Clinical features
Two hundred and seven children were finally included in our study. According to the inclusion criteria of this study, there were 58 included children requiring IMV, while 149 included children not requiring IMV. Thus, there were 58 cases in IMV group and 149 cases in control group. The flow chart of including patients is shown in Fig. 1.
All included children were admitted to the general wards and no one had severe extrapulmonary complications such as meningitis, pericarditis, endocarditis, osteomyelitis, arthritis, sepsis, hemolytic uremia syndrome, etc. Children were younger in IMV group compared with control group(p = 0.002). There were more children complicated by atelectasis in IMV group than control group (p = 0.02). Hemoglobin (Hb) (p = 0.006) and level of albumin(ALB) (p < 0.001) were significantly lower in IMV group than control group, while creatine kinase MB isoenzyme(CK-MB) was significantly higher in IMV group than control group(p = 0.02). Score of PRISM III was significant higher in IMV group than control group(p = 0.005), while children in IMV group had lower score of PCIS than those in control group(p < 0.001).
PRISM III, PEWS and PCIS had an AUC of the probability for IMV of 0.69 (95% CI, 0.62–0.77), 0.53 (95% CI, 0.44–0.62) and 0.67 (95% CI, 0.58–0.76), respectively(Figure S1). The descriptive statistics of all included children are presented in Table 1. There were 145 children finally divided into training set and 62 into validating set after random sampling. Summarized information in sub-sets is shown in Table S1.
Table 1
The clinical characteristics between IMV and control groups
|
Total(n = 207)
|
IMV group (n = 58)
|
Control group (n = 149)
|
p
|
Demographic characteristics
|
Age(months)
|
12.0(7.0–24.0)
|
8.5(6.0–16.0)
|
14.0(8.0–26.0)
|
0.002
|
Sex[girls(boys),n]
|
71(136)
|
20(38)
|
51(98)
|
0.97
|
Weight(kg)
|
9.5(8.0-11.3)
|
9.00(7.5–11.0)
|
9.5(8.0-11.5)
|
0.09
|
Clinical presentations, n(%)
|
|
|
|
Crackles
|
76(36.7%)
|
23(39.7%)
|
53(35.6%)
|
0.58
|
Length of fever
|
5.0(2.0–10.0)
|
6.0(2.0–11.0)
|
5.0(2.0–9.0)
|
0.31
|
Accompanied with virus infection
|
22(10.6%)
|
7(12.1%)
|
15(10.1%)
|
0.68
|
Accompanied with bacteria infection
|
113(54.6%)
|
33(56.9%)
|
80(53.7%)
|
0.68
|
Accompanied with mycoplasma pneumoniae
|
35(16.9%)
|
7(12.1%)
|
28(18.8%)
|
0.25
|
Length of disease before admission(days)
|
12.0(7.0–20.0)
|
11.5(6.0–20.0)
|
12.0(7.0–20.0)
|
0.65
|
Underlying conditions, n(%)
|
|
|
|
|
Congenital heart disease
|
71(34.3%)
|
24(41.4%)
|
47(31.5%)
|
0.18
|
Moderate-to-severe malnutrition
|
5(2.4%)
|
1(1.7%)
|
4(2.7%)
|
1.00
|
Bronchopulmonary dysplasia
|
4(1.9%)
|
1(1.7%)
|
3(2.0%)
|
1.00
|
Airway structure abnormality
|
36(17.4%)
|
10(17.2%)
|
26(17.4%)
|
0.97
|
Complications, n(%)
|
|
|
|
|
Atelectasis
|
31(15.0%)
|
14(24.1%)
|
17(11.4%)
|
0.02
|
Pulmonary consolidation
|
154(74.4%)
|
47(81.0%)
|
107(71.8%)
|
0.28
|
Pleurisy
|
46(22.2%)
|
17(29.3%)
|
29(19.5%)
|
0.13
|
Mediastinal or cutaneous emphysema
|
2(1.0%)
|
1(1.7%)
|
1(0.7%)
|
0.48
|
Gastrointestinal dysfunction
|
103(49.8%)
|
31(53.4%)
|
72(48.3%)
|
0.51
|
Laboratory values
|
|
|
|
|
WBC(*10^9)
|
9.8(6.6–13.4)
|
10.5(6.6–14.0)
|
9.6(6.6–13.1)
|
0.99
|
L(%)
