Alveolar crest

The alveolar crest is the alveolar bone crest — the interproximal bone landmark whose distance from the cemento-enamel junction is measured on bitewings. The distance between crest and CEJ is what tells you whether bone has been lost. Advanced loss is obvious. Early loss is the hard call.

Synonyms: alveolar bone crest; crestal bone; interproximal bone crest; bone crest

What is the alveolar crest, and where do you find it on a bitewing?

Follow the bone up between two teeth until it stops. That upper edge is the crest.

You never read it on its own. You read it against the cemento-enamel junction — the CEJ — which is the stable landmark for measuring clinical attachment loss and assessing periodontal damage. Periapical and bitewing radiographs are both standard for locating the CEJ, and severity is worked out from the crest level and the distance between the bone and the CEJ.

Bitewings are where most people meet this landmark, because the interproximal bone sits in the middle of the image. The ADA and FDA put the reading obligation plainly: care should be taken to examine all radiographs for any evidence of caries, bone loss from periodontal disease, developmental anomalies and occult disease. If you are still sorting out which structures read pale and which read dark, start with radiolucent vs radiopaque and come back.

How far below the CEJ does a normal crest sit?

There is no single agreed figure, and the published numbers are not measuring the same thing. Here is what the primary studies actually report.

Three things are going on, and they explain the spread.

The populations differ. One of these studies measured adolescents and one measured adults across four decades of life. The adult figure is higher.

Smoking shifts it. In Bergström's cohort, smokers sat at 1.82 mm at baseline against 1.16 mm for non-smokers.

The reference standard differs. Källestål and Matsson started from clinical gingival health and excluded sites with bleeding on probing, attachment loss, defective fillings or calculus. Hausmann and colleagues started from sites with no clinical attachment loss and reported confidence limits rather than a mean. One is describing healthy teenagers; the other is defining a range that means "not diseased."

So quote the range and say which study you are quoting. Do not average them. A tidy "0.5 to 2 mm" is a number none of these papers reports.

If you want the single most useful anchor: in Källestål and Matsson's healthy adolescents, 94–100% of maxillary sites and 100% of mandibular sites measured 1.5 mm or less.

What does early crestal bone loss look like against normal variation?

Start with what the literature admits. The paper that set the reference range opens by saying that in advanced periodontal disease crestal loss is readily apparent on radiographs, but in incipient disease there is a lack of consensus regarding what constitutes crestal bone loss on a radiograph. That was written in 1991 and nothing retrieved since settles it.

What you have instead of a threshold is a reference range and three habits.

Use the range as a prior, not a verdict. A CEJ-to-crest distance between 0.4 and 1.9 mm is consistent with no crestal bone loss at sites with no attachment loss. A site outside it is worth attention. A site inside it is not proof of health.

Look at the bone, not only the gap. When Källestål and Matsson recorded these sites they scored the appearance of the marginal bone as normal, sclerotic, notched and/or irregular, or showing widening of the periodontal ligament space. Those are the descriptors they recorded, not a validated grading scale — but they are the right things to be looking at while you measure.

Remember what a stage is measured against. In the 2018 staging and grading framework, stages I to IV are defined on severity — primarily periodontal breakdown with reference to root length — alongside complexity of management and extent. A number on a bitewing is one input to that, not the diagnosis.

Change over time is the strongest signal you can get, and it is slow. Over ten years, Bergström measured a mean bone height reduction of 0.27 mm in non-smokers and 0.74 mm in smokers. Hold that next to the section below, where changing the beam angle by ten degrees moved the measured crest by more than a millimetre. No study compares those two directly, but the reasoning is worth carrying to the light box: a crest that looks a couple of millimetres lower than last time is worth checking the geometry on before it earns a diagnosis.

Why does the crest look different in the anterior and the posterior?

Two answers, and only one of them is measured.

The distance itself varies by arch and by surface. Källestål and Matsson found means of 0.9–1.0 mm at maxillary surfaces against 0.7–0.8 mm at mandibular surfaces, with the mesial of the mandibular first premolar the shallowest at 0.5 mm. Small differences, but they are regional.

The projection varies more. Hausmann and colleagues, radiographing human skulls at known beam angles, found that measured beam angulation ranged between 90 and 80 degrees for posterior bitewings against 90 to 70 degrees for posterior periapicals — and that in the premolar and incisor regions, deviation of up to 20 degrees from perpendicular changed the crest level little if at all. So the molar region is where projection error bites hardest, and the anterior is comparatively forgiving.

You will also hear that the crest is pointed and sharp between anterior teeth and flatter between posterior teeth. Treat that as a teaching description rather than a measurement: nothing retrievable measures crest shape by region. What is measured is the distance and the geometry.

What it's confused with

Cervical burnout and the Mach band effect are both recognised as factors that influence the radiographic interpretation of caries.

Which technique errors make the crest look worse than it is?

Vertical angulation is the big one. In a study of five human mandibles, bitewings taken at −10°, 0° and +10° produced a wide range of over- and under-estimation of the alveolar crest level, with mean differences at individual sites from 1.84 mm to 3.70 mm. That is an in-vitro figure from cadaveric mandibles, not a measured chairside error rate — but the direction is not in doubt. Working from skulls, Hausmann and colleagues showed the inaccuracy of a crest measurement is proportional to how far the beam deviates from 90 degrees.

Horizontal angulation costs you differently: it hides the crest behind the neighbouring tooth.

And a malaligned beam can remove the crest from the image entirely. The ADA and FDA describe cone-cutting as non-exposure of part of the image receptor due to malalignment of the x-ray beam.

How to be sure

Hold the receptor properly. Holders that align the receptor precisely with the collimated beam are recommended for periapical and bitewing radiographs, and a receptor-holding device minimises the risk of cone-cutting. In the mandible study, the authors' own conclusion was that bitewing film holders should be used to minimise beam angulation error in general dental practice.

Prefer the bitewing for crest measurement. For longitudinal measurement of the crest-to-CEJ distance, particularly in the molar region, Hausmann and colleagues recommended bitewings over periapicals — that was a recommendation for research measurement, but the geometry behind it is the same geometry you are working in.

Repeat before you call it. If the crest looks low and the geometry looked wrong, the apparent level is the thing most likely to change on a properly held retake. Bone loss will still be there.

Then use the radiograph as one input. The ADA and FDA recommend that clinical judgment be used in determining the need for, and type of radiographic images necessary for, evaluation of periodontal disease, and that imaging may consist of, but is not limited to, selected bitewing and/or periapical images of areas where periodontal disease (other than nonspecific gingivitis) can be identified clinically.

Common errors

Averaging the published ranges into one number. The figures come from different populations under different exclusion rules. There is no mean of them that any paper reports.

Reading one radiograph as a measurement. Without knowing the beam angulation, a crest level is an estimate with a couple of millimetres of play in it.

Calling burnout at the cervix bone loss at the crest. Different location, different cause, and often a different appearance on a different view of the same tooth.

Treating in-vitro angulation figures as clinical error rates. The 1.84 to 3.70 mm range came from mandibles on a bench. It tells you the mechanism is real, not how wrong your last bitewing was.