DOES A BOLUS OF MANNITOL INITIALLY AGGRAVATE INTRACRANIAL HYPERTENSION?

Transcription

1 Br. J. Anaesth. 1987, 59, DOES A BOLUS OF MANNITOL INITIALLY AGGRAVATE INTRACRANIAL HYPERTENSION? A Study at Various Pa co Tensions in Dogs^ M. ABOU-MADI, D. TROP, N. ABOU-MADI AND P. RAVUSSIN Ever since hyperosmotic agents, and in particular mannitol, were first introduced for clinical use, the mechanism by which they reduce intracranial pressure (ICP), increase cerebral perfusion pressure (CPP), and enhance cerebral blood flow (CBF) has remained controversial (Johnston and Harper, 1973; Leech and Miller, 1974; Borgstrom, Johannsson and Siesjo, 1975; Brown et al., 1979; Dempsey and Kindt, 1982; Kassell et al., 1982; Muizelaar et al., 1983). Recent evidence suggests, however, that apart from their known water-drawing effect, they may operate by improving microcirculatory dynamics and decreasing cerebrovascular resistance (CVR) (Brown et al., 1979; Burke et al., 1981; Kassell et al., 1981, 1982; Muizelaar et al., 1983; Muizelaar, Lutz and Becker, 1984). Rapidly expanding mass lesions and traumatic head injuries cause profound alterations in the vasomotor tone of the cerebral blood vessels, which progress occasionally to complete vasoparalysis (Langfitt, Weinstein and Kassell, 1965; Langfitt, Tannanbaum and Kassell, 1966; Obrist et al., 1984). In association with such impaired vasomotor tone, autoregulation may be affected and cerebral vasodilatation, accompanied by severe hyperaemia, may ensue (Obrist et al., 1984). As a result, CBF and cerebral blood volume (CBV) are passively augmented, thereby increas- MOUNIR ABOU-MADI, M.B. CH.B., D.A., F.R.C.P.(C), D.A.B.A., F.A.CA.; DAVY TROP, M.D., M.SC., F.R.C.P.(C). F.A.C.A.; Department of Neuro-anaesthejia, Montreal Neurological Institute, McGill University, 3801 University Street, Montreal, Quebec, Canada H3A 2B4. NOHA ABOU-MADI, D.V.M., M.SC.; Department of Medical Sciences, University of Florida, Gainesville, Florida, U.S.A. PATRICK RAVUSSIN, M.D., F.M.H.; Service d'anesthesiologie, 1011-Centre Hospitalier Universitaire Vaudois, Lauzanne, Switzerland. Accepted for Publication: September 3, t A brief account of this work was presented to the sixth International Symposium on Intracranial Pressure, June 9-13, 1985, Glasgow. SUMMARY In two groups of anaesthetized dogs, with (n = 28) orwithout (n = 28) induced intracranial hypertension, we compared the effects on intracranial pressure (ICP) of the rapid administration of mannitol 2 g kg- 1 i.v. at Pa COl 2.7, 4.0,, and 6.7kPa (n = 7). In dogs with no induced intracranial hypertension, ICP increased during the administration of mannitol, reached a peak at 2 min after infusion, and then gradually decreased (P<0.05). More marked changes in ICP were observed in response to higher values of Pa COt (P < 0.05). In dogs with induced intracranial hypertension, the rapid infusion of mannitol caused an exponential decrease in ICP, without initial increase, which was significantly steeper at higher values of Pa COt (P < 0.05). This was followed by a more gradual decrease which achieved pre-balloon inflation values 10 min after infusion. We postulate that the absence of the initial increase in ICP is the result of (1) a concomitant decrease in arterial pressure, (2) a reduction in the volume-pressure response of the brain, (3) the failure of mannitol to dilate further the cerebral arterial vascular bed and (4) a hitherto unnoticed early water-drawing effect. Our study confirmed the safety of rapidly expanding the circulating blood volume with mannitol in circumstances of increased ICP in dogs. ing ICP (Miller, 1975; Obrist et al., 1984). Recently, caution has been urged in regard to the use of mannitol during states of increased ICP and lost autoregulation such as occur following traumatic head injury (Jennett and Johnston, 1971; Wilkinson, Wepsic and Austin, 1971; Johnston and Harper, 1973; Cottrell et al., 1977;

