Systematic review and meta-regression analysis of the prevalence of tick-borne pathogens in ticks and livestock in Uganda from 1980 to 2024

Abstract

Tick-borne pathogens (TBPs) cause diseases of significant impact on public health and livestock production in sub-Saharan Africa. In Uganda, livestock production contributes more than 17% of financial share of the agricultural sector. However, this venture is hampered by livestock diseases, especially tick-borne diseases (TBDs). A systematic review and meta-analysis were conducted to estimate the prevalence and distribution of TBPs in order to inform TBD control in Uganda. We retrieved published TBP prevalence articles (n = 6,446) for Uganda (January 1980 - August 2024) from five scientific databases, namely: Scopus, PubMed, Springer Nature, ScienceDirect and Web of Science. Following PRISMA guidelines, 36 studies were included in the meta-analysis. Heterogeneity was assessed using the Cochran’s Q statistic. The I² statistic and publication bias were evaluated using the Luis Furuya-Nakamori (LFK) index. Occurrence of TBPs was reported in 24 of the 146 districts of Uganda. Theileria parva (T. parva) was the most predominantly studied TBP representing 55.6% (20/36) of the studies with a national cattle prevalence of 29.2% (95% CI: 18.3–41.4), 62.6% (95% CI: 36.7–85.2) and 15.4 (95% CI: 0.78–42.5) based on DNA, serology and other parasite detection methods, respectively. Majority of the variation in T. parva prevalence in cattle could be explained by the diagnostic method used and sub-national region from which cattle originated. About 71% (17/24) of the districts represented across retrieved publications were located within Uganda’s cattle corridor. Data on TBP prevalence in small ruminants (goats and sheep) were very limited, representing only 13.8% (5/36) of the retrieved studies, despite the constraints posed by TBDs on small ruminant production in Uganda. More than half of all the studies retrieved reported T. parva, making it the most frequently reported TBP with an average prevalence of 29.2% in cattle. This is due to the fact that T. parva is the most important cattle TBP in Uganda. Further research on TBPs in data deficient districts, especially in the cattle corridor, is needed in order to support TBD control in Uganda.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-33956-x.

Background

Ticks and tick-borne diseases (TTBDs) pose a significant impact on nearly 80% of the 1.2 billion cattle population globally¹. In Sub-Saharan Africa (SSA), TTBDs contribute to an annual financial loss of about 20 billion USD especially in developing countries^2–4. In Uganda, where about 72.8% of the 9.3 million households engage in livestock farming contributing at least 17% of the total household income⁵, economic losses associated with TBDs are estimated to be over 1.1 billion USD per year⁶. Majority of losses are indirectly incured through reduced milk (187 million USD), and meat production (472 million USD), and blood loss (26 million USD). Additionally, importation of 378,000 L of acaricides and 83,000 L of associated drugs for treatment and control of TTBDs results in an annual foreign exchange outflow of over USD 83.3 million in Uganda⁷. Approximately 80% of total annual costs of controlling livestock disease are estimated to be linked to TTBD control⁸.

In Uganda, TBPs of veterinary importance include Theileria parva, Anaplasma marginale, Anaplasma centrale, Babesia bovis, Babesia bigemina and Ehrlichia ruminatium^9,10. Of these, Theileria parva (causative agent of East Coast Fever-ECF), is the most pathogenic and economically important TBP^6,11,12. Moreover, ECF alone threatens up to 28 million cattle in endemic areas in SSA^13,14. In Uganda alone, ECF is responsible for mortality in 30% of the calves and almost 100% in naïve exotic or crossbreed cattle¹⁵. Anaplasma infections rank the second most frequently reported TBD in Uganda¹⁶. Anaplasma spp. are primarily transmitted by the R. decoloratus ticks causing significant mortalities among adult cattle, goats and sheep^17,18. Additionally, Babesia infections transmitted by blue ticks [R. decolaratus and R. microplus ] are one of the most significant parasitic TBDs in Uganda^10,12,19 while Ehrlichia ruminatium that causes Heart Water also constraints cattle and goat production¹³.

