The history of total joint replacement has been about basically three things:

Firstly, finding the right materials that can function safely and effectively as biomaterials in long-term implants. Early attempts have used such disparate materials as ivory, cellophane, fascia lata, chromacized pig bladder, skin, other plastics, glass and metals in various forms.

The interface between the implant and bone, termed the fixation, has been the second issue. Press-fit, cement and bone ingrowth/ongrowth have been the principal approaches that have been tried.

Finally, the bearing surface has been an issue since the single side replacement was abandoned early on in favor of 'total' replacement.

The term 'total joint replacement' (TJR) is a bit of a misnomer, since the totality of a joint includes ligaments, tendons, muscle, synovium and bone. It would be more accurate to label what currently passes as TJR something that is more descriptive of what is done, such as 'complete bearing surface renewal', but everybody understands that this is what the term 'total joint replacement' actually means.


In the initial attempts to renew the painful or damaged joint surface, interpositional membranes with various tissues or materials listed above, were employed with the intent of guiding reformation of the natural articular surfaces. Surprisingly, many of these did enjoy limited success, but too frequently, were failures.

This was followed by single sided resurfacing or replacement, of which the Austen-Moore endoprosthesis, and the Aufranc Cup in the hip, and the MGH distal femoral resurfacing, the McKeever and Macintosh prostheses in the knee were examples.

Although resurfacing or replacing one side of the joint with a hard bearing surface enjoyed greater success than pliable interpositional membranes, this approach still did not produce the high level durable success that was being sought.

The modern era of TJR was entered in the late 1950's with the McKee-Farrar metal on metal (MOM) prosthesis and most notably with the introduction of the low friction arthroplasty employing a metal stem and high-density polyethylene cup, fixed to host bone with polymethylmethacrylate (PMMA) by Sir John Charnley. Over a very short time, this became the 'gold standard' for joint replacement, starting with total hip replacement and quickly spilling over to total knee replacement.

Over the next 20 years, the vast majority of prostheses consisted of metal or metal and plastic prosthetic components, fixed to bone by PMMA. Short-term results rose almost immediately from the 70% range to the mid 90%, as measured by lack of need for a second (revision) operation in 2-5 years. A great wave of enthusiasm for joint replacement swept the developed world in the 70's, as THE solution for previously insolvable problems seemed to be at hand.


However, it wasn't long before 2 longer term problems raised their collective heads. As joint replacements were being offered to younger and more active patients, component loosening, and osteolysis become clinically evident, with 50% failure with a minimum 5 year follow-up of THR's done in patients younger than 40 reported in 1980. Most of these failures were associated with loosening, and also showed in some cases, significant bone loss.

It was assumed that the loosening, that is, degradation of fixation, was the problem and that the bone loss from osteolysis was secondary. Attention turned to finding ways to eliminate loosening. This proceeded along 2 paths: understanding how to use cement more effectively and cementless fixation.

Both paths have produced dramatic success and aseptic loosening, of either cemented or cementless components designed for bone ingrowth or direct bone attachment, has largely been eliminated. Permanent fixation of prosthesis to bone is still technique dependent, but those techniques are well understood and largely practiced.

Along the way it was discovered that loosening was not an obligatory prelude to osteolysis. This had been missed for probably a decade because initially most of the osteolysis was associated with loosening, and also because the polyethylene particles were so small that the majority of them could not be visualized by conventional microscopy, whereas the fragments of PMMA were readily apparent.

There were a few other problems along the way that were also resolved. Metal stem breakage disappeared as metallurgical advances produced 'unbreakable' prostheses; metal sensitivity abated with the discontinued use of high nickel content stainless steel and the introduction of titanium based alloys for non-cemented devices, while techniques and instruments advanced to the point where attentive surgeons could achieve proper alignment and fixation of the components. Five and 10-year results were widely reported in the 90+% range for both hip and knee replacements.

