Rescue clazosentan for symptomatic cerebral vasospasm following subarachnoid hemorrhage due to anterior cranial fossa dural arteriovenous fistula: A case report with serial angiographic evaluation Authors: Takahiro Morita, Yu Nomura, Takao Sasaki, Tomohiro Kaji, Kohsuke Katayama, Atsushi Saito Published: 2026 ABSTRACT Background: Vasospasms associated with dural arteriovenous fistula (dAVF)-related subarachnoid hemorrhage (SAH) are rare, and their management has not been well established. Clazosentan, a selective endothelin A receptor antagonist, is typically administered soon after an SAH to prevent cerebral vasospasms. However, the optimal timing remains controversial because of the risk of fluid retention and related complications. Here, we report the case of dAVF-related SAH in which clazosentan was introduced as rescue therapy to treat cerebral vasospasm.Case Description: A 61-year-old man presented with SAH caused by the rupture of a Borden type III dAVF in the anterior cranial fossa. Following surgical occlusion, intravenous fasudil hydrochloride was administered for vasospasm prophylaxis. However, severe bilateral anterior and middle cerebral artery vasospasms developed on day 10, accompanied by progressive somnolence. Intra-arterial fasudil hydrochloride administration provided limited and transient improvements. Clazosentan was subsequently introduced as rescue therapy and was continuously administered for 4 days. Angiography revealed marked vasodilatation, particularly in the distal vessels, following clazosentan administration. Although fluid retention was observed, the patient was administered diuretics without clinical respiratory deterioration. No cerebral infarction occurred, and the patient was transferred with a favorable outcome (modified Rankin Scale score of 1).Conclusion: This case suggests that rescue administration of clazosentan may have contributed to improvement as part of multimodal vasospasm management following dAVF-related SAH, with vasodilatory effects comparable to those observed in aneurysmal SAH cases. Furthermore, short-term rescue administration of clazosentan may be tolerable in patients at high risk of fluid retention-related complications when careful fluid management and appropriate diuretic use are implemented. INTRODUCTION Clazosentan is a selective endothelin A receptor antagonist that was developed to prevent cerebral vasospasm following subarachnoid hemorrhage (SAH) due to ruptured cerebral aneurysms.[25] The CONSCIOUS-3 trial[10] and a phase 3 clinical trial conducted in Japan[5] demonstrated its significant efficacy in suppressing cerebral vasospasm. It was approved and launched for clinical use in Japan in April 2022. Based on the design of clinical trials,[5] clazosentan is recommended to be initiated within 48 h of SAH onset. Accordingly, in many cases, it is promptly initiated after definitive aneurysm treatment and continued throughout the vasospasm period.[14,15,18,21]. However, clazosentan has the notable adverse effect of promoting fluid retention, which may lead to pleural effusion, pulmonary edema, and other respiratory complications, occasionally necessitating discontinuation.[5,8,10,14-16,18,21] Previous reports have suggested that early administration may paradoxically increase the incidence of pulmonary edema[16] or treatment discontinuation,[8] indicating that early initiation is not always beneficial. During the clinical trial phase, a case series (REVERSE study) revealed the effectiveness of clazosentan as rescue therapy for cerebral vasospasm.[7] In patients at a high risk of fluid retention, delaying its initiation may be considered; however, there are very few studies evaluating the outcomes when clazosentan is not administered early, and its efficacy and safety in such settings remain unclear. Herein, we report a case of SAH caused by the rupture of a dural arteriovenous fistula (dAVF), in which symptomatic cerebral vasospasm developed despite the administration of prophylactic fasudil hydrochloride, and clazosentan was introduced as rescue therapy. We focus on the patient’s fluid dynamics and temporal changes in cerebral vasospasm. This case report was prepared according to the CARE guidelines.[20] CASE DESCRIPTION A 61-year-old man was transported to our emergency department because of the sudden onset of headache and vomiting. The patient’s past or family medical history was unremarkable. On admission, the patient complained of a moderate-to-severe headache. His level of consciousness was assessed using the Japan Coma Scale (score 1) and Glasgow Coma Scale (score 15), with no evident focal neurological deficits. Head computed tomography revealed a diffuse SAH and an intracerebral hematoma in the left frontal lobe [Figure 1A]. Cerebral angiography revealed a Borden type III dAVF[3] located in the anterior cranial fossa, bilaterally supplied by the anterior ethmoidal arteries, with cortical venous reflux. Imaging findings. (A) Computed tomography of the head on admission. (B) Lateral view of catheter angiography of the right internal carotid artery before the operation. The arrow indicates the shunt, which is the site of bleeding. (C) Lateral view of head radiography was formed after the operation. The arrow indicates the clip occluding the draining vein of