Main

The chronically cannulated ovine fetus has been very widely used to investigate fetal fluid balance (1), fetal renal function (2), development of the hypothalamic-pituitary-adrenal axis (3), fetal growth (4), and potential neuroprotective agents for use in fetal/neonatal brain injury (5). Most of the experiments have been carried out in the last third of gestation, from 90 to 100 d, until term (145-150 d). Some circulatory and metabolic studies have been carried out in mid-gestation fetuses (6) predominantly using umbilical vessels, but these are much more reactive and have a greater tendency to spasm, than the carotid artery or the jugular vein.

We were interested in the regulation of adrenocortical hormone production and fetal fluid balance at mid-gestation and cannulated a number of fetuses in the carotid artery and/or the jugular vein. In the control fetuses, infused only with small volumes of saline, a number of abnormalities of fluid balance occurred, sometimes accompanied by hydranencephaly-defined as the replacement of a normal brain by a membranous fluid-filled sac (7). Two fluid-filled compartments develop in the pregnant sheep; the amniotic and allantoic sacs. Amniotic fluid surrounds the developing fetus, whereas the allantoic fluid is contained in two interconnected compartments, one at the tip of the pregnant horn and one in nonpregnant horn in a singleton pregnancy. The abnormalities consisted of hydrallantois-excess accumulation of allantoic fluid, which is the most common fluid sac abnormality occurring in ruminants (8), and hydrops fetalis-the accumulation of excess fluid in the tissues and body cavities of the fetus (9).

In older fetuses (∼ 100 d of gestation) we had shown that hydranencephaly could be produced within 2 wk of the bilateral cannulation and occlusion of the carotid arteries (10). This study details the investigations of the fetal fluid balance in ovine fetuses cannulated at approximately 70 and 100 d of gestation. Previous theories as to the etiology of hydrallantois and/or hydrops fetalis include anemia of the fetus (11,12) and excess urine production secondary to increased glucocorticoid or atrial natriuretic peptide levels (13–16). The aims of the current study were 2-fold: 1) to see whether mid-gestational ovine fetuses are more prone to developing fluid abnormalities than older fetuses postvascular cannulation of the fetal neck and 2) to try to elucidate potential causes of the observed abnormalities.

METHODS

Group 1. All experiments were approved by the Howard Florey Animal Ethics Committee in accordance with guidelines set by the National Health and Medical Research Council of Australia. In the first group of animals (n = 12), ovine fetuses were cannulated at 70 ± 1 d of gestation. Silastic cannulas (internal diameter of 0.64 mm and outer diameter of 1.19 mm, Dow Corning, Midland, MI) were inserted into one jugular vein, one carotid artery, as well as the fetal bladder, via an abdominal incision, leaving urethral and urachal openings patent, while the ewe and fetus were under general anesthesia (2-3% halothane in oxygen). A silastic cannula (inside diameter of 1.57 mm, outside diameter of 3.18 mm) was placed in the amniotic compartment. In 4/12 fetuses vascular cannulation was performed on one jugular vein only. In five animals, with two cannulas, cannulation was accomplished on the same side of the fetal neck, and in three fetuses the cannulas were placed contralaterally as shown in Table 1 (see "Results"). One jugular vein of each ewe was cannulated at the time of surgery. Ewes were housed in individual metabolism cages with free access to food and water. Vascular cannulas were flushed with heparinized saline to maintain patency. In eight animals there were twins present.

Table 1 Effect of cannulation of the carotid artery and/or jugular vein on the development of fluid abnormalities in the ovine fetus at mid- and late-gestation
Full size table

Experimental protocol. Animals were allowed 4-5 d to recover from surgery. All fetuses received an infusion of isotonic saline (0.19 mL/h) via the jugular vein cannula for 72 h. A 5-mL fetal blood sample was taken before and at 24, 48, and 72 h after the start of saline infusion from an arterial cannula, where possible, or a venous cannula. This was used to measure fetal blood gases and hematocrit as well as plasma ACTH, aldosterone, and cortisol concentrations. At 72 h, an additional 5 mL of fetal blood was obtained for measurement of fetal plasma electrolytes and ANF. The volume of blood removed was replaced with an equal volume of isotonic saline. A 10-mL blood sample was taken concurrently from the ewe for measurement of cortisol and aldosterone. In addition, urinary flow rate was measured each day of the experimental protocol (4 d) for 2 h. The fetal bladder was emptied by a 1-h period of drainage, before the measurement of urinary flow.

