Seroprevalence of Rift Valley fever virus and factors associated with exposure in south-central Uganda

Abstract

Background

Rift Valley fever (RVF) poses a recurring public health and economic challenge in Uganda’s cattle corridor. It is a mosquito-borne zoonotic disease caused by the Rift Valley fever virus (RVFV), with high livestock density, favorable rainfall patterns, and close human–animal interaction facilitating viral transmission and disease outbreaks. Although RVF outbreaks have occurred, data on human exposure in south-central Uganda remain limited.

Methods

This cross-sectional study assessed the seroprevalence and associated factors of RVFV exposure among selected communities in the greater Masaka-Rakai region of south-central Uganda. Archived human sera (n = 1,199) retrospectively collected between June 2018 and October 2020 from the population-based Rakai Community Cohort Study were analyzed using both indirect in-house and commercial ID Screen^® RVF competition enzyme-linked immunosorbent assays (ELISAs) to detect RVFV IgG antibodies. Household census and socio-demographic data were linked to serological results, and logistic regression was performed to identify factors associated with seropositivity.

Results

RVFV IgG antibodies were detected in 86 of 1,199 (7.8%) individuals. In multivariable analysis, RVFV seropositivity was significantly associated with residence in agrarian communities and older age groups (25–34 and ≥ 35 years), while pig ownership was associated with lower risks of exposure.

Conclusion

These findings suggest RVFV human exposure in south-central Uganda and emphasize the need for strengthened surveillance, targeted prevention, and One Health interventions to protect high-risk groups.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12879-026-13923-1.

Background

Rift Valley fever (RVF) is a mosquito-borne zoonosis that poses a major public health and economic threat in Africa and beyond. The disease is caused by the Rift Valley fever virus (RVFV) and can progress to a viral hemorrhagic fever in humans [1]. Infected ruminants can experience mass abortions with mortality rates up to 100% [2, 3]. In humans, infections can be asymptomatic or have symptoms ranging from mild flu-like illness to severe hemorrhagic manifestations [4–6] with fatality rates of up to 50% in severe cases [7]. Comorbidities such as diabetes, chronic kidney disease, and HIV/AIDS further increase severe disease outcome [8]. Following infection, symptoms typically emerge within 2–6 days of exposure [9, 10]. The host mounts short- and long-term antibody responses. Immunoglobulin M (IgM) emerges within the first week and wanes after 2–3 months, marking recent infection, while IgG develops later and persists for years, indicating past exposure and long-term immunity [2]. Understanding the full spectrum of RVFV infections, including asymptomatic cases, is crucial for accurately estimating disease burden, as undiagnosed infections may lead to underestimation of population exposure, hindering effective surveillance, timely diagnosis, and control efforts [11].

Major RVF outbreaks have caused widespread morbidity and mortality across decades. The 2000 outbreak among humans on the Arabian Peninsula recorded a 14% case fatality rate (CFR) [12, 13]. This was the first occurrence outside of Africa, raising concerns about the wider spread of RVFV. In sub-Saharan Africa, RVFV is endemic in several countries, including Uganda, Kenya, Tanzania, Sudan, Mauritania, South Africa, and Senegal, where outbreaks often follow periods of heavy rainfall or flooding [14, 15]. Uganda’s cattle corridor, stretching diagonally from the northeastern to southwestern parts of the country, is characterized by extensive pastoral livestock systems, high livestock mobility, and frequent human–livestock interaction [16]. Environmental conditions such as high rainfall and proximity to water bodies with seasonal floods create favorable breeding sites for mosquito vectors and support livestock farming. Consequently, the region experiences recurrent RVFV outbreaks approximately every 2–4 years due to the favorable ecological conditions and proximity to endemic regions [17, 18]. Between 2016 and 2018, outbreaks across 10 districts resulted in 16 human cases and 7 deaths (CFR = 45%), all among laboratory-confirmed severe cases, consistent with reports of high mortality in severe RVF cases [19].

