Use of radiotelemetry to assess perinatal cardiac function in the ovine fetus and newborn

Abstract

The late gestation fetal ECG (fECG) has traditionally been difficult to characterize due to the low fECG signal relative to high maternal noise. Although new technologies have improved the feasibility of its acquisition and separation, little is known about its development in late gestation, a period in which the fetal heart undergoes extensive maturational changes. Here, we describe a method for the chronic implantation of radiotelemetry devices into late gestation ovine fetuses to characterize parameters of the fECG following surgery, throughout late gestation, and in the perinatal period. We found no significant changes in mean aortic pressure (MAP), heart rate (HR), or ECG in the 5 days following implantation; however, HR decreased in the first 24 h following the end of surgery, with associated increases in RR, PR, and QRS intervals. Over the last 14 days of fetal life, fetal MAP significantly increased, and HR significantly decreased, as expected. MAP and HR increased as labor progressed. Although there were no significant changes over time in the ECG during late gestation, the duration of the PR interval initially decreased and then increased as birth approached. These results indicate that although critical maturational changes occur in the late gestation fetal myocardium, the mechanisms that control the cardiac conduction are relatively mature in late gestation. The study demonstrates that radiotelemetry can be successfully used to assess fetal cardiac function, in particular conduction, through the process of labor and delivery, and may therefore be a useful tool for study of peripartum cardiac events.

the ovine fetus has long been used as a model of fetal development, particularly because of the similar developmental trajectory to the human fetus and the relative quiescence of the ovine uterus, reducing the risks inherent with open fetal surgery. These properties have made the chronically instrumented ovine fetus invaluable in the characterization and manipulation of physiological processes. The ovine fetus has been particularly useful as a model for study of fetal organ maturation, as the ovine fetus has similarities in maturational trajectory to the human, owing in part to the similar surge in fetal cortisol production in the days preceding birth (8, 19). Cortisol contributes to maturation of many fetal tissues including the heart, lung, gastrointestinal tract, liver, and brain (4, 29, 42, 48, 52).

Our laboratory has been interested in the study fetal cardiac maturation and in particular the adverse effects of elevated maternal cortisol on the fetal heart (13, 23, 44, 47) As we have found an increase in peripartum stillbirth in our model of late gestation maternal stress, with transcriptomic effects indicative of altered cardiac metabolism, we wanted to develop a refinement in the methods for assessing fetal cardiac function, and in particular ECG, during labor, delivery, and the immediate perinatal period. Many previous studies have documented the late gestation changes in fetal blood pressure and heart rate (HR), as well as responses of fetal HR and ECG to manipulations such as hypoxia, cord occlusion, or chronic growth restriction (14, 27, 30, 36, 62, 64). Fetal ECG (fECG) is suggested as a good measure for assessing fetal brain ischemia, and this has been widely studied in that context as an index of acute fetal stress (38, 51). However, very chronic study of fECG and study during labor and delivery are complicated by the need to limit maternal movements during signal acquisition. Telemetric devices have been successfully used in the ovine fetus to measure renal sympathetic nerve activity (5). The feasibility of chronic implantation of radiotelemetry devices for measurement of fetal blood pressure or ECG in the late gestation ovine fetus is supported by short-term limited studies by others (1, 17) but has not been rigorously tested. In the current study, we use surgically implanted radiotelemetry devices in late gestation ovine fetuses and assess the use of this method to characterize changes in the fECG occurring after surgery, throughout late gestation, and throughout labor and delivery.

METHODS

Ewes of known gestational age and their lambs were studied. Ewes were of either Rambouillet or Dorset breed, averaging 71.0 ± 7.1 kg at the time of surgery; lambs were 4.18 ± 0.33 kg at birth (143 ± 1; days 139–148 gestation). All animals were housed in a facility with temperature- and light-controlled (lights on 0700 to 1900) rooms throughout the study period; the University of Florida Institutional Animal Care and Use Committee approved all animal use for this study. Ewes were housed in individual pens of ~2.25 m² each throughout the study. Husbandry, including cleaning of the pens and feeding, typically lasted ~30 min and was completed between 0700 and 0800. Ewes were fed a diet of pelleted feed (Nutrena Goat and Sheep Feed: Cargill) according to National Research Council guidelines based on weight of the ewe and the gestational age. This diet was fed each morning and was supplemented with loose alfalfa hay and/or alfalfa cubes in morning and evening. An entry sheet was used to record times for husbandry and feeding as well as postoperative care; these times were excluded from the analysis.

