Main

It is still difficult to decide whether a newborn baby has suffered perinatal hypoxia or asphyxia of a severity likely to cause permanent brain damage. Currently available clinical and biochemical indicators, including cord blood pH and other measures of acid-base status, do not predict neurologic outcome well(1–3). Because many metabolic processes are oxygen-dependent, severe tissue hypoxia causes wide-ranging biochemical disturbances with accumulation of a variety of intermediary metabolites. These include organic acids, which are excreted by the kidneys and accumulate in urine. We previously observed pronounced, transient, abnormalities in the urinary organic acid profiles of infants during metabolic stress(4), and speculated that organic acid analysis might have potential value for detecting severe perinatal hypoxia. We went on to study 50 term newborns with fetal distress or birth asphyxia and 27 controls, and observed statistically significant abnormalities for the group of 12 severely asphyxiated babies. These predicted long-term outcome successfully in nine of them(5). They were not seen in less severely distressed babies. Organic acid analysis of the first samples of urine passed after birth might be a useful adjunct to other measures of birth asphyxia. However, we needed to know more about the origin of the organic acid disturbance to define its clinical significance.

ATP is degraded rapidly during hypoxia, producing adenosine, a vasodilator and putative neuroprotective agent(6) and, in muscle, inosine monophosphate. Hypoxanthine is a metabolite of both compounds. Adenosine has been measured in cord blood(7) but its short half-life (<10 s) precludes its use diagnostically. Hypoxanthine has a longer half-life (40 min in pigs)(8) and has been measured in blood(9–13), cerebrospinal fluid(11), amniotic fluid(11, 14), and urine(11, 15–17) to detect perinatal asphyxia. Because the placenta clears hypoxanthine efficiently(18), levels in cord blood may underestimate severe self-limiting intrauterine hypoxic events. Although often harmless, these sometimes cause problems neonatally and may even result in brain damage. Once hypoxanthine has been excreted by the kidneys it accumulates in urine. Thus urine provides a cumulative record of events(11), and its analysis is more likely to detect a previous serious episode. It is known that arterial plasma hypoxanthine levels increase in fetal hypoxia(18, 19), but urinary excretion has not been studied in utero.

In this observational study, we monitored late gestation fetal sheepin utero stressed by moderately severe isocapnic hypoxia for 60 min. We aimed to investigate first whether fetal hypoxia induced under controlled conditions produces an organic aciduria similar to that of severely birth-asphyxiated human newborns and second, the origin and timing of changes in urinary organic acid and hypoxanthine excretion with hypoxia. Our hypothesis was that biochemical disturbances are not due to hypoxia per se, but to chemoreceptor-driven reflexes that decrease blood flow to peripheral tissues, particularly skeletal muscle, causing ischemia. To this end, we analyzed urine from acutely hypoxic fetal sheep after section of both carotid sinus nerves, which reduces hypoxia-induced peripheral vasoconstriction dramatically(20, 21). For the study, we developed a new specific assay for urinary hypoxanthine using gas chromatography-mass spectrometry.

METHODS

Surgical preparation. All procedures were conducted under license and in accordance with the 1986 Home Office regulations on animal experimentation. Twelve cross-bred pregnant ewes were studied at 119-133 d(median 126 d) of gestation (term is 147 d) using general anesthesia (1 g of sodium thiopentone for induction; 2-3% halothane in oxygen for maintenance) and sterile techniques. The fetus was partially exteriorized through a uterine incision. Multistranded stainless steel wire electrodes (Cooner Wire Co., Chatsworth, OH, USA) were sewn into the fetal chest to record the ECG. Polyvinyl catheters (inside diameter, 1.0 mm; outside diameter, 2.0 mm; Portex Ltd., Hythe, Kent, UK) were placed in a carotid artery, a jugular vein, and the amniotic sac, and a catheter was placed in the bladder via a puncture incision and secured by a purse-string suture. A maternal femoral vein was catheterized. All leads were exteriorized through a maternal flank and secured to the ewe's back in a plastic bag. All catheters had a blunt needle in the distal end which was attached to a 3-way stopcock. The animals studied were primarily involved in other research which involved slightly different additional instrumentation. In the six intact fetuses, ultrasonic flow transducers (Transonics Inc, Ithaca, NY) were implanted around one femoral artery. In the six denervated fetuses, both carotid sinus nerves were sectioned(21). Treatment of the animals before our study did not differ. Postoperatively, the vascular catheters were maintained patent by a slow infusion of heparinized saline (50 IU mL-1 at 0.125 mL h-1). Sodium benzylpenicillin (Crystapen, Glaxo Ltd., Middlesex, UK) was administered i.v., 600 mg to the ewe, 200 mg to the fetus, and into the amniotic sac (300 mg). The antibiotics were given daily for 4-5 d postoperatively, but not on the day of experimentation. Results are also presented for femoral blood flow of 12 denervated fetuses that we studied previously under identical experimental conditions. Details of their surgical preparation have been reported(21).

