Abstract
Aims
To describe the demographic, clinical, and laboratory characteristics as predictors of mortality of children with severe community-acquired pneumonia (CAP) admitted to paediatric intensive care units in Sri Lanka.
Methods
We conducted a multi-center retrospective study among children with severe CAP admitted to all paediatric intensive care units (PICUs) from April 2018 to March 2020. The primary outcomes were mortality and length of PICU stay. Univariate and multivariate logistic regression analyses were used to identify the potential predictors of mortality.
Results
Records of 266 children with community-acquired pneumonia were studied. Mortality rate was 30.8%. 72% had at least one comorbidity. Mortality was higher among children < 1 year compared to children ≥ 5 years (aOR:5.69, CI:1.297,24.931, p = 0.02). Also, septic shock (aOR:12.04, CI:5.287,27.416, p < 0.001), invasive ventilation (aOR:4.92, CI: 2.34, 10.387, p < 0.001), comorbidities (aOR:7.75, CI:2.563,23.437, p < 0.001), leucopenia/leukocytosis (< 4 or > 15 × 109) (aOR:2.16, CI:1.030,4.531, p = 0.04) and multilobar involvement (aOR:3.44, CI:1.128,10.470, p = 0.03) increased the risk of mortality. The effect size was highest in cases involving septic shock unresponsive to fluid boluses and the presence of one or more comorbid conditions, as identified through regression analysis.
Conclusions
The mortality rate among children with community-acquired pneumonia was high in this study population. Younger age, the need for invasive ventilation, septic shock, underlying comorbidities and multilobar involvement were identified as statistically significant independent predictors of mortality. However, the high mortality rate observed may also reflect other contributing factors not captured within the scope of this study, highlighting the need for further research to explore these influences.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12887-026-06517-y.
Keywords: Child, Intensive care units, Paediatric, Pneumonia, Sri lanka
Introduction
Pneumonia remains a significant cause of hospitalisation and a leading cause of death among children, particularly those under five years of age. According to the World Health Organisation (WHO), deaths are highest in southern Asia and sub-Saharan Africa [1]. Most of these deaths have occurred in resource-poor settings where adequate access to health resources and interventions is not optimal. In Sri Lanka, out of the total number of 1969 deaths in children from one month to 14 years in 2019, 166 (8.4%) were due to pneumonia [2]. Most of these pneumonia-related deaths (70%) occurred in children under five years, underscoring the vulnerability of this group [2].
Severe pneumonia can lead to complications such as hypoxemia, respiratory failure, pleural effusion, lung abscess, empyema and even death [3]. Treating such patients can be challenging and requires hospitalisation, hemodynamic support, and respiratory management with invasive/noninvasive ventilation. Despite these challenges, with timely intervention, most children with pneumonia recover without complications [3].
While many studies worldwide have explored predictors of mortality from severe pneumonia in children, there is a relative scarcity of such research specific to the Sri Lankan context [4, 5]. The Infectious Diseases Society of America (IDSA) issued CAP management recommendations that included adult intensive care unit admission criteria in 2007 [6]. In 2011, the IDSA and the Paediatric Infectious Diseases Society (PIDS) inferred severity criteria from the adult guidelines for paediatric use [7]. However, factors such as age and comorbid conditions, crucial in determining the severity of pneumonia, need to be adequately addressed when it is used for children. Furthermore, mortality rates and local admission criteria and facilities differ significantly between developing and high-income countries [8], necessitating the development of simple models for predicting outcomes in severe CAP patients.
Sri Lanka is a lower middle-income country with an impressive healthcare system with some health indices comparable to those of developed countries. This has been achieved with relatively low healthcare spending. Sri Lanka is divided into nine administrative provinces, each served by an efficient network of hospitals that provide universal free healthcare. Despite the clinical significance of severe community-acquired pneumonia, there is a lack of data on its burden. Thus, gaining a clearer understanding of the caseload and severity of paediatric cases is crucial for policymakers. Such insights can guide the improvement of paediatric intensive care facilities nationwide.
