Introduction
CHD is the most common congenital malformation in infants and young children. Reference Liu, Chen and Zühlke1 Haemodynamically significant CHD is a well-known high-risk factor for severe pulmonary infection. Reference Kang, Lee and Kim2 . Viruses and bacteria are the most common pathogens responsible for hospital-acquired pulmonary infectious diseases. Reference Posten and Reed3 For children with CHD, pneumonia is prone to recurrence, is difficult to control after infection, and has a high mortality rate. Reference Zhao, Cui, Zhao and Yao4 Therefore, research into such children is of great significance.
The reported mortality rate of children with CHD complicated by pneumonia is 11.9–56%, often due to uncontrolled heart failure during conservative treatment. Reference Mohsin, Haque, Shaikh, Bano and Hasan5,Reference Luo, Qin and Wang6 . Surgery in infants with CHD and pneumonia is typically postponed until the pneumonia is fully treated. However, some infants die of respiratory and cardiac failure before surgical treatment can be administered. With the advancements in surgical techniques, myocardial protection methods, general anaesthesia, and intensive care management, the effectiveness of paediatric cardiac surgery has significantly improved. The current approach promotes early surgery. Reference Pagowska-Klimek, Pychynska-Pokorska, Krajewski and Moll7
A considerable number of children with CHD and preoperative pneumonia require longer stays in the ICU after surgery, which poses a significant challenge to intensive care providers. Prolonged hospitalisation also brings a substantial economic burden to hospitals. However, few studies have investigated the risk factors associated with prolonged ICU stays in these children. Reference Shi, Zhao and Liu8 In addition, differences in endpoint definitions (i.e., ICU stays) can make comparing these studies challenging. In some studies, the postoperative ICU stays after CHD surgery were relatively long. For instance, in the study by Shi et al., the median ICU stay was 7 days, Reference Shi, Zhao and Liu8 while in the study by Manrique et al., the median ICU stay was 9.5 days. Reference Manrique, Feingold, Di Filippo, Orr, Kuch and Munoz9 Recognising these gaps, this study conducted a single-centre review of the pre-, intra-, and postoperative risk factors in children with CHD complicated by pneumonia, focussing on those with postoperative ICU stays ≥7 days (representing the 50th percentile in this study).
Materials and methods
Ethical statement
This study was approved by the Medical Ethics Committee of Beijing Children’s Hospital (No. 2023-E-079-R). This retrospective analysis was conducted at the Beijing Children’s Hospital, affiliated with the Capital Medical University in China. Due to the retrospective nature of this study, the Institutional Review Committee waived the need for informed consent.
Study population
This study included 496 children with CHD and preoperative pneumonia who underwent cardiac surgical correction between 2017 and 2022. Most children were initially hospitalised for pneumonia. The criteria for ICU discharge were as follows: stable cardiac and pulmonary function, sufficient oxygen support through nasal catheters, and no severe infection indicator abnormalities (body temperature, white blood cell, C-reactive protein, etc.). The median ICU stay for all patients was 7 days (range, 1–111 days). The patients were divided into two groups: the short-term group included 296 children who stayed in the ICU for less than 7 days, and the long-term group included 200 children who were in the ICU for 7 days or more. The diagnosis of pneumonia was based on abnormalities such as cough, fever, shortness of breath (respiratory rate >50/min for infants aged 3–12 months), wet rales or wheezing detected by auscultation during chest examination, and spots or plaque infiltration on chest radiographs before admission or within 72 h after admission. Reference Shi, Zhao and Liu8 The diagnosis of severe community-acquired pneumonia was based on the diagnostic criteria proposed in the clinical practice guidelines of the Pediatric Infectious Disease Society and the American Infectious Disease Society. Reference Manrique, Feingold, Di Filippo, Orr, Kuch and Munoz9
Procedural details
The assessment of pre-, intra-, and postoperative risk factors was based on the completeness of previously published analyses, clinical judgements, and data collection. Reference Luo, Qin and Wang6,Reference Shi, Zhao and Liu8,Reference Lozano-Espinosa, Huertas-Quiñones and Rodríguez-Martínez10
Preoperative risk factors included age, sex, weight, chromosomal abnormalities, premature birth, malnutrition, pulmonary hypertension, tracheal stenosis, bacterial infection, viral infection, severe pneumonia, and bronchopulmonary dysplasia. Intraoperative risk factors assessed included congenital heart surgery risk (Risk Adjustment for Congenital Heart Surgery-1 grade) Reference Jenkins, Gauvreau, Newburger, Spray, Moller and Iezzoni11 and cardiopulmonary bypass. Postoperative risk factors included delayed sternal closure (the sternum remained open in the operating room or was reopened in the ICU), extracorporeal membrane oxygenation assistance, and failure to extubate. Pulmonary hypertension was defined as an average pressure of greater than 20 mmHg in the pulmonary artery. Reference Mocumbi, Humbert and Saxena12
According to the guidelines of the Centers for Disease Control and Prevention, sputum samples were collected using the sterile negative-pressure suction method. Samples collected within 24 h of admission were tested for common respiratory virus antigens using direct immunofluorescence assays. Each sample was immersed in 1 mL of sterile physiological saline, after which 0.1 mL of this solution was inoculated into a culture dish. The samples were then incubated at 37 °C for 24 h, colonies were isolated and counted, and a microbial detector was used for bacterial identification. The positive cut-off point for quantitative cultivation of pathogenic bacteria was ≥106 CFU/mL. Reference Horan, Andrus and Dudeck13
Statistical analysis
Continuous variables were expressed as the medians with interquartile ranges, and categorical variables were expressed as numbers and percentage (%). Before analysis, the normal distribution of all continuous variables was evaluated. As most variables exhibited a non-normal distribution, we conducted nonparametric univariate analysis using the Kruskal–Wallis test and the Mann–Whitney U test. For qualitative variables, the chi-square test was performed. Statistical significance was defined as p < 0.05.