Experimental procedure. At least 4 d elapsed after surgery before the experiments. These were based on a 3-h protocol(21). After a control period of 1 h, fetal isocapnic hypoxia was induced for 1 h by reducing the maternal inspired oxygen fraction to 0.09 (18 L min-1 of air, 22 L min-1 nitrogen and 1.2 L min-1 of carbon dioxide). The ewes were then returned to normoxia, and monitoring continued for 1 h. During hypoxia, fetal Pao2 was reduced from 2.86 ± 0.12 kPa to 1.55 ± 0.04 kPa (mean ± SEM). At 15-min intervals throughout the protocol, arterial blood samples (0.5 mL) were taken anaerobically from the fetus for determination of pH, blood gases, and hematocrit (model 1302; Instrumentation Laboratory, Warrington, Cheshire, UK; measurements corrected to 39.5°C) and whole blood lactate (11 fetuses)(YS1 model 2300 glucose and L-lactate analyzer; Yellow Springs Instrument Co., Yellow Springs, OH). In addition, up to five 1-mL arterial blood samples were collected for the other studies (results not reported here). To collect urine samples the stopcock was opened, and urine was drained into a sterile measuring container. Up to 3 mL were removed for analysis, and this volume was replaced with isotonic saline. Urine and saline were returned to the amniotic sac. Urine was collected before hypoxia, after 15 and 45 min of hypoxia, and 15 and 45 min (intact fetuses) posthypoxia. Arterial blood pressure, fetal heart rate, and FBF were recorded continuously on a chart recorder(Linearcorder FWR 3701, Graphtec, UK).

Postmortem examination. On completion of the experiments, ewes and fetuses were killed with an overdose of sodium pentobarbitone to determine fetal and organ weights and to examine the organs macroscopically for evidence of damage.

Analytical methods. Creatinine was measured by a kinetic alkaline picrate method using a discretionary autoanalyzer (CX7 analyzer; Beckman Instruments, High Wycombe, Bucks, UK). Organic acids were measured by a published gas chromatographic method(22), but using a heated split injector, split ratio 20:1, instead of cool on-column injection. An aliquot of urine equivalent to 1 μmol of creatinine was oximated with hydroxylamine hydrochloride, acidified to pH 1.0, saturated with sodium chloride, then extracted successively with ethyl acetate and diethyl ether. After evaporation, the urinary organic acids were derivatized with bis(trimethylsilyl)-trifluoroacetamide and analyzed by capillary gas chromatography with flame ionization detection. Unknown compounds were identified by combined gas chromatography-electron impact mass spectrometry(5892A series II gas chromatograph linked to a 59731A quadrupole mass spectrometer; Hewlett Packard, Bracknell, UK). This was particularly important for 3-hydroxyisobutyric acid which coelutes with 3-hydroxybutyric acid. To enable comparisons, acids were quantified from their peak areas usingn-tetracosane as the internal standard. The detection limit for most of the nonpolar acids was <1 mg/mmol of creatinine. Concentrations were related to the urinary creatinine which, although not ideal(6), corrects for differences in water diuresis.

