Failure Analysis

Since the inception of the DBEC retrieval library in 1976, our understanding of the complexities of joint replacement technology has improved substantially.  Retrieval laboratories around the world, including Dartmouth’s, have greatly facilitated the overall optimization of design and materials for orthopedic implants, which in turn produces profound quality-of-life improvements for joint-replacement patients worldwide. 

The laboratory has evolved scientifically and financially to weather dramatic changes in the orthopedic industry, making significant impacts on the care of patients worldwide.   In the last four decades, Thayer School's implant retrieval analysis played a key role in identifying failure modes and relating them to various designs and materials being used in the industry. Indeed, in 2000, NIH's Consensus Development Program produced a technology assessment statement acknowledging the value of implant retrieval programs. The statement drew the following conclusions:

  1. Implant retrieval and analysis is of critical importance in the process of improving care of patients in need of implants.
  2. Attention needs to be directed toward reducing various obstacles to implant retrieval and analysis, particularly legal and economic disincentives.
  3. The failure to appreciate the value of implant retrieval and analysis is a serious impediment to research in devices. A focused educational program will provide the information necessary for improving the quality of future devices.

While most people understand that medical implants improve quality of life, far fewer recognize the importance of retrieving and analyzing implants when they fail or are no longer in use.

Examples of the Value of Retrieval Analysis Over the Decades

1970’s-1980’s:  Retrieval studies show porous coating for bone ingrowth is a successful fixation technique.  Additional studies show porous coating needs specific metallurgy, structure, and adhesion to orthopedic components.  Industry/Clinical Change:  Porous coatings are widely used and the primary fixation technique for total hip arthroplasty.

1980’s-1990’s:  Retrieval studies show corrosion as a potential problem in metal components.  This is particularly true for modular devices. Industry/Clinical Change:  Manufacturers have changed metallurgical processing and attempted to minimize modularity in high-stress areas of devices.

1990’s:  Retrieval studies illuminate challenges and benefits of modularity in patellar components.  Industry/Clinical Change:  Metal-baked patellae no longer include thin polyethylene or snap-together components.  4mm is established as a minimum thickness.

1990’s:  Retrieval studies show benefits and drawbacks of design strategies in the knee and hip.  Industry/Clinical Change:  Numerous – Hydroxyapatite proved to be a successful fixation system; Rough titanium trays are becoming less common in modular total knee systems; Thin polyethylene leads to fracture; etc.

1990’s:  Retrieval studies show gamma in air oxidation leads to fatigue failure and increased wear of orthopedic devices.  Industry/Clinical Change:  All polyethylene components in the United States are sterilized in a gamma-barrier package or using a non-ionizing radiation source.

2000’s:  Retrieval studies show biomechanics contribute to wear and performance of devices. Industry/Clinical Change:  Rotating platform knees see a resurgence; Highly conforming knees see redesigns to minimize backside wear or accommodate torque from the femur; Surgeons are cautioned against using femoral stems where there is a lack of bone support.

2000’s:  Retrieval studies show contribution of material type to performance of bearings.  Industry/Clinical Change:  Many companies switch away from calcium-stearate-containing polymers in the knee and the hip.

2000’s:  Retrieval studies show performance of novel materials in hip and knee applications.  Industry/Clinical Change:  Many companies move to highly crosslinked materials in the knee and the hip, but change to lower crosslinking doses in the knee.  Surgeons are advised to minimize stress in acetabular cups with higher crosslinking doses.

2000’s:  Retrieval studies document in vivo oxidation of gamma barrier components. Industry/Clinical Change:  Many companies move away from gamma barrier devices, some institute antioxidant technologies.

2010’s:  Retrieval studies document in vivo oxidation of highly crosslinked devices. Industry/Clinical Change:  Companies are moving to antioxidant technologies.

2010’s:  Retrieval studies document squeaking phenomenon in ceramic hips.  Industry/Clinical Change:  Acoustic considerations are weighted heavily in bearing selection.

2010’s:  Retrieval studies illuminate biomechanics, damage, and wear associated with metal on metal hips. Industry/Clinical Change:  Studies are in progress.  We expect that failure analysis of all devices will inform future design of hard-on-hard hips.

Bearing Function

Retrieved knee devices sent to us by orthopaedic surgeons are assessed for damage, and also quantitatively assessed for wear. Dimensions of retrievals are compared to design specifications or shorter in-vivo duration devices to calculate both articular and backside wear. Wear and wear rate are correlated with variables including polyethylene pedigree, articular bearing geometry, device fixation, and patient factors.

