Thirty-six studies (7222 patients) demonstrated a significant association between pharyngeal viral load at onset of symptoms with severity of COVID -19 and ICU care[4-9,13,15,17-20,24,26,27,29,30,32,33,36-38,41,42,44,45,48,49,51-56,58,59]. The majority of these studies reported highest viral load at onset of symptoms.

Most studies consistently defined severity of illness and sepsis as: Respiratory rate ≥ 30 beats/min, resting-state oxygen saturation ≤ 93%, arterial partial pressure of oxygen/oxygen concentration ≤ 300 mm Hg or mechanical ventilation, shock, or multiple organ failure requiring care in ICU[4,8,29,65].

There is variation observed in kinetics, tissue distribution and antibody response between mild and severe infections. Wang et al[13] analyzed a cohort of 12 severe and 11 mildly ill patients and demonstrated a significant difference in the initial nasopharyngeal peak viral load (P < 0.001) between two groups. Subsequent prolonged viral shedding in other body fluids and stool occurred with detectable viral load for up to 40 d (days) in severely ill compared to 15 d in mildly ill group. Viral RNA was detected from respiratory tract, stool, plasma and urine samples in the severe group. Mildly ill patients had viral shedding restricted to respiratory tract and no virus was detected 10 d after onset of symptoms[13].

Yu et al[20] analyzed their cohort of 92 patients and observed that high viral load in baseline sputum samples was linearly associated with severity and risk of disease progression (P < 0.017).

Another cohort of 96 patients with mild and severe infections demonstrated similar viral kinetics. Respiratory viral load remained elevated in the severe group up to the third and fourth week after disease onset, compared to milder group where viral load peaked in the second week followed by a decline. Subsequent viral detection in serum samples was also higher in patients with severe disease than in patients with mild disease (45% vs 27%, P < 0.03)[15].

In general nasopharyngeal viral levels remained high in severe group and, begin to decrease after 14 d of symptom onset[4,15,65]. Subsequently, samples from other sites may also test positive for the virus. For example, viral load from stool samples were found to peak during the third and fourth weeks after disease onset and continue to remain positive during convalescence[9,13,15,19,25,31]. Some studies also reported presence of high viral load in stool up to 50 d after onset of COVID-19 symptoms[31,38].

Significance of viral load in stool remains unclear, whether it represents a true infection or residual viral nucleic acid and not transmissible live virus. Gastrointestinal epithelium also expresses angiotensin-converting enzyme II (ACE-2) receptors. Infection of gastrointestinal (GI) tract may occur primarily from swallowed nasopharyngeal secretions or due to dissemination to GI tract from viremia[23]. Eighteen studies (5479 patients) demonstrated a statistically significant (P value < 0.005) association between higher viral load in different samples and severity of disease[4,7,8,13,17,27,29,30,32,36,45,48,49,51,52,54-56].

Liu et al[4] analyzed their cohort of 46 mild and 30 severely ill patients with elevated nasopharyngeal viral load and demonstrated an association with severity. Viral load was 60 times higher in severe cases and with severe clinical outcomes (P < 0.005). Mild cases had viral clearance, with 90% of patients testing negative after 10 d. In contrast, all severe cases had persistently elevated viral load beyond 10 d of symptoms were elderly and required ICU care.

In a cohort of patients on dialysis, Schwierzeck et al[41] also demonstrated a similar association with severity. Ct values of symptomatic cases were significantly lower compared to asymptomatic cases (22.55, 29.94, respectively, P = 0.007), indicating approximately 200-fold higher viral load[41]. Similarly other authors from their cohorts from different countries Bermejo-Martin et al[48]; Spain, Shlomai et al[49]; Israel^, Chen et al[52]; China, Zhou et al[54]; China, Maltezou et al[55]; Greece have demonstrated a statistically significant association between admission high viral load and intubation, ICU care and multi-organ dysfunction.

Collectively these data from different cohort of patients suggests that severe COVID-19 patients with a high viral load correlate with higher risk for severe infection with ICU admission and multi-organ dysfunction. Factors common to these cohorts was increased age, and active preexisting medical co-morbidities.