Hypoxanthine was measured by a new method developed in this laboratory using gas chromatography-mass spectrometry with single ion monitoring. Urine with creatinine greater than 1 mmol/L was diluted to this concentration with deionized water. Other samples were analyzed neat. To an aliquot of urine equivalent to 1 μmol of creatinine was added 60 μL of 0.799 mmol/L 2,3,7,8-tetra-deuterated hypoxanthine as the internal standard (MSD Isotopes Ltd., Croydon, UK). Half of this mixture was applied to 1 mL of cation exchange resin (AG 50W-XB resin, 200-400 mesh, hydrogen form, Bio-Rad Laboratories, Hemel Hempstead, UK) in a disposable 2-mL plastic syringe plugged with cotton wool. After washing twice with 3 mL of water, hypoxanthine was eluted with 3 × 0.5 mL of 5 mol/L ammonium hydroxide and collected into a 1.8-mL glass vial. The pooled eluates were freeze dried. Hypoxanthine was derivatized by heating the residue for 30 min at 60°C with 100 μL ofN-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide(Sigma Chemical Co., Dorset, UK) and 10 μL of Analar pyridine as a catalyst. One microliter was injected into a gas chromatograph-mass spectrometer (5892A series II gas chromatograph linked to a 59731A quadrupole mass spectrometer, Hewlett Packard, Bracknell, UK), fitted with an SE-30 capillary gas chromatography column, 30 mm × 0.25 mm inside diameter, film thickness 0.25 μm (Econo-Cap, Alltech UK, Carnforth, UK). The split ratio was 13:1. Helium was the carrier gas; flow rate was 1 mL min-1. The injector was maintained at 250°C. The column temperature was held at 150°C for 5 min, then increased by 5°C min-1 to 270°C and held for 10 min. The retention time of hypoxanthine was 22 min. Two of the four original deuterium atoms of the internal standard were replaced during derivatization. Quantification of deuterated hypoxanthine and hypoxanthine extracted from urine was achieved by monitoring the ions at m/z 309 and 307, respectively, resulting from loss of a tert-butyl group, [M- 57]+, from the molecular ion [M]+. The concentration of urinary hypoxanthine (μmol mmol-1 creatinine) was calculated from the integrated peak areas. The method was sensitive to 0.5 μmol mmol-1 creatinine and linear to 85 μmol mmol-1 creatinine. Recoveries of standard hypoxanthine added to urine were: for 2.9 μmol mmol-1 creatinine added, mean 101.5% (range 99.8-102.7%); for 10.1μmol mmol-1 creatinine, 97.3% (96.0-99.9%); for 52.9 μmol mmol-1 creatinine, 99.8% (99.4-100.6%) (n = 6). Between-batch imprecision was low, the coefficient of variation being 3.7% at 8.0 μmol mmol-1 creatinine (n = 8) and 2.1% at 47.5 μmol mmol-1 creatinine (n = 6).

Statistical analysis. The Mann-Whitney U test was used for comparisons.

RESULTS

All of the fetuses recovered from the hypoxic challenge. None was excluded from the study.

Blood gases, pH, and lactate. During normoxia, arterial blood gases, pH, and blood lactate of intact and denervated fetuses were not significantly different. Pao2 fell from a control value of 2.86± 0.12 kPa (mean ± SEM) to 1.55 ± 0.04 kPa (p< 0.001) during hypoxia. Basal Paco2 was 6.65 ± 0.20 kPa, and levels were maintained during hypoxia (6.50 ± 0.22 kPa after 45 min). The pH fell significantly from 7.36 ± 0.01 to 7.22 ± 0.03(intact; p < 0.01) and to 7.31 ± 0.02 (denervated;p < 0.01). Despite rapid recovery of Pao2 with reoxygenation, the pH of intact fetuses fell further to 7.18 ± 0.03, 15 min posthypoxia, when it was significantly lower than for denervated fetuses(7.31 ± 0.02; p < 0.01), and remained significantly depressed after 45 min (7.20 ± 0.03). Blood lactate of intact fetuses increased from 1.02 ± 0.08 mmol/L basally to 2.98 ± 0.44 mmol/L after 15 min (p < 0.01) and 6.41 ± 0.51 after 45 min of hypoxia (p < 0.01). Levels increased further with reoxygenation to 7.53 ± 1.06 and 7.31 ± 0.75 at 15 and 45 min, respectively. The basal lactate concentration of denervated fetuses was 0.96 ± 0.15 mmol/L. Although levels increased significantly during hypoxia to 1.89± 0.19 (p < 0.05) and 3.77 ± 0.53 (p < 0.01) at 15 and 45 min, respectively, they did not increase with reoxygenation(3.77 ± 0.75 after 15 min). Lactate concentrations of intact fetuses were significantly higher than denervated fetuses both during and posthypoxia(p < 0.05).

