Rebuilding a Failed 4-Unit Anterior PFM Bridge | Case Study

The case at a glance

One sentence summary. The dentist sent us a PFM bridge rebuild on two canine abutments after the previous bridge developed an open margin on #11. We asked for one scan to be redone, built the frame on a noble alloy, layered the porcelain, caught our own mistake on the incisal effect of #8, remade that layer in-house, and shipped. No charge for the remake, because the mistake was ours.

Patient background

The patient is in her early thirties. Years ago she lost her four upper front teeth and had a PFM fixed bridge placed on the two canines — #6 and #11 as abutments, #7 through #10 as pontics. For most of five years she had no complaint. Day to day, the bridge felt fine.

At a routine visit with a new dentist, a bitewing picked up a visible gap between the #11 retainer and the prepared tooth. No pain, no swelling, no decay yet. But the margin was open enough that the dentist read it as "not if, but when." A failed cervical seal on a canine abutment is a slow problem that becomes a fast one — once secondary caries sets in under the retainer, the whole bridge is in trouble.

A few things made this harder than a standard remake:

• The canine on #11 sits with a deep subgingival margin — about 1.5 mm below the gingival crest on the facial. There is a reason this margin was placed there originally (crown lengthening history, root prominence, aesthetic transition). We do not get to change that.
• The two canines are already prepared. You cannot un-prepare a canine. If the rebuild fails, the next step is implants, which is a different conversation. So this bridge has to work on the first fit.
• The patient has no interest in a full-arch zirconia rebuild. She wants her dental bridge PFM replaced with a PFM bridge she can trust, on the teeth she already has.
• She does grind at night, mildly. Not a severe bruxer, but enough that connector geometry matters.

What the dentist asked for

The Rx on this case was short, which is the way we prefer it. Short Rx means the dentist knows which decisions are hers and which are ours.

Written requirements from the dentist:

1. Replace the 4-unit anterior PFM bridge. Keep #6 and #11 as abutments. No implants, no change in design philosophy.
2. Material: PFM, noble alloy. D6751, batch certificate to be sent with the shipment. The dentist chose PFM specifically because the margin on #11 is deep subgingival, which is one of the clinical scenarios where a metal substructure is still a strong default. The published literature backs this on long-term technical performance — see the systematic review summary at PMID 25935732 for metal-ceramic multi-unit FDPs.
3. Margins. #11 — shoulder, subgingival ~1.5 mm, matching the prep that is already there. #6 — chamfer, equigingival to slightly supragingival on the facial.
4. Shade. A2 body, slight translucency in the incisal third, a soft halo on the mesio-incisal edge of #8 to break the midline symmetry.
5. Occlusion. Shallow anterior guidance, no interference in lateral excursion, no steep contact against the lower incisors.
6. Connector area. Keep the connectors generous. The dentist wrote one line here: "patient is a light grinder, don't go minimum."
7. No rush. Standard turnaround. She wanted the case right, not fast.

What was not written. The dentist did not specify metal collar vs porcelain butt margin on #11. She did not specify framework extension to the incisal edge. Those are the two calls where she wanted us to decide and tell her.

That split is the thing — written items are instructions, unwritten ones are permissions to use our judgment. We handle those two categories differently in the design review.

The technical problem we had to solve

Three real problems, in order of risk.

Problem 1. The margin on #11 is the whole case.

A PFM bridge on canines does not fail at the pontics. It fails at the retainer margin. The physics are not subtle. Below ~150 μm of marginal gap, the risk of cervical porcelain cracking rises sharply — a 2017 in-vitro pilot on five-unit implant-supported metal-ceramic FDPs showed that a 150 μm misfit at the distal abutment took the incidence of cervical porcelain cracks from 1 in 10 to 9 in 10 (PMID 27995653, p<0.001). That study used implant-supported frames, not tooth-supported ones, so we read it as directional rather than a direct transfer — the mechanism is the same either way: an open gap leaves a span of porcelain with nothing under it.