Higher viral load on admission samples were also associated with elevated levels of IL-6, cytokines, lactate dehydrogenase (LDH), lymphopenia and elevated neutrophil/lymphocyte ratio; indicative of poor sequential organ failure assessment (SOFA) scores and associated with hyper-inflammatory state contributing to the severity of sepsis[8,19,24,29,33,36,37,42,48,49,52,65,66].

In a cohort of 48 patients, Chen et al[8] reported an association between high viral load in serum with elevated Il-6 Levels (≥ 100 pg/mL) and cytokine storm in critical compared to mildly ill patients (P < 0.001). These patients had a higher incidence of multi-organ failure and mortality.

Similarly Xia et al[42] in their cohort of 10 patients with severe illness and elevated nasopharyngeal viral load reported severe lymphopenia with CD4+ lymphocyte counts as low as 61 cells/uL (reference value: 355-1213 cells/µL). Neutrophil to lymphocyte ratio was also elevated in this group.

Liu et al[65] reported their cohort of 46 patients with severe illness and elevated nasopharyngeal viral load. CD4⁺ and CD8⁺ T lymphocyte count displayed a linear negative correlation (P < 0.001) with high viral count; and positively correlated with IL-2R, prothrombin time, lactate dehydrogenase, and hypersensitive troponin T (P = 0.002, P = 0.009, and P < 0.001, respectively). Also elevated, were levels of inflammatory factors, IL-2R, IL-6, IL-8 Levels in the severe compared to mild group (P = 0.022, 0.026, and 0.012, respectively)[65].

Blot et al[33] in their series of 14 patients demonstrated a positive correlation of high nasopharyngeal viral load on admission with risk of hypoxemia, increased oxygen requirements and SOFA score in respiratory distress syndrome patients (P = 0.013). Similar association with increase in severity of sepsis, organ damage and mortality was also reported by Xu et al[36].

Lucas et al[66] in their series of 113 patients with COVID-19 patients demonstrated an overall increase in cells of innate lineage and a reduction in T lymphocytic cell counts. High viral load correlated significantly with levels of IFNα, IFNγ, TNF and tumor necrosis factor-related apoptosis-inducing ligand. Chemokines responsible for monocyte recruitment correlated significantly with viral load in severe disease. Inflammasome associated cytokines were also elevated, including IL-1α, IL-1β, IL-6, IL-18 and TNF[66].

Similarly Han et al[67] in their series of 60 critical patients demonstrated high levels of cytokines TNF-α, IFN-γ, IL-2, IL-4, IL-6, IL-10 and C reactive protein (CRP). Serum IL-6 and IL-10 Levels were significantly higher in critically ill compared to moderately ill group. The levels of IL-10 positively correlated with CRP (r = 0.41, P < 0.01)[67].

Collectively these studies provide evidence that high viral load may be a surrogate marker for predicting inflammation and severity in COVID-19 infection.

Subgroup analysis of 20 studies (7183 patients) demonstrated an association of admission viral load with in hospital mortality[8,9,27,29,30,36,37,45,46,48,49,51-56,58-60]. Majority of patients in this category were older (median > 65 years) and with medical comorbidities[8,9,29,30,33,36,37,45,46,48,49,58-60]. High admission viral load was an independent risk factor for in hospital mortality (P < 0.005)[8,27,29,30,36,46,48,49,51,52,54,59,60].

Pujadas et al[27] demonstrated an association of viral load as an independent predictor of mortality in a cohort of 1145 hospitalized patients. Mean log₁₀ viral loads significantly differed between patients who survived [n = 807; mean log₁₀ viral load 5.2 copies/mL (SD 3)] vs those who succumbed [n = 338; 6.4 copies/mL (SD2.7)]. Cox proportional hazards model was adjusted for age, sex, asthma, atrial fibrillation, coronary artery disease, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, heart failure, hypertension, stroke, and race. The results demonstrate a significant independent association between viral load and mortality [hazard ratio 1.07 [95% confidence interval (CI): 1.03–1.11), P = 0.0014], and 7% increase in hazard for each log transformed copy/mL. Univariate survival analysis also demonstrated a significant difference in survival probability between high and with low viral load (P = 0.0003), with a mean follow-up of 13 d and a maximum follow-up of 67 d[27].

