Age-specific malaria seroprevalence rates: a cross-sectional analysis of malaria transmission in the Ouest and Sud-Est departments of Haiti

Abstract

Background: Malaria transmission continues to occur in Haiti, with 25,423 confirmed cases of Plasmodium falciparum and 161,236 suspected infections reported in 2012. At low prevalence levels, passive surveillance measures, which rely primarily on reports from health systems, becomes less appropriate for capturing annual malaria incidence. To improve understanding of malaria transmission in Haiti, participants from the Ouest and Sud-Est departments were screened using a highly sensitive enzyme-linked immunosorbent assay (ELISA).

Methods: Between February and May 2013, samples were collected from four different sites including a rural community, two schools, and a clinic located in the Ouest and Sud-Est departments of Haiti. A total of 815 serum samples were screened for malaria antibodies using an indirect ELISA coated with vaccine candidates apical membrane antigen (AMA-1) and merozoite surface protein-1 (MSP-119). The classification of previous exposure was established by using a threshold value that fell three standard deviations above the mean absorbance for suspected seronegative population members (OD of 0.32 and 0.26 for AMA-1 and MSP-1, respectively). The observed seroprevalence values were used to fit a modified reverse catalytic model to yield estimates of seroconversion rates.

Results: Of the samples screened, 172 of 815 (21.1%) were AMA-1 positive, 179 of 759 (23.6%) were MSP-119 positive, and 247 of 815 (30.3%) were positive for either AMA-1 or MSP-1; indicating rates of previous infections between 21.1% and 30.3%. Not surprisingly, age was highly associated with the likelihood of previous infection (p-value <0.001). After stratification by age, the estimated seroconversion rate indicated that the annual malaria transmission in the Ouest and Sud-Est department is approximately 2.5% (95% CI SCR: 2.2%, 2.8%).

Conclusions: These findings suggest that despite the absence of sustained malaria control efforts in Haiti, transmission has remained relatively low over multiple decades. Elimination in Haiti appears to be feasible; however, surveillance must continue to be strengthened in order to respond to areas with high transmission and measure the impact of future interventions.

Figure 1

Map of Haiti with the location of the study enrollment sites. The four study enrollment sites located inside the Ouest and Sud-Est Departments of Haiti in the communes of Gressier and Jacmel. Participants were enrolled from Christianville School in Gressier and from Hosana Baptist School and Portail Leogane Clinic, and the rural community of Chabin in Jacmel. Along with an inset of the entire country of Haiti, the enrollment sites (red circles) appear relative to the study commune (yellow), the national capital (star) and national highway systems (pink lines).

Figure 2

Distribution of ELISA responses in absorbance units (λ = 450 nm). The distribution of ELISA responses from the study participants in absorbance units (at 450 nm) appear for AMA-1 and MSP-1 on the left and right panels of Figure 2, respectively. The upper panels show the histograms of the suspected seronegative ELISA results overlaid with a normal distribution function. The sample mean (thick black line) and sample standard deviation of these functions were used to determine minimum absorbance values (thresholds) for the classification of a sample as seropositive using the sample mean for the suspected negative population members plus three to five sample standard deviations (gold, orange, and red dashed lines).

Figure 3

Seroprevalence by age class for participants ranging from 2 to 80 years of age as determined by ELISA using AMA-1 and MSP-1 antigens. The prevalence of samples that had an response to the AMA-1 or MSP-1 antigens are presented by age class for those classified as being seropositive with AMA-1 (white), MSP-1 (charcoal), or either (black dotted) antibodies.

Figure 4

Comparison of samples with positive AMA-1 or MSP-1 responses. The scatterplot of AMA-1 and MSP-1 responses (average absorbance at 450 nm) is shown along with the reference lines denoting positive AMA-1 or MSP-1 responses (dotted lines) using the threshold absorbance values for previous infection of 0.319 and 0.260 absorbance units, respectively. Samples were positive for AMA-1 and MSP-1 (region I), MSP-1 positive and AMA-1 negative (region II), AMA-1 positive and MSP-1 negative (region III), or negative for both AMA-1 and MSP-1 (region IV). For clarity, samples with an absorbance value above 1.8 AU are shown as having a maximum value of 1.8 AU and those with no positive response to either antigen (region IV) are not shown.

Figure 5

Seroprevalence estimates for the presence of AMA-1 or MSP-1 antibodies by age for the entire study population. The actual seroprevalence for AMA-1 and MSP-1 antibodies (circles) appear by age along with the probability of infection in each age class for AMA-1 and MSP-1 (black lines) and the respective 95% confidence limits (dotted lines), derived from the model estimated seroconversion rate (λ). The top, middle, and lower panels show the incremental increases in seroprevalence for AMA-1, MSP-1, and either AMA-1 or MSP-1 with age, respectively. The left panels show the model fits using data from participants 2 years to 20 years of age, while the right panels show the model fits using the entire data set including participants from 2 to 80 years.