Pco 2 reactivity in childhood strangulation

CBF and CBF/Pco2 Reactivity in Childhood Strangulation S t e p h e n A s h w a l , M D * , R o n a l d M. P e r k i n , MD+, J o s e p h R. T h o m p s o n , MD*, L a w r e n c e G. T o m a s i , M D * , D a v e d van Stralen, M D t , a n d S a n f o r d Schneider, M D *

Four children with self-inflicted strangulation injuries had cerebral blood flow determined by stable xenon computed tomography (XeCTCBF) within 24 hours of admission. All had suffered a severe hypoxic-ischemic cerebral injury; 3 initially had fixed pupils, all were apneic with varying bradyarrhythmias, and the initial mean arterial pH was 7.26 (+ 0.18). The initial blood glucose values were > 300 mg/dl (334 and 351 mg/dl) in the 2 patients who died compared to the 2 who survived (104 and 295 mg/dl). The cardiac index was depressed during the first several days of hospitalization in the 2 patients who died (< 2.0 L/min/m 2) compared to the 2 who survived. Total CBF was normal (63 + 8 ml/min/ 100 gm) and local variations in CBF were present. Pco2 reactivity was determined by hyperventilating the 4 patients for 20 min from an end tidal Pco2 of 39 _+3 torr to 29 _+ 1 torr and then repeating the XeCTCBF study. Marked regional variability in the CBF/PCO2 response was observed, ranging from 0.5-5.5 ml/min/ 100 gm/torr Pco2. In the 2 patients who died, the CBF/Pco2 was decreased (1.2 ml/min/100 gm/torr Pco2) compared to the 2 patients who survived (2.1 ml/min/100 gm/torr Pco2). Although CBF was normal in these 4 children, the hyperventilation response was depressed, variable, and even paradoxical which may be important in the evolution of further brain injury and is a critical factor in deciding whether hyperventilation may be of clinical benefit. Three factors correlated with poor outcome and ultimately death in 2 patients: initial blood glucose > 300 mg/dl; CBF/Pco2 response < 2.0 ml/min/100 gm/torr Pco2; and cardiac index < 2.0 L/min/m2 for longer than 12-24 hours.

child abuse in the younger pediatric population and suicide in older children and adolescents [1,2]. Survivors of such acute insults will frequently have serious, persistent neurologic deficits. Cerebral congestion due to jugular venous compression and cerebral infarction due to carotid artery occlusion are common mechanisms of injury. Increased intracranial pressure (ICP) due to cerebral edema has been observed in 36% of hospitalized children suffering strangulation injuries and is associated with a mortality of 75% [3]. In that study, the delayed onset of cerebral herniation was reported in an additional 27% of children and was associated with a 67% mortality and 33% serious neurologic morbidity. We recently studied total and local CBF in 4 children with self-inflicted strangulation to determine whether reduced CBF in these patients could account for their global cerebral deficits. Because hyperventilation is commonly used to treat increased ICP in these patients we also examined the changes in CBF after hyperventilation to determine whether this remains an effective mode of therapy. In addition, because the previous study of pediatric near-drowning patients by Ashwal et al. demonstrated a correlation between initial blood glucose values greater than 300 mg/dl and poor outcome [4], we examined this variable in these 4 children. Other recent studies examined the relationship between poor cardiac output and survival in the post-cardiac arrest pediatric patient and this was also investigated in our study [5,6].

Methods

Intentional childhood strangulation is a devastating and usually unexpected form of pediatric trauma that is due to

Stable xenon computed tomographic cerebral blood flow (XeCTCBF) determinations were performed in 4 children with self-inflicted strangulation admitted to the pediatric intensive care unit (ICU) at Loma Linda University Medical Center. XeCTCBF scans were obtained within the first 24 hours after admission using the General Electric 9800 Xenon Blood Flow System with the patient breathing a 33% nonradioactive xenon-oxygen mixture as previously reported [4,7]. Three EMI-plane flow maps were spaced 1-2 cm apart to allow coverage of the upper brainstem, subcortical nuclei, and peripheral cortex in the lower and mid cerebrum. Flow values (ml/min/100 gin) were determined with the region-of-interest cursor for both hemispheres on each of the 3 sections; these flows then were averaged to obtain an approximate total CBF. In addition, the region-of-interest cursor was used

From the Department of Pediatrics, Divisions of *Child Neurology and tCritical Care; *Department of Radiation Sciences, Division of Neuroradiology; Loma Linda University School of Medicine; Loma Linda, California.

