Rebleeding, secondary ischemia, and timing of operation in patients with subarachnoid hemorrhage

Methods.

From the SAH database of the University Medical Center Utrecht, we selected patients admitted to our service between 1993 and November 1997, with an aneurysmal origin of SAH, who were admitted within 72 hours after the onset of the bleed. The occurrence and time of ischemia, rebleeds, operation, and death are recorded in the SAH database.

The diagnosis of aneurysmal SAH was based on positive CT scan, or in patients with a normal CT scan, on xanthochromia of the CSF, and an aneurysm on angiography. In patients who died before angiography could be performed, the diagnosis was based on an aneurysmal pattern of hemorrhage on CT. Patients with a nonaneurysmal source were excluded.

Secondary ischemia was considered definite if patients developed focal signs, a decline in the level of consciousness, or both, and if it was confirmed by hypodense lesions on CT, compatible with the clinical signs, or by postmortem confirmation of cerebral infarction. Probable ischemia was diagnosed in the presence of aforementioned clinical features, with exclusion of other causes by CT scanning and metabolic screening. Asymptomatic hypodensities on postoperative follow-up CT scans and episodes of ischemia that were undoubtedly caused by the operation—for example, if they developed in consequence of a temporary occlusion of the middle cerebral artery during operation—were not considered secondary ischemia. For the purpose of this study episodes of probable and definite ischemia were combined. Definite rebleeding was diagnosed in case of a sudden decrease in the level of consciousness, with or without new focal signs, with CT showing an increase of subarachnoid, intracerebral, or intraventricular blood in comparison with the preceding scan. Probable rebleeding was diagnosed when similar symptoms and signs occurred, but in patients who died before CT could be performed and in whom postmortem examination was refused. Probable and definite rebleeds were also combined for the purpose of this analysis. The day of ischemia, rebleeding, operation, or death refers to the number of days passed since the index hemorrhage that had led to admission.

Criteria for performing early operation (within 4 days after onset of SAH) were 1) normal level of consciousness on admission without focal deficits and without parenchymal extension of the hemorrhage, 2) occurrence of acute hydrocephalus necessitating ventricular drainage, or 3) occurrence of an early rebleed.

Patients were not operated when they were in a very poor clinical condition and no benefit from operation could be expected. Lavage of CSF during operation was not used.

Seventy-nine of the 346 patients participated in a randomized trial that compared tranexamic acid with placebo with respect to the rates of recurrent hemorrhage and secondary ischemia.4 Fifty-one of these patients had late operation (after 10 days). Forty-one other patients participated in a placebo-controlled randomized pilot study on the postoperative administration of aspirin in patients who underwent operation within 4 days after onset of SAH.5

We used life table methods to assess the daily rates of secondary ischemia and of first rebleeding from day 0 (the day of onset of SAH) to day 42. The population at risk on each day included all patients at the beginning of the 24-hour interval minus half of the patients censored that day and half of the patients with an event on that day. For the analysis on rates of secondary ischemia, patients were censored if they died. For the analysis on rates of rebleeding patients were censored if they died, or if the aneurysm was treated by clipping or embolization. We calculated the rates of ischemia and of rebleeding for four time intervals: day 0, day 1 to 3, day 4 to 10, day 11 to 14, and day 15 to 21. We used the rates on day 4 to 10 (the period in which ischemia is most likely to develop6) as reference and assessed the rate ratio of ischemia and rebleeding in each of the other intervals.

For each interval we calculated the cumulative proportion of all patients entering the time interval who did not develop secondary ischemia between the first day of the interval and day 21 after the initial SAH. One minus this proportion presents the cumulative risk of developing secondary ischemia between the first day of the time interval and day 21. In the same way the cumulative risks for recurrent hemorrhage were calculated.

We compared the time course of secondary ischemia between patients operated within 4 days of SAH and those operated after 10 days.

We performed a sensitivity analysis by recalculating the rates of ischemia and of rebleeding after exclusion of patients who had participated in the trial on tranexamic acid or in the trial on aspirin.

To assess the influence of rebleeding on the development of ischemia, we calculated the daily rates of ischemia from the day of the rebleed and compared this with the rates of ischemia in patients without a rebleed. In addition we recalculated the rates of secondary ischemia in patients with early operation and patients with late operation after exclusion of patients with a rebleed.

