PRIMA Retinal Implant for Macular Degeneration Explained, From the NEJM Trial to Europe's First Approval

PRIMA is a wireless chip, two millimetres square, implanted beneath the retina of people who have lost their central vision to geographic atrophy, the advanced "dry" form of age-related macular degeneration. It works with camera glasses that project an image onto the chip in near-infrared light. The chip converts that light into small electrical currents that stimulate the retinal cells still alive above it.

In the pivotal trial, published in the New England Journal of Medicine, most participants regained enough central vision to read letters and words. In July 2026 the system received a CE mark, which permits it to be sold across the European Economic Area. That makes it the first neural implant approved in Europe for restoring form vision, meaning the ability to see shapes such as letters rather than just light and dark.

That is the short version. The longer version matters, because PRIMA does not restore normal sight, does not stop the disease, and requires eye surgery with real risks. It is also not approved in the United States. It is a clinically validated neuroprosthesis, a device that restores a lost function by stimulating the nervous system, and it is best understood on those terms.

The problem PRIMA was built for

Age-related macular degeneration damages the macula, the small central region of the retina responsible for reading, recognising faces and seeing fine detail. In geographic atrophy, patches of the retina's supporting layer and the photoreceptors above it die and do not come back. The central visual field develops a growing blind area, while peripheral vision is often largely preserved. Geographic atrophy affects roughly five million people worldwide.

Drugs approved in recent years for geographic atrophy aim to slow how fast the atrophic patches grow. They do not return vision that has already been lost. PRIMA addresses that gap: it does not repair the retina, but tries to take over the job of the missing photoreceptors.

The device relies on a biological fact. When photoreceptors die in AMD, the next layer of neurons in the retina, the bipolar cells, often survives for years. So do the ganglion cells that carry visual signals to the brain through the optic nerve. The visual pathway from the bipolar cells onward stays largely intact, even though the input stage has failed. If those bipolar cells can be stimulated electrically in a spatial pattern, the rest of the visual system can process the result.

How the PRIMA system works

The system has three parts.

The implant. A thin crystalline silicon chip, 2 × 2 mm across and about 30 micrometres thick, placed under the retina inside the atrophic area. It carries 378 hexagonal pixels, each 100 micrometres wide. Each pixel contains photodiodes that turn incoming light into electrical current, plus an active electrode and a return electrode. The implant has no wires, no battery and no cable passing through the wall of the eye. The light that carries the image also powers it.

The glasses. A camera on the frame captures the scene in front of the wearer. After processing, a projector in the glasses shines the image into the eye as pulsed near-infrared light at a wavelength of 880 nanometres. Any surviving photoreceptors are essentially insensitive to this wavelength, so the projected image stimulates the implant without disturbing the natural vision around it.

The processor. A pocket-sized unit handles image processing, including a zoom function that lets users magnify text. Zooming makes smaller print readable, at the cost of shrinking how much of the scene fits on the chip at once.

The resulting signal pattern enters the visual pathway at the bipolar cell level, which lets it use the retina's own processing before signals reach the brain. This is the main difference from earlier approaches. The best known of them, the Argus II, used an array placed on the surface of the retina and connected by cable to electronics outside the eye, and it stimulated ganglion cells further along the pathway. Argus II was discontinued around 2020. A wireless design placed under the retina avoids a permanent opening in the eye and makes the surgery simpler.

PRIMA is a direct output of academic research. The photovoltaic retinal prosthesis concept was developed in Daniel Palanker's laboratory at Stanford University and first tested in animals, then in small human feasibility studies. The French company Pixium Vision took it into clinical trials. After Pixium ran into financial difficulties in 2024, Science Corporation acquired PRIMA and completed its path to regulatory approval.

What the PRIMAvera trial showed

The key evidence comes from PRIMAvera (ClinicalTrials.gov NCT04676854), led by Frank G. Holz of the University Hospital Bonn. It was published in the New England Journal of Medicine in October 2025, with a 2026 print date. The study was conducted in humans, funded by Science Corporation and by Moorfields' NIHR Biomedical Research Centre in the UK.

How the study was designed

• Design: open-label, multicentre, prospective, single-group and baseline-controlled. Each participant's vision was compared with their own vision before implantation. There was no randomised control group and no sham procedure.

