Revision Allograft After Failed Primary — The Central Disagreement

→ Half Raymond's doctors recommend a second allograft (Gentile, UCSF Thuillier); half advise against it (Choung, Collman, Salk). The literature is thin but mostly supports the "advise against" camp when read rigorously.

The Only Direct Study: Gaul/Bugbee 2018

This is the single published case series of revision OCA of the ankle — from Bugbee's Scripps group, the world's largest OCA registry [190]:

• 20 patients — still the only dedicated series of revision talar OCA anywhere. Softened August 6, 2026: this page used to call it “the entire ankle revision OCA literature.” That is very nearly true but not exactly: a 14-patient prospective bipolar OCAT registry from the Missouri group contains 2 further revision ankle cases, with 13 of 14 successful at a median 43 months [375]. So the honest phrasing is n=20 in the only dedicated series, plus 2 scattered cases
• 57% reported moderate-to-severe pain even among surviving grafts
• Of failures: 4 went to arthrodesis, 1 to TAR, 1 to amputation
• Corrected August 6, 2026 — and the correction runs in his favour. This line previously read “primary talar OCA shows 89% 10-year survival — revision is roughly 24 percentage points worse,” with no citation attached. The 89% is the five-year number, mislabelled as ten. Bugbee's own primary talar OCA series — same institution, same registry, same era as the revision series above — reports survivorship “88.7% at 5 years and 81.3% at 10 years” [374]. Set against revision's 84% and 65%, the real penalty for being a revision is about 5 points at five years and about 16 at ten — not 24. Revision allograft is still clearly worse than primary, and nothing here makes it a good bet; but this page has been overstating the gap by roughly half, in the direction that argues against the option he is most curious about

Why Primary Allografts Fail (Williams Histopathology Study)

8 retrieved failed talar allografts examined [191]:

• Mean time to failure: 31 months (most failures are early, not late)
• All grafts: subchondral collapse and lucencies
• Histology: substantial loss of glycosaminoglycans, osteocalcin "nearly absent" (non-viable bone)
• CD4+ and CD8+ T-cell infiltrates at the graft-host interface — failure is primarily biologic (immune-mediated), not mechanical

→ Raymond's 12-year pain-free interval is unusually long — most failures happen in the first 3-5 years. This suggests his biology supported initial incorporation, which is favorable but doesn't tell us how a second graft will fare.

Head-to-Head: Revision Options Comparison

Juels systematic review of failed bulk talar allograft revision [192] — attribution corrected August 6, 2026; this page and the reference list both credited it to “Shah,” and there is no author of that name on it. The first author is Juels CA, with So E, Seidenstricker C, Holmes J and Scott RT:

Fusion outperforms revision allograft by ~27 percentage points in this systematic review.

The "Bulk Dead Bone" Problem (Van Dijk Editorial)

Van Dijk's commentary [193] argues:

• Most of the failed graft remains dead bone (per Williams histopathology)
• Bulk grafts compromise future salvage to fusion or TAR
• Each prior procedure adds ~2.5x failure risk per multivariable analysis
• Revision typically requires a LARGER graft (debridement back to bleeding bone)
• Larger grafts (>8 cm²): 2.12x failure risk (Bugbee 2025, n=560 knees) [194]

Salvage After Failed Revision OCA (Gaul 2019)

What happens if revision allograft fails [195]:

• 13 fusion patients: 88% satisfied, AAOS-FAM 83, but 63% reported continued ankle problems
• 11 TAR patients: only 50% satisfied, 40% continued problems
• Both fusion and TAR after failed OCA perform worse than primary fusion/TAR — the joint is compromised regardless of salvage strategy

Immunologic Considerations

• OCA is NOT HLA or ABO matched in clinical practice; chondrocytes are immunoprivileged [196]
• Blood type mismatch study (n=103): no association between mismatch and failure
• No published data on whether prior allograft exposure sensitizes for a second allograft
• The 12-year pain-free interval argues against aggressive immune response in this specific patient

Added August 6, 2026 — the knee has the data the ankle does not, and it points both ways

→ This page has always said the ankle revision evidence is one series of twenty people. That is still true. But revision allograft is done far more often in the knee, and that literature has moved a long way while this page was not watching it. It is a different joint, so treat it as mechanism rather than prognosis — but it is the only place where “what happens when you regraft a failed graft” has been studied at scale.