With this background, we aimed to describe the outcome and potential predictors of mortality of children with severe CAP admitted to paediatric intensive care units in Sri Lanka, fill the knowledge gap and enhance understanding of this issue in a resource-poor setting.
Methods
Study design and settings
This was a retrospective analytical observational study investigating outcomes and potential predictors of mortality among children aged > 1 month to 14 years with community-acquired pneumonia admitted to paediatric intensive care units (PICUs) from April 2018 to March 2020. The sample size was calculated based on the rule of having at least 10 cases per variable [9]. With 18 variables being assessed, the minimum required sample size was determined to be 180.
This study included all designated PICUs except for a PICU at the National Cancer Institute in Sri Lanka. These units are located in four of the nine provinces and are staffed by paediatric intensivists: Western Province (Lady Ridgeway Hospital), Southern Province (Teaching Hospital Karapitiya), Central Province (Sirimavo Bandaranayke Specialised Children’s Hospital) and North-Central Province (Teaching Hospital Anuradhapura). The total bed capacity of all these centres is 41. The total number of PICU admissions during the study period was 3783. However, we acknowledge that some critically ill children, particularly in regions without dedicated PICUs, are managed in adult ICUs. There may be one or more beds allocated for paediatric patients within adult ICUs, which are not formally categorised as PICUs, and we have not included those children with severe pneumonia managed in such units into this sample. A wide range of critically ill children aged one month to 14 years including all acute medical emergencies, surgical emergencies, and preoperative and postoperative patients requiring intensive monitoring and care are admitted to these units. In addition, selected oncology-related patients and all post-transplant patients are managed in these units, depending on the clinical need and the infrastructure available at each hospital.
The exclusion criteria were children with hospital-acquired pneumonia, hospital admissions within the previous two weeks, and incomplete/misplaced case records. Ethical clearance was obtained from the Ethical Review Committee of the Sri Lanka College of Paediatricians (SLCP/ ERC/2022/24).
The proforma was developed based on similar previous studies in literature [10, 11]. Its purpose was to extract socio-demographic, clinical (symptoms, chronic medical/surgical conditions, vital signs, and physical examination findings), laboratory, and radiological data. Also, the respiratory support received within the first 24 h after admission was noted. The adaptation process of the proforma involved careful consideration of the research objectives, ensuring that the instrument was appropriate to capture the data for the present study. A panel of experts (paediatricians, pulmonologists and intensivists) reviewed the questionnaire to ensure its comprehensibility and relevance to the study objectives. Randomly selected patient records with the diagnosis of pneumonia outside the study period at the same study sites were used for the pilot study. It was conducted to confirm the clarity, structure, ease of use, and overall usability of the questions. All these steps ensured reliability and validity.
Variables and definitions
Community-acquired pneumonia (CAP) is characterised by an infection of the lung tissue acquired outside of hospital settings. CAP was characterised by the presence of fever, tachypnoea, breathlessness or difficulty in breathing, cough, wheeze or chest pain, with or without laboratory investigations and chest X-ray evidence of pneumonia acquired outside of a hospital setting [12, 13]. Tachypnea was defined as a respiratory rate of > 60 in 1–2 months, > 50 in 2–12 months,>40 in 1–5 years and > 20 in > 5 years [12].
Bronchiolitis was defined as coryzal prodrome lasting for 1 to 3 days, with persistent cough, tachypnoea or chest recessions (or both) and wheeze or crackles on auscultation (or both) [13]. Children diagnosed with bronchiolitis or exacerbations of asthma were not categorised as having CAP unless there was a coexisting clinical or radiological diagnosis of pneumonia, as per the treating clinician’s judgment and documented in the patient records.
In children with underlying cardiac disease, differentiation between cardiac failure and pneumonia was based on a combination of clinical features, radiological findings, and echocardiographic data when available. Features such as hepatomegaly, gallop rhythm, and cardiomegaly and pitting oedema in older children were considered to support a diagnosis of cardiac failure [14].
Children who showed symptoms of pneumonia after 48 h of admission or had been hospitalised in the two weeks prior were considered to have hospital-acquired pneumonia (HAP) and not included in the sample [13].
