The fatal periangiographic rerupture of a posterior inferior cerebellar (PICA) aneurysm in a 20-year-old woman was published as a pictorial essay in 1998 (1). The angiographic image showed retrograde opacification of the contralateral vertebral artery, indicating vigorous injection of the diseased vessel. At the time, we sent a Letter to the Editor (2), emphasizing that the injection volume and rate should be kept low when vertebral arteries are evaluated in patients with a subarachnoid hemorrhage. The survey published by Yousem and Trinh (3), apparently designed as a reaction to this correspondence, is based on an e-mail questionnaire sent to 90 neuroradiology program directors. In addition to injection volumes and rates, Yousem and Trinh asked the potential respondents whether they thought that they could reduce their complication rates by decreasing injection rates, “within reasonable injection rates.” The authors did not indicate what they meant by “reasonable” rates. Perhaps the rates used by the surveyed group were assumed to be reasonable a priori. Forty-eight of 59 respondents “did not believe that injection rates substantially contribute to catheter-based complications.” The most precise information provided about periangiographic aneurysm rerupture was that it is “rare and anecdotally reported by those responding to the survey.” At the same time, a study (4) quoted by Yousem and Trinh in support of this opinion, and in which the injection volumes and rates are said “to correspond well with the results of this survey,” reports a rerupture rate of 1.4%, increasing to 4.8% for angiograms performed within 6 hours of the subarachnoid hemorrhage (ie, almost one rerupture in every 20 patients). This incidence does not qualify as anecdotal evidence, particularly when one considers that the prognosis of periangiographic rerupture is notably poor (even worse than that of spontaneous rerupture, with published mortality rates of greater than 90%).
High injection volumes and rates are inherited from the era before digital subtraction angiography and should be reconsidered. The assumption that lower injection rates introduce a risk of false-negative findings, although reasonable, is purely conjectural. On the other hand, the risk of missing a PICA aneurysm when the contralateral vertebral artery examined by means of reflux only is mentioned in the neurosurgical literature. More importantly, the elevation of distal intra-arterial pressure during the injection of contrast material has been clearly demonstrated. The increase in distal intracarotid pressure was shown to correlate with the injection rates and doses in dogs (2). In humans, intraaneurysmal pressure measurements obtained during angiography reveal “abruptly elevated intraaneurysmal pressure by injection of contrast medium” and that this increase “might cause rerupture of an aneurysm, ” as Gailloud and Murphy (2) quoted. Contrary to the impression conveyed by Yousem and Trinh’s publication, the suggestion of a possible correlation between injection volumes and rates and angiographic complications is nothing new. Even a leading authority such as Weir (5), who is inclined to believe that early periangiographic aneurysmal rerupture rates “are more a reflection of the natural history of rebleeding than a response to [catheter angiography],” states that it is “prudent to use the minimum pressures of injection and volumes of injectate in the early angiographic investigation of subarachnoid hemorrhage.”
In summary, Yousem and Trinh’s survey is based on the retrospective recollections and opinions of 62 neuroradiology program directors obtained through an e-mail questionnaire. The study provides no incidence of aneurysmal rerupture and no indication of the rate and volume used during these “rare” cases. The authors provide no information about the angiographic practice of the survey respondents (eg, angiographic case loads or overall complication rates) or their subspecialization (diagnostic neuroradiology versus interventional neuroradiology). On the basis of the findings from this survey, Yousem and Trinh feel that they are authorized to “provide industry norms for injections in the common carotid, internal carotid, and vertebral arteries.” We believe that the publication of guidelines regarding patient-safety issues requires more than the reporting of a selected collection of subjective opinions with no statistical value or clinical relevance. At this stage, we continue to recommend the use of low injection volumes and rates during cerebral angiography, particularly in the evaluation of patients with subarachnoid hemorrhage.
Reply:
Drs Gailloud and Murphy make valid points regarding the limitations of the article we published in the AJNR (1). Philosophically, we agree with Drs Murphy and Gailloud in that the smaller the pressures and amount of contrast agent used, the better. The emphasis of this article was not about the incidence of periangiographic aneurysm rupture; the data provided on this point was purely subjective and anecdotal just as Drs Gailloud and Murphy state. The biases of the program directors, their subspecializations, and their complication rates were not considered. We are not sure how valid the subjective self-reporting of complication rates would be in this arena.
The article does, however, provide the current standard of care, as judged by fellowship program directors, with respect to injection rates used in neuroangiography. Again, we believe that one should be cautious about greatly deviating from the results cited in the article. The mean values were the following: 7.2 mL/s (SD, 1.8) for a total of 9.9 mL (SD, 2.0) in a typical common carotid artery, 5.8 mL/s (SD, 1.4) for a total of 7.9 mL (SD, 1.5) in a typical internal carotid artery, and 5.4 mL/s (SD, 1.2) for a total of 7.8 mL (SD, 1.7) in a typical vertebral artery. These values reflect the injection rates taught by neuroradiology program directors to neuroradiology fellows. These injection rates are currently in use in 63 institutions in the United States and Canada at which neuroradiologists are trained. Clearly, one must judge each vessel individually, but the values reflect routine injection rates.
The injection rate cited in the letter published in the New England Journal of Medicine (2) for the evaluation of the vertebral artery (ie, 3.0 mL/s) is 2 SDs below the vertebral arterial values published in our article.
- Copyright © American Society of Neuroradiology
