Most smokers who contacted the research center met eligibility criteria and were enrolled. Most enrollees participated successfully (, ).
Baseline participant characteristics.
Depiction of participant recruitment, eligibility assessment, allocation to treatment conditions and disposition.
Logistic regression of data from the primary outcome, 10-wk continuous abstinence, yielded a significant three-way interaction between nicotine dose, dependence and quit-success genotype score (P = 0.015), exponentiated coefficient = 0.56 (95% CI = 0.35–0.91). The interaction noted in interim analysis of the initial set of subjects also was noted in separate analysis of data from the second half of the subject group (data not shown). In the entire group of subjects, the two-way interaction of nicotine dose and FTND score also was significant (P = 0.02), exponentiated coefficient = 1.78 (95% CI = 1.10–2.87). displays the rate of 10-wk continuous abstinence as a function of nicotine dose, dependence level and genotype score. The 42-mg nicotine dose condition tended to enhance treatment outcome for highly dependent smokers with low v1.0 scores (P = 0.06). By contrast, this high-dose patch impeded abstinence for less dependent smokers with low v1.0 scores (P = 0.009). The nicotine dose × dependence × genotype score interaction also was significant for 4-wk continuous abstinence at the end of treatment (P = 0.02, exponentiated coefficient = 0.62, 95% CI = 0.42–0.93), a secondary outcome (Supplementary Figure S2). Moreover, the 6-month-point abstinence results showed the same pattern of three-way interaction (P = 0.03, exponentiated coefficient = 0.62, 95% CI = 0.40–0.96); see Supplementary Figure S2.
Smoking abstinence as a function of nicotine patch dose, dependence and quit-success genotype score.
There were concordant results from analyses that separated European American and African-American smokers. For smokers with European ancestry (n = 369), the 10-wk continuous abstinence outcome showed a nicotine dose × dependence × v1.0 genotype score interaction (P = 0.05, exponentiated coefficient = 0.60, 95% CI = 0.35–1.00). With respect to 4-wk continuous abstinence and 6-month-point abstinence, the subsample showed a similar, though nonsignificant, trend (P = 0.12, exponentiated coefficient = 0.71, 95% CI = 0.46–1.09; P = 0.19, exponentiated coefficient = 0.72, 95% CI = 0.45–1.17). For smokers with African ancestry (n = 88), this three-way interaction displayed a trend similar to that noted for the entire sample for both 10-wk (P = 0.15, exponentiated coefficient = 0.33, 95% CI = 0.07–1.48) and 4-wk continuous abstinence (P = 0.08, exponentiated coefficient = 0.38, 95% CI = 0.13–1.12). The three-way interaction did reach significance in this sub-sample for 6-month-point abstinence (P = 0.02, exponentiated coefficient = 0.20, 95% CI = 0.05–0.78).
Abstinence outcomes also were predicted by individuals’ abilities to reduce smoking during the 2-wk precessation period, as measured by decreases in end–expired air CO levels. Smokers whose CO dropped by more than the median (55.6% decrease) showed a substantially higher rate of 10-wk continuous abstinence than smokers who did not show this decrease: 37.6 % versus 11.2% (P < 0.001, odds ratio = 4.77, 95% CI = 2.81–8.11; for results from each treatment group, see ). A similar result was obtained for 4-wk continuous abstinence (52.1% versus 26.5%; P < 0.001, odds ratio = 3.01, 95% CI = 1.97–4.60) and 6-month-point abstinence (40.2% versus 18.9%; P < 0.001, odds ratio = 2.89, 95% CI = 1.83–4.57).
Continuous 10-wk smoking abstinence as a function of end–expired air CO reduction during the 2-wk precessation period.
However, the precessation decrease in CO did not explain all of the influences on quit success provided by other measures of individual differences. The three-way interaction of nicotine dose, FTND score and quit-success genotype score remained statistically significant, even after inclusion of a term in the logistic regression model that reflected CO decreases (P = 0.04 for 10-wk abstinence).
Treatment was generally well tolerated. Three percent of participants withdrew from the study because of adverse effects. The most frequently reported patch-related adverse effects were: vivid dreams (83.5%), itching or burning (55.3%), insomnia (47.2%), rash (20.7%), nausea (15.9%), headache (26.1%), and vomiting (2.5%). There was a trend toward more frequent nausea in the 42-mg dose condition (P = 0.06). Each of two women who received 42-mg doses reported syncope on one occasion, each in conjunction with nausea or vomiting; neither reported any sequelae.