Role of emergency surgery to reduce mortality from rebleed in patients with aneurysmal SAH.
Potentially salvageable patients with aneurysmal subarachnoid hemorrhage may rebleed and die after admission to a hospital, awaiting an angiogram or surgery. In an attempt to reduce the number of patients thus lost, we are operating on such patients on an emergency basis. This report is a retrospective analysis of patients operated early after aneurysmal SAH, but during routine working hours, versus patients operated on an emergency basis. During an 18 month period 109 patients underwent surgery for intracranial aneurysm. Fifty two (Group 1) patients were subjected to emergency clipping of the aneurysm, and 57 (Group 2) patients underwent early clipping but during routine working hours. There was no selection bias between the two groups. Another 10 patients in Hunt and Hess Grade II and III rebled and died, after admission, awaiting an angiogram or surgery. The overall mortality for the two groups (Grade I to IV) was 11.9%. Mortality in Groups 1 and 2 was 9.6% and 14.0% respectively. At 1 month, good outcome was recorded in 86.6% patients in grade I, 69.2% patients in grade II and 42.8% patients in grade III in group I compared to 68.7% patients in grade I, 80% patients in grade II and 31.5% patients in grade III in group 2. However, the difference was not statistically significant. There was also no significant difference between the incidence of intraoperative brain swelling and delayed ischaemic neurological deficit between the two groups. Grade IV patients had a poor outcome in both the groups. Angiographic vasospasm was associated with poor out come, in either group, in grade IV patients. Patients in Grades I to III should undergo emergency surgery, if the surgeon is experienced and willing to operate at odd hours and necessary support facilities of neuroradiology and neuroanaesthesiology are available. This would avoid deaths in patients awaiting angiograms or surgery.
After an aneurysmal rupture, the peak period for mortality and morbidity is within the first 14 days, mainly due to rebleeding and vasospasm. In the early days of aneurysm surgery, concerns about rebleeding often prompted the treating neurosurgeon to operate upon the patient as soon as possible. However, surgeons were often confronted with swollen brains making retraction and visualisation of vital structures very difficult. As a result operative mortalities were unacceptably high. In 1953, Norlen and Olivecrona published a series of 100 patients with aneurysmal subrachnoid haemorrhage (SAH) who underwent delayed surgery, more than 3 weeks after the ictus. They reported a 3% mortality. This paper influenced many neurosurgeons to adopt delayed surgery for aneurysmal SAH. After the introduction of operating microscope and refinement in the techniques of neuroanaesthesia and brain protection , surgery for intracranial aneurysms became much easier and safer. In the past 15 years the results of early surgery have begun to compare,,,,,, with those of delayed surgery, particularly for alert patients. Early surgery has generally been accepted as surgery within 72 hours of aneurysmal SAH in grade I and II patients. Surgery is postponed if the patients in grade I and II are admitted from days 4 to 11 post SAH.,,,,,,, Freckmann et al, argued in favour of surgery as soon as possible after admission in grade I to III. At the neurosurgical services of All India Institute of Medical Sciences (AIIMS), grade I to III patients undergo surgery soon after admission, irrespective of the time the patient presents after the ictus. Grade IV and V are taken up for immediate surgery, only if the patient has a large intracranial haematoma. However we have experienced that despite subjecting patients to early surgical intervention, irrespective of when the ictus was, patients with a potentially good prognosis are still lost due to rebleeding while awaiting an angiogram or surgery. To minimise the number of patients who rebleed after admission, the senior author (VSM) has been operating upon patients with aneurysmal SAH on an emergency basis. This paper is a retrospective analysis of the results of emergency aneurysmal surgery versus early aneurysmal surgery performed in the same department during the period January 1995 to June 1996.
