Hydroxychloroquine Dose per Ophthalmology Guidelines and the Risk of Systemic Lupus Erythematosus Flares

Abstract

This study of an academic medical center patient cohort with systemic lupus erythematosus (SLE) investigates the association between hydroxychloroquine dose based on current guidelines and risk of lupus flares.

Hydroxychloroquine prevents systemic lupus erythematosus (SLE) flares and is recommended for all patients with SLE.¹ However, its long-term toxicity includes retinopathy, particularly at doses greater than 5 mg/kg per day.² This dose-dependent risk led to 2016 ophthalmology guidelines³ and subsequent rheumatology recommendations¹ to avoid prescribing hydroxychloroquine doses greater than 5 mg/kg per day. We assessed the association of this dose threshold on lupus flares.

Methods

We identified patients with SLE from Massachusetts General Hospital who had 2 or more rheumatology visits and used hydroxychloroquine between January 1, 2016, and December 31, 2020, the 5-year period immediately following the 2016 hydroxychloroquine dosing guidelines.³ This study was approved by the Mass General Brigham Institutional Review Board, which waived informed consent.

We conducted a case-crossover study for which each patient served as their own control. We defined case periods as the 6 months prior to a lupus flare. Control periods were defined as each 6-month period prior to a visit with no flare (eFigure in the Supplement). Each patient could have multiple nonoverlapping case and control periods. On medical record review, we assessed the outcomes of all lupus flares documented at each rheumatology encounter. Flares were defined using the revised Safety of Estrogens in Lupus National Assessment (SELENA)–Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) flare index (rSFI), including grades mild, moderate, and severe.⁴

The exposure of interest was the mean weight-based hydroxychloroquine dose assessed during each 6-month case or control period, categorized as either less than or equal to 5 mg/kg per day or greater than 5 mg/kg per day.^1,3

We used conditional logistic regression analysis, matched within a patient, to assess the association of hydroxychloroquine dose over the prior 6 months with the risk of lupus flares, adjusting for glucocorticoid use, immunosuppressant use, and disease activity (clinical SLEDAI score,⁵ continuous scale, range 0-101; ≤4 considered low disease activity) prior to each period. We repeated the analysis limited to moderate or severe flares, with mild flares included as control periods. We conducted subgroup analyses by SLE duration, history of lupus nephritis, and baseline SLEDAI score. To characterize the dose-response trend for the risk of lupus flares, we used restricted cubic splines to generate a smoothed dose odds ratio (OR) curve. We used SAS version 9.4 (SAS Institute Inc). Statistical significance was set at a 2-sided P < .05.

Results

Of 342 patients with SLE who used hydroxychloroquine during the study period, 168 (49%) had at least 1 lupus flare and were included in the case-crossover analysis; 80 (23%) had a moderate or severe flare. The mean age was 43 years (SD, 14.7); 148 (88%) were women. The mean baseline clinical SLEDAI score was 3.5 (SD, 3.4); 25 (21%) had lupus nephritis, 72 (43%) used glucocorticoids, and 80 (48%) used oral immunosuppressants.

There were 308 case periods (mean, 1.8 period per patient), including 141 moderate or severe flare periods, and 420 control periods. The adjusted OR (AOR) for any lupus flare associated with taking hydroxychloroquine doses of 5 mg/kg per day or less vs more than 5 mg/kg per day was 1.98 (95% CI, 1.03-3.79) and for moderate or severe lupus flares the AOR was 6.04 (95% CI, 1.71-21.30) (Table). The corresponding AORs were similar across subgroups but only reached significance for patients with baseline SLEDAI scores of 4 or less (AOR, 2.41 [95% CI, 1.17-4.96]). The smoothed dose OR curve indicated an apparent threshold near 5 mg/kg per day of hydroxychloroquine for an increased risk of flares (Figure).

Table. Association of Hydroxychloroquine Dose With Risk of Systemic Lupus Erythematosus Flares.

