7.13 Management of the infected total hip replacement

Infection is a major complication of arthroplasty surgery with a reported incidence after primary operation of 0.2–1%. Its successful management depends on clear understanding of certain principles summarized in this chapter.

All surgical wounds are contaminated by bacteria and it is surprising that relatively so few become clinically infected. Over 95% of prosthetic infections that occur during the first year after implantation are due to perioperative contamination. Infection can also occur through the blood stream; such haematogenous infections tend to manifest later. Irrespective of the route, infection becomes established when the dose and virulence of bacteria overcome the host defences. The degree of virulence determines the speed of infection. In addition to surgical and environmen-tal factors, patient factors play an important role, e.g. old age, chronic disease, immunosuppression, and a history of previous joint infection.

After implantation, body fluids immediately coat all surfaces of the prosthesis with a layer of serum proteins and platelets. Bacteria approach the surface of the prosthesis through an interaction of physical (van der Waals forces and hydrophobic interactions) and chemical forces (covalent and hydrogen ion bonds). There is then a race for colonization of the implant surface between the host’s cells and the bacteria. If this race is won by bacteria, adhesion progresses to aggregation on the surface of the prosthesis. In doing so, bacteria change from being planktonic to sessile: they become encased in a matrix of polysaccharides and proteins to form a slimy layer known as the biofilm. This has the following properties:

Gram-positive cocci dominate the list of pathogens identified accounting for over 50% of infections within several units. The majority are coagulase-negative Staphylococcus and Staph. aureus. Aerobic Gram-negative bacteria cause 10–20% of all deep infections and anaerobic bacteria are responsible for another 10–15%.

Infection can be superficial or deep.

As far as is known, haematogenous infection is relatively rare, many late infections being acquired from perioperative contamination.

Sometimes diagnosis is obvious, but not always: there is no single investigation which can conclusively confirm or refute its presence; no single test is 100% sensitive or specific and multiple tests are required.

Most patients following hip arthroplasty have a very good clinical result. Any deviation from this, particularly in the first 12–18 months should raise suspicion of infection. The nature of the pain should be ascertained; is it mechanical in nature or present at rest, is it associated with fever or night sweats, in fact, has the hip ever felt right? Did the wound heal quickly or did it require antibiotics and district nurse input?

Pain and reduced range of movement may be present in acute infection otherwise the joint may be only mildly irritable. Frank pus, abscess, or sinuses may present or sometimes more subtle changes like skin discolouration and induration. In the majority of cases there will be nothing specific to find on examination.

It is normal practice to measure the ‘inflammatory markers’ in the blood: the white cell count (WCC), the erythrocyte sedimentation rate (ESR), and the C-reactive protein (CRP). Full blood count often shows a normal WCC but patients with long-standing infection can develop anaemia of chronic disease and a raised platelet count.

The ESR takes up to 1 year from the index operation to return to normal levels but a raised level of 30mm/h or greater raises the clinical suspicion of infection. Other causes for a raised ESR must be taken into consideration.

The CRP is acutely raised after the index operation for 48–72h then falls rapidly over the next 2–3 weeks in the absence of infection. However, it may remain elevated in patients with rheumatoid arthritis for up to 8 weeks. After this time a level of 10mg/L or greater raises the clinical suspicion of infection.

Gaston has shown that serial measurements of CRP and ESR are more helpful than isolated results. A single elevated CRP and ESR was in his series, associated with a 50% chance of infection. This rose to 80% for persistently elevated test results. Normal levels—both isolated and in series—had a stronger predictive value in ruling out infection (90% and 94% respectively).

Plain radiographs in isolation may show radiolucent lines, osteolysis, scalloping, and periosteal reaction but these can also be present in aseptic loosening. Serial radiographs are more useful for assessing changes over time as infective loosening tends to be more rapid in its progression.

