Round 9: Update on Hip and Knee Arthroplasty

by Harpal Singh Khanuja, MD
Assistant Professor, Department of Orthopedic Surgery
Johns Hopkins University School of Medicine.

Release Date: July 14, 2006
Expiration Date: July 14, 2008

Dr. Khanuja has no significant financial interest or relationships to disclose.

In this update on hip and knee replacement, we’ll talk about:

  • Different types of prosthetic joints;
  • Indications for the various types of joints; and
  • Emerging techniques and technologies.

A tremendous number of hip and knee replacements are done each year — probably more than 500,000. In 2001, 326,000 total knee replacements were done; they are increasing at faster rate than total hip. In the next 20 years, close to three quarters of a million joints — primary joints — will be replaced each year. The incidence of revision also is increasing pretty significantly. But total joint replacement is really one of the best replacement surgeries that we have in terms of patient outcome. The durability of these implants is really apparent.

The Mayo Clinic has reported an 80% to 86% survivorship rate for the 2,000 hip replacements they’ve put in over 30 years(1). From an orthopaedic standpoint, survivorship means the implants have not been revised and do not need to be. It is not always a comment on their functional outcomes, but I think most functional outcomes are pretty good as well. If you look at 30-year post-ops from the University of Iowa(2), only 10% of the femurs have been revised. The acetabulum has a much higher revision rate, at 26%. However, these are 30-year results, which are pretty good. For patients younger than 50, the results are not nearly as good; they have about the same revision rate, but at 25 years. Thus we know that joints have not been holding up as well in younger patients, meaning people younger than 50. But the outcomes are really improving with the newer prosthetic designs. That’s particularly true since the introduction 15 years ago of cementless arthroplasty, which has yielded a 97% survivorship rate on both the femoral and the acetabular side.

Hip Replacement
To clarify some of the terminology: When we talk about hip replacement, we typically think of total hip replacement, which involves a femoral component and an acetabular component.



We’re replacing both sides of the joint. As for fixation, there are two types. Originally, we had cemented fixation, (see below) where the cement acts as a grout and is really an interface that does not stick to the prosthesis as much as it provides an area in which it can wedge in.


Now we have biologic fixation, (below)where there are various coatings on the implants.


Initially it’s pressed together tightly; eventually, the bone will grow onto the prostheses. It’s a biological interface, which means they may remodel in time and there is one less interface — just bone and prosthesis rather than bone and cement, and then cement and prosthesis. In 1999, leaders in the field agreed that a hybrid fixation is probably best.


A hybrid is cementless on the acetabular side and cemented on the femoral side. In the past six to seven years, we have begun to see the durability of cementless implants, and we are moving more to non-cemented fixation. For me, the primary determinant is bone quality; non-cemented implants need strong bones to wedge into and grow to. A bipolar hip replacement is called bipolar not because it is on both sides of the joint but because the setup is a ball within a ball.

That ball within a ball sits just on the femoral side, so this is just a femoral implant. The acetabulum is not done on this side; to have that, you need to have normal acetabular cartilage. So the bipolar articulates against normal cartilage. When might you have to replace a femoral head when you have normal articular cartilage? When you have a hip fracture in an elderly patient. In the elderly, the femoral head has at least a 20% likelihood of osteonecrosis (or dying) and subsequent fixation failure. So, we generally use age 65 or 70 as a cutoff for when we’ll take this out and put in a bipolar component that will articulate against the native acetabulum.

Another instance in which we would replace just the femoral side is in avascular necrosis. This is a resurfacing hip arthroplasty.

If you have a subchondral fracture or just a deficit in the femoral head, you can take off the damaged bone and the articular cartilage and basically put in a metal cap that then articulates with the normal cartilage. This is really the only indication for this prosthesis. The results are okay. But this is a bridging procedure. Both with the bipolar arthroplasty and with this, the acetabular cartilage eventually will wear. The typical bipolar patient is elderly and less active, so they may not wear as much. But in a younger patient, this is really a bridging procedure. At five years, about 80% are doing well; at 10 years it drops off to about 70%. If someone in their mid-30s comes to me with avascular necrosis, I may consider doing this, telling them they may have a little bit of pain and that eventually they will need a hip replacement. But it buys them time.

A resurfacing arthroplasty can do a total resurfacing when you replace both the femoral head and the acetabular component.

