INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BLOOD PRESSURE IS AN IMPORTANT "vital sign" that is carefully monitored in preterm infants. Hypotension is believed to be an important risk factor for cerebral ischemic injury and intracranial hemorrhage, and so preterm infants often receive volume expanders (even in the absence of hypovolemia) and/or inotropic drugs (most commonly, dopamine) to increase blood pressure (1, 18). However, concern has been raised about whether these management practices could be detrimental (29). Seri et al. (25) recently reported that low-dose dopamine infusion does not affect cerebral blood flow (CBF) velocity in sick, normotensive preterm infants. However, little else is known about the cerebrovascular responses of the developing brain to intravenous dopamine and/or acute increases in blood pressure. In adult animals, intravenous dopamine at moderate doses causes cerebral vasodilation, presumably by stimulation of specific vasodilatory dopaminergic receptors (32). In immature animals, dopamine is a less effective inotropic agent, presumably because of immaturity of cardiac ₁-adrenergic receptors. Minimal cerebral vasodilatory, or possibly vasoconstrictive, effects might therefore be expected at low or moderate dopamine doses, because development of vascular dopaminergic receptors is also likely to be incomplete, particularly compared with -adrenergic receptors (8, 12, 30, 31).

The objective of this study was to describe the cerebrovascular responses of developing fetuses at two gestational ages to increasing doses of intravenous dopamine with and without -adrenergic or dopaminergic (D₁) receptor blockade. We tested the hypothesis that dopamine causes dose-dependent cerebral vasoconstriction and that this response may be enhanced because of immaturity of dopaminergic receptor development and increased sensitivity and activation of -adrenergic receptors.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All surgical procedures and experimental protocols were approved by the Animal Care and Use Committees at both institutions.

Animals

Surgical Preparation

View this table: [in this window] [in a new window]   Table 6.   Cardiovascular, cerebrovascular, and oxygenation variables at baseline and in response to phenoxybenzamine and dopamine infusion with -adrenergic receptor blockade in near-term fetal sheep

Physiological Measurements

Regional brain blood flow was measured using the radiolabeled microsphere technique and the least-squares method of differential spectroscopy (14). Approximately 8 × 10⁵ (0.3 ml) microspheres (preterm fetuses) and 1 × 10⁶ (0.4 ml) microspheres (near-term fetuses) labeled with ¹⁵³Gd, ¹¹⁴In, ¹⁰³Ru, ⁹⁵Nb, ¹¹³Sn, or ⁴⁶Sc (DuPont NEN, Boston, MA) were injected into the IVC, followed by 3 ml of blood (preterm fetuses) or 5 ml of saline (near-term fetuses). Reference blood samples were withdrawn from the brachiocephalic artery at a rate of 1.0 ml/min (preterm fetuses) or 2.55 ml/min (near-term fetuses), beginning 30 s before the microsphere injection and continuing for 1 min after the injection was completed. The microsphere injections were not associated with changes in HR, blood pressure, or pulse pressure. After completion of the studies, the ewe and fetus were killed with an overdose of pentobarbital sodium followed by saturated KCl solution. Fetal catheter positions were checked, and the brain was removed at its base and divided at the cephalic border of the pons. All supratentorial tissue was pooled and counted to determine CBF. The cerebellum was removed at the peduncles and counted separately as was the brain stem (pons and medulla). The radioactivity in all tissue and blood samples was determined using a multichannel gamma counter (Packard Instrument, Dowers Grove, IL). All reference and tissue samples contained >400 microspheres.

Blood samples for pH, respiratory blood gases, hemoglobin (Hb) concentration, O₂ saturation, and lactate and glucose concentrations were withdrawn anaerobically into heparinized Natelson glass pipettes. Respiratory blood gases and pH were measured at 39.5°C using the ABL 30 (Radiometer, Copenhagen, Denmark). O₂ saturation and Hb concentration were measured using an OSM-3 hemoximeter (Radiometer). Whole blood lactate and glucose concentrations were measured using a glucose lactate analyzer (model 2300, Yellow Springs Instrument, Yellow Springs, OH).

