COVID-19 diagnosis: lessons to learn and hints for preparedness

KEYWORDS:

• Pandemics
• COVID-19
• SARS-CoV-2
• diagnosis
• RT-PCR
• gold standard

In December 2019, the emergence of a previously unknown coronavirus in Wuhan, China, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), became widespread and placed the world in an unprecedented emergency [1]. The virus has infected around 520 million people, caused over 6 million deaths, and undergone the healthcare industry billion dollars of catastrophic financial costs. Despite the many shared characteristics of global outbreaks, the coronavirus disease 2019 (COVID-19) pandemic noticeably differs in some areas, including diagnostics, benefiting from extensive and routine testing schemes not available in similar situations in the past [2]. Molecular tests, serology assays, and clinical experience with disease manifestations are employed to improve diagnostics. However, not all the recruited approaches have tracked the disease appropriately. Besides, socioeconomic inequities have rendered management problems almost insurmountable in some countries [1]. Today, there is an excellent opportunity to learn from the experiences gained throughout the crisis to hinder future outbreaks from taking heavy tolls and managing them more efficiently.

Amidst urgent responses to a pandemic, implementing a proper diagnostic strategy is vital in detecting and isolating potentially infectious individuals. For this purpose, the polymerase chain reaction (PCR)-based testing of SARS-CoV-2 was developed and qualified during previous coronavirus outbreaks and is now adopted as the primary diagnostic approach [3]. However, referring to a single method as the ‘gold standard’ could be affected by the shortage of diagnostic materials and saving policies, issues contributing to poor performance in acute, global outbreaks. It is also necessary to distinguish between the tests intended for clinical diagnosis and population-scale screening, as they possess different requirements and priorities [2]. The above-mentioned, combined with weaknesses of common diagnostic tests, lead us to employ a general method followed by complementary approaches as a standard package for tracking the disease more reliably and in various circumstances. In a well-received article published in April 2020 [4], we discussed issues affecting real-time reverse transcriptase (RT)-PCR results in COVID-19 diagnosis and practical solutions to address the flaws. Considering what we have learned during the pandemic, we will discuss other critical points necessary to control current and future outbreaks more efficiently.

At present, it seems inadvisable to opt for other than a PCR test as the mainstream method for detecting viral genomes, as many other parallel techniques are not yet thoroughly assessed and qualified in practice. However, the RT-PCR test occasionally returns false-negative and false-positive results, contributing to the under or over-diagnosis of the disease [4,5]. Although sensitivity and specificity are substantial in prioritizing a test, the positive and negative predictive values (PPV and NPV, respectively) inform us about its accuracy and ultimate reliability in a clinical setting [6]. Moreover, disease prevalence is another significant but somewhat disregarded variable influencing PPV, NPV, and interpretation of the results. Accordingly, excessive false-negative reports are more likely to occur in high-prevalence populations, whereas low-prevalence settings may represent more false-positive outcomes [5]. In retrospect, we notice that the COVID-19 prevalence in different regions has been considerably variable and subjected to factors such as emerging variants and preventive measures. This diversity demands flexible diagnostic strategies to provide reliable test formats and reduce misdiagnosis. In low-prevalence communities, for example, to avoid the impact of high-rate false-positive results on the test validity, only symptomatic or high-risk individuals should be tested. Therefore, such statistical viewpoints are crucial to defining practical diagnostic guidelines based on specific criteria for different populations.

Even the most reliable results should be interpreted according to the disease characteristics. Notably, a PCR-positive result may not necessarily indicate a transmissible infection. This is particularly remarkable regarding SARS-CoV-2 infection, where RT-PCR results may remain positive for an extended period after symptoms have entirely resolved [2]. As such outcomes are chiefly derived from traceable genomes of killed viruses or low untransmittable viral shedding, reporting a positive result can lead to incorrect decisions on healthy recovering individuals. For this reason, reporting the cycle threshold (C_T) value seems advisable to provide a more precise evaluation of the patient’s status. For instance, a convalescing person with a high C_T value should not be treated the same way as a contagious patient who may require isolation and repeated PCR monitoring. It, therefore, implies that not all the positives are the same, and supplementary data can assist clinicians in differentiating patients more confidently. Notwithstanding being informative, due to many non-normalized or insufficient samples, the C_T value is inaccurate as a surrogate marker to quantify replication-competent viral load [7]. Nevertheless, reporting the C_T value remains beneficial for contact tracing, identifying high-risk people for developing symptoms, and applying more evidence in clinical decision-making [7,8].

