Physiology & Behavior 149 (2015) 255–261

Contents lists available at ScienceDirect

Physiology & Behavior journal homepage: www.elsevier.com/locate/phb

A comparison of heart rate variability in women at the third trimester of pregnancy and during low-risk labour José Javier Reyes-Lagos a, Juan Carlos Echeverría-Arjonilla a, Miguel Ángel Peña-Castillo a, María Teresa García-González a, María del Rocío Ortiz-Pedroza a, Gustavo Pacheco-López b,e,⁎, Carlos Vargas-García c, Sergio Camal-Ugarte c, Ramón González-Camarena d a

Metropolitan Autonomous University (UAM), Campus Iztapalapa, Basic Sciences and Engineering Division, Mexico City 09340, Mexico UAM, Campus Lerma, Biological and Health Sciences Division, Lerma 52000, Mexico c Maternal and Childhood Research Center (CIMIGen), 09890 Mexico City, Mexico d UAM, Campus Iztapalapa, Biological and Health Sciences Division, 09340 Mexico City, Mexico e University of Leiden, Faculty of Social and Behavioural Sciences, Health, Medical and Neuropsychology Unit, 2333 AK Leiden, The Netherlands b

H I G H L I G H T S • • • •

A novel characterization of human heartbeat dynamics during labour. Vagal influence increased in labour in comparison to the third trimester of gestation. A cholinergic anti-inflammatory activity seems to be manifested in labour. Maternal heartbeat dynamics could indicate the physiological onset of labour.

a r t i c l e

i n f o

Article history: Received 6 January 2015 Received in revised form 21 April 2015 Accepted 31 May 2015 Available online 3 June 2015 Keywords: Heart rate variability Detrended fluctuation analysis Autonomic activity Anti-inflammatory cholinergic pathway Oxytocin Labour Pregnancy

a b s t r a c t Heart rate variability (HRV) has been recognised as a non-invasive method for assessing cardiac autonomic regulation. Aiming to characterize HRV changes at labour in women, we studied 10 minute ECG recordings from young mothers (n = 30) at the third trimester of pregnancy (P) or during augmentation of labour (L) (n = 30). Data of the L group were collected when no-contractions (L-NC) or the contractile activity (L-C) was manifested. Accordingly, the inter-beat interval (IBI) time series were processed to estimate relevant parameters of
 HRV such as the mean IBI (IBI), the mean heart rate HR , the root mean square of successive differences (RMSSD) in IBIs, the natural logarithm of high-frequency component (LnHF), the short-term scaling parameters from detrended fluctuation and magnitude and sign analyses such as (α1, α1(MAG), α1(SIGN)), and the sample entropy (SampEn). We found statistical differences (p b 0.05) for RMSSD among P and L-NC/L-C groups (25 ± 13 vs. 36 ± 14/34 ± 16 ms) and for LnHF between P and L-NC (5.37 ± 1.15 vs. 6.05 ± 0.86 ms2). Likewise, we identified statistical differences (p b 0.05) for α1(SIGN) among P and L-NC/L-C groups (0.19 ± 0.20 vs. 0.32 ± 0.17/0.39 ± 0.13). By contrast, L-NC and L-C groups showed statistical differences (p b 0.05) in α1(MAG) (0.67 ± 0.12 vs. 0.79 ± 0.12), and SampEn (1.62 ± 0.26 vs. 1.20 ± 0.44). These results suggest that during labour, despite preserving a concomitant non-linear influence, the maternal short-term cardiac autonomic regulation becomes weakly anticorrelated (as indicated by α1(SIGN)); furthermore, an increased vagally mediated activity is observed (as indicated by RMSSD and LnHF), which may reflect a cholinergic pathway activation owing to the use of oxytocin or the anti-inflammatory cholinergic response triggered during labour. © 2015 Published by Elsevier Inc.

1. Introduction

⁎ Corresponding author at: Universidad Autónoma Metropolitana (UAM), Unidad Lerma, División de Ciencias Biológicas y de la Salud, Departamento de Ciencias de la Salud, Av. De las Garzas 10, C.P. 52005 Lerma, Edo México, México. E-mail address: [email protected] (G. Pacheco-López).

http://dx.doi.org/10.1016/j.physbeh.2015.05.041 0031-9384/© 2015 Published by Elsevier Inc.

Contractions during labour are generally associated with an increased maternal heart rate and mean arterial pressure as well as the subsequent increase in cardiac output resulting from increments in both stroke volume and heart rate [1–5]. The details of how these haemodynamic changes are driven by autonomic adaptations and, moreover, the actual role of the autonomic nervous system (ANS) during

