Long COVID Doubles Heart Disease Risk After Mild Infection: What a 1.2-Million-Patient Swedish Cohort Reveals About Sex-Specific Pathways

Of the 1,222,058 working-age adults whose primary-care records were tracked across Stockholm for up to four years, only about 9,000 ever received a clinical long-COVID diagnosis. None of them had been sick enough with acute COVID to be hospitalized. None had a documented cardiovascular problem before the pandemic began. Two-thirds were women. And it is this small, self-selected, outpatient-only cohort — the ones who went home, recovered slowly, and kept going back to their GPs — that has just produced one of the clearest signals yet that long COVID is, quietly, a cardiovascular disease.

A prospective cohort analysis from Karolinska Institutet, published in eClinicalMedicine on 2 April 2026, reports that women carrying a long-COVID diagnosis had just over twice the adjusted risk of developing new cardiovascular disease over roughly four years of follow-up compared with women in the same registry who did not. Men had approximately one-third higher risk. Both signals held after adjustment for age, socioeconomic factors and other conventional cardiovascular risk factors, according to Karolinska Institutet. The study is the largest community-based investigation to date of what happens to the hearts of non-hospitalized long-COVID patients, and the way the risk splits by sex — and by diagnosis type — is not what the existing acute-COVID literature would predict.

The Population and What It Excludes

The analysis draws on the Multimorbidity Integrated Registry Across Care Levels in Stockholm — the MIRACLE-S cohort — a primary-care dataset covering essentially the entire adult working-age population of the Swedish capital. Between October 2020 and January 2025, researchers followed 1,222,058 individuals aged 18 to 65, 656,864 of them women, per the Karolinska release. Anyone with a cardiovascular diagnosis in the five years before study entry was excluded, as was anyone hospitalized for acute COVID-19. The remaining group is the version of the pandemic most households actually experienced: a positive test, a rough week or two at home, and a presumed recovery.

Within that population, about 9,000 people — 0.7 per cent — subsequently received a clinical long-COVID diagnosis, with women accounting for roughly two-thirds of cases, according to Medical Xpress's summary of the paper. That prevalence is lower than the figures household surveys typically report, a point the study's authors concede is a limitation. But what the diagnosis code captures is the subset of long-COVID sufferers sick enough and persistent enough to bring their complaints into the formal health system and be labelled by a clinician — the population, in other words, that primary-care physicians actually see and manage. If long COVID were merely a label for general post-viral fatigue, you would expect the cardiovascular event rates in this group to track closely with the rest of the cohort. They do not.

The Numbers by Sex

Across the four-year follow-up, 18.2 per cent of women with long COVID experienced a new cardiovascular event, compared with 8.4 per cent of women without the diagnosis. Among men, 20.6 per cent of the long-COVID group had an event, versus 11.1 per cent of men without — the figures come directly from the Karolinska Institutet announcement and are repeated by EurekAlert's release and Medical Xpress.

Two things stand out from those raw percentages. First, the absolute male event rate is higher in both groups, which is what baseline epidemiology would expect for this age band. Second, the ratio between the long-COVID and non-long-COVID groups is much wider in women: roughly 2.2 to 1 in women (18.2 vs 8.4), but closer to 1.9 to 1 in men (20.6 vs 11.1). Once the statistical model adjusts for demographic and clinical confounders, the authors' summary figures tighten to just over twice the risk for women and approximately one-third higher for men, per Karolinska. In other words, the adjusted female signal is larger than the crude ratio suggests, and the male signal is smaller — a pattern consistent with women in this cohort being younger, healthier and more socioeconomically heterogeneous than the men, which dampens their baseline risk.

Lead author Pia Lindberg, a nurse and PhD student at the Department of Medicine, Solna, summarized the core finding directly. "Our results show that long COVID can be a risk factor for cardiovascular disease, even in younger people who were previously healthy," she told Karolinska's press service.

The Arrhythmia Signal

Zoom in from overall cardiovascular risk to specific diagnoses and the sex split sharpens. The standout finding is for cardiac arrhythmias — irregular heart rhythm, most commonly atrial fibrillation and related disorders — where women with long COVID had roughly three times the risk of receiving an arrhythmia diagnosis, and men had a 61 per cent higher risk, according to CIDRAP's reading of the paper. Coronary artery disease was elevated by about a quarter in both sexes.

Two other diagnoses — heart failure and peripheral arterial disease — showed a meaningful signal only in women, each running roughly 25 per cent higher than in female controls, again per CIDRAP. Stroke risk showed no clear association in either sex.

