Early Human Development 90 (2014) 81–85

Contents lists available at ScienceDirect

Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev

Quantitative comparison of entropy analysis of fetal heart rate variability related to the different stages of labor Jongil Lim a,1, Ji Young Kwon a,1, Juhee Song b,c, Hosoon Choi d, Jong Chul Shin a, In Yang Park a,⁎ a

Department of Obstetrics and Gynecology, The Catholic University of Korea, Seoul, South Korea Department of Biostatistics, Scott & White Health Care, TX, USA Department of Medicine, Texas A&M Health Science Center, TX, USA d Institute for Regenerative Medicine, Texas A&M Health Science Center, TX, USA b c

a r t i c l e

i n f o

Article history: Received 31 December 2012 Received in revised form 13 November 2013 Accepted 16 December 2013 Keywords: Approximate entropy Sample entropy Labor progression Fetal heart rate Cardiotocography

a b s t r a c t Background: The interpretation of the fetal heart rate (FHR) signal considering labor progression may improve perinatal morbidity and mortality. However, there have been few studies that evaluate the fetus in each labor stage quantitatively. Aim: To evaluate whether the entropy indices of FHR are different according to labor progression. Study design: A retrospective comparative study of FHR recordings in three groups: 280 recordings in the second stage of labor before vaginal delivery, 31 recordings in the first stage of labor before emergency cesarean delivery, and 23 recordings in the pre-labor before elective cesarean delivery. Subjects: The stored FHR recordings of external cardiotocography during labor. Outcome measures: Approximate entropy (ApEn) and sample entropy (SampEn) for the final 2000 RR intervals. Results: The median ApEn and SampEn for the 2000 RR intervals showed the lowest values in the second stage of labor, followed by the emergency cesarean group and the elective cesarean group for all time segments (all P b 0.001). Also, in the second stage of labor, the final 5 min of 2000 RR intervals had a significantly lower median ApEn (0.49 vs. 0.44, P = 0.001) and lower median SampEn (0.34 vs. 0.29, P b 0.001) than the initial 5 min of 2000 RR intervals. Conclusions: Entropy indices of FHR were significantly different according to labor progression. This result supports the necessity of considering labor progression when developing intrapartum fetal monitoring using the entropy indices of FHR. © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Fetal heart rate (FHR) analysis is one of the important methods for evaluating the fetal condition. However, despite the widespread use of electronic FHR monitoring, its effect in decreasing fetal mortality and morbidity has not been established [1]. To reduce inter- and intraobserver variability of visual analysis, computerized analysis was developed, but it did not result in significant clinical improvement [2]. Accordingly, there have been many efforts to develop new monitoring methods more responsive in differentiating between normal and pathological fetal conditions. Entropy analysis that measures the correction and persistence of a signal is a nonlinear mathematical approach to quantify the irregularity and complexity of a system [3]. Entropy analysis of heart rate is based on a systematical biological theory that suggests weak connections

⁎ Corresponding author at: Department of Obstetrics and Gynecology, College of Medicine, The Catholic University of Korea, Seoul St. Mary's Hospital, #505 Banpodong, Seochogu, 137-040, Seoul Korea. Tel.: +82 222 582 813; fax: +82 259 51549. E-mail address: [email protected] (I.Y. Park). 1 Jongil Lim and Ji Young Kwon contributed equally to this work. 0378-3782/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.earlhumdev.2013.12.007

between systems or within a system that are associated with the mechanism of disease [4]. Approximate entropy (ApEn) was considered to provide a measurement of feedback and regularity, and a time series containing many repetitive patterns has relatively low ApEn, and a less predictable process has higher ApEn [5]. Since it has been found that the underlying mechanisms involved in the control of the FHR are mainly nonlinear [4], several studies have presented the use of nonlinear analyses to characterize the presence of nonlinear features in FHR variability [6–9]. Li et al. [9] suggested that the lower ApEn of FHR was associated not only with fetal distress and hypoxia, but also with respiratory and metabolic acidosis in women at term pregnancy. During labor, fetuses suffer considerable stress due to uterine contraction, repeated circulation insufficiency and resultant hypoxic environment, and head compression in the passage through the pelvic cavity. However, labor does not appear to impair fetal well being significantly, maybe due to a fetal compensation mechanism to stress. For intrapartum FHR monitoring, it is essential to understand the normal pattern of the nervous system regulating the stability of cardinal rhythm under such conditions of physiologic stress during labor. The accurate interpretation of the FHR signal according to labor progression may reduce unnecessary cesarean delivery and improve perinatal

