Do negative CCT and CSF findings exclude a subarachnoid haemorrhage? A retrospective analysis of 220 patients with subarachnoid haemorrhage.

OBJECTIVE: Subarachnoid haemorrhage (SAH) constitutes a neurological emergency. In most cases, the diagnosis is easy to establish; however, in rare cases, verification of the diagnosis is difficult. In this retrospective analysis, we report the clinical characteristics of patients with SAH who were admitted to our neurological intensive care unit. We focus on the additional diagnostic approaches in patients with a high suspicion of SAH but failure of the 'classic' diagnostic tools. METHODS: A retrospective chart review was performed for all patients in whom SAH was diagnosed between 1996 and 2008. Two hundred and twenty patients were analysed for presenting symptoms, radiological and laboratory findings, hospital course and outcome. RESULTS: A total of 220 patients were identified (mean age 50.5 years, 127 women). In 217 patients, the diagnosis was based upon cerebral computed tomography (CCT) or lumbar puncture. In three patients, the diagnostic work-up was continued because of distinct clinical signs even though CCT and cerebrospinal fluid (CSF) were negative for SAH. In these patients, vasospasm was detected by transcranial doppler sonography (TCD) and/or diagnosis of aneurysm was confirmed by conventional angiography. CONCLUSION: Subarachnoid haemorrhage with negative CCT and CSF is a rare presentation of a severe acute neurological emergency. Further diagnostic as TCD/computed tomography (CT)-A or MR-A should be considered in all patients with typical clinical presentation for SAH but unremarkable CCT and CSF as an additional diagnostic tool. Ultimately, a conventional angiography should be performed if distinct clinical signs of SAH are presented.

High spatial resolution and high contrast visualization of brain arteries and veins: impact of blood pool contrast agent and water-selective excitation imaging at 3T.

PURPOSE: To investigate a blood pool contrast agent and water-selective excitation imaging at 3 T for high spatial and high contrast imaging of brain vessels including the veins. METHODS AND RESULTS: 48 clinical patients (47 ± 18 years old) were included. Based on clinical findings, twenty-four patients received a single dose of standard extracellular Gadoterate-meglumine (Dotarem®) and 24 received the blood pool contrast agent Gadofosveset (Vasovist®). After finishing routine MR protocols, all patients were investigated with two high spatial resolution (0.15 mm (3) voxel size) gradient echo sequences in random order in the equilibrium phase (steady-state) as approved by the review board: A standard RF-spoiled gradient-echo sequence (HR-SS, TR/TE 5.1/2.3 msec, FA 30°) and a fat-suppressed gradient-echo sequence with water-selective excitation (HR-FS, 1331 binominal-pulse, TR/TE 8.8/3.8 msec, FA 30°). The images were subjectively assessed (image quality with vessel contrast, artifacts, depiction of lesions) by two investigators and contrast-to-noise ratios (CNR) were compared using the Student's t-test. The image quality and CNR in the HR-FS were significantly superior compared to the HR-SS for both contrast agents (p < 0.05). The CNR was also improved when using the blood pool agent but only to a minor extent while the subjective image quality was similar for both contrast agents. CONCLUSION: The utilized sequence with water-selective excitation improved image quality and CNR properties in high spatial resolution imaging of brain arteries and veins. The used blood pool contrast agent improved the CNR only to a minor extent over the extracellular contrast agent.

Prediction of hemorrhagic transformation after thrombolytic therapy of clot embolism: an MRI investigation in rat brain.

BACKGROUND AND PURPOSE: Thrombolytic treatment of stroke carries the risk of hemorrhagic transformation. Therefore, the potential of MRI for prediction of recombinant tissue plasminogen activator (rtPA)-induced bleeding is explored to identify patients in whom rtPA treatment may provoke such complications. METHODS: Spontaneously hypertensive rats (SHR) (n=9) were submitted to middle cerebral artery (MCA) clot embolism, followed 3 hours later by intra-arterial infusion of 10 mg/kg rtPA. Untreated SHR (n=9) were infused with saline. MRI imaging was performed before treatment and included apparent diffusion coefficient (ADC), T2, and perfusion mapping and contrast enhancement with gadolinium-DTPA. The distribution of intracerebral hemorrhages was studied 3 days later by histological staining. RESULTS: Clot embolism led to the rapid decline of ADC in the territory of the occluded artery. Tissue lesion volume derived from ADC imaging increased by 155+/-69% in the untreated animals and by 168+/-87% in the treated animals (P=NS), determined on the histological sections after 3 days. This same lesion growth in both groups indicated absence of therapeutic effect after 3-hour treatment delay. Hemorrhagic transformations were significantly more frequent in treated SHR (P<0.05). In untreated rats, hemorrhages were found in the border zone of the ischemic territory; in treated animals, hemorrhagic transformations occurred in the ischemic core region. rtPA-induced hemorrhages were predicted by a disturbance of the blood-brain barrier in 3 of 4 animals before treatment by Gd-DTPA contrast enhancement but not by ADC, T2, or perfusion imaging. The region of contrast enhancement colocalized with subsequent bleeding in these animals. CONCLUSIONS: The disturbance of blood-brain barrier but not of other MR parameters allows risk assessment for hemorrhagic transformation induced by subsequent thrombolytic treatment.

