Background: Lumpy skin disease (LSD) is a devastating transboundary viral disease of cattle which causes significant loss in production. Although this disease has been reported in Uganda and throughout East Africa, there is almost no information about its epidemiology, spatial or spatio-temporal distribution. We carried out a retrospective study on the epidemiology of LSD in Uganda between the years 2002 and 2016, using data on reported outbreaks collected monthly by the central government veterinary administration. Descriptive statistics were computed on frequency of outbreaks, number of cases, vaccinations and deaths. We evaluated differences in the number of reported outbreaks across different regions (agro-ecological zones), districts, months and years. Spatial, temporal and space-time scan statistics were used to identify possible epidemiological clusters of LSD outbreaks. Results: A total of 1161 outbreaks and 319,355 cases of LSD were reported from 55 out of 56 districts of Uganda. There was a significant difference in incidence between years (P = 0.007) and across different regions. However, there was no significant difference in the number of outbreaks per month (P = 0.443). The Central region reported the highest number of outbreaks (n = 418, 36%) followed by Eastern (n = 372, 32%), Southwestern (n = 140, 12%), Northern (n = 131, 11%), Northeastern (n= 37, 3%), Western (n= 41, 4%) and Northwestern (n = 22, 2%) regions. Several endemic hotspots for the circulation of LSD were identified in the Central and Eastern regions using spatial cluster analyses. Outbreaks in endemic hotspots were less seasonal and had strikingly lower mortality and case- fatality rates than the other regions, suggesting an underlying difference in the epidemiology and impact of LSD in these different zones. Conclusion: Lumpy Skin disease is endemic in Uganda, with outbreaks occurring annually in all regions of the country. We identified potential spatial hotspots for LSD outbreaks, underlining the need for risk-based surveillance to establish the actual disease prevalence and risk factors for disease maintenance. Space-time analysis revealed that sporadic LSD outbreaks tend to occur both within and outside of endemic areas. The findings from this study will be used as a baseline for further epidemiological studies for the development of sustainable programmes towards the control of LSD in Uganda. Keywords: Lumpy skin disease, Epidemiology, Agro-ecological zones, Spatio-temporal epidemiology, Uganda * Correspondence: email@example.com College of Veterinary Medicine, University of Minnesota, 1365 Gortner Avenue, St. Paul, MN 55108, USA Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 2 of 12 Background occasionally be as high as 20 to 85%. The disease is Lumpy skin disease (LSD) is an acute to sub-acute viral therefore a serious threat to the cattle farming commu- disease of cattle that is defined by fever, increased nasal nity in endemic areas and is associated with trade re- secretions, enlarged lymph nodes, formation of nodules strictions following outbreaks. The livestock sector is on the skin, mucous membranes and internal organs, one of Uganda’s important growth sectors contributing edema of the skin and sometimes death [1–3]. Lumpy about US $290 million to the total GDP. Livestock con- skin disease virus (LSDV) is classified within the genus stitutes 17% of the agricultural GDP and is a source of Capripoxvirus in the family Poxviridae, which includes livelihood to about 4.5 million people in the country the closely related viruses of sheep pox and goat pox . . Trade in cattle hides generates about US $17 mil- Although cattle are the natural host of LSDV, clinical in- lion annually and has potential for continued growth if fection has been observed in the Asian water buffalo conditions like LSD, which lower the quality of hides from Egypt  and antibodies have been reported in and skins, can be managed . black and blue wildebeest, eland, giraffe, greater kudu, The success of any disease control program depends African buffalo, and other animal species [5, 6]. LSDV on a clear understanding of the epidemiology of the dis- neither infects nor is it transmitted between sheep and ease . This requires analysis of available data to goats . understand the distribution and patterns of spread of LSD has been reported in a number of regions of Africa, the disease . Little has been studied about the epi- where it is endemic, in the Middle East, and more recently demiology of LSD in Uganda, yet cattle farmers, district in parts of Europe. There is a potential risk that LSDV veterinary authorities, and monthly surveillance reports could spread further into Europe and eventually world- indicate the presence and impact of the disease in the wide [8–10]. In Eastern Africa, LSD was first reported in country. Therefore, this study was conducted to describe Kenya in 1957 , Sudan in 1972, and in Somalia in the temporal and spatial distribution of reported out- 1983 [12, 13]. There is no published literature about when breaks of LSD from 2002 to 2016 and to generate base- LSD was first identified in Uganda, however the disease is line epidemiological information on LSD in Uganda, thought to have spread from Southern Africa into Uganda which will facilitate further studies on disease prevalence between 1955 and 1960 . The disease is currently and risk factors. present in all geographical regions of the country, with several outbreaks reported annually. Methods Outbreaks of LSD tend to be sporadic, and are likely Study area dependent on animal movements, immune status of ani- Uganda is a landlocked country located on the