Combination of procalcitonin and C‐reactive protein levels in the early diagnosis of bacterial co‐infections in children with H1N1 influenza

Objective This study evaluated the diagnostic value of measuring the levels of procalcitonin (PCT) and C‐reactive protein (CRP) to differentiate children co‐infected with H1N1 influenza and bacteria from children infected with H1N1 influenza alone. Methods Consecutive patients (children aged < 5 years) with laboratory‐confirmed H1N1 influenza who were hospitalized or received outpatient care from a tertiary‐care hospital in Canton, China, between January 1, 2012, and September 1, 2017, were included in the present study. Laboratory results, including serum PCT and CRP levels, white blood cell (WBC) counts, and bacterial cultures, were analyzed. The predictive value of the combination of biomarkers versus any of the biomarkers alone for diagnosing bacterial co‐infections was evaluated using logistic regression analyses. Results Significantly higher PCT (1.46 vs 0.21 ng/mL, P < 0.001) and CRP (19.20 vs 5.10 mg/dL, P < 0.001) levels were detected in the bacterial co‐infection group than in the H1N1 infection‐alone group. Using PCT or CRP levels alone, the areas under the curves (AUCs) for predicting bacterial co‐infections were 0.801 (95% CI, 0.772‐0.855) and 0.762 (95% CI, 0.722‐0.803), respectively. Using a combination of PCT and CRP, the logistic regression‐based model, Logit(P) = −1.912 + 0.546 PCT + 0.087 CRP, showed significantly greater accuracy (AUC: 0.893, 95% CI: 0.842‐0.934) than did the other three biomarkers. Conclusions The combination of PCT and CRP levels could provide a useful method of distinguishing bacterial co‐infections from an H1N1 influenza infection alone in children during the early disease phase. After further validation, the flexible model derived here could assist clinicians in decision‐making processes.

monly used to differentiate between bacterial and viral etiologies.
Although previous studies have focused on using CRP levels to detect bacterial co-infections in patients with H1N1 infections, the evidence from these studies is inconsistent. Studies suggested serum CRP as a potential diagnostic biomarker, [9][10][11] whereas Piacentini et al 12 found that CRP levels were unable to distinguish bacterial co-infections from H1N1 infections. Another interesting biomarker is procalcitonin (PCT), the prohormone of calcitonin produced by C cells in the thyroid. Plasma PCT concentrations are low in healthy individuals and increase during bacterial, parasitic, or fungal infections, whereas they remain at normal levels during viral infections or non-infectious inflammatory reactions. 13 Studies attempting to assess PCT levels in patients with H1N1 infection have found that PCT helped to distinguish bacterial co-infections from H1N1 infections alone. 14,15 Nevertheless, to the best of our knowledge, previous studies published to date have focused on adults 14,15 and patients with severe disease 16 but have included few pediatric patients with H1N1 infections.
Thus, in the present study, we conducted a retrospective analysis of 3180 children with H1N1 infection to evaluate the diagnostic levels of serum PCT, CRP and WBC alone and in combination for differentiating bacterial co-infections from H1N1 influenza infections alone in children. These results could be used to provide a reliable clinical diagnostic support system for improving diagnostic accuracy and enabling the early treatment of bacterial co-infections accompanying H1N1 influenza infections.

| Settings and participants
We performed a retrospective cohort study of consecutive patients with laboratory-confirmed H1N1 influenza infections, all of whom were children <5 years old who were hospitalized or received out-

| Definitions
Patients diagnosed with H1N1 influenza infection confirmed by realtime reverse transcriptase polymerase chain reaction (RT-PCR) 17  Bacterial co-infection was defined as a positive H1N1 influenza viral PCR result with one or more bacterial pathogens detected. Bacterial cultures were obtained from blood, valid sputum, and BAL fluid within the first 48 hours of hospitalization. We excluded patients who received antibiotics prior to hospitalization or who lacked bacterial cultures.

