Cardiac responses to environmental heat exposure in young and older adults

Abstract

Older individuals are at a greater risk for adverse cardiovascular events during extreme heat exposure. However, detailed characterization of their cardiac responses to environmental heat exposure are lacking. In 20 young (18 to 39 years) and 20 older (> 65 years) adults (50% male in both groups), we document the echocardiography assessed left-ventricular responses to a very hot and dry (DRY, 47°C and 15% RH) and hot humid (HUMID, 41°C and 40% RH) 3-hour heat exposure, with intermittent bouts of light physical activity throughout. In both climates and in both age groups, heat stress a) increased cardiac output by ~0.7 ± 0.8 L/min, b) decreased stroke volume by ~7 ± 10 mL and c) augmented diastolic function through increased atrial contribution to filling by 5 ± 5%. In the DRY climate, mitral annual systolic velocity (s’) increased to a greater extent in older subjects (Δ3.3 ± 2.1 vs Δ1.5 ± 1.5 cm/s, p = 0.002), with less difference in HUMID (Δ2.1 ± 1.3 cm/s vs 1.4 ± 1.3 cm/s, p = 0.096). Despite these adjustments, systolic blood pressure was only maintained in the younger group and fell consistently in older individuals (0 ± 8 mmHg) in DRY (Δ-11 ± 14 mmHg vs 1 ± 8 mmHg, p = 0.001) and HUMID (Δ-9 ± 15 mmHg vs −1 ± 8 mmHg, p = 0.030). In summary, older adults rely on a greater augmentation of systolic function during extreme heat exposure, but the magnitude of depends on the heat stress severity.

New and noteworthy

Comparing healthy young and older adults, we assessed left ventricular cardiac function (using echocardiography) during two separate 3-hour extreme heat exposures in a very hot dry or hot humid climate type. While the augmentation of diastolic function and cardiac output were similar between age groups, older adults showed a greater increase in mitral annular systolic velocity with heat exposure, indicating a stronger reliance on systolic mechanisms to maintain stroke volume.

Graphical Abstract

Introduction

Heat waves can cause mass morbidity and mortality in older adults and other vulnerable populations (1). Retrospective analyses utilizing ICD codes indicate that most heat-related deaths among older people are cardiovascular in origin (e.g., heart attacks, strokes) (2). However, ICD classifications provide only an estimated cause of death, as most heat-related fatalities occur outside of hospital settings without autopsy confirmation (1). Therefore, high-resolution physiological studies are critical to accurately determine how the heart responds to extreme environmental heat and how aging modifies these responses. Such information can help guide the development of targeted preventative strategies and be incorporated into next generation thermophysiological models (3).

Cardiovascular related risks to heat exposure is difficult to predict on an individual basis, particularly since details on how the human heart responds to extreme heat is lacking, especially when comparing young and older adults. Moreover, the information that we do have is based largely on studies using a water-perfused garment to raise core temperature by 1-1.5 °C (4–7). Data from these studies are difficult to translate into to real-world extreme heat events, since the water-perfused suit causes a rapid increase in skin temperature to supraphysiological values, causing strong physiological responses which are not mirrored in environmental chamber studies (8). Moreover, by design, in studies using water-perfused suits, measurements are taken at the same increase in core temperature from baseline, meaning that known thermoregulatory impairments in older people (9) are minimized on the studies’ outcomes.

Determining the acute cardiac adaptations in older adults exposed to environmental heating may help identify which comorbidities amplify these risks. For example, if healthy older adults primarily rely on increasing cardiac systolic function during heat exposure, as shown using the water-perfused suit approach (5), those with systolic heart failure are likely at greater risk of heat-induced decompensation. Moreover, even if heat-induced increases in heart rate or systolic and diastolic function are comparable between age groups, older adults operate closer to their physiological limits (5, 10, 11). As a result, even healthy older individuals may face a heightened risk of adverse cardiac events simply because they have less physiological reserve to draw upon. Despite the relevance of understanding these acute cardiac responses to extreme heat in older adults, such data remain unavailable.

The current study builds on this previous work in several ways. First, we employed two environmental heat stress exposures representative of historically deadly U.S. heat waves: the exceptionally hot and dry 2021 Pacific Northwest heat wave (DRY, 47°C and 15% humidity) and the notably humid 1995 Chicago heat wave (HUMID, 41°C and 40% humidity) (12, 13). Each climate poses a unique thermoregulatory challenge, either through strong convective heat gain (DRY) or impaired evaporative efficiency (HUMID). Second, the current study involved a 3-hour exposure and included intermittent bouts of light activity, reflecting heat production associated with typical daily tasks. Lastly, we did not artificially control the increase in participants’ core temperatures, thus enabling physiological assessments reflective of real-world heat exposure scenarios where body temperature is higher in older people. For example, our companion paper, utilizing the same dataset, reports a greater increase in body temperature among older participants compared to younger individuals to these heat exposures (9).

