Computerised analysis of fetal heart rate recordings in maternal type I diabetes mellitus Douglas Tincello a, * , Sarah White b , Stephen Walkinshaw c Objective 1. To study computerised cardiotocograph parameters from women with type I diabetes; 2. to examine the signi®cance of observed differences from the expected normal values. Design Prospective observational study in the third trimester of pregnancy. Setting The medical antenatal clinic of a tertiary referral centre. Population Twenty-six women with type I diabetes mellitus with a singleton pregnancy. Methods Computerised cardiotocograph recordings were made weekly from 28 to 39 weeks. Derived para- meters were compared with the published ®gures for uncomplicated singleton pregnancies. Details of maternal blood sugar, labour and delivery and neonatal outcome were recorded. Data were compared between groups according to the computerised analysis of the antenatal CTGs. Results One-hundred and thirty-one recordings were made with a median of ®ve per patient (range 1±12). 11.3% showed absent episodes of high variation compared with the expected value of 0.8%, a difference of 9.5% (95% CI 4.5-15.3). Differences in short term variation, basal heart rate, frequency of fetal movements and heart rate accelerations were also found which changed with gestation. Overall these changes represented a more immature form of fetal heart rate than that which would be expected. No relationship between the changes and adverse fetal outcome could be identi®ed. Conclusions Signi®cant differences exist in cardiotocographs in maternal type I diabetes compared with normal fetuses. The changes may represent a delay in fetal maturation. The analysis mode of the computer will register these as abnormal features, but there is no evidence that they are pathological. We would recommend that computerised analysis is not used to assess pregnancies complicated by maternal type I diabetes mellitus. INTRODUCTION Maternal insulin dependent diabetes mellitus (type I) poses particular risks for the fetus during pregnancy including fetal abnormalities (particularly cardiac malformations), macrosomia, and intrauterine death. Neonatal complications include polycythaemia, hypogly- caemia, hypocalcaemia, seizures and an increased inci- dence of infant respiratory distress syndrome. These complications are largely a consequence of poor maternal blood sugar control causing fetal hyperinsulinaemia. The risk of complications can be reduced signi®cantly by tight glycaemic control, both preconceptually and throughout pregnancy in the face of rising insulin requirement. As a consequence, despite a lack of evidence, many clinicians institute some form of fetal surveillance. This may include both ultrasound and cardi- otocography (CTG). The bene®t from regular CTG moni- toring remains unclear 1 and may be a result of considerable intra- and inter-observer error 2 . Dawes and Redman devised a computerised system of CTG analysis which eliminates observer variability and makes analysis much more reproducible 3 . Their system identi®ed several parameters which are markers of fetal wellbeing (Table 1) 4±8 . The computer CTG continuously compares the current recording with an internal database of normal data from 20,000 healthy recordings and self- terminates when suf®cient data have been recorded to con®rm fetal health (the analysis mode). After 60 minutes the recording is terminated regardless. A printed record is provided explaining in what areas the record is defective 9 . Computerised CTGs were ®rst introduced into the Liverpool Women's Hospital in 1992. Many diabetic women had recordings which failed the computer analy- sis on the basis of absent episodes of high variation. These are thought to be sensitive markers of fetal health 4 . This observation prompted us to conduct a retrospective review of 38 pregnancies 10 . We found signi®cant differ- ences in the proportion of patients with absent episodes of high variation, but no relationship with any outcome measure. 30.5% of CTGs from women with type I diabetes showed absent episodes compared with 0.8% of normal CTGs. q RCOG 2001 British Journal of Obstetrics and Gynaecology PII: S0306-5456(00)00208-4 British Journal of Obstetrics and Gynaecology August 2001, Vol. 108, pp. 853±857 www.bjog-elsevier.com * Correspondence: Dr D. Tincello, University Department of Obstetrics and Gynaecology, Liverpool Women's Hospital, Crown Street, Liverpool L8 7SS, UK. a University Department of Obstetrics and Gynaecology, University of Liverpool, UK b Clinical Sciences Centre for Research and Education, University Hospital Aintree, Liverpool, UK c Liverpool Women's Hospital, Liverpool, UK