Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

The Burden of Cholera in Uganda

The Burden of Cholera in Uganda Introduction: In 2010, the World Health Organization released a new cholera vaccine position paper, which recommended the use of cholera vaccines in high-risk endemic areas. However, there is a paucity of data on the burden of cholera in endemic countries. This article reviewed available cholera surveillance data from Uganda and assessed the sufficiency of these data to inform country-specific strategies for cholera vaccination. Methods: The Uganda Ministry of Health conducts cholera surveillance to guide cholera outbreak control activities. This includes reporting the number of cases based on a standardized clinical definition plus systematic laboratory testing of stool samples from suspected cases at the outset and conclusion of outbreaks. This retrospective study analyzes available data by district and by age to estimate incidence rates. Since surveillance activities focus on more severe hospitalized cases and deaths, a sensitivity analysis was conducted to estimate the number of non-severe cases and unrecognized deaths that may not have been captured. Results: Cholera affected all ages, but the geographic distribution of the disease was very heterogeneous in Uganda. We estimated that an average of about 11,000 cholera cases occurred in Uganda each year, which led to approximately 61–182 deaths. The majority of these cases (81%) occurred in a relatively small number of districts comprising just 24% of Uganda’s total population. These districts included rural areas bordering the Democratic Republic of Congo, South Sudan, and Kenya as well as the slums of Kampala city. When outbreaks occurred, the average duration was about 15 weeks with a range of 4– 44 weeks. Discussion: There is a clear subdivision between high-risk and low-risk districts in Uganda. Vaccination efforts should be focused on the high-risk population. However, enhanced or sentinel surveillance activities should be undertaken to better quantify the endemic disease burden and high-risk populations prior to introducing the vaccine. Citation: Bwire G, Malimbo M, Maskery B, Kim YE, Mogasale V, et al. (2013) The Burden of Cholera in Uganda. PLoS Negl Trop Dis 7(12): e2545. doi:10.1371/ journal.pntd.0002545 Editor: Edward T. Ryan, Massachusetts General Hospital, United States of America Received April 13, 2012; Accepted October 4, 2013; Published December 5, 2013 Copyright:  2013 Bwire et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Financial support was provided by the Bill and Melinda Gates Foundation through the Cholera Vaccine Initiative (CHOVI) program, administered by the International Vaccine Institute (IVI), Seoul, Korea. Current donors providing unrestricted support to the IVI include the Governments of Kuwait, Republic of Korea and Sweden. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected] ¤ Current address: Division of Global Migration and Quarantine, United States Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America. occur any time, but is often associated with extreme rain events or Introduction water supply disruptions. Cholera was first reported in Uganda in 1971, when 757 cases The frequency of reported cholera cases varies among districts were reported to the World Health Organization (WHO). During in Uganda. The highest risk areas include the border areas with the subsequent years up to 1993, Uganda reported cholera cases the Democratic Republic of Congo (DRC), Sudan, and Kenya as every 2–4 years to the WHO. From 1994 to 1998, cholera was well as urban slums in Kampala. Displaced populations and their reported annually in Uganda [1]. In 1998, Uganda reported neighboring communities are at elevated risk. The ongoing almost 50,000 cases with incidence throughout the country [2]. migration of people into and within Uganda can lead to rapid The reported incidence has fluctuated between 250 and 5,000 spread of the disease. The African Great Lakes, including Lake cases every year since 2000 (Figure 1). The reported case fatality Albert and Lake Victoria on the border of Uganda, may provide a ratio has decreased from 4–7% in the late 1990s to about 2–3% reservoir for cholera bacteria. Further, increases in incidence during 2004–2010. among the nations bordering these lakes have been shown to be Cholera in Uganda appears to be largely an epidemic disease. correlated with El Nin˜o warm weather events [3]. The WHO [4] However, endemic cholera occurs in high-risk areas along the has recently revised its guidelines and states in a position paper southwestern border with DRC and in Kampala city slums. that cholera vaccines should be used in combination with other Endemic cholera is commonly noted before and during the rainy prevention and control strategies in areas where the disease is season, from December through March. Epidemic cholera can endemic. ‘‘Endemic’’ is defined as areas with occurrence of PLOS Neglected Tropical Diseases | www.plosntds.org 1 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda In endemic areas: ‘‘a patient aged 2 years or more develops Author Summary severe dehydration or dies from acute watery diarrhea.’’ [6] Uganda has reported cholera cases to the World Health The identification of such cases should have triggered labora- Organization every year since 1997. Thus, the country may tory investigation. A cholera outbreak was confirmed when Vibrio consider the introduction of a WHO-prequalified oral cholerae O1 or O139 was isolated from at least one stool sample. cholera vaccine. This article reviews cholera surveillance data from 1997–2010 with a focus on the 2005–2010 time Only cases meeting the standard case definition above were period to identify high risk populations that may be investigated and included in the official cholera data. targeted for preventive vaccination campaigns. We esti- mated that an average of about 61–182 deaths occur each Laboratory methods year. Most cases (81%) occurred in a relatively small Summary laboratory data were obtained from the Head of the number of districts comprising just 24% of Uganda’s total Central Public Health Laboratory from the Ministry of Health. population. While there is a clear distinction between low Prior to analysis, stool samples from suspected cholera patients and high-risk districts, sentinel surveillance would help to were transported from the field in Cary Blair transport media. better quantify the burden in endemic districts. An Culture plates were set at 37uC overnight (for 18–24 hours) using economic analysis should also be undertaken prior to three culture media: TCB, XLD and MacConkey. Biochemical making a decision to introduce a cholera vaccine. identification of cholera organisms were based on Oxidase or Indole tests. Polyvalent antisera were used to differentiate between the Inaba and Ogawa serotypes, and specific monovalent tests culture-confirmed cholera in at least three of the previous five further confirmed which of the Inaba, Ogawa or O139 (Bengal) years. The WHO also recommends that cholera vaccines should strains caused disease. Isolates were refrigerated at 280uC and be considered for preemptive use in areas at risk for epidemic sent to the WHO’s collaborating laboratory (Unite´ de La Rage, cholera as long as vaccination does not interfere with efforts to Institut Pasteur, Paris, France) for quality control. treat cholera cases, improve water and sanitation, and mobilize communities. The vaccine may also be considered for reactive use if local infrastructure is sufficient to conduct mass campaigns Data collection depending on the current and historical epidemiology. District-specific data were abstracted from the Uganda Ministry of Health, Health Management Information System disease There is a dearth of information about the burden of cholera in low-income countries such as Uganda. A more accurate picture of surveillance database for the period 2005–2010. The 2005–2010 period was chosen based on the WHO criteria [4] for identifying this burden is particularly important because it can be used to endemic cholera (i.e. areas in which cholera has been reported in inform cholera prevention and control intervention questions: three of the previous five years.) The districts shown are based on whether or not to introduce vaccination as a complement to other the 2002 district boundaries, which were in existence during the cholera prevention and control interventions, where and when it most recent census. Cases in new districts created after 2005 were would be most effective to do so, and what demographic apportioned back to the 2002 districts. population should be targeted. This article presents available Age-specific morbidity and mortality data are stored at the disease burden data for Uganda that may help inform such district level. These ‘line list’ records include: patient age, outcome questions. of treatment (i.e. discharge or death), and date of admission or death (for suspected cholera patients who die prior to seeking Methods treatment). We were able to obtain these data from 15 outbreaks, which occurred in 12 districts spanning the time period from Study design 2002–2010. In total, the line list data included records of 6,125 This is a retrospective study in which we collected data from cholera cases with at least 154 deaths. The actual number of Uganda’s health information management system and Diarrheal deaths could not be ascertained because some of the records lack Disease Control program. District-specific data were used to data on patient outcomes. These records included seven instances classify districts as endemic or non-endemic based on the WHO in which death occurred in the community, (i.e. prior to receiving criterion and to identify high-incidence districts. A convenience treatment). In addition, there were 923 records with data on sample of more detailed data from individual cholera outbreaks inpatient and outpatient treatment and the duration of inpatient were summarized to estimate the age distribution of reported treatment. cholera