The primary aim of this study was to compare the cardiac responses of healthy young versus older adults during a three-hour heat wave simulation in both DRY and HUMID. Given the greater thermal stress and impaired diastolic function in older adults, we hypothesized that older individuals would exhibit larger elevations in indices of systolic function.

Methodology

This research forms a component of a broader project that investigates the integrative physiological responses of young and older adults exposed to extreme heat conditions (Clinical Trials ID: NCT04538144). The study’s protocol and the informed consent documents were granted approval by the Institutional Review Boards at the University of Texas Southwestern Medical Center and Texas Health Presbyterian Hospital Dallas (STU-2019-1759). All research activities adhered to the ethical principles outlined in the Declaration of Helsinki. Healthy young adults aged 18 to 39 (n = 20, 10 male and 10 female) and older adults aged 65 years and above (n = 20, 10 male and 10 female), were recruited from the Dallas-Fort Worth metropolitan region. Experiments took place between February 2022 and March 2023. We chose 65 years as the minimum age for the older group because research suggests that beyond this age, the adverse health effects of heat exposure increase significantly (1, 14). We did not control for menstrual cycle phases for young female participants, as fluctuations in female sex hormones during the menstrual cycle do not impact thermoregulatory control (15, 16). Core temperature, renal, and gastrointestinal responses to these exposures, questions unrelated to the present work, have been published (9, 17, 18).

Exclusion criteria

The exclusion criteria comprised the following; i) individuals with known heart disease or other chronic medical conditions currently necessitating regular medical treatment, such as cancer, diabetes, uncontrolled hypertension, or uncontrolled hypercholesterolemia, ii) those taking tricyclic antidepressants, loop diuretics, centrally acting calcium channel blockers, or beta-blockers, iii) those exhibiting abnormalities indicative of provokable ischemia or undiagnosed cardiac disease or displaying a resting left bundle branch block on the initial electrocardiogram during screening, iv) current smokers or individuals who had regularly smoked within the past three years, v) individuals with a body mass index equal to or exceeding 31 kg/m⁻², and vi) pregnant individuals, confirmed through a urine pregnancy test in young females.

Preliminary visit

During the preliminary visit, subjects first completed a medical history form, followed by measures of height, body mass, brachial blood pressure, and heart rate and rhythm through electrocardiography. Additionally, we determined body composition using dual-energy X-ray absorptiometry. Finally, we conducted a metabolic heat production assessment, requiring a brief period of semi recumbent cycling lasting 10 to 15 minutes, while expired gases were analyzed using indirect calorimetry (PARVO Medics True-One Metabolic Measurement System, Parvo Medics, Salt Lake City, Utah). This assessment aimed to identify the workload that would result in a metabolic equivalent of 3 METs, which was used to replicate heat generation associated with activities of daily living during extreme heat exposure. Participant characteristics and medication use reported at screening are shown in Tables 1 and 2, respectively.

Table 1.

Participant characteristics.

	Young (10M/10F)	Older (10M/10F)	p value
Age, y	29 ± 5	70 ± 4	na
Height, cm	171 ± 8	170 ± 8	0.707
Mass, kg	68.8 ± 10.9	75.1 ± 11.3	0.083
BMI, kg/m2	24 ± 3	26 ± 3	0.011
BSA, m2	1.80 ± 0.16	1.86 ± 0.17	0.254
Body fat, %	28 ± 8	36 ± 7	0.002
Lean mass, kg	49 ± 10	48 ± 10	0.682

BMI, body mass index; BSA, body surface area; MET mins, total metabolic equivalent energy expenditure per week.

Table 2.

Number of participants using the indicated medications (self-reported at screening).

	Young	Older
5-alpha reductase inhibitor	0	1
ACE inhibitors/ARBs	0	7
Alpha 1 blockers	0	1
Non-tricyclic antidepressants	1	3
Antihistamines	3	2
Aspirin	0	3
Bisphosphonate derivatives	0	2
CNS stimulants	2	0
Non-loop diuretics	0	3
Birth control/estradiol	2	1
Gabapentin	0	2
Inhaled glucocorticoids	0	4
Thyroid drugs	0	1
Proton pump inhibitors	0	2
Semaglutide	0	1
Statins/fibric acid	0	12
Testosterone	0	1
Triptan	1	0
Vitamins	5	7

ACE, Angiotensin converting enzyme; ARB, Angiotensin receptor blocker; CNS, Central nervous system

Experimental Trials

The experimental sessions involved exposure to DRY and HUMID heat for three hours each, with a minimum of six days between each session, conducted in a randomized order. These extreme heat exposures took place in a controlled environmental chamber that precisely regulated air temperature and humidity (CANTROL, Ontario, Canada).

