Combination Therapy of Varenicline With Nicotine Replacement Therapy Is Better Than Varenicline Alone

A Systematic Review and Meta-analysis of Randomized Controlled Trials

Ping-Hsun Chang; Chien-Hsieh Chiang; Wei-Che Ho; Pei-Zu Wu; Jaw-Shiun Tsai; Fei-Ran Guo

Disclosures

BMC Public Health. 2015;15(689) 

In This Article

Results

Study Identification

The search strategy retrieved a total of 63 citations. Of these, three were appropriate for full-text review (Fig. 1). No additional study was identified after searching the reference lists of these three articles. After review, all three studies[9–11] were finally included in the systematic review and meta-analysis (Table 1). These three studies were all combination therapy versus varenicline alone therapy trials.

Figure 1.

Flow diagram of meta-analysis: inclusion and exclusion of studies

Characteristics of Included Studies

Table 1 summarizes the baseline demographics, number of participants, interventions, and outcome measurements of included studies. All three trials recruited smokers who were aged 18 and over, not breastfeeding or pregnant, and had no current psychiatric or other serious illness. The inclusion and exclusion criteria were well described in two studies,[9,11] requiring that the enrolled patients have no recent experience of other cessation medication or successful abstinence. The other study provided relatively simple criteria that volunteers seeking treatment with no contraindications could be enrolled.[10] The mean age of participants was similar among all studies. One study included more female subjects than males.[9] There were 38.3 % males in this study vs 66.7 and 57.8 % in the other two studies respectively. The Fagerström Test for Nicotine Dependence (FTND) score was higher in one study because it included smokers who smoked 20 or more cigarettes per day,[11] whereas the other two studies included participants with lower daily cigarettes consumption. Treatment interventions differed among these studies. One study administered trial patch two weeks before the TQD, while the other two studies started patch use on the TQD. Two studies used a 15 mg/16 hours patch, while the other[11] used a 21 mg/24 hours patch. On the other hand, the use of varenicline was similar among the studies. All started at 0.5 mg per day one week before TQD, with increase to 2 mg/day on TQD, and continued for 12 weeks. One study[9] tapered the dose of varenicline on the 13th week. All studies provided concurrent behavioral counseling during the treatment phase. One study[10] measured the early outcome at 4 weeks; the other two measured it at 12 weeks. All studies used exhaled carbon monoxide to confirm continuous abstinence. For the late outcome, one study[9] measured 9 to 24 weeks of continuous abstinence, one[11] measured 2 to 24 weeks of continuous abstinence, and the other one did not measure outcome after treatment phase. One study[10] measured self-reported point prevalence at 12 weeks, which was not included in our meta-analysis. All studies adopted our definition of early outcome as the primary outcome, while our defined late outcome as the secondary outcome.

Study Quality and Publication Bias

The Jadad score of included studies ranged from 4 to 5. Only one study was granted a score of 4 because the dropout of patients was not detailed in the article.[10] The overall quality of the included studies was high. The funnel plot of early outcome was not symmetrical due to lack of smaller studies with positive effect. However, the power of the funnel plot was compromised by the small number of studies. The publication bias assessed by Begg's rank correlation test and Egger's regression test was not significant (p = 0.30 and 0.28 respectively). The publication bias of late outcome was not assessed because there were only two studies included. The Comprehensive Meta-Analysis software could not generate a funnel plot or a test report under this situation.

Varenicline Plus Nicotine Patch Versus Varenicline Plus Placebo Patch: The Early Outcome

Three studies with a total of 904 participants were included in this meta-analysis (Fig. 2). The heterogeneity was minimal (I 2 = 0 %, p = 0.41), therefore fixed effect model was used. The results demonstrated a significant increase in abstinence rate (44.4 % vs 35.1 %, OR = 1.50, 95 % CI 1.14 to 1.97). All three studies showed a favorable effect of combination therapy, while only one of them reached statistical significance.[9]

Figure 2.

Varenicline plus nicotine patch vs varenicline plus placebo patch: the early outcome

Varenicline Plus Nicotine Patch Versus Varenicline Plus Placebo Patch: The Late Outcome

Two studies with a total of 787 participants were included in this meta-analysis (Fig. 3). The heterogeneity was moderate (I 2 = 54 %) without statistical significance (p = 0.14). The fixed effect model was also used. The results demonstrated a significant increase in abstinence rate (32.4 % vs 23.1 %, OR = 1.62, 95 % CI 1.18 to 2.23). Both studies showed a favorable effect of combination therapy, while only one of them reached statistical significance.[9]

Figure 3.

Varenicline plus nicotine patch vs varenicline plus placebo patch: the late outcome

The Safety of Combination Therapy

The case numbers of adverse events were aggregated. The pooled ORs were generated by fixed effect model ( Table 2 ). Combination therapy reported more nausea (28.4 % vs 25.7 %), insomnia (18.7 % vs 15.4 %), abnormal dreams (13.6 % vs 10.7 %), but less headache (7.1 % vs 7.8 %). There were no significant differences between nicotine and placebo patch groups. Only one study reported the adverse events of depression, especially in the nicotine patch group (2.3 % vs 1.4 %; p = 0.50).[9] This study also reported more skin reactions in the nicotine patch group (14.4 % vs 7.8 %; p = 0.03).

A total of eight serious adverse events (SAEs) were reported in the included studies, where only one of them was considered relevant to the study medications. This was a female participant who became pregnant during the treatment phase where she was randomized to receive placebo patch combined with varenicline. She later on gave birth to an infant with Down syndrome and congenital heart defects. Another SAE was a female participant who also became pregnant during treatment phase with placebo patch. She had an anembryonic pregnancy and this SAE was considered to be irrelevant to the study medications.

Subgroup and Sensitivity Analysis

We did not perform subgroup analysis because of the small number of studies, and there was no significant heterogeneity. One RCT was identified to be different in study design (pre-cessation treatment with patch) and participant characteristics (more females than males).[9] When we eliminated this RCT from the meta-analysis model of the early outcome, the favorable effect of combination therapy became insignificant (OR = 1.28, 95 % CI 0.87 to 1.87). We then preformed sensitivity analysis of the late outcome in the same manner and the favorable effect of combination therapy also diminished (OR = 1.26, 95 % CI 0.79 to 2.00). We compared the results of fixed effect to random effects model of both early and late outcomes. The OR was 1.50 (95 % CI 1.14 to 1.97) vs 1.50 (95 % CI 1.14 to 1.97) in the early outcomes, 1.62 (95 % CI 1.18 to 2.23) vs 1.61 (95 % CI 1.00 to 2.58) in the late outcomes respectively. The conclusion of favorable effect persisted even using different models.

processing....