Blood Vessel Damage, Not the Heart, May Drive Long COVID Exercise Problems Three Years On

Three years after a SARS-CoV-2 infection, many adults who developed long COVID still cannot exercise the way they used to, and a new study points to a surprising culprit: the blood vessels rather than the heart itself. Researchers at the University of Brasilia in Brazil report in Physiological Reports that people with persistent post-COVID symptoms show marked and lasting dysfunction of the endothelium, the thin layer of cells lining blood vessels, even though the pumping performance of their left ventricle remains essentially normal. The finding, drawn from a carefully standardized case-control comparison, suggests that the exercise intolerance that defines so much of long COVID may arise primarily from impaired delivery of oxygen-rich blood to working muscles, not from subtle damage to the heart muscle itself.

The research team recruited 21 adults aged 35 to 68, dividing them into 11 controls and 10 individuals with long COVID. Every participant had a documented history of SARS-CoV-2 infection, but the two groups differed sharply in the severity of the acute illness and in their recovery. The controls had all experienced asymptomatic or mild infections and reported no lingering symptoms. The long COVID group, by contrast, had suffered moderate-to-critical disease: 60 percent required hospitalization, 60 percent needed oxygen therapy, and one person required mechanical ventilation. At the time of testing, roughly three years after infection, they continued to report fatigue, exercise intolerance, breathlessness, and headaches, and their scores on the Post-COVID-19 Functional Status Scale indicated substantially greater functional limitation than the controls.

To disentangle the possible mechanisms behind this persistent limitation, the investigators designed a three-visit protocol that separated each measurement from the others. On the first visit, participants completed a symptom-limited cardiopulmonary exercise test on a cycle ergometer, with breath-by-breath gas exchange analysis, continuous electrocardiography, and blood pressure monitoring. The second visit was devoted to transthoracic echocardiography, including advanced speckle-tracking analyses of myocardial work. The third visit assessed endothelial function using brachial artery flow-mediated dilation, a noninvasive ultrasound technique in which a blood pressure cuff is inflated for five minutes to induce ischemia and the subsequent surge of blood upon release is used to probe how well the artery dilates. Participants abstained from caffeine, alcohol, and strenuous activity before each session to prevent interference between measurements.

The exercise testing revealed clear and quantifiable deficits. Compared with controls, the long COVID group achieved a lower peak oxygen uptake, a lower predicted peak oxygen uptake relative to reference values, and a lower oxygen uptake at the first ventilatory threshold, the point during incremental exercise where metabolism begins to shift toward anaerobic contribution. These differences carried large effect sizes, indicating that the impairment was not marginal. Because both maximal and submaximal aerobic capacity were reduced, the results suggest that everyday activities, not just athletic performance, demand more physiological effort from these individuals than from their fully recovered counterparts.

The vascular findings were even more striking. Flow-mediated dilation was dramatically lower in the long COVID group, with a Cohen’s d of 2.30, an effect size that is exceptionally large for a between-group comparison in human vascular physiology. Crucially, the team also measured the shear-rate stimulus, the frictional force that flowing blood exerts on the vessel wall and that normally triggers the endothelium to release nitric oxide and other vasodilators. Peak shear rate, the area under the shear-rate curve during the hyperemic period, and the area under the curve up to maximal vasodilation were all significantly reduced in the long COVID participants. This means the problem involves not only a blunted vasodilatory response but also a diminished hyperemic signal, a double hit to the mechanisms that normally match blood flow to metabolic demand.

The researchers took care to ensure that the difference in flow-mediated dilation was not an artifact of differing artery sizes. Allometric scaling of log-transformed baseline and peak diameters showed no significant group-by-diameter interaction, and the between-group difference remained significant after adjustment for baseline diameter. The difference also persisted after statistical adjustment for age, sex, body composition, and comorbidities, strengthening the conclusion that the vascular impairment is a genuine feature of the long COVID phenotype rather than a byproduct of the higher body mass index and obesity prevalence observed in this group.

