Reduced blood volume decreases cerebral blood flow in preterm piglets

Abstract

Key points

• Preterm infants often have poor cardiovascular function that is associated with adverse neurodevelopmental outcomes.

• Preterm infants may be vulnerable to hypovolaemia due to excessive vasodilatation and leaky capillaries.

• Following reduction in blood volume, cardiac output and mean arterial pressure were reduced to the same extent in term and preterm piglets.

• Cerebral blood flow was maintained following blood volume reduction in term but not in preterm piglets.

• Effective detection and treatment of functional hypovolaemia may reduce the risk of brain injury in preterm infants.

Abstract

Preterm infants often have impaired cardiovascular function that may contribute to poor neurodevelopmental outcomes. The study aimed to determine the effects of reduced blood volume on cardiovascular function, including cerebral blood flow, in preterm and term piglets. In preterm (97/115 days) and term piglets, up to 10% of the estimated blood volume was removed. Removal of blood was stopped if MAP dropped below 20 mmHg. Heart rate, cardiac contractility and relaxation, cardiac output, mean arterial pressure (MAP), and cerebral blood flow were measured at baseline and again after blood volume reduction. The volume of blood removed was less in preterm piglets than in term piglets (5.1 ± 1.8 vs. 7.7 ± 0.9 mL kg⁻¹, mean ± SD, P < 0.001). Cardiac output and MAP decreased to the same extent in term and preterm piglets. Cerebral blood flow decreased in preterm but not term piglets and cerebral vascular conductance increased in term piglets only. Compensatory responses to maintain cerebral blood flow after blood volume reduction are active in term piglets but not in preterm piglets. As a result, even a small reduction in blood volume, or an increase in the capacity of the circulatory system leading to functional hypovolaemia, may lead to a significant reduction in cerebral blood flow and contribute to brain injury in preterm neonates.

Key points

• Preterm infants often have poor cardiovascular function that is associated with adverse neurodevelopmental outcomes.

• Preterm infants may be vulnerable to hypovolaemia due to excessive vasodilatation and leaky capillaries.

• Following reduction in blood volume, cardiac output and mean arterial pressure were reduced to the same extent in term and preterm piglets.

• Cerebral blood flow was maintained following blood volume reduction in term but not in preterm piglets.

• Effective detection and treatment of functional hypovolaemia may reduce the risk of brain injury in preterm infants.

Introduction

Preterm infants are at high risk of cardiovascular compromise during the first day of life (Evans et al. 2005; Batton et al. 2016). Preterm infants also have increased rates of cerebral palsy, motor disability and cognitive impairment, and there is an association between cardiovascular compromise and adverse brain outcomes (Osborn et al. 2007b; Batton et al. 2009, 2016). More effective cardiovascular support may reduce brain injury in preterm infants.

A contributing factor to poor cardiovascular function in preterm infants may be low blood volume. In preterm infants, a reduction in blood volume may be common in the early hours after birth due to several factors. As the pulmonary circulation opens with expansion of the lungs, the capacity of the vascular compartment is increased. High levels of circulating catecholamines at birth (Faxelius et al. 1984) may assist the term neonate to successfully adapt by increasing systemic vasculature resistance but the preterm neonate is less responsive to catecholamines (Eiby et al. 2014, 2016) and has a reduced ability to vasoconstrict (Fletcher et al. 2006; Wassink et al. 2007). In fact, some very preterm infants may also be inappropriately and excessively vasodilated for the extrauterine environment (Stark et al. 2008) leading to relative hypovolaemia, even if the absolute circulating volume is maintained. In addition to these vascular factors, early loss of fluid from the vascular compartment due to excess leakage from capillaries results in significant reductions in blood volume. The rapid loss of labelled albumin from the circulation coupled with early increases in haemoglobin concentration ([Hb]) suggest whole plasma is lost from the circulation of the preterm infant (Clark & Gairdner, 1960; Steele, 1962).

