and Riives (1948) stated that in the end, probably most, if not all unoperated patients die of hemorrhage or are completely incapacitated. Olivecrona (1957) found that 25% of his untreated patients were to die of further bleeding, one third were to suffer serious morbidity, but that 25% were to remain healthy and almost asymptomatic many years after diagnosis. He stressed that the two groups of operated and non-operated patients could not be compared as they were no doubt selected according to differing points of view.
Botterell (1966) pointed out that progressive neurological deterioration may occur even though follow-up arteriography showed no change in the size of an AVM and he contributed this to progressive gliosis of surrounding cerebral tissue.
Paterson and McKissock (1956) stressed that the follow-up of untreated cases in their series did not contribute substantially to the knowledge of what might be termed the natural history of intracranial angiomas, largely because insufficient time had elapsed since their patients had been diagnosed. Yet it was worthy of note that few of them were severely incapacitated even after many years. Although angiomas may, through hemorrhage, lead to death or permanent incapacity, they felt they were much less likely to rupture than intracranial aneurysms. They felt that Potter (1955) had published figures which lent support to their view; 27% of his patients survived for more than 20 years after the onset of symptoms and more than half of them had only slight disability or none at all.
The increasingly aggressive attitude toward the surgical treatment of AVMs raised two controversial and interrelated issues:
1. Many retrospective series, although based on relatively small numbers of patients, seemed to prove that unoperated patients had a better prognosis.
2. The mortality and morbidity rates of operated cases’ in some series were not low (up to 18%).
Svien and McRae (1965) from the Mayo Clinic considered that 85% of patients with angiomas were best treated conservatively.
Perret and Nishioka (1966) analyzed 545 cases of AVM in a Cooperative Study and stated that the operative mortality should be less than 10% and the postoperative morbidity should be better than, or comparable to patients treated conservatively, for surgery to be justifiable. They found that the mortality in unoperated cases was 5% and in operated cases was 12%. These authors included palliative surgical treatments such as carotid ligation or partial resection in their surgically treated group. Therefore it is clear that a true comparison between operated and unoperated groups is impossible.
Pool and Potts (1965) collected 523 cases from the available literature. They found that the mortality and morbidity of the conservatively treated group was 56%, while that for 187 patients treated by radical excision was only 26%. They suggested that “on the basis of these data, excision of a symptomatic AVM seems advisable whenever surgery can be done with reasonable safety since it offers the best chance of saving a patient from progressive neurological and mental deterioration, epileptic seizures, or death from hemorrhage. This is because excision improves the circulation to those areas deprived of normal blood supply by the arteriovenous malformation of normal blood supply and removes the threat of a serious hemorrhage. On the other hand, excision is obviously contraindicated if the patient may become worse after surgery due to his age, condition, or the location and extent of the lesions. Excision of lesions in the occipital lobe or the posterior part of the temporal lobe may, for example, lead to permanent homonymous hemianopsia, but this seems a small price to pay for relief from the threat of progressive brain damage or a fatal hemorrhage. Under favourable conditions, moreover, an arteriovenous malformation can usually be excised without sacrificing significant amounts of intact tissue, even if this does lie adjacent to the sensory or motor cortex or involves the dominant temporal lobe.”
“As a rule children tolerate excision particularly well. Good results can also be expected from excision if the arteriovenous malformation is small and the patient is under 50 years of age and in a good preoperative condition. In patients over 50 years of age the need for excision is less urgent than in younger ones, since progressive enlargement of the arteriovenous malformation is less likely.” They concluded: “In our experience, approximately 75% of patients recover well following excision while some reports indicate even better results.”
French et al. (1964) considered that “AVMs are not in any sense to be considered new growths, but they may enlarge with time by expansion of the vessels and dilatation of aneurysmal sacs”. This seemed an important factor to be included in the evaluation of the necessity of surgical excision. On this basis the practice of surgical intervention seemed unquestionably superior to nonsurgical management.
However, in 1970, Troupp reported on 137 patients seen between 1942 and 1969 with angiographically verified AVMs. These were not operated upon, except for exploratory craniotomy in 2 cases (Troupp et al. 1969, 1970). At that time, he felt their follow-up figures indicated a more cheerful outcome than that postulated by others, notably Olivecrona and Riives (1948), Pampus et al. (1960) and Norlén (1966).
After 7 years Troupp reviewed the same 137 cases and changed his mind somewhat, for only 27 cases were well. Fourteen were described as fair but 28 were disabled and 9 more had died from hemorrhage (total 23). The only factor of interest he could relate concerning the prognosis was the location of the AVM. Of those situated in the frontal, temporal and occipital regions, only 2 patients died from bleeding while 21 of those situated in the parietal, central and infratentorial regions ruptured fatally.
Drake’s observations (1979) concurred with the view that while the initial course of an AVM is reasonably benign, the long-term outcome is less favourable.
Waltimo (1973) discussed the increasing size of an AVM in terms of prognosis: In a study of 21 patients with AVM subjected to serial angiograms (with a median time between examinations of 44 months), it was noted that 12 of the lesions increased in size, 8 remained unchanged, and one became smaller. He also found that smaller AVMs were the most likely to increase in size and the largest were the most likely to reduce in size.
Höök and Johanson (1958) found that in 12 cases: 8 increased in size, 4 remained unchanged and 1 disappeared and Stein (1984) found in his series that one third increased, one third remained unchanged and one third became smaller.
Kelly et al. (1969) followed 33 patients for an average of 15.5 years after their initial hemorrhage from an AVM. There was a mortality rate of 28% and half of the patients had little, if any disability.
Fults and Kelly (1984) found that the prognosis for patients presenting with seizures was more favourable than for patients presenting with hemorrhage, and that the mortality associated with recurrent hemorrhage did not increase significantly with successive bleeds. Patients with posterior fossa AVMs fared considerably worse than patients with an AVM located elsewhere in the brain. Pellettieri (1979), however, found that the risk of death in an unoperated group of patients followed over a period at 18 years, to be 2.5 times greater than in an operated group. The prognosis for children with intracranial AVM was not different to that for adults. The size of the AVM did not dictate either the clinical outcome for the patient or the risk of hemorrhage. There seemed to be no correlation between pregnancy (or delivery) and hemorrhage from a cerebral AVM in a female patient.
The publication of Pellettieri et al. (1979) was a milestone in the literature on the natural history of the AVM as the authors applied a differential analysis to their cases. By relating the results to 6 or 7 variables (age, sex, AVM-size, AVM-location, symptoms at onset and neurological findings at admission) it was possible to grade each variable with respect to prognostic importance. The most favourable risk factors were age below 40 years, absence of a neurological deficit, a superficial and small AVM in a silent area, female sex, and SAH at onset. These variables were assigned numerical values, and grouped on a scale ranging from +16 to −16. The AVMs with values of −10 or below were found to have a poor prognosis whether the condition was treated surgically or conservatively.