That pattern matters because it argues against a simple long-COVID-makes-everyone's-heart-worse reading. If the driver were generalized endothelial damage, you would expect stroke — a quintessential vascular-wall event — to rise in parallel with coronary disease. It did not. And if the mechanism were purely inflammatory, you might expect heart failure and arterial disease to track in men as well as women. They did not, at least not strongly enough to clear statistical significance in this cohort. Something about the female phenotype of long COVID is producing a particular combination of rhythm and structural cardiac problems that the male phenotype — as captured by this registry — is not.

Lindberg's own summary in the Karolinska release is that cardiac arrhythmias and coronary artery disease were the shared outcomes across both sexes. The rhythm and the arteries are where the long-COVID male and female patterns converge; the pump and the periphery are where the female phenotype pulls ahead.

Why the Gap Between Women and Men

The paper itself is careful not to claim mechanism — it is an epidemiological observation in registry data, not a physiology experiment. But Lindberg and her senior co-authors, cardiologist Artur Fedorowski and researcher Axel C. Carlsson, have published an explicit candidate-pathway discussion in a companion essay in The Conversation. Three mechanisms are on their shortlist, and each one interacts with sex biology differently.

The first is endothelial dysfunction — injury or chronic activation of the cells lining blood vessels. Women generally enjoy more favourable endothelial function than men before menopause, thanks to oestrogen-driven nitric oxide signalling. That protective advantage can be preferentially eroded by prolonged inflammation, which may help explain why the long-COVID-versus-control gap is wider in women: the disease erases an advantage the female cardiovascular system was depending on, while men, starting from a less protected baseline, see a smaller relative jump.

The second is persistent low-grade inflammation and immune dysregulation. Female immune systems tend to mount stronger antibody and cellular responses than male ones, which is part of why women are overrepresented in classical autoimmune disease — and part of why long COVID itself is more common in women, a pattern echoed in this cohort's two-thirds-female long-COVID population. If the arrhythmia and cardiomyopathy signals in women are being driven by autoimmune or autoantibody mechanisms targeting cardiac tissue, the stronger female response to any given viral antigen load becomes a liability rather than an asset.

The third is autonomic nervous system disruption, the dysautonomia that shows up clinically as postural tachycardia syndrome and similar disorders. Autonomic dysfunction is disproportionately diagnosed in women even outside the COVID context, and small mechanistic studies have repeatedly documented abnormal heart-rate-variability patterns in long-COVID patients. A nervous system that is miscalibrating vagal tone is a nervous system that can plausibly produce both arrhythmia, as a direct rhythm destabilization, and secondary structural changes over time that edge toward heart failure.

None of this is proven in the MIRACLE-S data — the registry has diagnosis codes, not mechanistic readouts. But the combination of findings the paper does describe — rhythm disturbance in both sexes, pump and peripheral-artery disease concentrated in women, stroke untouched — is at least consistent with a story where chronic inflammation plus dysautonomia interacts with female vascular and immune biology to produce a different cardiological phenotype than the mostly-arterial phenotype seen in men.

What This Study Does Not Tell Us — Yet

Any honest reading of the paper has to flag four limits.

1. Observational design. This is a registry analysis, not a randomized trial. Long-COVID diagnosis is not assigned at random. People who end up with the diagnosis code differ from people who don't in ways the model can only partially adjust for — attention to symptoms, healthcare-seeking behaviour, underlying comorbidities not captured in the five-year look-back. Association, as ever, is not causation.

2. Under-diagnosis of long COVID. The 0.7 per cent registry prevalence is lower than the range typically reported in Western population surveys. The 9,000 diagnosed long-COVID patients in this cohort are the subset who reached the formal clinical system and received a specific diagnostic code — and whose cardiovascular outcomes therefore reflect the worst-affected rather than the median post-COVID patient.

3. The non-hospitalized frame. The study excluded anyone who had been hospitalized for acute COVID. That makes the findings relevant to the large majority of infected adults — the rough-week-and-back-to-work population — but says nothing direct about the sicker acute-COVID cohort, for whom prior studies have already documented much larger cardiovascular sequelae.

4. Vaccination and variant are not isolated. The follow-up window spans the wild-type, Delta, Omicron and post-Omicron eras, and covers a population in which Sweden rolled out mass vaccination concurrently. Teasing apart the independent contribution of vaccination status, reinfection count and dominant variant to the cardiovascular signal is not something this study design was built to do, and the authors do not claim otherwise.

Implications for Primary Care

The authors frame the clinical takeaway pragmatically. "Cardiovascular risk screening of individuals with long COVID should not be limited to previously hospitalised patients but may be warranted in broader community settings," is how CIDRAP summarizes the paper's stated conclusion. In practice that would mean a GP encountering an otherwise healthy 40-year-old with persistent post-COVID symptoms considers a baseline ECG, lipid panel and — if arrhythmia is suspected — an ambulatory rhythm monitor, rather than defaulting to watchful waiting.

