Search

Neuropathological examinations were carried out at necropsy on 83 very pre-term babies who died during their first hospital admission. Forty seven (57%) babies had evidence of cerebral damage-39 with ischaemic white matter damage. The time of onset of ischaemic lesions was thought to be prenatal in 12 cases (31%) and postnatal in a further 12 (31%). The exact timing of damage could not be determined in 15 (38%) cases. Maternal and neonatal case notes were reviewed to ascertain clinical associations of ischaemic white matter damage. There were no clear associations between adverse clinical factors and prenatal ischaemic white matter damage. In contrast, pre-eclampsia, intrauterine growth retardation, and delivery without labour were associated with postnatal damage as were neonatal sepsis, necrotising enterocolitis, and seizures. The absence of a clear association between the timing of adverse clinical factors and the timing of ischaemic cerebral damage suggests that cerebral damage in very preterm babies may result from a sequence of events rather than one specific insult.

Twelve normal and 32 asphyxiated neonates were studied using global and depth resolved phosphorus magnetic resonance spectroscopy (31PMRS). Eight of the asphyxiated group died or survived with major neurodevelopmental abnormalities. A global phosphocreatinine/inorganic phosphate (PCr/Pi) ratio below the range of values from normal infants predicted adverse outcome after asphyxia with a positive predictive value of 64%, sensitivity 88%, and specificity 83%. Corresponding values for global inorganic orthophosphate/adenosine triphosphate (Pi/ATP) ratios were positive predictive value 88%, sensitivity 96%, and specificity 88%. Spatially localised MRS data, obtained using phase modulated rotating frame imaging, showed cerebral energy metabolism to be more abnormal in deep than superficial regions after birth asphyxia. However, in this population of full term infants none of the regional metabolite concentrations were superior to global data for prediction of outcome.

MRS is a noninvasive technique that does not use ionizing radiation and can be used to measure relative metabolite concentrations in human tissues and organs in vivo. Phosphorus MRS can be used to study energy metabolites and intracellular pH. The first neonatal applications were described in 1983 in a study of cerebral metabolism. Since then, the value of cerebral MRS as research tool and an investigative technique has been confirmed, and its prognostic power in asphyxiated infants has been established. Techniques of spatial localization and quantitation have been developed, but studies of other organs and the use of other nuclei remain at a very preliminary stage. Considering the huge potential of MRS and the proliferation of high field magnets primarily designed for imaging, there has been a disappointing lack of progress in the development of clinical and research applications of spectroscopy. The logistic differences of studying sick infants in strong magnetic fields make MRS a time-consuming and labor-intensive investigation, which will inevitably limit its widespread routine use. Research studies are hampered by the diversity of spectroscopic and signal processing techniques, which make comparisons of data from different groups impossible. Some techniques for the assessment of cerebral hemodynamics such as doppler ultrasound and near infrared spectroscopy have the advantage of being available at the cotside, but MRS is unique in providing quantitative information about a wide range of intracellular metabolites. The altricial development of MRS as a clinical investigative tool in neonatology can be ascribed partly to practical difficulties, but these should not detract from the exciting possibilities opened up by a technique that gives a noninvasive insight into intracellular chemistry. The metabolic information from MRS is an invaluable addition to the information provided by other techniques and will certainly play an important role in unraveling the sequence of events between an hypoxic-ischemic insult and cell death. A better understanding of these mechanisms is a prerequisite to the development of rational therapeutic maneuvers following asphyxial insults.

Phase-modulated rotating frame imaging is a modification of magnetic resonance spectroscopy, which uses a linear radiofrequency field gradient to obtain spatially localized biochemical information. Phase-modulated rotating frame imaging was used to study regional cerebral energy metabolism in the brains of 9 normal newborns and 25 newborns after birth asphyxia. Relative concentrations of phosphorus-containing metabolites and intracellular pH were determined for brain tissue at three specified depths below the brain surface for all neonates. Wide variations in metabolite ratios were seen among normal neonates, and considerable metabolic heterogeneity was demonstrated in individual neonates by depth-resolved spectroscopy. Asphyxiated neonates with severe hypoxic-ischemic encephalopathy and a poor neurodevelopmental outcome showed the expected rise in inorganic orthophosphate and fall in phosphocreatine concentrations in both global and spatially localized spectra. Phase-modulated rotating frame imaging showed that metabolic derangement was less in superficial than in deeper brain tissue. The inorganic orthophosphate-adenosine triphosphate ratio from 1 to 2 cm below the brain surface was more accurate than any global metabolite ratio for the identification of neonates with a poor short-term outcome. These data are consistent with the known vulnerability of subcortical brain tissue to hypoxic-ischemic injury in the full-term neonate.