We hypothesized that reduced blood volume would reduce cardiac preload and thus lower cardiac contractility and cardiac output via the Starling relationship (Patterson & Starling, 1914). In the absence of effective compensatory mechanisms, this will lead to a reduction in blood pressure. Cerebral blood flow would normally be protected via cerebral autoregulation, but in the preterm infant autoregulation may not be fully functional (Fyfe et al. 2014) and cerebral blood flow may be compromised. Thus, hypovolaemia may contribute to reduced cerebral blood flow and brain injury in preterm infants. The beneficial effects of delayed cord clamping and placental transfusion support this hypothesis that hypovolaemia may contribute to preterm brain injury (Backes et al. 2014; Tarnow‐Mordi et al. 2017; Fogarty et al. 2018).

Given the propensity for preterm infants to be hypovolaemic, it is critical to understand how the preterm cardiovascular system responds to low blood volume. The aim of this study was to describe the effects of a reduction in blood volume in preterm piglets on cardiovascular function, including cardiac output, blood pressure and cerebral blood flow, and to compare the effects with those in term piglets.

Methods

Ethical approval

The care and handling of the animals was in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (8th edition, 2013) and was approved by The University of Queensland Animal Ethics Committee (AEC Approval No.: UQCCR/060/12). The investigators understand the ethical principles under which The Journal of Physiology operates and this work complies with the animal ethics checklist.

Animals

The piglets included in this study are a subset of the animals in which we described the response to dopamine and dobutamine under baseline conditions (Eiby et al. 2016). Piglets were delivered by Caesarean section at either 97 days (preterm; developmentally similar to a human neonate at 27 weeks’ gestation) or 113 days (term) where full term is 115 days (Eiby et al. 2013). All preterm piglets were exposed to maternal glucocorticoid treatment (intramuscular betamethasone sodium phosphate 0.13 mg kg⁻¹ and betamethasone acetate 0.1 mg kg⁻¹; Celestone Chronodose, Schering‐Plough, Kenilworth, NJ, USA) at 48 and 24 h prior to delivery. The timing and per kg dose were in line with standard clinical practice for women with threatened preterm labour. Piglets with birthweight >10th and <90th centile were studied. Piglets were resuscitated and maintained under standard neonatal intensive care conditions as previously described (Eiby et al. 2016).

A loading dose of 5 mg kg⁻¹ of propofol (Provive MCT‐LCT 1%; AFT Pharmaceuticals, Auckland, New Zealand) was given at birth. Sedation in term piglets commenced at 20 mg kg⁻¹ h⁻¹ of propofol and 0.11 mg kg⁻¹ h⁻¹ of fentanyl (Sublimaze; Janssen‐Cilag, Macquarie Park, NSW, Australia) and in preterm piglets commenced at 10 mg kg⁻¹ h⁻¹ of propofol and 0.06 mg kg⁻¹ h⁻¹ of fentanyl. Sedation was reduced after initial set‐up procedures and during the experiment was 5.6 ± 1.6 mg kg⁻¹ h⁻¹ of propofol and 0.031 ± 0.009 mg kg⁻¹ h⁻¹ of fentanyl in term piglets and 5.4 ± 1.4 mg kg⁻¹ h⁻¹ of propofol and 0.030 ± 0.008 mg kg⁻¹ h⁻¹ of fentanyl in preterm piglets. Piglets were intubated and surfactant administered to preterm piglets (2.5 mL endotracheal; Survanta; Abbott Australasia, Macquarie Park, NSW, Australia).

Ventilation adequacy was monitored using oxygen saturation ($S_{\mathrm{a}{\mathrm{O}}_2}$ ; measured using pulse oximetry, Masimo, Irvine, CA, USA) and intermittent arterial blood gas analysis (ABL800Flex Blood Gas Analyser; Radiometer, Brønshøj, Denmark). Fractional inspired oxygen $(F_{\mathrm{I}{\mathrm{O}}_2})$ and peak inspiratory pressure (PIP) were adjusted to maintain $P_{\mathrm{a}{\mathrm{O}}_2}$ 80–120 mmHg and $P_{\mathrm{aC}{\mathrm{O}}_2}$ 35–45 mmHg. Body temperature was maintained at 39.0 ± 0.5°C (normothermia for pigs).

All piglets were immediately euthanized following experimentation (pentobarbitone sodium 162 mg kg⁻¹ i.v.).

