Mortality in Persons with Autism Spectrum Disorders: A Danish Population-Based Cohort Study

Background: Previous studies consistently report elevated mortality in persons with autism compared to the general population, but these findings are limited to older cohorts, small sample sizes or lack of recent follow-up. Mortality patterns have not been investigated in large, population-based birth cohorts that reflect the recent dramatic autism prevalence increase and with follow-up to the present time.

Objectives: Conduct a population-based longitudinal cohort analysis of mortality in individuals diagnosed with autism spectrum disorders (ASD) born from 1980 and followed through 2012, considering age, sex, ASD diagnosis, and pre- and perinatal factors.

Methods:  The study cohort was comprised of all Danish births from 1980 through 2010 (N=1,934,672 births) identified in the Danish Medical Birth and Central Person Registers and study information for each cohort member was obtained through multiple national health registry linkages of individual-level data via a unique personal identifier.  ASD and select other developmental outcome diagnoses in the child and parent psychiatric history diagnoses were obtained from the Danish Psychiatric Central Research Register; medical comorbidity diagnoses were obtained from the Danish National Patient Register; perinatal and birth information were obtained from the Danish Medical Birth Register; family demographics and mortality information were obtained from the Danish Central Person Register; cause of death was obtained from the Cause of Death Register.  Mortality curves were derived using Kaplan-Meier methods. Hazard ratios of the relative risk of death in ASD versus non-ASD were estimated using Cox regression methods.

Results: By the end of follow-up (31 December 2012), 18,380 cohort children (1%) were diagnosed with ASD and 5187 (0.3%) were diagnosed with childhood autism, specifically. Fifty-eight deaths occurred in persons diagnosed with ASD, 15 deaths in persons with childhood autism.  The relative risk of death was generally 2-fold higher or more in persons with ASD than non-ASD overall, and by sex, age, and perinatal or parent characteristics. Cause of death will also be reported.

 Conclusions: Compared to their birth counterparts without ASD, persons born since 1980 with ASD are at increasing risk of death beginning in youth through early adulthood overall and the elevated risk is evident across a number of individual and parent characteristics present at birth. Understanding the contributions of long-term comorbidity on mortality risk is essential to guiding lifelong health care provision for individuals with ASD.