It’s less invasive, and there’s less bone removed. Later it could be converted to a total hip arthroplasty. Right now this is not FDA approved, but it will be soon and you are going to see more of these, for better or worse. It is actually a very old technology, but it had a high failure rate because the plastic between the two metal pieces would wear. The answer to that has been a metal-on-metal prosthesis. There’s only a three-year follow-up so far on the current generation, and the three-year survival rate is excellent, but we have 25 to 30 years of data on conventional hip arthroplasty. This also is technically more demanding than a conventional hip replacement. With a conventional hip replacement, there’s a lot of room to play with the modular components to increase leg length and balance tissue tension, which is important for dislocation and stability of the hip. You do not have that ability with resurfacing and you have to match the patient’s anatomy, which makes it technically more challenging.

To repeat: Bipolar arthroplasty would be used for a femoral neck fracture. A hemi-resurfacing is just for the femoral side, and just for avascular necrosis without articular cartilage damage.

So, when do we use a total hip? The conventional treatment has been when you have symptoms and non-operative management has failed. For the most part, I still tell my patients, “You will tell me when it is time to have your hip replacement or your knee replacement,” because it is really a matter of what quality of life they want. We have moved away from delaying the procedure as long as possible, in part because people today are much more active and are not willing to accept as much debility as they did 20 years ago. We have seen that, with cementless arthroplasty, the long-term results are good. Also, a number of studies have shown that the cost of waiting, in terms of functional debility and quality of life, has increased. And, there is an economic aspect to it in terms of missed days from work. So, we are expanding our indications and doing them in younger people. There also may be advantages to getting people when they are less debilitated, as ultimately they will have better outcomes.

As for osteoarthritis: I do not have a problem replacing the hip of someone who is over 45, provided their expectations are realistic and their activity level is appropriate. If they want their hip replaced so they can run marathons, I won’t do it. But a patient with juvenile rheumatoid arthritis or someone who is significantly debilitated is not going to be as active. In patients younger than 45 or patients that are active, I might consider doing resurfacing on both sides, but certainly not for older patients because we have long-term follow up on the total hip replacement. There are consequences that accompany metal-on-metal replacements. In terms of restrictions after total hip replacement, I do not like people doing impact activities. There is no good data on how impact activities affect wear, but they just cannot be good for a prosthetic joint. I do not have a problem if people want to ski — if they are already good skiers. But it is amazing how many people will have a joint replacement, feel better, and say, “I want to try skiing” or “I want to try horse-back riding.” This is not the time for that. If they did those activities before the replacement, they can continue them in moderation. Most patients, however, will self-limit their activities.

Knee Replacement
On the knee side, the long-term results for total knee arthroplasty are very good as well. We have 15 to 20 years of data, and 90% survivorship. Some 25-year studies show 95%. Again, 25 years ago, we were doing these in less-active patients, so that should be taken into account. One thing that has fascinated me is that there is only one prosthesis that is identical to what it was 20 years ago; all the others have gone through various changes due to technology. That means we it will be another 20 years before we know how the current generation of protheses is holding up. However, all the design changes supposedly have been made to sustain the longevity of the prostheses. Also, we have excellent results from rheumatoid patients as well. Studies differ as to whether rheumatoid patients do better than osteoarthritic patients; part of it depends on their initial debility.

When talking about knee replacements, we refer to unicompartmental knee replacements, patellofemoral knee replacements, and total knee replacements.

Unicompartmental Replacement
A unicompartmental knee replacement replaces one compartment: a medial or lateral compartment. You basically resect the sclerotic bone and resurface it. Clearly, you do not want to do this for something that would affect both sides of the joint, so you would not do it for an inflammatory process or avascular necrosis (secondary to steroids or affecting both sides of the joint). A small spontaneous lesion of osteonecrosis can be treated with a unicompartmental approach, but you would not want to do this for chondrocalcinosis because you will load a diseased side. The benefit of this type is that it’s done through a smaller incision, and normal cartilage is maintained. It certainly is less invasive than a total knee replacement because there’s no replacing at all.