Perhaps the most informative result came from the correlation analyses. Within the long COVID group, flow-mediated dilation was positively and significantly associated with both peak oxygen uptake and predicted peak oxygen uptake, meaning that individuals with better endothelial function tended to have better aerobic capacity. No such association appeared among the controls, whose vascular function was uniformly good. This pattern supports the idea that, in long COVID, the state of the peripheral vasculature becomes a rate-limiting factor for exercise performance, constraining oxygen delivery to skeletal muscle in a way that does not occur in healthy vasculature with reserve capacity to spare.

The heart, meanwhile, told a different story. Conventional echocardiographic measures, including left ventricular ejection fraction, volumes, and global longitudinal strain, were indistinguishable between groups, as were the myocardial work indices derived from pressure-strain analysis: global work index, global constructive work, global wasted work, and global work efficiency. Myocardial work integrates ventricular deformation with afterload and can detect subtle dysfunction that conventional parameters miss, yet even this sensitive approach found no group differences. Exploratory analyses did reveal intriguing internal patterns: in the long COVID group, lower global work index was associated with greater ventilatory drive and poorer ventilatory efficiency during exercise, and with higher body mass index, hinting at interactions between metabolic status, breathing mechanics, and cardiac mechanics that deserve further study in larger cohorts.

Taken together, the data sketch a coherent physiological account of long COVID exercise intolerance. Persistent endothelial injury after SARS-CoV-2 infection, plausibly driven by chronic inflammation, oxidative stress, and reduced nitric oxide bioavailability, compromises the ability of conduit and resistance vessels to dilate appropriately during exercise. The result is a peripheral limit on oxygen delivery that manifests as reduced aerobic capacity, while the central pump continues to function within normal limits. This peripheral mechanism aligns with earlier studies conducted one and two years after infection, but the present work extends the timeline considerably, showing that these abnormalities persist well into the third year after the acute illness.

The authors are careful to note the limitations of their study. It was observational, cross-sectional, and single-center, with a modest sample size, so the findings are hypothesis-generating rather than definitive. The long COVID cohort consisted predominantly of people with previous moderate-to-critical illness, and the controls all had prior asymptomatic or mild infection, so the results may not generalize to the full spectrum of post-COVID conditions. The cohort was also predominantly composed of individuals with overweight or obesity, and the study was not powered to detect sex or race differences. Even so, the large effect sizes and the a priori power calculation lend credibility to the central conclusion. Clinically, the work argues for evaluating vascular health, not just cardiac function, in patients with persistent post-COVID exercise intolerance, and for cardiometabolic risk management alongside structured rehabilitation. Whether improving endothelial function can restore exercise tolerance remains a critical question for future interventional trials.

Subject of Research: Persistent endothelial dysfunction and exercise intolerance in long COVID three years after SARS-CoV-2 infection

Article Title: Persistent endothelial dysfunction is associated with exercise intolerance in adults despite preserved myocardial work 3 years after SARS‐CoV‐2 infection

Article References: de Araújo Alves, C. C., Goulart, C. D. L., D'Ávila, L., de Souza Silva, B. H. E., Alves, I. M., Aguiar, B. V. S., Conde, C., Stein, R., & Cipriano, G., Jr. (2026). Persistent endothelial dysfunction is associated with exercise intolerance in adults despite preserved myocardial work 3 years after SARS‐CoV‐2 infection. Physiological Reports, 14(19), Article e71103. https://doi.org/10.14814/phy2.71103

Image Credits: AI Generated

DOI: 10.14814/phy2.71103

Keywords: long COVID, endothelial dysfunction, flow-mediated dilation, exercise intolerance, myocardial work, cardiopulmonary exercise testing, SARS-CoV-2, vascular function, peak oxygen uptake, echocardiography, nitric oxide, post-COVID condition