Fatigue Failure in Biomaterials

Fatigue failure is the progressive structural damage that accumulates when a material is loaded cyclically at stresses below its yield strength, and so well below its ultimate tensile strength. A single load of that size may be safe, but 10^6 cycles of it cause damage to accumulate. Fatigue is responsible for the majority of mechanical failures in orthopaedic implants, among them plate fractures, screw breakage and prosthesis stem fractures in total hip arthroplasty.

Fatigue versus static failure. A static failure happens when a single load exceeds the material's strength, and it is predictable from the ultimate tensile strength: a ductile material yields before it fractures, a brittle one fractures suddenly. In fatigue, cyclic loads accumulate damage below yield and a crack grows progressively. The final fracture is sudden and looks brittle.

Three stages. The process runs through three stages in sequence:

1. Crack initiation (Stage I). Microstructural damage accumulates until a crack nucleates at a stress concentration: a defect, a notch or the surface. Initiation takes 10-90% of fatigue life.
2. Stable crack propagation (Stage II). The crack grows a stable amount with each cycle. This is the Paris law region, and the one in which beach marks form.
3. Final fracture (Stage III). Once the crack reaches its critical length, the remaining section fails rapidly and unstably.

How many cycles. A hip replacement sees on the order of one to two million load cycles a year of walking, about 10^7 over a decade, which is why 10^7 is the conventional runout for implant fatigue testing. An implant meant to last 10-20 years must therefore survive 10^7-10^8 cycles; at two million cycles a year that is 20-40 million. Design stresses have to sit well below the fatigue limit.

The loads. Each implant is loaded in its own way:

• Hip stem: 2-5 MPa of cyclic stress
• Plate across a nonunion: repeated bending over hundreds of thousands of cycles
• Screw: cyclic shear and tension

The S-N curve. The S-N (Wöhler) curve is the fundamental characterisation of fatigue: cyclic stress amplitude (S) plotted against the number of cycles to failure (N). It is generated by testing specimens at a range of stress levels and recording the cycles to failure at each. High stress gives few cycles and low stress many; a horizontal asymptote, where the material has one, is the endurance limit.

View larger

View larger

High- and low-cycle fatigue. Failure at low stress after more than 10^5 cycles is high-cycle fatigue. Failure at high stress in fewer than 10^5 cycles is low-cycle fatigue.

Endurance limit and fatigue limit. Ferrous metals such as steel have a true horizontal asymptote, the endurance limit: a stress below which infinite cycles can be sustained. In non-ferrous metals, titanium and aluminium among them, the curve keeps declining, so a practical fatigue limit is quoted instead at 10^6 or 10^7 cycles. For titanium alloys it is defined at 10^7 cycles, about 60% of UTS.

View larger

View larger

Mean stress and stress ratio. Fatigue life depends on the mean stress and the stress ratio as well as on the amplitude:

• Stress amplitude σ_a = (σ_max - σ_min) / 2
• Mean stress σ_m = (σ_max + σ_min) / 2
• Stress ratio R = σ_min / σ_max

The Goodman relationship. Higher mean stress reduces fatigue life. The Goodman diagram plots allowable stress amplitude against mean stress, and the safe region lies below the line.

From one stress level to many. The S-N curve and the 20-40 million cycle design figure both assume constant-amplitude loading. A real implant sees a spectrum instead: level walking, stair climbing, rising from a chair, and the occasional stumble or near-fall. The classical bridge from constant-amplitude laboratory data to this variable physiological loading is the Palmgren–Miner linear cumulative damage rule.

Miner's rule. Each block of n cycles at a stress level whose constant-amplitude life, read off the S-N curve, is N consumes a fraction n / N of the total fatigue life. Failure is predicted when the fractions sum to one:

Σ (n / N) = 1

Why a stumble matters more than the steps. The high-stress part of the S-N curve is steep, so at high stress the constant-amplitude life N is very small. A small number of high-amplitude events (a stumble, a fall onto the limb, stair-climbing in a heavy patient) therefore consumes a disproportionate share of fatigue life. Peak load, patient weight and activity level dominate fatigue life far more than raw step count.

Spectrum, not count. A count of gait cycles is necessary but not sufficient: the loading spectrum, meaning how many high-amplitude events there are, governs survival. That is why pre-market test standards specify a defined cyclic load amplitude and a run-out target, not merely a number of cycles.

A deliberate simplification. The rule is linear and ignores the order of loading and the interaction between loads, since an overload can locally retard or accelerate subsequent crack growth. The real damage sum at failure therefore scatters around one rather than equalling it exactly. It remains the standard first-pass life-prediction tool.

View larger

View larger

View larger

Paris law. In the stable propagation region (Stage II), the rate of crack growth per cycle follows Paris law:

da/dN = C (ΔK)^m

• da/dN is the crack growth rate, in metres per cycle
• ΔK is the stress intensity factor range, K_max - K_min
• C and m are material constants, m typically 2-4

The stress intensity factor. K = Y × σ × sqrt(π × a), where Y is a geometry factor, σ the applied stress and a the crack length. As the crack grows, a increases and K with it, so the growth rate accelerates until K reaches K_IC, the fracture toughness, and final fracture occurs.

What follows from the law. Small cracks grow very slowly because ΔK is low, and growth then accelerates as a power of ΔK, the exponent m. Fatigue life therefore depends heavily on the initial defect size, and inspection can detect a crack before it reaches critical size.

What speeds a crack. The environment, the surface, residual stress and the microstructure all change how fast a crack starts or grows:

View larger

View larger

View larger

Reading the surface. A fatigue fracture leaves a characteristic, diagnosable face on the retrieved implant or in a clinical photograph, and reading it is a high-yield viva skill. The surface has three zones.

