Climate change and its associated environmental stressors (e.g., increased frequency of heat waves, hurricanes, etc.) is a public health crisis (Watts et al., 2018). Without rapid, drastic reduction to global emissions, global heating is projected to exceed 1.5°C within 20 years and to surpass 2° within 30–40 years (Skea et al., 2022). Extreme heat events have already increased in frequency, and this trend is expected to continue (Intergovernmental Panel on Climate Change (IPCC), 2022). Given the evidence of the cardiovascular effects of temperature (Madrigano et al., 2013; Lavados et al., 2018; Li et al., 2018; Vered et al., 2020; Rowland et al., 2021) and the relationship between mental health and cardiovascular disease (Edmondson et al., 2012), incorporating temperature into the field of behavioral health is critical (Atwoli et al., 2021). Such research can help protect people, especially those experiencing cardiovascular disease such as acute coronary syndrome (ACS) from the impacts of climate change and inform health care in a rapidly-changing climatic regime.

Climate change-related temperature stressors have adverse implications for individuals with cardiovascular disease. In addition to associations of hotter temperatures with adverse outcomes such as incident ischemic stroke (Lavados et al., 2018; Li et al., 2018; Vered et al., 2020; Rowland et al., 2021) and risk for mortality after myocardial infarction (MI; Madrigano et al., 2013), warmer months and extreme heat events associated with a range of mental health outcomes, including more completed suicides (Galvão et al., 2018) and hospital admissions for schizophrenia (Shiloh et al., 2005), emergency psychiatric evaluation (Carlsen et al., 2019), and mania (Medici et al., 2016). ACS is already a highly distressing event, and posttraumatic stress symptoms (PTSS) after ACS are common (Edmondson et al., 2012). Because PTSS after ACS can double risk for adverse clinical outcomes (Edmondson et al., 2012), hotter temperatures may have additional, indirect associations with cardiovascular prognosis due to increased psychological distress. Finally, other temperature-related indicators of climate change also carry adverse implications for health, including associations of greater temperature variability with increased risk for incident MI (Rowland et al., 2021) and unusually cold weather with incident cardiovascular events (e.g., ischemic heart disease, hemorrhagic stroke; Madrigano et al., 2013; Li et al., 2018). Cold temperatures may influence adverse mental health outcomes as well (Carlsen et al., 2019), though this potential relationship is less established.

The impact of temperature stressors related to climate change on health is additionally distributed according to geographic differences in socioeconomic status (SES). Low SES is associated with greater disease burden, including cardiovascular conditions such as MI and stroke (Min et al., 2017), such that these populations are more susceptible to the health impacts of climate change-related stressors. Indeed, people with pre-existing cardiac conditions experience worse outcomes when exposed to extreme hot or cold temperatures (e.g., greater mortality risk for people with atrial fibrillation when exposed to extreme heat; Zanobetti et al., 2013). Some studies have also found that low SES is a risk factor for developing psychological distress, including PTSS, after large-scale disasters (e.g., earthquake and hurricane; Norris et al., 2002; Foa et al., 2006; Paxson et al., 2012; Hrabok et al., 2020) or following stroke (Bruggimann et al., 2006; Goldfinger et al., 2014), though research regarding the link between SES and PTSS broadly is more mixed. The exposure to warming, and the adverse health effects of warming, are more pronounced for individuals living in areas with lower SES (Zanobetti et al., 2013; Gronlund, 2014; Benmarhnia et al., 2015), and SES disparities in heat exposure are apparent even within small, urban geographical areas (Wong et al., 2016; Schinasi et al., 2018; Santamouris, 2020) like New York City (Rosenthal et al., 2014; Pioppi et al., 2020), an area where temperature increases have accelerated during recent decades (Horton et al., 2015; Storms, 2015) and heat waves have become more frequent (Depietri and McPhearson, 2018). Low SES additionally impedes an individual’s ability to cope with extreme temperatures (e.g., reduced access to air conditioning; Rosenthal et al., 2014; McDermott-Levy et al., 2021) and a review found that sleep disruption due to the effects of climate change – a known risk factor for adverse cardiovascular outcomes (Cappuccio et al., 2011), including re-hospitalization post-ACS (Romero et al., 2020) and PTSS following ACS and stroke (Rojas et al., 2018; Romero et al., 2020) – is most pronounced in low SES populations (Rifkin et al., 2018). Taken together, these findings suggest that low SES is both a multidimensional risk factor for exposure to extreme temperatures and a potential effect modifier, increasing exposure to climate change, reducing ability to cope with climate change, and increasing susceptibility to its effects due to pre-existing health disparities. Thus, the uneven distribution of morbidity and mortality secondary to a warming climate has great potential to further exacerbate health disparities.

