Reversal of Severe Cerebral Vasospasm in Three Patients after Aneurysmal Subarachnoid Hemorrhage: Initial Observations Regarding the Use of Intraventricular Sodium Nitroprusside in Humans Jeffrey E. Thomas, M.D., Robert H. Rosenwasser, M.D. Division of Cerebrovascular Surgery and Interventional Neuroradiology, Department of Neurological Surgery, Thomas )efferson University, Philadelphia, Pennsylvania C O N C L U S IO N : These preliminary observations suggest that sodium nitroprusside delivered by an intrathecal route of administration may be a useful treatment for severe vasospasm complicating SAH in humans. (Neurosurgery 44:48-58, 1999) Key words: Angiography, Human, Intrathecal, Nitric oxide, Treatment, Vasospasm reliably effective treatment for the chronic delayed type of cerebral vasoconstriction (CDCV) that occurs after aneurysmal subarachnoid hemorrhage (SAH) re­ mains elusive. The mainstay of treatment for this complica­ tion, now the most important cause of mortality and neuro­ logical morbidity after aneurysmal SAH, is hypertensive, A 48 hypervolemic, hemodilution (HHH) therapy. Because sev ere cases ot C. IK \ may be refractory to HHH therapy and be­ cause some patients may not tolerate HHH therapy for med­ ical i easons, an alternative treatment for C. DCV is desirable. Tluee cases ot the prompt and substantial reversal ot medi­ cally reftactorv C IX \ using an intrathecally administered Neurosurgery, Vol. 44, No. 1, l.inu.m 1440 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 OBJECTIVE AND IM PORTANCE: The chronic delayed type of cerebral vasoconstriction that occurs after aneurysmal subarachnoid hemorrhage (SAH) is now the most important cause of mortality and neurological morbidity for patients who initially survive the rupture of cerebral aneurysms. Although intravascular volume expansion and cardiac performance enhancement have had a profound impact on the treatment of the chronic delayed type of cerebral vasoconstriction, this form of treatment is not tolerated by all patients and is unhelpful in some. A more specific and more reliable treatment for this condition has not been previously reported. Previous work in an animal model has demonstrated the efficacy of nitric oxide-donating compounds in reversing severe cerebral vasoconstriction when delivered to the adventitial side of the blood vessel. A clinical study was initiated after receiving approval from the United States Food and Drug Administration and the institutional review board. CLIN IC A L PRESENTATION: Three cases of prompt and substantial reversal of medically refractory vasospasm occurring after aneurysmal SAH in humans using an intrathecally administered nitric oxide donor and clinical, angiographic, and ultrasonographic documentation are presented. All patients developed severe vasospasm refractory to medical treatment 5 to 12 days after sustaining aneurysmal SAH. All patients manifested stupor of new onset (Glasgow Coma Scale score of 7) and new focal neurological deficit (hemiplegia). The condition was angiographically demonstrated in all cases. INTERVENTION: The patients were treated with intrathecally administered sodium nitroprusside, which caused the reversal of vasospasm, which was documented by angiography and transcranial Doppler ultrasonography up to 54 hours later and also by dramatic clinical improvement. Complications related to intracranial pressure elevation, changes in vital signs, and hemodynamic parameters were not observed during or after the procedures. Radiographic evidence of the reversal of vasospasm and brain ischemia was obtained. The clinical outcomes of the treated patients were excellent. All patients presented with hemiplegia and stupor that resolved or markedly improved (within several days, two patients; within 12 hours, one patient). All three patients were discharged and were living at home at the time of manuscript submission. Reversal of Human Vasospasm nitric oxide (NO) donor and angiographic and ultrasono­ graphic documentation are presented. The effects of this treat­ ment were still angiographically and clinically evident more than 2 days after treatment. Changes in intracranial pressure (ICP), vital signs, and hem odynam ic parameters were mini­ mal or nonexistent. Radiographic evidence of the reversal of vasospasm and brain ischemia, by angiography, transcranial Doppler (TCD) ultrasonography, and computed tomography (CT), was obtained. These preliminary observations suggest that NO donors delivered by an intrathecal or adventitial route of administration may be a useful treatment for severe CDCV in humans. PATIENTS AN D M ETH O D S Approval and informed consent Inclusion and exclusion criteria Patients eligible for treatment were patients with sym ptom ­ atic aneurysmal SAH-induced CDCV refractory to HHH ther­ apy for 1 hour. TCD velocity m easurements (carotid index) were used to diagnose CD CV in the context of the clinical examination, and asymptomatic patients were not considered for treatment. For patients for whom the clinical examination was not available (e.g., those patients under pentobarbitalinduced electroencephalographic burst suppression for brain protection), a higher reliance was placed on TCD m easure­ ments, although other criteria were also used (increase in ICP or volume of cerebrospinal fluid [CSF] drainage for a given parameter per hour). In addition to asymptomatic patients, patients with obvious infarction or intraparenchymal hem a­ toma revealed by CT, patients with ICP not reducible below 20 mm Hg, and patients younger than 18 years were excluded from participation in this study. TCD measurements TCD velocity measurements were obtained by a single operator who was specifically trained in the technique. The values were normalized for varying hem odynam ic conditions by using the carotid index (mean middle cerebral artery [MCA] velocity/mean cervical internal carotid artery [ICA] velocity). Cerebral angiography was performed in the standard m an­ ner (transfemoral selective) before, during, and after treatment with sodium nitroprusside (SNP), using a high-resolution biplane digital subtraction technique. In two of the patients (Patients 1 and 2), additional post-treatment angiography was performed between 12 and 54 hours after treatment. Cerebral circulation times, defined as the interval in seconds between the first observed filling of the ipsilateral carotid siphon and the first observed filling of either transverse sinus on the anteroposterior angiogram, were estimated to the nearest halfsecond and determined by a single observer. All cerebral circu­ lation times were measured at constant mean arterial blood pressure (MABP). Blood vessel diameters were measured digi­ tally, using the computer paradigm of the Phillips Integris 3000 Biplane Digital Subtraction Angiography Unit (Gensia Pharma­ ceuticals, Irvine, CA), by a single observer using constant ana­ tomic references from the angiogram, such as bony landmarks or origins of readily observed branch vessels. Dosage and administration All SNP was administered via ventriculostomy, which was performed in the standard manner (cannulation of the lateral ventricle via the frontal lobe). SNP was delivered as a solution in the patient's own CSF, at concentrations between 1 and 4 mg/ml. The initial dosage range of the intrathecally adm in­ istered SNP (ITSNP) was based on previous experimental observations (10) and the clinical response as measured by intraprocedural monitoring and angiography (see text). Dos­ age was limited to 30 mg per procedure in accordance with the structure of a dose escalation study in progress. C LIN IC A L PRESENTATION Patient 1 A 42-year-old woman sustained a Hunt and Hess (2) Grade III SAH that resulted from the rupture of a left-sided ICA bifurcation aneu­ rysm. She underwent an uncomplicated craniotomy and clip ligation of the aneurysm within 24 hours of SAH at an outside institution. On the 5th day after SAH, she was noted to have a deterioration in mental status and right-sided weakness (Glasgow Coma Scale [GCS] score of 7) (8), which progressed to hemiplegia before she was transferred to our institution. The patient's condition failed to improve after trans­ fer, despite the initiation of HHH therapy and cardiac performance enhancement with optimization of pulmonary capillary wedge pres­ sure, MABP, and cardiac index as determined by Swan-Ganz cathe­ terization. CT demonstrated low attenuation areas of left and right cerebral cortex in anterior cerebral (right) and anterior and middle carotid (left) arterial territory distributions, consistent with ischemia. Cerebral angiography performed 1 hour after the initiation of aggres­ sive HHH therapy disclosed severe vasoconstriction of the anterior circulation vessels bilaterally, including the proximal and distal an­ terior cerebral artery (ACA) and posterior communicating artery (PComA) on the right and the proximal ACA and proximal and distal Neurosurgery, Vol. 44, No. 1, January 1999 49 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 Approval of this investigation from the institutional review board (Thomas Jefferson University, Jefferson Medical Col­ lege, 1RB Control No. 97.9035) and the Food and Drug A d­ ministration under the Investigational New Drug Treatment protocol (FDA IND No. 52307, 1997) was obtained by the principal investigator (JET), and informed consent was ob­ tained from each patient's immediate family. Informed con­ sent was obtained in advance (at admission) w henever pos­ sible for every patient diagnosed with aneurysmal SAFI. Angiography 50 Thomas and Rosenwasser TABLE 1. Measurement of Mean Arterial Blood Pressure and Intracranial Pressure for Patient 1 during Treatment with Intrathecally Administered Sodium Nitroprusside'* T im e of close (m in) M A B P (mm Hg) IC P (mm Hg) 0 125 7 30 114 7 60 122 7 90 110 7 120 110 7 140 123 9 a MABP, mean arterial blood pressure; ICP, intracranial blood pres­ sure. Measurements of MABP and ICP for Patient 1, obtained contin­ uously and recorded at times of administration of intraventricular sodium nitroprusside (see text), are shown. Entire treatment duration was 2 hours 20 minutes. F IG U R E 1. Patient 1. A , anteroposterior projection angio­ gram of the right cerebral ICA, obtained before treatment. Note that the A1 segment of the A C A is severely constricted, as is the PComA (7); distal PCom artery (2). B, lateral projec­ tion angiogram of the right cerebral ICA, obtained before angiogram of the left cerebral ICA, obtained before treat­ ment. Note that the A1 and A3 segments of the A C A and the M1 (7) and M2 segments of the M CA (2) are severely con­ stricted. D, lateral projection angiogram of the left cerebral ICA, obtained before treatment. The degree of vasospasm in the A3 (7) segment of the A C A (2) is apparent. MCA on the left {Fig. I, A -D ). Left- and right-sided cerebral circula­ tion times were 7 and 7.5 seconds, respectively. Control angiograms from the left ICA were obtained at 15-minute intervals, while a solution of SNP (1.0 mg/ml) was delivered via ventriculostomy into the right lateral ventricle in 5.0-ml aliquots. Immediately before in­ jection, a volume of CSF equal to or slightly greater than the injected volume was withdrawn from the ventriculostomy. ICP monitoring and continuous hemodynamic monitoring (MABP, Swan-Ganz intra­ cardiac catheter) and neurophysiological monitoring (electroencepha­ lography, somatosensory evoked potentials [SSEPs]) were performed simultaneously. A total dose of 30 mg of ITSNP was administered. Results A progressive and marked increase in the caliber of multiple pre­ viously constricted vessels was observed, beginning with the 30minute angiography. There was no significant change in MABP, ICP, or neurophysiological monitoring parameters from the baseline dur­ ing or after treatment (Tabic 1). TCD velocity measurements obtained after the procedure were within normal limits, with the exception of those of the left MCA, which were initially normal postprocedure and became moderately elevated (mean velocity, < 200 cm/s) 20 hours after the procedure. TCD measurements remained unchanged from these values 26 hours after the procedure (Table 2). A control cerebral angiogram obtained 24 hours after treatment demonstrated persis­ tent substantial dilation of the previously spastic vessels on the right (A l, PComA) and significant further increase in the dilation of those on the left (M l, M2, A l, A2), according to digital measurements (Fig. 2, A -D ); enhanced distal circulation was present (Fig. 2, E and F). The resulting percentage change from the baseline calibers of these Patient 2 A 50-year-old man presented with a clinical (2) Grade III and radiographic (1) Grade III SAH that resulted from a ruptured right­ sided MCA aneurysm. His medical history was significant for poorly controlled arterial hypertension and a 50 pack-year history of ciga­ rette smoking. The patient underwent an uncom plicated right pterional craniotomy and clip ligation of the aneurysm on hospital Day 2 (post-SAH Day 1). The surgery was remarkable for an extremely swollen, injected-appearing brain with copious SAH despite effective osmotic diuresis and ventricular drainage. The temporary occlusion time of the inferior division of the right MCA was 5 minutes under etomidate burst suppression