Fetal heart rate and blood pressure. In the intact fetuses, fetal heart rate fell from 165 ± 6 beats/min to 134 ± 4 beats/min after 5 min of hypoxia (p < 0.01). The rate returned to control levels toward the end of the hypoxic period, and tachycardia (204± 4 beats/min) developed with reoxygenation (p < 0.01 compared with basal). There was a small, but not significant, increase in arterial blood pressure from 42 ± 2 mm Hg to 52 ± 3 mm Hg with hypoxia. The response of denervated fetuses was similar to that of our earlier study(21). There was no early bradycardia, and the heart rate increased, but not significantly, toward the end of hypoxia, returning to basal posthypoxia. Arterial blood pressure did not increase significantly from basal levels of 50 ± 1 mm Hg.

FBF. In this study, this was monitored only in the intact fetuses. FBF decreased sharply from 47 ± 1 mL min-1 to 20± 3 mL min-1 after 5 min of hypoxia (p < 0.01) and 12 ± 2 mL min-1 after 45 min (p < 0.01). There was a rebound rise to 60 ± 4 mL min-1 45 min posthypoxia(p < 0.05, compared with basal) (Fig. 1). InFigure 1 we also present, for comparison, FBF data for 12 denervated fetuses measured under similar experimental conditions and taken from our previous study(21). In those fetuses, FBF did not change significantly with hypoxia.

Figure 1
figure 1

FBF of intact (•) fetuses during the experimental protocol. Femoral blood flow of twelve carotid sinus-denervated fetuses(□) in our earlier study(21) are presented for comparison. Values are means ± SEM.

Full size image

Postmortem findings. There were no differences in body or organ weights between the groups, and there was no macroscopic evidence of damage to the kidneys or other organs.

Urinary organic acids and hypoxanthine. One intact fetus became oliguric during hypoxia, and there was insufficient urine for analysis. In another, there were problems with the bladder catheter drainage, and urine was not collected posthypoxia.

Urinary organic acids: intact fetuses. In normoxia the predominant organic acids were lactic and pyruvic acids, the tricarboxylic acid cycle intermediates, citric, succinic, and 2-oxoglutaric acids, a tyrosine metabolite, 4-hydroxybenzoic acid, hippuric acid, and glyoxylic acid. There were small amounts of 3-hydroxyisobutyric acid, produced during valine catabolism, but another valine derivative, 2-hydroxyisovaleric acid, was not detectable in quantifiable amounts. 2-Hydroxybutyric was detected (in trace amounts) in one sample only.

During hypoxia, the profile changed. Disturbances were apparent by 15 min but were gross and statistically significant by 45 min (Table 1; Figs. 2a and 3a). The striking features were: markedly increased lactic acid excretion accompanied by raised pyruvic acid, large increases in the tricarboxylic acid cycle intermediates, and in 3-hydroxyisobutyric acid, excretion of 2-hydroxybutyric acid (all fetuses) and of 2-hydroxyisovaleric acid (four of six). With reoxygenation, the organic aciduria intensified. The disturbance was maximal 15 min posthypoxia, but acid concentrations were not statistically different from those during hypoxia, probably because of the small number of samples and wide interindividual variation. With mass spectrometry, four fetuses were found to excrete trace amounts of 2-oxoisocaproic acid (from leucine), and two, 2-oxo-3-methylvaleric acid (from isoleucine) during or after hypoxia. Excretion of 4-hydroxybenzoic acid was not altered by hypoxia.

Table 1 Urine organic acids and hypoxanthine of intact fetuses before, during, and after hypoxia
Full size table
Figure 2
figure 2

Urinary lactic acid of fetal sheep before, during, and after 60 min of hypoxia. (a) Values for six intact fetuses. Samples were not analyzed from one fetus during hypoxia and one fetus posthypoxia;(b) values for six carotid sinus-denervated fetuses.

