Criteria for Considering Studies for this Review
Types of Studies
We will include randomized controlled trials that compare secretagogues and artificial tears. Study inclusion will not be restricted on the basis of language or publication status.
Types of Participants
We will include trials where the study population comprises people 18 years old or older with dry eye disease.
Types of Interventions
Treatment groups: Secretagogues, including Pilocarpine, Cevimeline and Diquafosol
Control group: Artificial Tears intervention with different types.
Types of Outcome Measures
Primary Outcomes
Secondary Outcomes
Adverse Outcomes
We will compare the proportion of adverse outcomes between treatment groups. We will consider adverse outcomes as reported by included studies. Specific adverse outcomes of interest will include the following:
-
Eye discharge.
-
Eye irritation.
-
Eye pruritus.
-
Eye pain.
-
Foreign body sensation.
-
Allergic conjunctivitis.
Search Methods for Identification of Studies
Electronic Searches
The search of information will include the following electronic databases for randomized controlled trials. There will be no language or publication year restrictions.
-
Cochrane Central Register of Controlled Trials (CENTRAL) (which contains the CEV Trials Register) in the Cochrane Library (latest issue)
-
PubMed (1948 to present)
-
Scopus (2004 to present)
-
LILACS (Latin American and Caribbean Health Science Information Database (1982 to present)
-
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov
-
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP)
Searching Other Resources
We will search the reference lists of eligible studies identified from the electronic searches for additional relevant trials that may not have been identified from the electronic searches.
Data Collection and Analysis
Selection of Studies
The study selection will be performed by pairs of review authors, they will assess the titles and abstracts of articles identified through the literature search and will compare these against the inclusion criteria. Each article will be assessed independently by both authors and will be classified as either ’definitely relevant’, ’possibly relevant’, or ’definitely not relevant’.
Covidence software will be used to manage the screening process (25). Any disagreement will be resolved by a third review author.
Then we will obtain the full-text copies of all studies classified as ’definitely relevant’ or ’possibly relevant’. Each review author will independently assess each study for inclusion and will label it as either ’include’ or ’exclude’. A third review author will resolve any disagreement. We will document the reason for exclusion of each study excluded after reviewing the full report in a ’Characteristics of excluded studies’ table. We will use Google Translate to assess studies written in languages other than English and Spanish. We will illustrate the study selection process in a PRISMA diagram.
Data Extraction and Management
Pairs of review authors will independently extract data from the included studies using a data extraction form (Additional File 1) adapted from the Cochrane Eyes and Vision (CEV) data extraction form (26) developed by CEV and accessed via Covidence. A third review author will resolve any disagreements.
One review author will enter data into Review Manager 5 (RevMan 5) (27), and a second review author will verify the data entered.
Assessment of Risk of Bias in Included Studies
The Cochrane Collaboration Risk of Bias (RoB) tool will be used to assess the RoB in selection (random sequence generation and allocation concealment before randomization), performance (masking of study participants and personnel), detection (masking of outcome assessors), attrition (missing data and absence of an intention-to-treat analysis), reporting (selective outcome reporting), and other potential sources of bias (28).
Two authors of the review will classify the risk of bias as either ’low’, ’high’, or ’unclear’ (insufficient information for assessment). A third review author will resolve any disagreement between review authors.
Measures of Treatment Effect
We will calculate mean difference (MDs) with 95% confidence intervals (CIs) for continuous measures (e.g., the Schirmer test value, TBUT) and risk ratios (RRs) with the corresponding 95% CIs for dichotomous outcomes to estimate effects (e.g., adverse events, effective rate). We will choose a cut-off for ordinal outcomes and measurement scales to handle them as binary data or treat them as continuous data, as appropriate.
Unit of Analysis Issues
The participant will be the primary unit of analysis whenever: a) only one eye per participant is enrolled in the trial; or b) two eyes of an individual are treated as a single unit after being administered the same treatment (e.g. Pilocarpine, Cevimeline and Diquafosol). For studies that enrolled both eyes of participant and in which the eye was the unit of analysis, we will document whether the trial had a within-person design and analyzed the data appropriately.
Dealing with Missing Data
We will contact corresponding authors by email to obtain missing data or data reported unclearly in the study reports. We will allow three weeks for study authors to respond and will use the available information whenever there is no response. We will not impute missing participant data for analysis.
Assessment of Heterogeneity
We will compare the participant characteristics, study interventions, and outcomes across trials to assess for clinical and methodological heterogeneity.
We will use the I² statistic, which estimates the proportion of variation in observed effects not due to chance, to identify inconsistency among trials; an I² statistic value of greater than 50% (moderate heterogeneity)will represent substantial heterogeneity (28). Chi² test statistics will be used to assess the statistical heterogeneity among estimates of effect size from the included studies.
Assessment of Reporting Biases
If 10 studies, or more, are included we will visually inspect funnel plots of the intervention effect estimates for evidence of asymmetry to identify publication bias.
An asymmetric funnel plot may suggest small study effects, which could be the result of reporting bias, heterogeneity, or differences in the methodological quality of studies. We will assess selective outcome reporting as part of the ’Risk of bias’s assessment among individual studies.
Data Synthesis
If there is no substantial heterogeneity between studies (I² < 50%) we will apply the fixed effects model to synthesis the data and a random effects models if we encounter substantial heterogeneity. If we deem a meta-analysis as inappropriate, we will document the reasons and report findings from the individual studies narratively.
Subgroup Analysis and Investigation of Heterogeneity
If sufficient data are available from included studies, we will examine findings by the degree of DED severity at baseline among the study participants.
Sensitivity Analysis
Where possible, we will perform sensitivity analyses for primary and secondary outcomes to explore the effects of restricting our analyses to trials judged to have adequate allocation concealment, adequate masking of outcome assessors, and had at least 80% follow-up of participants in each group, to determine the robustness of the conclusion.
Summary of findings
When sufficient evidence is available, we will summarize the findings of the review using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach to assess the strengths and limitations of evidence for both primary and secondary outcomes.