Tick-borne pathogens of medical importance in Uganda include: rickettsial bacteria pathogens belonging to the spotted fever group (SFG)²⁰, Coxiella burneti²¹. and Rickettsia africae²². Rickettsia africae and Rickettsia conorii (R. conorii) have been previously detected in ticks collected from various Ugandan districts^23–27. Furthermore, a recent study detected R. conorii israelensis in Uganda for the first-time²⁵. Rickettsia conorii israelensis is the causative agent of Israel spotted fever in humans, and often more severe disease than MSF²⁸. The detection of spotted fever group rickettsioses in Uganda underscores a significant public health risk, highlighting a need for enhanced surveillance and control. On the other hand, Coxiella burneti, has also been previously detected in ticks from Uganda²⁹. Bacterial Coxiella burneti causes coxiellosis in animals and Q fever in humans. Abortions may occur in coxiellosis-infected animals, directly affecting livestock productivity²¹.

Prevention and control of TBDs in Uganda relies on the treatment of sick animals against TBDs and acaricides to control ticks. Treatment against ECF involves the use of buparvaquone and parvaquone³⁰ while for babesiosis and anaplasmosis, imidocarb dipropionate and diminazen diaceturate³¹ are the drugs of choice. However, due to high drug costs, farmers utilize oxytetracycline as a drug of choice for TBDs treatment³². Vaccination against ECF using the infection and treatment method (ITM) by the Muguga cocktail vaccine has been promoted³³. However, drawbacks, such as high costs of purchase and a need for cold chain, make the vaccine undesirable by farmers³⁴. As such, farmers have become more reliant on acaricide usage for the TTBDs control. This over-reliance on acaricides has resulted in tick acaricide resistance in Uganda³⁵. Widespread tick acaricide failure in Uganda has been reported to occur against the common classes of acaricides such as organophosphates, synthetic pyrethroids, co-formulations and amidines due to prolonged use and misuse by farmers^35,36. In a bid to solve the challenge of acaricide resistance, the control of livestock diseases through vaccine application is being promoted through development of anti-tick vaccines^37,38.

Prior to this study, previous systematic reviews investigated TBPs on the African continent. The occurrence of tick-borne pathogens in domestic ruminants and ticks in the southern African development community has been previously reviewed³⁹. More recently, the epidemiology of TBPs among cattle and ticks have been reviewed^40–43. However, to our knowledge, a comprehensive review on the prevalence of TBPs in Uganda is unavailable. Thus, the objective of this systematic review was to aggregate TBP prevalence data from multiple studies conducted across the country using a systematic approach. The information generated offers a comprehensive insight into the geographical distribution and prevalence of TBPs in Uganda, thus contributing to the development of targeted control strategies that can mitigate the impact of TBDs on animal and public health. More specifically, we estimated the prevalence of TBPs with explicit consideration of the cattle corridor⁴⁴ – an important and diverse regional stretch in Uganda where cattle density is > 50 cattle per km^{2 44} – due to its veterinary health implications.

Methods

Study design

This systematic review and meta-analysis was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines^45,46. A PRISMA flow chart showing the article selection procedure of studies has been developed and PRISMA checklist is included in the Supplementary information item 1.

A meta-analysis is a statistical approach that is used to combine and synthesize results from multiple independent studies on a common research question. The key steps in a meta-analysis are (1) the formulation of research question, (2) a literature search, (3) the selection of eligible studies, (4) determination the appropriate effect measure(s), (5) selection of the appropriate meta-analytic model (e.g., fixed-effects model or random effects model), (6) assessment of heterogeneity through subgroup analyses or meta-regression, (7) the conduction of formal analyses, and (8) a final reporting of pooled estimate results.

Literature search strategy

Peer-reviewed articles published between 1st January, 1980 and 31st August 2024 were searched and recovered from five electronic bibliographic databases, namely: Scopus, PubMed, Springer Nature, ScienceDirect and Web of Science. Literature searches across all databases were accomplished by combination of search terms to build queries. A list of search terms used is provided in the Supplementary Information Item 6.

Automated batch retrieval of peer reviewed articles was performed by accessing existing database application programming interfaces (APIs). Python modules PyMed⁴⁷ and Elpasy⁴⁸ were used to access PubMed, Scopus, and ScienceDirect databases. Web of Science and Springer Nature APIs were accessed through direct uniform resource locator (URL) requests^49,50. After removing duplicate records, the retrieved publications were then screened by the authors (Auther Tamale Wasswa and Karla Rascón-García), first by title and abstract with a final subset of articles screened full-text based on the set eligibility criteria. Data from included records were extracted by Auther Tamale Wasswa and entered into a spreadsheet.

