Temporal effects of prolonged hypoxaemia and reoxygenation on systemic, pulmonary and mesenteric perfusion in newborn piglets

Abstract

Objective: Temporal effects of prolonged hypoxaemia and reoxygenation, on the systemic pulmonary and mesenteric circulations in newborn piglets, were investigated. Methods: Two groups [control (n=5), hypoxaemic (n=7)] of 1–3 day old anaesthetised piglets were instrumented with ultrasound flow probes placed to measure cardiac, hepatic arterial flow and portal venous flow indices, and catheters inserted for measurements of systemic and pulmonary arterial pressures. Hypoxaemia with arterial oxygen saturation 40–50% was maintained for 3 h, followed by reoxygenation with 100% inspired oxygen. Results: Cardiac index was transiently elevated at 30–60 min of hypoxaemia (23% increase from baseline 158±39 ml/kg/min), along with increases in stroke volume but not heart rate. A significant decrease in systemic vascular resistance after 30 min of hypoxaemia was followed by hypotension at 180 min of hypoxaemia. Progressive pulmonary hypertension with significant vasoconstriction was found after 30 min of hypoxaemia. The hypoxaemic mesenteric vasoconstriction was transient with a 37% decrease in portal venous flow index at 15 min of hypoxaemia (29±12 vs. 46±18 ml/kg/min of baseline, p<0.05). The hepatic arterial to total hepatic oxygen delivery ratio increased significantly during hypoxaemia. In contrast to the significant increase in systemic oxygen extraction throughout hypoxaemia, elevation in mesenteric oxygen extraction decreased after 30 min of hypoxaemia associated with modest decreases in oxygen consumption. Following reoxygenation, the pulmonary hypertension was partially reversed. Cardiac index decreased further (130±39 ml/kg/min) with reduced stroke volume, persistent systemic hypotension and decreased systemic oxygen delivery. Conclusions: We demonstrated differential temporal changes in systemic, pulmonary and mesenteric circulatory responses during prolonged hypoxaemia. Cautions need to be taken upon reoxygenation because the neonates are at risk of developing myocardial stunning, persistent pulmonary hypertension and necrotising enterocolitis.

Time for primary review 8 days

1 Introduction

Acute hypoxaemia causes differential responses in regional and systemic circulations. The pulmonary vasculature constricts [1, 2], whereas the reported hypoxaemic systemic responses have been variable. These variations may be because of the degree of hypoxaemia used: moderate hypoxaemia usually causes systemic vasodilatation whereas severe hypoxaemia has been shown to cause vasoconstriction in some systemic arteries [3]. The discrepancies may also be related to differences in vessel size and species being studied [4–7], and the use of anaesthetics [2, 8]. Leach et al. recently reported responses to hypoxaemia in their in vitro preparation of pulmonary and mesenteric arteries derived from adult rats [9]. They demonstrated a persistent pulmonary vasoconstriction and a biphasic mesenteric response with initial vasoconstriction and subsequent vasodilatation. We are unaware of any data concerning the temporal effects of prolonged hypoxaemia on regional haemodynamic changes in the newborn.

The regional oxygen metabolism responses in relation to the prolonged hypoxaemic haemodynamic changes in the newborn have also not been studied. In particular, information on mesenteric haemodynamic and oxygen metabolic responses would be important for understanding the pathogenesis of necrotising enterocolitis, which has been related to hypoxic injury [10]. Furthermore, because of the unique circulation of the liver, the effect of prolonged hypoxaemia on hepatic oxygenation is different from other circulations.

We therefore designed the following experiment to compare the temporal responses of systemic, pulmonary and mesenteric vasculatures and their respective oxygen metabolism to 3 h of hypoxaemia in a newborn animal model. Animals were then reoxygenated with 100% inspired oxygen. We hypothesised that there would be differential temporal changes in systemic and regional circulations in response to systemic hypoxaemia and these changes would be reversed by reoxygenation.

2 Methods

The study conformed to the regulations of the Canadian Council of Animal Care (revised 1993) and was approved by the Health Sciences Animal Welfare Committee, University of Alberta.

