Testing Times

The COVID-19 pandemic is showing no sign of letting up. As countries consider relaxing their various states of lockdown, secondary outbreaks represent a real (and, in some cases, realized) threat to those attempting to return to normality. The only way we can follow the extent of the issue is by implementing effective testing regimes. And the only way to beat it? A treatment or vaccine – both of which remain elusive.

In the world’s mission to identify the virus and treat those infected, analytical scientists are a secret weapon – turning samples into answers, exploring new targets, and optimizing drug discovery efforts.

We spoke to Peter Kissinger and Rolf Halden, two scientists involved in these crucial activities, to explore the challenges we face – and what to expect in the months (and years) ahead…
 

Tracing Individual Exposure

Bioanalytical guru Peter Kissinger provides an overview of the current status and future prospects of clinical COVID-19 testing

How are we testing for SARS-CoV-2?

Reverse-transcription PCR (RT-PCR) based on viral RNA sequence is the key qualitative tool for SARS-CoV-2 detection. These tests act primarily as a tool to prevent further transmission, but , being new as they are, many are yet to be validated with statistical rigor. The current situation, however, allowed for Emergency Use Authorization (EUA), meaning we have taken a shortcut, and statistical rigor will be accounted for later in time. But this must not be ignored for long.

A second approach is to detect a viral surface antigen via an antibody. Antigen tests could be conducted more simply and faster than typical PCR methods, but more data are needed to fully understand their utility. A third method then promises viral detection via CRISPR technology – developers say this can be accomplished quickly and at low cost, with the potential for home use based on nasal swabs or saliva.

Serological antibody measurement can indicate ongoing or previous infection. These can be applied in homes and instrument-free public health settings, and one from Roche was recently approved for use in the UK, but we are still not sure if reinfection remains possible despite the presence of antibodies. If not, for how long is this the case? This would be useful to know, especially given the potential for the virus to alter and overcome immune defences over time, but – from my reading at least – we have no definitive answer yet.

Sampling is also very important. Should we use a nasopharyngeal swab, a nasal swab, saliva, sputum, or blood? The first of these is closer to the action for a respiratory disease, but swabs are not without some “art” and luck. This topic has been a source of worry, and has added cost due to the need to employ experienced and (PPE) protected individuals. Home collection has recently begun, saving costly labour and travel expenses, but adding the risk of a poor collection.

The direction that disease-identifying tests are following is excellent – we just need to ensure they’re accurate – and fast. Movement towards testing for specific biomarkers to track infection and treatment response will also be essential for us to account for the wide range of ways in which patients respond to the virus, and improved cytokine measurements will help us understand these responses. Labs were overwhelmed early in the outbreak due to backlogs and the lack of supplies, but we are now seeing this situation begin to stabilize.
 

Who conducts the tests?

Analysts! The American Association for Clinical Chemistry (AACC) worries that the profession is generally undervalued. This is reflected in the fact that clinical measurements make up two percent of healthcare costs, while driving over 70 percent of clinical decisions. Pharma has noted for decades that a modest 10 to 12 percent of healthcare is spent on drugs – so it’s strange that those working in the field are viewed as villains in wider society. This latter figure has gone up recently – most likely due to the rise of protein therapies. Enough politics…

Hospital laboratories are relatively hidden, as are commercial reference laboratories. They are appropriately certified, and personnel are licensed. Samples go in, answers come out. But there are other testing environments, such as a nursing station, physician’s office, pharmacy, or even at home. In all cases, the test must be prescribed. It would be very attractive to have a reliable OTC test that would reduce cost and improve speed across these settings, perhaps via a finger prick test of the blood or saliva.


What is challenging about COVID-19 testing?

Specificity with regards to both the virus and associated disease is tough. The timing of disease progression is extremely important in both cases, and so a key question is “how often should a test be performed?” PCR tests can detect the virus prior to symptoms, while serological tests cannot, but both tests have utility in distinguishing COVID-19 from other febrile diseases as symptoms appear.

