Aortic coarctation, hypertension, and what machine learning now tells us about long-term risk

If you were born with aortic coarctation -- a narrowing of the main artery leaving your heart -- and you have had it repaired, you may have been told the problem is solved. Structurally, the repair is real. But the cardiovascular story rarely ends there. A significant number of people with repaired coarctation develop persistently high blood pressure, and some develop restenosis (a re-narrowing at or near the repair site) years or even decades later. A 2025 study published on PubMed took a careful look at this problem and asked a sharp question: can machine-learning models predict which patients will develop hypertension or restenosis while they are being managed under current guidelines? (https://pubmed.ncbi.nlm.nih.gov/42389462/) The answer, it turns out, is promising -- and it has real implications for how adults with congenital heart disease should be followed over a lifetime.

At NoMi Beach Health, we work with adults who have complex cardiovascular histories and who need a primary care partner who can keep up. This post breaks down what coarctation is, why the post-repair period is so often underestimated, and what the latest predictive modeling research means for you in practical terms.

What coarctation actually does to your blood pressure

Aortic coarctation is a congenital narrowing of the aorta, usually located just past the left subclavian artery near the site where the ductus arteriosus closes after birth. The narrowing forces the left ventricle to generate higher pressure to push blood through the bottleneck, which drives up blood pressure in the upper body. Below the narrowing, perfusion pressure drops -- which is why a blood pressure difference between the arms and legs is one of the classic diagnostic clues.

Repair -- whether by surgical resection, patch aortoplasty, or catheter-based stenting -- removes or dilates the structural obstruction. But the vascular biology upstream does not simply reset. The aortic wall proximal to the original coarctation has been exposed to years of elevated pressure and develops lasting changes in stiffness and compliance. The renin-angiotensin-aldosterone system, activated for years by reduced renal perfusion, does not switch off cleanly after repair. Collateral vessels that formed to bypass the narrowing persist and alter hemodynamics. The result is that a meaningful proportion of patients -- estimates in the literature range from 30 to 80 percent depending on the cohort and the definition of hypertension used -- have persistently elevated blood pressure after technically successful repair (https://pubmed.ncbi.nlm.nih.gov/22840753/).

This is not a niche problem. Adults with repaired coarctation carry a meaningfully elevated risk of premature coronary artery disease, aortic aneurysm, stroke, and heart failure compared to the general population (https://pubmed.ncbi.nlm.nih.gov/26673315/). Blood pressure control is one of the most modifiable levers in that risk picture.

Why guidelines alone are not enough -- and what the new research adds

Current guidelines from the American Heart Association, the American College of Cardiology, and the European Society of Cardiology recommend lifelong surveillance for adults with repaired coarctation (https://doi.org/10.1016/j.jacc.2018.08.1029, https://doi.org/10.1093/eurheartj/ehaa554). That includes periodic imaging of the repair site, ambulatory blood pressure monitoring -- not just office readings, which routinely underestimate post-coarctation hypertension -- and cardiology review at a center with adult congenital heart disease (ACHD) expertise.

What the guidelines cannot do on their own is tell you, at the individual patient level, how likely you are to develop hypertension or restenosis over the next five or ten years. That is the gap the 2025 machine-learning study set out to close.

The researchers developed ML models trained on clinical, hemodynamic, and imaging variables from coarctation patients managed under guideline-directed care, then validated those models in separate cohorts. The models integrated factors such as age at repair, pre-repair and post-repair pressure gradients, aortic arch geometry, measures of aortic stiffness, left ventricular mass, and collateral burden. When tested against conventional regression approaches, the ML models showed improved discrimination -- meaning they were better at correctly identifying which patients would and would not develop the outcomes of interest.

This matters because the variables that matter most in a machine-learning model are not always the ones clinicians intuitively weight most heavily. Aortic stiffness, for example, has been independently associated with residual hypertension in coarctation patients in earlier work (https://pubmed.ncbi.nlm.nih.gov/23812270/), but it is not consistently captured in routine follow-up. A model that surfaces stiffness as a high-weight predictor gives clinicians a reason to measure it systematically.

What this means for adults with repaired coarctation

If you had coarctation repaired in childhood and you are now an adult, here is what the evidence suggests you should expect from your care team -- regardless of whether predictive models are part of your center's workflow yet.

