This study presents population seroprevalence estimates of selected NTDs, VPDs, and arboviruses in Samoa in 2018. Key findings were high seroprevalence estimates for LF and arboviruses, particularly among older participants, and low seroprevalence to VPDs among participants from purposively selected PSUs. The highest degree of clustering of seropositivity for all Abs (i.e., immunity from infection or vaccination) was seen at the household level. Lastly, strong correlation was seen both within and between disease groups, particularly between Ab against vector-borne diseases, suggesting the potential for integrated public health prevention, control, and elimination activities, particularly integrated vector surveillance and control.
Of the NTDs investigated, the highest population-level seroprevalence was for the LF Abs. The WHO has targeted LF for global elimination as a public health problem by 203014, though several countries including Samoa remain endemic despite multiple MDA rounds over decades15. We have previously described the epidemiology of LF Ab from SaMELFS 2018 in detail9. Trachoma, the leading cause of infectious blindness globally, has also been targeted for elimination as a public health problem by 203014. Samoa is not considered endemic for trachoma16, and though some cases were identified through community surveys in the 1940s, 50s, and 80s17, no trachoma was identified in community screening in 200218 or reported from outreach clinics between 2011 and 201319. IgG Ab responses to Ct694 and Pgp3 are sensitive and specific markers of trachoma with alignment seen between Ab responses, PCR and active trachoma20. In this study, trachoma was defined as seropositivity to both Ct694 Ab and Pgp3 Ab and analyses were restricted to 5-9-year-olds due to difficulty in using Ab responses to different trachoma from chlamydia infection. We found that 5.5% of 5-9-year-olds were seropositive for trachoma, with significant associations between seropositivity and residing in ROU and SAV. The possibility of undetected trachoma transmission is concerning and warrants further investigation in Samoa.
Similarly, yaws, primarily a chronic childhood NTD that affects skin, cartilage, and bone21, is targeted for eradication by 203014. However, as of 2023, only India has successfully eliminated yaws21. Samoa was previously endemic for yaws, but its current status is unknown21. Seroprevalence of Rp17 Ab is believed to represent historical infection with yaws, whilst TmpA Ab represent recent infection5. Among 5-9-year-olds, we found Rp17 Ab and TmpA Ab seroprevalence of 0.2% and 1.0%, respectively. These low numbers suggest that there is no, or very limited, active yaws transmission in Samoa. However, as yaws is targeted for eradication, any possible pockets of transmission should be further investigated.
Over 90% of the study population were seropositive for at least one dengue serotype, > 85% for Zika, and 57% for chikungunya, confirming high circulation of arboviruses in Samoa. Large outbreaks of dengue, chikungunya, and Zika are common in the Pacific22,23. With the increase in arboviral outbreaks globally, there is a need for strengthened surveillance systems to guide resource allocation and control efforts24. However, data on arbovirus outbreaks in the Pacific tend to be based on syndromic surveillance due to limited lab capacity23. Utilising MBA to determine population-level arboviral infection/immunity has the potential to provide a more complete and comprehensive picture of circulating arboviruses, target hotpots of transmission for vector control, and inform decisions about the use of new dengue vaccines, which may only be recommended for populations with high seroprevalence25.
Regular VPD serosurveys can estimate population-level immunity levels, identify immunisation coverage gaps and/or waning immunity for top-up campaigns, and potentially serve as an early warning system for epidemic-prone diseases. MBAs provide a further advantage of determining immunity levels to several VPDs simultaneously. Measles is highly infectious and vaccine coverage of at least 95% is required to maintain herd immunity26. As of 2022, elimination of measles has been verified in eight countries in the WHO Western Pacific Region (WPRO)27; however, between 2022 and 23, measles cases in WPRO increased by 255% due to gaps in vaccination coverage and disease surveillance, and travel to/from endemic areas28. We found concerningly low measles seroprevalence (43.6%), reflecting the low reported vaccination rates in Samoa around the time of our survey29. These findings are particularly pertinent given the 2019 measles outbreak in Samoa that resulted in > 5,700 cases and 83 deaths (87% of deaths in children < 5 years old)29. We have since conducted serosurveys in Samoa in 2023 and 2024; the measles seroprevalence results (pending) will be important to determine whether further vaccination campaigns are required in Samoa.
