The Lamb Paper
A technical breakdown of a seminal paper in the extra-uterine device (EUD) literature.
In 2017, researchers from the Children’s Hospital of Philadelphia successfully grew fetal lambs inside extra-uterine devices (EUDs) for periods ranging from 20 to 28 weeks. Reports of this experiment were published in a Nature Communications paper that made headlines around the world.
What made these findings a breakthrough was the extreme premature fetuses’ outstanding duration of support in an ex-vivo uterine environment (i.e., outside the mother’s body) as well as positive fetal outcomes consisting of normal somatic growth, lung maturation, and brain growth and myelination, all without apparent physiologic derangement or organ failure.
The purpose of this piece is to break down the key elements of the Lamb Paper from a technical standpoint.
1. Why?
Extreme prematurity is the leading cause of infant morbidity and mortality in the United States.
The limit of viability is currently between 22 to 23 weeks of gestation. This being said, preterm infants are at higher risk of chronic lung disease and other complications of organ immaturity.
When preterm neonates are moved from liquid ventilation (in utero, when oxygen is delivered through the umbilical vein) to gas ventilation (ex utero, in incubators), their lungs are still structurally and functionally immature.
Because of this immaturity, premature babies often experience respiratory distress syndrome and require mechanical ventilation and/or oxygen therapy. However, high amounts of inhaled oxygen and pressure related to these treatments can cause inflammation and damage to the lungs, leading to bronchopulmonary dysplasia, which is an arrest in lung development.
Currently, there are no ways to keep fetuses and babies alive in extra-uterine environments using liquid ventilation. This calls for the development of EUDs that reproduce physiological conditions.
2. What?
2.1 Components and Parameters
Components of the EUD developed by the team include:
A pumpless arteriovenous circuit;
A closed sterile fluid environment;
Umbilical vascular access.
Parameters that were controlled for include:
Physiologic extracorporeal support of the fetus;
Nutrition.
2.1.1 Pumpless Arteriovenous Circuit
In the uterine environment, fetuses rely on the maternal circulation (through umbilical vessels) to provide oxygen and nutrients and eliminate carbon dioxide and waste materials. In this study, researchers attempted to replicate this system using extracorporeal membrane oxygenation (ECMO), a well-established technology that was developed to replace the heart and lungs via an oxygenator connected to tubes called cannulae. Nowadays, ECMO is widely used in critical care settings and it is increasingly utilized on neonates suffering from cardiac and respiratory failure.
In this study, the authors used a small-volume, near-zero-resistance oxygenator and short segments of circuit tubing consisting of either ECMO cannulae (Medtronic) or custom-made umbilical cannulae via 3/16′ ID × 1/16′ wall thickness BIOLINE-coated tubing (Maquet). This minimized surface area and priming volumes.
With most lambs, a Quadrox-ID Pediatric oxygenator (Maquet Quadrox-ID Pediatric Oxygenator: Maquet Cardiopulmonary AG, Rastatt, Germany) was used.
Priming volume: 81 ml.
With smaller lambs (0.5 to 1 kg), a modified Quadrox Neonatal oxygenator (Maquet Quadrox-I Neonatal and Pediatric Oxygenator: Maquet Cardiopulmonary AG) was used.
Priming volume: 38 ml.
In the end, the extra-uterine system was comparable to the volume of the placenta itself (for sheep, placental blood volume is estimated to be between 23.1 to 48.1 ml per kg). The circuit priming volume was within the normal placental volume range for lambs 1 to 3 kg.
2.1.2 Closed Sterile Fluid Environment
Researchers developed a single-use, closed system called the Biobag. Because the amniotic fluid is constantly recycled (using micropore filters) and never escapes the circuit, the system is deemed completely closed. The Biobag’s design reduced amniotic fluid volumes and could be customized in order to replicate the size and shape of the uterus.
A sterile synthetic amniotic fluid was used:
Electrolyte composition: Na (109 mM), Cl (104 mM), HCO3 (19 mM), K (6.5 mM), and Ca2 (1.6 mM);
pH: 7.0–7.1;
Osmolarity: 235.8 mOsm/kg water.
Temperature 38.5-40.5 degree Celsius.
Lambs were incubated in individually customized bag enclosures of 2- to 4-litre total volume made of antimicrobial polyethylene film (Wiman Custom Films & Laminates, Sauk Rapids, MN, USA) for UA/UV animal studies.
2.1.3 Umbilical Vascular Access
In the final study, following surgery to remove lamb fetuses from the pregnant ewes, researchers placed ECMO cannulae in one umbilical vein as well as two umbilical arteries. This constituted a double umbilical artery, single umbilical vein (UA/UV) cannulation.
