Their conclusions probably reflect the opinion of most neurosurgeons: “A favourable combination of variables gives relatively good results with both modes of therapy. Results deteriorate proportionally with falling values on the risk scale in both groups. Although surgery tends to give better results, the difference is only significant within a limited range on the risk scale. This probably explains the controversy between those who advocate surgery and those who prefer conservative treatment.”
Calica et al. (1984), recently took the idea of risk prediction a step further, using a complicated regression formula involving 14 variables to assess outcome (6). When they applied this formula to their 78 patients with intracerebral AVMs, it divided 85% of them into low (3% became impaired), medium (42% became impaired), and high (94% became impaired) risk groups.
Citing the paper of Calica et al., Wilkins (1985) has predicted that it may become possible with additional experience to use computerized paradigms to predict with greater accuracy the outcome of an intracranial AVM without surgery or with any of several possible treatment protocols so that the best approach can be planned. Until then, we must still rely on fragmentary published information about the “natural history” of such lesions and on a realistic assessment of our ever-changing abilities to deal with them surgically. Wilkins felt it had been difficult to assess the natural history of intracranial vascular malformations because they are varied in nature, they are frequently silent clinically, they are often treated when they are discovered and untreated lesions are not often followed in an organized way.
We would add to this argument the fact that many published data relate to unsufficiently analyzed cases. We agree also with the remarks of Mohr (1984), “The enormously accumulated studies concerning the natural history of the AVM is retrospective and the rarity of these lesions precludes any definitive prospective study of the natural history. Furthermore, the clinical picture of many of these lesions spans years if not decades. The remarkable variation of clinical material from center to center has become apparent and with it a hesitancy to offer such experience for publication.”
Most recently (1986) Crawford et al. reported upon 217 out of a total of 343 patients with cerebral AVMs, who were managed without surgery. He followed them for a mean of 10.4 years and, using life survival analyses, found that there was a 42% risk of hemorrhage, 29% risk of death, 18% risk of epilepsy and 27% risk of neurological deficit over a 20-year period. This represents the largest series of untreated cases studied over a long time span. He found, interestingly, that although small AVMs, as described by many authors, are more likely to present with hemorrhage in the first instance (82%) they did not subsequently carry a higher risk of recurrent hemorrhage. The operative rate in this series was only 34% with the tendency to leave untreated those AVMs which were large and deep, more posteriorly situated, in the left hemisphere or crossing the midline. However, the authors felt that the size, depth, and possibly the site of the arteriovenous malformations did not significantly affect outcome. The main influencing factors in their opinion were recurrent hemorrhage and increased age at diagnosis. Although the overall mortality at 20 years was 29%, only 65% of the deaths could be attributed directly to the AVM and then most commonly from hemorrhage. The risk of epilepsy is increased with temporal lobe lesions.
There are reasons other than those put forward by Mohr which make comparisons between operative and conservative management difficult:
a) The inclusion of ligation of extra- or intracranial vessels, coagulation and partial removal of the lesion and complete removal of the AVM under the term: “operated cases” is incorrect (Paterson and McKissock 1956, Pool and Potts 1965) (Tables 1.3 and 1.4).
The retrospective study of the 186 AVM cases of Krayenbühl (1936–1966) required separation of the patients into 3 groups: I Untreated, II Palliative treatment and III Complete removal of the AVM (Table 1.5).
A second analysis 15 years later (1984) clearly proved that the patients with complete removal of the cerebral AVM presented much better late results than the patients in group I–II (Table 1.5).
Twenty-one patients (20%) out of groups I–II had recurrent hemorrhages, whereas group III no case of recurrent hemorrhage occurred. Long-term clinical examinations have shown that only 15% of non-operated cases with large and moderate sized AVMs remain in a good clinical condition. The remainder of the cases develop within the following 10 to 15 years after diagnosis a progressive clinical deterioration characterized either by repeated hemorrhages or a progressive mental and neurological symptomatology ultimately leading to irreversible invalidism or even death. A precise analysis of these cases will be provided in Vol. III B.
b) Differences in retrospective studies are mainly caused by analyzing collected cases using different criteria applied to unoperated and operated cases. The statistics may give satisfactory information concerning age, sex, symptoms of the patient and size and site of the lesion, but they cannot necessarily provide a guide to treatment, as an “inoperable” lesion for some neurosurgeons, is deemed operable by others. It is remarkable that in some series 30–50% of AVMs are still deemed inoperable.
c) Authors with conservative attitudes may argue that the operated cases are “easy” lesions whereas the unoperated patient would be regarded as having more high risk characters (size, site etc.). This argument is only partially correct. Many operated cases are not elective “easy” lesions, but occur as emergency cases because of hemorrhage or progressive neurological and mental deficits. Some patients with “easy” operable lesions refuse surgery as they will not accept any operative risk. Some informed patients prefer to gamble on an early favourable clinical course in order to await further technological advances (Drake 1979).
d) Some patients accept surgical risk only after deterioration of their symptoms. Such cases which have been conservatively treated are, however, most often not included in the statistics of unoperated cases, but rather in those of operated cases. Without this recourse to surgery the statistical outlook of unoperated cases might be less favourable. Intracerebral hemorrhage is the most serious complication. Its frequency varies from about 40% to 68% in most series (Pia 1975) with those cases presenting initially with hemorrhage being at greater risk (Crawford et al. 1986).
e) As a result of discussion between neurologists, neuropathologists and particularly neurosurgeons, technical developments within the last 30 years (microsurgery, modern neuroanesthesia, high energy radiation, selective embolization) have offered new approaches in treatment. In many publications with large series of operated cases there is no clear separation of statistical data, as to which cases have been operated using conventional surgical technique, pure microsurgical techniques or using combined techniques such as embolization and microsurgery or surgery and radiation. These data are more clearly given in publications of smaller series, especially those concerning the surgery of “deep seated AVMs”.
f) Other variables have rarely been considered. It is necessary to indicate not just the size of an AVM, in cm2 or cm3, but also its precise construction (single or multiple niduses and compartments,