Monitoring

Physiological parameters including heart rate, cardiac contractility (dP/dt _max) and relaxation (dP/dt _min) and mean arterial blood pressure (MAP) were recorded continuously throughout the experiment using a 16‐channel Power Lab and LabChart 7 (ADInstruments Pty Ltd, Bella Vista, NSW, Australia) at a sampling frequency of 1 kHz. Arterial blood pressure was monitored using a pressure transducer (ICU Medical Inc., San Clemente, CA, USA) connected to an umbilical artery catheter (UAC). Heart rate was calculated from the arterial blood pressure wave. To measure cardiac contractility and relaxation, a solid‐state pressure catheter (Transonic SciSense, London, Ontario, Canada) was inserted into the left ventricle via the left carotid artery. Contractility (dP/dt _max) and relaxation (dP/dt _min) were calculated using LabChart.

Cardiac output and organ blood flows (cerebral hemispheres, gluteal muscle, abdominal skin, small intestine, large intestine) were measured using coloured microspheres (Dye‐Trak 15 μm microspheres, Triton Technologies, San Diego, CA, USA) injected into the left ventricle via a ventricular catheter (McDevitt & Nies, 1976; Walter et al. 1997). A reference blood sample was collected from the descending aorta in accordance with the validated method (Makowski et al. 1968). Microspheres were extracted from the tissue using the filtration method. Organ blood flows were calculated by comparison of the number of microspheres in the tissue with that in the reference blood sample. Cardiac output was calculated using the formula:

Cardiac output = total number of microspheres injected × reference sample withdrawal rate/number of microspheres in reference sample (McDevitt & Nies, 1976).

Systemic vascular resistance, a primary determinant of blood pressure, was calculated from the MAP/cardiac output. Vascular control at an organ level was assessed using conductance calculated from the blood flow/MAP. Cerebral oxygen delivery (c$D_{{\mathrm{O}}_2}$; mL O₂ (100 g)⁻¹ min⁻¹) = (1.39 × [Hb] (g mL⁻¹) × (%$S_{\mathrm{a}{\mathrm{O}}_2}$ /100)) × cerebral blood flow (mL (100 g)⁻¹ min⁻¹).

Experimental design

Baseline values for continuous measurements were calculated from the average over a 5 min period immediately before the first blood flow measurement. Approximately 8 h after birth and approximately 1.5 h after baseline measurements, blood volume was reduced by continuous withdrawal of approximately 10% of the total blood volume via the UAC over 10–15 min (based on 80 mL kg⁻¹; Strauss et al. 2003). MAP was checked when 2.5%, 5% and 7.5% of estimated blood volume had been withdrawn. If MAP fell below 20 mmHg, removal of blood was stopped as in our experience piglets are unlikely to survive for the duration of the experiment with MAP below this threshold. All measurements were repeated approximately 25 min after blood volume removal, when MAP had been stable for at least 5 min.

Following the baseline period and prior to blood volume reduction, some animals received inotropic support as part of a separate study (Eiby et al. 2016). A total of 19 preterm and 24 term piglets were studied of which seven preterm and 10 term piglets received dopamine, and six preterm and eight term piglets received dobutamine, with similar sex ratios across these drug groups. The half‐life of the inotropes used is 2 min and this treatment was stopped and parameters allowed to return to baseline before the blood volume was reduced approximately 40 min later. Heart rate, contractility, relaxation and blood pressures were similar across drug treatment groups prior to blood volume reduction (P > 0.1 for all). Changes in parameters in response to blood volume removal were similar across these drug treatment groups (P > 0.1 for all parameters) except for heart rate, contractility and cardiac output in term piglets, which all decreased more in the drug‐treated group compared to the control group (P < 0.05).

Data processing and analysis

Data were analysed using SPSS version 22 (IBM Corp., Armonk, NY, USA). Data were normally distributed except for blood flows, organ conductances, systemic vascular resistance and oxygen delivery, which were log‐transformed prior to analysis in order to achieve normality. Differences in body weight and the volume of blood removed between preterm and term piglets were sought using an independent sample Student's t test. The effect of reduced blood volume was examined using repeat measures two‐way ANOVA with volume and gestational age as fixed factors, along with their interaction. Differences in physiology between drug treatment groups prior to the bleed were sought using ANOVA within each gestational age. Differences in parameters between drug treatment groups in response to blood volume removal were sought using repeat measures two‐way ANOVA with volume and drug group as fixed factors, along with their interaction. A P value < 0.05 was considered statistically significant.