2 MANNITOL: EFFECTS ON ICP 631 Bruce, 1983; Cohen, 1983; Wilkinson and Rosenfeld, 1983). The reason for this concern is that, in patients with normal ICP, acute increases in ICP have been noted during and after the administration of hyperosmolar mannitol (Cottrell et al., 1977). Researchers have, therefore, suggested (but not demonstrated) that, in patients in whom ICP is increased, mannitol, by increasing CBV early in the infusion, may in fact transiently aggravate the intracranial hypertension (Johnston and Harper, 1973; Cottrell et al., 1977; Bruce, 1983; Cohen, 1983; Wilkinson and Rosenfeld, 1983). On the other hand, hyperventilation is well known to cause cerebral vasoconstriction, thereby decreasing CBF and CBV, and hence ICP. This effect has been shown to take place even in regions with impaired autoregulation (Paulson, Olesen and Christensem, 1972; Obrist et al., 1984). The present study was undertaken to compare the initial effects of a bolus infusion of mannitol on ICP at various values of Pa COt in two groups of dogs: one in which intracranial hypertension had been induced and one in which ICP was normal. MATERIALS AND METHODS Surgical preparation The design of the study was approved by the Montreal Neurological Institute Animal Care Committee. Sixty-two mongrel dogs weighing kg were studied. Anaesthesia was induced with a single i.v. dose of sodium thiopentone 12 mg kg" 1. A tracheal tube was passed and the lungs ventilated mechanically (Bird mark 7 pressure ventilator; Bird Corp., Palm Springs, Calif., U.S.A.) using compressed air. Ventilation was adjusted to achieve a Pa COl of ±0.27 kpa as measured with an infra-red end-tidal Pco 2 analyser. Anaesthesia was maintained with butorphanol tartrate 0.1 mg kg" 1 i.v. Pancuronium 0.1 mg kg" 1 i.v. was administered and drugs were repeated as required. Rectal temperature was measured and maintained at 38 ± 1 C with a thermal blanket. All surgical sites were infiltrated with a 0.25 % solution of bupivacaine. A 16-gauge catheter was introduced to the cephalic vein to allow for infusion of maintenance fluid (physiological saline solution 4 ml kg" 1 h" 1 ) and drugs. A 20-gauge catheter was placed in the femoral artery, under direct vision, to permit the continuous recording of systemic arterial pressure and the sampling of blood. A 20-gauge catheter was inserted through the atlanto-occipital membrane to the cisterna magna to monitor ICP. Arterial and intracranial pressures were measured using previously calibrated Bentley Trantec Model 800 pressure transducers (Bentley Lab., Irvine, Calif., U.S.A.). Continuous recording of heart rate (HR), mean arterial pressure (MAP), ICP and end-tidal Pco t were obtained on a Siemens Sirecust 300, four-channel recorder (Siemens AG, Erlangen, West Germany). Arterial bloodgas tensions were determined using a Corning 165/2 ph/blood-gas analyser (Corning Medical, Boston, Mass., U.S.A). Programme The animals were divided into three groups. Group A consisted of 28 dogs, subdivided into four subgroups of seven dogs each. Ventilation was readjusted to achieve, in subgroups A 1} A 2J A, and A 4, a Pa COl of 2.7, 4.0, and 6.7 kpa, respectively. Group B consisted of 28 dogs. In all group B dogs the skull was exposed surgically and an appropriate hole made over the parietal lobe for the placement into the extradural space of a 7-French size Fogarty (Edwards Laboratories Inc., Santa Ana, Calif., U.S.A.) venous embolectomy catheter. Group B animals were then subdivided into four subgroups of seven dogs each and their ventilation readjusted to achieve in subgroups Bu Bj, B 3 and B 4 apa COl of 2.7,4.0, and 6.7 kpa, respectively. Group C consisted of six control dogs. As in group B, their skull was surgically exposed and an appropriate hole made over the parietal lobe. Pa COt was maintained at ±0.27kPa. Group A. After a period of stabilization (30 min) baseline measurements of ICP, MAP, and HR were obtained and Pa COt values confirmed by the measurement of arterial blood-gas tensions. Following this, a 2-g kg" 1 dose of a solution of 20% mannitol was infused rapidly over a 5-min period. Data for all animals were collected at 2 min of infusion, at the end of the infusion, and at 2, 5, 10 and 15 min following the cessation of the infusion. Group B. A balloon was inserted to the extradural space and inflated gradually with 3 ml of air to exert brain compression. There was no effect of the distortion of the CSF pathway on the ICP recording secondary to the balloon inflation. ICP increased to peak values and then decreased gradually. Once a stable and sustained increase in ICP was observed over a min period,