With increasing use of uncemented prostheses, radiographic changes may be more difficult to interpret. Provided a prosthesis is not becoming loose overall, relatively marked signs somewhere at the bone–cement interface may be pain free.

Radioisotope scans although expensive and time consuming, may be useful. Indium-111 (¹¹¹In)-labelled white cell scans combined with technetium-99m (⁹⁹TC)-sulphur colloid marrow imaging have been found to be sensitive for loosening but not specific for infection. There is now interest in fluorodeoxyglucose positron emission tomography (PET scan) which may be more specific.

Hip aspiration is regarded by many as the gold standard. In addition to diagnosing the bacteria responsible it also offers an antibiotic sensitivity profile. Bacteria spend most of their time in the sessile phase and so to optimize accuracy of the aspiration the following steps should be taken:

Numerous authors have found hip aspiration to be highly specific (see Table 7.13.1) and therefore making more invasive tissue biopsy unnecessary.

A simple WCC from the aspirate can be a useful adjunct. There is a reported sensitivity of 97% and specificity of 98% when there are 1.7 × 10³ per microlitre white cells with a differential of 65% neutrophils.

Polymerase chain reaction (PCR) allows detection of the bacteria in small numbers by amplification of microbial DNA but its usefulness is limited by a high rate of false positives.

Although this investigation has been shown to be highly specific for the diagnosis of infection, it does not help with the preoperative diagnosis or yield antibiotic sensitivity profiles. It does, however, require an experienced pathologist. Whilst five polymorphs per high-power field is regarded as being consistent with a diagnosis of infection, the presence of ten polymorphs per high-power field gives an improved specificity of 99% without reducing sensitivity.

This simple test lacks any acceptable level of sensitivity and is not recommended for this purpose.

Multiple intraoperative samples should be taken and sent for microbiological analysis. A fresh scalpel and forceps are used for each specimen and antibiotics withheld until the samples have been taken. It is imperative that these samples are sent directly to the laboratory. Three or more cultures growing a consistent micro-organism are thought to be diagnostic of infection.

The probability of diagnosing infection is increased with increasing number of positive results from all the investigations described here.

Eradication of the infection is the fundamental necessity taking precedence in the sequence of other curative measures, usually, but not always reimplantation. A stable implant will most usually be pain free.

The options are:

Surgery is not always the appropriate choice. In cases where the patient may not be fit for, or does not want surgery then long-term suppression with antibiotics can be an acceptable form of treatment. The prosthesis, however, must be soundly fixed and the infective organism sensitive to the antibiotic used. Nevertheless, it must be recognized that antibiotics alone will not eradicate the infection—only slow the infective process to some degree.

Success rates of up to 84% have been reported with debridement and lavage with retention of the original prosthesis in cases of acute postoperative infection if performed within 2 weeks of the index procedure. After this time the chance of successful eradication of infection falls dramatically. A patient rarely presents during this time frame so making this treatment of later haematogenous infection much less successful.

Exchange arthroplasty surgery offers the best way of eradicating infection along with restoration of function. This can be undertaken in a direct or staged manner. Results in the literature of each pathway are difficult to compare, but the key to success of both treatment modalities depends upon:

The major advantage of direct exchange surgery is that it is one operative procedure with benefits to both the patient and the healthcare system. It removes the morbidity of a temporary pseudarthrosis and lessens the cost of treatment.

The generally accepted criteria for performing direct exchange are:

Direct exchange follows the principle that debridement removes all infected tissue and material, the dead space thus created is filled with the new implant which is fixed to the bone with cement acting as a depot for the release of local antibiotic. Recent published work has demonstrated the successful use of allograft bone as a delivery vehicle for antibiotics in single-stage surgery for infection using uncemented components. The combination of uncemented implants and an antibiotic bone compound used in a single-stage procedure has advantages and may improve long-term result.

