4

MOLECULAR DIAGNOSTICS: AN ALTERNATIVE TO HIGH CONTAINMENT?

Charles Chiu, MD, PhD, a member of the workshop planning committee, gave a presentation titled “Molecular Diagnostics in Low-Resource Settings.” He described the “classical” (i.e., non-molecular) microbiological testing methods as including:

• Culture
• Serology
• Microscopy
• Biochemical profiling
• Direct antigen testing: Lateral flow immunoassays and matrix assisted laser desorption/ionization (MALDI) for bacterial, viral, and fungal identification

Molecular diagnostics, or “DNA-Based detection,” include a variety of new, and even experimental, technologies, such as:

• Hybridization (probes), for example, clustered regularly interspersed short palindromic repeats (CRISPR)-Cas based assays
• Genotyping
• Sequencing, including nanopore sequencing
• Signal amplification
• Target amplification (polymerase chain reaction [PCR]): Singleplex and multiplex

Molecular diagnostic tests offer some advantages for low-resource settings. The pathogens are inactivated for testing, so handling them is safer than in methods that require the use of infectious live organisms, decreasing the potential for occupational exposures. They do not rely on culture-based amplification, which is important because many pathogens

are not culturable. Because such molecular-based testing enables performance of diagnostics with noninfectious inactivated pathogens, the need for costly and complex BSL-3 or -4 containment is obviated for diagnostic work on very hazardous pathogens. Molecular methods also offer faster turnaround time and do not require large sample volumes.

However, Dr. Chiu continued, molecular testing also has significant disadvantages. First, these methods are more expensive than classical techniques. One participant noted, for example, that the cost of one PCR kit equals about 1 year’s salary for a lab worker in low-resource settings. Second, performance assessment, validation, and regulatory approval of many of these methods are challenging, especially if the work is performed outside of highly controlled clinical laboratory environments, which may not be available in low-resource settings. Dr. Chiu reviewed some areas in which standards for molecular testing are lacking, particularly for environments that are not highly regulated: positive and negative controls, platforms, analytical performance, target pathogens, and reference databases. Because of this lack of standardization, the same assay run in two different labs may yield different results, and confirmatory testing is slow and costly.

In addition, the entities that normally certify and/or approve such tests (e.g., the U.S. Food and Drug Administration [FDA], the U.S. National Institute of Standards and Technology, the World Health Organization (WHO), and various nongovernmental organizations) have not yet done so for most of the new molecular technologies. In the United States, a regulatory framework, the Clinical Laboratory Improvement Amendments (CLIA), guides clinical laboratory testing. It sets minimum standards under which all clinical laboratories operate. CLIA laboratories are certified by inspection by an agency such as the College of American Pathology. Compliance with CLIA requires validation and quality assurance for all laboratory tests used in clinical care, including “laboratory-developed tests.” The Clinical and Laboratory Standards Institute (CLSI) also issues “Guidelines—CLSI Molecular Diagnostic Methods for Infectious Diseases” (CLSI, 2015). But certification under these frameworks is not the same as FDA approval. Furthermore, many laboratories in low-resource settings may not meet CLIA or similar regulatory standards for proficiency testing, incorporation of standardized controls, etc.

Pre-analytical, analytical, and post-analytical concerns exist for molecular diagnostics, Dr. Chiu explained. Pre-analytical concerns include the need for proper sample collection methods, appropriate timing, proper storage conditions for both organisms and assay components (e.g.,

maintenance of a cold chain, which is especially important in low-resource settings and with labile ribonucleic acid [RNA]), and control of contamination. Analytical concerns focus on test performance evaluation—sensitivity, specificity, precision, accuracy, linearity, matrix effect, interference, reproducibility, and limitations. Post-analytical concerns include proper reporting of results, copies/ml or IU/ml or Log IU/ml, positive and negative predictive values (PPV and NPV, respectively), and diagnostic value and clinical utility.