Human infection occurs through both mosquito bites and direct contact with infected animal materials (blood, tissues, carcasses, or aborted fetuses) during activities such as slaughter, birthing assistance, or veterinary procedures, with occupational exposure to livestock often representing the dominant risk factor [20]. Risk is heightened by keeping livestock near homes, ineffective mosquito control, and occupational exposure in farming, veterinary work, or slaughterhouse operations [21, 22]. Environmental drivers, especially heavy rainfall and flooding, create ideal breeding conditions for mosquitoes and can trigger outbreaks [23]. In south-central Uganda, episodic flooding is common, creating temporary bodies of water and expansive wetland conditions that sustain mosquito amplification cycles, thereby increasing the likelihood of RVFV circulation. In Uganda, transmission peaks during the March–May and the September–November rainy seasons [24]. Although livestock vaccines exist, there is no licensed human vaccine [25, 26] or specific antiviral treatment for RVF in humans [10]. Medical care remains supportive, and control focuses on mosquito prevention, livestock vaccination, and public health education [27, 28].

Although Uganda’s national One Health strategy emphasizes prevention of zoonotic diseases, responses remain largely reactive, triggered by outbreaks rather than proactive surveillance and early risk factor identification. The Food and Agriculture Organization (FAO) has forecasted high RVF risk across East Africa, recommending enhanced awareness and preparedness efforts at national, subnational, and community levels [29]. Further, metadata from 2008 to 2022 reveals geographic expansion of RVF clusters into previously unaffected regions, about 35% occurring in regions without prior RVFV activity, including south-central Uganda [30]. Here we retrospectively examined the seroprevalence of RVFV and factors associated with exposure among individuals in selected communities in south-central Uganda. This region lies within Uganda’s cattle corridor, extending from the western parts of Lake Victoria to the Mutukula border with Tanzania in the south. It is characterized by a tropical climate with distinct wet (March–May; September–November) and dry seasons, which influence livestock farming and RVFV epidemiology [18, 24, 31].

Methods

Study design, study setting, and study population

This was a cross-sectional study conducted using retrospectively collected data and serum samples from the Rakai Community Cohort Study (RCCS) participants [32, 33], of the survey round that was conducted between June 2018 and October 2020. We used demographic, behavioral, and disease surveillance data along with serum samples collected from participants of the RCCS in the greater Masaka–Rakai region. The RCCS is an ongoing longitudinal, population-based cohort study conducted by the Rakai Health Sciences Program (RHSP) since 1994. Approximately every 18 months, RCCS conducts a survey round that starts with a census of all permanent and transient residents across 21 agrarian, 16 trading, and 4 fishing communities, which are grouped according to the predominant economic activity [34]. Following the census, all eligible residents aged ≥ 15 years are invited to participate. Consenting individuals are then interviewed using structured, confidential tools and provide blood samples, GPS coordinates, and sociodemographic and health data. Each survey round recruits ~ 20,000 participants comprising new and follow-up residents from these communities. The sample size, community types, and demographics of this cohort were previously published [35].

Study variables

The primary outcome was serological evidence of exposure to RVFV, determined through antibody detection using enzyme-linked immunosorbent assay (ELISA). Potential confounders included socio-demographic characteristics (age, gender, and highest level of education attained) and household or community-level factors, including livestock ownership (cattle, goats, and pigs), bed net use, availability of cultivable land, occupation, and community type (agrarian, trading, or fishing).

Serological testing for Rift Valley fever virus

A serial testing algorithm was applied for RVFV IgG detection. Samples were classified as seropositive if they were positive in both the commercial ID Screen^® assay and the in-house ELISA. Agreement between the two assays was assessed by calculating percent agreement and Cohen’s kappa (κ) statistic, with percent agreement defined as the proportion of samples with concordant results. Analyses were performed using Stata/MP version 18 [36].

