Introduction

Vasospasm after rupture of cerebral aneurysms is the primary cause of heightened patient morbidity and mortality and is a relatively common complication of subarachnoid hemorrhage (SAH) seen by angiogram in two-thirds of patients with rupture and one-third of those with clinical symptoms [1, 2]. Despite rapid advancements made in surgical and endovascular treatments of cerebral aneurysms, mortality/morbidity due to vasospasm has remained unchanged. The etiology of vasospasm is thus poorly understood, leaving a void in definitive treatments for severely symptomatic patients who are then facing delayed cerebral ischemia.

Conservative treatment is a priority in instances of vasospasm, but oral nimodipine is the only agent to date with known benefits [1,2,3]. Other interventions, including intra-arterial (IA) chemoangioplasty, intrathecal drug infusion, balloon angioplasty, or angioplasty by retrievable stent, have also been performed in severely symptomatic patients refractory to medical treatment [2,3,4]. However, mechanical angioplasty via balloon and/or retrievable stent has very limited application if proximal arterial segments are involved [3, 5, 6]. Although recent use of intrathecal drug infusion is aimed at boosting intracerebral drug concentrations while reducing systemic hypotension [7], cerebrospinal fluid infection may be a related risk.

There are ample studies of chemoangioplasty by IA infusion of drugs, particularly calcium channel blockers (CCBs) and phosphodiesterase inhibitors (e.g., papaverine or milrinone) [3, 5, 6, 8,9,10,11,12,13,14,15,16,17,18,19,20,21]. Whereas most studies attest to their therapeutic efficacy, papaverine use is declining, given its paradoxical vasospasm and adverse reactions [21]. CCBs are otherwise relatively well known and are frequently used for their positive effects [2, 3, 5, 6, 8, 10,11,12,13, 22,23,24,25,26,27], although seldom directly compared. Moreover, there was no consensus that how long the vasodilatory action of IA chemoangioplasty is maintained and how many times a day should the chemoangioplasty be performed to avoid cerebral ischemia. The present investigation was launched to examine characteristics of the three most popular CCBs used for chemoangioplasty: nicardipine, nimodipine, and verapamil. A rabbit animal model enabled side-by-side comparisons, focusing on duration of the vasodilatory action based on angiography.

Materials and Methods

Preparation of Experimental Animals

This study was approved by the Committee for Ethical Animal Experimentation at our institution and was in compliance with all recommendations made. New Zealand white rabbits were selected for testing.

A total of 36 experimental rabbits weighing 2.5–3.0 kg were commercially obtained, without regard to gender. Each fasted for 6 h prior to actual testing, injecting a combination of ketamine (35 mg/kg; Huons Co, Seoul, South Korea) and xylazine (5 mg/kg; Bayer AG, Leverkusen, Germany) intramuscularly at the thigh for anesthesia. Supplemental injections took place every 20 min during experiments, as dictated by animal conditions. Once anesthetized, depilatory cream was applied to inguinal and occipital areas for hair removal, and each animal was secured to the examination table in supine position. All rabbits were skin tested and given intramuscular injections (50 mg/day) of a first-generation cephalosporin (cefazolin; Yuhan Co, Seoul, South Korea) until 3 days after the experiment.

Puncture of Femoral Artery and Insertion of Femoral Introducer

Each anesthetized animal was kept in supine position, using 75% alcohol to sterilize the inguinal area. Femoral artery was first palpated and then exposed within the femoral neurovascular bundle by inguinal incision (~ 20 mm) made under local anesthesia (1% lidocaine; Daihan Pharm Co, Ansan, South Korea). After puncturing the femoral artery (Micropuncture Introducer Set; Cook Medical, Bloomington, IN, USA), a 4-Fr introducer (Fast-Cath; St. Jude Medical, St Paul, MN, USA) was installed via Seldinger technique, secured to femoral artery, and filled with heparin/normal saline solution for subsequent transducer passage (IntelliVue MX700; Philips, Amsterdam, the Netherlands). Blood pressure readings were thus achieved. Once cerebral angiography was completed, the introducer was positioned within subcutaneous space for later follow-up angiography, using 5-0 nylon suture (Ethicon LLC, Guaynabo, Puerto Rico) for skin closure. Three days after inducing SAH of the brain, angiography was again conducted, expedited by the introducer still in place.

