In medieval times a Friar was walking, deep in thought, along the bank of the River Thames. Suddenly he became aware of two knights arguing loudly. They drew their swords and were about to fight to the death. The Friar intervened. “Pray, gentlemen, what are you arguing about?” The knights lowered their swords and turned towards him. “It is well known that this is the shortest crossing point of the river, but this disreputable man believes that a shorter route lies around that bend”, one of them explained, pointing along the river.

The Friar realized that he could intervene and prevent one of the knights dying needlessly. He said “But gentlemen, you could settle this dispute by simply measuring the two distances, to test which of you is correct.” He felt quite superior about this undeniable logic and tried not to look too smug about it. He did not expect what happened next: the knights turned their swords towards him, felled him to the ground, and resumed their argument.

Tests are the unsung heroes of science, engineering and medicine. Even Einstein’s famous thought experiments had to be tested. Usually they assume a low profile, but occasionally they receive intense focus from the public, as now. Suddenly, governments and the media are intensely interested in tests for Covid-19 (SARS-CoV-2), for its immediate presence during infection and for evidence of immunity acquired through a previous infection. It is not just the WHO that comprehends the importance of testing, testing and more testing right now.

You can read Chapter 1.1: Immunoassay for Beginners from The Immunoassay Handbook for a limited time now on ScienceDirect

As a practical example, in the midst of the Covid-19 outbreak, a man in Taiwan had a cough for a week, and then woke up with terrible chest pains. He went to the hospital, fearing he was infected with coronavirus, but was diagnosed as just having a bad cold. According to this man, who reported his experience on Quora (the question and answer website), the Taiwanese National Health Command Centre tests and quarantines anyone with symptoms. If a person tests positive, they trace their previous interactions to identify potential victims so they can test them. Only 348 of the 23 million residents had tested positive as of April 2nd, with 5 deaths, despite a huge regular passage of workers to and from countries with high infection rates.

Insufficient Covid-19 virus tests are available, even in California, a hub of the biotechnology industry, exacerbating an international crisis resulting in many deaths and significant economic damage. Tests for antibody to the Covid-19 virus (anti-SARS-CoV-2) will also be required in great quantities. It is essential that both types of test be developed, manufactured and carried out with relevant skills and extensive experience. Organizational capability and financing are also required. Ideally, existing platforms should be used, as this avoids reinventing the wheel.

The fields of infectious disease management and blood donor screening have been revolutionized by tests for viruses and the antibodies raised against them, but there are risks and pitfalls that need to be recognized. The current effectiveness of tests for HIV and hepatitis, for example, took years to achieve and several generations of tests.

The clinical performance of diagnostic tests for infectious diseases can be evaluated in a number of ways. Two of the simplest measures to understand are clinical sensitivity and clinical specificity.

Clinical sensitivity (or the detection rate) is the percentage of truly infected patients who are detected by the test. This is critical. If the sensitivity is 99%, 1% of infected patients will be missed i.e. are false negatives, and it cannot be assumed that every person who tests negative is free of the disease. In some cases, patients with symptoms of Covid-19 infection have been tested two or three times before a positive result is indicated. The sensitivity achieved by the test depends on the stage of the disease. Asymptomatic individuals who can infect others may be more difficult to detect in the test. The clinical sensitivity should therefore be estimated using a patient population mix appropriate to the intended application.

Clinical specificity is the percentage of truly uninfected test subjects who are correctly identified as being free of the virus. If the specificity is 99%, 1% of uninfected subjects show up as being infected in the test (false positives). A clinical specificity of 99% may be acceptable if the test is being applied to symptomatic individuals, but if applied to the general population with a real infection rate of 0.1%, 91% of the infections reported would be false positives.

In theory, reverse transcription polymerase chain reaction (RT-PCR) tests should achieve high levels of clinical sensitivity and specificity when performed by experienced personnel, but they are complex and there are multiple factors to consider.