Using strict criteria, 5% or 10% of total knee replacement patients are candidates for a unicompartmental replacement. Criteria include:

  • Weight. About 80 kilograms often is used as a cut-off. The surrounding cortical bone is the strongest bone, and this is not on cortical bone. A total knee replacement might be more durable for someone who is heavier.
  • Range of motion. Typically, you do not want more than a 15-degree flexion-contraction. You want at least 90 degrees of motion and no deformities. If you use more strict criteria, you will have better results.
  • Age. The outcomes are better in patients who are older because they are less active. For an older patient, this is going to be their definitive treatment; we’ll do a unicompartmental arthroplasty, and that will be it. But for a 40- or 50-year-old, this is more of a bridging procedure in that, down the road, it can be easily converted to a total knee replacement.

In some patients we use a uni-spacer — a metal disc that is shaped like an articular condyle and is placed between the sclerotic bone to unload the diseased compartment.

There’s been some initial enthusiasm over this, but studies have been conflicting. The designer reported an increase in function in 71% of his patients, but 21% needed to be revised within two years(3). In another study(4), 32 of 34 patients still had pain with ambulation, and there was no statistically significant improvement in functional or pain scores pre- and post-operative. I am not a fan of this implant.

Patellofemoral Replacement
The patellofemoral replacement is a much less common procedure than the unicompartmental tibiofemoral replacement, partly because it is very rare to see isolated patellofemoral arthritis. It is a finicky operation that requires very important alignment and tracking of the patellofemoral joint. Also, it can be very design-sensitive, with different designs having markedly different rates of success. Tibiofemoral disease typically progresses over time, so there is a high re-operation rate — probably 25% at 15 years, which in selected indications may be okay, but it certainly does not compare with total knee arthroplasty. You might consider this in someone younger than 50 who has isolated patellofemoral arthritis, again as a bridging procedure. The reason I would not use it in anyone else is because a total knee arthroplasty works well, so a 70-year-old who has patellofemoral arthritis is going to do very well with a total knee arthroplasty. I would rather take known results than risk doing a patellofemoral arthroplasty.

Total Knee Replacement
Total knee arthroplasty replaces both sides of the joint, at both the femoral and the tibial side. I routinely replace the knee cap, although it does not have to be replaced if you have up to grade 3 chondromalacia. Another technique you will be hearing more about is fixed-bearing or mobile-bearing knee replacements. The plastic is fixed to the base plate on the fixed-bearing knee; in the mobile-bearing knee, it will move. The difference is whether it rotates around the base plate. The reason for the mobile design is that the rotation actually conforms more with a person’s range of motion and flexion. It increases the contact area and decreases the stress on the polyethylene, which should lead to less wear. That has absolutely been the case in in vitro studies and in knee simulators. The survivorship of this implant is excellent, comparable to a fixed-bearing knee replacement. The one problem with these mobile bearings, however, is that they are mobile; they can dislocate or pop out of place. This problem has decreased some with newer designs.

What are the indications for total knee? It is a last resort, but again, patients are more active now and the indications are expanding.

  • Age. I have no problem with patients over 55, as long as they have appropriate understanding and expectations of what a knee replacement will do. It is important to maximize your conservative treatment, yet not wait too long. The problem is that the studies do not indicate what is too long.
  • Range of motion. One of the important predictors of range of motion post-operatively is pre-operative range of motion. Therefore, if a patient, in waiting for knee replacement, loses range of motion, that patient is going to have less range of motion once it is replaced.
  • Valgus arthritis. When I see a patient with valgus arthritis, I tell them they should consider having their knee replaced sooner than later. If I see this medial joint space start to open or widen, then you are almost getting to the point where it is too late.

Emerging Techniques and Technologies
Now, to look ahead: The future is here, but it’s not necessarily exciting. Three factors have changed what we are doing in joint replacement:

  1. Outcome studies and long-term follow up
  2. Innovations and advances in technology
  3. Direct-to-consumer advertising

Patients are being influenced by advertising. Jack Nicklaus tells us that the ceramic hip is the best thing to have. But this information comes from all sides — from industry, from hospitals, from physicians. Everyone is at fault.

As for the technological advantages: We’ve already mentioned the move to cementless design. Certainly, with computer assistance we are getting better at molding anatomy, which enhances the longevity of prostheses. Bearing surfaces also are very important; they have been the limiting factor in the longevity of total joint replacements. The No. 1 reason for failure of joint replacements is the aseptic loosening of the replacement — not infection, but aseptic loosening caused by prosthetic wear. The femoral head can wear out and the plastic become fractured, and then these little plastic particles are ingested by macrophages, producing an osteolytic reaction. They actually activate osteoclasts that cause what we call osteolysis.