SAH Induction in Animal Model

Each anesthetized animal was kept supine, with head rotated 90° to the left, for sterilization at occipital region (75% alcohol). Adjoining occipital–cervical areas were epilated and the head bent for puncture of dura mater under local anesthesia (1% lidocaine), using a 23-G spinal needle under fluoroscopy to pierce occipito-atlantal membrane. Upon first appearance of cerebrospinal fluid, a mixture of a contrast medium and a normal saline (1:1 ratio) was delivered to the cistern to identify subarachnoid space (Fig. 1a). Approximately 2 cc of cerebrospinal fluid was then removed, and 2 cc of arterial blood taken from femoral artery was introduced into subarachnoid space via spinal needle. The animal’s head was thereafter lowered about 30° and stabilized for 20 min [24, 26]. Ultimately, CT (Xper Computed Tomography) was performed to confirm SAH, using an angiography equipment (Fig. 1b).

Fig. 1
figure 1

Induction and confirmation of cerebral subarachnoid hemorrhage (SAH): a cisternography of rabbit through occipito-atlantal membrane and b sagittal reconstruction of computed tomography brain image in rabbit (black arrow denoting SAH and intraventricular hemorrhage)

Full size image

Cerebral Angiography and IA CCB Infusion

Iso-osmolar contrast (Visipaque 270; GE Healthcare, Chicago, IL, USA) was utilized for cerebral angiography (Allura Xper FD20/10; Philips, Eindhoven, the Netherlands). Both microcatheters (Excelsior SL-10; Stryker Neurovascular, Fremont, CA, USA) and microguidewires (Synchro14; Stryker Neurovascular, Fremont, CA, USA) were engaged for IA drug infusion.

Prior to induction of SAH, cerebral angiography was routinely performed to determine vertebral and basilar arterial calibers at baseline for later referencing (Fig. 2a). On Day 3 of the SAH animal model, angiography was repeated to verify vasospasm of basilar artery (Fig. 2b), defined as > 30% luminal compromise relative to baseline status. Any animals with < 30% arterial narrowing were ineligible for further participation.

Fig. 2
figure 2

Digital subtraction angiogram of vertebrobasilar artery in rabbit: a angiographic image prior to hemorrhage; b follow-up angiographic image 3 days after subarachnoid hemorrhage (black arrow at most severe point of vasospasm); and c angiographic image immediately after intra-arterial (IA) drug infusion

Full size image

Rabbits showing acceptable vasospasm were randomly assigned to IA drug infusion as follows: group C, nicardipine (0.06 mg/kg; Huons Co, Ansan, South Korea); group M, nimodipine (0.05 mg/kg; Reyon Pharm Co, Seoul, South Korea); or V group, verapamil (0.1 mg/kg; Ilsung Pharm Co, Seoul, South Korea). Each CCB capacity previously determined to have a maximum physiologic effect on arterial vasodilation was selected for this study [28,29,30,31]. The various drugs were individually mixed with normal saline (total volume of 5 cc) for IA infusion by microcatheter over a 10-min period. If blood pressure fell by 20% (relative to baseline), infusion was reduced by 50%. Infusion was stopped if blood pressure decline continued. After administration, initial drug efficacies were assessed by comparing degrees of vasospastic easing at points of severest luminal compromise (see Fig. 2b). Follow-up angiography was performed hourly for 5 h after IA infusion, measuring luminal calibers (Fig. 2c) to confirm that original vasospastic states had returned.

