Introduction

The outcome of patients with non-traumatic subarachnoid haemorrhage (SAH) has improved because of advances in neurosurgery, interventional neuro-radiology and neuro-intensive care [29]. Cerebral vasospasm is a known and feared complication of SAH and considered as one of the major factors contributing to morbidity and mortality after the haemorrhage giving rise to ischemia in the whole brain or parts of it. The frequency of vasospasm is reported to be as high as 50%. The pathophysiological mechanism behind the vasospasm is multifactorial. Several treatment options, including HHH (hypertension, hypervolemia, hemodilution), calcium antagonists, angioplasty, endothelin-receptor antagonist and statins have been used in order to prevent and treat vasospasm [13, 14, 20, 26–28]. None of the treatments are fully effective in counteracting the cerebral vasospasm.

The diagnosis of cerebral vasospasm in the clinical setting is either based on neurological deterioration of the patient or on increased flow velocities in cerebral vessels monitored by transcranial Doppler sonography (TCD) or diagnosed by cerebral angiography. A flow velocity exceeding 120 cm/s in the middle cerebral artery (MCA) is considered to be a sign of cerebral vasospasm.

Prostacyclin (PGI2), produced in the vascular endothelium, is a potent vasodilator and is a strong inhibitor of platelet aggregation [8, 15, 16]. Tromboxane A2 (TXA2) is on the other hand a powerful vasoconstrictor and promoter of platelet aggregation [15]. An imbalance of PGI2 and TXA2 may result in vasospasm [23]. Experimental and clinical data show that prostacyclin may reduce vasoconstriction elicited by SAH [2–4, 19]. Beneficial effects of low-dose prostacyclin (0.5 ng/kg/min) infusion have been reported in severe traumatic brain injury [11, 18].

In this pilot study we treated five SAH subjects with prostacyclin suffering from cerebral vasospasm refractory to high-dose nimodipine treatment. The flow velocity in the MCA was registered. Outcome was evaluated at 3 and 12 months after the SAH.

Patients and methods

This study is a prospective open pilot study. Inclusion criteria were proven SAH because of ruptured aneurysm and development of vasospasm as defined by transcranial Doppler (TCD) measurement with mean flow velocity (MFV) exceeding 120 cm/s in MCA. The velocity in the MCA was monitored with the TC2-64 transcranial Doppler device (EME, Überlingen, Germany). As the Doppler equipment used is not suitable for calculation of the Lindegaard index, it was a prerequisite that a pronounced difference should exist between the systolic and diastolic blood flow velocity, and the rise and fall of the Doppler signal should be steep. Further, the vasospasm should be resistant to high-dose i.v. (15 ml/h = 3 mg/h) nimodipine infusion and the subjects in severe condition in need of sedation and artificial ventilation and neuro-intensive care.

The diagnosis of SAH was based on typical symptoms, CT scan and if necessary lumbar tap. CT angiography was performed as the first investigation to identify the bleeding source. If no bleeding source was detected, conventional digital subtraction angiography was performed. The patients were treated with either endovascular treatment or open surgery with clip ligature of the aneurysm. Early treatment of the aneurysm was employed.

The patients were treated in an intensive care setting, sedated due to unconsciousness or high intracranial pressure (ICP) and artificially ventilated. All subjects had invasive arterial lines for monitoring of systemic arterial blood pressure and central venous lines for monitoring of central venous blood pressure. The patients had either a Codman MicroSensor™ (Johnson & Johnson Professional Inc., Raynham, MA) or a ventricular drain with an external pressure transducer for monitoring of the ICP. The MicroSensor™ was calibrated according to the manufacturer’s instruction. The zero level of the ventricular catheter was at the pre-auricular level, and the zero level for the blood pressure was set at the heart level. Mean arterial blood pressure (MAP), ICP and cerebral perfusion pressure (CPP) and all vital parameters were continuously monitored and displayed on a bedside monitor and digitally stored. When ICP was measured using the ventricular catheter, it was closed for about 10 min in order to get reliable values. The preset goal of the CPP level was ≥70 mmHg and ICP <20 mmHg. The patient was initially treated in a supine and flat position. Later on, a light head elevation was accepted. Subjects were kept normo-volemic and normo-tensive, using packet red blood cells (Hb >110 g/l) and albumin (S-Alb >40 g/l). In order to prevent the development of cerebral vasospasm, patients were initially treated with a normal dose of nimodipine infusion i.v. (2 mg/h). When vasospasm was observed in the MCA, the nimodipine dosage was increased to 3 mg/h. This has been shown to bring some vasospastic patients under control [31]. Epoprostenol (Flolan, GlaxoSmithKline) was infused i.v. in a dose of 0.5 ng/kg/min for 72 h if the vasospasm was not successfully treated by the above-mentioned measures.

Results are reported as mean ± SEM or median with range. For testing statistically significant differences, the non-parametric Kruskal–Wallis test was applied. The local ethics committee approved the study (dnr 03–474).

Results

Five patients fulfilled the criteria, three females and two males. Mean age was 49 ± 6 years, median Hunt–Hess 4 (2–5) and Fisher 4 (3–4). Figure 1 depicts the ICP, CPP and MAP 24 h before and during prostacyclin treatment. No statistically significant effect of prostacyclin treatment was observed on these parameters. Typically, the vasospasm developed between the 3rd to 4th day after the SAH. This was confirmed by cerebral angiography. The MFV was 163 ± 24 cm/s during the period of 2 mg/h nimodipine infusion. Nimodipine infusion was increased to 3 mg/h. The MFV was 199 ± 31 cm/s 24 h later. In two of the patients, selective intra-arterial infusion of 2 mg of nimodipine was administered into the vasospastic region without any effect on the radiological vasospasm. In this situation prostacyclin infusion was started, and 48 h later the MFV decreased to 122 ± 27 cm/s and after 72 h 92 ± 6 cm/s. Although the material is very small, the effect is statistically significant as evaluated by the Kruskal–Wallis test (p = 0.048). At 3-month follow-up the median GOS was 4 (3–4) and at 12 months 4 (3–4).