The purpose of this unique design was to reproduce fetal circulation and flow dynamics in utero as much as possible. Just as in the uterine environment, deoxygenated blood carrying metabolic waste left the fetus through the 2 umbilical arteries and returned through the single umbilical vein after being charged with nutrients and oxygenated — in this case, using ECMO.
2.1.4 Physiologic Extracorporeal Support
Researchers used a blend of room air and nitrogen to ensure the oxygenator delivered an oxygen concentration of 11-14%:
[…] sweep gas supplied to the oxygenator was a blended mixture of medical air, nitrogen and oxygen titrated to achieve fetal blood gas values (target PaO2 20–30 mm Hg, target PaCO2 35–45 mm Hg).
Erythropoietin (a naturally produced hormone that stimulates red blood cell production) was also administered to prevent the risk of transfusion due to anemia.
Stored whole maternal blood was transfused as required (10–20 ml/kg) to maintain fetal Hgb levels above 9 g/dl. In a subset of lambs (Prototype IV lambs 4–6), erythropoietin (400 U/kg) was administered intravenously once daily to promote fetal erythropoiesis, and was held for Hgb >16 g/dl.
2.1.5 Nutrition
The nutrition provided to the animals was based on substrate uptake of late-gestation fetal lambs and was predominantly composed of carbohydrate and amino acids with trace lipid.
For UA/UV lambs, researchers used a combination of:
Amino acids (TrophAmine 10%, titrated to blood urea nitrogen target level 30 mg/dl);
Lipids (Intralipid 20%, 0.1–0.2 g/kg per day);
Dextrose (titrated to blood glucose target 30–40 mg/dl);
Iron (1.0–1.5 mg/kg per day, titrated to plasma iron target 200–300 μg/dl).
The team ensured dextrose and amino acid levels were similar to those in plasma glucose (< 40 mg/dl) and blood urea nitrogen (< 30 mg/dl).
Insulin infusion was used on the last two UA/UV lambs, which improved substrate utilization and fetal growth and allowed the animals to tolerate higher caloric loads.
3. How?
Researchers removed 23 fetal lambs from pregnant ewes (female sheep) and placed them into incubators (open circuit, semi-closed circuit, Biobag) using different cannulation strategies (CA/JV, CA/UV, UA/UV; CA = carotid artery, JV = jugular vein, UA = umbilical artery, UV = umbilical vein).
5 lambs were placed in an open circuit incubator using CA/JV cannulation;
5 lambs were placed in a semi-closed circuit incubator using CA/JV cannulation;
5 lambs were placed in a Biobag incubator using CA/UV cannulation;
8 lambs were placed in a Biobag incubator using UA/UV cannulation. This group produced the most remarkable results — more on this later.
Fetal lambs were transferred to EUDs at gestational ages (age of the pregnancy measured from the date of last menstruation as opposed to the date of conception) of 104 to 125 days (term is ∼145 days). At 100-115 days of gestation, lamb fetuses are biologically equivalent to 22-24 week premature human infants from the perspective of lung development, being in the mid to late canalicular phase, which is the period during which respiratory bronchioles develop.
3.1 Closed Problems
How did the team come to this design? This section contains an analysis of major challenges paired with solutions that allowed researchers to “close” these problems.
Prior to the paper, the development of “artificial placentas” was met with three major obstacles:
Progressive circulatory failure due to preload or afterload imbalance imposed on the fetal heart by oxygenator resistance and pump-supported circuits;
Use of open fluid incubators resulting in contamination and fetal sepsis;
Problems related to umbilical vascular access leading to vascular spasm.
These obstacles were addressed by the system’s three main components: the pumpless arteriovenous circuit, the closed fluid environment with continuous fluid exchange, and the umbilical vascular access. Researchers iterated the design of the EUD over multiple pilot studies:
A series of pilot studies leading to our final device were performed that identified potential obstacles and allowed sequential design modifications. All pilot studies utilized the pumpless arterial–venous (AV) circuit […] The primary design modifications were related to the fluid environment and our approach to vascular access.
3.1.1 Preload or Afterload Imbalance → Pumpless Arteriovenous Circuit
Now for a bit of cardiac physiology.
The heart constantly alternates between two states, also called phases of the cardiac cycle: contraction (systole) and relaxation (diastole). During diastole, both right and left ventricles stretch as blood returns from the body into the right ventricle and from the lungs into the left ventricle. During systole, both right and left ventricles contract as blood is pushed from the right ventricle to the lungs (pulmonary circulation) and from the left ventricle to the rest of the body (systemic circulation).
Preload refers to the initial stretching of cardiomyocytes (muscle cells of the heart) at the end of diastole, right before ventricular filling. It is also called the left ventricular end-diastolic pressure (LVEDP).
Afterload refers to the amount of resistance (force, load) the left ventricle has to overcome in order to eject the blood into the systemic circulation. It is also called the systemic vascular resistance (SVR).
