Exercise, Escitalopram, and Anxiety in Heart Disease

For people with stable coronary heart disease and clinically significant anxiety, supervised exercise improved symptoms over 12 weeks. So did escitalopram. So did placebo. That last result deserves top billing: the structured clinical setting itself coincided with a large share of the improvement.

In the UNWIND randomized clinical trial, escitalopram produced the largest reduction in anxiety scores. Exercise reduced anxiety by only about 0.5 points more than placebo, a difference that did not reach statistical significance. Both active treatments, however, reduced 24-hour urinary catecholamines relative to placebo, a marker of sympathetic nervous-system activity. That is biologically interesting for heart disease, but it does not prove fewer heart attacks, arrhythmias, or deaths.

The practical case for exercise remains strong: it can build fitness and may help anxiety while avoiding medication-specific concerns. But the trial does not show that exercise equals escitalopram for anxiety, and it certainly does not show that a lower catecholamine reading guarantees better cardiovascular outcomes.

What UNWIND actually tested

UNWIND enrolled 128 sedentary adults older than 40 with stable coronary heart disease and elevated anxiety symptoms or a diagnosed anxiety disorder. Their mean age was 64.6 years, and 71% were men. None were receiving mental health treatment when they entered the trial.

Participants were randomly assigned to one of three 12-week treatments: supervised aerobic exercise three times weekly at 70% to 85% of heart-rate reserve; escitalopram titrated to 20 mg per day; or a matched placebo pill. The study measured anxiety with the Hospital Anxiety and Depression Scale Anxiety subscale, or HADS-A, and also assessed heart-rate variability, baroreflex sensitivity, flow-mediated dilation, and 24-hour urinary catecholamines.

This design matters. Exercise was not compared with doing nothing or a waitlist. It faced a placebo pill within a structured research program, which makes the comparison more useful—and makes a small difference harder to interpret as a meaningful treatment effect.

Escitalopram had the largest anxiety effect

All three groups improved on HADS-A. Scores fell by 3.5 points with placebo (95% CI, −4.5 to −2.4), 4.0 points with exercise (95% CI, −4.7 to −3.2), and 5.7 points with escitalopram (95% CI, −6.4 to −5.0). A patient looking only at change from baseline might reasonably conclude that every option worked.

Randomized trials ask a tougher question: how much did each active treatment add beyond the experience shared by the placebo group? Exercise added roughly 0.5 HADS-A points beyond placebo, and that difference was not statistically significant. Escitalopram outperformed exercise by 1.67 points (95% CI, −2.68 to −0.66; P=.002) and also outperformed placebo by a comparable margin.

The most defensible reading is straightforward. Escitalopram reduced anxiety more than exercise in this study. Exercise produced a favorable point estimate, but the trial could not rule out a small effect—or no advantage over placebo at all.

A 3.5-point reduction in the placebo group means treatment content was only part of the story. Regular clinical contact, monitoring, expectations, and the tendency for unusually high symptom scores to move closer to average values may all have contributed.

That does not make the relief imaginary. It means the trial cannot cleanly separate medication effects, exercise effects, regression toward the mean, and the therapeutic effects of a supportive clinical program. The placebo result is a reminder that anxiety care involves more than the active ingredient printed on a label.

The cardiac biomarker result is promising but incomplete

Both exercise and escitalopram reduced 24-hour urinary catecholamines compared with placebo (F_1,127=4.93; P=.01). Escitalopram also outperformed exercise on this measure (F_1,127=4.37; P=.04). Catecholamines reflect sympathetic activation, so a reduction fits with the idea that both treatments may reduce physiological stress signaling.

UNWIND found no between-group differences in heart-rate variability, baroreflex sensitivity, or flow-mediated dilation. Every group changed similarly on those measures. With roughly 43 people per arm and biomarkers that vary considerably within an individual, these null results are inconclusive rather than proof that exercise has no autonomic or vascular effect.