cases and to develop epidemic curves. Because the For this retrospective analysis, the study team compiled data Ugandan surveillance system is designed primarily to identify from samples that were previously collected and analyzed as part and respond to cholera outbreaks, a sensitivity analysis was of routine surveillance activities. performed to explore the potential limitations of the existing surveillance system to identify cases outside of recognized outbreaks. Data analysis The incidence of hospitalized cholera was estimated by district based on the annual average number of cases reported over the Current surveillance practices in Uganda six-year period from 2005–2010. The district-specific reporting These data were used to compile national statistics and for does not include data by age group. Thus, the age distribution of reporting to the World Health Organization’s (WHO) Weekly cases was estimated based on the 15 line lists. It was assumed that Epidemiological Record [5]. The cholera case definition was these 15 outbreaks were representative of the age distribution of based on WHO criteria, depending on whether or not cholera is cholera incidence in Uganda. The numbers of cases by age group endemic in the area: were calculated from the product of total cases and the national In non-endemic areas: ‘‘a patient aged 5 years or more average percentage distribution of cases by age from the line list develops severe dehydration or dies from acute watery data. Age-specific incidence rates were then calculated by dividing diarrhea.’’ the age-specific cases by the age-specific populations (2010 UN PLOS Neglected Tropical Diseases | www.plosntds.org 2 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 1. Annual number of cholera cases and deaths reported in Uganda 1997–2010. There was a major outbreak in 1998 and a fluctuating, but persistent burden of cholera in the years since the outbreak. doi:10.1371/journal.pntd.0002545.g001 population data). All analyses and graphs were produced with reported case. Estimation of asymptomatic cholera infections were Microsoft Excel (Microsoft, Redmond, WA) and maps were omitted from this analysis. created with ArcGIS (ESRI, Redlands, CA). Cholera case fatality Since the reported numbers of deaths were based on rates were estimated from the number of reported cases and deaths individually-identified cholera patients, these reports should be a by age group from the line list data. The Fisher’s exact test was lower bound. While deaths that occurred outside treatment used to compare case fatality rates across age groups. Statistical facilities were included in official reports when identified, it analyses were performed using STATA software (Version 8, remains likely that some cholera deaths were missed and not reported in official statistics. In a recent study in neighboring College Station, TX, USA). Kenya, an active case finding exercise identified a 200% increase in the number of cholera deaths that occurred during a 2008 Sensitivity analysis cholera outbreak [9]. This is a worst-case-scenario, since the In addition to hospitalized cases, we also estimated the number outbreak occurred during a chaotic period of post-election of non-hospitalized or non-reported cholera cases in Uganda. In a violence. However, in addition to deaths that were missed during recent analysis, Kirigia et al. [7] estimated that 10% of cholera outbreaks, isolated cholera deaths that occur outside of recognized cases could be classified as severe and require hospitalization. In outbreaks may also contribute to underreporting in official addition, Poulos et al. [8] reported that 22–38% of cholera statistics. As an upper bound estimate of the annual number of patients were hospitalized during multi-site surveillance studies cholera deaths in Uganda, we applied the 200% correction factor conducted in Jakarta, Indonesia and Kolkata, India. In Uganda, from the Kenya study to the number of cholera deaths identified in patients with mild diarrhea often do not seek formal seek care, but Uganda. In addition, for an upper bound estimate of the number do receive oral rehydration therapy at home. In this analysis, we of hospitalized cases, we assumed that a number of severe acute assumed that the official statistics include 25% of cholera patients watery diarrheal cases that occurred outside of recognized with severe cholera who would seek treatment and be reported in outbreaks were the result of infection by Vibrio cholerae. These official statistics and that 75% took oral rehydration therapy at endemic cholera cases have frequently been omitted from totals in home. This assumption is greater than that assumed by Kirigia et other cholera endemic countries [10,11]. Thus, we assumed the al., but falls at the lower end of the actual data presented by Poulos number of reported hospitalized cases may have only been about et al. Thus, we estimated that there were three non-hospitalized 50% of the actual cases although this rate is difficult to estimate in cholera cases requiring treatment at home per one officially the absence of sentinel surveillance data. PLOS Neglected Tropical Diseases | www.plosntds.org 3 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 2. Districts with reported cholera cases from 2005 to 2010 (endemic and non-endemic districts). Some districts may be considered endemic for cholera based on a history of cholera incidence during three of the five years from 2005–2010. Other districts have demonstrated cholera on a less consistent basis, while a few districts did not report any cases between 2005–2010. doi:10.1371/journal.pntd.0002545.g002 There were relatively few districts within Uganda in which Results cholera was confirmed every year. Kasese district, on the border The Ministry used laboratory confirmation for a sub-sample of with DRC, and the Kampala slums both reported cholera during at suspected cholera cases. In 2008, the national laboratory least five of the six years between 2005–2010. Figure 2 identifies confirmed that 71 cases out of 150 tested samples were due to districts that met the WHO definition of endemicity for cholera, i.e. V. cholerae (42%). The relatively low level of confirmed cases could where cholera was confirmed in three of the previous five years. be due to poor specimen collection, transportation or the less Districts in which cholera was reported, but in fewer than three of specific standard case definition, which tends to include severe the previous five years, were identified as non-endemic. The forms of other acute watery diarrhea. Importantly, the laboratory remaining districts did not report cholera from 2005 through 2010. also ruled out cholera as the causative organism for at least three In addition to the slums of Kampala, the cholera-outbreak-prone diarrheal disease outbreaks [personal communication: October 31, districts were mostly located along the western and northern borders. 2011, Mr. Ateki Kagirita, Head Central Public Health Labora- In addition to cholera, these districts have also been prone to tory]. These findings are comparable to those reported during the outbreaks of other waterborne diseases such as typhoid, shigellosis, 2002–03 cholera outbreaks in Uganda in which researchers found and hepatitis E. Cholera affected all age groups in Uganda. Based on that 52% of the suspected cholera patients had positive stool existing surveillance data, the age distribution of cholera followed the samples [12]. During 2008–09, about 82% of the confirmed age distribution of the population (Figure 3). The gender distribution cholera cases were attributed to the Ogawa serotype and the remaining 18% to the Inaba serotype. among cholera cases in Uganda was 54% female, 46% male. PLOS Neglected Tropical Diseases | www.plosntds.org 4 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda An annual average incidence rate of cholera was estimated for each district based on the average number of cases reported during 2005–2010. In Figure 5, the districts are subdivided into three categories, a high incidence category in which the average annual reported incidence was greater than 15 cases per 100,000 persons, a low incidence category in which the average annual incidence was greater than zero but less than 15 cases per 100,000 persons, and, finally, a category in which no cases were reported during the previous six years. Based on this threshold, 13 of the 56 districts that existed in 2005 were at high risk for cholera. These high-risk districts included an estimated 2010 population of 7.6 million people (24% of the population). Another 31 districts (18.0 million people, 57% of the population) were considered at ‘‘low risk’’ and 12 districts (6.1 million, 19% of the population) did not report any cases in the past six years (Table 1). Thus, the majority of Uganda’s population (76%) resides in districts that can be considered at ‘‘low-risk’’ for cholera. When considering district-specific incidence rates, it is impor- tant to note that cholera risk varied even within districts. During the 2010 cholera outbreak in the northeastern districts of Moroto and Kotido, more than 85% of cases in each district occurred in fewer than 27–33% of the sub-counties [13]. Figure 3. Age breakdown of cholera cases from non-random Estimated annual numbers of cases and deaths sample of cholera outbreaks, 2002–2010. The percentage distribution of cholera cases in Uganda is similar to the overall age The estimated annual number of cholera cases by age and by distribution of the population. district risk group is shown in Table 1. Inclusive of unreported cases doi:10.1371/journal.pntd.0002545.g003 treated at home, our estimated annual average number of cases was around 11,000, with around 81% of the cases occurring in the high risk districts. On average, about 61 cholera deaths were reported per year The distribution of cholera deaths by age group and the average during 2005–2010. Using the 200% correction factor reported case fatality rates during cholera outbreaks are summarized in from the Kenyan study [9], the potential range of annual