Here is another explanation that might provide clarity:
Preload and afterload can be thought of as the wall stress that is present at the end of diastole and during left ventricular ejection, respectively. Wall stress is a useful concept because it includes preload, afterload, and the energy required to generate contraction. Wall stress and heart rate are probably the two most relevant indices that account for changes in myocardial O2 demand. Laplace’s law states that wall stress (σ) is the product of pressure (P) and radius (R) divided by wall thickness (h):
σ=P×R/2h
Inadequate preload or afterload can affect wall stress, damaging cardiomyocytes over time and ultimately causing heart failure. Instead of straining to find the optimal preload and afterload for each lamb fetus, the authors instead opted for a pumpless circulation system:
From the inception of the study, we reasoned that a pumpless circuit—in which blood flow is driven exclusively by the fetal heart—combined with a very low resistance oxygenator would most closely mimic the normal fetal/placental circulation.
While previous studies have also explored pumpless circuits, they were hindered by the risk of cardiac failure:
If the circuit or oxygenator shows supraphysiologic (higher than what is naturally found) resistance, this causes increased afterload (afterload imbalance), eventually leading to left-sided heart failure.
If the circuit oxygenator shows subphysiologic (lower than what is naturally found) resistance, this causes a fall in systemic arterial blood pressure and subsequent neurohormonal activation, eventually leading to high-output heart failure. This is the more commonly encountered difficulty.
So far, other teams have used vasopressors (drugs that constrict blood vessels and allow systemic resistance to rise), pumps, and external flow regulators to counter subphysiologic resistance and supraphysiologic flow. These measures produced limited results with subjects requiring dialysis (renal replacement therapy) and continuous paralysis and presenting hydrops (abnormal fluid buildup) and respiratory failure.
For this study, researchers used near-zero-resistance oxygenators that were developed using hollow fibre plate technology. No instances of cardiac failure was recorded.
3.1.2 Contamination and Fetal Sepsis → Biobag
In the initial pilot studies, researchers used an open circuit (open fluid bath with continuous recirculation of the amniotic fluid), then a semi-closed circuit (continuous exchange of amniotic fluid as opposed to recirculation). Both led to bacterial overgrowth in the amniotic fluid and sepsis, although the second design allowed lambs to remain in the system for longer periods of time.
In the end, the authors opted for a novel, closed fluid circuit: the Biobag.
The Biobag’s design allows it to remain sterile, adapt to the lambs’ size, and efficiently use space and fluid volume.
It is made of a polyethylene film that is translucent and flexible. Because it is also sonolucent, meaning that it permits the passage of ultrasonic waves without the production of echoes, the film allows for easy monitoring via ultrasound. Manipulation and scanning are also made easier.
Water-tight ports accommodated cannulas, temperature probes, and sterile suction tubing, while an open, sealable side allowed for easy insertion of the fetus.
The development of the Biobag essentially solved the problem of gross fluid contamination, and has eliminated pneumonia on lung pathology. Throughout the subsequent experiments, low-level amniotic fluid contamination was observed only in circumstances where Biobag re-entry was required. When this occurred, contamination could be cleared by increasing the fluid exchange rate and injecting antibiotics into the bag fluid on a daily basis.
3.1.3 Vascular Spasm → Adapted UA/UV Cannulation
Vascular spasm is a well-known challenge for neonatal intensive care physicians, as the umbilical artery begins to constrict shortly after birth. To avoid this problem, researchers used the carotid artery (CA) and jugular vein (JV) for vascular access for the first 5 lambs (CA/JV cannulation). However, this group suffered sepsis and cannula-related complications, in part due to the open circuit design of the EUD.
These problems lessened but persisted with the second EUD design, a semi-closed circuit that was used on another group of 5 lambs using CA/JV cannulation. Compared with in utero controls, nevertheless, mean arterial pressure (related to blood flowing from the left ventricle into the systemic circulation) was lower and right-sided venous pressure (related to blood flowing from the systemic circulation into the right ventricle) was higher. This led to a weakened circuit flow.
For the third group of 5 lambs, the jugular venous (JV) access was replaced with an umbilical venous (UV) access in order to offload the right atrium (CA/UV cannulation). For this group, researchers introduced the final EUD design, which was the Biobag. Hemodynamic outcomes were more favorable, but the arterial flow remained below physiologic values, mostly due to the small size of the carotid artery (CA).
For the fourth group of 8 lambs, the authors opted for the more physiologically faithful umbilical artery (UA), umbilical vein (UV) model (UA/UV cannulation), also using the Biobag. Measures were taken to reduce the risk of umbilical cord spasm:
Researchers used topical papaverine (an antispasmodic drug), atraumatic operative technique, and maintained warmth and physiologic oxygen saturation of the blood entering the umbilical vein;
The team also developed a technique for cannulation that minimized vascular spasm, decannulation events, and the risk of mechanical obstruction, allowing for a length of native (belonging to the fetus) umbilical cord of 5-10 cm between the end of the cannulas and the abdominal wall.
3.1.4 Oxygenation Concerns and Nutrition
Researchers noted that supraphysiologic oxygen content in the blood delivered to the fetus via the umbilical vein impaired erythropoietin production in the fetal liver. This led to diminished erythrocyte (red blood cells) production, which itself resulted in lower levels of hemoglobin, causing anemia and eventually requiring transfusion. Two strategies allowed the team to mitigate this issue.
Throughout the studies, researchers made efforts to maintain normal fetal oxygen tension and carbon dioxide exchange while providing normal oxygen delivery.
Researchers also noted a correlation between oxygen delivery and substrate tolerance: UA/UV lambs tolerated physiologic levels of substrate delivery when oxygen delivery was relatively higher.
Daily erythropoietin administration in the last 5 UA/UV lambs completely eliminated the need for blood transfusion in the last 3 animals.
3.2 Open Problems
What challenges did the authors of this paper identify but did not succeed in solving, leaving these problems “open”? What problems do researchers envision in future studies?
3.2.1 Intracranial Brain Hemorrhage
A major concern in premature infants is intracranial hemorrhage, which can be problematic when extracorporeal support systems such as ECMO are used as these require anticoagulation.
To counter this, researchers used “substantially reduced” heparin doses compared with conventional ECMO in order to preserve activated coagulation time between 150-180s. This was made possible due to the reduced surface area of the circuit as well as a heparin-bound coating on all components of the system that were in contact with blood.
Authors advance that non-heparin-based coatings will eventually be developed, further enhancing safety. They also propose that intracranial brain hemorrhage is related to interventions that were not performed in this study, such as positive pressure ventilation and inotrope drugs:
Thus, physiologic support in a extracorporeal system without ventilation or pressors may, in itself, reduce the likelihood of haemorrhage, making prediction of the impact of our system on intracranial haemorrhage difficult.
3.2.2 Amniotic Fluid
While the authors used a simple electrolyte solution, amniotic fluid contains trophic factors and other components such as hormones. Therefore, more research in this area is warranted.
3.2.3 Circuit Flow Autoregulation
While the pumpless circuit design relies on the fetal heart to regulate flow, researchers noted that very premature lambs were unable to autoregulate circuit flow, resulting in increased flow leading to fluid buildup (hydrops):
This suggests that there is a delicate balance between adequate and excessive circuit flow and that the ability to compensate for increased flow and supraphysiologic right atrial pressures may be dependent on developmental maturity.
4. Outcomes / Results
The authors measured a myriad of outcomes. Here are the main ones in our opinion.
Overall, the UA/UV lambs (all contained in Biobags) demonstrated the most impressive results.
5 fetuses of 105-108 days of gestation were incubated for 25-28 days and 3 fetuses of 115-120 days of gestation were incubated for 20-28 days:
The longest runs were terminated at 28 days due to animal protocol limitations rather than any instability, suggesting that support of these early gestational animals could be maintained beyond 4 weeks.
Growth and maturation was noted in lambs that underwent prolonged runs:
Animals opened their eyes, became more active, had apparently normal breathing and swallowing movements, grew wool and clearly occupied a greater proportion of space within the bags […].
Hemodynamic and circuit flow parameters were measured for all lambs. In contrast with other lambs, the UA/UV lambs demonstrated levels of circuit flow comparable to normal placental flow at 150–250 ml/(kg · min). Daily echocardiography confirmed adequate cardiac function in terms of ductus arteriosus flow, ductus venosus patency and flow, right to left shunting through the foramen ovale, cardiac contractility, and chamber and vena caval size.
Lung maturation was assessed in UA/UV lambs by detailed morphometric analysis, histologic assessment, surfactant protein B analysis, and analysis of function after birth. Progression from the canalicular to saccular stages of lung development was noted.
Despite known maternal contributions to hepatic and renal function, metabolic parameters related to organ function and nutritional status remained surprisingly stable.
Neurologic growth and maturation were also assessed in UA/UV lambs using brain-to-body weight ratio, gyral width, and biparietal diameter. Researchers also evaluated ischemic brain injury using multiple modalities:
Postnatal T1-, T2-, and diffusion-weighted magnetic resonance imaging (MRI) sequences at 6 months of age;
Routine hematoxylin and eosin (H&E) staining on critical brain regions;
Densitometry of myelin-stained critical brain regions.
No evidence of hemorrhage, ischemia, infarction, or demyelination was found. However, it is important to note that fetal lamb brain maturation cannot be directly compared with human brain maturation, notably due to the earlier maturation of the germinal matrix in the lamb (70 days). Thus, any conclusion regarding neurologic development should be considered with caution.
Pumped for the working groups to begin!!!!