Discussion

Dr Roy Fujitani (Orange, Calif). Good morning, members and guests. Alex, that was a very well presented paper. The USC vascular surgery group has very nicely reported their continued experience with the use of carbon dioxide as an alternative to iodinated contrast material. Their experience dates back to 1991 when it was first reported by Fred Weaver and the authors at USC who had published their clinical applications of CO2 angiography in the Journal of Vascular Surgery. In this series, the authors report contemporary utility of CO2 digital subtraction angiography in imaging 16 of 100 patients who underwent endovascular aortic aneurysm repair with associated renal insufficiency over a 2-year period of time. The use of CO2 DSA helped protect renal function by lessening the need for iodinated contrast material with effective imaging quality but resulted in longer fluoroscopic and operating times.

Now Dr Chao, as part of my discussion, I will intersperse four questions for you to consider. Since carbon dioxide is approximately 20 times more soluble than oxygen, the carbon dioxide bubbles completely dissolve within 2 to 3 minutes after being injected into the vessel and it is eliminated through the respiratory system, but I would be concerned that in using carbon dioxide in patients with very advanced pulmonary compromise or insufficiency, such as someone with COPD who is a carbon dioxide retainer or those patients who have pulmonary hypertension, the diagnostic doses of CO2 may increase pulmonary arterial pressure. Additionally, patients with right-to-left shunts may be at risk for paradoxical gas embolism. This leads to my first question. Are there any absolute contraindications in the use of carbon dioxide as a contrast agent?

As you alluded, the incorrect technique may result in contamination of CO2 gas with atmospheric air and this will result in very serious complications since atmospheric air is much less soluble in the blood than carbon dioxide. Additionally, even without contamination, if enough carbon dioxide gas is trapped within a large abdominal aortic aneurysm and it persists, it may allow gas exchange between the carbon dioxide and nitrogen within the blood. This exchange may result in visceral ischemia due to vacuum locking within the mesenteric arteries. In your manuscript, you alluded to one patient having developed pancreatitis, and I wonder if this may not have been the mechanism that had occurred due to ischemia of the visceral circulation.

There is need to be very careful in timing of consecutive injections of carbon dioxide—and I think you mentioned 2 minutes—to prevent the localized accumulation of gas bubbles, which may produce a clinically significant gas embolism. My next question, therefore is, is there a maximum volume in a single injection that prevents gas accumulation that may result in nitrogen dissolution and/or vapor locking?

Since carbon dioxide is lighter than blood plasma and floats on the surface of the blood, once injected into large vessels such as the aorta or inferior vena cava, carbon dioxide bubbles flow along the anterior surface of the vessel with incomplete blood disbursement along the posterior portion. Because of the anterior origin the celiac vessels, the superior mesenteric arteries may be very well visualized even with smaller volumes. It has been reported that perhaps vessels that are smaller than 10 mm are best imaged with CO2 angiography because then it disperses within 80% of the volume of blood. My next question therefore would be, is there a critical maximal diameter of the vessel or aneurysm where the buoyancy of carbon dioxide gas can be problematic, not allowing complete visualization of the lumen and, therefore, having a deterioration in the quality of the imaging?

All in all, carbon dioxide seems to have very notable advantages compared to iodinated contrast material. It has no allergic reactions. There is no renal toxicity. There does not appear to be any hepatic toxicity, and you can go on to inject unlimited amounts of carbon dioxide in vascular imaging because the gas is effectively eliminated by means of respiration. Finally, compared to nonionic contrast agents and gadolinium-based types of agents, it is relatively cheap. I checked with our purchasing department and contrast agents tend to be a little over $1 per cc, whereas CO2 is less than a penny.

My final question: with so many advantages of CO2 DSA, has your group at USC considered performing all of its angiographic studies using this agent instead of iodinated contrast?

I want to thank the program committee for the privilege and opportunity to presents questions on this paper. Thank you.

Dr Alexander Chao. Thank you, Dr Fujitani, for your comments. In regards to your first question regarding any absolute contraindications to the use of CO2 : although there are contraindications to its use in certain procedures such as cerebral angiography, we have not found CO2 to be contraindicated in any single patient population. As you mentioned, there are descriptions of concern with use of CO2 in patients with severe COPD. Although we did not look at the severity of the COPD or any associated pulmonary hypertension, we have not found it to be a problem either in this study or in our previous experience.

