Background
Pertussis is a highly contagious acute respiratory infection disease that is one of the main reasons of infectious disease-related deaths in children [
1]. Since the spread of infant and childhood pertussis vaccination in the 1940s in the world, the incidence of pertussis has decreased more than 80% [
2]. However, during the last two decades, pertussis infections have re-emerged worldwide. The WHO estimated that pertussis affects nearly 240,000,000 children aged < 5 years old each year and causes 160,700 deaths in this age, with the mortality of 4% [
1,
3]. In a study of infants, the mortality was 70% and higher in infants younger than 6 weeks (84%) [
4]. In China, 30,027 children were diagnosed with pertussis in 2019, and the morbidity of pertussis was lower than that in other countries (21.5/million) [
5]. The mortality of pertussis in China is not very clear. This may be due to the limitations of the laboratory tests in China, and it does not reflect the actual incidence.
In this study, we analysed 144 hospitalized paediatric patients with pertussis, including 38 children with severe pertussis who were admitted to the paediatric intensive care unit (PICU) or neonatal intensive care unit (NICU). The purpose of this study was to identify the mortality risk factors in hospitalized patients with severe pertussis. This information may be beneficial to effectively prevent and to institute management strategies early for severe cases.
Discussion
According to the previous literature, higher rates of pertussis, hospitalizations, complications, and mortality were in infants than any other age group [
10]. In the present study, pertussis mainly occurred in children aged < 6 months, especially those aged < 3 months, which is in line with previous reports [
10,
11], and most of the patients were unvaccinated with regard to diphtheria-tetanus-pertussis (DTP). The outbreaks of pertussis have been reported periodically every two to five years [
12]. In our study, there was an obvious increase in the numbers of patients and deaths in 2018–2019. This may be due to the increasing laboratory tests of pertussis in our hospital in the last two years. The mortality of severe pertussis has been reported to be very high in developed countries, ranging from 19.7 to 31% [
13,
14]. The mortality of pertussis was 9.0% in our study, and the mortality of severe pertussis in the PICU/NICU was 34.2%. The mortality of severe pertussis might be lower than that in reality because of the limitations of the laboratory tests and the inadequate recognition of this disease in its early stage. Some pertussis patients who died were not tested for
B. pertussis and might be misdiagnosed. Previous reports have indicated that the mortality of pertussis in infants younger than 6 weeks was much higher [
4]. In our study, 56.9% of patients aged < 3 months died, 92.3% were younger than 3 months and 76.9% were younger than 6 weeks. This means that these children were not protected by the vaccine, because the recommended schedule for DTP vaccination is from 3 months of age in our country. In 2015, the WHO recommended that the primary DTP vaccination should be given at 6 weeks, at least not later than 2 months [
3]. From the results of our research, we also recommend the first dose of DTP should advance to the age of 2 months or even 6 weeks in China.
Based on other publications, younger age, lower birth weight and younger gestational age were the risk factors for death [
15,
16], and age < 3 months and underlying comorbid conditions were the independent risk factors for death [
17]. In our study, we identified that host factors such as younger age and lower birth weight were significantly associated with the death of patients with severe pertussis. However, it was an unexpected finding that younger age was not found to be an independent risk factor, likely because most of the pertussis patients in our study aged < 6 months, especially those aged < 3 months, would affect the data analyses, leading to bias om the result. On the other hand, we identified that clinical manifestations, such as fever ≥ 38.5℃, cough ≥ 14 days, cyanosis, shortness of breath, fatigue with feeding, seizures, altered sensorium, oxygen saturation < 90%, increased respiratory rate, increased heart rate, crackles, and capillary filling time(CRT) > 2 s were significantly associated with the death of severe pertussis patients. These factors would be some important hints to help paediatric doctors recognize severe pertussis patients in the early stage of the illness.
Severe pertussis was accompanied by a wide range of complications, such as pneumonia, pneumothorax, PH, haemorrhage from the gastrointestinal or respiratory tract, toxic encephalopathy, and septic shock being the most common reported [
18‐
20]. In our study, the presence of pneumonia, pleural effusion, pneumothorax, respiratory failure, ARDS, pneumorrhagia, PH, heart failure, sepsis, toxic encephalopathy, and acute renal failure were associated with death. Pneumonia was the most common complication in severe pertussis and was significantly associated with death [
18,
20]. All our dead cases presented with more severe pneumonia effusion, which was more likely to develop to ARDS or pneumorrhagia. ARDS and pneumorrhagia were associated with death from severe pertussis in our univariable analysis, which coincides with prior evidence. In our study, bacterial coinfections were detected in 26.4% besides
B. pertussis and 31.9% patients had fever. This might indicate secondary bacterial infections, mainly pneumonia, pleural effusion, or even septic shock. Beside macrolide antibiotics, some severe patients in our study needed other antibiotics such as cefoperazone, meropenem, and vancomycin et in the ICU.