anesthesia. No changes in neurophysi­ ological monitoring parameters (electroencephalography, SSEPs) were observed during the operation. The bone plate was not replaced at the termination of the operation in anticipation of cerebral swelling. The patient awakened uneventfully, however, and was extubated the day after surgery. He had no demonstrable focal neurological deficit for 5 days, and his GCS score was consistently 14 to 15. On postop­ erative Day 5 (SAH Day 6), the patient manifested an aggravated increase in IC D velocities on the left ( Table 4) simultaneously with a profound depression in mental status (GCS score of 7) and failure to move the light side of his body. He was emergently intubated. CT of the biain demonstiated right-sided cerebral swelling, moderate venti iculomegaly relative to the computed tomographic scan obtained at admission, and persistent residual SAH. No foca 1 injury to the left cerebral hemisphere was radiographieally discernible. A ventriculos- Neurosurgery, Vol. 44, No. /, /tmuarv 1444 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 treatment. Note that the PComA (7) is severely constricted; the A2 artery segment (2). C, anteroposterior projection vessels ranged from 90 to 115% (mean, 104%) for the right and from 26 to 132% (mean, 98%) for the left (Table 3). The cerebral circulation time was reduced to 5 seconds for both hemispheres. CT of the brain performed 48 hours after treatment showed im provem ent in the low attenuation regions. Cerebral angiography was repeated at 54 hours after treatment because of increased TCD velocities measured in the left M l. The dilated appearance of the right-sided vessels and the left-sided A l and A2 persisted, whereas a small degree of focal narrowing was observed in the left M l and proximal M2. The left Ml underwent angioplasty 54 hours after the SNP treatment because the constricted segment was unique and proximal, and its caliber was restored to that produced by SNP treatment. The patient subse­ quently achieved a dramatic clinical recovery. She emerged from coma and within several days began to demonstrate useful function of the previously plegic right side (the leg more than the arm). Serum cyanate levels drawn on the evening after treatment were undetect­ able. She was discharged to a rehabilitation facility within 2 weeks in stable condition, with a GCS score of 15. At post-treatment Day 129, she was able to walk with the aid of a walker at home, able to feed herself, and able to converse, with a moderate pure motor dysphasia. Her right upper extremity remained densely paretic, although she was able to move it willfully. I Reversal of Human Vasospasm si TABLE 2. Mean Transcranial Doppler Velocities and Calculated Velocity Indices of Patient 1 after Treatment with Intraventricular Sodium Nitroprusside*1 Days after Treatment Hours after Treatment Vessel/Index 1 4 20 26 46 51 85 47 93 54 113 107 100 73 73 57 44 29 3 .9 0 28 3.82 RM CA 71 73 RA C A 162 50 RICA R C IX 6 31 2 .2 9 36 2.03 44 LM CA 73 42 76 170 169 230 230 59 82 38 1.92 31 2 .4 5 54 33 5.15 57 59 32 128 41 129 42 5.2 8 5.61 100 5 .4 8 54 LA C A LICA LC IX 6 ACom A 45 1.93 2.11 42 70.5 16 17 18 136 82 97 97 112 93 128 56 50 1.64 47 2 .0 6 78 NM NM 188 123 143 132 42 100 69 40 90 46 4.53 138 4 .0 9 97 38 3 .2 4 40 3 .5 8 38 3 .4 7 90 100 NM 100 75 95 100 71 135 140 143 62 1.61 52 73 1.85 143 67 128 54 145 82 162 85 181 92 57 35 4 .1 4 38 4 .2 6 107 162 2.2 5 53 1.92 2 .0 9 120 2 .4 3 a RMCA, right middle cerebral artery; RACA, right anterior cerebral artery; RICA, right internal carotid artery; RCIX, right carotid index; LMCA, left middle cerebral artery; LACA, left anterior cerebral artery; LICA, left internal carotid artery; LCIX, left carotid index; ACoA, anterior communicating artery; NM, not measured. Mean transcranial Doppler velocities of the cerebral arteries of Patient 1, measured after treatment with intraventricular sodium nitroprusside (in cm/s), are shown. No measurements were obtained before treatment, because the patient was received in transfer in extremis. Carotid index was calculated as follows: mean MCA velocity/mean cervical ICA velocity. Results Improvement in the previously prolonged latencies of the left cortical SSEPs was observed within 23 minutes of the administration of SNP, dropping from 23.9 to 20.0 milliseconds for the upper extrem ­ ity and from 51.5 to 47.7 milliseconds for the lower extremity (Fig. 4). No significant change in the caliber of the larger conductance vessels (left Al and M l) was observed at first, and concurrent cerebral angioplasty was attempted. Because of remarkable tortuosity of the patient's left ICA, the balloon catheter could not be navigated beyond the carotid siphon. SNP administration continued, and progressive dilation of M l became apparent. Although administration was stopped at 30 mg, angiographic improvement continued, and there­ fore, no further attempt to perform angioplasty was made. The left­ sided cerebral circulation time improved from 8 to 6 seconds. Fiftyfive minutes after the final SNP dose was delivered, the baseline diameter of the left M l appeared to have been nearly restored and the PComA showed improved filling (Fig. 5, A and B). The circulation time on the right was also reduced (5.5 s). Large-caliber vessels on the right showed minimal dilation, but the distal A l segment and distal ACA in the interhemispheric fissure were better visualized. No im­ portant systemic hemodynamic changes were observed. Hypotension did not occur. ICP elevation to 21 mm Hg occurred transiently after the first dose and responded promptly to 3 cc of ventricular drain­ age. It did not occur subsequently, remaining at or below 17 mm Hg (Table 5). The neurophysiological parameters remained stable, and the procedures were terminated. The patient was returned to the neuro­ surgical intensive care unit, where TCD velocity measurements were noted to be dramatically improved (Table 4). The same evening, the patient demonstrated consistent purposeful movement of the previ­ ously plegic right side. Subsequent elevation of TCD velocities on the contralateral (right) side led to repeat cerebral angiography with angioplasty of the right supraclinoid carotid artery approximately 12 hours after the SNP treatment. Angiography was also performed on the left and demonstrated recurrent narrowing of the A l and M l segments that had been previously dilated by SNP. The cerebral circulation times were 5.5 and 6 seconds for right and left hem i­ spheres, respectively. Endovascular maneuvers, using a smaller bal­ loon catheter and a hydrophilic guidewire, permitted navigation of the patient's tortuous