Study design and objective
The Ocular Rosacea microBiome Study (ORBS) is a three-armed, randomized, masked, placebo-controlled clinical trial. Participants with a diagnosis of symptomatic ocular rosacea (n = 50) at the University of California, San Francisco in San Francisco, California, United States, will be prospectively enrolled and randomized to one of three arms in a 2:2:1 fashion. 1) Arm A, a SDD dose of doxycycline (40 mg) taken for eight weeks, 2) Arm B, an antibiotic dose of doxycycline (200 mg) taken for eight weeks and 3) Arm C, a placebo pill taken for eight weeks. The trial will begin enrollment in September 2022. This protocol adheres to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines (Supplementary Table S1) and is registered on clinicaltrials.org (NCT05296837).10 This study was reviewed and approved by the Institutional Review Board at the University of California, San Francisco (21-34725) and will be done in accordance with the Declaration of Helsinki.
The objectives of this trial are to investigate if there are detectable alterations in tetracycline antimicrobial resistance gene expression and microbiome of the ocular surface and gut with administration of SDD for the treatment of ocular rosacea. Our specific aims are to investigate 1) alterations of antimicrobial resistance (AMR) in the gut and ocular surface, 2) alterations to the gut and ocular surface microbiota, and 3) changes in conjunctiva immune response expression. We hypothesize that doxycycline at both the antibiotic treatment dose (200 mg) and SDD dose (40 mg) will increase tetracycline resistance gene expression, alter microbiota, and alter immune response genes relative to placebo.
Setting, participants, and eligibility:
The study setting for this trial is the ophthalmology clinics at the F.I. Proctor Foundation and Department of Ophthalmology and the dermatology clinics at the University of California, San Francisco (UCSF). Participant recruitment will include treating clinicians referring patients with ocular rosacea to one of the study staff listed on the protocol. The study coordinator will review the patient’s clinical information to determine if they are eligible for the study. The study staff members will approach, and consent interested participants for the study.
Inclusion criteria for this study include if the participant has 1) symptomatic ocular disease attributed to ocular rosacea as the primary diagnosis, 2) the ability the give informed consent, and 3) is age eighteen years old or older as doxycycline is contraindicated in children under ten years old. Patients with ocular rosacea will be excluded from study participation if they have 1) an active ocular or systemic infection, 2) a known allergy or intolerance to tetracycline antibiotics, 3) prior use of oral antibiotics within the last three months, or 4) pregnancy or the possibility of becoming pregnant within the eight-week study medication timeline as tetracyclines are considered teratogenic.
Randomization and masking
The clinical study coordinator will perform study randomization according to the 2:2:1 randomization schedule using the REDCap randomization module to allocate patients to groups. Given the modest trial size, we will use permuted block randomization to prevent overt imbalance. All study arm medications will be centrally located in sealed containers that the study coordinator will hand out according to randomized allocation. We will use letters, A, B, and C, to mask treatment arms. Each day’s dose will be marked in individual AM/PM packets. All patients will be masked to their SDD allocation and informed that they will be taking a pill by mouth twice a day. Participants taking 40mg doxycycline will be taking one 20mg doxycycline pill twice a day. Participants taking 200mg doxycycline will be taking 100mg twice a day. Participants allocated to placebo will be taking a placebo pill twice daily. All examining physicians responsible for tear breakup time (TBUT) determination and ocular photography will be masked to patient allocation assignment. One biostatistician will generate the randomization key and another will perform outcome analysis masked to patient allocation assignment. Only the study coordinator, one biostatistician, and one study coordinator from a separate project will have access to the randomization key. Emergency unblinding is acceptable if a participant experiences an adverse event.
Doxycycline at 40mg and 200mg and placebo will be purchased from the vendor, Amerisource Bergen.
Intervention and sample size
The study coordinator will offer a sealed container to each participant during their primary visit. They will instruct the patients to take both doses, AM and PM, daily for eight weeks. The initial dose will not be offered under clinical observation as the pills for each study arm will look different and can unmask the participant to the study team. Monitoring for adherence and side effects will be done at study follow up visits.
