GWGeorge M. Warimwe

George M. Warimwe

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


George M. Warimwe has a total of 464 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Temporal trends of dengue exposure among children in coastal Kenya

    by Daisy Mugo, Henry K. Karanja, Benedict Orindi, John N. Gitonga, Kennedy Mwai, Doris K. Nyamwaya, Donwilliams O. Omuoyo, Anthony K. Nyamache, Barnes S. Kitsao, Jennifer N. Musyoki, et al.

    Background: Serosurveys suggest widespread dengue virus (DENV) transmission in Africa but there is limited information on the temporal patterns of exposure. Methods: Using repeated cross-sectional serosurvey data, we estimated the prevalence and incidence of DENV infections in coastal Kenya over 20 years (1998–2018). Stored sera ( n = 8025), collected annually from an open longitudinal cohort of 1834 children aged 15 years and below, were screened for anti-DENV IgG antibodies. Results: We show that, of 8025 samples, 2287 (28.5%, 95% CI 27.5–29.5%) were seropositive for DENV IgG antibodies. Seroprevalence increased with age, from 2% among children aged below 1 year to 40% among those aged 9 to 10 years, and peaked during outbreak years. Among 1343 children who were seronegative at recruitment, the overall incidence (seroconversion) rate was 111.5 DENV infections per 1000 person-years (95% CI 102.5–121.2). The highest incidence occurred in 2013 at 439.7 infections per 1000 person-years (95% CI 375.1–515.5), coinciding with a large outbreak in coastal Kenya. Conclusions: Our findings suggest long-term DENV exposure among children in coastal Kenya and highlight an urgent need for clinical surveillance to determine the associated health burden.


  • Open Access

    Seroepidemiology of Chikungunya, Dengue and Rift Valley Fever Viruses across Five Health and Demographic Surveillance System Sites in Kenya

    by Augustine Masinde, James Nyagwange, Godfrey Bigogo, Patrick Munywoki, Abdhalah Ziraba, Makobu Kimani, Shirine Voller, Rose Ombati, Bernadette Kutima, John Gitonga, et al.

    Background: Arboviral infections are important public health concerns in Kenya, yet representative estimates of population-level exposure remain limited. We employed serological surveillance to investigate the epidemiology of chikungunya, dengue, and Rift Valley Fever (RVFV) across health and demographic surveillance system (HDSS) sites in Kenya.Methods: We retrospectively analysed 4,165 plasma samples collected in 2022 during COVID-19 serosurveys among randomly selected residents of five HDSS sites in Kilifi, Korogocho/Viwandani (Nairobi), Kibera (Nairobi), Manyatta (Kisumu) and Asembo (Siaya). We tested the samples for chikungunya, dengue, and RVFV immunoglobulin G (IgG) antibodies using a validated multiplex Luminex assay. We estimated population-level seroprevalence using Bayesian logistic regression models adjusting for assay performance and post-stratification by age and sex distributions specific to each site. Associations between serostatus and age and sex were assessed using logistic regression. We used mixture models of anti-chikungunya IgG levels as an alternate approach for estimating the prevalence of past chikungunya infection.Findings: Adjusted chikungunya seroprevalence ranged from 71·7% (95% credible interval 65·0-79·1) in Asembo to 5·5% (2·0-9·1) in Korogocho/Viwandani. It was >50% among all age groups in Asembo. In Manyatta, Kibera, and Kilifi, chikungunya seroprevalence was <20% among children aged <15 years and increased with age. Mixture model estimates of past chikungunya infection were 1·2-3·3 times higher than adjusted seroprevalence across sites. Adjusted dengue seroprevalence was 24·4% (20·0-29·4) in Kilifi and below 1% elsewhere, while RVFV seroprevalence was below 4% across all sites.Interpretation: These findings suggest intense childhood chikungunya exposure with rapid acquisition of population immunity in Siaya, cumulative lifetime exposure in Kilifi and Kisumu, and limited historical transmission in Nairobi implying a larger susceptible population while highlighting co-circulation of chikungunya and dengue viruses in Kilifi. Rational disease prevention and control strategies may include prioritizing case-based surveillance for chikungunya and sub-national targeting for future chikungunya and dengue vaccination programs.