cholera Figure 4. The case fatality rate was greater for elderly persons mortality is 61–182 deaths per year. (p,0.001) than other groups. However, mortality occurred in all age groups. The breakdown for inpatient versus outpatient treatment and Epidemic cholera length of hospitalization was not typically available in the line list The epidemic curves for fifteen cholera outbreaks that occurred data. Among the 923 records with such data, the hospitalization in Uganda between 2002 and 2010 are shown in Figure 6. The rate was about 90% and average duration of inpatient treatment average duration of the 15 outbreaks was about 15 weeks from the was 2.4 days. identification of the first case through the identification of the last Figure 4. Distribution of deaths by age group and average case fatality rate by age group. The age distribution of cholera deaths varies slightly from age distribution of cases because the case fatality rates are higher infants and the elderly relative to other age groups. doi:10.1371/journal.pntd.0002545.g004 PLOS Neglected Tropical Diseases | www.plosntds.org 5 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 5. Map of reported cholera incidence by district, hospitalized cases, 2005–10. The estimated incidence by district varies considerably in Uganda with a distinct pattern of low-incidence versus high incidence districts. doi:10.1371/journal.pntd.0002545.g005 case. The range of outbreak duration was between 4 weeks and 44 cholera vaccination campaigns in combination with other cholera weeks (Table 2). Almost half of the observed cases (43%) occurred control activities. within six weeks of the first case. It should be noted that the cases Cholera affects all age groups in Uganda. The age distribution reported for Kasese were more likely to be representative of of cases matched the population distribution. This may be due to endemic disease, as this district is one of the few that report cases low levels of background immunity, so that the entire population is on an ongoing basis. Arua district reported four outbreaks between equally susceptible. This age distribution deviates from age 2005 and 2008, but weekly cases declined from the peak observed distributions in other cholera-endemic areas, where young in early 2008. children tended to be at greater risk when systematic surveillance was conducted [14]. Systematic sampling of diarrheal cases from Discussion endemic areas has never been attempted in Uganda and may reveal that outbreak-based surveillance findings are not represen- This retrospective analysis showed that there is a clear tative of the true cholera burden. A comparison of age-specific subdivision between high-risk districts and low-risk districts in cholera incidence rates from Bangladesh demonstrated that the Uganda with about 24% of the population residing in high-risk average age of cholera infection was much higher during districts accounting for 81% of the average reported cholera outbreaks [15] than for endemic cholera [16]. burden. These high-risk districts may be considered for preventive PLOS Neglected Tropical Diseases | www.plosntds.org 6 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Outbreak data from the sub-district level suggests that there may be considerable heterogeneity of cholera incidence. Thus, surveillance efforts and reporting should be improved to facilitate better epidemiological characterization of cholera incidence and improved targeting of interventions to reach those at greatest risk. The estimate of 61 deaths per year involves accreditation of all cholera deaths to specific individuals, either at treatment centers or in the community. This is a relatively small fraction of the estimated 30,000 diarrheal deaths per year in Uganda (exclusive of deaths attributed to cholera and bloody diarrhea) [17]. It is certainly possible that a significant proportion of these 30,000 deaths was caused by unrecognized cholera than would be estimated from individually identified deaths. Although the outbreak response focus of cholera surveillance in Uganda may be insufficient to accurately estimate the numbers of cases and deaths caused by cholera, these data are very useful for identifying areas to target for surveillance in consideration of future vaccine introduction. In order to better quantify the burden of cholera in Uganda, sentinel site surveillance should be undertaken in at least two regions with districts at high risk for cholera for a period of at least two years. It would be better to continue surveillance for at least five years, since cholera incidence is highly variable from year to year. The Ministry of Health is participating in the AFRICHOL cholera surveillance in Africa project (www.africhol.org), led by Agence de Me´decine Pre´ventive (AMP) and the African Field Epidemiological Network (AFENET), which is an African-based non-government organization working to improve field epidemi- ology and public health laboratories in sub-Saharan Africa. As part of this project, enhanced cholera surveillance is being conducted in five districts in Eastern Uganda (Mbale, Tororo, Manafwa, Butaleja and Busia) and, whenever outbreaks occur, throughout the country. Such data may be combined with the available national reporting statistics to better model cholera burden within Uganda, which in turn may be used to conduct economic analyses (e.g., cost effectiveness or cost benefit studies) of the potential use of cholera vaccines in Uganda. Given the health challenges facing Uganda, the decision to pursue cholera vaccination must be weighed against the introduction of other health interventions that may have a greater impact on mortality (e.g., pneumococcal conjugate vaccines, rotavirus vaccines, future malaria vaccines or other interventions). In addition to targeting high-risk endemic populations, Uganda may consider using cholera vaccines from a recently established international stockpile to mitigate epidemic cholera. A review of 15 epidemic curves showed that about 57% of the cases occurred after six weeks across all outbreaks. This 57% figure may represent an upper bound on the number of cases that could be averted via reactive use of cholera vaccines from a global stockpile, assuming it would take at least three weeks to diagnose an outbreak and prepare for a mass vaccination campaign plus three weeks to generate immunity from the two-dose vaccine. This stockpile may also be used to prevent the spread of cholera to neighboring districts, such as when it was used in an Adjumani district refugee camp in 1997 [18]. While cholera incidence in Uganda has been manageable over the past decade, elimination of the disease is likely to take time especially given the slow progress on provision of safe water and sanitation among other risk factors. Most of the areas with the highest incidence rates either border countries with political instability and endemic cholera (e.g., DRC and Sudan) or contain semi-nomadic populations. For these districts, it would be difficult to prevent cholera-infected persons from crossing borders, achieve PLOS Neglected Tropical Diseases | www.plosntds.org 7 December 2013 | Volume 7 | Issue 12 | e2545 Table 1. Estimated annual cholera incidence rates, number of cases and death by district risk category and age group. Total annual Estimated incidence Annual deaths Annual deaths District risk category Age group Population Reported cases cases (range) per 1,000 (range) (reported) (range) High ,1 369,000 50 50–200 14–55 1 1–4 1–4 1,208,000 363 360–1,400 30–120 1 1–4 5–14 2,386,000 538 540–2,200 23–90 8 8–24 15+ 4,172,000 1,285 1,300–5,100 31–120 38 38–115 All ages 8,135,000 2,235 2,200–8,900 28–110 49 49–148 Low or Unknown ,1 1,164,761 12 12–48 0.2–0.8 0.23 0–1 1–4 3,809,393 85 85–340 1.5–6 2 2–5 5–14 7,527,670 126 130–550 2–8 3 3–8 15+ 13,160,134 301 300–1,200 5–20 7 7–20 All 25,661,958 524 520–2,100 8.8–35 11 11–34 All districts ,1 1,534,000 62 62–250 1–4 1 1–4 1–4 5,017,000 448 450–1,800 7.5–30 3 3–10 5–14 9,914,000 664 660–2,700 11–44 11 11–33 15+ 17,332,000 1,586 1,600–6,300 26–105 45 45–135 All 33,797,000 2,760 2,800–11,000 46–180 61 61–182 doi:10.1371/journal.pntd.0002545.t001 The Burden of Cholera in Uganda Figure 6. Epidemic curves of cholera outbreaks in Uganda, 2002–2010. The epidemic curves from a convenience sample of fifteen outbreaks shows both large and small outbreak have occurred during the period 2002–2010. There is considerable variation in the duration of outbreaks with longer outbreaks in districts with higher average annual incidence rates. doi:10.1371/journal.pntd.0002545.g006 PLOS Neglected Tropical Diseases | www.plosntds.org 8 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Table 2. A non-random sample of outbreaks in Uganda, 2002–2010. Time period Time period Time period District (month) (week) Cases District Time period (month) (week) Cases Amudat May 2010–Aug 2010 14 weeks 185 Kasese Jun 2009–Apr 2010 44 weeks 148 Arua Jun 2005–Sep 2005 16 weeks 289 Kitgum Feb 2006–Oct 2006 37 weeks 1,245 Arua Jul 2006–Oct 2006 16 weeks 135 Kotido May 2010–Jul 2010 12 weeks 491 Arua Dec 2007–Mar 2008 18 weeks 663 Moroto Apr 2010–Jun 2010 10 weeks 610 Arua Jun 2008–Nov 2008 21 weeks 73 Nebbi Dec 2002–Mar 2003 17 weeks 437 Bugiri Sep 2009–Oct 2009 8 weeks 188 Nebbi Oct 2007–Jan 2008 16 weeks 551 Bulisa Nov 2007–Dec 2007 6 weeks 291 Yumbe Nov 2007–Dec 2007 4 weeks 75 Kampala Nov 2006–Feb 2007 15 weeks 726 doi:10.1371/journal.pntd.0002545.t002 high vaccination coverage rates, or to construct reliable water and does not reduce the incidence of cholera cases, it does reduce the sanitation infrastructure for semi-nomadic populations [19]. social and economic burden of the disease. Improvements in access to improved water and sanitation Some global trends in cholera disease burden may lead to an increase in the number of cases and should be considered in would also lead to a concomitant decrease in cholera incidence. These cholera incidence data may also be used to target priority cholera control planning. At present, cholera is more prevalent in districts for improvements in