Abstract
Hydroxychloroquine (HCQ) is a pharmacological agent in the management of various rheumatological disorders, including systemic lupus erythematosus and rheumatoid arthritis. Its therapeutic effects are attributed to its immunomodulatory and antimicrobial properties, and it is widely used due to its efficacy, relatively low cost, and generally favorable safety profile. Despite its well-established clinical utility, chronic use of HCQ is associated with serious ocular side effects. This review covers the history, pharmacokinetics, and ocular safety profile of HCQ, with a focus on the risk, diagnosis, and screening for ocular toxicity. Retinopathy, the most significant adverse ocular side effect, may lead to irreversible vision loss if not promptly detected.
Hydroxychloroquine (HCQ) is derived from quinine, quinacrine, and chloroquine. The drug quinine is itself made from the bark of cinchona trees, and quinine and its derivatives have found many medicinal applications. As early as 1630, quinine was used as an antipyretic. Quinacrine was then synthesized from quinine and in the 1930s found use as an antimalarial.1 In the Second World War, antimalarial drugs gained prominence and became industrialized on a large scale because malaria was a leading cause of disease and morbidity among soldiers, especially those deployed in the South Pacific.2 The widespread use of antimalarials serendipitously led to the discovery of the value of quinine-derived antimalarials in rheumatological disease. An incidental finding of improvement in cutaneous lupus erythematosus and inflammatory arthritis while the Second World War soldiers were taking quinacrine for malaria led to the use of quinacrine to treat systemic lupus erythematosus (SLE).3,4 However, numerous side effects from quinacrine, including yellowing of the skin, were unsatisfactory. This led to further drug development by the US Army in 1943 to produce chloroquine. In the 1940s, there was concern regarding the emergence of chloroquine-resistant malaria strains, resulting in chloroquine being used more broadly to treat other infectious diseases.5
HCQ was first synthesized by American chemists Alexander Surry and Henry Hammer in 1946. They derived HCQ from chloroquine, creating a new drug that was cost-effective, safe, and better tolerated.1,6 HCQ is now the most prescribed antimalarial used to treat autoimmune disease. HCQ is approved by the US Food and Drug Administration to treat malaria, discoid lupus erythematosus, SLE, and rheumatoid arthritis. There are other uses that are off-label but find common clinical application, including cutaneous dermatomyositis, extraglandular manifestations of Sjögren disease, sarcoidosis, antiphospholipid antibody syndrome, porphyria cutanea tarda, and Q fever.7-9 Its versatility as a drug may be attributed to its multiple proposed mechanisms of action and its unique and highly variable pharmacokinetic profile.
Pharmacokinetics
HCQ is primarily absorbed in the gastrointestinal tract, with oral bioavailability of approximately 70%, although there is wide variability between patients on the extent of absorption.10-12 Drug metabolism is mediated through the hepatic cytochrome P450 enzyme pathway, where multiple active metabolites are produced and up to a quarter of unchanged drug is then renally eliminated.12,13 Thus, liver and renal impairment can significantly affect drug metabolism. Although the liver participates in metabolism, there is no clear association between liver disease and ocular toxicity. However, the American Academy of Ophthalmology (AAO) does recognize renal disease as a major risk factor for retinal toxicity.14 HCQ accumulates in several tissues and it has an affinity for areas of high melanin content, making the retinal pigment epithelium an area of concern.15 There can be up to 10-fold variability in blood levels of HCQ between patients, due to differences in absorption, metabolism, excretion, and other less well-understood physiologic processes.16,17 Thus, HCQ serum levels are an unreliable indicator of either medical efficacy or toxicity.18-21 This pharmacokinetic variability, along with an unpredictable and gradual accumulation within the retinal pigment epithelium, may lead to differences in how each individual patient responds to HCQ.
Mechanism of action
The immunomodulatory effects of HCQ include inhibition of molecular pathways and cellular function. On a molecular level, HCQ can interfere with lysosome functioning and autophagy. This can then affect membrane stability, inhibit major histocompatibility complex class II expression, antigen presentation, and immune activation. There is also disruption of Toll-like receptor signaling, leading to decreased stimulation to produce type 1 interferons and proinflammatory cytokines that are overactive in disease like SLE. Thus, HCQ reduces inappropriate activation of the immune system. The exact mechanism of action is likely dependent on the specific disease and the affected organs and tissues, making disease context an important factor.11.
