Introduction

As of August 3, 2022, children accounted for 20.7% (8.7% in the 0–9 cohort and 12% in the 10–19 cohort) of the 21,243,890 confirmed cases of COVID-19 in Italy, according to the latest report from the Italian National Institute of Health.1 The clinical features of SARS-CoV-2 infection in children have been well described. The most recent systematic review and meta-analysis, including 37 studies on pediatric COVID-19, agrees with previous findings on the relatively mild course of the disease in younger children, with fever (48.5%) and cough (40.6%) as the most reported symptoms by 2874 children.2 In the absence of more specific signs and/or symptoms of pediatric COVID-19, previous exposure to a person infected with SARS-CoV-2 becomes a critical discriminator for the diagnosis of the disease in children.3 While we have learned much about the diagnosis and management of pediatric COVID-19, we know little about the relationship between clinical presentation and viral load of SARS-CoV-2 in children. Most studies have focused on the association between viral load and infectivity and/or age4,5 rather than symptom severity. However, it has recently been shown that viral load is significantly higher in adults with severe disease than in those with mild disease6 and that prolonged viral shedding with high viral load has been associated with adverse outcomes in hospitalized patients.7 This finding is interesting because no significant differences in SARS-CoV-2 nasopharyngeal viral loads were found between adults and children.8 Only one recent study evaluated the association between viral load, age, and symptom severity in children, showing no association between the three factors.9 However, this study did not focus on the relationship between SARS-CoV-2 viral load and the specific signs and symptoms of pediatric COVID-19.

Nucleic acid amplification tests, based on real-time reverse transcription polymerase chain reaction (RT-PCR), are the reference method for the diagnosis of acute SARS-CoV-2 infection.10 Cycle threshold (CT) levels obtained at the end of RT-PCR are inversely proportional to the amount of target nucleic acid in the sample and can be used as a proxy marker of viral quantification11 of the viral RNA load of SARS-CoV-2.12 WHO and CDC reports have shown comparable sensitivity among different target genes13,14 and a national study of more than 100 Korean laboratories showed overlapping results for CT values obtained from different RT-PCR assays.15 The determination of SARS-CoV-2 viral load is critical for the interpretation of laboratory tests and the management of isolation and contact tracing protocols.16 Attempts are being made to use the value of CT to understand the association between viral load and disease severity or to predict the evolution of SARS-CoV-2 infection. Indeed, studies based on clinical outcomes have shown that higher SARS-CoV-2 viral load is associated with higher mortality17,18 and increased risk of intubation in COVID-19 patients.19 Recently, it has been shown that low viral load (CT > 35) is not substantially associated with ongoing virus replication.13,20 Despite these findings, the clinical association with high or low viral load is still debated. A study involving 5830 patients with COVID-19 determined that the viral load of symptomatic and asymptomatic patients was not significantly different.21 On the other hand, some studies have shown that CT levels are a robust predictor of mortality rates among patients after the initial diagnosis of SARS-CoV-2 infection.22,23,24 In a retrospective study of 461 adult patients hospitalized with COVID-19, a 4.5-fold increase in the odds of death was reported for patients with a high SARS-CoV-2 viral load.24 However, no study has yet evaluated the clinical association of different SARS-CoV-2 viral loads in the pediatric population.

The purpose of this study is to evaluate the association between SARS-CoV-2 viral load in respiratory secretions of infected children and the specific signs and symptoms of pediatric COVID-19.

Materials and methods

Study design and setting

We retrospectively reviewed the cases of all children with a positive nasopharyngeal swab (NPS) for SARS-CoV-2 who were admitted to the emergency department (ED) of the Bambino Gesù Children’s Hospital, Research and Healthcare Institute, Rome, Italy, between May 2020 and April 2021. The ED provides free 24/7 urgent medical care. During the study period, there were 40,431 visits to the ER, 7487 of which resulted in admission to a ward. Patients were tested for SARS-CoV-2 whenever they presented with fever or any respiratory symptoms, including cough, shortness of breath, and rhinorrhea. All patients tested by procedure before scheduled admission to the ward were also included. These patients, although not triaged, were still admitted to the ED and tested for SARS-CoV-2 before admission to any ward. For each patient included, the clinical characteristics at the time of admission to the ED and the SARS-CoV-2 viral load detected by one of the molecular tests in use at our laboratory were reported.

