Molecular, serological, and biochemical diagnosis and monitoring of COVID-19: IFCC taskforce evaluation of the latest evidence

Preanalytical considerations

Analytical considerations

Clinical performance of molecular tests compared to chest CT

Most regulatory agencies require some level of analytical (e.g. limit of detection [LoD], inclusivity, and cross-reactivity evaluations) and clinical validation prior to authorization [23]. Despite the recommended validation procedures, SARS-CoV-2 diagnostic assays present the potential for reduced sensitivity and false-negative results as in the rush to put testing to use, evaluations have not been completed as well as usually done by laboratory professionals.

There are few studies comparing the diagnostic sensitivity of molecular tests to chest CT, which is considered an ancillary method in most regions. Specifically, Fang et al. in Wuhan, China first demonstrated significantly lower sensitivity of RT-PCR in OP and NP swabs (71%) compared to chest CT (98%) in a small retrospective analysis of 51 patients [24]. In a similar examination of 1014 cases, Ai et al. observed lower sensitivity in RT-PCR in the same specimens compared to chest CT. In fact, 48% of patients with negative RT-PCR results for SARS-CoV-2 were deemed as highly likely cases based on CT imaging [25]. Other reports with a higher prevalence of milder cases have determined comparable sensitivity rates for RT-PCR on OP and NP specimens and chest CT [26]. However, these reports lack information about the sensitivity of the used RT-PCR assays as well as details of the time intervals between symptom onset and sample collection for RT-PCR testing, especially in cases of negative results. Thus, interpretation of the observations remains challenging, in particular as a few reported intervals between chest CT images and corresponding RT-PCT suggest timeframes of > 1 week between symptom onset and RT-PCR [25]. As Wölfel et al. had shown, viral loads in throat swabs of some patients may become undetectable after this period [15].

In addition to comparison with chest CT, some studies have compared commercially available RT-PCR assays to ‘gold standard’ assays recommended by the WHO and CDC and against each other. These studies are increasingly important as many labs start to adopt commercial tests for clinical practice [27], [28], [29], [30], [31], [32]. For example, Poljak and colleagues [28] compared the commercial Roche Cobas SARS-CoV-2 test to the assay described by Corman et al. [33] and found good (>98%) overall agreement. Zhen et al. also compared four molecular tests (New York SARS-CoV-2 Real-time Reverse Transcriptase [RT]-PCR Diagnostic Panel [Modified CDC], the Simplexa COVID-19 Direct [Diasorin Molecular], GenMark ePlex SARS-CoV-2 assay [GenMark], and the Hologic Panther Fusion® SARS-CoV-2 assay [Hologic]), demonstrating variable LoDs (analytical sensitivity), but rather similar clinical performance [32]. In addition, Uhteg et al. reported similar analytical performance among three different assays (the RealStar® SARS-CoV-2 RT-PCR, ePlex® SARS-CoV-2, and the CDC COVID-19 RT-PCR tests) [29], while Lieberman et al. compared four commercially available assays (Cepheid, DiaSorin, Hologic Panther, and Roche Cobas) and demonstrated some variability in analytical sensitivity [30]. Importantly, a recent report also suggests increased potential in false-negative rates at the claimed lower limit of detection (LLoD) in six commercial assays [31]. Whether molecular assays are carried out at the POC or using automated clinical laboratory assays, the quality of these tests is paramount and more studies would be needed to further elucidate the relative performance of newly emerging tests. With variable study conclusions, it is important to consider both the analytical and preanalytical factors that impact the performance of molecular testing.

Regulatory perspectives and assay authorization

The WHO maintains a list of molecular assay protocols [5], which includes the Charité (Germany) assay described by Corman et al. [33], China Center for Disease Control and Prevention (China CDC) assay, United States Centers for Disease Control and Prevention (US CDC) assay, Institut Pasteur (France) assay, National Institute of Infectious Disease (Japan) assay, HKU (Hong Kong) assay, and the National Institute of Health (Thailand) assay. The WHO has also issued an Emergency Use List (EUL) pipeline and maintains a comprehensive list of assays authorized by different regulatory bodies [34]. The European Commission refers to the European Centre for Disease Prevention and Control (ECDC) and the WHO for guidance on molecular testing, although they have published their own directive regarding authorization, which is denoted when a product is ‘Community European (CE)-marked’. Unfortunately, a comprehensive list of ‘CE-marked’ assays is not available at this time. In the US, only assays that have received Emergency Use Authorization (EUA) from the FDA are permitted for use as diagnostic tests. Similarly, authorization from Health Canada must be received before assays can be used for diagnostic purposes in the country. Other countries, including Australia, Singapore, China, Korea, Russia, and Brazil, also maintain lists of authorized assays. In all cases, laboratory-based assays greatly outnumber the available POC assays on the market; however the latter are beginning to become more available in response to increasing demand for rapid test results. While the use of rapid POC assays for molecular diagnosis of COVID-19 is highly attractive, independent validation data are urgently needed prior to clinical use at locations beyond healthcare institutions. A list of POC NAATs that are authorized for use in the US and Canada is presented in Table 1.