the dural arteriovenous fistula. On the same day, the draining vein was surgically occluded through craniotomy [Figures 1B, C and 2 A]]. Bilateral anteroposterior view of catheter angiography. (A) Performed on day 0 (before surgery). (B) Performed on day 10 (before the administration of intra-arterial fasudil hydrochloride). The image shows severe diffuse vasospasm. (C) Performed on day 10 (immediately after the administration of intra-arterial fasudil hydrochloride). This image shows improvement in vasospasm. (D) Performed on day 11 (before the administration of intra-arterial fasudil hydrochloride). Vasospasm recurred, and we subsequently performed second intra-arterial infusion of fasudil hydrochloride and intravenous administration of clazosentan. (E) Performed on day 12 (before the administration of intra-arterial fasudil hydrochloride). Compared with the previous day, marked vasodilation is observed, particularly in the peripheral arteries. Immediately after this angiography, third intra-arterial infusion of fasudil hydrochloride was performed. (F) Performed on day 17 (3 days after discontinuation of vasodilator administration). Although the vasospasm phase ended, the vessel diameter remained narrower than that observed during clazosentan administration. Postoperatively, no deterioration of consciousness or new neurological deficits were observed. Prophylaxis for cerebral vasospasm was initiated on postoperative day 1 using intravenous fasudil hydrochloride (30 mg 3 times daily). Fluid administration was managed between 3000 and 7000 mL/d based on three daily fluid balance assessments and body weight measurements. The patient initially remained stable; however, progressive somnolence developed on postoperative day 6. No focal neurological symptoms were observed, and various examinations did not identify any underlying causes such as hydrocephalus, epilepsy, or metabolic disorders. Furthermore, no medications affecting consciousness were used. On day 10, follow-up angiography [Figure 2B] revealed severe bilateral anterior and middle cerebral artery vasospasms. Intra-arterial infusion of fasudil hydrochloride (15 mg/vessel) through proximal M1 segments resulted in slight vasodilatation [Figure 2C]. Following intra-arterial fasudil administration, somnolence improved slightly for a few hours, but the unstable trend persisted. The following day, mild abnormalities in language comprehension and speech were observed. Repeat angiography on the following day (day 11) showed no significant improvement [Figure 2D], and a second intra-arterial fasudil hydrochloride infusion was performed. Clazosentan was administered intravenously at an approved dose of 10 mg/h. On day 12 (the day after clazosentan initiation), symptoms such as somnolence and speech difficulties remained largely unchanged, although angiography revealed marked improvement in vasospasm [Figure 2E]. Owing to persistently unstable deficits, an additional dose of intra-arterial fasudil hydrochloride was administered. On the following day, speech impairment had nearly normalized, although somnolence remained unstable; overall, the patient showed an improving trend. Fasudil hydrochloride and clazosentan were discontinued on day 14. Somnolence did not worsen after discontinuing the vasodilator, and by day 18, it had improved to a level that appeared nearly normal. Clazosentan was administered for 4 days. The urine output decreased markedly after initiation, and the maximum fluid balance reached +2034 mL/d. For a significantly positive fluid balance, furosemide (20 mg 3 times daily) was administered for 1 day. The mean fluid balance was 168.7 mL/d during fasudil hydrochloride monotherapy and 888.4 mL/d during combined therapy with clazosentan. On day 17, angiography revealed vasospasm recurrence [Figure 2F]; however, as the patient remained asymptomatic, no further intervention was performed. A chest radiography performed three days after clazosentan administration revealed reduced hilar transparency, however, no respiratory symptoms were observed and the finding improved after discontinuation of clazosentan [Figure 3]. Ultimately, no cerebral infarction developed [Figure 4], and the patient was transferred on day 26 with a modified Rankin Scale score of 1. Chest radiography. (A) Performed on day 0 (on admission). (B) Performed on day 11 (immediately before clazosentan administration). (C) Performed on day 15 (immediately before discontinuation of clazosentan administration). The circle indicates reduced hilar transparency, suggesting pulmonary edema. (D) Performed on day 17. The decrease in transparency in the hilar region improved. Magnetic resonance imaging before discharge. (A) T2-weighted image of fluid-attenuated inversion recovery. (B) Diffusion-weighted image. (C) Magnetic resonance angiography. These images show no abnormal findings, except for the initial brain injury. The trends in fluid management, medication use, body weight, and fluid balance during the vasospasm period are shown in Figure 5. Clinical course. Surgery was performed on day 0, and intravenous fasudil hydrochloride was administered from days 1 to 14. To enhance treatment for symptomatic vasospasm, fasudil hydrochloride was administered intra-arterially from days 10 to 12, and clazosentan was administered