Patients were grouped and compared by age (<1 year, ≥1 and <6 years, ≥6 and <10 years, ≥10 years). For each patient, the following variables were collected: age; sex; coexisting conditions; history of previous exposure to SARS-CoV-2; fever (defined as the temperature above 37.8 °C); respiratory symptoms, i.e., cough, shortness of breath, and rhinorrhea; other symptoms, i.e., vomiting, diarrhea, abdominal pain, rushing, asthenia and inappetence, conjunctivitis and, in patients aged 6 years and older, headache, chest pain, anosmia and ageusia; and admission to the ward. Fever and other signs and symptoms were recorded if the child presented them at the time of admission or if the parents reported them up to 48 h before admission to the ED. History of exposure to SARS-CoV-2-infected individuals was recorded only in case of direct contact. More specifically, all patients and/or their parents were asked whether they had had contact with someone in the household or someone with whom they spent a lot of time in close contact who was later found to be infected with SARS-CoV-2. Both comorbidities and other pathological conditions (such as a history of wheezing or recurrent infections) were included in the term “coexisting conditions.” Vaccination status was not recovered because the children were not yet vaccinated at the time of the study period. All variables in the four age groups were compared with the χ2 test.

Median CT values of patients divided dichotomously according to clinically relevant variables were reported and compared. Adopting the CT value as a continuous dependent variable, a linear regression analysis of significant variables was performed to identify the signs and symptoms most associated with SARS-CoV-2 viral load, at all ages and by age group. Finally, we looked for the threshold value of CT for each of the following criteria: fever; respiratory symptoms; fever and respiratory symptoms; and absence of fever and respiratory symptoms, considering patients as a set of groups and grouped into age groups.

SARS-CoV-2 assays

NPSs were collected in 3 mL of the universal transport medium (Copan Italia, Brescia, Italy) and analyzed for molecular detection of SARS-CoV-2. Samples from patients with acute conditions whose results were needed in less than 1 h were considered “urgent” (U) and analyzed with Cepheid GeneXpert Xpress SARS-CoV-2 (Cepheid, Sunnyvale, CA); samples from patients who needed a result within 2–5 h were considered “priority” (P) and analyzed with the Simplexa Covid-19 direct assay on LIAISON® MDX (DiaSorin Molecular, CA); samples from patients whose result was expected within 12 h were considered “standard” (S) and analyzed with the AllplexTM SARS-CoV2 assay (Seegene, Seoul, South Korea).

For the Cepheid GeneXpert Xpress SARS-CoV-2 assay, 300 µL of the sample was transferred into the cartridge and loaded into the GeneXpert Instrument System platform. The target genes are E and N. For Simplexa Covid-19 direct assay, 50 µL of the sample and 50 µL of the reaction mixture were transferred to the Direct Amplification Disc and loaded into the LIAISON® MDX. The target genes are S and ORF1ab. For the Seegene Starlet-Tanbead platform, a total of 300 µL per sample was extracted and eluted with 80 µL of elution buffer. RT-PCR was performed on CFX96 (Bio-Rad Laboratories) with AllplexTM SARS-CoV-2 Assay. For detection, 5 µL of extracted RNA was used in a final volume of 20 µL. Results were analyzed automatically using Seegene software (Seegene SARS-CoV-2 Viewer). The target genes are E, RdRp/S, and N.

An internal control was included in each reaction to check for PCR inhibition. Samples were considered positive when one or more genes were detected. For each sample, the average CT was calculated based on the genes detected.

