Under current filoviral hemorrhagic fever outbreak operation protocols several activities are undertaken where accurate and rapid diagnostic testing can have significant impact:
- Suspected patients are to be admitted to the isolation ward and managed by watching clinical progression and response; discharging those who responded quickly to empiric treatment, and following the rest until they recover or succumb to disease.
- Contact tracing of suspected cases requiring daily monitoring of family members and close contacts by field teams dedicated to this essential outbreak control function.
- Unexplained deaths are routinely treated as possible cases and bodies are buried safely to limit further spread but denying the family their customary burial practices.
- Those fortunate to recover can be shunned by their family, if any remain, or the community at large due to the fear they can act as a new source of infection.
To obtain diagnostic testing, specimens have normally been shipped to an international reference laboratory such as the Institut de recherche pour le développement (IRD), Franceville, Gabon; NCID in Sandringham, South Africa; or the US-CDC in Atlanta, United States resulting in a significant delay (days to weeks due to shipment issues) in laboratory diagnosis with limited or no benefit for acute case patient or outbreak management 
. Therefore, such operation protocols require a fairly large infrastructure, longer hospitalization periods, and more staff and consequently increase resources and exposure risks. An MLU, providing testing results in a 4 hour turn around, can be an integral part of the outbreak response and simplify lessen many of the efforts needed to quickly contain and control the outbreak.
Laboratory testing of a symptomatic individuals during triage will allow the team to quickly assess if the person is a case or not. Confirmed cases can be appropriately isolated and supportive care initiated. Symptomatic individuals with negative test results can be maintained separate from confirmed cases either by releasing to another ward or kept in an observation ward for follow up testing or discharging. In Uige, and to a lesser extent also at previous outbreak locations, the isolation ward was largely unacceptable to the local population and significant resistance was present to have family members admitted 
. However a positive test result for MARV was normally sufficient to convince people of the necessity for admission to the ward. Isolating only those individuals who require it will reduce the infrastructure needed for isolation, minimize the hospitalization time for non-cases, reduce the number of staff and consequently reduce the risk of exposure for both staff and non-cases.
Cases that can be confirmed or excluded by laboratory testing can significantly contribute to one of the most important outbreak control measures, contact tracing. The current protocols call for the follow-up of contacts of suspected cases for 21 consecutive days. The presence of a field laboratory can help to arrive at a rapid confirmed final diagnosis for each suspected case, thereby decreasing the burden of field teams who may frequently be conducting contract tracing of cases with uncertain diagnosis.
Testing in this outbreak found that oral swabs from severely ill or deceased patients were a suitable sample for MARV testing by Q-RT-PCR. This allowed the MLU to safely test samples from corpses of unknown cause and when possible, to release MARV-negative bodies to the family members for traditional and religious burial procedures, a sensitive issue with almost all local communities in endemic areas. The value of swabs from corpses for diagnostic purposes needs to be further evaluated in future outbreaks and perhaps confirmed by other technologies such as immunohistochemistry 
. Post mortem RNA degradation might render a test falsely negative even so infectious Ebola virus has been detected in blood samples more than a month after blood draw and storage at room temperature 
. Any test results should take clinical presentation and epidemiology into account.
A growing concern is the return of negative and convalescent patients to the community, which may increase with the implementation of more advanced case patient care and the perspective of treatment options in the future 
. These people are often shunned by their families and neighbours and a timely negative test result as provided through the MLU may aid in their re-acceptance and safe re-introduction into the community.
In Angola, field diagnostic support was used for the first time in response to a MHF outbreak. Also the first time, the combined operation of a field and reference laboratory allowed for a unique evaluation of field diagnostic capacity under difficult circumstances and proved it to be accurate, efficient and safe in operation. There have been previous attempts to provide field laboratory diagnostics for outbreaks of Ebola hemorrhagic fever. In 1976 during the Zaire ebolavirus
outbreak an immunofluorescence assay was used for acute case identification but the results were considered poor 
. In 2000 during the Ebola outbreak (Sudan ebolavirus
) the US-CDC operated a laboratory within the Gulu district at St. Mary's Lacor Hospital, Uganda, and used antigen capture and reverse transcription nested PCR (RT-PCR) to successfully diagnose infection in suspected patients 
. In 2003 during the Ebola outbreak (Zaire ebolavirus
) in Mbomo, The Republic of the Congo, NML together with partners from the IRD, Franceville, Gabon, and the Bundeswehr Institute of Microbiology, Munich, Germany, operated a small field laboratory under the lead of WHO using antigen capture and Q-RT-PCR to diagnose acute cases 
In general, the usefulness of on-site laboratory support during filovirus outbreaks is not really questioned 
, and, in particular, the positive experience from this MHF outbreak demonstrate that rapid turn-around RT-PCR diagnostics can clearly aid in surveillance and case management 
. PCR-based techniques can be prone to contamination resulting in false positive results. Here we used a technique that did not require opening of tubes largely reducing the risk of contamination. Other concerns have been raised towards the reliability of RT-PCR assays during early disease stages and for survivors in the early convalescent stage, the consequences of false-positive and false-negative results of RT-PCR assays could be dire to outbreak management 
. Indeed, PCR-based assays, like other diagnostic tests, have weaknesses and do not produce reliable results under all circumstances. Therefore, independent, methodologically different, confirmatory assay such as antigen capture to support RT-PCR should be mandatory. However, nowadays most laboratories depend on PCR detection as their first and most rapid diagnostic methods and there are good reasons to support that choice 
. If a confirmatory assay is not available or unsuccessful, alternatives for RT-PCR confirmation include sample re-extraction, a second clinical specimen and/or an assay with independent targets (). Nevertheless, any diagnostics should not replace general and common sense precautions in case patient management and on-site laboratory diagnostics should be in close proximity to the ward allowing for continuous interaction between physicians/nurses and laboratory personnel 
. Importantly, during this field laboratory deployment, Q-RT-PCR proved to be very sensitive and reliable even in this challenging environment. Patient samples were positive in our testing beginning on the day of onset of symptoms but we did see that detection in swab samples could be delayed by a few hours when compared to blood this early in the course of illness.