Blood samples for catecholamines were withdrawn into heparinized 3-ml syringes, kept on ice, and then centrifuged at 4°C for 20 min. The plasma was removed, and samples were stored at 10°C until ready for shipment on dry ice to Dr. Michael Heymann's laboratory (University of California, San Francisco) for analysis using high-performance liquid chromatography with electrochemical detection. Briefly, the thawed plasma samples were extracted by absorption onto acid-washed alumina at pH 8.6, washed, and eluted with 200 µl of 0.1 M perchloric acid. A precise aliquot of the eluate was injected into the high-performance liquid chromatograph for analysis. The mobile phase consisted of 0.1 M sodium phosphate, 100 mg/l EDTA, and 80 mg/l sodium dodecyl sulfate (pH 3.8), flowing at a rate of 1.0 ml/min. The catecholamines were separated on an 8 cm × 4.6 mm ID 3-µm C₁₈ reverse-phase column (ESA, Bedford, MA). The detector (Coulochem Electrochemical Detector, ESA), with a cutoff voltage of 30 mV and detector voltage of 330 mV, detects catecholamines in amounts of 20-25 pg per injected sample. The average recovery of each catecholamine is quite reproducible from day to day, but the recovery of one catecholamine is quite different from that of another. Thus each assay is run with a complete external (unextracted) and internal (extracted from charcoal-treated plasma) standard curve of five concentrations of each catecholamine of interest. The chromatographs of plasma samples are compared with the internal standard curve to determine plasma catecholamine concentrations. The intra-assay coefficients of variation for norepinephrine (NE), epinephrine, and dopamine are 1.4, 2.7, and 14%, respectively.

Experimental Protocol

Dopamine infusions. Dopamine infusion doses were calculated on the basis of fetal weight estimated at surgery. Actual weights obtained at necropsy were then used to determine the actual infusion doses administered, which are listed in Tables 1-6. Dopamine hydrochloride is available only in a 40 mg/ml solution (Elkins-Sinn, Cherry Hill, NJ). Two intravenous dopamine solutions (4 and 0.4 mg/ml) were prepared to provide similar infusion volumes as the dose was increased. Dopamine doses of 2.5 and 25 µg · kg¹ · min¹ were infused at 1 ml/h and doses of 7.5 and 75 µg · kg¹ · min¹ at 3 ml/h.

Blocker infusions. Doses of the -adrenergic blocker phenoxybenzamine and the D₁-receptor blocker SCH-23390 were calculated as described above. Phenoxybenzamine (Research Biochemicals, Natick, MA) was dissolved in ethanol as a 10% solution and administered as a single 5 mg/kg iv bolus injection. SCH-23390 was dissolved in 5% dextrose in water and infused intravenously at 30 µg · kg¹ · min¹.

Fetal blood replacement. All fetal blood withdrawn (except for the last microsphere withdrawal) was replaced with an equal volume of normal saline (near-term fetuses) or fetal blood (preterm fetuses). For preterm fetuses, fetal blood (9 ml) was obtained from the study fetus by a partial exchange transfusion with maternal blood 1 h before the study began. Blood was withdrawn from the fetal axillary artery while warmed maternal blood was simultaneously infused into the fetal IVC. We previously showed that we can obtain enough fetal blood for replacement without causing a significant increase in fetal O₂ half-saturation pressure of Hb, which can occur when maternal blood is used for replacement during the study (10). We did not use this procedure in near-term fetuses, because a much smaller percentage of their blood volume was withdrawn during the study.

Study Protocol

For the dopamine-response experiments, after a single baseline measurement was obtained, dopamine hydrochloride infusion was begun into the fetal IVC. The catheter dead space was estimated by drawing blood back into a syringe, and the catheter was then filled with dopamine solution before the infusion pump was started at the desired rate. Dopamine was infused at four sequential doses: 2.5, 7.5, 25, and 75 µg · kg¹ · min¹. Because we used two concentrations, infusion rates were 1 ml/h for 2.5 and 25 µg · kg¹ · min¹ doses and 3 ml/h for 7.5 and 75 µg · kg¹ · min¹ doses. Measurements were obtained after 15-20 min at each infusion rate. After the 7.5 µg · kg¹ · min¹ measurement, the dopamine solution was changed by first withdrawing all fluid in the dead space of the catheter and then refilling the catheter with the new concentration solution.

