Overview of Management

We conducted a prospective study of consecutive patients presenting with new onset of uveitis to the Singapore National Eye Centre Ocular Inflammation and Immunology Service over a 2-year period (January 1, 2009, through December 31, 2010). Ethical approval was obtained from the Singapore Health Services Centralized Institutional Review Board, and our study adhered to the tenets of the Declaration of Helsinki. After we obtained informed consent, all patients underwent a full systemic review, ocular examination and standard baseline investigations as previously described. We included all patients who were undergoing systemic review for acute uveitis and who gave informed consent to be enrolled in the study. We excluded patients who did not consent to the minimum follow-up period of 1 year after completion of ocular and systemic therapy.

Investigations

At presentation, all patients were tested with a standard panel of investigations as described, essentially: a complete blood count, erythrocyte sedimentation rate analysis, liver enzyme panel analysis, and infectious-disease screening, which included a venereal disease research laboratory test for syphilis, a tuberculin skin test, a urine microscopy, and a chest X-ray. Other tests, such as an HLA-B27 screen, an acid- and alcohol-fast bacilli smear from throat swabs, and PCR assays for TB DNA from ocular samples were performed if the patient consented to the procedure. Blood was taken for QuantiFERON-TB Gold In-Tube and T-SPOT.TB testing before the tuberculin skin test was performed so as to avoid any boosting effect (although it has been shown that this is unlikely to be significant). The tuberculin skin test was performed using the standard Mantoux method: intradermal injection of 0.1 mL (2 tuberculin units) purified protein derivative (RT23 SSI-2T.U./0.1 mL [Statens Serum Institut, Copenhagen, Denmark]).

Induration was measured at 72 hours with a ruler and was considered positive if it was more than or equal to 15 mm (as validated in our population).

T-SPOT.TB was performed according to the manufacturer’s instructions. For each patient, 8 mL of blood was collected in lithium heparin tubes and processed within 8 hours of sampling. Peripheral blood mononuclear cells were prepared by density-gradient centrifugation over Ficoll Paque Plus (GE Healthcare, Princeton, NJ). We seeded 250 000 cells in each of 4 wells of the assay plate. The cells were stimulated for 16 to 20 hours (under 5% carbon dioxide at 37°C) with GIBCO AIM-Vmedium (nil control), phytohaemagglutinin (mitogen-positive control, Oxford Immunotec, Oxford, United Kingdom), or the TB-specific peptide antigens (peptide pools forESAT-6 and CFP-10 in separate wells) in a total volume of 150 μL per well. Two readers quantified the number of interferon gamma spot-forming T cells visually, and a third reader was consulted if the results were disparate. The T-SPOT.TB test was considered positive if there were >8 spots compared to the negative control well; negative if there were <4 spots compared to the control well; or equivocal if the test wells had 5 to 7 spots more than the control. If the negative control well had >10 spots and/or <20 spots in the mitogen-positive control wells, the result was considered to be indeterminate.

QuantiFERON-TB Gold In-Tube was performed according to the recommended guidelines. Whole blood from each patient was divided into 3 tubes of 1 mL each (nil control, positive control, and TB-specific antigens [ESAT-6, CFP-10 and TB7.7]). Samples were incubated with the stimulating antigens for 16 to 24 hours at 37°C. Afterward, plasma samples were harvested and the amount of interferon gamma released was measured via ELISA. The result obtained in the nil control was subtracted from the mitogen control and the antigen-stimulated samples. The result was considered positive if the response to the specific antigens was ≥0.35 IU/mL, regardless of the level of the positive control; negative if the response to the specific antigens was <0.35 IU/mL and the interferon-gamma level of the positive control was ≥0.5 IU/mL; and indeterminate if both antigen-stimulated samples were <0.35 IU/mL and the level of the positive control was <0.5 IU/mL.

Treatment and Management of patients

The infectious-diseases physicians at the Singapore General Hospital independently evaluated all patients with a high clinical index of suspicion for TB. Those found to have associated systemic or pulmonary TB infection received anti-tuberculosis therapy (ATT), while uveitis patients with latent TB were advised about the risk-benefit ratio of ATT. Patients consenting to treatment received standard ATT according to the guidelines of the Centers for Disease Control and Prevention (isoniazid 5 mg/kg daily, rifampicin 450 to 600 mg daily, pyrazinamide 30 mg/kg daily, and ethambutol 15 mg/kg daily for 2 months, followed by 2 drugs for a 4-month continuation phase, for a total minimum of 6 months’ duration). In patients with posterior segment inflammation where ATT was not indicated, oral prednisolone was used at a starting dose of 1 mg/kg body weight, tapering slowly over the clinical course. Any anterior segment inflammation was treated with topical corticosteroids. The therapeutic response was monitored by 1 ophthalmologist (SPC), where a 2-step decrease in inflammation (Standardization of Uveitis Nomenclature working group activity score) was considered an improvement in clinical activity and a positive response to treatment.