To reduce this problem, the plastic has been made stronger and wearless. The determinant is cross-linked polyethylene, which cross-links the polyethylene chain and gives it a higher molecular weight. This has been shown to decrease wear. Simulators have shown negligible wear after 12 years; clinically, they’ve been out for about five years, and there has been a reduction in what we assume is wear. In a patient, wear is measured by the femoral ball’s penetration into the plastic over time. There has been about a 60% to 70% reduction in that penetration compared with the conventional polyethylene(5). What that is going to mean in 15 or 20 years, though, is unknown.

The other development has been the use of metal-on-metal bearings, which are harder and have a significantly lower wear rate. The newer designs, though, were FDA-approved only five to seven years ago so we do not yet have long-term follow-up data. However, the engineers have learned much more about the tolerance between the acetabular component and the femoral head, and have made adjustments to maximize the wear of these prostheses. Metal particles are much smaller than polyethylene particles, and when you have metallic wear, you do not get an osteolytic reaction. The main concern with metal bearings is the systemic concentration of ions. Metal ions chromium is known to cause mutagenesis to various cells. We did a study at Good Samaritan that showed it caused T-cell death in the tissues surrounding the metal. Delayed hypersensitivity reactions also have been described. These reactions are probably 1 in 15,000, but we do not know why people have them.

Another solution — and one that has even less wear than metal — is to use ceramic. This is what Jack Nicklaus had. The problem with ceramic is the fracture rate. These are very brittle, and when they fracture, they just explode in the hip. The fracture rate was probably 1 in 2,000 five years ago; now it has gone to about 0.01%. Therefore, it is really not that prevalent. But if you have a ceramic fracture, you can never clear those little pieces of ceramic, and it actually significantly decreases the revision survival. The position of the components is critical with ceramics. If you do not have a perfect position, if there is any point where the ceramic head butts against the acetabular component, you may get a chip, and that can start the process of fatiguing the ceramic, leading to subsequent fractures. Some hip simulator studies have shown that if you do not have a perfect coupling, if there is some room for play, you will get increased wear in the ceramic, and we know from gait studies with fluoroscopy that there is some uncoupling of the joint replacement. So, while the current generation — which is four years old — looks good, the question is what the situation will be in another five or 10 years.

When people want these new bearings, we have to caution them. These developments are not tried and true, and the patient really has to be willing to take risks. At this point in time, what is tried and true is polyethylene. So, I would use metal on cross-link polyethylene for most cases. For patients younger than 50, I will consider metal-on-metal. I am more comfortable with metal than ceramics simply because you have such a horrible problem if the ceramic shatters. However, metal would not be a good choice for women of childbearing age, as we do not know what impact these ions might have on a developing fetus. And metal is not recommended for people with poor renal function. The ions are found in the urine, in the liver and the kidneys. They are cleared by the kidneys, so people with decreased renal function would not be good candidates.

Computer-Assisted Arthroplasty
Computer-assisted joint arthroplasty is another development we hear a lot about.

It is passive navigation. There are location devices on the instruments and on the patients, which help with spatial orientation. The goal of this is to enhance the accuracy of your cuts, and how you put your prostheses in. A colleague who was criticized by a patient for not using computers when he did his total knee said: “You have navigation when you are in your car?” And she said, “Yeah.” He goes, “Do you use it to go from work to home, and home to work?” His point was that if you know what you are doing, you do not necessarily need it. Now, I have absolutely no doubt that computers will become an essential part of total joint replacement, but we’re just not there yet. Certainly the technology helps when you have a significant amount of deformity and you cannot find the landmarks that we normally use, but I don’t see a gain for routine cases. And it does increase your operative time. But it’s an excellent research tool to help figure out how to make knees work better. Also, we need to use it to look at knees dynamically. Right now, we are looking at cuts, but if we can figure out the kinematics with the computer, that will help long-term outcome and function.

Minimally Invasive Surgery So this brings us to minimally invasive surgery. Everyone does this surgery now. But what I want to address is “mini-incision surgery” — which I’m going to call “less-invasive surgery.”

The theoretical benefits are:

  • Less tissue damage;
  • Decreased blood loss;
  • Less pain; and
  • A quicker recovery.

It makes sense that there is less tissue damage. We do the same total hip through a smaller incision.

