ISO 14801

What the test does

An implant is clamped in a holder inclined 30 degrees to the machine axis, with the clamping plane 3 mm below the point representing bone level, leaving a defined length unsupported as bone loss would. The abutment and screw are fitted, and a hemispherical cap sits 11 mm from the clamping plane. A flat platen, free to slide sideways, presses on that cap and cycles sinusoidally between a chosen peak load and a tenth of it. The tilt and the lever turn axial force into bending, and cycling continues until the specimen fractures or reaches run-out.

What it measures, and why it matters

Each specimen gives a pair of numbers — peak load applied and cycles survived — and a set run at descending loads builds the load-cycle diagram from which a fatigue limit is read. What is really being applied is a bending moment: in the standard geometry it is 5.5 times the force at the cap, so a few hundred newtons becomes a substantial moment at the implant neck. That is where the method earns its keep, since chewing loads arrive off-axis and the neck and the abutment screw joint are where implants fail. It is not a measure of the base material's fatigue properties, and does not predict clinical survival.

Specimen

The specimen is an implant with the abutment and screw supplied for it, tightened with the manufacturer's instruments to the recommended torque, or within 5 % of it where those are unavailable: preload is part of the specimen, not a detail of assembly. The implant is held in a rigid clamp or embedded in a medium stiffer than 3 GPa, so the holder adds no compliance. Angulated abutments are tested with a further 10 degrees of tilt, representing a connection left undercorrected.

The method does not apply to implants with endosseous lengths below 8 mm, nor to magnetic attachments. Several specimens at each of several peak loads are needed to place the curve; the sample count the standard itself sets could not be confirmed from public sources.

What the machine must be capable of

Peak loads sit in the low hundreds of newtons, so capacity is not the constraint: a few hundred newtons will do the work, and one to two kilonewtons of dynamic capability leaves headroom. Accuracy is specified instead — force error must stay within 5 % at maximum load, verified to ISO 7500-1.

Loading is sinusoidal at a load ratio of 0.1, so the trough never reaches zero and the joint is never unloaded. Frequency is capped at 15 Hz in air and 2 Hz in liquid, the lower cap acknowledging that a bath cannot follow a fast cycle. Run-out is generally taken as five million cycles in air and two million in liquid — roughly four and twelve days at those caps. Both are in universal use but sit beyond the freely published part of the standard, so confirm them against the purchased text. Unattended running of that length makes automatic failure detection a requirement, not a convenience.

No extensometer is used. The fixture carries the method: the inclination, the clamping offset and the lever length are all held to tight tolerances, and the platen must stay free to move transversely — one that grips the hemisphere adds a side constraint and quietly changes the moment. Where corrosion fatigue is expected or polymeric parts are present, the specimen is immersed in saline or another physiological medium at 37 °C ± 2 °C; otherwise air at 20 °C ± 10 °C serves. The choice must be justified and reported.

What goes wrong in practice

Movement in the embedding spoils otherwise quiet data. If the implant settles in the pot, or the clamp relaxes, the free length grows and the moment arm with it, so the specimen sees more than the load recorded — visible only as a curve that drifts across a batch.

Abutment screw loosening is the second. Preload lost during cycling lets the joint micro-move, and the result is often reported as an implant fracture when it is really a joint failure; retightening between runs to keep a test going destroys comparability.

The third is thermal, and belongs to fast dry testing: near the upper frequency cap, polymeric components and the embedding medium warm, and the specimen that fails is no longer the one that was fitted.

Related and equivalent standards

ANSI/ADA Standard No. 127 is this method as adopted in the United States, EN ISO 14801 and DIN EN ISO 14801 the European and German adoptions; a laboratory quoting any of them is running the same test. Within ISO, the method leans on ISO 7500-1 for force verification and ISO 1099 for fatigue conventions, and takes its vocabulary from ISO 16443. It should not be confused with a single-cycle static load to failure on the same fixture, which gives a much higher figure and describes a different event.