In part due to a lack of large datasets that combine geospatial, medical, and psychosocial assessments, and despite the documented association between indicators of climate change and mental health, pathways relating temperature to cardiovascular health outcomes via psychological distress remain unexplored. Yet, psychological distress is a plausible mechanism that may explain the association of temperature with patient outcomes. Patients experiencing a cardiovascular event, such as ACS often develop posttraumatic stress symptoms (PTSS) secondary to the ACS experience (Edmondson et al., 2012). A recent meta-analysis found that as many as 13–20% of ACS patients develop clinically significant symptoms; this is non-trivial, because patients with elevated PTSS have doubled risk for cardiac event recurrence and mortality (Edmondson et al., 2012).

There is already theoretical support for a pathway from temperature to cardiovascular health via distress. The Enduring Somatic Threat model posits that attention to ongoing, internal cues, such as rapid heartrate when exercising, serve as ongoing reminders of the ACS, which can cause significant psychological distress (Edmondson, 2014). Exposure to extreme temperature triggers physiological symptoms like rapid heart rate (Ravanelli et al., 2015), which could potentiate greater somatic threat and increased risk for incident PTSS. Finally, while risk of PTSS appears to increase for those with low SES after large-scale disasters (e.g., earthquake; Norris et al., 2002; Foa et al., 2006); whether this extends to more subtle exposures, like temperature extremes, is uncertain.

The present study extends prior work on climate, SES, and cardiovascular health by asking whether climate vulnerability and socioeconomic vulnerability interact as causal effects (Hernán, 2018) of incident PTSS in patients evaluated for ACS. Utilizing a sample of patients evaluated for ACS at an urban academic medical center in New York City, we hypothesized the following:

Because people living in low SES areas may also be less equipped to deal with changes in temperature, we further hypothesized that:

A total of 1,741 participants enrolled in REACH, 1,436 of whom provided valid addresses in New York City and were eligible for this study. Of these participants, 1,195 were matched by date and location to temperature data and 956 provided data for the present analysis (missing outcome, n = 54; n = 0; missing SES, n = 3; missing other [e.g., Charlson], n = 182). The median cardiac-induced PTSS score was 20.0 (IQR = 10.0; Range 17.0–76.0), and 17.7% of the sample met the clinical cutoff for a probable diagnosis of PTSD (score ≥ 33). Full demographic details are available in Table 1, correlations of temperature variables are in Table 2. Histograms for census tract-level SES, temperature variables, and PTSS at 1 month are in Supplementary Figures S1A–F.

This study contributes to our understanding of the multifaceted impacts of climate change on health by testing potential associations of temperature and temperature variability with PTSS secondary to an ACS 1 month after ED evaluation. Results showed no evidence for the association of temperature with PTSS; however, living in a census tract with greater socioeconomic vulnerability was associated with significantly more PTSS symptoms.