At approximately day 118 (±1) of gestation, survival surgery was performed. Ewes were induced with isoflurane by mask, intubated, and maintained on isoflurane anesthesia with controlled ventilation at 8–12 breaths/min with 2–3 l/min of oxygen. The total period of anesthesia, including preperation and surgery, was 4.5 ± 0.3 h. Maternal end-expiratory Pco₂, O₂ saturation, blood pressure, and HR were monitored throughout the procedure. A flow probe (6-mm 6PSS; Transonics, Ithaca, NY) was placed on the main uterine artery for assessment of labor. To study changes in fetal aortic pressure and ECG, a radiotelemetry device (DSI PA- D70 PCTP; Data Sciences International, Minneapolis, MN) was placed within the fetus to allow continuous measurement of fetal aortic pressure, amniotic pressure, fECG, and temperature. Briefly, the head and neck of the fetus were located in the uterus and exposed. A midline incision was made in the fetal neck and the telemetry device was placed subcutaneously at the level of the clavicle and sutured in place. The grounding lead for the ECG was tunneled subcutaneously using a forcep and attached to the inner surface of the skin of the thorax using a sterilized fishing hook (size 6; Baitholder; Eagle Claw, Denver, CO). The solid tip ECG probe was placed into the right jugular vein of the fetus, advanced into the superior vena cava until a P wave was visualized and optimized in the acquisition software (Ponemah 5.00), and secured at the entry point in the jugular with sterile tissue adhesive (Vetbond; 3M, St. Paul, MN). The two telemetry device catheters were then placed for measurement of aortic pressure and amniotic pressure. For placement of the aortic pressure catheter, purse-string sutures (5–0 Prolene Suture, Ethicon; Somerville, NJ) were placed in the left carotid artery of the fetus, and the catheter was inserted in the carotid and advanced into the aorta, using the acquisition software to assure placement in the aorta outside of the left ventricle. The catheter was secured in place in the carotid using the purse string sutures and a drop of tissue adhesive. The other pressure catheter was tunneled underneath the fetal skin and sutured to the skin of the fetal neck with the tip exposed. Therefore, all parts of the telemetry device were implanted within the fetus. All aortic pressures were corrected by subtraction of amniotic fluid pressure. Catheters for blood sample collection were also placed in the fetal saphenous arteries and veins and advanced to the fetal femoral arteries and veins (24) and in the maternal femoral arteries and veins. At the end of surgery, elastic surgical dressing was placed around the abdomen of the ewe (size 9 and 10 Surgilast; Derma Sciences, Princeton, NJ) and these fetal and maternal catheters were placed in a pocket made from sterilization wrap (Kimguard; Kimberly Clark, Roswell, GA) which was placed under the Surgilast on the flank of the ewes. A repeater device for the telemetry device was secured in place on the flank of the ewe after wrapping the unit in bandage tape (Vetwrap; 3M) and fastened with a cable tie to the Surgilast over the maternal abdomen.

Six fetuses died because of surgical complications; all of these deaths appeared to be attributable to cord occlusion. Subsequently, we altered the surgery to minimize the extent to which the fetus was displaced during surgery; only the fetal head and neck was exposed and ~500 ml of sterile saline were infused back into the amniotic cavity.