Statistical analysis

All continuous variables were compared by the Kruskal–Wallis test, since they were mostly non-normally distributed. All categorical variables were compared by the χ2 test or Fisher’s exact test. A p value less than 0.05 was considered statistically significant. When multiple comparisons were performed, a Bonferroni-corrected p value was provided. A linear regression of the relevant variables of the univariate analysis was performed to identify the signs and symptoms most associated with molecular detection of SARS-CoV-2, adopting the CT value as a continuous dependent variable. CT value cut-offs for different combinations of signs and symptoms were derived by receiver-operating characteristic (ROC) curve analysis. All analyses were conducted using IBM SPSS statistical software, version 23.0.

Ethical approval

The study was approved by the local Ethics Committee. Informed consent from the participants’ legal guardian or next of kin was obtained at the time of admission to the pediatric emergency room.

Results

The study included 570 patients admitted to the ED of our hospital and tested positive for SARS-CoV-2 RNA detection. According to clinical conditions, three different molecular workflows were appropriately set up: 262 were tested as “U”, 90 as “P” and 218 as “S”.

Their clinical characteristics are shown in Table 1.

Table 1 Demographic and clinical characteristics of the study sample grouped per SARS-CoV-2 testa.
Full size table

The median age of the subjects was 4.75 years (IQR 1.06–11.32), with most patients aged ≥1 and <6 years. Overall, 308 (54%) were males and 262 (46%) were females, with a female-to-male ratio of 1:1.2. Among pediatric patients admitted to the emergency room, 246/570 (43.2%) had a proven history of previous exposure to a SARS-CoV-2-infected person. A total of 244/570 (42.8%) patients were febrile on admission, 187/570 (32.8%) patients had respiratory symptoms, including cough, shortness of breath, or rhinorrhea. In our population, 238/570 (41.8%) had no fever or respiratory symptoms) and 178/570 (31.2%) patients had mostly other symptoms, among which vomiting and diarrhea were the most frequent. Among the 570 patients, 144 (25.3%) were completely asymptomatic. Of all patients, 153/570 (26.8%) required admission to the ward.

The clinical characteristics of the patients, grouped by age and compared with the χ2 test, are shown in Table 2.

Table 2 Demographic and clinical characteristics of the study sample per age groupa.
Full size table

Children younger than 1 year old (136/570) were significantly more febrile (80/136, 58.8%; p < 0.001) and had more respiratory symptoms (61/136, 44.9%; p < 0.001), especially cough (41/136, 30.1%; p < 0.001) and rhinorrhea (32/136, 23.5%; p < 0.001) than the other subgroups. Children aged ≥1 and <6 years (190/570) had no significant clinical features. Children aged 6–10 years (75/570) were mostly free of fever and respiratory symptoms compared with the other age groups (48/75, 64.0%; p < 0.001), and among other symptoms, abdominal pain was most frequently reported (12/75, 16.0%; p < 0.001). Finally, children aged ≥10 years (169/570) presented more often with headache (19/169, 11.2%; p < 0.001) and coexisting conditions (39/169, 23.1%; p < 0.001).

Comparison of CT values of 10 validation samples on the three platforms showed a slight statistically significant difference (p = 0.023), with a maximum variation between mean CT values of 1.2 (Supplementary Table 1). Considering this variation was not clinically significant, the results of the three assays were analyzed together.

The median CT values of patients according to clinically relevant dichotomous variables in the whole sample and by age group are shown and compared with the Kruskal–Wallis test in Tables 3 and 4.