Participants arrived between 8:00 and 10:00 AM, to avoid any potential core temperature fluctuations due to the time of day. In preparation for these trials, participants received instructions to have a light breakfast with no caffeinated beverages on the morning of the testing (verbally confirmed upon arrival). Additionally, they were required to abstain from consuming alcoholic beverages and engaging in strenuous exercise within 24 hours before the testing (verbally confirmed upon arrival). Upon arrival, participants provided a urine sample. To confirm euhydration, a urine specific gravity value of ≤1.020 (measured with equipment from Atago Inc., Bellevue, WA) was required. If a participanťs urine specific gravity fell between 1.020 and 1.024, they drank 500 mL of water before proceeding with repeated testing ~30 minutes later. If their urine specific gravity was ≥ 1.025, the trial was postponed to a later date (n = 1). After confirming euhydration, participants rested in a controlled environment with a temperature ~22°C and ~40% humidity for at least one hour. During this period baseline cardiovascular ultrasound measurements were obtained (details provided below).

Following the completion of baseline measurements, participants were exposed to one of two extreme heat conditions within the environmental chamber: i) DRY (47°C and 15% humidity; reflective of the 2021 Pacific Northwest heat wave), and ii) HUMID (41°C and 40%humidity; reflective of the 1995 Chicago heat wave). The actual conditions during the protocol were 46.4 ± 0.4 °C and 15.9 ± 2.3 %RH for DRY, and 40.8 ± 0.5 °C and 40.3 ± 1.6 %RH for HUMID. The duration of each exposure was 3 h, which represents the approximate duration of peak daily environmental temperatures during a heat wave (19, 20). Before instrumentation, participants measured and annotated their nude body mass using a precision balance scale with ±10 g accuracy (Mettler Toledo, OH). After a brief period (~5 min) of instrumentation, participants were seated in the upright position on a chair with a breathable fabric, thereby reducing impediments to thermal exchange from the regions in contact with the chair. Participants wore athletic shorts (males) or athletic shorts and a sports bra (females) in addition to shoes.

To mitigate the impact of dehydration during heat exposure, individuals drank 3 mL per kilogram of their body mass per hour of temperature-controlled (14°C to 16°C) bottled water. This fluid intake amount was chosen based on preliminary data and was designed to keep participants euhydrated during heat exposure. The selected water temperature aimed to replicate the temperature of water accessible to people during a heat wave. After completing the 3-hour heat simulation, monitoring equipment was removed, and subjects were provided with a towel to remove excess sweat before re-weighing themselves in a nude state (to calculate whole-body sweat loss). An experimental protocol is shown in Figure 1.

Figure 1. Experimental protocol.

Healthy young (n = 20) and older (n = 20) participants were exposed to two 3-h extreme heat exposures on different days: 1) DRY (47°C and 15% humidity) and 2) HUMID (41°C and 40% humidity) in a randomised order. To mimic heat generation comparable with activities of daily living, participants performed seven 5 min bouts of recumbent cycling (∼3 METs) dispersed throughout the heat exposure. Participants consumed 3 mL/kg body mass of tap temperature water (~15°C) every hour during the heat exposures. At baseline and ~20 min prior to the end of heat exposure, ultrasound was used to assess left ventricular systolic and diastolic function, cardiac output, stroke volume, and brachial artery blood flow. Venous blood samples were taken at baseline, at the end of heat exposure, and 2-hours following recovery in a thermoneutral environment. Figure generated using Biorender.

Metabolic Heat Generation

During the 3-hour extreme heat exposures, participants engaged in seven 5-minute sessions of light physical activity. These exercise bouts took place at minutes 15, 35, 55, 75, 115, 135, and 155. As such, at least 15 minutes of rest elapsed between the final exercise bout and the echocardiography acquisition. These activities consisted of either cycling (n = 39) or walking (n = 1), performed at an intensity equivalent to approximately 3 metabolic equivalents (METs). This intensity was chosen to simulate common daily activities like cooking, cleaning, etc (21). To confirm the metabolic heat generated during these activities, indirect calorimetry was employed (PARVO Medics True-One Metabolic Measurement System, Parvo Medics, Salt Lake City, UT). Expired gases were collected during the first two 5-minute exercise sessions, and the workload determined during this phase was subsequently applied to all subsequent exercise sessions. The average oxygen consumption rate, respiratory exchange ratio, and workload were used to calculate metabolic heat production, utilizing standard formulae (22).

Instrumentation

Core temperature was assessed by monitoring rectal temperature, employing either a general-purpose thermocouple probe inserted 10 cm beyond the anal sphincter (Mon-a-therm, Mallinckrodt Medical, St. Louis, MO) for a subset of participants (Young: n = 17, Older: n = 8), or a telemetric pill (e-Celsius performance pill, BodyCap, Caen, France) administered as a rectal suppository for two participants (Young: n = 1, Older: n = 1). In cases where participants declined rectal temperature measurement or faced technical difficulties with the rectal thermocouple probe, core temperature was instead measured by utilizing an orally ingested telemetric pill (Young: n = 2, Older: n = 11). For these participants, the ingestible telemetric pill was administered at least one hour before the commencement of the extreme heat exposure. The duration following the ingestion of the telemetric pill does not impact the accuracy of using pill temperature as a measure of core temperature (23). Furthermore, there is a high degree of agreement between the telemetric pill and rectal temperature measurements (24, 25). Among participants who underwent simultaneous measurements of ingested pill and rectal temperature in this study (n = 25), the intraclass correlation coefficient was 0.87. Mean skin temperature was calculated as a weighted average recorded at various body locations, including the chest (22%), upper back (21%), lower back (19%), abdomen (14%), anterior thigh (13%), and calf (11%) (26).