The collection of appropriate clinical specimens for diagnostic testing has become an increasing problem during filovirus outbreaks. The reasons for this can include the lack of properly trained personnel, fear of personnel to apply invasive procedures, cultural objections to bleeding and any other invasive pre- and post mortem sampling procedure, and insufficient infrastructure for sampling and transportation 
. In that respect, the MHF outbreak in Angola was not different from previous outbreaks. In particular, resistance in the community to bleeding and post mortem invasive procedures, such as cardiac puncture or liver biopsy, and the increasing resistance of aid personnel to apply invasive procedures in the field (community) made oral swabs the predominant clinical specimen available for testing. As demonstrated here on paired blood/oral swab samples, in general there was no significant difference in viral load between oral swabs and EDTA blood taken at the same time (). This supported oral swabs as an alternative diagnostic specimen to blood. The few incidences when oral swabs were less suitable than EDTA blood related to early disease stage and early convalescent stage samples. Lower viral loads in oral swabs compared to EDTA blood, at these stages, are likely to explain this discrepancy. Additionally, there are inherent sampling variables associated with oral swabs (the technique and efficiency of swabbing; moisture level of the oral cavity) that are not present in a blood draw, which may also have a role in these differences. However, despite the fact that oral swabs seemed to have been an appropriate specimen source for laboratory testing during this outbreak, and oral/nasal swabs are valuable alternatives in cases of resistance in the affected population to invasive procedures, EDTA blood should remain the priority choice for a clinical specimen due to the longer period of detectable viremia, the suitability to serological-based testing, and the value for monitoring potential point of care therapies in future.
While this study is not a detailed epidemiologic study, brief mention of some of the data is warranted as it has not been yet published elsewhere. This MHF outbreak was unique in regards to its location, case number and case fatalities, but also showed a large proportion of paediatric cases and cases among woman in the child bearing ages 
. Since MARV, as Ebola virus, are usually transmitted through close contact with blood, secretions or excretions from infected patients, family members and medical personnel caring for patients or preparing bodies for burials are considered high risk exposure groups 
. It has been proposed that because women provide the majority of in-home care that this was the reason for the preponderance of cases in women 
. Certainly women provide the majority of care for the children and since, especially early in the outbreak, children less than 5 years of age represented the largest single age group affected may also be reflective of this fact. Furthermore, the detection of MARV in breast milk during this outbreak indicates that breastfeeding might have played a role in virus transmission. This is supported by epidemiological data indicating transmission from infected mothers to their nursing babies followed, after death of the mothers, by virus transmission from the infected babies to wet nurses who subsequently infected their own nursing child (authors, unpublished observation). Other factors may have come into play including the alleged lack of appropriate infection control within the paediatric ward prior to the identification of the outbreak 
. It is very unlikely that the predilection of women and young children represents a biological predisposition, given that the demographics of the outbreak changed through the course of the outbreak (i.e. early in the outbreak a very high percentage were paediatric cases whereas later cases became more evenly distributed by age), and yet the virus changed very little 
. Without more detailed epidemiologic data, it remains unclear which of these transmission routes constituted significant mechanisms for virus spread in the Uige outbreak.
Offering differential diagnosis significantly increases the value of on-site diagnostics. This is much harder to achieve in the field and requires variable clinical specimen (in particular blood or stool), more manpower and more extensive and continuous supplies. At a minimum, malaria diagnostics (e.g. commercially available rapid dipstick tests) and diagnosis for severe gastrointestinal infections should be available. Proper case patient management including intravenous fluid administration would also require blood chemistry and haematology analysis, another capacity that needs to be considered for expansion of a field laboratory response capacity.
Most of what constitutes the MLU can be sourced from equipment that most reference laboratories would have access to from their normal compliment of equipment and supplies, however a dedicated MLU would likely require the investment of approximately $100 000 and a weekly deployment cost of $2000 for reagents and supplies. Logistic needs and costs during a mission can be best managed through a close working relationship with other organizations including the WHO and Médecins Sans Frontières (MSF). The greatest challenge to the operation of the MLU was the lack of consistent electrical power and our reliance on portable generators. This necessitated the use of battery backup systems for thermocyclers and did not allow for storage of samples or reagents at freezing temperatures as freeze-thaw cycles could not be avoided. Fortunately, all reagents were relatively stable at 4°C over a three week rotation period before replacement teams replenished the reagents. We were able to efficiently operate the MLU using teams of two members as the workload and workflow rarely justified additional staff. We have since recommended that teams of three be deployed to allow for rest and health issues.
In conclusion, the combined operation of a field and reference laboratory in this outbreak allowed for a unique evaluation of field diagnostic capacity under difficult circumstances. Rapid MARV-specific Q-RT-PCR was useful for triage and assessing the need for isolation. The quick turn-around of laboratory diagnosis on the basis of Q-RT-PCR assays significantly improved outbreak response efforts. Therefore we propose: “On-site laboratory diagnosis should become a routine part of any future filovirus outbreak response as it provides all responders with valuable information to help minimize the extent and durations of these events”.