After surgery, ewes were returned to their housing pens (24 ft² each) in which they were allowed to move freely. The ECG and aortic and amniotic pressure signals were continuously acquired from the repeater device by a radioreciever fastened to the front of the pen, which was connected to the DSI exchange matrix and computer. The system used in these studies used a repeater device to assure that animals in adjacent pens could have data collected at unique frequencies; the current systems available from DSI do not require use of the repeater and instead use multiple receivers placed around the pen. Although the signal strength of the telemetry device is sufficient to allow complete implantation into the fetus, the repeater device assured that the rebroadcast signal allowed for uninterrupted communication with the receiver without interference with the signal from adjacent pens. This arrangement allows for the acquisition of data during labor, delivery, and thereafter from multiple animals simultaneously. Each signal (aortic pressure, amniotic pressure, and ECG) was sampled at 500 Hz by the Dataquest ART software. The catheters were used to collect samples for maternal and fetal cortisol (EA65; Oxford Biomedical, Rochester Hills, MI), glucose and lactate (YSI Model 2700 glucose/lactate analyzer, Yellow Springs, OH), measurements, and for fetal blood gases (iSTAT Handheld; Abbott Point of Care, Princeton, NJ). The cortisol ELISA is run after extraction of plasma with 100% ethanol; the assay has a minimal detectable concentration of 0.5 ng/ml and the coefficient of variation for the pools of nonpregnant, pregnant, and fetal plasma were 12.2, 8.7, and 5.2%, respectively. These samples were collected without restraint of the ewe and were collected in the morning at least 1 h after completion of the daily husbandry activities. Measurements for maternal and fetal cortisol concentrations as well as fetal blood gases are presented in Table 1.

Table 1.

Maternal and fetal plasma cortisol and fetal blood gases

	Plasma Cortisol, ng/ml		Fetal
	Maternal	Fetal*	Po2	Pco2	pH*
125 days	10.4 ± 3.3 (8)	4.5 ± 1.5 (8)	18 ± 1	49.5 ± 3.0	7.36 ± 0.02
130 days	5.6 ± 1.4 (8)	6.5 ± 2.2 (8)	17 ± 1	54.5 ± 1.5	7.37 ± 0.02
135 days	4.9 ± 0.9 (7)	16.4 ± 4.9 (7)	19 ± 1	54.0 ± 1.6	7.35 ± 0.01
138 days	7.4 ± 1.9 (8)	24.6 ± 6.4 (7)	18 ± 1	52.4 ± 1.6	7.37 ± 0.01
140 days	10.3 ± 3.1 (7)	33.6 ± 13.3 (7)	18 ± 2	54.4 ± 1.4	7.35 ± 0.02
Day of birth	17.4 ± 3.8 (7)	72.3 ± 12.6 (7)	17 ± 2	59.2 ± 2.8	7.32 ± 0.02

Data are shown as means ± SE. Number in parenthese indicates number of animals; N for fetal cortisol and fetal blood gases is the same.

^*Significant change over time.

Ewes were treated at the end of surgery and for 2 days postoperatively with analgesic (flunixin meglamine;1 mg/kg sid; Merck Animal Health) and for 5 days postoperatively with antibiotic (Polyflex; 12–15 mg/kg bid; Boehringer Ingelheim Vetmedica, St. Joseph, MO); rectal temperature was measured twice a day for 5 days. Ewes were fed a diet of pelleted feed per National Research Council standards adjusted for the ewe’s body weight and fetal gestational age.

One ewe was euthanized in labor due to dystocia, one delivered a live lamb that died shortly following birth after suspected dystocia, and one fetus was hypoxic throughout the pregnancy and died. The chronically hypoxic fetus (Po₂ of 10–15 mmHg) was not included in the analysis. In this study, a total of four male and four female lambs were born to eight ewes at 143 ± 1 days (range 139–148 days) of gestation.

Analysis of the acquired aortic pressure, HR, and fECG was performed using analysis modules in DSI Dataquest Open A.R.T 4.31 and Ponemah 5.00 Software Blood Pressure Analysis Module and ECG PRO, respectively. For calculation of MAP, amniotic pressure was used as the reference pressure. Hourly means of aortic and amniotic pressures and HR were calculated; HR calculation used the peak to peak intervals in the aortic pressure waveform. For the fECG analysis, data were imported into the Ponemah Analysis modules, and the software was assigned elements of the ECG (i.e., start and end of P wave, QRS, and T wave peak and end); these template cycles were then finely adjusted by the operator, and added to a library for each animal that was used to match the remaining cycles. Unmatched cycles were excluded from the analysis. About 80% of cycles in each period chosen for analysis could be matched.

Mean aortic pressure, HR, and fECG characteristics after surgery.

