Restoration overhang

A restoration overhang is an extension of restorative material beyond the confines of the cavity preparation. On a bitewing radiograph it reads as a radiopaque projection of filling material extending past the radiographic outline of the cavity margin into the proximal space. Periodontal guidance treats it as a plaque-retentive factor to be corrected.

Synonyms: overhanging restoration; overhanging dental restoration; marginal overhang; ledge

What is a restoration overhang?

The definition has been stable for decades: an extension of restorative material beyond the confines of a cavity preparation [1]. That's it. Not a badly shaped filling, not a rough margin — material that has gone somewhere the preparation didn't.

What makes it worth a whole entry is that it sits in two categories at once. It's a finding you report, and it's a plaque-retentive factor that periodontal guidance tells you to remove [2, 3]. Most of what you look for on a bitewing is a consequence of something. This one is a cause you can see.

The mechanism was demonstrated deliberately. In a 1983 cross-over experiment, nine dental students with clinically healthy gingivae had gold onlays with 1 mm proximal overhanging margins placed in mandibular molars for 19 to 27 weeks. The subgingival flora shifted to one that "closely resembled that of chronic periodontitis," with increased Gram-negative anaerobes and black-pigmented Bacteroides. When the same teeth carried clinically perfect margins instead, the flora looked like health or early gingivitis [4]. It's a small, old, film-era study — but it's an experiment, and almost everything since has been observational.

What it looks like on a bitewing

Look at the margin, not the body of the restoration.

A restoration counts as having an overhang when the restorative material extends beyond the radiographic outline of the cavity margin, forming a radiopaque projection into the proximal space [5]. So you're tracing one line: the outline of the preparation. Where the material stops following that line and juts out past it, that's the ledge.

Two calibrated examiners applying that criterion to bitewings agreed almost completely (Cohen's kappa = 0.90) [5]. Once the rule is explicit, it's a repeatable call.

Where to look. Molars, mostly. Millar and Blake found overhang prevalence greatest on upper molar teeth in a UK general-practice audit [6]. A fractal-analysis study of periapical images found them most commonly in the maxillary molar regions and in disto-occlusal cavity types [7], and a panoramic study found them more frequent in molars than premolars, and on distal than mesial surfaces [8]. Surface preference is less settled than the tooth preference: the bitewing study of student-placed composites found mesial surfaces slightly ahead of distal (27.9% versus 22%), and the difference wasn't significant (p = 0.529) [5]. Treat "molars first" as the useful heuristic and check both surfaces.

How big. In that general-practice audit the mean overhang measured 0.9 mm, with a range of 0.4 to 2.0 mm [6]. You are looking for something under a millimetre wide most of the time.

Prevalence varies enormously, and the variation is informative rather than confusing. Reported figures include 14.1% of filled interdental sites in Swiss Army recruits [9], 25% of restored approximal surfaces in young periodontal referrals [10], 25% of Class II composites placed by undergraduates [5], 32.4% of approximal restored surfaces on CBCT [11], and 56% of restored surfaces placed by final-year dental students [12]. Who placed the restoration, who was sampled, and when, all move the number. Quality has improved over time: in Jönköping, 85–90% of restorations had no proximal overhangs by 1993, against roughly 60% in 1973 [13].

What it's confused with

Density alone will not separate the first row. It is tempting to reduce this to "radiopaque means restoration, radiolucent means caries." The literature doesn't support that shortcut: in an in-vitro study validating bitewings against sectioned teeth, both radiolucencies and radio-opacities adjacent to the restoration were scored as caries [14]. Use continuity and position — is this material carrying on from the filling, and does it cross the cavity outline?

This is also genuinely hard, and the literature says so. Among the criteria examiners use when deciding whether to replace a restoration, deficient anatomic form and marginal overhangs produced the most frequent diagnostic difficulties and the most disagreement between them [15]. That study looked at Class I restorations, so it doesn't transfer cleanly to proximal surfaces — but it's a fair warning about which call is slippery.

Two more calibration points. In vitro, marginal overhangs were significantly easier to diagnose radiographically than marginal openings (ROC area 0.90 versus 0.63) [16] — so if you find gaps harder than ledges, that matches the evidence. And bitewing sensitivity for secondary caries under the approximal parts of Class II amalgams was 73% for molars and 80% for premolars in vitro, with specificity 90% and 95% [14]. In-vitro figures are a ceiling, not chairside performance.

