Total hip replacement for patients with active tuberculosis of the hip

The incidence of osteoarticular tuberculosis (TB) in western countries is rising as a result of an increase in immigration from regions where TB is endemic, an increase in the number of people with immune suppression, an ageing population and the development of drug-resistant strains of Mycobacterium tuberculosis.^1-4 TB of the hip constitutes approximately 15% of all cases of osteoarticular TB and is the most frequent site of bone involvement after the spine.⁵

Traditionally, two forms of surgical treatment were available for patients with advanced active TB of the hip: arthrodesis in a position of function and Girdlestone’s excision arthroplasty,^5-7 which may relieve pain and control infection, but the function of the hip is unsatisfactory.⁸ Some consider total hip replacement (THR) in these patients to be contraindicated because the risks of further activation of infection are too high.⁵ Others recommend a long interval between the treatment of the active infection and THR.^9,10 Our current knowledge of the outcome of THR in patients with active TB is confined to a few reviews of experience from a single institution or multicentre registries in one country. The incidence of arthritis of the hip secondary to TB and the outcome following treatment differ between people of different ethnic backgrounds.^5,11,12 Among all cases of osteoarticular TB, the incidence of TB of the hip has been reported to be 15% in India⁵ and 21% in Denmark.¹³ Further understanding of the place of THR in the management of patients with active TB would need a large database to generate adequate power.

The purpose of this study was to review the literature in order to assess the outcome of THR in patients with active TB. On the basis of previous literature,¹⁴ our hypothesis was that THR in these patients is a safe procedure providing symptomatic relief and functional improvement.

Materials and Methods

Literature review

We conducted a systematic review of the available literature using various search strategies. The following databases were searched: MEDLINE, PubMed, CINAHL and Cochrane systematic reviews, using the terms ‘arthroplasty’, ‘hip’ and ‘tuberculosis’. Two authors (SJK and JHK) completed the search separately and the results were obtained twice by each. The search was performed on 15 July 2012, and repeated four weeks later to ensure accuracy. No additional study was identified by repeating the search. The title, abstract and full text were reviewed when the title or abstract suggested that the papers were appropriate. These were then discussed among the authors and a decision was made regarding inclusion.

The inclusion criteria were papers published between 1 January 1950 and 1 July 2012; papers written in English about TB in humans; electronic publications that reported cases of TB of the hip; both retrospective and prospective series; only patients with active TB who underwent THR; and only papers that reported the outcome assessed by Harris hip scores (HHS),¹⁵ reinfection rates and complications.

The exclusion criteria included patients with quiescent TB; papers dealing with the technique of THR; papers that did not include the HHS; evaluation of any joint other than the hip, including knee and shoulder; and papers not written in English. No limits on the number of patients in each study or the minimum follow-up were included.

A first search of the MEDLINE database yielded 113 papers and a second search of PubMed using the same search strategy yielded 106 papers. The literature search is summarised in Figure 1. There were 121 papers that appeared in more than one of the four searches, yielding a total of 286 unique papers. We selected the most relevant papers from retrospective cross-sectional studies, clinical registries or prospective studies. If there was a disagreement among the authors regarding the inclusion of a paper, the senior author (JHK) made the final decision. A manual search was also performed from the references of the selected papers to identify any important reports that had been missed. Finally, the full texts of the six papers were obtained. Owing to a lack of prospective randomised studies, most of the larger cohorts giving an answer or at least an insight to the clinical problems were selected for this review. A meta-analysis was not performed because most papers involved a heterogeneous group of patients with multiple comorbidities. Studies of THR in patients with active TB started mainly after the paper by Yoon et al¹⁶ in 2001, although there were few isolated earlier studies that included patients with a history of TB.^10,17

Data extraction

The following data were extracted: demographics, including age, gender, a history of TB, the diagnosis and management of TB, post-operative antibiotic treatment, the outcome following THR including clinical resolution, post-operative ESR and HHS, reactivation of TB, and other complications including dislocation of the THR, infection, heterotopic ossification, and neurological or vascular injury.

Statistical analysis

Data were recorded using Microsoft Excel 2007 (Microsoft Corporation, Redmond, Washington) and analysed using SPSS software, v15.0 for Windows (SPSS Inc, Chicago, Illinois). The pre-operative HHS¹⁵ was not available in one paper,¹⁸ so these data were not included in the statistical analysis. Pre- and post-operative HHSs were obtained from the other five papers and compared using Wilcoxon’s signed ranks test. A p-value < 0.05 was considered to be statistically significant.

Results

A review of MEDLINE, PubMed, CINAHL and Cochrane literature searches revealed a total of 65 patients from the six selected papers.^8,14,18-21 Although complete data were not available, data such as gender, underlying medical/surgical conditions, reactivation and complications were collected. The ranges of age and follow-up were clearly identified in all the papers.

Demographic characteristics

There were 40 male and 25 female patients. The mean follow-up was 53.2 months (24 to 108). A total of 16 patients (25%) had associated pulmonary TB. Demographic information is given in Table I. In two patients the diagnosis was proven by pre-operative biopsy,¹⁸ whereas the diagnosis in the others was made on a clinicoradiological basis with confirmation obtained histologically and by the polymerase chain reaction (PCR) of tissue samples taken at the time of surgery.

