A retrospective cohort study of patients treated with anti-tuberculous therapy for presumed ocular tuberculosis
© The Author(s). 2017
Received: 19 September 2017
Accepted: 20 November 2017
Published: 4 December 2017
Uveitis involving the posterior segment is a significant and potentially blinding condition. The diagnosis and treatment of patients with uveitis associated with tuberculosis remains controversial, and commonly, patients are systemically well. Use of the interferon-gamma release assays has added to the controversy, as the significance of a positive test may be uncertain. We aim to report the outcomes of anti-tuberculous treatment in a cohort of patients treated in Birmingham, for presumed “ocular tuberculosis”, based on clinical findings, systemic assessment and specific testing for tuberculosis.
We found that in our cohort of 41 patients treated between 2010 and 2014, the majority achieved disease-free remission, even in cases where anti-tuberculous treatment was delayed.
Despite controversy, this study strongly supports the use of anti-tuberculous therapy in such patients and highlights the need for formal prospective trials and treatment protocols.
Intraocular inflammation, generally termed “uveitis”, may result from a hugely diverse range of causes, both infectious and non-infectious. Uveitis is classified by the predominant anatomical site of inflammation within the eye , the cause and whether or not it is related to an infectious agent or an underlying systemic condition . Uveitis is an important cause of visual loss, often affecting the working age population [3, 4].
Uveitis related to tuberculosis is a well-recognized clinical entity, with an extensive literature dating back more than 100 years . In some parts of the world, tuberculosis is one of the principle causes of uveitis [6–9].
Currently, a large proportion of patients in the developed world is described as having “non-infectious”, “idiopathic” or “undifferentiated” uveitis [10, 11]. Often no underlying systemic association or infection is identified despite extensive investigation. If such patients have significant or “sight-threatening” uveitis, they are generally treated with systemic corticosteroid, and sometimes systemic immunosuppression, including biologics [12, 13]. Such therapy is aimed at preventing relapses, as each relapse carries a risk of irreversible visual loss, morbidity and ocular complications .
The association between uveitis and “latent” or occult tuberculosis is not new yet has become increasingly relevant. Many patients display clinical appearances identical to those seen in patients with active TB, without demonstrating systemic manifestations of infection. Ocular appearances associated with TB are heterogeneous, making diagnosis challenging, and include (among others) choroiditis, serpiginous-like choroiditis, granulomatous uveitis, retinal vasculitis and intermediate uveitis . In addition, the condition known as “Eales” disease, a bilateral occlusive retinal vasculitis, with minimal inflammation and often complicated by vitreous haemorrhage, is observed more frequently in TB-endemic populations.
A large proportion of the global population will have immunological evidence of latent TB, and consequently, a significant proportion of patients with ocular inflammation will have evidence of latent TB, whether or not it is causal.
Studies investigating intraocular samples provide support for a central role of TB. Analysis of the aqueous humor, vitreous gel or epiretinal membranes using polymerase chain reaction (PCR) methods [16–20] demonstrates that a higher proportion of samples from uveitis patients are positive for the TB genome compared with samples taken from patients in a similar population undergoing surgery for other reasons, and a positive PCR result correlates with a clinical response to anti-tuberculous treatment (ATT) [19, 20].
Traditionally, the tuberculin skin test (Mantoux) has been used to determine whether a patient is infected with TB. The test may be affected by multiple factors, making it hard to determine its significance. More recently, interferon gamma release assays (IGRAs) have added significant controversy to the diagnosis and management of “ocular TB” [21–26]. Such assays, including the T-spot or QuantiFERON GOLD tests rely on the observation that T-cells will release interferon gamma when exposed to a specific TB antigen. The test is highly specific and not affected by previous BCG vaccinations. A positive IGRA test usually indicates that a patient has been exposed to TB; however, it cannot distinguish between latent infection and active disease.
Multiple recent reports support the use of anti-tuberculous therapy (ATT) in patients with a consistent uveitic phenotype and positive TB investigations [27–30]. There is a lack of consensus regarding treatment indications, treatment regimens, investigation protocols and what constitutes a successful outcome. Even the diagnosis of “ocular TB” is not standardized .