82

J. Lim et al. / Early Human Development 90 (2014) 81–85

morbidity and mortality. However, there have been few studies that evaluate the fetus in each labor stage quantitatively. Most of studies that used linear and nonlinear parameters of FHR variability for fetal monitoring focused on fetuses in specific stages of labor, and did not analyze those parameters according to the stages of labor [5–8]. Thus, this study was conducted to see whether the entropy indices of FHR are significantly different according to labor progression. For this purpose, we compared the entropy indices of FHR in fetuses at different stages of labor using FHR recordings just before delivery of fetuses who were delivered in three different modes: vaginal delivery in the second stage of labor, emergency cesarean delivery in the middle of the first stage of labor, and planned cesarean delivery in pre-labor. The results of this study are expected to be useful for understanding the characteristics of the cortical nervous controlling system of FHR with labor progression and for improving the clinical application of entropy indices for intrapartum fetal monitoring.

2. Materials and methods We analyzed the stored and digitalized FHR recordings of external cardiotocography measured before delivery at a tertiary hospital. Between August 2004 and December 2006, consecutive cases of vaginal delivery were enrolled. Also, between June 2006 and December 2006, consecutive cases of cesarean delivery, both elective and emergency, were enrolled. Elective cesareans were performed in pre-labor and emergency cesareans were performed in the first stage of labor with an indication of dystocia. Cesarean deliveries with an indication of fetal distress were excluded in this study. Among deliveries of singleton fetuses N37 weeks of gestation, we included fetuses with at least a half an hour of records of cardiotocographic fetal monitoring before delivery. Fetuses with intrauterine growth restriction or major congenital anomalies were excluded. As a result, 450 consecutive deliveries were enrolled in this study. After excluding 116 recordings with missing data of a large segment over 15%, finally, all 334 recordings were included for this analysis. Based on delivery modes and stages of labor, all recordings were divided into 3 groups: recordings in the second stage of labor before vaginal delivery, recordings in the middle of the first stage of labor before emergency cesarean delivery, and recordings in the prelabor before elective cesarean delivery. There were 280, 31, and 23 recordings in the vaginal delivery group, emergency cesarean group, and elective cesarean group, respectively. This study was approved by the Institutional Review Board of the Catholic University of Korea.

2.2. Entropy analysis Approximate entropy (ApEn) quantifies the complexity of FHR variability. Low ApEn values correspond to lower complexity while higher ApEn indicates higher complexity. All mathematical equations have been described in detail elsewhere [3,6]. Briefly, the algorithms are summarized as follows. A time series of size (i.e. the number of RR in the input sequence) N, x ¼ x1 ; x2 ; ……xN : We select an embedding dimension m (sometimes also referred in the literature as pattern length) and we construct a new vector series: ! ! ! x N−mþ1 x ¼ x 1 ; x 2 ; ……! where ! xi

  ¼ xi ; xiþ1 ; xiþ2 ……xiþm−1 :

! The size of the new time series x is N − m + 1 and the size of each vector is m. We also select a threshold distance or comparison length r. ! x i and x j of size m is smaller than r when: The distance of two vectors !     xiþk −x jþk br

0≤k ≤m−1

in which case we consider the vectors to be similar. Given the distance r, the probability of a vector ! x i of size m of being similar with a vector ! x j of the same size is N−mþ1 X m C i ðr Þ

¼

Θði; j; m; rÞ

j¼1

N−m þ 1 (

Θði; j; m; r Þ ¼

    xiþk −x jþk br

1 0

Otherwise

We define as N−mþ1 X m

Φ ðr Þ ¼

m

lnC i ðr Þ

i¼1

N−m þ 1

:

