Mental Illness in Relation to Cardiovascular Disease Mortality and Out-of-Hospital Cardiac Arrest Characteristics and Survival

Abstract

Abstract Background People with mental illnesses have a 10–20 years shorter life expectancy than the general population. This mortality gap is primarily attributed to the increased risk and worse outcome of cardiovascular disease (CVD) among individuals with mental illnesses. Socioeconomic factors and healthcare accessibility, combined with systemic deficiencies in healthcare infrastructure and new policies aimed at reducing healthcare utilisation, may have exacerbated these disparities. Despite efforts to prioritise equitable care delivery, stigma and a lack of training among healthcare providers about mental health conditions persist, contributing to the ongoing challenges of addressing this issue. Although the link between mental illnesses and CVD is well established, the relationship between mental illnesses and out-of-hospital cardiac arrest (OHCA) characteristics and outcomes has been comparatively underresearched. The majority of OHCAs have an underlying cardiac aetiology, and hence, risk factors for CVD, such as depression, can influence the probability of experiencing OHCA and surviving it. A better understanding of how mental illnesses affect OHCA outcomes may help explain part of the overall observed lower life expectancy in patients with mental illnesses. Therefore, this thesis aims to explore the mortality gap between individuals with and without mental illnesses, with a specific focus on OHCA as a potential contributor to this gap. Methods In this thesis, I conducted a systematic literature review and three data analysis projects using different datasets and methods. In the first research project, I used data from the Scottish Mental Health Inpatient and Day Case Scottish Morbidity Record

dataset to identify individuals who were 18 years of age or older and admitted with depression, as coded by the International Classification of Diseases (ICD)-10 F32– F33, between 2000 and 2019. I subsequently linked this dataset with death records to ascertain the causes of death. To estimate standardised mortality ratios (SMRs), I calculated the person-time at risk for people with severe depression by sex, 5-year age group, and calendar-year group. I estimated SMRs among individuals with severe depression using the whole Scottish population as the standard. To identify gaps in the current research on the relationship between mental illnesses and OHCA, I conducted a systematic review search of studies that looked into this topic, starting from the inception of relevant databases and ending on December 11, 2021, with an update on February 9, 2024. The search objective was to identify observational studies reporting on OHCA incidence, characteristics, or outcomes among individuals with mental illnesses. I and a second reviewer (another PhD student) independently screened titles and abstracts for eligibility. Both reviewers then independently assessed the full texts of potentially eligible articles against the eligibility criteria and independently conducted the data extraction. Following these initial investigations, I conducted two retrospective cohort studies using OHCA event data to further explore the survival and characteristics of OHCA in individuals with depression in one study and with unscheduled hospital admissions where a mental illness was recorded in the other study. In the first study, I used data from the ARREST (Amsterdam Resuscitation Studies) registry, spanning the period from 2008 to 2018. I extracted relevant comorbidity information, including depression, from patients' general practitioner records. OHCA survival was defined as survival 30 days after the OHCA or hospital discharge. To estimate the association between history of depression and OHCA survival, I used logistic regression modelling.

Additionally, I explored the mediating effects of initial heart rhythm and bystander cardiopulmonary resuscitation (bCPR) provision. In the second study, I linked the Scottish Ambulance Services (SAS) OHCA dataset with the Unscheduled Care Data Mart (UCD). The UCD includes patient demographics, mortality data, emergency and urgent care service use across Scotland, and unscheduled hospital admission records to acute and psychiatric hospitals from January 1, 2011 onward. The analysis excluded individuals below 18 years of age, those with traumatic OHCA incidents, and those without community health index numbers. I classified exposed patients based on whether they had a documented history of an unscheduled hospital admission where a mental health illness was recorded (ICD-10 codes F00–F99) or had previously experienced unscheduled psychiatric hospital admissions before the OHCA event. I assessed the association between unscheduled hospital admission record of mental illness and 30-day survival following OHCA using both unadjusted and adjusted logistic regression models. I performed subgroup analyses to examine this association across some demographic and clinical strata, including age groups, sex, socioeconomic status, year of arrest, and prevalent types of mental illnesses in the study population. Results In the initial study, I found that among 28,808 individuals with severe depression, 27.4% died over a median follow-up of 8.7 years, with all-cause relative mortality being over three times higher than expected. Circulatory diseases were the leading causes of death, with excess mortality particularly pronounced for cerebrovascular disease. Both males and females with severe depression experienced higher all -cause and cause-specific mortality compared to the general population, with deaths from suicide having the highest SMR.

For the literature review, the full search yielded 11,380 studies, with 10 meeting inclusion criteria (eight retrospective cohort studies and two nested case-control studies). Three of the included studies focused on depression, while seven encompassed various psychiatric conditions. People with mental illnesses were more likely to have arrests in private residences, unwitnessed arrests, more comorbidities, less bCPR, and non-shockable initial heart rhythms. The few studies that were identified also reported a higher OHCA incidence risk and lower survival rates in people with mental illnesses compared to people without mental illnesses. The analysis of the ARREST dataset included 5,594 OHCA cases, of which 582 individuals had pre-existing depression. These patients exhibited less favourable patient and resuscitation characteristics and lower odds of survival, even after adjusting for age, sex, and comorbidities. Initial heart rhythm and bCPR partially mediated the observed association, but they did not fully account for it. The final study, which used national Scottish data, initially identified 36,775 OHCA patients that were attended by SAS. After excluding those who were not eligible, the study included 30,523 patients with OHCA. Among the included cases, 12.8% had a history of an unscheduled hospital admission where a mental health illness was recorded. Those with a record of a mental illness had substantially lower 30-day survival odds compared to those without. This association persisted across various demographic and clinical subgroups. Conclusion This thesis addresses shortcomings in our understanding of the mortality gap between individuals with and without mental illnesses, particularly focussing on OHCA as a contributing factor. The findings show that patients with depression have higher

mortality rates, mainly due to circulatory diseases. A systematic review highlighted the scarcity of research on OHCA among people with mental illnesses. Subsequent studies showed that patients with depression in the Netherlands and those with an unscheduled hospital admission record of mental illness in Scotland had lower OHCA survival rates, regardless of event characteristics or comorbidities. This thesis also discusses common barriers to pre-hospital and mental health research, emphasising the necessity for improved data linkage and comprehensive national datasets. The association between mental illness and lower OHCA survival persisted after adjusting for key OHCA characteristics that could potentially explain the link, suggesting that this association may be independent of these factors. Future research should further investigate the underlying mechanisms to help reduce premature mortality in people with mental illness, including factors not discussed in this thesis, such as the potential impact of psychotropic medications on OHCA incidence and outcomes.