Heart rate data were acquired through a single-lead electrocardiogram (GE Medical Systems, Madison, WI). Age predicted maximum heart rate was calculated using the Tanaka equation (27). Blood pressure was determined through automated auscultation of the brachial artery (Tango+, SunTech Medical, Morrisville, NC). Rate pressure product was calculated as:

$$RatePressureProduct=\frac{SBP\times HR}{100}$$

where SBP is systolic blood pressure in mmHg, and HR is heart rate in beats per minute.

Echocardiography

Cardiac ultrasounds were performed by authors Drs. Sarma or MacNamara, cardiologists highly experienced and trained in medical imaging using echocardiography. Cardiac ultrasound measurements were made with commercially available ultrasound devices (matched within participant; iE33 or EPIQ 7, Philips Medical Systems, Andover, MA), using an X-5 transducer (Philips Medical Systems, Andover, MA). Echocardiography data were analyzed using TomTec software.

Tissue Doppler imaging

We measured peak early diastolic (e′), late diastolic (a′), and systolic (s′) velocities at the septal and lateral mitral annular regions. At each time point, quadruplicate measurements from the septal and lateral regions were averaged (i.e., means of four assessments from e’, a’, and s’) and served as indicators of early left ventricular diastolic relaxation (28), late diastolic mitral annular velocity, and left ventricular systolic contraction velocity, respectively. Standard tissue Doppler imaging techniques were utilized, as previously outlined (4, 28). The tissue Doppler measurements were acquired with the individual in a semi-recumbent position from the apical four-chamber view, with a 5.0 mm sample volume placed on both the septal and lateral mitral annuli.

Mitral inflow velocities

Mitral inflow velocities were assessed with pulsed wave Doppler from the apical four-chamber view, employing a sample volume of 2.0 mm positioned just above the mitral valve leaflet tips. We obtained measurements of the peak inflow velocity during the early phase of left ventricular relaxation (E wave), which serves as an indicator of early left ventricular diastolic filling (29, 30), as well as during left atrial contraction (A wave), which provides insight into late diastolic filling (29). These values were then utilized to quantify diastolic filling patterns by calculating the E/A ratio (29, 31). To assess the relative atrial contribution (late diastolic filling) to left ventricular diastolic filling, we used the formula: [A/(E + A)] × 100 (4).

Global Longitudinal Systolic Strain (GLSS)

Peak GLSS was obtained from 2D apical two (AP2), AP3, and AP4 views and analyzed using TomTec’s semi-automated wall tracking software (AutoSTRAIN© package), in which the package automatically detects endocardial borders. In most cases, the investigator made manual adjustments in end-systole and end-diastole to ensure appropriate tracking, from which peak systolic strain (average of A2C, A3C, and A4C) values were recorded.

Left Ventricular Volumes

Left ventricular end-diastolic (EDV) and systolic (ESV) volumes, twist (°) and torsion (°/cm) were quantified using TomTec’s 3D speckle tracking software (4D LV-analysis© package). The software provides a semi-automated detection of the endocardial border of the left ventricle after the operator manually provides the information on the location of the mitral valve and apex, while also correcting the plane and timing (ensuring true end-diastolic and end-systolic frames) of the image where necessary. Manual adjustments were made to ensure proper delineation of the LV endocardial border.

High sensitivity cardiac troponin I

We assessed hs-cTn I as an index of myocardial cell injury in the venous plasma (Dimension EXL integrated chemistry system with LOCI module, Siemens Healthcare Diagnostics, assessed by Inova laboratories). Samples were taken at baseline, end 3-hr heating, and 2-h post recovery. According to the assay manufacturer, the 99^th percentile value for hs-cTn I, derived from 2020 apparently healthy individuals with no signs of cardiac disease or injury, was 60.4 pg/ml.

Sample Power

The elevation in mitral annular systolic velocity (s’) was used for the power analysis. In both younger and older individuals, we previously showed that an increase in core temperature of ~1 °C increased s’ by 3.8 ± 1.3 cm/s, representative of a 56 ± 20% increase from baseline (5). We hypothesized that due to known thermoregulatory impairments with aging (9), this population will exhibit greater elevations in core temperature and therefore cardiac stress (i.e., Δ s’) from pre to post 3-hour heat exposure. The power calculation was performed to detect a 2 cm/s difference in the magnitude of the increase in s’. Based on these estimates, 24 total subjects (12 subjects per age group) would provide a statistical power 98% to detect a difference between age groups for s’ (alpha value = 0.05) (power analyses from: NCSS PASS 2019). Nevertheless, we enrolled 40 total subjects (20 per age group, with each group containing a 50:50 split between male and females) to allow for the assessment of variables that may have a greater degree of variability to the extreme heat exposure when compared to s’.

