[Skip to Navigation]
Sign In
Research Letter
July 30, 2020

Age-Related Differences in Nasopharyngeal Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Levels in Patients With Mild to Moderate Coronavirus Disease 2019 (COVID-19)

Author Affiliations
  • 1Division of Infectious Diseases, Department of Pediatrics, Ann & Robert H. Lurie Children’s Hospital, Chicago, Illinois
  • 2Northwestern University Feinberg School of Medicine, Chicago, Illinois
JAMA Pediatr. 2020;174(9):902-903. doi:10.1001/jamapediatrics.2020.3651

Children are susceptible to infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) but generally present with mild symptoms compared with adults.1 Children drive spread of respiratory and gastrointestinal illnesses in the population,2 but data on children as sources of SARS-CoV-2 spread are sparse.

Early reports did not find strong evidence of children as major contributors to SARS-CoV-2 spread,3 but school closures early in pandemic responses thwarted larger-scale investigations of schools as a source of community transmission. As public health systems look to reopen schools and day cares, understanding transmission potential in children will be important to guide public health measures. Here, we report that replication of SARS-CoV-2 in older children leads to similar levels of viral nucleic acid as adults, but significantly greater amounts of viral nucleic acid are detected in children younger than 5 years.


Between March 23 and April 27, 2020, we performed SARS-CoV-2 reverse transcriptase–polymerase chain reaction (PCR) on nasopharyngeal swabs collected at various inpatient, outpatient, emergency department, and drive-through testing sites at a pediatric tertiary medical center in Chicago, Illinois. The Ann & Robert H. Lurie Children’s Hospital of Chicago Institutional Review Board provided an exemption and full waiver of HIPAA authorization and informed consent. A Clinical Laboratory Improvement Amendments–certified laboratory analyzed samples using a US Food and Drug Administration Emergency Use Authorization PCR assay (Abbot RealTime SARS-CoV-2 Assay performed on the m2000 RealTime System [Abbott Laboratories]). PCR amplification cycle threshold (CT) values were recorded, with lower values indicating higher amounts of viral nucleic acid.

This cohort included all individuals aged younger than 1 month to 65 years who tested positive for SARS-CoV-2. Patients with symptoms suggestive of a COVID-19–compatible illness and/or high-risk exposures were tested. We included the first sample tested for patients with multiple samples. Because patients with severe infection have lower CT values,4 we excluded 7 children who required supplemental oxygen support. We also excluded 7 asymptomatic patients, 29 patients with unknown duration of symptoms, and 19 patients whose symptoms started more than 1 week prior to testing. Swabs were collected using a standard bilateral nasopharyngeal sampling procedure. Several controls, including samples with known copy numbers, were included in each PCR run. Median and interquartile ranges for each group were measured and compared using the nonparametric Wilcoxon rank sum test. Two-sided P values less than .05 were considered statistically significant. Analyses were performed using Stata/IC statistical software version 16.0 (StataCorp).


Our final cohort included 145 patients with mild to moderate illness within 1 week of symptom onset. We compared 3 groups: young children younger than 5 years (n = 46), older children aged 5 to 17 years (n = 51), and adults aged 18 to 65 years (n = 48). We found similar median (interquartile range) CT values for older children (11.1 [6.3-15.7]) and adults (11.0 [6.9-17.5]). However, young children had significantly lower median (interquartile range) CT values (6.5 [4.8-12.0]), indicating that young children have equivalent or more viral nucleic acid in their upper respiratory tract compared with older children and adults (Figure). The observed differences in median CT values between young children and adults approximate a 10-fold to 100-fold greater amount of SARS-CoV-2 in the upper respiratory tract of young children. We performed a sensitivity analysis and observed a similar statistical difference between groups when including those with unknown symptom duration. Additionally, we identified only a very weak correlation between symptom duration and CT in the overall cohort (Spearman ρ = 0.22) and in each subgroup (young children, Spearman ρ = 0.20; older children, Spearman ρ = 0.19; and adults, Spearman ρ = 0.10).

Figure.  Distribution of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Reverse Transcriptase–Polymerase Chain Reaction (RT-PCR) Amplification Cycle Threshold (CT) Values From Nasopharyngeal Swabs Collected From Patients With Coronavirus Disease 2019
Distribution of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Reverse Transcriptase–Polymerase Chain Reaction (RT-PCR) Amplification Cycle Threshold (CT) Values From Nasopharyngeal Swabs Collected From Patients With Coronavirus Disease 2019

Children younger than 5 years had significantly lower CT values compared with children aged 5 to 17 years (P = .02) and adults 18 years and older (P = .001). CT values were similar between children aged 5 to 17 years and adults 18 years and older (P = .34). Midlines indicate the median, boxes indicate interquartile ranges, whiskers indicate the upper and lower adjacent values (within 1.5-fold the interquartile range), and isolated data points indicate outliers.