left ICA, and the M l segment underwent an­ gioplasty concurrently with the administration of additional SNP (12 mg) to enhance distal circulation in resistance vessels. Neither the supraclinoid carotid nor the A l segment underwent angioplasty. The cerebral circulation times became 5 seconds bilaterally within 60 minutes of concurrent SNP treatment and angioplasty. Again, no hemodynamic or physiological changes were observed. M ABP re­ mained constant at 125 mm Hg. The next day, the patient was awake and alert with a GCS score of 10 (intubated), and he followed com ­ mands briskly with all extremities. Serum cyanate levels drawn on the evening after treatment were undetectable. CT of the brain performed 5 days after treatment demonstrated improvement in cerebral swell­ ing. This patient was subsequently uneventfully discharged with no demonstrable neurological deficit (GCS score of 15). Patient 3 A 38-year-old woman presented with aneurysmal SAH from a ruptured right carotid ophthalmic segment aneurysm. The clinical SAH grade (2) was III, and the radiographic SAH grade (1) was III (Fig. 6). The patient had no significant medical history but demon- Neurosurgery ; Vol. 4* , No. 7, January 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 tomy was placed for ventricular decompression. The ICP remained between 15 and 18 mm Hg. No significant improvement in the patient's neurological status was observed after 1 hour, despite opti­ mization of HHH therapy. The patient was taken emergently for cerebral angiography, which demonstrated severe narrowing of the right A l, left A l, and left M l arterial segments (Fig. 3, A -C ). The cerebral circulation time was 8 seconds on the left and 6.5 seconds on the right. Prolonged SSEP latencies representing the right upper and lower extremities observed at baseline before intervention were thought to be consistent with the angiographic demonstration of vasospasm in the left MCA and ACA distribution (Fig. 4). ITSNP (4.0 mg/ml) was delivered in aliquots of 1.0 ml, admixed with the pa­ tient's own CSF from the ventriculostomy, by means of a closed system using two syringes at intervals of 5 to 10 minutes for 53 minutes. The first dose was exceptional in that 2.0 ml (8.0 mg) was administered and 20 minutes were allowed to elapse before adm in­ istration of the second dose (Table 5). The total dose delivered was 30 mg. The circumstances of administration in terms of neuroanes­ thetic and neurophysiological monitoring were identical to those for Patient 1. 52 Thomas and Rosenwasser TABLE 3. Relative Diameters of Cerebral Vessels of Patient 1 before and after Treatment with Intraventricular Sodium Pretreatment RA1 RPComA LA1 LM1 LM2 LA3 0.96 0.93 0.93 1.96 0.93 0.93 (estimated) Post-treatment (24 h) 1.93 1.94 2.00 (estimated) 2.48 2.16 (estimated) 2.16 % Baseline 201 208 215 127 232 232 a RA1, right At segment; RPComA, right posterior communicating artery; LA1, left A1 segment; LM1, left M1 segment; LM2, left M2 segment; LA3, left A3 segment. Relative diameters of cerebral vessels of Patient 1, measured digitally before and after treatment with intrathe­ cal sodium nitroprusside, are shown. Absolute diameters, and therefore units of measurement, are omitted because calibration was not per­ formed on the original angiograms. gram of the right cerebral ICA, obtained 24 hours after treat­ ment. Persistent dilation of A1 (7) and PComA (2) can be observed. B, lateral projection angiogram of the right cere­ bral ICA, obtained 24 hours after treatment. Persistent dila­ tion of PComA (7) with a focal region of recurrent vasocon­ striction can be observed; A2 artery segment (2). C, anteroposterior projection angiogram of the left cerebral ICA, obtained 24 hours after treatment. Further dilation of the M1 (7) and M2 segments of the M CA (2) and A3 segment of the A C A (arrow) can be observed, compared with the same frame of the pretreatment angiogram (see Fig. 1C). D, lateral projection angiogram of the left cerebral ICA, obtained 24 hours after treatment. The dilation of the A3 segment of the A C A (2) can be seen; A2 artery segment (7). E, smaller anteroposterior projection angiogram (for compari­ son with Panel F) of the left cerebral ICA, obtained before treatment. The cerebral circulation time was 7.0 seconds. F, anteroposterior projection angiogram of the left cerebral ICA, obtained 24 hours after treatment. Note the enhanced distal circulation, compared with the same frame of the pretreatment angiogram (see Panel E). The cerebral circulation time was 5.0 seconds. strated clinical evidence of significant cardiac dysfunction of unclear origin (ejection fraction, 35%). Coil embolization of the aneurysm was recommended and performed in an uncomplicated fashion on hospi­ tal Day 2 (SAH Day 4). Full-scale heparinization was instituted during Neurosurgery , Vol. 44, No. I, lanuar\ 1900 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 F IG U R E 2. Patient 1. A, anteroposterior projection angio­ the course of coil embolization and continued for 24 hours in routine fashion. Nimodipine, Decadron, and Dilantin were administered in routine fashion. The patient remained in the neurosurgical intensive care unit without complication. TCD velocity m easurem ents (carotid index) were observed to increase on postoperative Day 2 (SAH Day 6) to 8.5 on the right and 9.2 on the left. No other clinical changes were observed. Intravascular volume expansion was performed in routine fashion using Swan-Ganz catheterization. On SAH Day 8, TCD values were noted to have once again increased (carotid indices, 8.5 and 9.3 on the right and left, respectively). There were no neurological svmptoms. Cardiac performance was optimized, and pulm onary capillar}7 wedge pressure was maintained at 14 mm Hg. Because of the pa­ tient's intrinsic cardiac dysfunction, however, the cardiac index did not surpass 3.25. On SAH Day 9, TCD values once again increased (carotid indices, 9.6 and 5.8 on the right and left, respectively) and Neosynephrine (American Regent Labs., Shirley, NY) and dopamine were added to the regimen. M ABP w as maintained between 110 and 120 mm Hg. On SAH Day 10, the patient experienced right hemiparesis and aphasia in the setting of a reduction of the M ABP to 80 mm Hg, precipitated by bacteremia. The cerebral sym ptom s resolved promptly with restoration of the M ABP to greater than 90 mm Hg. Antibiotic treatment was begun. On SAH Day 11, an aspiration pneu­ monia was diagnosed. In the evening, the patient again sustained a transient hemiparesis and aphasia in the setting of a transient de­ crease in the MABP. On SAH Day 12, TCD values remained markedly elevated (Table 6). The patient demonstrated no neurological symp­ toms. A temperature increase to 39°C occurred, and the antibiotic regimen was adjusted. Chest radiographs demonstrated pulmonary edema, and arterial oxygen saturation was marginal. Intubation was performed electively. Subsequently, the patient manifested right hemiplegia and aphasia. She was stuporous, with a G C S score of 7. The MABP was 110 mm Hg, and the cardiac index was 4.0. A frontal was placed, and the patient was taken emergentlv for C T of the brain. The results of the C4 were unremarkable, and the patient was taken directly to the angiography suite. Cerebral angiog­ raphy disclosed severe vasoconstriction of the left M l, right M l, left and iight supiaclinoid carotid, and right Al segm ents (F/g. 7, .