Should We Screen for Familial Intracranial Aneurysm?
To the Editor:
I read with interest the analysis of Crawley et al1 regarding screening for familial aneurysms. This mathematical modeling concluded that screening is not an effective way of reducing morbidity and mortality from ruptured intracranial aneurysms in individuals with a history of more than one affected first-degree relative with a ruptured intracranial aneurysm. I take exception to this paper on one very important account, namely, the authors’ assumption of surgical morbidity for the clipping of intracranial aneurysms. They used a figure of 8.0% for the risk of morbidity that resulted in death or dependence in everyday living from surgery on unruptured aneurysms. They state that this figure was taken from a meta-analysis published in 1998 by Raaymakers et al.2
The study of Raaymakers et al shows that for patients with non-giant anterior circulation aneurysms, the mortality rate was only 0.8% and morbidity was only 1.9%. It was the patients with giant posterior circulation aneurysms who had a mortality rate of 9.6% and a morbidity of 37.9% that skewed the overall figures to the 8% number used by Crawley et al in this study. Statistically, between 92% and 95% of aneurysms identified by routine screening methods will be non-giant anterior circulation aneurysms. The vast majority of the other 5% to 8% of aneurysms will be non-giant posterior circulation aneurysms. The projected morbidity and mortality rates for patients who were considered in this mathematical screening model will have expected surgical morbidity and mortality of very close to the 1.9% and 0.8% range.
In my own published series,3 looking at the outcome of patients with unruptured aneurysms, we found that morbidity and mortality was totally dependent on size. In all non-giant aneurysms, morbidity was seen in 5.9% of the cases, with mortality in 1.3% of the cases. However, if we looked at all totally incidental aneurysms that were asymptomatic at the time of presentation, the total morbidity and mortality rate for non-giant aneurysms fell to 3%. Similarly, looking at all aneurysms <1 cm in diameter, surgical morbidity was <2%.3
I suggest that the modeling used by Crawley et al is based on an invalid assumption, because the vast majority (probably in the order of 98%) of aneurysms that might be identified by MR and digital subtraction angiography screening for familial aneurysms are likely to be non-giant aneurysms. Almost all of these are likely to be non-giant anterior circulation aneurysms. Because none of these aneurysms are likely to be symptomatic, the great preponderance of aneurysms that are identified also tend to be small aneurysms, probably <1 cm in diameter.
I therefore suggest that if the correct morbidity and mortality figures for surgical treatment of intracranial aneurysms were used, screening for familial intracranial aneurysms in a high-risk population would be found to be highly effective in reducing both morbidity and mortality and cost savings.
- Copyright © 1999 by American Heart Association
We thank Dr Solomon for his interest in our model designed to assess the value of screening for unruptured cerebral aneurysms.R1 We accept that the figures used in the calculations used to model the outcome of health related interventions are somewhat arbitrary, and surgeons who believe they can do better than average will understandably take exception to the findings. However, to provide figures that allow physicians and healthcare providers to decide on the appropriateness of introducing a general policy of screening for aneurysm, it is necessary to use figures based on an overall average assessment of risks. To be applicable to health care in general, the average figures should be taken from as large a representative sample of different centers as possible. Hence, we used data taken from the Raaymaker et alR2 systematic analysis of 61 separate studies, which is likely to provide a good sample of both representative patients and representative centers. Our conclusions can therefore be applied across the wide range of centers and types of patients included in their analysis.
Dr Solomon takes exception to the fact that we used an overall average figure of 8%, taken from the study of Raaymaker et al, for the risk of morbidity resulting in death or dependency from surgery on unruptured aneurysms, on the grounds that the figure includes the risks of surgery for patients with giant aneurysms as well as those with smaller aneurysms. It is certainly true that the risk of surgery is greater with larger aneurysms than smaller ones, but when carrying out a screening procedure on an asymptomatic patient, the physician does not know the size of any aneurysm that may be detected. Hence, we were correct to use the average risk calculated from a series that included all comers and all sizes of aneurysms. The figures Dr Solomon quoted from Raaymaker et al,R2 namely, a combined mortality and morbidity of 2.7% for surgery on non-giant anterior circulation aneurysms, were estimated according to a multivariate analysis that included only 29 of the 61 studies included in that review. As the authors state,R2 this provides only a rough estimate, and the figures we used based on the overall analysis of all 61 studies are much more robust. The figures we used are supported by the recent report from the International Study of Unruptured Intracranial Aneurysm Investigators,R4 in which the overall rate of surgery related morbidity and mortality was 17.5% at 30 days and 15.7% at 1 year in their group of patients having no previous history of subarachnoid hemorrhage.
Our conclusions may not apply to individual patients (eg, if their risks of rupture are much higher than average) or individual surgeons (eg, if their results are very much better than average). Similarly, our conclusions may also not apply if an aneurysm has already been detected during CT scanning, when the figures for the risk of surgery and of spontaneous rupture will need to be adjusted according to the size of the known aneurysm. However, even with very safe surgery, screening for asymptomatic aneurysm is not cost effective. This is because the main determinant of the utility of screening for aneurysm is not the risk of surgery, but the risk of spontaneous rupture. Thus, our model demonstrates that if Dr Solomon adopts a policy of screening with MR angiography and operating on all asymptomatic unruptured aneurysms, and he maintains his publishedR3 combined mortality and morbidity rate of 3%, he will save only 1.7 strokes over 10 years for every thousand patients screened, at a cost of over £1.8 million ($3 million) per stroke saved.
We therefore reaffirm our conclusion that a general policy of screening for aneurysms should not be advocated, even in units with a low surgical morbidity. This does not mean that exceptions can not be made for individual cases.
Crawley F, Clifton A, Brown MM. Should we screen for familial intracranial aneurysm? Stroke. 1999;30:312–316.
Raaymakers TWM, Rinkel GJ, Limburg M, Algra A. Mortality and morbidity of surgery for unruptured intracranial aneurysm: a meta-analysis. Stroke. 1998;29:1531–1538.
Solomon RA, Fink ME, Pile-Spellman J. Surgical management of unruptured intracranial aneurysms. J Neurosurg. 1994;80:440–446.