Cardiovascular Safety Warnings and the Prescribing of Varenicline

To the Editor The cardiovascular safety of the smoking cessation drug varenicline is controversial. Initial concerns arose following a signal observed in a 2010 randomized controlled trial conducted in patients with cardiovascular disease.1 This led to the launch of a safety review2 by the Food and Drug Administration in June 2011 and a labeling change in July 2011 indicating a potential increased risk in patients with cardiovascular disease.3 At the time, the Food and Drug Administration and the European Medicines Agency concluded that varenicline’s benefits outweighed any risk in this population.3,4 An increased risk was also reported in a July 2011 meta-analysis of randomized controlled trials, which included all populations.5 Although this signal was not replicated,6,7 it is unclear if these concerns impacted clinical practice. Our objective was to examine the impact of these safety concerns on the prescribing patterns of varenicline. We used data from the Clinical Practice Research Datalink (CPRD).8 This study was approved by the Research Ethics Board of the Jewish General Hospital and the Independent Scientific Advisory Committee of the CPRD (protocol number 13_045RA). Interrupted time-series analysis with segmented autoregression compared the prescription rates of varenicline pre- (January 2007–June 2011) and post-warning (July 2011–December 2012). The denominator for all prescription rates was all patients aged ≥18 years registered at a CPRD practice. The fitted model estimated the immediate impact of the safety warning and the predicted rate that would have been observed in the postwarning period in its absence. To determine if the warning resulted in a change in the characteristics of patients who were prescribed varenicline, we compared the characteristics of first-time users in the year before the safety warning to those in the year after, restricting to patients with ≥1 year of CPRD observation time. We also examined the prescribing patterns of two other smoking cessation drugs (bupropion and nicotine replacement therapies). To examine the impact of the early increase in prescription rate that occurred upon varenicline’s regulatory approval, we repeated our analyses excluding data before September 2007. The observed and predicted prescription rates of varenicline are shown in the Figure. In the prewarning period, the estimated prescription rate increased by 27.4 (95% confidence interval [CI]: 17.0, 37.7) prescriptions per month per 100,000 person-years (PYs), resulting in a rate of 2,464 prescriptions per 100,000 PYs in June 2011. The warning resulted in an immediate drop in the prescription rate of 513.4 (95% CI: −1,020.9, −5.9) prescriptions per 100,000 PYs, representing a 21% decrease in the prescription rate (R2 = 0.82). In the postwarning period, there was a change in direction in the trend, with the prescription rate decreasing by 42.4 prescriptions per month per 100,000 PYs (95% CI: −92.9, 8.1) during this period. There were no important differences in patient characteristics before or after the warning (data not shown).FIGURE: Observed and predicted prescribing rate of varenicline in the United Kingdom before and after the cardiovascular safety warning. The autoregression analysis of order 12 was used to account for positive correlation between data values.Our bupropion analysis revealed no difference in the observed and predicted prescription rate (R2 = 0.94; data not shown), with no immediate decrease or postwarning rate change. Similarly, no difference was present for nicotine replacement therapies (data not shown). Our analysis excluding data before September 2007 revealed a similar but attenuated pattern for varenicline (prewarning rate: 14.4 [95% CI: 7.2, 21.6] prescriptions per month per 100,000 PYs; immediate change: −279.4 (95% CI: −469.8, −89.0) prescriptions per 100,000 PYs; postwarning rate: −30.8 (95% CI: −46.9, −14.6) prescriptions per month per 100,000 PYs; R2 = 0.77). Safety concerns regarding the cardiovascular effects of varenicline were associated with a decrease in its prescription in the United Kingdom. Although the labeling change was specific to cardiovascular patients, this decrease was independent of patient characteristics. The decreased rate was not accompanied by an increase in the prescription of other cessation drugs. Although evidence suggests that the benefits of varenicline outweigh any increased cardiovascular risk,3 this warning was sufficient to alter clinical practice. Kristian B. Filion Maria Eberg Laurent Azoulay Center for Clinical Epidemiology Lady Davis Institute, Jewish General Hospital McGill University Montreal, QC, Canada [email protected]

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.