  • Open Access

    Low-dose yellow fever vaccination in infants: a randomised, double-blind, non-inferiority trial

    by Derick Kimathi, Aitana Juan-Giner, Ndeye S Bob, Benedict Orindi, Maria L Namulwana, Antoine Diatta, Stanley Cheruiyot, Gamou Fall, Moussa Dia, Mainga M Hamaluba, et al.


  • Open Access

    Evaluating Orthoflavivirus Vaccines with PRNT: Lessons from Yellow Fever Clinical Studies

    by Moussa Dia, Ndeye Sakha Bob, Derick Kimathi, Aitana Juan-Giner, George Warimwe, Rebecca F. Grais, Ndongo Dia, Oumar Faye, Gamou Fall

    Flaviviruses represent a major global health concern, with yellow fever virus (YFV) and dengue virus (DENV) among the most studied members of the Orthoflavivirus genus. Transmitted by Aedes mosquitoes, these viruses are responsible for severe disease outbreaks worldwide. While YFV benefits from an effective vaccine, it remains a threat in endemic regions. Dengue, with its four distinct serotypes, continues to complicate vaccine development and disease management. Plaque assays and the plaque reduction neutralization test (PRNT) are key tools for assessing viral replication and immune responses. PRNT50, commonly used in postvaccination studies, quantifies the serum dilution needed to neutralize 50% of the virus, providing insights into vaccine-induced immunity. In contrast, PRNT90, with a higher neutralization threshold, is more appropriate for diagnostics and epidemiological surveillance, offering a stringent measure of past exposure and protection. This chapter focuses on the application of PRNT in yellow fever (YF) vaccine research, highlighting PRNT50 for evaluating immunogenicity and PRNT90 for population-level immunity assessments. We present findings from a core study validating the PRNT assay, followed by clinical research supporting fractional dosing strategies in various populations, including children and immunocompromised individuals. Together, these studies underscore the value of PRNT in optimizing vaccine use and advancing public health strategies against Orthoflavivirus infections.

  • Open Access

    Assessing the Impact of SARS-CoV-2 Spike Mutations on Antibody Binding: A Comparative Assessment of the Wuhan and JN.1 Variants’ Full-Length Spikes in a Multiplex Luminex Assay

    by Gerald Waweru, Ruth Nyakundi, Bernadette Kutima, Sharon Owuor, Gloria Konyino, John Gitonga, Doreen Lugano, Angela Maina, Jennifer Musyoki, Lucy Ochola, et al.

    Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) continues to evolve, with mutations leading to the emergence of new variants. JN.1, a subvariant of omicron BA.2.86, has demonstrated marked immune escape and is now included in updated vaccine formulations. While reduced sensitivity has been reported for antibody assays using ancestral spike protein subunits to detect omicron-induced responses, the performance of full-length spike-based assays against omicron sublineages remains unclear. We aimed to compare the sensitivity of ELISA and Luminex assays using full-length spike proteins from the ancestral Wuhan strain and the JN.1 variant. Methods: Wuhan and JN.1 full-length spike protein constructs were designed and expressed in Expi293F mammalian cells. In-house ELISAs based on previously validated protocols were used to measure anti-spike IgG levels. Additionally, a Luminex-based assay for anti-spike antibody detection was developed and validated. Both assays were applied to the following sample groups: pre-pandemic samples (designated “gold standard negatives”); PCR confirmed 2020 positives (“gold standard wildtype positives”); PCR confirmed 2024 positives (“gold standard omicron positives”); 2022 vaccinated individuals with verbal confirmed infection (“gold standard hybrid positives”); and 2024 household samples (“unknowns”). Results: Wuhan spike protein showed a sensitivity of 100% (95% CI: 0.88–1.0) in detecting omicron-specific antibodies using gold standard omicron positives with JN.1 spike protein as a reference assay. Overall, across all samples, in ELISA, the Wuhan antigen had a sensitivity of 0.93 (95% CI: 0.89–0.95) and a specificity of 0.98 (95% CI: 0.94–0.99). The JN.1 antigen showed a sensitivity of 0.91 (95% CI: 0.87–0.94) and a specificity of 0.97 (95% CI: 0.93–0.99). In Luminex, sensitivity was 0.95 (95% CI: 0.91–0.97) for Wuhan and 0.94 (95% CI: 0.91–0.96) for JN.1. Specificity for both antigens in Luminex was 0.98 (95% CI: 0.94–0.99). Conclusions: Both ELISA and Luminex assays showed comparable sensitivity and specificity for both Wuhan and JN.1 antigens, indicating that mutations in the JN.1 variant do not significantly impact assay performance. This suggests preserved antigenic recognition across variants.