Fig. 1
figure 1

Hourly intracranial pressure (ICP), cerebral perfusion pressure (CPP) and mean arterial blood pressure (MAP) 24 h before and during prostacyclin treatment started at zero time

Full size image

Discussion

The mechanisms behind cerebral vasospasm elicited by SAH are complex and intriguing. However, it seems that blood has to enter the basal cisterns, and the oxy-heamoglobin formed seems to be crucial for the development of vasospasm [13, 14]. Effects on many neurotransmitter systems as well as on endothelial function have been proposed. Both direct and indirect neurogenic effects have also been suggested as the cause of the vasospasm. Experimental and clinical research has suggested that inflammation may play a role in the development of vasospasm [9, 22].

In our study we treated vasospastic subjects with low-dose prostacyclin infusion with severe nimodipine refractory vasospasm after SAH. The vasospasm markedly decreased, defined as a decrease in flow velocity measured by TCD. As typical for SAH patients, ICP can often be brought under control by sedation, artificial ventilation and cerebro-spinal fluid drainage. Our preset goal of ICP <20 mmHg was achieved as well as a CPP ≥70 mmHg. No episodes of statistically significant hypotension were observed during the prostacyclin treatment. The subjects represented a group with a very severe SAH normally presenting an unfavourable outcome. Fortunately, all patients recovered to a considerably good outcome. Our results confirm an initial pilot study [24] showing a remarkable clinical improvement in SAH patients with clinical vasospasm treated with 1 ng/kg/min prostacyclin. In fact, we have also observed a beneficial side effect of prostacyclin in some SAH patients with severe respiratory problems reversing the need of high inspiratory oxygen concentration during treatment in the ventilator. However, this effect has not been systematically studied jet.

Prostacyclin has been shown to possess several potent biological effects, including platelet aggregation inhibition, prevention of leukocyte adhesion to the endothelium, inhibition of blood–brain-barrier leakage and a dose-dependent vasodilator effect [17]. These physiological effects of prostacyclin would be beneficial in preventing and perhaps treating vasospasm. A decrease of the endothelial-related prostacyclin could result in aggregation of platelets and vasoconstriction, finally eliciting a delayed cerebral ischemia. In isolated cerebral arteries, prostacyclin causes a relaxation and counteracts a vasoconstricting effect of cerebrospinal fluid from SAH subjects [2–4]. A disproportionate elevation of prostanoids in the CSF after experimental SAH has been reported with an overweight of constricting prostanoids [22]. Interestingly, intraventricular blood or re-bleeding in humans suffering SAH has markedly increased levels of prostanoids in the CSF [21].

Still more than 25 years after the discovery of prostacyclin, the clinical use is mostly as an anticoagulant during haemodialysis and as a vasodilator in patients with pulmonary hypertension [7]. The explanation for the sparse clinical use of prostacyclin may be the fear of inducing hypotension in critically ill patients. A favourable effect of low-dose prostacyclin (epoprostenol 0.5 ng/kg/min) has been reported in severe traumatic brain injury [11], and no adverse effects of this dose have been detected [18]. This dose is significantly lower that the dose recommended for the treatment of pulmonary hypertension. The biological effects of nitric oxide and prostacyclin are similar, and the release of the two substances is coupled. Indeed, nitric oxide donors have been proposed as a pharmacological treatment for cerebral vasospasm [5].

Several treatment options to improve cerebral blood flow in order to prevent or treat the cerebral ischemia resulting from the vasospasm have been applied. The so-called HHH treatment, including hypervolemia, hypertension and haemodilution, has been extensively used. However, the efficacy of HHH treatment has been questioned [6, 12, 25]. During the late 1980s, the calcium antagonist nimodipine was introduced as a treatment of vasospasm in order to reduce the neurological deficits because of delayed cerebral vasospasms [20]. Several reports have shown beneficial effects of nimodipine, and the drug seems also to have a neuroprotective effect [13]. Other treatments for cerebral vasospasm include angioplasty, endothelin-receptor antagonists and statins. However, despite the beneficial effects of the above-mentioned measures and a more sophisticated neuro-intensive care treatment with multi-modal monitoring of the patient, vasospasm is still an existing problem after SAH.

A weakness of the present study is the few subjects studied. However, the study intended to study only patients that were not responding to other measures. One can also question the TCD method used to detect vasospasm and not using the Lindegaard index. Several publications have shown TCD’s usefulness, particularly in detecting vasospasm in MCA [1]. It has also been demonstrated that the use of the Lindegaard index does not improve the predictive value of TCD monitoring [10, 30].

In conclusion, partially effective ICU regimes and pharmacological treatments have improved the outcome, but no absolute preventive measures for vasospasm are available. High-dose nimodipine may decrease the cerebral vasospasm within 24 h [31]. This was not the case in the presented subjects. In this study we showed that a low dose of prostacyclin may have a beneficial effect in reducing established nimodipine-resistant vasospasm. Indeed, a prospective, randomised, blinded study is needed to definitely show whether the effect of prostacyclin can reduce vasospasm after SAH. Thus, a previously shown effective treatment of cerebral vasospasm can be a good alternative to newer treatment measures.