\—C)The ceiebial ciiculation time ot the lett hemisphere was prolonged to 8 seconds and, ot the tight, to /.5 seconds. The ICP was 10 mm Hg Balloon angioplasty ot the left M l and both supraclinoid carotid segments was performed. Angioplasty of the right M l and Al could not be accomplished for technical reasons. SNP was administered via Reversal of Human Vasospasm 53 TABLE 4. Mean Transcranial Doppler Velocities and Velocity Indices of the Cerebral Arteries of Patient 2 before and after Treatment with Intraventricular Sodium Nitroprusside*' Hours after Treatment Vessel/Index -2 1 RMCA RACA RICA RCIXb LMCA LACA LICA LCIXi> AComA 109 85 23 1.74 211 69 25 8.44 NM -1 5.5 19 100 38 28 3.57 252 40 20 12.60 NM 179 42 20 8.95 81 85 23 3.52 NM 208 83 20 10.4 145 164 34 4.26 NM 46 49 179 81 22 8.14 150 169 33 4.55 NM 164 85 22 7.46 152 121 33 4.61 NM 26 200 64 22 9.09 145 151 33 4.39 NM 75.5 91.5 97 116 122 135 138 23 5.87 169 169 33 5.12 NM 138 85 25 5.52 169 152 33 5.12 NM 140 85 25 5.6 133 100 27 4.93 NM 157 104 26 6.03 126 145 29 4.34 NM 151 95 26 5.81 123 181 28 4.39 NM 69 152 93 22 6.9 151 110 33 4.58 NM a RMCA, right middle cerebral artery; RACA, right anterior cerebral artery; RICA, right internal carotid artery; RCIX, right carotid index; LMCA, left middle cerebral artery; LACA, left anterior cerebral artery; LICA, left internal carotid artery; LCIX, left carotid index; AComA, anterior communicating artery; NM, not measured. Mean transcranial Doppler velocities of the cerebral arteries of Patient 2, measured before and after treatment with intraventricular sodium nitroprusside (in cm/s), are shown. b Carotid index was calculated as follows: mean MCA velocity/mean cervical ICA velocity. 1 -J ‘ ■- .-A • 50 V.-, !N : L j ’ v' ' F IG U R E 3. Patient 2. A , anteroposterior - A / >, ■rx -w t f o r T o x ■ F r e e uV Rmp 600 (. MRV ■- O 1 0 m s . N o r m 3 0 .y N?0 . \ A ..3 'N : uV Pmp 30 I 7 -a 0 . . x " , \ -O .. 1 / U V Rtap- 8 projection angiogram of the right cerebral ICA, obtained before treatment with SNP. Note that the A1 segment is severely constricted. B, antero­ posterior projection angiogram of the left cerebral ICA, obtained before treatment with SNP. Note that the A1 and M1 F IG U R E 4. Patient 2. SSEPs from cortex of the left hem i­ sphere (wave N20, representing the right upper extremity), demonstrating a leftward shift to 20.0 milliseconds (double arrow) from the baseline (lower trace , 23.9 ms) within m in­ utes after administration of the first dose of SNP. The latency of the P40 wave representing the lower extremity was sim i­ larly decreased, from 51.5 to 47.7 milliseconds (comparison figure not shown). segments are severely constricted. The cerebral circulation time was prolonged (8.0 s, see text). C, lateral projection angiogram of the left cerebral ICA, obtained before treatment with SNP. The PCom A is poorly visualized. the ventriculostomy in two doses to a total of 10 mg. The patient's blood pressure and ICP remained stable (Table 7). The left cerebral circulation time had improved to 6 seconds within 20 minutes of completion of SNP administration and angioplasty (Fig. 8, B and C). The right-sided cerebral circulation time had improved to 6.5 seconds after administration of SNP without angioplasty in the A l or M l segments and without appreciable change in the caliber of the larger conductance vessels on that side (Fig. 8A). The patient was returned to the neurosurgical intensive care unit, and pentobarbital coma was ■nstituted to maintain an electroencephalographic burst suppression pattern. TCD values measured 7 hours later demonstrated a dramatic reduction in velocities in the left MCA but no change in the right MCA and a slight increase in velocities in the right and left ACAs (Table 6). No intervention was added. TCD measurements 14 hours later demonstrated a substantial reduction in velocities of the vessels that had not undergone angioplasty (right MCA and ACA as well as left ACA) and a further reduction in left MCA velocity (Table 6). The low velocities persisted, and the patient was brought out of pentobarbital coma after 5 days. The patient demonstrated continual improvement without recurrence of symptomatic vasospasm and without a trace of neurological deficit (GCS score of 15). Volume expansion was main­ tained; pressors were gradually and uneventfully withdrawn. Cere­ bral angiography was not repeated. The ventriculostomy was discon- Neurosurgery, Vol. 44 , No. 7, January 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 N4S 54 Thomas and Rosenwasser TABLE 5. Measurements of Mean Arterial Blood Pressure and Intracranial Pressure of Patient 2 during Treatment with Intraventricular Sodium Nitroprusside'1 Time of dose (min) M ABP (mm Hg) ICP (mm Hg) 0 96 21 20 131 15 26 128 16 36 136 16 41 135 16 46 139 16 53 126 17 a MABP, mean arterial blood pressure; ICP, intracranial pressure. Measurements of MABP and ICP for Patient 2, obtained continuously during the first treatment (without angioplasty, see text) and recorded at times of administration of intraventricular sodium nitroprusside (4.0 mg/ml), are shown. Sodium nitroprusside was administered in 1.0- or 1.5-ml aliquots (see text). Entire treatment duration was 53 minutes. F IG U R E 6. Patient 3. Different slices of unenhanced com ­ puted tomographic scans of the brain, obtained at admission, demonstrating diffuse SAH. gram of the left common carotid artery, obtained 55 minutes after termination of SNP treatment (115 min after initiation) (magnification slightly greater than that of the pretreatment image). Note that the baseline diameter of the M1 segment appears nearly restored and that the proximal A1 segment also shows mild dilation. The PComA (arrow) can be clearly visu­ alized, compared with the same frame of the pretreatment angiogram (selective ICA injection). A washout of dye occurred slightly earlier. The cerebral