Group 2. In the second group of animals, one fetus from each of 6 ewes had bilateral carotid artery occlusion at 100 ± 0.4 d of gestation. The common carotid artery was ligated in the mid-cervical region, caudal to the superior thyroid artery, while the ewe and fetus were under general anesthesia (2-3% halothane in oxygen). Silastic cannulas (inside diameter of 0.76 mm; outside diameter of 1.65 mm, Dow Corning, Midland, MI) were inserted into both carotid arteries, below the ligatures, the right jugular vein and transabdominally into the fetal bladder. In addition, two cannulae (inside diameter of 1.57 mm; outside diameter of 3.18 mm) were inserted into the amniotic compartment. There were noncannulated twins in four ewes. Fetuses were killed 2 wk postsurgery. Control data were collected from age-matched control animals (n = 10) that were involved in other experimental protocols, and had one carotid artery and one jugular vein were cannulated. Urine flow was measured once per week, and 10 mL of fetal blood was obtained for the assessment of fetal plasma cortisol and aldosterone concentrations. The volume of blood removed was replaced with an equal volume of isotonic saline. In addition, a 10-mL blood sample was taken from the ewe for measurement of plasma cortisol and aldosterone concentrations. Animals were then killed with an overdose of pentobarbitone sodium (Lethobarb, Arnolds of Reading, Peakhurst, Australia). At postmortem, samples of amniotic and allantoic fluids were taken for analysis of electrolytes, and the total fluid volume of each sac was measured in graduated cylinders.

Sample analysis. Fetal blood gases were measured on a Ciba Corning 278 blood gas system and Hb on a Corning 2500 Co-oximeter (Australian Diagnostic Corp. Melbourne, Australia). Osmolality was measured using an Advanced osmometer (Advanced Instruments, Needham Heights, MA). In all fluids obtained (fetal plasma, urine, and amniotic and allantoic fluids) a Synchron CX-5 clinical system was used to measure sodium, potassium, chloride, urea, creatinine, glucose, fructose, lactate, and total protein. Coefficients of variation for all parameters have been described previously and are generally <10% (13).

Hormone assays. ACTH was measured using a commercially available kit (Dyno, B.R.A.H.M.S. Diagnostica, Henning, Berlin, GMBH). This immunoradiometric assay recognizes human and ovine ACTH (1-39), but does not cross-react with other ACTH fragments. The sensitivity of the assay was 2 pg/mL, and the intraassay coefficient of variation was 5.2% at 20 pg/mL and 6.3% at 10 pg/mL. Plasma cortisol was measured in fetal and maternal plasma using an RIA, which has been described in detail and validated for use with ovine fetal blood (17). Intra- and interassay coefficients of variation were 10.3 and 13%, respectively. Aldosterone and ANF were also measured using established RIAs (18,19). For the aldosterone assay the intra- and interassay coefficients of variation were 7.8 and 11%, respectively and for the ANF assay, 9 and 14%, respectively.

Statistics. Values for amniotic and allantoic fluid volumes and composition, and hormone levels are expressed as mean ± SEM. Comparisons of data were made using a Student's unpaired t test. Only the data for total protein in allantoic fluid (see Table 3) was log transformed before the t test was used. Also, the Fisher-Freeman-Halton test was used to analyze the association between the type of cannulation and the occurrence of fluid abnormalities.

Table 3 Composition of allantoic and amniotic fluid in group 1 normal fetuses (n = 7) and fetuses with fluid abnormalities (n = 5)
Full size table

RESULTS

To study hormone regulation and function in sheep, cannulation was attempted in 12 ovine fetuses at 70 ± 1 d of gestation (group 1) (term 150 d). This study reports unexpected abnormalities observed in five fetuses as a result of vascular cannulation of the fetal neck. In this study the group that did not develop abnormalities (n = 7) had the same cannulation procedure and treatment.