Results

Demographics and baseline function

Preterm piglets had lower body weights (P < 0.001) but similar sex ratios compared to term piglets (P = 1.00, Table 1). Preterm piglets had significantly lower baseline cardiac contractility and relaxation, cardiac output, mean arterial pressure, cerebral blood flow (per 100 g) and skin and muscle blood flows (per 100 g) compared with term piglets (all P < 0.05, Table 1 and Fig. 1).

Table 1.

Cardiovascular function during baseline and reduced blood volume conditions in preterm and term piglets

Parameters	Preterm (n = 19a)		Term (n = 24a)		P
Male (%)	58		58
Body weight (g)	1032 ± 128#		1470 ± 191		<0.001
Blood volume	Baseline	Reduced	Baseline	Reduced	GA	Volume	Interaction
Heart rate (bpm)	146 ± 19	145 ± 30	131 ± 11	123 ± 17	0.001	0.23	0.32
dP/dt max a (mmHg s−1)	1184 ± 344	714 ± 292	1783 ± 315	1257 ± 401	<0.001	<0.001	0.64
dP/dt min a (mmHg s−1)	−1292 ± 425	−833 ± 379	−1935 ± 760	−1395 ± 623	0.003	<0.001	0.58
Systolic BP (mmHg)	44.9 ± 11.1	30.5 ± 12.1	59.7 ± 7.1	45.9 ± 11.3	<0.001	<0.001	0.83
Diastolic BP (mmHg)	22.4 ± 6.9	14.6 ± 6.6	28.5 ± 4.7	22.0 ± 7.4	0.001	<0.001	0.40
Skin BFa (mL min−1 (100 g)−1)	7.5 (6.5–10.0)	5.4 (4.0–8.0)	11.5 (8.6–17.1)	7.9 (6.2–12.3)	0.009	0.002	0.70
Skin VCa (mL min−1 (100 g)−1 mmHg−1)	0.30 (0.22–0.38)	0.39 (0.18–0.64)	0.29 (0.23–0.48)	0.32 (0.23–0.43)	0.79	0.29	0.23
Muscle BFa (mL min−1 (100 g)−1)	2.3 (1.9–3.0)	2.5 (1.9–4.5)	5.9 (4.0–9.4)	4.1 (3.1–5.0)	<0.001	0.40	0.006
Muscle VCa (mL min−1 (100 g)−1 mmHg−1)	0.09 (0.06–0.13)	0.19 (0.09–0.24)	0.15 (0.11–0.21)	0.15 (0.10–0.20)	0.26	0.003	0.004
Intestinal BFa (mL min−1 (100 g)−1)	50.5 (40.1–55.6)	24.4 (13.5–33.0)	44.0 (36.7–61.1)	24.7 (20.4–36.4)	0.76	<0.001	0.48
Intestinal VCa (mL min−1 (100 g)−1 mmHg−1)	1.91 (1.20–2.42)	1.34 (1.05–2.17)	1.27 (0.86–1.62)	0.89 (0.67–1.47)	0.99	0.007	0.99
SVRa (mmHg mL−1 min−1)	0.197 (0.154–0.246)	0.214 (0.170–0.248)	0.139 (0.114–0.213)	0.212 (0.126–0.273)	0.12	0.46	0.25
c$D_{{\mathrm{O}}_2}$ a (mL O2 (100 g)−1 min−1)	2.1 (0.9–2.6)	1.5 (0.1–2.5)	2.4 (2.0–3.2)	2.9 (1.6–4.3)	0.006	0.058	0.045

Values are mean ± SD or median (IQR). ^aIncomplete data due to poor quality blood pressure wave used for analysis or technical failure in microsphere measurements such that n is 12–15 preterm and 17–20 term piglets. BF, blood flow; BP, blood pressure; c$D_{{\mathrm{O}}_2}$, cerebral oxygen delivery; dP/dt _max, cardiac contractility; dP/dt _min, cardiac relaxation; GA, gestational age; SVR, systemic vascular resistance; VC, vascular conductance.

Figure 1. Effect of reduced blood volume on cardiovascular function in term and preterm piglets.

Cardiac output, mean arterial pressure (MAP), cerebral blood flow and cerebral vascular conductance in preterm and term piglets during baseline and reduced blood volume conditions. ^*indicates a significant effect of volume (blood volume), gestation (gestational age) and volume x gestation (the interaction between blood volume and gestational age).