Measuring Arterial Diameters

Arterial diameters were determined via single-blind computer application (Vascular Quantification Software, Philips) for cerebral angiography. Images were selected in arterial phases of angiography with the least motional artifacts, equating diameter with magnitude in transverse plane. Diameters of normal arteries and arteries most affected (i.e., severest vasospasm) were each measured twice within specific time intervals without any information, recording average values. Luminal spasms were graded, based on percentage of baseline arterial diameter lost, whereas vasodilative effects after IA infusion represented percentages of vasospastic reversal.

Statistical Analysis

Continuous variables were each shown as median value (range). For group-wise analyses of vasodilative effects and blood pressure change after IA infusion, the Wilcoxon signed-rank test was used. The Kruskal–Wallis test was applied to group comparisons of vasodilatory degree, duration of action, and blood pressure change after IA infusion. If statistically significant, the Bonferroni correction for multiple comparisons was invoked. All computations were driven by standard software (SPSS v21.0; IBM Corp, Armonk, NY, USA) setting significance at p < 0.05.

Results

Data Summary

Of the 36 available rabbits, 22 served for experimentation (M group, n = 8; C group, n = 7; V group, n = 7). Fourteen rabbits were excluded owing to six deaths and eight induction failures (insufficient vasospasm). No animals required interruption of drug infusion due to blood pressure declines. Compared with basilar artery at baseline (prior to hemorrhage), the mean degree of vasoconstriction overall in the animal model of vasospasm animal was 41.6% (range 31–55%) and the median was 40%. Separately, the medians were as follows: group C, 37% (range 32–55%); group M, 41% (range 31–55%); and group V, 38% (range 37–53%) (see Table 1).

Table 1 Changes in mean arterial pressure after IA CCB infusion, shown by group
Full size table

Efficacy of IA CCBs Infusion

Once vasospasm was induced, changes in basilar artery diameters (relative to baseline values) were measured immediately and at hourly intervals after IA infusion, as summarized in Table 1. Immediate vasodilatory effects were regularly observed (group C, 27.0%; group M, 27.1%; group V, 16.1%) and proved statistically significant (group C, p = 0.017; group M, p = 0.012; group V, p = 0.018). One hour later, repeat angiograms showed that vasodilation peaked in groups C (41.3%; p = 0.018) and M (35.6%; p = 0.012) but dissipated in group V, returning to original vasospastic states. At 2 h, significant vasodilative effects were sustained in groups C and M (p = 0.046 and p = 0.051, respectively), only to disappear completely in both groups at the next interval, 3 h after IA infusions (Fig. 3).

Fig. 3
figure 3

Luminal diameters of arteries before and after IA drug infusion. \({\text{Diameter}}\,({\text{\%}}) = \frac{{{\text{Diameter}}\,{\text{of}}\,{\text{basilar}}\,{\text{artery}}\,{\text{in}}\,{\text{each}}\,{\text{time}}\,{\text{zone}}\,{\text{after}}\,{\text{vasospasm}}}}{{{\text{The}}\,{\text{normal}}\,{\text{diameter}}\,{\text{of}}\,{\text{basilar}}\,{\text{artery}}\,{\text{before }}\,{\text{vasospasm}}}}\, \times \, 100\), IA, intra-arterial; C-group, nicardipine; M-group, nimodipine; V-group, verapamil; IADI, immediately after IA drug infusion. *p < 0.05

Full size image

In group-wise comparisons of maximum vasodilatory efficacy, groups C and M were similar at every periodic testing. Although groups C and M surpassed group V in terms of potency, the difference was not statistically significant (p = 0.193). Vasodilatation was also sustained for 2 h in groups C and M, compared with < 1 h in group V, constituting a substantial departure in duration of action.

Change in Mean Arterial Pressure During and After IA Infusion

Blood pressure changes during and after IA infusion were observed in C group, the average arterial pressure falling by 10 mmHg within 1 h after IA infusion (Table 2). However, there was gradual recovery after 2 h, with normalization of blood pressure after 3 h. Almost no change in blood pressure was observed in groups M and V groups after IA infusion.