Statistical Analysis

In the diagnosis of tubercular uveitis, there is currently no gold standard, ie, a diagnostic test with 100% sensitivity and specificity. Bayesian statistics are used to compute probability distributions (ie, posteriors) for parameters of interest in our statistical models by updating prior knowledge with new data, expressed formally by integrating the prior distribution and the likelihood function. We used Bayesian latent class models to evaluate the frequencies of true-positives derived from diagnostic tests as well as their sensitivities and specificities. Because the true numbers of patients with tuberculous uveitis are unknown, these were termed latent data . In this study, we performed an extended form of analysis using the Bayesian approach (previously explained in detail for the diagnosis of tuberculous uveitis in the absence of a gold standard), with some modifications. First, we collected data for another diagnostic test, the QuantiFERON-TB Gold In-Tube test, in addition to the tuberculin skin test and the T-SPOT.TB for a new group of patients with uveitis. Second, and most important, IGRAs T-SPOT.TB and QuantiFERON-TB Gold In-Tube are closely related tests, so our modified model takes into consideration their dependency in the estimation of sensitivities and specificities. Furthermore, in addition to using noninformative prior for the prevalence of tuberculous uveitis in our model, we have conducted a separate meta-analysis to estimate pooled prevalence, to incorporate informed prior knowledge as part of our sensitivity analysis. In summary, we performed the analysis in 4 major parts: (1) estimation of prevalence, sensitivity, specificity, and negative and positive predictive values; (2) analysis of the tuberculin skin test, T-SPOT.TB, and QuantiFERON-TB Gold In-Tube in combination; (3) Optimal choice of diagnostic test; and (4) Sensitivity analysis, where technical details are described in the Supplementary text . The Gibbs sampler algorithm, an iterative Markov-chain Monte Carlo technique, was used for estimations using the R and JAGS program. We used the JAGS software (v 3.3.0), running from R version 3.0.2 (R Development Core Team, 2013) to implement the Gibbs sampler, using specific marginal posterior densities. Convergence of estimation was checked and confirmed using the Gelman-Rubin convergence diagnostic.

The sensitivity (S) and specificity (C) of each test, as well as the prevalence of tuberculous uveitis ( π
π
), were the proposed model parameters to be estimated. Positive predictive value (PPV) and negative predictive value (NPV) can then be calculated with the 3 estimated parameters using the Bayes formula (refer to Supplementary text). Prior information for sensitivity and specificity of T-SPOT.TB, QuantiFERON-TB Gold In-Tube, and tuberculin skin test were obtained from a previous meta-analysis of tests for latent tuberculosis infection. To obtain relevant prior information from similar studies on the prevalence of tuberculous uveitis for patients with uveitis, we performed a meta-analysis of published literature using a similar Bayesian approach ( Table 1 , Supplementary Table 1 ) and updated the information in the model as part of our sensitivity analysis.

Patients’ demographics and clinical results

In this prospective study, we enrolled 120 patients (of whom 106 patients completed follow-up) with valid QuantiFERON-TB Gold In-Tube and T-SPOT.TB test results. The mean age of our patients was 48 ± 17 years, with an equal gender ratio (1:1, n = 53 males). The majority of the patients in our study were of Chinese ethnicity (65/106, 61.3%), followed by Indian (22/106, 20.8%) and Malay (7/106, 6.6%), reflecting the racial distribution in our Southeast Asian population. Of the 106 patients (152 eyes), 46 patients (43.4%) presented with bilateral uveitis. Uveitis was predominantly anterior (91/152, 59.9%), intermediate (4/152, 2.6%), or posterior (23/152, 15.1%); whereas 34 eyes (22.4%) presented with panuveitis. Suggestive clinical features of tubercular causes, such as granulomatous inflammation (38 eyes, 25.0%); extensive posterior synechiae (29 eyes, 19.1%); vasculitis (19 eyes, 12.5%); or single nodular or serpinginous choroiditis (1 eye, 0.7%) were observed in our study cohort. We found no significant differences in terms of age, gender, race or anatomic classification of uveitis when we compared the different QuantiFERON-TB Gold In-Tube, T-SPOT.TB, and tuberculin skin test results of the patients. We also did not have any definite cases of ocular TB infection, ie, culture-positive TB from ocular samples, in this study cohort. A single patient (0.9%) had Mycobacterium tuberculosis smear-positive sputum samples, and 1 patient (0.9%) had positive PCR results for Mycobacterium fortuitum based on the urine sample (tuberculin skin test positive, interferon-gamma release assay negative). The majority of patients (n = 90, 84.9%) had chest x-ray findings that were not suggestive of pulmonary TB infection. None of the study subjects were found to be immunocompromised, nor had they had BCG vaccinations within 10 years of the study enrolment.

