Secondary caries: when an old filling really needs replacing, and when it can be left alone

What decides the fate of an old filling

• Intervention is justified by a cavity reaching the dentine, a fractured wall of the tooth or caries under the filling confirmed by combined clinical and radiographic assessment; the choice between repair and replacement is decided by the extent of the lesion.
• Secondary caries is a cavity at the margin of a filling, not a dark rim around it. Old decay under a filling does not grow back - a new cavity starts at the junction of tooth and material.
• None of the detection methods studied reaches satisfactory accuracy: a radiograph picks up roughly six cavities in ten, and a probe fewer still.
• The way old fillings are assessed changes the number of procedures: criteria that take staining and small marginal gaps into account produced five times as many indications for replacement as criteria looking at the cavity alone.
• Those extra procedures did not extend the life of the fillings. In a separate trial with random allocation, after not quite six years, the proportion of fillings that had failed was similar in both groups.
• Repairing only the damaged part extends the life of a filling much as replacing the whole does, and it costs less sound tooth.

A dozen or so years after a filling is placed, something happens that nobody warned the patient about: the margin darkens while the filling holds firm. The check-up radiograph shows five fillings of that age. None hurts, none gets in the way when biting, none has changed colour for a long time. Only the margin has changed: at two of them it is darker than the rest of the tooth.

The dentist then says that these fillings ought to be replaced, and the patient usually falls asks: are we replacing all of them?

Every reopening of a tooth takes away a little sound tissue, and that cannot be put back. So more than one appointment is at stake. Secondary caries is sometimes the reason a replacement is needed. It is also sometimes a diagnosis made when nothing bad is happening under the filling.

Where does secondary caries under a filling come from?

Not from the old decay, because that does not grow back.

Secondary caries is a cavity at the margin of existing work: a filling, a crown or a sealant, that is the material that closes off the fissures. The mechanism is the same as in the successive stages of tooth decay on an untouched tooth: bacteria in plaque turn sugars into acids, and the acids leach minerals out of the enamel. In the dentine, the tissue beneath the enamel, progress speeds up. The material itself does not fall ill. Whether a cavity forms is decided by circumstances rather than by the substance. Three are usually weighed: the gap between tooth and filling and how wide it is, the patient's own caries risk, and the skill of the dentist who placed that filling. That is how they are summarised by a review prepared for a joint statement of two European dental societies. The authors add a caveat rarely met in textbooks: secondary caries is described surprisingly sparsely for a complication regarded as the most common one.

The annual proportion of fillings that failed ranged between centres from 0.08 to 6.3 per cent, that is from fewer than one filling in a thousand a year to six in a hundred, and that came out of 33 studies in which fillings were followed for at least five years. An average from studies like these says little about one particular tooth.

Why the margin of a filling is its weakest point

Because it is a border between two substances, and over time it works and darkens. The darkening of the margin is a consequence of that working. In itself it does not prove that a cavity sits under the filling. Composite, the white filling material, shrinks slightly as it sets. On top of that it takes chewing forces for a dozen or so years, until a gap appears at the junction of material and tooth - invisible to the naked eye, but wide enough for bacteria. At this stage a leaking filling is a place where plaque stays longer. It is not yet disease - secondary caries merely has a chance of starting there.

A stained margin agreed with the assessment of caries most weakly of all three features studied; marginal adaptation came out in the middle, and the cavity itself best. In all, 718 fillings in 185 patients were assessed. The authors write plainly that marginal staining is no longer regarded as an indicator of caries, although in the surgery it is still sometimes read that way. Staining was not assessed at amalgam fillings, that is the older metal material, because amalgam itself stains the tooth, and such fillings made up more than two fifths of this group.

A molar works like a vault: four sides, the walls, support the chewing surface with its raised points, the cusps. The more walls decay has taken, the fewer supports are left, and extensive fillings fail more often than small ones. The risk rose markedly for fillings involving three or more surfaces of the tooth and in patients over 60; this was calculated on 900 fillings in posterior teeth placed by dental students, from records covering twelve years. How much tooth is left weighs more in this calculation than the choice of restorative material.