Reduced blood volume

Measures of cardiovascular function for preterm and term piglets before and after blood volume reduction are shown in Table 1 and Fig. 1. The volume of blood removed was significantly less in preterm piglets than in term piglets (5.2 ± 1.8 vs. 7.7 ± 0.9 mL kg⁻¹, mean ± SD, P < 0.001) as removal of blood was frequently terminated prematurely in preterm piglets due to MAP <20 mmHg. The mean blood volume removed was approximately 6.5% of the estimated blood volume in preterm piglets and 9.6% in term piglets. When blood volume was reduced, heart rate was unchanged in either group (P = 0.226). Cardiac contractility, relaxation, cardiac output and blood pressure were reduced to the same extent in both groups after blood volume reduction (P < 0.05 for all).

The change in cerebral blood flow following blood volume reduction was significantly different in term and preterm piglets (P = 0.023) such that term piglets maintained cerebral blood flow during reduced blood volume but in preterm piglets cerebral blood flow was reduced by an average of 42%. Half the preterm piglets had cerebral blood flow after blood removal less than 30% of baseline values. The change in cerebral vascular conductance was also significantly different in term and preterm piglets (P = 0.049) such that conductance increased in term piglets but not in preterm piglets.

Systemic vascular resistance did not differ between term and preterm piglets and was not altered following reduced blood volume. Skin blood flow was lower in preterm piglets compared to term piglets and both groups decreased skin blood flow to the same extent following blood volume removal (Table 1). No change in skin vascular conductance was detected. Muscle blood flow was also lower in preterm piglets compared to term piglets but only decreased in the term group following blood volume removal. Muscle vascular conductance increased in the preterm piglets but not in the term piglets. Intestinal blood flows decreased to the same extent in both groups following blood volume removal.

$P_{\mathrm{aC}{\mathrm{O}}_2}$ was similar between term and preterm piglets at baseline (40 ± 19 vs. 38 ± 10 mmHg, P = 0.93) and did not change across the experiment (44 ± 10 vs. 37 ± 12 mmHg during reduced blood volume, P = 0.28 for baseline vs. reduced volume). The rate of anaesthetic drug administration was unchanged during the experiment (P = 0.55) and was similar between the term and preterm piglets (P = 0.86). Arterial haemoglobin levels decreased in both groups following blood removal (P = 0.001) but the decrease was greater in term piglets (from 77 ± 6 to 66 ± 8 g L⁻¹) compared to preterm piglets (from 84 ± 13 to 80 ± 15 g L⁻¹) (P = 0.049). Cerebral oxygen delivery (c$D_{{\mathrm{O}}_2}$) was significantly lower in the preterm piglets compared to term piglets (P = 0.006); a significant interaction indicates that c$D_{{\mathrm{O}}_2}$ decreased following blood volume in the preterm group but not the term group (P = 0.045; Table 1).

Discussion

We have previously reported that preterm piglets have lower cardiac outputs and MAPs than term piglets (Eiby et al. 2016). This study shows that, in preterm piglets, blood volume removal leads to further reductions in cardiac output, MAP and cerebral blood flow. In term piglets, removal of blood also decreases cardiac output and MAP but, in contrast to preterm piglets, cerebral blood flow is maintained. As often seen clinically, there was no change in heart rate in response to hypovolaemia. The volume of blood removed from preterm piglets before hypotension developed was approximately 6.5% of total blood volume (compared with 9.6% in term piglets), suggesting that even small reductions in blood volume may lead to reduced cerebral blood flow in preterm neonates.

The reduction in contractility seen following blood removal in term piglets is consistent with predictions from Starling's Law – decreased filling results in decreased stretch and decreased contractility (Patterson & Starling, 1914). In contrast, contractility does not fall in preterm piglets. However, contractility in preterm piglets is already significantly lower than in term piglets at baseline, likely due to a combination of immature heart structure/organization and pre‐existing hypovolaemia. Perhaps the preterm heart is already functioning at the low end of the Starling curve with little further reduction possible. Low contractility was associated with reduced cardiac output and arterial blood pressure in both term and preterm piglets. These variables were reduced to a similar extent in both groups and are likely to impact organ blood flows including the brain.