During the period January 1995 to June 1996, out of a total of 153 patients admitted, under the neurosurgical services of AIIMS, with a clinical diagnosis of SAH, 127 patients had angiographically confirmed aneurysms. In 23 patients no angiogram was performed because of poor clinical status (Hunt and Hess Grade IV-19, and Grade V-4). The angiogram was withheld in 19 out of these 23 patients, who were admitted in grade IV, because they further deteriorated to grade V after admission. In 3 patients (all Grade III) an angiogram was planned for the next working day, but it could not be performed as they rebled and deteriorated to grade V while waiting for an angiogram. Hundred and nine patients underwent surgical treatment of the aneurysm, 10 patients underwent endovascular therapy of the aneurysm and one patient refused any form of therapy. Amongst these 109 patients, a giant basilar top aneurysm was found to be unclippable at surgery in one patient, and in a second patient a giant carotico-ophthalmic giant aneurysm had to be trap ligated. Successful clipping of the aneurysm was carried out in the remaining 107 patients. Seven patients suffered a fresh rebleed and died while waiting for elective surgery (grade II-I and grade III-6). Grade V patients were not taken up for angiograghy. Some selected grade IV patients were taken up for angiography, followed by surgery, if the treating surgeon felt that the patient was clinically improving.
The 109 patients who underwent surgical clipping were divided into two groups. Division into eithergroup was decided solely by the fact whether the patient underwent emergency surgery or not -which in turn was decided by the fact that whether the treating surgeon was operating in emergency (VSM) or not (AV, VPS, RP). Age, sex, neurological status and presence of medical illness like diabetes mellitus or hypertension did not influence the division of patients into two groups.
Group 1 (52 patients) underwent angiogram followed by emergency surgery. If the patient was admitted during routine working hours, angiogram was performed the same day followed by immediate surgery. Patients who were admitted after working hours, underwent angiography the next morning followed by immediate surgery. The only factor that decided how soon the patient was operated after admission was the availability of cerebral angiogram laboratory. Only in patients admitted in grade IV, the anglogram was deferred till their clinical grade improved. But they were immediately taken up for surgery, once the angiogram was performed. Group 2 (57 patients) patients underwent angiography on the next working day after admission and surgical clipping in next available routine operation theatre time. The preoperative management , anaesthesia and post operative management protocols were the same for the two groups. Briefly each patient, after admission was clinically categorized according to Hunt and Hess grading. Before surgery, all patients were started on steroids, anti convulsants, a benzodiazepine tranquilizer, opiold analgesics, a laxative, and nimodipine. Post operatively steroids, anticonvulsants and nimodipine were continued. Patients in grade II, III and IV were also vigorously treated with triple - H (hypertension, hypervolemia and haemodilution) therapy.,
Outcome was assessed at 1 month after surgery, according to Glasgow Outcome Scale and each patient was allotted to one of the categories: i) good no deficit or minor deficit but normal life, ii) moderate disability-neurological deficit present, but independent, ii) severe disability-dependent with severe neurological deficit, iv) persistent vegetative state and v) death.
Forty one of 52 patients in group 1 [Table I] were in the third to sixth decade of life (mean age- 40.5 years). There were 30 male patients. At the time of admission, 15 patients were in grade I, 26 in grade II, 7 in grade III and 4 in grade IV. Twenty one patients were admitted within 3 days of lctus, and 35 patients within 1 week of ictus. Seven patients were admitted in the second week and 10 after 14 days of ictus. Forty patients presented with one ictus where as twelve patients had more than one ictus prior to admission. There were 61 aneurysms in these 52 patients [Table II]. Eight patients harboured multiple aneurysms (2 aneurysms in 7 patients and 3 aneurysms in 1 patient). Seven of these patients had bilateral aneurysms. There were 6 giant aneurysms.
