Introduction: Disparate outcomes have been reported in patients (pts) with myeloproliferative neoplasms (MPNs), such as increased risk of thrombotic complications in Black pts with cellular-phase MPNs and poorer survival outcomes in myelofibrosis (MF) when incorporating census-tract measures of disadvantage (Khan et al, CLML 2016; Palmer et al, ASH 2024). The impact of environmental exposures in pts with myeloid malignancies is being increasingly studied and one potential driver of disparities may be air pollutant exposure. Proposed mechanisms for the impact of air pollution on myeloid malignancies include oxidative stress, epigenetic modifications, and global inflammation (Bhattarai et al, Exp Mol Med 2024). Our group previously investigated the impact of air pollutant exposure on a cohort of pts with acute myeloid leukemia (AML) (Palmer et al, Blood Advances 2024) and aimed to study the impact of air pollutant exposure on MF. We investigated the putative impact of environmental air pollutant exposure in pts with MF with regards to disparate exposure levels within our cohort and potential impact on disease characteristics along with survival outcomes.

Methods: Adult pts with MF seen at participating institutions from 2011 onwards in the Chicagoland area were included in this retrospective study. Pollution data (including 1,3-butadiene, benzene, diesel particulate matter (PM), polycyclic aromatic hydrocarbons and polycyclic organic matter (PAHPOM), and PM 2.5 µm, and Ozone) from the Center for Disease Control National Environmental Public Health Tracking Network and averaged estimated pollution from the National Air Toxics Assessment for years 2014 and 2017-2019, were joined with pt census tracts. Kruskall-Wallis rank sum tests and pairwise Wilcoxon rank sum tests were used to evaluate pollutant exposures by racial/ethnic groups. Kaplan-Meier survival analysis was used to compare overall survival between pts in the most and least polluted census tracts for each pollutant. Dynamic International Prognostic Scoring System Plus (DIPSS +) and pollution was evaluated using multinomial logistic regression.

Results: Our dataset contained 533 pts joined with tract-level pollution data. The median age at diagnosis was 65 yrs. Our data contained 71% non-Hispanic White (NHW), 17% non-Hispanic Black (NHB), 6% Hispanic, and 6 % non-Hispanic Other (NHO) pts. NHB pts were exposed to more Diesel PM, Benzene, 1,3-butadiene, PAHPOM, and PM 2.5 relative to NHW pts (p<0.05). Hispanic pts were more exposed to Benzene, Diesel PM, and PM 2.5 relative to NHW (p<0.05). NHO pts were exposed to more 1,3-butdiene, Diesel PM and PAHPOM relative to NHW (p<0.05). NHW pts were exposed to more ozone than other race/ethnicities (p<0.05). We found no statistically significant relationship between measured air pollutant exposure and DIPSS + scores. We evaluated median overall survival (mOS) between the most (quartile 4, Q4) and least (Q1) polluted tracts. mOS for 1,3-butadiene was 13.8 yrs vs 8.9 yrs in Q1 vs Q4 respectively (p=0.04). mOS for benzene was 13.8 yrs vs 9.5 yrs in Q1 vs Q4 respectively (p=0.21). mOS for Diesel PM was 13.8 yrs vs 9.5 yrs in Q1 vs Q4 respectively (p=0.06). mOS for PAHPOM was not reached in Q1 and was 7.4 yrs in Q4 (p=0.02). mOS for PM 2.5 was 13.8 and 6.6 yrs for Q1 vs Q4 respectively (p<0.01). Conversely, mOS for Ozone was 7.4 vs 13.8 for Q1 vs Q4 respectively (p=0.02).Conclusions: We found that pts who are not NHW are disproportionately exposed to environmental air pollutants except ozone. While pollutant exposure was not significantly related to DIPSS+ score, we found pts living in the most polluted census tracts for 1,3-butadiene, PAHPOM, and PM 2.5 had significantly decreased survival relative to those in the least polluted quartiles. Interestingly, NHW pts had greater exposure to ozone and pts in tracts most polluted with ozone had greater survival. Air pollution may be a significant driver of disparate outcomes in myelofibrosis. Further analysis is necessary to disentangle the effects of other social determinants on these outcomes. Pollutant misclassification from our data source may limit the ability to detect true associations. Multivariable analyses adjusting for other potential geospatial risk factors (tract-level segregation, disadvantage, affuence, etc.) along with age and gender will be conducted to better understand the impact of specific pollutant exposures with survival.

This content is only available as a PDF.
2025
Sign in via your Institution