Hip fusion is an uncommon procedure. Hip fusion takedown, therefore, is equally an uncommon procedure. What is of considerable interest is that the results, which I achieved in 20 cases in a paper published in 1987 are considerably superior to the results, which I am achieving today. This suggests that no simple case is now fused. It also equally suggests that there is little sense in looking at literature more than 10 or 15 years old on fusion takedowns as the two conditions are likely completely different.

Most patients do not like a hip fusion. There are long-term problems with low back pain, ipsilateral global instability and contralateral patellofemoral osteoarthritis. A stiff hip produces a poor quality of life, especially in a tall person. The main problem in doing a hip fusion takedown is the condition of the abductors muscles. If fused before growth was complete, there may be pelvic hypoplasia. If the pelvis is small, the glutei will also be small. Sometimes, the glutei may have undergone fatty degeneration. This can be assessed by means of an MRI. If the abductors were damaged during fusion, a limp may persist. Other problems are that leg lengthening is difficult to achieve any longstanding hip fusion. Lengthening of 1–2 cm is usually about all that can safely be achieved. If the hip was fused in childhood, there is likely to be femoral hypoplasia. There is also likely absence of proximal cancellous bone and the proximal femur is a thin brittle cortical tube. The greater trochanter should not be detached as it is difficult to obtain union under such circumstances. The approach, which I prefer for a fusion takedown is an anterior Smith Peterson. The glutei are slid off the pelvis sidewall and then the upper part of the fusion can be exposed, blunt Hohmans can then be passed around the femoral neck prior to transection. Obviously, if any AO cobra plate has been used for a fusion, a trochanteric osteotomy may be required to preserve any glutei left. Old hardware can be removed either concurrently or as an interval procedure. In 1986, I published the results of 20 cases with a five to 40-year fusion time (mean 19). I used a variety of implants. Flexion was achieved to 90 degrees at 12 months in about 88% of people. Seventy-five percent ceased to limp by year one, although the elderly limp when tired. One patient was dissatisfied with the procedure. One was revised for pain.

I have reviewed the cases done in the last 20 years. These were 28 cases, two bilateral. Seven were spontaneous fusions. Twenty-one were formal hip fusions. One was an AO fusion with a cobra plate. There were various intra-operative complications including two calcar cracks, which were wired, three femoral shaft fractures, which necessitated the use of long stems. There was one drop foot, which recovered. At review, a limp was absent in 20%, mild in 12% and severe, i.e. Trendelenburg positive in 68%. Harris hip scores were excellent in 28%, good in 32%, fair in 16% and poor in 24%. Four patients only, however, continued to use canes. The eventual range of movement was good. In 80%, more than 90 degrees of flexion was obtained, but it took up to two years to obtain maximum flexion. In 12%, the range of motion was poor at being 50 degrees to 85 degrees. The range of motion was poor, i.e. less than 45 degrees in one bilateral case of athrogryposis. This was a stiff arthrogrypotic. Further surgery is required in several cases. An ipsilateral total knee replacement and one a supracondylar femoral osteotomy. One cup loosened and was revised at seven years and one liner was exchanged at ten years.

Persistent post-surgical pain remains a problem after knee replacement with some studies reporting up to 20% incidence. Pain is usually felt by those who do not operate to be a monolithic entity. All orthopaedic surgeons know that this is not the case. At its most basic level, pain can be divided into two categories, mechanical and non-mechanical.

Mechanical pain is like the pain of a fresh fracture. If the patient does not move, the pain is less. This type of pain is relieved by opiates. Mechanical pain is seen following knee replacement, but is fortunately becoming less frequent. It is caused by a combination of malrotations and maltranslations, often minor, which on their own would not produce problems. The combination of them, however, may produce a knee in which there is overload of the extensor mechanism or of the medial stabilizing structures. If these minor mechanical problems can be identified, then corrective surgery will help.