In group 2 (n=57), there were 32 male patients. Forty nine [Table I] patients were in the third to sixth decade of life (mean age-38.6 years). At the time of admission [Table III], 16 patients were in grade I, 15 patients in grade II, 19 patients in grade III and 7 patients in grade IV. Nineteen patients were admitted within 3 days of ictus, and 31 patients within 1 week of ictus. Eleven patients suffered from more than 1 ictus prior to admission. Twelve harboured multiple aneurysms of which 7 had bilateral aneurysms. Ten patients had giant aneurysm. There were 72 aneurysms in this group. Anterior cerebral artery-anterior communicating artery junction was the commonest site [Table II]. The management results of the 2 groups were compared using the Fisher's exact test.
The results have been analyzed for patients falling within grades I to IV [Table III]. Grade V patients have been excluded. 19 patients who were admitted in grade IV, but deteriorated to grade V after admission have also been excluded. In the 2 groups there were 13 operative deaths out of a total of 109 patients (11.9%)
[7. 1 % mortality for grades I,II and III]. Adding the 10 patients who died waiting for an angiogram or surgery, the overall mortality rises to 19.3%, and 8.4% patients died because of a delay in angiogram or surgery. At 1 month, good outcome was recorded in 64 (58.7%) patients from both the groups, including 24 (77.4%) patients in grade I, 30 (73%) patients in grade II, 9 (34.6%) patients in grade III and 1 patient (9%) in grade IV.
Overall mortality in group I was 9.6% compared to 14.0% in group 2. Mortality for the two groups excluding grade IV, was 4.1% and 10% respectively. At 1 month good outcome was recorded in 86.6% patients in grade I, 69.2% patients in grade II and 42.8% patients in grade III in group 1 [Table III] compared to 68.7% patients in grade I, 80% patients in grade II and 31.5 % patients in grade III in group 2 [Table III]. However none of the differences in Figures was statistically significant.
Angiographic vasospasm was seen in 27(25 %) patients, 11(21 %) patients were in group 1 and 16 (28%) patients in group 2. Ten (37%) patients of these 27 patients had a good outcome at 1 month. Five patients in group 1 (45%) and in group 2 (31%) each recorded a good outcome. The combined good outcome for grades I, II and III for both the groups was 48%. However the number in each group is too small to make any statistical comparison. Total combined mortality amongst patients with angiographic vasospasm was 11%, all the deaths being contributed by grade IV patients. None of the grade IV patients in either group with angiographic vasospasm had a good outcome at 1 month. The mortality was 50% in this category.
Delayed ischaemic neurological deterioration (DIND) attributable to ischaemia or vasospasm was encountered in a total of 55 patients during the post operative period. Nineteen (35%) of these patients also had angiographic vasospasm on the preoperative angiogram. Twenty five patients [angiographic vasospasm in 7 (28%)] were in group 1 and 30 [angiographic vasospasm in 12 (40%)] in group 2. Seventeen patients in group 1, and 10 patients in group 2 (i.e. 27 out of 55) made satisfactory recovery and were left with minimal or no deficit at the end of 1 month after surgery.
SAH or intraparenchymal clot on CT scan was seen in 43 (83%) patients in group 1 and 51 (89%) patients in group 2. Six (11%) patients in group 1 and 11(19%) patients in group 2 developed hydrocephalus following surgery and required a CSF shunt. Varying degrees of brain swelling was encountered at surgery in 52 patients, 30 in group 1 and 22 in group 2. In most of the patients brain swelling could be controlled with a combination of CSF drainage, diuretics and controlled hyperventilation. Severe brain swelling, requiring brain resection, was encountered in 3 patients in each group. One patient in group 2 required intra operative ventricular tapping. However in no case surgery had to be abandoned in either group.
The waiting period between admission and surgery ranged from 2 hours to 96 hours with a mean of 28 hours in group 1, and ranged from 22 hours to 428 hours with a mean of 95 hours in group 2. The waiting period exceeded 24 hours in group 1 if the next day after admission was not a working day or the angiogram laboratory was not available for any reason. The ictus to surgery interval for both groups is shown in [table IV]. Amongst the 10 patients who died waiting for an angiogram or surgery, 7 deaths occurred within 3 days of admission [Table III]. The duration of hospital stay ranged from 5 to 39 days in group 1 with an average of 12 days, and 6 to 63 days in group 2, with an average of 16 days.