Newborn piglets of mixed western breed were obtained, 1–3 days of age, weighing 1.4–2.4 kg (mean 1.84 kg). Anaesthesia was induced with inhaled halothane 5% and then decreased to 2%. A double lumen external jugular catheter was positioned at the right atrium and a common carotid arterial line was inserted. Following tracheotomy and commencement of assisted ventilation, halothane was discontinued after a maximum of 20 min. Subsequently, anaesthesia was maintained by fentanyl infusion (4 μg/kg/h) and the piglets were paralysed with 0.1 mg/kg doses of pancuronium as needed. Two doses of fentanyl (10 μg/kg) and acepromazine (0.1 mg/kg) were given prior to thoracotomy and laparotomy. Dextrose–saline solution was infused at a rate of 15–20 ml/kg/h. Piglets were ventilated at pressures of 16/4 cm H₂O at a rate of 12 to 18 breaths per min with inspired oxygen concentration of 21–30%. A left thoracotomy was then performed in the fourth intercostal space. The pericardium was opened and a 20-gauge catheter was inserted into the root of the pulmonary artery for the measurement of pulmonary artery pressure. A six-millimetre transit time ultrasound flow probe (Transonic, Ithaca, NY) was placed around the main pulmonary artery to measure cardiac output. A midline laparotomy was then performed. A 5-Fr Argyle catheter was inserted through the umbilical vein into the portal venous system. A 1-mm and a 2-mm Transonic transit time ultrasound flow probe were placed around the common hepatic artery and the portal vein. The neck incision, thoracotomy and laparotomy were closed with sutures after these procedures finished. Blood gases were drawn and 15 min of recording were done to ensure the animal was stable. Stability was defined as (1) heart rate and blood pressure within 10% of the postanaesthetic prethoracotomy values, (2) right atrial pressure of 3–8 mm Hg, (3) arterial PaO₂ 75–100 mm Hg, PaCO₂ between 35 and 55 mm Hg and pH between 7.35 and 7.45. The surgical procedure usually finished within 75 min. The rectal temperature was maintained between 38.0–38.5°C by an electrical heating blanket and an infrared-heating lamp.

After a baseline monitoring period of at least 15 min, simultaneous blood samples were drawn for arterial, mixed venous and portal venous oxygen saturation and haemoglobin concentration determinations by co-oximeter (OSM2 Hemoximeter, Radiometer, Copenhagen). The following haemodynamic variables were monitored continuously throughout the study period: mean arterial blood pressure (SAP), mean pulmonary arterial pressure (PAP), right atrial pressure (RAP), heart rate, pulse oximetry (Nellcor, Hayward, CA), cardiac output, portal venous flow and hepatic arterial flow. Analogue outputs of the pressure amplifiers and flow monitors were digitised by a DT 2801-A analogue to digital converter board (Data Translation, ON) in a Dell 425E personal computer. Software was custom written using the Asyst programming environment. All signals were continuously acquired at 24 Hz and saved on hard disk. These variables were monitored continuously for 4 h after stabilisation in the 5 normoxaemic controls. For the 7 piglets in the hypoxaemic group, hypoxaemia was induced by reducing FiO₂ to between 0.10 and 0.15, in order to maintain arterial oxygen saturation between 40–50% (PaO₂ between 30–40 mm Hg) after stabilisation. The haemodynamic variables at 15, 30, 60, 120 and 180 min of hypoxaemia were averaged over 5 min for analysis. Cardiac index (CI), portal venous flow index (PVFI) and hepatic arterial flow index (HAFI) were calculated by dividing the nonindexed variables by body weights. At 15, 30, 60, 120 and 180 min of hypoxaemia, simultaneous blood samples were taken from common carotid artery, pulmonary artery and portal venous catheters. Oxygen saturations of arterial (SaO₂), mixed venous (SvO₂) and portal venous (SpO₂) blood were measured. Arterial blood gases and plasma lactate were determined at 180 min of hypoxaemia. No bicarbonate solution was given during the experiment. All measurements were obtained and analysed in all animals.

After 180 min of hypoxaemia, the inspired gas was switched to a FiO₂ of 1.0 and the animal was monitored for a further 30 min after which blood sampling was repeated. The animals were then euthanised with an intravenous pentobarbital overdose.

We calculated the following variables at each phase of hypoxaemia and reoxygenation:

1. 1.

   Stroke volume (SV)=CI÷heart rate

2. 2.

   Total hepatic flow index (THFI)=PVFI+HAFI

3. 3.

   Systemic vascular resistance index (SVRI)=(SAP−RAP)÷CI

4. 4.

