Overall findings of the six-panel respiratory pathogen detection assay
Among the 11,056 collected samples from pediatric patients, 3,228 cases tested positive for at least one of the six common respiratory pathogens, with an overall detection rate of 29.20% (95% confidence interval [CI]: 28.20–30.22%). Among the 3,228 positive cases, 3,078 (95.35%; 95%CI: 94.57–96.05%) were single-pathogen positive, while 150 cases (4.65%; 95%CI: 3.95–5.43%) were mixed-pathogen positive. The detection rates from highest to lowest of each of the six pathogens was in the order RSV, FluA, PIV-3, FluB, ADV, and PIV-1, with detection rates of 13.58% (95%CI: 12.90–14.28%), 4.79% (95%CI: 4.39–5.21%), 4.52% (95%CI: 4.14–4.94%), 3.12% (95%CI: 2.80–3.47%), 2.67% (95%CI: 2.37–2.99%), and 1.90% (95%CI: 1.65–2.17%), respectively. Among the 150 patients with mixed-pathogen infections, 148 showed dual positivity, whereas two showed triple positivity. The top-three combinations of mixed infections were RSV + PIV-3 (32.67%, 49/150), RSV + ADV (20.67%, 31/150), and RSV + FluB (9.33%, 14/150) (Table 1 and Additional file 1).
Table 1
General characteristics of enrolled children.
Characteristics | RSV | FluA | PIV-3 | FluB | ADV | PIV-1 | Total |
Total positive samples | 1501 (13.58) | 529 (4.79) | 500 (4.52) | 345 (3.12) | 295 (2.67) | 210 (1.90) | 3228 (29.20) |
Single infections | 1391 | 512 | 432 | 315 | 239 | 189 | 3078 |
Co-infections | 110 | 17 | 68 | 30 | 56 | 21 | 150 |
Sex | | | | | | | |
Boys (n = 5862) | 817 (13.94) | 297 (5.07) | 264 (4.50) | 191 (3.26) | 169 (2.88) | 113 (1.93) | 1778 (30.33) |
Girls (n = 5194) | 684 (13.17) | 232 (4.47) | 236 (4.54) | 154 (2.96) | 126 (2.43) | 97 (1.87) | 1450 (27.92) |
χ2 | 1.320 | 2.045 | 0.003 | 0.690 | 2.043 | 0.025 | 7.648 |
P | 0.251 | 0.153 | 0.956 | 0.406 | 0.153 | 0.872 | 0.006 |
Age Group (Years) | | | | | | | |
Infant group (age < 1 year, n = 1,204) | 378 (31.39) | 52 (4.32) | 141 (11.71) | 31 (2.57) | 26 (2.16) | 27 (2.24) | 620 (51.50) |
Toddler group (age 1 to < 3 years, n = 2359) | 582 (24.67) | 103 (4.37) | 201 (8.52) | 53 (2.25) | 65 (2.76) | 38 (1.61) | 991 (42.01) |
Preschool group (age 3 to < 7 years, n = 4534) | 495 (10.92) | 221 (4.88) | 137 (3.02) | 136 (3.00) | 135 (2.98) | 128 (2.82) | 1201 (26.49) |
School-age group (age 7 to 17 years, n = 2960) | 46 (1.55) | 153 (5.17) | 21 (0.71) | 125 (4.22) | 69 (2.33) | 17 (0.57) | 416 (14.05) |
χ2 | 965.212 | 2.521 | 354.712 | 19.262 | 4.241 | 50.492 | 821.261 |
P | < 0.001 | 0.472 | < 0.001 | < 0.001 | 0.237 | < 0.001 | < 0.001 |
Notes: Data are presented as n (%) and were analyzed using the chi-square test for categorical variables. Abbreviations: FluA Influenza Virus A, FluB Influenza Virus B, RSV Respiratory Syncytial Virus, PIV Parainfluenza Virus, ADV Adenovirus |
Age and sex characteristics of six respiratory pathogen infections in children
The detection rates of six common respiratory pathogens varied among different age groups in children, with the highest detection rate of 51.50% (95%CI: 47.52–55.71%) in the infant group, followed by 42.01% (95%CI: 39.43–44.71%) in the toddler group, 26.49% (95%CI: 25.02–28.04%) in the preschool group, and 14.05% (95%CI: 12.74–15.47%) in the school-age group. The differences were statistically significant (χ2 = 821.261, P < 0.001).
