The ATTR-CM Treater Is Often in the Same Building. Many Patients Never See One.

Most demand reporting in ATTR-CM starts at the prescription. New-to-brand counts, specialty pharmacy data, hub enrollments, and prescriber deciles all begin when a provider writes one. A patient who is diagnosed and never prescribed doesn't appear in any of them.

We started at the diagnosis instead. We followed every patient whose first wild-type ATTR-CM diagnosis code appeared between June 2023 and June 2025, with no prior therapy. To leave out codes that may have only meant a suspected case, we kept only patients whose diagnosis was coded a second time at least a month later, or coded after a scan. We also kept only patients who were still in the data a year later. Then we sorted them by the clinician who made the diagnosis, not the one who wrote the script.

About one in four started therapy within a year. Two in three were diagnosed by a first-time diagnoser, a clinician who had never treated ATTR-CM. Their patients started at 23%. Patients diagnosed by a clinician who had treated five or more started at 40%, and clinicians with one to four fell in between. The problem isn't denials at the pharmacy. For about three in four of these patients, there is no claim for therapy at all, paid or rejected.

When these patients are treated, someone else usually writes the prescription

When one of these patients does start therapy, the first prescription usually comes from someone else. The diagnosing clinician writes it only about one time in ten. Most of the rest are written by a treater, meaning a clinician who has prescribed ATTR-CM therapy before.

So the key question is whether the patient ever reaches a treater. Just under half do. Those who reach one start therapy at about one in three. Those who never reach one start at about one in six. The two groups were the same age.

You might wonder whether experienced clinicians just get easier patients, ones already worked up somewhere else. That doesn't seem to be it. Their patients who arrived with no outside cardiology visits or amyloid testing in the prior year started therapy at nearly the same rate as the rest of their patients.

The neuropathic form works differently. There, the diagnosing doctor often treats the patient too. A neurologist seeing their first case writes the first prescription about half the time. A cardiologist seeing their first cardiomyopathy case writes it fewer than one time in five. A cardiologist who has treated five or more writes it about three times in four.

It matters which treater the patient sees, and how soon

Seeing a treater is not enough on its own. Here we counted only visits coded for amyloid, so these were visits about the disease, not a routine checkup. Patients who saw a treater with one to four prior ATTR-CM patients started therapy at 19%. That is barely more than the 16% for patients who saw no treater at all. Patients who saw a treater with five or more started at 41%, about the same as when an experienced clinician makes the diagnosis (40%).

A fair question is whether doctors simply refer the patients they already plan to treat. If that were true, any referral would help. It barely did. Patients sent to a less experienced treater were treated only slightly more often than patients who were never sent anywhere. Patients sent to an experienced treater were treated more than twice as often.

Timing matters too. The sooner patients saw a treater, the more likely they were to start. Among patients who saw one within two months of the diagnosis, about 27% started therapy in the six months after that visit. For patients first seen at two to four months, it was 19%. At four to six months, it was 11%.

Why: first-time diagnosers rarely start anyone on therapy, and the patients themselves don't explain the gap

We followed clinicians who made their first ATTR-CM diagnosis in the first half of 2024 and made no other new diagnosis through mid-2025. By then, fewer than one in ten had started anyone on ATTR-CM therapy. The few who did start a patient usually kept going. Just over half went on to treat another patient, likely one referred to them. The first prescription is the step most of them never take.

The difference isn't about testing. In both groups, patients who had both key tests were the most likely to start. Those are the nuclear scan and the blood test that rules out the other type of amyloid. But even with both tests done, the gap remained. First-time diagnosers' patients started at 27%. Experienced clinicians' patients started at 49%.

The difference isn't about the patient or the setting either. Heart failure history, atrial fibrillation, where the diagnosis was made, and the diagnosing clinician's specialty each changed the treatment rate by only a few points. The diagnosing clinician's experience changed it by about seventeen points.

Untreated patients of first-time diagnosers were somewhat more likely to have a visible reason not to treat, such as a nursing-facility stay, dialysis, or dementia. Even after removing those patients, the gap stayed about the same.

