What is your approach to the management of chronic GI bleeding from AVMs in an elderly patient on DOAC for atrial fibrillation?

I would definitely strongly consider the left atrial appendage occlusion device in these patients. While usually these devices (such as Watchman) do require anticoagulation for about 45 days until the device has an endothelial layer form on it (we usually confirm with a CT scan or TEE), there are some high-risk bleeding patients in whom we can also try just a DAPT for that period. Newer generation Watchman devices, such as Watchman FLX, is safer from the peri-procedural complications standpoint. Needs to have a comprehensive risk/benefit discussion with the patient before proceeding with Watchman, though.

Watchman or other LAA occluders would be the best choice for stroke prevention if GI bleeding cannot be satisfactorily corrected, as can be the case with AVM’s. Watchman and Amulet don’t mandate anticoagulation immediately after implant. DAPT is an option in Watchman and the correct approach in Amulet. There is an ongoing trial of different approaches to stroke prevention post Watchman. SIMPLAAFY is evaluating DAPT vs half dose NOAC vs ASA alone, potentially giving us more choices as part of the trial (randomized) and according to results, once it’s completed.

AVMs in a fragile elderly patient can be as life-threatening as AFib. There are several endoscopic, angiographic, surgical, and pharmacological procedures to treat and "cure" the condition. That possibility should be considered and discussed with the patient, besides LAAO obliteration/closure.

Octreotide depot injections have worked well for some of my patients. Cost is definitely a concern, but I think there may be some coupons being offered to a select few patients by the pharmaceutical company itself. Requires some back and forth from the insurance company and the drug company.

This suggests an elevated HAS-BLED score, making LAAO a reasonable option; post-LAAO DAPT may need to be truncated if post-implant TEE or CTA reveals a sealed LAA with adequate endothelialization.

I am a little different with these patients than previous responses. If we plug the LAA patients should really be on ASA long term, and we know from recent studies that ASA has similar bleeding to Eliquis. We do not know the effectiveness of LAAO devices without ASA. Therefore, I would consider no AC (we know the life expectancy of these patients is low from registry data), LAA clipping if they are more robust (again, lack of data but no foreign device in the body), or consideration of the Factor XI trial.

The only time I would consider perc LAAO closure would be if I accidentally isolated the LAA during ablation and maintained the patient on full-dose AC.

There is evidence for thalidomide for recurrent GI bleeding from angiodysplasias (Chen et al., PMID 37913505). May be good to get a GI perspective on this use.