There is also much we are yet to learn. Given that no proven treatment has emerged, either in early or late disease states, we face the challenge of applying the right test at the right time. Antivirals effective against HIV and Ebola are being explored, as are anti-inflammatories for malaria and arthritis. Remdesivir from Gilead has gotten the most ink thus far, but it’s too early to say how useful it will be. Regeneron, Lily, Pfizer and Merck are also contributing to a great community effort. But none of the options put forward cure or prevent the disease – they show promise for reducing symptoms.

Much depends on the quality and consistency of reagents, some of which have been in short supply. Several recently announced tests appear to demonstrate good specificity, but some earlier tests were highly deficient, and gave a high percentage of false results. Reagents vary in type and source, and can also suffer differences from time to time. What’s more, throwaway tests cannot be individually calibrated, and quality control can only be achieved by sampling from a manufactured batch, or periodically on a continuous production line. In other words, expect some wrong answers.


What are the limitations?

Every analytical measurement will suffer constraints at the detection limit, and that will vary with the “handle,” or the label established to produce a signal. Often this is some form of fluorescence or chemiluminescence, but could alternatively utilize electrochemistry and other means of transducing a signal. And, below a certain degree of infection, a signal will not be detected. The higher we set the threshold for a positive conclusion, the more false negatives we will have and vice versa. Using instrument-free methods, the human eye becomes the detector.

We also face issues regarding selectivity and cross reactivity. Is the sequence we are targeting with RT-PCR unique to this virus? Could there be some cross reactivity with other coronaviruses? And has any of this been established? The answers to these questions, at present, are maybe, possibly, and not quite yet, respectively.

In fact, a report from South Korea described an interesting ambiguity regarding the PCR tests. Some patients were giving positive results after recovering from symptoms. Apparently, they were shedding non-infectious or dead virus particles and were not infectious. It is thus possible to have a positive test for the virus and not be required to quarantine; the same recovered patients also had a positive serological test. This nuance also reduces the worry that fully recovered patients could continue to spread the virus.


Will these tests bring us closer to helping patients?

Yes, but it will be a long journey. As analytical chemists, we are often detached from the context of our samples, and analyze them weeks after collection and storage at -80 ^oC. Clearly, an ICU patient on mechanical ventilation will not benefit from these results.

Regarding treatment, each patient reacts to a drug in very different ways, and we are terrible at optimizing dosage for most drugs. This is, in part, due to “the tyranny of averages” from clinical trials. We also cannot determine drug concentrations repeatedly outside of clinical trials. We have the tools to do this, but the logistics are not currently available to support this – not even in an ICU. Our group is working on it!

Movement towards a treatment that fights the virus safely will require lots of data – acquired faster, better, and cheaper than ever before. Naturally, there are also many challenges integral to interpreting data that span the diversity of our 8 billion-strong populous. But you can’t change something you can’t measure, and the opportunity for analytical chemists to participate here is so much more than a supportive role – it is central.


How will analysts support the ongoing fight against COVID-19?

We can determine just about any chemical substance at any meaningful concentration given enough time and budget. But speed matters, as does context. Every stage of our fight against this pandemic, from testing to drug development (from discovery through manufacturing), relies on analytical chemistry.

Every hospital is an analytical laboratory, too. Consider a critical care nurse at 3 AM with a patient on mechanical ventilation suffering a cytokine storm. There will be a lot going on. A bunch of drugs will have already been administered. Will they interact? Will the ventilator settings be optimum? What is the right dose to be administered by infusion pump? A point-of-care device could potentially provide reliable, guiding data to help care for critically ill patients. Medicine is non-linear and intuitive, and guidance systems are always needed. Experience also guides, just as it does for a good Ferrari mechanic.


