Predictive and easy-to-use prognostic tools, which stratify the risk of EVD mortality at or after EVD triage.
February 4th, 2017PLOS: Hartley M-A, Young A, Tran A-M, Okoni-Williams HH, Suma M, Mancuso B, et al. (2017) Predicting Ebola Severity: A Clinical Prioritization Score for Ebola Virus Disease. PLoS Negl Trop Dis 11(2): e0005265. doi:10.1371/journal.pntd.0005265
- 158 Ebola patients: Study population
- The authors were able to accurately predict death at triage 91% of the time and death after triage 97% of the time.
- Co-infection with malaria was associated with a 2.5-fold increase in the odds of death.
- Disorientation, hiccups, diarrhea, conjunctivitis, shortness of breath, and muscle aches were also strong predictors of death.
- Age was also a predictor of mortality.
- The patient group aged between 5 and 24 years had the lowest mortality rate of 42.5%
- The over-45’s and under-5’s were particularly vulnerable, being 11.6 and 5.4 fold more likely to die, respectively.
Background
Methods/Principal Findings
This retrospective cohort study analyses the clinical characteristics of 158 EVD(+) patients at the GOAL-Mathaska Ebola Treatment Centre, Sierra Leone. The prognostic potential of each characteristic was assessed and incorporated into a statistically weighted disease score. The mortality rate among EVD(+) patients was 60.8% and highest in those aged <5 or >25 years (p<0.05). Death was significantly associated with malaria co-infection (OR = 2.5, p = 0.01). However, this observation was abrogated after adjustment to Ebola viral load (p = 0.1), potentially indicating a pathologic synergy between the infections. Similarly, referral-time interacted with viral load, and adjustment revealed referral-time as a significant determinant of mortality, thus quantifying the benefits of early reporting as a 12% mortality risk reduction per day (p = 0.012). Disorientation was the strongest unadjusted predictor of death (OR = 13.1, p = 0.014) followed by hiccups, diarrhoea, conjunctivitis, dyspnoea and myalgia. Including these characteristics in multivariate prognostic scores, we obtained a 91% and 97% ability to discriminate death at or after triage respectively (area under ROC curve).