Translation. If we cannot read the margin line cleanly in the scan, we cannot build a frame that seats cleanly on the prep. If the frame does not seat, we have already lost this case.

Problem 2. 4 units is not a long span, but it is not a short span either.

Five units is the usual threshold for "long span." We are at four, with canine abutments on both ends — which is biomechanically favorable (long roots, good bone support) but still puts shear and torsion through the two connectors between #7/#8 and #10/#11. For context, a long-term cohort on tooth-supported FDPs reported 20-year survival of 70.8% for short-span versus 52.8% for long-span bridges (PMID 18350953). Our case sits closer to the short-span end — which is good — but the connector cross-section still has to be sized for the real load, not a nominal one.

Problem 3. The patient is a mild grinder, and this is PFM.

Metal-ceramic failure modes for bridges are documented — framework fracture is rare (around 0.6% at 5 years in the systematic review at PMID 25935732), ceramic chipping is the more common one. The dentist flagged the grinding; we have to build the ceramic layer thickness so that it does not become the complication.

Three problems, three design decisions, in that order.

How we rebuilt the PFM bridge

Day 1 — Case intake and the first thing we found

The data came in as a digital case (STL + DCM + bite scan + patient photos). On the day it arrived, our case manager ran the intake check against the standard list. Prep edges, scan completeness, bite registration, shade photos, Rx match. Standard flow (Istar Q&A G081, G121).

Two issues on the first pass:

• The margin line on #11 facial had a shadow across about 0.6 mm of the preparation. The scanner had read the subgingival sulcus as geometry, not as tissue. On the mesh, the finish line disappeared into noise for a short run.
• No alloy certificate on file. The Rx said D6751 but did not specify a supplier preference.

We paused the case and emailed back the same day. For the margin, we asked the dentist to retract the tissue again, use a short cord, re-scan #11 only (one abutment, 20 seconds of chair time), and send the updated STL. For the alloy, we proposed a Bego noble alloy that matches D6751 and committed to attaching the batch certificate with the shipment.

She approved in the WhatsApp case group four hours later. New scan arrived the next morning.

Note on timing. The ~24 hours we spent waiting for the re-scan did not count against the turnaround clock. We restart the clock from the moment a case is complete (G157). This is a policy we document because it matters for how we quote.

Day 2 — CAD design, with the margin drawn by hand

Our CAD designer opened the case and did not trust the software's auto-margin detection. On a subgingival shoulder with any tissue memory, auto-detection will cut corners it should not cut. The designer manually traced the finish line on both abutments and locked it before placing the retainer walls.

Design decisions on this case:

• Framework. Full metal substructure under the retainers; the framework extends to within 1.0 mm of the incisal edge on #7–#10 pontics. This is a trade — a longer framework reduces unsupported porcelain and lowers chipping risk, at the cost of a thin band of less translucent incisal area. We chose this because the patient grinds. The dentist had left this decision to us.
• Metal thickness. 0.4 mm uniform at the retainers, stepping up to 0.5 mm across the connectors.
• Connector cross-section. Sized for the span and the grinder note. Rounded corners at the gingival embrasure (sharp corners initiate cracks in the overlying porcelain). The connector sits at the mid-level of the pontic height, not at the gingival third.
• Margin design on #11. Porcelain butt margin, not a metal collar. The dentist wanted the aesthetic transition clean at the gingival line, and the margin is deep enough that a grey halo would show if we ran metal to the edge. The trade is a slightly less forgiving fit at the margin; we offset that with tighter frame tolerances.
• Margin design on #6. Classic chamfer with a 0.3 mm metal collar on the facial, because the margin is near equigingival and the gingiva is thin — porcelain butt here would be more fragile with no benefit.

Design preview went to the dentist in the case group. She approved with one minor comment on the mesial contour of #8. Approved design locked.

