Romosozumab After a Prior Cardiac Event: Weighing the Boxed Warning
A single patient whose fracture risk and cardiovascular history point in opposite directions on the same drug. The question isn't whether romosozumab works — it does — it's whether her cardiac history rules it out for her specifically.
Dolores R., a 71-year-old woman, has spent the past three years dividing her time between cardiac rehab classes and babysitting her granddaughter two afternoons a week, proud that the stents placed after her myocardial infarction three years ago have held up well enough that her cardiologist cleared her for the rehab program's most demanding track. Her sister, the same age, broke a hip from what she described as barely a stumble, and that is what finally moved Dolores to get the DXA her own doctor had suggested twice before. It came back severe: a femoral neck T-score of −3.2, with no fracture of her own yet but a FRAX-estimated 10-year hip fracture risk of 6.1 percent — double the 3 percent threshold at which treatment is conventionally recommended, and reached without her ever having broken anything.
Romosozumab is the agent her endocrinologist would ordinarily reach for at that severity, given its dual anabolic-then-antiresorptive action, and it is also the one agent in the room carrying an FDA boxed warning. The warning descends from ARCH, which compared romosozumab directly against alendronate and found adjudicated cardiac ischemic events in 2.5 percent of the romosozumab arm against 1.9 percent of the alendronate arm at twelve months. Two details about that signal bear on her specifically and pull in opposite directions. The first is that FRAME, romosozumab's placebo-controlled trial, found no such difference — 0.8 percent in both arms — which makes the ARCH result at least partly a question about whether alendronate looked protective rather than romosozumab looked harmful. The second is that the label records those ARCH events as having occurred in patients both with and without a prior myocardial infarction or stroke, so the signal does not confine itself tidily to her subgroup the way a cardiologist reading her chart might assume.
Her own timeline sits outside the warning's operative clause. The label bars initiation within one year of a myocardial infarction or stroke; hers was three years ago, her stents have functioned, and her most recent stress test was reassuring. Nothing in the boxed warning forbids her this drug. What the warning does instead is leave the decision entirely to judgment, at a distance from the event where the label has stopped speaking and the trial data cannot say whether three years is different from one.
Reading the boxed warning against her actual timeline
I'd steer away from romosozumab here. ARCH randomized patients to romosozumab or alendronate and found a real excess of adjudicated cardiac ischemic events in the romosozumab arm — 2.5 percent against 1.9 percent at twelve months. One caveat I'd state plainly rather than lean on: the label records that those events occurred in patients both with and without a prior myocardial infarction or stroke, so this isn't a signal that cleanly confines itself to her subgroup.
She doesn't meet the label's strict one-year contraindication window — her MI was three years back — so this isn't a hard stop by the label's own language. It's a judgment call about how much residual caution a three-year-old event still deserves, and my honest answer is more than zero.
Worth naming directly: FRAME, romosozumab's placebo-controlled trial, did not find the same cardiovascular signal ARCH did — the difference showed up specifically against an active bisphosphonate comparator, not against no treatment. That doesn't make the signal fake, but it does make it harder to read as romosozumab causing harm outright rather than alendronate happening to look protective in that comparison.
I take the "not a hard stop by the label" point seriously — that's a real distinction, not one I'd wave away for someone at her severity of fracture risk with a comparably effective alternative sitting right there.
Then don't make her the test case for how the signal should be read. Zoledronic acid has no cardiovascular signal in any trial and is a genuinely reasonable first agent at her risk level — not as good as romosozumab's dual action on paper, but not a controversy either.
If her fracture trajectory doesn't improve on the bisphosphonate, romosozumab is still available to revisit later with a clearer sense of how her cardiac status is trending — this isn't a door closing, it's picking the option that doesn't require anyone to resolve today whether three years is long enough.
Agreed: start zoledronic acid now, with a documented plan to reassess fracture trajectory and revisit romosozumab if the bisphosphonate proves inadequate at her severity.
Not agreed: whether three years past a myocardial infarction is functionally equivalent to the label's one-year cutoff for future purposes — the cardiologist and endocrinologist left with different comfort thresholds, deliberately not forced to a shared number today since the immediate choice didn't require resolving it.