• The most on-point number anywhere, and it is not encouraging. A prospective registry of 182 primary versus 70 revision osteochondral allografts in the knee found overall revision survival of 71.4% against 79.1% primary at a mean 43 months — a far narrower gap than the ankle numbers above. But the authors subclassified revisions by what had failed first, and the split is stark: revisions after a prior cell-matrix procedure had 100% graft survival, against 58.6% for revisions after a prior osteochondral procedure (P = .049) [391]. He is in the second category. Regrafting after a failed bone-containing graft is the worst-performing revision subgroup they identified. Balanced against that, satisfaction with revision was still 75% and 84.1% said they would undergo it again
• And a study that cuts the other way, recorded deliberately. Fifteen revision OCAs of the femoral condyle matched 1:2 against 30 primary OCAs, mean follow-up 9.3 years, found no significant difference in survivorship — failure 13% in the revision group versus 20% in the primary group — albeit with 53% requiring some reoperation [392]. n=15 is small enough that the absence of a difference may be low power rather than equivalence, and it is the knee. It is here because a page that records only the discouraging studies is not an honest page

How to hold these two together without fooling himself. The registry finding [391] is larger, more directly subclassified, and says his specific situation — regraft after a failed osteochondral graft — is the weak case. The matched study [392] says revision in general is not obviously worse than primary. Both are knee data in a joint that tolerates allograft better than the talus does. The defensible reading is unchanged: a second allograft is a real option with a genuinely worse expected value than a first one, and the honest argument in its favour remains his own twelve-and-a-half-year result rather than anything in the literature.

Added August 8, 2026 — if a second graft ever happens, what goes under it now has Level I evidence, and this page had none of it

→ Whether to do a second allograft is the big question, and nothing today changes it. But there is a smaller question hiding inside it that this page had never asked: if a graft is done, should the surgeon soak it in concentrated bone marrow first? Two randomized trials — the strongest kind of study — now exist on exactly that, and this site carried neither.

How they surfaced is the usual story: an evidence-summary service deposited a DOI on August 7, which pointed at a four-month-old trial, which pointed at a nine-month-old one. Both are knee, both are primary transplants rather than revisions — read them as mechanism, exactly as this page already reads the knee registry data above.

• Barcelona RCT (n=36, AJSM June 2026): BMAC-augmented fresh allografts showed better osseous integration at the host–graft junction at 3 months, with the difference gone by 6 and 12 months, and no patient-reported difference at 2 years [404]. One sobering detail recorded deliberately: at 12 months, graft signal density was better in the non-BMAC group
• Rush RCT (n=36, double-blind with a sham incision, Arthroscopy Nov 2025): equal integration on 6-month CT, a trend toward fewer large cysts — and the number that matters: reoperation for graft debridement or revision was 5.3% with BMAC versus 35.3% without (P = .02), again with no PROM differences [405]. Funding note: supported by a graft vendor

Why this belongs on this page. The Williams histopathology above says allograft failure is a biology-of-incorporation problem — dead bone that never integrates, immune infiltrate at the interface. BMAC augmentation is the one intervention with randomized evidence aimed at precisely that step, it costs one extra draw from the iliac crest during an operation that is happening anyway, and in the Missouri bipolar-OCAT protocol that produced the 2 revision-ankle cases above [375], BMAC-soaking the graft is already standard technique. The honest ceiling on the claim: two small knee trials, primary grafts, neither showing a symptom benefit — what they show is faster early incorporation and, in one, dramatically fewer reoperations. The practical output is a new question for whoever would do the revision — Bugbee, Palanca, or anyone else: do you augment the allograft with BMAC, and if not, why not? It has been added to the specialist question list on the front page.

Most Defensible Framing for Decision-Making

"Revision allograft is the only joint-preserving option but carries roughly 35% chance of outright failure at 10 years and only ~50% chance of being satisfied with the result even if the graft survives. Fusion outperforms it in direct comparison (77% vs 50% satisfactory). The central trade-off: 'one more shot at preserving motion with high uncertainty' vs 'committing to a definitive but non-motion-preserving salvage now with better-documented outcomes.'"