East Afri- mals, and changes in weather patterns which affect vector can Plateau; it lies between latitudes 4°N and , and longi- populations. The main mode of transmission of LSDV is by tudes 29°35°E, with an area of about 241,038 km .It is mechanical arthropod vectors, such as biting flies, Aedes bordered by Kenya to the east, the Democratic Republic aegypti mosquitoes and three tick species belonging to the of the Congo to the west, South Sudan to the north, and family Ixodidae (Rhipicephalus (Boophilus) decoloratus, R. Rwanda and Tanzania to the south. The southern part of appendiculatus and Amblyomma hebraeum). Preda- the country includes a considerable portion of Lake tors, vermin and wild birds might also act as mechanical Victoria, which is shared with Kenya and Tanzania (Fig. 1). carriers of the virus [15, 16]. The virus can also be transmit- Uganda lies within the Nile basin as well as the African ted by fomites, such as equipment, clothing, and personnel Great lakes region, and has a diverse but generally equa- . Epidemics of LSD in non-endemic regions are re- torial climate. It is on average about 1100 m (3,609 ft) ported to be associated with hot and wet seasons, as well as above sea level. Currently, the country is divided into 112 areas close to water bodies, swamps and marshlands that districts; each district is sub-divided into counties and are conducive for breeding and multiplication of insects sub-counties; each sub-county consists of several parishes . Spread of LSDV between farms and districts might be and villages. Uganda has a number of national parks, how- due to the lack of complete restriction of animal move- ever the major parks considered in this study are: Bwindi ments [18, 19]. Impenetrable National Park (BINP), Kibale National Park LSD is known to cause substantial economic losses in (KINP), Kidepo Valley National Park (KVNP), Lake the form of severe emaciation, lowered milk production, Mburo National Park (LMNP), Mount Elgon National abortion, secondary mastitis, loss of fertility, extensive Park (MENP), Murchison Falls National Park (MFNP), damage to hides leading to low quality of leather and and Queen Elizabeth National Park (QENP). loss of draught power from lameness [13, 20]. The mor- bidity rate in cattle can vary from 3 to 85% depending Data source and collection on the presence of insect vectors and host susceptibility. RetrospectivedataonLSD outbreaks inUgandaduring Mortality usually ranges between 1 to 5% , but can 2002–2016 were retrieved from the Ministry of Agriculture Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 3 of 12 Fig. 1 Map of Uganda showing the location of Uganda in Africa (inset), national parks, international borders and the regions in this study (Source: This study) Animal Industry and Fisheries (MAAIF), Uganda. This in- grazing areas were also considered part of the same out- formation is based on the monthly disease surveillance re- break. New outbreaks were defined as those occurring in ports submitted to MAAIF by District Veterinary Officers a herd separated from other herds by a fence or physical (DVOs). For the period considered in this study, Uganda barrier such as hills, water bodies, forests or mountains. had a varying number of districts (56–112). Reporting of Each outbreak report contained data on the number of LSD outbreaks was at the district level, thus making it im- affected animals, susceptible animals, vaccinated ani- possible to disaggregate data from earlier years (56 districts) mals, and deaths. Since the exact locations of the into the present 112 districts. Thus, all the data analysed affected herds were not recorded, geographic coordi- here were aggregated into 56 districts consistent with 2002 nates of each district were defined as the centroid of the boundaries. Districts were classified as adjacent to inter- district. national borders and national parks. For data handling and presentation, districts were also grouped into seven geo- Data analysis graphical agro-ecological regions, which typically vary by Descriptive analysis was performed on outbreaks, cases, rainfall and farming production systems (Fig. 1). Data and vaccination data. Count data on number of outbreaks on livestock numbers were obtained from the Uganda Na- and cases were tested for normal distribution using tional livestock census report 2009 , and used to calcu- Shapiro-Wilk test and qqplots and found to be late cattle density. over-dispersed and positively skewed, with variance much In this paper, a case was defined as an animal with larger than the mean. Therefore, Kruskal-wallis chi-squared clinical signs or nodular lesions characteristic of LSD tests and post-hoc Dunn’s tests with Bonferroni corrections (with or without laboratory confirmatory diagnosis). An for p-values were carried out to assess whether the differ- outbreak was defined as the occurrence of one or more ences in number of outbreaks between regions, months cases of LSD in a particular herd. However, cases from and years were statistically significant. Spearman’scorrel- nearby herds with frequent animal contact or shared ation was used to investigate the relationship between cattle Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 4 of 12 density and number of outbreaks and cases. QGIS version Spatial distribution of LSD 2.18.9 with GRASS 7.2.1  was used to plot the distribu- The distribution of LSD at the district and regional level tion of LSD outbreaks per district (2002–2016) and to cre- was mapped (Fig. 2) to represent the spatial pattern of ate maps of the spatial and temporal distribution of LSD in outbreaks (2002–2016). The disease was reported in 55 