Table 2 Changes in mean arterial pressure after IA CCB infusion, shown by group
Full size table

Discussion

The therapeutic imperative for treating vasospasm after SAH is preservation of cerebral blood flow, ostensibly through acceptably sustained central perfusion pressures. However, our understanding of the mechanisms that lead to delayed cerebral ischemia in this setting is limited. Until recently, a number of factors have been implicated in vasospasm, but one very important issue reportedly is cellular influx of calcium [10, 32]. CCBs suppress this influx, which is key in preventing and treating vasospasm and explains why oral intake of nimodipine is strongly recommended [2]. Triple-H therapy (i.e., induced hypertension, hypervolemia, and hemodilution) and IA papaverine infusion have been commonly used in the past, but the benefits and side effects of doing so are still under debate [10, 16, 33,34,35]. Specifically, IA papaverine use has declined in recent years due to paradoxical vasospasm, systemic hypotension, and drug toxicity [10, 16, 35]. Still, other vasodilators have shown efficacy by this route in instances of medically intractable cerebral vasospasm [2].

Of the available vasodilators for IA infusion, CCBs are the most commonly used in treating vasospasm after SAH. Nimodipine, nicardipine, and verapamil have actually found broad clinical use in the treatment of vasospasm. Nimodipine is an L-type CCB that acts directly on smooth muscle of cerebral vessels [32,33,34,35,36]. Compared with similar drugs, nimodipine is characterized as follows: (1) it is lipophilic, readily crossing the blood–brain barrier (BBB), and is long-acting [32], (2) its vascular relaxation properties are more prominent in arterioles than in large arteries [37], (3) cerebral blood flow is enhanced, without loss of systemic blood pressure [25], and (4) there are neuroprotective effects, the mechanisms of which remain uncertain [38]. Nicardipine belongs to the same drug class as nimodipine and has likewise shown promise in treating vasospasm [13, 23, 35]. However, Huang et al. [23] found the efficacy of IA nicardipine infusion to be uncertain in terms of avoiding delayed cerebral ischemia. In our animal model, nicardipine triggered systemic hypotension during and after IA infusion, unlike the other CCBs. It may be thus inferior in maintaining cerebral perfusion. In the class of benzothiazepines, verapamil (amphiphilic nature) reacts selectively with myocardium and is useful for coronary vasospasm. The reason for its utility in the brain is that both coronary and cerebral arteries are structurally similar, and both are subject to vasospasm. On the other hand, each differs distinctly in the nature of vasospasm, which may be transient in coronary arteries but prolonged cerebral arteries, lasting for 2 weeks. Despite various reports of its usage for cerebral vasospasm [5, 11, 18, 39], verapamil has a short half-life compared with other drugs and may be better suited for treating coronary vasospasm [5].

In drugs given by IA infusion for vasospasm, the extent and duration of vasodilatory effects have been not well documented individually [24, 36, 40]. In the present study, all CCBs exerted vasodilatory effects immediately after IA infusions, but they did not differ overall in degrees of vascular relaxation (p = 0.466). Verapamil recipients showed maximum effects immediately after infusion, whereas the other agents (groups C and M) peaked 1 h after infusion. They also proved more potent than verapamil, falling short of significance (p = 0.193) (see Table 1). The vasodilatory effects observed in groups C and M were sustained for 2 h, as opposed to < 1 h in group V, corresponding with respective drug half-lives (8-9 h for nimodipine and nicardipine; 3-7 h for verapamil) [25, 34, 41]. In our study, the vasodilatory effects achieved by IA infusion using CCB were not sustained beyond 2 h in any of the CCB groups, and it may be a limitation of IA chemoangioplasty in vasospasm. Due to poor durability of the chemoangioplasty, repeated procedure may be required in severe vasospasm. Moreover, there has been no standard for how many procedures a day should be performed in order to relieve radiologic vasospasm or to stabilize clinical symptoms.