The brains of 30 normal preterm and term infants whose birth wt were appropriate for gestational age and 13 who were small for gestational age but healthy were studied by phosphorus magnetic resonance spectroscopy to determine values for metabolite concentration ratios and intracellular pH. In the AGA infants, phosphocreatine/inorganic orthophosphate increased between 28 and 42 wk of gestational plus postnatal age, suggesting a rise in the phosphorylation potential of brain tissue. At the same time, the concentration of phosphomonoester (mainly phosphoethanolamine) fell and that of phosphodiester (including phosphatidylethanolamine and phosphatidylcholine) increased. These changes reflected myelination and proliferation of membranes. Intracellular pH was approximately 7.1 and did not change with brain maturation. No differences were detected in these variables between the infants who were small for gestational age and those who were appropriate for gestational age.

Intracellular energy metabolism was studied by phosphorus magnetic resonance spectroscopy in the brains of 27 preterm and term infants with increased echodensities consistent with hypoxic-ischaemic injury and 18 comparable normal infants. In the normal infants the phosphocreatine (PCr)/inorganic orthophosphate (Pi) ratio increased significantly from 0.77 +/- 0.24 (95% confidence limits) at a gestational plus postnatal age of 28 weeks to 1.09 +/- 0.24 at 42 weeks. 9 of the 15 infants with increased echodensities whose PCr/Pi ratios fell below the normal range died; in all 6 survivors cerebral atrophy developed (cysts in brain tissue or microcephaly). In contrast, all 12 infants with increased echodensities whose PCr/Pi ratios remained within the normal range survived, although cerebral atrophy developed in 3 with ratios towards the lower limit of normal.

1. The effects of hypercapnia and hypocapnia on brain intracellular pH (pHi) and metabolism were investigated in new-born lambs under barbiturate anaesthesia. 2. 31P nuclear magnetic resonance (n.m.r.) spectroscopy was used to determine brain pHi and the relative concentrations of compounds containing mobile phosphorus nuclei including phosphocreatine (PCr), nucleoside triphosphates (NTP) and inorganic phosphate (Pi). Simultaneous measurements were made of the molar ratio of glucose to oxygen uptake by the brain. 3. During normocapnia (arterial partial pressure of CO2 Pa, CO2, 39 +/- 1 mmHg mean +/- S.E. of mean, n = 9) brain pHi was 7.13 +/- 0.02. Hypercapnia (Pa, CO2, 98 +/- 3 mmHg) was associated with a fall in brain pHi to 6.94 +/- 0.03 (n = 19, P less than 0.001), whereas no significant change in brain pHi occurred during hypocapnia (Pa, CO2, 16 +/- 1 mmHg; brain pHi 7.15 +/- 0.01). 4. During hypercapnia there was an increase in the ratio of Pi to NTP from 1.09 +/- 0.08 to 1.47 +/- 0.06 (P less than 0.001) and a decrease in the ratio PCr/Pi from 1.60 +/- 0.08 to 0.93 +/- 0.04 (P less than 0.001). There was a linear correlation between Pi/NTP and brain pHi. 5. Alterations in arterial PCO2 had no significant effect on the molar ratio of glucose to oxygen uptake by the brain, which remained close to unity. 6. The change in brain pHi observed during hypercapnia can be accounted for by the known physico-chemical buffering capacity of brain tissue. Homoeostasis of brain pHi during hypocapnia provides further evidence that additional regulatory mechanisms operate in these circumstances. 7. The observed changes in PCr and Pi can be accounted for in part by the [H+] dependence of the creatine kinase reaction.