Communications should be addressed to: Dr. Ashwal; Department of Pediatrics; Loma Linda University School of Medicine; Loma Linda, CA 92350. Received January 7, 1991; accepted March 26, 1991.

Ashwal S, Perkin RM, Thompson JR, Tomasi LG, van Stralen D, Schneider S. CBF and CBF/Pco2 reactivity in childhood strangulation. Pediatr Neurol 1991;7:369-74.

Introduction

Ashwal et al: CBF and Childhood Strangulation

369

to measure local flows for the following brain regions: frontal cortex and white matter, temporal, parietal, and occipital cortex, thalamus, caudate nucleus, lentiform nucleus, midbrain, and cerebellar cortex. Flows through the medulla and pons were not obtained because of unreliable blood flow maps due to Hounsfield artifacts from the cranial bones of the skull base. Normal CBF values using the XeCTCBF method have not been well established in children; however, based on our experience and that of other investigators, total CBF ranged from 40-80 ml/min/100 gm [4,7]. All patients were paralyzed with vecuronium (Norcuron®) and had 2 sequential XeCTCBF studies, initially at a carbon dioxide tension of 39 _+3 torr and then immediately after 20 min of hyperventilation with a Pco2 of 29 + 1 torr. When the second XeCTCBF study was performed, identical planes of section could be obtained because these patients were paralyzed and ventilator-dependent. The change in CBF per tort CO2 (ml/min/100 gm/torr Pco2) was calculated for each of the patients and then averaged. Data are reported as the mean _+S.D. A thermodilution cardiac output catheter was placed in all patients to monitor continuously the cardiac index (CI).

Results Patient 1. An 11-year-old boy was found hanging by a rope in his room. His mother immediately initiated cardiopulmonary resuscitation (CPR) and he was found by the paramedics to be apneic with palpable pulses and a low heart rate. On arrival in the outlying emergency room the child had a pulse of 68 beats per min, blood pressure 130/86 m m Hg, spontaneous respirations, and Glasgow C o m a Scale (GCS) of 4-5. The first arterial blood gas revealed a pH of 7.30, Pco2 38 torr, and Po2 538 torr. Initial serum glucose was 104 mg/dl. He was intubated, transported and, upon arrival to the ICU, 3 hours after discovery, had spontaneous respirations, reactive pupils, and decorticate posturing with stimulation. His initial CI was 4.23 L/min/m 2. This finding was associated with a normal blood pressure and normal systemic vascular resistance. All subsequent measurements of CI remained in the normal range. The child required no inotropic or vasoactive medications. He was subsequently transferred to a pediatric rehabilitation service because of severe neurologic deficits. He had myoclonic seizures, was nonambulatory, could only speak isolated, single words and short p h r a s e s , and had axial h y p o t o n i a with g e n e r a l i z e d spasticity. Patient 2. A 12-year-old boy was found hanging by a rope in his room. Initial CPR was delayed but eventually begun by a neighbor. Paramedics arrived and found the child to be apneic and pulseless but with a heart rate of 20/min. They obtained intravenous access, administered epinephrine, and began ventilation and closed chest compressions. On arrival to an outlying emergency room the child was flaccid, unresponsive with fixed, dilated pupils, had a blood pressure o f 70 m m Hg, and a heart rate of 123/min. He was immediately intubated and his initial arterial blood gas demonstrated a pH of 7.18, P c o 2 51 torr, and Po2 42 tort. Initial serum glucose level was 334 mg/dl. Later examination in the ICU revealed that the child was in myoclonic status. Blood pressure was 120/76 m m Hg and his capillary refill was rapid. The patient's CI remained low (range: 1.8-2.9 L/min/m 2) during the entire