• Participants: 38 people with geographic atrophy and very poor central vision, with visual acuity of 1.2 logMAR or worse (about 20/320 on the Snellen scale). They were implanted at 17 sites in five European countries.

• Measurements: vision was tested with and without the PRIMA glasses at 6 and 12 months.

• Primary endpoints: the share of participants whose visual acuity improved by a clinically meaningful amount, defined as at least 0.2 logMAR (about 10 letters, or two lines, on a standard ETDRS eye chart), and the number and severity of serious adverse events linked to the procedure or device.

The results

Of the 38 participants, 32 were assessed at 12 months. Of the six who were not, three had died (unsurprising in an elderly population), one had withdrawn and two were unavailable for testing.

• 26 of the 32 participants (81%) reached the primary efficacy endpoint.

• The researchers used statistical imputation to account for the missing six and estimated that 80% of all 38 implanted participants would have met the endpoint. The result was statistically significant.

• The average natural peripheral vision after surgery was equivalent to baseline. Implanting the chip did not measurably reduce the vision people had around their blind central area.

Science Corporation's summaries report an average improvement of about 25.5 letters, more than five lines on the eye chart. They also say that 84% of participants reported being able to read letters, numbers and words at home. Those figures come from the same trial, but readers should know that they are the company's presentation of the data. The peer-reviewed primary endpoint is the proportion who improved by at least 0.2 logMAR.

An independent NEJM editorial by Jacque Duncan of the University of California, San Francisco, who was not involved in the study, described PRIMA as the first treatment to restore vision in geographic atrophy. That independent assessment of the peer-reviewed data carries more weight than any press release.

The safety record

The trial recorded 26 serious adverse events in 19 of the 38 participants, half the group. The most common was ocular hypertension (raised pressure inside the eye). Others included peripheral retinal tears, bleeding under the retina during implantation and full-thickness macular holes, and one participant developed retinal detachment with proliferative vitreoretinopathy. Four events were graded severe. About 81% occurred within two months of surgery, and 95% of those resolved within two months.

The investigators linked all serious adverse events to the surgery itself or to the surgery combined with the device. None were attributed to the device alone. The profile matches what surgeons expect from vitrectomy and surgery beneath the retina. That is reassuring, but it also means complications are a real part of the treatment, and patients and surgeons have to weigh them against the expected benefit.

What "restored vision" means here

This point is most often lost in coverage. PRIMA produces prosthetic vision, which differs from natural sight in several ways.

Resolution is limited by the pixel size. Each pixel is 100 micrometres wide. On the human retina, this spacing corresponds to a theoretical best acuity of about 20/420. Earlier clinical studies found that patients' measured prosthetic acuity closely matched this limit. An eye seeing at 20/420 is still legally blind by most definitions. What changes is that people with an empty central field become able to see letters again, especially with magnification.

The field is small. The chip covers only a few degrees at the centre of vision. Reading means scanning line by line through a small window, and it is slow. Clinicians involved in the trial have described participants reading again, but not quickly.

It is not colour vision. Participants describe the stimulation as monochrome patterns rather than a natural colour image.

It requires training. Using prosthetic vision well takes rehabilitation. People have to learn to interpret the signal and to combine it with the natural peripheral vision they still have. In the trial, prosthetic central vision and natural peripheral vision were perceived at the same time.

It does not stop the disease. Geographic atrophy continues to progress. PRIMA replaces function in the area it covers but does not change the underlying biology.

For someone who can no longer read at all, these trade-offs may be worthwhile. For someone with moderate vision loss, the balance of benefit and risk is very different, which is why eligibility is limited to severe central vision loss.

Status in Europe and the United States

Europe: CE marked, rollout beginning

On 22 July 2026, Science Corporation announced that PRIMA had received a CE mark under the EU Medical Device Regulation. It was issued by the German notified body DEKRA, one of the independent organisations authorised to certify medical devices under EU rules. The CE mark allows the device to be sold in about 30 European countries. It is indicated for severely sight-impaired adults with geographic atrophy due to AMD.