Hydroxychloroquine dose, mg/kg/d	Periods		OR (95% CI)
	Lupus flare cases (n = 308)	Nonflare controls (n = 420)	Unadjusted	Adjusteda
Overall analysis
All rSFI flaresb
>5	131	185	1 [Reference]	1 [Reference]
≤5	177	235	1.86 (1.00-3.47)	1.98 (1.03-3.79)
Moderate or severe flares	141	243
>5	52	101	1 [Reference]	1 [Reference]
≤5	89	142	4.04 (1.35-12.09)	6.04 (1.71-21.30)
Subgroup analyses
Duration of SLE <10 y
>5	79	106	1 [Reference]	1 [Reference]
≤5	85	101	1.56 (0.70-3.49)	1.75 (0.75-4.09)
Duration of SLE ≥10 y
>5	52	79	1 [Reference]	1 [Reference]
≤5	92	134	2.40 (0.88-6.54)	2.37 (0.83-6.72)
Lupus nephritis
>5	15	27	1 [Reference]	1 [Reference]
≤5	48	64	1.99 (0.47-8.56)	2.14 (0.47-9.70)
No lupus nephritis
>5	116	158	1 [Reference]	1 [Reference]
≤5	129	171	1.83 (0.92-3.65)	1.95 (0.95-4.01)
Baseline SLEDAI-2K ≤4c
>5	97	160	1 [Reference]	1 [Reference]
≤5	127	185	2.34 (1.16-4.72)	2.41 (1.17-4.96)

Abbreviations: OR, odds ratio; rSFI, revised Safety of Estrogens in Lupus National Assessment (SELENA)–Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) flare index; SLE, systemic lupus erythematosus; SLEDAI-2K, Clinical Systemic Lupus Erythematosus Disease Activity Index-2000.

^a Odds ratios were calculated using conditional logistic regression analysis. Analyses were adjusted for SLEDAI-2K, glucocorticoid use, and SLE immunosuppressant use at the visit proceeding each period.

^b A validated instrument for measuring lupus flares that are graded as mild, moderate, or severe. One hundred sixty-eight patients were included in the all rSFI flares analysis, and 80 patients were included in the moderate or severe flares analysis. Mild flare periods were considered nonflare periods in the moderate or severe flares analysis. After excluding 32 case and 26 control periods without hydroxychloroquine use, there were 728 total 6-month periods.

^c A validated measure of lupus disease activity, with scores ranging from 0 to 101. Scores of 4 or less are considered low disease activity.

Figure. Estimated Odds Ratio for Lupus Flares According to the Average Weight-Based Hydroxychloroquine Dose.

The dose odds ratio curve was generated using the conditional logistic regression model adjusting for Systemic Lupus Erythematosus Disease Activity Index (SLEDAI), glucocorticoid use, and systemic lupus erythematosus (SLE) immunosuppressant use at the visit proceeding each 6-month period. The smoothed curve was fitted with restricted cubic splines with knots at 2.5 mg/kg per day, 5 mg/kg per day, and 6 mg/kg per day. The reference hydroxychloroquine dose was 5 mg/kg per day.

Discussion

In this case-crossover study within a contemporary SLE cohort, hydroxychloroquine dosing of 5 mg/kg per day or less, in accordance with current ophthalmology and rheumatology guidelines, was associated with a higher risk of lupus flares, including moderate or severe flares. Furthermore, the dose trend analysis also suggested an apparent threshold near 5 mg/kg per day. A recent study demonstrated that a higher risk of lupus flares was associated with any dose reduction, irrespective of the specific dose.⁶ Together, these findings suggest reduced efficacy of lower hydroxychloroquine dosing for lupus disease activity, particularly around doses of 5 mg/kg per day or less.

Study limitations include the retrospective study design and the lack of information on medication adherence within each period. Additionally, this study used an academic medical center cohort, and future studies should replicate these findings in diverse settings.

This study highlights the need to consider individualized risks and benefits in choosing the optimal dose of hydroxychloroquine, an important medication in lupus care.

Section Editors: Jody W. Zylke, MD, Deputy Editor; Kristin Walter, MD, Senior Editor.

Supplement.

eFigure. Case-crossover Study Design