Hip resurfacing in women

Thomas P Gross MD        8/24/2013              grossortho.com

A recent opinion posted on surface hippy correctly pointed out that women have a higher risk than men. However this is changing as we learn the nuances of hip resurfacing and improve our techniques. Some surgeons avoid women; in my practice odds of success are now virtually the same in women and men. When making a decision regarding hip resurfacing (HRA) vs. standard hip replacement  (THR) there are pros and cons that a patient should consider, some of which were identified by the author.

No surgeon has perfect results. The most important factor in determining the odds of a successful outcome in THR is surgeon experience and skill. This should be obvious, but it has also been demonstrated in multiple scientific studies. This is even truer for HRA because it is a newer procedure and surgeons are not trained to do this operation and therefore must learn it on their own. (The so-called Birmingham training program is a simple a site visit to observe). Because surgeon skill and experience are so varied reported results are extremely variable. Be sure to review outcome results in writing before choosing a surgeon. Mine are posted on my website.

I will now address the issue of women in resurfacing based on my personal results in over 3300 cases backed up by references to papers we have published. With my earliest resurfacing experience with the Corin device we published a 10-year survivorship of 93%; with men achieving 95% and women 88%. In the most recent Biomet uncemented resurfacing we have reported an overall 5-year survivorship of 98% with men achieving 99% and women 96%. Currently, 2-year survivorship for men and women are equal. This could be interpreted in 2 ways: women have more late complications, or women’s outcomes are improving. I believe we are solving the problems in women. More on this later.

Now let us review the advantages of resurfacing:

Stable Hip:

The overall instability rate for THR is 5% within the first year postop. Instability is the most common reason for revision THR accounting for 22% of revisions. My published rate of dislocation with HRA is 0.3% with a 0.06% rate of revision for instability (three series combined patient number: 2134, follow-up 2-12 years).  With THR there are lifelong positional restrictions, there are none in HRA after 6 months. The main source of instability in THA is the smaller bearing. HRA reconstructs the hip with normal bearing size and biomechanics. Some selected surgeons can achieve similar stability data with an anterior approach THR. But you need to look at individual surgeon written (best would be published) results. A recent randomized THR study  (Howie) showed 1-year dislocation rate of 5% with 28mm bearings and 1% with 36mm bearings. But some surgeons are now avoiding larger bearing THA due to recently reported trunion (where the head connects to stem) failures with larger heads. There is no trunion in an HRA.

Higher activity Level:

For older patients no difference can be detected. You cannot tell if a patient walking has a THR or HRA. The difference in functional ability only manifests itself at higher activity levels such as impact sports. 4/5 gait studies in a gait lab show superiority of muscle function in HRA vs. THR. Function is normal in HRA, but not in THR. Still a normal observer can’t see the difference when a patient walks. Two studies that compare sporting function  (Barrack; Noble) demonstrate a much higher return to sports in HRA vs. THR. Durability studies show minimal implant failure ascribed to extreme impact activities in HRA (Amstutz) but significantly higher long-term failure in THR (Argenson).

Bone Preservation:

The femoral head is largely preserved in HRA. In THR the head and half of the neck are resected, and the marrow canal is cleared for placement of the stem. On the socket side bone removal is the same for both HRA and THR. Thus the technical complexity of socket revision is the same between THR and HRA, but femoral revision is quite different. In THR the femur may need to be spit in half to remove a well-fixed stem that requires revision. Actual published results of revision HRA are mixed. Disastrous results with revision for pseudotumors in HRA have been published by some (Oxford), while others have reported better results (DeSmet, Pritchett). Our revision results are also extremely good (98% 5-year survivorship with few complications; accepted for presentation at ISTA 10/2013 and AAHKS 11/2013 and published thereafter).

No thigh pain:

The incidence of residual thigh pain in uncemented THR is 3-5%. This does not occur in HRA. It is caused by the irritation of a stiff metal stem in the femoral canal. This occurs with a well-fixed implant and cannot be corrected. It usually is not severe but may limit vigorous activity. Sometimes it resolves after several years, mostly patients just get used to it. They are usually satisfied with their THA because overall pain is less than before surgery.