| Statistical analysis
Categorical variables are summarized using absolute values and percentages, and continuous variables are presented as medians and interquartile ranges (IQR). The chi-square test (for nominal variables) or the Wilcoxon rank-sum test (for continuous variables) was employed for between-group comparisons. Univariate logistic regression analysis was used to assess the ability of each biomarker (PCT, CRP, and WBC) to diagnose bacterial co-infections.
Furthermore, iterative biomarker(s) were selected (including biomarkers with P < 0.10) using automatic forward stepwise regression, and the multivariate logistic regression model was built.
The performance of the models was then assessed by calculating the area under the receiver operating characteristic (ROC) curve (AUC). The AUC values were compared for each biomarker individually and in conjunction with biomarker models using the Hanley and McNeil method. 19 The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were also reported. The Youden index (sensitivity + specificity − 1) was used to determine the optimal ROC cutoff value. Moreover, 10fold cross-validation to evaluate the robustness of the estimates obtained from the constructed model was performed as described previously. 20 Then, we averaged the AUC, sensitivity, and specificity values obtained from the 10-fold cross-validations to generate summary performance estimates.
All statistical analyses were performed using R Software, version 3.4.2 (www.r-project.org). A two-tailed P value < 0.05 was considered significant.  (Table S1). Eight children (3.5%) displayed two positive respiratory tract bacterial cultures.

| Study population and bacterial pathogen characteristics
When the baseline characteristics and clinical outcomes of the H1N1 plus bacterial co-infection group were compared, children in the H1N1-alone group were older, but this result was not significant (median age, 2.5 vs 2.4 years, P = 0.197). Differences in gender or weight were not observed between the two groups; however, the bacterial co-infection group showed significantly higher inpatient admission (14.3% vs 50.4%, P < 0.001) and ICU admission rates (2.6% vs 36.3%, P < 0.001) than patients in the H1N1-alone group. The bacterial co-infection group also required longer hospital stays (5 vs 10 days, P = 0.003) than H1N1-alone group and thus had much higher hospital costs (median hospital cost, 1213.2 vs 3467.3 RMB, P < 0.001). Moreover, a higher inhospital mortality rate was noted for the bacterial co-infection group than the H1N1-alone group (0.1% vs 4.8%, P < 0.001; Table 1).

| Comparison of serum PCT, CRP, and WBC levels between H1N1 alone and H1N1 with bacterial co-infection groups
Serum PCT, CRP, and WBC levels were analyzed to identify potential biomarkers that distinguished between H1N1 infections and H1N1 and bacterial co-infections. The median serum PCT, CRP, and WBC TA B L E 1 Demographic and clinical characteristics of patients with H1N1 influenza who presented with and without bacterial co-infections  Figure 1).

| Univariate and multivariate logistic regression analyses
Univariate analysis revealed significant associations of serum PCT, CRP, and WBC levels with co-infections with H1N1 and bacteria (odds ratio F I G U R E 1 Serum PCT (A), CRP (B), and WBC (C) levels in patients with H1N1 influenza who presented with and without bacterial coinfections. The differences between the H1N1-alone group and bacterial co-infection group were examined using the Wilcoxon rank-sum test

| Comparison and validation of the model's diagnostic ability
Because serum PCT and CRP levels were independent predictors that differentiated patients with bacterial co-infections from pa-  reliably and accurately reduce inappropriate antibiotic exposure. 14 Our results showed that serum PCT levels were significantly higher in patients with bacterial co-infection compared with those infected with H1N1 alone, confirming that PCT is associated with The potential limitations of our study should be mentioned.

| D ISCUSS I ON
First, the levels of selected biomarkers (PCT, CRP, and WBC) were evaluated only once. Second, our diagnostic model was derived and validated at a single hospital center and should be validated in a multicenter trial before its broad application. Finally, we also acknowledge that we may have created bias, as bacterial organisms cannot be confirmed solely through blood, sputum, and BAL culture.

| CON CLUS ION
In conclusion, we detected serum PCT and CRP levels and revealed that they represent promising biomarkers and useful clinical tools for differentiating pediatric patients with bacterial co-infections from those infected with H1N1 alone. Furthermore, the combination of PCT and CRP levels could represent a useful method for screening bacterial co-infections from H1N1 influenza infections alone in children during the early disease phase. After further validation, the flexible model reported here may assist clinicians with decisionmaking processes.

ACK N OWLED G EM ENTS
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.