In the univariate chi-square test, the highest RSV detection rate was found in the infant group at 31.39% (95%CI: 28.31–34.73%), followed by 24.67% (95%CI: 22.71–26.76%) in the toddler group, 10.92% (95%CI: 9.98–11.93%) in the preschool group, and 1.55% (95%CI: 1.14–2.07%) in the school-age group. The differences in positive detection rates among the age groups were statistically significant (χ2 = 965.212, P < 0.001).
The PIV-3 detection rate showed a similar pattern to RSV, with higher detection rates in the infant and toddler groups (11.71%, 95%CI: 9.86–13.81%; 8.52%, 95%CI: 7.38–9.78%), and a significant decrease in detection rate in children above 3 years old. The differences in detection rates among the four age groups were statistically significant (χ2 = 354.712, P < 0.001). PIV-1 was relatively more common in the infant, toddler, and preschool groups, and less common in the school-age group (χ2 = 50.492, P < 0.001).
FluB showed a higher detection rate in school-age children and a lower detection rate in infants and toddlers, with statistically significant differences among the age groups (χ2 = 19.262, P < 0.001). FluA and ADV showed no statistically significant differences in detection rates among different age groups in children (χ2 = 2.521, P = 0.472; χ2 = 4.241, P = 0.237) (Table 1).
Using multivariate logistic regression analysis, we examined the relationship between each pathogen and age. For RSV, the odds ratio (OR) for age was 0.624 (95%CI: 0.603–0.646), indicating that a younger age was associated with a higher risk of RSV infection. For PIV-3, the OR for age was 0.626 (95%CI: 0.591–0.664), also suggesting that younger age is associated with a higher risk of PIV-3 infection. For PIV-1, the OR for age was 0.791 (95%CI: 0.735–0.851), indicating that a younger age was associated with a higher risk of PIV-1 infection. For FluA and ADV, after multivariate logistic regression analysis, the ORs for age were 0.968 (95%CI: 0.938–0.998) and 0.951 (95%CI: 0.911–0.994), respectively, suggesting that younger children were only slightly more at risk of FluA and ADV infections, although the differences were still statistically significant (Wald = 4.3920, P = 0.036; Wald = 4.9672, P = 0.026, respectively). However, in the case of FluB, no significant correlation with age was found by logistic regression analysis, indicating that age does not notably affect the risk of FluB infection.
The overall detection rate in male children was 30.33% (95%CI: 28.94–31.77%), while that in female children was 27.92% (95%CI: 26.50–29.39%). The difference in overall detection rates between the sexes was statistically significant (χ2 = 7.648, P = 0.006). However, no significant difference in the detection rates of individual pathogens were found between males and females in univariate chi-square analysis. Multivariate logistic analysis showed that the OR for sex in relation to FluB was 0.793 (95%CI: 0.629–0.999), indicating that female children were less likely to be affected by FluB, whereas male children were more susceptible to this viral infection.
