Patients and methods

All patients with space occupying supratentorial infarctions older than 55 years treated in our clinic between January 1998 and July 1999 were included in this analysis. Patients were not eligible for hemicraniectomy if they were older than 75 years or had severe medical or neurological comorbidity. All patients were treated conservatively following a fixed protocol with maintenance of a mean arterial pressure of 90–110 mm Hg, glucose serum concentrations of 5–6 mmol/l, and care for sufficient oxygenation and protection from aspiration.

The decision for surgery was based on a large hypodense infarct demarcation in the MCA territory regardless of lateralisation as visible on cranial CT with a midline shift of 5 mm or more indicating space occupying oedema. Clinically, deterioration of consciousness as a first sign of rising ICP had to be present. Hemicraniectomy was performed before other clinical signs of herniation. If additional signs of lateral herniation developed before surgery, an antioedematous treatment with mannitol infusions and controlled hyperventilation (paCO₂ 30 mm Hg) was started. Hemicraniectomy was performed with a diameter of at least 12 cm and an augmented dura plastic was inserted as described previously.5

Follow up data of all patients refer to the period of 3 to 9 months after ischaemic infarction. For assessment of clinical outcome and activities of daily living, the Barthel index and modified Rankin scale were used.

Results

Twelve out of 24 elderly patients with space occupying MCA infarction underwent hemicraniectomy (for summarised data see table 2); four patients were not operated on due to advanced age and eight because of pre-existing comorbidity (table3).

The mean (SD) age of the operated patients was 64.9 (6.7) years. Six patients had an ischaemic infarction restricted to the MCA territory, five patients had additional anterior cerebral artery (ACA) and one additional posterior cerebral artery (PCA) infarction. One patient had a progressive stroke, with an interval of 9 days between the initial symptoms and decompressive surgery. Without this patient the average time between onset of symptoms and craniectomy was 42.4 (SD 37.6) hours (range 13–130 hours). Four patients died during the period of follow up, but only one death resulted from transtentorial herniation after progression of an additional partial ACA infarction. Another 70 year old male patient died 17 days after hemicraniectomy from septic circulatory failure and two further patients died, one in a rehabilitation centre and one in a nursing home, presumably from heart failure and pulmonary embolism, respectively.

Of the eight surviving patients operated on (66%) none had a Barthel index score above 60 (range 0–60) and no patient had a Rankin score below 4 (range 4–5).

Mean age of the 12 medically treated patients was 72.5 (SD 8.9) years. Nine of these patients (75 %) died of transtentorial herniation in the first days after infarction. Another patient died shortly after ischaemia of pneumonia. The two surviving patients had Barthel index scores of 25 and 10 and Rankin score of 4.

Discussion

Current data suggest that hemicraniectomy is an effective treatment for large space occupying MCA infarctions in terms of survival rate and functional outcome. Indeed, mortality from herniation was low in our group of patients operated on, although overall mortality increased to 33% due to non-neurological complications. In line with published data on conservatively treated patients, mortality due to transtentorial herniation was much higher in the non-operated patients (75%).1 2

During follow up, none of the surviving patients, either surgically or medically treated, showed a Barthel index score above 60 and a Rankin score below 4; this indicates severely disabled and completely dependent patients in both groups. The poor functional results in the surgically treated patients contrast markedly with all other studies of hemicraniectomy for space occupying supratentorial infarctions.5-11

Factors with possible influence on the outcome–other than advanced age–may be the high proportion of patients with additional ACA or PCA infarctions or a delay between onset of symptoms and decompressive surgery.

In our series, six patients operated on had combined MCA/ACA or MCA/PCA infarctions. However, only one death in these patients could be attributed to herniation. In the surviving subgroup, the three patients with combined infarctions did not show worse Barthel index scores compared with the five patients with isolated MCA infarctions. In a recent large hemicraniectomy study,5 more than one third of patients had combined infarctions, but no negative impact on the functional outcome in this subgroup was reported. These findings argue against a major negative influence of combined infarctions on the outcome in our series compared with others.

Previous studies suggest a better outcome of patients treated with early rather than delayed hemicraniectomy.6 9 11However, no inverse correlation between time to surgery and outcome was detected in our patients. The average delay of 42 hours was shorter than in most earlier studies.6 8-11 Thus, the poor outcome in our series cannot be explained by a delay between stroke and hemicraniectomy.

In summary, advanced age is the only differing variable when comparing our patients who were operated on with earlier investigations. Although mortality from herniation was low after hemicraniectomy, clinical outcome and level of independence were poor. Even the patient with the best Barthel index score of 60 still needed massive assistance for his daily life. Our data suggest that in elderly patients with large space occupying supratentorial infarctions decompressive surgery should not be performed.