3 632 baseline measurements were obtained and mannitol administered as in group A animals. Group C. The extradural balloon was inflated to induce intracranial hypertension. However, mannitol was not infused. Measurements of ICP, MAP and HR were taken every 5 min throughout the study period. Statistical analysis The results of repeated measures and multiple groups were analysed by one-way analysis of variance. Multiple pairwise comparisons between and within subgroups were assessed by a Bonferroni t test (Wallenstein, Zucher and Fleiss, 1980). BRITISH JOURNAL OF ANAESTHESIA P < 0.05 was considered significant. P values are reported only when significant. Data are expressed as the mean ± SEM. RESULTS The three groups of animals behaved similarly up to the inflation of the extradural balloon and the administration of the mannitol. Group A : Dogs with no induced intracranial hypertension The effects of the infusion of mannitol on ICP, MAP, and HR are detailed in tables I III and in figures 1-3. Results showed a significant linear TABLE I. Changes in ICP (mm Hg) (mean± SEM) following infusion of mannitol at different Pa C o, *" d g> with normal ICP (group A (n = 28)). Analysis of variance: P < Multiple pairwise comparisons (Bonferrom): *P < 0.05 within subgroups; fp < 0.05 between subgroups (kpa) Initial ICP Subgroup Baseline 2 min 5 min 2 min 5 min 10 min 15 min 1 (n = 7) 2 («= 7) 3 («= 7) 4 (n = 7) ± ± ± ± ± ± ± ± ± ±0.6t 7.0±0.8f 7.3± ± ± ± ± ±0.4f 10.1±0.5*f 10.3±0.6*+ 9.1±0.5*f 7.3±0.4f 5.4± ± ± ±0.6*t 14.4±0.6*t 13.7±0.6*t 10.4±0.4t 6.4± ± ±0.7f 23.1±0.9*t 23.6±1.1*+ 19.9±0.9*t 14.7±0.9f 12.8±0.6+ s a. o Group A " = 28 /?,. \ I r N ^ ' >a coj Adjusted Mass Baseline 2min Smin 40mmHg Pa^^ lesion > 'SubgroupI (/?a CO2 20mmHg) < 'SubgroupII (Pa CO2 30mmHg) " * Subgroup III (Paco 2 40mmHg) «-- Subgroup IvCPacoj 50 mm Hg) n* 7 in each subgroup After mannitol Infusion I i I 2min Smln 10mln 15min FIG. 1. Mean changes in ICP in groups A and B following mannitol infusion at different values of P*<x> Analysis of variance: P< 0.001; multiple pairwise comparisons (Bonferrom) between A and B: P < 0.05.

4 MANNITOL: EFFECTS ON ICP 633 TABLE II. Changes in mean arterial pressure {MAP) {mm Hg) {mean ± SEM) following infusion of marmiiol at different Pico, *" dogs with normal ICP (group A (n = 28)). Analysis of variance: P < Multiple pairwise comparisons (Bonferroni): *P < 0.05 urithin subgroups; fp < 0.05 between subgroups Sub- P^co, Initial Basegroup (kpa) MAP line 2 min 5 min 2 min 5 min 10 min 15 min ±2.3 (n = 7) ± ± ± ±2.1* 114.3±2.7* 114.7±2.4* 107.0± ± ±2.6 (n = 7) ± ± ± ±2.4* 113.4±2.5* 110.0±2.3* 105.1± ± ±2.7 (n = 7) 101.0± ± ± ±2.6*t 107.6±2.5*t 106.1±2.3*t 102.1± ± ±2.2 (n = 7) ± ± ± ±2.7*f 126.1±2.5*t 121.1±2.3*t 113.8± ± *» 80 E ~ 180 I Group A n*28 Group B n.28 ^ C O j M\\nlwi Mass Baseline 2 min 40 mm Hg Pa C02 lesion Subgroup I (/ 3 a C o 2 2 mm Hfl) "Subgroup II (Pa^30mmHg) 'Subgroup III(Pa CO2 40mmHg) - -'Subgroup I V C P a c c ^ ) n' 7 In each subgroup I i i i i i 5 min 2 min 5 min 10mln 1Smln FIG. 2. Mean changes in mean arterial pressure in groups A and B following mannitol infusion at different values of P*c Ot. Analysis of variance: P < 0.001; multiple pairwise comparisons (Bonferroni) between A and B: P < trend in subgroups A,, A 2, A 3 and A 4. ICP increased during the infusion of mannitol, reached a peak at 2 min after the end of the infusion, and then decreased gradually (table I, fig. 1). Higher values of ICP were observed in response to higher Pa COt values (P < 0.05). Mannitol significantly increased MAP in all animals (P < 0.05) (table II, fig. 2); HR transiently increased in group A 4 (table III, fig. 3). Since the effects of mannitol were maximum at 2 min after the end of the infusion, only results derived from that period of time are discussed. Results were compared with baseline measurements. Group A l (Pa COt 2.7 kpa). Mannitol caused a mean 0.9 mm Hg (14%) increase in ICP. MAP increased by 10 mm Hg (9.6%) (P < 0.05), and HR declined by 7 beat min" 1 (5.9%). Group A 2 (Pa COl 4.0 kpd). Mannitol induced a mean increase of 3.4 mm Hg (49.3%) in ICP (P<0.01). MAP increased by 12.3 mm Hg (12.2%) (P<0.05). HR remained virtually unchanged.

