A Phosphoserine Bone Adhesive in Sheep Spines: The 2026 Tetranite Lumbar Fusion Study (Foley et al.)

In July 2026 the International Journal of Spine Surgery published a year-long sheep study of a bone adhesive. The first author is Kevin T. Foley, and the co-authors are Eric J. Woodard, Jonathan R. Slotkin, Ben Gadomski, Caroline D. Woodard, Michael C. Brown and Brian J. Hess. The material is tetracalcium phosphate-phosphoserine (TTCP-PS), sold as Tetranite Bone Adhesive by RevBio, Inc. It is one of the phosphoserine cements that the 2023 review of the field singled out for high bench adhesion strength (see What a Real Bone Glue Would Have to Do: The 2023 Bench-to-Bedside Review of Bone Adhesives (Bingol et al.)). The paper does not mention Bone-02, and it gives no reason to think the two products share a chemistry. It matters to this dossier for a different reason: it shows what a peer-reviewed animal study of a bone glue looks like, and how much such a study can and cannot settle.

What the material is said to be and do

The authors describe TTCP-PS as a mineral-organic material that sets by itself in a wet field. They say it is osteoconductive (new bone can grow along it) and bioresorbable. In this study the powder was tetracalcium phosphate and phosphoserine plus porous beta-tricalcium phosphate granules and poly(lactic-co-glycolic) acid fibres. It was sterilised by gamma irradiation and mixed with water into a putty. The paper says this putty could be worked for about 2 minutes and hardened within 10 minutes into a scaffold with 35% porosity. Surgeons reported that it handled as intended and gave off no heat as it set. Rapid setting in a wet field is one of the review's core requirements, and heat release is one of its complaints about cyanoacrylates.

The authors offer two explanations for any benefit. One is chemical: tetracalcium phosphate cements turn into hydroxyapatite, the mineral of bone, as they set. The other is mechanical. A graft that sticks to the bone stays pressed against it and may not drift away, and the authors cite earlier sheep work linking graft migration with lower fusion rates. Both explanations are offered as interpretation. The study did not test either one directly.

How the study was designed

Posterolateral lumbar fusion joins two vertebrae with screws and rods, plus graft material laid between the bony side projections (transverse processes) so that bone grows across. The paper explains that synthetic graft substitutes can resorb before a solid fusion forms, and the fusion then fails (nonunion).

Twenty healthy female sheep (53–73 kg, older than 2 years) had one level fused, at L3–L4 or L4–L5, with Stryker Xia pedicle screws and titanium rods. TTCP-PS went on one side of the spine and the control went on the other. The side was chosen by drawing from a hat, so each sheep served as its own control. The control was Actifuse ABX, a marketed graft substitute made of silicate-substituted hydroxyapatite granules. Six sheep were assigned to each of three end points (12, 26 and 52 weeks), with 2 spares. The work was done at Biomatech, an independent laboratory in France owned by NAMSA, under Good Laboratory Practice rules. The authors note one exception: the CT scans did not follow those rules. Outcomes were x-ray scores, CT density and volume, gross inspection, independent histopathology graded under ISO 10993-6, and histomorphometry. Histomorphometry measures what share of each tissue section is bone, fibrous tissue, leftover implant or empty space.

What was measured

The headline numbers come from histomorphometry. The share of the fusion region that was new bone was:

Only the 26-week difference was statistically significant. At that point the control sites were still almost half fibrous tissue (about 49%), against 15.6% for TTCP-PS (P < 0.001). At 12 weeks more TTCP-PS was left in place (40.4% vs 24.8%, P = 0.007). Both materials then lost roughly half of what remained by 26 weeks. At 52 weeks the differences in leftover implant, fibrous tissue and empty space were all non-significant.

On x-rays scored from 0 (large gaps in the bridge) to 2 (no gaps), both sides scored 0.0 at 2 weeks. At 10 weeks the scores were 0.7 for TTCP-PS and 0.2 for the control. From 26 weeks most sites scored 1.7–2.0, but half of the control sites still showed small gaps at 26 or 50 weeks. CT measured a larger fusion mass for TTCP-PS (6.05 ± 0.52 against 3.68 ± 0.53 mm³ across all time points, P < 0.01). The authors concede that on CT the adhesive was hard to tell apart from surrounding bone, because the two have similar density, and that the outlines were traced by hand. Lymph nodes showed no damage, and the liver and abdominal masses seen in a few sheep were put down to parasites and unrelated causes.

What the evidence does and does not show

The paper concludes that TTCP-PS was both safe and effective in this model, and that it produced "greater levels of bone formation at 1 year". The table above qualifies the second phrase. The one-year average was higher, but the difference was not significant, and by then the control had largely caught up. The supportable reading is narrower: in this model the adhesive was associated with faster early bone growth across the gap and less fibrous tissue at 26 weeks.

Several limits come from the paper itself. Each material had about six implant sites per time point. The animals were sheep, and the authors say human use needs further study. Most important for a dossier about bone glue, the authors state that the material's load-bearing and adhesive properties "were not the subject of this investigation". Screws and rods provided the fixation, and the adhesive served as a graft substitute. The study therefore says nothing about whether glue alone can hold bone together. The authors also leave open whether the material could reduce the hardware needed, saying this remains to be studied.

Funding and conflicts

Stryker funded the study. The disclosures are extensive. Two authors are RevBio employees who own RevBio stock, and one of them also owns Stryker stock. Foley holds patents with RevBio, sits on its board and owns its stock. Eric Woodard and Slotkin hold RevBio stock and sit on its advisory board. Gadomski received research funding from RevBio for this work, and only Caroline Woodard reports no conflicts. Using an outside GLP laboratory and an independent histopathologist reduces, but does not remove, the need to read the conclusions with these ties in mind.

How it connects to the dossier

The Bone-02 coverage offers press figures with no study design (see Bone Glue Claims vs. Evidence: Press Numbers, Patent Abstracts and the Absence of Primary Data). This paper sets out exactly the details those reports leave out: a named control, randomised sides, stated sample sizes, prespecified time points, statistics and declared conflicts. It also shows how modest the result of even a well-documented animal study can be. For the general idea of a bone glue, see Bone Adhesive (Bio-Glue).