Takotsubo Syndrome and Air Pollution

Background

Takotsubo syndrome (TTS) is an acute and reversible form of myocardial injury characterized by typical regional wall motion abnormalities in the absence of culprit epicardial coronary artery disease frequently precipitated by significant emotional stress or serious physical illness. The clinical presentation is usually similar to acute myocardial infarction (MI), with chest pain and/or dyspnea, ST-segment elevation or depression and/or T-wave inversion on the resting electrocardiogram (ECG) and elevation of serum cardiac troponin. Although previously considered a benign disease, it is now clear that TTS is associated with severe acute complications during the acute phase including hemodynamic and electrical instability and up to 5% of in-hospital mortality.

Notably, despite substantial research, the risk factors and pathophysiological mechanisms of TTS are not completely understood, with several hypotheses proposed but none offering a comprehensive explanation. Catecholamine-induced myocardial injury is likely to play a central role, as an emotionally or physically triggering event precipitating the syndrome can be identified in most cases and TTS has been associated with conditions of catecholamine excess (e.g.: pheochromocytoma, central nervous system disorders) and activated specific cerebral regions. Similarly, a markedly reduced parasympathetic activity has been reported during the acute phase of TTS. Moreover, recent studies reported abnormalities in both the functional structure and activity in the areas of the brain related to both emotions and the sympathetic nervous system including the basal ganglia, the hippocampus, the amygdala, and the insula, supporting the role of limbic system dysfunction as a potential mechanism in patients with TTS. Furthermore, it has been proposed that the impaired cardiac function could be the result of an acute coronary microvascular dysfunction with impaired microvascular perfusion leading to a demand-supply mismatch and an ischemic stunning. Therefore, the risk factors for endothelial dysfunction would predispose to the occurrence of TTS.

Air pollution is a complex mixture of unwanted particulate and gaseous material released into the environment by human activities and the world's fourth leading cause of disease and death. Of interest, accumulating evidence supports a consistent relationship between increased exposure to air pollution and CV diseases such as MI and heart failure. Urban ambient air pollution, in particular combustion-derived PM, has received the greatest scientific attention due to the high density of urban populations and increasing levels of traffic-derived emissions and urbanization of societies worldwide. PM includes both organic and inorganic particles (e.g.: dust, pollen, soot, smoke, liquid droplets) and is categorized according to the aerodynamic diameter into coarse particles (2.5-10 μm in diameter; PM10), fine particles (<2.5 μm in diameter; PM2.5), and ultrafine particles (<0.1 μm in diameter; PM0.1). The smallest particles, such as PM2.5 and PM0.1, may contribute disproportionately to the CV toxic effects due to their large reactive surface area and their ability to penetrate deeply into the alveoli and potentially directly into the bloodstream, causing damage and dysfunction of various tissues and cells far from the lung. Gaseous pollutants, in particular nitric dioxide (NO2), ozone (O3), carbon monoxide (CO) and sulphur dioxide (SO2), have been linked to increased morbidity and mortality from CV diseases, likely in an additive manner to PM2.5, but data are still scarce and frequently inconsistent. Mechanistically, the pathogenetic mechanism of air pollution toxicity on the CV system includes oxidative stress, systemic and vascular inflammation, endothelial dysfunction, autonomic and neuroendocrine disruption, metabolic alterations, transcriptional and epigenetic reprogramming. Furthermore, the acute responses to short-term (hours) air pollution exposure include sympathoadrenal activation, release of circulating inflammatory biomarkers, alterations of endothelial function, and acute vascular modifications, such as arterial vasoconstriction and impaired vascular reactivity. The acute effects of air pollution are even more significant in the context of chronic long-term (years) exposure. Indeed, chronic air pollution exposure, by promoting the development of a vulnerable systemic state, can exponentially increase the risk of acute CV events that are likely to be precipitated by acute variations in air pollution exposure. Notably, the investigators recently demonstrated that the exposure to higher concentrations of air pollutants (especially PM2.5) is associated with the presence of vulnerable plaque features and with plaque rupture as a mechanism of coronary instability assessed by optical coherence tomography (OCT) and, moreover, with an enhanced systemic and plaque inflammatory activation. In addition, the investigators also demonstrated that a higher exposure to PM2.5 and PM10 in patients with myocardial ischemia and non-obstructive coronary artery disease is associated with coronary vasomotor abnormalities, and PM2.5 is an independent risk factor for the occurrence of epicardial spasm and MINOCA as clinical presentation.