Our analyses suggest children younger than 5 years with mild to moderate COVID-19 have high amounts of SARS-CoV-2 viral RNA in their nasopharynx compared with older children and adults. Our study is limited to detection of viral nucleic acid, rather than infectious virus, although SARS-CoV-2 pediatric studies reported a correlation between higher nucleic acid levels and the ability to culture infectious virus.5 Thus, young children can potentially be important drivers of SARS-CoV-2 spread in the general population, as has been demonstrated with respiratory syncytial virus, where children with high viral loads are more likely to transmit.6 Behavioral habits of young children and close quarters in school and day care settings raise concern for SARS-CoV-2 amplification in this population as public health restrictions are eased. In addition to public health implications, this population will be important for targeting immunization efforts as SARS-CoV-2 vaccines become available.

Back to top
Article Information

Accepted for Publication: May 31, 2020.

Corresponding Author: Taylor Heald-Sargent, MD, PhD, Division of Infectious Diseases, Department of Pediatrics, Ann & Robert H. Lurie Children’s Hospital, 225 E Chicago Ave, Box 20, Chicago, IL 60611 (thsargent@luriechildrens.org).

Published Online: July 30, 2020. doi:10.1001/jamapediatrics.2020.3651

Author Contributions: Drs Heald-Sargent and Kociolek had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Heald-Sargent, Muller, Kociolek.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Heald-Sargent, Muller, Kociolek.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Heald-Sargent, Kociolek.

Administrative, technical, or material support: All authors.

Study supervision: Muller, Patel, Kociolek.

Conflict of Interest Disclosures: Dr Muller has received grants from Ansun BioPharma, Astellas Pharma, AstraZeneca, Abbott Laboratories, Janssen Pharmaceuticals, Karius, Merck, Melinta Therapeutics, Roche, Tetraphase Pharmaceuticals, and Seqirus. Dr Patel has received grants from Aqua Pharmaceuticals. Dr Kociolek has received grants from Merck. No other disclosures were reported.

Funding/Support: Drs Heald-Sargent and Kociolek were supported by grants T32AI095207 and K23AI123525, respectively, from the National Institute of Allergy and Infectious Diseases.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Lu  X, Zhang  L, Du  H,  et al; Chinese Pediatric Novel Coronavirus Study Team.  SARS-CoV-2 infection in children.   N Engl J Med. 2020;382(17):1663-1665. doi:10.1056/NEJMc2005073PubMedGoogle ScholarCrossref
Paules  C, Subbarao  K.  Influenza.   Lancet. 2017;390(10095):697-708. doi:10.1016/S0140-6736(17)30129-0PubMedGoogle ScholarCrossref
Cao  Q, Chen  YC, Chen  CL, Chiu  CH.  SARS-CoV-2 infection in children: transmission dynamics and clinical characteristics.   J Formos Med Assoc. 2020;119(3):670-673. doi:10.1016/j.jfma.2020.02.009PubMedGoogle ScholarCrossref
Liu  Y, Yan  LM, Wan  L,  et al.  Viral dynamics in mild and severe cases of COVID-19.   Lancet Infect Dis. 2020;20(6):656-657. doi:10.1016/S1473-3099(20)30232-2PubMedGoogle ScholarCrossref
L’Huillier  AG, Torriani  G, Pigny  F, Kaiser  L, Eckerle  I.  Culture-competent SARS-CoV-2 in nasopharynx of symptomatic neonates, children, and adolescents.   Emerg Infect Dis. 2020;26(10):26.PubMedGoogle ScholarCrossref
Moreira  LP, Watanabe  ASA, Camargo  CN, Melchior  TB, Granato  C, Bellei  N.  Respiratory syncytial virus evaluation among asymptomatic and symptomatic subjects in a university hospital in Sao Paulo, Brazil, in the period of 2009-2013.   Influenza Other Respir Viruses. 2018;12(3):326-330. doi:10.1111/irv.12518PubMedGoogle ScholarCrossref
11 Comments for this article
Age/CT relationship between 5-17 y.o.
S Eraly, MD, PhD |
Thank you and your group for your important contribution to our understanding of Covid-19 epidemiology! Recognizing that the analysis might be limited by the relatively small sample size, can you comment on the relationship between age and PCR amplification cycle threshold (CT) within the age range 5 to 17 y.o.?

It does not seem that the Age/CT relationship could be monotonic in this range, since one would in that case expect the median CT to be intermediate between those of younger children and adults. Is the lesser CT that would be expected for those at
the young end of the 5-17 range counterbalanced by greater CT in the rest of this group?