Uganda. Software used for data analysis were Microsoft Of- out of 56 districts during this period. Lira (n = 84, 6.3%) fice Excel, 2013 and R version 3.4.2. and Tororo (n = 83, 6.3%) had the highest number of Purely spatial, purely temporal, and space-time scan outbreaks (2002–2016) while Kisoro (n = 1, 0.075%), statistical analyses were performed using SaTScan™ Mayuge (n= 1, 0.075%) and Ntungamo (n= 1, 0.075%) v9.4.4 [28, 29]. The purely spatial scan statistic imposes had the lowest numbers of outbreaks during the period a circular window of varying size upon the locations of studied. No LSD outbreaks were reported in Yumbe dis- possible outbreaks. The space-time scan statistic utilizes trict. There was a significant difference between the a dynamic cylindrical window, with a circular geographic numbers of outbreaks by region (P < 0.002), with the base and with height corresponding to time. The purely Central region (n = 418, 36%) reporting the highest num- temporal scan statistic uses a window with varying ber of outbreaks followed by the Eastern region (n = 372, height corresponding to time, in the same way the 32%), Southwestern region (n = 140, 12%), Northern re- height of the cylinder is used in the space-time scan. For gion (n = 131, 11%), Western region (n= 41, 4%), North- each scan, the number of outbreaks in the window is re- eastern region (n= 37, 3%), and Northwestern (n = 22, corded and compared to the null hypothesis of a random 2%) region. An additional file shows this in more detail Poisson distribution, accounting for population size. A [see additional file 1]. We found significant differences relative risk is calculated as the number of observed out- in number of outbreaks by region for the following pairs breaks within a window divided by the number of ex- of regions; Central-West (Dunn’s test, p = 0.004), pected outbreaks across the study area. The window Central-West Nile (Dunn’s test, p = 0.013), North-West with the maximum log likelihood ratio (LLR) is defined (Dunn’s test, p = 0.02), and North-West Nile (Dunn’s as the most likely cluster. LLR is calculated by test, p = 0.02). Spearman’s correlation showed a signifi- cant correlation between cattle density and number of outbreaks, and no significant correlation between cattle 2 ðÞ N−n n N−n density and cases respectively (r = 0.27, p-value = 0.04; LLR ¼ log I EnðÞ N−EnðÞ r = 0.12, p-value = 0.37 respectively). Seventeen (17) out of 56 districts adjacent to national parks reported only 45 (3.9%) outbreaks, while 332 (28.6%) outbreaks were where N is the total number of cases; n is the observed reported in districts adjacent to international borders. number of cases within the scan window; E(n) and N – E(n) When outbreaks in districts adjacent to national parks are the expected number of cases within and outside the were compared according to which national park they window under the null hypothesis, respectively, and I is an bordered, we observed that districts bordering Queen indicator function (equal to 1 when the window has more Elizabeth National Park (QENP) reported a higher num- cases than expected under the null hypothesis and 0 other- ber of outbreaks than those reported by districts border- wise). Here, scans were conducted for areas of high rates, ing the other six national parks (an additional file shows testing for elevated risk within a window as compared to this in more detail [see additional file 2]). outside. District centroids were tested as potential outbreak locations, and the maximum possible spatial and/or tem- Incidence of LSD outbreaks adjacent to the international poral cluster size was set to 50% of the total population at Borders risk. Monte Carlo simulation (n = 999 permutations) was The 22 of 56 districts adjacent to the international bor- used to determine the significance of detected clusters . ders reported 332 (28.6%) LSD outbreaks as compared to 829 (71.4%) outbreaks from districts with no inter- Results national border. The number of LSD outbreaks varied A total of 1161 LSD outbreaks were reported at the dis- between the different international borders, the highest trict level from January 1, 2002 to December 31, 2016, being adjacent with Kenya (157 outbreaks in 6 districts) with an average of 77 (± 51.4 SD) outbreaks per year and DRC borders (87 outbreaks in 7 districts), while 55 and a median of 70 outbreaks per year. During this outbreaks were reported in 2 districts bordering 15-year period, 319,552 cases were recorded, with an Tanzania and 21 outbreaks were reported in 4 districts average of 21,303 ± 4121 SD cases per year, and 2169 re- bordering South Sudan. The lowest number of LSD out- corded deaths (average of 146 ± 17 SD deaths per year) breaks was reported among the districts bordering attributed to LSD. Morbidity, mortality and case fatality Rwanda (12 outbreaks in 3 districts). Analysis of these rates were 4.77, 0.03 and 0.72%, respectively (Table 1). differences by Kruskal Wallis test however revealed no Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 5 of 12 Table 1 Average annual number of outbreaks, morbidity, mortality and case fatality rates in different regions of Uganda. Population at risk refers to the number of susceptible cattle in herds where at least one case was reported Region No. of Outbreaks Population at Risk No. of Sick No. of Dead Morbidity rate (%) Mortality rate (%) Case fatality rate (%) Central 28 306,452 5746 49 1.88 0.02 0.86 East 25 75,323 12,112 16 16.08 0.02 0.13 North 9 18,878 1440 15 7.63 0.08 1.02 North East 2 268 61 2 22.90 0.87 3.80 South West 9 27,786 1228 24 4.42 0.09 1.95 West 3 17,222 639 43 3.71 0.25 6.69 West Nile 1 647 77 4 11.97 