   Pulmonary vascular resistance index (PVRI)²=PAP÷CI

5. 5.

   Systemic oxygen extraction (sysEO₂)=(SaO₂−SvO₂)÷SaO₂×100%

6. 6.

   Systemic oxygen delivery (sysDO₂)=CI×SaO₂×1.34×[Hb]

7. 7.

   Systemic oxygen consumption (sysVO₂)=CI×(SaO₂−SvO₂)×1.34×[Hb]

8. 8.

   Mesenteric oxygen extraction (mesEO₂)=(SaO₂−SpO₂)÷SaO₂×100%

9. 9.

   Mesenteric oxygen delivery (mesDO₂)=PVFI×SaO₂×1.34×[Hb]

10. 10.

    Mesenteric oxygen consumption (mesVO₂)=PVFI×(SaO₂−SpO₂)×1.34×[Hb]

11. 11.

    Total hepatic oxygen delivery (HDO₂)=(HAFI×SaO₂+PVFI×SpO₂)×1.34×[Hb]

12. 12.

    Ratio of hepatic arterial oxygen delivery to total hepatic oxygen delivery (HADO₂ ratio)=HAFI×SaO₂÷(HAFI×SaO₂+PVFI×SpO₂)×100%

In a separate series of experiment, 5 piglets were instrumented as above and induced with the same degree of hypoxaemia for 30 min followed by reoxygenation with 100% inspired oxygen. Changes in systemic and mesenteric haemodynamics and oxygen metabolism were studied.

3 Statistical analysis

One-way repeated measures analysis of variance and repeated measures analysis of variance on ranks were used to analyse the differences at various phases of hypoxaemia and reoxygenation for parametric and nonparametric variables, respectively (Sigma Stat 1.01 version, Jandel Scientific, San Rafael, CA). Dunnett's post-hoc test was used to compare the difference with baseline. A p value of <0.05 was considered as significant. The results are expressed as mean±standard deviation.

4 Results

4.1 Normoxaemic control (n=5)

The control animals had no significant change in any of the recorded haemodynamic and oxygen variables over 4 h of the study.

4.2 Hypoxaemic group (n=7)

4.2.1 Systemic haemodynamic changes during hypoxaemia (Table 1, Figs. 1 and 2)

By 180 min of hypoxaemia with systemic oxygen saturations being kept between 40–50%, all piglets were acidotic with pH 7.06–7.27 (mean 7.17±0.09), serum bicarbonate levels 14–23 (mean 19±3.8) mmol/l and plasma lactate levels 7.8–18.9 (mean 11.6±3.9) mmol/l. CI was significantly increased at 30–120 min of hypoxaemia, and then started to decline afterwards, being not significantly different to baseline at 180 min. Maximum increase in CI, along with SV, was at 30 min of hypoxaemia (CI: 194±47 vs. 158±39 ml/kg/min at baseline, SV: 0.57±0.19 vs. 0.48±0.18 ml/kg at baseline, both p<0.05). No significant changes in heart rate at different time intervals were noted. SAP decreased progressively after hypoxaemia was initiated, and became significantly lower than baseline at 180 min of hypoxaemia (61±16.8 vs. 82±12.9 mm Hg, respectively, p<0.05). A persistent and significant decrease in SVRI was noted after 30 min of hypoxaemia.

4.2.2 Pulmonary haemodynamic changes during hypoxaemia (Table 1Fig. 2)

The pulmonary vasculature showed progressive vasoconstriction with PAP increasing from 25±3.4 mm Hg to a maximum of 45±7.1 mm Hg at 120 min of hypoxaemia, and then remaining unchanged for the last 60 min of hypoxaemia. There were progressive increases in PVRI from 30 min to 180 min of hypoxaemia.

4.2.3 Mesenteric haemodynamic changes during hypoxaemia (Table 2Fig. 3)

In the mesenteric circulation, there was a transient hypoxaemic vasoconstriction with a 37% decrease in PVFI at 15 min of hypoxaemia (29±11.8 vs. 46±18.1 ml/kg/min at baseline, p<0.05). The PVFI returned to values indifferent from baseline from 30 min of hypoxaemia onwards. Portal venous flow contributed the major proportion of hepatic blood flow during normoxaemia and hypoxaemia, corresponding changes in THFI were therefore noted. HAFI was transiently elevated at 120 min of hypoxaemia (4.7±3.8 vs. 3.2±3.5 ml/kg/min at baseline, p<0.05).

