In this analysis, two mainstream platforms for amplification and quantification of SARS-CoV-2 displayed similar analytic characteristics. Both the CDC and the Fisher platforms produced highly linear Ct correlations with coefficient of determination close to 1 between their corresponding two viral targets (N1 and N2, or N and ORF1ab) used. Moreover, this analysis of 14,231 individual positive test results confirmed previous findings showing a wide spectrum of PCR Ct distribution in nasopharyngeal swab samples (Ct range 8–39, Fig. 1). The corresponding viral titers can range from few copies to billions of copies. When Ct distribution patterns were examined, samples appeared to form two peaks along the Ct gradients produced by either of the platforms (Fig. 3). However, the separation of the two peaks was less pronounced with the Fisher platform, presumably influenced by the multiplex chemistry owning to the self-inhibition among sets of primers and reduced amplification efficiency 12.
Notably, the distribution of Ct numbers observed in our series was bimodal (Table 1 and Fig. 3–5), suggesting contribution from two distinct subsets of samples. This effect is likely not an artifact of sample quality or preparation. The potential contributory factor pertaining to sample quality variation to the bimodality Ct distribution is ruled out as CDC platform included host RNase P as an internal control 13. Previous studies have demonstrated that the viral titer can be associated with inoculum size, tropism or replication in specific tissue or cell types, and risk of onward transmission 7. On the other hand, viral titers are not correlated with age or disease severity 14,15. Importantly, high levels of viral shedding may occur in asymptomatic hosts, posing substantial challenges to infection control efforts 16,17. However, there is currently little published information on COVID-19 Ct value distribution patterns or their significance to virus-host interactions in SARS-CoV-2 infection. A few studies that did note Ct distribution properties outside of normality did not analyze its significance in microbial and host relations distinctively associated with SARS-CoV-2 8,18. We explored whether the pattern of viral levels at the population level could provide insight into the nature of SARS-CoV-2 shedding difference potentially useful for infection prevention.
When the Ct distribution pattern was examined by age groups, the heterogeneous non-unimodal distribution was evident. For age group of < 5, <21 (by N2 only), 21–64, and 65 + years, their Ct distributions have met the bimodality coefficient criteria (Table 1). However, the non-bimodal nor unimodal Ct distribution pattern associated with the age groups of < 12 and < 17 years remains puzzling, when that of age group < 5 years was clearly bimodal. This result suggests there may be underlying differences between viral replication in very young patient’s vs teen’s. Otherwise, the bimodal nature of the Ct distribution was unaffected by gender, or calendar time-period, during which several different variants predominated. Notably, the Ct distribution of the 878 omicron samples appeared to show the two positive peaks skew closer into each other (Supplemental Fig. 1). The putative Omicron Ct distribution curves failed Bimodality Coefficient test. It is again possible that the Fisher multiplex chemistry suppressed the expression of bimodality as seen in the overall 52-week analysis (Table 1). However, at this point we believe there is still sufficient evidence to support the finding of this dichotomous distribution of viral replication pattern in the host population. More studies using other test platforms are needed to confirm this finding.
Host factors must play a role in heterogeneous viral replication properties. SARS-CoV-2 cell entry is mediated by human angiotensin-converting enzyme II (ACE2) and ACE2 polymorphisms, which may affect the risk for SARS-CoV-2 infection and the course of COVID-19 19. In a multivariable analysis by Nikiforuk et al., the researchers showed that the greatest viral RNA loads were observed in participants with high transmembrane ACE2 transcription, while transcription of the soluble isoform appears to protect against high viral RNA load in the upper respiratory tract 20. It is possible that multiple host genetic factors, innate and adaptive immunity, and respiratory microbiota may all play roles in viral titers and disease outcomes 21,22.
Our data support the notion that high viral load carriers may contribute most to new transmissions in the community 17,23. An operational categorization separating high from low/moderate viral shedding could therefore be relevant to isolation requirements after infection, and infection control efforts. Using a cutoff value of Ct < 22–24, corresponding to the upper bound of 0.5 SD − 1 SD of the first peak, representing 47% − 56% of individuals in this cohort, could be used as indicators separating levels of respiratory tract viral shedding potentials. Ideally any categorization would be tested against presence of culturable virus and risk of transmission in clinical studies.
There are several limitations of this analysis. This study did not include host information on clinical course, vaccination status, or immune responses at the time samples were collected. We are therefore unable to explicitly relate Ct values with these clinical factors. In addition, we do not know the identities of viral strain or variant associated with most of the Ct values obtained, and this information would be helpful in formally evaluating the role of infecting variant on Ct values. With the data set size, it is likely that that the samples collectively represent a random distribution along the clinical course of the viral infection. This dichotomous Ct distribution can be deployed in test reports as a relatively simple indicator that can be useful for the management of infected patients.
It has been well acknowledged that the PCR Ct values or their associated viral titers do not correlate well with the intensity of symptoms during SARS-CoV-2 infection, nor are they predictive of disease outcome; thus, they are currently not routinely used in clinical management. (https://www.aphl.org/programs/preparedness/Crisis-Management/Documents/APHL-COVID19-Ct-Values.pdf and https://www.idsociety.org/globalassets/idsa/public-health/covid-19/idsa-amp-statement.pdf). However, it has been suggested that those presenting with higher Ct values may require shorter periods of isolation to prevent onward transmission 24. Since the implications of super-spreading events are well known, we thus cautiously propose that in samples with a very low Ct finding (Ct < 22–24 or viral titers predicated to be in millions), a notation in test report could be considered. Clinical recommendations surrounding clinical standard actions such as isolation and period of quarantine with repeat test assurances for deisolation have to be developed before a reporting notation can be implemented. With the emergence of iterations highly transmissible SARS-CoV-2 variants of concern, it is prudent to empower existing testing and reporting strategies to reduce ongoing community transmission in order to control the case growth rate, healthcare burden, and workforce preservation.