Trachoma, caused by repeated ocular infection with Chlamydia trachomatis (Ct), has been targeted by the World Health Organization (WHO) for elimination as a public health problem by the year 2020. As part of the strategy to achieve elimination, WHO recommends annual mass drug administration (MDA) of azithromycin in endemic districts [1]. Program targets related to MDA focus on the district-level prevalence of trachomatous inflammation—follicular (TF) amongst children aged 1–9 years. To monitor progress towards elimination, population-based impact surveys are recommended to evaluate whether a district has reached the threshold of less than 5% TF prevalence in 1–9-year-olds and can cease azithromycin distribution. Two years after cessation of MDA, a surveillance survey to ensure that district-wide TF prevalence in 1–9-year-olds remains below 5% is conducted prior to the validation process. Currently, there are no guidelines for post-validation surveillance.

These surveys rely on a clinical grading scheme that is relatively inexpensive and simple to perform, but is poorly correlated with ocular Ct infection in low-prevalence settings [2]. Following MDA, the clinical sign trachomatous inflammation—intense (TI) has been shown to correlate better with infection than TF does [3]. However the measurement of clinical signs is subject to inter-grader variability and lack of real-time auditing since grading is performed in the field and thus can only later be validated or audited if images are taken. As trachoma elimination programs stand to benefit from an accurate, reproducible assessment of trachoma prevalence, other testing methods may be useful to help guide program decisions. These include tests of infection (polymerase chain reaction [PCR] testing of ocular swabs) and antibody-based testing [4–7].

Antibodies to Ct antigens may act as markers of cumulative exposure to Ct. Two previously described Ct antigens, Pgp3 and CT694, have been shown to be reactive against sera in young children living in trachoma-endemic communities [4,7,8]. At the individual level, antibodies to these proteins demonstrate high sensitivity to ocular infection and high specificity against non-endemic control specimens [8–10]. However, individual associations may not always hold at the community level, and trachoma elimination programs treat ocular Ct infection on a population level. Additionally, as antibody markers are not yet widely used to assess for Ct prevalence, better characterization of how seropositivity compares to other methods of assessing trachoma prevalence is necessary. Here, we evaluate the association between seropositivity, PCR positivity, and clinical signs of active trachoma (TF and TI) at the individual and community level in a region of Niger where some trachoma transmission is occurring (TF prevalence approximately 25% at baseline). Data were collected during the final follow-up visit of the Partnership for the Rapid Elimination of Trachoma (PRET)-Niger trial, in which communities were randomized to receive annual or biannual oral azithromycin for 3 years in order to assess the impact of treatment frequency on ocular chlamydia infection [11].

At the 36-month follow-up, 988 1–5-year-old children from 24 communities had clinical trachoma and serology data available. TF prevalence was 7.8% (95% CI 6.1 to 9.5%) and TI prevalence was 1.6% (95% CI 0.9 to 2.6%) (Table 1). The overall prevalence of antibody positivity to Pgp3 was 27.2% (95% CI 24.5 to 30%), and to CT694 was 23.7% (95% CI 21 to 26.2%). Community prevalence of ocular chlamydia in 0-5-year-olds was 5.2% (95% CI 2.8 to 7.6%).

The presence of antibodies to Ct antigens was correlated with both TF and ocular Ct infection following a 36-month annual or biannual mass azithromycin distribution program in a trachoma-endemic region of Niger. However, there was no difference in serologic outcomes by study arm, consistent with clinical data suggesting that biannual treatment did not significantly alter transmission of ocular chlamydia compared to annual treatment [11]. That serologic outcomes were consistent with other trachoma indicators supports the finding that antibodies to Ct antigens are correlated with TF and ocular Ct infection and provide complementary information. Future work evaluating serologic outcomes in a trial with a significant effect on ocular Ct infection or TF would provide additional evidence about the relationship between these indicators.