|
0.4(0.3–0.5)
|
0.3(0.3–0.5)
|
0.4(0.2–0.5)
|
0.31
|
PLT(*10^9)
|
325.0(249.0-436.0)
|
296.5(234.0-478.0)
|
333.0(253.0-416.0)
|
0.99
|
Hb(g/L)
|
108.0(98.5–118.0)
|
103.5(95.0-116.0)
|
110.0(100.0-121.0)
|
0.006
|
PCT(ng/ml)
|
0.6(0.1–1.8)
|
1.2(0.3–2.6)
|
0.5(0.1–1.5)
|
0.17
|
CRP(mg/L)
|
4.0(4.0-25.5)
|
11.0(4.0–29.0)
|
4.0(4.0–24.0)
|
0.39
|
ALT(U/L)
|
26.3(18.6–37.7)
|
30.2(18.4–45.0)
|
25.3(19.1–35.6)
|
0.14
|
AST(U/L)
|
49.1(38.2–77.5)
|
61.0(42.2–97.1)
|
47.8(37.5–69.0)
|
0.07
|
ALB(g/L)
|
37.2(32.2–41.9)
|
33.0(26.9–37.9)
|
38.0(33.8–43.0)
|
< 0.001
|
LDH(U/L)
|
450.0(315.5-674.5)
|
432.3(313.9–858.0)
|
455.0(317.0-647.8)
|
0.10
|
CK-MB(U/L)
|
1.3(0.7–2.8)
|
2.4(1.2-4.0)
|
1.0(0.6-2.0)
|
0.02
|
HAdV load (lnPCR)
|
16.2(11.7–18.2)
|
16.3(11.8–17.8)
|
16.2(11.6–18.3)
|
0.85
|
Note: The data were collected right after the admission to the general wards (clinical symptoms and signs, laboratory values were obtained within 24 hours, and the radiological data were obtained within 48 hours after admission). No one had severe extrapulmonary complications such as meningitis, pericarditis, endocarditis, osteomyelitis, arthritis, sepsis, hemolytic uremia syndrome, etc.
Abbreviations: ALB, albumin; ALT, alanine aminotransferase; AST, aspartate transaminase; CK-MB, creatine kinase MB isoenzyme; CRP, C-reactive protein; Hb, hemoglobin; IMV, invasive mechanical ventilation; L, lymphocyte ratio; LDH, lactic dehydrogenase; ln, Napierian logarithm; N, neutrophil ratio; PCR, polymerase chain reaction; PCT, procalcitonin; PLT, platelet; WBC, white blood cell.
|
Developing a predictive nomogram in training set
In the univariate logistic analysis, eight out of thirty variables were chosen to be included in the multivariate logistic regression analysis. Finally, four predictors were selected for high risk factors for need for IMV in training set including age (odds ratio [OR]: 0.96; 95% CI:0.92–0.99; p < 0.05), level of ALB(OR: 0.87;95% CI: 0.80–0.93; p < 0.001), CK-MB(OR: 1.32 ; 95% CI: 1.06–1.80; p < 0.05) and atelectasis(OR: 4.83; 95% CI:1.69–14.51; p < 0.01) (Table 2).
Table 2
Logistic regression analysis of predictors of IMV in training set
Univariate logistic analysis
|
Multivariate logistic analysis
|
Variables
|
OR(95%CI)
|
p-value
|
OR(95%CI)
|
p-value
|
Age
|
0.95(0.92–0.99)
|
0.005
|
0.96(0.92–0.99)
|
0.023
|
Sex
|
0.75(0.35–1.59)
|
0.45
|
|
|
Weight
|
0.88(0.79–0.99)
|
0.03
|
|
|
Crackles
|
1.28(0.61–2.67)
|
0.17
|
|
|
Length of fever
|
1.03(0.98–1.09)
|
0.24
|
|
|
Accompanied with virus infection
|
1.52(0.48–4.86)
|
0.48
|
|
|
Accompanied with bacteria infection
|
1.00(0.48–2.10)
|
1.00
|
|
|
Accompanied with mycoplasma pneumoniae
|
0.74(0.25–2.16)
|
0.58
|
|
|
Length of disease before admission
|
1.02(0.98–1.05)
|
0.38
|
|
|
Congenital heart disease
|
1.41(0.67–2.99)
|
0.36
|
|
|
Moderate-to-severe malnutrition
|
0.87(0.09–8.64)
|
0.91
|
|
|
Bronchopulmonary dysplasia
|
-
|
-
|
|
|
Airway structure abnormality
|
0.65(0.22–1.87)
|
0.42
|
|
|
Pulmonary consolidation
|
1.80(0.72–4.52)
|
0.21
|
|
|
Atelectasis
|
2.68(1.15–6.26)
|
0.02
|
4.83(1.69–14.51)
|
0.004
|
Pleurisy
|
1.56(0.71–3.41)
|
0.27
|
|
|
Mediastinal or cutaneous emphysema
|
2.67(0.16–43.68)
|
0.49
|
|
|