Treatment

The presence of a sinus tract into the pelvis or thigh was not a contraindication for THR.^8,19 A total of 16 patients (25%) did not receive antituberculosis medication pre-operatively but the remaining 49 received treatment for between two weeks and three months pre-operatively. The inflamed soft tissues and the necrotic bone were completely excised at operation. THR involved cementless press-fit components in 38 patients (58%) and cemented components in 27 (42%). All patients received antibiotics for between six and 15 months post-operatively. This involved treatment with rifampicin, isoniazid, ethambutol and pyrazinamide. In one paper streptomycin was used instead of pyrazinamide.⁸ Cephalosporins were given post-operatively in four studies to prevent superadded infection.^8,14,19,20

Outcomes

The ESR became normal (< 20 mm/h) between four and 12 months post-operatively and no patient had an elevated ESR at subsequent follow-up examinations. The mean HHS improved from 35.5 (4 to 46) pre-operatively to a mean of 91.7 (56 to 98) at the final follow-up (mean 53.2 months (24 to 108)) (p < 0.001). The post-operative HHS ranged from 86 to 98 excluding one patient with reactivation, whose score at follow-up (35 months) was 56. This was the only patient with reactivation of the infection.¹⁹ This patient had a cemented femoral component, was non-compliant with antituberculosis treatment, had a post-operative infection with Staphylococcus aureus, and was lost to follow-up two months post-operatively. She subsequently underwent a resection arthroplasty. One patient had a dislocation that was managed successfully by closed reduction and immobilisation for two weeks.¹⁸ Two patients developed Brooker grade II²² heterotopic ossification but the function of the hip was not affected.^18,21 There were no neurological or vascular complications in any of the patients. The outcomes after THR are shown in Table II.

Discussion

THR has been used successfully in patients with quiescent TB¹⁷ but it is unclear whether it should be performed in patients with active infection. To date, only a few studies describing THR in patients with active TB have been reported.^8,14 We aimed to evaluate the outcome and to assess whether THR should be offered to patients with active TB. Additionally, the treatment protocols of the reported studies were reviewed.

THR has previously been considered to be contraindicated in these patients owing to the risks of further activation of infection.^5,23 Hardinge et al²⁴ recommended postponing THR until any sinuses had not discharged for 20 years, or until the affected hip had been ankylosed for more than ten years.^24,25 Reactivation or superinfection can be a devastating complication of THR. Mycobacterium tuberculosis has particular biological and behavioural characteristics that differ from pyogenic bacteria.⁸ Metal implants have been used in patients with spinal tuberculosis with good results without reactivation.²⁶In vitro studies comparing the adherence and biofilm properties of Mycobacterium tuberculosis and Staph. epidermidis suggest that the former rarely adheres to a metal surface and has little or no biofilm formation, thus making it more susceptible to antibiotic treatment.^27,28 Therefore, antituberculosis treatment is crucial following THR in patients with active TB.⁸ Rifampicin, which was used in combination with isoniazid, ethambutol and pyrazinamide, either separately or as quadruple therapy, is well known for its intracellular action and its ability to enter the biofilm.²⁹ In our pooled analysis the TB was successfully managed with one-stage THR and antituberculosis chemotherapy. The functional results of THR as assessed by the HHS were satisfactory. There was reactivation of TB in only one of 65 patients. The complication rate after THR was relatively low (Table II). Therefore, THR can be confidently recommended for patients with active TB provided that adequate antibiotic treatment is given post-operatively. Although most authors continued post-operative treatment for one year, the duration can probably be reduced to six months¹⁴ if the post-operative inflammatory markers are normal and clinical findings satisfactory, in accordance with guidelines of the Infectious Diseases Society of America, which recommend six to nine months’ antituberculosis treatment for patients with osteoarticular TB.³⁰ Complete curettage and debridement of the infected tissues is required for a successful outcome.^14,16,20

The choice of cemented or cementless components appears to have no influence on the rate of reactivation.^10,20,31 Harris¹⁷ and Kim et al¹⁰ have reported similar reactivation rates for both cemented and cementless THRs in patients with quiescent TB. In our study, one patient with a cemented THR had reactivation as a result of non-compliance with chemotherapy. There is little information on the value of antituberculosis drugs mixed with cement.^32,33 Isoniazid, rifampicin and ethambutol are heat stable but the risk of toxicity and poor elution currently precludes their use.⁴ The reasons for the reactivation of cured or quiescent disease are uncertain. It may be due to the local effect of trauma, the arthroplasty itself, or the activation of an immune inflammatory response.²¹ It may also be as a result of the immunosuppressive effect of major surgery and anaesthesia.²¹ The treatment of patients with reactivated infection after THR remains controversial. Resection arthroplasty, and treatment with antituberculosis chemotherapy without removing the components, have both been reported.^18,34-36

Our study has some limitations. First, the short follow-up and the retrospective design means that diagnostic criteria, surgical approaches (e.g. posterior vs lateral), the medical management and the post-operative rehabilitation of the patients were not standardised. Second, this was a pooled analysis of the studies from large international administrative databases (MEDLINE, PubMed, CINAHL and Cochrane) that were not designed for a clinical research study, and potentially more detailed useful information was not available.

In conclusion, THR in patients with active TB of the hip is a safe procedure providing symptomatic relief and functional improvement. There is a low risk of reactivation of infection and few complications. There is no need to delay surgery until the quiescent stage of the disease.

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

This article was primary edited by J. Scott and first-proof edited by G. Scott.