31P NMR spectroscopy can be safely and successfully used to study cerebral energy metabolism in sick newborn infants, provided a suitable life support and monitoring system is available. Studies of normal infants have shown characteristic spectral peaks assigned to adenosine triphosphate, phosphocreatine, phosphodiesters plus phospholipids, inorganic orthophosphate, and a large peak at the phosphomonoester resonance frequency that is attributable mainly to phosphoethanolamine, a precursor of membrane phospholipids. Both the PCr/Pi and PCr/ATP ratios were lower in normal newborn infants than reported in studies using invasive techniques in adult animals. Tissue hypoxia-ischemia was associated with a fall in PCr levels and a rise in Pi. The PCr/Pi ratio, which is an index of cerebral energy status, was normal during the first day of life in infants who had suffered severe birth asphyxia, but subsequently fell to levels well below the normal range. The presence of this latent period raises the possibility of effective early treatment following birth asphyxia, before irreversible impairment of cellular energy metabolism occurs. Low PCr/Pi ratios have also been demonstrated in infants with idiopathic cerebral infarction, periventricular leukomalacia and inborn errors of metabolism. Preliminary experience suggests that ratios below 0.8 are associated with irrecoverable failure of energy metabolism and cellular necrosis. 31P NMRS and other spectroscopic NMR techniques utilizing, for instance, 1H and 13C nuclei, have exciting potential for exploring the mechanisms and methods of prevention of various forms of cerebral damage in the newborn period.

Phosphorus magnetic resonance spectroscopy (MRS) and near infrared spectroscopy (NIRS) have been used to study the brains of normal newborn infants and infants with cerebral disorders admitted to a neonatal intensive care unit. MRS, which involves transporting the infant to the spectrometer, allows measurement of mobile phosphorus compounds such as adenosine triphosphate and phosphocreatine in brain tissue, and has been performed on over 160 babies. NIRS gives cotside information about cerebral oxygenation and haemodynamics and has recently been introduced. These techniques, especially when used together, show promise of providing important information about the mechanisms and prognostic significance of hypoxic-ischaemic damage to the brain--the most important cause of permanent neurodevelopmental disabilities in infants who require intensive care.

Intracellular metabolism in the brains of seven infants, born at 33-40 weeks' gestation and aged 44 h to 17 days, was studied on fourteen occasions by phosphorus nuclear magnetic resonance spectroscopy (31P NMRS). The characteristic spectral peaks of ATP, phosphocreatine (PCr), phosphodiesters, and inorganic orthophosphate (Pi) were always detected, together with a large peak attributed mainly to ribose-5-phosphate. The ratio of PCr to Pi NMRS signals (which are related to concentration) in one infant thought to have a normal brain was 1.7. In three infants who had severe birth asphyxia the PCr/Pi ratio ranged from 0.2 to 1.0 but increased as their clinical condition improved: infusions of mannitol solution caused a rapid increase in the ratio on four occasions in two of these infants. The PCr/Pi ratio was 1.4 in an infant with congenital cerebral atrophy and 0.7 in an infant with meningitis. Grossly abnormal 31P spectra antedated the detection by ultrasound of large porencephalic cysts in two infants. No systematic changes in intracellular pH (calculated from the chemical shift of the Pi resonance) were observed: the mean value for all observations was 7.2 +/- SD 0.1 (n = 14).

The brains of 158 consecutively admitted very preterm infants were repeatedly examined with real time ultrasound. Abnormalities, most commonly periventricular haemorrhage, were detected in 79 (50%). The 109 infants who survived were followed up until they were 16-23 months old. Major or minor neurological or developmental sequelae were found in 5 of 62 infants (8%) with normal ultrasound scans and in an identical proportion, 2 of 25 infants (8%), with uncomplicated periventricular haemorrhage. By contrast, 15 of 21 infants (71%) whose ventricles became enlarged (with or without periventricular haemorrhage) had abnormalities at follow up. The proportion with sequelae depended on the cause and extent of the enlargement. Three of 8 infants (38%) with mild (usually transient) ventricular distension had sequelae, compared with 3 of 4 (75%) with hydrocephalus and 9 of 9 (100%) with cerebral atrophy (2 of whom also had hydrocephalus). Adverse neurodevelopmental sequelae at follow up appeared more often to be attributable to cerebral ischaemia and infarction than to periventricular haemorrhage.