  • Open Access

    Comparative performance of the InBios SCoV-2 DetectTM IgG ELISA and the in-house KWTRP ELISA in detecting SARS-CoV-2 spike IgG antibodies in Kenyan populations

    by Bernadette Kutima, Eunice Wageci Kagucia, Kennedy Mwai, Makobu Kimani, Antipa Sigilai, Daisy Mugo, Henry Karanja, John N Gitonga, Angela Karani, Donald Akech, et al.

    Background The InBios SCoV-2 Detect™ IgG ELISA (InBios) and the in-house KWTRP ELISA (KWTRP) have both been used in the estimation of SARS-CoV-2 seroprevalence in Kenya. Whereas the latter has been validated extensively using local samples, the former has not. Such validation is important for informing the comparability of data across the sites and populations where seroprevalence has been reported. Methods We compared the assays directly using pre-pandemic serum/plasma collected in 2018 from 454 blood donors and 173 malaria cross-sectional survey participants, designated gold standard negatives. As gold standard SARS-CoV-2 positive samples: we assayed serum/plasma from 159 SARS-CoV-2 PCR-positive patients and 166 vaccination-confirmed participants. Results The overall agreement on correctly classified samples was >0.87 for both assays. The overall specificity was 0.89 (95% CI, 0.87–0.91) for InBios and 0.99 (95% CI, 0.97–0.99) for KWTRP among the gold standard negative samples while the overall sensitivity was 0.97 (95% CI, 0.94–0.98) and 0.93 (95% CI, 0.90– 0.95) for InBios and KWTRP ELISAs respectively, among the gold standard positive samples. In all, the positive predictive value for InBios was 0.83 (95% CI, 0.79-0.87) and 0.98 (95% CI, 0.96-0.99) for KWTRP while the negative predictive value was 0.98 (95% CI, 0.97- 0.99) and 0.97 (95% CI, 0.95-0.98) for InBios and KWTRP respectively. Conclusions Overall, both assays showed sufficient sensitivity and specificity to estimate SARS-CoV-2 antibodies in different populations in Kenya.


  • Open Access

    The Burden of Chikungunya and Onyong’nyong Viral Infections in Uganda: A Hospital-Based Sentinel Surveillance Study

    by Doris Kemunto Nyamwaya, Nyamai Mutono, John Kayiwa, Julius Lutwama, Augustine Masinde, Sm Thumbi, Martin Antonio, George Warimwe, Pontiano Kaleebu, Thumbi Mwangi

