Reply to Goulden et al.: Risk of Bias in Study of Varenicline and Cardiovascular Outcomes.