Non-mechanical pain is present on a constant basis. It is not significantly worsened by activities. Opiates may make the patient feel better, but they do not change the essential nature of the pain. Non-mechanical pain falls into three broad groups, infection, neuropathic and perceived pain. Infection pain is usually relieved by opiates. Since some of this pain is probably due to pressure, its inclusion in the non-mechanical pain group is questionable, but it is better left there so that the surgeon always considers it. Low grade chronic infection can be extremely difficult to diagnose. Loosening of noncemented knee components is so rare that when it is noted radiologically, infection should be very high on the list of suspicions. The name neuropathic pain suggests that we know much more about it than we do in reality. Causalgia or CRPS-type two is rare following knee replacement. CRPS type one or reflex sympathetic dystrophy probably does exist, but it is probably over-diagnosed especially by the author of this abstract. The optimum treatment I have found is lumbar sympathetic blocks. Perceived pain is the largest group. It does not matter what you tell the patient, some believe a new knee should be like a new car, i.e. you step into it and drive away. The fact that they have to work to make it work is horrifying. Some of this pain is actually mechanical, especially in those with no benefits such as hairstylists. Perceived pain is widespread. The classic treatise on this is Dr. Ian McNabb's book “Backache”. It should be studied by all orthopaedic surgeons, who wish to understand pain complaints.

Any experienced knee surgeon will have his list of red flags or caveats. These are often politically incorrect and this information is transferred to young surgeons, usually in dim bars late at night. I will list only a few. If the patient comes in with a form asking for a disability pension on the first visit. If the patient's mother answers the questions. If the patient comes in taking massive doses of opiates. If the patient is referred to you by a surgeon, who does more knee replacements than you do. There is also the recently described Fern Silverman's syndrome.

There is no mathematical relationship between the internal diameter of the femoral metaphysis and diaphysis. Unless an infinite number of monolithic stems are available with variable metaphyseal and diaphyseal diameters, which is not economically possible, even in virgin cases, the surgeon has to decide if the stem is going to fit in the metaphysis or the diaphysis. It is not possible to match both.

In revision cases with a hollowed out metaphysis, the situation is much worse. As it is obviously easier to fit the diaphysis, this is what stems such as the AML and Wagner stem have done. They completely ignore the metaphysis and obtain fixation in the diaphysis. This is all well and good, but it means that the proximal femur is unloaded, like an astronaut in space. While, there will be some recovery due to removal of the toxins and local muscle pull, it will be incomplete. Furthermore, should sepsis occur, one is faced with the horror of removing a distally fixed implant.

Clearly, if proximal fixation, i.e. above the level of lesser trochanter could reliably be achieved, this would be preferable in terms of proximal loading leading to bone recovery and ease of removal should it be required. The only way that proximal loading can be achieved is if the metaphyseal and diaphyseal parts of the component can be varied infinitely. This clearly can only be achieved by using a modular stem.

The concern with modularity always has been fretting at the sleeve-stem locking mechanism with release of metal ions. The stem, which I have been using for the last 25 years, is the SROM stem. Fretting and ion release had never been an issue. As the components are made of a relatively soft titanium alloy, it is likely that the sleeve and the stem cold weld, thus, eliminating any movement and eliminating friction.

I have a follow-up of roughly 120 revision cases with a minimum follow-up of 5 years and a maximum follow-up of 22 years. I have no loosening in easy revision cases where a primary stem was used. I have had some loosenings in extremely difficult revision situations where a long bowed stem was required, but even then, the loosening rate is less than 3%.

I use this stem in primary situations, i.e. in about 80% of all the primaries I have done. This means I have done roughly 1500 cases or more. Other than some late infections, I have never, ever had any stem loosening in a simple case. Obviously, I have had loosenings in some cases, where we have been doing fancy shortening or de-rotation osteotomies, but none in simple primary cases. I would, therefore, suggest that the surgeon, if he wishes to use this stem, please try it out on some simple primary cases.

The ability to vary distal and proximal internal diameters and proximal geometry makes for easy surgery. I have been using this stem for 25 years and continue to use it in all my primary noncemented cases. I believe in the adage of “train hard and fight easy.” I think that surgeons should not get themselves into a situation where they are forced in a difficult case to use something they have never seen before.

Persistent post-surgical pain (PPSP) remains a problem after knee replacement with some studies reporting up to 20% incidence. At its most basic level, pain can be divided into two categories, mechanical and non-mechanical.

Mechanical pain is like the pain of a fresh fracture. If the patient does not move, the pain is less. This type of pain is relieved by opiates. Mechanical pain is seen following knee replacement, but is fortunately becoming less frequent. It is caused by a combination of malrotations and maltranslations, often minor, which on their own would not produce problems. The combination of them, however, may produce a knee in which there is overload of the extensor mechanism or of the medial stabilizing structures. If these minor mechanical problems can be identified, then corrective surgery will help.

Non-mechanical pain is present on a constant basis. It is not significantly worsened by activities. Opiates may make the patient feel better, but they do not change the essential nature of the pain. Non-mechanical pain falls into three broad groups, infection, neuropathic and perceived pain.

