SUBJECTS AND METHODS

The Neurosciences department admits around 350 patients with cerebrovascular diseases of various etiologies. Between June 1997 and June 2000, 19 patients were selected for decompressive hemicraniectomy following progressive swelling after MCA occlusion. All patients had routine hematological and biochemical profile, ECG and trans-thoracic echocardiography, cranial CT scan, and Doppler studies of the extra-cranial cerebral vessels. None had significant co-morbid medical or neurological disease. The neurological status was measured on the National Institute of Health Stroke Scale (NIHSS)¹² on admission. All patients received maximal medical management of intracranial hypertension and systemic support. Surgery was performed when there was evidence of transtentorial herniation. A wide fronto-temporo-parietal craniectomy was performed. The dura was incised but no duroplasty performed. Temporalis fascia was placed over the arachnoid overlying the dural defect. Infarcted brain tissue was not excised. Post-operatively, the patients were managed in the intensive care unit. The NIHSS scoring was repeated one week after surgery. Patients were assessed three months after surgery and the outcome was quantified on the Barthel Index (BI)¹³ and the Rankin Scale (RS).¹⁴ Cranioplasty was done about three months after discharge.

STATISTICAL ANALYSIS

SPSS 10 for Windows package was used for statistical analysis. All values are expressed as means. Statistical dispersion is expressed as standard deviation. Kruskal-Wallis and Mann-Whitney non-parametric tests were used to compare means between groups. Statistical significance was assigned to p value of <0.05.

RESULTS

There were 19 patients, 15 males and 4 females, with mean age of 46.5 yrs (SD 11.8; range 27–76 yrs). Fifteen patients (79%) were under 50 years of age. Ten patients (53%) had left-sided hemicraniectomy while the rest had right-sided decompression. The mean interval between stroke-onset and admission to hospital was 25.3 hours (SD 26.8; range 1–100hrs). The mean interval between stroke-onset and decompressive surgery was 60.3hrs (SD 23.9; range 20–103 hrs). The mean NIHSS score prior to surgery was 20.5 (SD 2.6; range 17–26). The NIHSS score a week after surgery was 10.5 (SD 4; range 6–22). The mean duration of hospital stay was 35.3 days (SD 16; range 19–88 days). At 3 month follow up, the mean BI was 56.3 (SD 19.2; median 55; range 25–90). The RS showed severe handicap in only 21%, while the majority showed significant improvement. One patient (5.2%) died due to post-operative meningitis. Among the patients with dominant hemisphere stroke, the speech function improved in most. We compared the duration of hospital stay, stroke onset to admission times, stroke onset to surgery time, NIHSS scores and BI among patients under 50 and those above 50 years of age. There were no significant differences between the groups for duration of hospital stay, stroke onset to admission time, stroke onset to decompressive surgery time, NIHSS scores or BI. The BI, however, was significantly higher (60.7; SD 18.9; median 57.5; range 30–90) for the below 50 years group as compared to 41.3 (SD 11.8; median 45; range 25–50) for the below 50 years group (p=<0.048).

DISCUSSION

We have described a series of patients with massive hemispherical infarction who underwent decompressive hemicraniectomy. The principal aim of surgery was to save life, since the majority (79%) of them were under 50 years and did not have concomitant serious medical illness. Even though decompressive craniectomy for supratentorial infarction has been done in a few patients sporadically over four decades, it is only in the last two decades that the treatment has been studied systematically.^3,15 A recent Cochrane review, however, has shown that there is no evidence from the extant literature of an evidence basis for the use of this procedure.¹⁶

Our study however, add to the previous findings that suggest both mortality and functional outcome may be better after surgery. This is supported by experimental data from rat models.^17,18 Functional outcome is better in younger patients^10,19 and when surgery is performed early in the course of the stroke.^3,9 It has generally been felt that hemicraniectomy must be offered only to patients with non-dominant hemispherical strokes, the reason being that speech is unaffected and functional outcome would therefore be better.^4–7,10 Several studies have however reported good language recovery in some patients.^3,8 We also found that all patients with aphasia had improved and could communicate usefully. Even though most had a degree of expressive dysphasia, they regained their ability to comprehend.

In our study, only one patient died (mortality rate 5.2%), 50 days after surgery because of postoperative meningitis and abscess formation and not due to herniation. The mortality rates in various reports vary from 0% to 27% (see table 1). The problem with our study, like many studies, is that it is retrospective and may be subject to selection bias. Nevertheless, we have demonstrated that the procedure can be performed safely in our set up in a developing country.

We conclude that although hemicraniectomy can be very effective, randomised controlled trials are needed to properly evaluate its place in the management of patients with space-occupying hemispheric stroke.