A Bleeding Source With No Strong Drug Behind It
A single patient whose bleeding source has no strong medical treatment behind it at all. The disagreement is honestly about whether a weak option is still worth trying when nothing stronger exists.
B.K., a 71-year-old retired church organist, has needed a blood transfusion roughly every six weeks for the past year, chronic enough now that his hepatologist and he have both started scheduling around it the way one schedules around a recurring appointment rather than an emergency. His cirrhosis, from decades of well-controlled but real alcohol use he stopped nine years ago, has produced gastric antral vascular ectasia — the classic "watermelon stomach" endoscopic pattern of longitudinal red stripes radiating from the pylorus — identified as the actual bleeding source after his previously banded varices were confirmed not to be the culprit on repeat endoscopy.
GAVE is mechanistically distinct from the portal hypertensive gastropathy that sometimes gets confused with it — GAVE is a primary vascular ectatic lesion, not simply a consequence of elevated portal pressure the way PHG is, and that distinction matters directly for treatment, because it means the nonselective beta-blockade that helps portal hypertensive gastropathy has little proven benefit here; some series even suggest NSBBs may worsen GAVE-related bleeding, an unresolved but real enough concern that it's shaped how his own regimen has been built.
Endoscopic argon plasma coagulation is the actual mainstay of GAVE treatment and B.K. has had it multiple times, with symptomatic but incomplete and temporary benefit each time — the lesions recur. The honest state of medical (non-endoscopic) therapy for GAVE is that it's genuinely weak: older hormonal approaches (estrogen-progesterone combinations) have been largely abandoned for poor efficacy and real hormonal side-effect burden, and octreotide has scattered case-series-level evidence of benefit, not a randomized trial base, in a mechanism — broad gut-hormone and possibly direct vasoregulatory suppression — that remains more plausible than proven for this specific lesion. The APC sessions themselves aren't free of their own real cost either: each requires sedation and a half-day of his time, in a 71-year-old already managing a chronic illness on a fixed retirement income, which is part of why a medical option that could even modestly space out how often he needs the procedure carries real practical value beyond the transfusion numbers alone.
Weighing a weak option against no option at all
I want to be honest about where the evidence actually sits: nothing in medical therapy for GAVE is strong. The old hormonal approaches are largely abandoned, and octreotide's own evidence is case-series level, not a randomized trial.
Given that, and given he's transfusing roughly every six weeks with incomplete benefit from repeated APC, I'd still add octreotide as an adjunct. Even a modest, unproven chance of reducing his bleeding frequency is worth trying when the alternative is continuing exactly what we're doing now.
I don't disagree about how limited the options are, and I take his transfusion burden seriously.
What I'd push back on is treating "nothing else works better" as itself sufficient justification to add a real drug, with a real side-effect and injection burden, to an already frail 71-year-old with Child-Pugh B cirrhosis. Case-series evidence in a condition like this is exactly the kind of signal that's looked promising before and not held up under closer study — I'd want a higher bar before adding therapy rather than continuing supportive care.
That's a genuinely fair caution, and I won't claim the octreotide evidence is stronger than it is just because I want another option to offer him.
I'd propose a time-limited trial — twelve weeks of octreotide with transfusion frequency tracked explicitly as the outcome measure — rather than an open-ended addition. If it doesn't measurably reduce his transfusion needs in that window, we stop it and don't carry an unproven therapy forward indefinitely.
A 12-week, explicitly time-limited trial of octreotide added to his continued APC sessions and supportive transfusion, with his transfusion frequency tracked directly as the outcome measure rather than a subjective symptom read.
Left genuinely unresolved: whether the trial will show a real effect distinguishable from natural variation in his bleeding pattern, given how thin the underlying evidence is — both physicians agreed the bounded trial design, not the drug itself, is what makes trying it defensible.