Based on these and other considerations, Dr. Chiu provided a list of relevant questions that should be addressed in the context of molecular testing in low-resource settings:

• Who will pay for a test, and who is trained and certified to run it?
• How often will the test be run? What volumes of material are required? Do the particular circumstances justify the costs?
• Can clinically significant organisms be identified and quantitated in patient specimens or from culture?
• If culture is not possible, molecular methods may be justified. But if they are to be used for prognosis, surveillance to guide public health interventions, or diagnosis to guide therapy for individual patients, will they provide the necessary accurate information?
• Where will a test be run and in what settings? Are these settings appropriate for achieving accurate results?

Dr. Chiu then briefly described each technology category that he listed at the start of his presentation and commented on their states of development (see Box 4.1 at the end of this chapter).

Dr. Chiu explained that direct detection methods, such as sequencing, cannot fully replace serology, the branch of laboratory medicine that investigates blood serum to detect antibodies and antigens.¹ He views molecular testing as complementary to, but not a replacement for, classical testing methods. He also described the stage of development and use for each of the molecular technologies:

• PCR is in place, but remains challenging because of lack of standardization.
• Next-generation sequencing is still limited. Although not yet FDA approved, some nanopore sequencing is being used in the field.
• CRISPR-Cas is very promising but remains in the research phase.

___________________

¹ See Medical Dictionary, https://medical-dictionary.thefreedictionary.com/serology.

• MALDI is generally too expensive for low-resource settings.
• Multiplex PCR is available but is also expensive.
• Host response-based assays are likely to evolve rapidly in the future.
• Metagenomic sequencing is promising but also expensive and not yet widely available.

Dr. Chiu offered the following takeaway messages:

1. A combination of traditional methods (e.g., immunofluorescent strips, real-time PCR) and state-of-the-art approaches (e.g., nanopore sequencing, CRISPR-Cas assays, multiplexed PCR) will likely be needed moving forward.
2. Cost and other practical considerations favor true point-of-care molecular diagnostics (e.g., lateral flow immunoassays, CRISPR-Cas).
3. It will be important to decide whether the focus should be on diagnostic testing or surveillance. Who (in loco, in country, international) should be doing what? Emerging infectious diseases do not respect borders.
4. Sequencing has made the greatest impact in genomic surveillance, but not yet in molecular diagnostics.
5. MALDI, multiplexed PCR platforms (e.g., BioFire, Luminex), and even single-plex PCR instruments (e.g., Cepheid GeneXPert) remain too expensive for use in diagnosis, but may be acceptable for targeted surveillance, such as during outbreaks.
6. Inexpensive, field-ready multiplexed diagnostics are urgently needed but do not exist.
7. Direct detection approaches likely will not replace serology (e.g., lateral flow assays) anytime soon.
8. Complex data from genomic sequencing and other methods will require cloud computing resources to disseminate results quickly, which is critical in public health scenarios.

Dr. Chiu ended his presentation by stating that the effectiveness and accuracy of many of the molecular technologies must be demonstrated before they become widely usable in low-resource settings. Although some testing is occurring in low-resource settings, much work remains to be done.

During the discussion that followed, one participant said that his group is working to develop non-probe PCR techniques, trying to use multiplex immunoassays for serology, and providing Sanger sequencing using

remote analysis, where needed, as a backup to PCR field applications. He noted that the real questions are how to deliver the new tools to the field and train people in all these new skills, especially data handling, storage, and security. Some of the tools now available require no maintenance and have disposable cartridges. One approach for data is to use cloud-based bioinformatics to analyze data and return results so that no local bioinformatics talent is needed for this purpose, provided internet connectivity exists. However, another participant stated that communications are a real problem in the field because bandwidth is insufficient. Therefore, cloud approaches may not work in an outbreak situation.

A participant asked whether a case can be made for aiming for reagent self-sufficiency. Dr. Chiu replied that there is because the reagent market is not very competitive and competition would likely drive down costs. The same participant said that a cost-benefit analysis, which does not exist but is needed, could help donors to decide what type of support they should provide.

