Comparison of the Accuracy of Four Malaria Diagnostic Methods in a High Transmission Setting in Coastal Cameroon

Abstract

Background. Despite recommendation from the World Health Organization that all malaria suspected patients undergo a parasitological confirmation using rapid diagnostic test or light microscopy prior to treatment, health facilities in remote malaria endemic settings sometimes resort to presumptive diagnosis of malaria for clinical management for various reasons. Following observation of this practice, we undertook a cross-sectional study aimed at comparing presumptive diagnosis based on axillary temperature, SD Bioline™ rapid test, and light microscopy as strategies for malaria diagnosis in the coastal region of Mutengene in the South West of Cameroon with the overall goal of supporting improved malaria diagnosis at local levels. Methodology. Venous blood from 320 participants was used to detect the presence of malaria parasite using SD Bioline™ mRDT and Giemsa stained microscopy or spotted on filter paper for PCR amplification of the 18s rRNA gene of Plasmodium sp following standard procedures. The axillary temperature of each participant was also measured. The sensitivity, specificity, and predictive values and their confidence intervals were determined for each of the methods with PCR as the reference. The area under the curve was used to estimate accuracy of diagnostic method and compared between test method using the X2 test with P<0.05 considered significant. Results. The overall diagnostic sensitivities of presumptive diagnosis using axillary temperature, light microscopy, and SD Bioline™ were observed to be 74.30% (95%CI: 67.90-80.01), 94.86% (95%CI: 90.99-97.41), and 95.33% (95%CI: 91.57-97.74), respectively, and their respective diagnostic specificities were 53.77% (95%CI: 43.82-63.51), 94.34% (95%CI: 88.09-97.87), and 94.34%(95%CI: 88.09-97.89). SD Bioline™ had a diagnostic sensitivity of 91.80% [95%CI: 81.90-97.28] at a parasitaemia of less than 500 parasites/μl of blood but a sensitivity of 100% for parasite counts above 500 parasites/μl of blood. The predictive values of the positive test were highly comparable between light microscopy (90.09%, [95%CI: 83.61-94.18]) and SD Bioline™ mRDT (90.91%, [95%CI: 84.50-94.83]), P=0.98 with kappa values of 0.898 but lower for presumptive diagnosis (50.89%, [95%CI: 43.72-58.03]), P<0.0001, and kappa value of 0.277. Perfect agreement was observed between SD Bioline™ mRDT and light microscopy (Cohen kappa= 0.924). Conclusions. The study showed that SD Bioline™ was as good as light microscopy in the diagnosis of malaria in remote areas of perennial transmission in South West Cameroon. This study equally revealed the limitations of presumptive diagnosis of malaria (as opposed to the use of RDTs or microscopy). Efforts should be made in such areas to promote parasitological confirmation of malaria using quality assured rapid tests or light microscopy for case management of malaria. The presence of nonnegligible levels of Plasmodium ovale in this study area indicate that treatment guidelines may require revision if same trend is proven in several other areas of same ecology.