Statistical Analyses

Statistical analyses were conducted in Prism version 10 (GraphPad Software Inc., La Jolla, CA). Before analyses, data were assessed to confirm that they were normally distributed. All variables met these statistical assumptions, such that data in text, tables, and figures are presented as means ± standard deviations and analysed with parametric tests. Within each heat exposure (DRY and HUMID), we analyzed data using linear mixed effect models with main effects of time (within factor; baseline vs post) and group (between factor; older vs. young). For plasma cardiac troponin, the main effects of time had three factors (baseline, post, and 2-hours post). In the event of a significant interaction term, multiple comparison tests with Bonferroni correction were used to identify specific group and time differences. The alpha value for significance testing was set as p < 0.05. alpha values shown in the results text refer to the ANOVA interaction effect.

Exploratory regression analyses were performed to help provide a mechanistic framework underpinning our results. We explored how a) thermal stress (core temperature) links to ventricular filling pressure (end diastolic volume), b) how filling pressure relates to heart rate (% age predicted max), and c) how age-predicted max heart rate links to the augmentation of diastolic function (Figure 5). Data from the DRY condition only were used for these analyses since it resulted in a wider spread of core temperature responses. Moreover, combining both the DRY and HUMID data would violate the assumption of independence between data points.

Figure 5. Regression analysis between multiple variables at the end of the DRY heat exposure.

High core temperature is indirectly related to end diastolic volume, a marker of left ventricular filling pressure. The prevailing end diastolic volume is associated with heart rate. That rise in heart rate is then linked to the augmentation of diastolic function. No such links were found which help explain the augmentation of systolic function. Hatched symbols show the female subjects. Linear regression p < 0.001.

Results

Group differences in baseline characteristics, thermoregulatory responses (core and skin temperature), forearm blood flow, heart rate, sweat rate and mean arterial pressure are shown in our companion paper using the same dataset (9). Briefly, that paper shows body fat% and BMI were higher in the older group, in line with expected population level differences. Sweat rate and skin blood flow were similar between age groups. That work also showed greater absolute core temperature responses in the older subjects in the DRY (Young: 37.81 ± 0.26°C vs. Older: 38.15 ± 0.43°C; p = 0.005), but not HUMID (Young: 37.67 ± 0.34°C vs. Older: 37.83 ± 0.35°C; p = 0.151) condition. Differences in core temperature must be considered when interpreting the results, since any reported differences in cardiac responses may have occurred in conjunction with a greater core temperature.

Heart rate, systolic and diastolic blood pressure, cardiac output, stroke volume, and ejection fraction

Volume data indexed to body surface area, cardiac minute work and stroke work, and body mass changes are available in the online supplement (Table S1).

DRY

Heart rate increased to a greater extent in the older adults in the DRY condition (Young: Δ18 ± 9 bpm vs. Older: Δ25 ± 11 bpm; p = 0.039). Age predicted maximum heart rate increased to a greater extent in the older adults (Young: Δ9 ± 4 % vs. Older: Δ15 ± 6 %; p = 0.001). Systolic blood pressure decreased in older adults only and was well maintained in young adults (Young: Δ1 ± 8 mmHg vs. Older: Δ-11 ± 14 mmHg; p = 0.001). Diastolic blood pressure decreased to a similar extent between young and older adults (Young: Δ-14 ± 10 mmHg vs. Older: Δ-11 ± 10 mmHg; p = 0.421). Cardiac output increased to a similar extent between young and older adults (Young: Δ0.88 ± 0.77 L/min vs. Older: Δ0.75 ± 0.82 L/min; p = 0.469). Stroke volume decreased similarly between young and older adults (Young: Δ-5 ± 9 mL vs. Older: Δ-9 ± 11 mL; p = 0.240). Ejection fraction was not different between age groups and was unaffected by heating in both climates.

HUMID

Absolute heart rate increased similarly between age groups (Young: Δ15 ± 9 bpm vs. Older: Δ18 ± 8 bpm; p = 0.252). Age predicted maximum heart rate increased to a greater extent in the older adults (Young: Δ7 ± 5 % vs. Older: Δ11 ± 6 %; p = 0.047). Systolic blood pressure decreased in older adults only and was well maintained in young adults (Young: Δ-1 ± 8 mmHg vs. Older: Δ-9 ± 15 mmHg; p = 0.224). Diastolic blood pressure decreased to a similar extent between young and older adults (Young: Δ-12 ± 7 mmHg vs. Older: Δ-9 ± 9 mmHg; p =). Cardiac output increased to a similar extent between young and older adults (Young: Δ0.88 ± 0.77 L/min vs. Older: Δ0.75 ± 0.82 L/min; p = 0.180). Stroke volume decreased to a similar extent between young and older adults (Young: Δ-7 ± 6 mL vs. Older: Δ-9 ± 6 mL; p = 0.278). Ejection fraction was not different between age groups and was unaffected by heating in both climates.