Of interest, even though the pathogenetic mechanism of air pollution are likely to predispose to the occurrence as well as to mediate a worse clinical presentation and outcome of TTS, the relationship between air pollution and the risk of TTS as well as its clinical course has never been assessed. Furthermore, given the poor understanding of the underlying pathophysiology, there is a lack of evidence-based prevention strategies as well as interventions to reduce the incidence as well as the acute complications of TTS.

Upon this background, the investigators hypothesized that:

1. the exposure to higher levels of air pollutants in the days (short-term exposure) or years (long-term exposure) before the TTS onset could be considered as a risk factor for the occurrence of TTS;
2. the exposure to higher levels of air pollutants in the days (short-term exposure) or years (long-term exposure) before the TTS onset could be associated with a worse clinical course in terms of in-hospital complications and mortality;
3. the exposure to higher levels of air pollutants in the days (short-term exposure) or years (long-term exposure) before the TTS onset could be associated with a worse prognosis at follow-up in terms of major adverse cardiovascular events (MACE) and/or TTS recurrence.

In this context, as largely reported throughout the last ten years, time-stratified case-crossover studies represent the best model strategy. In fact, in this study design, instead of comparing exposure between people experiencing a TTS (case) and people who did not (control), pollutant concentrations before TTS diagnosis (case period) were compared with other randomly selected periods, when the subject has not experienced TTS yet (control periods). As such, each patient represents its own control. Hence, time-independent confounders, such as age, comorbidities, and smoking status are controlled by the case-crossover study design, whilst variables that change between case and control time periods are possible confounders (e.g., weather, temperature …). Stratifying by year and month is adequate for most studies and can be done as well by day of the week and temperature.

Primary objective

The primary objective of the study is to assess the relationship between either short-term or long-term exposure to increased levels of air pollutants (PM10, PM2.5, O3, NO2, benzene [C6H6], SO2 e CO) and onset of TTS.

Secondary objectives

• To assess the relationship between either short-term or long-term exposure to increased levels of air pollutants (PM10, PM2.5, O3, NO2, benzene [C6H6], SO2 e CO) and a higher rate of in-hospital complications in patients with TTS.
• To assess the relationship between either a short-term or long-term exposure to increased levels of air pollutants (PM10, PM2.5, O3, NO2, C6H6, SO2 e CO) and a worse clinical outcome at follow-up in patients with TTS.

Study design

Ambispective observational case-crossover pilot study.

Study duration

The study will last 48 months from the approval of the present protocol by the local ethics committee. Enrolment will last 18 months, which will serve also to retrieve the retrospective data from the internal archives. The follow-up period will last 24 months and 6 will be dedicated to the data analysis and interpretation and drafting of scientific reports.

Air pollution data collection

The exposure of patients to air pollution compounds in the two years prior to the occurrence of TTS will be analysed. The investigators will assess: PM10, PM2.5, O3, NO2, C6H6, SO2 e CO. Residential addresses will be obtained from medical records. Annual average 24-h of pollutants levels will be measured matching each individual's home address, obtained through hospital file archive, and the "ArpaLazio" website (http://www.arpalazio.net/main/aria/sci/basedati/chimici/chimici.php). This website provides the concentration values of the following pollutants monitored by the regional automatic network since 1999 and are available for download, expressed as a concentration in micrograms per cubic meter (µg/m3): NO, NO2, NOx, PM10, PM2.5, O3, CO, C6H6, SO2. Hourly data are available for all gaseous pollutants, while the levels of PM10 and PM2.5 are expressed on a daily basis. In addition to the elementary data, the following standard calculations are available: daily averages, typical day, monthly averages and annual averages. The control units are localizable by latitude and longitude coordinates (Sampling Point of Latitude and Longitude) expressed in decimal degrees (DD) and approximated up to the 15th decimal digit, as well as uniquely identifiable by an alphanumeric code (Station Location ID). Data will be obtained from the air quality monitor closest to each participant's residence that was active for the entire year, and short-term (daily and weekly) and long-term (annual) air pollution exposure will be quantified as daily, weekly, and annual average 24-h pollutants level of measurements before TTS. In particular, daily exposure will be assessed as the average 24-h exposure to pollutants the day of TTS onset (0-day lag time between exposure and TTS), the day before (1-day lag time between exposure and TTS), 2 days before (2-day lag time between exposure and TTS), 3 days before (3-day lag time between exposure and TTS), 4 days before (4-day lag time between exposure and TTS), 5 days before (5-day lag time between exposure and TTS), 6 days before (6-day lag time between exposure and TTS) and 7 days before (7-day lag time between exposure and TTS). Similarly, weekly exposure will be assessed as the average 24-h exposure to pollutants the week of TTS onset (0-week lag time between exposure and TTS), the two week before (1-week lag time between exposure and TTS), the three weeks before (2-week lag time between exposure and TTS), and the four week before (3-week lag time between exposure and TTS). Finally, long-term exposure will be assessed as the annual average 24-h pollutants level of measurements of 2 years before TTS. Of note, this methodology to assess air pollution has been extensively validated and used in previous studies. Only patients with 2 years or more of available data on air pollution exposure prior to TTS will be included.