    Background: Chikungunya and Onyong’nyong viruses are mosquito-borne alphaviruses endemic to Africa, causing febrile illness with severe joint pain. While Chikungunya virus is known for its potential to cause life-threatening complications in neonates and vulnerable populations, Onyong’nyong virus remains vastly underreported due to weak surveillance and diagnostic overlap. As a result, Onyong’nyong virus’s true burden and epidemic potential remain dangerously underestimated. Methods: We conducted sentinel hospital-based surveillance in Uganda to assess the occurrence, geographical distribution, and clinical manifestation of Chikungunya and Onyong’nyong viruses. Serological and molecular diagnostic methods were employed to detect active and recent infections. Epidemiological data were analyzed to characterize clinical manifestations and determine risk factors for severe disease. Results: A total of 2,756 serum samples were collected from febrile patients across 63 districts in Uganda between 2018 and 2022. While no Chikungunya infection was detected by PCR, two cases tested positive for Onyong’nyong virus. Serological evidence of recent alphavirus infection was found in 8.79% of samples, with peak prevalence in 2020. Most positive cases reported fever, joint and muscle pain, and other non-specific symptoms. PRNT analysis revealed stronger neutralizing antibody responses to Onyong’nyong virus than Chikungunya virus, suggesting ongoing but under-recognized Onyong’nyong virus transmission across Uganda. Conclusion: This study underscores the endemic nature of Onyong’nyong virus in Uganda and the need for enhanced surveillance, improved diagnostic capacity, and targeted public health interventions. These findings will inform clinical management strategies and contribute to preparedness against future outbreaks.

  • Open Access

    Low-Dose Yellow Fever Vaccine in Adults in Africa

    by Derick Kimathi, Aitana Juan-Giner, Ndeye S. Bob, Benedict Orindi, Maria L. Namulwana, Antoine Diatta, Stanley Cheruiyot, Gamou Fall, Moussa Dia, Mainga M. Hamaluba, et al.

    Yellow fever vaccine is highly effective with a single dose, but vaccine supply is limited. The minimum dose requirements for seroconversion remain unknown. In this double-blind, randomized, noninferiority trial in Uganda and Kenya, we assigned adults with no history of yellow fever vaccination or infection to receive vaccination with the Institut Pasteur de Dakar 17D-204 yellow fever vaccine at a standard dose (13,803 IU) or at a fractional dose of 1000 IU, 500 IU, or 250 IU. The primary outcome was seroconversion at 28 days after vaccination with each fractional dose as compared with the standard dose, evaluated in a noninferiority analysis. Seroconversion was defined as an antibody titer at day 28 that was at least four times as high as the antibody titer before vaccination, as measured by a plaque reduction neutralization test. We conducted noninferiority analyses in the per-protocol and intention-to-treat populations. Noninferiority was shown if the lower boundary of the 95% confidence interval for the difference in the incidence of seroconversion between the fractional dose and the standard dose was higher than -10 percentage points. A total of 480 participants underwent randomization (120 participants in each group). The incidence of seroconversion was 98% (95% confidence interval [CI], 94 to 100) with the standard dose. The difference in the incidence of seroconversion between the 1000-IU dose and the standard dose was 0.01 percentage points (95% CI, -5.0 to 5.1) in the intention-to-treat population and -1.9 percentage points (95% CI, -7.0 to 3.2) in the per-protocol population; the corresponding differences between the 500-IU dose and the standard dose were 0.01 percentage points (95% CI, -5.0 to 5.1) and -1.8 percentage points (95% CI, -6.7 to 3.2), and those between the 250-IU dose and the standard dose were -4.4 percentage points (95% CI, -9.4 to 0.7) and -6.7 percentage points (95% CI, -11.7 to 1.6). A total of 111 vaccine-related adverse events were reported: 103 were mild in severity, 7 were moderate, and 1 was severe. The incidence of adverse events was similar in the four groups. A yellow fever vaccination dose as low as 500 IU was noninferior to the standard dose of 13,803 IU for producing seroconversion within 28 days. (Funded by the European and Developing Countries Clinical Trials Partnership and the Wellcome Trust; NIFTY ClinicalTrials.gov number, NCT04059471.).