Eligibility criteria

Publications were considered eligible if they met the following criteria: (i) the study was conducted in Uganda, (ii) written in English, (iii) published between 1st January 1980 and 31st August 2024, (iv) was a peer-reviewed published article, (v) was conducted on livestock of agricultural importance (cattle, goats and sheep) and ticks, (vi) was a study where a full-text document was available, (vii) the study reported sample size for livestock or total number of ticks tested, (viii) either the total number of animals positive for TBPs or the proportion of animals positive for TBPs, and (ix) specified the diagnostic technique used for detecting TBPs. Studies which did not meet these criteria were excluded. Additionally, articles that were reviews, meta-analyses, experimental or genetic-based studies, involving non-livestock (i.e., wildlife) were excluded. Grey literature was excluded to ensure data quality and reproducibility.

Assessment of study quality

To assess the quality of articles included in analyses, a critical appraisal tool developed by the methodological working group of the Joanna Briggs Institute (JBI)⁵¹, was adapted to assess the quality of articles reporting prevalence data. The JBI appraisal tool is composed of “yes”, “no”, “unclear” and “not applicable” responses to 9 critically framed questions. Responses were scored as 2 for “yes”, 1 for “unclear” and 0 for “no”. A scoring of “unclear” resulted when authors did not explicitly provide information for each appraisal criterion. For instance, when assessing whether a study population was sampled in an appropriate way, if authors did not explicitly state their sampling method (convenient sampling, randomized sampling, or cluster sampling) then a score of “unclear” was recorded. Studies were assessed across 8 of the 9 questions in the JBI appraisal tool; the modified tool used dropped one question which was not applicable to this work. Furthermore, average quality scores were calculated for each study included in analyses and reported in Supplementary information Item 3.

Data extraction

A standardized data extraction form was developed in Microsoft Excel^® version 16.0 (Microsoft Corporation, Redmond, WA, United States). Appropriate data were extracted from each eligible full text article and entered into the data extraction form. Data extracted included the following; name of publication’s first-author, publication year, the year when the study was conducted, host species, number of animals sampled, number of animals which tested positive, the tick-borne pathogen species detected, and diagnostic test used. Diagnostic methods used were categorized into three possible groups: parasite detection (microscopy), serology (enzyme-linked immunosorbent assay, immunochromatographic test strip, capillary agglutination test, and rapid card agglutination test) and DNA detection methods (polymerase chain reaction and reverse line hybridization). Furthermore, the tick species, number of ticks collected, and total positive tick pools were also recorded.

Data analysis

Meta-analysis

In this study, pooled prevalence estimates and their associated 95% confidence intervals were calculated for each TBP species reported and stratified by diagnostic method using a random effects model⁵² fit using a restricted maximum likelihood method (REML)⁵³. Data were transformed using a Freeman-Tukey Double arcsine transformation to stabilize the variance of proportions⁵⁴. Heterogeneity (between-study variation), was evaluated using the inverse variance index statistic (I²)⁵⁵ and significance determined using the Cochran’s Q test⁵⁶. Heterogeneity levels (I² index) range from 0 to 100%, where values ≤ 25%, 50%, and ≥75% describe low, moderate and high heterogeneity, respectively. Moreover, an I² of 0 is indicative of no heterogeneity or variation between studies. In addition, heterogeneity was considered significant when the Cochran’s Q test p-value was less than 0.05. Subgroup analyses were performed for each host species and further stratified by diagnostic method in order to obtain TBP species-specific prevalence estimates. Additionally, forest plots were generated to visualize pooled estimates for each meta-analysis.

Meta-regressions models

Univariable and multivariable meta-regression analyses were conducted on TBP scenarios with at least 5 publications (i.e., T. parva and A. marginale in cattle) to explore sources of heterogeneity. Moderators such as regions (Central, Western, Northern, and Eastern), study period and sample size were explored as possible explanatory variables. For T. parva in cattle, study periods were grouped into four classes (1998–2004, 2005–2011, 2012–2018, > 2019) while A. marginale in cattle were grouped into three classes (1998–2010, 2011–2017, > 2018). Additionally, study sample sizes were grouped into four classes for T. parva in cattle (< 200, 201–400, 401–600, > 601) and A. marginale in cattle were grouped in three classes (> 2018, 1998–2010, 2011–2017). Due to inconsistency in frequency of distribution of studies across years, the classes for study period and sample size were determined based on visual inspection of the frequency of articles over the years. For studies where the study year was not specified, the year of publication was assumed as the most probable year of collection. Covariates with a p-value less than 0.25 in univariable meta-regression analyses were explored in multivariable analyses.