256

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261

pregnancy remain to be fully elucidated [6]. In this regard, several findings prompt to consider labour as an inflammatory event that is not only initiated by hormonal factors [7]. In consequence, it can be assumed that labour may also introduce an anti-inflammatory response carried out by the autonomic activity [8]. Insights about the ANS activity of pregnant women have been studied by analysing the variation between the successive cardiac inter-beat intervals (IBIs), also referred to as heart rate variability (HRV) [9–11], which is a non-invasive approach to quantify the autonomous cardiac response to adrenergic and cholinergic influences [12]. In general, timedomain and spectral analysis, but also scaling methods (i.e. detrended fluctuation analysis: DFA) are applied to obtain measures that are used to estimate the autonomic response. Although some authors have assumed that during pregnancy maternal autonomic conditions mainly reflect a sympathetic involvement [9–11] to the aortocaval compression of late gestation [13], the dynamic patterns of maternal heartbeat fluctuations have not been thoroughly explored, in particular during labour. In this regard, previous studies have mainly focused on changes of the mean maternal heart rate in relation to uterine contractions [2,4,5]. Recently, Suzuki et al. assessed maternal HRV during labour using a power spectral analysis to describe beat-to-beat changes associated with the particular presence of uterine contractions. Despite that no differences in high-frequency (HF) components between uterine contraction and non-contraction periods were found, the low-frequency (LF) and very-low-frequency (VLF) components during uterine contractions were significantly stronger during contraction periods. Authors concluded that the maternal sympathetic activity was apparently increased during uterine contraction periods [14]. In this study, we evaluate the use of linear and non-linear parameters to analyse heartbeat fluctuations during labour. In particular, we aimed to compare short-term IBI fluctuations registered from low-risk women during augmentation of labour with data collected during the third trimester of pregnancy to identify dynamic HRV changes. We hypothesised that the autonomic adaptations during labour, either to facilitate haemodynamic requirements or even to restrain or counteract inflammation, are manifested on heartbeat fluctuations.

Table 2 Mean (±SD) of labour characteristics of the study population (n = 30) during periods with contractions (L-C) and without contractions (L-NC).

Dilatation (cm) Effacement (%) Number of contractions (per 10 min)

L-NC

L-C

4.2 ± 2.9 53 ± 31 1.7 ± 1.3*

4.1 ± 2.6 54 ± 28 3.7 ± 0.3

⁎ p b 10−6 between L-C and L-NC.

women maintained a semi-Fowler's position. The general characteristics and the pregnancy outcome of this group are presented in Table 1; no major complications occurred in newborns as indicated by weights and Apgar scores. The first and second stages of labour were identified by the presence of regular uterine contractions and cervix effacement and dilatation [15] (see values at Table 2). The ECG recordings during labour were classified into two distinct classes: the labour-contraction (L-C) included 10 min segments where the uterine activity was observable (with three or more contractions), and the other class, labour-no contraction (L-NC), included 10 min segments involving fewer uterine activity or no contractions at all. All studied women received intravenous oxytocin to improve contractility according to the Mexican guidelines for augmentation of labour [16]. For comparison, we also recorded 10 min segments of data collected during the last trimester of gestation (37–39 weeks) of other 30 women at the semi-Fowler's position not showing any clinical manifestation of the initiation of labour (P) (Table 1). ECG recordings were visually analysed by using the software of the device, which displays values of maternal heart rate in conjunction with an ECG-derived uterine activity signal (Fig. 1). To ensure the proper selection of L-C segments accompanied by uterine contractions, we also identified unequivocal increments of the displayed maternal heart rate because during well-established frequent contractions the haemodynamic adjustments should generally involve a compensation of cardiac output by an increased heat rate [4,5,17]. L-C and L-NC segments were only selected if these segments involve none or few maternal gross movements. 2.2. Data analysis

2. Methods 2.1. Subjects and data collection Electrocardiogram (ECG) recordings were collected from 30 women, who developed term pregnancy (39.6 ± 1.2 weeks) and underwent lowrisk labour. All recordings and procedures took place at Maternal and Childhood Research Center (CIMIGen) having obtained consent from each patient on a voluntary basis. For data collection we used an ECG portable device (Monica AN24 monitor, Monica Healthcare, Nottingham, UK), with a sampling frequency of 900 Hz. During recordings, subjects were free to choose their preferred posture due to the portability of the system, yet ECG segments to be analysed were only selected when

Table 1 Mean (±SD) of the general characteristics of the study population during labour (L) and at the third trimester of pregnancy (P).

Age (years) Weight (kg) Height (cm) Gestational age (weeks) Newborn's birth weight (g)a Newborn's 5-minute Apgar (points)a a

n = 15 (for P group).

L n = 30

P n = 30

26 ± 5 64 ± 13 156 ± 6 39.6 ± 1.2 3175 ± 297 8.9 ± 0.6

23 ± 6 73 ± 11 156 ± 5 34.8 ± 1.6 3186 ± 251 9.0 ± 0.0

Raw maternal ECG recordings were then processed using previous validated algorithms to generate IBI series which corresponded to episodes of contraction (L-C), no contraction (L-NC) or pregnancy (P) segments [18]. All series (L-C = 30, L-NC = 30 and P = 30) consisted of 600 IBIs for each subject (7 ± 1 minute duration), were reconditioned by a filtering approach and tested in line with previous studies [19] to exclude for ectopic beats and artefacts. The mean value of IBIs and HR, IBI and HR respectively, as well as the root mean square of successive differences (RMSSD) in IBIs and the natural logarithm of high-frequency (LnHF, 0.15–0.4 Hz) parameters, which quantify vagally-mediated high frequency fluctuations [12], were calculated for all series. The scaling parameters of all series were evaluated by applying detrended fluctuation analysis (DFA) and the magnitude and sign analyses (MSA). Accordingly, the DFA provided the scaling exponent, α1, as follows [20]. The maternal IBI fluctuation series were integrated by: k   Y ðkÞ ¼ ∑ IBI ðiÞ  IBI i¼1