One hour means of mean aortic pressure (MAP), HR, and fECG parameters [P duration (atrial depolarization), PR interval from the start of atrial depolarization to the start of ventricular depolarization), QRS duration (ventricular depolarization), corrected QT interval (depolarization and repolarization of the ventricles, QT interval corrected for RR interval, QTc), and ST interval (isoelectric period between ventricular depolarization and repolarization)] were collected and analyzed over the first 24 h following the end of surgery. After the first postoperative day, 6-h means were calculated over the next four postoperative days for statistical analysis.

MAP, HR, and fECG characteristics in late gestation.

MAP and HR were calculated as 24-h means over the 14 days before birth; this period was chosen to allow inclusion of data from all fetuses following the 5 days of recovery from surgery. The parameters of the fECG (P and QRS duration; PR, QR, QRS, QTc, and ST intervals) were calculated for the 1-h interval between 0600 and 0700 for the 14 days before birth.

MAP, HR, and fECG characteristics 24 h before birth.

In the 24 h before birth, MAP, HR, and fECG parameters were calculated as hourly mean values; in the last hour before delivery and, when possible, in the first 10 min after birth data were calculated as 1-min means. In most cases, the signal was disrupted by the final process of delivery, although intermittent reading over a number of beats was still possible between maternal “pushes.” Because the repeater was affixed to the ewes abdomen and needs to be within ~40 cm of the fetal transmitter, postnatal signals were not reliably collected unless someone was present at the time of birth to move the repeater to the neck of the lamb. However, in many cases the newborn stayed close enough to the ewe to collect data in the immediate postpartum period. In all animals, the time of birth was confirmed using the telemetry record for fetal/neonatal temperature, which in all cases showed a decrease in temperature measured by the telemetry device at the time of birth.

Analysis of diurnal rhythms for MAP, HR, and fECG characteristics.

On days 11, 10, and 9 before birth, hourly means for MAP, HR, and fECG were calculated across the 72 h to test for diurnal rhythms. The cosinor [Michael Sachs. (2014) cosinor: Tools for estimating and predicting the cosinor model. https://cran.r-project.org/web/packages/psych/index.html; version 1.6.9.) and psych [Revelle, W. (2016) psych: Procedures for Personality and Psychological Research, Northwestern University, Evanston, IL; https://cran.r-project.org/web/packages/cosinor/index.html] packages were used in R software (http://www.R-project.org/) to fit hourly mean data to a cosine function, and depict the 24 h variation graphically. The Metacycle package was used in R software to test for significance of rhythmicity in individual fetuses, and to determine acrophase (time of the peak of the rhythm), amplitude (the difference between the peak and the mean value), and MESOR (estimate of the average value of the oscillating variable) describing the rhythms (67).

Statistical analysis.

Analyses of fetal MAP, HR, and fECG parameters following surgery and in late gestation were performed using one-way ANOVA corrected for repeated measures across time in IBM SPSS version 23.0 (IBM, Armonk, NY). For all analyses, statistical significance was set for P < 0.05. Values represent means ± SE unless otherwise stated.

RESULTS

MAP, HR, and fECG characteristics after surgery.

In the 24 h following the end of surgery, fetal MAP did not significantly differ over time (Fig. 1B). The change in HR significantly declined over the first 24 h following surgery (from 216 ± 10 to 176 ± 5 beats/min in the first 13 h) (Fig. 1A); likewise, the duration of the RR interval calculated from the fECG significantly increased (281 ± 14 to 335 ± 18 ms in the first 5 h). The duration of the QRS and ST intervals significantly increased in the hours after the end of surgery (QRS: 24.9 ± 1.8 to 27.1 ± 2.1 ms in the first 4 h, and ST: 139 ± 13 to 160 ± 9 ms in the first 9 h, respectively) (Fig. 1, C–G). After end of the first day, there were no significant changes over time in MAP and HR nor in any parameter of the ECG over the next 4 days (data not shown).

Fig. 1.

Fetal heart rate (HR; A), aortic blood pressure (BP; B), and fetal ECG (fECG) parameters (C–G) in the 24 h following the end of surgery (n = 6). BPM, beats/min. Data are means ± SE. *Significant change over time.

MAP, HR, and fECG characteristics in late gestation.

