Effects of a single dose of dantrolene in patients with cerebral vasospasm after subarachnoid hemorrhage –a prospective pilot study

Abstract

Background and Purpose

New therapies for cerebral vasospasm (CVSP) after subarachnoid hemorrhage (SAH) are needed because of its high morbidity and mortality. We investigated the feasibility and safety of a single dose of intravenous (IV) dantrolene and its effect on transcranial Doppler (TCD) in CVSP after SAH.

Methods

In a prospective open label single dose ascending safety trial, five patients received IV-dantrolene 1.25mg/kg and the next five patients 2.5mg/kg over 60 minutes. All other infusions were kept steady and hemodynamic parameters were recorded. TCDs were performed at t₀, t_45min, t_90min and t_135min relative to infusion start. Basic chemistries, serum osmolality, arterial blood gas and liver enzymes were measured before and after.

Results

Laboratory values and hemodynamic parameters remained unchanged except for a decrease in the systolic blood pressure in the low dose group (-8 mmHg; 95% CI [-26 to 10 mmHg]; p=0.027). After correcting for this decrease in blood pressure, peak systolic TCD velocities (PSV) decreased significantly (-26 cm/s; 95% CI [-47 to -5 cm/s]; p=0.02), with a borderline change in mean velocities in the low dose (-16 cm/s; 95% CI [-36 to 4 cm/s]; p=0.07), and PSV in the high dose group (-26 cm/s; 95% CI [-56 to 5 cm/s]; p=0.05).

Conclusions

In this pilot study, a single dose of IV-dantrolene in CVSP after SAH appears feasible while inhibiting vasoconstriction in the low dose group, but it may lower blood pressure. Our study provides useful data for the design of larger future studies.

Introduction

Subarachnoid hemorrhage (SAH) is one of the most devastating forms of stroke affecting primarily young patients before the age of 65, and has a fatality of 50% in the first 30 days^1-2. For survivors of the initial insult, cerebral vasospasm (CVSP) is the leading cause of disability and death^2-3, and therefore points to an inviting therapeutic target.

CVSP occurs in 70% of patients with SAH, and one third develops neurologic deficits⁴. Available treatments for CVSP are limited, consisting of hypervolemic, hypertensive (HHT) therapy, and cerebral angiography with intervention, which carry the risks of pulmonary edema, myocardial infarction, stroke, and vessel rupture with death, and are labor-intensive and expensive^5-7. Alternative treatments for CVSP offer the hope of improving outcome.

Although CVSP is a multi-factorial disease process, the common pathway of vasoconstriction is the continuous elevation of intracellular Ca²⁺-levels due to a combination of influx from extracellular Ca²⁺, and release from the largest intracellular Ca²⁺ store, the endo/sarcoplasmatic reticulum mediated by the ryanodine receptor (RyR)^8-11. Nimodipine and nicardipine, both L-type specific Ca²⁺-channel blockers, have been used for CVSP^12-13, but are only partially effective, possibly because they affect only the influx of extracellular Ca²⁺, and have no effect on RyR-mediated intracellular Ca²⁺-release. Dantrolene is a known RyR inhibitor and already U.S. Food and Drug Administration (FDA) approved for other indications^14-15. There is evidence that dantrolene is neuroprotective^16-20. Furthermore, in an ex-vivo rat model, dantrolene has been shown to inhibit cerebral vasoconstriction alone as well as in combination with nimodipine²¹. Similarly, a small human study has suggested that dantrolene may attenuate CVSP after SAH²². Combining L-type specific Ca²⁺-channel blockers with a RyR blocker may be the key therapeutic target in CVSP.

Up to 30-40% of SAH patients may experience hyponatremia for various reasons²³, which may be exacerbated by intravenous (IV) dantrolene given its solution in free water. Further, the combination of nimodipine with dantrolene may theoretically cause hypotension. Therefore, we conducted a prospective open label single dose ascending safety trial in patients with transcranial Doppler (TCD) as an efficacy biomarker. We proposed that only if this pilot study showed any effect, future larger studies of dantrolene in SAH-related CVSP would be warranted.