ð1Þ

where Y(k) represents the k-th value of the resulting integration (k = 1,2…L), IBI(i) is the i-th RR interval, and IBI is the mean RR value of the IBI series (Eq. (1)). Next, the integrated series were divided into boxes having equal numbers of n samples. The local trends Yn were obtained for all boxes

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261

257

Fig. 1. Maternal heart rate traces and uterine activity signals during labour, including episodes of no-contractions: L-NC (a) and contraction episodes: L-C (b). Vertical marks from an accelerometer are depicted to indicate the presence of maternal gross movements, which were used as references to help in identifying and selecting episodes. Arrows indicate the manifestation of well-established regular contractions.

by a least-squared line fit and subtracted from Y(k) to reduce, in principle, the non-stationary trends. The average root-mean-square fluctuation, F(n), can then be calculated as Eq. (2): sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 L 2 F ðnÞ ¼ ∑ ½Y ðkÞ  Y n ðkÞ L k¼1

ð2Þ

The relationship on a double-log graph between F(n) and time scales n was approximated by a linear model F(n) ~ nα, so providing the scaling exponent α1 as the slope of the plot covering the short-term range of n from 4 to 11 IBIs (in accordance with Peña et al. [21] and Yeh et al. [22]). The scaling exponent may vary from 0.5 (uncorrelated fluctuations, white noise) to 1.5 (smoother fluctuations) where a value near 1 indicates the existence of fractal-like correlations. The long-term scaling exponent (α2) was not calculated in this study owing to well-known potential modifications of the scaling properties by periodic trends as reported by Hu et al. [23] (in our case uterine contractions during labour as well as the Braxton Hicks contractions presented at the third trimester of pregnancy). Nonlinearity and linear time ordering scaling were assessed by MSA following earlier studies [24,25]. The original IBI series were processed to obtain new sequences of increments by taking the differences between adjacent intervals (IBIi + 1–IBIi). These sequences (ΔIBI) were decomposed into magnitude, |ΔIBI|, and sign series, sign (ΔIBI). After subtracting their respective means, magnitude and sign series were integrated and DFA was again applied as described above. The slope of F(n)/n covering the range from 4 to 11 IBIs then provided magnitude and sign scaling exponents (α1(MAG) and α1(SIGN), respectively). Positive correlations in the magnitude series (i.e. finding α1(MAG) N 0.5) were identified as reliable markers of nonlinear properties [25]. The α1(SIGN) exponent provides information about the temporal directionality of the original series in relation to how series increments alternate, indicating if a positive or negative subsequent increment (decrement) is more likely to occur given a current increment (decrement) [24,25]. To assess the regularity/irregularity in relation to uterine contractions we also estimated the sample entropy (SampEn) calculated with m = 2 and r = 0.2, as described by Richman & Moorman [26]. 2.3. Statistics A discordant test was applied to the resulting values of all linear, scaling and nonlinear parameters for identifying potential outliers [27]. Having verified normality (by skewness, kurtosis and omnibus tests) and equal variance (using a modified-Levene equal-variance test), statistical differences of IBI, HR, RMSSD, LnHF, α1, α1(MAG) and α1(SIGN) between L-NC and L-C segments were evaluated by a paired t-test with the NCSS software. Differences of same parameters between L-C and P segments (or L-NC and P) were evaluated by a ttest for independent measures. Because the test for SampEn reported

no normality and rejected equal variance, a Kolmogórov–Smirnov test was applied to this parameter. Significance was considered by p b 0.05 for the paired t-test, and by p b 0.025 for the multiple independent comparisons. 3. Results We found ANS-related differences in cardiac activity during labour vs. last trimester of pregnancy (L-C/L-NC vs. P), as well as during labour on the presence of contractions vs. no-contractions (L-C vs. L-NC). Fig. 2 shows representative examples of the analysis of maternal heartbeat fluctuations during labour including segments of L-NC (Fig. 2a & c) and L-C (Fig. 2b & d). Whereas the raw IBI series are presented at the top of the figure, the log–log F(n) vs. n and log–log F(n)/ n vs. n relationships, providing the α1, α1(MAG) and α1(SIGN) exponents from a linear best-fit within the n range 4 to 11, are depicted below. This figure also includes results of the HR and RMSSD parameters (Fig. 2a & b). An increment of the HR owing to the presence of regular uterine contractions during the L-C segment is noticed. Yet the scaling parameters α1, and α1(SIGN) maintain similar values for both segments regardless of such contractions, with exception of α1(MAG). Fig. 3 shows a sign decomposition of typical data for P and L-NC segments. This figure illustrates a less anticorrelated pattern in heartbeat fluctuations during labour (L-NC) in comparison to data from the third trimester of pregnancy (P). After applying a discordant test to all parameters, we found four RMSSD and LnHF discordant values from the P group and these values were discarded. Table 3 summarises principal findings for all studied parameters in relation with L-NC, L-C and P segments. Whereas the IBI and HR parameters show differences between L-C and L-NC segments, in accordance with our proposed experimental design for the selection of segments during labour, α1(SIGN) was the only scaling parameter that showed a dynamic effect of labour on heartbeat fluctuations as shown by differences between P and L-C/L-NC segments. In accordance with earlier findings, we found statistical differences in RMSSD among LC/L-NC and P groups [28] as well as in LnHF among L-NC and P groups. By contrast, the parameters α1(MAG) and SampEn were influenced by the presence or absence of uterine contractions (Table 3). 4. Discussion Our main finding is the identification of subtle but significant dynamic changes in the directionality of heartbeat fluctuations during childbirth at term, as indicated by the parameter α1(SIGN), which were not directly associated with the presence of uterine contractions. Thus, regardless of the uterine activity, maternal HRV at labour shows weaker anticorrelations in comparison with the stronger anticorrelated autonomic activity manifested by pregnant women at the third trimester.