Publication bias

Funnel plots were produced for each TBP species to investigate publication bias. More specifically, publication bias was assessed both, by visually inspecting funnel plot symmetry and by calculating the Luis Furuya-Kanamori (LFK) index values for each group⁵⁷. The LFK index is more sensitive and easier to interpret compared to funnel plot symmetry, which can be misleading when few studies are evaluated. LFK index values < -2 and > + 2 are interpreted as having detected major asymmetry, suggesting a possible publication bias.

Analyses were completed in R statistical software version 4.3.3⁵⁸ using the following packages: ‘meta’⁵⁹, ‘metafor’⁶⁰ and ‘metasens’⁶¹. District-level TBP prevalence estimates were mapped using Q-GIS version 3.16.14-Hannover⁶². Three maps were created using natural breaks (Jenks) with either 3 or 4 classes, one describing T. parva prevalence among cattle, A. marginale prevalence among cattle, and Rickettsiae spp. prevalence in ticks. These three maps constituted the epidemiologic scenarios for which there were at least 4 studies available to contribute data.

Results

Literature search results

Our literature search retrieved a total of 6,446 articles from 5 bibliographic databases (PubMed, Web of science, Scopus, ScienceDirect and Springer Nature). Following removal of duplicates (n = 191), two independent reviewers screened titles and abstracts of 6,255 articles, with disagreements resolved by consensus. A total of 6,022 articles were excluded by title and 167 by abstract, remaining with 66 articles which were eligible for full-text evaluation. During full-text evaluation, 30 articles were excluded leaving 36 articles which were included in analyses (Figure. 1).

Fig. 1.

Flowchart of literature search selection process in accordance to PRISMA guidelines.

Assessment of publication quality

When evaluating the quality of publications included in analyses, mean quality scores of each article ranged from 1.5 to 2.0; the mean score across all articles was 1.86 (Supplementary information item 3). Majority (72.2%; 26/36) of articles were unclear about whether an adequate sample size was used. Additionally, a considerable number of articles (47.2%;17/36) were unclear whether the study population was sampled appropriately. Overall, all articles included in this study were considered to be of appropriate quality standard. As such, no articles were excluded based on our quality assessment.

General attributes of studies included in the meta-analysis

Majority of studies reported TBP prevalence among cattle (26/36) followed by ticks (6/36), goats (5/36), and sheep (2/36) (Supplementary Table S1). Furthermore, all 36 studies sampled from 24 of the 146 districts were distributed across all four national sub-regions (central, northern, western and eastern) (Fig. 2). Of the 24 districts sampled, only seven districts (Gulu, Kampala, Mityana, Jinja, Iganga, Tororo and Butaleja) were outside the cattle corridor. Sample sizes across all host species (cattle, goats and sheep) ranged from 62 to 2,658 blood samples. In ticks, pools ranged from 20 to 859 pools with the total number of ticks ranging from 20 to 3,668 ticks per pool. The most prevalent diagnostic detection method used was DNA detection (n = 24) followed by serology detection methods (n = 8), and other parasite detection methods (n = 7). In cattle, DNA detection method accounted for the most test used to screen for TBP in blood samples (Table 1). Similarly, in goats and sheep DNA detection method was the most used followed by parasite and serology detection method while in ticks only DNA detection method was used (Supplementary Table S2).

Fig. 2.

Districts with published records of the occurrence of TBPs among livestock and ticks in Uganda between 1980–2024. The cattle corridor represents a region with cattle density of > 50 cattle per Km^{2 44}.

Table 1.

Pooled prevalence estimates of TBPs in cattle by diagnostic method.