The most inclusive manner in which to visualize the Age/CT relationship could be to present a scatterplot of Age vs CT (both as continuous variables). Such a figure has doubtless already been generated by the research team – could it be provided here in the Comments section?
How Was the Adult Population Chosen?
Marc Hainaut, MD, PhD | CHU Saint-Pierre, Université Libre de Bruxelles, Brussels, Belgium
Thank you for this contribution to our understanding of SARS-CoV-2 infection in children. Such data are really important even if we know the transmission of a communicable disease in a group does not depend solely on the number of viruses present in the secretions (especially when counted by PCR).

I understand the format of a « Research Letter » does not allow to put many details but the analyzed population remains unclear to me.

In the methods section it is written "all individuals aged younger than 1 month to 65 years who tested positive for SARS-CoV-2" were included.
As the study was conducted at a pediatric tertiary medical center in Chicago, where did the adult patients in the study come from? How were they chosen? Were they the staff members who contracted the disease?

In all published series of SARS-CoV-2 infection, from all locations, children are significantly fewer in number than adults and have a less severe disease. How to explain that in your series no adult had a severe disease when 7 children did?

Is it possible that the adult population of the study is not comparable to a randomly selected adult population?
Understanding the selection of the population is critical to the analysis of these data.

We would also have liked to know if the difference between the children aged <5 years and 5-17 years is not explained by higher viral load only in very young infants.

It is unfortunate that data of such importance are not presented in the form of a detailed article, especially in view of the suggestion that young children could be a very important source for the spread of the epidemic, which is contrary to the majority of the data published to date.
Author Reply: Age vs PCR Amplification Cycle Threshold (CT)
Larry Kociolek, MD | Ann & Robert H. Lurie Children's Hospital of Chicago
Thank you for the comments. The sample size did not permit a robust analysis of CT by age for each pediatric age group, but we did assess age vs CT in the 5-17 year age group, and there was no relationship. R2 of that plot was 0.002. We chose to dichotomize our pediatric groups based on ages of school attendance.

Similar to the pediatric groups, all of the adults tested had symptoms for fewer than 7 days and none had severe disease. While the adult population in our study may not be generalizable, that does not
impact the findings of young children having higher viral loads than older children. Further, other studies have shown equivalent viral loads between children and adults, and our data are consistent when comparing older children and adults.
CONFLICT OF INTEREST: Senior Author of Paper
Pediatric Incidence Rates
Jason Bhardwaj, BSE, MBA | None
If the general incidence rate of mild to moderate disease is lower in the younger pediatric age cohort, is it possible that an alternate conclusion from the research is that greater viral levels are required to achieve the same severity of symptoms in the younger age cohort?
Aimee Renaud |
This research letter underemphasizes the fact that this analysis relies on not just a small sample but also one primarily obtained from a hospital setting as well as from mild to moderate patients with no additional weighting for the incredibly small incidence of this population within the under-10 population. This study has been interpreted as young people are more likely to carry higher viral load and that in no way is the outcome of this dataset. Please include additional data to quantify the actual population or caveat the limitations of this analysis.
Quantitative Assay?
Mark Vieyra | University of Pennsylvania
The Abbot RealTime SARS-CoV-2 Assay that used in this paper says in its package insert that it is "intended for the qualitative detection of nucleic acid from SARS-CoV-2". I understand that one would expect the CT values to be lower with higher viral loads but the assay is not really designed to draw quantitative conclusions. I get that this is potentially interesting preliminary data but shouldn't this be confirmed with an assay that is designed to quantify viral RNA before we draw the conclusion that "young children can potentially be important drivers of SARS-CoV-2 spread in the general population"?
qPCR Not Performed State of the Art
C X, PhD |
Your data seem plausible, however the qPCR data are not performed how they should.

- The Abbott test is only validated for qualitative purposes. A quantitative validation with a standard curve should have been included. Furthermore, this is a multiplex, which makes quantitation even harder. Cycle quantification values (Cqs) from different targets do influence each other.
- Getting quantitative data from such early PCR amplification cycle thresholds (Cts) is hard. Did you validate your linear range with a standard curve? I mean, at a Cq = 5 f.e. it gets really hard to correctly subtract your baseline. /> - Thirdly, Cq is not a quantitative measure. If you took a calibrator sample, you should use the values or you should work with (d)dCq taking into account efficiencies from your standard curve.
- An analysis with normalized qPCR data would be more correct. Normalization for volume/internal control is a minimum, as extraction can be a variable factor. Even better is a human control, to assess for RT efficiency and for swabbing efficiency.

In summary, you cannot take a quantitative assay and use in in a qualitative way.

Kind Regards
Substantial Technical Issues
Leor Weinberger, Ph.D. | University of California, San Francisco
I agree with the critiques above; there are substantial technical issues with this report. The lack of a standard curve is a critical omission that must be corrected before conclusions can be drawn. This is standard practice and it is surprising that the reviewers did not ask for it.