An important early outcome from this effort is the distinction between damage and wear of knee bearing inserts. Damage of joint arthroplasty bearings can be visually striking, can be described semi-quantitatively according to published techniques, and can impact mechanical performance and kinematics of the implant. Wear that occurs by abrasive/adhesive processes can be very challenging to discern and quantify, yet can produce large volumes of small debris particles that can lead to osteolysis. Damage and wear are important but distinct phenomena that can have different impacts on clinical performance.

Highly cross-linked (HXL) polyethylene has proven to be a wear-resistant acetabular bearing material in total hip arthroplasty (THA). In vitro wear testing has predicted a tenfold reduction in the wear rate of HXL polyethylene, as compared to conventional, non-HXL bearings. To date, radiographic studies of head penetration represent the state-of-the-art in determining clinical wear of polyethylene hip liners. However, as the amount of wear drops to very low levels, it becomes important to develop a precise and reliable method for measuring wear, facilitating a comparison of clinical results to laboratory expectations.  Fixed-magnitude errors associated with digital imaging necessitate increasingly large studies to statistically elucidate the low wear rates.  Retrieval analysis provides much better precision, but is subject to different sources of error.

Our current work focuses on locating and quantifying the maximum linear wear of retrieved acetabular liners using a coordinate measuring machine (CMM) and a reverse-engineering algorithm. Specifically, HXL liner wear can be assessed as a function of radiation dose and compared to a baseline of conventional, non-HXL bearings.

Few retrievals studies have comparatively examined wear of both reverse and total shoulder arthroplasty. The lack of literature on this topic prevents organizations from standardizing and publishing methods for wear testing of shoulder components.  In order to create relevant wear testing standards, it is crucial to understand how components wear in vivo including the modes and locations of wear.  One goal of our current work is to examine series of reverse and total shoulders to determine the incidence of abrasive and adhesive wear and determine typical locations for these wear patterns on polyethylene components.

Material Behavior

Medical grade ultra-high molecular weight polyethylene (UHMWPE) is the current gold standard for joint bearing materials used in TJA. Although TJA involving UHMWPE as a bearing surface has been one of the most successful procedures of the last century, issues of wear, oxidation, and fatigue failure remain obstacles to the longevity of joint replacements. As a failure mode, wear is biologically compounded, because wear debris can trigger a series of reactions leading to osteolysis, a condition resulting in long term resorption of the bone around the implant.

Radiation crosslinking of UHMWPE significantly improves its wear resistance, as evidenced by in vivo clinical studies and in vitro hip simulator studies. During irradiation, crosslinks are formed between polymer chains through homolytic cleavage of C-H and C-C bonds. However, ionizing radiation also produces free radicals randomly throughout UHMWPE as part of the crosslinking process. These long-lived species can react with oxygen, triggering a cyclic complex cascade of chemical reactions. While free radical oxidation involves a number of possible pathways with different mechanisms and end products, the overall outcome includes polymer chain scissions which reduce the molecular weight, and various oxidative products such as hydroperoxides, ketones, alcohols, and carboxylic acids. Overall, this cascading oxidative reaction is responsible for progressive embrittlement of the material. Oxidative degradation thus manifests as a reduction of wear resistance and mechanical properties.

In the late 1990’s, oxidation of UHMWPE was identified as a serious concern as it limits the overall lifetime and success of a joint replacement. Hence, the elimination of free radicals in UHMWPE has been an important focus of orthopedic manufacturers ever since the industry’s response to shelf storage oxidation occurring in gamma sterilized devices. In collaboration with materials research laboratories, device manufacturers have developed a variety of post-irradiation thermal treatments with the dual goals of promoting oxidative stability and enhancing the crosslink density of bearing materials.

One thermal treatment approach utilizes heating above the melting point of the crosslinked polymer following irradiation. This melts the crystalline regions and allows recombination of trapped free radicals in these domains. After the polymer recrystallizes, the residual free radicals have been quenched and the material is both wear resistant and more oxidatively stable in shelf-aging and artificial aging according to current ASTM Standards. However, this improved wear and oxidation resistance comes at a cost because post-irradiation melting further decreases the fatigue strength of UHMWPE, already reduced by radiation crosslinking; the melting step results in a decrease in crystallinity and ductility. Thus, upon cooling, the extent of recrystallization for crosslinked UHMWPE is inferior to that of UHMWPE without crosslinks.

