However, the use of azithromycin in children aged <6 months is not approved in many trachoma-endemic countries. Instead, parents of infants are provided with tubes of 1% topical tetracycline to be administered one to two times a day for 4 to 6 weeks. Research suggests that adherence to this protracted regimen of topical tetracycline is less than with a single dose of azithromycin.² In a study of a single village, some of the highest loads of infection were found in a few infants before mass treatment.³ Thus, there is concern that these infants may be a source of re-emergent infection in trachoma-endemic communities after mass treatment.

Conversely, if infants are largely cared for by mothers they may have less exposure to infection from other children, and less opportunity to transmit infection, and thus they may not represent a major source of re-emergent infection. If so, the risk of significant re-emergent infection from the few infants that were infected post MDA could be very low. There are few data on infection and disease in this very young age group of infants, and we sought to investigate the role of infants as a source of re-emergent infection in trachoma-endemic communities after mass treatment.

In the context of a longitudinal cohort study of all children aged under 10 years in four communities undergoing mass treatment in Tanzania, we evaluated the impact of residing in a household with an infant aged <6 months on the risk of re-emergent infection after mass treatment. We hypothesized that children residing in households with infants aged <6 months at baseline would have higher rates of infection with C. trachomatis 6 months post MDA compared with children residing in households with no infants.

Detailed study methods on the ocular examination and specimen analysis are described elsewhere.⁴ In summary, trained trachoma graders performed an ocular examination using 2.5× loupes and a torch to determine active trachoma. The trachoma graders were standardized during a 2 week training exercise before the start of the study. Interobserver agreement, and agreement with a senior grader had to be above kappa = 0.6 for TF and inflammatory trachoma (TI). Graders are monitored every 6 months by comparison of grades against the senior grader using a series of trachoma images.⁴ Active trachoma was graded using the World Health Organization simple grading scheme.⁵

An eyelid-conjunctival swab was obtained to assess infection with C. trachomatis. These swabs were taken using strict protocols to avoid contamination. Both the trachoma grader and eyelid flipper used gloves and changed them between each child examined. In addition, “air control” swabs were taken on a random sample of 5% of children. The swabs were waved in the air above the child but do not touch anything, marked as per any other sample, and sent for processing. If positive, further investigations were carried out to determine the source of contamination. The samples were kept at −20°C until they were sent to the International Chlamydia Laboratory at Johns Hopkins University for processing. The samples were analyzed using polymerase chain reaction (Amplicor; Roche Molecular Systems, Indianapolis, IN) according to manufacturer's directions. Optical density was used to identify positive results and defined as >0.7. Optical densities between 0.2 and 0.7 were considered equivocal and were retested. Results of these “air control” swabs showed some evidence of laboratory contamination that affected 34 samples at baseline. These contaminated samples were excluded; all further tests for contamination were negative at baseline and follow-up.

At the time of MDA, parents of infants aged <6 months were provided with tubes of 1% topical tetracycline with instruction to administer 1 to 2 times a day for 6 weeks. Children aged greater than 6 months were given a single dose of azithromycin (20 mg/kg up to 1 g). Community treatment assistants directly observed treatment with azithromycin; however, compliance with topical tetracycline was not monitored. Children and infants were re-evaluated at 6 months for the presence of active trachoma and C. trachomatis.

The association between infection in children at 6 months and presence of infants in the house was examined using a logistic model with infection as the outcome and presence of infants in the house as the predictor; the model accounted for the clustering of infection within children residents of the same household. An expanded model of infection was used to examine the independent contribution of age, sex, baseline chlamydial infection, and treatment at baseline, as well as presence of infants in the household. Odds ratios (ORs) and 95% confidence intervals (CIs) are presented where the generalized estimating equation (GEE) approach was used to correct the standard errors to account for the correlation among members of the same house, using the procedure GENMOD in SAS (Cary, NC) with binomial distribution, logit link function, and exchangeable correlation structure.

At baseline, the prevalence of chlamydia infection in infants was 5.9% and 1.2% had follicular trachoma (Table 1). The prevalence of infection and active trachoma was lower in infants than in older children. The prevalence of TF in those less than 3 months of age was 1.8% (1/55) compared with 0% (0/31) in those 3 to 6 months (Fisher's exact test P = 1.0).

With increasing age, the likelihood of signs of trachoma lingering after infection was gone also increased (Table 2). In children aged <1 year, clinical signs were highly associated with infection.

Of the 91 infants in the census at baseline, seven were not seen during MDA and thus were not provided with topical tetracycline. Of those seven, six infants were found to be not infected at baseline and one had an unknown infection status. Mass drug administation (MDA) coverage of children older than 6 months with azithromycin was overall very high and averaged 94%.

Six months post mass treatment, 82 of the 91 infants were examined and two (1.2%) were found to be infected (both were negative for infection at baseline). None of the infants with infection at baseline had infection at 6 months (data not shown). The prevalence of infection among children aged <10 years who resided in households with infants was 6.0% compared with 11.1% for children who resided in households without infants (P = 0.18; Table 3). In fact, contrary to our hypothesis, when controlling for age, sex, treatment status, and baseline infection status, children residing in a household with an infant had a reduced, but not statistically significant, risk of infection at 6 months [odds ratio [95% confidence interval] 0.50 [0.20–1.22]; Table 4).

It is possible that compliance with topical tetracycline in the infants was better than expected. We did not monitor compliance with topical treatment to confirm compliance, which we did for azithromycin. Coverage with azithromycin was very high, and was likely instrumental in preventing cross infection of the infants up to 6 months. Other studies have reported topical tetracycline to be equally effective as oral azithromycin in decreasing infection when compliance is high.⁶

We have shown that treatment with azithromycin decreases the load of infection, but if the load of infection pretreatment is high, eliminating infection with a single dose is less likely.⁷ We did not measure the load of infection in these infants, but we would not expect the load of infection pre-MDA to be lower than observed elsewhere. Thus, if the infants had high organism loads, as reported in other studies,³ and topical tetracycline was not as effective in such cases, the infants would be a risk to other children in the households. Our finding of no increased risk suggests that topical tetracycline may have been adequate treatment and/or infants are not a significant source of infection for older children in the family.

Only one infant <6 months old presented at baseline with follicular trachoma. The infection rate was higher, 6%, suggesting most children in this age group still cannot mount a clinical response to infection. Few children aged <1 year had clinical signs of trachoma that were not associated with infection, as contrasted with older children whose follicular response tends to linger. The fact that 6% of infants <6 months old had infection indicates how early in life exposure to C. trachomatis may occur. Consequently, researchers working on a chlamydia vaccine ought to give consideration toward the provision of protection at an early age to those in trachoma-endemic communities.