We go through the same muscular plane and, importantly, it is the same hip replacement. But when you do a procedure where you tell the patient, “You have a risk of pulmonary embolism, infection, dislocation, and death,” I do not think you can call it a “minimally invasive” procedure. I say they are “less invasive” than they used to be. Minimally invasive, on the other hand, is when you actually go through muscular planes; you spread the muscular planes, move them out of the way (see image below). To me, that is less invasive because you are not cutting muscles. Intuitively, patients should do better with that.

The truth is, you are still putting in a big component no matter how you do it, so you are still maximally invasive on the bone. That has not changed — and I do not really think it should change too much because the outcomes are so good now. If we were to tailor our prostheses, as some have suggested, to these small incisions, we might be taking a step backward, but we would not know for another 15 to 20 years. My conclusion at this point, honestly, is that a smaller incision is nice, but it is not medically beneficial.

Two-incision Approach
For less-invasive procedures, I use the two-incision approach. Instead of cutting muscle, you try to get around muscle so it is not too damaging. You might wonder why we need two incisions when it can be done with one, especially when the two incisions add up to more length than the one. But the incision size does not matter; what matters is what you do under the skin. The front incision on the anterior aspect of the thigh lets you get to the acetabulum and prepare the acetabular component.

To get to the femur, you need a second incision if you want to minimize the amount of dissection.

The incisions can be pretty small. You have basically a mobile window where you move your window of tissue around to get the visualization you need. Because it is limited with the two-incision approach, most of us use fluoroscopy with this approach. The femoral preparation is much the same way when it’s done posteriorly through incision. You can see what you are doing on the femur through the front incision, and then direct your broach and prosthesis through the back.

The results have been phenomenal. In the beginning, 91% of these patients were discharged within 24 hours of surgery(6). One surgeon reported that for the hundred cases he presented in 2004, all of them went home the next day, and all were in outpatient therapy by two weeks. People were driving on average after one week. However, early on these procedures were done on highly selective groups; participants had single major joint disease, were otherwise healthy and lean, and were all highly motivated and had a good support system. Your average 70-year-old — with, say, a history of chronic obstructive pulmonary disease, congestive heart failure, or other co-morbidities — is probably not going to go home the next day. If my father told me he was having a two-incision hip replacement, at this point I would be pretty reluctant about it — and that’s what I tell patients, too.

It’s worth remembering that not all surgeons have the same skills, and that can come into play. In the hands of people who are not doing a lot of these, patients’ results are compromised. As an orthopaedic community, I think we need to temper it a little. And that is happening; the journals are starting to focus on the problem of total hip replacements done through two-incision approaches(7).

As for knees, they, too, can be done through a smaller incision. But again, the size of the incision does not matter; it just has to be done properly. On average, patients who had a mini-incision were smaller. Studies have shown that mini-invasive knee surgery is feasible, but others have indicated that the procedure increases the risk of malalignment of the prosthesis(8) and there is no increase in short-term recovery(9).

Overall, what has been most beneficial for short-term recovery has been the anesthesia protocol; that’s what has changed the most. We are all doing smaller incisions, we all have changed our postoperative routine, and we are all changing our postoperative pain protocols. I think patients are more comfortable. And, we have people moving more quickly; we are trying to rehabilitate them faster. That is probably the single most important thing in terms of increase in short-term recovery.


  1. Berry DJ, et al. J Bone Joint Surg Am. 2002 Feb;84-A(2):171-7
  2. Callaghan JJ, et al. The Journal of Bone and Joint Surgery. 2004;86:690-695.
  3. Hallock RH, Fell BM. Clin Orthop Relat Res. 2003;416:154-63.
  4. Sisto DJ, Mitchell IL. J Bone Joint Surg Am. 2005;87(8):1706-11
  5. D’Antonio JA, et al. Clin Orthop Relat Res. 2005;441:143-50.
  6. Berry DJ, et al.. The Journal of Bone and Joint Surgery 2003;85:2235-2246
  7. Bal BS, Haltom D, Aleto T, Barrett M. J Bone Joint Surg Am. 2005;87(11):2432-8
  8. Dalury DF, Dennis DA. Clin Orthop Relat Res. 2005;440:77-81
  9. Tenholder M, Clarke HD, Scuderi GR. Clin Orthop Relat Res. 2005 Nov;440:67-76

Updated: August 16, 2012

Arthritis Center

About Arthritis Center

Founded in 1998, the Arthritis Center at Johns Hopkins is dedicated to providing quality education to patients and healthcare providers alike.