Sampling and evaluation of specific absorption rates during patient examinations performed on 1.5-Tesla MR systems.

It was the purpose of present study, to evaluate a large number of exposure-time courses measured during patient examinations in clinical routine in relation to the current IEC standard and the draft version of the revised standard and, moreover, to investigate whether there is a correlation between the subjective heat perception of the patients during the MR examination and the intensity of RF power deposition. To this end, radiofrequency exposure to 591 patients undergoing MR examinations performed on 1.5-Tesla MR systems was monitored in five clinics and evaluated in accordance with both IEC standards. For each of the 7902 sequences applied, whole body and partial body SARs were estimated on the basis of a simple patient model. Following the examinations, 149 patients were willing to provide information in a questionnaire regarding their body weight and their subjective heat perception during the examination. Although patient masses entered into the MR system were in some cases too high, reliable masses could be estimated by the SAR monitor. In relation to our data, the revision of the IEC standard results in a tightening of the restrictions, but still more than 96% of the examinations did not exceed the SAR limits recommended for the normal operating mode. For the exposure conditions examined, no statistically significant correlation was found between the subjective heat perception of the patients and the intensity of power deposition. Taking advantage of the possibility to compute running SAR averages, MR sequences can be employed in clinical practice for which SAR levels exceed the defined IEC limits, if the acquisition time is short in relation to the averaging period and energy deposition has been low previous to the applied high-power sequence.

Predictive value of neurochemical monitoring in large middle cerebral artery infarction.

BACKGROUND AND PURPOSE: Space-occupying brain edema is a life-threatening complication in patients with large hemispheric stroke. Early identification of patients at risk is necessary to decide on invasive therapies such as decompressive hemicraniectomy or hypothermia. To assess potential predictors of malignant brain edema by measurement of intracranial pressure (ICP) and microdialysis in patients with large hemispheric stroke and different clinical course. METHODS: In an ongoing prospective clinical study, an ICP and microdialysis probe were placed into the parenchyma of the ipsilateral frontal lobe of 10 patients. Extracellular concentrations of glutamate, lactate, pyruvate, and glycerol were measured continuously. Repeated cranial CT scans were scrutinized for size of infarction and presence of mass effect. RESULTS: The dynamics of the different substances varied in accordance with the clinical course, size of infarction, and local brain edema: Increase in ICP and in glutamate concentration and lactate-pyruvate ratio was followed by massive edema and large infarcts; generally low and stable ICP and substrate concentrations were found in patients without progressive space-occupying infarcts. CONCLUSIONS: In patients with large hemispheric infarction, bedside monitoring with microdialysis is feasible and might be helpful together with ICP recording to follow the development of malignant brain edema.

Magnetic resonance prediction of outcome after thrombolytic treatment.

