The first documented successful implementation of a transcutaneous skeletal device may have been Malgaigne’s double-sided hook. Designed in 1840, this construct penetrated the patient’s skin and clamped the superior and inferior poles of the patella, providing compression through a fracture. In 1843, Malgaigne¹ developed an early type of external fixation by applying a transdermal implant to fix a fracture, in which metal penetrated both skin and bone. Later, a more recognizable external fixation instrumentation was described by Codivilla and Steinmann between 1903 and 1910.^2,3,4,5,6 External fixation is an accepted form of treatment of some fractures, as well as a mainstay in the field of deformity correction and limb lengthening, albeit for a temporary period of time. The problem is not necessarily the bone-implant interface, which remains stable after true osseointegration occurs. Rather, difficulties remain with the soft tissue-implant interface, particularly if excessive motion exists between them.

As reviewed by Murphy,⁷ experimental work on bone fixation of external limb prostheses in the past has been sporadic and performed on a small scale. The first documented skeletally linked transcutaneous prosthetic attempts were likely the pilot studies performed by Cutler and Blodgett⁸ as early as 1942 at Harvard University. These surgeons tested stainless steel and vitallium screws inserted
into the intramedullary canal of 18 dogs (Figure 1). Vitallium retained better stability, and the researchers determined that the implant must remain motionless relative to the bone to prevent loosening.^7,8 By 1949, the United States Veterans Administration felt the surgical challenges for success in humans were too great and suspended further investigation.

In 1946, Dümmer in Pinneberg, Germany,⁷ studied metal-implanted limb extensions in sheep and extended the study to human patients, four of whom had prostheses attached to metal implants. When one implant became infected, the study was discontinued, and all the implants were removed.

In the United States, Mooney et al⁹ fitted a prosthesis to the humerus of an individual in 1967 who had three amputations. The implant was removed around 8 months later, after signs of deep infection developed. Esslinger¹⁰ performed a further study of penetrating implants in dogs; Hall,¹¹ and Hall et al¹² achieved considerable success with penetrating implants that were inserted into the hindlimbs of Spanish goats. In 20 goats, only two implants failed over 14 months, during which time the animals were very active.

These data provided little encouragement for further studies until an unrelated observation by Brånemark¹³ suggested the possibility of long-term tolerance of a transdermal metal implant. Brånemark¹³ was studying the microcirculation within bone, which involved the insertion of an observation tube into the bone marrow of the tibia of a rabbit. The optical device was made of titanium and was found to be almost impossible to remove after several weeks of study because the titanium appeared to have “integrated” with the bone. This was a breakthrough discovery in the material science of osseointegration, which led to extensive studies involving a team of researchers led by Brånemark in Gothenburg, Sweden, in association with the Department of Orthopaedics at the University of Gothenburg showing that commercially pure titanium was well tolerated within living bone and relatively resistant to infection.^14,15,16,17 Titanium forms a resistant surface oxide layer, which in living tissue can be augmented by a peroxide layer. This peroxide layer is thought to be a hydrated titanium peroxy matrix that does not tolerate surface pathogenic activity, does not inhibit endothelial cells or osteocytes, but inhibits macrophages. This characteristic allows the material to be tolerated in living bone without the interposition of a fibrous tissue layer, so, according to Brånemark et al¹⁴ and Linder et al,¹⁸ this close contact has been termed osseointegration with the implanted material.

Another important property to achieve osseointegration is the implant surface roughness divided into three levels depending on the scale of the features.^19,20 Macrolevel roughness is related to the implant geometry which is essential for early mechanical stability. Microlevel roughness maximizes the interlocking between mineralized bone and the surface of the implant.

At the nano-scale, the more textured surface increases surface energy, which increases osteoblastic cell proliferation, differentiation, and adherence. Based on the abovementioned characteristics, two different implant systems were established; both systems share the microlevel and nano-scale features but differ in the macrostructures. Early in the 1980s, a screw-fixation device was used for dental implants, and the same structure was implanted in 1990 into the femoral residuum of an amputee. The osseointegrated prostheses for the rehabilitation of amputees (OPRA) screw fixation implant system remains the oldest commercially available osseointegration device to date. On the other hand, hip arthroplasty uses a press-fit implant design in the femoral shaft relying on a specific macrostructure geometry that allows early axial and rotational stability of the implant. This led to the development of the osseointegrated prosthetic limb (OPL) implant system in Australia in 2010, which is the most widely used osseointegration implant to date worldwide. These are the predominant design philosophies of the most widely used osseointegration implants.

