Residual caries

Residual caries is demineralised carious tissue that was left in place when a restoration was placed. Secondary caries is a lesion that developed next to the restoration afterwards. The difference is timing, not appearance — on a bitewing the two can look identical, and no radiographic test reliably separates them.

Synonyms: retained caries; caries left under a restoration; residual decay

What is residual caries?

Cariology has an agreed vocabulary for this. Primary caries is a lesion in a previously sound tooth surface. Secondary — or recurrent — caries is a lesion that has developed adjacent to a restoration. Residual caries is demineralised carious tissue left in place before the restoration was placed. Those definitions come from a consensus workshop run by ORCA and the IADR Cariology Research Group, and they are the definitions this entry uses.[1][2]

Read them again and notice what they hang on. Not depth. Not shape. Not position. When the tissue became demineralised, relative to the moment the restoration went in. Residual caries was already there and stayed. Secondary caries arrived later, associated with the restoration or sealant that was already in place.[3]

That is a clean distinction on paper and a messy one at the chairside. A 2022 review of recurrent and residual caries around resin composites put it bluntly: an exact distinction between new recurrent caries and residual caries beneath an old restoration is missing, and the lack of a clear distinction is worst for wall lesions.[1] The same review notes that clinical discrimination between the two is impractical, even though reports and reviews continue to treat them as separate things.[1]

So the terminology is settled and the diagnosis is not. Worth knowing which of those two problems you're actually facing.

What does residual caries look like on a bitewing?

By definition it sits beneath the restoration — it's the tissue the bur didn't take, sealed in when the filling went in. What you see is a radiolucency at or under the restoration–dentine interface.

Here is the honest part: that description also fits a secondary wall lesion. Secondary caries occurs either as a surface lesion at or near the restoration margin, or internally as a wall lesion at the tooth–restoration interface.[1] The second of those is the same place residual caries lives. There's no border character, no shape, no density gradient that separates them, and the literature doesn't offer one.

You also aren't looking through a clear window. A systematic review of 23 studies of secondary caries detection found that visual, radiographic and laser fluorescence assessment all performed similarly, with sensitivities of 0.50 to 0.59 and specificities of 0.78 to 0.83.[4] Most of those studies were in vitro, on permanent posterior teeth, 14 of 23 looked at proximal surfaces and 16 of 23 at lesions adjacent to amalgam — and the review flagged high risk of bias or applicability concerns in most of them, plus probable publication bias.[4] Treat that band as an order of magnitude, not a number to quote at a viva.

One thing position does tell you. The greatest risk of secondary caries is at the gingival margins of restorations, irrespective of restoration type, and that risk climbs as the margin extends further apically from enamel toward the CEJ and cementum.[1] A radiolucency sitting right at a deep gingival margin is in the highest-risk location for a new lesion. That shifts the odds. It doesn't settle the question.

What is residual caries confused with?

Restoration overhangs have an entry of their own. Working through the mimics in a fixed order is what the bitewing caries reading sequence is for.

Why does residual caries happen?

Mostly on purpose.

The International Caries Consensus Collaboration is explicit that carious tissue is removed purely to create conditions for a long-lasting restoration, and that bacterially contaminated or demineralised tissue close to the pulp does not need to be removed.[9] Non-selective removal to hard dentine — what used to be taught as complete caries removal — is no longer recommended.[10] For shallow or moderately deep cavitated lesions the recommendation is selective removal to firm dentine; for deep cavitated lesions, selective removal to soft dentine, with stepwise removal as an option in permanent teeth.[9]

Follow that guidance and you create residual caries deliberately, every time you restore a deep lesion. That is the intended outcome, not a slip.

It isn't universal practice, though. In a 2016 review, across nine studies surveying 4,199 dentists, the pooled proportion using selective or stepwise removal for deep lesions was 53% (95% CI 44% to 62%).[11] So as of 2016 roughly half of dentists were working to the newer strategy and half were not — and the radiograph won't tell you which of them placed the restoration in front of you.

What does residual caries mean clinically?

The question behind the question is usually: does leaving it matter?