How is a cavity under an old filling recognised?

Not by the filling itself. A cavity gives itself away through a change in the tooth tissue at the margin or under it, and it hides beneath the material, so no instrument sees it well.

An intraoral radiograph detects roughly six cavities out of ten that are present. In eight fillings out of ten with no cavity underneath, it correctly finds none. A probe detects three in ten. That is how a comparison of seven methods came out, drawn from 25 clinical and laboratory studies of detecting caries around fillings in posterior teeth. The authors' conclusion is unambiguous: none of the methods studied reaches satisfactory accuracy.

A negative radiograph does not rule out a cavity, and one suspicious feature is not enough to open a tooth. The diagnosis is made up of several observations at once:

• whether the surface at the margin is dull and soft;
• whether the lesion has grown since the previous visit;
• how high the patient's caries risk is;
• what can be seen on the bitewing radiograph, that is the one showing the crowns of several teeth in both arches at once and revealing best the surfaces that contact the neighbouring tooth.

This set of observations is gathered at a check-up that includes an inspection of the fillings' margins.

The scale of diagnostic error has been measured.

A cavity was confirmed under 80 of 120 composite fillings that had previously been listed for repair or replacement; the state of the dentine was checked once the material had been removed. Under the remaining 40 the dentine was sound. The group studied was, however, selected in advance as problematic, so this figure must not be carried over to all old fillings.

When secondary caries calls for replacing the filling

Opening the tooth is justified by a cavity reaching the dentine, a fractured wall of the tooth or caries under the filling confirmed by combined clinical and radiographic assessment. Whether a repair will do is decided only afterwards, by the extent of the lesion.

What is seen, or what the patient feels	What that usually means	Decision
Dark rim at a composite, surface hard and smooth	marginal staining	observation, without touching the filling
Grey shadow in the enamel around an amalgam (an older metal filling)	the colour of the material showing through the tooth	follow-up
Narrow gap at the margin, tissue around it hard	a leaking filling with no cavity	smoothing the margin, or sealing it
Dull, whitish or brown lesion at the margin, surface soft	an active cavity in the enamel	local preparation, usually a repair
Cavity with exposed dentine, the probe sinks into soft tissue	a cavity reaching the dentine	repair or replacement
Shadow under the filling on a bitewing radiograph	suspected caries in the dentine	assessed together with the clinical examination; what follows depends on confirmation and on the extent of the lesion
Fractured wall of the tooth, or a cracked filling	loss of support	replacement, sometimes a laboratory-made restoration

Smoothing the margin and sealing it are procedures that take no bur into the tooth: the dentist polishes the border of the material or covers it with a thin layer of resin, and a leaking filling is not an indication for replacement as long as the tissue around it stays hard.

The table does not replace an examination.

What justifies replacement is whatever cannot be removed without taking off the whole filling. Replacing old fillings is therefore justified where the extent of the lesion does not allow the work to stop at a repair.

Pain is a different matter. A tooth that responds to cold for longer than a dozen or so seconds after the stimulus, hurts on its own or wakes you at night needs an assessment of the pulp, the living tissue inside the tooth, rather than a replacement filling. A symptom like that goes beyond what secondary caries causes, and it is sometimes the first sign of inflammation. Pain that persists once the stimulus is gone, or that wakes you at night, is a matter for the next few days, not for the next check-up - what separates the reversible state from the irreversible one is covered by a separate article on pulpitis.

When does replacement do more harm than good?

Usually when the reason for the procedure is the age of the filling or the colour of its margin alone, while the tissue around it stays hard. How much tooth is already left counts separately. Replacement then sets off what the literature calls the restorative cycle: every opening of the tooth widens the cavity, because the bur takes sound tissue around the old material as well. Less is known about the phenomenon that is the main reason for replacing fillings than the confidence with which it is discussed suggests.

The typical course looks like this:

• a filling on one chewing surface;
• replacement years later: the cavity now involves two walls;
• the next replacement: three walls, a cusp with no support, a laboratory-made restoration or a crown needed;
• root canal treatment, when the bur comes close to the pulp;
• extraction of the tooth.

