MY9700720 Neonatal Intracranial Hemorrhage

MY9700720 PRI - PAEDIATRIC RADIOLOGY

Neonatal Intracranial Hemorrhage In the fetal brain, the superficial cortical veins are poorly developed, and the highly vascular germinal matrix drains into the great vein of Galen and its Charles A. Gooding tributaries via thin-walled terminal branches of the thalamostriate vein. It is postulated that perinatal University of California, San Francisco, hypoxia causes circulatory failure, with generalized San Francisco, California, USA visceral venous congestion. Thrombosis with venous infarction or venous hemorrhage without infarction occurs in the terminal branches of the thalamostriate 6> Intracranial hemorrhage (ICH) is the most frequently vein within the germinal matrix. 8 Using chromium- 9 encountered neurological disturbance in neonatal 50-labelled red blood cells, Tsiantos et al. showed 10 medicine. Almost exclusively, it occurs in premature that most IVHs occur in the first 3 days of life. infants with birth weights less than 2,500 gm. The ltrasound and computed tomographic (CT) studies incidence of ICH is approximately 45% among have confirmed this. premature infants of less than 35 weeks' gestation1 or in those weighing less than 1,500 gm at birth. The Extensive experimental evidence has accumulated potential catastrophic nature of this entity is revealed attesting to the multi-factorial etiology of ICH. by the fact that ICH is the most common central Hypoxia and asphyxia seem to be specific precipitat- nervous system lesion found on postmortem exami- ing factors. It is known that the normal newborn nation of newborn infants.2 Bada et al.3 reported an maintains a constant cerebral blood flow despite 89% mortality rate for infants less than 1,000 gm and fluctuations in systemic blood pressure, a phenome- having intraventricular hemorrhage (IVH). In another non known as autoregulation. Studies have suggested study, IVH was the primary cause of death in one- that asphyxia may abolish this autoregulation, i.e., the third of infants weighing less than 1,000 gm.* It is cerebral blood flow parallels changes in systemic quite likely that, currently, proportionally even more blood pressure. Asphyxizcepisodes consist of a infants are dying from ICH as modern medical and combination of hypoxia, hypercapnia and acidosis. surgical techniques have reduced the death rates from other causes.5 Statistically linked risk factors in neonatal periventricular hemorrhage include prematurity, respiratory The purpose of this presentation is to review current distress syndrome, patentductusarteriosus, mechani- concepts regarding etiology, diagnosis, prevention, cal ventilation, hypercapnia, acidosis, and rapidcolloid treatment, and follow-up of infants with ICH. infusion. Pneumothorax has been strongly linked to subsequent hemorrhage,11 probably through a combi- Etiology nation of blood gas abnormality and hypertension. Approximately 75% to 90% of ICHs originate in the 6 germinal matrix. The germinal matrix is a highly The relationship of pneumothorax to the occurrence vascular subependymal structure adjacent to the lat- of IVH in the premature newborn was studied by Hill, eral ventricles, in the region of the caudothalamic et al.11 These investigators measured blood flow groove. It is a source of spongioblasts and neuro- velocity in the anterior cerebral artery by a transcuta- blasts, which migrate peripherally to form portions of neous Doppler technique.12 They found that at the the cerebral cortex, basal ganglia, and other forebrain time of pneumothorax there was a marked increase in structures. The germinal matrix is most pronounced flow velocity, especially during diastole, and that in the fetus of 6 and 8 months' gestation, and it with resolution of the pneumothorax, flow velocity decreases in size as the fetus matures. It is virtually 7 returned to normal levels over the ensuing hours. The absent in the full-term infant. Therefore it is not changes in flow velocity correlated closely with surpris,ng' that postmortem studies show a low systemic hemodynamic changes that occurred with .ncidence of ICH in stillbirths and that the majority of pneumothorax, i.e., an increase in mean systemic subependymal hemorrhages (SEHs) originate in in- blood pressure. Intraventricular hemorrhage docu- fants of between 28 to 34 weeks' gestation. mented by serial ultrasound scans was observed shortly