Missing pharmacy data doesn't appear to explain the difference either. We checked four ways: using only patients and clinicians whose prescriptions show up well in the data, looking for signs of earlier fills we couldn't see, looking for breaks in refill history, and checking the office-given injectable, which is billed as a medical claim instead of through a pharmacy. None of these accounted for the gap.

A treater on staff changes nothing unless the patient is sent to them

Having a treater on staff didn't help by itself. Take a patient diagnosed by a first-time diagnoser at an organization with a treater on staff. If that treater never saw the patient, the chance of starting therapy was 10.3%. At organizations with no treater at all, it was 10.8%.

The difference was whether the organization sent these patients on to a treater, either in-house or somewhere else. At organizations that sent none of them, 8% started therapy. At organizations that sent more than half, 27% did. This was true at both small and large organizations.

The class grew in 2025. Patients who never saw a treater barely moved.

Patients diagnosed in early 2025 had more treatment options than those diagnosed in early 2024. A second oral stabilizer launched, the injectable gained a cardiomyopathy label, and the Part D out-of-pocket cap took effect.

For first-time diagnosers' patients who saw a treater, treatment rates rose a lot, from 19% to 34%. Part of that rise may not be new treatment. Before the Part D cap, some Medicare patients got these drugs through free-drug programs, and free drug doesn't show up in claims. Once the cap lowered what patients pay, some of them moved into regular coverage, where we can see them. Both this shift and the new options likely played a part.

For patients who never saw a treater, almost nothing changed. Their treatment rate went from 13% to 15%. The share of patients who reached a treater didn't go up at all. And the gap between first-time and experienced diagnosers' patients stayed the same.

A real-life account example

Two hospitals in Florida, in two different health systems, each had ten patients diagnosed with ATTR-CM by a clinician who had never treated it. Both hospitals have ATTR-CM treaters on staff. Both perform the diagnostic scan in-house.

At the first hospital, all ten patients had a completed workup, seven were diagnosed by a cardiologist, and six had been hospitalized for heart failure before the diagnosis. None started therapy within a year. The hospital's own treaters saw none of them. Six were later seen at an organization with a treater, four of them inside the same system, and none were treated there either.

At the second hospital, half had been hospitalized for heart failure. Every patient was seen at an organization with a treater, a median of 18 days after the diagnosis. Six of the ten saw one of the hospital's own treaters, and five of those six started therapy. Of the four who didn't see one, one started. A treater inside the hospital wrote most of the prescriptions.

The split holds at each system's other sites. At the first system's other hospital, one of 15 such patients started. At the second system's two other sites, about four in ten did. We checked whether missing specialty pharmacy data could explain the difference between the two hospitals. It doesn't.

What this is worth and who owns it

Anyone working in ATTR-CM knows the disease is underdiagnosed and that a small group of experienced clinicians writes most of the prescriptions. What the data adds is where the diagnosed and untreated patients come from, and that the growth in the class hasn't reached them.

A scan let almost any clinician diagnose ATTR-CM. The treatment decision stayed with the few who treat it.

About a thousand new first-time diagnosers appear every six months, and the number is growing. Each has about one patient. Many are cardiologists and many are not. None appear on a prescriber list, because they have not prescribed. The experienced treaters are already on most call plans.

Not every patient should start therapy. Even experienced clinicians start fewer than half of the patients they diagnose. So the realistic aim isn't to treat everyone. It's for patients of first-time diagnosers to be treated about as often as patients of experienced clinicians.

None of this depends on which therapy the patient ends up on. The most useful step is making sure the patient sees an experienced treater within two months of being diagnosed. The organizations that don't send these patients on are visible in claims. For about half of these patients, a treater already works in the same organization.

What else the quarter covered

The same scan looked at where the injectable gets started now that it carries a cardiomyopathy label. Hospitals started it in fewer than one in four new patients and put the rest on an oral. Offices started it in four in ten.

What predicted it was whether the site already gave injections. Offices that had given patients another injected or infused drug in the past year started the injectable about half the time. Offices that hadn't given one started it about one time in five. Offices that already gave the office-administered cholesterol siRNA started it in more than eight in ten.