Materials and Methods

Study Design

We conducted a prospective, open-label, single-center study using a single infusion of IV dantrolene with two dose tiers (http://clinicaltrials.gov NCT00964548). The study was approved by the Massachusetts General Hospital (MGH)/Partners Health Care institutional review board. Written, informed consent was obtained from all patients or the health-care proxies. Using a convenience sample of ten patients, the first five patients received 1.25 mg/kg, while the subsequent five patients received 2.5mg/kg. We chose 2.5mg/kg as the highest dose for two reasons: in prior studies on healthy volunteers maximum grip strength weakness occurred at approximately this dose (2.2mg/kg)²⁴ and higher doses may pose too high risk for respiratory muscle weakness for our mostly non-intubated patient population; furthermore, based on its pharmacological properties, IV dantrolene infusions every 6 hours result in steady dantrolene serum levels²⁴. Therefore, in concordance with the maximum FDA approved daily cumulative total dose of 10mg/kg, an every-6-hour dosing regimen equates 2.5mg/kg IV every 6 hours. In order to investigate the effect of a lower dose on TCD velocities, we chose half of the high dose (1.25mg/kg).

Study Protocol

Patients were approached for consent if they met the following inclusion criteria: aneurysmal SAH, as proven by head computer tomography (CT) and CT-angiogram, MR-angiogram or cerebral angiography, admitted to our neuroscience intensive care unit (neuroICU), undergoing standard-of-care daily TCD, and at risk to develop cerebral vasospasm. Exclusion criteria were inability to obtain consent from the patient or health care proxy, age < 18 years, pregnancy, traumatic SAH, patients on verapamil (possible drug-drug interaction with dantrolene), prior history of cirrhosis or hepatitis B or C, or any two of the following liver enzymes elevated to greater than: alanine aminotransferase (ALT) >165 Units/L, aspartate transaminase (AST) >120 Units/L, alkaline phosphatase (AlkPhos) >345 Units/L (three times upper limit of normal). As requested by our IRB, patients were consented prior to the occurrence of vasospasm in order to increase the likelihood of patients self-consenting for the study.

Figure 1 summarizes the study design. All patients received nimodipine 60 mg orally every 4 hours. When the daily routine TCD exam suggested evidence of vasospasm, patients were started on HHT as standard-of-care at the MGH neuroICU: normal saline infusion at 150-200ml/hr, MAP goal >100 mmHg using phenylephrine with the addition of norepinephrine if needed in order to achieve the MAP goal. Likewise, 250ml 5% albumin was infused every 6 hours as needed. An arterial line for continuous blood pressure monitoring and a central venous line for medication application and central venous pressure monitoring were routinely placed. Patients only received study drug if they had CVSP on a confirmatory repeat TCD exam by a study physician (S.M. or J.S.).

Figure 1. Study design.

The study is described in detail in the methods section. Abbreviation: Chem7, basic chemistries.

Cerebral Vasospasm criteria

Cerebral vasospasm was defined by the following TCD criteria:

1. a >50% mean flow velocity (MFV) increase from the baseline MFV (baseline was the first TCD measurement, usually within 24 hours of admission), or

2. peak systolic velocities (PSV) of 200 cm/s or higher in the middle cerebral artery (MCA) or anterior cerebral artery (ACA) (for MCA with a concurrent ipsilateral Lindegaard ratio of 3 or higher)²⁵, or

3. PSV of 120 cm/s or higher in the posterior cerebral artery (PCA) or basilar artery, or

4. any daily 100 cm/s PSV increase from the previous day, or

5. any longitudinal MFV increase of 80 cm/s or more.

Infusion of study drug

Once the repeat TCD confirmed elevated velocities suggestive of vasospasm, a single dose of IV dantrolene was infused over 60 minutes at a dose of 1.25 mg/kg for the first five patients and 2.5 mg/kg for the subsequent five patients.

Laboratory measurements

Baseline serum laboratory values were examined prior to the infusion at t₀ (ALT, AST, AlkPhos, arterial blood gas [ABG], chemistry and osmolality). At the end of the study period (t_135min) all serum laboratories were repeated except ALT, AST and AlkPhos which were repeated 24 hours after the baseline (t_24h).