circulation time was 6.0 seconds. B, lateral projection angiogram of the left ICA, obtained 55 minutes after termination of SNP treatment (115 min after initiation). The PComA (arrow) can be clearly visualized, compared with the same frame of the pretreat­ ment angiogram (see Fig. 3C) . tinued after a trial of clamping demonstrated no further ventricuIomegalv■ CT of the brain performed 10 days after treatment showed no residual abnormality. The patient was discharged on SAH Day 32 to a rehabili­ tation facility for a brief time for reconditioning purposes without neu­ rological deficit. She then returned home, and her condition was excel­ lent. She was independent at 7 months after treatment. D ISCU SSIO N N O and cerebral vasoconstriction NO is the smallest biologically active molecule known and is the mediator of an extraordinary range of physiological processes (6, 9). It is also a known physiological antagonist of endothelin 1, which is the most potent known mammalian vasoconstrictor (at least 10 times the vasoconstrictor potency of angiotensin II) (11), which has been implicated in CDCV by many investigations (3, 4, 7). NO donors have been dem on­ strated by direct observation to rapidly reverse induced cere­ Neurosurgery, Vol. 44, No. /, lanuar\ 199 9 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 F IG U R E 5. Patient 2. A, anteroposterior projection angio­ bral vasoconstriction in an animal model in which NO was delivered intrathecally to the adventitial side of the blood vessel (10). Systemic hypotension did not occur in this model, presumably because the biological half-life of N O is extremely short (5, 6), and the route of administration tended to favor local vasodilatory action. 1 These preliminary observations regarding three patients suggest that intrathecal delivery of NO donor compounds may be effective and durable in the treatment of severe re­ fractory cerebral vasospasm that occurs after aneurysmal SAH in humans. Of three patients presenting in grave condi­ tion with profound neurological deficit, two have no residual deficit and one is greatly improved and living at hom e since relatively early (69 d) after the ischemic deterioration. Treat­ ment of the latter patient, furthermore, was uncharacteristi­ cally delayed, because she had been hem iplegic from vaso­ spasm for at least 12 hours before arriving at our hospital for further treatment. The observed pattern of reversal of vasoconstriction by SNP and delayed segmental recurrence of the condition is consis­ tent with a mechanism that depends on the presence and availability of a presumably finite substrate and that can apparently be at least temporarily overw helm ed. Although the durability of this treatment is not known, the effect ob­ tained by the reported intraventricular doses seems to have been sufficiently lengthy to be useful under the dire circum­ stances of critical brain ischemia, the possibility of repeated treatments and increased dosage notwithstanding. Although a substantial vasodilation of large-caliber conductance vessels in response to SNP was dem onstrated, the im provem ent in collateral circulation, dilation of sm aller resistance vessels, and shortening of circulation time, som etim es without a dra­ matic change in the size of conductance vessels, seem s to be an equally important effect of this treatment. Improved circu­ lation time alone, at a constant cerebral perfusion pressure, may imply enhanced circulation at anatom ic levels below angiographic resolution. NO is known to effect vasodilation thiough a cyclic guanosine m onophosphate-m ediated protein kinase activation mechanism (6, 9) that may be presumed to be functional at the level of the arteriole, because it is a muscu lari zed vessel. The NO molecule derived from 1TSNP Reversal of Human Vasospasm 55 TABLE 6. Mean Transcranial Doppler Velocities and Velocity Indices of the Cerebral Arteries of Patient 3 before and after Treatment with Intraventricular Sodium Nitroprusside'1 Days after Treatment Hours after Treatment Vessel/Index -3 4 312 RACA 303 214 RICA RCIX'’ LM CA RM CA LACA LICA LC IX fa ACoA 14.5 20.5 68 89 98 16 17 18 132 176 167 82 85 31 31 212 200 52 186 47 26 5 .0 8 44 33 5 .0 6 50 3 .7 2 100 47 4 .5 2 29 6 .0 7 159 47 93 107 78 42 69 81 124 1.81 124 33 81 62 176 150 38 40 1.73 167 50 1.62 52 50 2 .4 8 38 1 .8 7 64 131 138 140 39 52 140 136 224 136 64 162 113 112 35 20 44 8 .6 5 1 5 .6 107 3 .0 9 152 33 4 .1 2 25 4 .5 2 97 45 45 109 45 42 38 38 4 .0 33 2 .9 4 69 33 . 54 33 157 179 239 71 33 7 .2 4 NM 2 .3 8 171 1.3 6 140 1.3 6 78 64.5 66 42 1.05 202 2 .0 5 164 1 .7 4 2 .3 8 45 2 .0 7 71 a RMCA, right middle cerebral artery; RACA, right anterior cerebral artery; RICA, right internal carotid artery; RCIX, right carotid index; LMCA, left middle cerebral artery; LACA, left anterior cerebral artery; LICA, left internal carotid artery; LCIX, left carotid index; AComA, anterior communicating artery; NM, not measured. Mean transcranial Doppler velocities of the cerebral arteries of Patient 3, measured before and after treatment with intraventricular sodium nitroprusside (in cm/s), are shown. b Carotid index was calculated as follows: mean MCA velocity/mean cervical ICA velocity. TABLE 7. Measurements of Mean Arterial Blood Pressure B Time of dose (min) MABP (mm Hg) ICP (mm Hg) j \7* jfri * F IG U R E 7. Patient 3. A, anteroposterior projection angiogram of the right ICA, obtained before treatment. Note that the M1 and supraclinoid ICA segments and III the A1 segment are constricted. The cerebral circulation time was 7.5 seconds. The coil mass (arrow) i is demonstrated over the supraclinoid carotid. B, anteroposterior projection angiogram of the left ICA, obtained before treatment. Note that the A1, M1, and supraclinoid ICA segments are constricted. There is reflux of contrast into the external carotid artery. The cerebral circulation time was 8.0 seconds. C, lateral projection angiogram of the left ICA , obtained before treatment. The subtracted coil mass (arrow) on the right is superimposed over the left ICA. having p e n e tra te d th e a d v e n titia l s id e o f th e b lo o d v e s s e l, activates the s o lu b le fo r m o f g u a n y la t e c y c la s e , w h ic h in tu rn converts g u a n o s in e 5 '- t r i p h o s p h a t e to c y c lic g u a n y lic a cid . Through this la tter s e c o n d m e s s e n g e r , p r o te in k in a s e a c tiv a - 0 100 10 10 20 100 9 100 11 a MABP, mean arterial blood pressure; ICP, intracranial blood pres­ sure. Measurements of MABP and ICP for Patient 2, obtained contin­ uously during the first treatment (without angioplasty, see text) and recorded at times of administration