Treatment of clinical stroke with recombinant tissue plasminogen activator (rt-PA) carries the risk of hemorrhagic complications. Hence, predictors of therapeutic outcome with respect to (a) reperfusion and (b) tissue recovery would be very useful to identify potentially salvageable brain tissue. Magnetic resonance (MR) parameters, especially the apparent diffusion coefficient of water (ADC), perfusion-weighted imaging (PWI) and T(2) relaxometry are thought to provide this information. We evaluated the prognostic implications of ADC, PWI and T(2) relaxometry immediately before initiation of thrombolytic treatment in a model of clot embolism in rats. Animals (n = 14) were treated with intraarterial rt-PA (10 mg/kg) at 90 min after embolism. MR imaging was repeatedly performed at 4.7 T before and up to 5.5 h after embolism. ADC was calculated from diffusion-weighted images (b-values: 30, 765, 1500 s/mm(2)), arterial spin tagging was used for PWI, and quantitative T(2) relaxometry was performed with a Carr-Purcell-Meiboom-Gill (CPMG) sequence. A reperfusion index was calculated to assess the quality of thrombolytic recanalization. The decline of ADC at the end of the experiment to below 80% of control was defined as unfavorable outcome. The probability of tissue injury at the end of the experiments increased with the severity of ADC changes before the initiation of treatment (probability of unfavorable outcome: 21%, 44%, 65% for ADC values of 80-90%, 70-80% and <70% of control, respectively). Pretreatment PWI or T(2) relaxometry also correlated with outcome but-alone or in combination with pretreatment ADC maps-did not improve injury prediction over that obtained by ADC alone. Outcome was influenced positively by successful reperfusion the quality of which, however, could not be predicted by pre-treatment MR characteristics. The data demonstrate that ADC mapping performed before the initiation of thrombolytic treatment provides reliable risk assessment of impeding brain injury but due to uncertainties of postischemic reperfusion does not allow precise outcome prediction in individual experiments.

Treatment with an endothelin type A receptor-antagonist after cardiac arrest and resuscitation improves cerebral hemodynamic and functional recovery in rats.

OBJECTIVE: Successful resuscitation of the brain after cardiac arrest requires unimpaired microcirculatory reperfusion. Postischemic cerebral hypoperfusion presumably is mediated through activation of endothelin type A receptors (ET(A)). The effect of the selective ET(A) antagonist BQ123 on cerebral blood flow and function was studied in a rat model of cardiac arrest. DESIGN: Prospective, randomized trial. SETTING: Experimental animal laboratory. SUBJECTS: Twelve male Sprague-Dawley rats (290-350 g). INTERVENTIONS: Cardiac arrest for 12 mins was induced by electrical fibrillation of the heart, followed by standardized cardiopulmonary resuscitation. BQ123 (0.8 mg/kg; n = 6) or its vehicle (saline; n = 6) was injected intravenously at 15 mins after the return of spontaneous circulation. MEASUREMENTS: Cortical blood flow was measured by laser-Doppler flowmetry, electrophysiological function by recording the amplitude of somatosensory evoked potentials, vascular reactivity by ventilation with 6% CO2, and the functional coupling of blood flow by recording the laser-Doppler flow (LDF) changes during somatosensory stimulation. Hemodynamic and functional cerebral recovery was monitored for 3 hrs after the return of spontaneous circulation. MAIN RESULTS: Forty-five minutes after the return of spontaneous circulation, postischemic hypoperfusion developed in both groups, as reflected by a decrease of the LDF signal to about 60% of the preischemic level. In untreated animals, hypoperfusion persisted throughout the observation time, but in animals receiving BQ123, LDF gradually returned to normal. CO2 reactivity in untreated animals was severely reduced for 2-3 hrs after the onset of recirculation, whereas after BQ123 treatment it returned to normal and after 2 hrs even above normal. The ET(A) antagonist also induced a more rapid recovery of the somatosensory evoked potentials amplitude and of the functional blood flow response to somatosensory stimulation, but these parameters did not recover completely within the observation period. CONCLUSIONS: Application of the ET(A) antagonist BQ123 during the early reperfusion period after cardiac arrest shortens postischemic cerebral hypoperfusion and accelerates the restoration of the cerebrovascular CO2 reactivity and the recovery of electrophysiologic function.

Endothelin type A-antagonist improves long-term neurological recovery after cardiac arrest in rats.

OBJECTIVE: Antagonists of endothelin (ET(A)) receptors improve postischemic hypoperfusion. In this study we investigated whether the selective ET(A)-antagonist BQ123 also improves postischemic functional recovery. STUDY DESIGN: Cardiac arrest of 12 mins duration was induced in rats by electrical fibrillation of the heart, followed by advanced cardiopulmonary resuscitation. BQ123 (0.8 mg/kg; n = 9) or its vehicle (saline; n = 9) was injected intravenously at 15 mins after the return of spontaneous circulation. The neurologic deficit was scored daily for 7 days after resuscitation by rating consciousness, various sensory and motor functions, and coordination tests. On day 7, we measured functional coupling of cerebral blood flow under halothane anesthesia by recording laser-Doppler flow during electrical forepaw stimulation, and we measured vascular reactivity to CO2 by measuring the laser-Doppler flow change during ventilation with 6% CO2. The brains were perfusion-fixated with 4% paraformaldehyde, and the histopathologic damage was evaluated in the CA1 sector of hippocampus, in the motor cortex, and in the cerebellum. RESULTS: Treatment with BQ123 had no effect on histopathologic damage, but it significantly improved neurologic recovery. In all nine treated rats, neurologic performance returned to near normal within 2 days whereas four of nine untreated animals developed spastic paralysis of the hind limbs and severe coordination deficits. BQ123 also normalized CO2 reactivity and improved the functional cerebral blood flow response to somatosensory stimulation. CONCLUSIONS: The ET(A)-antagonist BQ123 significantly improves neurologic outcome after 12 mins of cardiac arrest. The apparent restoration of vascular reactivity demonstrates a correlation between hemodynamic factors and functional recovery.

