Main

Studies of both human and ovine fetuses near term have identified the clustering of physiologic and behavioral parameters into states(1). In the near-term human fetus, four behavioral states can be distinguished. States 1F and 2F are predominant and correspond to non-REM and REM sleep, respectively(2). In a recently published study in the human fetus using real-time ultrasonography, a considerable decrease (47%) was observed in fetal urine production during behavioral state 2F compared with 1F(3). This finding was later confirmed by others(4). It has been suggested that an increase in fetal renal artery resistance during 2F compared with 1F is responsible for the observed reduction in fetal urine production rate. However, Doppler studies of the renal artery in the human fetus during state 1F and 2F are not consistent(3–5).

In the ovine fetus at a gestational age of 120 d, two behavioral states can be identified(6–8). The LV ECoG state is accompanied by REM activity, rapid irregular breathing movements, and episodes of swallowing, and is regarded as paradoxical or REM sleep. The HV ECoG state corresponds to slow wave or non-REM sleep during which the fetal movements as described for LV ECoG, for the most part, are absent(1, 6–10).

In the sheep fetus, many physiologic variables have been shown to be state-dependent, in particular those of the cardiovascular system. During LV ECoG, increased blood flow due to vasodilatation in brain stem areas and in the gastrointestinal tract was shown(11–13). During this state a decreased blood pressure and heart rate are reported(12, 13). Micturition is also state-dependent and is associated with LV ECoG(7). To our knowledge, data on urine production rate in relation to behavioral states in the sheep fetus have not been published.

In view of the observations in the near-term human fetus, we investigated the relation between fetal urine production rate and ECoG in the near-term ovine fetus. The simultaneous measurement of renal artery blood flow enabled us to study also the relationship between renal blood flow and electrocortical states in the near-term fetal lamb. We hypothesized that in the ovine fetus urine production and renal blood flow during LV ECoG would be lower than during HV ECoG.

METHODS

Surgical procedures. Studies were performed on eight fetal sheep of mixed breed between 123 and 127 d of gestation (term 147 d). Fetuses underwent the surgical procedure at least 7 d before the initiation of the measurements. Food but not water was withheld for 16 h before surgery. Surgical procedures were conducted using aseptic techniques. General anesthesia was induced by 1 mg/kg diazepam, 10 mg/kg ketamine, and 0.4 mg of glycopyrulate i.v. The ewe was intubated and mechanically ventilated, while anesthesia was maintained with ketamine (14 mg/min) i.v. and halothane (0.5%) in oxygen inhalation anesthetics. Maternal catheters were placed in the left femoral artery and vein for blood sampling and administration of antibiotics.

The uterus was exposed and opened through a paramedian incision in the lower abdomen. Fetal hind limbs and abdomen were exteriorized. A closed tip catheter was inserted through a small abdominal incision in the bladder(duodenum feeding tube 391.16, Vygon Nederland BV, Veenendaal, The Netherlands). The urethra and urachus were ligated to ensure all produced urine was collected by the catheter. Through a paravertebral incision, the right renal artery was carefully exposed and an ultrasonic flow-transducer(Transonic Systems Inc., Ithaca, NY) was placed around the renal artery. Care was taken not to kink or occlude the artery by choosing a transducer size with a loose fit to the artery (2S or 3S transducers). The flow probes were tested in a warm saline solution, and during the surgical procedure the blood flow signal of the probes was also checked in situ. The fetus was returned into the uterus where-after a second uterine incision was made, and the fetal head and the forelimbs were exteriorized. Polyvinyl chloride catheters (0.9 mm outside diameter, 1.75 mm inside diameter) were inserted in the fetal right axillary artery and vein and in the left axillary artery for blood sampling and continuous blood pressure recording. A catheter was also placed in the fetal trachea to record breathing movements, and two double lumen catheters were placed in the amniotic cavity, one for continuous recording of intraamniotic pressure and one for returning the collected urine. Two pairs of Teflon-insulated stainless steel wire electrodes (N12-50F-405-0, New England Wire Corporation, Lisbon, NH) were placed bi-parietally on the fetal dura to measure ECoG. The uterus was closed in layers, and two pairs of electrodes were sewn in the myometrium wall to record myometrial contractility. All catheters and probe leads were tunneled to the ewes left flank, and the abdomen was closed in layers.