Indirect in-house ELISA

Serum samples were heat-inactivated at 56 °C for 1 h and screened by an in-house ELISA, with samples run in duplicate. Nunc Maxisorp 96-well plates were coated overnight at 4 °C with RVFV nucleoprotein antigen (strain ZH548, Native Antigen REC31640-500) diluted at 1:12,800 in PBS. The following morning, the excess antigen was removed, plates were blocked with 0.1% milk in PBS-Tween for 1 h at room temperature (RT). Plates were washed 3x with 250µL of PBS-Tween. Serum samples diluted 1:400 in 0.1% milk were added to the plates (100µL per well) in duplicate. Blank wells, known RVFV IgG-positive human sera, and negative pooled human sera were added in duplicate to each plate as controls. Plates were incubated at RT for 1 h, then washed 3x with 250µL of PBS–Tween. A secondary human antibody biotin conjugate (Goat Anti-Human IgG-BIOT, SouthernBiotech, Cat. No.: 2040-08), diluted 1:4000 in PBS-Tween, was added (100µL per well), followed by incubation at RT for 1 h. Plates were washed 3x with 250µL of PBS-Tween. Streptavidin–HRP conjugate (SouthernBiotech, Cat. No.: 7105-05), diluted 1:4000 in PBS, was then added (100µL per well) and incubated at RT for 30 min. Plates were subsequently washed 6x with 250µL of PBS-Tween. ABTS^® 2-Component Microwell Peroxidase Substrate containing 2, 2’-azino-di (3-ethylbenzthiazoline-6-sulfonate, SeraCare, Material No. 5120-0032) was added at a 1:1 dilution (100µL per well), and plates were incubated in the dark for 30 min. Optical density (OD) was measured and read spectrophotometrically at 405 nm using a SmartReader™ 96 Absorbance Plate Reader (model 9600), and values were averaged across duplicate wells. The IgG cut-off value was defined as the mean + 2 standard deviations of the negative control serum pool; samples above this threshold were classified as positive. The assay was optimized before use by determining optimal antigen and antibody dilutions and assay performance was monitored using known negative and positive controls to ensure assay performance. Agreement between the assays was assessed using percentage agreement and Cohen’s kappa statistic to assess agreement beyond chance.

ID screen^® RVF competition ELISA

Samples were also run for secondary assay confirmation by the ID Screen^® RVF Competition Multi-species ELISA (Innovative Diagnostics) for RVFV nucleoprotein IgG antibodies, following the manufacturer’s instructions. This assay is based on a competitive binding principle, in which antibodies present in the sample compete with labeled antibodies for binding to the antigen; therefore, lower signal intensity corresponds to higher antibody levels. This kit uses recombinant RVFV nucleoprotein as the coating antigen and has demonstrated a sensitivity of 98% and a specificity of 100% [37]. Results were expressed as competition percentage (S/N%), calculated as: S/N% = (OD_S of the sample / OD_NC of the negative control) × 100. Samples with S/N% ≤ 40% were classified as positive, while those with S/N% > 50% were considered negative, according to the kit instructions. The test validity was assessed in each plate run using kit-provided controls. In addition, known RVFV IgG-positive human serum was used as a positive control, while pooled normal human serum (Innovative Research) and blank wells were used as negative controls. All samples and controls were run in duplicate on each plate.

Statistical analysis

Frequencies and percentages were used to describe the distributions of potential confounding variables, while chi-square or Fisher’s exact tests were used to test differences in distributions between the community types. Categorical variables were described using frequencies and proportions, and relevant results were presented in tables to show distribution patterns across demographic groups and visualize community types. Seroprevalence with 95% confidence intervals (CIs) was estimated for each community type from the proportion of the RVFV-positive samples, and overall seroprevalence was computed from the population-weighted proportion of RVFV-positive samples. Significant association between potential confounders and the prevalence level of RVFV was assessed using odds ratios with corresponding 95% confidence intervals (CIs) that were computed from bivariable and multivariable logistic regression analysis. Variables with p < 0.20 in the bivariable analysis were included in the multivariable model; significance was set at p < 0.05. Interaction was assessed by generating interaction terms with significant variables and using the chunk test to identify any significant interactions. Unadjusted and adjusted odds ratios (ORs) with 95% CI were reported; model fit was analyzed using the Hosmer-Lemeshow test. Dual seropositivity for RVFV and CCHFV was determined by linking individual-level serological results from the current RVFV analysis with CCHFV IgG results from the same cohort [35] and identifying participants who were positive for both infections. To determine whether exposure risk was driven primarily by direct animal contact, the broader ecological environment, or both, livestock ownership and community type interaction effect to the outcome was assessed; subgroup-specific adjusted odds ratios were reported. Data was analyzed using Stata/MP version 18 (StataCorp LLC, College Station, TX, USA) [36].