When diagnostic uncertainty existed, cases were reviewed by the same panel of experts who reviewed the proforma, to enhance diagnostic accuracy. During this review process, standard clinical definitions and diagnostic criteria mentioned above to differentiate CAP from similar conditions such as heart failure, bronchiolitis, and hospital-acquired pneumonia were used.
In addition, for patients with chronic lung disease, asthma, or cardiac conditions, the diagnosis of CAP was made based on a change from baseline clinical status, new or worsening respiratory symptoms, and findings supportive of an acute infectious process. In instances where the diagnosis was uncertain, the same expert committee that reviewed the questionnaire also evaluated the clinical and radiological findings to reach a conclusion.
The primary outcome variable was mortality and length of PICU stay. The predictor variables were defined based on the PIDS-IDSA [7] and World Health Organisation severity assessment guidelines [15]. Among the clinical variables on admission to ICU, respiratory rate, effort of breathing (e.g., retractions, nasal flaring, grunting, dyspnea), hypoxia (documented oxygen saturation of < 90%), fluid-unresponsive septic shock (defined as the presence of sustained tachycardia and hypotension relative to age norms, or the requirement for inotropes to support circulation despite receiving adequate fluid resuscitation, i.e. ≥ 40 mL/kg of isotonic fluids and altered mental status (drowsiness and decreased level of consciousness as recorded by the treating physicians) were noted.
Underweight was defined as weight-for-age < − 2 SD [16]. Laboratory variables included were C-reactive protein level (< 100 mg/L vs. ≥ 100 mg/L) [17], leucopenia/leukocytosis (< 4 or > 15 × 103 /dL) [18], blood urea nitrogen (> 7.1mmol/L), aspartate transaminase levels (> 33 units/L), metabolic acidosis (HCO3 < 15 with pH < 7.35) and serum sodium (< 135 mEq/L). Laboratory parameters (liver and renal functions, inflammatory markers) were extracted if they were available on admission to PICU or performed within the first 24 h after admission.
Need for invasive ventilation via a nonpermanent artificial airway and noninvasive ventilation (e.g., high-flow nasal cannula therapy, continuous positive airway pressure, or bilevel positive airway pressure) within the first 24 h of admission, radiological evidence of multilobar involvement (radiological evidence of infiltrates, opacities, or consolidations noted in more than one lobe) and effusions were noted. Chest X-rays were interpreted by a consultant radiologist based on WHO criteria for interpreting chest X-rays [19]. PIDS-IDSA, Paediatric Early Warning Score (PEWS) [20] and the ratio of partial pressure of oxygen in arterial blood to the fraction of inspiratory oxygen concentration (PaO2/FiO2 ratio) were not assessed as there were more than 10% missing data.
Data collection procedure
At the time of a patient’s discharge/transfer/death, the standard procedure is to record the diagnosis/diagnoses in the ICU register and forward the bed head ticket/medical record to the medical record room of the hospital within 24 h. Thus, the eligible subjects were identified through a two-step process. First, the ICU register was screened to identify all cases of pneumonia. Subsequently, a review of bed head tickets/medical records was done to extract information on the children with community-acquired pneumonia.
The data was collected by four trained enumerators who were medical graduates awaiting their internship training. Emphasis was placed on the importance of data collection consistency, minimising biases, and addressing common challenges during data collection. Data gathering was conducted using Kobo Collector, a mobile data collection platform. The variables were defined, and a coding guide was provided. During data collection, the enumerators were closely monitored and guided. The data collected was reported to the principal investigator regularly.