To date, seven countries in WPRO have eliminated rubella27. The goal of rubella vaccination campaigns are to prevent congenital rubella syndrome (CRS), with one dose of the vaccine considered sufficient for lifelong immunity; however, several WPRO countries have reported CRS suggesting immunity gaps among women of reproductive age27. Recent data has shown higher proportions of rubella cases in individuals > 14 years in WPRO compared to global data27. We found rubella seroprevalence approaching 80% in this study, though significantly lower odds of seropositivity among 10-19-year-olds vs. 5-9-year-olds. This could be indicative of low vaccination coverage in this age group or waning rubella-specific immunity in this age group. Interestingly, waning rubella seropositivity has been reported in reproductive-aged women and there are recommendations to administer of a third dose of measles-mumps-rubella vaccine during pre-pregnancy planning30. Given the severity of CRS, further monitoring of rubella seroprevalence among reproductive-aged women in Samoa would be prudent.
We found much higher seroprevalence to rubella than measles, and no association between their seroprevalence at the PSU level despite the routine co-administration of measles and rubella vaccines (e.g., in the measles-mumps-rubella or measles-rubella vaccines). A likely explanation of the higher rubella seroprevalence is local circulation of the virus, resulting in higher levels of immunity to rubella than measles. Acute rubella infection is generally a milder disease than measles and may not be diagnosed or reported. Surveillance of CRS is therefore important even if few acute cases of rubella are reported.
In contrast, tetanus and diphtheria seroprevalence were 91.0% and 83.5%, respectively, with a moderate positive association seen at the PSU level. In 2018, coverage of the first dose of diphtheria, tetanus toxoid and pertussis-containing vaccine (DTP)-1 and the third DTP does (DTP-3) in Samoa had fallen to 67% and 44%31, respectively, likely reflective of general vaccine hesitancy at the time. Encouragingly, seroprevalence estimates from our study show high population coverage, though this should be monitored closely.
Of the 12 major diseases transmitted by mosquitoes, nine are NTDs32. The WHO Global Vector Control Response Strategy 2017–2030 recommends an integrated vector management approach with the viewpoint that several NTDs can be controlled the same time32. We found significantly higher seroprevalence to LF and DENV-1 in purposively selected PSUs (suspected LF hotspots), suggesting that LF hotspots could align with areas of high arbovirus transmission. We also found correlations between seroprevalence of LF, chikungunya, DENV-1, and Zika. Given the shared epidemiology of LF and arboviruses, with these diseases often sharing a mosquito vector, further investment in integrated vector management and control would be recommend. As the highest ICC values at the household level were seen for LF, dengue, and Zika, encouraging vector management and control at the household level could be an effective means to reduce transmission and of these diseases.
WHO has defined minimum MDA coverage thresholds for LF of at least 65% of the total population14. However, MDA might not be taken by individuals for several reasons including intentional factors (e.g. failure to ingest, fear of side-effects) or unintentional factors (e.g. away at the time, not offered medications, ineligible [due to comorbidities, pregnancy]) resulting in “never treated” individuals33, among whom LF seroprevalence is higher than among those who have taken at least one dose of MDA34. It is conceivable that households with never treated individuals are also more likely to have zero-dosed (unvaccinated) children if factors such as poor access to/by public health programmes, hesitancy to participate, or poor trust in public health programmes exist in these communities. If individuals in purposively selected PSUs (suspected LF hotspots) had lower MDA uptake, we could expect to see lower vaccine coverage in these communities. Indeed, Willis et al.. reported lower coverage of MDA in Samoa in 2018 in purposively vs. randomly sampled PSUs (77% vs. 81%), though this difference was not statistically significant35. We found significantly lower odds of VPD seropositivity in purposively selected PSUs and negative correlations between measles and LF seroprevalence. Interventions directed at these communities to increase MDA and vaccination uptake alongside vector control could be beneficial.
This is the first population-representative serosurvey that has been conducted in Samoa. A particular strength of this study is that, though this survey was designed to estimate LF prevalence, we have shown that collecting DBS alongside routine LF surveillance has enabled the integration of LF monitoring with other NTDs, VPDs, and arboviruses. Nonetheless, there are some key limitations that should be borne in mind with the interpretation of these results. Serological surveys cannot distinguish between immunity from vaccination or natural infection, limiting the utility of serosurveillance to identify areas of low/uncertain vaccination coverage. Further, laboratory assays used for serosurveillance can vary in specificity and sensitivity due to variability in individual IgG responses36 and the risk of cross-reactivity with Ab from other pathogens5. Possible cross-reactions between dengue virus and other arboviruses (chikungunya and Zika) have been reported in the literature37,38; therefore, it is possible that such cross-reactions in our sample may have caused overestimation of arbovirus seroprevalence. Additionally, interpretation challenges such as non-specific positivity and uncertain cut-off values may impact seroprevalence estimates. These limitations can affect the accuracy of serosurvey results, potentially resulting in misclassification of serostatus. This is a particularly concerning in low-transmission settings, where false positives/negatives can substantially influence prevalence estimates. Ongoing research is required to identify sensitive and specific recombinant Ag that can be used to confidently indicate seropositivity.