Sampling technique is important, to ensure that the infected materials are collected and preserved. This may involve training hundreds of sample collection personnel. Good sample processing is critical to avoid patient mix-ups between sampling and the communication of results. The avoidance of cross-contamination is absolutely crucial for RT-PCR tests as the RNA from minute traces of virus from another sample or air-borne droplet can be amplified causing a false-positive result. False negative results may be caused by contamination, e.g. from powder in gloves when they are changed over, or reagents that have lost potency or become contaminated. Quality control samples should be run frequently to check assay performance.

The other type of test receiving much attention is testing for antibodies to the virus to establish prior infection and, hopefully, immunity. This is a simple concept to grasp, but the development of a reliable test, with high levels of clinical sensitivity and specificity, is demanding. It might be considered that false negatives, i.e. inability to detect Covid-19 antibodies when they are present, is less serious than false positives, which could lead a person to become exposed to infection, because they mistakenly think they are immune. But if a test consistently underestimates the number of immune patients it could mislead public health organizations that are tracking the outbreak and lead to policy errors. Detectable antibodies do not necessarily equate to full immunity from infection, and the duration of immunity after detection by this test will not be known for some time. Immunity can be short-lived or non-existent after some coronavirus infections. For SARS, immunity typically lasted about 2-3 years.

Soon after infection, the body generates IgM antibodies to fight the virus. For SARS, these were detectable in many patients after 1-2 weeks, but undetectable in some patients. Lagging behind this, IgG antibodies are produced, which provide immunity or partial immunity to future infections. Initial detectability of IgG antibodies in SARS patients varied from 1-14 weeks after infection. To achieve good clinical sensitivity the test must detect IgM and IgG antibodies (although sometimes separate tests are used) and detect a large number of different variants in the complementarity-determining regions of the subjects’ antibodies, as each immunized patient has a unique antibody population, with varying degrees of affinity for the virus. RNA viruses like SARS-CoV-2 have a higher rate of mutation than DNA viruses so the virus against which the patient antibodies are generated may be subtly different from one patient to another.

To achieve good specificity, antibody tests must not give a positive result due to the presence of antibodies against closely related coronaviruses. This is avoided by choosing clearly differentiated surface antigens or inactivated whole virus as the capture reagent(s) in the tests.

The development of Covid-19 tests involves the handling of positive, infectious samples and in some cases, live virus, so specialized containment facilities are required.

There are various versions of RT-PCR tests, but within each version, there is very little difference between tests for different viral RNA sequences. This is fundamentally different to the situation with tests for antibodies, which require significant skill and experience to develop, and very thorough validation. Antibody tests may be simpler to perform than RT-PCR, but tests for antibodies to new viral infections are much more challenging to develop.

Many of the chapters in my book, The Immunoassay Handbook, provide detailed background information. Immunoassays are tests for antigens, using antibodies as reagents, or tests for antibodies (as in this article) that use antigens as capture reagents. The chapter on antibodies also covers antibody technology relevant to some of the new treatments under development by the pharmaceutical industry. Although this book is directly applicable to antibody tests (or tests for virus using antibody-based reagents), the chapter on Clinical Concepts is also applicable to RT-PCR tests for virus. You can read Chapter 1.1: Immunoassay for Beginners for a limited time now on ScienceDirect

About the book

• Provides a unique mix of theory, practical advice and applications, with numerous examples
• Offers explanations of technologies under development and practical insider tips that are sometimes omitted from scientific papers
• Includes a comprehensive troubleshooting guide, useful for solving problems and improving assay performancee
• Provides valuable chapter updates, now available on immunoassayhandbook.com

The book is available now on ScienceDirect. Want your own copy? Enter code STC320 when you order via the Elsevier store to save up to 30%

To support the extraordinary efforts of the health and research communities combatting coronavirus, Elsevier has created a range of free resources, including textbooks, evidence-based clinical guidance, and more than 20,000 research articles to read, download and data mine. Find out more here.