Thrombocytosis and thrombocytopenia are markers of poor outcome in pediatric patients with community-acquired pneumonia

Abstract Background: This study aimed to investigate the prevalence of thrombocytopenia and in with (CAP), and determine whether thrombocytopenia and thrombocytosis are associated with patient outcome. Methods: A total of 9,372 consecutive patients, who were 1-168 months old, diagnosed with CAP and admitted in the Children’s Hospital of Soochow University, were enrolled in the present retrospective observational study. Their clinical and laboratory data were collected. According to the platelet count on admission, these patients were divided into three groups: thrombocytopenia, normal platelet count, and thrombocytosis groups. The clinical characteristics and etiologic pathogens were compared among these groups. The multivariate logistic regression model was applied to identify risk factors for severe CAP, length of hospitalization ≥10 days and respiratory complications. The correlations between platelet count and clinical features were determined by Spearman’s correlation. Results: Thrombocytosis and thrombocytopenia were found in 3,376 (36.0%) and 43 (0.5%) patients, respectively. Normal platelet count was observed in 5,953 (63.5%) patients. Thrombocytopenia was an independent risk factor of severe CAP (OR, 6.206; 95% CI, 2.209-17.436; P=0.001), while thrombocytosis was associated with length of hospitalization of ≥10 days (OR, 1.315; 95% CI, 1.177-1.470; P<0.001). In addition, thrombocytosis was associated with respiratory complications (OR, 1.658; 95% CI, 1.171-2.346; P=0.004). Platelet count (median 350.0 IQR 270.2-447.0 × 109/L) was positively correlated with length of hospitalization (median 7.0 IQR 6.0-9.0 days) (r = 0.101, P<0.001), but negatively correlated with age (median 12.0 IQR 3.0-36.0 months) (r = -0.401,

. In the process of infection and inflammation, thrombocytes may secrete various substances, such as cytokines, pro-coagulants, oxidants and antimicrobial peptides, which are involved in beneficial or harmful activities [4][5][6]. Therefore, an abnormal platelet count may be a critical biomarker for assessing disease severity.
The impact of thrombocytopenia or thrombocytosis on the outcomes of pediatric patients with community-acquired pneumonia (CAP) has been scarcely studied, although previous studies have confirmed that thrombocytopenia and/or thrombocytosis were associated with mortality in adult CAP patients [7][8][9][10]. Thrombocytosis is actually more common in childhood [11]. Two studies on pediatric CAP revealed that thrombocytosis is frequently observed in severe bacterial infections, and that thrombocytosis is associated with more severe and protracted diseases [12,13]. However, the sample size of these studies was limited. Furthermore, the differences in clinical and etiologic characteristics between patients with thrombocytosis and normal platelet count have not been clarified.
To the best of our knowledge, few large-scale studies have investigated the impact of abnormal plate count on the outcomes of pediatric patients with CAP. The present study evaluates the prevalence of thrombocytopenia and thrombocytosis in pediatric CAP patients, and determines whether thrombocytopenia and thrombocytosis is associated with clinical outcomes.

Methods
In the present retrospective observational study, consecutively admitted pediatric patients, who were diagnosed with CAP at the Children's Hospital of Soochow University (a tertiary teaching hospital), Suzhou, China, between January 2012 and December 2017, were included.
The inclusion criteria were as follows: (1) patients within 1-168 months old; (2) patients with newonset pulmonary infiltrate on chest X-ray at admission, which may be accompanied by other signs consistent with pneumonia, such as fever, cough, or auscultatory findings.
The present study was approved by the Ethics Committee of the Children's Hospital of Soochow University. A signed consent was obtained from the patient's guardian.

Data collection and microbiologic evaluation
Clinical and laboratory data were collected. The collected variables included the patient's demographical information, basic diseases, clinical signs and symptoms, comorbidity, laboratory and chest X-ray findings. The laboratory results were obtained within six hours after admission. The criteria for the microbiologic diagnosis have been previously described [14,15].

Study definitions
Thrombocytopenia was defined as a platelet count of <100,000/L, while thrombocytosis was defined as a platelet count of > 400,000/L. CAP was defined the appearance of a new pulmonary infiltrate on chest X-ray at admission, in combination with fever, cough, or auscultatory findings [16]. Severe CAP was determined when the World Health Organization criteria for severe pneumonia was met [17].
Pleural effusion or empyema and lung abscess were considered as respiratory complications in the present study [16]. Fever was defined as an axillary temperature exceeding 38.0°C. A diagnosis of tachypnea was made when the respiratory rate was higher than the WHO classification categorized by age: 0-2 months old: >60 breaths/min; 2-12 months old: >50 breaths/min; 1-5 years old: >40 breaths/min; >5 years old: >20 breaths/min [16].
The cut-off for length of hospitalization was 10 days, considering the sample size and the percentage of children hospitalized for more or less than 10 days. The overall median length of hospitalization was 7.0 days (range: 3.0-56.0 days; interquartile range [IQR]: 6.0-9.0). The percentage of patients with a length of hospitalization of ≥10 days was 19.34% (n=1,813).

Statistical analysis
Statistical analyses were performed using SPSS 21.0 software. Data normality was tested using the Kolmogorov-Smirnov test. Continuous variables with non-normal distribution were expressed as median and IQR (25 th -75 th percentile). Non-normally distributed continuous variables were analyzed 5 using the Mann-Whitney U-test or kruskal-Wallis test. Categorical variables were compared using chisquared test or Fisher's exact test. The correlations between variables were determined by spearman correlation. Variables with a univariate P-value ≤0.1 were included in the multivariate logistic regression models. P<0.05 was considered statistically significant.

