Was the FDA correct in updating its warning of increased CV events with Chantix/Champix?

Reports of neuropsychiatric AEs prompted Pfizer to focus Champix/Chantix on smokers at high risk if they do not quit.  This cohort includes those with CV disease.

Rigotti et al (2010) initially reported efficacy vs placebo but also a non-significant elevation in myocardial infarction (2.0% vs 0.9%).

Singh et al (2011) subsequently reported a meta analysis in which CV risk remained (1.06% vs 0.82%). However, specific CV events could not be identified and were possibly driven by Rigotti’s study.

Rigotti’s study prompted an FDA safety alert while Pfizer’s responded to the meta analysis.  At the time Pfizer stated it was discussing with the FDA its own meta analysis using more reliable MACE to measure risk. This study has now been reported. 

Pfizer's meta analysis reports higher rates of MACE (0.21→0.31%). Although non-significant, the FDA has chosen to issue an updated safety review and update the product label.

We question the FDA’s approach in this instance.  The Agency’s duty is to inform of risk, however it has issued the update in the absence of significance of risk.  While it justifies actions by reporting similar risk when analysed many different ways it fails to report this expanded analysis.  

We also note data were censored at 30d.  Although consistent with many CV studies, we ask whether 12 or 24wks would have been a better cut reflecting varenicline treatment duration 

[Source: UpdatesPlus-Nicotinic Receptors and Addictive Disorders]