Importantly, the way in which patients are managed is believed to differ significantly between units, with some uveitis specialists unable to persuade the respiratory physician of the relevance of tuberculosis, whilst other specialists are able to access anti-TB treatment easily and as a result have a very low treatment threshold . Some uveitis specialists elect to refer patients for ATT only if the uveitis is severe, whilst continuing to treat “mild” recurrent anterior uveitis with topical steroid.
The principle indication for using ATT in uveitis is to treat the underlying systemic drive, which is believed to come from occult infection, antigenic mimicry or a hypersensitivity-type reaction to TB antigen. A secondary indication may be to ensure that systemic immunosuppression is safe in the face of a positive T-spot result. This is especially relevant for patients started with anti-TNF therapy .
The Birmingham and Midland Eye Centre is located in an inner-city part of Birmingham where the large surrounding population comprises of ethnically diverse communities with a majority of people originating from Southeast Asia, especially countries such as Pakistan, India and Bangladesh, where the prevalence of TB is high.
Aims and objectives
The aims of this study were first, to define and report the treatment outcomes of patients, who received anti-tuberculous therapy for uveitis, and second, to characterize the patient cohort seen at the Birmingham and Midland Eye Centre who were diagnosed as having ocular TB.
This was a retrospective evaluation of patient notes. Ethical approval was not required as the project was deemed to be a service evaluation.
Using the regional TB disease register obtained from the regional infectious diseases department, patients who had been diagnosed as having uveitis related to TB and who were started on ATT between 2010 and 2014 were identified. Patients who completed a course of treatment and were followed up for at least 12 months were included in the study.
A retrospective analysis of patient notes was undertaken. Ethical approval was not required as this was deemed to be a service evaluation and made use of anonymized retrospective patient information. Data pertaining to patient characteristics, uveitis phenotype, vision, duration of uveitis prior to therapy, date of commencement of ATT, treatment details, duration of follow-up and disease activity were collected. Data were also collected regarding systemic investigations for TB.
At presentation to the uveitis service, patients underwent a full ophthalmic examination and were assessed with regard to the site and severity of their uveitis, using the SUN criteria . A full systemic workup was undertaken, and patients underwent an extensive panel of investigations, with the aim of excluding other causes. All underwent chest radiograph or CT scan, with most also undergoing Mantoux testing or interferon gamma release assay, usually using T-spot. Investigations were not protocolized, and patients were investigated appropriately in an individualized fashion.
A consistent ocular phenotype, together with findings on chest imaging and/or positive T-spot or Mantoux testing, with the exclusion of other causes of uveitis, led to a diagnosis of presumed ocular TB.
Following this diagnosis, patients were referred urgently to the respiratory team where they were investigated further as necessary. Throughout the course of the study, the availability and use of interferon gamma release assays increased. This meant that some patients were not diagnosed with TB-related uveitis until a significant period of ocular inflammation had lapsed.
In most patients, disease activity was observed as a number of relapses, rather than continual inflammation. Treatment with anti-inflammatory medication, mainly corticosteroids, was used. These were administered topically, peri-ocularly or systemically. Relapse on withdrawal of therapy warranted repeat treatment.
Ocular disease activity was quantified using the number of “flare-ups” of uveitis. A uveitis “flare” was defined as the need for augmentation of treatment, with high-dose oral steroid (usually at least 40 mg daily tapered over several weeks), intravenous steroid, periocular steroid or hourly topical steroid. Where possible, “flare rate” prior to ATT was defined as the number of flare-ups per unit time of disease. Calculating a “flare rate” was not possible if a patient was treated with ATT immediately or had a very short duration of disease prior to ATT.
Data were reviewed using descriptive statistics. The “time to flare” post-treatment was calculated using Kaplan-Meier survival curves. Remission was defined as the absence of flare-ups for at least 6 months whilst long-term remission was defined as at least 12 months of disease quiescence after completion of therapy. Flare rates were compared before and after therapy using the Wilcoxon signed rank test.