Comparison of detection rates of six respiratory pathogens in children during 2022–2023
In 2022, of a total of 1,625 cases, 593 cases were positive for at least one of the six common respiratory pathogens, with a detection rate of 36.49% (95%CI: 33.61–39.55%). In 2023, of a total of 9,431 cases, 2,635 cases were positive for at least one of the six common respiratory pathogens, with a detection rate of 27.94% (95%CI: 26.88–29.03%). The overall detection rate in 2022 was statistically significantly higher than that in 2023 (χ2 = 48.634, P < 0.001). Further analysis of the detection rates of individual pathogens between these two years showed that the detection rates of RSV, ADV, and PIV1 were significantly higher in 2022 than in 2023 (χ2 = 31.772, P < 0.001; χ2 = 25.237, P < 0.001; χ2 = 151.887, P < 0.001). The detection rate of influenza B virus was statistically significantly higher in 2023 (3.66%) than in 2022 (0.00%, 0/1625; χ2 = 60.156, P < 0.001). There were no statistically significant differences in the detection rates of influenza A virus and parainfluenza virus type 3 between these two years (Table 2).
Seasonal characteristics of six respiratory pathogen infections in children
In the analysis of the seasonal prevalence characteristics of various pathogens in 2022 and 2023, we found that RSV had a high prevalence during the autumn and winter seasons of 2022, with an extremely low prevalence during the summer season, displaying a typical seasonal pattern. However, in 2023, RSV showed a year-round prevalence, with the highest detection rate in the summer season at 16.18% (95%CI: 14.65–17.83%), followed by 14.61% (95%CI: 13.16–16.18%) in winter, 13.09% (95%CI: 10.85–15.65%) in spring, and 8.97% (95%CI: 8.00–10.03%) in autumn. The seasonal prevalence pattern of RSV differed significantly between 2022 and 2023 (χ2 = 273.807, P < 0.001) (Table 2).
FluA showed a normal prevalence in the winter of 2022, with a relatively low prevalence in other seasons. However, in 2023, FluA had a high prevalence in the spring, at 13.20% (95%CI: 10.95–15.77%), followed by 6.77% (95%CI: 5.79–7.86%) in the winter, and the lowest prevalence, 0.28% (95%CI: 0.11–0.57%), in the summer. The seasonal prevalence pattern of FluA differed significantly between 2022 and 2023 (χ2 = 33.117, P < 0.001).
FluB was not detected in 2022, but in 2023, a sudden outbreak of this virus occurred during the winter season, with a prevalence of 13.42% (95%CI: 12.03–14.93%), but with an extremely low prevalence in the other three seasons. The seasonal prevalence of FluB differed significantly between 2022 and 2023 (χ2 = 61.360, P < 0.001).
PIV-3 was prevalent during the summer and autumn seasons of 2022, but in 2023, its prevalence was 10.08% (95%CI: 8.87–11.39%) during only the summer season, with a low prevalence in the other three seasons. The seasonal prevalence pattern of PIV-3 differed significantly between 2022 and 2023 (χ2 = 126.978, P < 0.001).
PIV-1 was prevalent during the autumn seasons of 2022 and 2023: 9.47% (95%CI: 7.52–11.78%) and 2.13% (95%CI: 1.68–2.68%), respectively. The seasonal prevalence pattern of PIV-1 differed significantly between these years (χ2 = 191.810, P < 0.001).
ADV had a relatively high prevalence during the autumn and winter seasons of 2022, but in 2023, it had a high prevalence during the winter season, at 4.71% (95%CI: 3.90–5.64%), whereas its prevalence was low in the other three seasons. The seasonal prevalence pattern of ADV differed significantly between the two years (χ2 = 90.288, P < 0.001) (Additional file 2).
To investigate the relationship between infection by various pathogens and seasonal prevalence characteristics, a binary multiple logistic regression analysis was conducted using data from 9,431 pediatric patients in 2023. Respiratory pathogens and seasons were included as independent variables in the analysis. The results of multiple logistic regression analysis showed that the OR for RSV infection in the summer and winter seasons were 1.410 (95%CI: 1.206–1.648) and 1.943 (95%CI: 1.646–2.295), respectively. Thus, RSV exhibited a bimodal seasonal pattern, with peaks in the summer and winter of 2023. For FluA virus, the ORs for seasonal infection were 0.019 (95%CI: 0.009–0.041) in summer, 0.307 (95%CI: 0.238–0.396) in autumn, and 0.611 (95%CI: 0.473–0.789) in winter. Compared with the peak of infections in the spring season, the other three seasons were considered low-risk fact periods. For FluB, the OR for winter infection was 199.145 (95%CI: 88.658–447.322), indicating a high risk of outbreaks. For human PIV-3, the ORs for the summer and autumn seasons were 6.954 (95%CI: 5.121–9.443) and 2.594 (95%CI: 1.849–3.640), respectively, with the highest risk of PIV-3 infection occurring in the summer season. Human PIV-1 predominantly circulated during autumn, with a high OR of 5.395 (95%CI: 3.159–9.216). ADV had a high prevalence in the winter season of 2023, with an OR of 3.871 (95%CI: 2.937–5.103).