5 634 BRITISH JOURNAL OF ANAESTHESIA TABLE III. Changes in heart rate (HR) (beat min ') (mean±sem) following infusion of mannitol at different Pa co, in dogs with normal ICP (group A (n = 28)). Analysis of variance: P < Multiple pairwise comparisons (Bonferrom): *P < within subgroups; fp < 0.05 between subgroups Sub- P&co, Initial Basegroup (kpa) HR line 2min 5 min 2 min 5 min 10 min 15 min ±4.9 (n = 7) ± ± ± ± ± ± ± ± ±3.6 (n = 7) ± ± ± ±4.6 12± ± ± ± ± ± ± ±3.0t 118.5±3.7f 119.7±3.6f 120.7± ± ± ±3.6 (n = 7) ± ± ±5.1f 144.0±4.8t 132.4±4.3f 131.8± ± ± T c 120 E S Group A n I Group B n.28 - Subgroup I (Paa>2 20 mm Hg) Subgroup II (/'acc^somm Hg) Subgroup III (Pacc^^mm Hg) Subgroup IV (Pacoj 50mm Hg) n=7 in each subgroup I PB C0. Adjusted Mass Baseline 2 min 5min 2 min 5 min 10min 1Smin 40mm Hg p acoj 'es^n FIG. 3. Mean changes in heart rate in groups A and B following mannitol infusion at different values of Pico t. Analysis of variance: P < 0.001; multiple pairwise comparisons (Bonferroni) between A and B:P<0.01. Group A 3 (Pa COt kpa). Mannitol caused a mean 5.7 mm Hg (65.5%) increase in ICP (P <0.001).MAP increased by 6.6 mm Hg(6.5%) (P < 0.05), and HR decreased by a 4 beat min~ l (3.3%). Group A 4 (Pa COt 6.7 kpa). Increasing the P^coi from to 6.7 kpa was associated with a mean increase of 5.2 mm Hg (58.4%) in ICP (P < 0.001) and mannitol increased it further by 9.5 mm Hg (67.4%) (P < 0.001). MAP increased by 12.8 mm Hg (11.3%) <P < 0.05), and HR decreased by 7 beat min (5.2%). The results of multiple pairwise comparisons for ICP differed significantly between the subgroups for up to 10 min after the cessation of the infusion (P < 0.05). Group B: Dogs with induced intracramal hypertension Changes in ICP, MAP, and HR are detailed in tables IV-VI, and in figures 1-3. An immediate decrease in ICP without any initial increase characterized the response of this group of animals to the infusion of mannitol (P < 0.05) (table IV,

6 MANNITOL: EFFECTS ON ICP 635 fig. 1). Initially, exponential decreases which were significantly steeper at the higher Pa COl were observed; these were followed by a more gradual decrease, reaching pre-inflation values 10 min after the discontinuation of the infusion. The systemic haemodynamic response to mannitol consisted of an early, statistically significant, decrease in MAP in subgroups B 3 and B 4 which closely correlated with the early decrease in ICP (P < 0.05) (table V, fig. 2). At 2 min after the infusion, MAP increased in subgroups B, and B 3 and then gradually decreased in all four subgroups (table V,fig.2). HR increased significantly in the four subgroups (P < 0.02) (table VI,fig.3). As the most important changes in ICP, MAP, and HR immediately followed the onset of the mannitol infusion, only the results derived during the infusion (at 2 and 5 min) are discussed. Results were compared with baseline measurements. Group B t (Pa COt 2.7 kpa). The ICP remained TABLE IV. Changes in ICP (mm Hg) (mean ± SEM) following inflation of cxiradural balloon and infusion of mannitol at different Pa co in dogs with maintained intracranial hypertension (group B (n = 28)). Analysis of variance: P < Multiple pairwise comparisons (Bonferroni): *P < 0.05 within subgroups; +P < 0.05 between subgroups (kpa) Initial ICP Mass lesion Subgroup Baseline 2 min 5 min 2 min 5 min 10 min 15 min 1 (n -7) 2 (" -T) 3 (n -7) 4 ( n = 7) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±0.9* 21.4± ±0.7* 23.7±l.l*t 21.1±1.0*t 35.7±1.3*t 29.5±1.4*t 14.8±1.2* 12.8±0.9* 11.5±0.8* 11.2±0.7* 17.1±1.0* 14.5±1.1* 12.2±1.0* 11.5±0.9* 18.0±0.8* 15.2±0.6* 13.4±0.5* 12.7±0.5* 25.2±1.2*f 23.2±l.l*t 21.4±1.2*t 21.0±1.2*t TABLE V. Changes in mean arterial pressure (MAP) (mmhg) (mean ± SEM) following inflation of extradural balloon and infusion of mannitol at different Pa COt in dogs with maintained intracranial hypertension (group B (n = 28)). Analysis of variance: V < Multiple pairwise comparisons (Bonferroni): *P < 0.05 within subgroups; fp < 0.05 between subgroups P*co, (kpa) Initial MAP Mass lesion Subgroup baseline 2 min 5 min 2 min 5 min 10 min 15 min ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±1.5f 117.7±1.7f 116.2±1.9f 112.5± ± ±3.1f 126.4±3.1f 130.7±2.9f 120.7±4.1* 109.0±3.2* 99.4±2.0* 136.4±2.5*t 136.0±2.1*t 141.7±2.0t ±2.3*t 114.7±2.3* 111.0±3.1* 150.4±3.0*t 148.7±2.7*f 149.5±2.9*f 138.7±2.2*t ±1.8*t 129.2±2.6*t