Presented at the Neonatal Society 2003 Autumn Meeting.
Mallaiah R, Counsell SJ, Rutherford MA, Maalouf E, Edwards D
Robert Steiner MR Unit, Imaging Sciences Department and Division of Paediatrics, Obstetrics & Gynaecology, Imperial College, Hammersmith Hospital, Du Cane Road London W12 OHS.
Introduction: The contrast in brain MRI is primarily influenced by water and fat content, but can also be affected by the presence of paramagnetic substances such as manganese (Mn), iron and copper. MRI studies in adults and older children receiving Mn as part of long term total parenteral nutrition (TPN) have demonstrated hyperintense signals in the globus pallidus of the basal ganglia thought to be secondary to Mn deposition. This is a potential source of image misinterpretation, and high signal in the globus pallidus might be taken to imply cerebral injury. It has been suggested that images should not be obtained while patients are receiving TPN.
Aim: To determine whether administration of TPN to premature infants is associated with an increase in MR signal intensity in the globus pallidus.
Patients and methods: We examined 25 MR brain images, obtained between 25 and 36 weeks gestational age from 20 preterm infants born at gestational age of 24 to 29 weeks (median 27 weeks) with birth weight of 590 to 1600 g (median 960 g). These infants were receiving TPN on the day of imaging, and for median of 8 days and a range of 1 to 46 days previously; the quantity of TPN administered each day was determined by clinical protocols. A further 22 images form 21 infants who had never received TPN, of gestational age 25 to 29 weeks (median 28 weeks) and birthweight of 580 to 1444 g (median 928 g) were also analysed.
T1 weighted MR images were analysed. The signal intensity from the right and left globus pallidus were measured and normalised by reference to cerebrospinal fluid (CSF) signal intensity; both values were used for statistical analysis.
Data were log transformed to normality and signal intensity in infants receiving TPN compared with those who never received TPN using Students t test. To allow for confounding variables a multiple linear regression model was then constructed which accounted for: repeated study in some infants, gestational age of infant at scan; right and left side; number of days of TPN administration.
Results: There was no significant difference in normalized signal intensity in the globus pallidus between infants who had or had not received TPN. Linear regression showed that signal intensity varied with gestational age, but there was no significant relation between signal intensity and days of TPN administration. There was no difference between left and right sides of the brain.
Conclusion: Using our archive of images we could not show that administration of TPN altered the intensity of the MR signal in the globus pallidus.