Hydroxychloroquine screening

Almony et al report in this issue of the BJO (p 569) the use of a threshold Amsler grid (TAG) as a screening tool for asymptomatic patients taking hydroxychloroquine (HCQ). They studied 56 patients taking HCQ and 12 controls. Patients were tested with a “white on black” Amsler grid (AG), a “red on black” AG (RAG), and the threshold AG (TAG). TAG uses cross polarising filters to reduce the perceived luminance of the grid. Scotomas were detected in two patients (3.6%) with the standard AG and five patients (8.9%) with RAG, but 25 (45%) patients with TAG. The TAG testing detected the two positive AG screens and the five positive RAG screens. The authors concluded that TAG has increased sensitivity to the detection of subtle scotomas in patients taking HCQ. Unfortunately, because there is no “gold standard” for HCQ retinopathy in asymptomatic or presymptomatic patients without visible retinopathy the specificity of the TAG results is unknown.

If the scotomas seen on TAG were also detected reproducibly in the same location using another central field test (for example, Humphrey 10-2) this would provide evidence for the specificity of the TAG findings. I would encourage the authors to continue to follow their cohort of HCQ treated patients and perhaps even test the patients with the abnormal TAG findings again with an automated (Humphrey) 10-2 strategy or even a multifocal electroretinogram (MERG). Although the sensitivity and specificity of MERG in HCQ toxicity continues to be explored it may be that objective electrophysiological testing might be superior to subjective tests of visual function like the AG.¹

One of the patients in this study (case 63) was only taking HCQ for 1 month and yet had large bilateral central scotomas. It is unlikely that this represented HCQ retinopathy and this patient did not have a baseline eye examination. This case demonstrates the limitations of not specifically excluding from the study any patients who did not have complete ophthalmological examinations before starting HCQ. It may be that false positive screens may be a significant limiting factor for the TAG. Pluennke and Blomquist reported that 6–11% of HCQ and control patients tested with RAG had a false positive result.² The false positive rate for the TAG is not known from the study by Almony et al.

Although there have been many guidelines in the United States, Canada, and the United Kingdom for screening examinations for patients taking HCQ, the cost effectiveness and diagnostic yield of these recommendations have not been evaluated in a rigorous and critical evidence based manner.^3–5 The risk of HCQ toxicity is exceedingly rare for low risk patients and over one million patients up to 2002 have been treated with HCQ with only 20 cases of toxicity at the “subthreshold” dose of <6.5 mg/kg/day. All of these 20 cases had taken the drug for more than 5 years. In addition, there still remains controversy as to the timing and content of screening examinations for these patients. The American Academy of Ophthalmology (AAO) has provided a screening strategy composed of three parts: (1) informed consent obtained by the prescribing primary physician with explicit written documentation in the medical record; (2) detection and minimisation of toxicity rather than prevention itself; (3) definition of high and low risk patients (see table 1); and (4) stratification of screening based upon risk factors. If a baseline eye examination is normal and the patient is taking a low dose (<6.5 mg/kg/day) of HCQ then the recommended screening interval follows the AAO screening recommendations for regular eye examinations in the general population. Annual screening was recommended for patients with higher or unknown dose or duration (>5 years) of HCQ therapy.³ Almony et al recorded several of the risk factors proposed by the AAO (see table 1) for HCQ retinopathy including weight adjusted doses, duration of HCQ therapy, and the age of the patients. They did not however include data on renal or hepatic insufficiency and no patients had documented other macular pathology.

Table 1.

 Criteria of low and higher risk for developing retinopathy (modified from Marmor et al³)

Criteria	Low risk	Higher risk
Dosage	<6.5 mg/kg	>6.5 mg/kg
Duration	<5 years	>5 years
Habitus	Lean or average fat	High fat level
Renal/liver disease	Absent	Present
Concomitant retinal disease	Absent	Present
Age	<60 years	>60 years

The specific recommendation of the American Academy of Ophthalmology is for a baseline examination (listed in table 2) for all patients starting HCQ treatment. Unfortunately, there is no “gold standard” for identification of toxicity before the development of the ophthalmoscopic changes (that is, pigmentary changes and “bull’s eye maculopathy”). Despite the recommendation of the AAO, it is not clear that a baseline examination is cost effective given the large numbers of patients on HCQ and the relatively low incidence of retinopathy. In the United Kingdom, the Royal College of Ophthalmologists, the British Association of Dermatologists, and the British Society for Rheumatology recommend baseline assessment of renal and liver function, inquiry about visual symptoms, and recording of near visual acuity with inquiry about visual symptoms at each visit and measurement of visual acuity annually. Buckley et al in the April 2004 guidelines from the United Kingdom for screening suggest that a baseline eye examination and regular ophthalmological screening may not be required in patients taking low (<6.5 mg/kg) doses of HCQ.⁵ These guidelines do recommend referral to an ophthalmologist for patients with ocular disease at baseline or for those who develop visual symptoms on treatment. Interestingly the Amsler grid is not included in the annual evaluation recommended for the rheumatology and dermatology clinics but is included in the assessment by ophthalmology.⁵ Publication of these recommendations and national guidelines may not ensure compliance however. Samanta et al reported wide variation among consultant rheumatologists in the United Kingdom and nearly half of surveyed respondents did not assess either baseline visual symptoms or visual acuity.⁶

Table 2.

 Baseline examination for all patients treated with hydroxychloroquine recommended by the American Academy of Ophthalmology (AAO)

• Complete ophthalmological exam including dilated fundus exam

• Baseline visual field testing (eg, Amsler or Humphrey 10–2)

• Optional colour vision testing (eg, screen for pretreatment colour blindness)

• Optional fundus photography (eg, if pretreatment macular pigmentary changes)

• Optional specialised tests (eg, fluorescein angiography or multifocal electroretinogram)

In summary, despite the limitations of the study by Almony et al, TAG may be a more sensitive means for detecting subtle scotomas in patients taking HCQ. The specificity of the TAG however remains to be defined. High risk and low risk features of the individual patient should determine the timing of screening for HCQ retinopathy. Appropriate informed consent, adequate documentation in the medical record, and an appropriate baseline assessment by the prescribing physician are important for medicolegal as well as medical reasons. Because the incidence of HCQ toxicity is extremely low at doses <6.5 mg/kg in asymptomatic and otherwise visually healthy patients, the need and cost effectiveness of baseline and more frequent screening examinations by an ophthalmologist remains debatable. The rationale for examining a patient within the first year of HCQ treatment is to establish a baseline and to document any pretreatment eye disease. The TAG however may be a more sensitive tool for detecting patients in the non-ophthalmology clinic setting who may need a full ophthalmology examination. More frequent screening should be performed in patients taking HCQ with high risk characteristics.