A CE mark is not the same as access. Science has said that the first commercial implantation would take place in Germany, and that country-specific reimbursement applications and activation of clinical sites were under way across Europe. In practice, availability depends on which specialist centres offer the procedure, on training surgeons, and on national decisions about payment. Those processes usually take months and differ widely between countries.

Science has not published a price in its public materials, and this article does not estimate one.

United States: not yet approved

PRIMA is not authorised for sale in the United States. It received FDA Breakthrough Device designation in 2023, which gives a device priority interaction with the agency during development. In July 2026 it received two Humanitarian Use Device designations. A HUD designation recognises a device intended for a small patient population. It is a step towards a Humanitarian Device Exemption, which requires proof of safety and of probable benefit rather than the full effectiveness standard of a conventional approval.

A designation is not an approval. Science must still submit a marketing application, and the FDA must review it. In the US, PRIMA remains an investigational device.

What remains unproven

The NEJM results are strong for a study of this kind, but several questions remain open.

No randomised controls. Because each participant served as their own comparison, the study cannot fully separate the effect of the device from factors such as practice with eye charts or the rehabilitation training. The size of the improvement and the fact that acuity was measured with and without the glasses make a pure practice effect an unlikely explanation. Still, the design is weaker than a randomised trial.

Long-term durability. The main trial reports 12 months. Smaller earlier feasibility studies followed a few patients for up to four years and found that the implant still worked and caused no obvious structural damage to the nearby retina. Those studies were very small. Long-term data from a larger population will come only from continued follow-up and post-market surveillance.

Real-world use. Performance in a trial, with selected patients, experienced surgeons and close support, often differs from routine practice. How well patients use the system at home, how many continue to wear the glasses, and how outcomes vary between surgical centres are questions that the commercial rollout will begin to answer.

Other diseases. Science has said it is exploring PRIMA for Stargardt disease and inherited retinal conditions such as retinitis pigmentosa. These are research directions only. PRIMA's approval covers geographic atrophy due to AMD.

The next generation, and where the physics sets limits

The obvious next step is higher resolution, and the physics shows why it is difficult. Smaller pixels generate weaker currents. The light needed to drive them must stay within safety limits on heating the retina. Stanford researchers have published modelling and preclinical work, in animals and in computer simulations rather than in patients, on three-dimensional electrode designs with pixels as small as 20 micrometres. In theory, that would correspond to about 20/80 on the human retina, five times better than today's device. That result is from engineering analysis and animal research, and is not a clinical outcome.

Science has also said it is developing lighter, more integrated glasses with eye tracking, and larger implants to widen the field of view. Because the implant is wireless, the PRIMAvera investigators noted that it could in principle be replaced by a higher-resolution version, or several chips could be placed side by side. No such procedure has been performed in humans.

Why PRIMA matters beyond ophthalmology

For the neuroprosthetics field, PRIMA's significance goes beyond one eye disease. It shows a neural interface meeting peer-reviewed clinical endpoints in dozens of patients and then passing European regulatory review for real use. That is the path many brain-computer interface companies describe but few have completed.

It also follows the same pattern as the cochlear implant, the most successful neuroprosthesis so far. PRIMA restores a lost function. It uses surviving biology instead of trying to replace it. And it succeeds through incremental engineering and long clinical follow-up rather than dramatic claims. Its limits are clear, its evidence is published, and its availability is only beginning. That combination makes it a meaningful development worth following, not a cure.

Frequently asked questions

Does the PRIMA implant cure macular degeneration?
No. It replaces part of the function of lost photoreceptors in the implanted area. It does not repair the retina or stop geographic atrophy from progressing.

Can PRIMA help wet AMD or early dry AMD?
It was tested and approved for geographic atrophy, the advanced atrophic form of dry AMD, in people with severe central vision loss. It is not a treatment for early AMD or for neovascular ("wet") AMD.

Do you need the glasses all the time?
The implant works only when the glasses project the near-infrared image onto it. Without the glasses, the person has only their natural remaining vision.

Is PRIMA available in the US?
Not yet. It has FDA Breakthrough Device and Humanitarian Use Device designations, but it has not received marketing authorisation.

---

This article is for general information and describes published research and regulatory status as of September 2026. It is not medical advice. Anyone with macular degeneration should discuss treatment options with a qualified ophthalmologist or retina specialist.