Results

Participant demographics

The study included 1,199 archived serum samples collected across 21 agrarian (n = 400), 16 trading (n = 399), and 4 fishing (n = 400) communities. Demographic characteristics are presented in Supplemental Table 1. Participant characteristics varied significantly by community type (p < 0.05), reflecting heterogeneity in potential risk factors for RVFV exposure. Females predominated in agrarian and trading communities, while males were more common in fishing areas. Outdoor or animal-related occupations were frequent, particularly in agrarian and fishing communities. Livestock ownership and land cultivation were highest in agrarian communities and lowest in fishing communities. As RVFV is a mosquito-borne virus, we also evaluated bed net usage. Bed net use was generally high but slightly lower in fishing areas (Supplemental Table 1).

Seroprevalence of RVFV by community type

Among the 1,199 samples analyzed, RVFV seroprevalence was 7.8% (95% CI: 1.12–16.57). Agreement between the in-house ELISA and the commercial ID Screen^® assay was high, at 96.8%, with a Cohen’s kappa (κ) of 0.812, indicating almost perfect agreement between the two tests. The highest seropositivity was observed in agrarian communities (10.5%, 95% CI: 7.85–13.91), followed by fishing (6%, 95% CI: 4.05–8.80) and trading (5%, 95% CI: 3.25–7.64) communities (Table 1). The spatial distribution of community seropositivity and GPS coordinates is shown in Fig. 1. Overall, seroprevalence differed significantly across the three community types (p = 0.006). Trading and fishing communities had lower and statistically similar seroprevalence compared to agrarian communities (p = 0.541). Since their serostatus did not differ significantly, they were pooled to increase statistical power for subsequent variable-specific analyses (Table 2). In the pooled comparison, agrarian communities had a significantly higher seroprevalence (10.5%) than the combined trading/fishing communities (5.5%, p = 0.002).

Table 1.

Seroprevalence of RVFV across community types

	Community type			Overall n (%)
Exposure variable	Agrarian n (%)	Trading n (%)	Fishing n (%)
RVFV status
Negative	358(89.5)	379(95.0)	376(94.0)	1113(92.2)
Positive	42(10.5)	20(5.0)	24(6.0)	86(7.8)
Chi-squared p-values
Agrarian vs Trading vs Fishing	0.006
Trading vs Fishing	-	0.541

Fig. 1.

A thematic point map showing the spatial distribution of the 41 RCCS communities included in the study across Rakai, Kyotera, Masaka, and Lyantonde districts, Uganda. Each point represents a community, with shape indicating community type (agrarian, trading, and fishing landing sites) and color representing RVFV seropositivity levels. Major water bodies and district boundaries are shown to provide environmental and geographic context. Figure adapted from Mihalakakos et al. [35]

Table 2.

Pooled comparison of RVFV seroprevalence

Pooled comparison
Exposure variable	Agrarian n (%)	Trading or Fishing n (%)
RVFV status
Negative	358(89.5)	755(94.5)
Positive	42(10.5)	44(5.5)
Agrarian vs. (Trading/Fishing)	0.002