Infection pain is usually relieved by opiates. Since some of this pain is probably due to pressure, its inclusion in the non-mechanical pain group is questionable, but it is better left there so that the surgeon always considers it. Low grade chronic infection can be extremely difficult to diagnose. Loosening of noncemented knee components is so rare that when it is noted radiologically, infection should be very high on the list of suspicions.

The name neurogenic pain suggests that we know much more about it than we do in reality. Causalgia or CRPS-type two is rare following knee replacement. CRPS type one or reflex sympathetic dystrophy probably does exist, but it is probably over-diagnosed especially by the author of this abstract. The optimum treatment I have found is lumbar sympathetic blocks. Lyrica, Gabapentin and Cymbalta may also help.

Perceived pain is the largest group. It does not matter what you tell patient, some believe a new knee should be like a new car, i.e. you step into it and drive away. The fact that they have to work to make it work is horrifying. Some of this pain is actually mechanical, especially in those with no benefits such as hairstylists. Perceived pain is widespread. The classic treatment on this is Dr. Ian McNabb's book “Backache”. It should be studied by all orthopaedic surgeons, who wish to understand pain complaints.

There are other issues such as good old fibromyalgia, which appears to have gone the way of the dodo. It has been replaced by something equally silly called central sensitization. The theory of central sensitization is that if one has pain somewhere or other for three months or six months or whatever, there are going to be changes in the brain and spinal cord. It then does not matter what happens to the original pain, i.e. whether or not it goes away, the pain will persist because of the changes in the brain, hence, the title of the pain in the brain syndrome.

If this theory was correct, we might as well all go home because we have all been wasting our time for the last 30 years because none of our patients would get any better. After all, all of our patients have had pain for a lot longer than three months, many of them have been involved in trauma and sometimes, compensation is at issue. The pain in the brain theory, therefore, sounds about as realistic as the flat earth society or the treatment of Galileo.

In revision surgery of a neglected loose cemented stem, especially if long-standing infection is present, the bone may become extremely thin and fragile. A fracture during implant removal, therefore, occurs readily and the bone may simply disintegrate so that it cannot be reconstructed to produce intrinsic stability. In such cases, rather than try to put together a house of cards, the bail and nail technique may be used.

I first described this in 1992. A large bore intramedullary nail is inserted into what is left of the femur and the bone fragments proximally or cerclage wires are around the nail. The patient is left with a Girdlestone. Most patients cannot weight-bear on a Girdlestone for about two months. After that time, they are encouraged to do so as load bearing speeds healing and bone regeneration.

Generally, within six to nine months, the femur is reconstituted and a revision can be carried out. This is a simple operation as the medullary canal is already open and the regenerated bone is often of surprising good quality. It may even be possible to use a primary stem as opposed to a long revision stem, but this probably should not be done under nine months.

The presence of an IM nail and cerclage wires does not prevent the elimination of infection. As our brothers, the trauma surgeons, will attest the treatment of an infected non-union of the femur is an intramedullary nail.

This technique is seldom required. The author has used it in eight cases in the last 30 years. It is, however, extremely valuable in bailing out of an impossible revision, especially if strut allografts or modular long stems are not immediately available.

Small canals are usually in small people, but occasionally some normal-sized people have huge cortical thickness with a corresponding small canal.

To give adequate strength to the cement mantle, either pure cement or cement/cancellous bone, it must be at least 2mm thick. If the medullary canal is 9mm or less, then the thickest stem, which can be used with cement will be 5mm. This stem is so small that under load, it may deform repetitively, i.e. cycle. If it does cycle, it will break up the cement mantle. In order to get in a stem large enough to prevent cycling, hard reaming will be required, thus, removing most of the cancellous bone so the cement interlock is poor.

Small stems are also usually fairly short stems. With a follow-up of more than 15 years, inevitably, some lucency between the cement and bone occurs in zones one and seven. If the stem is long enough, that is of no significance. If the stem is short, i.e. 120mm or less, then the area of distal fixation becomes precariously small.

For these reasons, if the canal is small, it is preferable to use a non-cemented stem. The reaming technique for a non-cemented stem is to reach endosteal cortical contact either circumferentially with a canal-filling stem or at the point of wedge for a wedge-shape stem. The metaphyseal bone may be poor quality in the elderly, but it is going to be removed anyway to load the endosteal cortex.

This means that a large stem can be used, which is, therefore, stronger and less likely to undergo mechanical failure and fixation failure.