Magleby et al[30] in their cohort of 678 patients demonstrated that higher viral load was associated with increased age, comorbidities, smoking status, and recent chemotherapy. Mortality was highest, 35.0% in the high viral (Ct < 25; n = 220) followed by 17.6% in the medium viral (Ct 25-30; n = 216) and 6.2% with a low viral load (Ct > 30; n = 242; P < 0.001). The need for mechanical ventilation was also highest in the high viral (29.1%), compared to medium (20.8%) and low viral load (14.9%; P < 0.001) group. High viral load was independently associated with mortality [adjusted odds ratio (OR) 6.05; 95%CI: 2.92-12.52; P < 0.001] and intubation (adjusted OR 2.73; 95%CI: 1.68-4.44; P < 0.001) in multivariate models.

Similarly Huang et al[29] in their analysis of 308 patients demonstrated a high viral load associated with in-hospital mortality in (6/16) of critical patients, while no mortality was observed in the low viral load group (P < 0.0001). High viral load was associated with myocardial damage, elevated troponins, coagulopathy, abnormal liver and renal functions. Elevated IL-6, LDH, and elevated neutrophil counts and reduced CD4+, CD8+ lymphocytes were noted in deceased patients P < 0.0001)[29].

In a cohort of 109 patients Bryan et al[59] demonstrated high viral load on admission was associated with a significantly increased 30-d mortality (OR, 4.20; 95%CI, 1.62–10.86. Their data suggested that a CT value of 22 may serve as a useful discrete cutoff for significant viral replication that is associated with mortality[59].

In a cohort of 1044 patients, Choudhuri et al[60] demonstrated a statistical correlation of Ct value at admission was higher for survivors (28.6, SD = 5.8) compared to non-survivors (24.8, SD = 6.0, P < 0.001). After adjusting for age, gender, body mass index, hypertension and diabetes, increased cycle threshold was associated with decreased odds of in-hospital mortality (0.91, CI: 0.89-0.94, P < 0.001)[60].

Collectively these multiple cohort of patients from different studies shows a trend of the association of high viral load and mortality in hospitalized patients.

Although not statistically significant, 20 studies (1857 patients) indicated the importance of high viral load dynamics with infectiousness and transmissibility (P > 0.05-0.53)[1-3,10-12,14,16,21,22,23,25,28,31,34,39,40,43,47,50].

Association between viral load and infectivity remains unclear, but earlier peak in viral load in SARS-CoV-2 infection suggests that infectivity may be higher earlier in the course than would be expected based on the SARS model[5,62,63].

Subgroup analysis suggests these patients are younger and had milder disease and may be highly infectious and transmit virus to the population given their asymptomatic or presymptomatic nature of illness. These studies shed light on high viral load and its association with infectivity and transmissibility. Highest respiratory viral load was noted at pre-symptomatic stage and infectiousness peaked before symptom onset[1,2,3,5,12,14,16,22,34,40,47,50].

He et al[1] demonstrated an infectiousness profile on 77 infector–infected transmission pairs. Highest viral load in oropharynx at the time of symptom onset correlated with infectiousness. Presymptomatic transmission was 44% (95%CI, 30%–57%) whereas infectiousness started at 12.3 d (95%CI, 5.9-17 d) before symptom onset and peaked at onset (95%CI: –0.9 to 0.9 d). They estimated that proportion of presymptomatic transmission was 37%-48%[1].