Factors associated with RVFV seropositivity

Multivariable analysis identified several significant positive predictors of RVFV seropositivity: individuals residing in agrarian communities had higher odds of seropositivity than those in trading/fishing communities (adjusted odds ratio [aOR] = 2.34; 95% CI: 1.41–3.88, p = 0.001). Age 25–34 years (aOR = 2.00; 95% CI:1.03–3.89, p = 0.042) and ≥ 35 years (aOR = 2.91; 95% CI: 1.53–5.53, p = 0.001), compared with 15–24 years, were also associated with higher odds. Pig ownership was associated with a lower likelihood of RVFV exposure compared with non-owners (aOR = 0.40; 95% CI: 0.23–0.69, p = 0.001). Other factors, including gender, education, occupation type, bed net use, cattle and goat ownership, and cultivable land, were not significantly associated with seropositivity in the multivariable model (Table 3). Although RVFV is primarily mosquito-borne, and bed net use was generally high across communities, they offer limited protection against RVFV vectors, mainly Aedes and Culex mosquitoes, which often bite outdoors or during daylight and dusk hours. Since RVFV seroprevalence did not differ significantly between trading and fishing communities (p = 0.541), these two were combined into a single reference category for regression analyses to increase statistical power.

Table 3.

Variable-specific burden of RVFV and factors associated with RVFV

Variable	RVFV status % (n/N)	Univariate		Multivariable
		uOR (95% CIs)	p-value	aOR (95% CIs)	p-value
Community type
Trading/Fishing	5.5% (44/799)	Ref		Ref
Agrarian	10.5% (42/400)	2.01(1.30–3.13)	0.002	2.34(1.41–3.88)	0.001
Gender
Female	6.3% (38/607)	Ref		Ref
Male	8.1% (48/592)	1.32(0.85–2.05)	0.216	1.25(0.78-2.00)	0.352
Age (years)
15–24	3.4% (17/449)	Ref		Ref
25–34	7.4% (28/378)	2.03(1.09–3.77)	0.025	2.00(1.03–3.89)	0.042
≥ 35	11.0% (41/372)	3.15(1.76–5.64)	< 0.001	2.91(1.53–5.53)	0.001
Education*
None	10.5% (4/38)	Ref		Ref
Primary	9.0% (54/598)	0.84(0.29–2.47)	0.756	0.97(0.32–2.93)	0.955
Secondary/ Tertiary	5.0% (28/563)	0.44(0.15–1.34)	0.150	0.61(0.19–1.95)	0.407
Occupation
Outdoor/Animal-Related	8.3% (64/773)	Ref		Ref
Formal/Structured	4.8% (9/189)	0.55(0.27–1.13)	0.106	0.96(0.42–2.20)	0.920
Informal/Unstructured	5.5% (13/237)	0.64(0.35–1.19)	0.159	0.82(0.42–1.58)	0.554
Livestock Ownership
Own goats
No	7.1% (66/926)	Ref		Ref
Yes	7.3% (20/273)	1.03(0.61–1.73)	0.911	1.02(0.57–1.84)	0.934
Own Cattle
No	6.9% (73/1046)	Ref		Ref
Yes	8.5% (13/153)	1.24(0.67–2.29)	0.498	1.42(0.73–2.77)	0.300
Own pigs
No	8.5% (65/768)	Ref		Ref
Yes	4.9% (21/431)	0.55(0.33–0.92)	0.022	0.40(0.23–0.69)	0.001
Bed net use
Uses bed net	7.0% (75/1073)	Ref		Ref
Does not use	8.7% (11/126)	1.27(0.66–2.47)	0.475	0.85(0.43–1.68)	0.634
Cultivable land**
Has cultivable land	7.8% (66/846)	Ref		Ref
No land	5.7% (20/353)	0.71(0.42–1.19)	0.193	0.78(0.42–1.42)	0.410

uOR = Unadjusted Odds Ratio; aOR = adjusted Odds Ratio, CI = Confidence Interval; Ref = Reference category* Education is a confounder to age; **Cultivable land is a confounder to community typePost-hoc pairwise comparisons (Bonferroni-adjusted) for age, education, and occupation at the multivariable level did not reveal statistically significant differences between non-reference subgroups (all p > 0.05), indicating that the observed primary differences in exposure occur between the reference and the other categories, rather than between non-reference subgroups

Livestock ownership and community type

Livestock ownership (cows, goats, or pigs) and community type were assessed to examine how these factors relate to RVFV seropositivity (Table 4), given that livestock contact is a key exposure pathway for human infection. Individuals in Agrarian communities had significantly higher prevalence compared with those in Trading/Fishing areas, independent of livestock ownership. Goat and cattle ownership were not associated with seropositivity. Pig ownership was associated with a lower burden in Trading/Fishing areas, while no significant effect was observed in Agrarian communities.