4.2.4 Systemic and mesenteric oxygen metabolism changes during hypoxaemia (Tables 1 and 2)

During hypoxaemia there were ≈50% decreases in sys DO₂ as per protocol design. Mes DO₂ values were reduced by between 30% and 42% of baseline mes DO₂. These decreases were accompanied by significant increases in sys EO₂ and mes EO₂. The decreases in sys VO₂ and mes VO₂ were not significant (p=0.07 and 0.06 respectively, β=0.4). HDO₂ was also decreased corresponding to that of mes DO₂. Significant increases in HADO₂ ratio were found between 15 to 120 min of hypoxaemia when the total hepatic oxygen delivery was decreased during hypoxaemia (i.e. a greater proportion of hepatic oxygen delivery was derived from the hepatic arterial supply).

4.2.5 Effect of reoxygenation on haemodynamics and oxygen metabolism (Tables 1 and 2Figs. 1–3)

Upon 30 min of reoxygenation, there was a further fall in CI that was significantly lower than baseline (130±39 ml/kg/min, p<0.05). This was accompanied by a further decrease in SV (0.41±0.14 ml/kg) and persistent systemic hypotension (59±10.0 mm Hg) (both p<0.05 vs. baseline values). Despite 100% saturated arterial blood, sys DO₂ remained significantly diminished due to the persistently decreased CI after reoxygenation. While there was a partial normalisation of PAP (29±5.3 vs. 25±3.4 mm Hg at baseline, p>0.05), the PVRI remained significantly above baseline, with a persistent decrease in SAP/PAP ratio despite reoxygenation (p<0.05). The changes in mesenteric perfusion and oxygen delivery during 180 min of hypoxaemia were completely reversed after 30 min of reoxygenation. Although the sys VO₂ and mes VO₂ were diminished and remained similar to that of 180 min of hypoxaemia, they were not significantly different from the respective baseline values. Sys EO₂ and mes EO₂ returned to baseline values.

4.2.6 Effect of 30-min hypoxaemia followed by reoxygenation (n=5)

During 30 min of hypoxaemia, piglets were acidotic with pH 7.25–7.30 (mean 7.27±0.02) and similar changes in systemic and mesenteric haemodynamics and oxygen metabolism were observed (data not shown). Upon 30 min of reoxygenation with 100% inspired oxygen, CI gradually returned to baseline (174±50 vs. 146±16 ml/kg/min, respectively, p>0.05). SAP decreased modestly but was not different from baseline (61±19 vs. 73±4 mm Hg, respectively, p>0.05). In regard to oxygen consumption, both sys VO₂ and mes VO₂ returned to respective baseline values (9.6±5.84 and 1.4±0.69 ml/kg/min at 30 min of reoxygenation vs. 10.8±1.44 and 1.4±0.79 ml/kg/min at baseline, respectively).

5 Discussion

In this experiment, we maintained severe hypoxaemia with systemic oxygen saturations being kept between 40–50% for 3 h. We showed contrasting temporal haemodynamic responses of the systemic, pulmonary and mesenteric circulations during prolonged alveolar hypoxaemia and upon reoxygenation with 100% inspired O₂.

5.1 Systemic and pulmonary haemodynamic responses

The effect of systemic hypoxaemia on cardiac output in the newborn varies between studies depending on the degree of hypoxaemia induced [11, 12]. No increase in cardiac output was demonstrated during severe hypoxaemia with oxygen saturation less than 40%, while an increase in cardiac output has been shown with moderate hypoxaemia. We showed that cardiac output also varied with the duration of hypoxaemia. This is in part consistent with some reports that studied the effect of hypoxaemia on cardiac output for a maximum duration of 90 min. In our study, in response to a ≥50% decrease in the arterial oxygen content, CI initially increased and peaked at 30–60 min of hypoxaemia. This compensatory increase in CI, which was attributed to the increase in SV but not heart rate, was transient and was insufficient to maintain systemic oxygen delivery with this severity of hypoxaemia. After 1 hour of hypoxaemia, CI decreased gradually and was then not significantly different from baseline. The inability to maintain increased CI and SV may be related to progressive myocardial dysfunction due to acidosis [13], hyperlactataemia [14], and ATP exhaustion [15], which occur during sustained hypoxaemia. The systemic vasodilatation, which is the usual response to hypoxaemia that improves perfusion of hypoxaemic tissues, may be related to vasodilator metabolite accumulation [16] and increased nitric oxide release [17, 18]. Since SAP reflects the combined effect of CI and SVRI, it was only significantly decreased at the late stage of prolonged hypoxaemia which had resulted in a decrease in CI along with marked vasodilatation.

