Background:
Although Quitlines have provided services for cessation with easy accessibility, convenient counseling, and high effectiveness, little attention has been paid to research on predictors of delayed relapse.

Methods:
Among the data of 45,368 quit attempts from 1st January 2007 to 31 December 2015, only 16,729 data which is the first quit attempt of each user aged more than 19 years old and with the complete data on baseline and followed up until 31 December 2012 were included in the final analysis.
At the baseline, information on socio-demographic characters, smoking related behaviors, and motivational characteristics were collected. Undergoing the Quitline protocol, self-reported cessation status and reasons of relapse have been asked during the 1 year of follow up. Cumulative abstinence and the effect size of related factor on relapse were estimated by Kaplan-Meier method and Multinominal logistic regression.

Results:
In the end of prospective follow up, 80.0% of participants were in relapse, and 15.5% of them were relapsed 4 weeks after starting their quit. The most common reason for late relapse was the stress caused by social context of daily living. Subjects induced to smoke when they drink coffee or alcohol (OR 1.359, 95% CI 1.218-1.516) or when they saw others' smoking scenes (OR 1.422, 95% CI 1.193-1.695) were more relapsed with the reason of temptation of environment, and subject who were induced to smoke when they were under stress (OR .479, 95% CI 1.289-1.696) were more relapsed with the reason of stress caused by social context of daily living than withdrawal symptom. Both reasons of relapse were also more frequent in higher education and more frequent drinking group, while lesser in high nicotine dependence and low self-efficacy groups.

Conclusions:
Development of relapse prevention program regarding the meaningful portion of late relapse and its' social environmental predictors is in need for effective Quitline service.