Clearly, there is still much work to be done in the area of molecular testing in COVID-19, particularly as new commercial assays continue to be developed and/or modified. Furthermore, a better understanding is required during which phase of COVID-19 the application of NAAT testing in which kind of sample type is most beneficial for patient management. Thus, based on current evidence, an initial negative NAAT result does not rule out SARS-CoV-2 infection due to potential pre-analytical and analytical issues or time of testing. Accordingly, combining NAAT testing with other methods may be key to improved patient diagnosis and therapy monitoring. Finally, for reliable test results the availability and timeliness of provision of external quality assessment (EQA) programs are key, as they are necessary to identify systematic testing errors, provide objective evidence of testing quality, and compare laboratory performance. The Royal College of Pathologists Australasia Quality Assurance Programs (RCPAQAP) first produced a Proficiency Testing Program (PTP) for the detection of SARS-CoV-2 by PCR in early March 2020; the participants were a selection of Australian and NZ laboratories that were performing testing for COVID-19 at the time. A second round of proficiency testing involving many more laboratories is currently under way, with a third round planned for later in the year. The WHO has a global influenza proficiency testing scheme that will include SARS-CoV-2 in a testing panel targeting national reference laboratories and this is produced by the Centre for Health Protection (CHP)in Hong Kong. That panel is expected to be dispatched soon. Quality Control for Molecular Diagnostics (QCMD) in Scotland has a pilot EQA scheme for coronaviruses, the Coronavirus Outbreak Preparedness EQA Pilot Study, which is currently open and will also be dispatched soon. The ECDC is currently organizing EQA programs. In addition, the National Center for Clinical Laboratories in China has also started conducting EQA programs among 855 laboratories. Two Canadian-based companies, Microbix Biosystems Inc. and Oneworld Accuracy, also recently announced a joint collaboration to provide a global EQA program. In lieu of these programs, more independent validation data are urgently needed.

Preanalytical considerations

Analytical considerations

Clinical performance of serological tests compared to virus neutralization assays

It is important to highlight that current serology assays available from commercial diagnostic manufacturers do not provide any definitive information regarding patient immunity. For this analysis, lab-based neutralization assays in a cell culture system are needed to determine the presence of active antibodies and relative protection against future infection. To date, our understanding of whether the presence of antibodies confers protective immunity remains incomplete. Preliminary studies have shown evidence of neutralizing antibodies in serum of recovered middle-aged and elderly patients, suggesting an age-related innate immune response [56]. The demonstration of antibody neutralization is not only important for determining protective immunity, but also for feasibility and efficacy of therapeutic treatment with convalescent (“hyper-immune”) plasma [50], [57]. Currently, neutralizing antibodies from convalescent plasma donors have been used to treat COVID-19 in non-randomized settings [56]. Specifically, Shen and colleagues completed an uncontrolled case series of five critically ill patients with COVID-19 that had progressed to acute respiratory distress syndrome (ARDS) [58]. A larger study in 10 severe adult cases demonstrated that one dose of convalescent plasma could lead to clinical improvement within 3 days [59]. In either study, administration of convalescent plasma resulted in general albeit non-specific improvement in clinical status [58], [59]. These preliminary findings are encouraging for the safety and effectiveness of convalescent plasma treatment. However, they are not randomized trials and warrant further study to ensure that antibody-dependent enhancement of viral infection does not occur [60].

