Recently, Carroll and colleagues (1) concluded that filter ventilation was not associated with biomarkers of exposure (BOE) or biomarkers of potential harm (BOPH). When one assesses exposure and harm from different cigarette types, one usually consider results not only in terms of harm to the smoker who uses them, but also on a per cigarette basis. The analyses reported, which did not control for the number of cigarettes smoked per day (CPD) in their filter ventilation groups Q1 to Q4, do not provide the essential information on exposure per cigarette.From the results in Table 3 of the article, we could generate trend probability values quite similar to those presented. We then adjusted for the mean CPD per filter ventilation group shown in Table 2, Ci (i = 1, 2, 3, 4) in various ways, by analyzing loge (Bi/Ci), (loge Bi)/Ci, or loge (Bi)/loge (Ci) where the Bi indicate the biomarker values in the ventilation groups (Supplementary Materials and Methods). On the basis of the results adjusted for age, sex, race, education, smoking duration, smoking status, quit effort, and CPD, we consistently found clear negative trends, often highly statistically significant, of increasing ventilation with decreasing levels of all five BOPH except IL6, and with some BOE, notably total nicotine equivalents (Supplementary Results, Tables 1–4).As we do not have available the individual person biomarker data, we could only conduct approximate analyses. However, we believe our conclusions—that most BOPH and some BOE are reduced per cigarette—would hold were precise analyses conducted, and we ask the authors to report such analyses as they have all the data available. While of course their original conclusion is important, and the accuracy of their analysis undisputed, it is also important to be aware of exposure and harm on a per cigarette basis. Adding this significant covariate into the multivariate analysis allows for more direct comparison between the groups, and shows the interesting contrast with the published results.In conducting such analyses, Carroll and colleagues should realize that the PATH study, on which their analyses were based, includes erroneous CPD data for some individuals, due to a confusion between packs per day and CPD. Thus, CPD in one individual was reported as 10 at Wave 2, 15 at Wave 3, and 300 at Wave 1, the data being entered as 15 packs per day rather than 15 CPD. Such errors need correction before conducting analyses.P. Lee reports other support from Japan Tobacco International S.A. during the conduct of the study; and consultancy with other tobacco companies. J. Fry reports other support from Japan Tobacco International SA during the conduct of the study; and consultancy with other tobacco firms.See the Response, p. 1450
Read full abstract