At present it is well accepted that early surgical intervention has a number of advantages over delayed surgery in case of cerebral aneurysms. These include prevention of rebleeding, amelioration of vasospasm, prevention and treatment of ischaemic complications, prevention of medical complications, decreased psychosocial stress and shorter hospitalization.
SAH from ruptured intracranial aneurysms is a serious disease with mortality of 50% to 60%, and morbidity of 20% to 30%. Patients with aneurysmal rebleed have an even poorer prognosis, with a mortality rate of 70% to 90%. The primary purpose of neurosurgical treatment of ruptured aneurysms is to prevent a rebleed. The 1983 cooperative aneurysm study, that analyzed 2265 patients, concluded that the rate of rebleeding was maximum (4.1%) during the first 24 hours. The cumulative rate of rebleeding was 19% during the first two weeks. In the Danish aneurysm study group report of 1987, covering 1076 patients the daily rate of rebleeds during the first two weeks, calculated using a life table method, varied from 0.2% to 2.1%. The maximum rate of rebleeds was 2.1% between days 4 and 9. The cumulative rate of rebleeding was 16.8% during the first 2 weeks. A notable finding of this study was that a significantly higher rate of rebleeding was demonstrated in patients in poor clinical grades (Hunt and Hess grades III to V), compared to those in good clinical grades (Hunt and Hess grade I and II). A number of studies reported in the literature have now shown that early surgery does not expose the patient to increased risk of surgical complications.,,,,,, It also minimises the hazard posed by rebleeding and vasospasm. It is a common experience in almost all neurosurgical centres that even when the accepted policy is of early surgery, patients are still lost due to rebleed while waiting for angiogram or surgery. Roos et al reported that, in Dutch neurosurgical centres practising early surgery, rebleeding was the cause of poor outcome in 35% of all the patients with a poor outcome. The International cooperative study reported that number of patients with rebleeding increased progressively with longer intervals of planned surgery from approximately 6% for days 0 to 3 to approximately 22% for days 15 to 32. Of all the patients admitted to a neurosurgical unit within 3 days of aneurysmal SAH, 25% died. Vasospasm and rebleeding are the leading causes of death in addition to the effect of the initial haemorrhage'. Most favorable outcome and lowest mortality rate is achieved in alert patients (Grades I and II) in whom surgery is performed on days 0 to 3 or after day 10 post SAH. Patients who were drowsy preoperatively (Grades III to V) have the most satisfactory results when the surgery was performed after day 10. However, the advantage gained by late surgery is neutralized by an increasing incidence of rebleeding during the waiting period.
This avoidable delay in treatment is a factor that pushes up the mortality in alert patients has also been recognized by others. Sved et a1 have stated that 'delayed diagnosis and referral remain the major preventable problems in the management of patients with aneurysmal SAH'. Ljunggren et al analyzed 23 patients with aneurysmal SAH who were suitable candidates for early surgery, but early surgery could not be performed because of any reason. 13 patients died and 9 of these expired due to rebleed. They attributed the delay in treatment to patient delay, doctor's delay (delay at the local hospital) and neurosurgical delay (organisational problem at the tertiary centre). In an another study of 102 cases, 8 potentially salvageable patients rebled and died waiting for angiography or surgery.
Some neurosurgeons still prefer to delay surgical treatment in their patients. In their series of 400 patients with aneurysmal SAH, managed according to a delayed surgery protocol, Maurice - Williams and Wadley reported 88% good outcome at 1 year. But there were also 51 patients (12.5%) who experienced rebleeding and 33 of these died. Twelve patients, out of these 33 patients could have benefited from early surgery. Excluding the two patients who rebled and died soon after admission, still leaves 2.5-3% of all admissions or 3.5-4% of those in grades I-III at the time of admission who could have been saved by resorting to early or emergency surgery.