Xu et al[2] reported on 51 symptomatic patients, demonstrating transmission from primary (patients who visited the epicenter, Wuhan), to secondary (patients who came into contact with primary) and tertiary (patients who came into contact with only secondary cases). Their findings suggested incubation period in tertiary group was longer compared to primary and secondary groups (both P < 0.05). Ct values detected in tertiary were similar to those for the imported and secondary patients at the time of admission (both P > 0.05). For tertiary group, the viral load was undetectable in half of patients (52.63%) on day 7 and in all patients on day 14. One third of patients in imported and secondary groups remained positive on day 14 after admission. They concluded that infectivity of SARS-CoV-2 may gradually decrease in tertiary patients[2]. This study emphasizes that early quarantine and lock down measures may have mitigated the spread of disease in countries that enforced it strictly. The reason for decrease in infectivity from secondary to tertiary exposed patient remains unclear. Although speculative, this may be due to reduced quantitative viral load transmitted and other strict mask and quarantine measures[2,44].

Some reports demonstrated an association of high viral load and risk of transmission in a closed knit population[28,40]. In a cohort of 80 patients including both health care workers and nursing home residents from COVID-19 outbreak in Washington State, high viral load in unrecognized asymptomatic and presymptomatic patients contributed to infectiousness and transmission. Although the mortality was high in these patients, it did not correlate statistically with the viral load[28]. Similarly Kimball et al[40] analyzed their cohort of 23 patients from a long term care facility. Ten (43%) had symptoms on testing, and 13 (57%) were asymptomatic. Seven days after testing, 10 of these 13 previously asymptomatic residents had developed symptoms and were inferred as presymptomatic at time of testing. The Ct values indicated large quantities of viral RNA in asymptomatic, presymptomatic, and symptomatic residents, suggesting potential for transmission regardless of symptoms[40].

There are at present limits to our understanding and evidence in determining infectiousness and the risk of transmissibility. As described earlier, there is evidence of ongoing viral shedding in various body fluids after symptom resolution in COVID infection and may be prolonged, especially in stool samples compared to respiratory secretions (P < 0.001-0.5)[9,13,15,19,25,31,38,67]. Currently there is no reported evidence of fecal –oral transmission. Further the severity of illness also appears to extend the duration of viral shedding. However, based on current data, there is no convincing evidence that duration of shedding correlates with duration of infectivity. The viral nucleic acid detected in various body fluids later in the course of infection may represent non-viable fragments of virions.

Wölfel et al[14] demonstrated that live virus can be cultured from respiratory samples in patients with positive SARS-CoV-2 RT-PCR. However, the percentage of positive cultures declined and no live virus was successfully isolated after day 8 from symptom onset despite ongoing high quantitative viral load. Additionally, virus could not be isolated from samples less than 10⁵ copies/mL. However a caveat with this cohort was that patients had mild symptoms and were young and middle aged adults. This emphasizes the point that elevated high viral load in convalescing patients may be suggestive but not a definitive factor in infectiousness and transmissibility[14].

There is evidence that children are susceptible to SARS-CoV-2 infection, but frequently do not have symptoms, raising possibility that children could be facilitators of viral transmission. Reports comparing viral kinetics in adults and pediatric patients have demonstrated that children, adolescents and adults can have same variation of viral load, but higher risk of transmission and asymptomatic illness in children may have other contributing factors[16,47,50].

The immune responses of the host to COVID-19 and its relation to infectivity and transmission remain unclear and data is emerging[5,13,59,68,69]. Most patients seroconvert by day 15 after symptom onset and Anti-SARS-CoV-2-NP or anti-SARS-CoV-2-RBD IgG levels correlate with virus neutralization[5]. While risk of transmission after symptom resolution and the presence of antibodies may be lower, it cannot be ruled out with available evidence[1-3,5]. Transmission by asymptomatic or minimally symptomatic individuals also appears likely and highlights the importance of contact tracing and isolation of exposed individuals, especially as transmission potential may be maximal early in course of infection as depicted in the nursing home cohort[28,40]. In their large series of 100 patients Li et al[68] demonstrated specific anti SARS-CoV-2 (IgM, IgG, IgA) antibodies to S-1, N, and RBD viral proteins in the serum within two weeks after onset and reached a peak in 17 d and maintained high levels up to 50 d post infection.