Recently, new alternatives to extend the vasodilatory effect in severe vasospasm were suggested. Hafeez et al. [7] reported that intrathecal nicardipine injections for a SAH-related cerebral vasospasm appear efficacious using systematic review. Administration of 4 mg of nicardipine every 12 h was the most commonly reported dosing regimen. Intrathecal nicardipine decreases mean flow velocities on transcranial Doppler (TCD) and reduces angiographic and clinical vasospasm. They insisted that duration of vasodilatory action in the intrathecal infusion seems to be longer than that in IA infusion. However, they did not present the concrete vasodilatory maintenance duration based on angiography or cerebral perfusion, and cerebrospinal fluid infection was demonstrated in 6%. Ortiz Torres et al. [42] and Majidi et al. [43] suggested that vasodilatory effect after IA dantrolene infusion maintained longer than 1 or 2 days, although the reports were case series.

According to related studies, various kinds of animals such as rats, rabbits, dog, and monkeys have been used as animal model of vasospasm [44]. The reason why we chose rabbits as experimental animals in our study was that the rabbit’s brain artery was large enough to measure its size and we had to experiment dozens of animals repeatedly and had advantages in terms of cost. Endo and Firat et al. also used rabbits as vasospasm model [29, 45]. Most of the rabbits used in our experiment were male, because it was not easy to find a female that weighed just as much as it could be used in the experiment. Mo et al. [46] used only male rabbit in animal vasospasm study, and Firat et al. [29] conducted the experiment regardless of the sex of the animal. It has not been well known about the equivalent dosage of each CCB drug. However, the dose of each drug used in our study was greater than the maximum vasodilatory dose in the dose–response curve of other studies [30, 31]. Therefore, we judged that the maximum relaxation effects of each CCB could be compared with each other. Nakajima et al. [47] argued that the more severe the vasospasm and the older the patient, the shorter the vasodilatory effects of CCBs will be.

As a final note, it must be emphasized that angiographic improvements in vasospasm do not necessarily mirror clinical improvements [6, 35]. Hoh et al. [6] showed that although vasospasm improved by 83% in follow-up angiograms, clinical improvement after IA infusion was much less (43%). Thus, angiographic findings in the setting of vasospasm may regularly correlate with the status of cerebral perfusion but not accurately align with clinical manifestations of cerebral ischemia. In addition, some authors insist that TCD is an inadvisable means of screening patients with potential cerebral vasospasm [48]. Brain perfusion testing and conventional angiography should accompany TCD [15, 35, 48].

This study has some limitations. Diagnosis and degree of vasospasm, as well as later improvement, were based on the most severely compromised arterial diameters. Because the vasodilatory effects of IA-infused drugs are expressed by both large arteries and small arterioles, such measurements may not reflect actual perfusion status. We must also acknowledge that the number of animals participating in this experiment was marginal for a three-group comparison. Furthermore, there are no established dosing equivalencies when treating cerebral blood vessels with these three CCBs, and studies focused on drug potencies have produced conflicting results. Hence, the IA drug dosages used herein may not be equivalent. Finally, we magnified photographs to determine basilar artery diameters, making it difficult to distinguish the boundaries of small vessels and perhaps introducing measurement errors. In aggregate, these drawbacks are considerable and should be appropriately resolved in future experimentation.

Conclusion

After IA infusion for vasospasm, all CCBs tested produced effective vasodilation based on angiography. Compared with verapamil, nimodipine and nicardipine exhibited longer durations of action, but only the nicardipine group displayed systemic hypotension during and after infusion. Unfortunately, the vasodilatory effects achieved by IA chemoangioplasty were not sustained for > 2 h in any of the treatment groups. Therefore, the new alternative should be found and verified to extend the duration of vasodilatory action in severe vasospasm, besides IA infusions of CCBs.