water, sanitation, and hygiene efforts. rural areas than in urban areas within Uganda. This may change if Cholera incidence is likely to be associated with high prevalence of present urbanization trends continue and the maintenance and other enteric diseases, for which cholera vaccination would have expansion of water and sanitation infrastructure cannot keep pace no effect. Considering that an estimated 30,000 persons die from with the rapidly growing urban population. The urban population diarrheal disease every year in Uganda, improved water, in Uganda is projected to increase more than seven-fold from 4.5 sanitation, and hygiene are urgently needed even if cholera million in 2010 to 31 million by 2050 [20]. vaccine is deployed. Some studies have found multidrug resistant V. cholerae in In conclusion, the existing surveillance system is geared Uganda, including strains resistant to trimethoprim, sulfonamides, toward mitigating the impacts of cholera outbreaks, not ampicillin, tetracycline, chloramphenical and streptomycin [21]. quantifying the burden of endemic cholera. Cholera control In addition, there is evidence that the severity of clinical cases of activities have been effective in slowing the spread of cholera and cholera in Asia and Africa is increasing, especially during reducing cholera fatalities. However, cholera cases continue to outbreaks. Some scientists attribute the increase in severity of be reported on an annual basis. The combination of sentinel cholera cases seeking treatment to the emergence of a new altered surveillance with national cholera incidence data could be used strain of V. cholerae O1 El Tor that secretes the classical cholera to develop an economic analysis to inform cholera vaccination toxin, making it more virulent [22]. It is not presently known if this policy. strain is present in Uganda. However, it has been isolated from recent African outbreaks in Mozambique and Zimbabwe [23,24]. Acknowledgments Due to global warming, the average temperature in Uganda is estimated to increase by up to 1.5 degrees over the next 20 years The authors wish to acknowledge Mr. Ateki Kagirita, Head Central Public Health Laboratory (CPHL), for providing laboratory methods and results; [25]. Recent research suggests a strong correlation between Mr. Jung Seok Lee of the International Vaccine Institute for assistance in increased rainfall and elevated temperatures with higher cholera creating GIS maps; and Drs. John D. Clemens, Anna Lena Lopez, and incidence [26,27,28,29,30]. This may pose an elevated risk for Thomas F. Wierzba of the International Vaccine Institute for assistance in districts bordering Lake Albert and Lake Victoria, which may developing and coordinating the study. The manuscript was edited by the provide an endemic reservoir of V. cholerae [31,32]. IVI Communications and Advocacy Unit. There are also trends suggesting a reduced need for cholera vaccination in Uganda. The multidisciplinary cholera outbreak Author Contributions response activities have been effective in mitigating the severity of Conceived and designed the experiments: GB MM BM AL. Performed the outbreaks, both in terms of morbidity and mortality. The cholera experiments: GB MM BM YEK. Analyzed the data: GB MM BM YEK case fatality rate has steadily declined since the large, nationwide AL. Contributed reagents/materials/analysis tools: GB MM. Wrote the outbreak in 1998 (refer to Figure 1). While improving treatment paper: GB BM AL VM. References 1. World Health Organization (2000) Report on Global Surveillance on epidemic 4. World Health Organization (2010) Cholera vaccines: WHO position paper. prone infectious diseases, WHO/CDS/CSR/ISR/2000.1 http://www.who.int/ Weekly Epidemiological Record 13: 117–128. csr/resources/publications/surveillance/en/cholera.pdf accessed on July 1, 5. World Health Organization (2010) Cholera, 2009. Weekly Epidemiological 2012. Record 85: 293–308. 2. Legros D, McCormick M, Mugero C, Skinnider M, Bek’obita D, et al. (2000) 6. The Republic of Uganda Ministry of Health (2007) Prevention and control of Epidemiology of cholera outbreak in Kampala, Uganda. East African Medical cholera, operational guidelines for the district health workers and planners. Journal 77: 347–349. Control of Diarrheoal Diseases Section CHD,. Kampala. 3. Nkoko DB, Giraudoux P, Plisnier P-D, Tinda AM, Piarroux M, et al. (2011) 7. Kirigia JM, Sambo LG, Yokouide A, Soumbey-Alley E, Muthuri LK, et al. Dynamics of Cholera Outbreaks in Great Lakes Region of Africa, 1978–2008. (2009) Economic burden of cholera in the WHO African region. BMC Emerging Infectious Disease 17: 2026–2034. International Health and Human Rights 9: 1–14. PLOS Neglected Tropical Diseases | www.plosntds.org 9 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda 8. Poulos C, Riewpaiboon A, Stewart JF, Clemens J, Guh S, et al. (2011) Costs of 21. Kruse H, Sørum H, Tenover F, Olsvik O (1995) A transferable multiple drug Illness Due to Endemic Cholera. Epidemiology and Infection Apr 18: 1–10. resistance plasmid from Vibrio cholerae O1. Microbial Drug Resistance 1: 203– 9. Shikanga O-T, Mutonga D, Abade M, Amwayi S, Ope M, et al. (2009) High Mortality in a Cholera Outbreak in Western Kenya after Post-Election Violence 22. Siddique AK, Nair GB, Alam M, Sack DA, Huq A, et al. (2009) El Tor cholera in 2008. American Journal of Tropical Medicine and Hygiene 81: 1085–1090. with severe disease: a new threat to Asia and beyond. Epidemiology and 10. Kuttiat VS, Lodha R, Das B, Kohli U (2010) Prevalence of Cholera in Pediatric Infection 138: 347–352. Patients with Acute Dehydrating Diarrhea. Indian Journal of Pediatrics 77: 67– 23. World Health Organization (2009) Cholera, Zimbabwe – update. Weekly Epidemiological Record 84: 109–110. 11. Nair GB, Ramamurthy T, Bhattacharya MK, Krishnan T, Ganguly S, et al. 24. Ansaruzzaman M, Bhuiyan NA, Nair B, Sack DA, Lucas M, et al. (2004) (2010) Emerging trends in the etiology of enteric pathogens as evidenced from an Cholera in Mozambique, Variant of Vibrio cholerae. Emerging Infectious active surveillance of hospitalized diarrhoeal patients in Kolkata, India. Gut Disease 10: 2057–2059. Pathogens 2: 4. 25. Hepworth N (2010) Climate Change Vulnerability and Adaption Preparedness 12. Alajo SO, Nakavuma J, Erume J (2006) Cholera in endemic districts in Uganda in Uganda. Nairobi, Kenya. during El Nin˜o rains: 2002–2003. African Health Sciences 6: 93–97. 26. Constantin de Magny G, Murtugudde R, Sapiano M, Nizam A, Brown C, et al. 13. United Nations Office for the Coordination of Humanitarian Affairs (2010) (2008) Environmental signatures associated with cholera epidemics. Proceedings Uganda: Cholera Affected Areas & Cases by Parish (as of 12 July 2010). Relief of National Academies of Science USA 105: 17676–81. Web report, http://reliefwebint/node/17116 (accessed on 1 Nov 2010). 27. Emch M, Feldacker C, Yunus M, Streatfield P, Thiem V, et al. (2008) Local 14. Deen J, von Seidlein L, Sur D, Agtini M, Lucas M, et al. (2008) The high burden environmental predictors of cholera in Bangladesh and Vietnam. American of cholera in children: Comparison of incidence from endemic areas in Asia and Journal of Tropical Medicine and Hygiene 78: 823–832. Africa. PLoS Neglected Tropical Diseases 2: e173. 28. Ferna´ndez MAL, Bauernfein A, Jime´nez JD, Gil CL, Omeiria NE, et al. (2009) 15. Harris A, Chowdhury F, Begum Y, Khan A, Faruque A, et al. (2008) Shifting Influence of temperature and rainfall on the evolution of cholera epidemics in prevalence of major diarrheal pathogens in patients seeking hospital care during Lusaka, Zambia, 2003—2006: analysis of a time series. Tropical Medicine and floods in 1998, 2004, and 2007 in Dhaka, Bangladesh. American Journal of Hygiene 103: 137–143. Tropical Medicine and Hygiene 79: 708–714. 29. Lobitz B, Beck L, Huq A, Wood B, Fuchs G, et al. (2000) Climate and infectious 16. Sack RB, Siddique AK, Longini Jr IM, Nizam A, Yunus M, et al. (2003) A 4- disease: use of remote sensing for detection of Vibrio cholerae by indirect year study of the epidemiology of Vibrio cholerae in four rural areas of measurement. Proceedings of National Academies of Science USA 97: 1438– Bangladesh. Journal of Infectious Diseases 187: 96–101. 17. The Republic of Uganda Ministry of Health (2009) Ministry of Health Statistical 30. Reyburn R, Kim DR, Emch M, Khatib A, Seidlein Lv, et al. (2011) Climate Abstract 2008/09. Variability and the Outbreaks of Cholera in Zanzibar, East Africa: A Time 18. Dorlencourt F, Legros D, Paquet C, Neira M, Ivanoff B, et al. (1999) Series Analysis. American Journal of Tropical Medicine and Hygiene 84: 862– Effectiveness of mass vaccination with WC/rBS cholera vaccine in Adjumani district, Uganda. Bulletin of the World Health Organization 77: 949–950. 31. Bompangue D, Giraudoux P, Handschumacher P, Piarroux M, Sudre B, et al. 19. Cummings MJ, Wamala JF, Eyura M, MalimboM, Omeke EM, et al. (2011) A (2008) Lakes as source of cholera outbreaks, Democratic Republic of Congo. . cholera outbreak among semi-nomadic pastoralists in northeastern Uganda: epidemiology and interventions. Epidemiology and Infection 140: 1376–85. Emerging Infectious Disease 14: 798–800. doi:10.1017/S0950268811001956. 32. Shapiro R, Otieno M, Adcock P, Phillips-Howard P, Hawley W, et al. (1999) 20. Population Division of the Department of Economic and Social Affairs of the Transmission of epidemic Vibrio cholerae O1 in rural western Kenya associated United Nations Secretariat (2007) World Population Prospects: The 2006 with drinking water from Lake Victoria: an environmental reservoir for cholera? Revision and World Urbanization Prospects: The 2007 Revision. . American Journal of Tropical Medicine and Hygiene 60: 271–276. PLOS Neglected Tropical Diseases | www.plosntds.org 10 December 2013 | Volume 7 | Issue 12 | e2545 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png PLoS Neglected Tropical Diseases Public Library of Science (PLoS) Journal

The Burden of Cholera in Uganda

PLoS Neglected Tropical Diseases , Volume 7 (12) – Dec 5, 2013
Download PDF
10 pages

Loading next page...