There are two types of pain, mechanical and non-mechanical. Mechanical pain hurts with movement/use, is not constant and is helped by morphine-type products. Non-mechanical pain is different. It is present 24 hours a day, often worse at night, and except for the pain of infection, is not relieved by morphine-type products.

If the cause of mechanical pain can be determined, it can be corrected by an operation. The usual cause of postoperative mechanical knee pain nowadays is multifactorial, i.e. a combination of minor errors, none of which on their own would require revision.

Non-mechanical pain, other than infection, is much more difficult to handle. The commonest cause is not really a pain complaint, it is disappointment due to a failure of expectation. It does not matter how often you tell patients, some patients still think they should step in a drive away. A lot of these failures of expectations become much more realistic by the end of year one.

There are several other categories. Incipient osteoarthritis or sensitive people (The Princess and the Pea). If the pain complaints were severe with minimal arthritis, an operation is not likely to help.

The patient on disability for no clear reason is unlikely to get a good result and Workmen's Compensation Board and motor vehicle accident patients are also a very bad prognostic sign and will often produce the postoperative painful knee. Preoperative use of large doses of morphine is also a very bad sign. It is not clear if it is the morphine, which influences the patient or the patient, who influences the morphine.

There are several pain syndromes, some of which are purely psychiatric such as Conversion Disorders and Somatoform Pain Disorders. Treatment of purely psychiatric conditions should be a referral to a psychiatrist is in order.

Complex regional pain syndrome is an organic pain disorder. Type 2 is causalgia or an actual nerve injury. This is unusual following knee replacement other than the odd drop foot, which even after recovery, leaves an area of dysaethesia on the dorsum of the foot. Type 1 used to be called reflex sympathetic dystrophy. This is not uncommon after total knee replacement. I managed to collect more than 40 cases. One problem is that the diagnosis to some extent is a diagnosis of exclusion. If the diagnosis can be made, then treatment is available including Cymbalta, Lyrica or Gabapentin. I have found most success with lumbar sympathetic blocks, but it is difficult to find someone, who can do these. Some patients have been treated with implantable electrical spinal stimulators with variable results.

The current flavour of the month pain syndrome is called central sensitization. The theory is that if someone has pain for more than six months, then there will be changes in the brain, which will remain after the original pain goes away, hence, the title the pain in the brain syndrome. If this theory were correct, then we as arthroplasty surgeons have been collectively wasting our time for the last 40 years as no patient would have recovered. The likelihood, therefore, of this theory having any basis in reality is pretty remote.

Fortunately, by the end of year one, the vast majority of our knee replacement patients are reasonably content with the procedure.

Modular necks arrived in North America in the late 1970s. The purpose was to allow ceramic balls to be attached to metal stems. The advantages of modularity were so obvious that it was universally adopted with almost no untoward consequences.

A double-taper neck was developed in Italy by the Cremascoli Company and was used extensively with few reports of problems.

Recently, problems have been reported not only with double-taper necks, but also with head-neck junction tapers. Something would appear to have changed recently.

Some of these changes were shortening of the taper, lengthening of the neck, version angles were increased, and head sizes bigger than 32mm were introduced. Surface finish on some of the tapers was changed and they were ridged. This produces a better fit for a ceramic ball, but facilitates crevice corrosion with a metal ball.

The author used the original OTI cemented stem with a double-taper neck between 2002 and 2005. It was a cobalt chrome construct. The Cremascoli was a titanium hip. The OTI used the classic Morse taper with cogs for increased rotational resistance.

One hundred forty five stems were inserted. The neck stem taper broke in two cases and dislocated in one. All of these cases had a long neck and long heads and thus produced maximum moment arm. The stem was withdrawn from the market, the taper lengthened and the strength doubled. It was reintroduced in 2007. From 2007 to 2011, the author has done 188 cases, all cemented stems. No taper problems have occurred.

The conclusion is that a properly designed double taper neck does not appear to have a downside. The upside is the ability to change neck length and especially version after head insertion. Leg length and hip centering can, therefore, be fine-tuned as never before. The author continues to use this stem enthusiastically for all cemented cases.

Hip fusion used to be a common procedure in children and young adults, but it is now exceedingly rare.

My results of hip fusion takedown more than 20 years ago were quite acceptable. Of 20 cases, 88% achieved more than 90 degrees of flexion and 75% stopped limping by the end of one year. The elderly would revert to limping when tired.