Full size image
Figure 3
figure 3

Urinary 3-hydroxyisobutyric acid of fetal sheep before, during, and after 60 min of hypoxia. (a) Values for six intact fetuses. Samples were not analyzed from one fetus during hypoxia and one fetus posthypoxia; (b) values for six carotid sinus-denervated fetuses.

Full size image

Urinary organic acids: denervated fetuses. In normoxia the urinary organic acid concentrations were not significantly different from those of intact fetuses. Excretion had not increased after 15 min of hypoxia, but by 45 min lactic acid concentrations were significantly higher than basal, and smaller, significant, increases were found for pyruvic, succinic, fumaric, and 3-hydroxyisobutyric acids. 2-Hydroxybutyric acid was detectable in five of six samples and 2-hydroxyisovaleric acid in four (Table 2; Figs. 2b and 3b). In contrast to intact fetuses, there was no surge in excretion posthypoxia when there were highly significant differences between the two groups (Table 2).

Table 2 Urine organic acids and hypoxanthine of carotid sinus denervated fetuses before, during, and after hypoxia
Full size table

Urine hypoxanthine. The urinary hypoxanthine of intact animals was similar to basal (11.1; 1.5-25.6 μmol mmol-1 creatinine, median and range) after 15 min of hypoxia, but by 45 min had increased significantly to 68.1 (30.0-184.4) μmol mmol-1 creatinine (Table 1). Fifteen-minute posthypoxia excretion was still significantly elevated compared with basal, reaching peak values in two fetuses. Levels were declining by 45 min. Basal hypoxanthine concentrations of denervated fetuses were significantly lower than for the intact group (p < 0.05)(Table 2). They had increased significantly by 45 min of hypoxia to a median concentration of 8.5 (range 2.8-32.8) μmol mmol-1 creatinine (p < 0.05) but at this peak were only within the range found basally for intact fetuses. Concentrations declined posthypoxia. Figure 4,a and b, shows the striking difference between groups.

Figure 4
figure 4

Urinary hypoxanthine of fetal sheep before, during, and after 60 min of hypoxia. (a) Value for five intact fetuses. Samples were not analyzed from one fetus posthypoxia; (b) values for six carotid sinus-denervated fetuses.

Full size image

DISCUSSION

In the late gestation fetal lamb, acute hypoxemia triggers a rapid chemoreflex response characterized by an initial bradycardia followed by an increase in heart rate and blood pressure with a major redistribution of blood flow(20, 21, 23–28). This is largely eliminated by bilateral section of the carotid sinus nerves(20, 21). The physiologic responses of intact and denervated fetal lambs in this study were typical of those reported. Femoral blood flows of denervated animals were not recorded here, but were probably similar to those of our earlier study under the same conditions(21) (see Fig. 1), because the heart rate responses were the same, indicating complete denervation.

In normoxia, the urinary organic acid excretion of intact and denervated fetuses was not significantly different. However, the intact and denervated fetuses had significantly different responses to hypoxia. In intact fetuses changes in organic acid excretion were apparent by 15 min, statistically significant by 45 min, and maximal 15 min after reoxygenation. In contrast, organic acids of denervated fetuses were increased significantly only after 45 min of hypoxia, and there was no surge in excretion posthypoxia. Because the changes were large in animals in which we documented a dramatic decrease in FBF and barely significant in denervated fetuses in which FBF would decrease more slowly and to a lesser extent(21), it is probable that a major contribution to the organic acid disturbances was from altered metabolism of skeletal muscle, which may become almost ischemic during the severe vasoconstriction induced by hypoxia in intact fetuses. The further increases posthypoxia were probably due to a “wash-out” of metabolites during the transient vasodilatation seen with recovery(23), which was evident in the increased FBF of intact fetuses but which does not occur in denervated fetuses. Altered liver and renal tubular metabolism probably also contributed to the organic aciduria. Brain was not the source because cerebral blood flow increases in fetal hypoxia(27).

Lactate is the main end product of anaerobic glucose metabolism. Although 5-10% of blood lactate is cleared by a single passage through the placenta in fetal asphyxia(18), arterial blood lactate increases rapidly and then falls with reoxygenation. Efflux of lactate from the fetal hind limb was found to increase with moderate hypoxia and was highest during recovery(18). This probably accounts for much of the increase. Decreased clearance by the liver(28) could also contribute. Blood lactate concentrations increased in both groups of fetuses, but were significantly higher in the intact group. However, the difference was small in comparison with the very large differences in their urinary lactate concentrations, explained by cumulative lactate excretion during the sampling periods and indicating considerably greater lactate production by hypoxic, intact, fetuses.