Over the last 14 days of fetal life, fetal MAP significantly increased (from 45.2 ± 2.2 to 55.6 ± 3.4 mmHg) and HR significantly decreased (from 156 ± 16 to 145 ± 15 beats/min) (Fig. 2, A and D). There were no significant changes over time in in the measured parameters of the fECG except for the RR interval (Fig. 3, A, C, E, G, and I).

Fig. 2.

Fetal heart rate (A–C) and aortic blood pressure (D–F) in the final 14 days before birth (A and D), in the final 24 h (B and E), and from 60 min before birth to 10 min after birth (C and F). Data are means ± SE of 1 h of data in the morning in A and D (n = 8): 1-h data over 24 h in B and E (n = 8) and 1 min of data in C and F (n = 6). Note that the y-axis range in C is different than A and B. *Significant change over time.

Fig. 3.

fECG parameters in the final 14 days before birth (A, C, E, G, and I) and 24 h before birth shown as hourly means (B, D, F, H, and J). Data are means ± SE (n = 8). *Significant change over time.

MAP, HR, and fECG characteristics before and after birth.

In the final 24 h before birth, the hourly mean in aortic pressure significantly increased (from 51.8 ± 3.3 to 56.6 ± 2.1 mmHg) and the hourly mean HR significantly increased (from 154 ± 4 to 160 ± 6 beats/min) as labor progressed (Fig. 2, B and E). The PR and RR intervals were the only parameters of the ECG to significantly change over this period (Fig. 3, B, D, F, H, and J). The PR interval initially decreased as labor progressed, with a small increase shortly before birth (Fig. 3B). In the immediate perinatal period, the fetal HR decreased in the final hour before birth from 173 to 144 beats/min at birth and averaged 217 beats/min over the first 10 min of ex utero life (Fig. 2C). Similarly, the fetal blood pressure rose from 58 to 66 mmHg in the final hour and averaged 68 mmHg in the early postnatal period (Fig. 2F).

Analysis of diurnal rhythms for MAP, HR, and fECG characteristics.

Analysis for 24 h patterns over 3 consecutive days in late gestation indicated significant patterns in HR and the RR interval in four out of seven fetuses; in the duration of the ST interval in three out of seven fetuses; in the duration of the QRS, PR, and QTc intervals in two out of seven fetuses; but in MAP in only one out of seven fetuses (Fig. 4, A–F; acrophases, amplitudes, and MESOR for each parameter are presented in Table 2).

Fig. 4.

Twenty-four hour diurnal patterns in fetal HR, BP, and fECG parameters in fetuses (n = 7). Circles indicate the hourly mean values on the 3 consecutive days (days 11, 10, and 9 before birth); lines indicate the cosine fit of these mean values; black bars on x-axis indicate period of lights off.

Table 2.

Values of parameters of the 24-h patterns in the late gestation ovine fetus

	Baseline	Amplitude	Acrophase	No. of Fetuses With Significant Rhythm
HR, beats/min	172	3.9	0159	4/7
MAP, mmHg	46	0.8	2241	1/7
RR interval, ms	356	6.7	2236	4/7
ST interval, ms	173	3.4	1909	3/7
QRS interval, ms	28.4	0.9	0057	2/7
PR interval, ms	79.8	2.1	0239	2/7
QTc interval, ms	279	3.9	1831	2/7
QR interval, ms	12.8	3.6	0255	0/7
P interval, ms	36.3	1.6	2324	0/7

Data are mean values for 7 fetuses calculated using the Metacycle Package for R. HR, heart rate; MAP, mean aortic pressure.

DISCUSSION

In this study, we described a method for the successful implantation of radiotelemetry devices to chronically measure fetal MAP, HR, and fECG. We used this method to characterize changes over time in MAP, HR, and parameters of the fECG, following surgery, throughout late gestation, and in the immediate perinatal period. Furthermore, we examined these for 24-h patterns of rhythmicity in late gestation ovine fetus. However, the appreciable advantage of the use of telemetry in this animal is its ability to be used throughout labor and delivery and after birth to measure acute responses that may be affected by various treatments or interventions. Although other methods that require some maternal restraint or housing in metabolic cages can be used to effectively determine fetal cardiovascular responses over shorter periods of study, these present difficulties during labor and delivery.