Peter Kissinger is a Professor at Brown Laboratory of Chemistry, Purdue University, and a founder of Bioanalytical Systems (BASi), Prosolia, and Phlebotics


​​​​​​​

City-Wide Monitoring (via Sewage)

Rolf Halden and colleagues had just begun tracking viruses in US wastewater when the pandemic broke out. We caught up with Rolf to find out how his team is extracting population-level SARS-CoV-2 exposure information from our sewers – and how this can help us fight the virus’ spread locally and globally

How did you start measuring coronavirus in wastewater?

We have been continuously monitoring hundreds of chemicals and biological markers in wastewater from around the world for years – I discussed this in “The Great Sewage Census” in The Analytical Scientist in 2019. It was in the same year that we began applying our method to viral quantification. I was fortunate to find an opportunity to collaborate with Arvind Varsani (a virologist specializing in virus discovery) and Matthew Scotch (a bioinformatician focusing on zoonotic RNA viruses), and together we discovered over 3,000 viruses – mostly by studying US wastewater (1) (2) (3).

We approached the National Institute of Health with our method, presenting it as a convenient “atlas” of viruses across the US – and as a potential early warning system to monitor infectious disease. They liked it so much that they offered us a grant of $1.5 million, and we were ready to begin real-world viral tracking by October. Two months later, the COVID-19 pandemic began, and we quickly pivoted to track population-wide exposure to the SARS-CoV-2 virus across the US by applying qPCR and metagenomic approaches to wastewater and clinical samples.


How have people responded to your work?

People have always approached our work with a degree of scepticism. When we started our viral measurements, people mocked us – they said we were wasting our time. But now it turns out that everybody is jumping on the bandwagon. Hundreds of people across the US and further afield are now playing catch-up to conduct this type of viral monitoring themselves… And that’s great! Every dollar we invest in these early warning systems helps to prevent human suffering and associated costs downstream.


How can your approach help?

The beauty of public health engineering is that we can anticipate and detect a threat upon arrival, before lots of people are exposed and made sick. But we have to be realistic. Many people label new breakthroughs in health as “silver bullets” – people thought that sequencing the human genome would allow us to cure cancer, for example, and subsequently that proteomics and then epigenetics would provide us with all the answers required to solve pressing medical problems. I want to stress that our approach is useful but represents only one of many tools in our toolbox. We need to deploy multiple tools simultaneously in order to succeed. But wastewater analytics has immense utility in expanding our current view of viral exposure across populations, and it can do so at a very low cost.

Testing individuals is demanding, both in terms of labor and cost, and it’s simply not viable to repeat these measurements as often as they are needed. Our approach can test wastewater from 2 million people over a 24-hour period for the same cost as screening one single individual. If we were to divert only one percent of the money currently spent on individual testing toward wastewater analysis, we could rapidly screen approximately 70 percent of the US population once or twice per week.

This is the main advantage we are able to provide – producing a “radar” that detects coronavirus occurrence for large populations at negligible cost. In fact, 2.1 billion people are connected to approximately 105,000 centralized wastewater treatment plants worldwide (4). Imagine the data we could extract from this sample, equivalent to around a quarter of the global population (1) (4).


Could this information help handle the spread of COVID-19?

It definitely has the potential to help contain the pandemic. Our current method can identify COVID-19 hotspots through area-specific signal intensities, and we are currently achieving this on a three-times-a-week basis in Tempe, Arizona, across different parts of the city. These data are publicly available on the world’s first public dashboard, displaying information obtained by wastewater analytics without delay at https://covid19.tempe.gov – anyone can take a look!

This tracking has helped us to study the epidemic as it unfolded here in Tempe. During the first wave, uncontrolled community spread was rampant because people weren’t aware that they could spread the virus while being asymptomatic and we saw it tear through communities across the United States. Then came the stay-at-home ordinance from the governor’s office. Locally, we entered into a shutdown period and wastewater coronavirus concentrations soon fell to below the detection limit across all five monitored areas of Tempe. On 15^th May the economy reopened – so the question now is: “at what point do we see the virus emerge again from below-the-detection limit?”