Table 3 Median CT value for each variable.
Full size table
Table 4 Median CT value for each variable per age group.
Full size table

When analyzing the entire population, all signs and symptoms considered were found to be significantly associated with the change in CT value. Fever, respiratory symptoms, and previous exposure to an individual infected with SARS-CoV-2 were all associated with a lower CT value (with a median CT of about 20). Specifically, children with fever and/or respiratory symptoms showed a median CT of about 20 (fever 20.6, IQR 15.6–33.3; respiratory symptoms 20.2, IQR 16.1–28.4) compared with 34.2 (IQR 23.3–40.3) for asymptomatic patients. This difference in CT values was not evident in all age groups. In more detail: fever was significantly associated with a lower CT value at all ages except in children aged 1–6 years (15.9, 20.2, and 24.2 compared with 20.7, 34.9, and 33.9, in the <1 year, 6–10 years, and >10 years groups, p = 0.005, 0.005, and <0.001, respectively); respiratory symptoms and previous exposure to an individual infected with SARS-CoV-2 were significantly associated with a lower CT value in patients aged 1–6 years and >10 years (respiratory symptoms 25.6 and 21.8 compared with 33.1 and 33.9, p = 0.001 and <0.001, respectively; exposure to SARS-CoV-2 25.4 and 23.3 compared with 34.4 and 32.3, p < 0.001 and <0.001, respectively).

A linear regression analysis was conducted, adopting the CT value as the dependent variable and the clinically or statistically significant variables from the univariate analysis as the independent variables. Ward admission, although found to be slightly significant in previous analyses, was excluded from the regression model because it would be considered a clinical outcome, rather than a clinical predictor, of higher viral load. In addition, in many cases, patients were admitted on the basis of significant comorbidity rather than on the basis of COVID-19 infection alone. The results of linear regression analysis in the whole study sample and for each age group are shown in Tables 5 and 6, respectively.

Table 5 Linear regression analysis (dependent variable: CT value).
Full size table
Table 6 Linear regression analysis per age group (dependent variable: CT value).
Full size table

This analysis confirms that previous exposure to SARS-CoV-2-infected individuals, fever and respiratory symptoms, predicted a lower CT value (corresponding to a higher viral load), while older age, absence of fever and respiratory symptoms, and presence of other symptoms, predicted a higher CT value (corresponding to a lower viral load).

When stratified by age, only exposure history remained a parameter significantly associated with a lower CT value in all age groups except the 1- to 6-year-old group.

Using ROC curve analysis, we explored the cut-offs of the CT value for each of the following criteria: fever; respiratory symptoms; fever and respiratory symptoms; and absence of fever and respiratory symptoms. However, the ROC curves did not produce a reliable shape and area under the curve to define valid cut-offs (data not shown). Therefore, clinically meaningful cut-offs for the CT value could not be produced.

Discussion

Our study sought to associate the signs and symptoms of pediatric COVID-19 with the viral load of SARS-CoV-2 in the respiratory secretions of positive children accessing our ED. Samples were obtained using a NPS and measured by RT-PCR, the gold standard diagnostic test for the detection of SARS-CoV-2. In addition to its high sensitivity, the molecular test generates the CT, a proxy for viral load, which is particularly useful in defining the infectivity of positive patients.13 Although no clinical study has validated the use of CT values to define patient management,25 many efforts are aimed at defining a putative association between CT values and clinical setting.26

The clinical and demographic features of our study are comparable to those reported in the literature,27 with fever and respiratory symptoms dominating the presentation of symptomatic COVID-19 in children and with other nonrespiratory symptoms, such as vomiting, diarrhea, abdominal pain, and headache, accounting for less than 10% of all cases. Slightly less than half of the pediatric COVID-19 cases remain asymptomatic, consistent with the hypothesis of a milder course of the disease in children. Most studies on pediatric COVID-19 have focused on the pediatric population as a whole, without distinguishing between age groups.28,29 In this study, we divided patients according to age to see if age groups could be related to specific manifestations of COVID-19. Indeed, we found significant differences between the age groups in terms of clinical presentation (as shown in Table 2). More specifically, younger children (<6 years) appeared to be more febrile and presented with more cough than older children (>6 years), who in turn presented with more comorbidities and “atypical” nonrespiratory symptoms. The reason for this may be twofold: either the symptoms were more likely to be described by older and more “self-conscious” patients than by younger patients, or the reported symptoms were related to another illness, so the finding of SARS-CoV-2 was only occasional, as the later evidence of low viral load observed in this group seems to suggest.