Implant breakage:

Not possible with metal bearing HRA, no matter how heavily the patient uses the hip. The bone around the implant has the same risk of breakage as normal bone after the initial 6 month healing period is over.  Total hip stems are now so strong that fatigue failure is rare, but still possible with long-term heavy use. The plastic bearing is the most likely to break. Newer cross-linked polyethelene plastics are much more wear resistant, but they are more brittle and thus more prone to fatigue fracture.  Surgeons also realize that larger bearings are more stable. But larger bearings require thinner plastic liners (there is a finite space in the hip joint to accommodate the implant). Thinner cross-linked plastic does not have diminished wear properties, but a thinner more brittle plastic is more prone to fatigue fracture. Therefore impact sports are NOT recommended on plastic liners.

Cement is also a weak link. Breakage of the cement mantle leads to implant loosening. There are two methods to fix implants to bone, cementation and bone ingrowth. Cement provides immediate fixation, while bone ingrowth implants require a period of healing before the bond is secure. But cement is a brittle acrylic that suffers fatigue failure with time and repetitive impact. Once healed (6-12 months), bone ingrowth rarely loosens. We have reported a 3% rate of femoral cement failure in HRA by 10 years. We have had no failure of bone ingrowth or loosening in 2200 uncemented femoral components since 2007. Our data demonstrates at least equal short-term results with uncemented femoral fixation; we await the 10-year results to draw a final conclusion. Osteonecrosis cases are high risk for femoral HRA failure. We have had no femoral failures in uncemented HRA so far; significantly better than our results with cemented femoral components in this high-risk group.

We have used uncemented fixation for all HRA since 2007, most surgeons use hybrid fixation (cemented femur, uncemented socket). 90% of THR in the US employ uncemented fixation.

Next we will address the stated disadvantages of HRA:


Three factors have combined to lead to the impression that metal bearings are problematic. A very high failure rate of one implant brand (wear and loosening 20-50%, DePuy ASR total hip and resurfacing) due to design flaws. A very high failure rate of another brand (loosening 20%, Zimmer Durom total hip and resurfacing) An extremely high rate of “pseudotumors” reported by one group (4% @ 8 years, Oxford) using otherwise proven implants (not ASR or Durom). Furthermore this group reported that the problem was not related to implant position, but rather speculated that patients were having allergic reactions.  They were wrong on both counts. To make matters worse, this group reported disastrous results with revision of HRA for “pseudotumors”.  These problems were sensationalized in the media and by lawyers and have led most surgeons to conclude that implanting these bearings is too risky.

But it is undeniably true that large  (anatomic-sized) bearings are the single most important factor in achieving a stable hip. Large metal bearings function extremely well in most cases with extremely low wear rates. Bearing wear is so low that they should never “wear out” The same is true for ceramic and newer plastic bearings. A small amount of wear debris is tolerated by the patient. We now know that severe inflammatory soft tissue reactions can occur when the implant is wearing improperly. This occurs by a mechanism of “edge wear”. When an acetabular component design is too shallow and /or it is implanted too vertically, edge loading can sometimes occur resulting in extreme debris generation and subsequent soft tissue inflammation.  Oxford has called this problem “Pseudotumor”, and they speculate that it is caused by allergy to metal. The better term to use is Adverse Reaction to Metal Wear Debris (AWR). I have published the largest clinical series of HRA addressing this problem. My rate of AWF is1% @ 10 years in 2600 cases. All cases of AWF were in smaller bearing sizes (less than 50mm) positioned with an acetabular inclination angle (AIA) above 50 degrees. Of all cases with AIA> 500, 5% had AWF. No cases with bearing greater than 50mm had AWF. The reason AWF is more common in smaller bearings is that they are shallower by design than the larger implants (for every brand).

In another study we demonstrated a safe zone for placing acetabular components based on bearing size. This safe zone is called RAIL (relative acetabular inclination limit). If an implant is placed by these criteria, no AWF or dislocations occur and the rate of ion levels elevated above 10ug/L is less than 1%. Over 2 years ago we developed intraoperative x-ray techniques to be within RAIL 100% of the time.