For the -adrenergic and D₁-receptor blocker experiments, a baseline measurement was obtained, the blocker was administered (single 5 mg/kg bolus of phenoxybenzamine or infusion of SCH-23390 at 30 µg · kg¹ · min¹), and a measurement was obtained 30 min later. Dopamine infusion was then begun at 7.5 µg · kg¹ · min¹ and increased sequentially to 25 and then 75 µg · kg¹ · min¹. Measurements were obtained after 15-20 min at each infusion rate. At the end of the phenoxybenzamine experiments, a single dose of methoxamine (150 µg) was administered to confirm total -adrenergic blockade. We selected phenoxybenzamine as an -receptor blocker, because previous studies had been done in fetal sheep (23); we used a similar dose and observed the same cardiovascular responses. We selected the D₁-receptor blocker SCH-23390, because the D₁ receptors are believed to be the vascular dopaminergic receptors and are present in the brain (15). The dose we administered by continuous infusion was the same as that reported by Segar et al. (24) in preterm and near-term fetal sheep.

Data Analysis/Calculations

Measurements were calculated and data are reported as means ± SE for all study fetuses. Differences between groups were analyzed by repeated-measures analysis of variance unless a comparison was sought between only two values (or responses); in the latter case, a t-test was performed. For the ANOVA, if the F test was significant, specific differences were sought with the Newman-Keuls test. Significance was considered at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dopamine and blocker infusions were prepared on the basis of estimated fetal weights at surgery. Actual fetal weights were determined at necropsy and were used to determine actual dopamine infusion rates (see Tables 1-4). For the dopamine-response studies, we tended to underestimate the fetal weights at surgery (or perhaps the fetuses grew), and therefore the actual infusion rates were slightly less than desired, but there were no differences between the groups.

Cardiovascular variables (MAP and HR), arterial blood gases, and arterial glucose, lactate, and hemoglobin concentrations are shown at baseline and in response to dopamine infusion in Table 1. All baseline values are consistent with previous values in preterm and full-term fetal sheep. Dopamine caused a dose-dependent increase in blood lactate and glucose concentration and a decrease in arterial pH but no changes in arterial PO₂ (Pa_O2). The percent increase in glucose concentration was greater in the near-term than in the preterm fetuses.

Regional brain blood flows, myocardial flow, MAP, and cerebrovascular resistance (CVR) at baseline and in response to increasing dopamine infusion rates are shown in Table 2. Comparisons between preterm and near-term responses are also depicted in Fig. 1. Dopamine caused a dose-dependent increase in MAP that was greater in near-term than in preterm fetuses at 25 and 75 µg · kg¹ · min¹ (P < 0.05). In near-term fetuses, there was an increase in CBF at 7.5 µg · kg¹ · min¹. In preterm and near-term fetuses, there was a dose-dependent increase in CVR that was greater in preterm than in near-term fetuses at 7.5 and 25 µg · kg¹ · min¹. Cerebrovascular responses in the brain stem and cerebellum paralleled those in the cerebral hemispheres and midbrain. Myocardial blood flow did not increase.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Mean arterial blood pressure (MAP; A), cerebral blood flow (CBF; B), and cerebrovascular resistance (CVR; C) at baseline and in response to increasing doses of intravenous dopamine in 8 near-term (dashed lines) and 8 preterm (solid lines) unanesthetized, chronically catheterized fetal sheep. All measurements were ~30 min apart. *P < 0.05 compared with baseline.

Cerebral O₂ metabolism data are shown in Table 3. There were no changes in Ca_O2 or Pa_O2 in either group nor were there any changes in CMRO₂, cerebral O₂ delivery, or cerebral O₂ extraction.

Plasma catecholamine (norepinephrine and dopamine) levels (pg/ml) are shown in Table 4. Baseline catecholamine levels were similar in preterm and near-term fetuses. Serum dopamine levels increased at each infusion rate in a dose-dependent fashion. At 75 µg · kg¹ · min¹, the dopamine concentration was lower in preterm than in near-term fetuses.