 
/lp/public-library-of-science-plos-journal/the-burden-of-cholera-in-uganda-gtVUsS1d90

References (38)

Publisher
Public Library of Science (PLoS) Journal
Copyright
Copyright: © 2013 Bwire et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Financial support was provided by the Bill and Melinda Gates Foundation through the Cholera Vaccine Initiative (CHOVI) program, administered by the International Vaccine Institute (IVI), Seoul, Korea. Current donors providing unrestricted support to the IVI include the Governments of Kuwait, Republic of Korea and Sweden. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist.
eISSN
1935-2735
DOI
10.1371/journal.pntd.0002545
Publisher site
See Article on Publisher Site

Abstract

Introduction: In 2010, the World Health Organization released a new cholera vaccine position paper, which recommended the use of cholera vaccines in high-risk endemic areas. However, there is a paucity of data on the burden of cholera in endemic countries. This article reviewed available cholera surveillance data from Uganda and assessed the sufficiency of these data to inform country-specific strategies for cholera vaccination. Methods: The Uganda Ministry of Health conducts cholera surveillance to guide cholera outbreak control activities. This includes reporting the number of cases based on a standardized clinical definition plus systematic laboratory testing of stool samples from suspected cases at the outset and conclusion of outbreaks. This retrospective study analyzes available data by district and by age to estimate incidence rates. Since surveillance activities focus on more severe hospitalized cases and deaths, a sensitivity analysis was conducted to estimate the number of non-severe cases and unrecognized deaths that may not have been captured. Results: Cholera affected all ages, but the geographic distribution of the disease was very heterogeneous in Uganda. We estimated that an average of about 11,000 cholera cases occurred in Uganda each year, which led to approximately 61–182 deaths. The majority of these cases (81%) occurred in a relatively small number of districts comprising just 24% of Uganda’s total population. These districts included rural areas bordering the Democratic Republic of Congo, South Sudan, and Kenya as well as the slums of Kampala city. When outbreaks occurred, the average duration was about 15 weeks with a range of 4– 44 weeks. Discussion: There is a clear subdivision between high-risk and low-risk districts in Uganda. Vaccination efforts should be focused on the high-risk population. However, enhanced or sentinel surveillance activities should be undertaken to better quantify the endemic disease burden and high-risk populations prior to introducing the vaccine. Citation: Bwire G, Malimbo M, Maskery B, Kim YE, Mogasale V, et al. (2013) The Burden of Cholera in Uganda. PLoS Negl Trop Dis 7(12): e2545. doi:10.1371/ journal.pntd.0002545 Editor: Edward T. Ryan, Massachusetts General Hospital, United States of America Received April 13, 2012; Accepted October 4, 2013; Published December 5, 2013 Copyright:  2013 Bwire et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Financial support was provided by the Bill and Melinda Gates Foundation through the Cholera Vaccine Initiative (CHOVI) program, administered by the International Vaccine Institute (IVI), Seoul, Korea. Current donors providing unrestricted support to the IVI include the Governments of Kuwait, Republic of Korea and Sweden. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected] ¤ Current address: Division of Global Migration and Quarantine, United States Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America. occur any time, but is often associated with extreme rain events or Introduction water supply disruptions. Cholera was first reported in Uganda in 1971, when 757 cases The frequency of reported cholera cases varies among districts were reported to the World Health Organization (WHO). During in Uganda. The highest risk areas include the border areas with the subsequent years up to 1993, Uganda reported cholera cases the Democratic Republic of Congo (DRC), Sudan, and Kenya as every 2–4 years to the WHO. From 1994 to 1998, cholera was well as urban slums in Kampala. Displaced populations and their reported annually in Uganda [1]. In 1998, Uganda reported neighboring communities are at elevated risk. The ongoing almost 50,000 cases with incidence throughout the country [2]. migration of people into and within Uganda can lead to rapid The reported incidence has fluctuated between 250 and 5,000 spread of the disease. The African Great Lakes, including Lake cases every year since 2000 (Figure 1). The reported case fatality Albert and Lake Victoria on the border of Uganda, may provide a ratio has decreased from 4–7% in the late 1990s to about 2–3% reservoir for cholera bacteria. Further, increases in incidence during 2004–2010. among the nations bordering these lakes have been shown to be Cholera in Uganda appears to be largely an epidemic disease. correlated with El Nin˜o warm weather events [3]. The WHO [4] However, endemic cholera occurs in high-risk areas along the has recently revised its guidelines and states in a position paper southwestern border with DRC and in Kampala city slums. that cholera vaccines should be used in combination with other Endemic cholera is commonly noted before and during the rainy prevention and control strategies in areas where the disease is season, from December through March. Epidemic cholera can endemic. ‘‘Endemic’’ is defined as areas with occurrence of PLOS Neglected Tropical Diseases | www.plosntds.org 1 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda In endemic areas: ‘‘a patient aged 2 years or more develops Author Summary severe dehydration or dies from acute watery diarrhea.’’ [6] Uganda has reported cholera cases to the World Health The identification of such cases should have triggered labora- Organization every year since 1997. Thus, the country may tory investigation. A cholera outbreak was confirmed when Vibrio consider the introduction of a WHO-prequalified oral cholerae O1 or O139 was isolated from at least one stool sample. cholera vaccine. This article reviews cholera surveillance data from 1997–2010 with a focus on the 2005–2010 time Only cases meeting the standard case definition above were period to identify high risk populations that may be investigated and included in the official cholera data. targeted for preventive vaccination campaigns. We esti- mated that an average of about 61–182 deaths occur each Laboratory methods year. Most cases (81%) occurred in a relatively small Summary laboratory data were obtained from the Head of the number of districts comprising just 24% of Uganda’s total Central Public Health Laboratory from the Ministry of Health. population. While there is a clear distinction between low Prior to analysis, stool samples from suspected cholera patients and high-risk districts, sentinel surveillance would help to were transported from the field in Cary Blair transport media. better quantify the burden in endemic districts. An Culture plates were set at 37uC overnight (for 18–24 hours) using economic analysis should also be undertaken prior to three culture media: TCB, XLD and MacConkey. Biochemical making a decision to introduce a cholera vaccine. identification of cholera organisms were based on Oxidase or Indole tests. Polyvalent antisera were used to differentiate between the Inaba and Ogawa serotypes, and specific monovalent tests culture-confirmed cholera in at least three of the previous five further confirmed which of the Inaba, Ogawa or O139 (Bengal) years. The WHO also recommends that cholera vaccines should strains caused disease. Isolates were refrigerated at 280uC and be considered for preemptive use in areas at risk for epidemic sent to the WHO’s collaborating laboratory (Unite´ de La Rage, cholera as long as vaccination does not interfere with efforts to Institut Pasteur, Paris, France) for quality control. treat cholera cases, improve water and sanitation, and mobilize communities. The vaccine may also be considered for reactive use if local infrastructure is sufficient to conduct mass campaigns Data collection depending on the current and historical epidemiology. District-specific data were abstracted from the Uganda Ministry of Health, Health Management Information System disease There is a dearth of information about the burden of cholera in low-income countries such as Uganda. A more accurate picture of surveillance database for the period 2005–2010. The 2005–2010 period was chosen based on the WHO criteria [4] for identifying this burden is particularly important because it can be used to endemic cholera (i.e. areas in which cholera has been reported in inform cholera prevention and control intervention questions: three of the previous five years.) The districts shown are based on whether or not to introduce vaccination as a complement to other the 2002 district boundaries, which were in existence during the cholera prevention and control interventions, where and when it most recent census. Cases in new districts created after 2005 were would be most effective to do so, and what demographic apportioned back to the 2002 districts. population should be targeted. This article presents available Age-specific morbidity and mortality data are stored at the disease burden data for Uganda that may help inform such district level. These ‘line list’ records include: patient age, outcome questions. of treatment (i.e. discharge or death), and date of admission or death (for suspected cholera patients who die prior to seeking Methods treatment). We were able to obtain these data from 15 outbreaks, which occurred in 12 districts spanning the time period from Study design 2002–2010. In total, the line list data included records of 6,125 This is a retrospective study in which we collected data from cholera cases with at least 154 deaths. The actual number of Uganda’s health information management system and Diarrheal deaths could not be ascertained because some of the records lack Disease Control program. District-specific data were used to data on patient outcomes. These records included