With Dr. Chiu’s summary of the state of the art and his conclusions about technology readiness, the participants were ready to examine the practicalities of field deployment and use. Jonathan Towner, PhD, of the U.S. Centers for Disease Control and Prevention (CDC) gave a presentation on his field experiences during several hemorrhagic viral outbreaks, including the West African Ebola outbreak in 2014-2015. Dr. Towner’s experiences illustrate the application of available techniques to real-world responses to major disease outbreaks. His first field experience was in Uganda in 2000-2001, where CDC used both ELISA-based and PCR-based testing. The work was performed in a hospital lab, and more than 1,000 samples were processed over a 3-month period, with more than 280 testing positive for the virus. In 2005, he participated in the response to an outbreak of Marburg virus in Angola. ELISA and PCR were again used, and this time the work was performed in an existing lab that was established for HIV diagnostics. This time 180 of 505 samples from blood or serum, breast milk, or swabs were found positive for the virus. From 2010 to 2016, Dr. Towner participated in a program of enhanced viral hemorrhagic fever (VHF) surveillance and diagnostics in Uganda to provide training for the local medical and other staff. This program placed CDC personnel in-country and helped to achieve a greatly reduced number of later VHF cases as well as reducing the time to diagnoses.

Then in 2014, the huge Ebola outbreak in West Africa occurred, seriously affecting Guinea, Sierra Leone, and Liberia. This event attracted, in Dr. Towner’s words, a “United Nations” of field response with

Germany, France, Italy, Belgium, the Netherlands, England, Canada, the United States, Nigeria, South Africa, China, and Russia sending people, equipment, and other aid in an impressive response and collaboration effort. The U.S. agencies included CDC, the National Institutes of Health, and the Department of Defense. The aim was to provide rapid diagnostics in the field. In all, 27 field laboratories were set up across West Africa.

The scale of the response to the West African Ebola outbreak generated its own challenges. For example, many different real-time PCR assays were used in the large network of laboratories, creating a need for quality panels and attempts to standardize assays. For the panels distributed by CDC in Sierra Leone and Guinea, 2 of 6 laboratories were producing incorrect results at a rate of 10 percent, which required implementation of improvements. The need for a two-target Ebola assay, plus cell RNA PCR controls, emerged to reduce the rate of false positives or negatives when only one target was used. The lack of a cell RNA control also increased the risk of false-negative results.

There were also database, documentation, and reporting challenges. It was difficult to complete sample submission forms, so a considerable number of samples had little or no documentation. There was an absence of unique identifiers for samples and cases and difficulties with linking lab, clinical, and epidemiological data. Information on date of onset was often missing, as was knowledge of whether swabs were from corpses (appropriate) or live patients (inappropriate). Finally, there were problems with turnaround time for results, insufficient numbers of trained phlebotomists, and transport of samples.

Dr. Towner concluded that much was achieved, despite these problems. Peak testing occurred during the October-December 2014 period, with the highest number of samples tested at 180 per day in July 2015. On average, 71 percent of samples were tested on the day they arrived at the lab, and 99.9 percent were tested either the same or the next day. Samples were received for about 14 months from 12 of 14 districts and were mainly whole blood and cadaver oral swabs. Overall, more than 27,000 samples were tested. Dr. Towner’s lab remained operational for 406 days, with no days off or disruptions in testing. A pilot study testing for viral persistence in male survivors began on May 23 and resulted in the testing of more than 500 semen samples. The Sierra Leone vaccine trial, or STRIVE, began on May 24, and 51 samples from 30 participants were tested. Twenty-eight teams of personnel from 17 different branches throughout CDC were trained in Atlanta on the Bo lab protocols and procedures and then deployed to help keep the lab operational.

After the crisis phase of the epidemic was over, Dr. Aiah Lebbie of Njala University in Sierra Leone was selected as the recipient of a 3-year CDC cooperative agreement to conduct ecological VHF surveillance on the region’s bats. The University’s laboratory facilities underwent major lab renovations from March through August 2017. There are now stable and properly maintained electricity, freezers, and other working equipment. Purchasing is operational, although delivery of perishables remains an issue. This arrangement with the University will provide educational as well as public health benefits for the region and exemplifies what partnering can accomplish. Approximately 5,000 bat specimens have been collected, all from forested areas. There are two field stations, one on Tiwai Island in Sierra Leone, which is starting renovations, and another in Gola Rainforest National Park in Liberia.