Regulatory perspectives and assay authorization

Serology assays currently available from diagnostic manufacturers vary significantly in their methodology, antibody target, and acceptable specimen type [49]. More than 90 serology assays have been submitted to the FDA, with 12 receiving EUA (as of May 11, 2020). However, unlike molecular COVID-19 testing, the FDA did not require EUA for the clinical application of serological testing until recently [61]. Table 2 summarizes the characteristics of available commercial serological assays that have been granted EUA. While these assays have been marketed widely, there is currently minimal peer-reviewed data on their performance in clinical settings. Thus, our knowledge relies almost completely on manufacturer claims and limited peer-reviewed and some non-peer-reviewed data [62], [63], [64], limited by sample size, patient characteristics, and sampling time. The lack of peer-reviewed data and regulatory oversight combined with general misconceptions regarding the appropriate uses of serology testing in COVID-19 have highlighted the crucial role of the clinical laboratory in robustly validating these tests to ensure sufficient analytical and clinical performance.

Overall, the diagnostic performance of serology assays has not been systematically evaluated, and more data are urgently needed to support their clinical utility in different settings. As Farnsworth et al. recently emphasized, in low prevalence settings, even immunoassays with excellent specificity can lead to insufficient positive and negative predictive values, and could misinform practice and policy [65]. Algorithms to confirm an initial positive assay result are common practice in transfusion medicine, usually employing alternative assays and methods like immunoblots, and might be valuable in COVID-19 in low prevalence settings to lower falsely positive result. It will be up to laboratory professionals to advocate for rigorous validation of these tests in the setting they will be clinically used, and then appropriate application based on the findings.

Inflammatory biomarkers

Several inflammatory biomarkers have been implicated in severe COVID-19, suggesting an immunochemical profile consistent with the so-called “cytokine storm”. In brief, elevation of proinflammatory cytokines, particularly interleukin (IL)-6 and tumor necrosis factor-α (TNF-α), has been observed in patients with severe disease and also found to be significantly associated with mortality [66], [67], [68], [69]. It is important to note that proinflammatory cytokines appear to be not solely biomarkers, but also causative factors in COVID-19 progression and mortality. Although cytokine measurement is not common in clinical laboratory practice, surrogate biochemical markers of inflammation including ferritin, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) have all been implicated in severe COVID-19 and can be useful in assessing disease severity [67], [68], [69]. Higher levels of procalcitonin have also been shown in severe COVID-19, suggesting onset of bacterial co- or supra-infection in critically ill patients [70]. In addition to biochemical markers of inflammation, hematological findings suggest that lymphopenia, as well as elevated neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) have prognostic potential. Elevation in the D-dimer coagulation parameter has also been consistently reported and associated with worsening disease and a higher risk of developing a wide spectrum of thromboembolic events in COVID-19 patients, including in situ pulmonary micro-thrombosis, deep vein thrombosis, overt pulmonary embolism, and even disseminated intravascular coagulation (DIC) [67], [71], [72], [73].

Overall, these data confirm that severe cases of COVID-19 are characterized by a massive proinflammatory response or cytokine storm that is estimated to progress to multiple organ damage and failure (i.e. MOF) in severe cases. Biochemical monitoring of COVID-19 patients will thus involve assessing the inflammatory profile, as well as early recognition of cardiac, renal, and hepatic injury through routine laboratory testing.

Cardiac biomarkers

Accumulating clinical data suggest prominent cardiovascular dysfunction in severe COVID-19 patients [74], [75]. Initial reports in China found that 12% of all patients and 31% of intensive care unit (ICU) patients presented with acute myocardial injury, as determined by increased cardiac troponin I levels [76]. Notably, patients with higher cardiac troponin values were more likely to be admitted to intensive care [76], [77] and showed higher in-hospital mortality [66], [78], [79], [80], [81]. Another case study series of 187 confirmed COVID-19 patients reported that 27.8% of patients had myocardial injury, which resulted in cardiac dysfunction and arrhythmias [78]. From these clinical data, it is clear that monitoring cardiac biomarkers such as cardiac troponin and natriuretic peptides throughout disease progression will be essential for appropriate patient risk stratification. The American College of Cardiology (ACC) recently published a statement on the role of cardiovascular biomarker monitoring in patients with COVID-19. They stated that “clinicians are advised to only measure troponin if the diagnosis of acute myocardial infarction is being considered on clinical grounds” [82]. This clinical advice is likely to avoid unwarranted diagnostic testing in COVID-19 patients and minimize downstream consultation and procedures, including bedside echocardiography and angiography [83]. However, many groups have rebutted this opinion, emphasizing that cardiac troponin should not be solely considered a binary test for diagnosing myocardial infarction, but a useful prognostic indicator of both ischemic and non-ischemic causes of cardiac dysfunction that can be extremely helpful in patient triage and proper treatment selection [83]. Suspected mechanisms for cardiovascular complications of COVID-19 include: (a) viral myocarditis, (b) direct myocardial injury, (c) cytokine-driven myocardial damage, (d) microangiopathy, and (e) exacerbation of coronary artery disease [75], [84]. Currently, none of these proposed mechanisms have been described as the main contributor to the cardiovascular complications observed in severe COVID-19. Further clinical studies will be important to pinpoint the main driver of myocardial damage for treatment purposes and for guiding laboratory assessment.