  • Open Access

    Kinetics of naturally induced binding and neutralising anti-SARS-CoV-2 antibody levels and potencies among SARS-CoV-2 infected Kenyans with diverse grades of COVID-19 severity: an observational study

    by John Kimotho, Yiakon Sein, Shahin Sayed, Reena Shah, Kennedy Mwai, Mansoor Saleh, Perpetual Wanjiku, Jedidah Mwacharo, James Nyagwange, Henry Karanja, et al.

    Background Given the low levels of coronavirus disease 2019 (COVID-19) vaccine coverage in sub-Saharan Africa (sSA), despite high levels of natural severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) exposures, strategies for extending the breadth and longevity of naturally acquired immunity are warranted. Designing such strategies will require a good understanding of naturally acquired immunity. Methods We measured whole-spike immunoglobulin G (IgG) and spike-receptor binding domain (RBD) total immunoglobulins (Igs) on 585 plasma samples collected longitudinally over five successive time points within six months of COVID-19 diagnosis in 309 COVID-19 patients. We measured antibody-neutralising potency against the wild-type (Wuhan) SARS-CoV-2 pseudovirus in a subset of 51 patients over three successive time points. Binding and neutralising antibody levels and potencies were then tested for correlations with COVID-19 severities. Results Rates of seroconversion increased from day 0 (day of PCR testing) to day 180 (six months) (63.6% to 100 %) and (69.3 % to 97%) for anti-spike-IgG and anti-spike-RBD binding Igs, respectively. Levels of these binding antibodies peaked at day 28 (p0.99). Similarly, antibody-neutralising potencies peaked at day 28 (p<0.01) but declined by three-fold, six months after COVID-19 diagnosis (p0.60, p<0.01). Levels and potencies of binding and neutralising antibodies increased with disease severity. Conclusions Most COVID-19 patients generated SARS-CoV-2 specific binding antibodies that remained stable in the first six months of infection. However, the respective neutralising antibodies decayed three-fold by month-six of COVID-19 diagnosis suggesting that they are short-lived, consistent with what has been observed elsewhere in the world. Thus, regular vaccination boosters are required to sustain the high levels of anti-SARS-CoV-2 naturally acquired neutralising antibody potencies in our population.


  • Open Access

    Comparative performance of the InBios SCoV-2 DetectTM IgG ELISA and the in-house KWTRP ELISA in detecting SARS-CoV-2 spike IgG antibodies in Kenyan populations

    by Bernadette Kutima, Eunice Wageci Kagucia, Kennedy Mwai, Makobu Kimani, Antipa Sigilai, Daisy Mugo, Henry Karanja, John N Gitonga, Angela Karani, Donald Akech, et al.

    Background The InBios SCoV-2 DetectTM IgG ELISA (InBios) and the in-house KWTRP ELISA (KWTRP) have both been used in the estimation of SARS-CoV-2 seroprevalence in Kenya. Whereas the latter has been validated extensively using local samples, the former has not. Such validation is important for informing the comparability of data across the sites and populations where seroprevalence has been reported. Methods We compared the assays directly using pre-pandemic serum/plasma collected in 2018 from 454 blood donors and 173 malaria cross-sectional survey participants, designated gold standard negatives. As gold standard SARS-CoV-2 positive samples: we assayed serum/plasma from 159 SARS-CoV-2 PCR-positive patients and 166 vaccination-confirmed participants. Results The overall agreement on correctly classified samples was >0.87 for both assays. The overall specificity was 0.89 (95% CI, 0.87–0.91) for InBios and 0.99 (95% CI, 0.97–0.99) for KWTRP among the gold standard negative samples while the overall sensitivity was 0.97 (95% CI, 0.94–0.98) and 0.93 (95% CI, 0.90– 0.95) for InBios and KWTRP ELISAs respectively, among the gold standard positive samples. Conclusions Overall, both assays showed sufficient sensitivity and specificity to estimate SARS-CoV-2 antibodies in different populations in Kenya.