As no simple hips are currently fused, the results of hip fusion takedown in the last 20 years are very much inferior. Of 28 cases, limp is absent in 20%, mild in 12% and severe in 68%. Range of motion is acceptable with 80% eventually achieving more than 90 degrees of flexion.

There are complications, but these are quite manageable. The aseptic loosening rate is small and the longevity is high. Current implants, therefore, can easily handle the hip fusion takedown.

As the incidence of limp is prohibitively high, additional techniques to reinforce the hip abductors either concurrently or more likely as a secondary procedure as suggested by Whiteside should be learned by all those proposing to carry out hip fusion takedown.

The S-ROM stem is distally circular canal filling with thin sharp flutes which engage the endosteal cortex. The rotational stability produced by this is 37 Nm, which exceeds the service loads on the hip of 22 Nm. The distal canal fill prevents varus and valgus displacement. The porous-coated proximal sleeve provides resistance to vertical sink and also excludes the distal stem from the effective joint space.

The primary stem is straight and the long stem is bowed with a 15 degrees anteversion twist proximally. The neck comes in lengths from 30 to 46mm with varying offset. The sleeves come in variable size and geometry.

The stem choice in revision surgery is based on the Scoot Diamond Classification. Type 1 (this is going to be easy) is a primary stem. Type 2 (this is going to be difficult) implies diaphyseal bone loss and will require a long stem. Type 3 (Oh My God), implies more than 70mm of completely missing proximal femur and will require a structural allograft cemented to the sleeve.

Hinges were used early in total knee replacement history. The stems were too short, the trochlear groove was absent or inadequate and the mechanism lacked strength. Hinges, therefore, acquired a very bad reputation.

As the only stems in early knees were on hinges, I used them when necessary, i.e. a completely missing medial collateral ligament, a flexion gap of more than three centimeters with a normal extension gap and missing bone. I used a Guepar II and my results were good up to 20 years when the plastic wore out and the spindle was damaged. By that time, new spindles were not available and the cases had to be revised. Revision of a cemented bowed stem is a nightmare. One would wish, therefore, to have a design, where the bearings could be detached from the stems and a new bearing inserted.

The main problem with hinges nowadays is that they all rotate, or at any rate, all the hinges which I can get, rotate.

The commonest indication for a hinge is the multiply revised knee due to missed tibial torsion. A rotating hinge used in a situation like this results in the patient externally rotating the tibia and subluxating or dislocating the patella. They, therefore, cannot use this knee as the leading leg on any activity and they may have instability or falls as a result of the patellar dislocation.

It is not difficult to block rotation but, so far as I know, no one has done it.

Is there a role for hinge? Very seldom and not with the existing designs.

Femoral components in total hip replacements fail in well-known ways. There is vertical sink, posterior rotation and pivot, either distal or mid-stem. In order to sink, the stem moves into valgus and then slides down the inside of the calcar. It does not cut through the calcar.

To prevent sink and pivot, a canal filling stem is required. Canal fill prevents the stem from moving into valgus and, therefore, it will not sink. Two centimeters with complete canal fill is adequate in a primary stem. A long stem will give longer canal fill in a revision. Sharp distal flutes will prevent rotation. The distal end of the stem should be polished. One is looking for a distal stability, not distal fixation.

If the isthmus is intact, a primary stem can be used. If the isthmus is damaged, a long stem is necessary. If the calcar is intact, a primary neck is adequate. If the calcar is missing down to the level of the lesser trochanter, a calcar replacement neck is required. If there is more than 70 millimeters of completely missing proximal femur, a structural allograft is required.

If the proximal femur is damaged, the ability to place a sleeve or collar to seek the best bone available independently of the stem version is very helpful.

No matter how poor the proximal bone quality is, it can be supplemented by cerclage wires. The implant will sink only if the cerclage wires break.

The advantage of proximal fixation is that loading the proximal femur speeds recovery. The huge disadvantage of distal fixation is removal of the implant should it become necessary.

My long term results for the S-ROM stem used in revision are now out over 20 years. There were 119 primary stems with a minimum follow up of 5 years with no revisions for aseptic loosening.

There were 262 long stems used. Nine (3.7%) underwent aseptic loosening. Most of these were due to technical errors due to my inexperience in the learning process of revision surgery. Four were dependent on strut-grafts and should have been treated with structural allografts.

There were seven cases with structural allografts. Three were revised. Again, these were largely from problems arising from inexperience.

I believe proximal modularity with distal stability allows the vast majority of revision cases to be treated with proximal fixation.