intravenously from days 11 to 14. Line graphs show daily fluid balance and changes in body weight. Before clazosentan administration, the patient received a relatively large volume of fluid; however, minimal fluid retention was observed. However, significant fluid retention was observed after clazosentan administration. Therefore, on the 2nd day after administration, furosemide was administered to induce diuresis and regulate fluid balance. GCS: Glasgow Coma Scale. Trend of vascular diameter In this case, cerebral angiography was performed 5 times (on days 0, 10, 11, 12, and 17) to evaluate vasospasm. A 4-Fr catheter was placed in the cervical internal carotid artery (ICA), and 5–7 mL of 100% Omnipaque 300 (GE Healthcare Pharma, Tokyo, Japan) was injected manually. The changes in vessel diameter were quantitatively evaluated. The measurements were performed by one physician who was not directly involved in the treatment itself but understood its general outline. Measurements were taken at the proximal (ICA terminal portion, M1, A1) and distal arteries (M2, A3) immediately after angiography. The diameter of each artery at each time point was calculated as a ratio relative to the ICA petrous portion. Furthermore, using day 0 as the baseline, the ratio of the change in vascular diameter before intra-arterial fasudil hydrochloride administration at each time point was calculated. The measurement sites are shown in Figure 6A, the trend of change in the distal arteries in Figure 6B, and the trend of change in the proximal arteries in Figure 6C. Trends in arterial diameter on serial angiography. (A) Overview of the arterial diameter measurement points. (B) Trends in the diameters of the distal arteries (M2, A3). Cerebral vasospasm was observed on day 10, and clazosentan was initiated on day 11. Marked vasodilation was observed on day 12 during clazosentan administration, but the vessel diameter narrowed again by day 17 after clazosentan administration was completed. (C) Trends in the diameters of the proximal arteries (internal carotid artery, M1, A1). Similar to the distal vessels, cerebral vasospasm was observed in the M1 and A1 arteries on day 10. However, unlike in the distal vessels, vasodilation following clazosentan administration was not pronounced. ICA: Internal carotid artery. Despite prophylactic intravenous fasudil hydrochloride administration, vasospasms distal to A1 and M1 were observed on day 10 [Figures 2B, 6B and C]. The intra-arterial administration of fasudil hydrochloride resulted in transient global vasodilatation [Figure 2C]. However, on day 11, only a slight improvement was noted in the right A1 and bilateral M1, with minimal overall change [Figures 2D and 6C]. Therefore, clazosentan therapy was initiated. On day 12, marked dilation was observed in the bilateral M2 segments, whereas the diameters of the ICA and M1 remained largely unchanged. The anterior cerebral artery diameter increased by approximately 10–20% compared with the measurement before clazosentan administration [Figures 2E, 6B and C]. All vasodilators were discontinued on day 14. On day 17, follow-up angiography showed that the M2 diameters had returned to preclazosentan levels, and the diameters of other vessels (ICA, M1, and anterior cerebral artery) were similar to those on day 11 [Figures 2F, 6B and C]. DISCUSSION We describe the clinical course of a patient with SAH due to rupture of the anterior cranial fossa dAVF, in whom prophylactic intravenous fasudil hydrochloride was insufficient to prevent vasospasm; however, the combination of intra-arterial fasudil hydrochloride and intravenous clazosentan successfully prevented ischemic complications. Although SAH due to dAVF is common, reports of associated vasospasms are scarce.[4,6] Therefore, the responsiveness, efficacy, and appropriate use of vasodilators, including clazosentan, remain unclear. Clazosentan is typically administered continuously for 14 days, starting in the early phase of SAH onset. In this case, rescue administration after vasospasm may have contributed to improved vessel diameter and infarction prevention. Serial angiography indicated characteristic vasodilatation, particularly in distal vessels, during clazosentan administration. These findings may provide insights into the treatment of vasospasm following dAVF-related SAH and the optimal timing of clazosentan administration. Clinical essence when using clazosentan and adaptation to this case In Japan, clazosentan is generally administered from the early postoperative phase until approximately 14 days after SAH onset, when vasospasm is expected to resolve; most postmarketing reports have adhered to this recommendation.[8,9,14-16,18,21] However, a major limitation of clazosentan is its potent fluid-retaining effect, which may necessitate discontinuation owing to respiratory complications such as pleural effusion and pulmonary edema.[5,8-10,14-16,18,21] To avoid these complications, a dry-side management approach has been proposed for patients with aneurysmal SAH receiving clazosentan. This approach differs from conventional management by limiting fluid administration to 1.0–1.5 mL/kg/h[5] and maintaining a fluid balance of <500 mL/d[21] to avoid complications associated with fluid retention. Furthermore, if fluid retention is observed, management strategies such as actively using diuretics to promote fluid excretion are recommended.