TABLE VI. Changes in heart rate(hr) (beat min' 1 ) (mean ± SEM) following inflation of extradural balloon and infusion of mannitol at different Pa co in dogs with maintained intracranial hypertension (group B (n = 28)). Analysis of variance: P < Multiple pairwise comparisons (Bonferroni): *P < 0.05 within subgroups; fp < 0.05 between subgroups Paco, (kp«) Initial HR Man lesion Subgroup Baseline 2 min 5 min 2 min 5 min 10 min 15 min ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±4.9* 182.4±5.9* 183.0±5.5* ±5.2* 184.2±4.2* 170.5± ±4.9* 181.2±4.6* 183.2±4.3* 184.5±5.1* 181.8±3.6* 161.4±3.0*t 177.0±4.0*t 185.0±3.4*f 188.4±2.8*t ±2.6*t 190.5±2.9*t 147.2± ±4.7*t 160.4±4.3*t ±5.0*t 153.7±5.0*t 150.8±4.8f

7 636 BRITISH JOURNAL OF ANAESTHESIA TABLE VII. Mean (±SEM) changes in ICP, MAP, and HR following inflation of extradural balloon in control dogs with maintained intracranial hypertension (group C (n = 6)). Multiple pairwise comparisons (Bonferroni): no significant differences within the group Initial value Mass lesion Baseline 5min 10 min 15 min 20 min ICP 9.3± ± ± ± ± ± ±1.8 (mmhg) MAP 88.6± ± ± ± ± ± ±2.6 (mmhg) HR 102.2±4.1 16O.8± ± ± ± ± ±3.7 (beat min" 1 ) unchanged during the first 2 min of infusion. At 5 min ICP decreased by a mean of 2.9 mm Hg (14.2%) (P < 0.05); this decrease was not associated with any significant change in MAP. HR increased at the end of the infusion by 27 beat min- 1 (17.3%) (P < 0.02). Group B t (Pa COt 4.0 kpa). At 2 and 5 min ICP decreased by means of 1.7 mmhg (7.4%) and 4 mm Hg (17.3 %) (P < 0.05), MAP decreased by 4.2 mm Hg (3.2%) and 4.1 mm Hg (3.1 %), and HR increased by 14 beat min- 1 (9.1 %) and 25 beat min" 1 (16%) (P < 0.02), respectively. Group B 3 (Pa COl kpa). At 2 and 5 min ICP decreased by a mean of 3.8 mm Hg (13.8%) (P<0.05) and 6.4 mm Hg (23.2%) (P < 0.01), MAP decreased by 6.1 mm Hg (4.3 %) (P < 0.05) and 6.5 mmhg (4.5%) (P < 0.05), and HR increased by 19 beat min" 1 (13.1%) (P < 0.01) and 34 beat min" 1 (24%) (P < 0.01), respectively. Group B t (Pa COt 6.7 kpa). Marked decreases in ICP and MAP characterized the onset of the mannitol infusion in this subgroup. At 2 and 5 min, ICP decreased by a mean of 5.7 mm Hg (13.8%) (P<0.02) and 11.9 mmhg (28.8%) (P< 0.001), MAP decreased by 6.6 mm Hg (4.2%) (P<0.02) and 8.3 mm Hg (%) (P < 0.01), and HR increased by 11 beat min" 1 (8.7%) and 23 beat min" 1 (16.7%) (P < 0.01), respectively. The results of multiple pairwise comparisons for ICP were not significant between subgroups Bj and B,, but differed significantly between B s and B 3 (P < 0.05) and between B 3 and B 4 (P < 0.001). Comparing pre-and post-infusion differences in ICP between groups A and B at corresponding values of Pa COl revealed consistently a high level of significance (P < 0.01). Control group At baseline measurements, the untreated control dogs were comparable to group B 3 (Pa COl kpa) and showed minimal fluctuations in ICP (1-2 mm Hg) over the subsequent 20-min study period (table VII). DISCUSSION The dose of mannitol (2 g kg" 1 ) used in this study approximates to the amount recommended clinically for the treatment of head injuries in dogs (Kirk and Bistner, 1981). The use of a small control group was beneficial in that we were able to demonstrate prolonged and sustained stability in ICP, following the induction of the intracranial hypertension at Pac Ot 2.7, 4.0, and 6.7 kpa. Along with the fact that other authors have used similar models to study increased ICP (Sullivan, Miller and Searle, 1980; Dempsey and Kindt, 1982; Wilkinson and Rosenfeld, 1983), this gives additional support to the validity of our method. All animals studied showed a stable ICP for min before the mannitol was administered. Thus we were convinced that a reasonable steady-state should have resulted during that period and that the changes in ICP that followed the administration of mannitol were not influenced by haemodynamic events attendent upon the inflation of the balloon. Many authors have written extensively on the relationship between the rapid administration of mannitol and the possible early aggravation of intracranial hypertension (Jennett and Johnston, 1971; Cottrell et al., 1977; Bruce, 1983; Cohen, 1983; Wilkinson and Rosenfeld, 1983). In the present study, we found important differences in the effect on ICP of the rapid administration of mannitol in dogs with induced intracranial hypertension and in those without. Cottrell and colleagues {1977) compared the ICP-reducing effect of frusemide and mannitol in man, and demonstrated a significant initial increase in ICP in the group of patients to whom mannitol