Characteristics of the patient cohort
Gender (number of patients (%))
Age, years, mean (range)
Ethnicity (number of patients (%))
Diabetic (number of patients (%))
Site of uveitis (number of patients (%))
Bilateral disease (number of patients (%))
Uveitis features present (NB some patients may have had more than one feature)
Granulomatous anterior uveitis/ mutton fat KP
Maintenance treatment prior to ATT
Number of patients (%)
Topical steroid alone
Oral steroid alone
Oral steroid and immunosuppression
Topical steroid and immunosuppression
Topical and systemic steroid
Uveitis was predominantly bilateral with most having disease affecting the posterior segment. Isolated retinal vasculitis, including the “Eales’ phenotype”, was categorized as posterior uveitis. A wide range of uveitic phenotypes was observed, including granulomatous anterior uveitis, retinal vasculitis, nodular scleritis, choroiditis and intermediate uveitis.
The date of first presentation with uveitis preceded the diagnosis of TB-related eye disease in the majority, with four patients starting ATT at the same time as their uveitis was diagnosed. The duration of disease prior to ATT ranged from 8 to 4495 days (more than 12 years) with a median of 12 months. Prior to ATT, 31 patients were on maintenance therapy for uveitis as described in Table 1.
Investigations to support the diagnosis of active or latent tuberculosis
Investigations to support TB
Number of patients
No details available
Chest CT changes
With positive T-spot or Mantoux test
Positive Mantoux alone
Positive T-spot alone
Positive T-spot and Mantoux test
Typical X-ray changes present
With a positive T-spot and/or Mantoux test
Systemic symptoms presenta
With a positive T-spot and/or Mantoux test
With consistent CT or Chest X-ray findings
ATT regimens, duration and compliance
R + E
R + Z
Anti-TB (unknown specific treatment)
Mean duration of therapy
6.8 months (range 4–12)
Disease activity following completion of therapy
Of the 47 patients who complied with ATT, follow-up data were available for 41, with the remaining patients not attending follow-up or having missing data. Duration of clinic follow-up ranged from 20 to 2192 (median 557) days with five patients being lost to follow-up or who had moved elsewhere. Patients completing therapy were instructed to attend eye casualty in the event of flare-up, and it was assumed that their disease remained quiescent if no attendances were recorded. The mean clinic follow-up was 808 days.
A survival curve including patients for which both pre- and post-treatment data was available is shown in Fig. 1b. This highlights that 94% of patients were flare free at 6 months and 86% were flare free at 12 months.
The data set was then divided into two groups, those patients who were treated with ATT within 8 months of their first uveitis episode (group 1) and the remaining (group 2) who had had a diagnosis of uveitis for more than 8 months. This was to see whether patients with a longer duration of uveitis were less likely to respond favorably to ATT. Kaplan-Meier and log-rank tests (shown in Fig. 1c) were used for analysis. No significant difference in the time to flare was found (P = 0.565 log-rank) between the two groups.
The monthly flare rate pre-ATT treatment was calculated from the first uveitis episode to initiation of ATT. It was assumed the nine patients with a very short time to treatment (< 4 months) would have had just one flare in the time period. Post-treatment flare rates were calculated from time of ATT completion to last day of data collection (December 2016). Flare rates after ATT were significantly lower than those before (Wilcoxon signed rank P = 0.000).
The data was further split into two groups; group 1 included patients treated within 8 months of their first uveitis flare. Group 2 included those who were treated more than 8 months after their first uveitis flare. Flare rates post-ATT treatment for both groups were significantly lower than pre-ATT (Wilcoxon signed rank P = 0.000, P = 0.000 respectively).
To assess whether the short time to treatment following first uveitis episode had an overall effect on the significant difference in pre- and post-ATT flare rates, the nine patients with less than 4 months follow-up were excluded and the data re-analysed. The post-ATT flare rates remaining significantly lower following the exclusion of these patients (P = 0.000 Wilcoxon signed rank). Of the nine patients excluded, all were flare free at 12 months.
Disease flare rates were also calculated using rates per person year (PPY). Prior to ATT treatment, the flare rate was 0.69 while after ATT, this reduced to 0.14. The difference was statistically significant (P = 0.000 mid-P exact).
The principal finding from this study is that ATT in a real-life clinical setting appears to significantly reduce the number of flare-ups and to enable long-term remission in patients with presumed ocular tuberculosis. This beneficial effect appears to persist even in patients who experienced a long delay between the onset of uveitis and the diagnosis of ocular TB.