[1] However, only two cases of dAVF-related SAH have been reported in the past, and both were treated with intra-arterial verapamil administration[6] or balloon angioplasty[4,6] after the onset of symptomatic vasospasm. Therefore, it remains unclear whether these recommendations also apply to dAVF-related SAH, and the extent of their effectiveness and side effects in this context has not been established. For this reason, these drugs were administered in the same manner as in aneurysmal SAH in the present case. Regarding drug selection, fasudil hydrochloride was selected as the first-line treatment from a safety perspective because, although it is less effective in preventing vasospasm, it is associated with fewer complications.[1,13,18,21] Clazosentan was then used as a rescue drug when symptomatic vasospasm occurred. As shown in Figure 5, when clazosentan was introduced, the daily fluid intake was reduced from approximately 2000 mL to approximately 4000 mL/d. This volume exceeded the recommended fluid intake described above. The intention was to avoid abrupt fluid reduction during active vasospasm. In addition, because a negative fluid balance had persisted before clazosentan initiation and dehydration could not be ruled out, a slightly higher infusion volume was administered. However, this over-infusion may have resulted in excessive fluid retention, which may have contributed to the tendency toward pulmonary edema observed on chest radiography. Nevertheless, as observed in aneurysmal SAH, the patient responded well to diuretics. After clazosentan was discontinued on day 14, fluid retention resolved, a negative fluid balance was rapidly achieved, and treatment was completed without symptoms. In addition, vasospasm recurred after clazosentan discontinuation in this case. Fortunately, the vasospasm resolved spontaneously and did not become symptomatic. However, when clazosentan is discontinued, the loss of its vasodilatory effect, together with its fluid-retaining effect, may lead to recurrent cerebral vasospasm and dehydration, potentially resulting in cerebral infarction. Ultimately, the response to clazosentan observed in this case, including its efficacy and side effects, was almost identical to the responses conventionally observed in aneurysmal SAH. Nonetheless, this case was a special case in which clazosentan was used as rescue treatment for dAVF-related SAH; therefore, the findings cannot be readily generalized to other cases. Clazosentan as rescue therapy As mentioned above, clazosentan carries a risk of fluid retention-related complications, which is a major factor limiting its use. Reported risk factors for discontinuation include advanced age,[14] cardiothoracic ratio enlargement on admission,[16,24] and anemia.[24] In addition, management-related factors have been suggested, including the association between early postoperative initiation and increased complications. Nishii et al. reported that the time from SAH onset to clazosentan initiation was shorter in the discontinuation group than in the completion group (median, 49 vs. 57 h).[16] Itazu et al. demonstrated that promoting fluid excretion before clazosentan administration significantly reduced pulmonary edema on chest radiography, and the median time to initiation was 0.8 days shorter in the fluid-managed group than in the group without fluid management (median, 3.1 vs. 3.9 days).[8] Although the precise relationship between these patient and management factors and fluid retention remains unclear, the timing of clazosentan initiation is a modifiable clinical factor. In patients with a substantial risk of complications, an intentional delay in administration may be a reasonable option. A key question is when to initiate clazosentan administration if it is used in the delayed phase. Respiratory complications and fluid retention have been reported to occur at a median of 3 days after clazosentan initiation (95% confidence interval: 2–3 days),[2] within 3 days,[12] or at approximately 4 days (interquartile range: 2.0–5.0 days).[23] In the REVERSE study, clazosentan was administered for a mean of 68.1 h (range: 3.5–180.5 h), and pulmonary complications occurred in 21.4% of patients, which is comparable to previous reports.[7] In this case, clazosentan was administered for 96 h (days 11–14), and fluid retention was observed immediately after initiation but was successfully managed with diuretics, without clinical deterioration. These findings suggest that when clazosentan is used as rescue therapy, limiting its administration to approximately 4 days under careful fluid management and appropriate diuretic use may help avoid clinically significant fluid retention. However, early discontinuation of clazosentan may worsen vasospasm. Previous reports have described clinical deterioration following premature discontinuation.[1,13] At the molecular level, upregulation of endothelin A receptors was observed following clazosentan administration in rat models, and discontinuation of clazosentan on day 7 resulted in exacerbation of vasospasm.