The approximate entropy is expressed as 2.1. Signal acquisition and pre-processing

m

mþ1

ApEnðm; r; NÞ ¼ Φ ðr Þ−Φ For FHR signal acquisition, a Corometrics 150 (Corometrics, CT, USA), and a Doppler ultrasound cardiotocography with an autocorrelation function were used. Pulse repetition frequency of 2 Hz, pulse duration of 92 μs, and heart rate counting range of 50–210 bpm were used. All records were stored in the linked personal computer for further offline analysis. We digitalized the FHR recordings of the last 30 min before delivery of the study group through the Catholic computer-assisted obstetric diagnosis system (CCAOD; DoBe Tech, Seoul, Korea). FHR signals measured during the last minutes of delivery are likely to be lost or contaminated. A previously published pre-processing algorithm [10] was utilized for signal pre-processing. Heart beats lower than 60 beats per minute (bpm) and beat-to-beat differences of higher than 25 bpm were identified and filtered. Spline interpolation was utilized to substitute for the filtered beats for periods of signal loss of 2 s or less. Longer periods were replaced with the most recent segment of equal length with no signal loss. Finally, all heart rate data in bpm were converted into RR intervals for entropy calculations. Heart rate (bpm) can be expressed as 60/RR interval (second) [11,12].

ðr Þc:

As explained in the algorithms, three parameters were used in computing the ApEn value: the embedding dimension (m), the comparison length (r), and the number of RR in the input sequence (N). One of the disadvantages in ApEn is that ApEn depends on the number of input sequences. Additionally, the value of 1000 input sequences has been suggested to be sufficient for statistical validity with m = 2 and r ranging from 0.1 to 0.2 (i.e. 10% to 20% of the standard deviation (STD) of the input sequence, RR) [3]. The number of RR intervals as input sequence was 2000 in this study. The embedding dimension (m) was empirically set to 2 and the comparison length (r) was calculated for each data as 15% of the STD (i.e. r = 0.15) of RR intervals [3]. In order to confirm our ApEn calculations, another nonlinear measure, SampEn (Sample Entropy), was provided and compared with the results of ApEn. The algorithm of SampEn has been explained in detail elsewhere [13]. As mentioned earlier, in ApEn calculations, a pre-processing algorithm based on previous studies was implemented on the FHR signals, with conversion to RR intervals [12]. The final 2000 consecutive RR intervals (16.7 min), the initial 600 consecutive RR intervals (5 min) of the 2000

J. Lim et al. / Early Human Development 90 (2014) 81–85

83

Table 1 Maternal and fetal characteristics according to the stage of labor and delivery mode. Variable

Vaginal delivery group (The second stage of labor) (N = 280)

Emergency cesarean group (The first stage of labor) (N = 31)

Elective cesarean group (Pre-labor) (N = 23)

P-value

Maternal age (years) Gestational age (weeks) Gender (male) Birth weight (g) Apgar score 1 min ≤ 6 5 min ≤ 7 Umbilical arterial blood pH Hco3 Base excess (mmol/L)

31.0 ± 3.2 39.3 ± 1.1 148 (52.86%) 3296 ± 368

32.5 ± 4.3 39.7 ± 1.2 18 (58.06%) 3495 ± 374

32.4 ± 3.4 38.8 ± 1.3 14 (60.87%) 3220 ± 379

0.008 0.016 0.674 0.009

7 (2.5%) 3 (1.1%)

2 (6.5%) 0 (0.0%)

1 (4.3%) 1 (4.3%)

0.260 0.287

7.37 (7.3–7.4) 18.0 (0.2–23.3) 0.9 (−3.7–7.4)

7.36 (7.3–7.4) 11.1 (−1.2–22.5) 7.17(−3.8–7.4)

7.37 (7.32–7.39) 13.9(1.9–22.6) −0.45(−7.1–7.3)

0.407 0.890 0.410

All values are expressed as median (25%–75%), mean (±standard deviation) or number (%). P-values for comparison of three groups.