Diagnostic method	TBP species/taxon	No. of Studies	Cases	Total samples	Prevalence % [95% CI]	I2	p-value
	Theileria parva	14	1634	8576	29.2 [18.3–41.4]	99.3	< 0.0001
	Anaplasma marginale	5	477	1241	36.7 [8.3–71.6]	99.4	< 0.0001
	Theileria spp.	3	653	1733	26.4 [1.5–66.5]	99.3	< 0.0001
	Theileria mutans	3	279	747	39.4 [1.70–88.5]	99.6	< 0.0001
	Theileria taurotragi	3	95	747	13.3 [0.00–11.2]	98.8	< 0.0001
	Theileria verifera	3	213	747	26.5 [0.5–70.7]	99.4	< 0.0001
	Theileria buffeli	2	29	507	5.7 [3.8–7.9]	0.0	0.3741
	Theileria sp. sable	1	19	363	5.2 [3.2–8.1]	-	-
	Theileria sp. buffalo	1	7	363	1.9 [0.8–3.9]	-	-
	Theileria bicornis	1	5	363	1.3 [0.5–3.2	-	-
	Theileria/Babesia genus-specific only	1	4	240	1.7 [0.5–4.2]	-	-
DNA	Ehrlichia/Anaplasma genus-specific only	1	12	240	5.0 [2.6–8.6]	-	-
	Anaplasma centrale	2	148	480	30.3 [12.3–52.2]	96.0	< 0.0001
	Anaplasma phagocytophilum	1	1	240	0.4 [0.01–2.3]	-	-
	Anaplasma bovis	1	13	240	5.4 [2.9–9.1]	-	-
	Anaplasma spp. Omatjenne	1	152	240	63.3 [56.9–69.4]	-	-
	Anaplasma platys-like	1	24	208	11.5 [0.8–16.7]	-	-
	Babesia spp.	1	338	1285	26.3 [23.9–28.8]	-	-
	Babesia vogelli	1	3	363	0.8 [0.02–2.4]	-	-
	Babesia bigemina	3	86	857	8.9 [4.5–14.5]	85.8	0.0009
	Anaplasma spp.	1	581	1285	45.2 [42.5–47.9]	-	-
	Theileria parva	6	1839	2817	62.6 [36.7–85.2]	99.4	< 0.0001
	Anaplasma marginale	2	718	1332	65.2 [18.0–98.4]	99.7	< 0.0001
Serology	Theileria mutans	1	374	935	40.0 [36.8–43.2]	-	-
	Babesia bigemina	2	151	857	10.6 [6.9–14.8]	78.9	0.0297
	Babesia bovis	1	25	401	6.2 [4.1–9.1]	-	-
	Anaplasma spp.	1	198	320	61.8 [56.3–67.3]	-	-
	Theileria parva	4	173	1091	15.4 [0.78–42.5]	99.1	< 0.0001
	Anaplasma marginale	2	16	552	4.9 [0.00–24.0]	96.3	< 0.0001
Parasite	Anaplasma spp.	4	1048	4736	14.9 [4.3–30.4]	97.5	< 0.0001
	Theileria spp.	1	28	186	15.1 [10.2–21.0]	-	-
	Babesia spp.	2	794	4230	12.2 [2.3–28.3]	97.9	< 0.0001

Tick-borne pathogen pooled prevalence in livestock and ticks

Tick-borne pathogen prevalence estimates were stratified by host species and diagnostic method. However, the pooled TBP prevalence estimates are majorly based on cattle, as few studies (5/36 studies) involving small ruminants (goats and sheep) were included in the analyses.

Tick-borne pathogen prevalence among cattle

In cattle, T. parva was the most predominantly studied TBP (n = 20). DNA detection methods estimated a national pooled prevalence of T. parva among cattle of 29.2% (95% CI: 18.3–41.4). Serological detection methods estimated a higher prevalence of 62.6% (95% CI: 36.7–85.2) and parasite detection methods a much lower prevalence (see Table 1) of 15.4% (95% CI: 0.78–42.5).

Similarly, A. marginale among cattle estimated a high prevalence when using serological methods (65.2%, 95% CI: 18.0-98.4), and 36.7% (95% CI: 8.3–71.6) by DNA detection. Babesia bigemina estimated a lower prevalence of 8.9% (95% CI: 4.5–14.5) by DNA and 10.6% (95% CI: 6.9–14.8) by serology. Subgroup analyses by diagnostic method for T. parva, (A) marginale and (B) bigemina among cattle were conducted to estimate the pooled prevalence of these TBPs in the different districts (Supplementary Table S3 – S5).

Tick-borne pathogen prevalence in ticks

Ticks were reported to harbor a variety of TBPs. Rickettsia spp. was the most predominant TBP detected with a prevalence of 55.6% (95% CI: 18.2–89.8) among ticks, followed by Theileria parva (25.0%, 95% CI: 12.6–41.2) and Babesia microti (10.0%, 95% CI: 2.8–23.7) all using DNA diagnostic methods. Other TBPs of significance detected in ticks included Ehrlichia ruminantium, Anaplasma centrale, Babesia bigemina and Babesia rossi (Table 1).

Subgroup analyses were conducted to evaluate the pooled prevalence of Rickettsia sp. by district. Rickettsia prevalence ranged from 14.0 to 100% (Supplementary Table S6). Tick-borne pathogens from ticks were detected in fourteen tick species belonging to the Ixodidae family namely: Rhipicephalus appendiculatus, Rhipicephalus evertsi evertsi, Rhipicephalus praetextatus, Rhipicephalus sanguineus, Rhipicephalus decoloratus, Rhipicephalus pulchellus, Rhipicephalus turanicus, Amblyomma variegatum, Hyalomma truncatum, Rhipicephalus pravus, Amblyomma gemma, Hyalomma rufipes, Amblyomma lepidum, and Haemaphysalis elliptica (Supplementary Table S7).