Probability of metabolic tissue recovery after thrombolytic treatment of experimental stroke: a magnetic resonance spectroscopic imaging study in rat brain.

The effect of thrombolytic therapy on metabolic changes was studied in rats submitted to thromboembolic stroke. Reperfusion was initiated at three different time points, 1.5, 3, and 4.5 hours after embolism (n = 3 each), by injection of recombinant tissue-type plasminogen activator (rt-PA). Recovery was observed during 5 hours of reperfusion using perfusion-weighted images and a two-dimensional 1H magnetic resonance spectroscopic imaging (MRSI) technique. Temporal evolution of the cerebral metabolites lactate and N-acetyl-aspartate (NAA) was determined. To analyze the chances of metabolic tissue recovery, the outcome of treatment, defined by a reversal of lactate concentration, was compared with the lactate intensity before treatment. In untreated animals (n = 4), clot embolism resulted in a drop of perfusion signal intensity in the occluded hemisphere followed by an increase of lactate concentration and a decrease of NAA that persisted throughout the observation period. Thrombolysis partially restored blood flow, but the mean lactate concentration decreased only slightly after successful lysis in animals treated 1.5 hours after embolism. If treatment was initiated later, no decline of lactate level was observed. Five hours after initiation of thrombolysis, the average tissue lactate amounted to 95 +/- 6, 111 +/- 17, and 139 +/- 60% of the early ischemic value (40 minutes after embolization) if treatment began 1.5, 3, and 4.5 hours after embolism, respectively. The NAA level declined slightly but never showed a recovery after rt-PA treatment. In individual pixels, the probability of metabolic tissue recovery clearly declined with increasing lactate concentration before thrombolysis. Interestingly, this probability was independent of treatment delay, but the number of pixels with low lactate declined with increasing ischemia time. Potential clinical applications of MRSI include monitoring of therapeutic intervention as well as support for prognosis of outcome after rt-PA treatment.

Thrombolysis of cerebral clot embolism in rat: effect of treatment delay.

Rats submitted to focal cerebral ischemia by middle cerebral artery clot embolism were treated with recombinant tissue plasminogen activator (rt-PA) at increasing delays (1.5, 3 and 4.5 h) after the onset of ischemia. Treatment efficacy was evaluated by NMR imaging of the apparent diffusion coefficient of water (ADC). In untreated animals the size of the ADC-detectable lesion gradually increased after clot embolism, expanding over 8 h to 174 +/- 17% of the volume visible at 30 min. Thrombolysis initiated 1.5 h after embolism did not reverse the ischemic lesion but reduced its growth to 113 +/- 19% (p < 0.05). Lesion size increased to 135 +/- 14% after 3 h (NS) and to 214 +/- 35% after 4.5 h delay (NS). Thrombolysis with rt-PA attenuates infarct expansion but does not reverse ischemic injury.

Brain hemorrhages after rt-PA treatment of embolic stroke in spontaneously hypertensive rats.

Intracerebral hemorrhage is a major complicating factor of thrombolytic therapy of stroke. To investigate the incidence of bleeding in animals with a diseased vascular system, thrombolysis was carried out in spontaneously hypertensive rats (SHR) after clot embolism of the right middle cerebral artery (MCA). Three hours after embolism SHR were treated with either recombinant tissue-plasminogen activator (rt-PA) or saline, and neurological deficits and intracerebral hemorrhages were evaluated after 3 days survival. Rt-PA-treated SHR exhibited a significantly higher incidence of hemorrhages than untreated rats but neurological deficits and survival rates showed a non-significant trend for improvement. This model offers the possibility of investigating the pathophysiology of post-thrombolytic complications in a clinically relevant small animal model.