RR intervals, the middle 600 consecutive RR intervals of the 2000 RR intervals, and the final 600 consecutive RR intervals of the 2000 RR intervals were tested in the analysis of FHR variability. All calculations were performed using MATLAB 7.10 (R2010a, The MathWorks, Natick, MA). 2.3. Statistical analysis All maternal and fetal characteristics were summarized according to the stages of labor using descriptive statistics: mean (±STD) or median (Q1–Q3) for continuous variables and frequency (percent) for categorical variables. Three groups were compared in each of the characteristics. The Kruskal–Wallis test was utilized for comparisons between 3 groups in continuous variables, and the Chi-square test or Fisher's exact test was utilized for comparisons between 3 groups in categorical variables. Pairwise comparisons of 3 groups were done utilizing the Wilcoxon rank-sum tests with the Bonferroni adjusted α. A P-value of less than 0.05 indicated a statistical significance. SAS 9.2 (SAS Institute INC, Cary, NC) was used for statistical data analysis. 3. Results Maternal and fetal characteristics in the vaginal delivery group, emergent cesarean group, and elective cesarean group are presented in Table 1. Significant differences were not detected between the 3

groups in the 1-minute and 5-minute Apgar scores, median umbilical arterial pH and base deficit at delivery. One hundred and eight infants were males and 154 were females, and no significant gender difference was observed. Maternal age, gestational week, and fetal weight showed statistically significant differences between the 3 groups. However, the differences in mean values of each group seemed to be clinically insignificant. Neither SGA fetuses nor preterm fetuses were included in this study. In Table 2, linear indices including mean FHR, STD, coefficient of variation (CV), and dispersion index (DI) were significantly different between 3 groups in all time segments (2000 RR intervals before delivery, initial 5 min of 2000 RR intervals, and final 5 min of 2000 RR intervals) (all P b 0. 001 except mean FHR of the initial 5 min of 2000 RR intervals (P = 0.006)). Pairwise comparisons showed significant differences in linear indices for total 2000 RR intervals and final 5 min of 2000 RR intervals between the vaginal delivery group and emergency cesarean group and between the vaginal delivery group and elective cesarean group (all Bonferroni adjusted P b 0.001 except mean FHR comparison for total 2000 RR intervals between vaginal delivery and emergency cesarean groups (Bonferroni adjusted P = 0.040)). In addition, there were significant differences in linear indices for initial 5 min of 2000 RR intervals between the vaginal delivery group and emergency cesarean group and between the vaginal delivery group and elective cesarean group (all Bonferroni adjusted P b 0.05), except mean index for initial 5 min of 2000 RR intervals between the vaginal delivery group

Table 2 Linear statistics of RR intervals (seconds) according to the stage of labor in the final 2000 consecutive RR intervals preceding delivery, the initial 5 min of the 2000 RR intervals, and the final 5 min of the 2000 RR intervals.

Mean

STD

CV

DI

Variable

Vaginal delivery group (The second stage of labor) (N = 280)

Emergency cesarean group (The first stage of labor) (N = 31)

Elective cesarean group (Pre-labor) (N = 23)

P-value

2000 RR intervalsa,b Initial 5 mina Final 5 mina,b 2000 RR intervals a,b Initial 5 mina,b Final 5 mina,b 2000 RR intervalsa,b Initial 5 mina,b Final 5 mina,b 2000 RR intervalsa,b Initial 5 mina,b Final 5 mina,b

0.45 (0.42–0.47) 0.44 (0.41–0.47) 0.45 (0.42–0.49) 0.05 (0.03–0.08) 0.03 (0.02–0.05) 0.05 (0.03–0.08) 0.11 (0.07–0.16) 0.07 (0.04–0.11) 0.11 (0.06–0.17) 0.006 (0.002–0.013) 0.002 (0.001–0.006) 0.006 (0.002–0.015)

0.42 (0.39–0.43) 0.41 (0.40–0.43) 0.42 (0.40–0.44) 0.02 (0.02–0.04) 0.02 (0.01–0.03) 0.02 (0.01–0.04) 0.05 (0.04–0.08) 0.05 (0.03–0.07) 0.05 (0.03–0.10) 0.001 (0.001–0.003) 0.001 (0.000–0.002) 0.001 (0.000–0.004)

0.43 (0.41–0.44) 0.43 (0.41–0.45) 0.43 (0.40–0.45) 0.02 (0.02–0.03) 0.02 (0.01–0.02) 0.02 (0.01–0.03) 0.05 (0.04–0.08) 0.04 (0.03–0.06) 0.04 (0.03–0.07) 0.001 (0.001–0.003) 0.001 (0.000–0.001) 0.001 (0.000–0.002)

b0.001 0.006 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001

All values are expressed as median (25%–75%). P-values for comparison of the three groups using the Kruskal–Wallis test. STD; standard deviation, CV; coefficient of variation, and DI; dispersion index. a Significant difference was detected between vaginal delivery group and emergency cesarean group after Bonferroni-adjustment. b Significant difference was detected between vaginal delivery group and elective cesarean group after Bonferroni-adjustment.