Your second question concerned the maximum volume of CO2 in a single injection that prevents accumulation. We currently use a 50-cc syringe bolus injection by hand with 2-minute intervals between injections. I am not aware of any human studies, but there were many dog studies that looked at volume. Rapid intravenous and intra-arterial injections of 7.5 milliliters per kilogram repetitively and continuous infusions of 100 milliliters per kilogram intravenous for up to 10 L of CO2 were safe. Even injections of 7 milliliters per kilogram directly injected into canine renal arteries caused no complications in the supine position.

You talked about the vapor lock. Yes, there have been sporadic reports in the literature of complications that are thought due to vapor lock, with the greatest number of cases seen in the smaller series. I think these reports may play a large part in preventing wider use of CO2 . I do not really have an explanation why there are sporadic reports. My impression is that it may be due to accidental contamination of room air into the injection. There are many different injection techniques currently in use. Notably, one group has described relying on automated injectors to prevent contamination by room air. We continue to have confidence in hand injection with attention to frequent purging. It is also important to ensure proper tightening of the seals on the lure-locks to, again, avoid unintentional withdrawal of air into the system.

Is there a critical maximum diameter of an aneurysm? I think that is going to be a hard one to answer. Your description of the physiology of CO2 gas in the vascular system was most thorough. Unfortunately, my answer is much less impressive. We routinely image large aortic aneurysms well over 5 cm in diameter, while on the other end of the spectrum, we have reported in our experience with studying renal arteries as well as distal tibials with CO2 .

I think a large determinant of image quality depends on the injection catheter positioning and imaging angle. Initially, we frequently tried patient repositioning in regards to the buoyancy you mentioned, but we really have abandoned that recently, and find that with proper catheter positioning, you can get excellent contrast imaging in large as well as smaller vessels.

Your final question, with the many advantages of the CO2 we purport, should we consider using it in all patients? I think that is a very good question. We have currently at USC initiated a prospective nonrandomized study looking at CO2 versus iodinated contrast in each patient in attempt to get a better idea of the exact sensitivity of CO2 gas in detecting endoleak during EVAR. We are considering increasing our use of CO2 in the elderly patient population where the creatinine level may not be as sensitive an indicator of true renal function. We are trying to more consistently use the calculated estimated GFR.

Dr Willis Wagner (Los Angeles, Calif). A very nice presentation. Based on the leadership role that Fred Weaver and Doug Hood have shown us in the use of CO2, we actually have a large experience with using CO2 in patients with renal insufficiency. However, 2 years ago, I did have a case that has tempered our enthusiasm, and I would like to get some feedback from you. We had a patient who was having a CO2 angiogram for occlusive disease who developed massive fatal emboli to everything below the diaphragm, unlike anything I had ever seen with patients having standard contrast angiograms. Due to lack of autopsy, we were unable to identify the cause of the embolization, but I am just wondering whether you have seen this and whether this was in fact related to the CO2 or just the fact that the patient had the angiogram.

Dr Chao. What you describe is actually very similar to a case report in the literature attributed to CO2 aortic angiography. This is something we thought about, but obviously, I cannot answer that definitively. We did not have any problems during this study, but our previous series of 600 CO2 angiograms in patients from 1985 to 1995 did have one fatality that was due to what was initially thought to be small bowel vapor lock. That patient actually had an autopsy and was found to have multiple severe cholesterol microemboli in the small bowel. Similar to the risks of high-pressure injection during regular iodinated contrast angiography, the power injection of the CO2 may also cause fragmentation of unstable aortic debris.

This brings up a technical point I failed to mention. Due to the compressible nature of the CO2 gas, during a rapid 60-cc manual injection the majority of the CO2 is not actually expelled until the syringe is nearly empty. It is a good idea to ease off a little at the end of the injection to avoid too high a pressure at the delivery end. Others have suggested that all injections be done with an automated injector to prevent overpressurization, but once again, we do not really find that to be a problem currently.

Dr J. Dennis Baker (Los Angeles, Calif). Why did you have such a substantial increase in your fluoroscopy time and radiation? Can you explain why it takes so much more radiation to complete the task with carbon dioxide?

