G381(P) Accuracy of neonatal hypoglycaemia measurement – blood gas analyser vs hand held glucose measurement device – a tricky situation!

Abstract

Introduction Hypoglycaemia is a common reason for admission for term and near term infants to neonatal units (NICU). Early recognition and management of neonatal hypoglycaemia has important clinical and economic consequences. In our hospital neonatal blood sugar is measured using both a standard ‘blood gas analyser’ and a branded hand-held glucose measurement device (HHGMD), and there is tendency to trust the HHGMD result in preference to blood gas result. Aim We aimed to compare the blood glucose results of both the blood gas analyser and HHGMD on same blood sample at same time to ensure no significant difference was detected. Methods Blood sugar results in individual neonates were recorded using both the HHGMD and blood gas analyser. We obtained basic demographics and route of sample for all babies. A quality control measure for the HHGMD was carried out most days by the biochemist and a calibration of the blood gas machine was done daily. Data are mainly descriptive; categorical variables are described as percentages with binomial 95% confidence intervals (95% CI). Wilcoxan Matched-Pairs Signed Ranks Test was used to test for statistical significance. Results There were 118 separate paired blood gas and HHGMD results logged from 32patients in the 8 week period of this study, of which 53% (n=17/32) were term and 47% (n=15/32) were preterm babies. One third (67%, n=79/118) of these samples were preterm. Only 0.3% (n=4/118) samples were taken from babies not currently admitted to NICU. 3% (n=4) of cases both HHGMD/blood gas results were same while in 6% (n=7) of cases HHGMD result was lower than the blood gas result. However in 107 cases HHGMD result was higher than blood gas result with a mean difference of 1.07 mmol (95% CI 0.94–1.15, p Conclusions A mean difference of approximately 1 mmol/L glucose between both measurement devices on the same blood sample has potential clinical and economic significance particularly in our population where hypoglycaemia and subsequent neonatal unit admission is common. A more robust way of standardising glucose measurement is required in this cohort of babies.