Brain metabolism and intracellular pH were studied during and after episodes of ischaemia and hypoxia-ischaemia in lambs anaesthetised with sodium pentobarbitone. 31P and 1H magnetic resonance spectroscopy methods were used to monitor brain pHi and brain concentrations of Pi, phosphocreatine (PCr), beta--nucleoside triphosphate (beta NTP), and lactate. Simultaneous measurements were made of cerebral blood flow and cerebral oxygen and glucose consumption. Cerebral ischaemia sufficient to reduce oxygen delivery to 75% of control values was associated with a fall in brain pHi and increase in brain Pi. Progressively severe hypoxia-ischaemia was associated with a progressive fall in brain pHi, PCr, and beta NTP and increase in brain Pi. In two animals the increase in brain lactate during hypoxia-ischaemia measured by 1H nuclear magnetic resonance (NMR) could be quantitatively accounted for by the increased net uptake of glucose by the brain in relation to oxygen, but was insufficient to account for the concomitant acidosis according to previous estimates of brain buffering capacity. In four animals brain pHi, PCr, Pi, and beta NTP had returned to normal 1 h after the hypoxic-ischaemic episode. In one animal brain pHi had reverted to normal at a time when 1H NMR indicated persistent elevation of brain lactate.

Brain metabolism and intracellular pH were studied during and after episodes of incomplete cerebral ischaemia in lambs under sodium pentobarbitone anaesthesia. 31P and 1H magnetic resonance spectroscopy was used to monitor brain pHi and brain concentrations of inorganic phosphate (Pi), phosphocreatine (PCr), beta-nucleoside triphosphate (beta NTP), and lactate. Simultaneous measurements were made of arterio-cerebral venous concentration differences (AVDs) for oxygen, glucose, and lactate. Cerebral ischaemia was induced by a combination of bilateral carotid clamping and hypotension, and the acute effects of systemic administration of glucose and sodium bicarbonate were examined. The molar ratio of glucose to oxygen uptake by the brain (6G/O2) increased above unity during cerebral ischaemia. Statistically significant AVDs for lactate were not observed. Cerebral ischaemia was associated with a reduction in brain pHi PCr/Pi ratio, and an increase in brain lactate. No effect of arterial plasma glucose on brain lactate concentration or brain pHi was evident during cerebral ischaemia or in the postischaemic period. Administration of sodium bicarbonate systemically in the postischaemic period was associated with a rise in arterial and brain tissue PCO2. A fall in brain pHi occurred which was attributable in part to coincidental brain lactate accumulation. The increase in brain lactate measured by 1H nuclear magnetic resonance in vivo during ischaemia was insufficient to account for the change in buffer base calculated to have occurred from previous estimates of brain buffering capacity.

Phosphorus (31P) nuclear magnetic resonance spectroscopy was used to study intracellular metabolism in the brains of 6 normal newborn infants and 10 infants who had been asphyxiated during delivery. In the normal infants spectral peaks mainly attributable to adenosine triphosphate, phosphocreatine (PCr), phosphodiesters plus phospholipids, and inorganic orthophosphate (Pi) were always detected, together with an additional large peak in the phosphomonoester region indicating the presence of a metabolite or metabolites (probably largely phosphoethanolamine) which may be involved in rapid growth of the brain. In the asphyxiated infants, data obtained on the first day of life showed no differences from those in normal infants, but by the second to ninth days inverse changes in the concentrations of PCr and Pi had caused a significant reduction in PCr/Pi. This latency suggest the possibility of effective early treatment before irreversible metabolic damage sets in. Mean intracellular pH when PCr/Pi was minimal was 7.17 +/- 0.10. Values for PCr/Pi below 0.80 were associated with a very bad prognosis for survival and early neuro-developmental outcome.