Lindberg explicitly flags sex-specific follow-up as an open gap. "This underlines the need for structured follow-up that takes gender differences into account, particularly as cardiovascular disease in women often presents with more diffuse symptoms," she said in the Karolinska release. The concern here is not abstract. Women with diffuse post-COVID fatigue, breathlessness and palpitations sit at the intersection of two under-diagnosis problems. Long COVID has historically been taken less seriously when the patient is a middle-aged woman; cardiovascular disease, especially non-ST-elevation coronary disease and heart failure with preserved ejection fraction, has historically been under-recognized in women. Stack those two biases and the specific patient group at highest relative risk in this study — women with long COVID and new cardiac symptoms — becomes the group most likely to be told their problems are psychosomatic.

There are four practical questions systems will have to answer.

First, whom to screen. If only 0.7 per cent of the registry carried a long-COVID diagnosis but cardiac-event risk was elevated across that group, is the right triage point the clinical code or the symptom pattern? A stricter code will miss people; a broader symptom-based criterion will over-screen. Either choice has cost and equity implications.

Second, what to screen for. The paper's sex-stratified pattern suggests cardiology workup for post-COVID women should not stop at coronary assessment. Rhythm monitoring and heart-failure-preserved-EF evaluation are on the differential; peripheral arterial disease is worth a pulse check even in a 40-year-old non-smoker — a combination most primary-care protocols do not currently index.

Third, for how long. The follow-up is about four years. Whether the excess risk plateaus, accumulates, or reverses with time is unknown. Systems designing long-COVID cardiology clinics will need a plan for indefinite review rather than a fixed discharge milestone.

Fourth, what counts as treatment. The paper is silent on intervention. No drug, rehab protocol or lifestyle regime has been shown in a randomized trial to reduce the long-COVID cardiovascular hazard documented here. Clinicians asked by patients what to do will be reaching for conventional secondary-prevention tools — statins, antiplatelets, rhythm control, beta-blockers where indicated — on extrapolation rather than direct evidence.

Why This Study Should Change the Conversation

The cardiovascular-sequelae literature on COVID-19 has, until now, been dominated by large administrative datasets from the United States Veterans Affairs system and similar hospital-centric registries, which are heavily male, heavily older, and heavily weighted toward severe acute illness. Those studies established the qualitative point — yes, COVID-19 raises cardiovascular risk — but they left two gaps. They could not speak directly to people who never saw the inside of a hospital, and they could not say much about women, who are underrepresented in their source populations.

The MIRACLE-S cohort fills both gaps. It is a community-level, primary-care-registered, majority-women-in-the-exposure-group study that confirms the earlier VA-era finding in a cleaner population, and then adds a structure to it: women are not just at elevated risk, they are at roughly twice the adjusted risk of men after confounder control, and the specific cardiac entities driving that excess — arrhythmia, heart failure, peripheral arterial disease — differ from the male pattern. That is an editorial upgrade, not a replication.

It is also an uncomfortable finding for health systems that have spent 2025 and 2026 treating long COVID as a resourcing nuisance rather than a cardiovascular-risk category. A twofold hazard in a population of working-age adults is not a rounding error; extrapolated to any mid-sized city, it translates into a public-health problem with a long tail measured in decades of potential life-years.

Key Takeaways

• In the 1.2-million-adult MIRACLE-S registry, women with a clinical long-COVID diagnosis had just over twice the adjusted four-year cardiovascular risk of women without the diagnosis; men had roughly one-third higher risk — none of these patients had been hospitalized for acute COVID, per Karolinska Institutet.
• Absolute four-year event rates reached 18.2 per cent in long-COVID women versus 8.4 per cent in controls, and 20.6 per cent in long-COVID men versus 11.1 per cent in controls.
• The strongest specific signal was for cardiac arrhythmias — roughly threefold in women and 61 per cent higher in men — with heart failure and peripheral arterial disease elevated only in women, per CIDRAP.
• Stroke risk was not elevated in either sex, arguing against a generalized vascular-damage explanation.
• The authors propose three candidate mechanisms — endothelial dysfunction, chronic inflammation, and autonomic nervous-system disruption — none of them proven in this study, all of them with plausible sex-differential effects, as discussed by Lindberg, Fedorowski and Carlsson in The Conversation.
• The clinical implication is that primary-care-level cardiovascular screening for long-COVID patients — not just the hospitalized ones — deserves structured follow-up with sex-specific protocols.