An extra-uterine system to physiologically support the extreme premature lamb

Abstract

In the developed world, extreme prematurity is the leading cause of neonatal mortality and morbidity due to a combination of organ immaturity and iatrogenic injury. Until now, efforts to extend gestation using extracorporeal systems have achieved limited success. Here we report the development of a system that incorporates a pumpless oxygenator circuit connected to the fetus of a lamb via an umbilical cord interface that is maintained within a closed 'amniotic fluid' circuit that closely reproduces the environment of the womb. We show that fetal lambs that are developmentally equivalent to the extreme premature human infant can be physiologically supported in this extra-uterine device for up to 4 weeks. Lambs on support maintain stable haemodynamics, have normal blood gas and oxygenation parameters and maintain patency of the fetal circulation. With appropriate nutritional support, lambs on the system demonstrate normal somatic growth, lung maturation and brain growth and myelination.

Figure 1. UA/UV Biobag system design.

(a) Circuit and system components consisting of a pumpless, low-resistance oxygenator circuit, a closed fluid environment with continuous fluid exchange and an umbilical vascular interface. (b) Representative lamb cannulated at 107 days of gestation and on day 4 of support. (c) The same lamb on day 28 of support illustrating somatic growth and maturation.

Figure 2. Haemodynamic and oxygen parameters in CA/UV lambs versus UA/UV lambs.

CA/UV lambs data represented by grey circles (n=5) and UA/UV lambs as black circles (n=8). (a) Heart rate. (b) Mean pre-membrane oxygenator pressure calculated as 1/3 systolic+2/3 diastolic. (c) Body weight-adjusted circuit flow. (d) Post-membrane oxygen saturation (sat). (e) Haemoglobin. (f) Oxygen delivery. Data are presented as mean±s.e.m. Statistically significant difference between groups in (a–f) is denoted by *P<0.05 (analysis of variance (ANOVA)).

Figure 3. Oxygen parameters in UA/UV lambs.

Data from n=8 lambs. (a) Haemodynamic and laboratory parameters. (b) Pre-membrane P_aCO₂ and P_aO₂. (c) Haemoglobin levels with and without erythropoietin (EPO). (d) Pre-membrane pH. (e) Plasma lactate. † In utero control values in (a) derived from measured data in ref. . Data in (b–e) are presented as mean±s.e.m. Statistically significant difference between groups in (c) denoted by *P<0.05 (analysis of variance (ANOVA)).

Figure 4. Echocardiographic parameters in UA/UV lambs.

Data from n=8 lambs. (a) Echocardiographic parameters. (b) Combined cardiac output. (c) Ductus arteriosus (DA) flow to right ventricular output ratio. Data in (b,c) are presented as mean±s.e.m. *In utero microsphere data in (a) derived from measured data in ref. . ^†Lamb 3 noted to have Tetralogy of Fallot (TOF) with restricted pulmonary flow before cannulation was excluded from mean value calculations (‡) due to TOF physiology.

Figure 5. Nutrition and growth in UA/UV lambs.

Data from n=8 lambs. (a) Nutritional substrate and laboratory parameters; data are presented for individual animals and as group averages (mean± s.e.m.). (b) Control growth curve using UA/UV and control lamb weights at hysterotomy (n=19). Solid line represents exponential best fit. (c) UA/UV growth curve. Solid line represents exponential best fit of UA/UV non-insulin lamb weights at hysterotomy and at end of studies. Dashed line represents control growth curve. ^†In utero control values in (a) derived from measured data in refs , , . *P value in (c) refers to between-group analysis of variance (ANOVA; statistical significance defined as P<0.05).

Figure 6. Structural and biochemical lung development and early neonatal pulmonary function in UA/UV lambs.

(a–e) Morphology of control ((f), 113d GA and (g),139d GA) and experimental lambs ((h), 132d GA, CA/UV lamb 3; (i), 144d GA, UA/UV lamb 5) following 19 and 28 days on circuit, respectively (scale bars, 50 μm). Ongoing alveolar formation on circuit is evidenced by increased density and height of secondary septae (arrowheads). Density of surfactant protein-B-positive alveolar cells ((j), arrowheads in k,l) and neonatal pulmonary function were not different from that of age-matched controls (m). *P value in (j) refers to difference between groups (Student's unpaired t-test, statistical significance defined as P<0.05). ^†Control group in (j) includes near-term lambs only (n=4, mean GA 141 days). ^§Mean control values in (m) derived from measured data in ref. . ^‡Near-term control lambs in (m) delivered by caesarean section and ventilated in the same manner as experimental animals. Data in (j,m) are presented as mean±s.e.m.

Figure 7. Neurologic development and maturation in experimental lambs.

(a) Post-mortem brain-to-body weight ratio of experimental (Exp.) and control (Ctrl.) animals; animals that received insulin are excluded. (b) Gyral width on haematoxylin and eosin (H&E) stained sections by region as compared with 140-day gestation controls. (c) Biparietal diameter in experimental (Exp.) versus predicted (equation derived from ref. 26) animals. (d) Postnatal T2-weighted coronal magnetic resonance imaging (MRI) at 6 months of age. Scale bar in (d) is 1 cm. Data in (a–c) are presented as mean±s.e.m. P values in (a–c) refer to difference between control and experimental groups (Student's t-test for each region/time point, with statistical significance defined as P<0.05). BPD, bronchopulmonary dysplasia.

Figure 8. Histologic and other parameters of brain development and maturation.

(a) Representative haematoxylin and eosin (H&E)-stained sections of post-therapy cerebrum/cerebellum and their respective cortices displaying normal brain parenchyma and absence of injury. (b) Representative Luxol fast blue myelin stain and digital pixel identification depicting myelin density (orange). (c) Percent positive pixels identified in the selected regions. (d) Maximum positivity (density of myelin) in the selected regions. (e) Ocular EMG activities in instrumented in utero control (Ctrl.) and experimental (Exp.) fetal lambs at 120 days (upper tracing) and 134 and 139 days, respectively (lower tracing). White area of bar represents activity and black areas quiescence. (f) Correlation of middle cerebral artery pulsatility index (MCA PI) and fetal oxygen delivery; solid line represents linear best fit. Data in (c,d) are presented as mean±s.e.m. The scale bars in (a,b) from left to right are 4 mm, 300 μm, 1.5 cm, 150 μm, 4 mm and 2 mm, respectively. P values in (c,d,f) refer to difference between control and experimental groups (Student's t-test in c,d and Pearson's correlation coefficient in f, with statistical significance defined as P<0.05).