Microscopy and CareStart™ HRP2/pLDH (Pf/PAN) Rapid Diagnostic Testing for Malaria among Symptomatic Patients at Federal Medical Centre Keffi, Nasarawa State, Nigeria

Abstract

Background: Microscopy and malaria rapid diagnostic tests (mRDTs) are routinely used to identify Plasmodium infection. Locally generated paired-test data can help characterise agreement between diagnostic approaches in symptomatic hospital attendees. Objective: To estimate malaria test positivity and compare the CareStart™ HRP2/pLDH (Pf/PAN) combination mRDT with light microscopy among symptomatic patients attending Federal Medical Centre Keffi, Nasarawa State, Nigeria. Methods: In this hospital-based cross-sectional study, venous blood specimens from 500 participants enrolled between August 2024 and September 2025 were examined by Giemsa-stained microscopy and the CareStart™ combination mRDT. Diagnostic accuracy indices were calculated with microscopy as the comparator, and agreement was assessed using Cohen’s kappa. Age- and sex-stratified test positivity and results across microscopy-reported parasite-density categories were summarised. Results: Microscopy identified Plasmodium infection in 128/500 (25.6%) participants, compared with 121/500 (24.2%) by mRDT. Relative to microscopy, there were 121 true-positive, 372 true-negative, seven false-negative, and no false-positive mRDT results. Sensitivity was 94.5% (95% confidence interval [CI]: 89.1–97.8%), specificity 100.0% (95% CI: 99.0–100.0%), positive predictive value 100.0%, negative predictive value 98.2%, and overall agreement 98.6% (Cohen’s κ = 0.963). Microscopy positivity was highest in the 11–15-year group (27/46; 58.7%). Sensitivity across the four recorded microscopy-density grades was 92.9%, 93.0%, 100.0%, and 94.4%, respectively. All microscopy-identified infections were classified as P. falciparum. Conclusion: The CareStart™ combination mRDT showed high agreement with microscopy in this selected symptomatic hospital population but missed seven microscopy-positive results. The findings describe performance relative to an imperfect comparator and do not establish community prevalence or the mechanisms of individual discordant results.