Hepatic biomarkers

Recent evidence suggests that liver dysfunction may also be commonplace in severe COVID-19 patients. Several large-scale hospital studies have reported elevations in select liver enzymes of both hepatic and mixed type, such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma glutamyl-transferase (GGT) [76], [77], [85], [86], [87], [88], [89], [90]. A recent study by Cai and colleagues completed a retrospective chart review of clinical and laboratory findings from 417 patients with laboratory-confirmed COVID-19 [89]. Of the 417 patients, 76.3% were reported to have abnormal liver marker results and 21.5% were reported to have liver injury during hospitalization [89]. Interestingly, while both GGT and alkaline phosphatase (ALP) are considered cholangiocyte-related enzymes, only GGT has been implicated in severe COVID-19, thus pinpointing that the damage is likely drug-induced rather than bile duct injury [89]. While it is clear that liver harm is associated with a higher likelihood of developing severe COVID-19, there is much debate regarding the cause of abnormalities in liver injury tests. Potential clinical mechanisms behind liver dysfunction in COVID-19 include: (a) immune-mediated damage due to the severe inflammatory response following infection, (b) direct cytotoxicity due to active viral replication in biliary epithelial cells which express ACE2, (c) hypoxic hepatitis due to anoxia, and (d) drug-induced liver injury [91], [92], [93]. As the contribution of underlying hepatic conditions and the clinical mechanism behind COVID-19-related liver dysfunction remains unclear, it is nevertheless evident that liver function tests should be ordered routinely to assess relative hepatic injury in COVID-19 patients within the clinical context, particularly for those receiving antiviral treatment.

Renal biomarkers

Previous data from the SARS epidemic in 2002–2003 reported that 6.7% of patients developed acute renal impairment, and that mortality of SARS patients with acute kidney injury (AKI) was as high as 91.7% [94]. Given the previously emphasized homology between the causative viruses, the evaluation of kidney dysfunction in COVID-19 appears to be quite important. Available data suggest that the prevalence of AKI is relatively low, but variable in COVID-19, ranging between 0.5 and 19.1% in various studies [76], [77], [79], [85], [87], [95], [96]. Both blood urea nitrogen and serum creatinine are known to be routine markers of kidney injury [96]. Indeed, the present method for detecting AKI is mainly based on acute changes in serum creatinine, whilst the frequency of serum creatinine testing has a substantial impact on detection rate [95], [97]. In a recent study by Cheng et al., patients with elevated baseline serum creatinine were more likely to be admitted to the ICU and to undergo mechanical ventilation [95]. These findings suggest that early identification and treatment of kidney deterioration, including adequate hemodynamic support and limiting nephrotoxic drugs, may be vital in COVID-19 care and increased frequency of creatinine testing or other kidney markers may be warranted. Serum and urinary albumin and total protein could also be useful as prognostic indicators in COVID-19. Similar to liver dysfunction, the pathophysiological mechanisms behind renal dysfunction in COVID-19 patients are unknown and likely multifactorial. Potential mechanisms include: (a) intrarenal inflammation and damage due to cytokine inflammatory response to infection, (b) direct cytopathic effects on kidney tissue, and (c) organ cross-talk (e.g. cardiomyopathy and acute viral myocarditis can contribute to renal vein congestion, hypotension, and renal hypoperfusion) [98].

Through these initial studies, it is evident that COVID-19 disease severity is associated with a strong inflammatory response leading to multi-organ failure. The clinical laboratory can play a critical role in assessing not only inflammatory markers but also closely monitoring indicators of potential organ failure together with typical measures in critical care (i.e. blood gas, electrolytes, etc.).