Year: 1997
Type: article
Abstract: We agree that the large chemoprevention trials of β-carotene supplementation produced surprising, if not shocking, results. Randomized trials are the acid test of inferences from observational epidemiology and clinical and laboratory mechanistic studies. Randomization is intended to avoid confounding. We have intensively analyzed CARET results for the influence of pack-years of smoking, cigarettes smoked per day, and variation in smoking during the intervention, as well as age, sex, occupational exposures to asbestos, alcohol intake, elapsed time since randomization, baseline serum and dietary β-carotene, and other factors. None of these variables changes the striking adverse results in CARET (or the ATBC Study). Moreover, none of these factors are disproportionate between the intervention and control groups. The data are more indicative that interaction of β-carotene/retinyl palmitate or β-carotene with active cigarette smoking leads to a carcinogenic effect through the interaction of the agents and the chemicals in cigarette smoke rather than through confounding on statistical grounds. Among the population of heavy smokers, the excess total mortality associated with the active intervention is not statistically significantly different for former and current smokers (relative risk [RR] of 1.06 for the former smokers versus RR of 1.15 for the current smokers) ( 1 ), and the former smokers have higher excess cardiovascular mortality in the active arm, versus placebo, than do the current smokers in CARET (our unpublished data). The lung cancer incidence rates are very similar in the ATBC Study and CARET, as Pastorino's Table 1 shows. It is no surprise that current smokers have an eightfold higher lung cancer risk (and twofold to threefold higher total mortality risk) than the mostly neversmokers or long-term former smokers in the Physicians' Health Study. The difference in total mortality rates of 201 versus 119 in the control arms, respectively, for the ATBC Study versus CARET may be attributed partly to inclusion of 44% women in the CARET smoker population. The total mortality rate for CARET male current smokers on the placebo arm, excluding those exposed to asbestos, is 158 per 10 000 person-years. This group had a lower average alcohol intake than the ATBC Study population (median, 3 g/day versus 11 g/day for the ATBC Study) and was similar with regard to age, average number of cigarettes smoked per day, and body mass index. The Physicians' Health Study had such low incidence rates for lung cancer that it is unwarranted to hold out their results of no detectable harm as balancing the ATBC Study and CARET lung cancer results, despite the longer duration of treatment. With regard to trials of specific agents, we look forward to results this year from the EUROSCAN trial ( 2 ) of retinyl palmitate alone and in combination with N -acetyl-cysteine in chemoprevention of recurrences and new primary tumors in patients with completely resected, early stage non-small-cell lung cancer, based on Pastorino's previous adjuvant trial ( 3 ). The International Agency for Research on Cancer is launching its Chemoprevention Programme with a formal evaluation of animal and human evidence of cancer prevention and carcinogenic effects of β-carotene and other carotenoids, with a full-scale conference to be held in December 1997. Finally, we agree that intensive tobacco-control policies are our first line of prevention of lung cancer and that former smokers are desirable participants for future chemoprevention trials.
Authors: G. S. Omenn, M. D. Thornquist
Institution: Fred Hutch Cancer Center
Language: English
Subfield: Biochemistry
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4237258276