Abstract Background: In 2015, tuberculosis (TB) infection resulted in 1.8 million deaths worldwide. Reducing the delay in initiating treatment for TB is often cited as an approach to reduce TB related deaths. However, there is limited reliable evidence available to support this. Aim: To determine if delay in initiating treatment for pulmonary TB leads to an increased risk of all-cause mortality for cases resident within Yorkshire and Humber.Methods: The cohort consisted of Yorkshire and Humber residents notified with pulmonary TB from 2000 to 2015. Patients were followed up from date of notification until date of death or final censoring date of 2 August 2016. Time to treatment was measured as the number of days from symptom onset to treatment start date, and its association with all-cause mortality assessed using flexible parametric survival models. After examining the martingale residuals a log transformation of time to treatment was undertaken.Results: The final cohort consisted of 4458 individuals of whom 1068 (24%) died during the follow-up period. The median time to treatment for the cohort was 72 days (IQR: 35 - 237). After adjustment for age, sex, social risk factor, previous diagnosis of TB and deprivation, a 13% reduction in risk of TB was observed for every unit increase in time to treatment following a [ln](x+0.1) transformation, although this was not statistically significant (HR: 0.87; 95% CI: 0.76-1.01).Conclusions: The results suggest a delay in treatment initiation was not associated with a change in all-cause mortality. This finding should be substantiated in the national dataset, which would allow focus on more recent data and inclusion of co-morbidities. Such an analyses has the potential to identify whether reducing treatment delay reduces the associated burden of pulmonary TB infection.