Simultaneous recording of evoked potentials and T2*-weighted MR images during somatosensory stimulation of rat.

Somatosensory evoked potentials (SEP) and T2*-weighted nuclear magnetic resonance (NMR) images were recorded simultaneously during somatosensory stimulation of rat to investigate the relationship between electrical activation of the brain tissue and the signal intensity change in functional NMR imaging. Electrical forepaw stimulation was performed in Wistar rats anesthetized with alpha-chloralose. SEPs were recorded with calomel electrodes at stimulation frequencies of 1.5, 3, 4.5, and 6 Hz. At the same time, T2*-weighted imaging was performed, and the signal intensity increase during stimulation was correlated with the mean amplitude of the SEP. Both the stimulation-evoked signal intensity increase in T2*-weighted images and the amplitude of SEPs were dependent on the stimulation frequency, with the largest signals at a stimulation frequency of 1.5 Hz and decreasing activations with increasing frequencies. The feasibility of simultaneous, artifact-free recordings of T2*-weighted NMR images and of evoked potentials is proved. Furthermore, the study demonstrates-in the intact brain-the validity of functional magnetic resonance imaging for estimating the intensity of electrocortical activation.

Brainmapping of alpha-chloralose anesthetized rats with T2*-weighted imaging: distinction between the representation of the forepaw and hindpaw in the somatosensory cortex.

T2*-weighted imaging at 4.7 T was used to identify the cortical areas activated by electrical stimulation of the forepaw and hindpaw of alpha-chloralose anesthetized rats. Variation of the coronal slice position relative to the bregma, showed that the forepaw representation in the somatosensory cortex is more frontal and lateral than that of the hindpaw. Overlap between both activation areas was observed only in a small region in the slice at the level of the bregma. Documented localizations of both representations are in good agreement with earlier observations using invasive techniques. The determination of the separate areas of both paws indicates the feasibility of more complex activation studies in anesthetized animals, such as combined stimulations for the investigation of potentiation or depression effects on individual stimuli.

A reproducible model of middle cerebral artery occlusion in mice: hemodynamic, biochemical, and magnetic resonance imaging.

A reproducible model of thread occlusion of the middle cerebral artery (MCA) was established in C57 Black/6J mice by matching the diameter of the thread to the weight of the animals. For this purpose, threads of different diameter (80 to 260 microns) were inserted into the MCA of animals of different weights (18 to 33 g), and the success of vascular occlusion was evaluated by imaging the ischemic territory on serial brain sections with carbon black. Successful occlusion of the MCA resulted in a linear relationship between body weight and thread diameter (r = 0.46, P < 0.01), allowing precise selection of the appropriate thread size. Laser-Doppler measurements of CBF, neurological scoring, and 2,3,5-triphenyltetrazolium chloride staining confirmed that matching of animal weight and suture diameter produced consistent cerebral infarction. Three hours after MCA occlusion, imaging of ATP, tissue pH, and cerebral protein synthesis allowed differentiation between the central infarct core, in which ATP was depleted, and a peripheral penumbra with reduced protein synthesis and tissue acidosis but preserved ATP content. Perfusion deficits and ischemic tissue alterations could also be detected by perfusion- and diffusion-weighted magnetic resonance imaging, demonstrating the feasibility of dynamic evaluations of infarct evolution. The use of multiparametric imaging techniques in this improved MCA occlusion model opens the way for advanced pathophysiological studies of stroke in gene-manipulated animals.

Induction of protein inhibitor of neuronal nitric oxide synthase/cytoplasmic dynein light chain following cerebral ischemia.

Administration of inhibitors of neuronal nitric oxide synthase or deletion of the encoding gene in rodents provided evidence that neuronal nitric oxide synthase activity may contribute to neuronal cell death following global and focal cerebral ischemia. In the present study, we investigated by in situ hybridization the expression of an endogenous inhibitor of neuronal nitric oxide synthase activity, designated protein inhibitor of neuronal nitric oxide synthase and homologous to cytoplasmic dynein light chain, in the post-ischemic rat brain. Following global ischemia induced by cardiac arrest, messenger RNA expression of protein inhibitor of neuronal nitric oxide synthase was rapidly induced in pyramidal neurons of the hippocampal CA3 region and granule cell of the dentate gyrus which are resistant to ischemic damage. In vulnerable CA1 pyramidal neurons however, protein inhibitor of neuronal nitric oxide synthase expression remained at basal level after global ischemia and was associated with an increase in nicotinamide adenine dinucleotide phosphate-diaphorase activity and subsequent DNA fragmentation indicating ischemia-mediated neuronal cell death. Following focal cerebral ischemia induced by permanent occlusion of the middle cerebral artery, transcripts of protein inhibitor of neuronal nitric oxide synthase progressively accumulated in cortical neurons bordering the infarct area. After transient middle cerebral artery occlusion however, messenger RNA levels of protein inhibitor of neuronal nitric oxide synthase increased in the reperfused neocortex. Our findings indicate that cerebral ischemia leads to an increase in neuronal expression of protein inhibitor of neuronal nitric oxide synthase in brain regions where sustained or "uncoupled" nitric oxide synthase activity may be detrimental to neurons. Lack of post-ischemic induction of protein inhibitor of neuronal nitric oxide synthase in CA1 pyramidal neurons may result in high nitric oxide synthase activity after global ischemia and could contribute to delayed neuronal cell death.