370 PEDIATRIC NEUROLOGY Vol. 7 No. 5

hospitalization. This low CI was associated with an elevated systemic vascular resistance and an elevated arterial blood pressure, but no evidence of metabolic acidosis and normal lactate levels. Efforts to increase cardiac output were unsuccessful. The child's neurologic status did nol improve and on the third day of hospitalization he was deemed clinically brain dead with an isoelectric electroencephalogram (EEG). The following evening he suffered a cardiac arrest and was not resuscitated. Patient 3. A 13-year-old girl was found hanging from an electrical cord in her garage. CPR was initiated at the scene and when the paramedics arrived they Ibund her apneic and pulseless but with a heart rate of 30/min. She was intubated and given epinephrine and atropine: on arrival to an outlying emergency room, heart rate was 120/min, blood pressure 84/40 m m Hg, and there were no spontaneous respirations. Her pupils were fixed and dilated and the GCS was 3. The initial arterial blood gas on 100% oxygen revealed a pH of 7.07, Pco2 37 tort, and PO2 218 torr. The initial serum glucose concentration was 351 mg/dl. Shortly after arrival to the ICU, she developed seizures and was treated with intravenous phenytoin. An ICP monitor was placed 12 hours after admission. Although the initial ICP was 14 mm Hg, during the first 24 hours of hospitalization, the ICP was very difficult to control and reached levels as high as 40 mm Hg, despite mannitol and hyperventilation. Phenobarbital (4 mg/kg) was administered prior to X e C T C B F study and afterward was followed by a continuous infusion of pentobarbitat. Her initial CI was only 2.0 L/min/m 2. This low CI was associated with elevated arterial blood pressure and elevated systemic vascular resistance. Efforts to improve her CI during the first 3-4 days of hospitalization included fluid infusion, inotropic support, and careful vasodilation, but it remained in the range of 1.8-2.1 L/min/m 2 without evidence of metabolic acidosis or elevated serum lactate levels. Despite medical efforts, her ICP continued to increase and reached a peak of 50-60 cm H20. By the sixth hospital day, she had no brainstem function, remained apneic, flaccid, comatose, had an isoelectric EEG, and was declared brain dead. Patient 4. A 9-year-old boy with a known history of depression was found hanging by a cord from a stairway railing. His aunt cut him loose and began CPR but he was found asystolic and pulseless with fixed dilated pupils by the paramedics. He was intubated and given intravenous epinephrine and atropine. In the emergency room, his blood pressure was 72/40 and heart rate 100/min but he was flaccid, had weak spontaneous respirations (although he required respirator assistance), and pinpoint, poorly reactive pupils. Initial arterial blood gas revealed a pH of 7.41, Pco2 14 ton', and Po2 590 ton'. Initial serum glucose was 295 mg/dl. Seizures were treated with intravenous phenytoin and his ICP was monitored and remained below 20 cm H20. His initial CI was low, 2.4 L/min/m 2, but after treatment with dobutamine, the CI increased to 3.26 L/min/m 2 within 6-8 hours. X e C T C B F study was per-

Table 1. Representative hemodynamic data at the time of blood flow XeCTCBF study

Heart Rate

SBP (S/D/M)

PBP (S/D/M)

Cardiac Index (L/min/m2)

CVP (mm Hg)

PCP (mm Hg)

SVRI (dynetisec/ cmS/m2)

PVRI (dynetisec/ cmS/m2)

Comments

Patient 1

108

134/79/95

18/12/15

4.23

3

6

1,737

170

Dobutamine 8 ~tg/kg/min

Patient

158

140/86/98

31/20/26

1.97

12

14

3,407

568

Dobutamine, dopamine (low dose); lactate 1.1 mM/L (0.3-1.3 mM/L)

135

124/86/96

20/8/14

2.02

4

10

3,401

395

Dobutamine, dopamine, nitroprusside, nembutal, lactate 0.9 mM/L (0.3-1.3 raM/L)

104

152/97/110 41/26/32

2.4

16

23

3,160

300

Dobutamine

2

Patient

3

Patient 4

Abbreviations: CVP = Central venous pressure PBP = Pulmonaryblood pressure PCP = Pulmonarycapillary pressure PVRI = Pulmonaryvascular resistance index

SBP = Systolic blood pressure S/D/M = Systolic/diastolic/mean SVRI = Systemic vascular resistance index

formed on the day of admission. EEG disclosed an alpha coma pattern. He remained comatose for the first 2 weeks and then evolved into a vegetative state. Eight months after admission he remains unchanged and has had a gastrostomy and tracheostomy. Discussion