295 after the pneumothorax in the nine infants studied, as Head rotation to the side is known to increase intrac- well as in six additional infants that they mention in ranial pressure in the infant, and this may contribute the addendum to their study. Their data therefore to neurological morbidity of newborn infants either demonstrate a marked increase in flow velocity in the directly, by causing vascular distention leading to cerebral circulation at the time of pneumothorax. capillary rupture and IVH, or indirectly by producing This increase is important in the genesis of IVH, as cerebral edema and impairment of cerebral perfusion is suggested by the occurrence of IVH soon after the following asphyxia or IVH.18 cerebral hemodynamic changes occur. Hormonal neuropeptides such as B-endorphins, vaso- The close association between pneumothorax and the pressin, and prolactin are other possible participants exacerbation of IVH has been reported previously.13 in the pathogenesis of IVH.19 Lipscomb, et al.w reported that of 14 ventilated premature infants who developed pneumothorax, 12 Diagnosis (86%) developed IVH. They also showed that the Computed tomographic (CT) scanning and serial presence of extra-alveolar air is a significant risk ultrasound scanning of the neonate's head have factor for the development of IVH. These studies generated much information about ICH. Based on CT raise the question of whether pneumothorax is related scanning, a grading system for ICH has been proposed to IVH in a causative manner or whether pneumotho- byPapileetal.:20-21 rax and IVH are separate consequences of some common event. The data presented by Hill, et al. Grade 1: Subependymal suggest that the relationship is a causative one and that the pathogenetic mechanism includes an abrupt Grade 11: Intraventricular hemorrhage with- increase in cerebral blood flow at the time of pneu- out ventricular dilatation mothorax. Grade 111: Intraventricular hemorrhage with ventricular dilatation Birth asphyxia or trauma producing blood loss and hypotension may interfere with cerebral autoregula- Grade IV: Intraventricular hemorrhage with tion. Endotracheal tube blockage or displacement has parenchymal hemorrhage also been linked to hemorrhage.15 Severe degrees of 16 atelectasis may also be a cause. Intraventricular hemorrhage is often a clinically silent event. Papile et al.2© showed that 78% of the surviv- Meticulous care with continuous blood gas monitoring infants with IVH had no clinical evidence of IVH. ing may avert the often brief episodes which may not Furthermore, 20% of the infants without IVH had even be clinically recognized and which can lead to clinical signs suggestive of IVH. The analysis of the major ICH. Different insults predominate at different cerebrospinal fluid for protein and red blood cells was postnatal ages. Perinatal management may have a not helpful in determining which of the infants had significant role in reducing the incidence and severity IVH. of ICH. A reduced incidence of ICH is associated with delivery by caesarean section and absence of Shankaran, et al.22 developed a classification of ICH premature labour prior to delivery. Active resuscita- based on the ultrasonographic appearance of the tion with use of early respiratory support, avoidance hemorrhage. They performed real-time ultrasonic of the complications of artificial ventilation, and examinations on admission and every 3 days during attention to maintenance of a normal blood pressure the first 3 weeks of life in neonates less than 1,500 gm may be the prime factors associated with reduction in and when clinically indicated in other neonates. The severity of ICH. real-time sector scan was obtained through the anterior fontanelle using a 3.5- or 5.0-MHz transducer. Hemorrhages were sonographically classified as: New evidence indicates that use of pupillary dilators during ophthalmologic examination of the newborn may precipitate increased intracranial pressure, which Mild: Hemorrhage confined to the subependymal might be related to ICH.17 periventricular region or accompanied by a small amount of blood in the normal-sized lateral ventricle.