Temperature extremes and variability resulting from the current climate crisis has increased cardiovascular risk for individuals with ACS (Madrigano et al., 2013; Lavados et al., 2018; Li et al., 2018; Vered et al., 2020; Rowland et al., 2021). Given the association of temperature with adverse mental health outcomes (Shiloh et al., 2005; Medici et al., 2016; Galvão et al., 2018; Carlsen et al., 2019) and cardiovascular symptoms (Ravanelli et al., 2015) that may serve as traumatic reminders of the ACS event (Edmondson, 2014), may also experience increased risk for PTSS secondary to the ACS. Because PTSS doubles risk for cardiac event recurrence and mortality (Edmondson et al., 2012), this may be one pathway through which individuals with ACS experience poor outcomes. Low SES may compound this risk, given geospatial inequities in the distribution of climate burden (Rosenthal et al., 2014; Wong et al., 2016; Schinasi et al., 2018; Pioppi et al., 2020; Santamouris, 2020; McDermott-Levy et al., 2021) and cardiovascular disease (Min et al., 2017), as well as a potential association of SES with PTSS following large-scale disasters (Norris et al., 2002; Foa et al., 2006).

The finding that census tract-level SES is associated with ACS-induced PTSS deserves greater attention. Literature linking SES to PTSS has been somewhat mixed and it can be hard to disentangle order effects (i.e., what came first – a diagnosis or lower SES?) (Koenen et al., 2017). This study disambiguates order effects by examining PTSS secondary to an acutely traumatic medical event at baseline enrollment; thus, all PTSS are necessarily newly acquired. It also shines a spotlight on health disparities by showing that, not only is socioeconomic vulnerability associated with a greater burden of cardiovascular disease (Min et al., 2017), but it also predicts the development of psychological trauma after cardiovascular events such as ACS. Thus, PTSS may be one mechanism of the association between low SES and adverse cardiovascular outcomes (Edmondson et al., 2012). Because exploratory interaction analyses suggested that this relationship is primarily true for patients with true ACS (v. those who had suspected ACS only), intervening early with individuals experiencing ACS who reside in areas with high socioeconomic vulnerability may be warranted to prevent psychological distress and its medical sequelae. A few promising interventions include those that utilize cognitive behavioral therapy or meaning making, but additional research is needed to confirm effectiveness and scalability (Birk et al., 2019). However, early intervention fails to address the root causes of socioeconomic disparities, and addressing disparities should be a priority.

Counter to our hypotheses, we did not observe association of temperature or any aspect of temperature variability with PTSS secondary to the ACS event; neither did interactions of temperature with census tract-level SES emerge. It is possible that there is simply no causal relationship between temperature and the development of these symptoms. However, it is also possible that the present study was unable to detect effects because they are small (and therefore the study may have been underpowered) or the timing of the assessments did not match the causal mechanisms as they unfold in real time. For example, the Enduring Somatic Threat model (Edmondson, 2014) would be better tested by examining the association of hotter days or hours with momentary symptoms and related intrusions (e.g., the somatic symptom of rapid heart rate during an extremely hot day (Ravanelli et al., 2015) might lead to heart-focused attention and intrusive, upsetting reminders of the ACS event). Future research should test this possibility. Indeed, even research on the association of temperature with event recurrence has highlighted the fine-grained associations of hourly temperature as a trigger for cardiovascular events such as stroke (Rowland et al., 2021). Other factors, such as changing locations or home air conditioning, may also have prevented the detection of temperature effects.

This study had numerous strengths, including integration of geospatial, medical, and psychosocial assessments to test novel and theory informed hypotheses with public health implications. PTSS is common after ACS and has significant adverse health consequences (Edmondson et al., 2012), so understanding the incidence of PTSS in this population has significant practical implications for improving patient health and well-being. There were also limitations. Participants were drawn from New York City and results may not generalize to other cities or to non-urban areas. Individuals living in other areas of the country may also be more vulnerable to other environmental stressors associated with climate change, such as catastrophic wildfires, with implications for both cardiovascular health (e.g., smoke exposure, and particulate matter concentration; Haikerwal et al., 2015; Reid et al., 2016; Jones et al., 2020; Chen et al., 2021) and PTSS (e.g., due to loss of life or property; Zhang et al., 2022). Incorporating these additional environmental stressors will be critical for conducting a more comprehensive assessment of the intersecting impacts of climate change and low SES for mental and physical health. Furthermore, although this study was completed prior to the COVID-19 pandemic, consideration of vulnerability and susceptibility to disease crises associated with both cardiovascular health (Li et al., 2020; Madjid et al., 2020; Richardson et al., 2020; Wu and McGoogan, 2020) and SES (Khanijahani et al., 2021) will be of interest in future work.

