Michael J. Acarregui, MD
Peer Review Status: Internally Peer Reviewed


There are four major types of intracranial hemorrhage which may affect the neonate. These include subdural hemorrhage, primary subarachnoid hemorrhage, intracerebellar hemorrhage and periventricular-intraventricular hemorrhage (PVH-IVH). In the Intensive Care Nursery PVH-IVH is the most common of the four and for the preterm infant represents the type of hemorrhage of greatest clinical significance.

The incidence of PVH-IVH varies considerably in the literature, the majority of centers reporting an incidence of 20-30% for infants with a birth weight <1500 g. Different incidence figures among centers reflects multiple factors such as the proportion of inborn and outborn births (the latter group have been shown to have a higher incidence of PVH-IVH compared to inborn infants), timing of sonography, and whether all eligible infants were evaluated.

There are several classifications to characterize the extent of PVH-IVH. A relatively simple classification which is often used is as follows:

  • Grade I: hemorrhage limited to the germinal matrix (subependymal hemorrhage)
  • Grade II: hemorrhage which has extended into the ventricular system but without dilation of the lateral ventricles.
  • Grade III: hemorrhage extending into the ventricular system with the blood resulting in ventricular dilatation.
  • Grade IV: hemorrhage which extends into the brain tissue (this grade is also referred to as PVH and associated with intraparenchymal echodensity (IPE) by some).

A problem with this grading system that needs acknowledgment is that objective determination of ventricular dilation is difficult. Determination of the extent of hemorrhage is important since most follow-up studies have found that the probability of neurologic morbidity (cognitive, motor, etc.) is high (>50% depending upon the study) for more extensive hemorrhage (grade III and IV). In contrast, it appears that the presence of a grade I or II PVH-IVH does not measurably increase the chance of neurologic morbidity. Lesions which occur in the periventricular white matter occur in 3-10% of infants with birth weight <1500 g, are frequently bilateral, are felt to be ischemic in origin, and will evolve into cystic lesions of the periventricular white matter (periventricular leukomalacia, PVL). The presence of PVL carries a high risk of neurologic morbidity (most often spastic diplegia).


As the name implies, PVH originate in the tissue abutting the lateral ventricle, e.g., germinal matrix. In most infants, PVH arise in the germinal matrix at the level of the foramen of Monroe, although in extremely preterm infants (<28 wks) PVH often arise further posteriorly in the germinal matrix. From multiple sonographic studies of the matural history of PVH-IVH, it is evident that most hemorrhages remain confined to the germinal matrix area (60-70% of PVH-IVH depending upon the study). Many hemorrhages will be clinically silent, and very few hemorrhages have a catastrophic presentation (e.g., profound alteration in neurologic state, hypotension, apnea, bulging fontanel, drop in hematocrit, etc.).

The pathogenesis of PVH-IVH remains unclear. A complex multifactorial etiology is likely. There have been a number of clinical trials to prevent the occurrence of PVH-IVH using phenobarbital, Vitamin E, indomethacin, Vitamin K, and ethamsylate, all without conclusive results. If PVH-IVH occurs, additional sonograms will be needed to monitor for extension of the hemorrhage and post-hemorrhagic complications (porencephaly, hydrocephalus). Using serial sonography, it has been shown that the occurrence of post-hemorrhagic hydrocephalus is relatively uncommon after PVH-IVH (~ 13%). However, using serial cranial sonography, it has been shown that enlargement of the lateral ventricles may precede change in head circumference. Thus, once PVH-IVH has occurred follow-up cranial ultrasound is indicated. Similarly, when post-hemorrhagic hydrocephalus (PHH) is evolving (50% of Grade III IVH will be complicated by PHH), management (if any) should be discussed with the Attending Staff.


Since it is difficult to predict the presence or absence of neonatal intracranial hemorrhage by clinical criteria, the following schedule is used for routine head ultrasounds for "all" infants ≤1500 g birth weight:

  • Ultrasound 1. 5-7 day
  • Ultrasound 2. 28-30 day or before discharge
  • If PVH-IVH is detected on ultrasound, should be obtained more frequently (weekly) to evaluate progression of ventricular dilitation or cystic change.

The timing of the above head ultrasound schedule takes into account that most PVH-IVH occurs in the first week of life. However, the presence of late PVH-IVH does occur and necessitates an ultrasound examination at a month of life.

Post Hemorrhagic Hydrocephalus

It has been well demonstrated that enlarging head circumference is an insensitive sign of hydrocephalus in the premature infant. Ventricular dilatation after neonatal intracranial hemorrhage probably begins soon after the hemorrhage in many infants and pre-dates the increase in the rate of head growth by days to weeks. Infants with hydrocephalus have a poor prognosis, and one important factor in their outcome may be a delay in the detection and treatment of hydrocephalus.


