Tuberculous Arthritis (Peripheral Joint TB)

Two routes to the joint. TB reaches the joint by haematogenous spread. In primary synovial TB the organism seeds the vascular synovium and causes a granulomatous synovitis; in secondary disease a juxta-articular or epiphyseal osteomyelitic focus erodes into the joint.

The marginal pannus. The granulomatous synovitis forms a pannus that arises from the synovial reflections and creeps over the cartilage from the joint margins, the peripheral 'bare areas' where synovium contacts bone not covered by cartilage. Marginal erosions therefore appear first, often on both opposing sides of the joint ('kissing' lesions), while the central weight-bearing cartilage is spared until late.

Why it is slow. Mycobacterial destruction is granulomatous and lacks the proteolytic enzyme load of pyogenic pus, hence the chronic, indolent course.

The Phemister triad. The classic radiograph shows three signs, each a product of the mechanism above:

• Juxta-articular osteopenia - from hyperaemia and disuse of the painful joint
• Peripheral (marginal) bony erosions - the pannus reaching the bare areas
• Gradual joint-space narrowing - the space relatively preserved until late

The preserved joint space distinguishes TB from pyogenic septic arthritis, which destroys cartilage and the joint space rapidly.

View larger

The history. The onset is insidious: weeks to months of pain, swelling, stiffness, muscle wasting and a limp in a weight-bearing joint, commonly the hip or knee. The swelling is often relatively 'cold', without the acute heat and redness of pyogenic infection, and there are low-grade constitutional symptoms such as night sweats, weight loss, malaise and low-grade fever.

Who to suspect. Because it mimics inflammatory or low-grade pyogenic arthritis, the diagnosis is often delayed. Consider it especially in patients from TB-endemic areas, those with a migration background and the immunocompromised. The age distribution is biphasic: elderly patients in low-incidence countries and younger patients with a migration background.

Get tissue. The diagnosis is confirmed on tissue. Always send samples for both histology and culture, and consider HIV testing.

• Synovial biopsy - caseating granulomas are the histological hallmark
• Mycobacterial culture - the gold standard for confirmation and drug sensitivities, but slow (several weeks)
• Molecular tests (NAAT/GeneXpert PCR) - rapid detection of M. tuberculosis and of rifampicin resistance
• Synovial fluid - typically lymphocytic with low glucose; the AFB smear has low yield

Supporting tests. Raised ESR and CRP and a tuberculin (Mantoux) test or IGRA support the diagnosis. Plain radiographs show the Phemister triad, MRI defines the synovitis, any abscess and the bone and soft-tissue extent, and a chest radiograph looks for pulmonary disease.

Not every cold joint is TB. TB is easily contrasted with acute pyogenic arthritis, but the harder task is the differential of a chronic, indolent monoarthritis, since several conditions mimic it. What confirms TB is caseating granuloma plus a positive mycobacterial culture or PCR, which is why getting tissue, not just aspirate, matters.

Pyogenic arthritis. The acute hot, red, exquisitely painful joint is a surgical emergency, treated with urgent washout and antibiotics; see septic arthritis pathophysiology.

Each of these is developed separately in brucellosis of the spine, fungal osteomyelitis, atypical mycobacterial infection, musculoskeletal sarcoidosis, PVNS, gout and crystal arthropathy and pseudogout and CPPD disease.

Chemotherapy is the mainstay. Standard multi-drug therapy runs an intensive phase of rifampicin, isoniazid, pyrazinamide and ethambutol, then a continuation phase of rifampicin and isoniazid, for a prolonged total course. Local and national protocols commonly specify 9-12 months for bone and joint TB, adjusted for drug resistance. Doses are weight-based and set by WHO and national programmes, so prescribe to the local protocol rather than from memory.

Surgery is adjunctive. Its roles:

• Biopsy for diagnosis
• Drainage of a cold abscess
• Synovectomy or debridement of extensive disease, or to relieve symptoms
• Arthrodesis for a painful destroyed joint
• Delayed total joint arthroplasty for end-stage destruction

Supportive care. Rest or splintage in the acute phase protects the joint and prevents deformity, followed by graded rehabilitation. Nutritional support and management of comorbidity (HIV, diabetes) go alongside.

The late joint. Untreated or late disease destroys the joint, leaving deformity and fibrous or bony ankylosis, especially of the hip and knee, and in children growth disturbance and limb-length discrepancy. A joint destroyed by past TB may present years later as painful, stiff or ankylosed, requiring arthrodesis takedown or arthroplasty.

The protocol range is the safe answer, but the duration of chemotherapy for osteoarticular tuberculosis has been tested. A candidate who knows what the trials showed, and what they did not cover, answers a level above one who quotes a range.

The spinal trials. A systematic review and meta-analysis of six randomised trials (877 patients) compared 6 months against 9 months or longer in spinal tuberculosis. Healing was achieved in 375 of 414 (90.6 percent) on the short course against 404 of 463 (87.3 percent) on the long course, a pooled relative risk of 0.98 (95 percent CI 0.92 to 1.04, p = 0.44): no detectable difference. An earlier literature review found relapse in 0 of 56 treated with 6 months of isoniazid, rifampicin and pyrazinamide against 4 of 218 (2 percent) treated for 9 months or more.

Reading them with care. Heterogeneity between the trials was substantial (I-squared 40.8 percent), and equivalence within a confidence interval is not proof of equivalence.

Why that does not settle the peripheral joint. Every one of those trials studied spinal disease, and no randomised comparison of treatment duration confined to peripheral joint tuberculosis was retrieved. Applying the six-month result to a tuberculous hip or knee is therefore an extrapolation: reasonable, because the organism and the drugs are the same, but one to declare rather than disguise.

Why the longer course persists. National programmes and local protocols still commonly specify it, and it persists on the grounds of poor drug penetration into avascular and necrotic bone, difficulty assessing healing, and the consequences of relapse in a weight-bearing joint. Drug resistance, HIV co-infection, poor adherence and extensive or multifocal disease all push towards the longer end. Treat to the local protocol; know why the debate exists.