The patient cohort is diverse and strict diagnostic criteria are absent, requiring a high degree of clinical judgment and collaboration between specialists when treating such patients.
The study has limitations in view of the “real-life” clinical setting. Many were lost to follow-up or non-compliant with ATT. “Flare rate” was used as a measure of disease activity but may not accurately reflect disease in patients with ongoing activity, where uveitis is chronically uncontrolled, or being suppressed with corticosteroid. ATT was not standardized, and consequently, patients received a range of regimens. The majority however received at least 6 months of therapy. There is evidence to support treating patients for at least 9 months  although a wide range of treatment regimens and durations is reported in the literature.
Recent studies investigate therapeutic outcomes and the beneficial effects of ATT. One meta-analysis concluded that ATT in the management of patients with presumed ocular TB enabled remission in 84% . Another large analysis explored whether any factors were associated with a poorer outcome and found that prior immunosuppression and/or a high absolute QuantiFERON GOLD level were associated with poorer resolution of inflammation post-ATT or an ongoing need for systemic steroid .
This study highlights challenges and controversies in defining outcomes in such cohorts. We attempted to look at therapeutic outcomes in a number of ways, including comparing flare rates pre and post, reporting time to flare using Kaplan-Meier survival curves and also calculating “flare rates” before and after treatment. We also recognize that such measures may introduce bias and favour a benefit in patients with a short period of disease prior to the use of ATT. For this reason, we compared results between patients with a long duration of disease and a shorter duration of disease prior to ATT. Despite some patients experiencing significant delays in being started on ATT, we showed that such patient still displayed a clear benefit from treatment.
We did not report visual outcomes or complications. Whilst we show that there is a significant treatment benefit, even when duration of uveitis is long, patients with longer disease courses prior to ATT will be more likely to suffer visual loss and complications. Therefore, early treatment should be recommended .
In conclusion, this retrospective study shows that despite the unresolved controversies regarding diagnosis, the relevance of interferon gamma assays and what exactly ocular tuberculosis is, if patients have a consistent uveitis phenotype and other uveitic causes are excluded, then ATT is likely to have a beneficial effect.
We would like to acknowledge Peter Nightingale for his assistance with the statistical analysis, Dr. Nazim Nathani, Miss Sahar Al Hussainy, Dr. Guy Hagan and Miss Panagiota Stavrou for their involvement in the clinical care of these patients.
No funding resources were used for this research project.
EMD conceived the original design, supervised data collection and statistical analysis and wrote the manuscript. SD participated in data collection and statistical analysis and contributed to the manuscript. CM participated in data collection and contributed to the manuscript. XL and JH participated in data collection and review of the manuscript. AD, EM, MD and PIM participated in clinical care of the patients, study design and review of the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
This study used retrospective observational clinical data. Ethical approval was not required as this was deemed to be an evaluation of service.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- SUN WORKING GROUP (2005) Standardization of uveitis nomenclature for reporting clinical data. Results of the first international workshop. Am J Ophthalmol 140(3):509–516. doi:10.1016/j.ajo.2005.03.057 View ArticleGoogle Scholar