There is no evidence of systemic vascular compensation for hypovolaemia in either preterm or term piglets as there was no increase in systemic vascular resistance or decrease in skin, muscle or intestinal vascular conductance. Therefore, the observed decreases in peripheral blood flows are probably secondary to decreased cardiac output rather than a result of vasoconstriction. This lack of vascular compensation may be due to poor peripheral vascular control even in term piglets (Fletcher et al. 2006; Wassink et al. 2007; Stark et al. 2008).

In contrast to the lack of compensatory changes in systemic resistance in response to removal of blood, there do appear to be protective changes in the cerebral circulation, although only in term piglets. In term piglets, the reduction in blood volume was associated with an increase in cerebral vascular conductance and thus no reduction in cerebral blood flow. These differences were not due to different administration of fentanyl and propofol as these were the same in preterm and term piglets. It is possible that the anaesthetic may have had a different effect at different ages. Likewise, there was no difference in $P_{\mathrm{aC}{\mathrm{O}}_2}$ between term and preterm animals or between baseline and post‐bleed time points suggesting that differing $P_{\mathrm{aC}{\mathrm{O}}_2}$ levels do not contribute to these observations. Results suggest that in term piglets autoregulation is effective and piglets are able to compensate for a reduction in cardiac output and cerebral perfusion pressure by increasing cerebral vascular conductance to maintain cerebral blood flow. This compensation did not occur in preterm piglets. Numerous studies have demonstrated that preterm infants have impaired autoregulation and fluctuating pressure‐passivity of the cerebral vasculature (Fyfe et al. 2014). Impaired autoregulation may contribute to the lack of compensation in preterm piglets. It has been also suggested that many preterm infants operate close to the low end of the autoregulatory plateau. In this case, even small reductions in blood pressure would result in a blood pressure outside the range of cerebral autoregulation and hence reduced cerebral blood flow (Fyfe et al. 2014).

The decrease in cerebral blood flow following blood volume reduction in the preterm piglets resulted in a decrease in cerebral oxygen delivery. There was no decrease in cerebral oxygen delivery in the term piglets where cerebral blood flow was maintained. We did not measure cerebral oxygen metabolism or extraction, or brain injury so we are not able to assess if the reduction in cerebral blood flow and cerebral oxygen delivery had detrimental effects on the brain. However, the repeated demonstration of a relationship between reduced superior vena cava flow or impaired autoregulation and adverse outcomes (Hunt et al. 2004; Osborn et al. 2007a; Fyfe et al. 2014) suggests that the reduced cerebral blood flow occurring in the presence of hypovolaemia may be a significant contributor to preterm brain injury. In preterm piglets, baseline cerebral blood flow was lower than in term piglets. This is comparable to the human where very preterm infants have resting cerebral blood flows that are approximately half those measured in term infants and one‐quarter of adult values (Greisen & Borch, 2001; Munro et al. 2004). This could reflect a lower cerebral oxygen requirement, but since cerebral oxygen extraction is increased and cerebral venous oxygen saturation is lower, it may be that in some preterm infants, cerebral blood flow and oxygen delivery are below those required to avoid cerebral tissue hypoxia (Greisen & Borch, 2001; Andersen & Stark, 2012; Balegar et al. 2014). Reduction in blood volume would exacerbate this possibility because of the limited ability of the preterm cerebral vasculature to compensate.

The most common method of volume expansion in preterm infants is administration of normal saline, but this is often ineffective in increasing blood pressure (Osborn & Evans, 2004), leading to the conclusion that hypovolaemia is not present. However, saline rapidly leaks from the circulation in adults, and probably more rapidly in neonates (Clark & Gairdner, 1960; Steele, 1962). In addition, there is little correlation between blood volume and blood pressure in the neonate (Wright & Goodall, 1994), and so normotension may not equate to normovolaemia. Since our study suggests hypovolaemia results in low cerebral blood flow in many individuals, it is essential to develop improved methods to detect and treat this condition in order to prevent preterm brain injury.

In the clinical setting, if volume expansion by i.v. saline is not effective in raising blood pressure, the next approach is to provide inotropic medication. This is often ineffective in elevating blood pressure (Osborn & Evans, 2004), possibly because the primary problem is undetected hypovolaemia. Inotropes do raise the plateau of the Starling curve, but if the right atrial pressure is low due to hypovolaemia, inotropes have much less effect when operating on lower parts of the curve (Sarnoff, 1955; Guyton et al. 1973). Thus, in the presence of hypovolaemia, inotropes will have limited capacity to increase contractility.

