Sensitive surveillance tools are important for the determination of exposure levels in low-prevalence settings by program managers who are faced with the decision as to whether or not to stop MDA, often in the absence of PCR data. The development of serological tools to detect antibody responses subsequent to trachoma exposure and infection might provide an alternative to clinical exams and PCR analysis for surveillance as an indicator of interruption of transmission. In principle, children born following MDA should experience fewer infections and this should be reflected by lower antibody responses. The absence of an antibody response to a trachoma antigen might indicate an interruption of transmission in formerly endemic areas. Although trachoma programs have not investigated the potential use of antibody tests for program surveillance, the use of serological markers has been shown to be useful in detecting exposure to C. trachomatis
in the context of genital infections 
. In this study, we used serological markers to screen for antibody responses in relation to ocular infections and clinical disease. Two antigens, pgp3 and CT694, were selected for the multiplex assay after an extensive literature search. Pgp3 is the only plasmid-encoded ORF (pORF) secreted into the host cell cytoplasm during infection 
. Pgp3's function remains unknown but it appears to play a role in pathogenesis. In addition, previous authors have suggested its potential use as a diagnostic marker for genital chlamydial infections 
. CT694 is expressed during infection as a T3SS effector 
and has also shown to be recognized by host antibodies 
. We screened bloodspot eluates from Tanzanian children with two chlamydial antigens to measure antibody responses to trachoma antigens after MDA and compared results to clinical exam and PCR analysis data. We have demonstrated that these antibody responses are related to both disease and infection status, suggesting that responses to these chlamydial antigens should be further explored for utility for trachoma surveillance after MDA.
In our assay, we were able to detect and correlate antibody response with clinical and PCR status both at the community and individual level. At the community level, communities with higher trachoma prevalence also had higher pgp3 and CT694 antibody responses, as seen in . Village 0401 had the highest levels of pgp3 and CT694 antibody along with the highest numbers of positive clinical exams. Village 1001 showed lowest prevalence of positivity in all three tests. At individual level, we were able to correlate both clinical signs of trachoma and PCR positivity for Chlamydia
with increased levels of pgp3 and CT694 antibody response. Of children with normal clinical exams, there were many with positive antibody responses. These likely represent children with previous infection 
. Additional support for this conclusion comes from our analysis of age-specific responses. Antibody-positive children with normal ocular exams were typically older than 3 years of age and, thus, potentially infected prior to the beginning of MDA. In contrast, most antibody-positive children (89% of pgp3 positives and 88% of CT694 positives) younger than three years of age were either PCR-positive or had ocular pathology, and antibody prevalence was significantly lower than among older children.
Based on these results, we suggest that antibody-based tools may be valuable for post-MDA surveillance of trachoma. Lack of antibody in young children may be indicative of interruption of transmission and protection by the SAFE strategy, as shown for responses to pgp3 and CT694 in children less than three years of age. Antibody responses may be especially useful as evaluation tools as they represent a cumulative measure of infection, unlike PCR positivity which may be more transient.
Six children (4%) who showed signs of trachoma through clinical examination exhibited no pgp3 or CT694 antibody. This result may have occurred due to imprecision of clinical diagnosis, in which follicle formation was due not to trachoma but to other causes, such as allergic conjunctivitis. Alternatively, there may be some genetic differences in antibody response at the individual level. Finally, as Ghaem-Maghami et al suggest, there may be some suppression of antibody response in severe cases of inflammation 
; if this is so, it was rare in our series. Additional studies are needed to define the kinetics of antibody response following infection and how antibody responses shift following repeated infections.
The US samples are unidentified, and, without knowing the child's country of origin and medical and travel history, we do not know their true disease status. There is a possibility that chlamydial infection was acquired at birth 
. The responses of Haitian children were not as easy to interpret. Eight of 86 children were positive for at least one antigen, and three of these children were positive for both. Even though Haiti is considered non-endemic for trachoma, it is difficult to determine from this sample group if antibody reactivity was due to chlamydial infection acquired at birth (not trachoma) or evidence of acquired chlamydia after birth. It may be that these responses reflect cross reactivity to other antigens. From our own analysis, we have demonstrated that the antibody responses to pgp3 and CT694 do not cross-react.
The multiplex bead assay may be a valuable antibody-based tool for trachoma surveillance. Confirmatory data can be generated by using multiple antigens within the same sample. In theory, antigens might also be selected to distinguish between current infections and exposure or to differentiate infections caused by various serovars. Along with trachoma-specific antigens, antigens for monitoring and evaluation from other NTD public health programs may be included in the multiplex assay. Community profiling can be accomplished through a panel of antigens representing a multitude of diseases, such as helminths, viruses, and waterborne and vaccine-preventable diseases, which can be screened from a single bloodspot or microliter of serum, thus greatly increasing cost-effectiveness of surveillance activities. Using a panel of antigens also facilitates measurement of the impact of MDA on multiple diseases, which might not otherwise be tracked through separate monitoring programs.
There are several limitations to consider in the context of this study. First, all villages have been treated by MDA prior to sample collection, so the baseline antibody responses are unknown for each antigen. Also, because samples were collected at only one time point, our understanding of how each child's antibody and disease status changed over time is limited. These points make it difficult to describe the kinetics and longevity of antibody responses with the antigens tested and, more specifically, whether antibody responses reflect current or previous infection. A longitudinal study, including a baseline collection is an important next step to confirm the value of the antibody tools described here. Because antigens were chosen based on literature specific to genital chlamydial infections 
, we cannot be sure that these antigen choices are the most appropriate in terms of ocular infection. For example, cases of anti-chlamydial responses but no ocular pathology may be due to a non-ocular strain of Chlamydia
. The use of serovar-specific antigens, such as the major outer membrane protein peptides, might distinguish between genital and ocular infections and better characterize the nature of the infecting bacteria.
In summary, antibody-based assays, in particular the multiplex assay, could be valuable tools to evaluate the impact of MDA programs and for post-MDA surveillance for trachoma.