This letter refers to the article “Impact of the Revised American Academy of Ophthalmology Guidelines Regarding Hydroxychloroquine Screening in Actual Practice,” by Browning, and to the accompanying editorial by Marmor. The problem of defining cost-effective and reliable detection of retinal toxicity confronts every ophthalmologic practice. However, it is just as important to avoid the unconfirmed recommendation that treatment be halted and the patient be switched to a more toxic medication. In his study, Browning attempted to concentrate on the sensitivity of the screening process, but not on the loss of specificity incurred, by limiting the screening process to daily dosage and avoiding objective testing. In fact, by following his recommendations, sensitivity remains inadequate.

Since the initial report using the ring ratio method, the number of patients screened for toxicity with both visual fields and multifocal electroretinography (mfERG) in my clinic is now more than 600. Of the 89 patients with retinal toxicity confirmed by Humphrey visual field 10-2 and mfERG testing, more than 50% had taken a daily dose of less than 6.5 mg/kg body weight, whereas none had a cumulative hydroxychloroquine dose of less than 870 g. This confirms the importance of using cumulative dosage levels as an important screening determinant, but shows the limitation of using daily dose as the primary screening tool, as recommended by Browning. These patient variables provide information regarding when to test and on which patients to concentrate screening efforts. Adding visual field testing markedly improves sensitivity. Adding objective testing verifies the defect, which improves specificity.

As Marmor has noted, we have the tools to avoid premature cessation of treatment by the addition of objective testing methods, which will confirm the presence or absence of toxicity. My recommendation for screening, which also seeks to maximize cost effectiveness, follows the revised American Academy of Ophthalmology guidelines, but with some variation from Marmor’s.

• 1.
  

  Baseline screening should include both white stimulus 10-2 visual fields using pattern deviation, mfERG, and spectral-domain ocular coherence tomography (SD OCT) where possible so that verifiable changes can be detected during treatment. There is sufficient variability in all these tests, particularly in visual fields, to warrant the expense.

  
  
• 2.
  

  Although toxicity rarely develops before 5 years of treatment (approximately 750 g cumulative dose), except with high daily doses (600 mg/day), low body weight, or poor drug elimination, patients with autoimmune disease usually are seen yearly. This makes it useful and cost effective to perform a visual field test at 1 or 2 of these visits to confirm reliability for each patient.

  
  
• 3.
  

  After 5 years, annual visual field testing should be carried out with referral for both mfERG testing and SD OCT testing if the visual field is unreliable or shows multiple loci with at least −4-dB loss on pattern deviation. Although Marmor’s recommendation to emphasize SD OCT in place of mfERG may turn out to be effective, quantitative criteria have not been defined for the SD OCT, making for a less precise determination of early toxicity.