Comparison of surveillance methods applied to a situation of low malaria prevalence at rural sites in The Gambia and Guinea Bissau

Abstract

Background: Health record-based observations from several parts of Africa indicate a major decline in malaria, but up-to-date information on parasite prevalence in West-Africa is sparse. This study aims to provide parasite prevalence data from three sites in the Gambia and Guinea Bissau, respectively, and compares the usefulness of PCR, rapid diagnostic tests (RDT), serology and slide-microscopy for surveillance.

Methods: Cross-sectional surveys in 12 villages at three rural sites were carried out in the Gambia and Guinea Bissau in January/February 2008, shortly following the annual transmission season.

Results: A surprisingly low microscopically detectable parasite prevalence was detected in the Gambia (Farafenni: 10.9%, CI95%: 8.7-13.1%; Basse: 9.0%, CI95%: 7.2-10.8%), and Guinea Bissau (Caio: 4%, CI95%: 2.6-5.4%), with low parasite densities (geometric mean: 104 parasites/microl, CI95%: 76-143/microl). In comparison, PCR detected a more than three times higher proportion of parasite carriers, indicating its usefulness to sensitively identify foci where malaria declines, whereas the RDT had very low sensitivity. Estimates of force of infection using age sero-conversion rates were equivalent to an EIR of approximately 1 infectious bite/person/year, significantly less than previous estimates. The sero-prevalence profiles suggest a gradual decline of malaria transmission, confirming their usefulness in providing information on longer term trends of transmission. A greater variability in parasite prevalence among villages within a site than between sites was observed with all methods. The fact that serology equally captured the inter-village variability, indicates that the observed heterogeneity represents a stable pattern.

Conclusion: PCR and serology may be used as complementary tools to survey malaria in areas of declining malaria prevalence such as the Gambia and Guinea Bissau.

Figure 1

Current and previous malaria parasite prevalence data from Basse, Farafenni and Caio. Data from our survey are presented alongside historical data extracted from published reports from A) Basse area, B) Farafenni area or C) Guinea Bissau. Where information on parasite densities was available, percentage of people with 1-4999 and ≥5000 parasites/μl is shown in white and grey blocks, respectively; where parasite densities were not reported, parasitaemia rates are shown as hatched bars. Bars show percentages with 95% CI. Information on age groups surveyed, bed net coverage and the reference are provided in a tabulated format underneath each bar. To enhance comparability our results are displayed for similar age groups as those predominantly reported for this area previously.

Figure 2

Continuous increase of MSP-1₁₉ antibody sero-prevalence with age. A-C) For each site the sero-prevalence of anti-MSP-1₁₉ antibodies is shown according to age. Triangles represent the observed values, the continuous line shows the values predicted by the catalytic conversion model [17] with 95%

Figure 3

Variability of parasite -- and sero-prevalence is greatest within study sites. The prevalence of P. falciparum parasitaemia as assessed by A) slide microscopy, or B) by PCR and the C) sero-prevalence of anti MSP-1₁₉ antibodies measured by ELISA as well as D) the percentage of children aged 1 to 15 years with a palpable spleen is shown for individual villages at each study site (Basse, Farafenni and Caio). Bars show percentages with 95%

Figure 4

Peak of parasite prevalence shifts to adolescence. Parasite prevalence assessed by A) PCR or B) slide microscopy is shown, stratified by age group, showing a significant linear increase from 0-15 years (p < 0.0001 [PCR]; p = 0.003 [slide]). In B) geometric mean parasite densities are added for each age group, showing a significant decline from 0-15 years p < 0.0001), remaining stable thereafter (p = 0.3). Plot C) shows age-stratified parasite prevalence as determined in October 1988 [38] from a village in the Farafenni area. D) Corresponding data from the Farafenni area (villages "D" "L" "S" and "W") from the survey in January 2008 (described in this paper) are shown. Bars show percentages with 95% CI.

Figure 5

Bed net use in under 5 year olds reduces parasite prevalence measured by slide but not by PCR. Age-stratified parasite prevalence assessed by A) slide microscopy or B) PCR, and C) MSP-1₁₉ antibody sero-prevalence for individuals sleeping under bed nets (open bars) or not (hatched bars) are shown. Bars represent percentages with 95% CI.