258

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261

Fig. 2. Representative data of woman heartbeat fluctuations on labour during episodes of no-contraction: L-NC (a), and during contraction episodes: L-C (b). Corresponding log-log F(n) (average root-mean-square fluctuation) vs. n and log-log F(n)/n vs. n, relationships are shown below (c and d). The values of the parameters HR, RMSSD, α1, α1(MAG) and α1(SIGN) calculated for those series are depicted as well.

Following our previous results [21], this stronger anticorrelated condition of maternal heartbeat fluctuations seems in fact to be maintained prenatally during low-risk pregnancies. Our findings for α1(SIGN) may indicate that during labour the maternal cardiac regulation becomes less antagonistic in relation with short-term directionality (see Fig. 3b). Implying that, in terms of the mathematical framework of “random walks” [29], the autonomic interplay is manifested with different attracting levels for the heart frequency as opposed to the resulting strongly anticorrelated dynamics introduced, for instance, by a sympathetic predominance during the third trimester [30]. The autonomic activity has been considered to offer an antiinflammatory response through the direct action of acetylcholine inhibiting synthesis of pro-inflammatory cytokines, or indirectly via catecholamines stimulating the release of the anti-inflammatory cytokines, such as IL-10 [8]. Recent findings have provided insights into the correlation between heartbeat fluctuation parameters and inflammatory markers in accordance with the consideration of the so called anti-inflammatory reflex, mainly mediated by the vagus nerve [8,31, 32]. In this regard, the weaker anticorrelations of the maternal IBI fluctuations at labour (L-NC vs. P) could then reflect the participation of such anti-inflammatory autonomic pathway. This consideration is in accordance with the particular conceptualization of labour as an inflammatory event, either driven by exogenous or endogenous stimuli [7,28, 33,34]. This position is supported by evidence identifying that some important labour processes, such as cervix remodelling, rupture of membranes, and rhythmic contractions, appear to be mediated by

inflammatory processes [34]. Besides, the release of inflammatory cytokines during labour probably contributes to the expelling of placenta's fragments and, eventually, to contend with pathogens in the postpartum period as well [35]. Furthermore, the activation of the antiinflammatory autonomic pathway may also be attributed to the role of exogenous oxytocin that embraces functional effects as cardiovascular and the autonomic modulating neuroimmunoendocrine peptide [36–39]. Clodi et al. have found that oxytocin decreases the neuroendocrine and cytokine activation caused by bacterial endotoxin in men. Given that the effect of oxytocin was not identified for monocytes and peripheral blood mononuclear cells in vitro, they concluded that such decrease is possible due to the modulation of the cholinergic antiinflammatory pathway [40]. Notwithstanding that the administration of intranasal oxytocin did not alter the circulating levels of proinflammatory cytokines or stress hormones, Norman et al. found that oxytocin also influences the autonomic cardiac control as indicated by an increased HF parameter and decreased pre-ejection period [37]. Similarly, Kemp et al. have reported that the acute administration of intranasal oxytocin increases the HRV in male humans during rest as indicated by changes in the power of the HF component at trend levels [41]. Oxytocin for the purposes of augmentation of labour is one of the most frequently used medications in obstetrics.1 Recent studies have

1 Augmentation of labour with oxytocin is part of the concept of “active management of labour” to reduce its duration and the rate of surgical delivery [42].

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261

259

Fig. 3. Inter-beat interval (IBI) fluctuations series from the third trimester group (P): α1(SIGN) = 0.15 (a) and from the labour no contraction group (L-NC): α1(SIGN) = 0.37 (b). An excerpt of their corresponding sign decomposition is shown below.

showed that oxytocin is used in 40% of labouring women in Latin American countries [42]. Our results for the RMSSD and LnHF parameters during labour seem also consistent with the possible manifestation of a cholinergic mechanism to attenuate inflammation as we found increased values of both parameters at L-NC episodes as compared with the P group, suggesting the Table 3 Means (±SD) of HRV parameters of women (n = 30) during labour with (L-C) or without uterine contractions (L-NC) and also from pregnant women (n = 30) at the third trimester (P). Parameter

L-NC

L-C

P

IBI (ms)