Forest plots displaying pooled prevalence estimates were produced for each TBP in cattle (A. marginale, T. parva, B. bigemina) and goats (Anaplasma spp.). Similarly, a forest plot for prevalence of Rickettsia spp. in ticks was produced (Fig. 3). Additionally, forest plots stratified by diagnostic method for each TBP are presented (Supplementary Figure S2).

Fig. 3.

TBP prevalence forest plots for (a) A. marginale in cattle (b) B. bigemina in cattle, (c) Rickettsia spp. in ticks, (d) T. parva in cattle and (e) Anaplasma spp. in goats. (b) Forest plot of B. bigemina prevalence estimates in cattle, (c) Forest plot of Rickettsia spp. prevalence in ticks, (d) Forest plot of T. parva prevalence estimates in cattle, (e) Forest plot for Anaplasma spp. prevalence estimates in goats.

(a) Forest plot of A. marginale prevalence estimates in cattle.

Forest plots visually display large variations in prevalence estimates across studies, ranging among animal hosts as low as 0.6% (Fig. 3a, A. marginale prevalence estimates in cattle) and as high as 79.13% (Fig. 3e, Anaplasma spp. prevalence estimates in goats). Similar heterogeneity was observed among ticks with prevalence estimates as low as 30.43% and as high as 97.14% (Fig. 3c, Rickettsia spp. prevalence in ticks).

Spatial distribution of TBP prevalence in cattle and ticks

The spatial distribution of TBP prevalence estimates by district, based on all diagnostic methods, was mapped using Q-GIS. Only T. parva and A. marginale among cattle and Rickettsiae spp. in ticks were mapped as they each had at least 4 studies contributing underlying data (Figure. 4).

Rickettsia species prevalence among ticks was highly variable with over 95% of ticks sampled from Soroti and Amuria districts estimating testing positive across studies (Fig. 4a). Though only seven publications informed A. marginale prevalence, the Karamoja region estimated a high prevalence with Kotido and Moroto districts estimating a prevalence over 70% among cattle (Fig. 4c).

Fig. 4.

TBP prevalence estimates by all diagnostic methods for (a) Rickettsia spp. in ticks (b) T. parva in cattle and (c) A. marginale in cattle. District-level prevalence estimate maps were produced using natural breaks (Jenks) with 3–4 classes. The cattle corridor represents a diverse geographical region with cattle density of > 50 cattle per km^{2 44}.

Publication bias

Publication bias was assessed using funnel plots and the LFK index. Moreover, publication bias was assessed by TBP and host, specifically among five groups: T. parva, B. bigemina, and A. marginale in cattle, Anaplasma spp. in goats and Rickettsia spp. in ticks. Minor publication bias (LFK index = 1.38) was observed for T. parva prevalence estimates in cattle and no publication bias was observed for (A) marginale in cattle and Anaplasma spp. in goats with LFK index values of -0.5 and 0.78, respectively. However, publication bias was detected for (B) bigemina in cattle and Rickettsia spp. in ticks (LFK index = -2.97 and 2.85, respectively). Funnel plots are presented in the Supplementary Figure S1.

Meta regression analyses

Meta-regression analyses were performed for TBPs which had more than 5 studies, specifically for T. parva and A. marginale in cattle. For other TBPs, meta-regression analysis was not possible because of few publications. A univariable meta-regression analysis for T. parva revealed diagnostic method and region informed most of the total variation in T. parva prevalence (R²) explaining 11.78% and 26.5%, respectively (Supplementary Table S8). Multivariable meta-regression analyses were performed by considering interactions and additive effects of independent variables (Table 2). The interaction between diagnostic method and region explained the highest proportion of heterogeneity in T. parva prevalence among cattle (R² = 52.49%, p < 0.0001).

Table 2.

Multivariable meta-regression analysis results for T. parva and A. marginale prevalence in cattle.

TBP	Variable	p-value	R2 (%)
T. parva	Region * Diagnostic method	< 0.0001	52.49
	Region + Diagnostic method	< 0.0001	45.50
A. marginale	Region * study period	< 0.0001	86.50
	Region * Diagnostic method	< 0.0001	96.17
	Region + Diagnostic method + study period	< 0.0001	85.28
	Region + Diagnostic method	< 0.0001	87.45
	Region + study period	< 0.0001	82.76
	Region * Diagnostic method* study period	< 0.0001	98.30

Note: an asterisk (*) indicates an interaction between variables, and plus (+) represents additive effect of variables.

