The Role of Hyperleukocytosis in Predicting the Severity of Pertussis

Background: Few reports have described the relationship between WBC count and the severity of pertussis or the timing of ET in patients with hyperleukocytosis. Methods: A retrospective clinical analysis of infants with pertussis and a WBC exceeding 30*10^9/L was performed. Results: A total of 158 patients were enrolled in the study. There were signicant differences in the clinical manifestations of cyanosis, fever, highest respiratory rate, and highest heart rate. There were signicant differences in all complications except for pulmonary hypertension. In laboratory ndings, there were signicant differences in organ damage (myocardial markers, ALT), increased inammation indicators (CRP, PCT), and the incidence of combined bacterial infections. There were signicant differences in ICU stay length, mechanical ventilation use, days hospitalized, days until cough relief and days until the WBC fell below 25*10^9/L. A WBC count >55.38 *10^9/L was calculated as the cutoff value with 88.2% sensitivity and 23.4% specicity in predicting ET. A respiratory rate of 59 breaths/min had 94.1% sensitivity and 36.7% specicity in predicting ET. A heart rate of 159 beats/min had 100% sensitivity and 38.1% specicity in predicting ET. Conclusion: WBC count is related to the severity of pertussis. We recommend that ET is considered when the patient’s WBC count exceeds 55*10^9/L, breathing exceeds 60 breaths/min, and/or heart rate exceeds 160 beats/min.


Impact Of This Article:
The WBC count is related to pertussis severity.
The higher the WBC count, the more severe the disease, the greater incidence of complications, the higher the occurrence of combined organ dysfunction, the higher the risk of combined bacterial infection, the slower the recovery and the worse the prognosis.
We recommend that ET be considered for infants under 1 year old, patients with a WBC count that has reached or exceeded 55*10^9/L, patients with a respiratory rate that has exceeded 60 (breaths/min), and/or patients with a heart rate that has exceeded 160 (beats/min).

Background
Pertussis is a highly contagious respiratory illness. Severe pertussis is the leading cause of death in children with pertussis. The occurrence of severe pertussis is affected by many factors, such as pneumonia, pulmonary hypertension and hyperleukocytosis. Leukocytosis is a predictor of mortality in patients with severe pertussis infection. Kathleen Winter's research has shown that WBC counts > 30*10^9/L are strongly associated with fatal pertussis in infants [1] ;research conducted by Berger [2] showed that children with a WBC count greater than 50 *10^9/L had a nearly ten times higher risk of death. WBC counts > 70*10^9/L are strongly associated with death [1] ,and hyperleukocytosis [WBC count > 100*10^9/L] is a rare complication that strongly predicts mortality in cases of severe pertussis [3] . Research indicates that elevated WBC counts are associated with increased mortality risk. However, there has been no systematic research analyzing the ability to predict the severity and risk of death by WBC count in patients with pertussis. The purpose of exchange transfusion (ET) is to lower the WBC count, and ET can reduce blood viscosity, pertussis toxin levels, and the pulmonary small blood vessel obstruction caused by white blood cells. In recent years, ET has been widely used in hyperleukocytosis treatment, but there is no uni ed understanding of the effect and standard use of ET in patients with pertussis. Cherry et al [4] suggested exchange blood transfusion and monitoring criteria for infants with pertussis who are ≤ 60 days of age. Tian et al [3] summarized the literature on ET from 2004 to 2017, and the survival rate of pertussis patients with hyperleukocytosis was signi cantly higher when ET therapy was used. However, the scattered ET reports have different conclusions. Therefore, we wanted to compare the relationship between disease severity and risk of death at different WBC count by collecting patients with WBC ≥ 30*10^9/L; at the same time, we also analyzed the data in transfusion and nontransfusion patients. We assessed WBC levels in relation to the severity of pertussis and the timing of ET.  43.6% in group B, and 21.7% in group C, and there were signi cant differences among the three groups (P = 0.032).

Patients
There were 12 patients who died in the study, 10 (43.5%) patients in group C, and only 2 (3.6%) patients in group B. No patient died in group A. There were signi cant differences in the mortality rate among the three groups (P = 0.000).
In the three groups, there were no signi cant differences in the times of vaccination, age of onset, or time from onset to admission (P = 0.479, 0.78, 0.312, respectively).