Indices of left-ventricular diastolic function

Mitral annular velocity during early (e’) and late (a’) filling, and the ratio of early to late peak diastolic filling velocity are available in the online supplement (Table S2).

DRY

Early filling velocity (E, cm/s) decreased to a similar extent in both young and older adults (Young: Δ-7 ± 11 cm/s vs. Older: Δ-6 ± 14 cm/s; p = 0.885). Late filling velocity (A, cm/s) increased to a similar extent in both young and older adults (Young: Δ11 ± 3 cm/s vs. Older: Δ9 ± 13 cm/s; p = 0.718). The contribution of atrial contraction to diastolic filling increased to a similar extent in both young and older adults (Young: Δ7 ± 5 % vs. Older: Δ5 ± 5 %; p = 0.420).

HUMID

Early filling velocity decreased to a similar extent in both young and older adults (Young: Δ-5 ± 12 cm/s vs. Older: Δ-6 ± 11 cm/s; p = 0.770). Late filling velocity increased to a similar extent in both young and older adults (Young: Δ6 ± 15 cm/s vs. Older: Δ5 ± 12 cm/s; p = 0.869). The contribution of atrial contraction to diastolic filling increased to a similar extent in both young and older adults (Young: Δ4 ± 5 % vs. Older: Δ4 ± 6 %; p = 0.917).

Indices of left-ventricular systolic function and plasma troponin

Additional data for LV twist and Torsion are available in the online supplements (Table S3).

DRY

Mitral annular systolic velocity (s’) increased to a greater extent in older adults (Young: Δ1.5 ± 1.5 cm/s vs. Older: Δ3.3 ± 2.1 cm/s; p = 0.002). Global longitudinal systolic strain (GLSS) decreased in young, but not older adults (Young: Δ5.2 ± 2.3% vs. Older: Δ0.6 ± 3.6 %; p < 0.001). Plasma troponin I was greater in older adults 2 hours into thermoneutral recovery (Young: Δ-0.61 ± 0.91 pg/ml vs. Older: Δ0.58 ± 1.12 pg/ml; p = 0.054). Left ventricular Twist (%) and Torsion (°/cm) were not different between age groups and were unaffected by heating in both climates.

HUMID

Mitral annular systolic velocity (s’) increased similarly in young and older adults (Young: Δ1.4 ± 1.3 cm/s vs Older: Δ2.1 ± 1.3 cm/s, p = 0.096). Global longitudinal systolic strain decreased in young, but not older adults (Young: Δ4.0 ± 3.0% vs. Older: Δ1.2 ± 2.8 %; p < 0.001). Left ventricular Twist (%), Torsion (°/cm), and plasma Troponin I were not different between age groups and were unaffected by heating in both climates.

Regression analyses at 3-hr timepoint in DRY

Core temperature was negatively related to end diastolic volume (R² = 0.39, p < 0.001). End diastolic volume (mL) was negatively related to heart rate (%max, R² = 0.69, p < 0.001). Heart rate (%max) was positively related to the atrial contribution to diastolic filling (R² = 0.55, p < 0.001).

Discussion

The aim of this study was to compare the cardiac responses to extreme environmental heat exposure between young and older adults. Age differences in these responses were compared in two distinct two environmental conditions; DRY (47 °C, 15% RH) reflective of the peak conditions during the Pacific Northwest 2021 heat wave, and HUMID (41 °C, 40% RH) to reflect the 1995 Chicago heat wave. To improve ecological validity, these participants performed intermittent (7 x 5 min) bouts of physical activity at 3 METS to simulate the extra heat produced during activities of daily living. While this paper focused on the acute cardiac adjustments to heat stress, such changes occurred in line with an elevated thermal strain in the older group. Those findings are discussed in more detail in our companion paper (9).

There are several novel findings from this study. First, older individuals showed a greater reliance on systolic mechanisms in the DRY condition, evidenced by a greater increase in systolic mitral annular velocity (s’). Second, while the absolute reliance on atrial contraction for diastolic filling (A%) was higher in the older individuals, the augmentation of diastolic function was similar between age groups, with A% increasing by ~ 10% in both groups. Third, we showed only modest increases in cardiac output (~0.5 to 1 L/min), and consistent reductions in stroke volume up to ~20 mL in both age groups. Despite the augmentation of systolic and diastolic function in older individuals, systolic blood pressure fell consistently in older adults but was well maintained in younger adults. Finally, regression analysis demonstrates several moderate-strong predictors of the cardiac response to heating. For example, we show that end diastolic volume is strongly linked to heart rate, and that heart rate links strongly to the magnitude of augmentation of diastolic function (i.e., atrial kick, Figure 5). These findings demonstrate the compensatory mechanisms that older adults use to maintain adequate cardiac output during heat stress, highlighting potential vulnerability to early decompensation in populations with systolic or diastolic heart failure, or perhaps those taking beta blockers.