Given the marginal interactions of census tract-level SES and temperature variability metrics with ACS status, a larger sample comprising more patients with true ACS is also warranted. A larger sample size would also allow for an examination of potential effect modification by ACS type. Effect modification by other factors related to the ACS (e.g., delay to care receipt, in-hospital intervention) should also be tested in these samples, given that health disparities in ACS exist along the care continuum (Simon and Ho, 2020). Temperature data were keyed to participants’ home addresses, but participants likely spent time in other areas that may have had varying temperatures (e.g., walking in Central Park or along the river, running errands in midtown) and home temperatures, including access to fans and air conditioning, were not assessed. More detailed assessments of actual temperature exposure are warranted. Our ability to detect significant effects may also have been limited due to a lack of extreme temperatures within the dataset. For example, one would not expect extreme heat events to commonly occur during spring or fall months; similarly, variability during these months may be less problematic if all temperatures are relatively mild. Restricting the sample to only comprise participants who enrolled during summer or winter months would have severely limited the sample size, so this was not a viable option. It is also not clear that this would address the issue because temperature metrics were averaged over 31 days, such that a single extreme event would be less influential. Residuals in models predicting PTSS were not normally distributed, which can violate model assumptions in small samples. That said, results replicated across a wide range of sensitivity analyses, including models that treated PTSS as a binary score using an established cutoff for a suspected diagnosis of clinically significant PTSS. Finally, this is a prospective observational study, thus, estimates of causal effects may be confounded due to unmeasured factors. We are explicit about the goal of estimating causal effects to lend clarity to the conduct of our study and analyses (Hernán, 2018).

The present study is among the first to link census tract-level SES to increased risk of developing PTSS secondary to evaluation for an acute cardiac event. Exploratory analyses suggested that this association occurs primarily among patients with confirmed ACS (v. those who ultimately did not receive an ACS diagnosis). Counter to hypotheses, temperature, and temperature variability were not associated with PTSS, nor did temperature variables appear to modify the relationship between SES and PTSS. These null findings may be due to imprecision in temperature estimates (e.g., not capturing in-home access to air conditioning, different locations throughout the day), a relatively small sample size, or a lack of any true causal effect, among other reasons. Because residing in low SES areas is associated with greater disease burden and increased morbidity and mortality, psychological interventions to prevent PTSS in at-risk patients experiencing ACS may be warranted. We should also prioritize protecting people residing in low SES areas from extreme temperatures and other climate change-related exposures.

The datasets presented in this article are not readily available because they contain geospatial information. Requests to access the datasets should be directed to tmc2184@cumc.columbia.edu.

The studies involving human participants were reviewed and approved by Columbia University Irving Medical Center Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.

TC: conceptualization, formal analysis, methodology, writing – original draft, and writing – review and editing. JC: methodology, supervision, and writing – review and editing. AJ: data curation, methodology, and writing – review and editing. SR: methodology and writing – review and editing. DE: funding acquisition, investigation, supervision, and writing – review and editing. All authors contributed to the article and approved the submitted version.

This work was supported by the National Institutes of Health Science of Behavior Change Program through an award administered by the National Institute on Aging (https://www.nia.nih.gov/; U24AG052175; PI, DE). Cornelius receives support from the National Institutes of Health National Center for Advancing Translational Sciences (https://ncats.nih.gov/; KL2TR001874). Just receives support from the National Institute of Environmental Health Sciences (https://www.niehs.nih.gov/; P30ES023515 and R01ES031295; PI, AJ). Rowland receives support from the National Institute of Environmental Health Sciences (https://www.niehs.nih.gov/; R01ES028805). The content is solely the responsibility of the authors and does not necessarily represent the official view of the NIH.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.