A new observation was that fetal hypoxia has a significant impact on valine metabolism evident from large increases in the intermediate, 3-hydroxyisobutyric acid, and excretion of 2-hydroxyisovaleric acid, not found in normoxia. Peak urine concentrations were much lower in denervated fetuses(p < 0.01). Skeletal muscle is very active in branched chain amino acid catabolism and has considerable capacity for valine oxidation(29–31). The observed abnormalities can be explained by inhibition of two catabolic enzymes: branched chain oxoacid dehydrogenase (EC 1.2.4.4) and 3-hydroxyisobutyric acid dehydrogenase (EC 1.1.1.31) by a high intramitochondrial NADH:NAD ratio in hypoxia(29, 32, 33). 3-Hydroxyisobutyric acid is effectively transported out of muscle(31) and was released during perfusion of rat hind quarters(29). Hypoxic muscle was probably the main source of the valine metabolites found here.

2-Hydroxybutyric acid is produced mainly in the liver by catabolism of the amino acids threonine and methionine to 2-oxobutyric acid. Decarboxylation of this compound by the pyruvate dehydrogenase complex (EC 1.2.4.1) is inhibited by a high NADH:NAD ratio in hypoxia(34). 2-Hydroxybutyric acid may have accumulated because of a combination of increased catabolism and reduced liver perfusion(28). The increased tricarboxylic acid cycle acids might have originated in any of the tissues with reduced oxygen delivery, including the kidneys, because these compounds are produced by renal metabolism and secreted by renal tubules(35).

The urinary organic acid excretion profile of fetal sheep in normoxia was similar to that of term human newborns after an uneventful delivery(5), except that the fetuses did not excrete detectable amounts of medium chain dicarboxylic acids or the ketone body 3-hydroxybutyric acid, derived from fat catabolism. Fetal hypoxia caused similar urinary organic acid disturbances to those of severe birth asphyxia, with large increases in lactic and pyruvic acids, and excretion of 2-hydroxybutyric acid. Both hypoxic fetuses and newborns excrete intermediates of branched chain amino acid catabolism, but the profiles differ: fetal excretion of 3-hydroxyisobutyric acid is proportionately higher, and branched chain oxoacids are excreted infrequently compared with those of newborns. This might be explained by a species or age difference in the sensitivity of 3-hydroxyisobutyric acid dehydrogenase to inhibition in hypoxia, to greater permeability of fetal tissues to 3-hydroxyisobutyric acid, or to the nature of the insult, because our fetuses were only moderately hyoxic, without hypercarbia or severe acidosis. A more notable difference is that, in contrast to fetal hypoxia, excretion of tricarboxylic acid cycle intermediates were not significantly increased after severe birth asphyxia, despite the fact that all affected babies had evidence of renal damage with hematuria and/or proteinuria. These differences may be due to decreased glomerular filtration after birth asphyxia, or to differences in renal tubular maturation, because much of this occurs late in gestation(36).

The basal urinary hypoxanthine concentration of five intact fetuses of 11.1(1.5-25.6) μmol mmol-1 creatinine (median, observed range) was similar to concentrations for active term human newborns on the 2nd d of life(mean 12.0, range 5.3-27.5) μmol mmol-1 creatinine(11). In response to hypoxia, excretion by all five fetuses increased markedly, and after 45 min the difference from basal was highly significant (p < 0.01). Concentrations were as high, or higher, after 15 min of reoxygenation, but had clearly fallen by 45 min. Urinary hypoxanthine excretion therefore responded rapidly to changes in tissue oxygen delivery. Our observations are compatible with those of Thiringer et al.(19) who found a substantial increase in arterial plasma hypoxanthine of exteriorized fetal lambs during graded hypoxia, and a close correlation with other indices of fetal asphyxia-pH, base deficit, arterial oxygen saturation, and blood lactate concentration. Our observations are similar to others who have shown increasing urinary hypoxanthine in moderate to severely hypoxic young pigs(37). Hypoxanthine metabolism differs among species but is probably similar in pigs and man(8).

