Comparison of Commercial ELISA Kits to Confirm the Absence of Transmission in Malaria Elimination Settings

Abstract

Background: Antimalarial antibody measurements are useful because they reflect historical and recent exposure to malaria. As such, they may provide additional information to assess ongoing transmission in low endemic or pre-elimination settings where cases are rare. In addition, the absence of antibody responses in certain individuals can indicate the cessation of transmission. Commercial malaria enzyme-linked immunosorbent assays (ELISA) detect antimalarial antibodies and are commonly used to screen blood donations for possible malaria infection. However, there is no standardized test to detect antimalarial antibodies for epidemiological use. Here we compared five commercially available ELISA kits (Trinity Biotech, newbio, DiaPro, Cellabs, and NovaTec) in search of a standardized tool for supporting claims of absence of malaria transmission. For comparison, a research-based (RB) ELISA protocol was performed alongside the commercial kits. Results: The commercial kits were first compared using serum samples from known malaria-unexposed individuals (n = 223) and Toxoplasma-infected individuals (n = 191) to assess specificity and cross-reactivity against non-malaria infections. In addition, 134 samples from ≥10-year-olds collected in a hyperendemic region in the Gambia in the early 1990s were used to assess sensitivity. Three out of five kits showed high sensitivity (90-92%), high specificity (98-99%), low cross-reactivity (0-3%) and were considered user-friendly (Trinity Biotech, newbio and NovaTec). Two of these kits (Trinity Biotech and NovaTec) were taken forward for epidemiological evaluation and results were compared to those using the RB-ELISA. Samples from two pre-elimination settings (Praia, Cape Verde; n = 1,396, and Bataan, the Philippines; n = 1,824) were tested. Serological results from both the Trinity Biotech kit and the RB-ELISA concurred with recent passively detected case counts in both settings. Results from the Trinity Biotech kit reflected a significant decrease in the number of reported cases in Bataan in the 1990s better than the RB-ELISA. Results from the NovaTec kit did not reflect transmission patterns in either setting. Conclusion: The Trinity Biotech commercial ELISA kit was considered reliable for epidemiological use and accurately described transmission patterns in two (previously) malaria endemic settings. The use of this simple and standardized serological tool may aid national control and elimination programs by confirming that regions are free from malaria.

Keywords: ELISA; IgG; antibody; commercial ELISA kits; elimination; immunoglobulin; malaria; pre-elimination.

Figure 1

Map of survey locations in Bataan, the Philippines (A) and Praia, Cape Verde (C). The square highlighting Cape Verde in (B) is enlarged in (C). The Philippines (A) and Cape Verde (B,C) are shown in green, study areas are in red and surrounding countries are in gray.

Figure 2

Reported malaria cases (a,b), seroconversion curves (c,d) and seroprevalence by self-reported history of malaria (e,f) using antibody responses recorded by commercial and the research-based (RB) enzyme-linked immunosorbent assays. In (a,b) counts of reported malaria cases at local health facilities are shown over time; in Bataan, data was available for 0 to 12 years prior to data collection (i.e., 2017 – 2005) (28) and 21–35 years prior to data collection (i.e., 1996 – 1982) (26). In Praia, data was available from 1 to 21 years prior to data collection (i.e., 2016 – 1996) (30) and 22 to 31 years prior to data collection (i.e., 1995 – 1986) (29). In (c,d) seroconversion curves of age-specific seroprevalence are shown; solid lines represent the fit of the reversible catalytic model (2), while dashed lines represent 95% confidence intervals (CIs). Seroconversion rate and change point estimates with 95% CIs are shown on plots. In (e,f) seroprevalence estimates and 95% CIs are shown by self-reported history of malaria. Results for commercial kits using a two-Gaussian mixture model for seropositivity thresholds are shown in Supplementary Figure 4.

Figure 3

Venn diagram showing the intersection of seropositivity recorded by commercial and the research-based (RB) enzyme-linked immunosorbent assays in those born since the decline in passively detected malaria cases in Bataan, the Philippines. Seropositivity is shown for those younger than 22 years corresponding to the marked decline in passively detected cases in 1995, see Figure 2a. Ages are shown on the plot for individuals who were seropositive in multiple assays.