Major findings of the study

Through your participation in this study, we have:

• found safe new ways to perform MRI scans on newborn babies.
• confirmed the major patterns of hypoxic-ischemic brain injury on MR imaging, and shown that MRI scans performed early in life yield useful information.
• evaluated contrast enhancement of the blood-brain barrier, describing the appearance in normal infants and finding that abnormal enhancement is associated with poor outcome.
• developed a MRI scoring system and evaluated which MRI images are most useful for predicting outcome.
• constructed a neuromotor score to summarize the result of neurological exams.
• developed an encephalopathy score to describe the baby's brain function in the first days of life.
• built a seizure scoring system to size up the severity of seizures in the newborns.
• studied the usefulness of MR spectroscopy in predicting future neurodevelopmental problems.
• confirmed that inflammation during pregnancy or birth is associated with later neurological damage of the child.

The MRI scan generates many different images of your baby's brain. Certain areas of the brain and certain sorts of injury show up better on different types of images. The appearance of an injury is also influenced by how long after the injury the scan is obtained (ref 1, 2). One of the most basic, but important, findings in the BAMRI study so far is that hypoxic-ischemic brain injury (i.e. injury due to lack of oxygen and/or blood supply) affects specific parts of the brain and shows up well on MRI scans in the first 10 days of life. It has been possible to obtain images safely even in very young babies (as young as 17 hours) by using simple sedation and specially designed incubators that fit inside the MR scanner (ref 3).

The blood-brain barrier (BBB) is a specialized layer of cells that separate brain cells from blood vessels and effectively filters anything passing from the bloodstream into the brain. In the early phase of the study, some children received injections of "contrast medium" into their bloodstream during the MRI scan. If the BBB is working efficiently, the contrast should remain in the blood vessels, but if it has become temporarily damaged (e.g. by lack of oxygen or blood supply) then contrast can leak across it. This leakage shows up on the MRI scan. Data from the BAMRI study confirmed that in infants with normal MRI scans the BBB works effectively (ref 4). Infants who had "leaky" BBB's were more likely to have more severe brain injury and to have more neurological/developmental problems as they grew up (ref 5). We do not give contrast medium to the new babies in the study any longer.

As the study has progressed, we have been able to include newer MR techniques. Since 1996 we have been able to perform MR spectroscopy, which measures levels of certain chemicals in specific regions of the brain by using a magnet. MR spectroscopy is performed at the same session when the MR scanner obtains MR images of the brain. We measure lactate, which is produced by body tissues when they are short of oxygen, and other chemicals (N-Acetyl Aspartate or NAA and choline), which are produced by healthy nerve cells in the brain. Due to the lower oxygen and/or blood flow in the babies' brains, we have found higher levels of lactate and reduced levels of NAA at the time of the scan in the babies who become neurologically impaired by 2.5 years of age, compared to children who are neurologically normal at that time. We have also found that children in families with lower household income are at higher risk of neurological worsening between one year and 2.5 years of age than children in families of higher household income.

For the past couple of years, we have also been able to perform diffusion MR imaging, which allows us to measure how much the water in the brain is moving and in what direction. Water stops moving very soon (in as little as 15 minutes) after brain cells are injured. Therefore, we are able to see if injury has occurred very soon after birth. We have started performing MRS and diffusion imaging within the first 24 hours after birth, and our preliminary findings in 10 babies have suggested that MR spectroscopy was always abnormal in children who had brain injury, whereas diffusion imaging usually identified, but consistently undersestimated the extent of the injury. Currently, we are working to see whether loss of anisotropy -directionality of water movement - is a better early marker of injury than the reduction in diffusion.

Before information from a study can be statistically analyzed, it has to be expressed as a number. Many pieces of information are simply numbers from the start (e.g. birth weights, head sizes). However, it is often necessary to describe more complicated information, such as a physical exam or a series of MRI pictures, as a single number. This requires the development of scoring systems and much information from the BAMRI study is used this way.