That is how fillings that have failed end. In the same twelve-year analysis of 900 fillings, 256 were judged failures. This is what became of them: more than half another filling, one in five root canal treatment, roughly one in seven a crown or another laboratory-made restoration, and one in twelve extraction of the tooth.

The strongest evidence comes from a trial with random allocation in which the same team tested two assessment strategies in a group of 185 adults. The first took into account the cavity, marginal adaptation and staining. The second the cavity alone. The first led to the repair or replacement of 113 of 356 fillings, the second 31 of 371 - close to a fourfold difference in the number of procedures. The figure of five given earlier comes from an earlier, cross-sectional study by the same team and counts indications for replacement. This trial counts procedures actually carried out, which is why the multiple is different.

After not quite six years of follow-up, 34 of 371 fillings had failed in the group assessed by the cavity alone, and 30 of 356 in the group assessed more broadly. Close to four times as many procedures at the outset did not translate into longer-lived fillings.

Seven fillings in ten could be reassessed during the follow-up. That is enough to say that no advantage of the more invasive strategy was demonstrated - and too little to say that there is none.

Four situations in which replacement costs more than it gives:

• the only ground is the age of the filling;
• the only ground is a stained margin at a hard and smooth surface;
• the filling does not hurt, it fits closely and it has not changed since the previous check;
• the tooth has two replacements behind it and little wall of its own.

This does not mean that replacement is harmful. What is harmful is replacement with no discernible reason at the tooth. In everyday clinical practice at Modern Dental & Orthodontics (Klinika MDO) we meet patients who arrive with a ready-made list of fillings to be replaced, drawn up solely by their age. We begin the conversation about such a list by checking what has changed at those fillings.

Can the filling be repaired instead of replaced?

Yes, in many situations. A repair consists of removing the damaged part and making it good with new material, without touching the rest of the work.

The estimate at ten years: slightly more than four repairs in ten and just under five replacements in ten survived with no further intervention at all, including a minor correction of the margin. This was calculated on 616 repaired and 264 replaced fillings, drawn from a database of 8542 composite restorations in 3239 patients. The figures relate to work already judged once to need intervention, so they are lower than for fillings placed anew in a sound tooth. The difference amounts to about five fillings in a hundred, and the authors sum it up by saying that repairs hold as long as replacements. With durability that close, what is left is the price in tissue: replacement costs sound tissue, a repair much less. Patients were not randomly allocated here, so the two groups may have differed from the outset.

Repair has limits. It will not solve the problem of a filling that leaks around its whole circumference, or of a tooth with a fractured wall, nor will it help when secondary caries has reached far under the work and its extent cannot be seen from the margin. It works for local damage: a chipped margin, a cavity at one wall. The decision is taken once the border has been exposed, because only then is it clear how far the lesion reaches.

How to extend the life of the fillings you already have

The life of a filling is extended outside the surgery.

• Cleaning the border of tooth and filling with floss or an interdental brush - plaque stays there longest.
• Reaching for sugar less often during the day; what counts is the number of episodes, not the amount at one time.
• Toothpaste with fluoride at adult strength - the packet then carries the figure 1450 ppm or something close to it, in small print on the back.
• Regular check-ups; the number of visits a year is associated with the durability of fillings.
• A night guard, if you grind your teeth.

It is worth raising that last item with the dentist yourself, because patients often do not notice night-time clenching in themselves. Lecturers who work on the longevity of restorations add one more factor to the list, rarely mentioned in conversation with a patient: the threshold at which the dentist reaches for the bur. That is expert opinion, not a study result, but it is consistent with what the two assessment strategies compared above showed.

Monitoring the condition of fillings belongs to restorative dentistry as much as treating new cavities does.

Frequently asked questions

My dentist is suggesting a replacement and I would rather wait. What am I risking?

It depends on what can be seen. With marginal staining alone, deferring does not change the prognosis; it is then worth asking what exactly the dentist sees at that filling. With a lesion that looks like an active cavity, the preparation may be larger at the next visit than it would be today. Do not defer, however, when the tooth hurts on its own, wakes you at night or responds to cold for longer than a dozen or so seconds.