296 Moderate: Intermediate amount of blood in enlarged rhage/subependymal hemorrhage were found in 90% lateral ventricle. of 113 infants of 34 weeks' gestation. The disagree- ments between ultrasound and CT studies done in the Severe: Hemorrhage filling the entire lateral ventri- first week of life were only in cases of small cle forming a cast, or intracerebral extension hemorrhages. of the hemorrhage, or both. Norman showed that CT visualization of blood 23 Fleischer, et al. demonstrated a close relationship depends on the hemoglobin concentration.2" A hemo- between the severity of the ICH and the development globin concentration of 7 to 8 gm/100 ml has the same of progressive ventricular dilation. Eighty percent of attenuation value as the brain and therefore appears premature infants with minor bleeds (SEH or SEH/ isodense with the rest of the brain tissue. Bejar, et small IVH) did not develop significant ventricular al.27 demonstrated that ultrasound could detect fluid dilatation, whereas all infants with IVH or intrapar- blood (cerebrospinal fluid hematocrit 2%), whereas enchymal hemorrhage, or both, developed moderate CT studies did not. Furthermore, in an autopsy study, to severe dilatation. Bejar, etal. showed that an occipital horn with bloody cerebrospinal fluid (hematrocrit 0.5%) was isodense Levene, et al.24 used ultrasound to study 146 infants in the simultaneous ultrasound study. Borders, et al29 of 34 weeks' gestation or less to diagnose the presence foundthat in vitro blood free of microbubbles could of IVH. They found IVH in 52 (36%) of these 146 produce echoes with a hematocrit as low as 0.02%. infants and in 32 (50%) of the 64 infants of 30 weeks' The lesser sensitivity of CT studies could partially gestation or less. The first incident of IVH was before explain why Papile, et al.20 found a lower incidenc of 72 hours in 32 (78%) of 41 infants studied early. In these hemorrhages (45% to 50%). Furthermore, in this study, clinical factors that significantly corre- their studies with CT, the first scans were performed lated with the development of IVH were outborn as several days after birth. The hemoglobin concentra- opposed to inborn, administration of sodium bicar- tions of the hemorrhage could have been modified by bonate, hypothermia, intermittent positive pressure the reticuloendothelial (glial) cells which rapidly ventilation, continuous positive airway pressure, organize the clot and by an increase in the water hypercapnia, severe acidosis, and respiratory distress content of the hemorrhagic area. These changes would syndrome. The latter was not the most important reduce the concentration of hemoglobin and further factor. The ultrasound diagnosis agreed with the decrease the possibility that a hemorrhage could be necropsy findings in 90% of infants. detected by CT studies.

Partridge, et al.25 believe, from a pragmatic stand- point, that the optimal timing for the ultrasonic diag- Very recently, nuclear magnetic resonance (NMR) 30 31 nosis of ICH is days 4 to 7, with follow-up at day 14. spectroscopy and NMR imaging have been They believe that the most efficient time for ultra- advocated as ICH investigational tools by London's sound to diagnose ventricular dilatation is day 14, University Hospital and Hammersmith Hospital with follow-up at 3 months. teams. Brain damage in newborn infants in the inten- sive care nursery commonly results from cerebral A comparison of ultrasound and CT findings has been hemorrhage or ischemia leading to infarction. reported by a number of investigators. Edwards, et Hemorrhages are easy to detect with ultrasound or al.26 found that in all 17 cases compared, there was no CT, but neither technique reliably identifies ischemic disagreement between ultrasound and CT on ven- iesions until loss of brain substance has occurred. tricular size. The advantages of ultrasound incl ude no ionizing radiation, less expense than CT, and capa- Phosphorus-31 NMR can be used to measure nonin- bility of being performed in the isolette in the nursery, vasively the relative concentration of mobile phos- thereby minimizing the risk to the infant of hypoxia phorus compounds involved in energy metabolism in and hypothermia. living animals. In rabbits having cerebral ischemia experimentally produced at University College Bejar, et al.27 found that ultrasound was more sensi- Hospital, a phosphorus-31 spectrum seemed consis- tive in diagnosing small IVH/SEH and organized tent in the experimental animals. Phosphocreatine clots than were CT studies. Intraventricular hemor- concentration fell faster, but in unison with that of