The optimal transcutaneous implant surface coating is another area of ongoing research. While current work is focusing on the significance of bioresponsive, endogenously triggered, surface additives to reduce infection risk at the intramedullary implant interface, the abutment-soft tissue interface is an important area of focus considering that most infections occur at this interface. Though published work to date has not yet revealed any important benefit from surface coating or treatment at the intramedullary implant level, the OPL system introduced a highly polished surface with a titanium niobium oxynitride coating to provide bacteria-repellant properties at the abutment-soft tissue interface.²¹

The skin-implant interface has been the greatest challenge for osseointegration. Numerous studies have been conducted at the interface between the skin and the penetrating metal. Two divergent philosophies have been tested, one attempting to provide a full seal of the
skin-implant interface,²² and the other recognizing the development of the long-term stoma, or skin penetration site. To seal the implant-tissue interface, several techniques have been used including acrylate adhesives which provided only temporary attachment to the metal surface, and titanium gauze and textile meshes of various polymers which provided a degree of cellular invasion into mesh materials, with no success. A porous titanium transdermal surface has also been used but was later abandoned because of poor results.²² None of these techniques has yet provided any clinical advantage over the simple skin perforation of surgically stabilized soft tissues over a smooth abutment²³ used in both the OPRA and OPL systems.

In a recent evaluation of 51 patient with transfemoral amputation who were treated with the OPRA (Integrum) implant system and followed for a minimum of 2 years, before-and-after assessments demonstrated statistically significant improvements in prosthetic use, mobility, and functional status—plus fewer problems—compared with their previous socket-based prostheses.²⁴ As expected, superficial infections were the most frequent complications. Twenty-eight of the 51 patients (55%) had one or more superficial infections that, without further complications, were successfully treated with antibiotics. Deep infections developed in four patients (8%), and in one of these patients, the infection loosened the fixture. However, in the remaining three patients, the fixtures were retained. In three of four patients, aseptic loosening was the reason for implant loosening. Most of the femurs demonstrated some degree of radiographic bone resorption. However, no periprosthetic fractures occurred, and no correlation to impaired rehabilitation results or implant loosening could be found. Instead, in accordance with previous experiences from oral applications, radiographic bone resorption seemed to diminish with time.

In another study, Al Muderis et al²⁵ reported on 86 patients (91 implants) using the pressfit system with a minimum 2-year follow-up examining the safety of osseointegrated implants for transfemoral amputees which demonstrated 36% uneventful course without complications, 34% one or more infections, and 30% had other complications but no infection. Hence a grading system for the classification of infection was devised, graded as 1 to 4 in severity, and categories A to C in recommended management (Table 1). The results show grade 1A in 27% of patients which is regarded as 87% of all infections, grades 1B and 1C in one patient each (1% of patients individually), and 5% grade 2C infection. There were no grade 3 or 4 infections. Other complications were described as stoma hypergranulation (20%), redundant soft tissues (16%), periprosthetic fractures (3%), hardware breakage (2%), and failure of osseointegration requiring revision (1%). A smaller retrospective study of 22 patients who received the OPL implant showed a significant improvement in short-term clinical outcomes in postoperative quality-of-life and function, with low associated risk of severe adverse events.²⁶

There are three design paradigms for currently relevant osseointegration implants: a threaded screw, a press-fit intramedullary stem, and a spring-loaded platform inducing constant compression.²⁷ The key design and surgical features of each are described below and summarized in Table 2.

Engineering alone is insufficient to mitigate all complications related to the use of a surgical device, and careful attention to the surgical technique is arguably as important. Given that the OPRA system by Brånemark and the OPL system by Al Muderis have well-defined protocols, outcomes, and published data, the focus will be on these two systems.

The OPRA system has been extensively used for dental prosthetics, maxillofacial reconstructions, prosthetic ears, and hearing aids. In the field of limb prosthetics, phalangeal and metacarpal implants have been used for patients with digital (finger and toe) amputations and have been particularly effective for thumb prostheses. Radial and ulnar implants have been used for patients with forearm amputations, with humeral implants being used for patients with upper arm amputations. All of these patients were treated with custom-designed implant systems.

In the lower limb, implants in the residual femur for patients with transfemoral amputations represent most of the patients treated, and the OPRA system (Figure 2) is available on an off-the-shelf basis for this amputation level (it is not currently available for other lower limb levels). Because of the high loading imposed by lower limb prosthetic use, load is slowly and progressively applied. From commencement of loading to the unrestricted use of a prosthesis without aids may take 6 to
12 months, depending on the patient’s body weight and initial bone quality.

All the implant components were originally fabricated from commercially pure titanium to avoid any potential electrolytic effects or effects from the alloy; for this reason, stronger titanium alloys were not used in the system. The instruments used to handle these components also were manufactured from the same titanium. More recently, based on basic research²⁸ and clinical practice over the past 10 years, stronger alloys are being used. In 2011, laser etching surface finishing was added to improve osseointegration.²⁹

The fixture, or implant, that is inserted into the medullary cavity of the residual long bone consists of a tubular component with a self-tapping thread on its surface to engage the endosteal cortex. The fixture serves as the connecting point for the abutment, which penetrates the skin of the residual limb and protrudes to permit attachment of the artificial limb.

The abutment is located in the lower end of the implant in a hexagonal recess and retained by a long retaining bolt, passing through the abutment, to engage an internal threaded section in the lower portion of the implant. When in place, the head of the retaining bolt forms the lowest point of the protruding abutment. This arrangement enables the abutment to be changed, from time to time, without disturbing the implant itself. A damaged abutment can be replaced in an outpatient setting, without anesthetic. If use of the system has to be discontinued, the abutment alone can be removed, and the penetration site heals rapidly.