8 MANNITOL: EFFECTS ON ICP g kg * was administered. Since no patient had pre-operative clinical evidence of increased ICP, they postulated that, in addition to its effect on CBF, mannitol might have direct vasodilating properties which could also lead to increases in CBV and pressure. Our data from group A animals confirm these findings and allow us to make the following conclusions: (1) The cerebral vasoconstriction induced by hyperventilation attenuated the magnitude of the increases in ICP (groups A x, A^). (2) Although there were significant increases in ICP, these never exceeded normal ICP values (15 mm Hg) in the normo- and hypocarbic dogs. (3) Hypercarbia increased significantly the induced increase in ICP. This may be explained by the fact that the increase in Pa COt dilates the cerebral vasculature, increases CBV, increases saggital sinus pressure, and alters the mechanical properties of tissues around the CSF spaces (Sullivan, Miller and Searle, 1980). As a result, spatial compensation by means of CSF translocation cannot occur, any replaceable intracranial volume becomes greatly diminished, and the ability of the brain to compensate for additional augmentation in cerebral vasodilatation or CBV becomes limited. The major finding of the present study was that, in experimental hypertension, bolus use of mannitol decreased ICP readily, did not cause any initial increase in pressure, and was devoid of intracranial side effects at all values of P& COl - Previous studies of the effect of mannitol on CBF in situations of increased ICP have given conflicting results (Johnston and Harper, 1973; Shapiro, 1975; Brown et al., 1979; Dempsey and Kindt, 1982; Muizelaar et al., 1983; Muizelaar, Lutz and Becker, 1984). Although our study does not address this issue specifically, careful consideration must be given to the impact on the clinical management of patients of the warnings by many authors concerning the effect of the rapid infusion of mannitol to patients with increased ICP (Jennett and Johnston, 1971; Cottrell et al., 1977; Bruce, 1983; Cohen, 1983; Wilkinson and Rosenfeld, 1983). Our results in dogs with intracranial hypertension showed an immediate decrease in ICP in response to the infusion of mannitol. Two separate processes, mediated by two different mechanisms, seem to be involved. First, we noted that a significant decrease in MAP preceded the early marked decrease in ICP. This may have been caused by the well documented vasodilator properties of hypertonic solutions (Gazitua et al., 1971; Atkins, Wildenthal and Horwitz, 1973; Cote, Greenhow and Marshall, 1979; Krishnamurty, Adams and Willerson, 1979; Stiff, Munch and Bromberger-Barnea, 1979; Findlay et al., 1981). Indeed, the vasodilatation in skeletal muscles, the transient hypotension and the marked shift in the distribution of the cardiac output (Cote, Greenhow and Marshall 1979) may have prevented the increase in CBF and CBV required to produce an increase in ICP. In addition, if the redistribution of cardiac output transiently reduces CBF and CBV, ICP will also be decreased (Davis and Sundt, 1980). In dogs with normal ICP, the injection of mannitol resulted in an increase in arterial pressure, while in the dogs in which ICP was increased, mannitol caused a decrease. The rapid administration of mannitol increases cardiac output (Miller and Sullivan, 1979) and decreases peripheral resistance (Willerson et al., 1975; Cote, Greenhow and Marshall, 1979; Stiff, Munch and Bromberger- Bernea, 1979). As MAP is determined by changes in cardiac output and peripheral resistance, and as profound increases in peripheral resistance have been observed immediately following experimental brain injury (Langfitt, Tannanbaum and Kassell, 1966), we can, therefore, speculate that in the dogs with normal ICP the effect of mannitol on cardiac output predominated and increased MAP, whereas in the dogs with increased ICP the effect of mannitol on the peripheral resistance overrode its effect on cardiac output and decreased MAP. This hypothesis is presently under investigation in our laboratories. Following this initial decrease, ICP continued to decrease