Fourati et al[69] demonstrated an inverse relationship of lower serum titer of neutralizing antibodies (anti-S1 Ig A and Ig G) with elevated nasopharyngeal viral load and severe COVID-19 sepsis. This may indicate an inability to clear infection and have a deleterious impact on survival. Patients who were alive at 28 d displayed higher titers of anti-S1 Ig A and Ig G on admission compared to those who succumbed[69]. Similar observation was demonstrated by Bryan et al[59]; this study demonstrated that detection of anti-SARS-CoV-2 nucleocapsid IgG is associated with lower viral loads in patients. They concluded that high viral loads almost never coexist with SARS-CoV-2 sera-positivity and suggest that persons with anti-SARS-CoV-2 antibodies on admission have reduced 30-d all-cause mortality[59]. Both these studies may suggest that presence of antibody titers on admission, coupled with molecular testing, may be particularly prognostic factor, helpful to assess the disease course for high risk patients who cannot provide a clinical history[59,69]. The mechanism may be due to lower host humoral immune response in the elderly patients with comorbidities.

The heterogeneity of the non-respiratory specimen’s limits its significance in explaining the risk of transmission and no correlation can be inferred. Further research is needed. In addition it is also important to determine viability of virus outside the respiratory and gastrointestinal tract at different stages of infection in both asymptomatic and symptomatic individuals. This will improve understanding of transmission risk and allow greater certainty around guidelines for appropriate contact tracing and quarantine periods[70].

This study is a large pooled, qualitative content analysis of 60 manuscripts with a cohort of 10514 patients’ from different cohorts and countries evaluating patterns of quantitative viral load in predicting disease severity, mortality, risk of infectiousness, transmissibility, and prognosis in patients with COVID-19. The author presents the relative merits and discusses the objective data presented in these studies. This a correlation study and does not imply causation.

However, there are certain limitations in this study. Since there is a high heterogeneity of samples and data in the majority of these manuscripts, the content analysis is qualitative (narrative) and these data should be interpreted with caution and considered only as trends. Differences in distribution of age, sex, definition of disease severity, and other confounding variables such as medical comorbidities, different virologic tests and heterogeneous samples may contribute to different clinical outcomes. For instance very few studies adjusted their statistic models for the other medical morbidities which could have increased the risk for morbidity and mortality[4,6,7,15,19,27,30]. The majority of these studies are on hospitalized patients which has a potential bias of analyzing the more severely ill amongst the overall infected population. Further variations of ACE 2 receptors and expression in various tissues in different ethnic populations may play a role in virulence and transmissibility of this virus[76]. A viral nucleic acid load from a particular sample assay may not represent an exact systemic viral load in the body; further viral load may also not represent viable virions and may be falsely misleading. In addition there is no consistent trajectory of why certain samples test positive with high virus loads and others do not. Another important point to consider is that, majority of studies is from one country: China and from a few medical centers around the epicenter of outbreak, possibly leading to overlapping of population data in reported manuscripts. Other limiting factors may include the testing protocol and standards, set for RT-PCR targets vary between different laboratories[68-70]. Finally there is always a possibility of observer (author bias) which is to be considered.

Although majority of studies showed a positive association between a high viral load and mortality there were three studies with (434 patients) suggestive of an inverse correlation between the two. Argyropoulos et al[35] in their report on 205 patients demonstrated an inverse correlation of admission nasopharyngeal viral load with duration, severity of sepsis and no correlation with survival (P < 0.001). The reason for low mortality in this study is unclear. One possible explanation could be due to the fact that viral loads detected from nasopharyngeal samples were obtained at a later time point in the disease course. As we have described earlier, that SARS-CoV-2 viral load peaks earlier in the infection followed by cytokine storm and hyper-inflammation when the innate immune system is unable to control the initial viral replication[61]. At these later times points the viral replication may start to defervesce but the multi-organ dysfunction is secondary to systemic hyper-inflammatory response. Similarly Hasanoglu et al[46] on their cohort of 60 patients demonstrated an inverse relationship of high viral load with mortality; however their study had a mean age of 32 signifying a younger age group, where mortality is lower compared to older patients. Another group of 169 patients, reported from Spain by Carrasquer et al[57] demonstrated no statistical association of high viral load with in hospital mortality when adjusted to age, gender and serum cardiac troponin levels. The conclusions from this study suggested myocardial damage with medical comorbidities as the cause for increased mortality in susceptible population and not high viral loads.

