This randomized controlled trial was conducted at the Department of Ophthalmology at Seoul National University Hospital from May 2011 through August 2013. Inclusion criteria were 1) adult (aged > 18 years) patients with pSS diagnosed on the basis of the 2002 American–European Consensus Group (AECG) criteria (3) and 2) subjects who had the ability to give informed, dated, and signed consent before the beginning of any proceedings related to the trial.

Exclusion criteria were 1) previous treatment with HCQ with insufficient discontinuation time (3 weeks) after discontinuation of treatment; 2) known cardiac disease, respiratory disease, renal disease, or gastrointestinal disease (except gastrooesophageal reflux disease); 3) diabetes mellitus; 4) psoriasis; 5) known drug allergy or hypersensitivity; 6) previous or ongoing treatment with any other drugs (including topical drugs) that may affect the lacrimal system with insufficient washout time after discontinuance of treatment (e.g., SSRI, anti-histamines, and pilocarpine); 7) closed-angle glaucoma; 8) previous intraocular surgery; 9) macular disease; 10) previous or ongoing treatment with drugs that may have an effect on the macula; 11) pregnancy; and 12) planning pregnancy.

A double-blind, randomized control study was conducted in 39 subjects of 153 pSS subjects for 2 years. We were unable to recruit the intended 60 subjects within the 2-year recruitment period because many subjects with pSS did not meet our inclusion criteria. Most of the subjects did not wish to discontinue previous oral drugs because of severe systemic symptoms. During the study period, 153 pSS patients visited our clinic. Among them, only 67 subjects met all eligibility criteria. Twenty-eight subjects declined to participate in the trial, and 39 subjects were finally enrolled in this study (Fig. 1).

During the study period of 2 years, a new criterion for SS was reported (11). Because this study was designed before the recent criterion was published, we enrolled subjects according to the previous criteria of the American–European Consensus Group to maintain consistency of the subject population.

All subjects underwent initial medical and ophthalmologic history taking and physical examination at the baseline visit (week 0). The study eye was selected as the eye that showed the higher corneal fluorescein staining score (on a scale of 0–15; National Eye Institute scale [12]) at the baseline visit. Subjects were randomly assigned to the HCQ or placebo group by using sealed randomization envelopes. The placebo was manufactured to be identical in appearance to the active drug, and all tablets were kindly supplied by Kyung Poong Pharma Co. Both the investigators and subjects were blinded to the treatment assignments. The subjects were instructed to take two tablets (one 200 mg tablet and one 100 mg tablet) of the study medication, giving a total of 300 mg, once daily (qd) for 12 weeks and to report missed doses and adverse events. All subjects used Hyalein Mini ophthalmic solution 0.1% (0.1% hyaluronic acid; Santen, Osaka, Japan) six times per day. Other topical medications were restricted. Participants returned to the study site at 6 and 12 weeks for efficacy and safety evaluations. After 12 weeks, oral HCQ and placebo medication were withheld, whereas topical medication was continued in all subjects for 4 weeks. The subjects revisited the clinic at 16 weeks. At each visit, the visual acuity, corneal fluorescein staining score using NEI grading, Schirmer test score without anesthesia, tear film break-up time (TBUT), Hardy Rand and Rittler (HRR) color vision test, and fundus examination were performed. Schirmer test was done with a Schirmer strips (Eagle Vision, Memphis, TN, USA) for five minutes without topical anesthesia. TBUT was measured after instillation of one drop of 0.25% fluorescein dye into the conjunctival sac, and the subjects were asked to blink several times. Subjective assessment was performed using the ocular surface disease index questionnaire (OSDI) (13). The OSDI questionnaire was translated to Korean from the original English. Tear and blood samples were collected for measurement of IL-6, BAFF, ESR, and IL-17 levels in addition to Th17 cells that secrete IL-17.

Tear fluid was obtained from the patients at the same time at baseline and at 6, 12, and 16 weeks. Tears were collected from the medial and lateral canthus. Topical anesthesia was not used. To minimize ocular surface irritation, we obtained tear samples by using a Merocel sponge (PVA 0525; Oasis, Glendora, CA, USA) (14). After collection, the sponge was inserted into a 0.5-mL tube (Eppendorf, Fremont, CA, USA) and the tear fluid was subsequently recovered by centrifugation at 10,000 rpm for 10 minutes.

ESR was measured by conventional methods in our hospital laboratory on the day of the patient visit. For BAFF and IL-6, 0.5 mL of plasma was separated from heparinized peripheral blood and stored together with the tear sample at -70°C until further examination.

The concentrations of BAFF, IL-6, and IL-17A were measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions. ELISA using serum was performed in duplicate to ensure the reproducibility of the data. However, ELISA using tears was performed only once because of the lack of a sufficient amount of tears. Tear samples were diluted 1:12.5 and 1:100 for BAFF and IL-17, respectively.

The concentration ranges used for the standard curve were 62.5–4,000 pg/mL, 9.38–600 pg/mL, and 15.6–1,000 pg/mL of human recombinant BAFF, IL-6, and IL-17, respectively (R&D Systems).