Table 4.

Interaction of livestock ownership and community type with RVFV status

Livestock x Community	RVFV status % (n/N)		Univariate				Multivariable
			uOR (95% CIs)		p-value		aOR (95% CIs)		p-value
Goat ownership versus community type
No goats: Trading/Fishing		5.6% (37/664)		Ref				Ref
No goats: Agrarian		11.1% (29/262)		2.11(1.27-3.51)		0.004		2.36(1.40-3.97)	0.001
Goats: Trading/Fishing		5.2% (7/135)		0.93(0.40-2.13)		0.857		1.20(0.51-2.80)	0.673
Goats: Agrarian		9.4% (13/138)		1.76(0.91-3.41)		0.093		1.85(0.95-3.64)	0.073
Cattle ownership versus community type
No cattle: Trading/Fishing		5.3% (37/704)		Ref				Ref
No cattle: Agrarian		10.5% (36/342)		2.12(1.31-3.42)		0.002		2.29(1.40-3.73)	0.001
Cattle: Trading/Fishing		7.4% (7/95)		1.43(0.62-3.31)		0.399		1.66(0.70-3.89)	0.248
Cattle: Agrarian		10.3% (6/58)		2.08(0.84-5.15)		0.114		2.15(0.85-5.43)	0.104
Pig ownership versus community type
No pigs: Trading/Fishing		6.6% (38/577)		Ref				Ref
No pigs: Agrarian		14.1% (27/191)		2.34(1.38-3.94)		0.001		2.57(1.50-4.40)	0.001
Pigs: Trading/Fishing		2.7% (6/222)		0.39(0.16-0.95)		0.037		0.42(0.17-1.01)	0.054
Pigs: Agrarian		7.2% (15/209)		1.10(0.59-2.04)		0.77		0.94(0.46-1.92)	0.875

CI = Confidence Interval; Ref = Reference category; uOR = univariate odds ratio. aOR = adjusted odds ratio adjusted for age, education, gender, and occupation

Discussion

The overall RVFV IgG seroprevalence observed in this study was 7.8% (95% CI: 1.12–16.57), indicating prior exposure despite the absence of reported RVF cases in the study area.

The 7.8% estimate observed aligns with previous Ugandan studies, reporting comparable seropositivity between 10% and 13% [18, 19, 38, 39], also similar to the pooled seroprevalence of 7.8% across Africa reported in a recent meta-analysis [40]. Similar levels have been observed among livestock-dependent communities across East Africa [41–44]. In contrast, higher seroprevalence (12.8–39%) has been reported in Kenya and Tanzania, particularly among high-risk groups like slaughterhouse workers and pastoralists [45–48]. These elevated levels likely reflect intensified ecological and occupational exposures, including mosquito density, proximity to wetlands or flood-prone areas, variations in livestock farming practices, cultural practices such as pastoralism, consumption of raw animal products [49], and the timing of sampling relative to RVF outbreak periods, when recent transmission may increase seroprevalence estimates.

RVFV-specific IgG antibodies typically appear 1–2 weeks post-infection and can persist for two years or longer, or even a lifetime [2]. Therefore, it is unclear when the seropositive individuals in our study were exposed to RVFV. The persistence of IgG, combined with reservoir–host dynamics, suggests silent, undetected transmission in this region, with many cases being asymptomatic, mild, misdiagnosed, or entirely missed by the health system. Livestock such as cattle and goats can serve as early indicators of RVFV circulation, where unusually high abortions or neonatal mortality often precede human cases. This may give a warning to health care systems to consider diagnoses of RVF when patients exhibiting febrile illness arrive at the clinic. However, in the absence of forewarning livestock disease, the clinical symptoms of RVFV infection overlap with those of malaria, typhoid, and other viral hemorrhagic fevers, increasing the likelihood of misdiagnosis in endemic settings [50–52]. For example, during the 2016 RVF outbreak in Kabale District, Uganda, initial cases presented with non-specific febrile symptoms and were not recognized as RVFV infections until molecular testing confirmed the diagnosis [53].