705 ± 100

705 ± 90

87 ± 14

656 ± 107⁎⁎ 94 ± 17⁎⁎

36 ± 14 6.05 ± 0.86 1.12 ± 0.23 0.67 ± .12 0.32 ± 0.17 1.62 ± 0.26

34 ± 16 5.85 ± 0.82 1.14 ± 0.20 0.79 ± 0.12⁎⁎ 0.39 ± 0.13 1.20 ± 0.44⁎⁎

25 ± 13⁎ 5.37 ± 1.15† 1.03 ± 0.24 0.72 ± 0.15 0.19 ± 0.20⁎ 1.48 ± 0.47⁎

HR (bpm) RMSSD (ms) LnHF (ms2) α1 α1(MAG) α1(SIGN) SampEn

87 ± 11

⁎ p b 0.025 between L-NC and P and also between L-C and P segments (n = 26 for RMSSD). ⁎⁎ p b 0.05 between L-NC and L-C segments. † p b 0.025 between L-NC and P (n = 26).

participation of the vagal cholinergic pathway [43], and the subsequent dynamic manifestation of diverse heart frequency attracting levels from the mathematical standpoint of “random walks” [29]. The autonomic condition of pregnant women has been mainly evaluated by quantifying the spectral power, or by assessing the structure of irregularity of the heartbeat fluctuations through the use of scaling exponents [21,22,44]. In this regard, some authors have identified autonomic changes in the heartbeat fluctuations during pregnancy; these have been generally associated with the particular response to undertake the aortocaval compression of late gestation [13]. Yeh et al. reported that the heartbeat fluctuations of late pregnant women involve lower magnitude and an increased short-term fractal scaling exponent α1 as compared with data collected 3 months after delivery, which in return show similar values of non-pregnant controls [22]. Similarly, Baumert et al. [45] have found increased α1 values in comparison with the first half of pregnancy at the end of gestation, perhaps coinciding with the reduced vagal outflow reported during pregnancy [46].Yet, no changes in low-risk pregnant women during gestation for the scaling exponents α1 and α1(SIGN) have also been reported [21]. Corresponding to these studies, we observed that no changes in α1, RMSSD and LnHF were in fact introduced by the uterine activity, despite provoking a significantly different mean heart rate. This finding is in accordance with a recent study reporting no differences in high-frequency HRV

260

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261

components during periods with uterine contractions [14]. Actually, α1 at labour showed here similar values with the abovementioned prenatal studies [21]. In the case of α1(MAG), we also failed to find differences when comparing L-C and P groups. Thus, according to the dynamic meaning that can be considered for these parameters, the short-term heartbeat fluctuations maintain irregular fractal-like correlations (α1 ≈ 1) and non-linearity (α1(MAG) N 0.5) regardless of the increased hemodynamic demands of labour. Using the interpretation adopted from our previous study [21], both dynamical conditions can thus be considered as evidence that the ANS regulation of the cardiovascular system does not become hampered at low-risk labour. Similar dissociation between changes in the magnitude of HRV parameters and their α1 scaling invariance has been described in other physiological scenarios. For example, the well-known reduction of the IBI fluctuations that occurs during ageing was not found to be accompanied by changes of the fractal (scale-invariant) behaviour [47]. This finding suggests that the alterations in the cardiac control mechanisms with advance age differ from the mechanistic changes in the autonomic regulation associated with pathological conditions. Such results appeared in accordance with Pikkujämsä et al. who reported that children show lower overall HRV, despite presenting similar IBI short-range correlation properties as healthy adults [48]. Likewise, despite that the heart rate and other short-term indexes of HRV are affected by physical activity during free-running ambulatory conditions [49], this influence, by contrast, did not appear significant for α1. In the same way, Tulppo et al. have reported that an enhanced vagal outflow did not change α1 values [50]. Regarding clear heartbeat dynamic effects provoked by the manifestation of contractions during labour, it is important to mention that, in addition to the differences in the mean heart rate and SampEn2 (see below), we did find differences in the α1(MAG) parameter, which probably indicate a contraction-driven increment of the nonlinear properties involved in the cardiovascular regulation. These properties have been linked to the complexity of the feedback mechanism of neurohormonal cardiac regulation [24,47], which increases for pregnant women after mid pregnancy possibly owing to new control influences or to modifications of feedback interactions [21]. Given the pain perception dynamics associated with uterine activity [52,53], the contraction-driven increment of complexity (indicated by α1(MAG)) that we report here appears relevant to be explored in more detail as we did not find any study yet reporting an association of this parameter with pain.

5. Limitations Labour introduces changes in ventilation that may also affect the respiratory-driven IBI fluctuations. All studied women here were under spontaneous respiration, and we calculated the central frequency of the HF band to estimate the mean respiration frequency between labour and third trimester groups. Yet, as no statistical differences were found [28], an important influence of ventilation in the results found for labour in this study is not likely. The mean time elapsed between the onset of labour and the beginning of the collection of ECG data was 5.8 ± 7.1 h. The total time of labour was 13 ± 8 h. As a high variation in such beginning and duration of labour was observed, it would be appropriate as well to consider their potential effects for our analysis. However, we found no significant time-dependence on the estimation of HRV parameters, which would be strongly associated with that variation.