0.64 5.34 Total 77 446,575 21,303 153 4.77 0.03 0.72 significant difference in the numbers of outbreaks per (n = 121) while the lowest annual incidence was reported international border (p-value = 0.41). in 2009 (n = 9) (Fig. 3). The highest incidence was re- ported in the month of January (n= 117 across all years), Temporal distribution of LSD which accounted for 10% of all outbreaks reported, and On average, 22 districts (± 9.8 SD) experienced out- the lowest in November (n = 80), accounting for 6.9% of breaks of LSD each year. High annual incidences of LSD all reported outbreaks. There was no significant differ- outbreaks were reported in 2002 (n = 182 outbreaks), ence in the incidence of outbreaks between months 2003 (n = 153), 2004 (n = 117), 2011 (n = 110) and 2012 (p = 0.443). When the overall data were grouped into Fig. 2 Map of Uganda showing district and regional distribution of LSD outbreaks (2002–2016), national parks and national borders. The size of the red circles indicate the respective number of LSD outbreaks in the areas marked (Source: This study) Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 6 of 12 Fig. 3 Total yearly Lumpy skin disease outbreaks in Uganda from 2002 to 2016 four seasons, two wet seasons and two dry seasons, the 31, 2005, with 383 observed outbreaks, compared to a highest incidence was reported in the first dry season calculated 137.97 expected outbreaks. The space-time (Dec–Feb, n = 312, 26.9%) followed by second dry sea- cluster is shown in Table 3 and Fig. 6. son (Jun–Aug, n = 300, 25.8%), first wet season (Mar– May, n = 286, 24.6%) and second wet season which had Purely temporal clusters of LSD the lowest incidence (Sep-Nov, n = 263, 22.7%). More Temporal cluster analysis of LSD outbreaks in Uganda marked intra-annual variation was observed when sub- showed one peak period with only one cluster identified dividing the analysis by region, with the northeastern during January 1, 2002 to December 31, 2004. The overall (Fig. 4d), western (Fig. 4f), and West Nile region show- RR within the cluster was 2.34 (LLR =85.92, P =0.001) ing more profound seasonal patterns (Fig. 4g). with 417 observed outbreaks compared to 226.24 ex- pected outbreaks. Purely spatial clusters of lumpy skin disease The spatial pattern of LSD was found to be nonrandom. Discussion A total of 7 clusters were identified, two (2) of which To understand the spatial epidemiology of lumpy skin were located in Central region, three (3) in Eastern re- disease (LSD) outbreaks in Uganda, we described the gion, one (1) in Southwestern region and one (1) in geographic and temporal occurrence of LSD and ana- Northern region (Fig. 5). The most likely cluster was ob- lyzed the data for spatial and temporal clusters using served in the Kalangala district in Central Uganda. The retrospective data collected between 2002 and 2016. radius of the cluster was 0 km, indicating that the cluster During this period, an average of 77 LSD outbreaks were only included Kalangala district. The relative risk (RR) reported across 22 (±9.8 SD) districts each year, demon- was 156.17, indicating that cattle within this district were strating that LSD is endemic in Uganda. around 156 times more likely to be affected by LSD than Incidence of reported LSD outbreaks differed between re- in areas outside the cluster (Table 2). The observed gions, with more outbreaks reported in the Central and number of outbreaks for this cluster was 66 compared Eastern regions as compared to the rest of the regions. The with a calculated 0.45 expected outbreaks. Secondary Central and Eastern regions represented more than half of clusters were located in (Luwero, Kayunga, Wakiso, and the reported LSD outbreaks during this time-frame. This Kampala), found in central Uganda; (Busia, Tororo), marked difference could be due to a number of factors in- Jinja, (Kapchorwa, Sironko, Mbale and Kumi) in Eastern cluding animal husbandry practices, presence of high num- Uganda; Kasese in Southwestern Uganda; and Lira in bers of insect vectors, higher frequency of exotic cattle Northern Uganda; (Table 2 and Fig. 5). RR for these breeds, awareness of disease control, uncontrolled animal clusters ranged from just over 1.8 to over 9. movements, and potentially biases in disease reporting re- lated to proximity to the central administrative center of Space-time clusters of lumpy skin disease MAAIF in Kampala [24, 31, 32]. However, the primary con- One space-time cluster was identified and it persisted tributor to the high rates of reported outbreaks in this re- for a duration of 3 years. This space-time cluster was lo- gion of Uganda may be climate, given that the Central and cated in 24 districts found in Eastern and Central region, Eastern regions of Uganda form part of the Lake Victoria and in 2 districts found in Northern region. This basin; these two regions also have other lakes (Kyoga, space-time cluster was from January 1, 2002 - December Opeta, and Bisina), rivers (Nile, Manafwa, Mpologoma, Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 7 of 12 ab cd ef Fig. 4 Spider plots showing the monthly distribution of LSD outbreaks per region from 2002 to 2016 Malaba) and wetlands which provide wet and humid distribution of LSD outbreaks and identified areas with micro-climates [33, 34]. This, coupled with an average high endemicity of LSD and clustering patterns using monthly temperature range of 22 °C–29 °C, provides suit- spatial scan statistics. We showed that in the period able conditions for multiplication of arthropod vectors for from 2002 to 2016 as a whole, the geographic distribu- LSD . The only published studies about arthropod vec- tion