Our results using telemetry agree with already published data on MAP and HR using other methods in “control” fetuses. We found that fetal MAP and HR increased over the final day as birth approached, with a concomitant decrease in the PR as well as in the RR interval, although no other parameters of the fECG were changed. In the ovine fetus, a gradual rise in fetal blood pressure normally occurs in the hour before parturition, which is associated with the occurrence of uterine contractions (10). In these control fetuses, aortic pressure steadily rose over the last hour of life, including the immediate postpartum period. These findings reflect the cardiovascular adaptations associated with birth and the transition to ex utero life. In the sheep, cardiac output approximately doubles following birth, as the ventricular circulation transitions from a parallel (fetal) to in-series circuit (infant/adult) (18). Systolic blood pressure also rises at birth in part due to the removal of the low resistance placental unit, increased systemic vascular resistance, and circulating vasoactive hormones including cortisol and catecholamines (19, 60). These adaptations are thought to provide for an increased metabolic demand and processes regulating breathing and thermogenesis, which is reflected by changes in blood flow to certain organs during this transition (2). Comline and Silver (10) also showed that fetal HR during labor varied, but decreased overall in the last hour before birth. Our data also reveal a gradual increase in HR; we observed a variable increase in HR at the time of birth that was not significant overall. Others have shown that HR increases immediately following birth in humans (12) and in mice is accompanied by a decrease in the duration of the QRS and QTc intervals (37).

This methodology also allowed us to examine the recovery of the fetus from our surgical manipulation, which involved fairly extensive manipulation of the fetus. Although there was no significant change in MAP in the hours following surgery, ST interval and QRS duration increased, while the HR (and RR interval) decreased after surgery and stabilized after several hours. These findings indicate that ventricular activation and relaxation times are affected by surgery; however, these changes were transient and there was a rapid recovery during the first postoperative day. These postoperative changes likely reflect the fetuses intraoperative exposure to isoflurane, which is known to decrease blood pressure and HR in the fetus, and to alter the ECG in adolescents and adults (22, 40, 50, 56).

We confirmed the increase in arterial pressure and decrease in HR during late gestation that was reported by Unno et al. (62); Unno et al. showed in their study that the fetal HR increases between 140 and 143 days of gestation, corresponding to the exponential rise in fetal cortisol concentrations. However, we found little change in the parameters of the fECG occurring over the same period. This indicates that while the cardiac tissue undergoes maturational changes throughout late gestation, the dynamics of the fetal cardiac action potentials and the conduction system are mature by the last 0.10 of gestation, suggesting that cardiac ion channels are present and mature by this time. This is consistent with our previous observation in transcriptomic analyses that genes associated with Purkinje fiber and ion channel maturation are not differentially expressed between 0.90 and 0.97 gestation in the ovine fetus (46). We found only 11 genes that were related to voltage-gated ion channels, ligand-gated ion channels, or other ion channels and significantly changed between 0.90 of gestation and postnatal day 14 (55); however, the change in expression of these genes occurred postnatally. Similar studies in the mouse have shown that although expression of various cardiac ion channel genes changes considerably from embryonic day (E)17.5 to adulthood, less change occurs from E17.5 to 1 day postnatal, as a relatively mature conduction system is likely necessary for survival at birth (16). Although studies using fECG and magnetocardiography indicate that considerable change in some parameters occur between mid and late gestation in the human fetus (particularly increasing durations of the P and T waves, PR, QRS, and QTc intervals), less change occurs during late gestation (9). It has been proposed that this is due to the growing myocardial mass during development, which creates a larger surface area that the electrical signal must traverse (9). In premature human neonates (26–37 wk of gestation), the duration of the P wave, QRS and QT intervals are shorter compared with the full-term neonate, and this is consistent with an association of the size of the chambers to the conduction times through the heart (58, 59). However, estimates of the duration of the P wave and QRS and PR intervals in the human fetus at ≥37 wk of gestation were considerably longer than measured in the sheep fetus from our study (P: 53 vs. 32 ms; PR:110 vs. 78 ms; QRS duration: 53 vs. 28 ms; Ref. 9). These discrepancies are likely not related to differences in heart size since it is similar in both species at birth (this study: 25.5 vs. 24.5 g; Refs. 15, 32). Instead, this might reflect differences in methodology, proteins involved in cardiac conduction, or perhaps, most likely, differences in autonomic activity between the human and sheep fetus. Measurements of the PR interval in our study are similar to those made invasively in other studies in the sheep fetus (65).