Flow cytometry was also performed on the same day as the patient visit. Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood by density gradient centrifugation (BIOCHROM Inc., Cambridge, UK). PBMCs were stimulated for 6 hours with 50 ng/mL 1-phorbol-12-myristate-13-acetate (PMA; Sigma-Aldrich, St Louis, MO, USA) and 1 μg/mL ionomycin (Sigma-Aldrich) in the presence of brefeldin A (BD Bioscience, San Jose, CA, USA) for the final 4 hours. The cells were fixed and permeabilized using a BD Cytofix/Cytoperm kit (BD Bioscience). The fixed cells were stained with anti-CD3-APC-Cy7 and anti-CD4-FITC for 30 minutes at 4°C for initial surface staining. For intracellular staining, the cells were incubated with anti-IFN-γ-PE-Cy7 (all from BD Bioscience) and anti-IL-17A-PE (eBioscience, SanDiego, CA, USA) mAb for 60 minutes at 4°C. Flow cytometric analysis was performed using a FACS BD LSR II (BD Bioscience), and the data were analyzed by FlowJo (Treestar, Ashland, OR, USA). Isotype mouse Ig G1-PE-Cy7 (eBioscience) and mouse Ig G1- PE (eBioscience) were used as controls.

After four-color compensation, lymphocytes were gated in the forward scatter (FSC)/side scatter (SSC) gate, and CD4 T cells were then gated by CD3^hi and CD4^hi expression; finally, the cells were gated by IFNγ and IL17 expression (Fig. 2). The percentage of Th17 cells was calculated by summing the amounts of IL17^hi cells and IFNγ/IL17 ^{double hi} cells.

A previous literature review on SS did not include results for the tear IL-17 concentration after HCQ medication. The tear IL-17 concentration of SS patients was 504.91 pg/mL and 352.45 pg/mL in the total and mild keratoconjunctivitis sicca groups, respectively, in this previous report (15). We assumed that the tear IL-17 concentration after HCQ medication would be similar to that of the mild keratoconjunctivitis sicca group from the previous study. A sample size of 60 patients (30 in each group) was needed to detect a significant difference between 504.91 ± 136.38 pg/mL and 352.45 ± 136.38 pg/mL in the 2 groups using a 2-sided test with a power of 80% and the significance level controlled at 5%. The Medical Research Collaborating Center of Seoul National University Hospital assisted in the sample size calculation.

Primary analyses of the data were based on the per-protocol population, which included all participants who took the 12-week course of study medication. Statistical analyses were performed using PASW software for Windows (v. 19.0; SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant. The Shapiro–Wilk test was applied to evaluate data normality. The two groups were compared with Student’s t-test (two-tailed) of the Mann–Whitney U test depending on normality. To compare changes within groups at different time intervals, a paired sample t-test and Wilcoxon signed-rank test were used depending on normality. Differences between the HCQ and placebo groups during the course of the treatment were analyzed by repeated-measures ANOVA and linear mixed models when data were missing.

Written informed consent was obtained from all subjects, and the study was granted ethical approval by the institutional review board of Seoul National University Hospital (IRB number: H-1104-083-359). This study was conducted in accordance with the tenets of the Declaration of Helsinki. This study was registered with ClinicalTrials.gov (Identifier: NCT01601028).

Of the 39 recruited subjects, 26 were treated per protocol. Thirteen subjects including five in the placebo group and nine in the HCQ group did not complete the study. The initial age, ESR, TBUT, Schirmer score, corneal staining score, and OSDI were not significantly different between complete and incomplete groups. In the placebo group, three subjects did not keep their follow-up appointments and were unavailable to be reached, one subject suffered dyspepsia, and one subject was found to be unsatisfactory for enrolment. In the HCQ group, two subjects did not keep their follow-up appointments and were unable to be contacted, three subjects suffered dyspepsia, one subject did not have the time for follow-up examinations, two subjects were found deficient with regard to the enrollment criteria, and one subject developed incidental subretinal hemorrhage because of occult myopic choroidal neovascularisation (CNV) with accompanying impaired vision 3 days after medication, which was not regarded as a drug-related complication. There were no other drug-related side effects with regard to ocular complications (optic neuropathy or maculopathy) or general subjective complications except dyspepsia (10.3%).

Of the 26 pSS subjects analyzed, 11 subjects were randomly assigned to receive HCQ and the remaining 15 subjects were randomly assigned to receive the placebo. All subjects were women with a mean age of 56.85 years (range, 35–74). The mean (range) age was 55.0 (35–73) years in the placebo group and 59.4 (44–74) years in the HCQ group, which was not significantly different (P = 0.263, independent t-test). The initial results for ESR, serum IL-6, BAFF, Th17, tear BAFF, TBUT, Schirmer score, corneal staining score, and OSDI are shown in Table 1. There were no significant differences between the groups for all of the demographic factors and parameters.

Changes in parameters are shown in Table 2 and Figs. 2 and 3. Regarding systemic inflammatory parameters, the mean ESR tended to continuously increase in the placebo group and decrease in the HCQ group, although there were no significant differences between the groups. The mean serum IL-6, BAFF, proportion of Th17 cells, and tear BAFF were not significantly different between the groups. Tear IL-17 was not detected in almost of all the samples.

With respect to the ocular signs and symptoms, TBUT did not change significantly from the baseline to week 12 in either group. The Schirmer test score did not significantly change in either group. The HCQ group tended to show continuous reduction of the corneal staining score, although the reduction was not significant. OSDI was significantly decreased between the baseline and weeks 6 and 12 only in the HCQ group (P < 0.05, baseline vs. 6 weeks; P < 0.01, baseline vs. 12 weeks; paired t-test; Fig. 3H), suggesting improvement of subjective symptoms. However, the HCQ and placebo groups were not significantly different (P = 0.209, repeated-measures ANOVA).

After discontinuation of the oral medication, only the proportion of Th17 cells was significantly increased in both groups (P < 0.01, repeated-measures ANOVA). However, there was no difference between the groups (P = 0.566). Other parameters did not show significant changes after the oral medications were discontinued (Table 2, Fig. 2).