Comparing the median CT value of the patients according to the single dichotomous variable considered (as shown in Tables 3 and 4), we found significant differences at all ages, suggesting that the presence or absence of specific signs and symptoms, as well as previous exposure to a SARS-CoV-2-infected person, may drive viral load. In more detail, patients with fever and/or respiratory symptoms showed a lower CT value than both patients negative for these parameters (CT ~20 vs. CT ~32; p < 0.001) and asymptomatic patients (CT ~34 vs. CT ~23, p < 0.001). Children younger than 1 year were more febrile than the other age subgroups, and their CT value (CT ~16) was also the lowest among all subgroups. A history of previous exposure to an individual infected with SARS-CoV-2 was also associated with a lower CT value (CT ~22 vs. CT ~34, p < 0.001). The reasons explaining this phenomenon may be related to Italian health policy, which allows easier and free access to treatment, or to an increased awareness of parents, who bring their children to the emergency room in the first phase of the disease, when the viral load is highest. In fact, Yonker et al. showed that children have a higher viral load in the first 5 days of illness.8 Of note, fever, respiratory symptoms, and previous exposure to SARS-CoV-2 were all associated with a lower CT value in children older than 10 years. Overall, patients with coexisting conditions were associated with a higher CT value (CT ~35 vs. CT ~24; p value < 0.001). This could be due to the occasional finding of SARS-CoV-2 positivity in patients who accessed the ED for other reasons or required hospitalization for which SARS-CoV-2 testing is performed as part of the admission protocol.

In the linear regression model (as shown in Tables 5 and 6), advanced age, absence of fever and respiratory symptoms, and presence of other symptoms were negatively associated with viral load; while temperature, respiratory symptoms, and exposure to SARS-CoV-2 were positively associated with it. All these results were significant. In age-specific linear regression analysis, exposure to SARS-CoV-2 was significantly associated with elevated viral load at all ages except in the 1- to 6-year-old group. In children younger than 1 year, fever was the best predictor of higher viral load.

These results strongly suggest that the lack of overt symptomatology in children may be associated with a lower SARS-CoV-2 viral load; in fact, completely asymptomatic patients, especially older ones, show lower CT values. Unfortunately, our analysis did not allow us to define stringent cut-offs for CT values associated with specific signs and symptoms.

The single-center retrospective design of our study prevents us from generalizing our hypotheses to the entire pediatric population. In addition, we did not include a comparison of clinical variables according to the type of test used to determine the CT value. However, validation of CT values among the platforms demonstrated a very low difference between the three, with a maximum mean difference of 1.2, despite a significant p value (p = 0.023), which we considered not clinically significant. The concordance between molecular tests for SARS-CoV-2 is further supported by a recent study by Raju et al. in which strong linearity in CT values is reported between different platforms, including the GeneXpert and Simplexa tests, two of the three tests used in this study.30 Another limitation of this study is that we could not consider the new SARS-CoV-2 variants of interest (VOC), which are currently emerging and could change the pediatric scenario of VOCID-19. During the study period, SARS-CoV-2 wild-type, followed by the UK variant (VOC 202012/01) since approximately February 2021, has been the most prevalent viral strain, while the omicron variant currently accounts for 99.8% of all new cases of COVID-19 in Italy.31

Conclusion

Our study is the first to address the association between SARS-CoV-2 viral load and clinical manifestations of COVID-19 in different pediatric ages. We found that a higher viral load was often associated with fever, SARS-CoV-2 exposure, and respiratory symptoms in children, especially infants, while age, the presence of nonrespiratory symptoms, or the absence of any symptoms were associated with a lower viral load. On the other hand, a lower viral load may be associated with the complete absence of fever and respiratory symptoms in older children. Thus, while a higher viral load may be associated with fever, respiratory symptoms, and recent infection, the lack of the latter is usually indicative of a lower viral load. Further studies are needed to demonstrate the association between SARS-CoV-2 viral load, infectivity, and clinical severity of pediatric COVID-19.