View larger

Beach marks versus striations. This is the classic trap. Beach marks (clamshell or arrest marks) are macroscopic, visible to the naked eye, and each marks a change in loading or a rest period, not one cycle. They reflect variable, real-world loading, so they are often absent on a constant-amplitude laboratory failure. Striations are microscopic, seen on electron microscopy, and in the classical model each striation represents one loading cycle: the physical record of Paris-law advance.

Reading the whole surface. Beyond identifying the zones, three features are read:

• The origin identifies the stress riser responsible
• The ratio of the smooth fatigue zone to the rough fast-fracture zone indicates the nominal stress. A large fatigue zone with a small final zone means a low applied stress over many cycles; a small fatigue zone with a large final zone means a high applied stress
• Multiple origins and ratchet marks (steps between adjacent crack fronts) indicate a high stress concentration or a high load

On a retrieved implant. This is how the Gilbert retrieval analysis (Evidence Base) identified intergranular corrosion-fatigue: scanning electron microscopy of the fracture face localised the origin to the corroded taper region.

View larger

View larger

View larger

Where implants fail in fatigue. The common clinical scenarios:

• Plate fracture in delayed union or nonunion: the plate bears cyclic bending for months
• Screw breakage, including in spinal instrumentation: stress concentration at the threads, especially if overtightened
• Hip stem fracture: rare with modern designs, seen with undersized stems. Loosening changes load transfer, and the loss of support turns bending into a fatigue problem
• Tibial baseplate failure in TKA: unsupported overhang creates cantilever bending
• Modular junction: taper fractures in hip stems from fretting and corrosion

Prevention. Proper implant design and early bone healing, before fatigue damage accumulates, are the defences:

• Size the implant properly and avoid undersizing
• Minimise stress concentrations: avoid sharp corners and notches
• Surface treatments: polishing, passivation
• Avoid stress shielding
• Achieve early bony union, which reduces the loading cycles on the implant
• Follow the manufacturer's guidelines and do not modify implants

Corrosion and fatigue. Corrosion dramatically reduces fatigue life, through four mechanisms:

• Fretting corrosion: micromotion creates wear particles and crevices
• Crevice corrosion: oxygen depletion in gaps accelerates oxidation
• Pitting corrosion: creates stress concentration sites for crack initiation
• Stress corrosion cracking: tensile stress combined with a corrosive environment

The modular taper. Modular taper junctions in THA are subject to fretting corrosion. Micromotion between head and stem creates debris and a crevice environment, with the potential for catastrophic taper fracture, so proper assembly (clean, dry, impaction) is critical.

Fatigue or something else? A fractured implant on a radiograph is not automatically a fatigue failure. The viva-critical skill is recognising the failure mode from the history, the surface appearance and the timing.

View larger

View larger

Global Epidemiology

• Implant fatigue fracture is rare with modern designs but contributes to revision burden, especially where union is delayed. Cyclic loading is universal: walking generates roughly 1 to 2.5 million gait cycles per year, so a 20-year implant must survive tens of millions of cycles.
• Fretting and corrosion at modular junctions (trunnionosis) emerged as a recognised mode of metal-ion release and component failure during the metal-on-metal and large-head era, prompting redesign and several device withdrawals worldwide.

Standards, Guidelines and Testing Frameworks (Side by Side)

Registry Evidence

• Joint registries (NJR England and Wales, AJRR USA, AOANJRR Australia, Swedish SHAR, Norwegian, NZJR) track revision causes. Pure stem fatigue fracture is uncommon, but registries flagged early failure of specific modular-neck and large-head metal-on-metal designs, providing the population signal that drove withdrawals.
• Registry data give indirect evidence of fatigue- and corrosion-related failure through device-specific revision rates and time-to-revision curves.

High- vs Limited-Resource Practice Variation

• In well-resourced settings, alloy selection, surface engineering and modular-junction surveillance (serum metal ions, cross-sectional imaging) are routine.
• In limited-resource settings, reuse of implants, off-label modification (contouring/cutting plates, which introduces stress risers) and delayed nonunion management increase fatigue-fracture risk. The universal mitigation everywhere is achieving timely union to transfer load from implant to bone.

Exam Relevance (Global)

S-N curves, endurance limit, Paris law and the clinical link to plate fracture in delayed union are core basic-science material in orthopaedic fellowship and board examinations worldwide. Material selection (titanium vs stainless steel vs cobalt-chrome) and fatigue properties are frequently examined.

Fatigue principles are well established, but several clinically important questions remain debated.

Does modularity do more harm than good? Modular necks and dual-taper stems add intra-operative flexibility but introduce extra crevice junctions vulnerable to fretting corrosion-fatigue. Several modular-neck designs have been withdrawn after high fracture and adverse-reaction rates, yet matched data on when modularity is genuinely needed remain limited.

Optimal construct stiffness. Very stiff locked constructs can suppress callus and shift cyclic load onto the implant, risking fatigue if union is delayed; constructs that are too flexible permit excess motion. The ideal working length, screw density and use of titanium versus stainless steel for a given fracture remain areas of active biomechanical debate.

View larger

Surface treatments in practice. Shot peening and laser shock peening raise fatigue limits in the laboratory (10 to 17 percent in Ti-6Al-4V), but whether this translates into measurable reductions in clinical implant fracture or revision is not established by registry-level evidence.

Predicting individual fatigue life. Fatigue life is highly sensitive to initial defect size, patient weight and activity, so population S-N data poorly predict any single patient. There is no validated clinical tool to forecast time-to-fatigue-fracture for an individual implant in vivo.