Bearing Surfaces

However after 10 years, polyethylene wear and osteolysis began to appear in increasing numbers and it started to look like 15 - 20 years was the most that could be expected from a TJR. Efforts have been made along the way to improve the wear capacity of ultrahigh molecular weight polyethylene (UHMWPE), which had broadly replaced HDPE (high density polyethylene), including heat-pressing, introduction of carbon fibers, and altered processing. None of these proved successful; in fact they all showed equivalent or worse (in the case of knee replacements) wear in the clinical setting in spite of the fact that laboratory mechanical testing predicted improvement in wear.

Ultrahigh Molecular Weight Polyethylene (UHMWPE)

UHMWPE is now sterilized by either gamma irradiation or gas sterilization. Gamma irradiation produces cross-linking of the polyethylene chains and increases resistance to wear. However it also produces free radicals, which, when combined with oxygen, produce chain scission reducing wear resistance. When components were irradiated and stored in air the integrity of the component began to deteriorate in the box on the shelf, and this continued in the body after implantation.

In the mid 90's 2 separate solutions to the problem were undertaken, 1. Sterilize with gas (in preference to gamma irradiation) and 2. Irradiate and store in an inert gas, such as nitrogen. The first avoided creation of free radicals, but failed to obtain cross-linking, and the second achieved cross-linking and prevented the free radicals from combining with oxygen to cause chain scission. In the mid 90's the concept of increasing cross-linking by increasing the irradiation was introduced and components implanted.

Those components have now been in clinical use for minimum 5-year studies to appear, documenting significantly reduced wear. But they have not been in use long enough to know that this translates into reduced osteolysis, because even with conventional UHMWPE, osteolysis is not usually seen at 5-year follow-up. However, numerous studies correlate increased wear with increased incidence of osteolysis, especially in the hip. Therefore the goal to minimize wear appears a reasonable means to reduce/eliminate osteolysis.

More recently advanced processing has been introduced which permits strengthening UHMWPE by cross-linking with little or no remaining free radicals. However, clinical use of these advanced materials is still too recent to reach any conclusions.

Other Solutions (MOM & COC)

There have been 2 other roads taken to reduce wear: MOM (Metal-on-Metal) bearing surfaces, and COC (Ceramic-Ceramic) bearing surfaces. Both these concepts have been around for a long time, but enjoyed limited popularity.

One of the very early MOM hip prostheses, the McKee-Farrar prosthesis, was a metal-metal couple using a cobalt based alloy as both the structural and bearing surface material. However manufacturing techniques of the time did not allow uniformity of prosthetic component production and many of these devices failed prematurely. The bearing surfaces wore excessively producing a black metal sludge. Also presumably high friction from prosthetic components that did not mate properly ('stiction friction'), led to an increased incidence of loosening as, presumably, the increased friction forces were translated to the bone-cement interface and degraded it.

MOM bearing surfaces have enjoyed resurgence lately as manufacturing tolerances and bearing finishes have been improved to provide a low friction surface with extremely low wear rates. Five year follow-up reports have also been published for this bearing couple with very low loosening and osteolysis rates.

COC bearing surfaces have also been in clinical use for many years. One formerly popular design, the Mittelmaier prosthesis, had a screw in all ceramic acetabular components that had an elevated aseptic loosening rate.

Ceramic as components have also been plagued periodically by fracture, which presents some unique problems that will be reviewed below. However, several large series, particularly from France, have reported outstanding multi-decade results with a low incidence of both fracture, loosening and wear.

Recently acetabular components have been introduced where the ceramic is embedded in a porous surfaced metal shell. This protects the ceramic liner from damage during insertion and provides an ingrowth surface. Again results at 5+ years show high success, low aseptic loosening and low osteolysis rates.

Today's Choices

It appears that in 2006 there are 4 candidates to choose from for an ultra-durable hip replacement prosthesis: MOM, COC and metal or ceramic on highly cross-linked UHMWPE. On the basis of both laboratory testing and available clinical evidence, it would appear that ANY of these could be the ultimate long-lasting THR.