We will need to keep a watchful eye on the virus as it becomes detectable once more and monitor the delay between its detection in community wastewater and people becoming seriously sick, being hospitalized, and dying. Having this information on the relationship between changing viral levels in community wastewater and individual and healthcare events will allow us to monitor the effectiveness of social distancing measures and lockdowns, and may even inform when these measures should be put in place or lifted in order to stagger the pandemic’s effects.

Of course, we must also consider the economy when making these decisions; America has a GDP of around $21 trillion, and shutting down the entire country may cost as much as $1 trillion per month. It could be advantageous to manage economic activity and virus spread locally rather than state-wide or nationwide, using data from both the healthcare and sewer system. This should allow us to keep the nation running at a minimal capacity of at least 70 percent at all times, with rolling shutdowns implemented only where they are truly needed and beneficial.


What’s the current status of your work on COVID-19?

We’re currently collecting data from over 100 cities across the US, and also from South and Central America, so our team is working day and night to deal with the huge sample throughput.

We have cloned the Tempe dashboard that we originally developed for tracking the opioid epidemic (https://arcg.is/ey0Ha) to give citizens access to information on coronavirus levels in Tempe and its neighbouring cities of Guadalupe and Gilbert (https://covid19.tempe.gov). And we’ve received requests from hundreds more locations across the world.

At this point, I would no longer consider what we’re doing to be exclusively research – we have fast-forwarded into the realm of routine monitoring. In addition to our research lab operations, paid for by tax dollars, we’ve also formed two additional entities to address the need for public health information. One is a non-profit called OneWaterOneHealth (https://onewateronehealth.org) that seeks to help underserved communities, and the other is Aquavitas LLC – a for-profit organization that provides not only measurements of the virus, but also helps in interpreting what those measurements mean (https://aquavitas.com/). The phone never stops ringing and we’re incredibly busy, but it’s very exciting and rewarding to work on a pressing public health issue!


Are there any limitations?

Yes – there are currently a lot of unknowns surrounding our measurements. We are able to measure the viral load passing through a wastewater treatment plant, but what we don’t know is exactly what this measurement means. How does our measurement correlate with the number of people who have been exposed to the virus? To what extent does each individual shed the virus in feces, and when does this occur over the course of infection? And how are the people shedding viral particles distributed throughout the region that the wastewater corresponds to? We have run simulations to gain a better insight into the potential answers to these important questions (4) and we will be able to gain a more accurate picture from our samples as the science and monitoring efforts advance.


What are the challenges in scaling up this approach?

Differences between regions, for one. For example, having 50,000 copies of viral RNA in a given amount of wastewater from the US would have a very different meaning to the same finding in wastewater from Europe, where they use approximately a third of the amount of water that we do each day. Clearly, we are dealing with different dilution factors. The overall chemistry of the samples also differs with the types and concentrations of detergents used in each region, and the presence of more industry in Europe versus the US may also have an impact, as do wastewater temperature and the average travel time of wastewater in municipal sewers.

If we were to use point-of-care tests for individuals only, it would take us a minimum of several months to test all American citizens. I equate this to of painting the Eiffel Tower – once you’re finished at the top, it’s time to start all over again at the bottom. But we need to test most Americans on a weekly basis to truly track the pandemic – wastewater analytics makes this an attainable and affordable goal. Associated numbers show this is feasible (4), and the approach thus promises to help prevent unnecessary human suffering, while also saving billions of dollars in individual screening costs. We said it before, both types of analyses are needed. The challenge right now is to find and implement the perfect balance between community wastewater screening to identify flare-ups and follow-up testing of individuals to isolate cases. Getting this right is a matter of life and death for both people and local economies. That’s what we are working towards right now.


Rolf Halden is Founding Director of the Biodesign Center for Environmental Health Engineering at the Biodesign Institute, Professor in the Ira A. Fulton School for Sustainable Engineering and the Built Environment, Senior Sustainability Scientist in the Global Institute of Sustainability, Arizona State University, Tempe, USA, and co-founder of the non-profit project OneWaterOneHealth and the ASU start-up company AquaVitas.