patterns of LSD outbreaks in Uganda were not ran- tors in this region are studies reporting distribution of Glos- dom. Spatial cluster analysis identified 7 clusters, which sina spp, [35, 36] which are known to transmit LSD were primarily located in the Central and Eastern re- mechanically . These studies have found high density of gions. The most likely spatial cluster was observed in these flies in the Central and Eastern regions of Uganda, Kalangala district in Central Uganda. High incidence in thus suggesting that the climatic conditions are suitable for this region is likely driven by climate and presence of arthropod vector multiplication and survival. wetlands in this district. Kalangala is a district made up Spatial, temporal and space-time scan statistics are of 84 islands surrounded by Lake Victoria, with 95% of tools used to detect aggregations of disease outbreaks or the district area covered by water bodies, and mean annual cases and identify whether these outbreaks or cases of rainfall ranging from 1125 to 2250 mm . The climate of disease in space or time can be explained by chance this district is generally moist and humid all through the alone or are statistically significant. Clusters may occur year with moderately small seasonal variations of due to local transmission of the disease or due to shared temperature, humidity, and wind throughout the year . risk factors within an area. We investigated the spatial These conditions are known to maintain arthropod vectors Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 8 of 12 Fig. 5 Purely spatial distribution of identified clusters of LSD cases with significantly higher incidences in Uganda from 2002 to 2016 (Source: This study) Table 2 SaTScan statistics for purely spatial clusters with significantly higher incidence of LSD in Uganda from 2002 to 2016 District/location Coordinates/radius Number of Expected Relative Log likelihood P-value outbreaks outbreaks risk ratio −17 Kalangala (0.320837 S, 32.293743 E) / 0 km 66 0.45 56.17 265.81 < 10 − 17 Busia, Tororo (0.470669 N, 34.091980 E) / 25.35 km 103 11.30 9.93 139.83 < 10 −17 Luwero, Kayunga, (0.840409 N, 32.497668 E) / 55.57 km 135 30.46 4.90 101.58 < 10 Wakiso, Kampala −10 Kasese (0.169899 N, 30.078078 E) / 0 km 34 7.50 4.64 25.19 < 2.3 × 10 −8 Jinja (0.447857 N, 33.202612 E) / 0 km 20 3.11 6.54 20.48 < 2.5 × 10 −6 Lira (2.258083 N, 32.887407 E) / 0 km 95 49.24 2.01 17.65 < 4.2 × 10 −5 Kapchorwa, Sironko, (1.335021 N, 34.397636 E) / 54.24 km 94 52.18 1.87 14.33 < 1.1 × 10 Mbale, Kumi Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 9 of 12 Table 3 SaTScan statistics for a space-time cluster with a significantly higher incidence of LSD in Uganda from 2002 to 2016 District/location Coordinates/radius Timeframe Number of Expected Relative Log likelihood P-value outbreaks outbreaks risk ratio −18 Kamuli, Kayunga, Iganga, Jinja, Pallisa, Luwero, (0.944785 N, 33.126717 E) / 2002.1.1 to 383 137.97 3.69 179.08 < 10 Mukono, Bugiri, Kaberamaido, Nakasongola, 168.37 km 2005.12.31 Kampala, Mayuge, Wakiso, Soroti, Kumi, Mbale, Busia, Mpigi, Tororo, Sironko, Apac, Kapchorwa, Lira, Katakwi, Kiboga, Kalangala which transmit LSD. District local government reports list thirteen of fourteen districts, identified as purely Lumpy skin disease among the most economically import- spatial clusters, were also identified as part of the ant Livestock diseases in the district , which is in agree- space-time cluster. However, one district (Kasese) iden- ment with our findings. Similar factors may also play a role tified by the purely spatial analysis did not appear in in creating the other hotspots identified in the spatial clus- the space-time cluster. The districts in the space-time ter analysis. cluster were found in Central, East and Northern parts We also conducted a space-time cluster analysis in of the country, and the duration of the associated addition to the purely spatial cluster analysis. A single space-time cluster was for four years. These areas thus space-time cluster was identified. When we compared appear to have experienced an epidemic wave of LSD the results of the purely spatial cluster analysis with for this four year period, occurring mainly within the those of the space-time cluster analysis, we found that endemic hotspots. Fig. 6 Space-time distribution of identified clusters (n = 5) of LSD cases with significantly higher incidences in Uganda from 2002 to 2016 (Source: This study) Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 10 of 12 We found that LSD occurs throughout the year with of the types of wildlife species present and the extent to outbreaks reported every month. In the more endemic which wildlife and livestock interact. It is notable that 20 areas around the Lake Victoria and Lake Kyoga basin of out of the total 45 outbreaks bordering national parks Uganda, there is rainfall throughout most of the year, (44.4%) were reported in districts bordering Queen providing hot and wet weather conditions which are Elizabeth National Park (QENP). QENP holds popula- conducive for breeding of biting flies which are known tions of African Buffalo (Syncerus caffer), kudu and to transmit LSD. During the dry season (December to waterbuck, which have previously been shown to have February), there is reduced availability of pasture and antibodies against LSDV [6, 39, 40] and therefore could water, so cattle are moved to swampy marsh lands which be potential hosts for the virus. It must however be are common in Central and Eastern regions of the coun- noted that three (3) other parks are inhabited by buffa- try and present in the other regions as well. These loes, and