seven instances classify districts as endemic or non-endemic based on the WHO in which death occurred in the community, (i.e. prior to receiving criterion and to identify high-incidence districts. A convenience treatment). In addition, there were 923 records with data on sample of more detailed data from individual cholera outbreaks inpatient and outpatient treatment and the duration of inpatient were summarized to estimate the age distribution of reported treatment. cholera cases and to develop epidemic curves. Because the For this retrospective analysis, the study team compiled data Ugandan surveillance system is designed primarily to identify from samples that were previously collected and analyzed as part and respond to cholera outbreaks, a sensitivity analysis was of routine surveillance activities. performed to explore the potential limitations of the existing surveillance system to identify cases outside of recognized outbreaks. Data analysis The incidence of hospitalized cholera was estimated by district based on the annual average number of cases reported over the Current surveillance practices in Uganda six-year period from 2005–2010. The district-specific reporting These data were used to compile national statistics and for does not include data by age group. Thus, the age distribution of reporting to the World Health Organization’s (WHO) Weekly cases was estimated based on the 15 line lists. It was assumed that Epidemiological Record [5]. The cholera case definition was these 15 outbreaks were representative of the age distribution of based on WHO criteria, depending on whether or not cholera is cholera incidence in Uganda. The numbers of cases by age group endemic in the area: were calculated from the product of total cases and the national In non-endemic areas: ‘‘a patient aged 5 years or more average percentage distribution of cases by age from the line list develops severe dehydration or dies from acute watery data. Age-specific incidence rates were then calculated by dividing diarrhea.’’ the age-specific cases by the age-specific populations (2010 UN PLOS Neglected Tropical Diseases | www.plosntds.org 2 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 1. Annual number of cholera cases and deaths reported in Uganda 1997–2010. There was a major outbreak in 1998 and a fluctuating, but persistent burden of cholera in the years since the outbreak. doi:10.1371/journal.pntd.0002545.g001 population data). All analyses and graphs were produced with reported case. Estimation of asymptomatic cholera infections were Microsoft Excel (Microsoft, Redmond, WA) and maps were omitted from this analysis. created with ArcGIS (ESRI, Redlands, CA). Cholera case fatality Since the reported numbers of deaths were based on rates were estimated from the number of reported cases and deaths individually-identified cholera patients, these reports should be a by age group from the line list data. The Fisher’s exact test was lower bound. While deaths that occurred outside treatment used to compare case fatality rates across age groups. Statistical facilities were included in official reports when identified, it analyses were performed using STATA software (Version 8, remains likely that some cholera deaths were missed and not reported in official statistics. In a recent study in neighboring College Station, TX, USA). Kenya, an active case finding exercise identified a 200% increase in the number of cholera deaths that occurred during a 2008 Sensitivity analysis cholera outbreak [9]. This is a worst-case-scenario, since the In addition to hospitalized cases, we also estimated the number outbreak occurred during a chaotic period of post-election of non-hospitalized or non-reported cholera cases in Uganda. In a violence. However, in addition to deaths that were missed during recent analysis, Kirigia et al. [7] estimated that 10% of cholera outbreaks, isolated cholera deaths that occur outside of recognized cases could be classified as severe and require hospitalization. In outbreaks may also contribute to underreporting in official addition, Poulos et al. [8] reported that 22–38% of cholera statistics. As an upper bound estimate of the annual number of patients were hospitalized during multi-site surveillance studies cholera deaths in Uganda, we applied the 200% correction factor conducted in Jakarta, Indonesia and Kolkata, India. In Uganda, from the Kenya study to the number of cholera deaths identified in patients with mild diarrhea often do not seek formal seek care, but Uganda. In addition, for an upper bound estimate of the number do receive oral rehydration therapy at home. In this analysis, we of hospitalized cases, we assumed that a number of severe acute assumed that the official statistics include 25% of cholera patients watery diarrheal cases that occurred outside of recognized with severe cholera who would seek treatment and be reported in outbreaks were the result of infection by Vibrio cholerae. These official statistics and that 75% took oral rehydration therapy at endemic cholera cases have frequently been omitted from totals in home. This assumption is greater than that assumed by Kirigia et other cholera endemic countries [10,11]. Thus, we assumed the al., but falls at the lower end of the actual data presented by Poulos number of reported hospitalized cases may have only been about et al. Thus, we estimated that there were three non-hospitalized 50% of the actual cases although this rate is difficult to estimate in cholera cases requiring treatment at home per one officially the absence of sentinel surveillance data. PLOS Neglected Tropical Diseases | www.plosntds.org 3 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 2. Districts with reported cholera cases from 2005 to 2010 (endemic and non-endemic districts). Some districts may be considered endemic for cholera based on a history of cholera incidence during three of the five years from 2005–2010. Other districts have demonstrated cholera on a less consistent basis, while a few districts did not report any cases between 2005–2010. doi:10.1371/journal.pntd.0002545.g002 There were relatively few districts within Uganda in which Results cholera was confirmed every year. Kasese district, on the border The Ministry used laboratory confirmation for a sub-sample of with DRC, and the Kampala slums both reported cholera during at suspected cholera cases. In 2008, the national laboratory least five of the six years between 2005–2010. Figure 2 identifies confirmed that 71 cases out of 150 tested samples were due to districts that met the WHO definition of endemicity for cholera, i.e. V. cholerae (42%). The relatively low level of confirmed cases could where cholera was confirmed in three of the previous five years. be due to poor specimen collection, transportation or the less Districts in which cholera was reported, but in fewer than three of specific standard case definition, which tends to include severe the previous five years, were identified as non-endemic. The forms of other acute watery diarrhea. Importantly, the laboratory remaining districts did not report cholera from 2005 through 2010. also ruled out cholera as the causative organism for at least three In addition to the slums of Kampala, the cholera-outbreak-prone diarrheal disease outbreaks [personal communication: October 31, districts were mostly located along the western and northern borders. 2011, Mr. Ateki Kagirita, Head Central Public Health Labora- In addition to cholera, these districts have also been prone to tory]. These findings are comparable to those reported during the outbreaks of other waterborne diseases such as typhoid, shigellosis, 2002–03 cholera outbreaks in Uganda in which researchers found and hepatitis E. Cholera affected all age groups in Uganda. Based on that 52% of the suspected cholera patients had positive stool existing surveillance data, the age distribution of cholera followed the samples [12]. During 2008–09, about 82% of the confirmed age distribution of the population (Figure 3). The gender distribution cholera cases were attributed to the Ogawa serotype and the remaining 18% to the Inaba serotype. among cholera cases in Uganda was 54% female, 46% male. PLOS Neglected Tropical Diseases | www.plosntds.org 4 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda An annual average incidence rate of cholera was estimated for each district based on the average number of cases reported during 2005–2010. In Figure 5, the districts are subdivided into three categories, a high incidence category in which the average annual reported incidence was greater than 15 cases per 100,000 persons, a low incidence category in which the average annual incidence was greater than zero but less than 15 cases per 100,000 persons, and, finally, a category in which no cases were reported during the previous six years. Based on this threshold, 13 of the 56 districts that existed in 2005 were at high risk for cholera. These high-risk districts included an estimated 2010 population of 7.6 million people (24% of the population). Another 31 districts (18.0 million people, 57% of the population) were considered at ‘‘low risk’’ and 12 districts (6.1 million, 19% of the population) did not report any cases in the past six years (Table 1). Thus, the majority of Uganda’s population (76%) resides in districts that can be considered at ‘‘low-risk’’ for cholera. When considering district-specific incidence rates, it is impor- tant to note that cholera risk varied even within districts. During the 2010 cholera outbreak in the northeastern districts of Moroto and Kotido, more than 85% of cases in each district occurred in fewer than 27–33% of the sub-counties [13]. Figure 3. Age breakdown of cholera cases from non-random Estimated annual numbers of cases and deaths sample of cholera outbreaks, 2002–2010. The percentage distribution of cholera cases in Uganda is similar to the overall age The estimated annual number of cholera cases by age and by distribution of the population. district risk group is shown in Table 1. Inclusive of unreported cases doi:10.1371/journal.pntd.0002545.g003 treated at home, our estimated annual average number of cases was around 11,000, with around 81% of