Of the abnormal fetuses, four had both cannulas (jugular and carotid) on the same side of the neck and the fifth fetus had a jugular cannula only. In the three fetuses with a combination of all observed abnormalities (hydrallantois, hydranencephaly, and hydrops fetalis), carotid and jugular cannulation was performed on the same side of the fetal neck. Of the seven normal fetuses, three had a single venous cannula and four had both cannulas (in 3/4 on the opposite side of the fetal neck) as shown in Table 1. By the Fisher-Freeman-Halton test there was no statistical association between type of cannulation and any fluid abnormality (p = 0.113).

Blood gas analysis revealed no difference between normal and abnormal fetuses (Table 2), and the same was true for the hematocrit, except that one hydropic fetus had a marked drop in hematocrit, secondary to hemorrhage, for 24 h before tissue collection (Fig. 1).

Table 2 Fetal arterial blood gas analysis and hematocrit in mid-gestational fetuses
Full size table
Figure 1
figure 1

(A and B) Hematocrit in group 1 fetuses. Hematocrit values in seven normal mid-gestation fetuses (A, ▪) and three mid-gestation fetuses with hydrops fetalis (B, ○), before (0) and at 24, 48, and 72 h after the start of saline infusion.

Full size image

A postmortem, all five abnormal fetuses had a large volume of fluid accumulated in the allantoic fluid compartment, hydrallantois, being 760 ± 141 mL. In this study, hydrallantois was defined as a volume of allantoic fluid greater than the mean + 2 SD of the volume found in the noncannulated twin fetuses (134 ± 61; mean ± SEM; n = 8). The volume of allantoic fluid in noncannulated twins was similar to that found in normal fetuses (104 ± 23; n = 7) (Fig. 2). The solute compositions of allantoic fluid and amniotic fluid samples are shown in Table 3. The allantoic fluid composition in abnormal fetuses differed from normal fetuses in having higher sodium (p < 0.05), chloride (p < 0.001), and lactate (p < 0.05) and lower concentrations of creatinine (p < 0.05), magnesium (p < 0.01), and total protein (p < 0.05). It is noteworthy that the measured solutes account for the majority of the osmolality in allantoic fluid from the abnormal fetuses, whereas they account for just over 60% in normal allantoic fluid. This suggests that some components of normal allantoic fluid have decreased in the abnormal animals. Although the volume of amniotic fluid was similar in these two groups of animals, the composition was altered. There was a significant increase in concentrations of urea (p < 0.01) and lactate (p < 0.001) in abnormal fetuses.

Figure 2
figure 2

Allantoic fluid volume in group 1 and group 2 fetuses. Allantoic fluid volume in the mid-gestation fetuses; (□) normal, n = 7; (□) abnormal, n = 5; (▪) noncannulated twins (n = 8) and later gestation hydranencephalic fetuses 2 wk after ligation of both carotid arteries and one jugular vein; (□) age-matched controls, n = 10; (□) hydranencephalic fetuses, n = 6; and (▪) noncannulated twins n = 4. ***p < 0.001.

Full size image

The excessive volume of allantoic fluid in abnormal fetuses was accompanied by a consistently high (on each day of a 4-d protocol) urine flow rate in two of the three fetuses in which a bladder cannula functioned (21 and 28 mL/h versus 7.5 ± 0.6 mL/h in seven normal fetuses). The third abnormal fetus had a urine flow rate of 8 mL/h. The three abnormal fetuses that had a patent bladder cannula had higher urine osmolality 216 ± 12 mOsmol/kg water than did normal fetuses (n = 7), 147 ± 2 mOsmol/kg water; p < 0.001. In addition, urine composition in abnormal fetuses differed from that of normal fetuses in having significantly higher sodium (87 ± 15 mmol/L versus 54 ± 2 mmol/L; p < 0.05) and chloride concentrations (63 ± 12 mmol/L versus 35 ± 1 mmol/L; p < 0.05).