297 ATP, and the concentration of inorganic phosphate Twenty-six of 35 surviving neonates had post- rose. Intracellular pH fell. Following reperfusion, the hemorrhagic hydrocephalus, and 11 infants required values returned to normal..30 shunt insertion.11 The survival rate of infants with hemorrhage and the incidence of post-hemorrhagic At Hammersmith Hospital, NMR brain scans were hydrocephalus correlated with the severity of the performed in 16 infants, of whom 14 had IVH. hemorrhage. The highest mortality rate was seen in Ultrasound provided more information than NMR on the group with ventricular cases, and all infants with size and position of hemorrhage as well as degree of ventricular cases developed hydrocephalus. The nfants ventricular dilatation. Nuclear magnetic resonance with ICH had a high percentage of neurological appears most useful in assessing degrees of myelini- deficits. All surviving neonates with ICH developed zation and detecting periventricular edema in acute porencephaly, probably as a result of liquefaction ne- hydrocephalus and following acute perinatal crosis of intracerebral hematoma. Sonographically asphyxia." this process is seen as a dense echogenic area which over the course of 2 to 6 weeks appears mottled. Later Prevention and Treatment this affected area becomes anechoic and communi- Aside from the obvious improvement in general cates with the ventricle, forming a porencephalic obstetrical and neonatal nursing care, several excit- dilatation of the ventricle. The incidence of post- ing specific new regimens have been proposed to hemorrhagic hydrocephalus in this study is greater minimize the incidence of ICH. Controversy and than that reported in other studies, possibly because contradictory evidence surround some of the newer these sonographic studies were done every 3 days, as proposals, particularly the use of phenobarbital. opposed to CT studies done at intervals of 7 to 14 days. Indomethacin, which inhibits prostaglandin synthetase, has been demonstrated to prevent cere- The post-hemorrhagic ventriculomegaly is probably bral microvascular permeability changes evoked the secondary to an obi iterative arachnoiditis in the acute onset of hypertension or asphyxia. Newborn posterior fossa, which impairs the flow of cerebro- beagle pups were given hemorrhagic hypotension spinal fluid either out of the fourth ventricle or through followed by volume expansion to produce IVH. Nine the tentorial notch. Ventricular dilatation does not percent of pups receiving indomethacin had IVH, always progress; surgical decompression of the compared with 80% of those who received saline.32 ventricles, therefore, should be reserved for those infants having progressive dilatation. Hydrocephalus per se did not influence prognosis. These findings are Ethamsylate is a synthetic, water-sol uble, nonsteroidal similar to those of Krishnamoorthy, et al.34 who found drug which has been used clinically to reduce capil- that the abnormal ventricular size on follow-up lary bleeding in ear surgery, buccal surgery, after examination did correlate with the severity of the prostatectomy, and in menorrhagia. It has no known hemorrhage and appears to persist. Forty-seven toxic effects. It increases platelet adhesiveness. It was percent with moderate IVH and 90% with severe IVH used in a double-blind study of 70 infants. Periven- continued to have abnormal ventricular size in follow- tricular hemorrhage developed in 9/.35 infants on up. However, abnormal ventricular size was notal ways ethamsylate and 18/35 who were on placebo." associated with poor neurological outcome.