Hypoxaemic pulmonary vasoconstriction has been extensively investigated. While the response is observed even in the absence of functional endothelium [19], many vasoactive agents, including endothelin [20, 21], have been implicated. Inhibition of nitric oxide production may modulate this response [22–24]. Calcium and potassium channel activities may also play critical roles in hypoxaemic pulmonary vasoconstriction [25, 26]. Some workers have demonstrated that hypoxaemic pulmonary vasoconstriction is a biphasic phenomenon [9, 15, 27]. However, we did not observe that in our in vivo model.

5.2 Mesenteric haemodynamic responses

At the anatomic site of application of the flow probe, the measurement of portal venous flow includes contributions from the bowel and also a small contribution from the splenic vein. In the neonate, the blood flow to the bowel is approximately 15 to 20% of the cardiac output whereas the flow to the spleen is only about 2 to 3% of the cardiac output [28, 29]. In a previous study in the newborn lamb of severe hypoxaemia of 20 min duration, the relative decreases in gastrointestinal and splenic blood flows were approximately the same [29]. Therefore inferences from the measurement of portal blood flow at this location can reasonably be said to apply to the mesenteric circulation. Furthermore, estimation of hepatic oxygen delivery requires the inclusion of all sources of hepatic blood flow and placement of the flow probe in the location chosen was technically the only feasible methodology. Attempts to cannulate hepatic veins for measurement of hepatic oxygen consumption had been unsuccessful due to prolonged surgical time, increased morbidity and mortality.

The hypoxaemic response of the mesenteric circulation has been variable, with investigators showing either vasodilatation or vasoconstriction [30–32]. This discrepancy may be related to differences in the severity and duration of hypoxaemia [2, 3, 9] vessel size and species being studied [4–7] and the use of anaesthesia [2, 8]. Previous reports using single arteries in a constant perfusion-pressure preparation demonstrated mesenteric vasodilation to hypoxaemia [32, 33]. In vivo experiments with a short duration of hypoxaemia have shown an opposite response [34]. In our in vivo experiment, hypercapnia or hypocapnia, which have been shown to enhance or mask hypoxaemic vasodilation [31, 35], were avoided. We demonstrated a transient hypoxaemic mesenteric vasoconstriction; an initial intense vasoconstriction followed by vasorelaxation to baseline; confirming the results of prolonged hypoxaemia (1 h) in an in vitro single artery model by Leach et al. [9]. Corresponding changes in THFI were observed because PVFI contributes the major proportion of hepatic blood supply in this newborn model (about 90% at baseline). We postulate that the transient hypoxaemic mesenteric response is due to an initial phase of sympathetically mediated vasoconstriction followed by a phase of vasorelaxation due to progressive accumulation of metabolites during hypoxaemia. But the washout effect of altered flow on local metabolites may further complicate the picture.

It is interesting that while the SVRI decreased after 30 min of hypoxaemia that HAFI increased only after 120 min of hypoxaemia. The apparently unchanged HAFI in the first 60 min of hypoxaemia is due to a decrease in the arterial pressure despite vasodilatation. However, we cannot exclude a genuine differential response time in the systemic and hepatic arterial vasculature nor an apparent difference due to detection limits of our experimental design.

5.3 Oxygen delivery and metabolism

Many tissues respond to diminished oxygen delivery with increasing oxygen extraction and decreasing oxygen consumption [36]. The increasing oxygen extraction in this study was able to partially compensate for the decrease in oxygen delivery. Although our experiment did not show a significant decrease in systemic and mesenteric oxygen consumption, there was a trend demonstrated (p=0.07 and 0.06, respectively) and failure to show statistical significance may be related to small sample size (β=0.4). With the initial mesenteric vasoconstriction, bowel oxygen extraction increased dramatically. As the transient increase in MVRI abated, there was a reduction in the oxygen extraction after 60 min of hypoxaemia. The progressive fall in mesenteric oxygen extraction was probably a response to the persistent hypoxaemia with failing oxygen consumption and extraction in tissue [37]. This indicates significant cellular dysfunction has resulted from prolonged hypoxaemia. The temporal cellular response to prolonged hypoxaemia is similar to the response to graded hypoxaemia [38].