the cases occurring in the high risk districts. On average, about 61 cholera deaths were reported per year The distribution of cholera deaths by age group and the average during 2005–2010. Using the 200% correction factor reported case fatality rates during cholera outbreaks are summarized in from the Kenyan study [9], the potential range of annual cholera Figure 4. The case fatality rate was greater for elderly persons mortality is 61–182 deaths per year. (p,0.001) than other groups. However, mortality occurred in all age groups. The breakdown for inpatient versus outpatient treatment and Epidemic cholera length of hospitalization was not typically available in the line list The epidemic curves for fifteen cholera outbreaks that occurred data. Among the 923 records with such data, the hospitalization in Uganda between 2002 and 2010 are shown in Figure 6. The rate was about 90% and average duration of inpatient treatment average duration of the 15 outbreaks was about 15 weeks from the was 2.4 days. identification of the first case through the identification of the last Figure 4. Distribution of deaths by age group and average case fatality rate by age group. The age distribution of cholera deaths varies slightly from age distribution of cases because the case fatality rates are higher infants and the elderly relative to other age groups. doi:10.1371/journal.pntd.0002545.g004 PLOS Neglected Tropical Diseases | www.plosntds.org 5 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Figure 5. Map of reported cholera incidence by district, hospitalized cases, 2005–10. The estimated incidence by district varies considerably in Uganda with a distinct pattern of low-incidence versus high incidence districts. doi:10.1371/journal.pntd.0002545.g005 case. The range of outbreak duration was between 4 weeks and 44 cholera vaccination campaigns in combination with other cholera weeks (Table 2). Almost half of the observed cases (43%) occurred control activities. within six weeks of the first case. It should be noted that the cases Cholera affects all age groups in Uganda. The age distribution reported for Kasese were more likely to be representative of of cases matched the population distribution. This may be due to endemic disease, as this district is one of the few that report cases low levels of background immunity, so that the entire population is on an ongoing basis. Arua district reported four outbreaks between equally susceptible. This age distribution deviates from age 2005 and 2008, but weekly cases declined from the peak observed distributions in other cholera-endemic areas, where young in early 2008. children tended to be at greater risk when systematic surveillance was conducted [14]. Systematic sampling of diarrheal cases from Discussion endemic areas has never been attempted in Uganda and may reveal that outbreak-based surveillance findings are not represen- This retrospective analysis showed that there is a clear tative of the true cholera burden. A comparison of age-specific subdivision between high-risk districts and low-risk districts in cholera incidence rates from Bangladesh demonstrated that the Uganda with about 24% of the population residing in high-risk average age of cholera infection was much higher during districts accounting for 81% of the average reported cholera outbreaks [15] than for endemic cholera [16]. burden. These high-risk districts may be considered for preventive PLOS Neglected Tropical Diseases | www.plosntds.org 6 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Outbreak data from the sub-district level suggests that there may be considerable heterogeneity of cholera incidence. Thus, surveillance efforts and reporting should be improved to facilitate better epidemiological characterization of cholera incidence and improved targeting of interventions to reach those at greatest risk. The estimate of 61 deaths per year involves accreditation of all cholera deaths to specific individuals, either at treatment centers or in the community. This is a relatively small fraction of the estimated 30,000 diarrheal deaths per year in Uganda (exclusive of deaths attributed to cholera and bloody diarrhea) [17]. It is certainly possible that a significant proportion of these 30,000 deaths was caused by unrecognized cholera than would be estimated from individually identified deaths. Although the outbreak response focus of cholera surveillance in Uganda may be insufficient to accurately estimate the numbers of cases and deaths caused by cholera, these data are very useful for identifying areas to target for surveillance in consideration of future vaccine introduction. In order to better quantify the burden of cholera in Uganda, sentinel site surveillance should be undertaken in at least two regions with districts at high risk for cholera for a period of at least two years. It would be better to continue surveillance for at least five years, since cholera incidence is highly variable from year to year. The Ministry of Health is participating in the AFRICHOL cholera surveillance in Africa project (www.africhol.org), led by Agence de Me´decine Pre´ventive (AMP) and the African Field Epidemiological Network (AFENET), which is an African-based non-government organization working to improve field epidemi- ology and public health laboratories in sub-Saharan Africa. As part of this project, enhanced cholera surveillance is being conducted in five districts in Eastern Uganda (Mbale, Tororo, Manafwa, Butaleja and Busia) and, whenever outbreaks occur, throughout the country. Such data may be combined with the available national reporting statistics to better model cholera burden within Uganda, which in turn may be used to conduct economic analyses (e.g., cost effectiveness or cost benefit studies) of the potential use of cholera vaccines in Uganda. Given the health challenges facing Uganda, the decision to pursue cholera vaccination must be weighed against the introduction of other health interventions that may have a greater impact on mortality (e.g., pneumococcal conjugate vaccines, rotavirus vaccines, future malaria vaccines or other interventions). In addition to targeting high-risk endemic populations, Uganda may consider using cholera vaccines from a recently established international stockpile to mitigate epidemic cholera. A review of 15 epidemic curves showed that about 57% of the cases occurred after six weeks across all outbreaks. This 57% figure may represent an upper bound on the number of cases that could be averted via reactive use of cholera vaccines from a global stockpile, assuming it would take at least three weeks to diagnose an outbreak and prepare for a mass vaccination campaign plus three weeks to generate immunity from the two-dose vaccine. This stockpile may also be used to prevent the spread of cholera to neighboring districts, such as when it was used in an Adjumani district refugee camp in 1997 [18]. While cholera incidence in Uganda has been manageable over the past decade, elimination of the disease is likely to take time especially given the slow progress on provision of safe water and sanitation among other risk factors. Most of the areas with the highest incidence rates either border countries with political instability and endemic cholera (e.g., DRC and Sudan) or contain semi-nomadic populations. For these districts, it would be difficult to prevent cholera-infected persons from crossing borders, achieve PLOS Neglected Tropical Diseases | www.plosntds.org 7 December 2013 | Volume 7 | Issue 12 | e2545 Table 1. Estimated annual cholera incidence rates, number of cases and death by district risk category and age group. Total annual Estimated incidence Annual deaths Annual deaths District risk category Age group Population Reported cases cases (range) per 1,000 (range) (reported) (range) High ,1 369,000 50 50–200 14–55 1 1–4 1–4 1,208,000 363 360–1,400 30–120 1 1–4 5–14 2,386,000 538 540–2,200 23–90 8 8–24 15+ 4,172,000 1,285 1,300–5,100 31–120 38 38–115 All ages 8,135,000 2,235 2,200–8,900 28–110 49 49–148 Low or Unknown ,1 1,164,761 12 12–48 0.2–0.8 0.23 0–1 1–4 3,809,393 85 85–340 1.5–6 2 2–5 5–14 7,527,670 126 130–550 2–8 3 3–8 15+ 13,160,134 301 300–1,200 5–20 7 7–20 All 25,661,958 524 520–2,100 8.8–35 11 11–34 All districts ,1 1,534,000 62 62–250 1–4 1 1–4 1–4 5,017,000 448 450–1,800 7.5–30 3 3–10 5–14 9,914,000 664 660–2,700 11–44 11 11–33 15+ 17,332,000 1,586 1,600–6,300 26–105 45 45–135 All 33,797,000 2,760 2,800–11,000 46–180 61 61–182 doi:10.1371/journal.pntd.0002545.t001 The Burden of Cholera in Uganda Figure 6. Epidemic curves of cholera outbreaks in Uganda, 2002–2010. The epidemic curves from a convenience sample of fifteen outbreaks shows both large and small outbreak have occurred during the period 2002–2010. There is considerable variation in the duration of outbreaks with longer outbreaks in districts with higher average annual incidence rates. doi:10.1371/journal.pntd.0002545.g006 PLOS Neglected Tropical Diseases | www.plosntds.org 8 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda Table 2. A non-random sample of outbreaks in Uganda, 2002–2010. Time period Time period Time period District (month) (week) Cases District Time period (month) (week) Cases Amudat May 2010–Aug 2010 14 weeks 185 Kasese Jun 2009–Apr 2010 44 weeks 148 Arua Jun 2005–Sep 2005 16 weeks 289 Kitgum Feb 2006–Oct 2006 37 weeks 1,245 Arua Jul 2006–Oct 2006 16 weeks 135 Kotido May 2010–Jul 2010 12 weeks 491 Arua Dec 2007–Mar 2008 18 weeks 663 Moroto Apr 2010–Jun 2010 10 weeks 610 Arua Jun 2008–Nov 2008 21 weeks 73 Nebbi Dec 2002–Mar 2003 17 weeks 437 Bugiri Sep 2009–Oct 2009 8 weeks 188 Nebbi Oct 2007–Jan 2008 16 weeks 551 Bulisa Nov 2007–Dec 2007 6 weeks 291 Yumbe Nov 2007–Dec 2007 4 weeks 75 Kampala Nov 2006–Feb 2007 15 weeks 726 doi:10.1371/journal.pntd.0002545.t002 high vaccination coverage rates, or to construct reliable water and does not reduce the incidence of cholera cases, it does reduce the sanitation infrastructure for semi-nomadic populations [19]. social and economic burden of the disease. Improvements in access to improved water and sanitation Some global trends in cholera disease burden may lead to an increase in the number of cases and should be considered