Results
During the study period, a total of 9,372 consecutive patients were enrolled in the present study. The study flowchart was presented in Figure 1. Thrombocytosis was found in 3,376 (36.0%) patients, while thrombocytopenia was found in 43 (0.5%) patients. Normal platelet count was observed in 5,953 (63.5%) patients. The patient characteristics are presented in Table 1. Factors associated with severe CAP, length of hospitalization ≥10 days and respiratory complications were analyzed (Table 3). Age, bronchopulmonary dysplasia, congenital heart disease, neuromuscular disorder, wheezing, fever, gastrointestinal symptoms, tachypnea and C-reactive protein were independent predictors of severe CAP. When platelet count was included as a categorical variable, the 6 association between thrombocytopenia and severe CAP was confirmed (OR, 6.206; 95% CI, 2.209-17.436; P=0.001). Asthma, bronchopulmonary dysplasia, congenital heart disease, neuromuscular disorder, wheezing, fever, gastrointestinal symptoms, tachypnea and C-reactive protein were independent predictors of length of hospitalization ≥10 days. When platelet count was taken as a categorical variable, the association between thrombocytosis and length of hospitalization ≥10 days was observed (OR, 1.315; 95% CI, 1.177-1.470; P<0.001). Age, asthma, wheezing, fever and Creactive protein were independent predictors of respiratory complications. Platelet count, as a categorical variable, was associated with respiratory complication (OR, 1.658; 95% CI, 1.171-2.346; P=0.004).
The correlation between platelet count at admission and clinical characteristics and outcomes were

Discussion
The present study evaluated the prevalence of thrombocytopenia and thrombocytosis in hospitalized pediatric patients with CAP, and the contributions of these two to clinical outcomes. These present results revealed that thrombocytopenia occurred in a considerable pediatric population, and that both were correlated with poorer outcome.
Thrombocytosis was identified in approximately one-third of the studied pediatric CAP patients, which is consistent with the previously reported percentage (9-48%) of thrombocytosis in patients with respiratory tract infections [18-20]. In addition, it was found that platelet count was negatively correlated with age, which could be explained by the fact that bone marrow precursor cells in young children are more sensitive to external stimuli, such as infection and inflammation [11].
The prevalence of thrombocytopenia was merely 0.5% in the present study, indicating that 7 thrombocytopenia is not common in pediatric patients with CAP. In a previous study, the researchers found that patients with thrombocytopenia tended to be older [8]. However, the exact underlying mechanism remains unclear and warrants further investigation. Thrombocytosis is associated with multiple pathogens. It was observed that parainfluenza virus, respiratory syncytial virus, human bocavirus and rhinovirus are more frequently identified in patients with thrombocytosis, when compared to patients with normal platelet counts. However, it has not been fully elucidated why infection could promote thrombocytosis. Similarly, the incidence of respiratory syncytial virus and rhinovirus was also reported to be greater in the thrombocytosis group, when compared to the normal group, in a previous study [21]. However, contrary to these previous In the present study, CAP patients with asthma were associated with a lower incidence of respiratory complication. This finding is consistent with that from previous studies [26]. Furthermore, the severity of CAP is largely correlated to comorbid conditions, such as cardiopulmonary and immune status [16].
In agreement with these previous findings, the investigators also found that the conditions of bronchopulmonary dysplasia, congenital heart disease and neuromuscular disorder were independent predictors of severe CAP in the present study.
Thrombocytopenia in adult severe CAP has already been studied [9,27,28]. However, few studies have explored the impact of thrombocytopenia on the severity of CAP in pediatric patients. In the 8 present study, it was demonstrated that thrombocytopenia is an independent predictor of severe CAP in pediatric patients.
These present findings revealed that thrombocytosis was independently associated with length of hospitalization ≥10 days and respiratory complications, which is consistent with two previous studies in adult patients, although the exact mechanism for the association of thrombocytosis with respiratory complication remains unclear [7,8]. As expected, it was found that platelet counts at admission were positively correlated with length of hospitalization. In addition, platelet count at admission was negatively correlated with C-reactive protein, indicating that lower platelet count might be associated with a stronger inflammatory response in patients with CAP. This was consistent with a previous study that revealed that sepsis shock patients were more likely to have lower platelet counts [29].
There are several limitations associated with the present study. First, the changes in platelet count were not dynamically monitored, and merely the values on admission were included. According to previous studies, the serial measurements of platelet counts during hospitalization could differentiate between a transient event and sustained derangements in platelet count [30][31][32]. Second, the present findings derived from a single medical center might not applicable to patients in other areas. Hence, a multi-center study is necessary in the future. Third, related cytokines were not concurrently analyzed, which limited our understanding of the underlying mechanism.

Conclusions
Thrombocytosis is highly prevalent in pediatric CAP patients, while thrombocytopenia less frequently occurs. Parainfluenza virus, respiratory syncytial virus, human bocavirus and rhinovirus are more frequently identified in CAP patients with thrombocytosis. Both thrombocytosis and thrombocytopenia may serve as useful prognostic markers for pediatric CAP.
Abbreviations CAP: community-acquired pneumonia IQR: interquartile ranges Declarations a Differences between the thrombocytopenia and normal platelet count group b Differences between the thrombocytosis and normal platelet count group  Figure 1 Flowchart of the selected population CAP: community-acquired pneumonia