How are surgeons and patients supposed to choose?

I do not believe that there is any evidence today that one is more likely to be more successful in the long run than the other three. It is possible that all 4 will be equally successful, as measured by clinical outcome, and that is what I think is a reasonable supposition.

It is possible that with 10 and 15 years follow-up that one of the 4 will be more successful than the other 3, but such documentation will have to await the definitive follow-up, which cannot be expected for another decade.

The question is what to do now?

Perhaps it doesn't make any difference, but there are differences.

Both MOM and COC solutions are significantly more expensive than metal on highly crosslinked UHMWPE, while substituting a ceramic head for a metal head is intermediate in cost. Translated to the whole population of TJR candidates, these cost increases could be quite significant.

Ceramic has a history of fracture. Manufacturers claim that this has been largely eliminated through the use of higher strength, smaller grain size materials, and certainly it is rare but not zero. However, ceramic fracture is a clinical disaster. Manufacturers strongly recommend replacing the fractured femoral head with a new ceramic component, only if a new trunion can be provided, and discourage use of a more conventional UHMWPE acetabular component after such failures. This means revising an otherwise perfectly good femoral stem (and an occasional acetabular cup) with all the potential complications that that engenders.

MOM prostheses have clearly demonstrated both high short-term success rates and low wear/osteolysis rates. However, all MOM bearings produce significantly higher metal ion levels in both blood and tissues, and these elevated levels, that can be 5 - 10 times normal, persist for as long as they have been followed, in some cases over decades. It does not appear to simply be an issue of 'bedding in', but rather ongoing wear of the bearing surface.

Promoters of the MOM solution cite the extremely low wear volume of the metal debris from the MOM bearing, noting that it is only a fraction of the volume of that produced even by the highly crosslinked UHMWPE, and this is certainly true. However the typical particle size of the metal wear debris is also only a tiny fraction of the particle size of UHMWPE.

The resulting high number of extremely small particles, even when present in such low total volume, exposes a very large surface area of metal to body fluids, which undoubtedly contributes to the increased chromium, and cobalt levels measured in local and distant tissues and body fluids. Both Chromium and Cobalt are presumptive human carcinogens. Epidemological studies of patients with older MOM hip replacements have produced mixed results: some show slight elevations of blood related malignancies and others show no adverse effects. The situation is further complicated since the very low dose rates produced by these devices present the possibility of late effects, due to the very long, multi-decade latency known for Chromium induced malignancies in industrial exposures.

There is no compelling indication for MOM prostheses in the average patient presenting for THR (average age 70) and the higher cost cannot be justified. However, the rationale for use in the young, because of low wear, is also problematic because of the potential for late stage malignancy.

There is also a new potential problem that has recently surfaced, one of metal allergy. This is an unknown long-term liability. Because the exposure to the metal ions is continuous, it is apparently something that can develop at any time, related to transient sensitization episodes produced by some illness or medications. Although conventional prostheses are often made of cobalt-based alloy, the incidence of allergy is extremely low and does not appear to increase with time. However the higher level of metal ion exposure with the MOM prosthesis apparently makes this a very real potential problem.

Finally, the introduction of ceramic femoral heads (to replace conventional metal heads) while apparently reducing clinical wear rates, does not appear to reduce the need for revision surgery, and thus, it seems hard to justify their higher cost and the rare risk of fracture.


Based on an overview of the available options it appears to me that metal-highly crosslinked UHMWPE bearing surface offers the best risk/benefit ratio of the 4 potentially long-wearing bearing surfaces in hip replacement. It is the least expensive, the easiest to revise, should that become necessary, has a long history of successful use as a bearing surface system and is well tolerated as long as wear is kept below a threshold that certainly appears to be achieved by the new highly crosslinked UHMWPE offerings.

Conventional UHMWPE, with durability now improved by better sterilization and storage techniques, bearing against cobalt-based alloy, remains the best choice for knee replacement.

David Hungerford, MD
January, 2006
Baltimore, Maryland