After surgery 500 mL of warm Ringer's lactate was infused into the amniotic cavity to replenish the lost amniotic fluid during surgery. During and after surgery antibiotics [cefuroxime (Zinacef ®), Glaxo, Zeist, The Netherlands] were given to the ewe and fetus (1200 mg i.v. to the ewe, 100 mg i.v. to the fetus, and 200 mg into the amniotic cavity).

Analgesics were administered to the ewe after surgery [metamizol (Novalgin®), Hoechst, München, Germany, 4 mL i.v. and 4 mL intramuscularly]. At the end of the experiments, the animals were killed using a veterinary euthanasia solution (pentobarbital). Ultrasonic flow probes were checked for their baseline (zero) value while still in situ, and in case of a value different from zero, data obtained during the experiments were corrected. The fetuses were autopsied to verify the condition and location of catheters. Guidelines for care and use of animals as approved by the local Animal Medical Ethics Committee were followed.

Maintenance. The ewes were housed individually always in the company of another ewe. Food and water was provided ad libitum. Lights were on between 0700 and 1900 h. The first 5 d after surgery antibiotics were given twice a day to the ewe and fetus, distributed as described above. At least once a day, blood gas analysis was performed with a temperature correction at 39°C. Arterial blood pH, Pao2, Paco2, and base excess were analyzed on an ABL 330 (Radiometer, Copenhagen, Denmark), Sao2, Hb, and O2 content were analyzed using a hemoximeter (OSM3, Radiometer, Copenhagen, Denmark), with a correction for fetal or maternal ovine blood. The blood catheters were flushed continuously by a heparinized physiologic saline solution (10 IU/mL) at a rate of approximately 0.6 mL/h. Fetal urine was collected by gravity into a sterile container continuously and was returned to the amniotic cavity by means of a hand-constructed Automatic Urine Controller. By means of an electronic liquid level controller, every 10 mL of urine were automatically returned to the amniotic cavity, and each cycle was recorded by means of a counter. To prevent amniotic fluid from being collected into the urine container, a one-way valve(Braun Medical, Uden, The Netherlands) was installed between the Automatic Urine Controller and the amniotic catheter.

Urine flow was measured continuously by means of an external placed ultrasonic flow probe (1N, Transonic System Inc., Ithaca, NY) situated in-line with the bladder catheter. Fetal tracheal, intraamniotic, and arterial blood pressure were measured by means of uniflow pressure transducers (Baxter Healthcare Corporation, Santa Ana, CA).

Physiologic measurements. The ewe and fetus were allowed to recover for 6 d after surgery. Previous studies have shown that an interval of at least 6 d is required after surgery to allow for full recovery of renal function(14). Thereafter, measurements were performed during 2 consecutive days. Recordings were made between 0830 and 1700 h for 6 h. Blood gas analysis was performed every hour. Urine osmolality was determined from samples taken during both HV ECoG and LV ECoG (six urine samples per animal per day). Osmolality was measured by freezing point depression (Osmomat 030 Gonotec, Salm en Kipp, Breukelen, The Netherlands). Electrocortical activity was filtered and amplified by means of a bioamplifier(Nihon-Khode, RM6000, Tokyo, Japan) (filtering high cutoff 100 Hz). Fetal renal blood flow, urine flow, tracheal pressure, intraamniotic pressure, and arterial blood pressure were measured continuously.

Fetal arterial blood pressure and tracheal pressure were corrected for amniotic pressure. All signals were recorded on an eight-channel recorder(Gould Inc. model 8802, Cleveland, OH) and after A/D conversion stored on a disk by an on-line computer system (Poly®, Inspector Research, Amsterdam, The Netherlands) at a sample rate of 10 Hz.

