Greenness exposure and all-cause mortality during multi-drug resistant tuberculosis treatment: A population-based cohort study.

Background: Living closer to greenness were thought to benefit various health outcomes. We aimed to assess the association between residential greenness and mortality among patients undergoing multidrug resistant tuberculosis (MDR-TB) treatment.

Methods: We enrolled all local MDR-TB patients reported in Zhejiang, China from 2009 to 2017 and followed them throughout the treatment. We calculated the contemporaneous normalized difference vegetation index (NDVI) in the 250 and 500 m radius around patient's residence. Cox proportional hazards regression models with time-varying NDVI were used to assess the impact of greenness exposure on all-cause mortality during MDR-TB treatment, adjusting for potential individual and contextual covariates.

Results: We ascertained 1,621 active MDR-TB cases, which contributed 3036 person-years at risk with an average follow-up of 684 days (s.d. 149 days) per patient. Among them, there were 163 deaths during follow-up, representing a crude mortality rate of 537 deaths per 10,000 person-years. Patients exposed to the second quintile (Q2) of greenness within the 500 m buffer had around 64% reduced mortality risk over the lowest quintile of greenness with hazard ratio (HR) = 0.364 (95% CI: 0.109-1.22). In lower nighttime light (NTL) areas, the hazard ratios (HR) per quintile increase in NDVI within the 500 m buffer were Q2: 0.35 (95% CI: 0.10-1.18), Q3: 0.24 (95% CI: 0.09-0.66), Q4: 0.26 (95% CI: 0.10-0.69), and Q5: 0.26 (95% CI: 0.10-0.71) relevant to the lowest quintile Q1, with a trend of p-value ≤0.01. Patients who were female, younger (<60 years), resided in urban areas, or had high PM (i.e. particles with diagram ≤2.5 μm) exposure were more likely to benefit from greenness exposure. Associations were neither observed with NDVI in the 250 m buffer nor for patients living in higher NTL areas. There was a non-linear exposure-response relationship between greenness and deaths with p-value ≤0.05.

Conclusion: Increasing greenness exposure along with medical treatment reduces all-cause mortality among patients living in lower NTL areas.