Sarah C Hiscock1, Martyn J Evans, Rosemary J Morton, David O Hall. 1. Department of Medical Physics and Bioengineering, Royal United Hospital, Bath , Guildford Medical Physics and Bioengineering, Bristol General Hospital, UK. sarah.hiscock@ruh.nhs.uk
Abstract
BACKGROUND: An investigation has been undertaken to calculate normal ranges for left ventricular ejection fraction (LVEF) measured by the gated blood pool (GBP) technique. A common set of normal studies was used at 11 hospitals within the south of England to assess the variability of results and normal ranges. METHODS: Normal studies were identified by retrospective review of patients who had undergone a GBP study and echocardiogram at the Royal United Hospital, Bath. Patients who had left ventricular function qualitatively identified as normal on echocardiogram and normal wall motion for the GBP were included. In total, 64 datasets were found to match the criteria. All the studies were made anonymous prior to being distributed to the participating hospitals. The upper and lower limits for normal ejection fraction were defined for each system using the 95% confidence limits around the mean value, before and after normalizing the results to remove systematic differences between the processing systems. RESULTS: The lower cut-off for normal function varied between 40 and 51%. Analysis of the individual operator results gave an inter-operator standard deviation of 4.2 and an intra-operator standard deviation of 2.7. CONCLUSIONS: It is recommended that all studies should be processed at least twice and an average taken to minimise these sources of uncertainty. If there is a significant difference the study should be reprocessed. Due to random differences between results from different systems it is suggested that an equivocal range be used between clearly normal and abnormal function.
BACKGROUND: An investigation has been undertaken to calculate normal ranges for left ventricular ejection fraction (LVEF) measured by the gated blood pool (GBP) technique. A common set of normal studies was used at 11 hospitals within the south of England to assess the variability of results and normal ranges. METHODS: Normal studies were identified by retrospective review of patients who had undergone a GBP study and echocardiogram at the Royal United Hospital, Bath. Patients who had left ventricular function qualitatively identified as normal on echocardiogram and normal wall motion for the GBP were included. In total, 64 datasets were found to match the criteria. All the studies were made anonymous prior to being distributed to the participating hospitals. The upper and lower limits for normal ejection fraction were defined for each system using the 95% confidence limits around the mean value, before and after normalizing the results to remove systematic differences between the processing systems. RESULTS: The lower cut-off for normal function varied between 40 and 51%. Analysis of the individual operator results gave an inter-operator standard deviation of 4.2 and an intra-operator standard deviation of 2.7. CONCLUSIONS: It is recommended that all studies should be processed at least twice and an average taken to minimise these sources of uncertainty. If there is a significant difference the study should be reprocessed. Due to random differences between results from different systems it is suggested that an equivocal range be used between clearly normal and abnormal function.