Without a standard curve, the PCR amplification cycle threshold (CT) values (median CT = 6) reported yield nonsensical viral loads: the Abbott limit of detection is 100 copies/ml carrier fluid with a cutoff CT = 39; so the viral load for a CT=6 is >100*2^(33)=8.6*10^11 copies/mL. Such viral levels
are rarely achieved in highly idealized lab settings after concentration/ultracentrifugation (most published papers are reporting CT values > 20 for SARS-CoV-2). I also agree that the internal control to a human reference gene is needed to correct for sampling differences.

I'd strongly caution against making conclusions until the above data are provided.
RE: Age-Related Differences in Nasopharyngeal SARS-CoV-2 Levels in Patients With Mild to Moderate COVID-19
Tomoyuki Kawada, MD | Nippon Medical School
I appreciate the authors' speculation on COVID-19 sero-epidemiology (1). Although the number of samples was limited, an appropriate statistical method was applied for the analysis.

Dr. Eraly recommended presenting the relationship (scattergram) between age and PCR amplification cycle threshold (CT) with range of age under 17 years old. Regarding this recommendation, the authors replied that there was no significant relationship in the 5-17 years age group with square value of regression coefficient of 0.002. I think that many factors would contribute to CT values, and age might be one of the contributing factors. Although the authors emphasized that children
< 5 years had specific biological characteristics of keeping higher amount of SARS-CoV-2 in the upper respiratory tract, there needs to be more evidence confirming their medical hypothesis. Anyway, presenting scattergram between age and CT values in children under 17 years old would present information regardless of statistical significance.

1. Heald-Sargent T, Muller WJ, Zheng X, Rippe J, Patel AB, Kociolek LK. Age-Related Differences in Nasopharyngeal Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Levels in Patients With Mild to Moderate Coronavirus Disease 2019 (COVID-19). JAMA Pediatr. 2020 doi:10.1001/jamapediatrics.2020.3651
Author Reply: Methodologic Questions
William Muller, MD, PhD | Ann & Robert H Lurie Children's Hospital of Chicago; Northwestern University Feinberg School of Medicine
We very much appreciate the attention this paper has received and the comments left by different readers.

Several readers have left comments on technical characteristics of the assay used to generate the reported data. While it is correct that the clinical application of the assay is for qualitative detection of SARS-CoV-2 RNA, the cycle threshold data reported in this study were gathered for research purposes. There are abundant data in the literature to support the use of cycle threshold values as a proxy for the level of viral RNA in a sample, including for SARS-CoV-2 (1-4). Indeed, the commenters
questioning the assay generally agree that cycle threshold values should correlate with the level of viral RNA in the samples. Although as many readers have noted there are caveats to doing so, some clinicians have argued that cycle threshold data could be applied in medical decision making based on correlations between these values and severity of COVID-19 disease in adults (4).

The Abbott assay used in our clinical laboratory includes an internal control (non-viral) sequence with every sample. Additionally, every clinical run included a control sample with 1000 copies of SARS-CoV-2 RNA per well. There are known differences between the cycle threshold generated with the Abbott assay and other platforms used for clinical detection of SARS-CoV-2 RNA that lead to significantly lower cycle thresholds for the assay we used (5), which do not alter either the results of our study or their interpretation. Using a rough rule-of-thumb that a one logarithm difference in template RNA corresponds to 3.3 cycle thresholds, the reported median CT for the youngest age group of about 6 cycles would correspond to about 10^8 copies/mL, not 10^11 as one reader suggests. Such levels are not at all dissimilar to those reported from upper respiratory tract samples for other viruses (6).


1. Zou N Engl J Med 2020, 382(12): 1177
2. Kam et al Clin Infect Dis 2020, 71(15): 847
3. Lee et al. JAMA Intern Med doi:10.1001/jamainternmed.2020.38622020
4. Magleby et al., Clin Infect Dis 2020; ciaa851
5. Degli-Angeli et al., J Clin Virol 2020; 129: 104474
6. Granados et al., J Clin Virol 2017; 86: 14
The Three Cohorts
Gregory Steinke, MD, MPH | Erlanger Health System
Thank you for your article.

You chose to divide your cohort at age 5 and age 18 which makes sense when comparing young children, school age children, and adults. I would like to know more about how you chose your participants. Why did you have a relatively even number of participants in each cohort? Was this coincidence or was this by design? Why did you choose to use a variety of settings to gather the participants? How many were from each setting? Was there a difference between the level of illness within the mild to moderate window between cohorts?
For example, were the young children more likely to be in the moderate illness category while the older children and adults were more likely to be in the mild category? How did you define level of illness? Did you use the CDC criteria?

I realize this is a short research letter. If a more extensive report is published it would be helpful.

Thank you for your research.