of intraventricular sodium nitro­ prusside (1.0 mg/ml), are shown. Sodium nitroprusside was adminis­ tered in 5.0-ml aliquots. Entire treatment duration was 20 minutes. tion takes place and subsequently causes relaxation of m us­ cular smooth muscle, resulting in vasodilation. Relatively thinner walls of smaller vessels could conceivably render them more readily susceptible to the effects of NO donors administered adventitially, causing them to respond faster and improving cerebral circulation time. Such effects are con­ sistent with the known physical properties of NO, which is a very small molecule capable of penetrating tissues directly. This is most evident in its interaction with blood vessels, which it may penetrate from the adventitial side (Mima et al. (4a), 1989, cited in Ref. 9), but would also be expected in brain tissue and blood clot. The computed tomographic scans of the patients in this study demonstrated substantial volum es of subarachnoid blood, which do not appear to have prevented the salutary penetration of ITSNP. It is also possible that ITSNP may penetrate the parenchyma by way of the ependyma of the ventricle, its effects in this way becom ing, to a degree, independent of ventricular CSF circulation. Considerations for route of administration These salutary effects of vasodilation may be obtainable without precipitating intracranial hypertension or systemic hypotension. If these observations are confirmed, thought may be given to administering these com pounds even more locally, by subarachnoid catheter or intraoperatively, under Neurosurgery , Vol. 44, No. 1, January 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 and Intracranial Pressure of Patient 56 Thomas and Rosenwasser was relatively good, it might have been better had she been B / I F IG U R E 8. Patient 3. A, anteroposterior projection angiogram of the right ICA, obtained 20 minutes after segment. Note that there is no appreciable change in the caliber of the larger conductance vessels (Al and M1). The cerebral circulation time was 6.5 seconds. Filling of the A CA and lenticulostriate vessels was slightly better, compared with the same frame of the pretreatment angiogram (Fig. 7A). The subtracted coil mass (arrow) can be visualized over the supraclinoid carotid. B, anteroposterior projection angiogram of the left ICA, obtained 20 minutes after treatment with SNP and angioplasty of the left and right supraclinoid carotid segments and of the left M1 segment. Note that although the A1 segment shows no change in caliber, the distal A CA in the interhemispheric fissure and its branches (arrows) can be better visualized, compared with the same frame of the pretreatment angiogram (Fig. 7B). The cerebral circulation time was 6.0 seconds at constant MABP. C, lateral projection angiogram of the left ICA, obtained 20 minutes after treatment with SNP and angioplasty of the left and right supraclinoid carotid segments and of the left M1 segment. Improved filling of the A C A (arrow), which did not undergo angioplasty, is demonstrated in comparison with the same frame of the pretreatment angiogram (Fig. 7C). the appropriate conditions of segmental vasospasm. We have observed a salutary clinical and angiographic effect of an NO donor administered locally in this manner. Targeted local therapy of this type would be a theoretical advantage where CSF pathways become compromised by brain swelling. Timing and dosage The timely delivery of this treatment for established CDCV also deserves emphasis. It is our practice to perform cerebral angiography and plan alternative therapy when medical treatment with HHH therapy has been maximized and inef­ fective for 1 hour. Although the clinical outcome for Patient 1 Caution The administration of these com pounds in this m anner is a serious undertaking that should not go unaccom panied by intense clinical, neurophysiological, and neuroanesthetic monitoring, preferably also under direct angiographic con­ trol, as in these studies, until proper parameters and any peculiar hazards have been identified. W e have noted a brief episode of hypotension occurring on one occasion after a relatively high one-time intraventricular dose (12 mg) admin­ istered for severe distal vasospasm in the setting of cerebral angioplasty of a proximal arterial segment. | Particular caution should also be exercised under circum­ stances of long-standing ( > 4 h) brain ischemia, because the demonstration of frank infarction or significant intraparenchymal hematoma is a relative contraindication to the use of these com pounds in the m anner described, not only because of the risk of hemorrhagic conversion of infarction by reper­ fusion but also because of potential excitotoxic effects of NO. We have refrained from the use of these m edications in the setting of medically refractory vasospasm w ith concurrent obvious infarction or intraparenchymal hem atom a. Our cur­ rent indications for including patients in this study are the same as those for cerebral angioplasty; patients with vaso­ spasm refractory to HHH therapy as manifested by a failure to respond clinically within 60 minutes are eligible for treat­ ment with or without cerebral angioplasty. If angiographic or neurophysiological improvem ent is not observed within 30 minutes ot drug administration, patients are submitted to angioplasty where it is feasible (proximal conductance vessels not distal to Ml and Al segments). We have observed neuro- Neurosurgery, Vol. 44, No. /, lanuary 1000 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 treatment with SNP and angioplasty of the left and right supraclinoid carotid segments and of the left M1 treated earlier. It is emphasized that these observations regarding the use of intrathecal NO donors are preliminary and that optimal dosage and efficacy have not been determined. Whether higher doses can prevent recurrence of established CDCV is unknown. The structure of our dose escalation study imposed a limitation of 30 mg per procedure, and all patients received angioplasty either subsequently (Patients 1 and 2) or concur­ rently (Patient 3). Of note, two patients initially received ITSNP with obvious effect without angioplasty. Although we think that the vasodilatory effects of SN P and those of angio­ plasty are distinguishable in these cases (because of their separation in time and the distribution of the vessels affected), the definitive dosing, safety, and efficacy of these compounds must be rigorously demonstrated in clinical trials. In obser­ vance of this limitation of the present report, we deliberately avoided specific assertions regarding dosage. The concentra­ tion of medication