2 Regarding the changes of SampEn introduced by contractions, it is convenient to take into consideration that owing to the fact that this parameter measures the time-series regularity, a decrease in its value can be related to very-low-frequency changes in the IBI series, see Fig. 2b, which may be influenced by the periodicity of electrohysterographic activity during labour [51].

Despite that it was not possible for us to make a longitudinal study to avoid potential individual differences between subjects in the HRV values, the general characteristics of the study population (5-minute Apgar score, birth weight, gestational age and others, reported at Table 1) support the manifestation of quite similar conditions between the L and P groups, both reflecting a low-risk pregnancy status. In future work, we will consider investigating HRV data obtained during pre- and post-partum stages at immediate periods before and after labour. 6. Conclusion In summary during normal childbirth the maternal short-term cardiac regulation shows a concomitant nonlinear dynamics that should provide stable and adaptive capabilities [54]. Yet such regulation becomes weakly anticorrelated (as indicated by α1(SIGN)) and involves an increased vagally-mediated fluctuations (as indicated by RMSSD and LnHF), which may reflect an activation of the cholinergic pathway owing to the use of oxytocin or the anti-inflammatory response triggered during labour. We consider that dynamic changes in maternal heartbeat fluctuations could then become potential features to identify the physiological onset of labour. Conflict of interest The authors declare no conflict of interests. Acknowledgements The support of volunteer mothers and staff of CIMIGen is gratefully acknowledged. Financial aid for this research was provided by the Mexican Council for Science and Technology (CONACyT, project I000383999). J. Javier Reyes-Lagos thanks the CONACyT for providing a scholarship (CVU/scholarship number: 381983/253449). References [1] S.C. Robson, W. Dunlop, R.J. Boys, S. Hunter, Cardiac output during labour, Br. Med. J. (Clin. Res. Ed.) 295 (6607) (1987 Nov 7) 1169–1172. [2] D.J. Sherman, E. Frenkel, Y. Kurzweil, A. Padua, S. Arieli, M. Bahar, Characteristics of maternal heart rate patterns during labor and delivery, Obstet. Gynecol. 99 (4) (2002 Apr) 542–547. [3] E.J. Bridges, S. Womble, M. Wallace, J. McCartney, Hemodynamic monitoring in high-risk obstetrics patients, I. Expected hemodynamic changes in pregnancy, Crit. Care Nurse 23 (4) (2003 Aug) 53–62. [4] N. Söhnchen, K. Melzer, B.M. Tejada, N. Jastrow-Meyer, V. Othenin-Girard, O. Irion, M. Boulvain, B. Kayser, Maternal heart rate changes during labour, Eur. J. Obstet. Gynecol. Reprod. Biol. 158 (2) (2011 Oct) 173–178. [5] T.R. VAN Veen, M.A. Belfort, S. Kofford, Maternal heart rate patterns in the first and second stages of labor, Acta Obstet. Gynecol. Scand. 91 (5) (2012 May) 598–604. [6] Q. Fu, B.D. Levine, Autonomic circulatory control during pregnancy in humans, Semin. Reprod. Med. 27 (4) (2009 Jul) 330–337. [7] E. Golightly, H.N. Jabbour, J.E. Norman, Endocrine immune interactions in human parturition, Mol. Cell. Endocrinol. 335 (1) (2011 Mar 15) 52–59. [8] K.J. Tracey, The inflammatory reflex, Nature 420 (6917) (2002 Dec 19–26) 853–859. [9] C.D. Kuo, G.Y. Chen, M.J. Yang, H.M. Lo, Y.S. Tsai, Biphasic changes in autonomic nervous activity during pregnancy, Br. J. Anaesth. 84 (3) (2000 Mar) 323–329. [10] E.M. Ekholm, R.U. Erkkola, Autonomic cardiovascular control in pregnancy, Eur. J. Obstet. Gynecol. Reprod. Biol. 64 (1) (1996 Jan) 29–36. [11] G. Speranza, G. Verlato, A. Albiero, Autonomic changes during pregnancy: assessment by spectral heart rate variability analysis, J. Electrocardiol. 31 (2) (1998 Apr) 101–109. [12] Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, Heart rate variability. Standards of measurement, physiological interpretation, and clinical use, Eur. Heart J. 17 (3) (1996 Mar) 354–381. [13] G.Y. Chen, C.D. Kuo, M.J. Yang, H.M. Lo, Y.S. Tsai, Comparison of supine and upright positions on autonomic nervous activity in late pregnancy: the role of aortocaval compression, Anaesthesia 54 (3) (1999 Mar) 215–219. [14] N. Suzuki, J. Sugawara, Y. Kimura, S. Nagase, K. Okamura, N. Yaegashi, Assessment of maternal heart-rate variability during labor using wavelet-based power spectral analysis, Gynecol. Obstet. Investig. 74 (1) (2012) 35–40.

