Three ebola outbreaks in Uganda 2000-2011

Abstract

Three separate outbreaks of Ebola associated with high fatality occurred in Uganda between 2000 and 2011. A country wide national response contained each epidemic with various degrees of success .The experiences challenges and successes are described in Gulu, Bundibugyo and Luwero outbreaks.Objectives: The study linked the following objectives: to describe the three Ebola outbreaks and the national response in Uganda from 2000-2011; to establish the risk factors associated with the Bundibugyo ebolavirus outbreak, 2007; to estimate the case fatality rate related to the Bundibugyo ebolavirus outbreak, 2007. Methods: A descriptive design study documented the three different outbreaks. The occurrence and epidemiological characteristics of each epidemic were defined and described basedd on the adapted case definition for Ebola haemorrhagic fever. Data and information was systematically collected from cases and contacts routinely using questionnaires covering personal, demographic and social parameters. Active case search, isolation, as well as community mobilisation for public education including the media were the major strategies used. Risk factors were studied using a case control design (Paper III) which compared cases and non-cases in a sub set of presumed cases identified at community level. A quantitative study of a subset of cases with only laboratory confirmed acute phase blood samples estimated the corresponding case fatality rate (Paper IV). Results: Two large outbreaks of the Sudan ebolavirus and the Bundibugyo ebolavirus occurred in Uganda in 2000 (425 cases) and 2007 (116 cases) respectively, followed in 2011by a single case outbreak in Luwero. Clinical characteristics were similar in all the three outbreaks: acute onset of high grade fever, severe headache and chest pain, abdominal pain, associated with some bleeding tendencies. Clustering of cases was common. The case definition helped in screening suspected cases but the major weakness of this approach was that it had a low specificity, and less than 50% of suspected cases identified by the community were confirmed as true Ebola patients. In Gulu, attack rates were higher among women than men (RR=1.6; 95% C.I. = 1.3; 1.9). Children between 5-14 years had the lowest attack rate. The risk increased with age and was highest at 60-64 years age group (RR =16.4; 95% CI = 9.4; 28.8). Case fatality was highest in the SEBOV subtype (53%; 955 CI = 47.8-57.5) and lowest in the BEBOV (33.6%; 95% CI 25.0-42.2; p=0.005). There was two fold increase in mortality (RR=1.8, p value <0.001) when bleeding manifestations occurred in patients with the Ebola Sudan subtype. Ebola is a highly fatal nosocomial infection. Delayed detection often resulted in spread of infection in health care settings. Some 31 health care workers in Gulu and another 14 in Bundibugyo were infected during the outbreaks. Direct contact with a known case (OR 7.4, 95% CI 2.9-19.3) was probably the major mode of spread as demonstrated in Bundibugyo. Sex differences were not associated with significant risk factors (OR 1.3, 95% CI 0.7-2.5) and age (OR 1.3, 95% CI 0.7-2.7) unlike in the Gulu observation. Seasonality was observed in the three outbreaks- erupting between May and December, which coincides with the rainy and fruit season. Evidence of asymptomatic past infection was demonstrated among individuals with Ebola positive IgG in the districts of Gulu, Luwero and Mpigi. The known primary cases that started the outbreaks came from rural remote areas. A zoonotic connection was apparent but unclear although one monkey specimen was found positive for Ebola IgG. The study demonstrated weakneses in infection control as isolation procedures were apparently less effective. Despite instituting isolation procedures, 64% of the 31 health care workers in Gulu were infected after the isolation units were established, thus showing gaps in procedures for infection control.1). Prompt detection and communication was demonstrated to be effective in containing the Luwero outbreak which resulted in the best outcomes. Strengthening laboratory capacity and surveillance, therefore, at national level and enhancing collaborative networks at regional and international levels is crucial for effective timely diagnosis and management. Ad hoc incentives improved staff commitment, demonstrating that better remuneration of health care workers may contribute to better performance. Involving local communities and the media in outbreak control activities supported community based surveillance and timely identification of cases, in areas without health care workers. Ethical principles were breached and waiver of informed consent was considered a practical option under these life threatening circumstances, demonstrating further the difficulties of doing research under emergency situations. Conclusion: The experiences and challenges from the three Ebola outbreaks in Uganda have been described. Attempts were also made to establish the risk factors and severity. Prompt detection and communication yielded the best ideal outcome and timely containment. Not all was one hundred percentage perfect, but the Ministry of Health working alongside its partners and the community contained the outbreaks against the constraints of the low resource settings, sometime with delays but once promptly and effectively done.