Diagnostic tests are critically important to define both infectious and noninfectious etiologies in uveitis.
As young ophthalmologists, we are all aware of the fact that performing diagnostic tests that are possibly unnecessary may create false positive results and incorrect diagnoses and lead to wrong treatment and loss of patient compliance.
On the other hand, we may recall at least one uveitis patient encounter where we struggled to pick the right tests, eventually taking a shotgun approach in order to cover all possible diagnoses, all of which really did not help. Don’t worry. You are not alone!
Even experienced ophthalmologists and uveitis specialists may have difficulties in making the smartest choices. How do you improve? There is no magic formula, but brushing up on some of the basic concepts always helps.
Pathway to Diagnosis
The classic method to diagnose uveitic diseases requires pattern recognition, a combination of detailed patient history and rigorous ophthalmic examination to recognize patterns that indicate particular conditions. When there is no pattern, establish a differential diagnosis of possible diagnoses (“probabilistic method”) through indirect or intuitive approaches that are guided by clinical experience.
Both pattern recognition and probabilistic methods should be supported with diagnostic tests that include serologic tests, imaging modalities and tissue sampling.
Understanding Bayesian statistics and Implications
Bayes’ theorem implies that the positive predictive value of each diagnostic test is related to the prevalence of the particular disease in the general population.
Below is an example of a disease, which has a 1:1,000 prevalence rate in the population and has a diagnostic test with 99% sensitivity and 95% specificity. Any patient with no previously known ocular or systemic finding would therefore have 0.1% pretest probability of having the disease. The positive predictive value (PPV) may be calculated as:
PPV = (0.99×0.001)/(0.95×0.001+0.05×0.999)=0.019
This indicates that only 1.9% of cases that were tested positive might actually have that disease. According to Bayes’ theorem, narrowing the general prevalence by performing the diagnostic test in cases with specific clinical findings would be the only way to increase the post-test probability.
Ordering Laboratory Tests
Infectious Uveitis
Tuberculosis. Hypothetically, if all patients were screened for tuberculosis with purified protein derivatives (PPD) or detection of IFN-γ expression following antigen stimulation (Quantiferon®) tests, PPVs would be less than 10%. PPVs of these tests would increase (up to 96%), only when performed at an endemic area or for a patient with clinical findings suggestive of tuberculosis such as serpiginous lesions.
Syphilis. Non-treponemal venereal disease research laboratories (VDRL) and rapid plasma reagin (RPR) are used to screen active syphilitic disease, whereas treponemal (FTA-ABS, MHA-TP, TPHA, EIA and syphilis IgG) tests recognize T. pallidum specific antibodies and demonstrate previous syphilitic exposure.
Thirty percent of RPR and VDRL tests may give false negative results for latent disease and neurosyphilis. In tertiary referral centers, where the prevalence is higher due to selection bias, initially a specific test (Syphilis IgG or FTA-ABS) is recommended in order to avoid false negative results.
Non-infectious Uveitis
Human Leukocyte Antigen B27 (HLA-B27). With 5% prevalence in a normal population, the expressivity of HLA-B27 increases from 50 to 80% in cases with unilateral acute anterior uveitis (AAU). PPV of the test varies depending on the anatomic location of uveitis, with anterior uveitis being highest.
Antinuclear antibodies (ANA). With a positive predictive value of 1%, it has very limited use in diagnosis of uveitic syndromes, which only encompasses juvenile inflammatory arthritis, scleritis, peripheral ulcerative keratitis and vasculitis.
Antineutrophil cytoplasmic antibodies (ANCA). These are exclusively beneficial for differential diagnosis of necrotizing scleritis, peripheral ulcerative keratitis and retinal vasculitis.
Angiotensin converting enzyme (ACE). ACE has a moderate sensitivity and specificity: an increase in ACE level has a PPV around 47% in diagnosing sarcoidosis-associated uveitis, which is thought to increase up to 72% when combined with increased serum lysozyme levels.
Urinary β2-microglobulin is a useful marker to screen for tubulointerstitial nephritis and uveitis syndrome (TINU) in bilateral AAU patients. It is positive in over 80% to 90% of TINU cases.
Quick Guide for Uveitis Testing
Anterior Uveitis
Tests to include: HLA-B27, syphilis IgG or FTA-ABS (if not available then VDRL), chest radiography, serum ACE and lysozyme and Quantiferon® (if not available, then PPD) tests. Urinary β2-microglobulin is also helpful to assess for TINU in the pediatric population.
Tests to omit: RF, ANA and ANCA are unlikely to be related to uveitis in the adult population unless specific findings are present, including PUK and scleritis.
Intermediate Uveitis
Tests to include: Syphilis IgG or FTA-ABS (if not available, then VDRL) , chest radiography, serum ACE and lysozyme, Lyme serology and Quantiferon® (if not available, then PPD) tests.
Tests to omit: HLA-B27, RF, ANCA and ANA.
Posterior/Panuveitis
Tests to include: Differential diagnosis of infectious etiologies are crucial. Syphilis IgG or FTA-ABS (if not available, then VDRL), chest radiography, serum ACE and lysozyme, Lyme serology and Quantiferon® (if not available, then PPD) are needed. In addition, other bacterial (cat-scratch disease), viral (HSV, VZV, CMV) and parasitic (toxoplasmosis, toxocariasis, oncocercosis) infections should be investigated. Hematuria and proteinuria are assessed for systemic involvement in retinal vasculitis, scleritis and PUK.
Tests to omit: RF, ANA, ANCA and HLA-B27 tests if no vasculitis or related systemic involvement.
Now you have an evidence-based approach to approaching uveitis patients. Please keep in mind that all of these laboratory workup is even more valuable with ancillary imaging.
Note: Some aspects of this article may vary geographically, and need to be adjusted based on local prevalence.
 |
About the author: Murat Hasanreisoglu, MD, FEBO, is a clinician scientist at Koç University School of Medicine and Koç University Research Center for Translational Medicine in Istanbul, Turkey. He is a committee member of the European Society of Ophthalmology – Young Ophthalmologists (SOE-YO). His main areas of clinical interest are uveitis and medical and surgical retinal diseases. |