Does a stained filling margin always mean decay?

No. The colour of the border is most often pigment from food and drink deposited in the microgap. What decides the diagnosis is the hardness and the appearance of the tissue right at the margin: dull and soft means a cavity, hard and smooth means staining alone. Staining alone is an indication for observation and does not call for a bur. The grey shadow around an old amalgam is likewise only the colour of the material.

Amalgam may no longer be used in the European Union. Does that mean my old fillings have to be removed?

No. There are no data justifying preventive replacement, and every replacement takes away sound tissue. The ban, in force in the European Union since 1 January 2025, concerns the placing of new amalgam fillings, apart from exceptions justified by a patient's medical needs, and it follows from EU legislation on protecting the environment from mercury. It does not require the removal of well-sealed fillings. What should decide the replacement of such a filling is its condition.

Does a filling have to be replaced after a certain number of years?

No. The age of a filling alone is not an indication for replacement. A filling that fits closely to the tooth, does not hurt and does not change between visits may stay where it is, whatever its age. What decides replacement is its condition assessed at a check-up. If the proposal rests on the age of the filling alone, it is worth asking what else can be seen at it.

How often should the condition of fillings be checked?

There is no single recommendation, and the studies give no clear answer here. The interval is set according to caries risk, and it is shortened by numerous fillings, a dry mouth, frequent snacking and a change noticed at the previous visit. The number of check-ups a year is among the factors associated with the durability of fillings in observational studies.

Does secondary caries hurt?

Usually not at the start. A cavity at the margin of a filling develops painlessly, and the first symptom that is felt is often a brief reaction to something sweet or cold. Pain that persists once the stimulus is gone, pain that comes on for no reason and pain that wakes you at night already mean the pulp is involved and require an urgent appointment.

Will a radiograph always show it?

No. In the comparison of 25 studies, an intraoral radiograph detected about six cavities in ten of those actually present, and so missed four. What it shows above all are lesions on the contacting surfaces and those that have reached the dentine. A negative result is not a certificate that nothing is happening under the filling.

How many times can a filling in the same tooth be replaced?

There is no set limit, but every replacement widens the cavity and brings closer the moment when an ordinary filling stops being enough. In the twelve-year analysis of 900 fillings, among the 256 judged failures, the most common outcome was another replacement, more common than root canal treatment or extraction. That is why repairing a part is sometimes better than taking off the whole work.

Can secondary caries also develop under a crown?

Yes, and it is harder to recognise. The margin of a crown often runs below the gum, so neither sight nor probe reaches there as it does at a filling, and metal or ceramic obscures the image on the radiograph. Indirect symptoms then matter: sensitivity at the neck of the tooth, a smell from under the crown.

When watching is the treatment

An old filling that does not hurt, fits closely to the tooth and does not change between visits has every right to stay where it is.

This principle has another side. A cavity left under a filling for too long ends in root canal treatment, and that is a greater cost than one replacement carried out as a precaution. A dentist who proposes a replacement is usually acting on what they were taught. This article is not an argument for refusing procedures, but for asking what grounds them.

Before a check-up it is worth establishing the same three things at every filling put forward for replacement: what exactly can be seen at it, whether that has changed since the previous visit, and whether it can be repaired without taking off the whole.

Risk to a tooth with a filling of many years is sometimes not the cavity, but a procedure done pre-emptively. A missed cavity leaves the next check-up to catch it; tissue once removed is never recovered.

Read more:

• Aesthetic and Restorative Dentistry in Warsaw
• Composite filling, ceramic inlay or crown: what decides is how much tooth is left
• Tooth decay stages: when can it be stopped without drilling
• Pulpitis - reversible or irreversible. Why this distinction decides the fate of the tooth
• Comprehensive dental review: What does it contain and why is it crucial for health?