gradually. There are three possible reasons for this further decrease. (1) A reduced volume pressure response may have permitted more blood volume to be readily accepted by the brain. Indeed, Leech and Miller (1974), in their study of the effects of mannitol on intracranial volume pressure relationships during experimental brain compression in primates, demonstrated that mannitol reduced the volume-pressure response considerably more than it reduced ventricular fluid pressure (VFP), suggesting that the shape of the volume-pressure curve was altered, thus permitting the intracranial contents to accept a given volume addition more readily. Since this effect appeared earlier and lasted longer than the decrease in VFP, Leech and Miller (1974) emphasized the potential benefit of the administration of mannitol in circumstances of increased

9 638 BRITISH JOURNAL OF ANAESTHESIA ICP, particularly where the volume-pressure curve is at or near the critical, steeply-increasing stage. (2) In the presence of a reduced cerebral vasomotor tone (Langfitt, Weinstein and Kassell, 1965; Langfitt, Tannanbaum and Kassell, 1966; Obrist, 1984), mannitol may have failed to dilate further the cerebral arterial vascular bed (Johnston and Harper, 1973). (3) In the face of an altered primary vascular response, an early water-drawing effect of the mannitol may have been unmasked (Pappius and Dayes, 1965). The usefulness of hyperventilation as an adjunct to osmotic diuresis in the treatment of intracranial hypertension is well established (Overguard and Tweed, 1974; Enevoldsen and Jensen, 1978; Obrist et al., 1984). However, of particular interest to our study is the finding by Paulson, Olesen and Christensem (1972) that hypocapnia restores CBF autoregulation in patients with cerebral hyperaemia. This phenomenon may explain the probable recovery of cerebral vasomotor tone in our hyperventilated dogs, which allowed mannitol to dilate directly the cerebral arterial vascular bed independently of changes in cardiac output and MAP, thereby reducing the steepness of the ICP reduction curve (fig. 1). In conclusion, the commonly held idea of the danger of rapidly infusing mannitol in the presence of intracranial hypertension must be reconsidered. Our animal data validate the safety of this therapeutic procedure and discount the need for any prior manoeuvre to decrease brain size. ACKNOWLEDGEMENTS The authors are grateful to Dr Victoria Lees and Ms Franca Sauro for their editorial assistance, to Mr Charles Hodge and his group for their photographic work, and to Mrs Dalila Arruda and Ms Angie Giannakopoulos for their secretarial help. REFERENCES Atkins, J. M., WUdenthal, K., and Horwitz, L. D. 1973). Cardiovascular responses to hyperosmotic mannitol in anesthetized and conscious dogs. Am. J. Pkysiol., 225, 132. Borgstrom, L., Johannsson, J., and Siesjo, B. K. (1975). The influence of acute normovolemic anemia on cerebral blood flow and oxygen consumption of anesthetized rats. Acta Physiol. Scand., 93, 505. Brown, F. D., Johns, L., Jafar, J. J., Crockard, H. A., and Mullan, S. (1979). Detailed monitoring of the effects of mannitol following experimental head injury. J. Neurosurg., 50,423. Bruce, D. (1983). Handbook of Neuroanesthesia: Clinical and Physiologic Essentials, 1st Edn, p Boston: Little, Brown. Burke, A. M., Quest, D. O., Chien, S., and Cerri, C. (1981). The effect of mannitol on blood viscosity. J. Neurosurg., 55, 550. Cohen, N. H. (1983). Handbook of Neuroanesthesia: Clinical and Physiologic Essentials, 1st Edn, p Boston: Little, Brown. Cote, C. J., Greenhow, D. E., and Marshall, B. E. (1979). The hypotensive response to rapid intravenous administration of hypertonic solutions in man and in the rabbit. Anesthesiology, 50,30. Cornell, J. E., Robustelli, A., Post, K., and Turndorf, H. (1977). Furosemide and mannitol-induced changes in intracranial pressure and scrum osmolality and electrolytes. Anesthesiology, 47, 28. Davis, D. H., and Sundt, T. M. jr (1980). Relationship of cerebral blood flow to cardiac output, mean arterial pressure, blood volume, and alpha and beta blockade in cats. J. Neurosurg., 52, 745. Dempsey, R. J., and Kindt, G. W. (1982). Experimental augmentation of cerebral blood flow by mannitol in epidural intracranial masses. J. Trauma, 22, 449. Enevoldsen, E. M., and Jensen, F. T. (1978). Autoregulation and CO, responses of cerebral