As shown in Figure 2, hydrallantois did not occur in older fetuses (group 2; 115 ± 1 d of gestation; n = 6) 2 wk after both carotid arteries were ligated and one jugular vein was cannulated, and fetuses become grossly hydranencephalic. Urine flow rate in the six hydranencephalic fetuses was 11 ± 3 mL/h, and that was similar to the urine flow rate found in 10 age-matched normal fetuses. The allantoic fluid composition of the six hydranencephalic fetuses differed from 10 control fetuses in having higher osmolality (p < 0.01) and potassium concentration (p < 0.01) and lower concentrations of fructose (p < 0.01) and glucose (p < 0.05) (Table 4). Changes in amniotic fluid composition between the six hydranencephalic fetuses and normal fetuses (n = 10) are shown in Table 4. The six hydranencephalic fetuses had higher amniotic fluid concentrations of sodium (p < 0.001), potassium (p < 0.05), phosphate (p < 0.001), and lactate (p < 0.05) and lower amniotic fluid concentrations of creatinine (p < 0.05) and fructose (p < 0.05) when compared with 10 normal fetuses.

Table 4 Composition of allantoic and amniotic fluid in group 2 normal fetuses (n = 10) and fetuses with experimental hydranencephaly (n = 6)
Full size table

Upon opening of the fetal skull, in mid-gestation fetuses (group 1), in 3/5 abnormal fetuses the fetal brain was found to have deteriorated and the brain cavity was filled with liquid-hydranencephaly. Also, the same animals were found to have developed a hydrops fetalis observed as accumulation of excess fluid within the tissue and body cavities of the fetus. The wet body weight of the hydropic fetuses was 317 ± 22 g (n = 3) and this was significantly greater than the body weight of normal fetuses (249 ± 17 g; n = 7) or noncannulated twins (245 ± 11 g; n = 8) (p < 0.05).

The composition of fetal plasma was very similar in both groups of mid-gestation fetuses except that abnormal fetuses had lower plasma CO2 (25 ± 0.8 mmol/L, n = 5 versus 28 ± 0.4 mmol/L, n = 7; p < 0.05) and higher plasma urea concentrations (6.4 ± 0.3 mmol/L, n = 5 versus 5.3 ± 0.2 mmol/L, n = 7; p < 0.01). As shown in Figure 3, fetal plasma concentrations of ACTH (10.6 ± 3.5 pg/mL), aldosterone (62 ± 10 pmol/L), cortisol (7.4 ± 1.7 nmol/L), and ANF (30 ± 5 pmol/L) in abnormal fetuses were not different from the plasma concentrations of these hormones in normal fetuses. The fetal plasma ANF concentration was measured at completion of the 72-h saline infusion in all normal fetuses (n = 7) and in four of the five abnormal fetuses. As shown in Figure 3 plasma ANF concentrations were similar in normal and abnormal fetuses, except that one abnormal fetus had an ANF concentration of 74 pmol/L, while having a normal urine flow rate of 8 mL/h. In addition, maternal plasma concentrations of cortisol (30 ± 5 nmol/L) and aldosterone (106 ± 11 pmol/L) of normal fetuses were not different from maternal plasma concentrations of cortisol (26 ± 3 nmol/L) and aldosterone (132 ± 19 pmol/L) in abnormal fetuses.

Figure 3
figure 3

Hormonal status in group 1 fetuses. Fetal plasma concentrations of ACTH, aldosterone, cortisol, and ANF in seven normal (□) and five abnormal (□) (except for ANF; n = 4) mid-gestation fetuses. Results are mean ± SEM.

Full size image

In group 2 the fetal plasma cortisol (5.5 ± 1.2 nmol/L) and aldosterone (106 ± 26 pmol/L) concentrations in hydranencephalic fetuses were similar to the plasma concentrations of cortisol (5.7 ± 0.9 nmol/L) and aldosterone (101 ± 13 pmol/L) in normal fetuses. Maternal plasma cortisol (29 ± 6 nmol/L) and aldosterone (160 ± 36 pmol/L) concentrations in control fetuses were not different from maternal plasma concentrations of cortisol (29 ± 4 nmol/L) and aldosterone (154 ± 33 pmol/L) in hydranencephalic fetuses.

DISCUSSION

The most interesting observation in the current study is the occurrence of fluid abnormalities (hydrallantois and hydrops), and hydranencephaly in the mid-gestational ovine fetus postvascular cannulation (neck). Fetal fluid abnormalities never occurred in older fetuses in which both carotid arteries and one jugular vein were ligated even though fetuses become hydranencephalic (10), suggesting that the mid-gestational ovine fetuses are more prone to abnormalities postvascular cannulation in the neck. Also, abnormalities are more likely to occur if both the carotid artery and the jugular vein are cannulated on the same side of the fetal neck.

