Different aspects of electronic fetal monitoring during labor
Background: Cardiotocography (CTG) is a tool to assess fetal well-being during labor and to detect early signs of fetal distress and thereby enable timely interventions to reduce neonatal morbidity and mortality. CTG is associated with shortcomings; poor reliability in interpretation, low specificity with a high proportion of false positive tracings indicating fetal distress when not accurate, no proven effect on rare severe outcomes such as mortality and cerebral palsy, but rather contributing to an increased risk of operative delivery. The aims of this thesis was to determine I) if an extended CTG education could lead to better reliability in interpretation compared to a national standard education, II) if a computerized algorithm could be developed with precision in detecting and quantitating decelerations on CTG, III) if deceleration area was a better predictor of fetal acidemia during labor than deceleration depth and duration, IV) the proportion of fetuses with undetected small for gestational age (SGA) in a low-risk population, comparing women that present with normal CTG at admission to labor (admCTG) to those with abnormal admCTG and to compare neonatal outcomes in the two groups stratified on SGA or non-SGA.
Material and methods: The CTG tracings used in paper I-III were extracted from a previous cohort of women in labor, from Karolinska University Hospital, Sweden. All women had undergone fetal blood sampling (FBS) during labor due to suspicious CTG patterns. Six obstetricians from two different hospitals were used as observers in paper I. Inter- and intra-observer reliability using Cohen’s and Fleiss kappa was determined for different parameters assessed on CTG. In paper II two obstetricians visually analyzed CTG tracings with variable decelerations and specified duration, depth and area for each deceleration. The computerized algorithm analyzed and quantified the same CTG traces and was compared to the observers using intra-class correlation and Bland-Altman analysis. In paper III the predictive value of deceleration area, duration, and depth for fetal acidemia, measured as lactate concentration at FBS, was explored using receiver operating characteristics, area under curve (ROC AUC). In paper IV, a register-based study, the risk of SGA in relation to the result of admCTG, normal vs abnormal was assessed in low-risk pregnancies. Neonatal outcomes were also determined by multiple logistic regression analysis.
Results: I) The inter- and intra-observer reliability was moderate to excellent at both departments, kappa 0.41-0.93. The department with extended education reached significantly higher interobserver agreement for two of six CTG parameters assessed. II) Computerized assessment of decelerations on CTG compared to visual observers reached excellent intraclass correlation (0.89-0.95) and low bias in Bland-Altman analysis, comparable to that between the two observers. III) The deceleration measures with the best prediction of fetal acidemia was cumulative deceleration area and duration, ROC AUC 0.682 and 0.683 respectively compared to deceleration depth 0.631. IV) The proportion of SGA was two-fold higher among neonates presenting with abnormal admCTG (18.6%) compared to normal admCTG (9.7%). The risk of composite severe adverse neonatal complications was substantially higher in the group with abnormal admCTG/SGA compared to normal admCTG/non-SGA, adjusted odds ratio 23.7 (95% confidence interval 9.8-57.3).
Conclusion: Inter- and intra-observer agreement was better than expected at both departments studied and extended education might have an impact on interpretation reliability. A novel computerized algorithm for CTG assessment has high precision in detecting and quantifying decelerations. Cumulative deceleration area and duration are better predictors of fetal acidemia than deceleration depth. In presumed low-risk pregnancies there is a group of undetected SGA fetuses that more often present with abnormal admCTG and are at higher risks of neonatal complications.
List of scientific papers
I. Gyllencreutz E, Hulthen Varli I, Lindqvist PG, Holzmann M. Reliability in cardiotocography interpretation - impact of extended on-site education in addition to web-based learning: an observational study. Acta obstetricia et gynecologica Scandinavica. 2017;96(4):496-502.
https://doi.org/10.1111/aogs.13090
II. Gyllencreutz E, Lu K, Lindecrantz K, Lindqvist PG, Nordstrom L, Holzmann M, et al. Validation of a computerized algorithm to quantify fetal heart rate deceleration area. Acta obstetricia et gynecologica Scandinavica. 2018;97(9):1137-47.
https://doi.org/10.1111/aogs.13370
III. Gyllencreutz E, Varli IH, Lindqvist PG, Holzmann M. Variable deceleration features and intrapartum fetal acidemia - The role of deceleration area. Eur J Obstet Gynecol Reprod Biol. 2021 Dec;267:192-197.
https://doi.org/10.1016/j.ejogrb.2021.11.009
IV. Gyllencreutz E, Hulthén Varli I, Johansson K, Lindqvist P G, Holzmann M. Admission cardiotocography in low-risk pregnancies and neonatal outcomes, the significance of undetected small-for-gestational age fetuses: a registerbased study. [Submitted]
History
Defence date
2022-12-02Department
- Department of Women's and Children's Health
Publisher/Institution
Karolinska InstitutetMain supervisor
Holzmann, MalinCo-supervisors
Lindqvist, Pelle; Hulthén Varli, IngelaPublication year
2022Thesis type
- Doctoral thesis
ISBN
978-91-8016-807-6Number of supporting papers
4Language
- eng