84

J. Lim et al. / Early Human Development 90 (2014) 81–85

In the vaginal delivery group, there were significant differences between the initial 5 min of 2000 RR intervals and final 5 min of 2000 RR intervals in the median ApEn and SampEn. The final 5 min had a significantly lower median ApEn (0.49 vs. 0.44, P = 0.001) and lower median SampEn (0.34 vs. 0.29, P b 0.001) than the initial 5 min in the vaginal delivery group. No significant differences were detected between different time segments for the elective cesarean and emergency cesarean groups. Fig. 2. shows the change of median ApEn and median SampEn according to the labor progression during the 2000 RR intervals before delivery in the vaginal delivery group. This graph presented a decreasing trend in the median ApEn and the median SampEn as labor proceeded in the second stage of labor.

Fig. 1. The mean RR intervals according to the stages of labor in the final 2000 consecutive RR intervals preceding delivery. The median values (Q1–Q3) of the mean RR intervals (seconds) are shown for each stage of labor. There were significantly different between 3 stages of labor (P b 0.001). Pairwise comparisons showed significant differences between the pre-labor group and the second stage of labor group (Bonferroni adjusted P b 0.001), and between the first stage of labor group and the second stage of labor group (Bonferroni adjusted P = 0.040).

and elective cesarean group (Bonferroni adjusted P = 0.774). Fig. 1 showed the mean RR intervals according to the stages of labor. Particularly, in the vaginal delivery group, the final 5 min of 2000 RR intervals had significantly greater linear indices than the initial 5 min of 2000 RR intervals (all P b 0.001). Nonlinear indices including ApEn and SampEn for different delivery modes are presented in Table 3. The vaginal delivery group had the lowest median ApEn, followed by the emergency cesarean group and the elective cesarean group for all time segments (P b 0.001 for 2000 RR intervals, P = 0.037 for initial 5 min of 2000 RR intervals, and P = 0.002 for final 5 min of 2000 RR intervals, respectively). Also, the median SampEn showed the lowest value in the vaginal delivery group, followed by the emergency cesarean group and the elective cesarean group for all time segments (P b 0.001 for 2000 RR intervals, P = 0.026 for initial 5 min of 2000 RR intervals, and P = 0.001 for final 5 min of 2000 RR intervals, respectively). In addition, pairwise comparisons detected significant differences in the median ApEn for the 2000 RR intervals between the vaginal delivery group and emergency cesarean group and between the vaginal delivery group and elective cesarean group after Bonferroni correction (Bonferroni adjusted P = 0.002, Bonferroni adjusted P = 0.026, respectively). In the median SampEn for the 2000 RR intervals, significant differences between the vaginal delivery group and emergency cesarean group and between the vaginal delivery group and the elective cesarean group were detected as well (Bonferroni adjusted P b 0.001, Bonferroni adjusted P = 0.010, respectively).

4. Discussion This study compared the entropy indices of FHR in fetuses according to the stage of labor, and the results showed that the FHR in the second stage of labor showed significantly the lowest ApEn and SampEn of the labor course. In addition, the linear indices of FHR were significantly higher in fetuses in final labor just before vaginal delivery than in those in the first stage of labor and in those in pre-labor. As this study analyzed only fetuses in a favorable condition, the decrease in the entropy indices of FHR along with labor progression is believed to be a physiologic response of FHR to the stressful environment during labor. Differences in FHR indices according to labor progression were also observed in the results of the comparison of linear and nonlinear FHR indices between the initial 5 min of 2000 RR interval before delivery and the final 5 min of 2000 RR interval before delivery in the vaginal delivery group. In this study, the entropy indices were significantly lower and the linear indices were significantly higher during the final 5 min than during the initial 5 min. In the emergency cesarean group, on the contrary, the linear and nonlinear FHR indices were not significantly different between the initial 5 min and the final 5 min. This suggests that the significant decrease in entropy indices resulting from the progression of delivery takes place only during the second stage of labor. From a maternal and fetal perspective, in the second stage of labor, the fetal circulatory system is likely to be changed due to fetal vagal nerve activation and maternal placental insufficiency. Interestingly, our results based on FHR variability were in agreement with this physiologic change. The stress to fetus was increased according to labor progression, thus the median ApEn significantly decreased. This phenomenon probably arises due to the highly stressful situation during labor which decreases the complex regulatory activity of the central nervous system in fetuses. Favorable fetal conditions can be maintained despite the decrease in cortical nervous activity probably because of compensatory modulation by the autonomic nervous system. In a hypoxic situation severe enough to cause fetal acidemia, the cortical network integral state will go weaker and accordingly lower entropy

