Rheumatologists: What if hydroxychloroquine eye tests show retinopathy [FEB 2023 Update]
Do not automatically stop hydroxychloroquine if your eye doctor tells you your patient has HCQ-retinopathy!
Heed my practical advice below
credit: By Wies6014 – Own work, CC BY-SA 4.0, httpscommons.wikimedia.orgwindex.phpcurid=77634880
FEB 2023 UPDATE: Good news from a 2023 study…
“our findings suggest that, under current screening recommendations and dosing patterns, loss of visual acuity from hydroxychloroquine retinopathy should be rare“
The hydroxychloroquine eye tests, SD-OCT and VF 10-2, can have false positives.
PEARL: Don’t stop your lupus patient’s most important medicine without doing the following:
– It is a good idea to have your staff call retinologists around your area and find out if they have this capability or not
Rationale is below:
(note that this statement is made with help from retinologists Dr. Jonathan Lyons and Dr. Reshma Katira in Silver Spring, MD and Alexandria, VA respectively. They specialize in HCQ-retinopathy. This section will appear in the 2nd edition of “The Lupus Encyclopedia,” but it is important to make it known widely to help our patients)
Mistakes in the medical literature
We rheumatologists, and eye doctors who do not specialize in antimalarial retinopathy (AMR) need to be careful about the subtleties of making this diagnosis and the potential flaws of the screening tests. The 2014 Browning Clinical Ophthalmology article (so often cited for its sensitivities and specificities), and their reproduction in reports (such as the 2020 Aduriz-Lorenzo article in Lupus) aimed at rheumatologists, need to be mindful of the weaknesses of these methods. HCQ is such an essential therapy for our SLE patients that stopping it due to incorrect diagnoses of HCQ retinopathy should be avoided as much as possible. We have seen this in our own practice where patients have been told they have HCQ retinopathy (based on VF 10-2 or SD-OCT testing), yet have an AMR expert identify an unrelated eye problem (using mfERG testing) as the cause for the test abnormalities. These patients have continued their HCQ with close follow-up by the retinologist, using mfERG technology.
There is a differential diagnosis for SD-OCT and VF 10-2 changes that can look like HCQ damage
It is essential to keep in mind the differential diagnosis for AMR based on these screening tests. Common causes of false-positive tests on SD-OCT and visual field testing include vitreomacular traction, retinal detachment, and age-related macular degeneration in our patients. Some less common causes of abnormalities on these tests include retinitis pigmentosa, infectious retinitis (syphilis, rubella), autoimmune (paraneoplastic) retinopathy, inherited retinal dystrophies (Stargardt disease, Bardet Biedl syndrome, enhanced S-cone syndrome, isolated bulls-eye maculopathy), pigmented paravenous chorioretinal atrophy, and traumatic retinopathy. Visual field testing can have false positives due to dry eye, glaucoma ,and cataracts.
The superiority of mfERG for accurately diagnosing HCQ-retinopathy
The Browning article states that the specificities of VF 10-2 is 92.5%, SD-OCT is 98.1%, while mfERG is 86.9%. These findings are misleading. First, in the Browning research study, HCQ was stopped in patients based upon a clinical judgment for AMR diagnosis. Second, VF 10-2 testing included both red target and white target testing. Red target testing is not as reliable as white target VF 10-2 testing. Also, a 20% error rate was allowed for VF testing, which is too high to be reliable. Third, the mfERG criteria used is not the current standard. Therefore, VF 10-2 and SD-OCT should not be used individually for diagnosis. However, they have excellent sensitivities for screening tests.
The most recent research shows that mfERG testing has better sensitivity and specificity for AMR than either VF 10-2 or SD-OCT testing. Using mfERG and SD-OCT has 100% sensitivity for identifying ARM, while the currently accepted use of VF 10-2 and SD-OCT has only an 86% sensitivity. A British study recently showed a 95% specificity and sensitivity for mfERG in the diagnosis as a singular test (far superior to VF 10-2 and SD-OCT). Also, mfERG testing is useful for monitoring patients who have other retina problems.
The highest standard of care for this issue is to use SD-OCT and VF 10-2 (white target) testing as screening tests (per the 2016 AAO recommendations). If abnormalities are present that can be seen in AMR, a referral should be given for mfERG testing by a retinologist experienced in AMR. We realize that mfERG testing is not available in all locations, so many patients will not have this luxury. In these cases, referral to a retinologist who does not have mfEFG technology would be the next best step before assuming a diagnosis of AMR. In the future, hopefully, mfERG will be more commonly available to allow even earlier identification of ARM (used as a screening test with SD-OCT) as well as allowing a more accurate diagnosis.
Aduriz-Lorenzo PM, Aduriz-Llaneza P, Araiz-Iribarren J, Khamashta MA. Current opinion on hydroxychloroquine-related retinal toxicity screening: where do we stand now? Lupus. 2020;29(7):671-675. doi:10.1177/0961203320919499
Broderick K, MD, Harrison Ngo, BS, Reshma Katira, MD [abstract]. Evaluation of SD-OCT Results in Screening Patients in Early and Later Stages of HCQ Maculopathy. American Academy of Ophth. AAO 2020 Vision Virtual Meeting, NOV 13-15.
Browning, DJ, Lee, C. Relative sensitivity and specificity of 10-2 visual fields, multifocal electroretinography, and spectral domain optical coherence tomography in detecting hydroxychloroquine and chloroquine retinopathy. Clin Ophthalmol 2014; 8: 1389–1399.
Browning DJ, Lee C. Scotoma analysis of 10-2 visual field testing with a red target in screening for hydroxychloroquine retinopathy. Clin Ophthalmol. 2015;9:1499-1509. Published 2015 Aug 20. doi:10.2147/OPTH.S87850
Lyons JS, Severns ML. Detection of early hydroxychloroquine retinal toxicity enhanced by ring ratio analysis of multifocal electroretinography. Am J Ophthalmol. 2007 May;143(5):801-809. doi: 10.1016/j.ajo.2006.12.042. Epub 2007 Mar 6. PMID: 17336914.