Calculated risk of ocular toxicity
The traditional recommendation of dosing of HCQ at < 6.5 mg/kg/day was based on older studies and used the patient’s ideal body weight. However, this number did not take into consideration the duration of treatment.22 The current dosing recommendations to minimize risk of ocular toxicity stem mainly from the population-based study of Melles and Marmor, where daily dosing and duration of treatment were taken into account to provide risk assessment.23 In that study, dosing by real body weight predicted risk better than ideal body weight and was used for all calculations. The conventional belief that used ideal body weight was based on the false assumption that these drugs were not retained in fat. However, laboratory studies show that it is indeed stored in melanotic tissue, liver, and kidney, with lower concentrations in muscle, fat, and other organs.24,25 Thus, very thin patients were at increased risk of retinal toxicity when doses were calculated with ideal body weight because they were overdosed. Using actual weight is more appropriate in treating patients with a broad range of body habitus.
In the study by Melles and Marmor,23 the average dose of HCQ was approximately 5.0 mg/kg of real weight, due to varying compliance and body habitus. As a result, 5.0 mg/kg/day of HCQ based on actual, not ideal, patient weight corresponds to current prescribing practices and should provide effective immunomodulation for most patients.
It should be noted that HCQ is sometimes prescribed in a manner where it is not taken every day. In these situations, an individual’s dose on medication days may exceed 5.0 mg/kg while remaining under the 5.0-mg/kg/day safe dosing threshold when averaged over an entire week. As an example, a dose of 400 mg Monday through Friday (not taken on Saturday or Sunday) in a 60-kg individual means that 6.67 mg/kg is taken Monday through Friday, but is 4.76 mg/kg/day when the daily average is calculated.
The overall prevalence of HCQ retinopathy is 7.5%. However, this figure requires further interpretation due to significant variability based on daily dosage and duration of use. The odds ratio for developing retinopathy was 5.67 (95% CI 4.14-7.79) for individuals taking > 5.0 mg/kg/day and 3.22 (95% CI 2.20-4.70) for those using the medication for over 10 years. For patients dosed at < 5.0 mg/kg/day, the risk is < 1% in the first 5 years of therapy and < 2% at 10 years. The risk rises sharply to approximately 20% after 20 years.23 This is similar to more recent data presented by the Systemic Lupus International Collaborating Clinics where their cumulative risk was < 1% within the first 10 years of continuous HCQ use.26
Looking at incremental annual risk of toxicity in a patient who does not show signs of toxicity, it is < 1% during the first 10 years of therapy if use is < 5 mg/kg/day and increases to only 4% after 20 years. The risk is therefore small with appropriately low doses, but it is not clear if there is a safe dose for longer durations of use. However, the dose of 5.0-mg/kg/day real body weight appears to have an acceptable level of ocular toxicity risk for being screened annually to detect change early.23
Risk factors for ocular toxicity
Daily dose, duration of use, and lifetime cumulative dose. Excessive daily dose by actual weight is the most important factor affecting risk for development of ocular HCQ toxicity. Doses > 5.0 mg/kg daily can increase population risk and annual incremental risk. Further, extremely high doses (eg, 800-1000 mg daily or up to 20 mg/kg) demonstrated 25-40% incidence of retinopathy within 1-2 years.27,28 The duration of use is important in that even with a recommended dose of < 5.0 mg/kg/day, there is increased risk of retinal toxicity after prolonged use. This is closely related to research that shows that a lifetime cumulative dose of > 1000 g is associated with an elevated retinal toxicity risk. Useful online calculators are available to aid the clinician in calculating a patient’s lifetime cumulative dose (such as https://www.eyedock.com/calcs/plaquenil-calcs).
Kidney disease. Decreased renal function leads to elevated serum concentrations of HCQ.18 Subnormal glomerular filtration rate with a drop of 50% in renal function leads to an approximate doubling of the risk of retinopathy.23 Many patients with SLE may also have decreased renal function, so careful note must be taken of effective glomerular filtrate rate and subsequently, adjustments in dosing and screening frequency may need to be made.
Tamoxifen use. Tamoxifen is a selective estrogen receptor modulator used in the treatment of estrogen receptor–positive breast cancer. Its concomitant use with HCQ is associated with an increase of approximately 5 times the risk of retinal toxicity.23 This finding was not consistent when the breast cancer treatment was anastrozole (an aromatase inhibitor that decreases endogenous estrogen production). Tamoxifen use itself can cause a retinopathy that typically manifests as crystalline deposits or macular edema.29 These findings are distinct from HCQ-related retinal toxicity. Patients with HCQ-related retinal toxicity did not exhibit any clinical features of concurrent tamoxifen retinopathy.