A logistic regression model was used for multivariate analysis. The Wald forward selection method was utilised, with an entry probability of 0.1 and an exit probability of 0.05. Severe illness was defined as a postoperative ICU stay of ≥7 days. The predictive variables included in the model were age, sex, body weight, cardiopulmonary bypass, delayed chest closure, viral infection, bacterial infection, severe pneumonia, pulmonary hypertension, tracheal stenosis, and CHD score. The results of multivariate analysis are expressed as the relative risk and 95% confidence interval. Multivariable linear regression analysis was also applied to assess the prognostic value of risk factors for ICU stays, expressed as linear regression coefficient (β). All statistical tests were two-tailed, and the significance level was set at P < 0.05. Data were analysed using the SPSS version 24.0 (IBM Corp., Armonk, NY, USA) statistical software package.
Results
Baseline characteristics
There were no significant differences in gender between the groups. The average age and weight values of the long-term ICU stay group were significantly lower than those of the short-term ICU stay group (p < 0.01). Half of the children in the long-term ICU stay group were under 3 months old and weighed less than 5 kg. The Risk Adjustment for Congenital Heart Surgery-1 grade revealed no significant differences between the groups (p = 0.76). However, significant differences were observed between the two groups in terms of cardiopulmonary bypass time, blockage time, hospitalisation time, ventilator time, pneumonia severity, pulmonary arterial hypertension, and extubation failure. There were no significant differences in proportion of cardiopulmonary bypass surgery, delayed chest closure, tracheal stenosis, premature birth, malnutrition, bronchopulmonary dysplasia, chromosomal abnormalities, or extracorporeal membrane oxygenation support. During the study period, the in-hospital mortality rates in the long-term and short-term ICU admission groups were similar and were 8.5% and 4.1%, respectively (p = 0.39). The main causes of death were severe infections and multiple organ system failure (Table 1).
Baseline clinical characteristics
RACHS-1 = Risk Adjustment for Congenital Heart Surgery-1; ECMO = extracorporeal membrane oxygenation; IQR = interquartile range.
Differences in bacterial infection were observed between the groups. The positive rate of bacterial infection culture in the long-term group was significantly higher than that in the short-term group (p < 0.01). In the long-term group, Klebsiella pneumoniae was the most common gram-negative bacteria, while Staphylococcus aureus was the most common gram-positive bacteria. In the short-term group, Haemophilus influenzae was the most common gram-negative bacteria, while S. aureus was the most common gram-positive bacteria. Viral infections showed no difference between the groups (p = 0.37), with respiratory syncytial virus being the most frequently identified. Additionally, there was no significant difference in the incidence of mixed infections between the groups (p = 0.28; Table 2, Figure 1).
Distribution of bacterial and viral infections.
Bacterial, viral, and mixed infections in two groups of patients
Risk factor analysis
Univariate analysis
Demographic data and preoperative, intraoperative, and postoperative variables were used as potential risk factors for prolonged ICU stay in the univariate analysis. Patients in the long-term ICU stay group were significantly younger, had lower body weights, longer cardiopulmonary bypass and blockage times, a higher incidence of severe pneumonia and pulmonary hypertension, and a higher risk of extubation failure. However, there was no significant correlation between sex distribution and prolonged ICU stay.
Multivariate analysis
Multivariate logistics regression analysis showed that low body weight (p = 0.001), bacterial infection (p = 0.001), severe pneumonia (p = 0.001), and Risk Adjustment for Congenital Heart Surgery-1 (p = 0.001) were independent risk factors for prolonged stay in ICU after cardiac surgery in children with pneumonia (Table 3). Multivariate linear regression analysis showed that bacterial infection (p = 0.001), severe pneumonia (p = 0.04), and Risk Adjustment for Congenital Heart Surgery-1 (p = 0.001) were risk factors for longer ICU stays (Table 4).