Using all diagnostic tests results, we had more data to improve estimations for individual tests ( Table 1 ) and found that the QuantiFERON-TB Gold In-Tube was estimated to be more specific (QuantiFERON-TB Gold In-Tube: 0.995, 0.988 to 0.999) than T-SPOT.TB (0.905, 0.879 to 0.926); and slightly less sensitive (QuantiFERON-TB Gold In-Tube: 0.64, 0.60 to 0.69) compared to T-SPOT.TB (0.67, 0.60 to 0.74). Tuberculin skin test, as a reference, has sensitivity (0.69, 0.64 to 0.74) and specificity (0.74, 0.60 to 0.85). The correlations of sensitivities between QuantiFERON-TB Gold In-Tube and T-SPOT.TB were estimated to be 0.62 (0.43 to 0.77), but the specificities were not found to be correlated. The pooled prevalence of tuberculous uveitis based on our meta-analysis of similar studies was 1.62 (95% CrI: 0.88, 2.81) ( Supplementary Figure 1 ). Comparisons between estimations in our final model and the model updated with prior knowledge from our meta-analysis for prevalence of tuberculous uveitis are also shown in Table 1 . Sensitivity and specificity estimates were similar. Predictive values differ greatly, but this was not surprising because they are dependent on the prevalence of disease.

Table 2 shows the estimated number of true-positives corresponding to each of the 4 possible outcomes of the diagnostic tests. All who tested positive on QuantiFERON-TB Gold In-Tube and T-SPOT.TB in our study were estimated to be true-positive. However, among discordant results, QuantiFERON-TB Gold In-Tube was significantly more accurate than T-SPOT.TB (QuantiFERON-TB Gold In-Tube positive 98% vs T-SPOT.TB positive 76%, with a ratio of 1.28, 95%Crl: 1.11 to 1.72, ie, 95%Crl >1.0 was statistically significant). Similarly, both QuantiFERON-TB Gold In-Tube (0.99 95%Crl 0.98 to 1.00) and T-SPOT.TB (0.90 95%Crl 0.82 to 0.96) were more accurate than the tuberculin skin test in the discordant results. Using the estimated sensitivity and the specificity of each diagnostic test (tuberculin skin test, T-SPOT.TB, and QuantiFERON-TB Gold In-Tube), we then calculated the risk of each test as expressed as the function of probability of tuberculous uveitis ( Figure 1 ). Based on statistical decision theory, QuantiFERON-TB Gold In-Tube is the first-line test and should be performed in preference to T-SPOT.TB and the tuberculin skin test for diagnosis of tuberculous uveitis. Our sensitivity analysis confirmed that posterior distribution of parameters with mock data varied appreciably with the changing likelihood functions, suggesting that our results were data driven and were not overly influenced by priors. Furthermore, the optimal choice of diagnostic test is QuantiFERON-TB Gold In-Tube as seen in Figure 1 , regardless of the influence of choice of prior for prevalence of tuberculous uveitis.

Table 2

Counts of actual and estimated true-positive cases of tuberculous uveitis in our study cohort, compared to test results of QuantiFERON-TB Gold In-tube, tuberculin skin test and T-SPOT.TB

Study Data a				Estimated true-positive Counts b
T-SPOT.TB	QFT		Total	T-SPOT.TB	QFT		Total
	+	−			+	−
+	72	16	88	+	72 (1.00, 1.00-1.00)	12 (0.76, 0.57-0.88)	84
−	7	57	64	−	7 (0.98, 0.93-1.00)	27 (0.46, 0.30-0.67)	34
Total	79	73	152	Total	79	39	118

T-SPOT.TB	TST		Total	T-SPOT.TB	TST		Total
	−	+			+	−
+	57	27	84	+	56 (0.98, 0.96-0.99)	24 (0.90, 0.82-0.96)	80
−	18	43	61	−	14 (0.75, 0.56-0.89)	14 (0.32, 0.18-0.54)	28
Total	75	70	145	Total	70	38	108

QFT	TST		Total	QFT	TST		Total
	−	+			+	−
+	55	20	75	+	55 (1.00, 1.00-1.00)	20 (0.99, 0.98-1.00)	75
−	20	50	70	−	15 (0.75, 0.56-0.88)	16 (0.31, 0.19-0.53)	31
Total	75	70	145	Total	70	36	106

QFT = QuantiFERON-TB Gold In-tube; TST = tuberculin skin test; + = positive test result; − = negative test result.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Acute Retinal Arterial Ischemia: An Emergency Often Ignored
2. Importance of Birth Weight as a Risk Factor for Severe Retinopathy of Prematurity When Gestational Age Is 30 or More Weeks
3. Reporting Visual Acuities
4. Reply

Stay updated, free articles. Join our Telegram channel

Join