The following is the recommended approach to the identification and care of these infants:

  1. All infants <1500 grams will be screened by ultrasound for evidence of intraventricular hemorrhage. The screening ultrasound will be done on or about the seventh day of life. The daytime nursery ward clerks will be responsible for identifying which babies are seven days of age from the census book and filling out an x-ray request. Ultrasounds will be performed daily Monday through Friday. If an infant's seventh day falls on the weekend, the scan should be done on Friday or Monday, whichever day is closer to the seventh day. Abnormal ultrasounds will be presented at the daily radiology conference.
  2. The house staff will be responsible for identifying and ordering ultrasounds on infants >1500 grams who might be at risk for significant hemorrhage.
  3. Once a hemorrhage has been identified by screening ultrasound, the pediatric house officer then becomes responsible for ordering ultrasounds on a weekly basis until it is clear that the hemorrhage has resolved and that there has been no progression of ventricular size.
  4. If there is clinical evidence of an intraventricular hemorrhage (drop in hematocrit, seizures, full fontanel, bloody CSF, unremitting acidosis, etc.) an ultrasound should be ordered by the resident on the day that it is desired. Emergency ultrasounds can be done whenever indicated.

Serial Lumbar Punctures:

Serial lumbar punctures have been used to control increased intracranial pressure when there is clinical evidence of rapidly progressive ventricular size.

There is little evidence in the literature of the efficacy of serial lumbar punctures for the prevention of hydrocephalus. However, it appears that serial lumbar punctures can be beneficial in protecting the cortical mantle in an infant with progressive hydrocephalus who is too small to be shunted. The care of each baby needs to be individualized, and there may be changes in our approach to these infants as new information becomes available. Decisions with regard to lumbar punctures should be made with the attending neonatologist. In general the following recommendations appear reasonable:

  1. Minimally dilated ventricles without progression do not warrant the use of serial lumbar punctures.
  2. If progressively enlarging ventricles are identified on ultrasound (with or without clinical signs of increased intracranial pressure), daily LP's may be indicated. Enough fluid should be removed to soften the fontanel, usually 10 to 15 ml. Those with significant intracranial pressure may need 20 to 30 ml removed. The taps should be continued until the ventricles stabilize or decrease in size, or until the infant is large enough to undergo a ventriculoperitoneal shunt. If taps are discontinued because the ventricles have decreased in size, a follow-up ultrasound should be obtained in about seven days to insure that ventricular size remains stable.
  3. Enlarged ventricles may be secondary to cerebral atrophy. In such cases, ventricular dilatation is a passive process and not related to change in CSF dynamics. Serial lumbar punctures are not indicated in such cases.
  4. In cases in which progressive ventricular enlargement and clinical signs of increased intracranial pressure cannot be controlled by periodic taps and in which the child's weight is still sufficiently low that shunting cannot be done, other modes of therapy should be considered after appropriate consultation.
  5. If the decision is made to undertake serial lumbar punctures, certain precautions must be observed. The lumbar puncture must be done with meticulous technique because meningitis is a potential risk. In addition, electrolytes should be measured periodically if large volumes of fluid are removed.
  6. Repeat head ultrasound exams should be done at weekly intervals whenever blood has been identified in the ventricles in order to monitor changes in ventricular size. Discontinuation of ultrasounds must be decided on an individual basis. In addition, consideration should be given to ordering a late head ultrasound exam at about 6 weeks of age in infants born at 26 weeks or less; the purpose of this late exam is to screen for periventricular leukomalacia.
  7. Depending upon the severity of the initial hemorrhage and the clinical presentation, a single CT scan might provide information about cerebral cortical and white matter pathology not available by other means of investigation. If desired for prognostic reasons, the CT scan should be performed near the time of discharge.
  8. At the time an infant with post hemorrhagic hydrocephalus is discharged, arrangements should be made for follow-up in the Neonatology Clinic in four weeks. In some cases a repeat ultrasound may be necessary at that time. The infant who has had a shunt procedure during the hospitalization should also be followed in the Neurosurgery Clinic.


Goldstein GW, Chaplin ER, Maitland J. Transient hydrocephalus in premature infants: treatment by lumbar puncture. Lancet 1976;1:512-514.

Papile LA, Burstein J, Burstein R, Koffler H, Koops BL, Johnson JD. Posthemorrhagic hydrocephalus in low-birth-weight infants: treatment by serial lumbar punctures. J Pediatr 1980;97:273-277.

Kreusser KL, Tarby TJ, Kovnar E, Taylor DA, Hill A, Volpe JJ. Serial lumbar punctures for at least temporary amelioration of neonatal posthemorrhagic hydrocephalus. Pediatrics 1985;75:719-724.

Szymonowicz W, Yu VYH, Lewis EA. Post-haemorrhagic hydrocephalus in the preterm infant. Aust Paediatr J 1985;21:175-179.

Volpe J. Intraventricular hemorrhage and brain injury in the premature infant: neuropathology and pathogenesis. Clinics in Perinatology 1989;16(2):361-386.

Volpe J. Intraventricular hemorrhage and brain injury in the premature infant: diagnosis, prognosis and prevention. Clinics in perinatology 1989;16(2):387-411.