A univariable meta-regression analysis for A. marginale prevalence among cattle suggested district and region as significant predictors influencing heterogeneity. Region alone informed 84.6% of the observed variation (Supplementary Table S9).). In a multivariable meta-regression, the interaction between region, study period and diagnostic method informed majority of the observed variations (98.3%) compared to other interactions (Table 2).

Discussion

Tick-borne pathogens cause TBDs that result in devasting effects to both animal and human health, particularly in endemic countries in SSA. This underscores the need for national-based studies which provide pooled estimates to better assess the burden of TBP in order to guide appropriate control measures. This systematic review and meta-analysis provide national pooled estimates for major TBP in livestock and ticks in Uganda and identifies specific geographic regions and host data inadequacies that can guide targeted control and surveillance of TBDs. This study identified 24 TBP species. Theileria parva was the most reported TBP in Uganda (20/36 studies) with a pooled prevalence in cattle of 29.2% (DNA detection methods), 62.6% (antibody detection) and 15.4% (parasite detection methods). This explains why ECF is most important TBD in Uganda^9,63.

Discrepancies in prevalence estimates reflect the differences in specificity and sensitivity of diagnostic methods used. Pathogen-specific prevalence estimates by host species and diagnostic tool were provided in Supplementary Table S2. Serological and DNA detection methods are more sensitive compared to parasite detection methods, which are usually hampered by low parasitemia and technical expertise⁶⁴. Therefore, studies relying exclusively on parasite detection methods may underestimate TBP infections. Furthermore, a variety of non-pathogenic TBPs such as Theileria spp. (T. mutans, T. verifera, T. separata), Babesia spp. (Babesia vogelli), and Anaplasma spp. (Anaplasma platys-like, Anaplasma spp. Omatjenne) have been detected in cattle and ticks in this study. Although these TBPs are known to be less pathogenic to cattle⁶⁵, their detection in mixed infections especially using serological based methods complicate diagnosis due to cross reactivity, thus overestimating the disease burden. As such, there is a need to integrate DNA based detection methods in TBD studies to enhance accurate diagnoses⁶⁶.

The high seroprevalence (62.6%) of T. parva in this study aligns with Uganda’s status, being situated in the ECF endemic zone. With ECF being endemic, coupled with suitable environmental conditions for tick vectors (R. appendiculatus) and the availability of wildlife reservoirs, such as African buffalos that can propagate transmission cycles. This, together with inadequate TTBDs control strategies have enabled the widespread occurrence of ECF in Uganda^19,67,68, thus stagnating livestock production in Uganda.

Anaplasma marginale estimated a moderate prevalence in cattle (36.7% DNA detection and 65.2% serology) underscoring its role in bovine anaplasmosis in Uganda. Similar to this study, previous systematic review of Anaplasma species detected a significant prevalence of A. marginale in cattle in Africa⁴², India⁶⁹, and SADC³⁹. Several tick species are known to biologically transmit A. marginale. In addition, mechanical transmission by hematophagous flies, such as Tabanus, Stomoxys and through contamination by blood fomites has also been reported^70,71. However, in Uganda, A. marginale is primarily transmitted by the Rhipicephalus (Boophilus) decoloratus ticks^13,16,19,72. Moreover, the displacement of R. decolaratus by a more vector competent and invasive R. microplus ticks, due to climate change and ecological drivers, may enhance the risk of bovine anaplasmosis in both endemic and non-endemic areas in Uganda⁷³.

Additionally, Babesia species cause serious public health challenges and pose economic constraints in livestock production. In the present study, B. bigemina prevalence in cattle was estimated at 8.9% by DNA detection and 10.5% by serology from 5/36 studies. The moderate prevalence of B. bigemina is likely influenced by multiple factors, including the extensive distribution of tick vectors (R. decolatarus)^19,74, seasonal, ecological and climatic changes, and acaricide resistance in Uganda³⁶. Noteworthy, information gaps still exist on the distribution, prevalence and transmission of B. bovis in cattle in Uganda.