Multivariate logistic regression analysis of the influence of perioperative risk factors on prolongation of ICU length of stay over 7 days
RACHS-1 = Risk Adjustment for Congenital Heart Surgery-1.
Linear regression regression analysis of the influence of perioperative risk factors on prolongation of ICU
RACHS-1 = Risk Adjustment for Congenital Heart Surgery-1.
Discussion
Zhang et al. Reference Ao14 reported that patients with severe pneumonia and concomitant CHD have higher mortality rates. Furthermore, community-acquired pneumonia has significant fatal consequences in children with various CHDs. Reference Zhang, Lu and Yang15 Moreover, CHD increases the risk of severe community-acquired pneumonia. These findings emphasise the need for preventive measures against pneumonia in children with chronic disease, particularly CHD. However, the impact of CHD on disease severity varies among different age groups, with CHD emerging as a crucial predictor of severity in paediatric patients under 2 years of age. Thus, in most cases, CHD is recommended to be surgically treated within 2 years of birth. Reference van der Bom, Luijendijk, Bouma, Koolbergen, de Groot and Mulder16
Children with CHD and pneumonia may undergo one of the two treatment strategies: (1) medication and corrective surgical intervention during the same hospitalisation period or (2) medical management, discharge, and elective surgical repair. Many physicians in developing countries choose the second strategy. Concerns about employing cardiopulmonary bypass in critically ill infants with active pulmonary infections arise from concerns regarding the exacerbating lung inflammation and increasing potential infections due to immune suppression.Reference Bhatt, Roth and Kumar17 In this context, components such as complements, kinins, endotoxins, proinflammatory cytokines, and tumour necrosis factors are believed to mediate lung injury and alter cellular and humoral immune responses, thereby exacerbating pneumonia symptoms. Consequently, pulmonary complications often occur in young children and lead to delayed recovery after cardiac surgery.
Most centres in developing countries have relatively limited human and material resources, preventing the adoption of proactive policies to promptly treat heart defects in young infants. Typically, limited resources are allocated to patients who are more likely to achieve better outcomes. In line with these considerations, we have adopted a policy of early surgical correction because we believe that, in cases of haemodynamically significant CHD combined with pneumonia, long-term attempts to completely resolve preoperative respiratory infections will not be successful.
Our study showed that the long-term ICU stay group mainly comprised young, low-weight infants, with significant differences in ICU stays among the different weight groups. Furthermore, low body weight was shown to be an important predictor of prolonged ICU stay in our multivariate logistics regression analysis, which is consistent with previous studies. Reference Shan, Shi and Chen18,Reference Williams, Zhu and Grijalva19 Specifically, we found that the number of infants aged under 3 months and weighing less than 5 kg who stayed in the long-term group was significantly higher than that of older and heavier children. This may be attributed to an underdeveloped immune system Reference Hill, Kadow and Long20 and immature respiratory system, which make the infant more susceptible to this serious disease.
Most children with severe pneumonia complicated by CHD have left-to-right shunts. Specifically, 95% of the cases in this group exhibit left-to-right shunts, including those associated with ventricular septal defects, patent ductus arteriosus, atrial septal defects, right pulmonary arteries that originated from the aorta, and other similar conditions. Meanwhile, the remaining 5% consisted of cases with right-to-left shunts or vascular rings. Due to the presence of shunts at the atrial, ventricular, or great artery levels, these children are prone to pulmonary circulation congestion, pulmonary capillary exudation, and varying degrees of oedema in the alveolar interstitium and airway mucosa, which can lead to secondary bacterial infection. It is worth noting that the increase in pulmonary blood flow is also conducive to bacterial growth. Furthermore, these children often encounter early disease onset, incomplete lung development, and a heightened susceptibility to immune dysfunction, especially in cases with chromosomal abnormalities. In clinical practice, many children with CHD are prone to recurrent pneumonia and are more likely to develop severe pneumonia. However, due to the frequent incidence of refractory heart failure and pulmonary arterial hypertension in these children, the treatment effectiveness rate tends to be low. Therefore, timely adjustment and selection of sensitive antibacterial drugs based on clinical efficacy and drug sensitivity results can effectively improve the success rate in children with severe pneumonia and CHD while also improving cardiopulmonary function. Ultimately, when the general condition of the child improves and surgery becomes tolerable, timely intervention should be carried out to achieve treatment during a single hospitalisation period.
