Ebola virus disease and critical illness

While the cumulative case fatality proportion in West Africa is approximately 40 % as of May 2016, it has varied substantially during the course of the outbreak (being higher near the beginning). In comparison, the cumulative case fatality proportion for patients treated in Western Europe and the USA during 2014–2015 was 18.5 % [40] (Table 2). Case fatality remains highest among young children and older adults [5, 6, 16, 43]. Pregnant women often experience spontaneous abortion followed by bleeding, as well as preterm labor and stillbirth if Ebola virus infection occurs later in pregnancy. Vertical transmission and subsequent neonatal mortality has been virtually uniform in the few documented live births by women with acute EVD [44]. A high Ebola viral load at time of admission is associated with more severe illness and mortality [5, 6, 29, 43, 45‐47], with other markers of organ dysfunction variably associated with outcomes [6, 8, 29, 48].

The management of critically ill EVD patients in a resource-constrained setting has historically been restricted to variable monitoring of daily clinical signs and symptoms without access to continuous assessment [8]. The need for strict IPC practices and separation of patient care areas in West Africa significantly limited documentation and review of daily clinical records both inside and outside of high-risk patient care. As the case burden decreased and the ratio of healthcare personnel to patients increased, assessments were performed more systematically at some Ebola treatment facilities, with temperature, heart rate, blood pressure, respiratory rate, pulse oximetry, qualitative descriptions of urine, and gastrointestinal output, as well as fluid balance estimation.

Laboratory data other than Ebola RT-PCR results were essentially unavailable at Ebola treatment facilities in West Africa early in the outbreak [33, 49, 50]. Initially, there was very limited attention to diagnostics other than Ebola RT-PCR. The point-of-care systems for monitoring biochemistry and hematology parameters, such as the i-STAT® or the Piccolo Xpress®, were inconsistently utilized inside Ebola treatment facilities [5, 6, 48], in part due to limited manufacturer-recommended temperature and humidity ranges. Over the course of the West African outbreak, with support from national and deployed international laboratories, basic biochemistry, blood counts, and coagulation profiles helped to characterize the course of illness, but remained inconsistently available and often with substantial delays in results reporting due to transport and processing time. Malaria rapid diagnostic tests, and less commonly RT-PCR, were available at most of the laboratories that supported Ebola treatment facilities in West Africa. However, testing for Lassa fever virus (endemic in EBOV-affected countries) or other causes of sepsis was not routine.

Bedside ultrasonography has not been widely deployed [51] but could help with assessing volume status, responsiveness to intravenous fluids [50], and assessment of challenging clinical signs such as abdominal distension [42]. Plain chest and abdomen radiography has been performed in European and North American settings [52] but has rarely been available to patients with Ebola in West Africa. Among patients treated in the USA and Europe, pulmonary edema has been reported in 44 % and acute respiratory distress syndrome in another 22 % [40].

The World Health Organization (WHO) recommends considering discharge of patients from isolation on the basis of a negative blood Ebola virus RT-PCR result taken at least 3 days after the resolution of symptoms [33, 41]. However, Ebola virus can persist in certain body fluids after it is undetectable in the blood and after clinical recovery from EVD [42, 53‐56]. Viable Ebola virus has been isolated from urine, semen, cerebrospinal fluid, and vitreous humor many months after blood clearance [42, 56‐61], suggesting that some activities (unprotected sex, invasive procedures, or penetrating eye trauma) confer a transmission risk even after symptoms and viremia resolve. It is therefore critical to counsel EVD survivors about the risks of Ebola virus persistence [59, 60] and appropriate precautions after discharge, including barrier protection during sexual intercourse [50] until semen has tested negative for Ebola virus twice, or for at least 6 months after EVD onset [62]. Health workers and others should continue to apply standard precautions, as with all patients, when evaluating EVD survivors in the convalescent period. Additional infection prevention and control practices, based upon an individual patient risk assessment, may be prudent for specific procedures in convalescence, even when there is no detectable EBOV in the blood (e.g., lumbar puncture or vitreous humor sampling) [63].

