Among the 103 patients included, 77 were randomized to and received HCQ, and 26 were randomized to and received placebo (3:1 ratio) (Figure 1 ). Ninety‐six patients (72 in the HCQ/HCQ group and 24 in the placebo/HCQ group) were included in the full analysis set. All 103 patients were included in the safety analysis set.

The baseline characteristics of patients in the full analysis set are shown in Table 1 . In total, 74.0% were female with a mean age of 42.7 years, and 58.3% had SLE (as defined by the 1997 American College of Rheumatology criteria for the classification of SLE) 25 . The mean ± SD CLASI activity scores at baseline were 13.5 ± 8.0 and 13.6 ± 7.5 in the HCQ/HCQ and placebo/HCQ groups, respectively. The characteristics of the 2 groups did not differ significantly.

Efficacy

The CLASI activity scores over time in the placebo/HCQ group and the HCQ/HCQ group are shown in Figure 2. The primary end point, a reduction in the CLASI activity score, was achieved after 16 weeks of HCQ therapy (−4.6; 95% confidence interval [95% CI] −6.1, −3.1 [P < 0.0001, based on the LOCF]). A reduction in the CLASI score (−3.2; 95% CI −5.1, −1.3 [P = 0.002, based on the LOCF]) was also observed in the group that received placebo for 16 weeks. The group difference was −1.6 (95% CI −4.3, 1.1 [P = 0.197]). The CLASI activity score in the placebo/HCQ group decreased further after switching from the placebo to HCQ and reached the level of the HCQ/HCQ group at week 32. The CLASI activity score in both groups continued to decrease gradually after week 32. Smoking status and treatment with systemic steroids or immunosuppressive agents were not associated with the end point (data not shown).

Figure 2 Open in figure viewerPowerPoint CLASI activity scores over time in the placebo/HCQ and HCQ/HCQ groups (observed cases). Values are the mean ± SE. See Figure 1 for definitions.

The percentages of patients in each of the 5 central photo evaluation categories are shown in Figure 3A. At week 16, this evaluation showed greater improvement from baseline in the HCQ/HCQ group compared with that in the placebo/HCQ group. At week 16, the proportions of patients shown to be “improved and remarkably improved” were 59.4% in the HCQ/HCQ group and 30.4% in the placebo/HCQ group (P = 0.029).

Figure 3 Open in figure viewerPowerPoint A, Percentages of patients in each central photo evaluation category (5 grades) at weeks 16, 32, and 52 (observed cases). B, Percentages of patients in each patient's global assessment category (7 grades) at weeks 16, 32, and 52 (observed cases). C, Change in the Skindex‐29 score from baseline to weeks 16, 32, and 52 (observed cases). Values are the mean ± SE. HCQ = hydroxychloroquine.

The 7‐point PGA scale data are shown in Figure 3B. At week 16, the proportions of patients shown to be “improved and remarkably improved” were 21.4% in the HCQ/HCQ group and 13.0% in the placebo/HCQ group (P = 0.546). However, the proportions of patients who were “slightly improved, improved, and remarkably improved” reached 72.9% in the HCQ/HCQ group and 47.8% in the placebo/HCQ group (P = 0.041).

Changes in the total Skindex‐29 score (a skin‐specific QoL measure) from baseline to weeks 16, 32, and 52 are shown in Figure 3C. Scores for all domains of the Skindex‐29 (emotions, symptoms, and functioning) and the overall score significantly improved from baseline to week 16 in the HCQ/HCQ group (for emotions, −10.2 [95% CI −14.3, −6.0] [P < 0.0001]; for symptoms, −6.9 [95% CI −11.2, −2.6] [P = 0.002]; for functioning, −5.5 [95% CI −9.0, −2.0] [P = 0.002]; and for overall score, −7.4 [95% CI −10.9, −4.0] [P < 0.0001]). In the placebo/HCQ group, scores for all domains of the Skindex‐29 decreased, although not to a statistically significant degree (for emotions, −5.8 [95% CI −14, 2.5] [P = 0.163]; for symptoms, −4.4 [95% CI −11.1, 2.4] [P = 0.195]; for functioning, −3.0 [95% CI −9.6, 3.6] [P = 0.354]; for overall score, −4.3 [95% CI −10.3, 1.7] [P = 0.150]; for differences between groups, P = 0.137, P = 0.583, P = 0.418, and P = 0.258, respectively).