All of these children had suffered severe global hypoxicischemic cerebral injuries. Initial arterial blood gases demonstrated a pH of 7.26 (+ 0.18), PO2 347 (+ 262 torr), and PCO2 35 (+ 15 tort). All were comatose on admission, 3 initially had fixed, dilated pupils, and all had depressed or absent corneal reflexes and impaired oculocephalic responses. All 4 patients were apneic or had insufficient respiratory effort and required intubation and assisted ventilation and had varying bradyarrhythmias that required inotropic support. The initial blood glucose values were greater than 300 mg/dl (334 and 351 mg/dl) in the 2 patients who died compared to 104 and 295 mg/dl in the 2 surviving patients. The initial elevation of blood glucose was unrelated to the amount of inotropic support, use of corticosteroids, or intravenous administration of dextrose. These elevated levels returned to the normal range within 24-48 hours. Sequential measurements of CI were obtained in all 4 patients after insertion of the thermodilution pulmonary catheter (Table 1). The CI was markedly decreased during the first several days of hospitalization in the 2 patients who died, averaging less than 2.0 L/min/m 2. In the 2 surviving children, one had a normal CI (4.2 L/min/m 2) and in the second patient, the CI returned to normal (3.75) within the first 8-16 hours of admission. Normal values for

children of this age at our institution range between 3.55.5 L/min/m 2 [8]. Total and regional pre- and post-hyperventilation CBF values are s u m m a r i z e d in Table 2. At the time o f XeCTCBF study, mean arterial blood pressure (MABP) was 96 (+ 11 mm Hg), pH 7.44 (+ .03), Poe 152 (+ 66 torr), and end tidal Pco2 25 (_+ 3 torr) before and 29 (+ 1 tort) after hyperventilation. Prior to the time of study, Patient 3 had received phenobarbital followed afterward by pentobarbital. The average CBF was normal (63 + 8 ml/min/100 gm) and in 1 patient there were brain regions with flows at the lower limits of normal. Local CBF values were similar in magnitude to those previously reported with similar flows in the different cortical and subcortical regions [4,7]. The Pco2 reactivity for each brain region in each of the 4 patients is summarized and averaged in Table 3. By convention, the expected decrease in CBF with hyperventilation is reported as positive, whereas negative values represent a paradoxical response (i.e., increased CBF with hyperventilation). The CBF/Pco2 response was depressed in many regions of the brain and the degree of reduction was quite heterogeneous, ranging from 0.5-4.9. In the 2 patients who died the change in CBF/PCo2 was decreased (1.2 ml/min/1 O0 gm]Pc02)compared to the 2 patients who survived (2.1 ml]min/lO0 gm/Pco2). In 3 patients, 2 of whom survived, paradoxical responses to hyperventilation were observed in several brain regions particularly the caudate nucleus, ranging from -1.1 to -2.6 ml/min/lO0 gm/torr Pco2. In 1 patient, local flows were at the lower limits of normal before hyperventilation in several brain regions (41 + 1 ml/min/lO0 grn) and after hyperventilation flow approached ischemic levels (16 + 6 ml/min/lO0 gm).

Ashwal et al: CBF and ChildhoodStrangulation 371

Table 2. Pre- and post-hyperventilation cerebral blood flow in 4 children with self-inflicted strangulation* Pre-hyperventilation

(Pco2 = 39 + 3 torr)

Post-hyperventilation (Pco2 = 29 + l torr)

Total cerebral blood flow

63 ± 8

46 ± 16:

Frontal gray

63 ± 6

41 ± 6~

Frontal white

51 ± 8

27 ± 71

Temporal

71 ± 22

50 ± 12

Parietal

57 _+ 11

44 ± 13t

Occipital

66 ± 8

54 ± 32

Caudate nucleus

68 _+21

70 ± 22

Thalamus

94 ± 30

78 ± 48

Lenticular nucleus

77 ± 26

63 ± 27*

Midbrain

85 _+ 18

39 ± 5t

Cerebellum

91 ± 11

78 ± 11

* Values are expressed as ml/min/100 gm and are mean ± S.D. for the 4 patients. * P < 0.05 (Student t test). S t r a n g u l a t i o n , as s e v e r a l o t h e r f o r m s o f c h i l d h o o d trauma, causes s e v e r e g l o b a l h y p o x i c - i s c h e m i c injury. This acute insult is frequently m a d e worse by carotid arterial c o m p r e s s i o n or d e l a y e d o c c l u s i o n w h i c h causes c e r e b r a l i s c h e m i a and by j u g u l a r v e n o u s c o m p r e s s i o n w h i c h leads to cerebral congestion. In contrast to other injuries, such as n e a r - d r o w n i n g accidents, increased I C P occurs m o r e frequently in c h i l d h o o d strangulation [9]. In one series, 36% o f hospitalized children suffering from