- Deschenes J, Murray PI, Rao NA, Nussenblatt RB (2008) International Uveitis Study Group International Uveitis Study Groupd (IUSG): clinical classification of uveitis. Ocul Immunol Inflamm 16(1):1–2.Google Scholar
- Durrani OM, Tehrani NN, Marr JE, Moradi P, Stavrou P, Murray PI (2004) Degree, duration, and causes of visual loss in uveitis. Br J Ophthalmol 88:1159–1162. doi:10.1136/bjo.2003.037226 View ArticlePubMedPubMed CentralGoogle Scholar
- de Smet MD, Taylor SRJ, Bodaghi B et al (2011) Understanding uveitis: the impact of research on visual outcomes. Prog Retin Eye Res 30(6):452–470. doi:10.1016/j.preteyeres.2011.06.005 View ArticlePubMedGoogle Scholar
- Davidson M (1929) The modern approach to the problem of ocular tuberculosis: Werdenberg, E. Schweiz. Am J Ophthalmol 59(6):126Google Scholar
- Al-Baker ZM, Bodaghi B, Khan SA (2016) Clinical patterns and causes of uveitis in a referral eye clinic in Qatar. Ocul Immunol Inflamm 0(0):1–10. doi:10.1080/09273948.2016.1206573
- Sabhapandit S, Murthy SI, Singh VM et al (2016) Epidemiology and clinical features of uveitis from urban populations in South India. Ocul Immunol Inflamm 0(0):1–7. doi:10.1080/09273948.2016.1236971
- Dogra M, Singh R, Agarwal A et al (2016) Epidemiology of uveitis in a tertiary-care referral institute in North India. Ocul Immunol Inflamm 0(0):1–8. doi:10.1080/09273948.2016.1255761
- Nguyen M, Siak J, Chee S-P, Diem VQH (2016) The spectrum of uveitis in Southern Vietnam. Ocul Immunol Inflamm 0(0):1–7. doi:10.1080/09273948.2016.1231826
- Wong A, McKelvie J, Slight C, Sims J (2017) Land of the Long White Cloud: The Spectrum of Uveitis at a Tertiary Referral Center in New Zealand. Ocul Immunol Inflamm 25(sup1):S115–S121.Google Scholar
- Jones NP (2015) The Manchester Uveitis Clinic: the first 3000 patients—epidemiology and casemix. Ocul Immunol Inflamm 23(2):118–126. doi:10.3109/09273948.2013.855799 View ArticlePubMedGoogle Scholar
- Levy-Clarke G, Jabs DA, Read RW, Rosenbaum JT, Vitale A, Van Gelder RN (2014) Expert panel recommendations for the use of anti-tumor necrosis factor biologic agents in patients with ocular inflammatory disorders. Ophthalmology 121(3):785–796. doi:10.1016/j.ophtha.2013.09.048 View ArticlePubMedGoogle Scholar
- Jabs DA, Rosenbaum JT, Foster CS et al (2000) Guidelines for the use of immunosuppressive drugs in patients with ocular inflammatory disorders: recommendations of an expert panel. Am J Ophthalmol 130(4):492–513 Available at: http://www.ncbi.nlm.nih.gov/pubmed/11024423 View ArticlePubMedGoogle Scholar
- Tomkins-Netzer O, Talat L, Bar A et al (2014) Long-term clinical outcome and causes of vision loss in patients with uveitis. Ophthalmology 121(12):2387–2392. doi:10.1016/j.ophtha.2014.07.007 View ArticlePubMedGoogle Scholar
- Varma D, Anand S, Reddy AR et al., Tuberculosis: an under-diagnosed aetiological agent in uveitis with an effective treatment. Eye (Lond) 20(9):1068–1073 doi:10.1038/sj.eye.6702):1068-73. doi:10.1038/sj.eye.6702093
- Singh R, Toor P, Parchand S, Sharma K, Gupta V, Gupta A (2012) Quantitative polymerase chain reaction for mycobacterium tuberculosis in so-called Eales’ disease. Ocul Immunol Inflamm 20(3):153–157. doi:10.3109/09273948.2012.658134 View ArticlePubMedGoogle Scholar
- Singh U, Mohapatra S, Wagh V, Porwal C, Kaushik A, Rajpal (2015) Association of mycobacterium tuberculosis in the causation of Eales’ disease: an institutional experience. Indian J Med Microbiol 33(5):43. doi:10.4103/0255-0857.148829
- Bansal R, Sharma K, Gupta A et al (2015) Detection of mycobacterium tuberculosis genome in vitreous fluid of eyes with multifocal serpiginoid choroiditis. Ophthalmology 122(4):840–850. doi:10.1016/j.ophtha.2014.11.021 View ArticlePubMedGoogle Scholar
