Mwende, Gabriel Miring’u2026-03-132026-03-132025-08https://ir-library.ku.ac.ke/handle/123456789/32743A Thesis Submitted in Partial Fulfillment of the Requirements for the Award of the Degree of Master of Science in Infectious Diseases (Medical Bacteriology) in the School of Health Sciences Of Kenyatta University, August 2025. Supervisor 1. Dr. Abednego Musyoki 2. Dr. Nelson MenzaSepticemia is mainly caused by Gram-negative bacteria (GNB) and misdiagnosis and/or delayed diagnosis can increase the disease burden. Blood culture is considered the ‘gold standard’ but it is insensitive and time consuming. Some molecular diagnostic tools have been developed to circumvent the blood culture shortcomings. However, they are costly for routine use and the existing multiplex PCR systems identify a limited bacterial spectrum. Therefore, this study developed a novel multiplex PCR assay for concurrent detection of the major GNB and determined the isolates’ phenotypic antimicrobial resistance. The multiplex PCR assays was developed at Nagasaki University laboratory, Kenya. The PCR assays were applied to diagnose septicemia alongside traditional culture method at Kiambu County Referral Hospital, Kenya, targeting children (<5 years) presenting with septicemia clinically. An experimental design was adopted for multiplex PCR development and a cross-sectional approach for its application. Blood samples were aseptically collected into biphasic blood culture bottle and in an EDTA tube. The EDTA blood sample was used for direct molecular bacterial DNA detection. Positive blood culture were processed to isolate causative bacterial pathogen, with an aliquot of culture fluid taken after every 4 hours of incubation, bacterial DNA extracted for multiplex PCR assays. Data obtained analyzed in STATA version 14. Descriptive statistics were used to analyze categorical variables, with the results presented in figures and tables. Blood culture was used as gold standard in assessing diagnostic performance of the multiplex PCRs. Sensitivity, specificity, overall agreement and area under the receiver operating characteristics (ROC) curve was calculated to establish the accuracy of the mPCR assays. Primers of species-specific were used to confirm known bacterial clinical isolates and standard strains. Two mPCR assays were developed and established an optimal annealing temperature of primers at 55°C which was used in the final amplification. All the targeted bacteria were detected, with a minimum detectable DNA concentration at 100 pg. No target bacteria were directly detected from whole blood, however, after 4 hours and 8 hours of primary incubation, 41% (5/12) and 100% (12/12) of the target bacteria were detected in culture fluids, respectively. The assays also identified Klebsiella pneumoniae and Salmonella spp. co-infections and extra pathogens (2 K. pneumoniae and 1 Escherichia coli) compared with blood culture. The multiplex PCR demonstrated a sensitivity of 100.0% (71.7-100.0) and specificity of 98.0% (90.7-99.0), indicating the assays have high ability to correctly identify individual without and with the disease respectively. Area under the ROC curve was 1.00 (0.00-1.00), indicating a perfect performance of the assays. K. pneumoniae were the only multidrug-resistant and extended-spectrum β-lactamase producing isolates. Given the high sensitivity, specificity, and perfect ROC of the mPCR, adoption of this rapid assays tool by the Ministry of Health for routine use to ensure timely diagnosis of bacterial septicemia in Kenya is recommended. Additionally, with early detection of blood bacterial pathogens by multiplex PCR, the assays are recommended for timely and continuous antimicrobial resistance surveillance to inform antibiotics policy guidelines and mitigate the resistance burden in the current study area and beyond.enThesis