Table 3.

strangulation had increased ICP due to cerebral e d e m a which was associated with a mortality o1" 75'~;~ 131. hI addition, the s a m e study found that another 27'.~ of patients who initially had normal ICP d e v e l o p e d a delayed onset o f cerebral herniation that was associated with a 675~: mortality with the remaining 33% suffering serious longterm neurologic sequelae. In our study, we found that after the acute injury, C B F returned to normal in most brain regions once cerebral perfusion and o x y g e n a t i o n were re-established. Our earliest X e C T C B F study was p e r f o r m e d 8 hours after admission (Patient l) and although we did not measure 1CR the initial cranial C T demonstrated no e v i d e n c e of cerebral edema, infarction, or hemorrhage. Occlusion of the extracranial or intracranial arteries was not observed in the 2 patients in w h o m autopsies were performed. Because these 2 patients had normal C B F results, our findings suggest limited prognostic value in the acute m e a s u r e m e n t of CBF, a finding that has been o b s e r v e d in other forms of acute global i s c h e m i a in adults 110]. In contrast, an impaired cerebrovascular response m hyperventilation has been shown in previous studies of adult head trauma to correlate with p o o r outcome. Nordstrom et al. reported that a global C B F / P c o 2 response less than 1 m l / m i n / 1 0 0 g m / t o r r P c o 2 was a s s o c i a t e d with p o o r neurologic o u t c o m e [11]. O f 11 patients with impaired C B F / P c o 2 reactivity, 7 died, 2 r e m a i n e d in a persistent v e g e t a t i v e state, and 1 m a d e a g o o d clinical recovery. Significant differences in regional C B F / P c o 2 reactivity were not reported in their study and this was b e l i e v e d to be related to m e t h o d o l o g i c problems associated with the 133Xe technique. Similar data for pediatric head trauma patients or other acute cerebral insults have not been re-

Total and regional cerebral blood flow Pc02/CO2 responsitivity*

Region

Pt l

Pt 2

Pt 3

Pt 4

Mean ± S.D.

Total cerebral blood flow

1.6

2.2

0.7

2.0

1.6 ± 0.7

Frontal gray

2.8

2.0

2.2

1.8

2.2 _+0.5

Frontal white

1.5

3.0

1.1

3.2

2.2 + I. 1

Temporal

4.1

0.7

3.6

0.8

2.3 + 1.8

Parietal

1,3

1.6

0.9

1.5

1.3 +_0.3

Occipital

2,0

2.5

-2.6

1.6

0.9 ± 2.3

Caudate nucleus

1.0

-1.1

~0.8

~).03

--0.2 ± 0.9

Thalamus

1,5

3.2

-1.9

2.4

1.3 _+2.2

Lenticular nucleus

1.2

0.5

1.9

1.9

1.4 ± 0.7

Midbrain

4.3

4.9

2.9

5.5

4.4 ± 1.2

Cerebellum

0.7

1.9

-1.9

2.9

0.9 ± 2.0

* All values are in ml/mirgl00 gm/torr.