Rickettsia spp. (55.6%) dominated the tick prevalence estimates. Similarly, a recent systematic review and meta-analysis on the distribution and prevalence of TBPs in Africa reported Rickettsia spp. with the highest prevalence in African ticks⁴⁰. Rickettsia species are known vectors of SFG rickettsioses. Moreover, R. africae the causative agent of African bite tick fever transmitted by A. variegatum, is reported as the most predominant Rickettsia spp. globally⁷⁵. Notably, the close cohabitation of livestock and humans in SSA predisposes farmers to tick bites. This highlights a strong risk of tick-borne rickettsioses, hence raising concerns about the potential public health impact.

Sub-national analyses revealed that T. parva prevalence in cattle was highest in central (45.9%) and western (45.3%) regions compared to northern and eastern regions of the country. This elevated prevalence of T. parva in cattle in central and western regions is due to the widespread tick acaricide resistance against the commonly used classes of acaricides, such as synthetic pyrethroids, co-formulation, organophosphates and amitraz³⁵. Notably, R. appendiculatus, the major vector of T. parva, has been reported to be resistant in at least 60% of the tick population in central and western Uganda³⁵. Additionally, agro-ecological factors, such as humid lowlands typical of these areas encourage tick survival and proliferation⁶. Suffice to say, these factors contribute to the increased transmission of T. parva in western and central regions of Uganda. Nonetheless, fewer studies have been conducted across districts in northern and eastern regions. As such, T. parva true prevalence is likely to be underestimated.

This study’s strengths are anchored in the nearly 44-year study time-window and the robustness of APIs which have allowed us to cast the widest net possible when requesting records related to TBDs in Uganda. The comprehensiveness of API pulls yield this work one of the most up-to-date representations of the epidemiologic state of TBDs in the country. Some study limitations in this study, however, were important findings as they identified knowledge gaps which future researchers should consider priority areas. Limitations, like very low publication records and incomplete spatial sampling for specific Anaplasma spp, Theileria spp, and Babesia spp among small ruminants and ticks, prevented us from being able to perform formal meta-analyses and consequently limited our ability to properly characterize the national epidemiologic state of TBDs in Uganda across all landscapes. These low publication counts influenced wide confidence intervals around prevalence estimates reported here while non-homogenous sampling of districts in Uganda, similarly, limited the generalizability of national or district-level estimates. Nevertheless, we recommend future researchers to use maps produced here to inform their future study regions as well as targeted host and pathogen species.

When evaluating publication bias, B. bigemina and Rickettsia species prevalence, publication bias was identified (LFK index > 2), but adjusted estimates were not explored through means like trim-and-fill because we believed doing so would still produce biased estimates as funnel plot asymmetry is more likely to be explained by alternative factors, like heterogeneity of land used in the underlying data. Thus, as it relates to prevalence estimates reported here, findings should not be considered robust as these results instead underscore a strong need to direct research towards these infectious pathogens across more regions in Uganda.

Only 17/24 districts sampled from were geographically located in the cattle corridor in Uganda. This stresses the need to screen livestock and ticks in the rest of cattle corridor districts with unavailable data on TBPs in Uganda. Subsequently, data on TBP prevalence in small ruminants (goats and sheep) was very limited (5/36 studies). In Uganda, small ruminants (21.8 million) constitute a large population than cattle (14.5 million) and greatly contribute to rural livelihoods and food security¹⁷. As such, there is need to conduct more research and surveillance on TBPs in small ruminants in order to support TBDs control in Uganda.

Conclusion

In this meta-analysis, we pooled prevalence estimates for TBPs in Uganda. Twenty-four (24) TBP species were reported in cattle, goats, sheep and ticks. Theileria parva in cattle was the most frequently studied TBP with a high prevalence, signaling its continued impact on livestock production in Uganda. Findings in this study corroborate earlier reports which highlight T. parva, (A) marginale and (B) bigemina as the most important TBPs in Uganda^6,10,18,63. However, there’s a need to conduct further epidemiological studies in data deficient districts, particularly in the cattle corridor, to provide baseline information to support TBD control in Uganda before actionable policies recommendations can be designed.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

K.R.G performed article search and retrieval; A.T.W screened articles, extracted data from included studies and performed the quality appraisal, conducted the data analysis and visualizations; K.R.G., A.T.W., and D.M. produced the original manuscript draft; A.T.W., K.R.G., and D.M conceptualized the study; JN, PMK., HA., and MK reviewed the manuscript and provided feedback on methods. All authors read and approved the final manuscript.

Funding

The authors received no funding for this work.

Data availability

The dataset analyzed in this study is available upon reasonable request from the corresponding author. Detailed descriptive statistics are provided in Supplementary Information.

Declarations

Competing interests

The authors declare no competing interests.

Supporting information

Supplementary information.