Ebola Symptoms

CDC's consumer page on Ebola symptoms is dated April 2024 and describes the disease in general terms; it does not distinguish Bundibugyo virus from the Zaire species that caused every large outbreak before 2026. The three columns below are the largest published series for each. The West Africa figures come from the WHO Ebola Response Team's analysis in NEJM of 3,343 confirmed and 667 probable Zaire-species cases through September 14, 2014. The Uganda 2026 column is from CDC's MMWR report of September 10, 2026 — the first peer-reviewed clinical description of the current outbreak's species — and covers the 18 confirmed Bundibugyo patients admitted to the Mulago Hospital treatment unit. The 2012 column is Kratz et al.'s prospective study of 18 treatment-centre patients in the Isiro (DRC) Bundibugyo outbreak.

Three things stand out. First, fever is not universal in the largest series: one in eight West African patients had no fever recorded before detection, which is one reason a fever-only screening rule misses cases. Second, the 2026 Bundibugyo patients in Uganda presented with a more "aches-and-pains" picture — headache, muscle and joint pain in three-quarters or more — and less vomiting and diarrhea than the West African average, although these were patients caught early enough to reach a treatment unit, and 18 patients is a very small base. Third, in every series bleeding was the exception, not the rule.

The 2007 Bundibugyo Series: the First Description of the Species

The species was identified during the 2007–2008 outbreak in Bundibugyo District, Uganda. Roddy et al. described the 26 laboratory-confirmed patients who were hospitalised: the most frequent symptoms were non-bloody diarrhoea (81%), severe headache (81%) and asthenia (77%). Seven of the 26 had haemorrhagic symptoms, and six of those seven died — a pattern the 2026 Uganda series repeated exactly (see below). The earlier field investigation by Wamala et al. found that most transmission in 2007 was associated with handling the dead without protection (adjusted odds ratio 3.83, 95% CI 1.78–8.23), which is why the safe-burial indicators in the 2026 response matter so much.

Because the early symptoms are non-specific, clinicians have looked for laboratory patterns that separate Ebola from malaria or typhoid before the PCR result is back. The 2026 Uganda series is the first to report this for Bundibugyo virus. On admission, all 18 patients had elevated liver enzymes (AST), low albumin, and low sodium; 78% had elevated ALT, 72% had a low white-cell count, 61% had raised C-reactive protein, and 28% had low platelets. No patient was anaemic on admission (CDC, MMWR 2026;75(35):551). The authors suggest these three universal findings "might be characteristic features of early BVD" — a hypothesis from 18 patients, not an established rule.

It matches what was seen in the 27 patients treated in US and European hospitals in 2014–2015, where the predominant findings during the illness were diarrhea, low albumin, low sodium, low potassium, low calcium and low magnesium; 52% developed low blood oxygen and a third had reduced urine output (Uyeki et al., NEJM 2016;374:636). Viral RNA in blood peaked a median of 7 days after symptom onset and cleared after a median of 17.5 days. Those patients were the best-monitored Ebola cases ever recorded, and 81.5% survived.

A 2024 meta-analysis of 42 outbreaks in 16 countries from 1976 to 2022 (Izudi et al., Journal of Infection and Public Health) is the best single summary of how lethal each species has been:

The 2026 DRC figure sits above the upper bound of the historical confidence interval for its own species. Three explanations are on the table, and the sources do not settle which dominates:

• Care is arriving late or not at all. CDC's operational review of the DRC response found that 59% of confirmed deaths were occurring outside an Ebola treatment unit and that only 15–20% of new cases had previously been identified as contacts (CDC MMWR 2026;75(35):554). The Uganda cases, nearly all of whom reached a treatment unit, had a 10% fatality rate with the same virus.
• Mild cases are being missed. A crude case fatality ratio divides deaths by confirmed cases. If infected people with milder illness never get tested — plausible where 72% of alerts were tested and test positivity was 24% — the denominator is too small and the ratio too high. WHO's 2014 analysis made the opposite point about crude ratios in a growing epidemic: deaths lag cases, so the crude ratio can also run low. Both biases are present at once.
• A genuine difference. WHO reported one province, North Kivu, with a CFR of 65.4% and said investigations into the reasons were ongoing (DON 617). No published analysis yet attributes the 2026 mortality to the virus itself.

Our calculation, for context: WHO reported 3,267 deaths and "at least" 1,590 recoveries among DRC's confirmed cases as of September 7, 2026. Counting only patients with a known outcome, 3,267 ÷ (3,267 + 1,590) = 67.3%. Because the recovery count is a floor, this is an upper bound that will fall as discharges are recorded; the true known-outcome rate lies somewhere between the 48.3% crude figure and this number. Neither is a property of the virus alone.

Compared to what? Ebola's fatality rate is extreme by any measure, but its exposure is rare. To see how it compares with 14 other viruses on exposure, hospitalization and death, see the network's Virus Risk Perspective and this site's Ebola risk perspective.