METHODS: Two real-time PCR methods currently used in Sabah for confirmatory malaria diagnosis and surveillance reporting were evaluated: the QuantiFast™ Multiplex PCR kit (Qiagen, Germany) targeting the P. knowlesi 18S SSU rRNA; and the abTES™ Malaria 5 qPCR II kit (AITbiotech, Singapore), with an undisclosed P. knowlesi gene target. Diagnostic accuracy was evaluated using 52 P. knowlesi, 25 Plasmodium vivax, 21 Plasmodium falciparum, and 10 Plasmodium malariae clinical isolates, and 26 malaria negative controls, and compared against a validated reference nested PCR assay. The limit of detection (LOD) for each PCR method and Plasmodium species was also evaluated.
RESULTS: The sensitivity of the QuantiFast™ and abTES™ assays for detecting P. knowlesi was comparable at 98.1% (95% CI 89.7-100) and 100% (95% CI 93.2-100), respectively. Specificity of the QuantiFast™ and abTES™ for P. knowlesi was high at 98.8% (95% CI 93.4-100) for both assays. The QuantiFast™ assay demonstrated falsely-positive mixed Plasmodium species at low parasitaemias in both the primary and LOD analysis. Diagnostic accuracy of both PCR kits for detecting P. vivax, P. falciparum, and P. malariae was comparable to P. knowlesi. The abTES™ assay demonstrated a lower LOD for P. knowlesi of ≤ 0.125 parasites/µL compared to QuantiFast™ with a LOD of 20 parasites/µL. Hospital microscopy demonstrated a sensitivity of 78.8% (95% CI 65.3-88.9) and specificity of 80.4% (95% CI 67.6-89.8) compared to reference PCR for detecting P. knowlesi.
CONCLUSION: The QuantiFast™ and abTES™ commercial PCR kits performed well for the accurate detection of P. knowlesi infections. Although the QuantiFast™ kit is cheaper, the abTES™ kit demonstrated a lower LOD, supporting its use as a second-line referral-laboratory diagnostic tool in Sabah, Malaysia.
METHODOLOGY/PRINCIPAL FINDINGS: We conducted comprehensive surveys in three areas where P. knowlesi transmission is reported: Limbuak, Pulau Banggi and Matunggung, Kudat, Sabah, Malaysia and Bacungan, Palawan, the Philippines. Infection prevalence was low with parasites detected by PCR in only 0.2% (4/2503) of the population. P. knowlesi PkSERA3 ag1 antibody responses were detected in 7.1% (95% CI: 6.2-8.2%) of the population, compared with 16.1% (14.6-17.7%) and 12.6% (11.2-14.1%) for P. falciparum and P. vivax. Sero-prevalence was low in individuals <10 years old for P. falciparum and P. vivax consistent with decreased transmission of non-zoonotic malaria species. Results indicated marked heterogeneity in transmission intensity between sites and P. knowlesi exposure was associated with agricultural work (OR 1.63; 95% CI 1.07-2.48) and higher levels of forest cover (OR 2.40; 95% CI 1.29-4.46) and clearing (OR 2.14; 95% CI 1.35-3.40) around houses. Spatial patterns of P. knowlesi exposure differed from exposure to non-zoonotic malaria and P. knowlesi exposed individuals were younger on average than individuals exposed to non-zoonotic malaria.
CONCLUSIONS/SIGNIFICANCE: This is the first study to describe serological exposure to P. knowlesi and associated risk factors within endemic communities. Results indicate community-level patterns of infection and exposure differ markedly from demographics of reported cases, with higher levels of exposure among women and children. Further work is needed to understand these variations in risk across a wider population and spatial scale.