the extent of wildlife-livestock interaction may swampy areas maintain a hot and wet micro-climate vary in these parks thus limiting cross species transmis- which support large populations of biting insects; this sion of LSD. Neutralizing antibodies have previously together with a surge in cattle herds competing for lim- been detected in African Buffalo sera from QENP . ited grazing areas may lead to spread of LSD and there- Though this was in the 1980s, these findings suggest fore an increase in the number of new outbreaks that wildlife may play a role in the maintenance cycle of reported. The Central and Eastern regions showed no LSD. More research is needed to clarify the role of buf- seasonal pattern of LSD outbreaks, however the North- falo. Genotyping of LSD at wildlife-livestock interfaces, east, West Nile and West showed more seasonal pat- as well as at international borders, should be performed terns of outbreaks. A slight increase in the number of to determine the molecular epidemiology of the disease outbreaks was observed around the month of August for and shed more light on the effect of wildlife and the West and West Nile regions, and an increase in out- cross-border animal movements. When we compared breaks was observed in January and February for the outbreaks in districts adjacent to international borders, Northeast and West Nile regions (Fig. 4). This suggests we found that even if 47.3% of these 332 outbreaks were that seasonal factors have greater effects on incidence reported at the Kenya-Uganda border, this difference for these regions. These results further substantiate the was found not statistically significant when compared suggestion that the epidemiology of LSD may differ in with outbreaks from the four other international borders the endemic hotspots of Central and Eastern Uganda, of Uganda. characterized by less seasonality, presence of spatial The findings of this study should be interpreted with clusters as well as space-time clusters of outbreaks, and caution because of the potential bias related to underre- non-endemic zones that experience sporadic outbreaks porting of outbreaks and cases . In addition, the but no persistent circulation. Interestingly, endemic hot- cases were determined based on clinical signs with no spots (Central and Eastern) had strikingly lower mortal- confirmatory diagnostic tests, which may have led to ity and case-fatality rates than the other regions, which biases occurring from nonreporting of sub-clinical cases. further suggests an underlying difference in the disease’s Outbreak location information was at the district level, epidemiology and impact in these different zones. How- which therefore prevented more elegant spatial analyses ever, management and ecological factors could also im- and made it difficult to more precisely assess the role of pact the fatality rate of the disease. spatial proximity to international boundaries or national There were two evident temporal waves of LSD spread parks. More purposeful sampling schemes based on ac- during which high number of outbreaks were reported, tive surveillance and molecular epidemiology are needed spaced about ten years apart (Fig. 3). Temporal cluster to better resolve risk factors and dynamics of LSD analysis also identified the first of these two temporal spread in Uganda. waves, January 2002 to December 2004 as a period with heightened occurrence of LSD in Uganda. Low numbers Conclusions of outbreaks were reported in 2009, but we do not have Uganda’s hot and wet climate provides a conducive en- sufficient data to propose factors responsible for this vironment for biting arthropods which are known to occurrence. transmit LSD. In this study, we demonstrate that LSD is Uganda has seven (7) major game parks; these parks endemic in Uganda, with annual outbreaks in all regions are not fenced and it is therefore common for livestock of the country, albeit in varying incidence. We identified to graze with wildlife. While there have been few studies potential endemic hotspots for LSD outbreaks, highlight- elsewhere in Africa investigating the role of wildlife in ing the need for risk-based surveillance in these areas to the transmission of LSD, the 17 districts bordering na- establish the actual disease prevalence and risk factors tional parks accounted for only 3.9% (n = 45) of the total for maintenance of the disease. Our space-time analysis LSD cases reported. However, parks may vary in terms also revealed that sporadic LSD outbreaks tend to occur Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 11 of 12 within endemic hotspot areas. Interestingly, endemic Department of Veterinary Population Medicine, University of Minnesota, CN is an MSc candidate with interest in epidemiology and molecular biology, hotspots had less seasonality in incidence and strikingly AM is a PhD candidate studying epidemiology of animal viruses in Uganda, lower mortality and case-fatality rates than the other re- ARAO is the assistant commissioner diagnostics and epidemiology at the gions, suggesting that epidemiology and impact of LSD Ministry of Agriculture Animal Industry and Fisheries Uganda, NN is the assistant commissioner disease control at the Ministry of Agriculture Animal may vary within and outside these hotspots. Based on Industry and Fisheries Uganda, RM is a Senior Veterinary Officer- our findings, we suggest that true prevalence of the dis- epidemiology at the Ministry of Agriculture Animal Industry and Fisheries ease, and viral genotypes, should be determined in order Uganda, FNM is a professor of Veterinary virology at Makerere University with vast experience in epidemiology and