In addition to technical considerations, mass testing on a pandemic scale requires leveraging all available tools. Although nucleic acid amplification tests (NAATs) generally comprise the primary detection method, antibody assays should not be ruled out from a comprehensive testing scheme. Serological tests are consequential for seroprevalence studies, from which valuable insights can be gained into infection decline or progression, the number of missed infected patients, and still-uninfected individuals in a population. Nonetheless, they are unsuitable for diagnosing ongoing SARS-CoV-2 infections and thus could not mitigate the heavy demand for COVID-19 testing. Furthermore, rapid antigen detection kits are not sensitive enough to be relied upon during the pandemic or as a replacement for NAATs; they could be useful adjuncts to RT-PCR in detecting active infections [9]. Radiology-based techniques such as X-rays, CT scans, and ultrasounds can also monitor lung abnormalities, especially in the early stages. However, their incapability to reveal disease progression in patients with low viral load makes them suitable only as auxiliary methods requiring improvements using deep learning and artificial intelligence (AI) [10,11]. Hence, this widespread use and dependence on a single technique have frequently resulted in a severe shortage of test kits and required materials from the onset of the pandemic. Perhaps the next critical step is providing the infrastructure for using a combination of tests with more reliance on various equipment and supply chains that secure delivery of required equipment for testing. In this regard, many suggested techniques, namely loop-mediated isothermal amplification (LAMP), clustered regularly interspaced short palindromic repeats (CRISPR)-based, transcription activator-like effector nuclease (TALEN), microarrays, and biosensors, are relatively accurate, cost-effective, and less labor-intensive compared to conventional options [10,12,13]. Utilizing novel technologies may contribute to higher accuracy and less inconsistent results by making it more feasible to sample from different anatomical sites and test various specimens [4]. Nonetheless, it seems impractical to test different specimens for hundreds of cases per day in acute phases of a pandemic unless for confirmation purposes. Moreover, modifying techniques and optimizing testing regimens are especially beneficial in diagnosing children and older adults, whose asymptomatic infections and lower viral load pose significant challenges in detecting the infected cases. For instance, pre-analytical errors could be minimized by developing extraction-free kits, and repetitive testing schedules could be optimized and incorporated into diagnostic programs for specific age groups [14]. While the feasibility of harnessing novel and optimized assays must still be assessed under different circumstances, progress in this area should favor decentralized diagnosis in non-hospital settings as a vital prerequisite for more timely and effective responses. To this end, online databases, mobile applications, and point of care (POC) or in-house diagnostic tools benefiting from AI and machine learning-based models are helpful in virtual diagnosis, surveillance, and disease management [10].

Lastly, other viruses capable of triggering new outbreaks should be monitored along with SARS-CoV-2 variants to spot unusual changes in their behavior and prevalence. Although a decline in other circulating viral respiratory infections was observed at some points in the current pandemic, recent out-of-season surges in respiratory syncytial virus (RSV) infections could be a harbinger of impending outbreaks of viruses such as influenza [15]. Moreover, the recent unexpected monkeypox and poliovirus outbreaks are perhaps sequelae of a post-pandemic situation in which restrictive measures, interrupted vaccination, and ensuing low immunity have rendered individuals more susceptible to endemic pathogens. Meanwhile, superspreading events can accelerate the out-of-ordinary spread of infections through huge gatherings, social activities, and sexual behaviors, to name a few. As part of global efforts to control a pandemic and prevent successive outbreaks, contact tracing measures should be maintained to identify and isolate potentially infectious patients. At the same time, diagnostic strategies should also target other viruses that may become serious candidates for future global disasters. As frequent mutations and emerging strains can lead to false-negative results and misdiagnosis, multiple target gene amplification and integration of genome sequencing within population-scale testing are necessary to identify emergent variants and newly-infected cases [2,4]. Monitoring human shedding of viruses in municipal wastewater and hospital sewage is a workable approach for capturing tiny changes in prevalence and genome mutations of seasonal threats such as RSV and Influenza as well as emerging pox, polio, and coronaviruses to take immediate action and prevent new outbreaks. Ultimately, while wildlife disease surveillance of zoonotic viruses is essential in preventing other viruses from crossing the species barrier, destroying the wildlife habitat should also be ceased to avoid the contact of new unknown pathogens with humans and further spillovers.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewers disclosure

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was not funded.