blood flow in patients with acute severe head injury. J. Neurosurg., 48, 689. Findlay, S. R., Dvorac, A. M., Kagey-Sobota, A., and Lichtenstein, L. M. (1981). Hyperosmolar triggering of histamine release from human basophils. J. Clin. Invest., 67, Gazitua, S., Scott, J. B., Swindall, B., and Haddy, F. J. (1971). Resistance responses to local changes in plasma osmolality in three vascular beds. Am. J. Physiol., 220, 384. Jennett, W. B., and Johnston, I. H. (1971). Brain Hypoxia, p London: Heinemann. Johnston, I. H., and Harper, A. M. (1973). The effect of mannitol on cerebral blood flow. An experimental study. J. Neurosurg., 38, 461. Kassell, N. F., Baumann, K. W., Hitchon, P. W., Gerk, M. K., Hill, T. R., and Sokoll, M. D. (1981). Influence of a continuous high dose infusion of mannitol on cerebral blood flow in normal dogs. Neurosurgery, 9, 283. (1982). The effects of high dose mannitol on cerebral blood flow in dogs with normal intracranial pressure. Stroke, 13, 59. Kirk, R. W., and Bistner, S. I. (1981). Handbook of Veterinary Procedures and Emergency Treatment, 3rd Edn, p Philadelphia: Saunders Company. Kristnamurty, V. S. R., Adams, H. R., and Willerson, J. T. (1979). Paradoxical inhibition of vasoconstrictor and vasodilator responses by hypertonic mannitol in isolated arterial smooth muscle. Eur. J. Pharmacol., 58, 379. Langfitt, T. W., Tannanbaum, H. M., and Kassell, N. F. (1966). The etiology of acute brain swelling following experimental head injury. J. Neurosurg., 24, 47. Weinstein, J. D., and Kassell, N. F. (1965). Cerebral vasomotor paralysis produced by intracranial hypertension. Neurology, (Mirmeap.), 15, 622. Leech, P., and Miller, J. D. (1974). Intracranial volumepressure relationships during experimental brain compression in primates. 3. Effects of mannitol and hyperventilation. J. Neurol. Neurosurg. Psychiatr., 37, Miller, J. D. (1975). Volume and pressure in the craniospinal axis. Clin. Neurosurg., 22, 76. Sullivan, H. G. (1979). International Anesthesia Clinics. Management of Acute Intracranial Disasters, Vol. 17, Nos 2 and 3, p 61. Boston: Little, Brown.

10 MANNITOL: EFFECTS ON ICP 639 Muizelaar, J. P., Lutz, H. A. in, and Becker, D. P. (1984). Effects of mannitol on ICP and CBF and correlation with pressure autoregulation in severely head-injured patients. J. Neurosurg., 61, 700. Wei, E. P., Kontos, H. A., and Becker, D. P. (1983). Mannitol causes compensatory cerebral vasoconstriction and vasodilation in response to blood viscosity changes. J. Neurosurg., 59, 822. Obrist, W. D., Langfitt, T. W., Jaggi, J. L., Cruz, J., and Gennarelli, T. A. (1984). Cerebral blood flow and metabolism in comatose patients with acute head injury. Relation to intracranial hypertension. J. Neurosurg., 61, 241. Overguard, J., and Tweed, W. A. (1974). Cerebral circulation after head injury. Part I: Cerebral blood flow and its regulation after closed head injury with emphasis on clinical correlations. J. Neurosurg., 41, 531. Pappius, H. M., and Dayes, L. A. (1965). Hypertonic urea. Its effect on the distribution of water and electrolytes in normal and edematous brain tissues. Arch. Neurol., 13, 395. Paulson, O. B., Olesen, J., and Christensem, M. S. (1972). Restoration of autoregulation of cerebral blood flow by hypocapnia. Neurology, 11, 286. Shapiro, H. M. (1975). Intracranial hypertension: therapeutic and anesthetic considerations. (Review article.) Anesthesiology, 43, 445. Stiff, J. L., Munch, D. F., and Bromberger-Barnea, B. (1979). Hypotension and respiratory distress caused by rapid infusion of mannitol or hypertonic saline. Aneslh. Analg., 58, 42. Sullivan, H. G., Miller, J. D., and Scarle, J. R. (1980). An interpretation of pressure/volume interaction in the craniospinal axis. Neurosurgery, 6, 453. Wallenstein, S., Zucher, C. L., and Fleiss, J. L. (1980). Some statistical methods useful in circulation research. Circ. Res., 47, 1. Wilkinson, H. A., and Rosenfeld, S. (1983). Furosemidc and mannitol in the treatment of acute experimental intracranial hypertension. Neurosurgery, 12, 405. Wepsic, J. G., and Austin, G. (1971). Diuretic synergy in the treatment of acute experimental cerebral edema. J. Neurosurg., 34, 203. Willerson, J. T., Curry, G. C, Atkins, J. M., Parkey, R., and Horwitz, L. D. (1975). Influence of hypertonic mannitol on ventricular performance and coronary blood flow in patients. Circulation, 51, 1095.