Fig. 2. The median (Q1–Q3) approximate entropy (ApEn) and sample entropy (SampEn) of fetuses in the vaginal delivery group at the initial 5 min, middle 5 min and final 5 min of the final 2000 RR intervals before delivery. Both ApEn and SampEn tended to decrease as delivery proceeded. The final 5 min of 2000 RR intervals had significantly lower median ApEn (0.49 vs. 0.44, P = 0.001) and lower median SampEn (0.34 vs. 0.29, P b 0.001) than the initial 5 min of 2000 RR intervals.

J. Lim et al. / Early Human Development 90 (2014) 81–85

85

Table 3 Nonlinear analysis using approximate entropy (ApEn) and sample entropy (SampEn) according to the stage of labor in the final 2000 consecutive RR intervals preceding delivery, the initial 5 min of the 2000 RR intervals, and the final 5 min of the 2000 RR intervals.

ApEn

SampEn

Variable

Vaginal delivery group (The second stage of labor) (N = 280)

Emergency cesarean group (The first stage of labor) (N = 31)

Elective cesarean group (Pre-labor) (N = 23)

P-value

2000 RR intervals a,b Initial 5 minb Final 5 minb 2000 RR intervals a,b Initial 5 minb Final 5 minb

0.43 (0.32–0.53) 0.49 (0.36–0.63) 0.44 (0.34–0.57) 0.26 (0.17–0.36) 0.34 (0.23–0.51) 0.29 (0.19–0.43)

0.52 (0.40–0.79) 0.53 (0.41–0.72) 0.54 (0.42–0.69) 0.34 (0.23–0.60) 0.39 (0.26–0.59) 0.39 (0.24–0.56)

0.65 (0.43–0.86) 0.60 (0.47–0.78) 0.60 (0.39–0.77) 0.41 (0.32–0.72) 0.45 (0.35–0.64) 0.44 (0.28–0.68)

b0.001 0.037 0.002 b0.001 0.026 0.001

All values are expressed as median (25%–75%). P-values for comparison of the three groups using the Kruskal–Wallis test. a Significant difference was detected between vaginal delivery group and emergency cesarean group after Bonferroni-adjustment. b Significant difference was detected between vaginal delivery group and elective cesarean group after Bonferroni-adjustment.