Table 2
Detection of respiratory viruses among different years and seasons.
Characteristics | RSV | FluA | PIV-3 | FluB | ADV | PIV-1 | Total |
Year | | | | | | | |
2022 | 293 (18.03) | 74 (4.55) | 86 (5.29) | 0 (0.00) | 74 (4.55) | 94 (5.78) | 593 (36.49) |
2023 | 1208 (12.81) | 455 (4.82) | 414 (4.39) | 345 (3.66) | 221 (2.34) | 116 (1.23) | 2635 (27.94) |
χ2 | 31.772 | 0.167 | 2.410 | 60.156 | 25.237 | 151.887 | 48.634 |
P | < 0.001 | 0.682 | 0.121 | < 0.001 | < 0.001 | < 0.001 | < 0.001 |
Seasons | | | | | | | |
Spring (n = 917) | 120 (13.09) | 121 (13.20) | 19 (2.07) | 2 (0.22) | 15 (1.64) | 1 (0.11) | 272 (29.66) |
Summer (n = 2,521) | 408 (16.18) | 7 (0.28) | 254 (10.08) | 0 (0.00) | 35 (1.39) | 16 (0.63) | 682 (27.05) |
Autumn (n = 3,467) | 311 (8.97) | 156 (4.50) | 107 (3.09) | 4 (0.12) | 52 (1.50) | 74 (2.13) | 682 (19.67) |
Winter (n = 2,526) | 369 (14.61) | 171 (6.77) | 34 (1.35) | 339 (13.42) | 119 (4.71) | 25 (0.99) | 999 (39.55) |
χ2 | 78.842 | 275.046 | 275.700 | 933.046 | 84.710 | 41.383 | 289.151 |
P | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 |
Notes: We exclusively examined the seasonal characteristics of viruses in 2023. Data are presented as n (%) and were analyzed using the chi-square test for categorical variables. |
Diagnostic value of the six-panel respiratory pathogen detection assay for LRTIs and severe pneumonia
In total, 9,431 children were enrolled as study participants for 2023. As shown in Additional File 3, the positivity rate for the six pathogens in patients with URTIs (32.62%, 95%CI: 28.67–36.97%) was significantly higher than that observed in patients with LRTIs (27.53%, 95%CI: 26.44–28.66%; P < 0.001). We designated one as the outcome for LRTIs and 0 as the outcome for URTIs, considering them as the dependent variables. Additionally, we considered the six pathogens, age, sex, and season of onset as independent variables for conducting binary logistic regression analysis. As shown in Fig. 1, RSV (OR 9.760, 95%CI: 6.280–15.169, P < 0.001) and PIV-3 (OR 1.842, 95%CI: 1.235–2.748, P = 0.003) were identified as significant risk factors for LRTIs, while FluA (OR 0.342, 95%CI: 0.259–0.451, P < 0.001), FluB (OR 0.415, 95%CI: 0.281–0.611, P < 0.001), and ADV (OR 0.248, 95%CI: 0.163–0.346, P < 0.001) were more likely to cause URTIs. In 2023, compared to spring, the likelihood of LRTIs was higher during summer (OR 3.974, 95%CI: 3.174–4.976, P < 0.001), autumn (OR 9.269, 95%CI: 7.240–11.866, P < 0.001), and winter (OR 5.607, 95%CI: 4.388–7.165, P < 0.001).