Sample size calculation

To the best of our knowledge, there is no previous literature focusing on air pollutants effects in Takotsubo syndrome (TTS), hence this reflects as a pilot study. As such, no formal sample size calculation is needed. Common rules of thumb for pilot studies by Browne (1995) refer to 30 subjects as a minimum sample size. Based on data retrievable from the archives since 2016 and on an estimated 30 subjects/year diagnosed with TTS at our Unit, the investigators plan to enroll 250 patients in 18 months, which is consistent with the sample size simulation study for case cross-over studies with Dupont's formula, which would imply a sample of at least 195 subjects.

Statistical analysis

The sample will be described in its demographic, anthropometric, clinical, and instrumental, variables through descriptive statistical techniques. In-hospital mortality will be defined as the number of actual (observed) deaths over the total of patients enrolled. In depth, qualitative variables will be expressed by absolute and relative percentage frequencies. Quantitative variables, indeed, will be reported either as mean and standard deviation (SD) or median and interquartile range (IQR), respectively in the case they were normally or not normally distributed. Their distribution will be previously assessed by the Shapiro-Wilk test.

Multiple imputation will be applied to handle missing data, by "imputeR" R package, by using Lasso regression methods centered on the mean for what concerns quantitative variables, whilst classification trees strategy centered on the mode, i.e. the most frequent class, will be applied don qualitative data.

Between groups differences in the demographic, clinical and laboratory features will be assessed by the Chi Square or the Fisher's exact test as for qualitative variables (with Freeman-Halton's extension when appropriate). Differences in the quantitative variables will be instead evaluated either by the Student's t test or the Mann- Withney U test, according to their distribution. Between-groups most significant differences will be further graphically represented by means of "violin plots" drawn with the aid of the R packages "ggstatsplot", "ggpubr" and "ggplot2".

This study will use a time-stratified design to control for time trend and other short-term varying confounders like weather and temperature, as it compares exposure levels between same weekdays/ within each month of each year. Exposure on days where TTS occurred (case days) will be compared with the exposure on days where TTS did not occur (control days). Control days will be chosen on the same day of the week earlier and later in the same month in the same year. Daily TTS counts approximately follow Poisson distribution, hence a linear model of conditional Poisson regression with time-stratified case-crossover design will be fitted to estimate the short- and long-term exposure of air pollutants and TTS risk, applying as potential confounders' parameters such as temperature and weather. Relative Risk (RR) and 95% confidence intervals (CIs) will be estimated based on the per 1 μg/m3 increase of air levels. In the exploratory analysis, the natural cubic splines with three degrees of freedom for temperature, relative humidity and weather will be introduced in new models to examine the nonlinear effect (i.e. the potential confounding by meteorological conditions) and Akaike's Information Criterion will be used to choose the best model. Relative risk increase (RRI) was estimated by RR-1. The RRI in for TTS per 10 μg/m3 increase of PM levels will be calculated as follows: RRI% = exp (β * 10) - 1 * 100%, where β is the exposure-response coefficient from conditional Poisson regression combined under the time-stratified case-crossover design, which refers to a unit increase in PM pollutants. Single-pollutant models will be further implemented to discern the effects of air pollution. When estimating the effects of some potential effect modifiers, time-stratified analyses will be further provided, linking with various subgroups (e.g., by season (warm season: April-September, and cold season: October-March), applying the above analyses for these subgroups. The statistic differences from stratified analyses (e.g., the difference between male and female) will be estimated by Z-test. The same model will be applied also on secondary endpoints of hospitalization rates and MACEs. All analysis will be conducted using R version 4.0.4 with gnm package for conditional Poisson regression combined under the time-stratified case-crossover design.

All statistical tests with p values of < 0.05 will be considered as statistically significant. P-values between 0.05 and 0.10 will be also reported as suggestive. All analyses will be performed by using R software (v. 4.2.0, R Core Team, Vienna, Austria, 2022).