groups. Compared to those with the lowest levels, patients with a mid-range level of NNAL had a 43 per cent increased risk of lung cancer, while those at the highest level had a more than two-fold increased risk of lung cancer after taking into account the effect of number of cigarettes per day, number of years of smoking, and urinary levels of cotinine on lung cancer risk.

Levels of nicotine in the urine were also calculated. Those with the highest levels of nicotine and NNAL had an 8.5-fold increase in the risk of lung cancer compared with smokers who had the lowest levels after accounting for smoking history.

"Smoking leads to lung cancer, but there are about 60 possible carcinogens in tobacco smoke, and the more accurately we can identify the culprit, the better we will become at predicting risk," said Yuan.