IGA, which is a major secondary end point based on CPE, PGA, and QoL, differed significantly in terms of the proportion of patients in the “improved” and remarkably improved" categories between the HCQ/HCQ group (51.4%) and the placebo/HCQ group (8.7%) at week 16 (P = 0.0002) (Figure 4). The proportion of patients in the “slightly improved,” “improved,” and “remarkably improved” categories reached 78.6% in the HCQ/HCQ group and 56.5% in the placebo/HCQ group (P = 0.057).

Figure 4 Open in figure viewerPowerPoint Percentages of patients in each investigator's global assessment category (7 grades) at weeks 16, 32, and 52 (observed cases). HCQ = hydroxychloroquine.

Regarding the efficacy end points for the assessment of patients with SLE (n = 56), the pain score decreased significantly from baseline to week 16 in the HCQ/HCQ group (−1.0 [95% CI −1.7, −0.3] [P = 0.005]), but increased in the placebo/HCQ group (0.3 [95% CI −1.1, 1.7] [P = 0.658]), demonstrating a significant difference between the 2 groups (−1.4 [95% CI −2.6, −0.1] [P = 0.033 by ANCOVA]). The decrease in the fatigue score from baseline to week 16 was significant in the HCQ/HCQ group (−1.1 [95% CI −1.9, −0.3] [P = 0.006]) but not in the placebo/HCQ group (−0.7 [95% CI −2.7, 1.3] [P = 0.452]). The group difference was −0.7 (95% CI −2.2, 0.8 [P = 0.367]).

The PGA of SLE decreased significantly from baseline to week 16 (−0.7 [95% CI −1.3, −0.1] [P = 0.03]) in the HCQ/HCQ group (n = 42) but did not change in the placebo/HCQ group (n = 12) (0.0 [95% CI −1.8, 1.7] [P = 0.959]). The group difference was −0.9 (95% CI −2.2, 0.4 [P = 0.165]).

RAPID‐3 scores decreased significantly from baseline to week 16 in the HCQ/HCQ group (−1.7 [95% CI −2.9, −0.4] [P = 0.009]) but increased in the placebo/HCQ group (0.2 [95% CI −2.7, 3.0] [P = 0.896]). The group difference was −2.1 (95% CI −4.6, 0.4 [P = 0.094]).

The proportion of IGA of SLE in the “improved” and “remarkably improved” categories was 47.4% in the HCQ/HCQ group and 36.4% in the placebo/HCQ group at week 16 (P = 0.733). The proportion of those in the “slightly improved, improved, and remarkably improved” categories reached 68.4% in the HCQ/HCQ group and 54.5% in the placebo/HCQ group (P = 0.480).

Among SLE patients with active musculoskeletal disease based on the BILAG grade (A–C in “musculoskeletal” system), 16 (42.1%) of those in the HCQ group (n = 38) improved by at least 1 grade, and 2 (5.3%) improved by at least 2 grades at week 16. In contrast, 4 patients (40.0%) in the placebo group (n = 10) improved by at least 1 grade, while none (0%) improved by 2 grades. Among SLE patients in the HCQ group who had active disease based on the BILAG grade (A–C in “constitutional” system) (n = 7), 4 (57.1%) improved by at least 1 grade and 1 (14.3%) improved by 2 grades at week 16. In contrast, among those in the placebo group (n = 2), 1 (50.0%) improved by at least 1 grade and none (0%) improved by 2 grades.

Changes in the anti‐dsDNA antibody level (IU/ml) and complement C3 and C4 levels (mg/dl) from baseline to week 16 were −3.8 (95% CI −11.1, 3.5 [P = 0.299]), 4.2 (95% CI 0.3, 8.0 [P = 0.034]), and 1.3 (95% CI 0.2, 2.4 [P = 0.022]), respectively, in the HCQ/HCQ group (n = 40). Corresponding changes in the placebo/HCQ group (n = 11) were −1.9 (95% CI −6.8, 3.1 [P = 0.419]), −1.2 (95% CI −7.8, 5.4 [P = 0.699]), and −0.3 (95% CI −3.0, 2.4 [P = 0.827]), respectively. The dosage of prednisolone (mg/day or equivalent) decreased by −1.0 (95% CI −1.8, −0.3 [P = 0.007]) in the HCQ/HCQ group (n = 29) and by −0.8 (95% CI −1.8, 0.2 [P = 0.095]) in the placebo/HCQ group (n = 11) from baseline to week 52.