Following Dr. Towner’s presentation, one participant noted that the new molecular technologies reduce risk, so minimal containment levels are needed if pathogens such as Ebola and Marburg are indigenous to a particular locality. In such a case, BSL-3 and -4 laboratories are probably not needed, but some recipients seek high-containment facilities for prestige rather than real needs. Another participant restated the need to distinguish between labs conducting surveillance and those conducting diagnostics, and noted that reference labs are necessary for identification of strains and for research. Perhaps low containment is adequate for the field while higher containment is required for reference labs.

Although Dr. Chiu said that the costs of the molecular technologies need to be reduced before they can be used in low-resource settings, one participant stated that acceptable cost may depend on the disease, the number of affected patients, the costs of care and treatment, and other factors. This translates into common diseases needing cheaper analytical capabilities. Some technology is already spreading. For example, 165 “Gene Expert” machines have already been deployed throughout the Democratic Republic of Congo, although, as pointed out by one participant, their throughput is low. In addition, these machines only detect the Ebola Zaire strain, which, if it mutates, might be undetected like other strains. Another participant noted that funders should account for already-deployed capabilities when making support decisions.

A participant highlighted the different needs for normal operating situations vs. responses to epidemics. His organization uses Luminex testing, but what is appropriate depends on whether the researcher is looking for one specific pathogen or more. Another participant stated that anything new that is built should be linked to existing facilities that are

already part of a network so that reagents and other resources can be shared.

Dr. Chiu stated that direct detection of the organism of interest is the “gold standard.” Clinicians are more conservative than researchers, so it might be 5 to 10 years before new test types are accepted as the basis for patient treatment. Because nasty incidents with drug and vaccine testing have already occurred in developing countries, there is a particular need to be conservative with molecular diagnostics as new tools for guiding medical treatment on these grounds as well.

A participant reiterated that the cost of reagents is an issue, especially because of the small budgets of low-resource countries. Another participant pointed out that the technologies themselves are very costly and donors should be advised to rely on proven technologies as a baseline. In light of cost factors, another participant said that it makes sense to use distributed sets of labs for basic testing and to reserve the expensive, high-throughput capabilities for central locations. Yet another participant noted that emergencies are special, but emergency response will improve if more laboratories are equipped to deal with normal business. Having a lab in place for normal operation also keeps personnel and supply chains trained and practiced. Donors should also inquire about quality management systems. In developing countries, reference labs and other resources may not be available to facilitate standardization, verification, and other necessary steps.

A participant said that the “Wild West” can result from a lack of regulatory oversight—even in the United States, but more so in some developing countries. Another participant noted, however, that France provides oversight in Francophone countries. Another participant stated that the lack of reagents and of equipment maintenance and repair is a donor’s nightmare and that training for these latter functions is greatly needed. However, another participant noted that the education levels in some localities are so low that the concept of maintenance does not exist and therefore providing such training for local personnel is very difficult. In addition, transportation poses a barrier to building reference capabilities. Partnerships are needed to provide fuel, dry ice, and other laboratory staples. Multinational corporations, such as Coca Cola, and oil companies, for example, have provided some of these supplies. The African Union, the Economic Community of West African States, and the South African Development Community have developed ways to address some of these problems,

A participant suggested that biobanks be sited in secure locations, such as military bases. Another participant noted that biobanks secure and

recognize the value of live samples, but may not have plans for what to do with them, raising questions about whether they should have collections without clearly defined purposes.

Participants called for a broader vision to ensure preparation for the next outbreak. They also posed the question of how to interest donors in enhancing what already exists instead of building new facilities. Finally, they wondered how donors could help with “leave behind” facilities after an outbreak, as was done in Sierra Leone after the 2014 Ebola outbreak.

This page intentionally left blank.