An assessment of the cardiovascular health of autistic individuals

Date of Approval: 
2020-11-13 00:00:00
Lay Summary: 
People with autism often have difficulties with social interaction and communication, which could impact on the frequency and quality of their healthcare visits. Further, recent research suggests that autistic people may be more likely to have a sedentary lifestyle, obesity, and mental health conditions (such as, anxiety and depression). Autistic women also often suffer from hormonal conditions, like polycystic ovarian syndrome (PCOS), which is associated with increased hormone levels and insulin resistance throughout their lives. This is particularly relevant for Covid-19-associated mortality, which significantly impacts males (who also have high testosterone throughout their life) and people with a history of heart conditions, obesity, and diabetes. Since the number of adults diagnosed with autism has increased, more research is needed in order to estimate the physical health risks to this population and to design appropriate, preventative guidelines. Using data from GP and hospital visits, we aim to calculate how often autistic people experience heart-related symptoms (e.g. angina), heart disease (e.g. heart attacks), death due to heart complications, or death due to Covid-19, and compare this to people without autism. To make sure our conclusion is not affected by causative factors other than autism, we will also take lifestyle conditions into account (e.g. smoking, socioeconomic status, age etc.). In addition, we will conduct the same analysis in mothers of autistic people, since recent data has shown that they are also more likely to have PCOS, abnormal hormone levels during pregnancy, and increased risk of other serious conditions (e.g. cancer).
Technical Summary: 
We propose to use the Clinical Practice Research Datalink (CPRD), linkages to the Hospital Episode Statistics (Admitted Patient Care data), Covid-19 linkages, and Office of National Statistics (ONS) Death data linkage to consider the cardiovascular health of autistic individuals, and its relationship to cardiovascular conditions and severe disease from Covid-19. We will use clinical, referral, therapy, and test records to determine whether individuals have increased glycosylated haemoglobin (HbA1C), HDL cholesterol, Non-HDL:HDL cholesterol ratio, BMI, use of medications (e.g. statins, beta blockers, ACE inhibitors, diuretics, and calcium-channel blockers), as well as whether or not individuals are more likely to exhibit hypertension, angina, arrhythmia, transient ischaemic attack (TIA), coronary revascularization procedures, and/or die from stroke, myocardial infarctions, heart failure, or Covid-19. Cohort 1 will consist of autistic individuals and will be matched 5:1 with non-autistic controls on age, sex, and practice (Cohort 2). We will use Cox Regression to determine risk of cardiovascular outcomes, as well as cause of death due to cardiovascular conditions or Covid-19, and composite risk scores will be investigated for association to autism, via linear regression (Pearson’s coefficient). We will covary for several confounding factors, and will use the Patient-Level Indices of Multiple Deprivation (IMD) linkage to control for socioeconomic status. Missing data on covariates will be imputed via use of Multiple Imputation by Chain Equations (MICE). Mixed effects modelling will be utilized to account for nested data structure. Statistical significance of the results will be adjusted with the application of Benjamini-Hochberg False Discovery Rate. We will use the Mother-to-Baby link to identify non-autistic mothers of Cohort 1; they will make up Cohort 3 and will be matched on age, sex, and practice to the mothers of Cohort 2 (Cohort 4). We will use the same procedure to test the same risk factors and outcomes in this population.
Health Outcomes to be Measured: 
Risk Score: HbA1C; HDL cholesterol; Non-HDL:HDL cholesterol ratio; BMI Cardiovascular treatments: Prescriptions for statins, beta blockers, ACE inhibitors, diuretics, and calcium-channel blockers Cardiovascular Signs & Symptoms: angina; hypertension; arrhythmias; transient ischaemic attack (TIA) Cardiovascular Outcomes: ischemic stroke; myocardial infarction; heart failure, coronary revascularization procedures Cause of Death: Cardiovascular outcomes (listed above) and Covid-19
Application Number: 

Simon Baron-Cohen - Chief Investigator - University of Cambridge
Elizabeth Weir - Corresponding Applicant - University of Cambridge
Adriana Cherskov - Collaborator - Yale University
Alex Tsompanidis - Collaborator - University of Cambridge
Carrie Allison - Collaborator - University of Cambridge
Rupert Payne - Collaborator - University of Bristol

CHESS (Hospitalisation in England Surveillance System);CPRD Mother-Baby Link;HES Admitted Patient Care;ICNARC (COVID-19 Intensive Care National Audit and Research Centre);ONS Death Registration Data;Patient Level Index of Multiple Deprivation;COVID-19 Linkages