Providing supportive care to critically ill patients with EVD in resource-poor settings is challenging [64] due to limited infrastructure, lack of materials and trained healthcare personnel, and uncertainty regarding the translation of modern sepsis treatment strategies [65] and optimal intravenous fluid management protocols in the absence of advanced monitoring used in resource-rich settings [66]. Respiratory symptoms such as cough are not a prominent feature of EVD and tachypnea likely represents respiratory compensation of severe metabolic acidosis [9, 30, 50]. Translating fluid resuscitation protocols used in a resource-rich setting [42] to settings where supplemental oxygen therapy is not routinely available [49], in a disease with possible vascular leak syndrome [38], could result in increased morbidity [67] and warrants further investigation [68‐70]. The use of antibiotics is common at Ebola treatment facilities [5, 6] before the diagnosis of EVD is confirmed in febrile patients and as empiric treatment of potential bacterial co-infection or gastrointestinal bacterial translocation in patients with confirmed EVD [9]. However, disruption of gastrointestinal flora due to broad-spectrum antibiotics could exacerbate diarrhea and fluid losses. Symptomatic treatment of severe diarrhea with loperamide was variably employed across Ebola treatment facilities. The risks and benefits of these practices warrant evaluation with observational studies and clinical trials [71].

Early and during the peak of the outbreak, clinical management was generally limited to supportive care focusing on aggressive oral and occasional intravenous volume resuscitation. As the case numbers decreased, advanced care became more common in some treatment facilities. Despite the limitations of working in PPE, feasibility and safety of central venous catheter placement was demonstrated at a UK military-supported treatment facility in Kerry Town, Sierra Leone [72]. Feasibility of transthoracic echocardiography was demonstrated at a military Ebola treatment facility in Conakry, Guinea [51]. By mid-December 2014, EMERGENCY, an Italian nongovernmental organization, established an Ebola critical care unit in Lakka and Goderich, Sierra Leone, the latter consisting of constant bedside nursing, continuous blood pressure, heart rate, respiratory rate monitoring, pulse oximetry, arterial and venous cannulation, nasogastric tube feeding, invasive ventilation, continuous renal replacement therapy, diagnostic biochemistry and hematology, ultrasonography, and plain radiography (Fig. 2). With waning case numbers, accurate evaluation of the impact of these interventions on patient outcomes has not been possible. Other sites, such as the GOAL-supported Mathaska Ebola Treatment Unit (ETU) and the Partners in Health-supported Maforki ETU in Sierra Leona also began using aspects of critical care procedures by February 2015, including nasogastric tube feeding, bedside ultrasound, as well as intraosseus cannulation for intravenous fluid resuscitation [73].

The historical philosophy of providing only oral fluids for EVD care has given way to the delivery of context-appropriate critical care [38, 42, 74‐79]. To date, 27 patients managed in nine countries outside of West Africa (Table 1) have been described, with a survival of 81.5 % [40] (Table 2). Thirteen detailed accounts of EVD management in modern healthcare settings in the USA, Germany, Spain, and Switzerland provide insight into the course of the illness [38, 42, 76‐82] (Table 3). These case reports confirm that intensive care monitoring in appropriately prepared centers is feasible. Noninvasive ventilation [38, 42], mechanical ventilation [38, 78, 81], central venous catheter insertion for vasopressor support [38, 42, 78, 79, 82], and renal replacement therapy [38, 78, 81, 82] can be provided effectively and safely (Table 3, Figs. 3 and 4) [83].

