Retinopathy With Long-Term HCQ Use: Large Study IDs Risk Factors

Retinal damage from taking hydroxychloroquine (HCQ) for more than 5 years occurred more frequently among people with certain characteristics, records from nearly 5,000 patients in the Kaiser Permanente system indicated.

Significant risk factors included the average daily dose and cumulative doses over time, as one would expect — but there were others, too, including female sex, use of tamoxifen, older age, and presence of moderate-to-severe chronic kidney disease (CKD), according to April M. Jorge, MD, of Massachusetts General Hospital and Harvard Medical School in Boston, and colleagues.

And at least as important, the study showed just how low the absolute risk is. With 4,677 patients tracked after starting HCQ, just 125 developed retinopathy during 5 to 15 years of follow-up, the researchers reported in JAMA Network Open.

HCQ is the chief maintenance treatment for systemic lupus erythematosus and is also commonly prescribed for rheumatoid arthritis (RA) and other rheumatologic conditions and connective-tissue disorders. Retinal damage with long-term use has been a recognized side effect for decades, but beyond the dosing level and duration, “[t]he potential association of other risk factors with hydroxychloroquine retinopathy is less clear,” Jorge and colleagues explained.

The Kaiser Permanente system in Northern California, with its many members and extensive records, including prescription fills, looked to be a good source of data for this topic. Jorge and colleagues pulled records for all patients prescribed HCQ from July 1997 to December 2014, with follow-up through December 2020. To be included, patients needed to have at least one HCQ prescription written more than 5 years after starting the drug (it was not required that patients had to be continuously supplied with HCQ, however) and to maintain enrollment in the Kaiser system for at least 5 years.

Mean age when starting HCQ was 52, and 83% were women. More than half had RA or other inflammatory arthritis as the primary indication; lupus and other connective-tissue disorders were the reason in some 34% of patients, and “dermatologic conditions or other” accounted for 8%. Just under 60% of patients were white, while the rest were about evenly divided among Black, Asian, and Hispanic people.

At HCQ initiation, the mean dose was 4.4 mg/kg, with 61% given doses greater than 5 mg/kg. At year 5, doses had been reduced in many patients: the mean was 3.2 mg/kg and only 34% were on doses higher than 5 mg/kg. The cumulative dose at year 5 averaged 455 g (almost exactly 1 pound).

At least numerically, the strongest risk factor for retinopathy turned out not to be the daily or cumulative doses, but patient age: after adjusting for covariates, patients 65 or older faced a hazard ratio of 5.68 relative to those younger than 45 (95% CI 2.99-10.79). But a daily dose greater than 6 mg/kg was a close second (HR 5.30, 95% CI 3.46-8.10).

Cumulative HCQ dose affected retinopathy risk to the tune of a 1.64-fold increase for each 100 g (95% CI 1.44-1.87).

Other significant risk factors included the following (adjusted for covariates including race and daily and cumulative HCQ doses):

• Female sex: HR 3.83 (95% CI 1.86-7.89)
• Tamoxifen use: HR 3.43 (95% CI 1.08-10.89)
• CKD grade 3 or more: HR 1.95 (95% CI 1.25-3.04)

Interestingly, patients with RA appeared to face less risk for retinal damage than those with lupus or other connective-tissue disorders, with a hazard ratio of 0.62 (95% CI 0.41-0.94).

Other findings involved different types of retinopathy. Of the 125 patients developing retinopathy, 102 showed a parafoveal pattern while the remaining 23 developed the pericentral type. The Kaiser data indicated that, after adjusting for HCQ dose, Black patients were markedly more likely than white patients to show pericentral damage (HR 5.51, 95% CI 1.22-24.97); and even more so with Asian patients (HR 15.02, 95% CI 4.82-46.87). Hispanic patients had a similar trend but it fell short of statistical significance.

Meanwhile, women were more likely than men to show parafoveal retinopathy (HR 5.50, 95% CI 2.23-13.59), and this pattern was also more common in patients with grade ≥3 CKD (HR 1.92, 95% CI 1.18-3.12) relative to those with no or less severe CKD.

Understanding these novel risk factors for HCQ retinopathy “should influence hydroxychloroquine dosing and monitoring for this complication,” Jorge and colleagues wrote. Noting the increased rates of pericentral retinopathy among non-white patients, the investigators also urged that “[b]ecause recognition of the pericentral pattern requires wider examination of the fundus than standard optical coherence tomography recordings or central 10-2 visual fields provide, specialists who interpret hydroxychloroquine retinopathy screening studies must be aware of this pattern to ensure timely detection.”

Limitations to the research included its reliance on administrative data and the small number of patients developing retinopathy, which meant that the study may have been underpowered for some analyses. Also, the authors noted that retinopathy detection “must be based on active screening,” and thus it’s possible that diagnosis might have come long after actual onset in some cases.

Please enable JavaScript to view the

comments powered by Disqus.