TCD measurements

At t₀, patients underwent a confirmatory TCD examination (baseline) by one of the investigators (S.M. or J.S.) selective to the vessel with elevated velocities, and the corresponding contralateral vessel (if applicable), or the vertebral arteries ([VA], for basilar artery). TCDs throughout the study were repeated by the same examiner and were exclusively selective in order to remain within the time constraints of the study. A full TCD exam may take ≥ 30 minutes, blurring the time intervals for the TCD aspect of the study, while the selective study TCD took between 5-10 minutes. TCD have been shown to be sensitive in detecting CVSP after SAH when performed by the same experienced examiner²⁶. TCDs were repeated three times every 45 minutes: once at t_45min during the dantrolene infusion, and twice after the dantrolene infusion at t_90min and t_135min. Vessels were insonated at pre-defined depths (in mm from the temporal bone for middle cerebral artery [MCA] and anterior cerebral artery [ACA] and occiput for VA and basilar artery [BA]) in order to standardize the TCD exam and allow calculations of mean change from baseline: the MCA was insonated at depths 65, 60, 55, 50 and 45 mm, ACA at 65, 70 and 75 mm, VA at 60, 65, 70, 75, and 80 mm and BA at 85, 90, 95, 100, 105 and 110 mm. PSV and MFV were recorded. For each patient, the baseline velocity at the corresponding depth for each vessel in vasospasm was set as zero. For each vessel at each corresponding depth, the change in PSV and MFV from baseline was calculated and averaged for each vessel in vasospasm. TCD were performed with a Spencer PMD 150^® TCD machine (Spencer Technologies, Seattle, WA).

Hemodynamic measurements

During the infusion period hemodynamic parameters were recorded every 10 minutes (heart rate [HR], blood pressure, central venous pressure [CVP], intracranial pressure [ICP] if applicable and cerebral perfusion pressure [CPP] if applicable). From t₀ until t_135min the vasopressor dose was left unchanged.

Statistical Analysis

Laboratory data were analyzed using the Wilcoxon sign rank test. Vital signs and TCD velocities were analyzed using repeated-measures ANOVA. P≤0.05 was considered significant. Post-ANOVA analysis was performed using Bonferroni's multiple comparisons test with adjusted p-values (p≤0.001). Analyses were performed using SAS^® 9.2, SAS Institute Inc., Cary, NC, and PRISM^® 5.03, GraphPad Software Inc., La Jolla, CA.

Results

Between June 2007 and October 2008, ten patients received the study drug and completed the study protocol (Figure 2). A total of 16 vessels found to be in TCD vasospasm were included in the analysis. Baseline characteristics are shown in Table 1 and Supplementary Table 1.

Figure 2. Screening and enrollment (CONSORT diagram).

A total of 28 patients were approached for consent, of which 24 consented. Ten patients received study drug, while 14 did not for various listed reasons.

Table. Baseline Characteristics.

	Low dose group (1.25 mg/kg) n=5	High dose group (2.5 mg/kg) n=5
Vessels of interest (in vasospasm)	6	10
Mean age (SD) [years]	42 (9)	44 (10)
Female	60%	40%
Ethnicity
Caucasian	40%	80%
Asian	20%
Hispanic	40%	20%
Admission Hunt & Hess Scale median (IQR)	2 (2;5)	3 (2;4)
Admission Fisher Group median (IQR)	3 (3;3)	3 (3;3)
Aneurysm location [n]
Middle cerebral artery	4	1
Anterior communicating artery		1
Basilar artery	1
Post inferior cerebellar artery		2
Internal carotid artery		1
Aneurysm treatment
Clipped	80%	60%
Coiled	20%	40%
Mechanically ventilated [n]	2	1

Laboratory values

No differences were noted in the basic chemistries, osmolality, ABG or liver function tests before and after the dantrolene infusion in neither dose group (Supplementary Table 2).

Hemodynamic values

In both dose groups, HR, MAP, ICP, CPP and CVP remained unchanged during the study period (Supplementary Figure). There was a significant decrease in the SBP in the low dose group (maximal -8 mmHg; 95% CI [-26 to 10 mmHg]; p=0.027), but not the high dose group. Post-hoc ANOVA analysis using Bonferroni's test revealed that there was not a specific time point during which this decrease in SBP occurred. This finding was driven by a single patient. Exclusion of this patient for this particular analysis showed no significant change in the SBP in the remaining patients (p=0.28).

