My ATTUD blog in April (“Is Varenicline More Effective than NRT?” – now on ATTUD website) concluded that varenicline was probably more effective than single NRT but whether it was better than dual NRT was unclear. Since then a new study and a new meta-analysis on this question have appeared that you may have heard about.
The study was a non-randomized study (Kralikova et al, Addiction, preprint available via internet) of Czech smokers attending a specialized smoking cessation clinic. The study compared those who chose to use varenicline (n=519), single NRT (n= 211) or dual NRT (e.g. patch + gum; n=125). Varenicline did better than both single NRT (OR = 1.4) and dual NRT (OR = 2.2). The major assets of the study were its real-world setting (e.g. smokers had to pay for meds) and its ability to statistically adjust for many differences in varenicline vs. NRT users. The major liability is that, unlike almost all prior studies, it did not find that dual use NRT was better than single NRT; thus, it was not surprising that varenicline outperformed dual NRT.
A new Cochrane meta-analysis (Cahill et al, www.cochranelibrary.org) published in 6/13 used a new method called “network meta-analysis” to better directly compare the efficacy of different treatments. That analysis found that varenicline was more effective that three single NRTs (ORs = 1.4, 1.5 and 1.7) but not more effective than dual NRT (OR = 1.1). It also found no evidence of increased psychiatric or cardiac events due to varenicline.
In summary, the analysis with the greatest external validity (i.e. generalizability) suggests varenicline is more effective than dual NRT but the analysis with the greatest internal validity (e.g. highest compliance) did not find this.