An alternative method of thermal treatment is post-irradiation annealing below the melting point of the crosslinked polymer. In the absence of a recrystallization step, annealed materials possess superior mechanical properties in comparison to fully remelted materials with the same radiation dose.  However, this approach reduces but does not completely eliminate free radicals as achieved by melting. As a result, this material is still susceptible to in vivo oxidation. Our lab and others have reported oxidation in irradiated and annealed UHMWPE both in retrieval analysis18 and in vitro accelerated aging studies.

Recent retrieval analyses conducted by our laboratory showed that despite the absence of free radicals (prior to implantation), highly crosslinked (HXL) remelted acetabular liners and tibial inserts showed signs of oxidation occurring in vivo, with greater oxidation rates occurring in TKA in comparison to THA. The crosslinking radiation dose significantly impacted the material’s oxidation potential. Additionally, the in vivo oxidation rate significantly correlated with transvinylene bond concentration (also referred to as unsaturations), and potentially to contact stress. Others have observed similar results, and have further shown potential connection between absorbed species and in vivo oxidation.  There thus exist several potential oxidation pathways, including: (1) the conventionally accepted free radical mediated mechanism; (2) A stress induced chain scission mechanism; (3) An absorbed pro-oxidative species mechanism; and (4) an irradiation-energy, chemical bond-related mechanism.

In an effort to improve oxidation resistance without compromising mechanical properties through thermal treatments, manufacturers have examined the use of antioxidants to prevent oxidation of free radicals.  The most prevalent antioxidant available today is a liquid-based alpha-tocopherol (Vitamin E). Two methods of adding vitamin E to the UHMWPE have been examined: combining liquid antioxidant into UHMWPE resin powder prior to compression molding, and diffusion of vitamin E into already cross-linked UHMWPE.  Adding vitamin E into the resin prior to crosslinking reduces the crosslink efficiency since the antioxidant scavenges free-radicals during irradiation, thereby reducing the effective amount of crosslinking.  Diffusion of vitamin E into the polymer following crosslinking does not inhibit crosslinking.  More recently, solid-state pentaerythritol tetrakis has been marketed in bearing materials for TKA.  Due to the novelty of these antioxidant materials, long duration clinical and retrieval studies haven’t yet been published, particularly with respect to in vivo oxidation prevention.

The documentation of unexpected in vivo oxidation of thermally stabilized UHMWPE may be of concern to clinicians, industry, and patients, due to the potential for polymer oxidation to lead to increased incidence of fatigue- and wear-related failures.  Moreover, the shift in recipient demographic to a younger, heavier, more active patient places greater mechanical demands on bearing surfaces.  Thus, attaining a better understanding of the in vivo oxidation process and rate is crucial to ensure device longevity.  With this understanding, it is important to develop a predictive capability to accelerate aging under controlled laboratory conditions to test new and existing materials before they are employed in humans.

Over the last decade, our understanding of oxidation has advanced through systematic review of retrieval oxidation in the laboratory.  Specifically, we have observed exponential oxidation rates in gamma-air, gamma-barrier, annealed, and remelted materials.  These rates appear to be influenced by several different factors, including stress, free radical concentration, absorbed species, and radiation source.  We believe that these factors may lead to oxidation through a number of competing pathways, each of which is a subject of research in our laboratory.


Wearable Technology

Rehabilitation following joint arthroplasty is most often a ‘one-size-fits-all’ approach: all patients receive the same physical therapy. However, it is not well established if this is the highest value approach to healthcare. It is conceivable that different patients would need variable levels of postoperative PT to achieve optimal recovery. In addition, postoperative progress is often only gauged via single data points as measured in clinic/laboratory settings (e.g. passive range of motion (ROM) via goniometer) whereas motion throughout the day is only assessed anecdotally.

Our laboratory has developed and implemented a novel method for monitoring continuous long term joint function using inertial measurement units (IMUs). Prospective studies are in progress to compare knee and shoulder function before and after arthroplasty.  This data can be compared to a cohort of healthy individuals with no known joint arthropathy. 

Shoulder Modeling

The number of total and reverse shoulder arthroplasty procedures performed annually in the United States has been growing steadily over the last decade. One aim of our current work is to understand the mechanical and tribological interactions of shoulder arthroplasty with the patient. The development of algorithms for analysis of explanted components, patient outcomes, in vitro wear testing, and finite element analysis will provide a better understanding of joint behavior and potential impacts to the patient. This knowledge may allow for device designs and implantation methods to be adjusted to account for potential failure mechanisms.

