Performance of Rapid Diagnostic Testing in Patients with Suspected Malaria in Cambodia, a Low-Endemicity Country Aiming for Malaria Elimination

This is the authors' abstract. We don't add key points for this paper.

To the Editor—We read with interest the article by Ranadive et al [1] assessing the performance of malaria rapid diagnostic testing (RDT) vs polymerase chain reaction (PCR) in Swaziland, a low-transmission country aiming at elimination. Through a large regional data set collected from 37 health facilities over 2 years, they demonstrated the poor sensitivity of RDT (First Response Malaria Ag P. falciparum HRP-2 Detection Rapid Card Test, Premier Medical) for Plasmodium falciparum (Pf) diagnosis (51.7%), due to a high proportion of low-density infections among symptomatic subjects (54/162 [33.3%]), along with a low positive predictive value (PPV) (67.3% for all samples and 62.3% for ≥100 parasites/µL samples), due to the high proportion of false positivity (32.4%). To overcome some of the limitations of the study (eg, the decision to include only 10% of negative RDTs samples), the authors called for more inclusive analyses. We would like to share our ongoing experience in Chey Saen district (population 22499, 27 villages), Preah Vihear province, Cambodia [2]. The district is served by 3 health centers, 2 health posts, and 28 village malaria workers. In 2014, the Pf prevalence detected by PCR was estimated at 0.73% [3]. The incidence of symptomatic Pf infections in 2016 was 3.6‰. Since 2014, a network of malaria RDT providers has been supported and trained by Médecins Sans Frontières, in providing national guidelines treatment and in the RDT use (SD FK80 p.f/P.v Malaria Antigen Rapid Test, Standard Diagnostics). Since October 2015, the network is routinely collecting filter paper blood spots for subsequent qualitative and quantitative (using parasite density-calibrated controls) real-time PCR diagnosis (Institut Pasteur in Cambodia) [2, 4]. We conducted an overall analysis of the data collected between October 2015 and March 2017. A total of 4382 patients with suspected malaria were tested with both RDT and PCR. Of the 168 PCR-positive Pf samples, 23.8% (40/168) had a parasite density <100/μL. Table 1 displays all RDT and PCR results either including (n = 4382) or excluding samples with parasitemia <100/μL (n = 4342). The false-positive and false-negative rates were 11.0% (15/136) and 1.1% (47/4246), respectively. The sensitivity of RDT (vs PCR) was 72.0% (95% confidence interval [CI], 64.5%–78.5%), compared to 90.6% (95% CI, 83.8%–94.8%) after exclusion of low parasitemia samples. The negative predictive value increased from 98.9% to 99.7% when low-density samples were excluded. In both analyses, specificity was 99.7%, and the PPV scored 89.0% and 88.5%, respectively. Low parasitemia was the main reason for false-negative RDT results (35/47 [74.5%]). For the remaining 25.5%, although rare in Asia [5], the deletions of pfhrp2 and/or pfhrp3 genes are currently investigated. Comparison of Rapid Diagnostic Test (RDT) and Polymerase Chain Reaction (PCR) Results Among All Samples and Samples From High Density (≥100 Parasites/µL) Infections—Diagnostic accuracy of RDTs Versus PCR as Gold Standard Abbreviations: CI, confidence interval; NPV, negative predictive value; PCR, polymerase chain reaction; PPV, positive predictive value; RDT, rapid diagnostic test. Comparison of Rapid Diagnostic Test (RDT) and Polymerase Chain Reaction (PCR) Results Among All Samples and Samples From High Density (≥100 Parasites/µL) Infections—Diagnostic accuracy of RDTs Versus PCR as Gold Standard Abbreviations: CI, confidence interval; NPV, negative predictive value; PCR, polymerase chain reaction; PPV, positive predictive value; RDT, rapid diagnostic test. In conclusion, the present study complements the previous findings by Ranadive [1]. In particular, it grants more accuracy to the RDT in terms of PPV (89.0% vs 67.3%). Moreover, it confirms that the sensitivity of RDT, although higher than previously calculated (72.0% vs 51.7%), remains insufficient to meaningfully detect Pf infection in low-transmission, preelimination areas. Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.