High resolution MRI and MRS: a feasibility study for the investigation of focal cerebral ischemia in mice.

Combined NMR imaging and spectroscopy have been applied to mouse brain during focal cerebral ischemia. The present study evaluated the feasibility of NMR measurements on mice in order to fine-tune the sequences and experimental setup for systematic investigations on stroke including future studies on transgenic animals. The acquisition of high quality diffusion-weighted, perfusion-weighted, and T2-weighted images (DWI, PWI, T2-WI, respectively) is demonstrated and complemented by measurements of 1H volume-selective spectroscopy and spectroscopic imaging (SI). Despite the small volume of the mouse brain, a satisfactory signal-to-noise ratio can be achieved with reasonably short measurement times. C57black/6J mice with an average body weight of 25 g were studied using state-of-the-art NMR sequences at 4.7 T. After induction of focal cerebral ischemia, the lesion was found clearly distinguishable in all imaging techniques. The apparent diffusion coefficient (ADC) was reduced in the ischemic region, and an expansion of the affected volume was observed with ongoing ischemia time. In the H spectra of ischemic animals a distinct change in the concentrations of NAA and lactate was visible. This is the first report on both SI data and perfusion-weighted imaging on mouse brain. It is demonstrated that the perfusion deficit during ischemia can be well demarcated. The spatial resolution of changes in metabolite concentrations allows the clear differentiation of elevated lactate levels in ischemic brain tissue.

Corticotropin-releasing hormone enables weaning from ventilator in a neurosurgical patient under long-term controlled mechanical respiratory support.

A patient with recurrent hemorrhage into the posterior fossa leading to loss of central respiratory drive is described. Repeated i.v. application of corticotropin-releasing hormone over a period of 24 days enabled weaning from long-term controlled mechanical ventilation by reinstitution and maintenance of spontaneous respiration.

Effect of osmotic blood-brain barrier disruption on gentamicin penetration into the cerebrospinal fluid and brains of normal rabbits.

Rapid infusion of hyperosmolar solutions into the internal carotid artery transiently disrupts the blood-brain barrier, permitting entry of substances that are ordinarily excluded from the nervous system. This study compared gentamicin concentrations in the cerebrospinal fluid (CSF) and brain tissue of rabbits receiving intracarotid infusions of 2 molal mannitol with those in three groups of control animals. After catheter placement and intravenous gentamicin administration (20 mg/kg), rabbits received either 2 molal mannitol in the internal carotid artery, 2 molal mannitol intravenously, 0.9% saline in the internal carotid artery, or 0.9% saline intravenously. Mannitol and saline were administered in 8-ml bolus injections over 40 s. After 2 h, serum, CSF, and brain specimens were obtained for antibiotic assay. Gentamicin concentrations in serum were comparable in all groups (mean concentrations ranged from 14.6 to 17.9 micrograms/ml at 60 min and from 5.7 to 7.4 micrograms/ml at 135 min), but gentamicin concentrations in CSF and brains were significantly higher in animals in the group receiving mannitol in the internal carotid artery. In this group the mean gentamicin concentration in CSF, 5.3 micrograms/ml, was twofold greater than those in the other three groups (range, 1.7 to 2.6 micrograms/ml). Similarly, the mean gentamicin concentration in the brains of animals in the group receiving mannitol in the internal carotid artery, 2.3 micrograms/g was significantly higher than those in the other groups (mean of measurable values, 1.4 micrograms/g, in all three control groups). Osmotic disruption of the blood-brain barrier enhanced the penetration of gentamicin into CSF and brain tissue.