[26] These findings suggest that discontinuation should ideally occur after vasospasm improvement, typically after day 14. Taken together, initiating clazosentan after 11–12 days and discontinuing it within 3–4 days – before significant fluid retention becomes clinically apparent – would result in treatment completion after day 14, potentially reducing fluid-related complications and the risk of rebound vasospasm. Nevertheless, multiple clinical trials[5,10] and postmarketing case series[1,2,8,9,12-16,18,21,23,24] have demonstrated the efficacy of early administration. In addition, prolonged duration of clazosentan therapy has been associated with improved outcomes,[9] and endothelin-1 levels increase early after SAH and correlate with vasospasms.[11] Therefore, early treatment initiation is the standard approach. Recent improvements in clinical management, including the aggressive use of diuretics[1] and optimization of fluid administration based on postmarketing experience,[8,22] have reduced the incidence of respiratory complications. Thus, clazosentan can now be used relatively safely in most patients, and rescue use should be considered primarily in patients at a high risk of fluid retention. Vasodilatation Another noteworthy aspect of this report is that cerebral angiography was performed relatively frequently from the onset to the end of the vasospasm period, allowing for temporal observation of vascular changes. To the best of our knowledge, this case represents the first attempt to serially evaluate vasospasms and the effects of vasodilators in SAH caused by dAVF. Although this evaluation was limited to a single case, imaging changes following the administration of intravenous fasudil hydrochloride, intra-arterial fasudil hydrochloride, and clazosentan were assessed. The observed vascular reactivity was comparable to that observed in aneurysmal SAH cases. Regarding vasodilatation following clazosentan administration, Ogawa et al. reported a detailed comparison of intravenous fasudil hydrochloride and clazosentan in clinical cases. While no apparent vasodilatation was observed on imaging in the intravenous fasudil hydrochloride group, the clazosentan group demonstrated no significant dilation in proximal vessels such as the ICA and M1 compared with baseline; however, marked dilatation of approximately 50% of the baseline was observed in distal vessels such as M2 and M3.[17] Distal vasodilatation following clazosentan administration was also demonstrated in the REVERSE study[7] and in a clinical trial by Endo et al.[5] In this case, marked dilation of the M2 segment was observed following clazosentan administration and subsequent narrowing occurred after discontinuation. These findings are consistent with those of previous studies and suggest that the observed vascular changes might be influenced by clazosentan. Limitations This study has some limitations. First, this was a case of rescue clazosentan administration for vasospasm after SAH caused by a dAVF. Therefore, the findings cannot be directly generalized to typical aneurysmal SAH. In addition, because this is a single-case report, no conclusions regarding generalizability can be drawn. However, similarities in pharmacological responses to those reported for aneurysmal SAH suggest that certain findings may be applicable. Second, cerebral angiography was performed as part of routine clinical practice, and the catheter positioning was not standardized. Because the contrast medium was injected manually, differences in contrast enhancement could not be completely excluded. Furthermore, measurements were performed manually, which may have introduced an observer bias. These factors may have influenced the assessment of vasodilatation and should be considered when interpreting imaging findings. Future studies should incorporate automated measurement systems, standardization of catheter position and diameter, and the use of an automatic injector to improve accuracy. Third, as vasodilators other than clazosentan were administered, it cannot be definitively concluded that the observed changes in vessel diameter were solely attributable to clazosentan. However, a previous report by Ogawa et al. revealed no significant imaging changes with intravenous fasudil hydrochloride alone,[17] suggesting that its effect in this case was minimal. In addition, intra-arterial fasudil hydrochloride generally produces only transient effects,[19] and its influence is likely to be diminished at the time of subsequent angiographic evaluation. Based on these considerations, clazosentan may have been the main contributor to the vasodilatation observed in this case. CONCLUSION Here, we report a case in which clazosentan was administered as rescue therapy for cerebral vasospasm that could not be prevented with intravenous fasudil hydrochloride administration following SAH caused by dAVF rupture. The observed vasodilatory effects were comparable to those observed in aneurysmal SAH cases. Although this is a rare case and further accumulation of cases is necessary, our findings suggest that both fasudil hydrochloride and clazosentan may contribute to the treatment of vasospasm associated with dAVF rupture. Furthermore, short-term rescue administration of clazosentan may be tolerable in patients at high risk of fluid retention-related complications when careful fluid management and appropriate diuretic use are implemented.