Hydrallantois, although quite common in the cow, is relatively rare in the sheep; however, it occurs when the ovine fetus is exposed to excess glucocorticoid (dexamethasone) administered to the ewe for 2-3 d around mid-gestation (13,14). The increase in volume of the allantoic fluid in previous studies could be accounted for by a large increase in fetal urine production, and the fact that fetal urine at that stage flows only into the allantoic sac. In the current study, hydrallantois cannot be explained always by a simple increase in fetal urine production, because although two fetuses of this group had a urine flow rate approximately 2-3 times that of normal over the course of the 3 d, one fetus produced what could be considered an appropriate amount of urine for fetuses of this age. It is noteworthy, however, that the composition of allantoic fluid in the mid-gestation fetuses with abnormalities varied significantly from that of normal fetuses. The major differences were higher sodium and chloride concentrations, so that more of the osmolality could be accounted for by the constituents measured in the abnormal fetuses. In normal fetuses measured solutes could account for just over 60% of allantoic fluid osmolality, suggesting that some unmeasured component of normal allantoic fluid has decreased in abnormal fetuses. Similar changes have been noted in hydrallantoic cows (20).

The presence of hydrops in some mid-gestation fetuses in this study was most unexpected as it was previously demonstrated that hydrallantois at this age, when induced by excessive glucocorticoids, was not accompanied by any gross fetal abnormalities, although subtle changes were observed in some organs (13). In older fetuses several experimental interventions produced hydropic fetuses with various reliability. It is reported that a small accumulation of thoracic or abdominal fluid (<50 mL) in 5/8 fetuses and a significant accumulation (650 mL) in only one fetus occurred when 120-d ovine fetuses were subjected to severe chronic anemia (oxygen content was reduced to 1.6 mL/dL and hematocrit to 13.3% in anemic fetuses compared with 6.9 mL/dL and 32% in controls) (11). Blair et al. (12) studied the effect of anemia on the occurrence of fluid abnormalities in similar gestational age ovine fetuses, and found that 6/12 fetuses developed modest ascites (78 mL). In the current study, mid-gestational ovine fetuses with hydrallantois and hydrops fetalis, which had developed postvascular cannulation, had normal blood gas analysis and hematocrit, indicating some other underlying mechanism must account for the development of these fluid abnormalities.

Stevens et al. (15) found gross hydrops fetalis in one ovine fetus and ascites in 12 fetuses (88 mL) that were subjected to induced supraventricular tachycardias. In a later study by Nimrod et al. (16) it was demonstrated that ovine fetuses with hydrops fetalis, produced by right atrial pacing had a 3-4-fold increase in plasma ANF concentration compared with control. We did not study arterial pressure and heart rate in fetuses with fluid abnormalities, but plasma levels of ANF were not altered in most fetuses.

The most reliable experimental manipulation in producing hydrops fetalis proved to be the excision of the major lymph vessels where 5/5 fetuses (120 d old) became hydropic. However, in the same study ligation of the same lymphatic vessels resulted in only 1 out of 11 animals demonstrating hydrops fetalis (21).

Hydrops fetalis was associated with hydranencephaly in the mid-gestation ovine fetus, but not in the more mature fetuses. Alterations of interstitial fluid pressure will normally reflect changes in interstitial fluid volume (22). Studies in rats suggested a link between the nervous system and interstitial fluid pressure. In these studies an increase in the negativity of the interstitial pressure and increased vascular permeability of the rat trachea was achieved by a stimulation of the sensory fibers in the vagal nerve (23,24), indicating that interstitial fluid pressure can be influenced by the nervous system. It is possible that in our study mid-gestation ovine fetuses developed fluid abnormalities due to some damage of the vagal nerve during vascular cannulation. However, no fluid abnormalities occurred when more mature fetuses were experimentally made hydranencephalic by ligation of both carotid arteries and cannulation of one jugular vein (10).

In conclusion, mid-gestation ovine fetuses have a much higher risk of developing fluid abnormalities postvascular cannulation than do later gestation fetuses. In mid-gestation ovine fetuses hydrallantois can occur without hypoxia, anemia, or an increase in hormones known to be diuretic and natriuretc.