Data and statistical analysis. To analyze the ECoG signal, a frequency distribution of the voltage amplitude was made of each set of data(6 h per animal per day). A cut-off amplitude of LV and HV values was determined on the basis of the bimodal distribution of the amplitudes (Fig. 1). All voltage values ±5% of the determined cut-off amplitude were excluded, to preclude the inclusion of overlapping values. A state was recognized only when the voltage distribution complied with the terms and lasted at least 3 min. The HV ECoG and LV ECoG states were visually verified, and those parts of the recordings which revealed signal artefacts or a bad signal-to-noise ratio (too much noise to obtain a reliable signal) were excluded from the study. Physiologic measurement during each of the electrocortical states were pooled over 2 d, resulting in one paired observation for both HV ECoG and LV ECoG state per fetal sheep. The results are presented as mean ± SEM. For statistical evaluation the Wilcoxon matched pair test was used, each fetal sheep serving as its own control. Data were considered significant at p < 0.05.

Figure 1
figure 1

Example of the frequency distribution of the ECoG in an ovine fetus. After a dividing between LV and HV ECoG is made, LV and HV limits are determined as 5% (of the maximal voltage scale) below and above the dividing point, respectively.

Full size image

RESULTS

Table 1 shows the mean values of arterial blood gases, arterial blood pressure, urine flow, urine osmolality, and renal blood flow of the eight fetuses during the study. All of these variables were within normal ranges. No significant differences in these variables were observed between the two consecutive days.

Table 1 Average fetal physiological data during HV ECoG and LV
Full size table

Fetal ECoG could be analyzed in HV ECoG and LV ECoG states in 63 ± 6% of the recording time. Fetuses spent 41.5 ± 4.2% in HV ECoG state and 58.5 ± 4.22% in LV ECoG state. A total number of 451 HV ECoG and LV ECoG states were included in the study (209 HV ECoG and 244 LV ECoG states). The average duration of the HV ECoG and LV ECoG states was 7.0 ± 0.8 min and 8.5 ± 1.0 min, respectively.

Figure 2 shows the urine flow rates per animal. There was no significant difference in urine flow rate between HV ECoG and LV ECoG states, mean urine flow values being 0.40 ± 0.03 mL/min and 0.41± 0.03 mL/min, respectively. Also, no significant changes in urine osmolality were seen between HV ECoG and LV ECoG states(Table 1).

Figure 2
figure 2

Average urine production rate in HV ECoG (solid bars) and LV ECoG (open bars) state per individual animal. Data show fluctuations in urine production rate between animals but no significant differences were seen between the two ECoG states per animal.

Full size image

Figure 3 shows the renal blood flow values of the right kidney in HV ECoG and LV ECoG of the individual animals. There was no significant difference in renal blood flow during states, the average flow rate in HV ECoG and LV ECoG being 11.4 ± 0.8 mL/min and 11.2 ± 0.8 mL/min, respectively. Fetal mean arterial blood pressure in HV ECoG and LV ECoG was 44.7 ± 1.7 and 43.4 ± 1.5 mm Hg, respectively(p > 0.05).

Figure 3
figure 3

Average renal blood flow of the right kidney was of the same magnitude in HV ECoG (solid bars) and LV ECoG (open bars) states per animal.

Full size image

DISCUSSION

Using real time ultrasound, it has been observed that the estimated urine production rate in the human fetus is highly dependent on the behavioral state. Hourly fetal urine production rate was almost reduced by half during state 2F, compared with state 1F(3). State 2F is considered to be similar to active or REM sleep. Interestingly, the reported findings in the human fetus were very similar to previously reported observations in adult male volunteers during REM sleep(15). This prompted us to investigate the influence of behavioral state in the near-term sheep fetus where urine flow rate can be measured exactly in a continuous way. In view of the inconsistent data on state dependence of renal blood flow in the human fetus(3–5), the sheep experiments also gave us the opportunity to evaluate the effects on renal blood flow.

In the present study an average urine production rate of 0.40 ± 0.03 mL/min was found, which is similar to the results reported by other investigators at this gestational age(16, 17). The arterial blood gas values of the eight fetuses studied were well within the normal ranges(8, 18). Although blood pressure is also considered to be state-dependent(12, 13), we and others(19) did not find significant differences between blood pressures during both states.

The time spent in HV ECoG and LV ECoG states (41.5 and 58.5%, respectively) is in agreement with the results of others(1, 7, 12, 20). Urine flow was measured continously by an in-line ultrasonic flow probe, which has the advantage that also very low flow rates can be recorded. Because the duration of a behavioral state can be short (in this study a 3-min limit was used), other techniques for urine measurement such as droplet counters can be inaccurate.