We estimate that twenty patients per arm will provide over 80% power to detect a two - fold increase in tetracycline resistance genetic determinants between Arms A and B, assuming the standard deviation in read number found previously, allowing for 10% drop out, and correcting for baseline.11 As a secondary analysis, we estimate that twenty patients in Arm A and ten in Arm C will provide over 80% power to detect a three-fold increase in tetracycline resistance genetic determinants.
We estimate that twenty patients per arm will provide over 80% power to detect a difference between Arms A and B in Shannon’s diversity effective number of 2 (from an effective number of 10. That is, 10 to 8 or 10 to 12 effective number), assuming 10% drop-out and a standard deviation in the effective number of 1.9, found in two previous studies.12,13 As a sensitivity analysis, we anticipate similar power using Simpson’s diversity (again, expressed as an effective number).
Outcomes assessment, schedule, and procedures
The primary outcome of interest is comparing the load, a quantitative measure, of AMR genetic determinants in lower lid and rectal swab samples between arms A (SDD dose) and arm B (placebo) when corrected for baseline. The secondary outcome of interest is microbiome diversity of the ocular surface and gut through Shannon’s Index and Simpson’s index. Other exploratory outcomes include comparative analysis of 1) host conjunctiva inflammasome expression, 2) Ocular Surface Disease Index (OSDI), with scores ranging from 0 to 100 with higher scores indicating greater symptom severity, and 3) TBUT scores with numbers representing how many seconds after a full blink a discontinuity in the tear film appears. A normal tear breakup time is greater than 10 seconds of tear stability between blinks. All outcomes will be compared between the three study arms, corrected for baseline.
Participants will be seen for four approximately 40-minute visits over the course of the study: at baseline, four weeks, eight weeks, and once four months after their last dose of study medication. The four-week visit will be 25–35 days after the initial visit, the eight-week visit 49–63 days after the initial visit, and four months 90–150 days after the initial visit. Participants will undergo ocular evaluation at every visit as well as have a lower lid margin swabs from each eye and a rectal swab collected at baseline, eight weeks, and three to six months after the first dose of study medication. Table 1 summarizes the procedure schedule for each visit under SPIRIT guidelines.10
The following ocular evaluation procedures should be performed in the order listed below to minimize disturbance of the ocular surface.
1) Tear breakup time (TBUT):
It is preferable to get an automatic TBUT measurement, such as can be obtained through an Oculus keratography topographer. If this topographer is not available, then the TBUT should be measured by a physician masked to the study assignment. TBUT should be measured prior to application of anesthetic.
If in the case a doctor is measuring the TBUT, they should utilize the following procedure:
-
A fluorescein strip, only gently moistened, should be touched lightly to the tear meniscus.
-
The patient instructed to blink several times.
-
After even distribution of the fluorescein, the patient should be asked to blink again and then hold the eye open and the doctor starts a stopwatch.
-
The tear film is examined under cobalt blue light at the slit lamp.
-
The doctor stops the stopwatch when the first break in the tear film occurs.
-
This measurement is repeated a total of three times, and all three times recorded.
2) Intraocular pressure (IOP):
Non-contact tomography, such as iCare, is preferred as it requires no anesthetic that will alter the tear film. If iCare is not available, then traditional Goldmann applanation tonometry, as is routine with all eye examinations, may be performed at the end of the examination.
3) Ocular Surface Disease Index (OSDI):
The patient will complete the standard 12-question questionnaire created by the Outcomes Research Group at Allergan Inc. (Supplementary S2).
4) Ocular Photography:
Lid margin and conjunctiva will be photographed either with a slit lamp camera or with the ocular keratography topographer that has lid imaging modalities.
Sample collection procedures for the bilateral lower lid and rectum are summarized below. Study coordinators will provide separate sample collection kits for the lower lid and rectal samples. Samples will be processed in the Ralph & Sophie Heintz Laboratory at the Proctor Foundation for RNA deep sequencing with subsequent bioinformatic analysis.
1) Bilateral Lower Lid Swab Collection:
The study coordinator will obtain patient consent before collection. A sample collection tube labeled with the patient’s ID sticker will be filled with 1ml of Zymo solution to stabilize the RNA/ DNA in the sample. The collection kit will also contain an applicator swab, tube rack, and parafilm to prevent leakage after collection.