A recent perspective by Rostal et al. proposes that RVFV transmission may exist as a spectrum with “epidemic” regions exhibiting notable but sporadic epidemics, whereas “hyperendemic regions” may have few or no recognized outbreaks despite sustained low-level circulation [54]. Our human seroprevalence data in the absence of reported RVFV outbreaks in the area suggests this region of Uganda may be “hyperendemic” for RVFV. However, further studies are needed to understand vector ecology and RVFV dynamics in livestock species in this region of Uganda. Such cryptic circulation contributes to a hidden disease burden, and reliance solely on reported or symptomatic cases may underestimate true prevalence and delay outbreak detection within this and similar regions.

In this study, community type was significantly associated with RVFV exposure. Individuals in agrarian communities had 2.34 times higher odds of seropositivity than those in trading or fishing areas, consistent with studies from Kenya, Tanzania, and Uganda linking elevated exposure to agricultural and livestock-related activities [39, 45, 49, 55]. Agrarian communities, by definition, rely on crop cultivation and livestock rearing, and are situated in areas with extensive arable land, introducing both occupational and ecological exposure risks. Routine practices such as herding, assisting with birthing, milking, and slaughter entail close, frequent contact with livestock and potentially infectious animal fluids. However, interaction analyses indicated that the increased odds observed in the agrarian areas persisted independent of individual livestock ownership. This suggests that additional factors, such as proximity to flood-prone areas, standing water, and dense vegetation that support year-round breeding of mosquitoes, also contribute to the risk [56–58]. Together, these occupational and ecological factors explain a high-risk interface for RVFV exposure observed in agrarian communities.

Individuals aged 25–34 years had 2 times the odds of exposure, and those aged 35 years and above had nearly 3 times the odds compared to the 15–24-year age group. This likely reflects that these age groups have had more cumulative exposure to infection, and not necessarily due to their increased biological susceptibility to RVFV. Similar findings have been reported across East Africa and Ethiopia [39, 59, 60], reflecting cumulative exposure through prolonged involvement in livestock activities and repeated contact with mosquito vectors. Pig ownership was associated with a lower burden of RVF, with households that kept pigs showing 0.4 times the odds of seropositivity compared with non-owners, particularly within Trading and Fishing communities. Although pigs are not considered major amplifying hosts for RVFV, studies report low susceptibility [61] and minimal viral replication in pigs [62, 63]. However, this pattern in our data may also reflect distinct household-level husbandry patterns for households with cattle and goats versus pigs, rather than a direct protective effect of pigs themselves.

In addition to RVFV, previous analyses of this cohort have revealed exposure to another zoonotic pathogen, CCHFV, with an overall seroprevalence of 4.25% [35]. Seropositivity for CCHFV was significantly associated with cow and pig ownership, reflecting the established role of livestock in CCHFV transmission [64, 65]. Community type also influenced CCHFV exposure, with fishing communities exhibiting higher seroprevalence than agrarian or trading communities. Notably, 11 individuals were seropositive for both CCHFV and RVFV (3, 7, and 1 in agrarian, fishing, and trading communities, respectively), highlighting overlapping exposure risks or shared ecological determinants.