The lack of almost any difference in urine flow rate between HV ECoG and LV ECoG in the ovine fetus is in sharp contrast with the reduction by almost half, previously reported in the human fetus. This raises the question to what extent differences in the study design or species differences are responsible for these divergent observations between the fetal sheep and the human fetus. Regarding the differences in methodology, the techniques used in the present study (i.e. the direct and continuous measurement of fetal urine flow and renal blood flow), are much more able to detect differences in real time than those that can be performed in the human fetus.

In the present study, no differentiation was made between the LV ECoG state with and without REM. Theoretically, it is possible that urine production during LV ECoG with REM would be lower than during LV ECoG without REM. However, because the time spent in LV ECoG without REM is less than 10% of the total time spent in LV ECoG(10), it is unlikely that this has a major impact on total urine production in LV ECoG.

In the ovine fetus, the behavioral states differ in some aspects from the behavioral states as observed in the human fetus. In contrast to the the ovine fetus, where the ECoG activity is recorded and used to determine a behavioral state, in the human fetus one has to rely on heart rate patterns and a variety of fetal movement activities to assess the various behavioral states. A major difference between the ovine and human fetus is the absence of breathing movements during HV ECoG in the ovine fetus, whereas in the human fetus regular breathing movements can be seen during HV ECoG(21). This raises the question whether behavioral states in the ovine fetus are similar to behavioral states in the human fetus.

It has been suggested that the lower urine production rate during 2F in the human fetus is the result of a decrease in renal blood flow during active sleep(3). It has been shown that fetal blood flow to the cerebrum and the skeletal musculature increases during state 2F compared with state 1F(22, 23). Because the combined ventricular output in the human fetus does not change between electrocortical states(24), it is reasonable to assume that blood flow to some organs is reduced, which could include the kidneys. In the human fetus, ultrasound-Doppler studies of renal arteries show conflicting results: in a previous study no changes were found in renal blood flow between states 1F and 2F(25), whereas others reported a decrease in the pulsatility index during 2F(4). The latter finding indicates a decrease in renal impedance to blood flow rather than suggesting an increase in renal blood flow during state 2F.

In the ovine fetus, behavioral states are known to affect the regional distribution of cardiac output(26), possibly affecting aortic blood flow and consequently renal blood flow. A more recent study, however, did not find changes in abdominal aortic blood flow between behavioral states in the ovine fetus(27), which supports the findings of the present study, where in the ovine fetus renal blood flow during HV ECoG and LV ECoG was found to be similar. Our findings corroborate those of Jensen et al.(11), who also found no significant difference in renal blood flow between electrocortical states in the sheep fetus. They used the microsphere technique, which limits the number of measurements per animal, whereas we measured renal blood flow continuously, providing precise real time data. In a study on the effect of sleep states on regional blood flow distribution in piglets, Cote and Haddad(28) also found no sleep state dependency on renal blood flow. Also in the adult cat, no significant changes in renal blood flow occurred during different sleep states(29).

The possible role of neuroendocrine changes during REM sleep has to be considered, as in human adults REM sleep is associated with considerable fluctuations in PRA; PRA reaches a nadir just before the beginning of REM sleep, whereas during REM sleep PRA is considerably reduced(30, 31). To our knowledge, data on behavioral state-related changes in fetal PRA have not been reported. In general, information about state-dependent hormonal changes in the ovine fetus is scarce. It has been shown that both plasma epinephrine and norepinephrine levels decrease during LV ECoG in the fetal sheep(12), but obviously this does not substantially influence urine flow rate and renal blood flow. Our finding that state changes are not accompanied by changes in urine osmolality in the ovine fetus, whereas in human adults urine osmolality increases during REM sleep(15), supports the concept that in this respect species differences exist.

In conclusion, urine production rate and urine osmolality in the sheep fetus show no difference between behavioral states, which is in sharp contrast to recent observations in the human fetus. Also, renal blood flow remained virtually unchanged between states in the sheep fetus. Species differences are the most likely explanation for the observed dissimilarities in renal responses between behavioral states in the human and sheep fetus, although the effect of methodologic differences cannot be completely ruled out.