In our study, carotid sinus denervation had a significant impact on urinary hypoxanthine excretion. Basally, urinary concentrations of denervated fetuses were low compared with intact fetuses (p < 0.05). To our knowledge, this has not been reported before but cannot be explained by our data. Perhaps intact fetuses excrete more adrenaline basally, with increased ATP degradation through metabolic effects(7). Alternatively, adenosine production may be one mechanism by which chemoreceptors influence vascular tone. Clearly this observation merits further study. Forty-five minutes of hypoxia led to a small increase above normoxic values (p < 0.05), but at their peak, levels were only as high as the basal concentrations of the intact group, and they fell rapidly posthypoxia. The findings indicate that peripheral vasoconstriction in intact fetuses underlies the normal hypoxanthine response to hypoxia. From the surge in urinary excretion that coincided with restored femoral blood flow, it seems likely that skeletal muscle contributes substantially to the increase posthypoxia. Thiringer et al.(18), from studies of exteriorized fetal lambs, proposed that, during hypoxia, the liver is the main source of increased plasma hypoxanthine, and that skeletal muscle releases hypoxanthine mainly during reoxygenation. Their earlier work had shown that brain was an unlikely source(38). The renal tubules may have contributed some hypoxanthine to urine, because renal blood flow decreases in response to acute fetal hypoxia due to increased renal vascular resistance(27, 28, 39). In contrast, after chemoreceptor denervation(40), or renal denervation(39), the initial response to acute hypoxia is transient vasodilatation with increased renal blood flow and renal vasoconstriction is delayed.

The potential value of hypoxanthine as an indicator of perinatal hypoxia has been explored clinically. So far, most investigators have measured hypoxanthine in cord blood of human newborns. Concentrations were raised in babies with severe birth asphyxia(9, 12, 13), but only in 52% in one study(10). Levels correlated with blood pH, base deficit, and lactate in two studies(12, 13), but not in a third(10). However, cord blood hypoxanthine levels of birth-asphyxiated babies predicted their 2-y-old developmental outcome poorly(12). The problem with measurements of plasma hypoxanthine is that increases may be transient in utero, because the placenta clears hypoxanthine from fetal blood very efficiently(18). Increases are more persistent in urine, and its analysis is more likely to detect a previous serious episode. Few have explored this potential so far, probably because of the practical difficulties of collecting the first voidings of newborns, coupled with analytical problems. Manzke et al.(15) found raised hypoxanthine in urine during the first 24 h of life in 58% of 45 babies with moderate or severe birth asphyxia. Harkness et al.(16) and Laing et al.(17) proposed that urine collected on the 2nd d of life would be more helpful, because raised hypoxanthine then could be explained by renal damage and would imply a serious circulatory shut-down. In their studies, babies who were neurologically abnormal for more than 48 h from birth had the highest excretion, but there was no correlation between urinary hypoxanthine and developmental quotient at 1 y of age(17).

Our fetal observations indicate that the organic acid disturbances and increased hypoxanthine excretion are due largely to peripheral vasoconstriction, particularly in skeletal muscle. They are transient and resolve when oxygen is restored. They are indicators of overall hypoxia and not of brain hypoxia. Neither do they imply organ damage. Nevertheless, they do indicate a serious hypoxic event that probably reduced arterial oxygen content to below 1.5 mM(26). However, it should be stressed that often even severe insults in utero do not cause brain damage, if self-limiting. The study does not define the duration of hypoxia needed to cause a significant biochemical disturbance. Although we did not observe significant changes until 45 min of hypoxia, they might be apparent after a shorter episode, during the wash-out of metabolites associated with reoxygenation. The findings for lactic acid support the view that an acute antepartum hypoxic event is more likely to be detected by analysis of urine, in which disturbances are cumulative, than of blood in which metabolite changes are minimized by placental clearance and may be transient if hypoxia is relieved before delivery. The same is probably true for hypoxanthine, although plasma levels were not measured. Finally, inasmuch as the largest disturbance was in lactate excretion, lactate analysis of the first urine samples after birth should be as informative as a full organic acid profile. The value of urinary hypoxanthine for assessing birth asphyxia needs to be further investigated.