Clinical manifestations and complications
Clinical manifestations and complication data are shown in Table 1. Regarding the clinical manifestations, there were no signi cant differences in the incidence of whooping and wheezing (P = 0.166, 0.091). The incidence of cyanosis was 36 (45%) in group A, 32 (58.2%) in group B, and 18 (78.3%) in group C, with signi cant differences (P = 0.015). There were signi cant differences among the three groups (P = 0.015). The incidence of fever in group A was 33 (41.3%), the incidence in group B was 33 (60%), and in the incidence in group C was 21 (91.3%); there was a signi cant difference (P = 0.000).
However, there were no signi cant differences in the highest temperature of fever among the three groups (P = 0.114). The highest heart rate (beats/min) in group A was 144 (136.5-161), in group B was 157 (148.75-172.25), in group C was 187.5 (166.75-201.25); the difference was statistically signi cant (P = 0.000). The highest respiratory rate (breaths/min) in group A was 53.9 ± 11.46, group B was 59.4 ± 12.33, and group C was 69.04 ± 14.08; the difference was also statistically signi cant (P = 0.000).
Regarding complications, the number of pneumonia cases was 42 (52.5%) in group A, 29 (52.7%) in group B and 3 (13%) in group C; the difference was statistically signi cant (P = 0.002). The number of severe pneumonia cases in group A was 16 (20%), that in group B was 23 (41.8%), and that in group C was 20 (87%); the difference was statistically signi cant (P = 0.002). The incidence of pneumonia and severe pneumonia was 72.5% in group A, 94.5% in group B, and 100% in group C. The incidence of complications such as respiratory failure, pulmonary consolidation or atelectasis, pulmonary hypertension, heart failure, and pertussis encephalopathy all increased with the increased WBC. The incidence rates of all complications were signi cantly different among the three groups (P = 0.000, 0.000, 0.000, 0.000, respectively), except for pulmonary hypertension among the three groups (P = 0.190).

Laboratory Data
Laboratory data are presented in table 1. On organ function assessment, the incidence of abnormal myocardial markers in group A was 13.8% (n=11), the incidence in group B was 21.8% (n=12), and the incidence in group C was 43.5% (n=10); the incidence of abnormal liver function in group A was 20% (n=16), the incidence in group B was 16.4% (n=9), and the incidence in group C was 78.3% (n=18). The incidence of myocardial and liver function abnormalities in the three groups were signi cantly different (P=0.013, 0.000).
In the examination of in ammatory indicators, the number of patients with increased CRP in group A was 5 (6.3%), in group B was 14 (25.5%), and in group C was 16 (69.6%); the number of patients with increased procalcitonin (PCT) in group A was 33 (44%).), in group B was 27 (50%), and in group C was 18 (81.8%). of the incidence of increased CRP and PCT were signi cantly different among the three groups (P=0.000, 0.007).
In the pathogenic examinations, 28 (35.9%) cases in group A were combined with bacterial infection and 14 (17.9%) were combined with viral infection; the numbers of cases combined with bacterial infection and viral infection were 25 (46.3%) and 16 (29.1%) in group B, and 19 (82.6%) and 8 (34.8%) in group C, respectively. The difference in combined bacterial infections among the three groups was statistically signi cant (P=0.000), but there was no signi cant difference in the incidence of combined viral infections (P=0.152).

Treatment
There was no signi cant difference in the days from illness onset to macrolide plus treatment or the total course of macrolides among the three groups of patients (P=0.465, 0.366). The incidence of gamma globulin, steroids, vasoactive drugs, other antibacterial drugs, and ET increased gradually with increasing WBC levels among the three groups, with the exception of steroid use (P=0.31), there were signi cant differences among the groups (P =0.000, 0.000, 0.003, and 0.000).
The number of patients admitted to the ICU was 5 (6.3%) in group A, 10 (18.2%) in group B, and 21 (91.3%) in group C. There was a signi cant difference in the rate of ICU admission within 48 hours of hospital admission among the three groups (P=0.000). Although the rate of admission to the ICU within 48 hours after hospital admission also increased with the increase in white blood cell count, there was no signi cant difference in the rate of admission to the ICU within 48 hours of hospital admission among the three groups (P=0.380).

Improvement after treatment
The length of hospital stay in group A was 9 days (6-13), in group B was 12 days (9-16), and in group C was 16 days . The number of days until cough relief in group A was 23 (17-27.75), in group B was 25.5 (21-31), and in group C was 29 (24.5-37). The number of days until the WBC fell below 25*10^9/L in group A was 6 (4-7), in group B was 9 (7-12.75), and in group C was 12 (9-13.25). The length of the hospital stay, the number of days until cough relief and the number of days until the WBC fell below 25*10^9/L were all prolonged as the WBC level increased, and the difference was statistically signi cant (P=0.011, 0.013, 0.013, respectively), as shown in Table 1. The area under the receiver operating characteristic (ROC) curve (AUC) value for WBC count with the cutoff point of 55.38 *10^9/L was 0.899 (95% con dence interval (CI), 0.834~0.963 P 0.000) in predicting exchange transfusion. When applying the ROC curve with the WBC count cutoff point of 55.38 *10^9/L, the analysis yielded 88.2% sensitivity and 23.4% speci city ( Figure 1). The area under the receiver operating characteristic (ROC) curve (AUC) value for respiratory rate with a cutoff point of 59 breaths/min was 0.795 (95% con dence interval (CI), 0.699~0.891 P 0.000) in predicting exchange transfusion. When applying an ROC curve with the respiratory rate cutoff point of 59 breaths/min, the analysis yielded 94.1% sensitivity and 36.7% speci city (Figure 2). The area under the receiver operating characteristic (ROC) curve (AUC) value for heart rate with a cutoff point of 159 beats/min was 0.813 (95% con dence interval (CI), 0.731~0.895 P 0.000) in predicting exchange transfusion. When applying the ROC curve with heart rate at the cutoff point of 159 beats/min, the analysis yielded 100% sensitivity and 38.1% speci city (Figure 3).

