Menu
Menu
Menu
Menu
Background: Since December 2009, Médecins Sans Frontières has diagnosed and treated patients with visceral leishmaniasis (VL) in Tabarak Allah Hospital, eastern Gedaref State, one of the main endemic foci of VL in Sudan. A survey was conducted to estimate the VL incidence in villages around Tabarak Allah. Methods: Between the 5th of May and the 17th of June 2011, we conducted an exhaustive door-to-door survey in 45 villages of Al-Gureisha locality. Deaths were investigated by verbal autopsies. All individuals with (i) fever of at least two weeks, (ii) VL diagnosed and treated in the previous year, and (iii) clinical suspicion of post-kala-azar dermal leishmaniasis (PKDL) were referred to medical teams for case ascertainment. A new case of VL was a clinical suspect with a positive rk39 rapid test or direct agglutination test (DAT). Results: In the 45 villages screened, 17,702 households were interviewed, for a population of 94,369 inhabitants. The crude mortality rate over the mean recall period of 409 days was 0.13/10’000 people per day. VL was a possible or probable cause for 19% of all deaths. The VL-specific mortality rate was estimated at 0.9/1000 per year. The medical teams examined 551 individuals referred for a history of fever of at least two weeks. Out of these, 16 were diagnosed with primary VL. The overall incidence of VL over the past year was 7.0/1000 persons per year, or 7.9/1000 per year when deaths possibly or probably due to VL were included. Overall, 12.5% (11,943/95,609) of the population reported a past VL treatment episode. Discussion and Conclusion: VL represents a significant health burden in eastern Gedaref State. Active VL case detection had a very low yield in this specific setting with adequate access to care and may not be the priority intervention to enhance control in similar contexts. © 2012 Mueller et al.
Ethical clearance was granted from the Sudanese National Ministry of Health’s Research Ethics Review Committee. Written authorization to conduct the study was obtained from the Gedaref Ministry of Health and each head of village. Each head of household provided oral informed consent to the collection of demographical data, history of VL treatment, skin rash after treatment, and presence of fever of at least two weeks among household members. A referral form was given for each individual presenting with fever of more than two weeks, with suspicion of PKDL or having been treated for VL in the last year. The information included in these forms was not identifying and individuals were free to reach or not the medical team for clinical investigation. An additional oral consent was obtained from clinical suspects before testing for VL. The choice of oral consent was made because of the low literacy rate in the study area and the unlikelihood to easily find an impartial literate witness for each household. The Sudanese National Ministry of Health’s Research Ethics Review Committee expressly approved the method of oral consent without use of a witness or written record of oral consent. Between the 5th of May and the 17th of June 2011, we conducted an exhaustive door to door survey in the 45 villages of Al-Gureisha locality, covering a population of about 85,000 inhabitants. The survey villages were grouped into four geographical areas. Each area was surveyed by four field teams and one medical team. Demographic information (age, sex, household composition on the day of survey and one year prior, number of births, deaths and movements within the past year) was collected by the field teams in each household. A household was defined as all people living together under the responsibility of one head of household and eating regularly together. For each household member, the history of VL treatment and possible subsequent PKDL was also recorded. The number and the causes of any death occurring in the past year were investigated in order to identify deaths possibly attributable to VL. Verbal autopsies were conducted for all reported deaths except for neonatal, delivery-related, and accidental deaths, as these were unlikely to be related to VL. Maternal deaths not directly related to delivery were investigated, as VL during pregnancy is known to be associated with increased treatment toxicity and mortality [12], [13]. Individuals with fever of at least two weeks duration, individuals diagnosed and treated for VL during the past year, and clinical suspects of either PKDL or VL relapse (independently of the time elapsed since treatment) were referred to the medical teams for clinical examination and case ascertainment. New clinical VL suspects (defined as fever for at least two weeks with at least one of the following: splenomegaly, lymphadenopathies or history of weight loss) were tested with an rK39 antigen-based rapid test (DiaMed IT-Leish) [14] and, if negative, with the direct agglutination test (DAT) [15], [16] for VL confirmation. A new VL case was defined as a clinical suspect who was confirmed either by the rK39 or the DAT. New VL cases, suspected VL relapses, and moderate and severe PKDL cases were referred to Tabarak Allah Hospital. Because of the self-healing nature of PKDL in Sudan and the potential toxicity of the recommended SSG treatment, mild PKDL cases were not offered SSG treatment [6] and therefore were not referred to Tabarak Allah Hospital. To estimate the incidence rates at the village level, the population was exhaustively screened. We calculated a sample size of 266 deaths to estimate a proportion of deaths due to VL of 30% with a 5% precision (alpha 0.05). Based on an expected total number of deaths around 1500 (corresponding to an annual mortality rate of 0.5/10’000 persons per day), we planned to investigate the cause of every fifth death through verbal autopsy, using a systematic sampling procedure. All deaths were recorded consecutively on a tally sheet, with the death to be investigated pre-highlighted. As the data collected during the first three weeks of the survey showed a number of deaths much lower than expected, we later conducted verbal autopsies for every reported death. The analysis of the causes of death was weighted accordingly. All verbal autopsies were reviewed independently by two clinicians experienced in VL and fluent in Arabic. In case of disagreement, the files were reviewed by a third expert clinician, with the help of a translator, and his verdict was final. Death was considered possibly due to VL if the respondent mentioned fever of at least two weeks duration and either one of the following: enlarged lymph nodes, a visible mass in the left upper part of the abdomen (spleen side), or weight loss, during the final illness of the deceased. Death was considered as probably due to VL if it occurred during treatment for VL (clearly mentioned by the relatives of the deceased) in a treatment facility offering reliable VL diagnosis (i.e. rk39 rapid test, DAT or microscopic examination of lymph node aspirate with quality control in place). If a death was reported to have occurred in another treatment facility during VL treatment, it was considered as possibly due to VL. The event chosen to define the start of the recall period (covering the past year) was the presidential elections in Sudan, which occurred on the 10th and 11th of April 2010. The average recall period (referred hereafter as the “the past year”) was therefore 409 days. The end of the sesame harvest (end of October 2010) was used to define a 6-month recall period. VL incidence rate over the period was calculated by summing the new VL cases detected during the survey, the VL cases and the deaths possibly/probably due to VL reported over the recall period, divided by the mid-year population. All documents were translated in Arabic and back-translated into English, and were subjected to pilot testing with subsequent update before the start of the survey. Data were entered in the EpiData software (EpiData, Odense, Denmark) by four data entry clerks. Data were analysed using the Stata 11 software (Stata Corporation, College Station, Texas, USA). Description of geographical information was performed using the QuantumGIS software, version 1.7.0. The coordinates of the Atbarah River were obtained by manually drawing along the river in Google Earth.