in would also lead to a concomitant decrease in cholera incidence. These cholera incidence data may also be used to target priority cholera control planning. At present, cholera is more prevalent in districts for improvements in water, sanitation, and hygiene efforts. rural areas than in urban areas within Uganda. This may change if Cholera incidence is likely to be associated with high prevalence of present urbanization trends continue and the maintenance and other enteric diseases, for which cholera vaccination would have expansion of water and sanitation infrastructure cannot keep pace no effect. Considering that an estimated 30,000 persons die from with the rapidly growing urban population. The urban population diarrheal disease every year in Uganda, improved water, in Uganda is projected to increase more than seven-fold from 4.5 sanitation, and hygiene are urgently needed even if cholera million in 2010 to 31 million by 2050 [20]. vaccine is deployed. Some studies have found multidrug resistant V. cholerae in In conclusion, the existing surveillance system is geared Uganda, including strains resistant to trimethoprim, sulfonamides, toward mitigating the impacts of cholera outbreaks, not ampicillin, tetracycline, chloramphenical and streptomycin [21]. quantifying the burden of endemic cholera. Cholera control In addition, there is evidence that the severity of clinical cases of activities have been effective in slowing the spread of cholera and cholera in Asia and Africa is increasing, especially during reducing cholera fatalities. However, cholera cases continue to outbreaks. Some scientists attribute the increase in severity of be reported on an annual basis. The combination of sentinel cholera cases seeking treatment to the emergence of a new altered surveillance with national cholera incidence data could be used strain of V. cholerae O1 El Tor that secretes the classical cholera to develop an economic analysis to inform cholera vaccination toxin, making it more virulent [22]. It is not presently known if this policy. strain is present in Uganda. However, it has been isolated from recent African outbreaks in Mozambique and Zimbabwe [23,24]. Acknowledgments Due to global warming, the average temperature in Uganda is estimated to increase by up to 1.5 degrees over the next 20 years The authors wish to acknowledge Mr. Ateki Kagirita, Head Central Public Health Laboratory (CPHL), for providing laboratory methods and results; [25]. Recent research suggests a strong correlation between Mr. Jung Seok Lee of the International Vaccine Institute for assistance in increased rainfall and elevated temperatures with higher cholera creating GIS maps; and Drs. John D. Clemens, Anna Lena Lopez, and incidence [26,27,28,29,30]. This may pose an elevated risk for Thomas F. Wierzba of the International Vaccine Institute for assistance in districts bordering Lake Albert and Lake Victoria, which may developing and coordinating the study. The manuscript was edited by the provide an endemic reservoir of V. cholerae [31,32]. IVI Communications and Advocacy Unit. There are also trends suggesting a reduced need for cholera vaccination in Uganda. The multidisciplinary cholera outbreak Author Contributions response activities have been effective in mitigating the severity of Conceived and designed the experiments: GB MM BM AL. Performed the outbreaks, both in terms of morbidity and mortality. The cholera experiments: GB MM BM YEK. Analyzed the data: GB MM BM YEK case fatality rate has steadily declined since the large, nationwide AL. Contributed reagents/materials/analysis tools: GB MM. Wrote the outbreak in 1998 (refer to Figure 1). While improving treatment paper: GB BM AL VM. References 1. World Health Organization (2000) Report on Global Surveillance on epidemic 4. World Health Organization (2010) Cholera vaccines: WHO position paper. prone infectious diseases, WHO/CDS/CSR/ISR/2000.1 http://www.who.int/ Weekly Epidemiological Record 13: 117–128. csr/resources/publications/surveillance/en/cholera.pdf accessed on July 1, 5. World Health Organization (2010) Cholera, 2009. Weekly Epidemiological 2012. Record 85: 293–308. 2. Legros D, McCormick M, Mugero C, Skinnider M, Bek’obita D, et al. (2000) 6. The Republic of Uganda Ministry of Health (2007) Prevention and control of Epidemiology of cholera outbreak in Kampala, Uganda. East African Medical cholera, operational guidelines for the district health workers and planners. Journal 77: 347–349. Control of Diarrheoal Diseases Section CHD,. Kampala. 3. Nkoko DB, Giraudoux P, Plisnier P-D, Tinda AM, Piarroux M, et al. (2011) 7. Kirigia JM, Sambo LG, Yokouide A, Soumbey-Alley E, Muthuri LK, et al. Dynamics of Cholera Outbreaks in Great Lakes Region of Africa, 1978–2008. (2009) Economic burden of cholera in the WHO African region. BMC Emerging Infectious Disease 17: 2026–2034. International Health and Human Rights 9: 1–14. PLOS Neglected Tropical Diseases | www.plosntds.org 9 December 2013 | Volume 7 | Issue 12 | e2545 The Burden of Cholera in Uganda 8. Poulos C, Riewpaiboon A, Stewart JF, Clemens J, Guh S, et al. (2011) Costs of 21. Kruse H, Sørum H, Tenover F, Olsvik O (1995) A transferable multiple drug Illness Due to Endemic Cholera. Epidemiology and Infection Apr 18: 1–10. resistance plasmid from Vibrio cholerae O1. Microbial Drug Resistance 1: 203– 9. Shikanga O-T, Mutonga D, Abade M, Amwayi S, Ope M, et al. (2009) High Mortality in a Cholera Outbreak in Western Kenya after Post-Election Violence 22. Siddique AK, Nair GB, Alam M, Sack DA, Huq A, et al. (2009) El Tor cholera in 2008. American Journal of Tropical Medicine and Hygiene 81: 1085–1090. with severe disease: a new threat to Asia and beyond. Epidemiology and 10. Kuttiat VS, Lodha R, Das B, Kohli U (2010) Prevalence of Cholera in Pediatric Infection 138: 347–352. Patients with Acute Dehydrating Diarrhea. Indian Journal of Pediatrics 77: 67– 23. World Health Organization (2009) Cholera, Zimbabwe – update. Weekly Epidemiological Record 84: 109–110. 11. Nair GB, Ramamurthy T, Bhattacharya MK, Krishnan T, Ganguly S, et al. 24. Ansaruzzaman M, Bhuiyan NA, Nair B, Sack DA, Lucas M, et al. (2004) (2010) Emerging trends in the etiology of enteric pathogens as evidenced from an Cholera in Mozambique, Variant of Vibrio cholerae. Emerging Infectious active surveillance of hospitalized diarrhoeal patients in Kolkata, India. Gut Disease 10: 2057–2059. Pathogens 2: 4. 25. Hepworth N (2010) Climate Change Vulnerability and Adaption Preparedness 12. Alajo SO, Nakavuma J, Erume J (2006) Cholera in endemic districts in Uganda in Uganda. Nairobi, Kenya. during El Nin˜o rains: 2002–2003. African Health Sciences 6: 93–97. 26. Constantin de Magny G, Murtugudde R, Sapiano M, Nizam A, Brown C, et al. 13. United Nations Office for the Coordination of Humanitarian Affairs (2010) (2008) Environmental signatures associated with cholera epidemics. Proceedings Uganda: Cholera Affected Areas & Cases by Parish (as of 12 July 2010). Relief of National Academies of Science USA 105: 17676–81. Web report, http://reliefwebint/node/17116 (accessed on 1 Nov 2010). 27. Emch M, Feldacker C, Yunus M, Streatfield P, Thiem V, et al. (2008) Local 14. Deen J, von Seidlein L, Sur D, Agtini M, Lucas M, et al. (2008) The high burden environmental predictors of cholera in Bangladesh and Vietnam. American of cholera in children: Comparison of incidence from endemic areas in Asia and Journal of Tropical Medicine and Hygiene 78: 823–832. Africa. PLoS Neglected Tropical Diseases 2: e173. 28. Ferna´ndez MAL, Bauernfein A, Jime´nez JD, Gil CL, Omeiria NE, et al. (2009) 15. Harris A, Chowdhury F, Begum Y, Khan A, Faruque A, et al. (2008) Shifting Influence of temperature and rainfall on the evolution of cholera epidemics in prevalence of major diarrheal pathogens in patients seeking hospital care during Lusaka, Zambia, 2003—2006: analysis of a time series. Tropical Medicine and floods in 1998, 2004, and 2007 in Dhaka, Bangladesh. American Journal of Hygiene 103: 137–143. Tropical Medicine and Hygiene 79: 708–714. 29. Lobitz B, Beck L, Huq A, Wood B, Fuchs G, et al. (2000) Climate and infectious 16. Sack RB, Siddique AK, Longini Jr IM, Nizam A, Yunus M, et al. (2003) A 4- disease: use of remote sensing for detection of Vibrio cholerae by indirect year study of the epidemiology of Vibrio cholerae in four rural areas of measurement. Proceedings of National Academies of Science USA 97: 1438– Bangladesh. Journal of Infectious Diseases 187: 96–101. 17. The Republic of Uganda Ministry of Health (2009) Ministry of Health Statistical 30. Reyburn R, Kim DR, Emch M, Khatib A, Seidlein Lv, et al. (2011) Climate Abstract 2008/09. Variability and the Outbreaks of Cholera in Zanzibar, East Africa: A Time 18. Dorlencourt F, Legros D, Paquet C, Neira M, Ivanoff B, et al. (1999) Series Analysis. American Journal of Tropical Medicine and Hygiene 84: 862– Effectiveness of mass vaccination with WC/rBS cholera vaccine in Adjumani district, Uganda. Bulletin of the World Health Organization 77: 949–950. 31. Bompangue D, Giraudoux P, Handschumacher P, Piarroux M, Sudre B, et al. 19. Cummings MJ, Wamala JF, Eyura M, MalimboM, Omeke EM, et al. (2011) A (2008) Lakes as source of cholera outbreaks, Democratic Republic of Congo. . cholera outbreak among semi-nomadic pastoralists in northeastern Uganda: epidemiology and interventions. Epidemiology and Infection 140: 1376–85. Emerging Infectious Disease 14: 798–800. doi:10.1017/S0950268811001956. 32. Shapiro R, Otieno M, Adcock P, Phillips-Howard P, Hawley W, et al. (1999) 20. Population Division of the Department of Economic and Social Affairs of the Transmission of epidemic Vibrio cholerae O1 in rural western Kenya associated United Nations Secretariat (2007) World Population Prospects: The 2006 with drinking water from Lake Victoria: an environmental reservoir for cholera? Revision and World Urbanization Prospects: The 2007 Revision. . American Journal of Tropical Medicine and Hygiene 60: 271–276. PLOS Neglected Tropical Diseases | www.plosntds.org 10 December 2013 | Volume 7 | Issue 12 | e2545

Journal

PLoS Neglected Tropical Diseases – Public Library of Science (PLoS) Journal

Published: Dec 5, 2013

There are no references for this article.