Cardiovascular, cerebrovascular, and oxygenation variables at baseline, in response to -adrenergic or D₁-receptor blockade, and in response to combined dopamine and blocker infusions are shown in Tables 5 and 6. Neither -adrenergic nor D₁-receptor blockade caused any changes in baseline values. Combined with -adrenergic receptor blockade, dopamine infusion at 7.5, 25, and 75 µg · kg¹ · min¹ caused no changes in blood pressure, HR, CBF, CVR, cerebral O₂ delivery, Ca_O2, or CMRO₂. When dopamine was infused during D₁-receptor blockade, the cardiovascular and cerebrovascular responses generally paralleled the responses of the near-term fetuses infused with dopamine alone, except CVR did not increase at 25 µg · kg¹ · min¹ (P = 0.06) and there was no further increase at 75 µg · kg¹ · min¹.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major finding of this study is that exogenous dopamine infusion is associated with cerebral vasoconstriction in hemodynamically stable, unanesthetized preterm and near-term fetal sheep. Overall, the pressor response to dopamine is less in the preterm fetus, but the cerebrovascular response is greater, particularly at lower infusion rates. These responses are blocked in near-term sheep under -adrenergic receptor blockade, while D₁-receptor blockade has no effect.

The most likely explanation for dopamine's cerebral vasoconstrictive effect is that it represents an autoregulatory response to an increase in blood pressure. This is supported by the fact that, in both groups, CVR only increased when blood pressure increased and CBF did not decrease. Furthermore, the cerebral vasoconstrictive effect was blocked by phenoxybenzamine, which also blocked the increase in MAP, thus effectively eliminating the "need" for an autoregulatory response. Studies in sheep have demonstrated that cerebral autoregulation is intact in preterm and near-term fetuses (13, 22), although the autoregulatory range is narrow and the lower limit is near the resting blood pressure. Thus it is not known whether this autoregulatory response could occur in hypotensive or sick animals. However, two recent clinical studies suggest that it may. Both studies measured cerebral artery blood flow velocity during dopamine infusion in sick preterm infants: in one study the infants were hypotensive (36), while in the other study the infants were normotensive (25). In either case, CBF velocity did not change when blood pressure increased.

These fetal cerebrovascular responses to dopamine are different from the vasodilatory responses observed in adult animals (32). One explanation for this unique fetal cerebrovascular response is developmental insensitivity of vascular receptors to dopamine because of immature dopamine receptor development or immature receptor function or both. In general, activation of cerebral -adrenergic and dopaminergic vascular receptors causes cerebral vasodilation, while activation of -adrenergic receptors causes cerebral vasoconstriction (7, 28, 33). Ontogenic changes in the sensitivity of vascular beds to catecholamines have been described in several species, all of which suggest that -adrenergic receptor development occurs first, followed by -adrenergic and, finally, dopaminergic receptors. Wagerle et al. (34) used the closed cranial window technique in developing sheep to demonstrate dose-dependent decreases in pial arteriolar diameter in response to NE. They showed that preterm fetuses are significantly more sensitive to NE than full-term fetuses and newborn lambs and that adult cerebral arterioles do not constrict at all (34). Vasodilatory dopaminergic receptor activity may thus be incompletely developed in newborns compared with -adrenergic receptor activity, supporting our hypothesis that dopamine stimulates vasoconstrictive -adrenergic receptors and has minimal vasodilatory dopaminergic effects in the immature brain. Our blocker studies also support this hypothesis, because although dopaminergic receptor blockade had no effect on the cerebrovascular responses to dopamine, the responses were completely blocked during -adrenergic receptor blockade.

Another explanation for the unique cerebral vasoconstrictive effects of dopamine in fetuses is enhanced release of NE from sympathetic nerves in fetuses (compared with adults) as well as increased cerebral arterial reactivity to NE in fetuses. The latter is supported by data from chronic pial window studies in developing sheep, in which NE caused a dose-dependent arteriolar vasoconstriction mediated by ₁-adrenergic receptors. This response was most significant in preterm fetuses (94-121 days) and was not present in adult sheep (34). In adults and children, ~20% of infused dopamine is converted to NE, and dopamine also triggers release of stored NE from nerve endings (27), although sick preterm infants may not exhibit dose-dependent increases in NE during dopamine infusion (20). We did observe a small increase in NE in our fetuses at increasing dopamine doses. Cerebral vasoconstriction may thus have been related to increased vascular reactivity to NE as well as to increased NE levels.