Therefore it is possible to place HRA with any bearing size in a position that avoids the complication of AWF. Therefore women (usually have smaller implant sizes) are no longer at higher risk for this complication.

Cancer risk:

There has long been a “concern” raised by proponents of plastic bearings about a higher risk of cancer in metal bearings. Metal bearings have been around for 60 years and still no higher risk of cancer has been demonstrated. Two large studies of various bearing type, one out to 30 years, have demonstrated no higher risk of cancer than in the general population (Visuri, Oswestry). One recent study based on the English Joint registry has even shown that patients with HRA live longer than patients with THR, even when the study groups have been adjusted for age and preoperative medical condition (HRA patients are usually younger and more interested in sports) (McMinn).

It is true that the wear debris of metal bearings can be measured in the blood, and the wear debris of plastic bearings cannot. Plastic debris remains in the body forever, metal debris passes out through the urine.

Early femoral failures (femoral neck fracture and head collapse [AVN]):

Femoral neck fractures and femoral head collapse (some call it avascular necrosis) within the first 2 years after surgery are early failure modes that do not occur with THR. The rate of early femoral failure was 2% in my first series of HRA. In my most recent cases the rate is down to 0.1%.  I have never seen a higher rate of this problem in women, although others have reported this. Many experts mention that low bone density is a contributing factor. We published the only papers that provide scientific data on this topic. We have shown that low bone density and high BMI (body mass index) are the only two factors that influence the rate of early femoral failure. Furthermore, we have published data demonstrating methods to prevent early femoral failure in high-risk groups. For many years we have routinely measured bone density using a DEXA scan preoperatively and calculated BMI on every patient. We do not exclude patients with high risk from HRA; rather we manage them differently to prevent these complications. This has led to our current very low rate of early femoral failure of 0.1%.

Dorr has reported trochanteric and femoral shaft fractures in THR at a rate of 0.7% in a large study. My personal experience with these complications in THR is 1.5% (all in women). We are now applying the same lessons we learned in HRA to THR. Hopefully a reduction in femoral fractures will also occur in THR.  It is interesting to note that the rate of femoral fractures in HRA is now much lower than in THR.

Inconsistent results

The results in hip replacement surgery vary tremendously depending on a variety of factors such as surgeon skill and experience and the prosthesis chosen. The surgeon has the responsibility for choosing the prosthesis. Unfortunately some have made poor choices such as the Durom and the ASR. An HRA is a significantly different and more difficult operation than a THR. There is a long learning curve. I have published a paper showing that my results improved significantly even after 200 cases. We have studied the problems of HRA carefully and have found solutions to lower the incidence of several complications. I previously described how the problems of AWR and early femoral failure have been solved. Experience and critical scientific analysis allow us to solve the problems of HRA and get better.

The Lancet journal published a highly flawed article by Smith earlier this year indicating that THR had better 5-year results than HRA. Even though the Lancet is theoretically a peer-reviewed journal, the peers were either completely biased or just asleep. They were derelict in their responsibility and did not require the author to discuss the weaknesses in their study. The study was based on the English and Wales Joint Registry. The major flaws were as follows:

  • No mention was made of the lack experience of the surgeons performing HRA. Every surgeon is taught how to perform a THR in residency and has performed many before graduating. Few learn HRA. Using the raw numbers provided in the article, I calculated that the average surgeon in this study performed less than 6 HRA per year. It has already been demonstrated in numerous other studies that there is a long learning curve with HRA. The only conclusion that you could possibly draw from this study is “ do not allow an inexperienced surgeon to perform your HRA”.
  • The lost to follow-up rate was 18%. This is inexcusable in a large registry study. The primary strength of registry studies is supposed to be near 100% inclusion. My studies typically have 92% or more follow-up maintained, which is considered high for large clinical series. Scandinavian and Australian registry studies are typically close to 100%. The high loss of follow-up of this Lancet article makes all conclusions very difficult to trust.
  • The primary weakness of all registry studies is the lack of detailed data. Only cases that are revised are considered failures. If a patient has had three dislocations but has not been revised they are still considered “successful”. Patients who have loose implants but have not been revised are also considered “successful”. Complication rates are not reported.