Although infection and inflammation begins with the respiratory tract, it also involves extra pulmonary organs[77]. Isolation of viral nucleic acid in multiple tissues, blood and body secretions are indicative of systemic spread and are indicative of severe infection. Evidence from these manuscripts suggests that high viral load occurs in respiratory tract samples during presymptomatic period and peaks at the onset of symptoms and gradually declines over the next one to three weeks[1,2,3,5,9,12,14,16,22,34,40]. Increased viral load in respiratory tract represents active viral replication and a surrogate marker for predicting severity[28,32,37,61]. This is in contrast to previous SARS-CoV epidemic in 2003 where the peak viral load occurred during second week after symptoms appeared and was positively correlated with increased mortality[5,62,63]. This fact explains the increased infectivity and rapid transmission of SARS-CoV-2 compared to previous SARS-CoV epidemic[5]. Along with comorbidities, assessment of viral load from nasopharynx or sputum may determine the risk of severity of sepsis in symptomatic, hospitalized elderly patients[4,5,18]. High viral load is also associated with elevated cytokine, lymphopenia i.e., markers for inflammation and portends poor prognosis[8,24,33,36,37,42,52,65,66]. Early determination of viral load also has therapeutic benefits, such as administration of convalescent plasma, neutralizing antibodies, antiviral medicines and corticosteroids in susceptible elderly patients[6,7,11].

SARS-CoV-2 pandemic continues to spread unabated in United States and worldwide. This is particularly evident after the end of lock down and social distancing measures with increased mobility of the population. A report from a reference laboratory evaluated 29713 de-identified samples from respiratory tract. 14.9% of samples tested positive. Highest positivity rate was identified in males born between1964-1974. Patients between ages of 11-25 had highest viral load (> 10 Log₁₀ copies/mL). The clinical symptoms or outcomes of these patients were not known. This study demonstrates that high viral load in younger group may be an important risk factor for infectivity and transmission in a community, regardless of their symptom status[78].

COVID -19 infections in younger asymptomatic patients, with high viral load may fare well due to their robust physiologic reserve. However, they are at highest risk for transmitting the disease and are called super spreaders. These infections generally appear asymptomatic or milder in younger population, but elderly patients bear the brunt of severe infection, hospitalization and mortality[61,62].

High viral load has an implication in the clinical outcomes in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. At present there is no Food and Drug Administration Emergency Use Authorization for quantitative viral load assay in the current pandemic. Currently the coronavirus disease 2019 (COVID-19) tests are reported as a binary assessment of either positive or negative test.

The intent of this research is to identify whether quantitative SARS-CoV-2 viral load assay correlates with severity of infection and mortality?

To assess high viral load and its association with the severity, mortality, infectiousness in COVID-19 infections.

A systematic literature search was undertaken for a period between December 30, 2019 to December 31, 2020 in PubMed/MEDLINE using combination of terms “COVID-19, SARS-CoV-2, Ct values, Log₁₀ copies, quantitative viral load, viral dynamics, kinetics, association with severity, sepsis, mortality and infectiousness’’. Data on age, number of patients, sample sites, real time reverse transcriptase polymerase chain reaction (RT-PCR) targets, disease severity, intensive care unit admission, mortality and conclusions of the studies was extracted, organized and is analyzed.

High SARS-CoV-2 viral load was found to be an independent predictor of disease severity and mortality in high proportion of studies, and may be useful in predicting the clinical course and prognosis of patients with COVID-19.

There is a wide heterogeneity in fluid samples and different phases of the disease and these data should be interpreted with caution and only considered as trends. In aggregate, these observations support the hypothesis of checking and reporting viral load by quantitative RT-PCR, instead of binary assessment of a test being positive or negative.

In future, longitudinal studies with viral load should be monitored and analyzed, so it can be considered in interpretation of outcome data. It may also be a guiding principle for therapy and infection control policies for current and future pandemics.