Results.

A total of 346 patients fulfilled the inclusion criteria for this study. The mean age was 53 years (SD 14.3 years); 239 (69%) were women. Most patients (245; 71%) were admitted on the day of the hemorrhage, 73 (21%) were admitted after 1 day, 15 (4%) after 2 days, and 13 (4%) patients after 3 days. The World Federation of Neurological Surgeons (WFNS) grade on admission was I in 111 patients (32%), II in 74 patients (21%), III in 32 patients (9%), IV in 35 patients (10%), and V in 94 patients (28%). Operation was performed in 220 patients (64%). In 131 patients, operation was performed within 4 days after the onset of the bleed. Fifteen patients were operated between day 5 and day 10. The median day of operation in the other 74 patients was day 14 (range day 10 to 53). In 212 patients the aneurysm was clipped. In seven patients the aneurysm could not be clipped; in two of these patients the parent vessel was ligated. In one patient no aneurysm was found during operation, although angiography had shown an aneurysm; this patient later died from rebleeding. Of 126 patients who were not operated, two patients were successfully treated by embolization with coils.

Of the 124 patients in whom the aneurysm was not obliterated, 101 (81%) died within 10 days after onset of SAH.

In 75 patients (22%) at least one rebleed occurred; in 26 it was fatal within 1 day. Figure 1 shows the rates of first rebleeding per time interval of 2 days, from the day of onset of the hemorrhage onward. The rate was highest on the first 2 days after onset of SAH. Thereafter the rebleed rate diminished but it slightly increased again during the second week.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Time course of recurrent hemorrhage in patients with subarachnoid hemorrhage (SAH); the rate of recurrent hemorrhage is presented per time interval of 2 days from the day of the initial hemorrhage (day 0) onward. Rates are presented per 2 days and not for each single day to reduce fluctuation.

Of the 346 patients, 96 developed ischemia (28%). Figure 2 shows the rates of secondary ischemia per 2-day period. The rate was highest on day 4, diminished after day 10, but peaked again from day 14 to 18 for patients who were operated later than 10 days after aneurysmal rupture.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Time course of secondary ischemia in patients with subarachnoid hemorrhage (SAH); the rate of secondary ischemia is presented per time interval of 2 days from the day of the initial hemorrhage (day 0) onward.

Table 1 shows the clinical condition on admission and the event rates of secondary ischemia between the day of the hemorrhage and day 21 for patients with early operation, for those with late operation, and for patients who were not operated. Although the clinical condition on admission was better in patients who were elected for early operation than in the 74 patients who were operated after 10 days, the event rate of secondary ischemia in these patients was as high as in the patients scheduled for postponed operation, who had higher risks for secondary ischemia from the outset. Of 131 patients who underwent early operation, 41 developed ischemia (31%); 2 before operation and 39 during operation or between the first and tenth day after operation. The median day of ischemia in these 39 patients was 2 days after operation. Fifteen of 41 patients with early operation and ischemia (37%) had a good outcome (Glasgow Outcome Scale 4 or 5). Twenty-four of 74 patients (32%) with late operation developed ischemia; in 9 of them this occurred during operation or between the first and third day after operation (median 1 day after operation). Six of these 24 patients (25%) had a good outcome. Figure 3 shows the time course of secondary ischemia in the subgroups of patients with early operation and with late operation. In both groups the rate of secondary ischemia peaked postoperatively but after early operation the peak was much higher. Patients operated early developed ischemia from day 1 until day 12 after the initial SAH (median day 4), whereas patients operated after 10 days developed ischemia from day 1 until day 22 (median day 9). After exclusion of patients with a rebleed the event rates and time courses of secondary ischemia in both groups were essentially the same.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Time course of secondary ischemia in the subgroups of patients who had early operation (within 4 days after onset of subarachnoid hemorrhage [SAH]; dark gray) and patients who had late operation (after 10 days; light gray); in each group the rate of secondary ischemia is presented per time interval of 2 days from the day of the hemorrhage (day 0) onward.

Of 271 patients without a rebleed, 76 developed ischemia, with peak frequencies on day 3, 4, and 5. In the 75 patients with a rebleed, 20 developed ischemia; in 16 of them this occurred after the recurrent hemorrhage, most frequently 3, 4, 5, or 10 days after the recurrence. In 12 of these 16 patients the rebleed occurred within 2 days after the hemorrhage that had led to admission.