372 PEDIATRIC NEUROLOGY Vol. 7 No. 5

ported; however, recent studies of childhood bacterial meningitis reported that the children who died had CBF/Pco2 reactivities below 1.0 and those who had poor outcomes had reactivities below 1.5-2.0 [12,13]. In our study, the 2 patients who had impaired reactivity averaging 1.2 ml/min/100 grn/torr died in contrast to the 2 who survived (2.1 ml/min/100 gm/torr). The patient who became brain dead had the most impaired CBF/Pco2 response (0.7 ml/min/100 gm/torr). At the time of XeCTCBF study her ICP was elevated (25 mm Hg) and cerebral perfusion pressure (60 mm Hg) and CBF (68 ml/min/100 gm) were normal; however, during the second 24 hours of hospitalization she no longer responded to hyperventilation therapy, her ICP remained quite elevated, and she continued to worsen, becoming brain dead on the fifth hospital day. She also had received phenobarbital (4 mg/kg) and a pentobarbital infusion (1 mg/kg/hr) to control the ICP but this was not successful. This phenomenon was also observed in the patients reported by Nordstrom et al. with impaired CBF/Pco2 reactivity [11]. Those patients with an impaired CBF/Pco2 response did not benefit from barbiturate or hyperventilation therapy in controlling elevated ICP. The second patient with an impaired CBF/Pco2 remained comatose and on the fourth hospital day suffered a cardiorespiratory arrest and died. He had not received any barbiturates and his repeated EEG one day before death was isoelectric suggesting neocortical death. Although ICP measurements were not obtained, neuropathology demonstrated diffuse cerebral edema and congestion suggesting that his ICP may have been elevated [14]. The findings in these 2 patients suggest that an impaired CBF/Pco2 response may be an early and relatively objective parameter that indicates that cerebral edema is likely to develop, that it may be difficult to control, and that ultimately it may be associated with death. The response to hyperventilation varied in the different brain regions and among the 4 patients (Table 3). Presumably this finding is a reflection of variable local metabolic activity, changes in extraceUular pH with hyperventilation, and the degree of local brain parenchymal and vascular injury. Several observations are important. The CBF/PCO2 response in the midbrain (4.4 + 1.2 ml/min/100 gm/torr PCO2) is much higher than in any other brain region, yet midbrain flow is similar to many of these regions (Tables 1,2) which raises concerns that flow to the midbrain, and presumably to other brainstem regions, during hyperventilation may be disproportionately reduced resulting in local ischemia. Because other investigations have documented that global brain arteriovenous oxygen extraction increases during hyperventilation to maintain oxygen delivery, it is probable that this also occurs in the brainstem to maintain local tissue oxygen requirements [15]. Of note, several brain regions, particularly the caudate nucleus, demonstrated a paradoxical response to hyperventilation. We do not believe that this manifestation was

a methodologic problem related to the stable XeCTCBF technique because the areas of paradoxical flow were large and we have observed this finding in several other patients, as have other investigators [16]. The significance of this finding is as yet uncertain. It may indicate an area of either transient or permanent localized tissue injury similar to that reported with luxury perfusion or loss of autoregulatory control. Of the 3 patients exhibiting this response, Patient 3 had no evidence at autopsy of a specific lesion in the caudate nucleus, while Patients 1 and 4 demonstrated diffuse central and cortical atrophy on follow-up magnetic resonance imaging scans performed 1 month after the insult. There were also several brain regions in one patient in which CBF was at the lower limits of normal with an intact CBF/Pco2 response. With prolonged hyperventilation, it is likely that CBF in such areas would fall below the ischemic threshold precipitating further injury which has been observed in several previous studies [17], including one by Ashwal et al. of childhood bacterial meningitis [12]. It is unknown whether this decreased flow is of clinical significance. The blood glucose levels in the 2 patients who died were above 300 mg/dl compared to the patient who demonstrated some recovery (104 mg/dl) but not that much greater than in Patient 4, who remains in a vegetative state (295 mg/dl). Elevated blood and brain glucose levels are believed to increase neurologic damage during hypoxic-ischemic injury and similar findings have been demonstrated in a variety of experimental animal models and several recent clinical studies in both adults and children [4]. Hyperglycemia augments ischemic brain injury by acceleration of anaerobic glycolysis and increased brain lactate formation leading to further cell injury. The greater cerebral injury associated with hyperglycemia, however, could simply be a manifestation of a more profound generalized ischemic insult; however, its presence in children should suggest the possibility of serious morbidity. In the 2 patients who died, the depressed CI was associated with elevated systemic and pulmonary vascular resistance and elevated filling pressures of both the right and left sides of the heart. Despite this low flow, highresistance state, CBF was maintained. Shock following hypoxic-ischemic events in infants and children has been described as cardiogenic [5]; however, this form of cardiogenic shock appears to differ from isolated myocardial failure. In the latter, the systemic vascular resistance index increases to return blood pressure to normal, but in our patients MABP was higher than normal which suggests that severe central nervous system injuries cause an exaggerated increase in systemic vascular resistance. Therapy to improve CI may require vasodilators; however, this must be performed carefully because inappropriate vasodilation may increase CBF and ICP. The results of our studies in these 4 children suggest that although CBF was normal, the response to hyperventilation may be of importance in the evolution of further brain