Currently, decision-making for MDA in trachoma elimination programs relies solely on clinical grading of TF. Grading of clinical trachoma is subjective, and prevalence surveys have demonstrated that there is poor agreement between clinical disease and ocular Ct infection, particularly after multiple rounds of antibiotic treatment [2,3,14,16]. Additionally, TF may be observed in the absence of infection, either as residual inflammation from the etiologic agent Ct[17,18], or due to other non-chlamydial bacteria [19]. Point prevalence of TF, or a test of infection, reflects disease or infection state at a current point in time, whereas serologic patterns may allow for identification of longer-term patterns in Ct transmission [7,8]. Anti-Ct antibody responses increased with age in this study, whereas TF prevalence was not significantly different across age groups, suggesting that antibody positivity rates represent the pool of exposed individuals rather than currently or recently-infected ones.

While PCR assessment of ocular Ct infection and clinical grading for TF and TI represent cross-sectional prevalence of trachoma, age-seroprevalence curves may provide additional insight to changes in transmission of ocular Ct over time. PCR or NAAT testing has historically been too costly for use in program settings[20] but cost-effective PCR tests are now being evaluated in program contexts [6]. In ocular swab specimens, PCR tests for Ct infection are a more specific indicator for the causative agent of trachoma than antibody testing in sera, as antibody responses to the antigens we studied cannot differentiate between exposure to ocular Ct and other chlamydial infections. Perinatal transmission of Ct from mothers with urogenital chlamydia could potentially lead to seropositivity among young children, as could sexual exposure in individuals after the age of sexual debut. However, focusing on the younger children and on the age-seroprevalence curve rather than absolute rates of seropositivity may allow for distinction between ongoing ocular Ct transmission and a single exposure. For example, in the Solomon Islands, a lack of increase in seropositivity to Pgp3 in 1–9-year-olds correlated with low infection rates, despite the 26% prevalence of TF, contrasting to the steep increase in seropositivity with age observed amongst 1–9-year-olds in Kiribati where TF prevalence was 28% and infection prevalence was 24% [21,22].

Antibody testing for chlamydial antigens has been conducted in a number of trachoma program settings. In treatment-naïve communities, the slope of the age seroprevalence curve increased with increasing community TF prevalence [4,21]. A significant decline in antibody responses has been shown following mass azithromycin distribution compared to pre-treatment levels in a cross-sectional study [8]. In this mesoendemic region of Niger, we noted more than 30% prevalence of antibodies to Pgp3 and CT694, consistent with results from mesoendemic communities in Tanzania [4]. In settings where surveillance surveys for trachoma elimination have been conducted, age seroprevalence curves corresponded to decreases in trachoma transmission [23]. Seroprevalence of antibodies to Pgp3 in children ages 1 to 9 years in these surveys ranged from less than 2% in some surveys[24,25] to as high as 7.5%[26] but without a steep increase in age seroprevalence curves seen in settings with ongoing transmission. The current data add to this body of knowledge by evaluating antibody responses at impact surveys after 3 rounds of MDA, a program setting for which limited data exist.

TF prevalence in this study was 7.5% after 3 years of MDA and thus antibody data do not come from a setting in which elimination thresholds have been achieved. Furthermore, the communities included in this analysis were treated with an enhanced coverage target, with up to four days of treatment. Antibiotic coverage may have been higher than is seen in programmatic or trial settings with lower coverage targets. The antibody responses and curves therefore may not be representative of what would be seen in a previously-endemic setting that has reached the elimination threshold. Trachoma programs typically include children up to age 9 in monitoring. Other studies have shown further increases in antibody responses in children aged 6–9 years[21,26], and inclusion of those ages here may have improved our ability to draw inferences from the shape of the curve. Since ocular swabs were pooled for PCR analysis, we were unable to obtain individual-level correlations between PCR and antibody positivity.

WHO currently recommends use of TF for deciding programmatic endpoints. Laboratory testing, including Ct serology, could be used as a supplement or replacement for TF when conducting surveillance after validation of the elimination of trachoma as a public health problem, given serology is generally inexpensive, objective, and provides estimates of exposure over time.[26] The elimination thresholds do not require a complete absence of ocular Ct infection, and therefore infection may still be present in communities that have reached the district-wide elimination threshold, as was seen in Tanzania where infection was present but did not lead to re-emergence.[27] Having a test of exposure, or of repeated infection, would allow more complete evaluation of the history of exposure to ocular Ct in children.

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.