- Biswas J, Kazi M, Agarwal V, Alam MS, Kl T (2016) Polymerase chain reaction for mycobacterium tuberculosis DNA detection from ocular fluids in patients with various types of choroiditis in a referral eye center in India. Indian J Ophthalmol 64(12):904. doi:10.4103/0301-4738.198857 View ArticlePubMedPubMed CentralGoogle Scholar
- Bhagya S, Lalitha P, Kumar AL, Rathinam S (2017) Polymerase chain reaction and its correlation with clinical features and treatment response in tubercular uveitis. Ocul Immunol Inflamm 0(0):1–8. doi:10.1080/09273948.2017.1287925
- Ang M, Wong W, Ngan CCL, Chee S-P (2012) Interferon-gamma release assay as a diagnostic test for tuberculosis-associated uveitis. Eye (Lond) 26(5):658–665. doi:10.1038/eye.2012.1 View ArticleGoogle Scholar
- Cordero-Coma M, Calleja S, Torres HE et al (2010) The value of an immune response to mycobacterium tuberculosis in patients with chronic posterior uveitis revisited: utility of the new IGRAs. Eye (Lond) 24(1):36–43. doi:10.1038/eye.2009.51 View ArticleGoogle Scholar
- Ang M, Chee S-P (2017) Controversies in ocular tuberculosis. Br J Ophthalmol 101(1):6–9. doi:10.1136/bjophthalmol-2016-309531 View ArticlePubMedGoogle Scholar
- Bramante CT, Talbot E a, Rathinam SR, Stevens R, Zegans ME (2007) Diagnosis of ocular tuberculosis: a role for new testing modalities? Int Ophthalmol Clin 47(3):45–62. doi:10.1097/IIO.0b013e318074de79
- Ang M, Vasconcelos-Santos DV, Sharma K et al (2016) Diagnosis of ocular tuberculosis. Ocul Immunol Inflamm 3948(July):1–9. doi:10.1080/09273948.2016.1178304 View ArticleGoogle Scholar
- Ang M, Htoon HM, Chee SP (2009) Diagnosis of tuberculous uveitis: clinical application of an interferon-gamma. OPHTHA 116(7):1391–1396. doi:10.1016/j.ophtha.2009.02.005 View ArticleGoogle Scholar
- Bansal R, Gupta A, Gupta V, Dogra MR, Bambery P, Arora SK (2008) Role of anti-tubercular therapy in uveitis with latent/manifest tuberculosis. Am J Ophthalmol 146(5). doi:10.1016/j.ajo.2008.06.011
- Agrawal R, Gupta B, Gonzalez-lopez FJJ, et al. The role of anti-tubercular therapy in patients with presumed ocular tuberculosis. 2015;23(October 2014):40-46. doi:10.3109/09273948.2014.986584
- Ang M, Hedayatfar A, Wong W, Chee S-P (2012) Duration of anti-tubercular therapy in uveitis associated with latent tuberculosis: a case–control study. Br J Ophthalmol 96(3):332–336. doi:10.1136/bjophthalmol-2011-300209 View ArticlePubMedGoogle Scholar
- Kee AR, Gonzalez-lopez JJ, Al-hity A et al (2016) Anti-tubercular therapy for intraocular tuberculosis: a systematic review and meta- analysis. Surv Ophthalmol 61(5):628–653. doi:10.1016/j.survophthal.2016.03.001.This
- Lou SM, Larkin KL, Winthrop K, Rosenbaum JT (2015) Lack of consensus in the diagnosis and treatment for ocular tuberculosis among uveitis specialists. Ocul Immunol Inflamm 23(1):25–31. doi:10.3109/09273948.2014.926936
- Conant MM, Vrasich CR, Wongskhaluang J V, et al. Role of the infectious disease consultant in management of patients with tuberculosis-associated ocular inflammation. 2015:1-5. doi:10.1093/ofid/ofv195
- Keane J, Gershon S, Wise R, et al. Tuberculosis associated with Infliximab,. 2001;345(15):1098-1104.Google Scholar
- Agrawal R, Gonzalez-Lopez J, Nobre-Cardoso J et al (2013) Predictive factors for treatment failure in patients with presumed ocular tuberculosis in an area of low endemic prevalence. Br J Ophthalmol 17(11):545149. doi:10.1136/bjophthalmol-2014-306474
- Gunasekeran DV, Gupta B, Cardoso J, Pavesio CE, Agrawal R (2017) Visual morbidity and ocular complications in presumed intraocular tuberculosis: an analysis of 354 cases from a non-endemic population. Ocul Immunol Inflamm 0(0):1–5. doi:10.1080/09273948.2017.1296580