For pulp preservation, the evidence is consistent and it favours leaving tissue behind. A Cochrane review found stepwise removal reduced the incidence of pulp exposure by 56% compared with complete removal (RR 0.44, 95% CI 0.33 to 0.60), with mean incidence of 34.7% in the complete-removal groups against 15.4% in the stepwise groups. Partial removal reduced it by 77% (RR 0.23, 95% CI 0.08 to 0.69) — 21.9% against 5%.[12] A separate meta-analysis of 10 studies and 1,257 patients found reduced odds of pulpal exposure with incomplete excavation (OR 0.31, 95% CI 0.19 to 0.49), and no clear difference in failure (OR 0.97, 95% CI 0.64 to 1.46).[13]

For overall success, the evidence is thinner and more equivocal than the pulp-exposure figures suggest. A 2021 Cochrane review of deep lesions in permanent teeth found higher odds of failure for conventional removal than stepwise (OR 2.06, 95% CI 1.34 to 3.17, moderate-certainty evidence) and than selective removal (OR 11.32, 95% CI 1.97 to 65.02) — but that second estimate is very low-certainty and its confidence interval runs from 1.97 to 65.02, which is another way of saying nobody knows.[14] A 2025 network meta-analysis ranked stepwise (RR 1.11, 95% CI 1.00 to 1.23) and selective removal (RR 1.06, 95% CI 0.97 to 1.16) above non-selective removal for success, but the differences were not statistically significant, more than half the included studies were at high risk of bias, and the authors state directly that the effect of remaining carious dentine on success could not be assessed.[15]

So: leaving demineralised dentine reliably protects the pulp, and does not clearly help or harm the restoration.

The mechanism is worth holding onto. There is little evidence that infected dentine must be removed before sealing the tooth, and leaving it does not seem to result in caries progression, pulpitis or pulp death — though some bacteria do survive.[16] Sealing infected dentine within the tooth, by partial removal under a well-sealed filling, arrests the lesion.[17]

One line in that review should change how you read these radiographs: the presence of radiolucent zones beneath composite restorations is not considered a justification for replacing the restoration.[1] Residual or remaining dentine caries beneath an existing restoration is not treated as a failure, because a sound peripheral seal is the more clinically relevant thing.[1]

Getting this wrong is expensive. Across 718 restorations in 185 adults, assessment using FDI criteria indicated five times more replacements than assessment using CARS criteria.[18] The criteria you apply, not the tooth, drove the treatment decision.

How can you be sure?

You can't, from one radiograph. That's the honest answer.

What the imaging gives you is a band of moderate performance — sensitivity around 0.50 to 0.59 across visual, radiographic and laser fluorescence methods in that systematic review, with the caveats above.[4] The same 2022 review reports sensitivity and specificity for radiolucent lines beneath a resin composite restoration at less than 80%.[1] Detection methods for secondary caries are only sparsely validated and are likely prone to over-detection, and it may be prudent to prioritise specific methods rather than sensitive ones, to avoid invasive overtreatment.[3]

What actually separates residual from secondary caries follows from the definitions rather than from any test: residual caries is defined by when the tissue was left, so the thing that identifies it is the record of what was done at placement — the notes, the material used, the removal strategy. If a deep lesion was managed by selective removal to soft dentine, a radiolucency under the body of that restoration is the expected finding. That is reasoning from the definition, not an evidence-based diagnostic protocol, and no retrieved study tests it.

The practical move is to stop trying to win the argument from the image and ask what the answer would change. If the seal is sound, the margins are intact and the patient is symptomless, the distinction rarely changes the plan.

Common errors

• Reading any radiolucency under a composite as a reason to replace. It isn't.[1] The criteria you apply change the replacement rate fivefold.[18]
• Using position as the discriminator. "At the margin means secondary, under the body means residual" fails exactly where it matters, because secondary caries occurs as a wall lesion at the tooth–restoration interface too.[1]
• Forgetting the material. A low-radiopacity liner or an unfilled adhesive can read as a lesion, because not every material is more radiopaque than dentine.[5][6]
• Quoting the detection figures as if they were chairside performance. Most of the underlying studies were in vitro and at high risk of bias.[4]
• Hunting for certainty a single radiograph can't supply. Detection methods here are sparsely validated and prone to over-detection.[3]