Longitudinal relationship of particulate matter and metabolic control and severe hypoglycaemia in children and adolescents with type 1 diabetes.

ABSTRACT

BACKGROUND: Evidence for the metabolic impact of long-term exposure to air pollution on diabetes is lacking. We investigated the association of particulate matter <10 µm (PM10) and <2.5 µm (PM2.5) with yearly averages of HbA1c, daily insulin dose (IU/kg) and rates of severe hypoglycaemia in type 1 diabetes (T1D). METHODS: We studied data of 44,383 individuals with T1D < 21 years which were documented in 377 German centres within the diabetes prospective follow-up registry (DPV) between 2009 and 2018. Outcomes were aggregated by year and by patient. PM10-and PM2.5-yearly averages prior to the respective treatment year were linked to individuals via the five-digit postcode areas of residency. Repeated measures linear and negative binomial regression were used to study the association between PM-quartiles (Q1 lowest, Q4 highest concentration) and yearly averages of HbA1c, daily insulin dose and rates of severe hypoglycaemia (confounders sex, time-dependent age, age at diabetes onset, time-dependent type of treatment, migratory background, degree of urbanisation and socioeconomic index of deprivation). RESULTS: Adjusted mean HbA1c increased with PM10 (Q1 7.96% [95%-CI 7.95-7.98], Q4 8.03% [8.02-8.05], p-value<0.001) and with PM2.5 (Q1 7.97% [7.95-7.99], Q4 8.02% [8.01-8.04], p < 0.001). Changes in daily insulin dose were inversely related to PM (PM10 and PM2.5 Q1 0.85 IU/kg [0.84-0.85], Q4 0.83 IU/kg [0.82-0.83], p < 0.001). Adjusted rates of severe hypoglycaemia increased with PM-quartile groups (PM10 Q111.2 events/100 PY [10.9-11.5], PM10 Q4 15.3 [14.9-15.7], p < 0.001; PM2.5 Q1 9.9 events/100 PY [9.6-10.2], PM2.5 Q4 14.2 [13.9-14.6], p < 0.001). DISCUSSION: Air pollution was associated with higher HbA1c levels and increased risk of severe hypoglycaemia in people with T1D, consequently indicating a higher risk of diabetes complications. Further studies are needed to explore causal pathways of the observed associations.