Follow-up Outcome data are not yet fully developed since most One hundred nine surviving infants were followed of the identified affected children are still quite young. until they were about 18 months old. Major or minor Surviving infants with grade I and 11 lesions had developmental or neurological sequelae developed in normal CT scans at 3 weeks. The finding of promi- 5 (8%) of 62 infants with normal ultrasound scans, nence of the interhemispheric fissure in 83% of these and 2 (8%) of 25 infants with uncomplicated periven- infants on CT exam at 6 months may indicate some tricular hemorrhage. In contrast, 15 (71%) of 21 degree of cerebral atrophy. Extensive IVH (grade 111 infants who had ventriculomegaly were abnormal on and IV) was associated with progressive ventricular follow-up.35 dilatation.20

298 Summary 13. Dykes FD, et ai: Intravenlricular hemorrhage: A prospective Intracranial hemorrhage in neonatal premature infants evaluation of etiopathogenesis. Pediatrics 66:42, 1980. is a potentially catastrophic compl ication that is m uch 14. LJpscomb AP, et al: Pneumothorax and cerebral hemorrhage more common than previously thought. It is readily in preterm infants. Lancet 1 -.414,1981. amenable to diagnosis by noninvasive ultrasound exam. Multifactorial causes have been proposed to 15. Cooke RAVI, Morgan MEI, Coach NAG: Pneumothorax, explain its occurrence, and several specific prophy- mechanical ventilation and periventricular haemorrhage. Lancel 1 :155,1981. lactic and therapeutic regimens are predicated upon these proposals. 16. Wilson-Davis SL, et al: Late periventricular hemorrhage in low birth weight infants: Etiology of new hemorrhage detected after ten postnatal days. (In preparation). References 17. Lees BJ, Cabal LA: Increased blood pressure following upillary dilation with 2.5% phenylephrine hydrochloride in preterm infants. Pediatrics 68:231, 1981. 1. Ahmann PA, el al.: Inlraventricular hemorrhage in ihe high- risk preterm infant: Incidence and outcome. Ann Neurol 18. Emery JR, Peabody JL: The effects of head positioning on 7:118,1980. intracranial pressure in newborn infants. In Syllabus of Ihe Second Special Ross Laboratories Conference on Perinatal 2. Towbin A: Central nervous system damage in the human Intracranial Hemorrhage, Vol. 1, pp. 118-130, December fetus and newborn infant. AmJDis Child 119:529,1970. 1982. 3. Bad* HS, et al.: Intracranial pressure and cerebral arterial 19. Stark Rl, et al: B-endorphin, vasopressin and prolactin: pulsatile flow measurements in neonatal intraventricular Possible contributors to the palhogenesis of inlravenlricular hemorrhage. J Pedialr 100:291, 1982. hemorrhage. In Syllabus of the Second Special Ross Labora- tories Conference on Perinatal Imracranial Hemorrhage. 4. Valdes-Dapena MA, Arey JB: The causes of neonatal Vol. l,pp. 158-188, December 1982. mortality: An analysis of 501 autopsies on newborn infants. } Pedialr 77:366,1970. 20. Papile L, Burstein J, Burstein R: Incidence and evolution of subependymal inlravenlricular hemorrhage: A study of infants 5. Pevsner PH, et al.: Subepcndymal and inlraventricular with birth weights less than 1500 gm. JPediatr92:529,\91i. hemorrhage in neonates. Radiology 119:111, 1976. 21. Burstein J, Papile L, Burstein R: Intravenlricular hemorrhage 6. Leech RW, Kohncn P: Subepcndymal and intraventricular and hydrocephalus in premature newborns: A prospective hemorrhages in the newborn. Am J Palhot 77:465, 1974. study with CT. AJR 132:631, 1979. 7. Burslein J, Papile L, Burstein R: Subependymal germinal 22. Shankaran S, et al: Sonographic classification of intracranial matrix and intraventricular hemorrhage in premature infants: hemorrhage: A prognostic indicator of mortality, morbidity, Diagnosis by CT. AJR 128:971, 1977. and short-term neurologic outcome. JPediatr 100:469,1982. 8. Coal VA, et al: Palhogenesis of intraventricular hemorrhage 23. Fleischer AC, et al: Serial sonographic monitoring of in newborn infants. Arch Dis Child 49:722,1974. posthemorrhagic ventricular dilatation and porencephaly occurring with inlracranial hemorrhage (ICH) in theprelerm 9. Tsiantos A,et al: Intracranial hemorrhage in the prematurely neonate. In Sylla bus of the Second Special Ross Laboratories bom infant: Timing of clots and evaluation of clinical signs Conference on Perinatal Intracranial Hemorrhage. Vol 11, and symptoms. J Pedialr 85:854, 1974. pp. 786-798, December 1982. 10. Emerson P, et al: Timing of intraventricular haemorrhage. 24. Levene Ml, Fawler CL, Lamonl RD: Rick factors in the Arch Dis Child 52:183,1977. development of intraventricular haemorrhage in the preterm neonate. Arch Dis Child 57:410,1982. 11. HillA,PerlmanJM,VolpeJJ:Relationshipofpneumothorax to occurrence of inlraventricular hemorrhage in the prema- 25. Partridge IC, el al: Optimal timing for diagnostic cranial ture newborn. Pediatrics 69:144, 1982. ultrasound in low-birth-weight infants: Detection of intracranial hemorrhage and ventricular dilatation. In Syllabus of 12. Perlman JM. Volpe JJ: Cerebral blood flow velocity in the Second Special Ross Laboratories Conference on Peri- relation to intravenlricular hemorrhage in the premature natal Intracranial Hemorrhage. Vol 1, pp. 614-639, December newborn infant. J Pedialr 100:956, 1982. 1982.