TCD results

We observed a significant decrease from baseline in the mean change of both PSV and MFV in the low dose group (-27 cm/s; 95% CI [-43 to -11 cm/s]; p=0.004 and -18 cm/s; 95% CI [-33 to -3 cm/s]; p=0.01 respectively; Figure 3). After excluding TCD data from the low blood pressure outlier, the mean change of PSV remained significant (-26 cm/s; 95% CI [-47 to -5 cm/s]; p=0.02), with a borderline change in MFV (-16 cm/s; 95% CI [-36 to 4 cm/s]; p=0.07). In the high dose group, there was a trend towards a decrease in PSV (-26 cm/s; 95% CI [-56 to 5 cm/s]; p=0.05), but no significant decrease in MFV (-13 cm/s; 95% CI [-34 to 8 cm/s]; p=0.14; Figure 3). Mean baseline Lindegaard ratios (calculated if the middle cerebral artery was in CVSP) were 3.2±1 for the low dose (n=5), and 3.8±0.8 for the high dose group (n=6). There was no significant change in the Lindegaard ratios in either dose group (at t₁₃₅: low dose mean 2.8; [95% CI 1.4 to 4.2] ; p=0.72; high dose mean 3.4; [95% CI 1.9 to 5] ; p=0.24).

Figure 3. Transcranial Doppler flow velocities.

Shown are mean changes in PSV and MFV with 95% CI at 45, 90 and 135 minutes from baseline in the vessel of interest (vessel in vasospasm) in the low dose group (all patients, panel A; single low blood pressure outlier excluded, panel B), and high dose group (panel C). Asterix (*) indicate significant changes by repeated measures ANOVA.

Discussion

Dantrolene is an inviting new therapeutic agent for CVSP as it has a plausible biologic mechanism, is already FDA approved, has a well-known and well-tolerated side-effect profile, is neuroprotective^16-20 and may inhibit cerebral vasoconstriction as shown previously^21-22. Giving IV-dantrolene to SAH patients on nimodipine and HHT may pose some safety concerns, as 20 mg dantrolene are diluted in 60 ml of sterile water with 5% mannitol and sodium hydroxide yielding a pH of 9.5²⁰. Especially in SAH patients, this may lead to hyponatremia, hypotension (particularly in combination with nimodipine), hypoosmolality and alkalosis. Furthermore, dantrolene may cause liver toxicity^{20, 27}. We conducted this pilot study to explore these safety concerns and the short term effects of a single dose of dantrolene in this specific patient population.

We did not observe any significant hyponatremia, hypoosmolality, changes in pH or liver enzyme abnormalities after a single dose. However, such side effects may only become apparent with cumulative doses, and future studies with repeated dantrolene doses should continue to monitor for these.

We observed a decrease in the SBP in the low dose group, but not in the high dose group. One single outlier in the low dose group likely explains this finding, as no changes in SBP were seen after exclusion of this patient. The etiology of hypotension in this particular may be idiosyncratic or a true systemic effect of nimodipine and dantrolene. No other human studies have linked dantrolene to hypotension, neither alone or in combination with calcium-channel blockers. In the ischemic swine heart, the combination of amlodipine and dantrolene did not worsen the hypotension and AV-nodal conduction delay induced by amlodipine²⁸. Our study is susceptible to outliers given the small sample size. Future studies warrant close observation of blood pressure during the study period.

Although this study was not powered to study efficacy, the observed changes in TCD velocities are interesting. We detected a significant decrease in the mean change of PSV with a trend towards decreased MFV in the low dose group, and a trend towards a decrease PSV in the high dose group. A possible explanation is that the receptor binding and drug kinetics may follow a U-shape curve with less inhibition of vasoconstriction with higher doses. It is unclear how repeated doses affect receptor binding and drug kinetics. We did not evaluate patients for cerebral autoregulation which may be impaired after SAH. After excluding the single patient with a blood pressure decrease in the low dose group from the TCD analysis, the mean change in PSV, but not MFV remained significant, underlining the importance of considering impaired cerebral autoregulation in the analysis of TCD data in SAH.

Limitations

Apparent limitations of our study include the small sample size which significantly limits statistical analysis and interpretation, the administration of a single dose as opposed to repeated dosing over a longer period of time, the unblinded design, the lack of radiological confirmation of the dynamic changes of CVSP as well as clinical outcome after dantrolene administration. Our data does not allow us to comment on possible accumulation, end-of-dose or rebound effects due to the short study period (135 minutes). Moreover, we did not study the effect of dantrolene on cerebral blood flow, blood volume, or metabolic changes on brain tissue by means of microdialysis or local tissue oxygenation.

Conclusions

Due to its plausible biologic mechanism, its neuroprotective and inhibitory effects on cerebral vasoconstriction, dantrolene might be well suited for the treatment or prevention of CVSP after SAH. This study shows feasibility and safety of a single infusion of dantrolene, as well as inhibition of TCD vasoconstriction in one dose tier. Future prospective studies using repeated doses of dantrolene over the entire vasospasm period are warranted.