Ashwal et al: CBF and Childhood Strangulation 373

injury and is a critical factor in deciding whether hyperventilation may be of clinical benefit. Also of importance is our observation that, in the face of presumed elevated ICE we did not demonstrate improved cerebral perfusion with hyperventilation in any of these patients. In addition, 3 factors correlated with poor outcome and ultimate death in 2 patients included the following: (1) Initial blood glucose greater than 300 mg/dl; (2) CBF/PCO2 response less than 2.0 ml/min/100 gm/torr PCO2; and, (3) CI that remained less than 2.0 L/min/m 2 for longer than 12-24 hours.

References

[1] Bird CR, McMahan JR, Gilles FH, Senac MO, Apthorp JS. Strangulation in child abuse: CT diagnosis. Radiology 1987;163:373-5. [2] McClure GM. Recent changes in suicide among adolescents in England and Wales. J Adolesc 1986;9:135-43. [3] Feldman KW, Simms RJ. Strangulation in childhood: Epidemiology and clinical course. Pediatrics 1980;65:1979-85. [4] Ashwal S, Schneider S, Tomasi L, Thompson J. Prognostic implications of hyperglycemia and reduced cerebral blood flow in childhood near-drowning. Neurology 1990;40:820-3. [5] Lucking SE, Pollack MM, Fields AI. Shock following generalized hypoxic-ischemic injury in previously healthy infants and children. J Pediatr 1985;108:359-64. [6] Hildebrand CA, Hartmann AG, Arcinue EL, Gomez RJ, Bing RJ. Cardiac performance in pediatric near-drowning. Crit Care Med 1988; 16:331-5.

374

PEDIATRIC NEUROLOGY

Vol. 7 No. 5

[7] Ashwal S, Schneider S, Thompson J. Xenon cc~mputed I~m~og raphy measuring cerebral blood flow in the determination of brain death in children. Ann Neurol 1988:25:539-46. [8] Perkin RM. lnvasive monitoring in the pediatric intensive care unit. In: Nussbaum E, ed. Pediatric intensive care. Mourn Kasco: Future Publishing, 1989;255-67. [9] Orlowski Jl~ Drowning, near-drowning, and Joe-water submersions. Pediatr Clin North Am 1987;34:75-92. [10] Grotta JC. Can raising cerebral blood flow improve outcome after acute cerebral infarction. Stroke 1987:18:264-7. [11] Nordstrom C-H, Messeter K, Sundbarg G, Schalen W, Wemer M, Ryding E. Cerebral blood flow, vasoreactivity, and oxygen consumption during barbiturate therapy in severe traumatic brain lesions. J Neurosurg 1988;68:424-31. [12] Ashwal S, Stringer W, Tomasi L, Schneider S, Thompson J, Perkin R. Cerebral blood flow and CO2 reactivity in children with bacterial meningitis. J Pediatr 1990;117:523-30. [13] Ashwal S, Tomasi L, Schneider S, Perkin R, Thompson J. Bacterial meningitis in children: Pathophysiology and treatment. Presented at 16 European Federation of Child Neurology Societies meeting, Brussels, December, 1991. [14] Simpson RK Jr, Goodman JC, Rouab E, Caraway N. Baskin DS. Late neuropathological consequences of strangulation. Resuscitation 1987;15:171-86. [15] Bruce DA. Effects of hyperventilation on cerebral blood flow and metabolism. In: Philips JB III, ed. Clinics in perinatology: Neonatal pulmonary hypertension. Philadelphia: WB Saunders, 1984;673-80. [16] Darby JM, Yonas H, Marion DW, Latchaw RE. Local "inverse steal" induced by hyperventilation in head injury. Neurosurgery 1988; 23:84-8. [17] Cold GE. Does acute hyperventilation provoke cerebral oligemia in comatose patients after acute head injury? Acta Neurochir 1989; 96:100-6.