Current EVD treatment focuses on supportive care [70] as there are no specific treatment options yet to be proven effective [70, 90, 91]. A number of Ebola-specific treatment strategies have undergone preliminary clinical trial investigation, including convalescent plasma, Favipiravir, Brincidofovir and TMK-130803 [92‐97]. Transfusion of convalescent whole blood or plasma donated by EVD survivors has been used in this and prior EVD outbreaks [98] in an uncontrolled or compassionate-use basis [25, 79, 81, 99], and in animal models [100, 101]. One of three clinical trials of convalescent plasma therapy [94] has been completed and reported [102]. In this nonrandomized comparison to historical controls, transfusion of up to 500 ml convalescent plasma with unknown levels of neutralizing antibodies in 84 patients with confirmed EVD was not associated with a significant improvement in survival. While there were no serious adverse reactions in this trial, transfusion-related acute lung injury was described during convalescent plasma therapy in Spain [25]. Favipiravir (Toyama, Japan) [103], a pre-existing influenza virus inhibitor, has been administered for compassionate use outside West Africa [37, 38, 42]. In a multicenter, nonrandomized clinical trial in Guinea [104], 111 patients receiving Favipiravir had similar survival to that based upon historical control patients. The trial authors suggested that Favipiravir should be further studied in patients with medium to high viremia, but not in those with very high viremia [105]. Brincidofovir (Chimerix, USA), a nucleotide analog that inhibits RNA-polymerase with in vitro activity against Ebola [106], was administered to a small number of EVD patients for uncontrolled compassionate use [42, 79, 81, 99] and was tested in a phase 2 clinical trial in Liberia [95] that was stopped after the manufacturer withdrew study support [107]. TKM-130803 is a formulation of lipid-nanoparticle-encapsulated small interfering ribonucleic acids (siRNA) targeting two proteins involved in Ebola virus transcription and replication (Tekmira, USA, Canada). It was used in nonhuman primate Ebola virus infection as a postexposure treatment strategy [108] and in patients medically evacuated from West Africa in uncontrolled compassionate use [79, 81]. However, a phase 2 clinical trial (RAPIDE-TKM) in Sierra Leone [96] was halted according to pre-established stopping rules [109].

ZMapp, a monoclonal antibody cocktail (Leafbio, USA) [110], has been used under the emergency investigational new drug approvals from the Food and Drug Administration in patients treated in the USA, West Africa, and Western Europe [40, 76, 111]. ZMapp treatment of rhesus macaques resulted in 100 % survival even when started 5 days after lethal EBOV infection [110]. In the only randomized controlled trial of an investigational therapeutic for EVD, ZMapp plus standard of care was compared to standard of care alone for EVD patients in four countries, including the three most impacted West African countries. Due to the decline in EVD cases, this unblinded ZMapp randomized controlled trial only enrolled 72 of the prespecified target goal of 200 EVD patients; data were analyzed for 71 EVD patients, and mortality in the ZMapp treatment group was 22 % versus 37 % in the untreated group, but this difference was not statistically significant [112, 113].

The open-label, uncontrolled, and selected administration of other agents such as amiodarone [114], HMG-CoA reductase inhibitors, and angiotensin II receptor antagonists [115], and therapies to counteract vascular leak (FX06) [38] preclude any conclusions. In an observational study examining temporal trends in mortality among patients with EVD in one ETU in Guinea, 125 of 194 (64.4 %) patients receiving artemether–lumefantrine for malaria prophylaxis died as compared with 36 of 71 patients receiving artesunate–amodiaquine (50.7 %). In adjusted analyses, the risk ratio was 0.69 (95 % confidence interval, 0.54 to 0.89), with a stronger effect observed among patients without malaria [116]. These findings have not been confirmed in a randomized clinical trial.

Two vaccine candidates demonstrated efficacy in nonhuman primates [92, 117, 118]. A recombinant, replication-competent vesicular stomatitis virus-based vaccine expressing a surface glycoprotein of Zaire ebolavirus rVSVΔG-EBOV-GP (rVSV) [118, 119] was evaluated in an open-label, ring vaccination trial involving 7651 people in 90 clusters, randomized to immediate or delayed (21 days) administration. The vaccine was well tolerated and in the immediate vaccination group there were no new EVD cases while in the delayed vaccination group there were 16 EVD cases [120]. Another candidate vaccine, cAd3-EBOV (cAd3) [117] remains under investigation [92, 121]. Other vaccine candidates are also under development and evaluation [122, 123].

Several healthcare personnel received post-exposure prophylaxis with different interventions, including a candidate Ebola vaccine, following potential high-risk exposures to Ebola virus; although Ebola virus disease did not occur in these individuals, no conclusions can be made about the effectiveness of these uncontrolled interventions [124‐126].