Content and liability disclaimer

This article is informational and educational in nature and does not constitute medical advice, a diagnosis or a therapeutic recommendation — it does not replace a consultation with a specialist. If you are experiencing symptoms, have doubts or are facing a decision about treatment, consult a dentist. The methods described and the data cited are general in nature. The results of scientific studies relate to populations, not to an individual patient. The course and outcome of therapy depend on the individual clinical situation and may vary between patients. No information contained in this article constitutes a guarantee of result. The content has been prepared with due care, based on publicly available medical knowledge and the scientific publications indicated in the Sources section. We do not advise taking or refraining from any health-related action solely on the basis of the content of this article, without prior consultation with a doctor.

Sources

Source 1

Links https://doi.org/10.1007/s00784-020-03268-7 │ https://pubmed.ncbi.nlm.nih.gov/32300980/

Description Askar H, Krois J, Göstemeyer G, Bottenberg P, Zero D, Banerjee A, Schwendicke F. „Secondary caries: what is it, and how it can be controlled, detected, and managed?” Clinical Oral Investigations. 2020;24(5):1869-1876.

Source 2

Links https://doi.org/10.1186/s12903-022-02112-6 │ https://pubmed.ncbi.nlm.nih.gov/35300657/

Description Signori C, Uehara JLS, Romero VHD, Moro BLP, Braga MM, Mendes FM, Cenci MS; CaCIA Collaborative Group. „Comparison of two clinical approaches based on visual criteria for secondary caries assessments and treatment decisions in permanent posterior teeth.” BMC Oral Health. 2022;22(1):77.

Source 3

Links https://doi.org/10.1177/00220345241247773 │ https://pubmed.ncbi.nlm.nih.gov/38752325/

Description Digmayer Romero VH, Signori C, Uehara JLS, Montagner AF, van de Sande FH, Maydana GS, Chaves ET, Schwendicke F, Braga MM, Huysmans MC, Mendes FM, Cenci MS; CaCIA Collaborative Group. „Diagnostic Strategies for Restorations Management: A 70-Month RCT.” Journal of Dental Research. 2024;103(7):697-704.

Source 4

Links https://doi.org/10.1016/j.jdent.2024.105541 │ https://pubmed.ncbi.nlm.nih.gov/39719157/

Description Ku JC, Lam WY, Li KY, Hsung RT, Chu CH, Yu OY. „Accuracy of detection methods for secondary caries around direct restorations: A systematic review and meta-analysis.” Journal of Dentistry. 2024;153:105541.

Source 5

Links https://doi.org/10.1016/j.dental.2022.11.009 │ https://pubmed.ncbi.nlm.nih.gov/36494241/

Description Demarco FF, Cenci MS, Montagner AF, de Lima VP, Correa MB, Moraes RR, Opdam NJM. „Longevity of composite restorations is definitely not only about materials.” Dental Materials. 2023;39(1):1-12.

Source 6

Links https://doi.org/10.1016/j.dental.2019.11.001 │ https://pubmed.ncbi.nlm.nih.gov/31784061/

Description Kanzow P, Wiegand A. „Retrospective analysis on the repair vs. replacement of composite restorations.” Dental Materials. 2020;36(1):108-118.

Source 7

Links https://doi.org/10.2341/24-095-B │ https://pubmed.ncbi.nlm.nih.gov/39420218/

Description Opdam NJM, Montagner AF, Cenci MS. „Buonocore Memorial Lecture 2023: Longevity of Dental Restorations or Longevity of Teeth: What Matters?” Operative Dentistry. 2024;49(6):655-664.

Source 8

Links https://doi.org/10.1007/s00784-024-05631-4 │ https://pubmed.ncbi.nlm.nih.gov/38630376/

Description Alonso ALL, Tirapelli C, Cruvinel PB, Cerqueira NM, Miranda CS, Corona SAM, Souza-Gabriel AE. „Longevity of composite restorations in posterior teeth placed by dental students: a 12-year retrospective study.” Clinical Oral Investigations. 2024;28(5):253.

Source 9

Links https://doi.org/10.1007/s00784-026-07038-9 │ https://pubmed.ncbi.nlm.nih.gov/42467381/

Description Bittar A, Alkan E, Tağtekin D. „Diagnostic accuracy of QrayPen fluorescence scoring for secondary caries in repair-/replacement-indicated posterior composite restorations.” Clinical Oral Investigations. 2026;30(8):342.