Gastrointestinal dysfunction
|
1.53(0.73–3.20)
|
0.26
|
|
|
WBC
|
1.01(0.96–1.06)
|
0.79
|
|
|
L
|
0.63(0.14–2.90)
|
0.55
|
|
|
PLT
|
1.00(1.00–1.00)
|
0.37
|
|
|
Hb
|
0.98(0.95-1.00)
|
0.05
|
|
|
PCT
|
1.03(0.99–1.06)
|
0.16
|
|
|
CRP
|
1.00(0.99–1.01)
|
0.86
|
|
|
ALT
|
1.01(1.00-1.02)
|
0.15
|
|
|
AST
|
1.00(1.00-1.01)
|
0.09
|
|
|
ALB
|
0.88(0.82–0.94)
|
< 0.001
|
0.87(0.80–0.93)
|
< 0.001
|
LDH
|
1.00(1.00–1.00)
|
0.12
|
|
|
CK-MB
|
1.45(1.15–1.83)
|
0.002
|
1.32(1.06–1.80)
|
0.04
|
lnPCR
|
1.01(0.93–1.10)
|
0.80
|
|
|
Abbreviations: ALB, albumin; ALT, alanine aminotransferase; AST, aspartate transaminase; CK-MB, creatine kinase MB isoenzyme; CRP, C-reactive protein; Hb, hemoglobin; ICU, intensive care unit; IMV, invasive mechanical ventilation; L, lymphocyte ratio; LDH, lactic dehydrogenase; ln,Napierian logarithm; N, neutrophil ratio; NPA, nasopharyngeal aspirate; PCIS, Pediatric Critical Illness Score; PCR, polymerase chain reaction of human adenovirus; PEWS, Pediatric Early Warning Score; PCT, procalcitonin; PLT, platelet; PRISM III, Pediatric Risk of Mortality III; WBC, white blood cell.
|
A predictive nomogram incorporating the 4 predictors was established by logistic regression (Fig. 2). The AUC of the probability of IMV for the predictive model was 0.85 (95% CI, 0.78–0.91) in training set, which is bigger than the 4 optimal predictors[age: 0.73 (95% CI, 0.63–0.82), level of ALB: 0.70 (95% CI, 0.60–0.79), atelectasis: 0.59 (95% CI, 0.51–0.67), CK-MB: 0.72 (95% CI, 0.62–0.83)] (Figure S2).
Validation of the predictive nomogram
Validation of the predictive nomogram was performed with a 1000-bootstrap analysis. Harrell’s concordance index in the validating set was 0.81. In the validating set, AUC of the probability for IMV was 0.81 (95% CI, 0.69–0.94). The calibration curves of the predictive nomogram showed good probability consistencies between the prediction and observation in training set (Fig. 3A) and validating set (Fig. 3B). Hosmer-Lemeshow goodness-of-fit test indicated no significant deviation between observed and predicted events in validating set (p = 0.55).
The clinical value of the predictive nomogram was conducted by DCA and CIC analysis. DCA is a novel method for evaluating alternative predictive strategies, which has advantages over the ROC. The DCA curve showed obvious net benefits of the predictive nomogram. The CIC of this predictive model visually showed the estimated number that would be declared as high risk for each risk threshold and the proportion of those who were cases (true positives). The DCA curve and CIC for the predictive model in the validating set is presented in Figure S3.
Comparison with three critical scores
In validating set, PRISM III, PEWS and PCIS had a AUC of the probability for IMV of 0.72 (95% CI, 0.58–0.85, 0.60 (95% CI, 0.43–0.76) and 0.58 (95% CI, 0.41–0.75), respectively(Fig. 4A). The AUC of the probability for IMV for the predictive model was bigger than all the three scales. Significant difference was obtained in PEWS, PCIS and PRISM III (all p < 0.05) in comparison with the new predictive model, respectively. The DCA curves showed the predictive nomogram had more obvious net benefits than the three critical scores(Fig. 4B).