diagnosis of livestock diseases. to inform appropriate control measures in these en- demic hotspots, such as vaccination, to prevent further Ethics approval and consent to participate spread of the disease. LSD should be included amongst Written consent was obtained from the Ministry of Agriculture Animal the priority cattle diseases in Uganda, where regular sur- Industry and Fisheries (MAAIF), department of Epidemiology to carry out this study, using data collected by them. veillance and vaccination are done by the government. Our findings provide a baseline for further studies into Competing interests the epidemiology of LSD in Uganda and East Africa. The authors of this paper do not have any financial or personal relationship with other people or organisations that could inappropriately influence or Additional files bias the content of the paper. The authors therefore declare that they have no competing interests in the publication of this paper. Additional file 1: Mean annual Lumpy skin disease outbreaks across different regions (agro-ecological zones) from 2002 to 2016. The mean Publisher’sNote annual Lumpy skin disease outbreaks reported in the Central, East, North, Springer Nature remains neutral with regard to jurisdictional claims in Northeast, Southwest, West and Westnile regions of Uganda from 2002 published maps and institutional affiliations. to 2016. (DOCX 33 kb) Additional file 2: Occurrence of LSD outbreaks in districts adjacent to Author details national parks in Uganda 2002–2016. This table shows the yearly number College of Veterinary Medicine, Animal resources and Biosecurity, Makerere of Lumpy skin disease outbreaks reported in districts bordering each of University, P.O.BOX 7062 Kampala, Uganda. College of Veterinary Medicine, the seven major national parks in Uganda. A total of forty five outbreaks University of Minnesota, 1365 Gortner Avenue, St. Paul, MN 55108, USA. were reported, notably twenty out of these forty five outbreaks are from Ministry of Agriculture Animal Industry & Fisheries, Berkley Ln, Entebbe, districts bordering Queen Elizabeth national park. (DOCX 13 kb) Uganda. Received: 22 January 2018 Accepted: 24 May 2018 Abbreviations BINP: Bwindi Impenetrable National Park; DVO: District Veterinary Officer; KINP: Kibale National Park; KVNP: Kidepo Valley National Park; LMNP: Lake Mburo National Park; LSD: Lumpy Skin Disease; MAAIF: Ministry of Agriculture References Animal Industry and Fisheries; MENP: Mount Elgon National Park; 1. Babiuk S, Bowden TR, Parkyn G, Dalman B, Manning L, Neufeld J, et al. MFNP: Murchison Falls National Park; QENP: Queen Elizabeth National Park Quantification of lumpy skin disease virus following experimental infection in cattle. Transbound Emerg Dis. 2008;55:299–307. Acknowledgements 2. OIE Terrestrial Manual. Aetiology epidemiology diagnosis prevention and The authors would like to acknowledge the staff of the Ministry of control references. Oie. 2012:1–5. Agriculture, Animal Industry and Fisheries for the valuable information on 3. Abutarbush SM, Ababneh MM, Al Zoubi IG, Al Sheyab OM, Al Zoubi MG, Lumpy skin disease especially Ms. Esther Nambo for valued support offered Alekish MO, et al. Lumpy skin disease in Jordan: disease emergence, clinical while accessing the data archives at the department of epidemiology. signs, complications and preliminary-associated economic losses. Transbound Emerg Dis. 2015;62:549–54. Funding 4. Ali AA, Esmat M, Attia H, Selim A, Abdel-Hamid YM. Clinical and pathological This research was funded by the University Of Minnesota Academic Health studies of lumpy skin disease in Egypt. Vet Rec. 1990;127:549–50. Center and they had no role in the design and execution of the study as 5. Hedger RSHC. Neutralising antibodies to lumpy skin disease virus in African well as the decision to publish this manuscript. wildlife. Comp Immunol Microbiol Infect Dis Microbiol Infect Dis. 1983;6: 209–13. Availability of data and materials 6. Fagbo S, JAW C, Venter EH. Seroprevalence of Rift Valley fever and lumpy The datasets used and/or analysed during the current study are available skin disease in African buffalo (<i>Syncerus caffer</i>) in the Kruger from the corresponding author on reasonable request. National Park and Hluhluwe-iMfolozi Park, South Africa. J S Afr Vet Assoc. 2014;85:1–8. Available from: http://www.jsava.co.za/index.php/jsava/article/ Authors’ contributions view/1075 SO contributed to the conception of the idea, design, and data collection, 7. OIE. Lumpy skin disease. OIE Terr. Anim. Heal. Code [Internet]. 2016;1–4. drafting and writing of the manuscript. KVW contributed to statistical analysis, Available from: http://www.oie.int/fileadmin/Home/eng/Animal_Health_in_ interpretation of results and manuscript preparation. CN contributed to statistical the_World/docs/pdf/Disease_cards/SHEEP_GOAT_POX.pdf. analysis and drafting of the manuscript. AM contributed to design and writing of 8. FG D. Lumpy skin disease. Virus Dis. Food Anim. Gibbs EPJ (ed), Acad. Press. the manuscript, ARAO contributed to data collection and writing of the manuscript. London 1981;2:751–764. NN contributed to data collection and writing of the manuscript. RM contributed 9. Al-Salihi KA, Hassan IQ. Lumpy skin disease in Iraq: study of the disease to data collection and writing of the manuscript. FNM contributed to conception of emergence. Transbound Emerg Dis. 2015;62:457–62. the idea, design and writing of the manuscript. All authors read and approved the 10. Tuppurainen ESM, Venter EH, Shisler JL, Gari G, Mekonnen GA, Juleff N, et al. manuscript. Review: Capripoxvirus diseases: current status and opportunities for control. Transbound Emerg Dis. 2017;64:729–45. Authors’ information 11. Burdin ML, Prydie J. Lumpy Skin Disease of Cattle in Kenya. Nature. 