Significant diurnal rhythms in HR were observed in the majority of the fetuses in the current study. Studies in several species, including the late gestation fetal baboon, sheep, and humans have shown significant 24-h rhythms in HR, HR variability, and MAP (6, 7, 41, 62). These variations can be influenced by behavioral state, rest and activity cycles, and body/breathing movements (31, 33, 34, 43, 57). The autonomic nervous system of the fetus also considerably influences HR. Sympathetic tone predominates in the fetus, but the relative contribution declines as gestation advances and parasympathetic tone increases; however, neither sympathetic nor parasympathetic input alone are essential for the diurnal rhythms of HR (11, 21, 63). Unno et al. (62) previously found an acrophase in HR occurring ~2330, using 14-h light and 10-h dark periods, compared with ~0200 in our study that used 12-h light and 12-h dark periods. In our study, there was no observed rhythm in MAP. Others have found a significant rhythm in MAP, with peak times occurring around the time of the HR peak (62). Maternal factors including daylight, feeding, and melatonin are known to contribute to the entrainment of fetal circadian rhythms, which could contribute to differences between studies (39, 45, 61). The absence of significant variation in mean pressure could reflect an absence of rhythm in fetal cortisol. In previous studies from our laboratory (3), we found no diurnal rhythm in fetuses in which ewes were fed ad libitum; in contrast, others have found a rhythm in animals when fed for a limited time in the morning (35, 53, 54).

We also found there to be significant 24-h rhythms for several parameters of the fECG occurring in some of the fetuses, including the duration of the PR, RR, ST, QTc, and QRS intervals. These are likely to be driven by the autonomic nervous system as it contributes to the regulation of the ECG through innervation of the sinoatrial and atrioventricular nodes and the ventricular myocardium. However, intrinsic factors that might influence the rhythms of the fECG include circadian expression of cardiac ion channels, as has been shown for adult human and rodent hearts (28, 49, 68) and in association with circadian changes in HR (20, 69) and metabolism (26).

There are some pitfalls that we encountered while implementing this method that should be considered, particularly instances of unexpected and sudden fetal death occurring in the first 48 h of implantation. These are easily detected post hoc as acute increases in aortic pressure and HR, followed by severe bradycardia and progressive hypotension until death. Several of these instances could be attributed to umbilical cord occlusion resulting from manipulation during surgery or insufficient replacement of the amniotic fluid with sterile saline. Because of these issues, we adjusted our protocol to minimize fetal manipulation during surgery and to infuse sterile saline into the amniotic cavity. Implementing these strategies greatly improved the likelihood of a successful preparation.

Perspectives and Significance

Our study demonstrates that the use of implanted radiotelemetry devices is a powerful tool that can be used to allow analysis of fECG and aortic pressures in the chronically instrumented late gestation ovine fetus. The method offers the ability to continuously measure fetal cardiac function over a relatively long period in gestation, in our study from day 118 to delivery at 140–147 days. Notably, we were able to measure ECG as well as fetal pressure during labor and delivery and in the first minutes after birth, even in instances when the investigator was not present. This methodology also allows study of the ovine fetus without maternal restraint and therefore with minimal stress to the ewe. This therefore offers an opportunity for refinement of the animal model. The results from this study also help establish baseline measures of ECG parameters in the late gestation ovine fetus and characterize fetal HR and MAP changes in the fetus in unrestrained and freely moving ewes.

GRANTS

This work was funded by National Institute of Child Health and Human Development Grant HD-057871 and the American Heart Association Grant 14GRNT20420048.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.A., C.E.W., and M.K.-W. conceived and designed research; A.A., C.E.W., and M.K.-W. performed experiments; A.A. and M.K.-W. analyzed data; A.A., C.E.W., and M.K.-W. interpreted results of experiments; A.A. prepared figures; A.A. and M.K.-W. drafted manuscript; A.A., C.E.W., and M.K.-W. edited and revised manuscript; A.A., C.E.W., and M.K.-W. approved final version of manuscript.