Impact of Two Alternative Dosing Strategies for Trachoma Control in Niger

Population

We will take advantage of the ongoing work in the ten villages currently being studied in Kornaka West. They have never had mass treatment with azithromycin, and the baseline trachoma and infection rates are greater than 20%. The final survey for that current study will occur in January, 2008. Within villages, we will use the updated complete village census lists generated in the January 2008 survey. The children for that survey were randomly selected from the baseline census to provide a sample of approximately 50-60 children in the village ages 0 to five years. They are now a cohort of children ages 2 to 7 years. We propose to add approximately 15 children ages 0-2 from the updated census list for a total of 65-70 children per village.

Sample

We propose to re-randomize the villages, stratified by baseline trachoma rates and former intervention, into treatment intervention (2 rounds of mass treatment) and control (one round of mass treatment) arms. The villages will be balanced by baseline trachoma rates and the original randomization to water and sanitation interventions. Within villages, we propose to use the same sample of children ages 2-7 years and add a random sample of 15 children ages 0 to 2 years.

It will also be important to determine the effect of the two mass drug administration arms on infection in adults, so we propose to randomly select one adult from each household where there is an index child. If the adult is out of the village at the time of the survey, then the next randomly assigned adult will be selected for the study.

Statistical plan including sample size justification and interim data analysis

We intend to analyze the data starting by determining comparability of sample children and adults in intervention and control villages. Village characteristics, household characteristics, and age and gender distributions will be compared by intervention and control status. Importantly, baseline assessment of trachoma, and C. trachomatis infection, will be used to assure comparability. Variables that differ will be used as potentially confounding factors. We will determine the change from baseline to one month and one year in the trachoma prevalence and prevalence of C. trachomatis in the sentinel sample, stratified by children and adults. We will compare the average prevalence in the intervention villages compared to the control villages, by way of preserving the unit of randomization. We will then use logistic regression models to predict trachoma/infection at each time point, adjusting for clustering within villages and other confounding factors. Coverage of mass treatment will also be included as a predictor of trachoma/infection.

We use our sample of children to estimate power, as they are the risk group with highest rates of infection and trachoma. With our sample size of 350 children per group, we have 80% power (at α=0.5) to detect a 15% difference in decline in active trachoma or infection, assuming modest village level clustering.

We will sample 140 subjects per village (70 children plus one randomly selected adult from the same household) for a total of 1400 subjects. Subjects will be sampled at baseline, one-month post-treatment, and at one-year post-treatment.

Prior to the surveys, a training program will take place to accomplish the following objectives for the survey team:

• All persons who will be grading trachoma are standardized against a senior grader, with reliability of kappa=0.65 for TF and for TI, at least. Consistency across graders is essential so that differences are not attributable to grader variations. In any case, all graders must work in all villages, so that the effect of variation by grader does not confound the effect of variation by village.
• All persons who will be taking or assisting with laboratory specimens are trained in proper techniques for taking and storing specimens in the field.
• Proper completion of the survey form, "Examen Oculaire" for each child and adult in the sample, and the completion of the census list on treatment receipt for all persons in the village

The Baseline survey for trachoma in the sample of children and adults will take place prior to any antibiotic intervention. The surveys will consist of the following steps:

1. Prior to the survey in the village, a member of the team will alert the village leadership that the survey team is coming, that mass treatment for all members of the community, as part of the Niger Trachoma Control Program will be part of the survey.
2. The day of the baseline survey, all members of the household will be asked to stay in their concession for the examination and mass treatment, which is done house to house. As they come, the name will be checked on the list of those in the sample survey, and those who are in need of treatment only. If the person in the house is part of the trachoma survey, a form is prepared and a single specimen label filled out for the examiner and the laboratory technician. The label consists of the type of visit (b=baseline, 1=one month, 2=one year) and the full study identification number of the sample person: (village number)-(concession number)-(person number). Thus, during the baseline survey, a person who lives in village number 2, in concession number A-034 and who is on the census list as person 16 would have a study identification number of 02-A-034-16, and a label for the vial of B-02-A-034-16
3. The trachoma grader will be everting the eyelids. Therefore, his fingers are the primary source of contamination for the laboratory specimen. He will change gloves between each exam (or wash his gloves with soap and water) between each child, even children in the same house and even if the child does not appear to have trachoma. This is because about 20% of children without trachoma can still have infection with C. trachomatis (sub-clinical infection). The trachoma grader, wearing 2.5X loupes and using a torche (or in sunlight), will assess the trachoma status of the tarsal plate, using the WHO Simplified grading scheme. The assistant will first evert the right eyelid, grade the tarsal plate, then evert the left eyelid and grade the trachoma status of the tarsal plate. A scribe will record the trachoma assessment on the "Examen Oculaire" form.
4. While the left eyelid is still everted, the laboratory technician, following careful procedures described in the training manual, will roll the swab three times across the tarsal plate to obtain a specimen. The swab must not touch anything other than the tarsal plate. The lab technician can also not touch anything other than the swab and the vial. Once the swab has been taken, it is inserted in an open NUC vial, broken off, and the NUC vial is closed. The sealed vial is labeled with a pre-printed label and placed in a cold box with ice packs while in the field. The scribe will record that the specimen has been taken, or any reason why it was not taken.
5. At this point, the sample person is eligible for azithromycin, which is administered at 20mg/kg. A height stick will be used to estimate dose. For children younger than 2 months, topical tetracycline will be used for 4-6 weeks. The form is reviewed for completeness, and stored safely for eventual data entry. Please note: The first azithromycin treatment in each study arm is administered as part of Niger's trachoma control program; only the second azithromycin treatment is provided to Arm 2 as part of the experimental protocol
6. The concession is then checked to see if all members of the sample have been examined and a swab obtained. If so, the remainder of the concession is treated with azithromycin in accordance with program guidelines.
7. The second treatment team assigned to the village will be providing azithromycin treatment to households who do NOT have anyone in the sample, in order to expedite treatment of the entire village.
8. At the end of each day in the field, all specimens are transferred to a freezer in World Vision in Maradi, awaiting return to the freezer in Niamey. During the drive to Niamey, specimens must be kept frozen as well, with ice packs.
9. The data entry clerk will enter the survey form into the "baseline" database if the survey is the baseline survey, or the "one month" or "one year" database, depending on the follow-up survey. The data entry clerk will also enter the data on those who received treatment at baseline, and at two months into the treatment data base.

The same sample of children and adults will be surveyed for trachoma and infection at one month post the last treatment, and at one year. No additional persons will be added to the sample to replace any who have died or moved away, as we will not have baseline data for any replacements. The procedures for the one month and one year follow up surveys are exactly the same as for the baseline survey, except the following: the laboratory label is changed from a "B" to a "1" or a "2" as the first part of the label, and the survey forms are entered into the one month or one year data bases.

All positive specimens will have the major outer membrane gene amplified and sequenced. The genovars will be mapped for location within villages and families and then their distribution will be followed over time, after treatment to provide a better understanding of the epidemiology of the infection. Results of the study will be used as data input for the generation of mathematical models to predict whether community-wide retreatment (or alternate strategies) will be needed, and the optimal timing for such retreatment.