delivered to Patient 2, for exam ple, was increased in the interest of m inim izing the fluid volume de­ livered intracranially (because the ICP was already > 1 5 mm Hg) and also the time to therapeutic effect, based on previous exper­ imental observations (10). Similarly, the ideal timing of this treatment is not established and exploring the possibility of prophylactic administration of NO donors to offset or prevent CDCV awaits the definitive demonstration of their safety. Reversal of Human Vasospasm physiological improvement (shortened SSEP latencies) where the only discernible angiographic improvement seems to be in collateral circulation and improved circulation time, without a dramatic increase in the diameter of spastic relatively large con­ ductance vessels. In these cases, we have combined cerebral angioplasty with intrathecal NO donor administration. SUMMARY ACKNOWLEDGMENTS We gratefully acknowledge the expert technical assistance and professional advice of Jack Scully, Roger Barone, and Bob Curtin of the Audiovisual Department of Wills Neurosensory Institute and Wills Eye Hospital who tirelessly provided m ed­ ical images of the highest quality. Received, January 27, 1998. Accepted, August 21, 1998. Reprint requests: Jeffrey E. Thomas, M.D., Division of Cerebrovascular Surgery and Interventional Neuroradiology, Department of Neurologi­ cal Surgery, Thomas Jefferson University, Philadelphia, PA 19107. REFERENCES 1. Fisher CM, Kistler JP, Davis JM: Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomo­ graphic scanning. Neurosurgery 6:1-9, 1980. 2. Hunt WE, Hess RM: Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg 28:14-20, 1968. 3. Kasuya H, Weir BKA, White DM, Stefansson K: Mechanism of oxyhemoglobin-induced release of endothelin-1 from cultured vascular endothelial cells and smooth-muscle cells. J Neurosurg 79:892-898, 1993. 4- Kobayashi H, Hayashi M, Kobayashi S, Kabuto M, Handa Y, Kawano H, Ide H: Cerebral vasospasm and vasoconstriction caused by endothelin. Neurosurgery 28:673-679, 1991. 4a. Mima T, Yanagisawa M, Shigeno T, Saito A, Goto K, Takakura K, Masaki T: Endothelin acts in feline and canine cerebral arteries from the adventitial side. Stroke 20:1553-1556, 1989. 3. Morris SM Jr, Billiar TR: New insights into the regulation of induc­ ible nitric oxide synthesis. Am J Physiol 266:E829-E839, 1994. 6. Nathan C: Nitric oxide: Roles, tolls and controls. Cell 78:915-918, 1994. 7. Suzuki R, Masaoka H, Hirata Y, Marumo F, Isotani E, Hirakawa K: The role of endothelin-1 in the origin of cerebral vasospasm in patients with aneurysmal subarachnoid hemorrhage. J Neuro­ surg 77:96-100, 1992. 8. Teasdale GF, Jennett B: Assessment of coma and impaired con­ sciousness. Lancet 2:81-84, 1974. 9. Thomas JE: Molecular biological considerations in cerebral vaso­ spasm following aneurysmal subarachnoid hemorrhage. Neuro­ surg Focus 3:1-5, 1997. 10. Thomas JE, Nemirovsky A, Zelman V, Giannotta SL: Rapid re­ versal of endothelin-l-induced cerebral vasoconstriction by intra­ thecal administration of nitric oxide donors. Neurosurgery 40: 1245-1249, 1997. 11. Yanagisawa M, Kurihara H, Kimura S, Tom obe Y, Kobayashi M, Mitsui Y, Yazaki Y, Goto K, Masaki T: A novel potent vasocon­ strictor peptide produced by vascular endothelial cells. Nature 332:411-415, 1988. CO M M EN TS The authors present their interesting and exciting p re­ lim inary experience with intrathecal nitric oxide donors for the treatm ent of vasospasm from aneu rysm al su barach noid hem orrhage. They state the im portant caveats regarding the prelim inary nature of their w ork and the possible n eg ­ ative effects of nitric oxide adm inistration. Som e questions rem ain regarding the ability of the treatm ent drug a d m in ­ istered by this route to effectively reach all vessels in spasm or all territories at risk for ischem ic injury. If the au thors' goal was to increase cerebral blood flow, it seems some method of blood flow determination may be helpful in future studies to answer questions regarding dose and efficacy. Fur­ thermore, the role of this treatment in relation to standard vol­ ume and cardiac manipulations, as well as other experimental techniques such as intrathecal thrombolytic therapy and intraaortic balloon counterpulsation, will need to be more carefully assessed in further studies. Warren R. Selman Cleveland, Ohio Thom as and R osenw asser report three interesting cases in which intraventricular sodium nitrop ru ssid e (SN P) p ro ­ duced im provem ent in arteriographic caliber and tran scra­ nial D oppler velocity in patients w ith cerebral vasospasm . Intravenously (2) or intra-arterially (3) adm inistered n itro ­ prusside has been show n to be effective in a canine m odel of vasospasm , but clinical use has been lim ited by u n to ­ w ard system ic effects (4). Previous reports of efficacy for intrathecally adm inistered SN P (ITSN P) in reversing v a so ­ spasm (1) suggested a potential benefit w ithou t co n co m i­ tant hypotension, as observed in the three patients d e­ scribed in this report. O f note, the SN P effect seem ed to be transient in som e cases, sim ilar to that seen w ith intraarterially adm inistered papaverine. In addition, in this re­ port, patients did not im m ediately im prove after the a d ­ m inistration of SN P, although angioplasty produced rapid clinical im provem ent. The concurrent use of angiop lasty in these patients com plicates the analysis. N evertheless, these Neurosurgery, Vol. 44, No. 1, January 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/44/1/48/2847548 by Karolinska Institutet Library user on 13 February 2019 In summary, this small preliminary experience with the intrathecal administration of NO donors for the treatment of cerebral vasospasm in hum ans seems to be favorable. Vaso­ dilation may be effected without precipitating intracranial hypertension or systemic hypotension. Although this report emphasizes angiographic and ultrasonographic docum enta­ tion of effect, an equally or more important effect of this treatment may be enhancem ent of cerebral blood flow at the level of the microcirculation. W e have observed no adverse outcome in the extremely judicious, cautious, and vigilant use of this treatment. Further clinical studies are indicated, par­ ticularly using diffusion magnetic resonance imaging and xenon-enhanced CT for the sensitive evaluation of cerebral ischemia before and after treatment with ITSNP. Such studies will help to elucidate the potential use of ITSNP earlier in the course of CDCV, perhaps before an aggravated neurological deficit occurs. Dose escalation studies are currently underway at this institution. 57