PH was present in only 12 patients in our study, but it was a strong predictor of death, as well as an independent risk factor for death. A previous study reported refractory PH in fatal pertussis, which is often associated with hyperleukocytosis [
21]. In our study, the leucocytosis in one of patient who died exceeded 100 × 10
9/L (103.23 × 10
9/L). We identified that leucocytosis > 70.0 × 10
9/L and absolute lymphocyte count > 20 × 10
9/L were also significantly associated with the deaths of patients with severe pertussis, and leucocytosis (> 70.0 × 10
9/L) was an independent risk factor for death. The mechanism of PH and hyperleukocytosis occurred in severe pertussis due to pertussis toxin (PT) [
18]. PT can affect cellular signalling and promote leucocytosis with lymphocytosis, which can result in a hyperviscosity syndrome [
22]. Previous reports have shown that luminal aggregates of leucocytes have been observed in pulmonary arterioles, veins and lymphatics of post-mortem lung tissue from infants who died from pertussis [
23]. Abnormal leucocyte aggregation in the lungs can diminish blood flow by increasing vascular resistance, which may lead to PH [
24]. In addition, pertussis pneumonia may trigger hypoxia, acidosis, acute pulmonary vasoconstriction, microcirculation disturbances, and clotting dysfunction. All these compounded effects produce markedly elevated pressures in the vasculature of the lung that could trigger irreversible PH [
24,
25]. Meanwhile, PT is a known inhibitor of G-proteins, which are cardioprotective [
26]. PT could alter the vagal control of the heart rate and respiratory rate through the regulation of G-proteins [
18,
27]. The rapid increase in heart and respiratory rates, which were associated with death in our cases. Future prospective studies are needed to assess the mechanism of severe pertussis.
The management of severe pertussis is extremely challenging, especially when accompanied by PH and hyperleukocytosis. Most of the patients received macrolide treatments, and some severe patients received immunoglobulin therapy. Previous literature has reported that using inhalation of NO to treat neonatal PH can significantly shorten treatment time and reduce mortality [
28]. However, in our study, PH did not improve in four infants who received inhalation of NO. The traditional approach to reduce pulmonary vascular resistance, such as inhalation of NO, may fail because of hyperviscosity and vascular obstruction [
22]. Our data suggest that inhalation of NO may not be useful in pertussis-related PH but more samples are needed to confirm this hypothesis. Exchange blood transfusion, which is frequently conducted in the NICU, was first published in a patient with severe pertussis [
29]; thereafter, exchange blood transfusion has been reported in multiple case series and case reports of severe pertussis to reduce the level of the total leucocyte count [
13,
18,
22]. In our study, 11 patients with severe pertussis in the PICU/NICU underwent exchange blood transfusion, eight patients survived and three died. Exchange blood transfusion can reduce the levels of the leucocytosis and thrombosis in patients with severe pertussis, thereby improving the severity of PH. In addition, two patients who underwent exchange blood transfusion in the early stages of disease also underwent ECMO therapy when exchange blood transfusion did not seem to work. One patient survived, while the other died because of refractory heart failure and septic shock. The use of ECMO in severe pertussis has been reported with some success in small series [
18,
22,
30], with a survival rate of only 30%, and the mortality remains higher than that for other indications for ECMO [
30]. Further and larger prospective studies are urgently needed to confirm the critical time of exchange blood transfusion and to define the optimal use of ECMO in severe pertussis in order to decrease its mortality.
The study limitation was lacking the attentions about the macrolide resistance of pertussis. The previous lectures showed a strikingly high rate of macrolide resistance in B. pertussis in China (85–91.9%), especially in erythromycin [
31,
32]. The macrolide resistance might be one reason for the high frequency of severe pertussis in our hospitalized patients. But B. pertussis culture and drug sensitivity were not tested in our hospital. Further investigations should be undergoing to test the macrolide resistant B. pertussis genes which may help to reduce the high frequency of severe pertussis and mortality in these hospitalized pertussis patients in China.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.