This is the first report to provide data on hepatic oxygen delivery in newborn animals during prolonged hypoxaemia. Our experiment showed the changes in the relative contribution of hepatic arterial and portal venous oxygen delivery in relation to the total hepatic oxygen delivery during prolonged hypoxaemia. We demonstrated a significant contribution from hepatic arterial flow to the hepatic oxygen delivery during severe systemic hypoxaemia. By 180 min of hypoxaemia, with the falling mesenteric oxygen extraction, portal venous oxygen content was increasing and the percentage portal venous oxygen delivery contribution to total hepatic oxygen delivery had increased. HADO₂ ratio decreased somewhat to a level not significantly different from baseline. These changes in the relative contributions of hepatic oxygen delivery support the significant role of hepatic arterial supplies in the unique hepatic circulation in the event of possible hypoxaemic hepatic injury during systemic hypoxaemia.

5.4 Reoxygenation effect and its clinical relevance

We should be cautious to extrapolate these results to human pathophysiological conditions. Upon reoxygenation, the mesenteric haemodynamic variables were completely reversed. But that was not the case for the systemic and pulmonary circulations. Reoxygenation with 100% inspired oxygen did not improve the myocardial dysfunction (reduced CI and SV) or persistent hypotension which occurred during the prolonged hypoxaemia. In addition to lactate acidosis, reoxygenation injury has been widely investigated and contributes to the development of myocardial stunning [39]. Following reoxygenation, the simultaneous production of nitric oxide and superoxide, hence, their product—peroxynitrite, may cause further damage to the already compromised myocardium [40, 41]. Thus, abrupt correction of hypoxaemia with 100% inspired oxygen may not lead to immediate improvement in cardiac performance, and may even be followed by further decreases as shown here.

There was only a partial recovery from the severe hypoxaemic pulmonary vasoconstriction and PAP was still elevated although not being significantly different from baseline. Acidosis may have contributed to the persistent pulmonary vasoconstriction [42]. The significant decrease in SAP/PAP ratio despite reoxygenation can explain in part, the mechanism of persistent fetal circulation after perinatal asphyxia.

On the other hand, sys VO₂ and mes VO₂ also did not improve upon reoxygenation. These findings contrast to those in 30 min of a similar degree of hypoxaemia, which showed a normalisation of oxygen consumption. We speculate that the prolonged ischaemia-hypoxia injury might have caused irreversible cellular damage. Also, free oxygen radicals and peroxynitrite generated during reoxygenation might prevent recovery of cellular function. Peroxynitrite has been shown to inhibit mitochondrial respiration in vitro [43, 44]. Therefore, the prolonged ischaemia-hypoxia injury may not be reversed or even aggravated by reoxygenation. Despite a normal PVFI, modestly decreased mes VO₂ along with the decrease in elevated mes EO₂ may suggest persistent cellular dysfunction. While necrotising enterocolitis has been associated with ischaemia-hypoxia injury and enteral feeding, caution is required for introduction of enteral feeds despite successful reoxygenation after asphyxia.

In this acute newborn animal experiment, we have demonstrated the contrasting effects of prolonged hypoxaemia on systemic, pulmonary and mesenteric perfusions. Prolonged hypoxaemia with oxygen saturations between 40–50% over 3 h caused progressive hypoxaemic pulmonary vasoconstriction and transient hypoxaemic mesenteric vasoconstriction. Following an initial rise, CI decreased with progressive systemic vasodilatation. In contrast to significant increases in sys EO₂ throughout hypoxaemia, elevations in mes EO₂ decreased after 30 min of hypoxaemia along with modest decreases in sys VO₂ and mes VO₂. Persistent systemic hypotension with a further decrease in CI was noted on reoxygenation with 100% inspired oxygen. The hypoxaemic pulmonary and mesenteric responses were partially or wholly reversed with reoxygenation, respectively. Cautions need to be taken upon reoxygenation because the neonates are at risk for the development of myocardial stunning, persistent pulmonary hypertension and necrotising enterocolitis.

Acknowledgements

This study was supported by the Heart and Stroke Foundation of Alberta and The Perinatal Research Centre, University of Alberta, Edmonton, Canada.

References

Author notes