Retinal and macular disease. There is no data to confirm if preexisting retinal or macular disease is a significant risk factor for HCQ-related retinal toxicity. However, macular disease can be a significant confounding variable when trying to interpret screening tests used for retinal toxicity monitoring. It also seems intuitive not to add a potentially toxic agent to an already vulnerable retina.
Nonsignificant risk factors. Although HCQ is metabolized by the liver, there is no clear relationship between liver disease and retinal toxicity.23 Age by itself does not seem to be a risk factor of ocular toxicity. However, increasing age can be associated with increased duration of HCQ use, decreased renal function, concurrent retinal disease, and also higher accumulative dose. Genetic factors, such as abnormalities in the ABCA4 gene (involved in transporting and clearing potentially toxic substances in the retina) and cytochrome P450 (involved in HCQ metabolism in the liver) have not been conclusively related to ocular toxicity.21,30-32
Clinical features of ocular toxicity
Within the eye, HCQ can affect a variety of structures, including the cornea, ciliary body, and retina. Corneal toxicity presents as intra-epithelial deposits in the cornea, but these rarely affect vision.33 The ciliary body is involved in accommodation and near vision; however, this is rarely affected. As retinal pathology is most significant in vision loss, subsequent discussion will focus on retinopathy.
The textbook clinical finding of ocular HCQ toxicity is a “bull’s eye retinopathy” that affects the macula, also called a maculopathy. There is depigmentation of the retinal pigment epithelium in the parafoveal region, leaving a preserved central island in the fovea. However, with appropriate dosing and proper screening, this finding should ideally no longer be seen since bull’s eye retinopathy is a late-stage finding. Any retinal toxicity should be discovered long before this occurs. In the early stages of toxicity, patients may be asymptomatic, but with progression they may complain of decreased vision, color vision changes, difficulty reading, paracentral scotomas, and metamorphopsia.
HCQ retinal toxicity is usually irreversible. Even with cessation of the drug, there can still be progressive cellular damage for a time before stability is reached. Thus, screening and early detection are of paramount importance. Early detection and recognition leading to drug cessation can limit subsequent damage.
Diagnosis and screening for ocular toxicity
Screening schedule. Four major medical societies (American College of Rheumatology, American Academy of Dermatology, Rheumatologic Society, and AAO) developed a joint statement outlining principles and guidelines on cooperative screening efforts. Further, there has been agreement on these screening guidelines by consensus on the use of HCQ in patients with SLE.34
Baseline eye examination is recommended before treatment or within the first year of initiating HCQ therapy. This visit identifies and documents any risk factors and existing ocular conditions, and establishes clinical baseline. Follow-up can then be determined based on this exam. The risk of ocular toxicity is low within the first 5 years of treatment if the drug dosing is low and there is an absence of other risk factors.23
After baseline examination, annual screening can be deferred in a reliable patient until after 5 years of drug exposure. Thereafter, annual screening is sufficient in most cases because toxicity develops slowly. With suspicious findings, repeat testing at a shorter interval can be performed to confirm reproducibility.35 However, if a patient has significant risk factors as listed above or concurrent ocular disease that may make interpretation of testing difficult, such as preexisting macular pathology from other retinal conditions, then more frequent screening should be considered. In addition, if a patient has significant weight loss (without dose adjustment) or change in risk factors (such as kidney disease), more frequent eye examinations may be warranted.
Screening tests. Every eye exam begins with baseline visual acuity. In addition, the AAO recommends the use of both functional testing, such as automated visual fields, and structural testing, such as spectral domain optical coherence tomography (SD-OCT), in routine HCQ screening.14
Automated visual field testing is a subjective, functional test but can be very sensitive in reliable patients. The visual field test in most patients should focus on the central 10° (10-2 protocol). This has high resolution for macular testing and should be used for non-Asian patients to detect parafoveal damage.36 However, in patients of Asian descent with ocular toxicity a pericentral pattern is more prevalent. Thus, the protocol for visual field testing needs to be adjusted. In Asian patients, a wider visual field test should be used, namely, the 24-2 protocol which focuses on the central 24°. Should any central pathology be identified in the wider 24-2 protocol, testing should be repeated using the 10-2 protocol. Visual fields are subjective tests and often vary from test to test, making interpretation difficult. If there is diagnostic uncertainty in the visual field results, repeat testing should be pursued in combination with other investigations.