Day 3 — Frame casting, oxidation, first porcelain bake

Bego noble alloy, cast on an induction furnace in our ceramic & metal department (70 technicians across this line, Istar company profile). After casting:

• Finished the frame on the stone model. Checked internal fit under a loupe. No high spots on either retainer.
• Air-abraded with 50 μm alumina at low pressure, ultrasonic cleaned, controlled oxidation firing in a vacuum porcelain furnace.
• Verified the CTE match between the Bego alloy and the Ivoclar porcelain layering system — target difference 0.5–1.0 × 10⁻⁶/°C. Porcelain picked for compatibility, not substituted.
• Opaque layer applied in two passes, slow cool after each. Fired.
• Dentine layer in a single pass. Fired.

So far so good. Frame seats, opaque is uniform, dentine is at target shape.

Day 4 — Incisal and effect layering, and the step we got wrong

This is the step where we made the mistake.

The dentist asked for a soft halo on the mesio-incisal of #8 to break the midline. Our ceramist is a seven-year technician — not junior, but not one of our top-tier anterior specialists either (we assign those to cases flagged as high-aesthetic, Istar Q&A G129, P151). On the first incisal bake, the opal and translucent layer on #8 went in a shade thinner than specification. Under daylight the halo was there, but soft enough that against #9 the contrast read as uneven, not as a feature. Side by side, #8 looked flatter than its mirror.

Our final-stage QC caught it before the glaze fire.

Specifically — this is from the QC report on the case — the inspector pulled the bridge under the standardized light box, held it next to the dentist's shade photograph, and flagged "incisal translucency on #8 reads under target; mirror asymmetry visible from frontal view." She stopped the case.

What we did:

• Did not glaze. If we had glazed, the fix would have required stripping back more porcelain.
• Returned the piece to the same ceramist with the QC note and the shade photograph side by side.
• Reground the incisal of #8 by about 0.6 mm, re-layered the opal and translucent in the correct volume, re-fired.
• Second QC pass with the same inspector. Approved.

Who paid for this. We did. Not the clinic. This was not a case of the Rx being unclear or the data being wrong. The dentist wrote what she wanted, we read it correctly, and the ceramist still under-shot the layer. On the policy side, we absorb the cost of a remake when the mistake is on our side. This one lived in that category.

What it added to the timeline. One working day. The case shipped on day 5 instead of day 4.

What we changed afterwards. We added a photographic side-by-side at the incisal layering step for every anterior PFM bridge on canine abutments going forward. The ceramist does this himself before firing. It is not a QC redundancy — it is a design review the technician does on his own work before the inspector gets near it.

Day 5 — Glaze, final QC, pack, ship

Final glaze fire. Micro-polishing on the facial surfaces of pontics #7–#10 to match the surface texture of the adjacent canines (patient's natural canines read as slightly glossy under the shade photographs — we matched that). Final QC against the standard checklist:

• Marginal fit on both retainers — model-check with loupe, no visible gap at the finish line.
• Internal surface — no bumps, no pits, no leftover investment.
• Interproximal contacts — snug but passes shim stock on the mesial of #6 and distal of #11.
• Occlusion — no premature contacts in centric; smooth guidance on lateral excursion with no interference.
• Shade match — side-by-side against the dentist's photograph under the standardized light box.
• Surface finish — passes under 10× magnification.

Packed in an individual foam-strip-lined box. In the package:

• The bridge.
• A try-in paste sample.
• The Bego alloy batch certificate.
• A single-page QC photo sheet with the final frontal and incisal shots.
• One sentence note from the technician: "incisal of #8 was remade in-house for shade; glad we caught it."

Shipped same day via the dentist's preferred carrier.

What the final PFM fixed bridge had to deliver

The dentist's acceptance is in her chair. We cannot promise what the clinical fit will do on the day of cementation — that is between her and the patient. What we can promise is that the piece we shipped meets a set of lab-side conditions.

A quick note on the published numbers for anyone reading this and asking "how long will it last." Honest answer: it depends on the patient and on the clinical maintenance. The best current pooled data on metal-ceramic bridges puts 5-year survival in the 91–95% band depending on which review you read (PMID 25935732, PMID 41650383), with technical complications more frequent than catastrophic failures. We do not quote "30 years" to clinics, because that is the tail of the distribution, not the expected value. If a lab tells you 30 years as a flat number, ask them which cohort.