Ambulatory blood pressure monitoring matters more than office readings. Exercise-provoked hypertension is common in this population even when resting readings are normal. A 24-hour ambulatory blood pressure monitor or a standardized exercise blood pressure test can catch what an office cuff misses. The 2018 AHA/ACC ACHD guidelines specifically flag this.

Imaging of the repair site should be periodic and protocolized. Restenosis can be gradual and clinically silent for years. Cardiovascular magnetic resonance (CMR) or CT angiography of the aorta, on a schedule agreed with your ACHD cardiologist, is the standard for catching it early.

Medication for post-coarctation hypertension is not the same as generic hypertension management. The neurohumoral component -- driven by persistent renin-angiotensin activation -- makes ACE inhibitors and angiotensin-receptor blockers (ARBs) logical first choices in many patients. Beta-blockers are often added when there is aortic dilation or a residual gradient. The right combination depends on your specific hemodynamic picture, which is why management should involve a cardiologist, not just a blood pressure target.

Your primary care clinician should know your coarctation history in detail. This sounds obvious, but it is frequently a gap. Adults who were repaired as children often have their congenital heart disease history siloed in a pediatric record that never transfers cleanly to adult care. A concierge primary care clinician who has actually read your operative report, your most recent echo, and your gradient data is a fundamentally different resource than one seeing your "history of CHD" as a checkbox on a problem list.

The role of predictive modeling in everyday clinical care

Machine-learning models for cardiovascular risk prediction are not yet part of standard outpatient workflow at most practices. The 2025 coarctation study is a development-and-validation paper -- an important step, but not yet a deployed clinical tool. What it represents, along with a broader shift in cardiovascular medicine, is a move toward individualized risk stratification rather than one-size-fits-all guideline application.

We have seen this pattern in other cardiovascular contexts. Blood pressure variability -- not just average blood pressure -- has emerged as an independent predictor of outcomes across antihypertensive regimens, a topic we explored in depth at [/blog/blood-pressure-variability-and-outcomes-across-antihypertensive-regimens]. The principle is similar: aggregate-level data tells you what to do on average; individual-level modeling tells you what to do for this person, with this history, at this point in time.

For coarctation specifically, the variables that will likely matter most in a predictive framework are ones that require deliberate measurement -- aortic stiffness indices, arch geometry on cross-sectional imaging, ambulatory blood pressure profiles, exercise gradients. If your current follow-up does not include these, the predictive models will have nothing useful to run on. The research is only as actionable as the data collection behind it.

Coordinating long-term care -- what we do at NBH

Adults with repaired coarctation need two things that do not always exist in the same place: a cardiologist who specializes in adult congenital heart disease for structural management, and a primary care clinician who can manage everything in between and make sure the pieces connect.

At NoMi Beach Health, Dr. Jezwah Harris (NP, JD, MBA, FNP-BC, MEP-C) works with adults who have complex cardiovascular histories -- including congenital conditions managed in childhood that are now showing up as adult hypertension, fatigue, or exercise intolerance. We take the time to actually read your records, understand your repair history, coordinate with your specialists, and manage your blood pressure with the nuance your specific hemodynamics require. We also track guideline updates so that your surveillance schedule reflects current evidence, not a protocol written a decade ago.

This kind of care does not happen in a 12-minute appointment. It happens when your clinician knows your history deeply enough to flag that your ambulatory blood pressure looks different this year, or that your most recent imaging is overdue, or that the medication combination you are on may not be the right one for your anatomy. If you want to understand more about how we approach complex primary care, our concierge primary care guide lays out the philosophy in full.

Long-term cardiovascular outcomes in coarctation are not fixed at the time of repair. They are shaped by decades of blood pressure management, imaging surveillance, and coordinated care. The new predictive modeling research reinforces something clinicians who follow this population have known for a while: the repair is the beginning, not the end.

---

If you have a history of aortic coarctation -- repaired or not -- and you want a primary care partner who will actually engage with your cardiovascular complexity, we are ready to help. Visit our primary care services page to learn what a new-patient visit looks like, or call us at (786) 744-5152. We will review your records before your first appointment, ask the right questions, and build a monitoring plan that keeps up with where the evidence is going.