To the Editor: Uveitis is an intraocular inflammation which potentially leads to permanent loss of vision.1,2 Tuberculosis is considered to be an infrequent infectious cause of uveitis in the developed world. However, its recurrence as a major public health problem raises the possibility that the incidence of tuberculosis-related uveitis in the developed world may rise.3,4 Uveitis in tuberculosis is presumed to result from either direct invasion or a hypersensitivity reaction.

At the ophthalmology departments of the Erasmus Medical Center and the Eye Hospital in Rotterdam, The Netherlands, all patients presenting with refractory uveitis undergo investigation for a systemic cause, including tuberculin skin testing. When ocular findings are consistent with intraocular tuberculosis, and the tuberculin skin test is positive, while no other cause of uveitis is suggested by symptoms, signs or ancillary testing, then a diagnosis of presumed intraocular tuberculosis is made. Using these criteria, eight cases of presumed intraocular tuberculosis were identified among 89 people referred with refractory uveitis between January 2002 and January 2004. Characteristics of the eight patients are shown in the Box. One patient (F) withdrew from clinical care, and another (A) later had a positive culture result for tuberculosis on lymph node biopsy. This patient had complete remission of uveitis after tuberculostatic treatment, but was excluded from this study as the aim was to assess whether antituberculosis treatment is warranted based solely on a positive tuberculin skin test.

We treated the patients with a complete tuberculostatic regimen (2 months of isoniazid, rifampicin, ethambutol and pyrazinamide, followed by 4 months of isoniazid, rifampicin and ethambutol). All had been previously treated for more than 3 years with immunosuppressive drugs (mainly corticosteroids), either local or systemic, or both, without adequate response.

Main outcome measures were visual acuity and degree of intraocular inflammation seen on ophthalmological examination before and on completion of antituberculosis therapy.

The predominant clinical finding was blurred vision. Five patients exhibited decreased intraocular inflammation and an increase in visual acuity after antituberculosis treatment, allowing tapering of the corticosteroid treatment. One patient had no response. Improvement as part of the natural history was regarded unlikely.

As our department is a tertiary referral centre for patients with uveitis, our patient population is not a representative sample of all patients with uveitis in The Netherlands. Nevertheless, our findings suggest that intraocular tuberculosis should be considered in the differential diagnois of uveitis, even in developed countries.

We believe that, given our results, antituberculosis therapy is justified in patients with uveitis even when a positive tuberculin skin test is the only argument for tuberculosis as the cause of the eye disease. An additional argument for antituberculosis treatment is that many patients with uveitis refractory to immunosuppressive therapy can be adequately treated with tumour necrosis factor-α (TNF-α) blocking drugs.5 However, as severe tuberculosis infection has been described after use of these agents, antituberculosis therapy is warranted in any patient with a positive tuberculin skin test who is a candidate for TNF-α blocking therapy.