This study was part of a larger study and so we acknowledge several limitations. We were not able to measure cardiac output and organ blood flows immediately before removal of blood as this would have required an additional colour of microspheres, which was not available to us. Also, we were not able to directly quantify the fraction of blood removed as we did not have direct measures of total blood volume. In calculating the volume of blood to be removed we did not use a higher value for the blood volume in preterm infants because the literature is not consistent on whether a difference exists or not (Linderkamp et al. 1980; Aladangady et al. 2006, 2008). This would, however, not have affected the total volume removed as blood removal in preterm piglets was usually terminated prematurely based on MAP. If circulating blood volume per kg body weight is higher in the preterm, the fraction removed would have been lower than reported, further underlining the extreme sensitivity of preterm piglets to a reduction in blood volume.

There was a significant delay between our baseline measurements and removal of blood and we cannot exclude the possibility that some of the differences observed are due to intervening events. However, the values for continuous measurement where data were available immediately before blood removal (blood pressures and contractility) were significantly higher than values after blood removal suggesting that the changes reported here are not due to intervening events. Intervening events included inotrope administration in some piglets. Heart rate, contractility, relaxation and blood pressures were similar across drug treatment groups prior to blood volume reduction (P > 0.1 for all). Changes in parameters in response to blood volume removal were similar across these drug treatment groups (P > 0.1 for all parameters) except for heart rate, contractility and cardiac output in term piglets, which all decreased more in the drug‐treated groups compared to the control group (P < 0.05). This suggests that the inotropes may still have had some cardiac effects in these animals; however, if the drug‐treated groups had responded like control animals this would have meant the drop in cardiac output with reduced blood volume would have been less in the term animals, magnifying the cardiac output differences observed between preterm and term groups.

Conclusions

Preterm piglets do not maintain cerebral blood flow or cerebral oxygen delivery when blood volume is reduced. Our findings suggest that many preterm human infants with low blood volume may have very low cerebral blood flow, potentially contributing to the high rate of adverse neurodevelopmental outcomes observed in this population. Better and clinically practical methods to detect and treat hypovolaemia may reduce preterm brain injury.

Additional information

Competing interests

The authors have no competing interests to declare.

Author contributions

Experiments were performed at Herston Medical Research Centre and UQ Centre for Clinical Research. Conception and design: Y.E., I.W., E.L., P.C., G.D., B.L. Analysis and interpretation of data: Y.E., N.S., E.L., P.C., B.L. Drafting or revising the manuscript: Y.E., N.S., E.L., I.W., P.C., G.D., B.L. All authors have read and approved the final version of this manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This project was funded by an Australian government National Health and Medical Research Council project grant (GNT1028888) and a Royal Brisbane and Women's Hospital Foundation project grant (2014002740).

Translational perspective.

Preterm infants often have impaired cardiovascular function. This may result from hypovolaemia due to excessive vasodilatation and leaky capillaries. Impaired cardiovascular function may contribute to the high rate of neurodevelopmental disability in preterm infants. This study tested the hypothesis that reduced blood volume leads to impaired cardiovascular function and reduced cerebral blood flow in preterm piglets. We found that cerebral blood flow was maintained following reduction in blood volume in term but not in preterm piglets. The inability to maintain cerebral blood flow during hypovolaemia may contribute to the high disability rate in children born preterm. Hypovolaemia is challenging to detect in preterm neonates and methods of measuring blood volume are currently not clinically feasible. More effective methods to detect and treat hypovolaemia in preterm neonates may prevent low cerebral blood flow and reduce the risk of brain injury.

Biography

Yvonne A. Eiby is a Research Fellow in the Systems Physiology Research Group at the Perinatal Research Centre, UQCCR. She used a pre‐clinical piglet model of the very preterm infant to investigate cardiovascular factors influencing outcomes in preterm infants with a focus on brain development. She has studied preterm cardiovascular structure and function with the aim of better understanding preterm cardiovascular compromise and the reasons for ineffectiveness of current treatments. Her overall objective is to improve cardiovascular support in preterm infants and reduce adverse long‐term outcomes.