For sample collection, the physician will wear gloves and a mask, and instruct the patient that neither of them should speak during swabbing as RNA-seq is very sensitive and can pick up oral flora contamination in the air. No anesthetic drops will be used to further limit contamination. The physician will swab the first eye using the polyester tipped applicator swab (Fig. 2a), pulling down the lower eyelid and rolling the swab across the length of the inferior lid margin twice, while rotating the swab, and then repeating in the second eye. The order of eyes swabbed is inconsequential as both eyes will be swabbed with the same applicator. The physician will then open the sterile capped plastic tube container, place the swab tip down, and carefully snap the swab will holding the tube steady with the other (Fig. 2b). Immediately after capping the plastic tube, they will stretch the parafilm over the top for further leak protection (Fig. 2c) before placing the tube back on the rack.
2) Rectal swab collection:
Study participants will collect their own rectal swabs. They will first collect the swab package with a cotton-tip swab and collection tube labeled “RECTAL,” as well as a biohazard bag to place the collection swab in, from the study coordinator. In a private location such as a restroom, the gloved participant will open the sterile flocked collection device and remove the swab from its wrapper without touching the cotton tip (Fig. 3a). Holding the swab between the thumb and forefinger for the best control, they can raise one foot onto the ledge of the toilet bowl and pull back the buttock with their left hand (Fig. 3b). They will then insert the cotton tip into the anus ½ to 1 inch deep (Fig. 3c) and rotate 360 degrees twice (Fig. 3d) before gently removing. They will then place the soft tip into collection tube, snapping the neck of the swab, firmly closing the lid, and covering with parafilm (Figs. 2b and 2c). After placing the tube into an empty cell of the UNLINE 6 absorbent sheet they should dispose of the gloves and thoroughly wash their hands.
Adverse events
All adverse events will be recorded and reported via the Adverse Event Reporting Form. Serious adverse event (SAE) collection begins after the patient has signed informed consent and has taken the first dose of the study medication. Study personnel must alert the medical monitor of any SAE within 24 hours of investigator awareness of the event. Alerts issued via telephone are to be immediately followed with official notification on the Serious Adverse Event Reporting form. All serious or unexpected Adverse Events judged to be definitely, possibly, or probably related to study participation will be reported to the UCSF Committee on Human Research within 5 working days of their occurrence. Causality of the adverse event will be determined by the study investigators and the Data safety monitoring committee. Participants will be compensated if harmed.
Data collection, management, and monitoring
Demographic data will be collected on enrolled study participants including ocular conditions, age, and gender. Collected sample material will be de-identified and analyzed in aggregate. Collected data will be recorded and managed electronically using REDCap electronic data capture tools hosted at the University of California, San Francisco. All devices used for electronic data collection will be password-protected, as will the mobile application itself.
Study data will only be accessible by study team members and investigators during the duration of the trial to protect confidentiality. Upon completion of the trial, de-identified data will be available upon request.
Study oversight
A Data and Safety Monitoring Committee (DMSC) created before the study begins will provide independent oversight of data quality and participant safety at annual committee meetings. The committee will review the number of patients and any significant adverse events. If necessary, the study analyst will provide unmasked treatment assignment to the committee. The DSMC will recommend modifications to the study protocol as necessary. Major protocol changes will be reported to the DSMC and Institutional Review Board (IRB) as required.
Statistical considerations
The primary analysis for AMR outcomes will be ANOVA to compare determinants of resistance to doxycycline across the three study arms. Power estimates for a difference in the overall microbiome composition between Arms A and B (L1-norm PERMANOVA) are more difficult, but it is worth noting that a similar design found significant differences in the gut microbiome after oral azithromycin.12 Individual organisms will be compared between arms using a Benjamin-Hochberg False Discovery Rate of 5%, with results considered exploratory. Both overall composition and individual species will be compared using stool and lower lid specimens. In addition, comparison of the host immune response in the lower lid between arms will be exploratory, as we do not have previous estimates of the necessary standard deviation. Although, note, this study will allow an estimate of that standard deviation for future studies.
All primary analysis will be intention-to-treat. Patients that do not attend their scheduled follow-up visits will be contacted by the study coordinator the following day via phone and offered rescheduling. Repeated attempts will be made weekly for four weeks to contact patients lost to follow up. Patients who discontinue one of the study treatments due to intolerability or side effects will be observed until the end of the study period if willing to return.
An interim analysis for efficacy is not planned for this trial given a small sample size.