4 Discussion
In the current study, the length of hospital stay in most pediatric patients with COVID-19 (average 13 days) was not significantly different from pneumonia caused by other pathogens (average 11·2 days)17. Interestingly, the number of pediatric patients younger than 6 months (11, 10%) was significantly less than that in other age groups, which may be related to residual protection from maternal immune factors. The number of patients in other two age groups (6 Months to 6 Ys, 47, 41%; >6Ys, 56, 49%) was not significantly different from each other. Similar to results reported in previous studies, according to the lung involvement, the majority of cases were mild cases and only conventional therapy for viral pneumonia was needed12,13,18. Severe and fatal cases in pediatric patients were very rare. Thus, early diagnosis to avoid further spread of disease would be much more important than treatment for individuals.
In the current study, chest CT had similar characteristic manifestations including ground-glass opacities and consolidations with bilateral inferior lobes, as was recently published13,19. Of all 114 COVID-19 positive children in this study, 32 (28%) had completely normal chest CT, 63 (55%) had no more than 2 lung lobe lesions, and 34 patients had concurrent infections (30%). Similar to previous studies, CT of the chest is often atypical, especially in the early stage, resulting in difficulty in diagnosing or ruling out COVID-19 20. In addition, in our study chest CT identified only 53 (46%) cases by day 7 of the onset of symptoms, and by day 14, only 76 (67%) cases of chest CT were positive. This is totally different from the results of related research in adults. In adults, as an important complementary tool for low-sensitive and time-consuming RT-PCR, chest CT has first been considered as diagnostic tool for clinical confirmed case of COVID-19 in China, where the epidemic is most severe21,22. Considering that the clinical manifestations and CT features of most children are mild, CT has limited diagnostic value for children (especially 0-7 days after onset). Therefore, CT of the chest is of limited value in early stages of the disease, and should be discouraged to reduce the radiation dose to children.
In the current study, throughout the entire course of COVID-19 pneumonia in children, the diagnostic positive rate of CT has been far lower than that of RT-PCR (All P <·05), and 28% of children have no obvious abnormal signs of CT. However, in adults, previous studies have shown that the positive rate of RT-PCR is only 59% to 61·3%, while the positive rate of chest CT is 88%4,5. These findings suggest that most infected pediatric patients have less lung involvement in the early stages of the COVID-19 infection. As some COVID-19 cases confirmed by RT-PCR could have no lesions on chest CT, pathogen identification by RT-PCR have more important role in management of infectious source. RT-PCR maybe is more reliable than CT in pediatric patients’ diagnosis, and repeated RT-PCR is the recommended screening for pediatric patients during the first 7 days.
In addition, the main indicator of children’s recovery is to determine the true RT-PCR negative examination results. For well-known reasons, we may inevitably get false negative RT-PCR results in treatment responds evaluation. Our research shows that only RT-PCR results obtained 7 days after the onset of symptoms are reliable. This indicates that for children who have disappeared clinical symptoms and may totally recover, the follow-up tests of RT-PCR to evaluate the efficacy must be performed at least after day 7.
To reduce radiation dose of pediatric patients included in this study, most cases (85%) had 1 or 2 chest CT scans during their hospital stay. In our study, it was found that the median values of CT scores were low before the 11th day from the onset of symptoms, with the median value of 1. The median value of CT score on lung involvement reached a peak at 12-17 days, with the median value of 2. After the 24th day from onset of symptoms, the median value of CT score fell back to 1. Comparing to median CT score of 5 on lung involvement in adults reported in previous study 16, it indicates that children with COVID-19 pneumonia are relatively mild. As reported, lung involvement peaking on 6 to 11 days from symptom onset in adults, delayed peaking in children may be related to different immune reaction to virus 16. In our study, CT score reached 14 in 1 case during the 0-5 days from the onset of symptoms, which suggested that rapid progression could also be observed in pediatric patients, even if it was rare.
There are several limitations in our study. First, even if the sample was the largest as we known, the overall cases in the two included hospitals was still limited. Second, as a retrospective study, selection bias could not be avoided. Third, during this outbreak period of COVID-19, delay in seeking care (more than 7 days from symptom onset) would influence the diagnosis and prognosis.
In conclusion, chest CT is not recommended as a primary method for early diagnosis in children with COVID-19, especially to avoid repeated CT scans, while RT-PCR may have a more valuable position. For treatment responds, reliable RT-PCR follow-up results are not available until at least 7 days after the onset of symptoms. Only if necessary, CT can be employed as a tool to assess lung involvement.