Table 4:
Summary of the literature on identification of IUGR fetuses using FHR parameters is presented in chronological order. Fetal magenetocardiogram (fMCG) and electrocardiogram (fECG) studies have been included as well as 1D-DUS for comparison.
Study | Modality | Method | Database | Main findings |
---|---|---|---|---|
Stroux et al. (2017); Stroux & Clifford (2016) | 1D-DUS/ CTG | Multiparameter behavioral state dependent metrics: LTV and STV averaged over high or low variability episodes, no. and average duration of high and low variability episodes, no. of minutes in high or low variability, onset of the first high variability episode, and estimated GA | 1163 IUGR and 1163 control cases at 23–42 weeks of gestation | LTV in active sleep was superior to STV (AUC of 72% vs. 71%). The number of minutes in high variation per hour (AUC of 75%) was the most predictive. The combined model improved the performance to 76%. |
Magenes et al. (2014) | 1D-DUS/CTG | multivariate: time domain (Rcov, STV, LTV), regu- larity/complexity (ApEn, Lempel Ziv, SE) and PRSA (APRS, DPRS) | 60 IUGR at 32.27 ± 2.79 weeks and 60 control cases at 34.78 ± 0.53 weeks of gestation | LR performed on ApEn, LTI, LZC and RCO achieved 92.5% accuracy of IUGR detection, with 93% sensitivity and 91.5% Specificity |
Gon¸calves et al. (2013) | 1D-DUS/CTG | Linear and entropy methods: mean FHR, LF, HF and MF, LF/(MF+HF) ApEn, SampEn, MSE. | 15 severe IUGR fetuses at 28–37 gestation weeks and 18 controls at 29–38 gestation weeks | significantly lower mFHR was only evident in IUGR males and lower entropy in IUGR females. Lower LF/(MF+HF) for IUGR females but not for males. Better detection of IUGR for male fetuses. |
Hoyer et al. (2013) | fMCG | FABAS: a multivariate model including: amplitude, skewness, generalized MSE, pNN5 and VLF/LF | 428 normal (113 quiet sleep, 286 active sleep, 29 active awake), and 19 IUGR cases, at 21- 40 gestation weeks | Classification of quiet and active sleep states (93.1%) and reduced fABAS for 11 IUGR fetuses preselected in active sleep. |
Fanelli et al. (2013) | 1D-DUS/CTG | PRSA, DPRS depending on the slope sign of the PRSA curve. | 61 IUGR and 61 control cases at 3435 weeks of gesta- tion | better discrimination by APRS (AUC=0.823) and DPRS (AUC=0.837) than STV (AUC=0.816). Significantly different STV, Delta, LTV, for IUGR. |
Lobmaier et al. (2012) | 1D-DUS/CTG | AAC to assess the dynamic capacity of the fetal ANS, and STV | 39 IUGR and 43 control cases, at 2638 weeks of gesta- tion | AAC differentiates better than STV, with higher AUC (97% vs. 85%), PPV (90% vs. 77%) and NPV (90% vs. 81%) |
Graatsma et al. (2012) | fECG | STV, AAC, ADC using PRSA | 30 small for GA, at 27–36 weeks and 90 control fetuses, 2140 gestation weeks | In small fetuses, both AAC and ADC z-scores were lower than the STV z-scores. |
Huhn et al. (2011) | 1D-DUS/CTG | Transformed PRSA, AAC | 74 IUGR and 161 normal cases at 2836 gestation weeks | Lower AAC and STV for IUGR. AUC of 81.4% for AAC and 70.5% for STV. |
Buscicchio et al. (2010) | 1D-DUS/CTG | baseline FHR, no. of small and large accelerations, no. of decelerations, duration of high and low variation in minutes, LTV, STV, no. of fetal movements per hour | 100 gestational diabetes cases on diet therapy and 100 on insulin therapy, 100 gestational hypertention, 100 IUGR, 100 premature rupture of membranes, 100 controls, all 35–36 gestation weeks | Baseline FHR, the duration of episodes of low variation and STV were significantly higher in all abnormal cases than in controls; significantly reduced fetal movement for IUGR, hypertention and premature rupture of membranes. |
Ferrario et al. (2009b) | 1D-DUS/CTG | time domain and frequency domain FHRV, and complexity parameters: ApEn, SampEn, MSE, LZC, DFA | 25 recordings from 6 IUGR cases, at 28–34 weeks, 4 subjects (13 recordings) with altered fluximetric indices | IUGR cases without fluximetry alterations, had reduced HRV amplitude and regularity, lower spectral components and complexity. |
Ferrario et al. (2006, 2009a) | 1D-DUS/CTG | LZC and MSE with k-mean cluster analysis | 23 severe IUGR, 19 non-severe IUGR and 17 control fetuses, at 27 to 34 weeks of gestation | LZC and MSE are significantly different for severe IUGR vs. non-sever and control (Se=77.8% and Ac=82.4%). |
Kikuchi et al. (2008) | 1D-DUS/CTG | DFA | 68 IUGR fetuses, at 24 to 40 weeks, and 119 control fetuses at 22 to 41 weeks of gestation | α2 exponent values of IUGR were significantly higher than control |
Serra et al. (2008) | 1D-DUS/CTG | STV | 257 IUGR cases within 24 hours of delivery (26–42 weeks) | Decreasing STV was correlated with earlier deliveries and worse postnatal outcome. |
Ferrario et al. (2007) | 1D-DUS/CTG | Time and frequency domain FHRV, LZC, ApEn, SampEn | 23 severe IUGR, 19 non-severe IUGR and 17 control fetuses, at 27 to 34 weeks of gestation | Only LZC, DELTA and STV were discriminative, no improvement by adding ApEn and SampEn |
Schneider et al. (2006) | fMCG | linear and nonlinear FHRV parameters | 36 IUGR and 29 control fetuses, at 28 to 39 weeks of gestation | Significantly lower SDNN, RMSSD, TP and LF/HF for IUGR |
Kikuchi et al. (2006) | 1D-DUS/ CTG | Nonlinear FHRV: attractor reconstruction, largest Lyapunov exponents and correlation dimension | 69 IUGR fetuses, at 24 to 40 weeks, and 119 control fetuses at 22 to 41 weeks of gestation | Decreased variability, less chaotic FHR dynamics and decreased complexity for IUGR. |
Anastasiadis et al. (2003) | fMCG | chaotic and periodic heart rate dynamics | 11 IUGR and 19 control fetuses at 34 to 37 weeks of gestation | Significantly lower correlation dimension and higher LF, HF powers for IUGR. |