Table 2 shows the rates, the rate ratios, and the cumulative risks of secondary ischemia and first rebleed on day 0, day 1 to 3, day 4 to 10, day 11 to 14, and day 15 to 21. The rate of ischemia on day 4 to 10 was higher than in each of the other intervals; the rate was lowest on the day of the hemorrhage. The rebleed rate was highest on the day of the hemorrhage (8%). The rebleed rates for the other time intervals were not significantly different from the rate on day 4 to 10. The cumulative risk of secondary ischemia abruptly falls after day 10, whereas the cumulative risk of recurrent hemorrhage gradually declines in the course of the first 3 weeks after the initial SAH.

After exclusion of patients who had participated in the trial on tranexamic acid or in the trial on aspirin, the rates of recurrent hemorrhage and ischemia were essentially the same.

Discussion.

The results of our study indicate that operation is an important risk factor for the development of secondary ischemia. The rate of ischemia peaked postoperatively, both after early and after late operation. Two observations indicate that particularly early operation is a strong risk factor for ischemia more than postponed operation. Firstly, episodes of ischemia in patients operated early were clustered in the first 4 days after the operation, whereas it was more evenly distributed over time in patients with postponed operation. Secondly, despite a better clinical condition on admission with inherently lower risk for ischemia in patients who underwent early operation,7-9⇓⇓ the rate of ischemia was as high as in patients with poor clinical condition in whom operation was postponed.

We did not assess the amount of subarachnoid blood on the initial CT scan, although this is a strong predictor for the development of secondary ischemia.7-9⇓⇓ During the study period the main reason for early operation was a good clinical condition on admission; that is, patients with the least severe hemorrhage. It is therefore unlikely that patients with early operation had more subarachnoid blood with inherently higher risk for ischemia than patients with late operation. Larger amounts of subarachnoid blood in patients with early operation is therefore an unlikely explanation for the relatively high rate of ischemia in early operated patients. In comatose patients ischemia can be harder to detect. Patients with late operation were more often in a poor clinical condition than patients operated early and this could have caused underdiagnosis of ischemia in this group. We found that in nonoperated patients, who most frequently were in a poor clinical condition, the overall rate of secondary ischemia was higher than in patients with late and with early operation. Moreover, in patients with late operation and in patients who were not operated a poorer clinical condition on admission correlated with a higher rate of ischemia. We therefore do not think that we have missed many instances of secondary ischemia in patients operated late.

We cannot exclude the possibility that patients who died from rebleeding may have had a relatively high or low risk of secondary ischemia (or vice versa), and that some bias may have been introduced by their disappearance from the group of patients at risk for the other event. However, in survivors of recurrent hemorrhage, the time course of ischemia after recurrent hemorrhage was essentially the same as that in patients without rebleeding. Because most rebleeds occurred within 2 days after the first bleed, we cannot separate the effects of the rebleed on the development of ischemia from those of the initial bleed. Besides, in the subgroups of patients with early operation and patients with late operation the event rates and time courses of secondary ischemia were essentially the same after exclusion of patients with a rebleed.

Because even operation on day 11 to 14 increases the risk of ischemia, further postponement of the operation might reduce this risk. Yet, between day 11 to 21, the rate of recurrent hemorrhage is much higher than that of ischemia; this argues in favor of planning the operation soon after day 10 in patients who were not candidates for early surgical clipping.

For the prevention and treatment of ischemia in general it is important to appreciate that operation is a strong risk factor. The intermediate factors through which operation increases the risk of ischemia have not been elucidated. Postoperative thrombosis may be an important cause of complications.10 The prevention and treatment of secondary ischemia should therefore aim not only at neuroprotection and vasospasm, but also at thromboembolism.

Our observations indicate that the initial phase of aneurysmal SAH, in which most complications occur, lasts about 3 weeks. Ischemia and first rebleeding rarely occurred after that period.

Earlier studies showed that early operation decreases the rebleed rate but is not more beneficial than late operation.2,3⇓ This is probably explained by the high rate of postoperative ischemia. Our observations support the notion that alternative treatment modalities should be explored. Acute stage embolization with coils can be expected to decrease the rebleed rate, and it may be less likely to induce ischemic complications than surgical clipping.