1959; SO is a PhD candidate with interest in epidemiology and diagnostics for 183:55–6. Available from: https://doi.org/10.1038/183055a0 Animal and zoonotic viruses. KVW is an assistant professor at the 12. Nawathe, Paden J, Confl R. In Nigeria. Pieleg Polozna 1982;36:19, 25. Ochwo et al. BMC Veterinary Research (2018) 14:174 Page 12 of 12 13. FG D. Lumpy skin disease of cattle: a growing problem in Africa and the 36. Albert M, Wardrop NA, Atkinson PM, Torr SJ. Tsetse Fly ( G . f . fuscipes ) near east [internet]. 1991. Available from: http://www.fao.org/ag/aGa/agap/ Distribution in the Lake Victoria Basin of Uganda. 2015;1–14. Available from: frg/feedback/war/u4900b/u4900b0d.htm. https://doi.org/10.1371/journal.pntd.0003705. 14. Lubinga JC, Clift SJ, Tuppurainen ESM, Stoltsz WH, Babiuk S, Coetzer JAW, 37. Smith J. Lumpy skin disease in Bulgaria and Greece. 2016. et al. Demonstration of lumpy skin disease virus infection in Amblyomma 38. Kalangala District Local Government. State of Environment Report [Internet]. hebraeum and Rhipicephalus appendiculatus ticks using 2005. Available from: http://www.nemaug.org/district_reports/Kalangala_ immunohistochemistry. Ticks Tick Borne Dis. 2014;5:113–20. Elsevier GmbH. DSOER_2004.pdf. Available from: https://doi.org/10.1016/j.ttbdis.2013.09.010 39. Davies FG. Observations on the epidemiology of lumpy skin disease in Kenya. J Hyg (Lond). 1982;88:95–102. Available from: https://www.ncbi.nlm. 15. Kitching RP, Mellor PS. Insect transmission of capripoxvirus. Res Vet Sci. nih.gov/pmc/articles/PMC2134151/. 1986;40:255–8. 40. Hedger RS, Hamblin C. Neutralising antibodies to lumpy skin disease virus in 16. Australian Veterinary Emergency Plan (AUSVETPLAN). Animal health african wildlife. Comp lmmun Microbiol Infect Dis. 1983;6(3):209–13. Australia. Disease strategy: Lumpy skin disease (Version 3.0). 2009. 41. Thrusfield M. Veterinary Epidemiology. Third Edit. Blackwell Sci. Ltd. Wiley; 17. European Food Safety Authority. Lumpy skin disease: I. Data collection and 2005. p. 170–171. analysis. EFSA J. 2017;15:54. 18. Tuppurainen ESM, Oura CAL. Review: lumpy skin disease: an emerging threat to Europe, the Middle East and Asia. Transbound Emerg Dis. 2012;59:40–8. 19. Tuppurainen ESM, Venter EH, Coetzer JAW. The detection of lumpy skin disease virus in samples of experimentally infected cattle using different diagnostic techniques. Onderstepoort J Vet Res. 2005;72:153–64. 20. Molla W, MCM d J, Gari G, Frankena K. Economic impact of lumpy skin disease and cost effectiveness of vaccination for the control of outbreaks in Ethiopia. Prev Vet Med. 2017;147:100–7. Elsevier. Available from: https:// www.sciencedirect.com/science/article/pii/S0167587717303999?via%3Dihub 21. Behnke R, Nakirya M. The contribution of livestock to the Ugandan economy. IGAD Livest. Policy Initiat. Work. Pap. [Internet]. 2012;1–37. Available from: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1. 366.4644&rep=rep1&type=pdf. 22. Uganda Investment Authority. Uganda livestock sector profile. In: Uganda Investiment Auth; 2014. 23. Dukpa K, Robertson ID, Edwards JR, Ellis TM. A retrospective study on the epidemiology of foot-and-mouth disease in Bhutan. Trop Anim Health Prod. 2011;43:495–502. 24. Ayebazibwe C, Tjørnehøj K, Mwiine FN, Muwanika VB, Ademun Okurut AR, Siegismund HR, et al. Patterns, risk factors and characteristics of reported and perceived foot-and-mouth disease (FMD) in Uganda. Trop Anim Health Prod. 2010;42:1547–59. 25. Basalirwa CPK. Delineation of Uganda into climatological rainfall zones using the method of principal component analysis. Int J Climatol. 1995;15: 1161–77. 26. The Ministry Of Agriculture AIAF. The National Livestock Census a Summary Report of the National Livestock Census. 2009. 27. QGIS Development Team. QGIS Geographic Information System. v 2.18.7- Las Palmas. Open Source Geospatial Found. Proj. 2015. 28. Kulldorff M. A spatial scan statistic. Commun. Stat - Theory Methods. 1997; 26:1481–96. 29. Kulldorff M, Heffernan R, Hartman J, Assunção R, Mostashari F. A space-time permutation scan statistic for disease outbreak detection. PLoS Med. 2005;2: 0216–24. 30. Meyer D. Modified Randomization Tests for Nonparametric Hypotheses. Ann Math Stat. 1957;28:181–7. Available from: http://www.jstor.org/stable/ 31. Alexandersen S, Mowat N. Foot-and-mouth disease: host range and pathogenesis. Curr Top Microbiol Immunol. 2005;288:9–42. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15648173 32. Kalenzi Atuhaire D, Ochwo S, Afayoa M, Norbert Mwiine F, Kokas I, Arinaitwe E, et al. Epidemiological overview of African swine fever in Uganda (2001– 2012). J Vet Med. 2013;2013:1–9. Available from: http://www.hindawi.com/ journals/jvm/2013/949638/. http://downloads.hindawi.com/journals/jvm/ 2013/949638.pdf%5Cn. 33. UBOS. 2004 Statistical Abstract 2004;256. 34. Nsubuga FNW, Namutebi EN, Nsubuga-Ssenfuma M. Water Resources of Uganda: An Assessment and Review. J. Water Resour. Prot. Water Resour. Uganda An Assess. Rev. J. Water Resour Prot. [Internet]. 2014;6:1297–315. Available from: http://www.scirp.org/journal/jwarp. http://dx.doi.org/10. 4236/jwarp.2014.614120. 35. Nakato T, Jegede OO, Ayansina A, Olaleye VF, Olufemi B. Mapping the Distribution of Tsetse Flies in Eastern Uganda. Geogr. Inf Syst [Internet]. 2013;938–51. Available from: http://sci-hub.tw/https://www.igi-global.com/ journal/international-journal-ictresearch-development/1172.
BMC Veterinary Research – Springer Journals
Published: Jun 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera