Nothing but fibrein the blood path
Every part added to a circuit adds volume to fill and surface to clot on. The baseline is a fibre bundle in a clear housing, with a port at its high point to let the air out.
Talk to us
A blood oxygenator for a 500 g patient: 8 mL to fill, no pump, driven by the infant’s own heart.
In development. What you read here are design targets, not test results.
The smallest oxygenator a hospital can buy holds 40 mL of blood. An infant born at 500 g has about 45 mL in total. Novanatomy is designing one that fills with 8.
Who it is for. Infants born at 22 to 25 weeks, weighing 400 to 1,000 g. Their lungs are not ready for air, and ventilating them can cause lasting injury. Support outside the body could spare the lungs, but no circuit on the market is sized for a patient this small.
What we are making. A single-use hollow-fibre cartridge with a matched cannula set. About 0.1 m² of membrane, 8 mL of fill or less, and resistance low enough that the infant’s own heart moves the blood. Nothing else sits in the blood path.
Where it stands. Requirements and sizing are drafted and prototypes come next. Every figure on this page is a design target, not a measured result.
Who we want to hear from. Neonatologists, fetal surgeons, ECMO coordinators and perfusionists who can tell us where these targets are wrong. Laboratories running preterm lamb or piglet models. Engineers who have built hollow-fibre devices.
500 g
A circuit has to be filled before it can run, and whatever fills it mixes with the patient’s blood. When the fill is nearly as large as the patient’s whole blood volume, the circuit must be primed with donor blood.
At 500 g, filling today’s smallest oxygenator takes 89% of the infant’s blood volume. The target cartridge takes 18%.
Areas are to scale. Blood volume is estimated at 90 mL per kg, a textbook approximation. The 40 mL figure is the published fill volume of the smallest commercial neonatal oxygenator, which has 0.38 m² of membrane.
The circuit copies the route blood already takes to the placenta. The cannulas use half of the pressure the heart can offer, so they are designed together with the cartridge.
Blood low in oxygen leaves through the two umbilical arteries. Only the infant’s heart pushes it: about 20 mmHg of driving pressure.
It passes once through a short, wide bundle of hollow fibres. Oxygen moves in and carbon dioxide out, for 5 mmHg or less.
Oxygenated blood returns through the umbilical vein. No pump, heat exchanger, filter or reservoir stands in the way.
8 mL
to fill the cartridge, against 40 mL for today’s smallest.
28 days
of continuous support without a change-out. The floor is 14.
Pressure, not gas exchange, is the hard constraint. Membrane area can shrink fourfold, but resistance has to stay as low as a device four times the size.
| Parameter | Must meet | Aim for |
|---|---|---|
| Patient weight | 400 to 1,000 g | Design point 500 g |
| Blood flow | 40 to 300 mL/min | Stable down to 30 mL/min |
| Oxygen transfer | 8 mL/min at 150 mL/min of flow | 12 mL/min |
| Carbon dioxide removal | 8 mL/min | 12 mL/min |
| Cartridge fill | 8 mL or less | 5 mL or less |
| Whole circuit fill | 15 mL or less | 10 mL or less |
| Pressure drop | 5 mmHg or less at 100 mL/min | 3 mmHg or less |
| Continuous run | 14 days | 28 days, no change-out |
| Blood thinners | Low-dose heparin | None |
| Red-cell damage | No worse than today’s neonatal device | Half |
Every part added to a circuit adds volume to fill and surface to clot on. The baseline is a fibre bundle in a clear housing, with a port at its high point to let the air out.
Talk to usPolymethylpentene hollow fibre, the one membrane proven to run for weeks without leaking plasma.
PMP fibreGas insideBlood outside
A short, wide stacked mat. A large face and a blood path of a few millimetres give the least resistance for the area.
Stacked matShort pathLow resistance
One 0.1 m² module serves the smallest patient. Three or four in parallel serve a 3 kg newborn on ECMO.
StackableOne partTwo uses
A passive anti-clotting coating. The aim is support with no blood thinners, because bleeding in the brain is the main risk.
Passive coatingNo drug releaseHeparin-free aim
Heat comes from the enclosure and the sweep gas, flow from a clamp-on sensor. Neither adds a millilitre.
Warmed enclosureClamp-on flowFlush pressure ports
Each prototype climbs the same ladder, cheapest test first, and only a design that passes one rung moves to the next. This is the plan, not a record. Results will be published here as they exist.
| Stage | Test | Method | Pass |
|---|---|---|---|
| Bench | Pressure and flow | Glycerol and water at blood viscosity, 20 to 400 mL/min | 5 mmHg or less at 100 mL/min |
| Fill and de-airing | Fill by weight; time to bubble-free | 8 mL or less, under 2 minutes | |
| Blood | Gas transfer | Ovine or bovine blood at standard and fetal conditions | 8 mL of oxygen a minute at 150 mL/min |
| Red-cell damage | Six-hour loop beside today’s neonatal device | No higher than the control | |
| Clotting | Low-heparin and no-heparin loops, each coating | No loss of flow over 6 hours | |
| Endurance | Pumpless behaviour | Pulsing mock circulation at 20 to 25 mmHg | 150 mL/kg/min for a 500 g model |
| Durability | 30 days of continuous running | No leak, under 20% loss of transfer | |
| Animal | Acute | Preterm lamb or piglet, 24 to 72 hours, with a partner laboratory | Stable flow and blood gases |
| Chronic | 14 days, with a partner laboratory | 14 days of stable pumpless support |
We want to hear from the people who would use this cartridge, test it, or tell us where it is wrong.
Run a NICU, an ECMO programme or a fetal lab?
Tell us where you work and which target you would change first.