indices are expected, but this study did not cover the change of entropy indices according to fetal acidemia. To our knowledge, only one study investigated the changes in linear or nonlinear FHR indices according to labor progression. Gonvalves et al. [8] reported that progression towards the last minutes of labor was associated with a significant decrease in nonlinear indices. This is in line with our analysis. In particular, our study is distinguished from the previous one in that it examined the change of entropy indices of FHR through different stages of labor by comparing the entropy indices of FHR in the pre-labor and in the first and second stages of labor. The results of this study, which showed that the entropy indices of FHR were different according to the stage of labor, highlight that when we use linear and nonlinear FHR indices as intrapartum fetal monitoring tools, we should consider the labor progression. That is, the change of FHR indices caused by labor is not expected in the first stage of labor, and therefore if the entropy indices decrease at this stage, the possibility of intrinsic fetal pathology should be considered. On the contrary, in the second stage of labor, stressful external stimuli are maximal and therefore some decrease in the entropy indices of FHR may be regarded as a physiologic response. In future studies, the cutoff values of FHR indices for discriminating between normal fetuses and pathologic fetuses need to be presented according to the stage of labor. As this study retrospectively analyzed FHR just before delivery among fetuses delivered with different delivery modes, selection bias should be taken into account. However, the bias might be limited to some degree because there were no significant differences in fetal characteristics such as fetal gender, weight, umbilical arterial blood gas, Apgar scores and fetal anomaly between the groups. Since we selected the FHR of fetuses whose deliveries were performed during required stage of labor and analyzed the FHR just before delivery, whether vaginal or cesarean, each median value of entropy indices of FHR in each group of different delivery mode could reflect the specific stage of labor. More studies that analyze FHR of fetuses who were delivered vaginally according to the stage of labor will be helpful to understand the change of FHR during labor. In conclusion, this study showed that the entropy indices of FHR of fetuses in the second stage of labor were significantly lower than those in pre-labor and those in the first stage of labor. In addition, during the final 2000 RR intervals preceding vaginal delivery, fetuses in the final 5 min of labor had significant lower entropy indices of FHR compared with those in the initial 5 min of labor. These results help our understanding of the fetal physiologic response to the highly stressful situation of labor, and emphasize the necessity of establishing different

cutoff values of entropy indices according to the stage of labor. Further studies that investigate acidemic fetuses are required to develop intrapartum fetal monitoring tools using the entropy indices of FHR. Conflict of interest statement All authors do not have any financial and personal relationships with other people or organizations that could inappropriately influence (bias) this work. There are no potential conflicts of interest include employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and grants or other funding of any of the authors. Acknowledgements Financial support for this study was provided by a grant from Seoul Saint Mary Hospital 2012 research fund. References [1] ACOG Practice Bulletin No.106. Intrapartum fetal heart rate monitoring: nomenclature, interpretation, and general management principles. Obstet Gynecol 2009;114:192–202. [2] Costa A, Ayres-de-Campos D, Costa F, Santos C, Bernardes J. Prediction of neonatal acidemia by computer analysis of fetal heart rate and ST event signals. Am J Obstet Gynecol 2009;201(464):e1–6. [3] Pincus S. Approximate entropy (ApEn) as a complex measure. Chaos 1995;5:110–7. [4] Buchman TG. The community of the self. Nature 2002;420:246–51. [5] Pincus SM. Assessing serial irregularity and its implications for health. Ann N Y Acad Sci 2001;954:245–67. [6] Pincus SM, Viscarello RR. Approximate entropy: a regularity measure for fetal heart rate analysis. Obstet Gynecol 1992;79:249–55. [7] Signorini MG, Magenes G, Cerutti S, Arduini D. Linear and nonlinear parameters for the analysis of fetal heart rate signal from cardiotocographic recordings. IEEE Trans Biomed Eng 2003;50:365–74. [8] Gonçalves H, Rocha AP, Ayres-de-Campos D, Bernardes J. Linear and nonlinear fetal heart rate analysis of normal and acidemic fetuses in the minutes preceding delivery. Med Biol Eng Comput 2006;44:847–55. [9] Li X, Zheng D, Zhou S, Tang D, Wang C, Wu G. Approximate entropy of fetal heart rate variability as a predictor of fetal distress in women at term pregnancy. Acta Obstet Gynecol Scand 2005;84:837–43. [10] Bernardes J, Moura C, de Sá JPM, Pereira-Leite L. The porto system for automated cardiotocographic signal analysis. J Perinat Med 1991;19:61–5. [11] Acharya UR, Kannathal N, Sing OW, Ping LY, Chua T. Heart rate analysis in normal subjects of various age groups. Biomed Eng Online 2004;3:24. [12] Khandoker AH, Jelinek HF, Palaniswami M. Identifying diabetic patients with cardiac autonomic neuropathy by heart rate complexity analysis. Biomed Eng Online 2009;8:3. [13] Richman JS, Moorman JR. Physiological time series analysis using approximate entropy and sample entropy. Am J Physiol Heart Circ Physiol 2000;278:H2039–49.

Quantitative comparison of entropy analysis of fetal heart rate variability related to the different stages of labor.

The interpretation of the fetal heart rate (FHR) signal considering labor progression may improve perinatal morbidity and mortality. However, there ha...
261KB Sizes 0 Downloads 0 Views