Still, the catecholamine result has an important limit. Urinary catecholamine reduction is a biomarker outcome, not a clinical event outcome. The trial did not test major adverse cardiovascular events, sudden cardiac death, or mortality, and it does not establish urinary catecholamines as a surrogate that predicts those outcomes in stable coronary heart disease.

Why broader exercise research does not settle this question

Exercise has substantial evidence behind it for depressive symptoms in many populations. A 2024 meta-analysis of HIIT trials, including 34 randomized trials and 1,607 participants, reported a small but statistically significant reduction in depressive symptoms versus control conditions (standardized mean difference −0.40, 95% CI, −0.60 to −0.20). A 2024 network meta-analysis in older adults, a 2025 analysis in menopausal women, and a 2024 perinatal-depression meta-analysis also found that exercise interventions can reduce depressive symptoms.

Those findings support exercise as a psychologically active intervention. They do not overturn UNWIND because they answer different questions. Most measured depression rather than anxiety, and none studied the older, sedentary population with stable coronary heart disease enrolled in UNWIND. Depression and anxiety overlap, but they are not interchangeable endpoints. A result in perinatal depression cannot be pasted onto anxiety in cardiac rehabilitation like a new label on an old jar.

Intensity and dose may also matter. A 2025 meta-analysis of high-intensity exercise for depression found meaningful symptom improvement across non-cardiac populations, while a 2025 network meta-analysis in adolescents reported that exercise type, intensity, and duration moderated depression outcomes. These studies raise a reasonable hypothesis: a longer, more intensive, or more varied program might have produced a larger anxiety effect in UNWIND. They cannot establish it. UNWIND used a medically supervised moderate-to-vigorous aerobic program in older adults with heart disease, not high-intensity protocols in younger or psychiatric cohorts.

Fitness has stronger support than anxiety relief in cardiac rehab

The exercise prescription in UNWIND was not pointless because it failed to beat placebo decisively on anxiety. A 2025 systematic review and meta-analysis by Collins et al. found dose-response relationships between aerobic exercise and improvements in peak oxygen uptake in cardiac-rehabilitation populations. Peak oxygen uptake is a fitness measure, not a measure of anxiety, catecholamines, heart-rate variability, baroreflex sensitivity, or vascular function. But it supports the established rationale for supervised aerobic training in people with coronary disease.

Likewise, a 2023 trial of acute upper-body exercise in cold conditions showed that baroreflex sensitivity can be measured and altered by an exercise-related hemodynamic challenge in stable coronary artery disease. It does not show that chronic exercise training improves baroreflex sensitivity, so it cannot rescue or refute UNWIND's null chronic result.

Who should apply these findings?

UNWIND applies most directly to sedentary, older adults with stable coronary heart disease and clinically important anxiety who are not already in mental health treatment. Its 71% male sample limits confidence about whether women would respond similarly. Its 12-week duration and modest per-group sample size also leave open whether a different exercise dose would change the biomarker or anxiety results.

People without coronary heart disease should not treat this trial as a personalized exercise-versus-SSRI decision tool. The broader exercise literature supports mood benefits, especially for depression, but this specific study tested a medically constrained cardiac population under supervision.

Escitalopram decisions also require clinical context that UNWIND does not resolve. Questions about QTc prolongation, platelet effects, and bleeding risk with concurrent antiplatelet treatment matter in coronary heart disease. Discuss them with the clinician managing both the anxiety and the cardiac regimen.

The honest bottom line

If you have stable coronary heart disease and anxiety, supervised aerobic exercise is a sensible part of care: it is compatible with cardiac rehabilitation, has established fitness benefits, and in UNWIND was associated with lower urinary catecholamines than placebo. It may also improve anxiety symptoms.

Do not choose exercise over escitalopram because this trial proved they work equally well for anxiety. It did not. Escitalopram produced the greater HADS-A reduction, while exercise did not significantly outperform placebo. The trial's larger lesson is less glamorous and more useful: anxiety treatment works better when it includes a structured, supportive plan—and the plan should be built around the patient, not a headline.