Early and repeated sputum cultures and pathogen susceptibility tests after admission are helpful for the rational and effective use of antibiotics in clinical practice. Notably, the main pathogens in these children are viruses and bacteria. Given that viral infection is usually transmitted from the upper respiratory tract and that viruses grow and reproduce in living cells, upper respiratory tract specimens can reflect lower respiratory tract viral antigens. In this group of patients (CHD combined with pneumonia), the total positivity rate of bacterial infections, mainly S. aureus and K. pneumoniae, was 36.1%. Moreover, the bacterial infection rate in the long-term ICU stay group showed significant differences compared to the short-term ICU stay group. Furthermore, gram-negative bacteria are the most common microorganisms, which is in line with other studies (42–65%). Reference Patria, Chidini and Ughi21
Among the participants in this study, the total positive rate of viral infection (mainly respiratory syncytial virus) was 10.2%, with no statistical difference between the two groups. Respiratory syncytial virus is a common pathogen that causes community-acquired pneumonia in children in the Chinese population. Reference Chen, Miao and Chen22 Additionally, this virus is the most common viral pathogen in children with severe community-acquired pneumonia. Reference Gong, Zhang and Luo23 In our study, respiratory syncytial virus accounted for 7.5% of the cases, with no significant difference between the two groups. Notably, respiratory syncytial virus infection has been previously reported as a risk factor for CHD-related death. Reference Hall, Weinberg and Iwane24 Therefore, for the majority of children infected with respiratory syncytial virus, we found that antipyretic and symptomatic treatments were administered for respiratory symptoms. After the pneumonia was cured, the patients were discharged for recovery and underwent elective surgical treatment. Hence, the proportion of children infected with the virus was relatively small in this study. These findings are consistent with those of Medrano et al. Reference Medrano López and García-Guereta25 , who evaluated the epidemiological behaviour of acute respiratory infections in 167 children under 2 years of age with CHD and haemodynamic responses. At the community level, most childhood pneumonia cases are caused by respiratory syncytial virus (approximately 29% of all pneumonia cases). Reference Rhedin, Lindstrand and Hjelmgren26 Additionally, it is worth noting that the mortality rate caused by viral pathogens in children with pneumonia is very low, with only 6.6% of deaths attributed to respiratory syncytial and influenza viruses. In this context, most mild-to-moderate pneumonia cases in children are caused by viruses, whereas the most severe and fatal cases are caused by bacterial infections.
Jenkins et al. published the Risk Adjustment for Congenital Heart Surgery-1 in 2002, which was one of the earliest standardised risk adjusted methods for CHD surgery. Reference Jenkins, Gauvreau, Newburger, Spray, Moller and Iezzoni11 Higher Risk Adjustment for Congenital Heart Surgery-1 classification was associated with longer ICU stays. Reference Larsen, Pedersen, Jacobsen, Johnsen, Hansen and Hjortdal27 In this study, multivariate logistics regression analysis and multiple linear analysis revealed that a higher Risk Adjustment for Congenital Heart Surgery-1 grade was an independent risk factor for prolonging hospitalisation time in the ICU.
In this study, the two groups showed significant difference in pulmonary hypertension. The determination of pulmonary hypertension in this group of children mainly relied on echocardiography, Reference Lopez, Colan and Frommelt28 although it was not an independent risk factor for longer ICU stay. However, there were significant differences in both the average and total length of ICU stay between the two groups.
Combining multivariate logistics regression analysis and multiple linear regression analysis, our results indicate that bacterial infections, severe preoperative pneumonia, and Risk Adjustment for Congenital Heart Surgery-1 can prolong the ICU duration in children with CHD and pneumonia. Better identification of the risk factors associated with long-term postoperative ICU stays in these children, along with timely diagnosis and treatment of respiratory infections in high-risk populations, can effectively shorten ICU stays and improve resource utilisation.
This study has some limitations. First, it was a single-centre, small-scale case study, which may hinder the application of the current results to other institutions. Second, this was a retrospective study that may introduce potential misclassification bias. This study is a preliminary systematic evaluation of the impact of perioperative variables on the duration of ICU stay in children with CHD complicated by pneumonia. Larger prospective multicentre studies are necessary to further investigate this important clinical issue.
Data availability statement
The data used to support the findings of this study are available from the corresponding author upon request.
Acknowledgements
We are grateful to the staff in the Department of Cardiovascular Surgery and Cardiac Intensive Care Unit, Beijing Children’s Hospital and Xinrui Liu for their research contributions.
Financial support
None.
Author contributions
Nan Ding contributed to all aspects of this study, including study concept and design, data acquisition, statistical analysis and interpretation, and drafting and revising the manuscript. Lei Shen, Zhiqiang Li, Jian Guo, Xiaofeng Li, and Hanlu Yi contributed to data acquisition and ethical issues. All authors have approved the final article.
Competing interests
None.
Ethical standard
This study was approved by the Medical Ethics Committee of Beijing Children’s Hospital (No. 2023-E-079-R).
