I like this case for the issues that you have raised.
Firstly, I would suggest that if you have not stented the VA, this lesion is still the likely culprit. Also, complete arteriogram may be needed to evaluate for adequate arterial inflow. Also, I would make certain the arterial anastomosis is nicely patent. Sometimes, this connection needs to be dilated without an obvious lesion on angiography. The stenosis can be difficult to see with fluoroscopy, but it will almost always be apparent on an angioplasty balloon.
Secondly, does this patient have adequate cardiac reserve. If he has systolic dysfunction, he may not be able to support the blood flow through the graft. If you can assess blood flow volumes through the graft, this might assist in assessing the arterial inflow when the graft is fully patent.
Thirdly, there is no good data on anticoagulation and graft thrombosis. There have been some suggestions that clopidogrel may be helpful, but this was really a trend and has not been rigorously studied. There is similar weak data for Aggrenox.

I agree that it ought to be a remanent stenosis in the vascular circuit.
However, it is curious to place a graft in a patient with a known coagulopathy. I am not very familiar with PTFE grafts, but in such case, I would place a fistula, with a surgery without stopping the coumadine obviously.

A good case to brain storm. I agree with previous comments about ensuring that this is not a recurrent venous anastomotic stenosis, though rather early.
The question really is- what are non stenotic causes of access thrombosis. Poor EF is certainly a cause and so is coagulopathy and hypotension. High hemoglobin has historically been thought to cause access thrombosis. What is not generally considered is high right heart pressure which can stagnate venous return. I have often felt that pulmonary hypertension is an important cause of thrombosis, especially when access has marginal flow. Does this patient, with h/o clinical (and possibly subclinical) PE due to coagulopathy have high right sided pressure? If he continues to have recurrent thrombosis without apparent cause, it might be reasonable to use a pressure transducer catheter to investigate the site of pressure gradient. Just a thought.

I cannot sleep - "Why do grafts fail?"
How many leaves are in the forest?

Thank you for your comments.
At the 2nd angioplasty, I placed a 7mm diameter balloon at the arterial anastomosis which did not show stenosis at arterial end. He has 30% EF and has AICD in place. No pulmonary hypertension. He had 2 AVF (right radiocephalic and right brachiobasilic) attempted which thrombosed spontaneously within 24 hrs of creation.

Should I place a catheter next?
Shoud apixiban be used instead of coumadin?

Agree with everyone's comments.

Given that you had a previously successful thrombectomy with a venous anastomoses PTA, the simplest explanation would be a recurrent lesion here.

Sometimes after you've initially opened a venous anastomosis lesion , a "soft" recoil can occur, which may be difficult to ascertain during a thrombectomy (with the presence of clot and minimal balloon waisting.). This is usually best managed with a venous anastomosis stent graft ( I personally use the Viabahn). This may be your simplest management option here, and you have little downside in attempting a stent-graft. If this works, you've saved the patient from a catheter and new access.

All the other confounding issues mentioned can be challenging to manage. However you already have a therapeutic INR, not sure any other anti-coagulant would be of benefit.

" He had 2 AVF (right radiocephalic and right brachiobasilic) attempted which thrombosed spontaneously within 24 hrs of creation."
Where these fistulas created with an INR of 2 or a little more? For me, any thrombosis of a fistula must have an explanation!
Since I understand that even with healthy vessels, a thrombosis can occur, I give coumadine immediately in such cases. And the patient is reoperated with an INR between 2 and 2,5, rarely a little more. The problem can be the opinion of the anesthesiologist! He can refuse to perform an axillary anesthesia, and ask for a local anesthesia.
Tomorrow, I have to transpose a basilic vein in such conditions.
Knowing this technique, any patient coming to me with an already coumadine treatment for whatever reason is operated in the same conditions, avoiding the risk of hypercoagulability, and the risk of early thrombosis! The surgery is a little longer, but I have very few thrombosis every year, and not too much hematomas.
We must not "attempt" to create a fistula, but simply create a fistula.

Thierry my friend - it has been too long - I am now with "The French Girl". We will be in Lille with my daughter for Thanksgiving, and I hope to introduce you.
Sometimes the explanation is ill-advised surgery, or an inexpert surgeon - this may be why we have a 30-40% failure rate in fistula maturation in the US, YOU do not have such a failure rate, I do not have such a failure rate, John Ross does not have such a failure rate, John Lucas III in Mississippi certainly does not have such a failure rate. Experience and insight matters.
I agree that we must not "try" to create a fistula, but rather to simply create a fistula. Having said that, I do fail. My failure rate is between 1-2 % per year.
We do not stop plavix or aspirin. I ask that Eliquis be held for 24 hours. I will do a transposition with an INR less than 2.5, a simple creation with an INR less than three, and a percutaneous procedure with an INR less than 3.5. I have a low threshold for admitting patients post-operatively for observation or a heparin drip. I evacuate a dozen or fewer hematomas a year.

Thank you everybody for your input. I will place a stent-graft across the venous anastomosis next time he thromboses. Will keep you updated.

I had two similar cases. I identified too much ultrafiltration as the cause of recurrent thrombosis. Once this was adjusted there were no repeat thrombosis episodes. Too much UF causes hypotension and hemoconcentration and these can lead to thrombosis.