It is generally accepted that direct exchange is less successful in eradicating infection than staged reconstruction. However, the difference is relatively small. The success of a one-stage procedure is highly dependent on the aggressiveness of the debridement, the experience of the surgeon, and close interaction with a microbiologist. Several European centres have reported excellent results treating all infections in this manner.

Although single-stage exchange has an obvious role we believe this should probably be restricted to the case of a single organism with known antibiotic sensitivities and reasonable bone stock.

The only difference between direct and staged exchange is the interval phase. Radical debridement is essential to both. Antibiotics need to be administered either intravenously or to elute from bone cement (usually pellets or a spacer inserted during the interval phase as described later in this section) and after a period of weeks or even months the hip is reconstructed.

Following debridement, the extent of bone loss will be evident thereby allowing the surgeon to plan subsequent reconstruction. If bone graft is required, it can be safely used at the time of second stage without there being an increased risk of infection.

The interval phase allows for the administration of antibiotic. This can be given systemically or by elution from bone cement in the form of beads or a spacer. The advantage of cement elution is that it produces a much higher local concentration of antibiotic than could be safely given systemically. Local levels of gentamycin can be 10–100 times higher than the minimal inhibitory concentration of the causative bacteria whereas only 10% of the antibiotic available from the serum is found within bone when given systemically.

Any antibiotic can be added to the bone cement prior to polymerization as long as it is a sterile, dry powder which is heat stable and water soluble. It is usually mixed by hand in theatre at the time of surgery. The elution properties relate directly to the water uptake by the bone cement, antibiotic dissolves in the body fluid and elutes out. Generally no more than 4g of antibiotic should be added to each 40g packet of the polymethylmethacrylate (PMMA) polymer (10%) as above this level the biomechanical properties of the cement are affected.

Whilst the role of prophylactic antibiotics is established, is there a need for prolonged systemic therapy when the mainstay of treatment is surgical and such high concentrations can be delivered locally from the bone cement? Review of the literature is not particularly helpful when trying to evaluate the success of two-stage regimens. In Sheffield in the early nineties we took the decision to treat our infections in a two-stage manner on the principle of radical debridement, local elution of antibiotic without the need for prolonged adjunctive systemic therapy. Our results with respect to eradication of infection are comparable with other authors who have used prolonged courses of antibiotics, but have the advantage of reduced cost and no associated morbidity of prolonged therapy.

The timing of reimplantation is based upon the clinical progress of the patient as evidenced by the quality of soft tissues and falling ESR and CRP. The minimum period between stages is usually 6 weeks and there is no need for a prolonged interval before reimplantation according to the monitoring tests. If the soft tissues do not improve or inflammatory markers remain elevated it is preferable to explore and debride as opposed to delay.

The long-term results for successful eradication of infection with two-stage surgery range from 87–92% at 5- and 10-year follow-up.

The results for successful eradication of infection following one stage exchange range from 86–91% at 3- and 10-year follow-up.

If, following debridement, hip reconstruction is not possible or the patient declines further surgery, then excision arthroplasty remains a viable option. Whilst the results for eradication of infection are good, lower limb function, particularly in the younger more active and the elderly frail patients, is poor.

Disarticulation is the final option available to the surgeon and should be reserved for life-threatening sepsis or irreparable neurovascular damage.

Diagnosis and treatment of infected hip arthroplasty requires a multidisciplinary team approach to eradicate infection, alleviate pain, and restore function. Each patient must be treated on an individual basis. The spectrum ranges from conservative suppressive therapy to radical ablation. Most cases that present will be chronically infected so the mainstay of treatment is surgical debridement, which must be radical and complete. Antibiotics obviously have an important but secondary role.

The surgery necessary for these infected cases does not involve any special techniques but it is extensive with respect to excision, becoming on occasion alarming unless the surgeon has at their disposal a comprehensive shelf of instruments and implants, adequate bone graft material, and time for long operative procedures. These are the main reasons why treatment of the infected arthroplasty should be conducted in specialist centres.