Why PFM still won this case

A few clinics ask us whether a 4-unit anterior case like this should have been monolithic zirconia instead. Fair question. We would answer it the way the dentist did on her Rx.

• The margin is deep subgingival. When a canine prep sits 1.5 mm below the crest, we want a substructure that is forgiving in that zone — and a cast metal frame reads a cleanly-scanned finish line more predictably than a ground zirconia inner surface in our hands.
• The abutments are already prepared. We are not re-prepping canines to accommodate a different material. The thicknesses that are available are the thicknesses that are available. PFM gives us a thinner retainer wall at the cervical than monolithic zirconia does, without sacrificing structural integrity, because the metal does the structural work.
• 5-year systematic review data for tooth-supported multi-unit metal-ceramic FDPs is strong — pooled survival of 94.4% in the 2015 review (PMID 25935732) and 91.3% in the 2026 update (PMID 41650383). Head-to-head meta-analysis of PFM vs zirconia for posterior multi-unit bridges reported 5-year survival of 96.9% vs 95.4% (PMID 32222175, p=0.364) — practically tied. The right material depends on the case, not the trend.
• The patient is a mild grinder. For a mild grinder on a 4-unit anterior, the framework-supported ceramic layer of a well-made PFM fixed bridge is a predictable choice. We back it with thicker opaque and a connector sized above minimum.

Zirconia is a good material. It is not the right material for every case. On this one, dental bridge PFM was the dentist's call, and after we built it, we agree.

What a clinic can use this case for

A few things you can take from this if you are considering sending us a PFM bridge case.

1. We will tell you when a scan needs to be redone. On this case we paused the file for one abutment. We did not try to build around a shadowed margin and hope for the best. If you send us a scan with a cervical read-error, you will hear from us the same working day, not after we have milled a frame that will not seat.

2. We will tell you when we got something wrong. The incisal of #8 was remade in-house on this case. The clinic was not billed. We did not fold it into the next invoice. The remake is in the production record, with the technician's name and the QC inspector's note attached — those records are retained. You can ask to see them on your own cases.

3. We ship the alloy certificate with the shipment. Every PFM case gets a batch cert, with the alloy class (noble / high noble / predominantly base), the brand (Bego on this case), and the batch number. If a question ever comes up about what alloy class is actually in your patient's mouth, the answer is on paper.

4. Standard lab production is 2–3 working days for a fixed case; this one took 5 because of the data re-request and the in-house remake. We do not count waiting-on-you time against the turnaround clock. Rush is available and is not charged under 10 cases per month.

5. We are a direct manufacturing lab. Not a middleman. The CAD design, the frame casting, the porcelain layering, the QC, and the packing all happened under one roof at our Shenzhen or Huizhou facilities (Istar company profile). The technician on #8's incisal is one of our seven-year ceramists in the Ceramic & Metal department. If you want to meet him, you can — we take factory visits and video audits.

6. We make mistakes. We catch them in-house. This case had one error, and our internal QC caught it before the piece left the door. That is the system working. The reason we are documenting it on a case page is not that it is unusual — it is that it is useful information for a clinic deciding where to send its next pfm bridge case, and we would rather you have it up front.

If any of the above matches what your practice needs from a lab, send us a case. One case is enough to see what we do. If the first one lands well, we scale from there.

About the lab

Istar Dental Lab. Founded 2012. Full-service dental manufacturing facility in Shenzhen and Huizhou, China. Around 460 in-house staff including a 70-strong Ceramic & Metal department, a dedicated 25-person QC team, and a 40-person case coordination group that handles the clinical back-and-forth. FDA listed, MDR notified, MHRA registered, ISO 13485 certified. Serving over 100 clinics, 400+ dentists, and 50+ dental laboratories across the United States and Europe. Full in-house workflow — no outsourcing of your case to a third-party production facility.

For a factory visit, a sample case, or a quote, contact our sales team.