During the first week of May 2003, the Early Warning and Response Network, established in 1999 in southern Sudan, reported an outbreak of fatal hemorrhagic fever of unknown etiology in the Imatong region of Torit County, which is near the Ugandan border, in a mountainous area covered with tropical rain forest. During the civil unrest in early 2002, many residents were relocated to an internally displaced persons camp in Ikotos County, but in 2003, a number of the residents moved back to the Imatong region. During April and May 2003 suspected cases of hemorrhagic illness were reported, and blood samples collected from Sarianga, Itohom, Lenyleny, Tarafafa, Lofi, and Locomo villages were tested at the Kenya Medical Research Institute (KEMRI), in Nairobi, where yellow fever virus was identified as the causative agent of the outbreak. Aliquots of specimens were submitted to the Special Pathogens Unit at the National Institute for Communicable Diseases in Johannesburg for confirmation of the diagnosis. A report dated June 19, 2003, from the Office of the United Nations Resident and Humanitarian Coordinator for the Sudan described a steady decline in the number of cases occurring during the 4 weeks before the report with a total of 162 suspected cases with 48 deaths in Torit County. However, the report also noted that the recording of suspected cases was inadequate, the figures quoted were an estimate, and final case numbers and deaths were uncertain. The details of the outbreak and laboratory investigations are reported in a separate publication.

Virus isolates were obtained from three patients, two from Locomo village and one from Lofi village. The viral isolations were performed by injecting serum samples from all the patients into suckling mice. RNA was extracted from harvested mouse brain tissue by using Trizol reagent (Gibco BRL, Life Technologies, Gaithersburg, MD), according to the manufacturer's directions. A 670-bp region of the polyprotein gene, which included the 3´ end of the premembrane protein gene, the complete membrane protein gene, and the 5´ end of the envelope protein gene, was amplified by using reverse transcription–polymerase chain reaction (RT-PCR) (2). The nucleotide sequence for each amplicon was determined by using Big Dye Terminator Sequencing Ready Reaction kits with AmpliTaq DNA polymerase FS, as described by the supplier (Applied Biosystems, Warrington, UK). A consensus sequence was established by aligning the sequences obtained from the Sudan 2003 outbreak of yellow fever with sequences obtained from GenBank for yellow fever isolates from previous outbreaks of the virus in Africa (Table). The sequences from GenBank were selected to represent the five distinct genotypes circulating in Africa (2).

Alignment of the nucleotide sequence data was performed by using DNASIS for Windows version 2.5 (Hitachi Software Engineering America, San Francisco, CA). The phylogenetic analysis was performed on a 572-bp region of the amplicons by using a weighted maximum parsimony method, transversion:transition weighting of 5:1, and Phylogenetic Analysis Using Parsimony (PAUP) software version 4.0b4a for Macintosh (3). Bootstrap confidence intervals were calculated from 100 heuristic search replicates. The tree (Figure) shows that the virus circulating during the recent outbreak in southern Sudan was closely related to an isolate from an outbreak in Kenya in 1993 that belonged to the East African genotype. Sequence divergence was determined by using Molecular Evolutionary Genetics Analysis (MEGA) version 2.1 (4) to calculate the average p-distances or the proportion of pairwise differences within and between groups. The recent Sudan isolates clustered with the East African isolates. A pairwise comparison of nucleotide sequences within the group showed an average distance of 0.111 and 0.179 between the Sudan isolates and an isolate from Kenya (1993) and isolates from Uganda (1948), respectively. No insertions or deletions were found in the Sudan nucleotide sequences compared with the reference strains; however, a number of nucleotide substitutions occurred, most of which were synonymous, and a pairwise comparison of the predicted amino acid sequences showed a high degree of homology. The predicted amino acid sequences were identical for the Kenyan and the Sudan isolates, while the p-distance calculated between the Sudan and Ugandan isolates from 1948 was 0.016. The nucleotide heterogeneity between isolates from East and Central Africa (1.6%–11.6%) supports the concept that different genotypes are circulating and that outbreaks are not the result of imported virus from urban areas (2).

Outbreaks of yellow fever virus have frequently been reported from areas within West Africa since the 18th century, with far fewer outbreaks being identified in East Africa. Serologic evidence for the presence of yellow fever virus in Sudan, Kenya, Uganda, and Tanzania was first reported in 1936 (5). However, not until 1940 was the first epidemic confirmed in East Africa, in central Sudan (5). Sporadic cases were identified annually in East Africa until 1959, when an outbreak was recorded in the Blue Nile region of Sudan and subsequently in the neighboring region of Ethiopia (6). From 1960 to 1962, the largest outbreak to date in Africa occurred in southwest Ethiopia. Additional serologic studies confirmed that yellow fever activity was widespread in Uganda, Somalia, Ethiopia, and Kenya (6). Although two possible cases of yellow fever in Kenya were reported in 1943 (7), not until 1992–1993 was a large outbreak confirmed in the Rift Valley province of Kenya (7,8). Subsequent sporadic isolations of virus have been made in East Africa since then, but no large epidemics were recognized until the outbreak in southern Sudan in 2003. Originally, researchers speculated that outbreaks in East Africa were less frequent and smaller than the large outbreaks recorded in West Africa because they were the result of virus's being introduced into the area at the time of the outbreak. However, the genetic data suggest that yellow fever virus is endemic in East and Central Africa, with outbreaks occurring as a result of favorable environmental conditions (2). The fact that the isolates from Sudan were closely related to an isolate obtained 10 years ago in Kenya supports the contention that yellow fever is endemic in East Africa and has the potential to cause large outbreaks when conditions favor transmission to humans.