The primary reason of emergency surgery in group 1 in the present study was to minimise rebleed while awaiting definitive treatment. In the present series 10 patients (1 in grade II and 9 in grade III) suddenly deteriorated and died before any treatment could be undertaken. The cause of death was rebleeding and not vasospasm as the deterioration was sudden and not a gradual drift in the level of consciousness. The rebleed was also confirmed in all by a CT scan. Seven patients had been admitted in the hospital for 3 days before suffering another rebleed. One patient in grade III rebled more than 2 weeks after admission, as surgery had to be postponed because of severe chest infection. Delay in diagnosis, and referral by the primary treating hospital and infrastructural inadequacies at the tertiary referral centre contribute to postponement of treatment. In developing countries delay in transportation of patient to a specialised neurosurgical centre, further defers the treatment. Though it was not possible to expedite the treatment to the maximum possible extent, but still the average admission to surgery interval was reduced to 22 hours in group 1, compared to 95 hours in group 2. In a large neurosurgical referral centre, like the AIIMS, more than one patient with clinical SAH may be admitted on any day. There is an enormous load on the cerebral angiography laboratory, where both diagnostic as well as therapeutic angiograms are performed. Therefore occasionally deserving patients with clinical SAH may have to wait for diagnostic angiogram. This leads to an unavoidable delay, over which neither the neurosurgeon nor the neuroradiologist can exercise any control.
The difference between the results of the two groups, in the present series, are not statistically significant. Probably a larger number of patients in each group may give a definitive answer. It can be concluded that, in this series, the results of emergency surgery were comparable to those of planned early surgery. Results in grades IV and V continue to be dismal. We recorded a good outcome in only 9% of all patients admitted in grade IV. Bailes et al, have identified a group of patients in, grades IV and V, who could benefit from emergency surgery. They reported a good outcome in 54.3% of such patients at 3 months, if the preoperative scan did not show evidence of irreversible brain damage (massive cerebral infarction with midline shift or large basal ganglia or brain stem haematomas), and the intracranial pressure was controllable with a ventriculostomy.
Angiographic vasospasm has a typical temporal course, with onset at 3 to 5 days after the haemorrhage, maximal narrowing at 5 to 14 days and gradual resolution over 2 to 4 weeks, In about 50% of the cases, vasospasm manifests by the occurrence of a delayed ischaemic neurological deficit (DIND), which may resolve or progress to infarction. Cerebral vasospasm remains the greatest treatable cause of mortality and morbidity in patients who survive the ictus of SAH. In most contemporary series, up to 15% of such patients suffer stroke or death from vasospasm despite maximal therapy.,,, In the present series 19 (70%) out of 27 patients with angiographic vasospasm manifested DIND during course of their illness. This high incidence of deterioration could probably be because of strict criteria applied for labelling a patient as having DIND. However, the 11 % mortality in patients with angiographic vasospasm compares favourably with other series. The international cooperative study on the timing of aneurysm surgery reported that permanent morbidity and mortality resulting from vasospasm declined with increasing intervals to planned surgery, but a number of recent studies have not reported a worsening of vasospam after early surgery.,, We also feel that angiographic vasospasm, or increased velocities on transcranial doppler did not significantly influence the outcome unless accompanied by poor clinical grade (grade IV) at the time of surgery.
Patients with aneurysmal SAH, in clinical grades I to III, should be operated upon, as soon as possible, irrespective of the time since ictus. Delay in definitive surgical treatment after admission to a neurosurgical unit, can lead to death from rebleeding in an additional 8.4% of patients, who are in a stable neurological state. This additional mortality can be reduced if facilities for emergency cerebral angiogram, optimal theatre facilities and neuroanaesthesia backup are available, coupled with surgeon's willingness, expertise and capability to give equally good results while operating on an aneurysm at odd hours.