Discussion
Pertussis is a highly contagious and vaccine-preventable respiratory illness. In the past few decades, immunization has greatly reduced the incidence of the disease [5] . However, pertussis has not disappeared due to widespread vaccination; it continues to threaten the health and even lives of humans, especially young children. In 2010, Black et al [6] . reported that 16 million cases of pertussis occurred in 2008 worldwide, resulting in 195,000 deaths. In our study, there were 12 patients who died among the 158 patients included. Ten (43.5%) patients had WBC counts exceeding 70*10^9/L, and only 2 (3.6%) patients in the 50*10^9/L ≤ WBC < 70*10^9/L group had WBC counts exceeding 70*10^9/L. No patient died in 30*10^9/L ≤ WBC < 50*10^9/L group. This shows that when the WBC counts is below 50*10^9/L, the risk of death is low. The risk of death is mainly in patients with WBC ≥ 70 *10^9/L. In our research, we also noticed that there was a signi cant difference in the sex ratio among the three groups of patients (P = 0.032). When the WBC counts was 30*10^9/L ≤ WBC < 50*10^9/L, there was no signi cant difference in sex, but when the WBC counts was ≥ 50*10^9/L, the higher the WBC count, the higher the proportion of female patients. When the WBC count was ≥ 70*10^9/L, the proportion of female patients was 78.3%.
However, Haberling et al showed that among severe pertussis and pertussis-related deaths, the rate of female patients may not be as high as stated in other reports [7] .
Severe pertussis is related to many factors, especially hyperleukocytosis [2,8] . In our study, according to the patient's WBC count, we divided the patients into three groups: 30*10^9/L ≤ WBC 50*10^9/L, 50*10^9/L ≤ WBC 70*10^9/L, and 70*10^9/L ≤ WBC. We found that the three groups of patients had signi cant differences in many clinical symptoms, such as the incidence of cyanosis (P = 0.015), the probability of fever (P = 0.000), the highest respiratory rate (P = 0.000), and the highest heart rate (P = 0.000). The results suggest that the higher the WBC count is, the more severe the disease may be. However, the highest temperature was not signi cantly different among the three groups (P = 0.114), indicating that the highest temperature does not necessarily re ect the severity of the disease. The incidence of complications such as severe pneumonia, respiratory failure, lung consolidation or atelectasis, pulmonary hypertension, heart failure, and pertussis encephalopathy in the three groups of patients all increased with increasing WBC count. The higher the WBC count was, the more severe the disease.
Comparing the laboratory ndings of patients with different WBC counts, we found that the incidence of organ damage (liver and heart) was also signi cantly different among the three groups of patients (P = 0.013, 0.000), which may be related to the severity of the disease. In the comparison of in ammatory indicators, the incidence of increased peripheral blood C-reactive protein (CRP) and procalcitonin (PCT) increased with the increase in white blood cell count, with signi cant differences (P = 0.000, 0.007). This may be caused by the presence of a combined bacterial infection because we also found that the three groups of patients had a signi cant difference in the incidence of combined bacterial infection (P = 0.000). The results still showed that the incidence of increased CRP, PCT and combined bacterial infections increased with the increased WBC count.
In addition, the patient's ICU admission rate, days hospitalized, days until cough relief (days), and days until the WBC fell below 25*10^9/L were signi cantly different at different WBC levels (P = 0.000, 0.011, 0.013, 0.013, respectively). This relationship is also consistent with the increase in the white blood cell count.
At present, people have recognized the effects of hyperleukocytosis in pertussis [9][10] . The most widely used treatment for hyperleukocytosis is ET. It can reduce the WBC count and improve the survival rate of patients. It is very important for the treatment of severe pertussis [11] .Although ET has been used in the treatment of patients with hyperleukocytosis for many years, there has been no systematic study to analyze the timing and indicators of ET. We drew the ROC curves using transfusion and nontransfusion data, and we found that ET may be considered for infants under 1 year old, patients with a WBC count that reached or exceeded 55*10^9/L, patients in whom the highest respiratory rate exceeded 60 (breaths/min), and/or patients in whom the highest heart rate exceeded 160 (beats/min).

Conclusion
Taken together, the level of WBCs is related to disease severity. The higher the WBC count, the more severe the disease, the greater incidence of complications, the higher occurrence of combined organ dysfunction, the higher risk of combined bacterial infection, the slower the recovery and the worse the prognosis. We recommend that ET is considered for infants under 1 year old, patients with a WBC count that has reached or exceeded 55*10^9/L, patients in whom the highest respiratory rate exceeded 60 (breaths/min), and/or patients in whom the highest heart rate exceeded 160 (beats/min).

Declarations Ethical Approval statement
This study was approved by the Ethics Committee of Children's Hospital of ChongQing Medical University. As it is a retrospective clinical analysis, it was decided by the ethics committee granted an exemption for informed consent.

Consent for publication
Not applicable.