An older age was found to be associated with a higher risk of developing LRTIs (OR 1.106, 95% CI: 1.015–1.205, P = 0.022). Girls were more likely to experience such infections (OR 1.254, 95%CI: 1.061–1.481, P = 0.008) (Fig. 1). Notably, PIV-1 did not show any differences in its association with LRTIs or URTIs.
We enrolled 8,680 pediatric patients clinically diagnosed with LRTIs, such as pneumonia, in 2023 and categorized them into two groups based on the severity of pneumonia: severe pneumonia and non-severe pneumonia. The overall positivity rate for the six pathogens in the severe pneumonia group was 25.23% (56/222), whereas that in the non-severe pneumonia group was 27.60% (2,334/8,458). The positivity rate for RSV in the severe pneumonia group was higher than that in the non-severe pneumonia group (χ2 = 4.051, P = 0.044). Single-factor analysis revealed that RSV infection was significantly associated with the development of severe pneumonia (OR 1.447, 95%CI: 1.025–2.042) (Additional File 3).
Binary logistic regression analysis was conducted to determine the factors associated with severe pneumonia. In this study, patients with severe pneumonia were defined as 1, and those with non-severe pneumonia, as 0. The independent variables included the six pathogens, along with age, sex, and season. PIV-3 infection was associated with non-severe pneumonia (OR 0.386, 95%CI: 0.168–0.886, P = 0.025). The incidence of severe pneumonia was influenced by both season and age, with a higher occurrence observed during summer (OR 2.036, 95%CI: 1.548–2.679, P < 0.001). Additionally, younger age groups exhibited greater susceptibility to this condition (OR 0.935, 95%CI: 0.893–0.979, P = 0.004) (Fig. 2). After adjusting for seasonal and age-related factors, multivariate regression analysis revealed no statistically significant difference in RSV infection between non-severe and severe pneumonia.
Evaluation of the diagnostic performance of the six-panel respiratory pathogen panel using tNGS
We randomly selected 138 patients from 8,680 clinical cases diagnosed with LRTI in 2023. Among them, 77 patients tested positive in the respiratory panel test (including 12 cases of mixed infections), while 61 tested negative. These cases were verified using tNGS of respiratory pathogens. The agreement between the multiplex respiratory panel test and tNGS overall and for each of the viruses are shown in Table 3.
Table 3
Evaluation of the diagnostic performance of multiplex RT-qPCR for six respiratory viruses
Virus | Se (%) [95%CI] | Sp (%) [95%CI] | PPV (%) [95%CI] | NPV (%) [95%CI] | Kappa |
RSV | 93.55[78.58–99.21] | 100[96.61–100.00] | 100[88.06–100.00] | 98.17[93.54–99.78] | 0.957 |
FluA | 94.12[71.31–99.85] | 99.17[95.48–99.98] | 94.12[71.31–99.85] | 99.17[95.48–99.98] | 0.933 |
PIV-3 | 100[86.77–100.00] | 100[96.76–100.00] | 100[86.77–100.00] | 100[96.76–100.00] | 1.000 |
FluB | 75.00[19.41–99.37] | 100[97.28–100.00] | 100[29.24–100.00] | 99.26[95.94–99.98] | 0.854 |
ADV | 88.89[51.75–99.72] | 100[97.18–100.00] | 100[63.06–100.00] | 99.23[95.79–99.98] | 0.937 |
PIV-1 | 100[54.07–100.00] | 100[97.24–100.00] | 100[54.07–100.00] | 100[97.24–100.00] | 1.000 |
Total | 94.87[87.38–98.58] | 98.33[91.05–99.96] | 98.67[92.80–99.97] | 93.65[84.53–98.24] | 0.927 |
Abbreviations: Se, sensitivity; Sp, specifitiy; PPV, positive-predictive value; NPV, negative-predictive value; CI, confidence interval |