The relation between aneurysm size and outcome in patients with subarachnoid hemorrhage

Material and methods.

From July 1995 to March 1997, 138 patients with SAH were admitted to the Utrecht University Hospital. The diagnosis of SAH was based on demonstration of extravasated blood on CT or, if CT was negative, by xanthochromia in the CSF. Seventeen patients were excluded: 10 were moribund on admission, and contrast injection was withheld, and in 7 patients no aneurysm was found.

The size of the ruptured aneurysm was measured on CTA by indicating on the workstation the outer borders of the aneurysm on the screen. The computer calculated the distance in millimeters. We defined aneurysms as large if the diameter was at least 10 mm. The radiologist measured the size of aneurysm during the clinical assessment of the CTA, and was thus unaware of the final outcome of the patient.

We used the World Federation of Neurologic Surgeons (WFNS) scale to assess the clinical condition on admission.6 In the analysis we dichotomized between good (WFNS 1–3) and poor condition on admission. The amount of cirsternal blood was assessed by means of the grading scale described by Hijdra et al.7 The total amount of blood on this semiquantative scale ranges from 0 to 30. In the analysis this variable was dichotomized at the median. We classified the operation as difficult if procedures such as temporary occlusion of a parent artery or a bypass procedure had to be undertaken. We defined secondary ischemia as a gradual decrease in level of consciousness, appearance of focal signs, or both, for which other causes had been excluded. A rehemorrhage was defined as a sudden decrease in consciousness and an increased amount of blood on CT, compared with the previous scan. We assessed outcome at 3 months by means of the Glasgow Outcome Scale and dichotomized into poor (death, persistent vegetative state, or dependency) and good outcome (independent or good recovery).

In the comparison of outcome for patients with large and small aneurysms, we used relative risks (RR) and corresponding 95% CI. We used stratified analyses and logistic regression analyses to calculate adjusted RR and corresponding 95% CI.

Results.

The sizes of the ruptured aneurysms in the 121 patients ranged from 1.5 to 24 mm. The 27 patients with large aneurysms more often had a poor clinical condition on admission (RR = 1.96; 95% CI 0.98 to 3.9) and tended to have smaller amounts of extravasated blood on admission CT (table 1).

Patients with large aneurysms tended to be operated on less frequently (table 2), to have more rehemorrhages and fewer episodes of ischemia, and to die more often from the effect of the initial bleeding and secondary ischemia.

Seventeen (63%) of the 27 patients with a large aneurysm had a poor outcome, compared with 39 (41%) of the 94 patients with a small aneurysm (RR = 1.5; 95% CI 1.0 to 2.2). In stratified analyses (table 3), the worse outcome in the group with the large aneurysms was not markedly influenced by age, gender, location of the aneurysm, or amount of cisternal blood. Logistics regression analyses with these baseline characteristics did not change the OR by more than 0.15.

Discussion.

Patients with large aneurysms have a greater chance of poor outcome after SAH than patients with small aneurysms. Patients with large aneurysms more often have a poor clinical condition on admission but the risk of clinical and surgical complications is essentially the same as in patients with small aneurysms. These data suggest that the worse outcome in patients with large aneurysms is mainly caused by a greater initial impact of the hemorrhage. The lesser amount of cisternal blood in patients with large aneurysms seems not to substantiate this suggestion, but the amount of blood is a less powerful predictor for poor outcome than the clinical condition.8 We did not use the clinical condition on admission as adjustment factor because it is a factor in the causative pathway to poor outcome, in which situation adjustment is incorrect.

A limitation of our study is that 10 patients were moribund on admission and CTA was therefore not performed. Exclusion of these patients may have introduced a selection bias, but even if these 10 patients all would have had small aneurysms and a poor outcome, the outcome in the entire series would still be worse for patients with large aneurysms (RR = 1.34; 95% CI 0.94 to 1.90). We found one other study on aneurysm size and outcome; in that study no relation was found.9 Two factors may have introduced a bias toward patients with good outcome in that study. First, the study was performed in a tertiary care center; secondly, aneurysm size was assessed on conventional angiograms, and thus not in patients who are not candidates for operation because of a poor condition. Yet, in that study, patients with large aneurysms had a longer period of unconsciousness after the hemorrhage and patients with small aneurysms had a greater chance of complete recovery. The data of this previous study are therefore in agreement with our data.

In our series, patients with large aneurysms were less frequently operated. This implies that the poor outcome in these patients might partly be explained by an increase in frequency of rehemorrhage, but this was not the case (see table 2). The larger proportion of patients with large aneurysms who have not been operated can be explained by the worse clinical condition from the outset, and is therefore a consequence and not so much a causal factor for the poor outcome.