This study has some limitations: (1) The cross-sectional design limits causal inferences; however, careful multivariable modeling was applied, and conclusions focused on associations rather than causality. (2) Using archived samples and data, some key variables like febrile illness history and mosquito bite frequency were not available for assessment, although, where possible, proxy variables that best approximated key exposures were carefully selected and interpreted within context. (3) Seroprevalence estimates were based solely on IgG detection, reflecting past exposure, including among symptomatic individuals, and were interpreted as cumulative exposure rather than current transmission intensity. (4) Selective survival bias may underestimate the true burden if severe or fatal cases that were absent could have died before sample collection. (5) Random error, especially in variables with low counts, may have diluted true associations; this was considered when interpreting exploratory analyses. (6) Our geographical scope was limited to selected communities in the greater Masaka-Rakai region, which restricts generalizability; however, ecological relevance to RVFV transmission was prioritized, and extrapolation was applied cautiously.

Conclusion

Our data indicate silent RVFV circulation in south-central Uganda without reported outbreaks or clinical RVF cases. Variation across community types suggests that exposure is strongly influenced by local ecological and livelihood-related factors. These findings highlight the need for strengthened, One-Health RVFV surveillance, integration of RVFV testing into routine febrile illness screening, improved clinical awareness, and integrated vector management strategies to enhance early detection and response in both outbreak and inter-epidemic periods. These findings also demonstrate the value of the RCCS as a platform for investigating zoonotic disease burden in Uganda.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

aOR

adjusted Odds Ratio

CCHFV

Crimean–Congo Hemorrhagic Fever Virus

CEU

Clinical Epidemiology Unit

CFR

Case Fatality Rate

CI

Confidence Interval

cELISA

competitive ELISA

ELISA

Enzyme–Linked Immunosorbent Assay

FAO

Food and Agriculture Organization

GPS

Global Positioning System

HRP

Horseradish peroxidase

IgG

Immunoglobulin G

IgM

Immunoglobulin M

OD

Optical Density

PBS

Phosphate Buffered Saline

RCCS

Rakai Community Cohort Study

RHSP

Rakai Health Sciences Program

RT

Room temperature

RVF

Rift Valley Fever

RVFV

Rift Valley Fever Virus

uOR

unadjusted Odds Ratio

Author contributions

DWH, HF, KR and DR developed the study concept. DWH, HF, DKB, SJR, JN and VS designed and coordinated the study. VS, SJR, RMG, RS, JN, GK, JO and EAM managed human sample collection and gathered study data. JN, VS, and GK did serological assays and performed data analyses. JN and DKB wrote the initial manuscript draft. DWH, KR, DR, HF, SJR, VS, JO, DKB, DM, MKG, and JN did substantive revision of the work. All authors contributed to writing and finalizing the manuscript and agreed to submit for publication. All authors reviewed the data.

Funding

This study has been supported in whole or in part by the National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) (U01AI075115, U01AI100031, ZIAAI001040, R01AI114438, K25AI114461, R01AI123002, K01AI125086, R01AI143333, R21AI145682, R01AI155080, AI167048-01), NIH National Institute of Child Health and Development (R01HD070769, R01HD091003), NIH National Heart, Lung, and Blood Institute (R01HL152813), National Institute of Mental Health (R01MH099733, F31MH095649, R01MH105313, R01MH107275, R01MH115799), National Institute on Alcohol Abuse and Alcoholism (K01AA024068), the Intramural Research Program of the NIH through the International Centers of Excellence in research (ICER), the Bill and Melinda Gates Foundation (OPP1084362), the Johns Hopkins University Center for AIDS Research (2P30AI094189), and the U.S. President’s Emergency Plan for AIDS Relief through the Centers for Disease Control and Prevention (NU2GGH00081). This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. 75N91019D00024. The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views or policies of the NIH or the U.S. Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

Data availability

The data used in this study can be requested from datarequests@rhsp.org.

Declarations

Ethics approval and consent to participate

The RCCS received approval from the Research and Ethics Committee of the Uganda Virus Research Institute (GC/127/19/03/709), the Uganda National Council for Science and Technology (HS-364), and the Johns Hopkins School of Medicine Institutional Review Board (IRB00204691) in accordance with the Declaration of Helsinki. All participants provided written informed consent prior to enrolment, and data was stripped of personally identifying information before it was transmitted from the RCCS servers to our study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.