Objective  To evaluate the medium and long-term treatment effects of tuberculosis (TB) based on the observation of the relapse of TB after successful treatment in newly registered patients.

  Methods  Based on the national tuberculosis information management system, all the registered TB patients registered in a county in 2010 were selected as the observation cohort, and the follow-up for the cohort was conducted during 2010−2019. Kaplan-Meier method was used for univariate analysis on the relapse, and timedependent Cox regression model was used for multivariate analysis.

  Results  The record of TB relapse started in the 2^nd year after successful treatment, the relapse rate reached the peak of 4.46% in the 3^rd year, remained ＞3.00% in the 4^th−6^th year, and then decreased year by year to 0.23% in the 10^th year. Men were more likely to have TB relapse and more relapses occurred in age group 50 years in all the years. The relapse TB cases were mainly farmers. The Kaplan Meier method and Cox proportional risk regression model analysis showed that older age, being farmer, being bacterial positive for the first time, non-active medical care seeking and diagnosis delay were the risk factors for the relapse of TB after successful treatment in newly registered patients. The average interval before the relapse was longest after the first registration, then the interval gradually decreased.

  Conclusion  TB prevention and control is a challenging project. If it is feasible, follow-up for newly diagnosed TB patients with high risk for relapse after successful treatment should be strengthened and, furthermore, life cycle health management in whole population can be conducted based on Health China Initiative.