The cardiovascular effects of dopamine that we observed are consistent, in part, with studies in developing animals suggesting that neonates exhibit relative insensitivity to inotropes, believed to be due to immaturity of cardiac ₁-adrenergic receptor activity (5, 6, 8, 12, 31). In human infants, the cardiovascular effects of dopamine are less consistent, presumably because of the variable physiology of critically ill neonates who require or receive pressors. Among hypotensive (but stable) preterm infants, those who had a positive inotropic response to dopamine were more mature than nonresponders (19). In sick preterm infants, an enhanced pressor response to low-dose (2-4 µg · kg¹ · min¹) dopamine was believed to be due to decreased dopamine clearance (26). In sick, but normotensive, preterm infants, dopamine increased blood pressure and urine output at a dose of 6 µg · kg¹ · min¹ (25). In our preterm fetal sheep, we needed much higher dopamine doses than those used clinically to elicit a pressor response (up to 75 µg · kg¹ · min¹), and only the preterm fetuses had a dose-dependent increase in HR. We did not measure cardiac output; however, myocardial blood flow did not increase in preterm or near-term fetuses, suggesting immaturity of cardiac ₁-adrenergic receptors.

Baseline plasma catecholamine levels were similar to those reported by Palmer et al. (21), with a similarly high degree of variability obscuring gestational age-related differences. However, an interesting finding was that, at increasing dopamine infusion rates, the plasma dopamine and NE concentrations were lower in preterm than in near-term fetuses. In addition, the levels in both groups of fetuses were lower than those reported for human infants receiving exogenous dopamine (2, 20, 27). The most likely explanation for this finding is enhanced placental transfer and clearance of catecholamines in more preterm fetuses. Previous studies in fetal sheep (<105 days of gestation) have demonstrated that the biological activities of NE, epinephrine, and angiotensin II (infused via umbilical artery or vein) are reduced by perfusion through the fetal placenta (16). In near-term fetal sheep, placental catecholamine clearance (via placental amine plasma membrane transporters) is an important metabolic function that maintains fetal homeostasis, despite high fetal catecholamine production rates (4). Furthermore, placental NE transporter binding decreases modestly in fetal sheep between 99 days of gestation and full term (3). If similar placental clearance mechanisms apply to dopamine also, then lower plasma dopamine levels in response to dopamine infusion might be expected in more immature fetuses.

The use of the fetal sheep as a relevant model for neonatal cerebrovascular studies has advantages and limitations. An important advantage for this study is the ability to examine cerebrovascular responses in unanesthetized very preterm animals who are hemodynamically and metabolically stable. Another advantage of the fetal sheep model is that physiological studies can be performed at a critical stage of brain development, i.e., 90 days of gestation, in which there is brain growth and onset of organized electrocortical activity preceding a period of intense neuroglial multiplication and myelination (17). Finally, previous cerebrovascular studies in immature fetal sheep demonstrate that they have lower CBF and CMRO₂ as well as limited autoregulation and blunted responses to hypoxemia compared with near-term fetuses (9, 10, 13). A limitation of this model is the potential difficulty in extrapolating results obtained in fetuses to newborns, particularly the influence of the uteroplacental circulation and fetal physiology on the cerebrovascular and systemic responses we observed. However, in support of our results and the fetal sheep model, Wagerle et al. (35) showed similar contractile responses to NE in fetal and neonatal ovine cerebral arteries that could be modified by nitric oxide and -adrenergic mechanisms. Another potential limitation of our model is the need to estimate fetal weight to calculate drug dosages. However, although we tended to underestimate weight slightly, there were no differences between the groups in our estimation error.

In summary, exogenous dopamine infusion is associated with dose-dependent cerebral vasoconstriction in preterm and near-term fetal sheep that is likely an autoregulatory response. Caution is advised in extrapolating these results to sick human infants, but the studies do provide new data regarding the cerebrovascular effects of dopamine infusion and/or moderate increases in blood pressure in unanesthetized normotensive, normovolemic immature animals.