What it means clinically

Periodontally, the association is consistent and the mechanism is not settled. The 2017 World Workshop consensus states that infringement of restorative margins within the supracrestal connective tissue attachment is associated with inflammation and/or loss of periodontal supporting tissue, and adds that it is not evident whether the negative effects are caused by dental plaque biofilm, trauma, toxicity of dental materials, or a combination [17]. The background review for that workshop puts it the same way: restorations "have often been associated with plaque retention and loss of attachment," and the evidence on biologic mechanisms "is not conclusive" [18].

Sites next to overhangs probe deeper and bleed more. In one study of restorations placed by final-year students, 64.3% of pockets adjacent to overhanging margins exceeded 3 mm, against 23.1% at unrestored surfaces and 49.2% at restorations without overhangs; 32% bled on probing, against 10.5% at unrestored surfaces [12]. In a periodontal referral population, pockets at overhang sites were 0.42 mm deeper than at restorations without overhangs [19].

Bone loss: the direction is consistent, the size of the effect is contested. Millar and Blake, using sequential bitewings over up to 25 years, measured mean bone loss of 0.16 mm/year adjacent to overhangs against 0.06 mm/year at unrestored control teeth (P = 0.01) [6]. On CBCT, 71.3% of overhanging surfaces showed bone loss against 49.1% of controls, with 2.28 mm of loss against 1.53 mm [11]. But the same CBCT study reports that the same trend applied to restored surfaces without overhangs and their controls [11] — which is a strong argument that the restoration, not only the ledge, carries the risk.

And there is a clear negative study. In 20-to-29-year-olds referred for periodontal treatment, overhangs were common but "were not associated with increased levels of bone destruction" [10]. Age is the likely explanation rather than contradiction: that cohort hadn't had time to accumulate loss, while the positive result came from up to 25 years of radiographs.

Whether size matters is unresolved. Jeffcoat and Howell found small overhangs, unlike medium and large ones, did not increase bone loss [20]; a 2021 review concludes that inflammation and bone loss occur, for both direct and indirect restorations, only with large overhangs [21]; Millar and Blake found bigger overhangs associated with greater loss, while noting "not all ledges were associated in bone loss" [6]. Against that, one study found overhang width did not affect the result [22], and the CBCT study found bone loss not correlated with overhang size [11]. The studies don't share a definition of "large," which is most of why they disagree.

Cariologically, the honest answer is that the evidence isn't there yet. Overhangs and secondary caries are both common and they turn up in the same mouths — in 502 radiographic records of patients with proximal fillings, more than 30% of individuals had overhanging restorations and more than 30% had secondary caries lesions [23]. But that study reports co-prevalence, not an association between the two, and no retrieved study demonstrates that an overhang raises the risk of caries at that margin. If you've been taught that overhangs cause recurrent caries, note that the periodontal case is well documented and the cariological one is currently an assumption.

How to be sure

• Use a bitewing, not a panoramic. Bitewings were significantly more effective than panoramic radiographs at detecting overhanging restorations, secondary caries, interproximal caries and caries depth (p < 0.001) — though that comparison came from only 11 patients [24].
• Don't rely on the radiograph alone. Many overhangs "are not detected on radiographs and are evident only by use of an explorer directed subgingivally" [1].
• Be careful with laser fluorescence at these margins. Stain at restoration margins, especially combined with filling overhangs, interfered with laser-fluorescence readings and produced false-positive measurements in vitro [25].
• Expect automated detection to help unevenly. A single-purpose model reached 90.9% precision and 85.3% sensitivity for overhangs on bitewings [26]. In an eight-class charting model, the overhanging-filling class scored F1 0.35, with secondary caries at 0.29 [27]. Same finding, different task — narrow tools do better than broad ones here.

Common errors

• Reading the body of the filling instead of its margin. The diagnosis lives entirely at the cavity outline.
• Calling density instead of continuity. Both radiolucencies and radio-opacities next to a restoration have been scored as caries in validation work [14].
• Treating a small ledge as automatically significant. Not all ledges are associated with bone loss [6], and the modern review position is that the periodontal cost attaches to large overhangs [21].
• Treating a large ledge as automatically causal. The consensus wording is "associated with," and the mechanism is explicitly unsettled [17].
• Assuming the caries link is established. It isn't, on the retrieved evidence [23].
• Stopping at detection. Correcting plaque-retentive factors — "remove overhanging restorations or alter denture design" — is a named component of supportive periodontal care [3], and removing overhangs improves the hygienic phase of periodontal therapy [1]. In a randomised split-mouth trial, recontouring 26 paired defective amalgams produced significant differences from the untreated side in plaque, gingival inflammation and pocket depth, with the greatest change in the first four weeks [28].