Gouider 2021.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 51 participants (72 eyes) total; 28 participants (38 eyes) in FML group and 23 participants (34 eyes) in CsA group Unit of randomization: unclear Number analyzed (total and per group): see Notes Exclusions and losses to follow‐up (total and per group): see Notes How were missing data handled?: NR Sample size calculation: "The sample size was measured by our research department, to obtain a conclusion of noninferiority of CsA compared with fluorometholone in the chronic phase of EKC" Method of analysis: the authors planned "intention‐to‐treat" analysis, but the denominators of the reported outcomes did not match the numbers randomized (see Notes) Study duration (planned and actual participant follow‐up time): planned and actual: 7 months Time points at which outcomes were measured: on admission to study, and at 1, 3, 6, 7 months Diagnosis of EKC: "The diagnosis of EKC was based on the constellation of clinical features"
Participants	Country: Tunisia Setting: military hospital ophthalmology department Baseline characteristics Intervention: fluorometholone Age, years: Mean (SD), (range): 35.8 (17.5) (9 to 70) Gender: Women and girls, n (%): 15 (54%) Race: n (%): NR Prevalence of symptoms at time of diagnosis: 100% Underlying disease conditions: NR Comparator: topical 0.5% CsA Age, years: Mean (SD), (range): 36 (17.9) (9 to 77) Gender: Women and girls, n (%): 11 (48%) Race, n (%): NR Prevalence of symptoms at time of diagnosis: 100% Underlying disease conditions: NR Overall Age, years: Mean (SD), (range): 35.9 (17.5) (9 to 77) Gender: Women and girls, n (%): 26 (51%) Race, n (%): NR Prevalence of symptoms at time of diagnosis (see protocol for list of symptoms): 100% Underlying disease conditions: NR Inclusion criteria: SEI after EKC persisting for 14 days or more. The diagnosis of EKC was based on the constellation of clinical features. Exclusion criteria: participants with a history of glaucoma, other anterior or posterior segment disease or surgery, and chronic use of topical or systemic medications, pregnant women, contact lens wearers, and those who could not attend the whole follow‐up
Interventions	Intervention: fluorometholone 4 times a day for 1 month, then 3 times a day for 1 month and 2 times a day for 4 months; no other medication was allowed during the follow‐up Comparator intervention 1: topical 0.5% CsA 4 times a day for 1 month, then 3 times a day for 1 month and 2 times a day for 4 months; no other medication was allowed during the follow‐up
Outcomes	Primary outcomes of the review: NR Secondary outcomes of the review: NR Outcomes reported: SEI number, persistence of SEI at 3 and 6 months, recurrence of SEI at 7 months, clinical score, satisfaction score, BCVA (2 Snellen lines, LogMAR), Schirmer type 1, burning sensation on eye drops instillation, conjunctival injection, decreased TBUT (< 10 s), and adverse events such as ocular hypertension (> 6 mmHg above baseline), cataract, and corneal ulcer or bacterial superinfection
Identification	Sponsorship source: NR Author's name: Dhouha Gouider, MD Institution: Department of Ophthalmology, Military University Hospital of Tunis Potential conflicts of interest: "The authors have no funding or conflicts of interest to disclose" Trial registration number: NCT04376970
Notes	"Patients who had less than 1 month of treatment or an uncontrolled treatment side effect or who presented a degenerative iron line were excluded from the rest of the regimen" but included in the analysis. Pretherapeutic period (week), mean (range): 17.4 (2 to 104) for FML, 10.5 (2 to 30) for CsA, P = 0.102 Mean SEI number at baseline: 25.84 +/‐15.07 vs 34.59 +/‐22.49, P = 0.060 The study randomized 28 participants (38 eyes) to the FML group and 23 participants (34 eyes) to the CsA group. The denominator of outcome reported at 3 months was the number of eyes (verified by P value). In the FML group 27 participants (37 eyes): 1 participant was excluded from the analysis up to month 3. Proportion with sign resolution reported was likely at the eye level. In the CsA group, 22 participants (33 eyes): 1 participant was excluded from the analysis up to month 3. Proportion with sign resolution reported was likely at the eye level. The denominator of outcome reported at 6 months was the number of eyes (via personal communication). In the FML group, 3 participants (4 eyes) discontinued treatment before month 6 (25 participants in the study at month 6); the denominator is 33 (37‐4) eyes => 23/33 = 70%. In the CsA group, 3 participants (3 eyes) discontinued treatment before month 6 (20 participants in the study at month 6); the denominator is 30 (33‐3) eyes => 14/30 = 47%. The denominator of adverse effect estimation was the number of participants based on ITT analysis. In the FML group: 27 participants In the CsA group: 22 participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The allocation sequence of the treatment was generated by chance and concealed until all outcomes were assessed. The allocation codes were provided by a Web‐based module and were stratified according to clinical center with a permuted block method with randomly chosen block sizes."
Allocation concealment (selection bias)	Low risk	"The allocation sequence of the treatment was generated by chance and concealed until all outcomes were assessed. The allocation codes were provided by a Web‐based module and were stratified according to clinical center with a permuted block method with randomly chosen block sizes. The generated codes were sealed and only opened by a pharmacist who was not involved in patient care."
Masking of participants and personnel (performance bias)	Low risk	"The subjects were masked to the contents. In monthly visits, they were instructed to return the empty tubes before getting new filled ones. Neither the patient nor the ophthalmologist knew the allocated treatment."
Masking of outcome assessment (detection bias)	Low risk	"Neither the patient nor the ophthalmologist knew the allocated treatment."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	In the publication, the authors stated that only 1 participant in each treatment group (3.6% and 4.3%) was excluded from the analysis. During personal communication, the authors corrected the information by providing additional numbers of participants who were withdrawn from the study after 1 month of treatment but were included in the analysis for outcomes up to month 3. For outcomes reported at month 6, only eye‐level outcomes were reported (30/ 34 eyes in the CsA group, 33/38 eyes in the FML group); the authors did not respond to further requests for additional clarification.
Selective reporting (reporting bias)	High risk	All outcomes specified in trial registry record were reported in the published article. However, the numbers of eyes being evaluated for the outcomes of interest were not always provided and did not match the percentages reported in the publication.