Cardiovascular
Eleven-year counterfactual risk of cardiovascular hospitalization under hypothetical sustained reductions in post-traumatic stress disorder symptom severity: the WTC-Heart Cohort, New York State, 2012–2023 Sean Andrew MacAllister* ALFREDO MORABIA MORABIA MORABIA MORABIA MORABIA MORABIA MORABIA MORABIA Barry Commoner Center, Queens College, City University of New York
Sustained reductions of post-traumatic stress disorder (PTSD) symptom severity may reduce the risk of cardiovascular disease (CVD). Randomized controlled trials are not well suited to assess the long-term effects of PTSD symptom duration (“load”) on risk for myocardial infarction (MI) and stroke, given the prolonged follow-up and large sample sizes required. To address this gap, we use longitudinal observational data to emulate key features of an hypothetical trial of sustained PTSD symptom reduction.
The WTC-Heart study is a prospective cohort initially comprising 6,481 first responders involved in World Trade Center debris cleanup. The cohort was assembled in 2012–2013 and followed through July 2025, with 85% retention. PTSD symptoms and conventional CVD risk factors were assessed repeatedly over follow-up. Incident CVD events were self-reported and validated through medical record review or, through 2023, ascertained via linkage to the New York Statewide Planning and Research Cooperative System (SPARCS). Dates and causes of death were confirmed through linkage with the National Death Index.
Among 5,446 New York State residents, we will estimate the 11-year counterfactual risk of hospitalization for MI (n=342), stroke (n=80), or both CVD events under hypothetical sustained reductions in PTSD severity, using repeated assessments of PTSD symptoms (PTSD Checklist; PCL). We will apply the parametric g-formula to estimate: (1) the counterfactual CVD risk under sustained reductions in PTSD severity of 10%, 20%, and 100%, compared with the natural course; and (2) the causal effect of reducing cumulative PTSD exposure by one PTSD-year (severity-time units) on CVD risk. Analyses will adjust for baseline and time-varying sociodemographic, cardiometabolic, trauma-related, and mental health confounders. We will report absolute risk differences and risk ratios at 11 years and present standardized cumulative incidence curves comparing counterfactual risks with the observed natural course.
Under standard assumptions, including no unmeasured confounding, correct model specification, and missing at random covariate data, the parametric g-formula will provide estimates of CVD risk under full adherence to sustained hypothetical reductions in PTSD severity. These findings have potentially major public health implications, given the high prevalence of PTSD among first responders to an increasing number of human-made or climate-related disasters.