Dr Chao. That was something that was a little bit of a surprise when we sat down and looked at the data, but the total radiation as measured by DAP is easy to explain. Normally for angiography we use either two or three frames per second, usually two frames per second. For the CO2, we really have found it necessary to go to six frames per second. As you know, the runs really count for most of the patient radiation exposure. The fluoroscopy time was also increased significantly, here almost doubled by number, and I do not really have a good explanation of that. When we looked, we did have—I think it was—38% in group 2 requiring hypogastric artery embolization and a smaller percentage in group 1. Group 2 patients I think overall tended to be somewhat more complicated. We saw a trend in a greater number of graft components used and the longer operative times. The CO2 itself does not increase the operative time at all, so I think it may have just been part of the patient population that explains the increase in fluoroscopy time.

Dr Benjamin Starnes (Seattle, Wash). Alex, that was a very nice presentation. I have a couple of questions. In my experience with CO2 angiography, especially with the patient awake, one injects the CO2 and the patient complains of severe abdominal pain and then moves all over the place and that certainly affects the imaging quality substantially. If you are doing these under local or regional anesthesia, perhaps that is an explanation as to why you have increased your fluoroscopy time. Are you having to put every single patient under general anesthesia to use this technique? One would think that this would add to the morbidity profile.

The second question is, why not use IVUS to precisely locate your pelvic vessels and be able to precisely place your graft? I use IVUS quite often for this purpose and I think that it is an attractive alternative over CO2 .

Dr Chao. We have had some experience with IVUS use in the past and again recently for thoracic endograft placement but have not enough experience to become proficient or have a personal opinion on its utility. Additionally, for routine EVAR, the acquisition cost versus patient benefit is somewhat prohibitive. All the patients in this series received general anesthesia. In regards to lower extremity CO2 angiography, we have found the DS to be no more or less than with iodinated contrast.

Dr Timothy Chuter (San Francisco, Calif). I notice you use on average 27 milliliters of contrast. We have been able to use even lower volumes of contrast without using CO2 . We perform selective angiograms of the renal arteries using bright-tip catheters and little puffs of half-strength contrast. In fact, we have inserted multibranched stent grafts for TAAA using as little as 25 milliliters of contrast. Have you tried selective angiography as an alternative to CO2 angiography, which seems to involve intubating the patients and spending a lot of extra time in the OR?

Dr. Chao. That is very impressive and I have heard descriptions from your institution regarding very low volumes of contrast.

Dr George Andros (Los Angeles, Calif). Under Fred’s goading and Nick Nelken’s persistence, we instituted a CO2 angiography program many years ago. Because we have a lot of diabetics in our practice, we use it very commonly for bilateral renal stenting and all sorts of infrainguinal work, but not for aneurysms so far. And it’s cheap!

Isn’t it a little disingenuous to say you only use 27 cc of contrast when nearly all patients are evaluated with followup CT angiography? Is this your imaging policy, or are you doing followup studies with duplex? We heard earlier today that, on average, patients treated with EVAR will have a measurable decline in renal function when checked 1 year postoperatively. If you are economical with iodinated contrast as you suggest, have you also observed this deterioration in kidney function or are the patients spared this complication by using CO2 and duplex?

Dr Chao. Those are excellent points, and I am glad we have an opportunity to discuss the importance of cumulative contrast toxicity. Greenberg’s recent data suggest a cumulative effect on the kidneys. Going by our protocol, in any patient with baseline renal insufficiency, we avoid contrast during preoperative or postoperative CTs unless absolutely necessary. We have previously demonstrated excellent correlation in our institution between duplex imaging and CT, so we either obtain duplex preoperatively or MR angiography. Preoperatively we follow by duplex only on these patients so there is no other cumulative contrast insult.

As far as the BUN and creatinine, it is a problem. In this day and age, it is difficult to get follow-up labs on patients. We only had post-op inpatient and discharge labs; however, we do know that no patient required dialysis at any point for contrast-induced renal failure.

Dr Chuter. I have to respond to the word disingenuous. I do not think anybody would advocate using CO2 outside of the setting of the stent graft implantation, so that is where I was focusing my comments and that is what the 25 milliliters refers to.