SD-OCT is a ubiquitous noninvasive imaging system that provides objective structural data of the macula. These imaging systems provide unparalleled details of retinal structure with near-cellular resolution (< 10 μm). Damage from HCQ will show localized thinning of the photoreceptor layers in the parafoveal region in non-Asian eyes or near the superior and inferior vascular arcades in Asian eyes. These areas of damage may correspond to visual field changes. Even though SD-OCT may not be as sensitive as visual field testing or electroretinography (discussed below), the advantage is that SD-OCT is an objective, structural test and indicates definitive result when characteristic regional thinning of retinal toxicity is seen.
Additional testing may include multifocal electroretinography (mfERG) and fundus autofluorescence (FAF). mfERG is an objective, functional test. The full-field ERG measures the electrical activity of the entire retina in response to a brief light stimulus. The mfERG measures more localized responses within the central 30° to produce a topographic map of central retinal cone photoreceptor function, thus enabling it to detect macular dysfunction. This test can be as sensitive as visual field testing.37 A limitation of ERG testing is that it requires specialized equipment and experienced technicians to interpret the data. It is usually available in large clinical or academic centers.
FAF is an objective and functional test. It produces a noninvasive image of the retina that provides a density map of lipofuscin, the predominant ocular fluorophore in the retinal pigment epithelium. Early photoreceptor damage due to HCQ retinal toxicity on FAF imaging can be seen as areas of increased autofluorescence due to retinal pigment epithelium dysfunction and an accumulation of lipofuscin. This finding may precede retinal thinning seen on SD-OCT38,39 (Figure 1).
Clinical findings of a 53-year-old patient with advanced hydroxychloroquine toxicity (left panels), compared to a patient of similar age with no ocular pathology (right panels). (A) Fundus photographs with bull’s eye pigmentary changes. (B) Fundus autofluorescence with abnormal macular hyper- and hypoautofluorescence in areas of photoreceptor loss. (C) Optical coherence tomography with parafoveal loss of the outer retinal layers (arrows). (D) 10-2 Humphrey visual field with severe constriction. DS: diopter sphere; MD: mean deviation; N/A: not applicable; PSD: pattern standard deviation; SITA: Swedish Interactive Thresholding Algorithm.
Even with visual field, OCT, mfERG, and FAF testing, there is still room to improve the early identification of HCQ retinal toxicity. Limitations to the established testing methods mentioned above include patient subjectivity, difficulty in the interpretation of subtle testing abnormalities, and accessibility of more specialized testing modalities. Early detection of toxicity is critical to avoid permanent visual impairment. Some novel retinal imaging techniques that show promise include OCT angiography, multicolor imaging, adaptive optics, and retromode imaging. These techniques are still under investigation and validation.40
Automated machine learning has been explored as a potential tool to identify patients with HCQ retinal toxicity. There is promising early work in this area. SD-OCT findings in HCQ retinal toxicity can be subtle and difficult to recognize, even for expert physicians. With the collection of more quantitative metrics, such as ellipsoid zone mapping and outer retinal layer parameters, the goal is to develop automated classification models. Validation of these evolving machine learning models is ongoing.41,42
Once HCQ retinal toxicity is suspected, strong consideration to stop this medication should be addressed with the prescribing physician. The conditions treated and stabilized with HCQ can be serious and carry significant morbidity and mortality; however, the retinopathy is irreversible, and vision loss can progress even with medication cessation if the retinopathy is severe. Early detection of retinopathy can minimize this loss and potential progression.
Conclusion
HCQ is a well-tolerated and cost-effective drug used to successfully manage many rheumatologic conditions. There is a risk of ocular toxicity, with visually significant retinopathy being the primary concern. Major risk factors for retinal toxicity include daily dose > 5 mg/kg, prolonged duration of treatment (< 1% up to 5 years if dosage is under threshold and < 2% up to 10 years, but rising to 20% after 20 years), cumulative lifetime dose of > 1000 g, concurrent tamoxifen use, and renal disease.
Screening should occur within 1 year of starting HCQ and then annually after 5 years of treatment; however, the screening schedule should be altered depending on individual patient risk factors. Routine screening includes visual field testing (10-2 automated visual field, except for patients of Asian descent who require 24-2 visual field) and SD-OCT testing. Other modalities, such as FAF and mfERG can also be useful in cases of diagnostic uncertainty. Immediate conversations about discontinuing the drug should be made with the prescribing physician upon detection of retinal toxicity due to the irreversible nature of retinal damage.
Footnotes
CONTRIBUTIONS
All authors: conceptualization, visualization, writing - draft, writing - review & editing.
FUNDING
The authors declare no funding or support for this work.
COMPETING INTERESTS
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication November 24, 2025.
- Copyright © 2026 by the Journal of Rheumatology
This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.