Heat exposure increased s’ in almost all participants, indicating an elevated systolic contractility in the heat. Using environmentally induced heating with intermittent activities of daily living, we show here that s’ increased to a greater extent in older participants in the DRY climate (Young: Δ1.5 ± 1.5 cm/s vs. Older: Δ3.3 ± 2.1 cm/s; p = 0.002). No such age differences were present in the HUMID condition, likely due to a less pronounced difference in their thermal strain (see Mckenna et al., 2023). These findings agree with our previous work using the water perfused suit, where the increase in s’ was similar between age groups when core temperature was matched (5). In conjunction with an elevated s’, older individuals were operating a greater percentage of their age-predicted maximum heart rate (Figure 2).

Figure 2. Stroke volume, cardiac output, systolic blood pressure, and percentage age-predicted maximum heart rate before and 3-hour into DRY (47°C, 15% RH, left) and HUMID (41°C, 40% RH, right) heat exposure in young (circles) and older (squares) adults.

Females shown with hatched symbols. Arrows indicate difference in average response from pre to post in each age group (dashed arrows used in older group). Significant age group (exact p values) and time differences (shown with a #) are shown in the event of a significant interaction term (p < 0.05).

The greater s’ in older people was likely due to a combination of impaired diastolic function and the attainment of a higher core temperature (increasing sympathetic stimulation). Although such differences in s’ may not be clinically relevant acutely, older adults are likely operating closer to their maximum contractile reserve. For example, in healthy heat-stressed subjects infused with a β1 adrenergic agonist, we found that the increase in s’ in older individuals was ~50% of the increase observed in younger individuals (10). When a maximum threshold is reached, older individuals will have limited capacity to maintain cardiac output. As such, in scenarios where peak systolic contractility is further limited (i.e., in systolic heart failure, medications), individuals may be at greater risk of heat-related adverse events. In addition, we found modest elevations in cardiac Troponin I in older individuals, but only in the DRY condition (see Figure 4), which may be due to a greater thermal strain in this environment. Although the increase of ~ 2 pg/ml is subclinical, it demonstrates an impact of age per se on myocardial damage during heat exposure in older people, which may be augmented in with a greater duration of heat stress and/or in individuals operating closer to their cardiovascular reserve. We explored whether the increase in Troponin I could be explained by changes in core temperature, heart rate, systolic and diastolic function, and body characteristics, finding no clear associations.

Figure 4. Indices of left ventricular systolic function (panel a) before and 3-hour into DRY (47°C, 15% RH, left) and HUMID (41°C, 40% RH, right) heat exposure in young (circles) and older (squares) adults. Plasma cardiac troponin I (panel b) was also measured 2-h into recovery in a thermoneutral environment.

Females shown with hatched symbols. Arrows indicate difference in average response from pre to post in each age group (dashed arrows used in older group). Significant age group (exact p values) and time differences (shown with a #) are shown in the event of a significant interaction term (p < 0.05).

An unexpected finding was the differential GLSS response between age groups. A decrease (i.e., becoming less negative) suggests reduced longitudinal systolic performance. For clarity, we use the term decrease to describe a change in GLSS closer to zero i.e., −20 to −15% equals an absolute decrease of 5%. In older adults, GLSS was maintained with heat stress in both the DRY and HUMID condition. However, in young adults, exposure to both climates caused GLSS to decrease (became less negative) by ~5% (Figure 4). Speculatively, the final GLSS value is determined by the balance between unloading and contractility. For example, cardiac unloading decreases GLSS, whereas increasing contractility elevates GLSS (i.e., GLSS becomes more negative) (32). Heat-induced unloading of the left ventricle was equivalent between age groups, demonstrated by similar decreases in end diastolic volume. However, s’ (a marker of contractility) increased to a greater extent in older adults, potentially countering the effects of unloading on decreasing GLSS. Our findings of a decreased GLSS with heating in younger individuals are consistent with studies using a water perfused suit (5, 33). In those studies, group average GLSS decreased by ~2-3% with passive heat stress, albeit non-significantly perhaps due to a relatively low sample size.

Both the DRY and HUMID climate increased cardiac output by ~ 1 L/min (Figure 2). The magnitude of the change in cardiac output broadly agrees with a recent meta-analysis demonstrating a 1 L/min increase on average per 1 °C increase in core temperature (8). Since heat exposure decreased stroke volume in most participants, the increase in cardiac output was driven by a combination chronotropic and inotropic mechanisms. Importantly, we show that absolute heart rate was similar between age groups, older individuals were operating at a greater percentage of their age predicted max, Figure 2) i.e., closer to their cardiovascular reserve. Figure 5 demonstrates that the increase in heart rate is strongly associated with the prevailing ventricular filling pressure.