Computational research in the laboratory occurs at various scales and couples the interaction between musculature and bone.  This helps identify how device implantation can affect musculature using dynamic modeling (top left), and in return, result in changes to stresses and strains experienced by the bone (bottom).  General or patient specific geometry and devices can be accommodated by utilizing CT scans and solid modeling (top right). 

New Materials

Angular Extrusion for Polymer Processing

Ultra high molecular weight polyethylene (UHMWPE), a common bearing surface in total joint arthroplasty, is subject to material property tradeoffs associated with conventional processing techniques. For orthopaedic applications, radiation-induced cross-linking is used to enhance the wear resistance of the material, but cross-linking also results in decreased relative chain movement in the amorphous regions and hence decreased toughness. Equal Channel Angular Extrusion (ECAE) is employed as a novel mechanism by which entanglements can be introduced to the polymer bulk during consolidation, imparting the same tribological benefits of conventional processing without complete inhibition of chain motion. ECAE processing at temperatures near the crystalline melt for UHMWPE yields increased entanglements over control materials, increasing entanglements with increasing temperature, and mechanical properties between never irradiated polyethylene and literature values for cross-linked polyethylene. These results support our additional research in ECAE-processed UHMWPE for joint arthroplasty and industrial applications.

Simulating In Vivo Behavior of Biomaterials

Total joint revisions due to infection pose significant burdens to the patients, hospitals, and the healthcare system. Transitioning from a two-stage infection treatment to a single stage procedure is one potential solution to these burdens.  Off-label use of a resorbable  calcium sulfate antibiotic carrier has been implemented in single stage and two-stage procedures in the United States.  It is unknown if adverse effects of calcium sulfate on the joint space during articulation exist.  Current studies in our lab seek to determine whether this new use of a biomaterial have the potential to change damage patterns or wear rates of artificial joints.

New Devices

Bearing Design

Total hip arthroplasty (THA) is an increasingly utilized and cost-effective treatment for osteoarthritis of the hip. An estimated 460,000 hip replacement procedures are performed in the United States annually.  Hip replacement has been shown to have the lowest cost per quality adjusted life year ($8,964 per QALY) compared to conservative treatments ($11,530 to $92,081 per QALY).  However, not all arthroplasty procedures have positive long term outcomes. Overall survivorship for THA devices is less than 93% at 7.5 years, with significantly worse survivorship for younger, more active patients (<90%). Overall, approximately 13% of all hip arthroplasty procedures are revisions, costing the U.S. health system approximately $3 billion in 2011.  Revision surgeries present more risk and morbidity to the patient and require higher utilization of healthcare resources.  Wear and/or failure of the bearing surfaces is one of the leading causes of revision, either directly because of poor bearing articulation or through the detrimental effects of wear debris on peri-prosthetic tissues, device fixation, and  the patient's immune system.

We are testing a new bi-material bearing to be employed in a THA device. It is possible that this innovative approach will reduce bearing surface damage and wear when compared to state of the art approaches in bearing design.

Algorithms to Determine Joint Alignment

Poor component alignment has the potential to increase the incidence of failure of total knee arthroplasty.  Thus, surgeons are eager to validate their surgical cuts and corresponding component placement.

Currently, there are several validation methods available to orthopedic surgeons, however each has its limitations. At the most basic level, a surgeon utilizes a surgical cutting jig to select the location and orientation of his or her cuts in two planes (e.g. sagittal and frontal plane on the tibia in TKA). The only way for a surgeon to validate their cuts utilizing this method is with intraoperative radiography, computed tomography (CT) scans, or fluoroscopy, adding additional cost and time to each surgical procedure. Moreover, these traditional methods are inaccurate, allowing several degrees of variability in the frontal plane. To achieve greater levels of accuracy, some surgeons have turned to computer navigated TKA procedures. Despite the purported benefits of navigation, there have been mixed results with respect to the accuracy of component placement when compared to traditional validation methods . Although some surgeons have seen marked improvements when using navigation techniques, the added time to surgery and high cost to entry are may be barriers to widespread use and adoption.

There exists a distinct need for an intraoperative method for quantifying the orientation of the prosthetic components used in TKA that is efficient, easy to use, cost effective, and quick with respect to total surgical time.  Recently, several companies have developed inertial measurement units (IMUs) to more effectively elucidate the orientation of surgical cuts. IMUs utilize gyroscopes, accelerometers, magnetometers, or some combination of all three to identify the orientation of the surgical cuts with respect to some known reference (e.g., gravity).

Current work in our laboratory centers on developing analytical and computational approaches to better measure surfaces cut by a surgeon. Our methods are derived from first principles, and are currently implemented in bench-top simulations and cadaveric models.