299 26. Edwards MK, et al: Cribside neurosonography: Real-lime sonography for incranial investigation of the neonate. AJR 136:271,1981.

27. Bejar R, et al: Diagnosis and follow-up of intraventricular and intncerebral hemorrhages by ultrasound studies of infant's brain through the fontanelles and sutures. Pediatrics 66:661,1980.

28. Norman D: Computed tomography in intracrania! hemorrhage in Computed TomographK Norman D, Korobkin M, Newton T (Eds.), pp. 279-299. St. Louis: C.V. Mosby, 1977.

29. Borders S.etal: Ultrasonic energy backscatlcred from blood: An experimental determination of the variation of sound energy with hematocril. Ann BiomedEng 6:83, 1978.

30. Delpy DP, el al: Non-invasive investigation of cerebral ischaemia by phosphorus nuclear magnetic resonance. In Syllabus of the SecondSpecial Ross Laboratories Conference on Perinatal Intracranial Hemorrhage. Vol. 1, pp. 579- 594, December 1982.

31. Levene Ml, et al: Nuclear magnetic resonance imaging of IVH and related disorders. In Syllabus of the Second Special Ross Laboratories conference on Perinatal Intracranial Hemorrhage. Vol 1, pp. 595-613, December 1982.

32. Mail LR, et al: Beagle puppy model of iniraventricular hemorrhage: Randomized indomethacin prevention trial. In Syllabus oftheSecondSpecialRossLaboratoriesConference on Perinatal Intracranial Hemorrhage. Vol. U, pp. 816-837, December 1982.

33. Morgan MEI, Benson JWT.Cooke RWI. Elhamsylate reduces the incidence of perivenlricular haemorrhage in very low birth-weight babies. In Syllabus of the Second Special Ross Laboratories Conference on Perinatal Intracranial Hemorrhage. Vol. U.pp. 838-839, December 1982.

34. Krishnamoorthy KS, et al: Neurologic sequelae in survivors of neonatal intraventricular hemorrhage. Pediatrics 64:233, 1979.

35. Stewart Al.etal: Relation between ultrasound-appearance of the brain in very prelerm infants and neurodevclopmcnlal outcome at 18 months of age. In Syllabus of the Second Special Ross Laboratories Conference on Perinatal Intracranial Hemorrhage. Vol U, pp. 1090-1116, December 1982.

300