It is the duty of peer reviewers to require the author to address these flaws in the discussion section of their article.

There are many surgeons who have dedicated their career to HRA who have published excellent results in large clinical series of HRA including McMinn, Treacy, Amstutz, DeSmet, Holland and myself. There is another generation of dedicated surgeons who have made HRA their career and have done large numbers of cases. We look forward to the publication of their clinical results.

My most recent clinical series published in the Journal of Arthroplasty reports a 97.4 % survivorship at 5 years with only 0.3% AWF in the first 1000 uncemented Biomet HRA devices that I implanted beginning in 2007.

The Dysplasia Effect

One additional fact o be considered in this discussion is how the diagnosis of dysplasia affects the results. Dysplasia is used to describe a spectrum of deformities of the hip that a child is born with. Occasionally children are born with a dislocated hip, the most extreme example of dysplasia. They are diagnosed as infants and treated. Most children do very well with modern treatment, but the hip is rarely made completely normal. The vast majority of dysplasia cases are never diagnosed in childhood because they are mild and do not cause any symptoms. 72% of these are in women who are typically hyper-flexible and gravitate towards activities such as ballet and gymnastics where they have an advantage due to their flexibility. Between ages 40-50 their biomechanically abnormal hip breaks down and becomes arthritic. Many of these women seek out resurfacing in order to continue their activities. The primary deformities of dysplasia are a shallow oval socket that is positioned too vertical as well as an oval femoral head with a high neck angle and narrow neck.

Several studies have shown that the risk of HRA is higher in dysplasia. Numerous other studies have shown the same for THA. The primary problems in THA are acetabular component fixation and a higher than normal dislocation rate. The problems in HRA are acetabular component fixation and AWF. No direct comparison study between THR and HRA in dysplasia exists. In my HRA cases 28% of women were in the high-risk dysplasia group while only 4% of men were in this group. We have shown that this is the primary reason that women have higher failure rates. Most clinical series of THR are in older patients where the incidence of dysplasia is much lower. I know of no clinical series of THR that has a 28% rate of dysplasia in the women. Therefore it is impossible to compare THR and HRA in women. There are separate series of THR in dysplasia typically with a higher failure rate than for other diagnoses.

Failure of acetabular ingrowth is the most common problem of HRA in dysplasia patients. This problem has been solved in 2007 with the introduction of the Biomet Trispike Magnum acetabular component. We have not had a single failure of socket fixation in dysplasia in 6 years. The Trispike socket has additional spikes for fixation. We use it selectively in the most deformed sockets that are at highest risk for fixation failure. The milder dysplasia cases do not require it.

AWF is the other problem seen more frequently in dysplasia. These patients typically have smaller bearing sizes, which are higher risk for AWF, if the acetabular component is placed too steeply. Most expert recommendations on component positions are too steep for the smaller bearings. We have published the first study that provides scientific evidence for a robust safe zone based on bearing size (see RAIL above). Because dysplasia sockets are shallow, oval and too steep to start with, the surgeon is also fooled into placing the socket too vertical. This leads to edge loading and subsequent AWF. If the surgeon takes poor quality x-rays and still uses older positional guidelines he may even believe the component is correctly positioned and he may be tempted to speculate that “allergy” is the problem. AWF (or Pseudotumors) are caused by shallow sockets that are placed too steeply; unfortunately many surgeons do not yet understand this. It is a problem that is now completely preventable by good surgical technique.

In conclusion, hip resurfacing is now a very reliable operation in patients of all ages and diagnoses if performed with a well-designed prosthesis by an experienced and skilled surgeon.  Women have traditionally had higher failure rates than men, but the gap is nearly closed. Hip resurfacing is unique because it reconstructs the hip with near normal biomechanics and therefore provides a stable hip joint. It is durable enough to be used without limitations after the initial 6-month healing period. Equal results can be achieved with low bone density, provided that early postoperative management is modified. Our data indicates that the problems in women have now largely been solved. Longer-term follow-up will prove this fact.

Thomas P. Gross, MD