While the current study and others show a reduced stroke volume during extreme heat exposure (7), other studies (including work from our lab) report either no changer or an increase in hyperthermia-induced stroke volume (6, 34). The reason for this discrepancy is difficult to explain but may be due to inconsistencies in the method used to measure stroke volume (i.e., modelflow, echocardiography, thermodilution, or MRI), and/or differences in body position. In most water perfused suit studies, participants are heated in the supine position, reducing the gravitational constraint on venous return. In the present study, echocardiography acquisitions were made with subjects in the semi-recumbent position, however, they were sat upright during most of the heat exposures. Such postures may have contributed to a slight decrease in stroke volume through fluid loss to the interstitial space (venous pooling). Moreover, it should be noted that the employed light intensity exercise was unlikely to have a direct effect on the obtained echocardiography-based indices of cardiac function given that those indices were obtained at least 15 min after the final light-intensity exercise bout. Nevertheless, these assertions are speculative and warrant validation.

As expected, a major difference between age groups were indices of diastolic function; i.e., our findings at baseline are consistent with expected age-related decreases in passive recoil, marked by a decrease early mitral valve filling velocity and increased filling pressure in older adults. In both environmental conditions and both age groups, heat stress increased the reliance on atrial contraction (A) for overall diastolic filling (Figure 2); the contribution of late atrial contraction to filling increased by ~5% with heat exposure in both age groups and both environmental conditions. Our findings are broadly in agreement with our previous study using the water perfused suit to increase core temperature by 1.2 °C in young and older adults (5). In that study, matched hyperthermia increased atrial contribution to filling in younger adults only, where the final percentage contribution with matched hyperthermia was ~50% in both age groups. In contrast, the present study showed that the atrial contribution to diastolic filling increased similarly in both age groups during both DRY and HUMID heat exposure trials (+ ~5%). The percentage maximum heart rate was linked directly to the magnitude of atrial contribution to diastolic filling regardless of age (A%, Figure 5).

In addition to the elevated atrial contribution with heat stress, both heat conditions decreased early mitral valve inflow velocity in both age groups. To our knowledge, only one previous study (from our lab) has compared the early mitral valve inflow velocity to heat stress between young and older individuals (5). In that study (matched increase in core temperature of 1.2 °C via the water perfused suit approach), early mitral valve inflow velocity was maintained with heat stress in young and older adults. The difference in heating modes (e.g., environmental heating versus water perfused suit heating) as well as differences in posture between studies may explain these results. Regarding the latter, in the prior study participants laid in the supine position throughout the protocol. In the present study, participants sat upright through the 3-hour exposure, although participants were moved to a semi recumbent position for echocardiography assessments.

Limitations

There are several limitations with our work that should be considered. First, the duration of heat exposure was relatively short compared to the possibility of several days heat exposure in real-world heat waves. That said, the primary purpose of the present work was to simulate the peak environmental heat stress experienced during the 2021 Pacific Northwest and 1995 Chicago heat waves. Thus, future studies should investigate detailed cardiac responses in older people exposed to prolonged heat wave-like conditions. Second, the use of euhydrated healthy older individuals limits the extent to which our data can be applied to the most vulnerable populations, such as frail older people with chronic diseases. Nevertheless, we show that even in healthy seniors, there is a greater reliance on systolic mechanisms to maintain stroke volume during heat stress. Future work should determine the extent to which cardiovascular function is affected by heat exposure in individuals with chronic diseases, including various cardiovascular or pulmonary diseases typical in older populations. Another limitation is that we did not control for or measure menstrual cycle phase or hormone levels in our female participants. We chose not to restrict testing to a single phase since prior evidence indicates only a minimal thermoregulatory impact across the cycle (16), and focusing on one phase would have reduced generalizability. In addition, variation in primary measures such as s′ and stroke volume was small within the female group, suggesting that any influence of menstrual cycle phase was likely minor in the context of this study’s findings. Finally, our cohort took several prescription medications primarily for the control of blood pressure and cholesterol, with the impact of these medications on our findings unexplored due to insufficient statistical power.

Conclusions

From a public health perspective, these findings highlight the compensatory mechanisms used to reach an appropriate cardiac output in healthy older adults exposed to extreme heat. We demonstrate that, similarly to young healthy adults, healthy seniors rely on augmentations of heart rate and systolic & diastolic function to meet the demands hyperthermia places on the cardiovascular system. However, it is noteworthy that an index of systolic function (s’) was augmented to a greater extent in the older group. Since these mechanisms cannot be augmented, or their augmentation is attenuated, with certain medications (i.e., beta blockers) or in specific clinical populations (i.e., systolic or diastolic heart failure), these groups may be more likely to experience early decompensation during extreme heat exposure. Future studies are needed to clarify these speculations.

Supplementary Material

Supplementary Tables S1-S3 https://doi.org/10.6084/m9.figshare.29298224

Figure 3. Indices of diastolic function. before and 3-hour into DRY (47°C, 15% RH, left) and HUMID (41°C, 40% RH, right) heat exposure in young (circles) and older (squares) adults.

Females shown with hatched symbols. Arrows indicate difference in average response from pre to post in each age group (dashed arrows used in older group). Significant age group (exact p values) and time differences (shown with a #) are shown in the event of a significant interaction term (p < 0.05).

Data availability

Data will be made available upon reasonable request to the corresponding author CGC.