First, we developed a scoring system for the extent and site of injury seen on the MRI scan. We found that it was often possible to predict which babies would go on to develop neurological or developmental problems by 1 year of age and that certain types of MRI images were most effective at predicting future problems (ref 7).

We also needed to score the neurological exam, and devised the "neuromotor score"(NMS). We found that babies with high scores (i.e. many neurological abnormalities) at 3 months of age continued to show major abnormalities when they reached their first birthday. However, babies with mild neurological abnormalities at 3 months of age frequently had completely normal exams at 1 year (ref 8).

Next we developed an "encephalopathy score" to describe how normally or abnormally a newborn baby's brain is performing all its functions. These include alertness and responsiveness, control of muscle tone, movement and reflexes, the presence of seizures, and the ability to feed and breathe without assistance. Previous studies have shown that babies with abnormal brain function persisting for a long period of time (weeks) after hypoxic-ischemic injury are very likely to have developmental problems. Using our scoring system during the first 3 days of life we were able to predict which children were likely to have neurodevelopmental problems at 1 year of age.

Finally, we developed a "seizure score" to be able to express the severity of newborn seizures as a number. We took several factors into consideration, such as how early the first seizure was noted, how many times the baby had seizures in the hospital, what the brain wave test (EEG) showed, and how many drugs needed to control the seizures. When we looked if this seizure score was associated with the lactate and NAA measured by MRS, we found that seizure severity was an independent marker of brain injury.

Although all these scoring systems seem to be useful measures, none of them are perfect. The ideal test separates out all the babies who will go on to have problems from those who will not and never gets it wrong. Whether we use MRI scans, spectroscopy, encephalopathy scores or early neurological exams it is easiest to pick out those babies who will have severe problems and, at the other extreme, those who will be completely unaffected. It is much harder to predict accurately what will happen to the babies in our study who have only mild/moderate abnormalities early in life. Maybe adding together all these separate pieces of information will be a better solution. All our scores are being published in medical journals and can then be used by other researchers studying similar babies.

Our scores still need to be tested in other groups of babies, and by other investigators, to see if they are truly reliable at predicting what will happen as babies grow up. For this reason we do not release any of our scores to parents or doctors outside the study. However, we are happy to share the information on which they are based (MR reports, neurological exam findings etc.)

We have asked the first 120 families of the study to let us retrieve some archived blood samples of their children that were drawn for screening purposes in the hospital when they were born. In those samples, we measured certain proteins, called cytokines, that are produced during inflammation. Inflammation is the body's set of responses to various conditions, such as infection, trauma or hypoxia. We have found that the blood levels of certain cytokines were much higher in children who had confirmed or suspected severe neurological damage at one year of age. We still do not know if these cytokines are causing the brain injury or only markers of it, and this question would most probably be answered by animal and laboratory studies. Currently, we are investigating if the levels of cytokines are associated with the lactate and NAA levels in the brain to see if the cytokines could be used in diagnosing brain injury in newborns.

So far, we have concentrated mainly on looking at how children are performing at 1 and 2.5 years of age. Unfortunately, since babies' brains are still immature at 1 year, the results based on these tests may not give us completely accurate results. To do meaningful statistical tests we need a fairly large group of older children, who have been tested when they are 2.5 and 4 years old. Our results from the BAMRI children who have had their 2.5 year examinations indicate that neurological and developmental findings can change significantly between 1 year and 2.5 years, so we really need to follow-up as many children as possible to figure out what our neonatal exams mean! Some families are having difficulty returning for visits. We hope to have more funding available soon to help with transportation and accommodation costs for families who live a long way from San Francisco. Please do what you can to get to your follow up exams! We will work with you and give you any help we can. You can get in touch with us by mail, telephone, fax, or e-mail.

Your participation in the study is of great value to us as we attempt to learn better ways to care for children with birth asphyxia. We appreciate your commitment to the BAMRI study and look forward to seeing you at the next follow-up visit!

back to top