J.J. Reyes-Lagos et al. / Physiology & Behavior 149 (2015) 255–261 [15] A. DeCherney, L. Nathan, T.M. Goodwin, T.M. Laufer, Current Diagnosis & Treatment Obstetrics & Gynecology, Tenth edition McGraw-Hill Companies, Incorporated, 2007. [16] The Mexican Ministry of Health. Monitoring and management of low-risk labour (in Spanish: Secretaría de Salud: Vigilancia y manejo del trabajo de parto en embarazos de bajo riesgo). Mexico, Mexico City, 2014. At: http://www.cenetec.salud.gob.mx/ descargas/gpc/CatalogoMaestro/052_GPC_VigilanciaManejodelParto/IMSS_052_ 08_EyR.pdf. Accessed March 3, 2015. [17] S. Hunter, S.C. Robson, Adaptation of the maternal heart in pregnancy, Br. Heart J. 68 (6) (1992 Dec) 540–543. [18] J.C. Echeverría, R. Ortiz, N. Ramirez, V. Medina, R. Gonzalez, A reliable method for abdominal ECG signal processing, Comput. Cardiol. (1998) 529–532. [19] N. Wessel, A. Voss, H. Malberg, C. Ziehmann, H.U. Schirdewan, U. Meyerfeldt, J. Kurths, Nonlinear analysis of complex phenomena in cardiological data, Herzschr. Elektrophys. 11 (2000) 159–173. [20] C.K. Peng, S. Havlin, H.E. Stanley, A.L. Goldberger, Quantification of scaling exponents and crossover phenomena in nonstationary heartbeat time series, Chaos 5 (1) (1995) 82–87. [21] M.A. Peña, J.C. Echeverría, M.T. García, M.R. Ortiz, A. Martínez, C. Vargas-Garcia, R. González-Camarena, Short-term heart rate dynamics of pregnant women, Auton. Neurosci. 159 (1–2) (2011 Jan 20) 117–122. [22] R.G. Yeh, J.S. Shieh, G.Y. Chen, C.D. Kuo, Detrended fluctuation analysis of short-term heart rate variability in late pregnant women, Auton. Neurosci. 150 (1–2) (2009 Oct 5) 122–126. [23] Hu K, Ivanov PC, Chen Z, Carpena P, Stanley HE. Effect of trends on detrended fluctuation analysis. Phys. Rev. E Stat. Nonlinear Soft Matter Phys.. (2001) Jul;64(1 Pt 1): 011114. [24] Y. Ashkenazy, P.C. Ivanov, S. Havlin, C.K. Peng, A.L. Goldberger, H.E. Stanley, Magnitude and sign correlations in heartbeat fluctuations, Phys. Rev. Lett. 86 (9) (2001 Feb 26) 1900–1903. [25] Y. Ashkenazy, S. Havlin, P.C. Ivanov, C.K. Peng, V. Schulte-Frohlinde, H.E. Stanley, Magnitude and sign scaling in power-law correlated time series, Physica A. 323 (2003) 19–41. [26] J.S. Richman, J.R. Moorman, Physiological time-series analysis using approximate entropy and sample entropy, Am. J. Physiol. Heart. Circ. Physiol. 278 (6) (2000 Jun) H2039–H2049. [27] S.P. Verma, L. Díaz-González, Application of the discordant outlier detection and separation system in the geosciences, Int. Geol. Rev. 54 (5) (2012) 1–22. [28] J.J. Reyes-Lagos, J.C. Echeverría-Arjonilla, M.A. Peña-Castillo, A.J. Montiel-Castro, G. Pacheco-López, Physiological, immunological and evolutionary perspectives of labor as an inflammatory process, Advances in Neuroimmune Biology 5 (2) (2014) 75–89. [29] P.C. Ivanov, L.A. Nunes Amaral, A.L. Goldberger, H.E. Stanley, Stochastic feedback and the regulation of biological rhythms, Europhys. Lett. 43 (4) (1998 Aug 15) 363–368. [30] R. Karasik, N. Sapir, Y. Ashkenazy, P.C. Ivanov, I. Dvir, P. Lavie, S. Havlin, Correlation differences in heartbeat fluctuations during rest and exercise, Phys. Rev. E Stat. Nonlinear Soft Matter Phys. 66 (6 Pt 1) (2002 Dec) 062902. [31] J.M. Huston, K.J. Tracey, The pulse of inflammation: heart rate variability, the cholinergic anti-inflammatory pathway and implications for therapy, J. Intern. Med. 269 (1) (2011 Jan) 45–53. [32] U. Andersson, K.J. Tracey, Reflex principles of immunological homeostasis, Annu. Rev. Immunol. 30 (2012) 313–335. [33] J.R. Challis, C.J. Lockwood, L. Myatt, J.E. Norman, J.F. Strauss 3rd, F. Petraglia, Inflammation and pregnancy, Reprod. Sci. 16 (2) (2009 Feb) 206–215. [34] M.L. Houben, P.G. Nikkels, G.M. van Bleek, G.H. Visser, M.M. Rovers, H. Kessel, W.J. de Waal, L. Schuijff, A. Evers, J.L. Kimpen, L. Bont, The association between intrauterine inflammation and spontaneous vaginal delivery at term: a cross-sectional study, PLoS One 4 (8) (2009 Aug 10) e6572. [35] M.R. Peltier, Immunology of term and preterm labor, Reprod. Biol. Endocrinol. 1 (2003 Dec 2) 122.

