Assessment of Prognosis and Indication for Surgery on the Basis of Cluster Analysis in Intracerebral Hematomas

Abstract

Although the surgical evacuation of intracerebral hematomas is rather common, there is still considerable controversy regarding the indications for this treatment [2, 4, 5, 7]. The clinical trial of McKissock et al. [3] more than 30 years ago suggested that conservative treatment is preferable in most cases. Their conclusion was mainly due to the rather poor surgical outcomes of that time and not attributable to particularly successful medical therapy. In the meantime, the diagnosis has been facilitated by the availability of the computed tomographic (CT) scan. This has led to the improvement of the operative treatment owing to the better assessment of size and location of the lesion by CT imaging. Less invasive techniques, such as small craniectomies combined with stereotaxic or ultrasonic retrieval necessitating only minimal cortical incisions, have been developed. It can be assumed that this progress has made the results of this early clinical trial obsolete. However, more recent randomized trials are not available. Therefore we performed a retrospective study to answer the question of when to operate and when to treat conservatively. Like all retrospective trials, our study has the inevitable deficiency that no randomization was performed. To compensate for this, we performed a cluster analysis. This statistical procedure allows for objective grouping of the patients on the basis of clinical data. Thus it is possible to compare the results of conservative and surgical treatment within homogeneous groups of patients.