Factors associated with the time to return negative RT-PCR from COVID-19 in paediatric patients: a retrospective cohort study

Objective This study aimed to describe the epidemiological and clinical features and potential factors related to the time to return negative reverse transcriptase (RT)-PCR in discharged paediatric patients with COVID-19. Design Retrospective cohort study. Setting Unscheduled admissions to 12 tertiary hospitals in China. Participants Two hundred and thirty-three clinical charts of paediatric patients with confirmed diagnosis of COVID-19 admitted from 1 January 2020 to 17 April 2020. Primary and secondary outcome measures Primary outcome measures: factors associated with the time to return negative RT-PCR from COVID-19 in paediatric patients. Secondary outcome measures: epidemiological and clinical features and laboratory results in paediatric patients. Results The median age of patients in our cohort was 7.50 (IQR: 2.92–12.17) years, and 133 (57.1%) patients were male. 42 (18.0%) patients were evaluated as asymptomatic, while 162 (69.5%) and 25 (10.7%) patients were classified as mild or moderate, respectively. In Cox regression analysis, longer time to negative RT-PCR was associated with the presence of confirmed infection in family members (HR (95% CI): 0.56 (0.41 to 0.79)). Paediatric patients with emesis symptom had a longer time to return negative (HR (95% CI): 0.33 (0.14 to 0.78)). During hospitalisation, the use of traditional Chinese medicine (TCM) and antiviral drugs at the same time is less conducive to return negative than antiviral drugs alone (HR (95% CI): 0.85 (0.64 to 1.13)). Conclusions The mode of transmission might be a critical factor determining the disease severity of COVID-19. Patients with emesis symptom, complications or confirmed infection in family members may have longer healing time than others. However, there were no significant favourable effects from TCM when the patients have received antiviral treatment.


Limitations
Authors: "This study was limited to a small number of pediatric COVID-19 patients, given that the prevalence of COVID-19 pediatric cases is relatively low compared with adults population". Reviewer: I don't think it's a limitation but a strength of the study. If the prevalence in children is low (thankfully) a retrospective study is ideal. The limit may be that the cases scattered over 12 hospitals in different cities can be very heterogeneous in terms of diagnosis and treatment characteristics, as there are no validated protocols on the subject, given the novelty. We all do what we can do. Please delete this limitation, also from the end of the paper and discuss on heterogeneity of management. Possible limits due to this aspect is the use of Chinese medicine, which makes the data hardly comparable with those of other nations. The worst part of the manuscript is the abstract, to be corrected as suggested for the syntax. The rest of the paper is much better written and interesting. The figures and tables are understandable and in an acceptable number. Therefore by modifying the abstract as suggested the work can be published. The problem of submission to the ethics committee is unclear. For retrospective studies, the Ethics Committee should be notified of the ongoing study. The main finding of this study is that the presence of confirmed patients in the family and patients with gastrointestinal symptoms are associated with longer time to achieve negative PT-PCR. This is an interesting and important epidemiological study, but I would like the authors to clarify several things before further decision.

Response:
We appreciate your generous comments on our manuscript.
Comment 2: Please explain the age range of the patients including in the study. This is critical in pediatric study.
Response: Thank you. As suggested, we have added the age range of the patients included in our study in the corresponding part. Please see page 9 line 52-54.

Comment 3:
The definition of abnormal findings such as myocardial injury or liver insufficiency are not clear. Please describe the definitions in the methods.
Response: Thank you for pointing out it. As suggested, we have added the definitions of myocardial injury and liver insufficiency in the corresponding part of methods. Please see page 7 line 26-31.

Comment 4:
Because the patients of this study are tested by either suspected symptoms or potential exposure, the authors did not capture all (especially asymptomatic) COVID-19 patients in the community. Please describe how this impacts the interpretation of this study.
Response: Thank you for your opinion and we agree with your point. We failed to capture all asymptomatic patients for these two reasons: a) the study was conducted based on hospitals which relied on passive reporting of patients by themselves; b) the patients without symptoms were much less likely to seek healthcare service than those with symptoms unless there was mass screening or became a close contact.
Although this likely causes an underestimation of the true prevalence of asymptomatic SARS-CoV-2 infection, estimation the prevalence isn't the purpose of our research. Besides, the feature of asymptomatic patients is without any clinical symptoms and no evidence exhibits asymptomatic patients need a longer time to return negative RT-PCR than other clinical types at present. Therefore, the failure to capture all asymptomatic patients have limited impact on the interpretation of our study. We have added this as a limitation in the "discussion" section. Please see page 20 line 4-15.
Comment 5: Despite the previous reports by Bielecki et al, I still do not understand why intimate transmission is associated with severe disease. The severity of the disease is determined by virulence and host factors.
Response: Thank you for the question. We agree that severity of the disease is determined by virulence and host factors. We are sorry for the misunderstanding and have corrected the inappropriate descriptions given that our study is unable to directly conclude that intimate transmission is associated with severe disease.
The Bielecki et al. 1 provided evidence that social distancing and hygiene measures had been shown to quantitatively reduce the viral inoculum during infection. Another study also suggested lower intensity of viral exposure might be another factor leading to less severe disease 2 . Besides, given that the SARS-CoV-2 is transmitted through contact, droplets, or aerosols, if there is an infected patient in the family environment and without any quarantine measures, the objects or spaces frequently touched and stayed by this infected patient will remain a high viral load and exposure which could bring greater health risks and threats to family members in closely contact with the space. Therefore, strict non-pharmaceutical measures should be implemented in intimate transmission scenario with confirmed COVID-19 patient, especially in family settings, to reduce the probability of infection or prevent developing to more severe disease. The details of revision are shown in page 19 has also shown that the occurrence of virus infection can affect gastrointestinal function and prone to gastrointestinal symptoms 3-5 . This could explain what we found at a certain degree.
In addition, given that there is no direct causal relationship between the time to turn negative and the severity of the disease, it's reasonable that emesis symptom is present in patients with both mild and severe disease, which is consistent with our data.
We have revised the descriptions of corresponding part. The details are shown in page 19 line 32-59.

patients. I think the authors mostly described my previous questions and comments.
My final comment is about liver insufficiency, which the authors added in line 9 of page 7. 'The inability of the liver to perform its normal synthetic and metabolic functions, etc.' is still vague. Please state specific criteria to define liver dysfunction.
Response: Thank you for pointing out it. As suggested, we have added the specific criteria of liver dysfunction definition in page 7 line 26-31. Response: Thank you for arising this important issue. In fact, the existence of individuals with 0 day between positive and negative result can probably be explained that: This is a hospital-based study.
As we did not have accurate date if a patient was diagnosed before admission to hospital, the starting point of positive diagnosis was therefore defined by the data of admission. We agree that this is a limitation and have clarified in the limitation section of the manuscript. The details showed in page 20 line 4-10.
However, the discharge criteria of each pediatric patient should meet two consecutive negative RT-PCR test results. Usually, the patient who have a positive RT-PCR before admission would have a second test to confirm the diagnosis, which reduce the possibility of false positivity. In addition, related literatures 1 have evaluated the nucleic acid detection methods which are commonly used in clinical practices; and found that the specificity results of these methods are as high as 100%. Therefore, the possibility of false positives is extremely low.
Finally, we did additional analysis based on removing the data with the negative time of 0 (three observations) and the results were showed below (Table R1). There were no substantially changes in the scenario to remove data of 0, which proved the reliability of the original results.