261

[36] K.T. Higa, E. Mori, F.F. Viana, M. Morris, L.C. Michelini, Baroreflex control of heart rate by oxytocin in the solitary–vagal complex, Am. J. Physiol. Regul. Integr. Comp. Physiol. 282 (2) (2002 Feb) R537–R545. [37] G.J. Norman, J.T. Cacioppo, J.S. Morris, W.B. Malarkey, G.G. Berntson, A.C. Devries, Oxytocin increases autonomic cardiac control: moderation by loneliness, Biol. Psychol. 86 (3) (2011 Mar) 174–180. [38] M. Gamer, C. Büchel, Oxytocin specifically enhances valence-dependent parasympathetic responses, Psychoneuroendocrinology 37 (1) (2012 Jan) 87–93. [39] J. Gutkowska, M. Jankowski, Oxytocin revisited: its role in cardiovascular regulation, J. Neuroendocrinol. 24 (4) (2012 Apr) 599–608. [40] M. Clodi, G. Vila, R. Geyeregger, M. Riedl, T.M. Stulnig, J. Struck, T.A. Luger, A. Luger, Oxytocin alleviates the neuroendocrine and cytokine response to bacterial endotoxin in healthy men, Am. J. Physiol. Endocrinol. Metab. 295 (3) (2008 Sep) E686–E691. [41] A.H. Kemp, D.S. Quintana, R.L. Kuhnert, K. Griffiths, I.B. Hickie, A.J. Guastella, Oxytocin increases heart rate variability in humans at rest: implications for social approach-related motivation and capacity for social engagement, PLoS One 7 (8) (2012) e44014. [42] M. do Carmo Leal, A.P. Pereira, R.M. Domingues, M.M. Theme Filha, M.A. Dias, M. Nakamura-Pereira, M.H. Bastos, S.G. da Gama, Obstetric interventions during labor and childbirth in Brazilian low-risk women, Cad Saude Publica 30 (Suppl 1) (2014 Aug) S1–16. [43] J.F. Thayer, Vagal tone and the inflammatory reflex, Cleve. Clin. J. Med. 76 (2009 Apr) S23–S26. [44] H. Malberg, R. Bauernschmitt, A. Voss, T. Walther, R. Faber, H. Stepan, N. Wessel, Analysis of cardiovascular oscillations: a new approach to the early prediction of pre-eclampsia, Chaos 17 (1) (2007 Mar) 015113. [45] M. Baumert, M. Javorka, A. Seeck, R. Faber, P. Sanders, A. Voss, Multiscale entropy and detrended fluctuation analysis of QT interval and heart rate variability during normal pregnancy, Comput. Biol. Med. 42 (3) (2012 Mar) 347–352. [46] A.V. Klinkenberg, U.M. Nater, A. Nierop, A. Bratsikas, R. Zimmermann, U. Ehlert, Heart rate variability changes in pregnant and non-pregnant women during standardized psychosocial stress, Acta Obstet. Gynecol. Scand. 88 (1) (2009) 77–82. [47] D.T. Schmitt, P.C. Ivanov, Fractal scale-invariant and nonlinear properties of cardiac dynamics remain stable with advanced age: a new mechanistic picture of cardiac control in healthy elderly, Am. J. Physiol. Regul. Integr. Comp. Physiol. 293 (5) (2007 Nov) R1923–R1937. [48] S.M. Pikkujämsä, T.H. Mäkikallio, L.B. Sourander, I.J. Räihä, P. Puukka, J. Skyttä, C.K. Peng, A.L. Goldberger, H.V. Huikuri, Cardiac interbeat interval dynamics from childhood to senescence: comparison of conventional and new measures based on fractals and chaos theory, Circulation 100 (4) (1999 Jul 27) 393–399. [49] A.J. Hautala, J. Karjalainen, A.M. Kiviniemi, H. Kinnunen, T.H. Mäkikallio, H.V. Huikuri, M.P. Tulppo, Physical activity and heart rate variability measured simultaneously during waking hours, Am. J. Physiol. Heart. Circ. Physiol. 298 (3) (2010 Mar) H874–H880. [50] M.P. Tulppo, A.M. Kiviniemi, A.J. Hautala, M. Kallio, T. Seppänen, T.H. Mäkikallio, H.V. Huikuri, Physiological background of the loss of fractal heart rate dynamics, Circulation 112 (3) (2005 Jul 19) 314–319. [51] J. Vrhovec, Evaluating the progress of the labour with sample entropy calculated from the uterine EMG activity, Elektrotehniški vestnik 76 (4) (2009 Jan) 165–170. [52] T. Saisto, R. Kaaja, O. Ylikorkala, E. Halmesmäki, Reduced pain tolerance during and after pregnancy in women suffering from fear of labor, Pain 93 (2) (2001 Aug) 123–127. [53] S. Alehagen, B. Wijma, U. Lundberg, K. Wijma, Fear, pain and stress hormones during childbirth, J. Psychosom. Obstet. Gynaecol. 26 (3) (2005 Sep) 153–165. [54] T.G. Buchman, The community of the self, Nature 420 (6912) (2002 Nov 14) 246–251.