Data analysis
Collected data was analysed, processed, and stored using Statistical Package for Social Sciences (SPSS) version 22. Multiple imputations were made if missing data was < 5% of the records. We used multiple imputation methods in SPSS. It helped to fill in missing values by using the information from other data points in the dataset. The descriptive statistics were summarised in tables. A bi-variable logistic regression was done for each predictor variable. Variables with a p-value of < 0.1 and other variables of interest were fitted into a multivariable logistic regression model with the forward selection method (Age, metabolic acidosis, invasive ventilation, septic shock, radiological evidence of effusion. multilobar involvement, aspartate transaminase > 33 IU/L, total white cell count < 4 × 109/L or > 15 × 109/L, blood urea nitrogen > 7.1 mmol/L, and comorbidities). Multi-collinearity was checked using variance inflation factor (VIF > 5) with pair-wise correlations (> 0.5). The Model fitness was checked using the Hosmer-Lemeshow goodness of fitness test and R2 values. In multivariable analysis, p < 0.05 was considered statistically significant.
Results
During the study period, 298 children with CAP were admitted to the four PICUs. Two hundred sixty-six children were included in the analysis, and thirty-two were excluded (Table 1). The percentages of children with CAP from Lady Ridgeway Hospital, Teaching Hospital Karapitiya, Sirimavo Bandaranayke Specialised Children’s Hospital and Teaching Hospital Anuradhapura) were 55.3%, 8.3%, 20.7%, and 15.8%, respectively. Lady Ridgeway Hospital is the premier tertiary care children’s hospital in Sri Lanka and receives referrals from across the country, resulting in a higher patient load. In contrast, Teaching Hospital Karapitiya serves primarily the Southern Province with a more limited catchment area. Sirimavo Bandaranaike Specialized Children’s Hospital serves a comparatively larger population than Karapitiya and Anuradhapura, which is likely to contribute to the higher proportion of cases from that site.
Table 1.
Recruitment flow
| Number of records | |
|---|---|
| The total number of PICU admissions from April 2018 to March 2020 | 3783 |
| Total number of admissions with physician-diagnosed CAP | 298 |
| Incomplete data | 8 |
| Hospital admission within the previous 2 weeks | 13 |
| Misplaced records | 11 |
| Final number analysed | 266 |
Most (60%) children were below one year. Almost half received invasive ventilation (42.8%). 72% had at least one comorbidity. Forty-five children (16.9%) had more than one comorbidity. A greater percentage of non-survivors had more than one comorbidity than survivors (28% vs. 12%). Underweight was the most common (52.6%), followed by cardiac anomalies (22.6%) (Table 2). A significantly high percentage of children with comorbidities were underweight compared to children without comorbidities (65% vs. 38%, p < 0.001). Additionally, 24% (n = 35) of underweight children were born prematurely, which likely contributed to their underweight status.
Table 2.
Demographics and comorbidities among survivors and non-survivors
| Variable | All subjects (n = 266) | Survivors (n = 184) | Non-survivors (n = 82) | p-value |
|---|---|---|---|---|
| Age | ||||
| < 12 months | 160 (60.1) | 102 (55.4) | 58 (70.7) | 0.02 |
| 12–59 months | 76 (28.6) | 56 (30.4) | 20 (24.3) | 0.31 |
| ≥ 60 months | 30 (11.2) | 26 (14.1) | 04 (4.8) | 0.03 |
| Gender (Male) | 147 (55.2) | 101 (54.8) | 46 (56.0) | 0.93 |
| Prematurity | 51 (19) | 32 (17.4) | 19 (23.2) | 0.27 |
| Comorbidities (≥ 1) | 191 (71.8) | 115 (62.5) | 76 (92.6) | < 0.001 |
| Comorbidities (> 1) | 45 (16.9) | 22 (11.9) | 23 (28.0) | < 0.001 |
| Underweight (WAZ <-2SD) | 140 (52.6) | 83 (45.1) | 57 (69.5) | - |
| Chronic lung disease | 19 (7.1) | 09 (4.8) | 10 (12.1) | - |
| Asthma | 19 (7.1) | 16 (8.6) | 03 (3.6) | |
| Neurological disorders | 25 (9.3) | 20 (10.8) | 05 (6.0) | - |
| Cardiac disorders | 60 (22.5) | 28 (15.2) | 32 (39.0) | - |
| Endocrine disorders | 05 (1.8) | 02 (1.0) | 03 (3.6) | - |
| Chronic kidney disease | 05 (1.8) | 02 (1.0) | 03 (3.6) | - |
| Immunodeficiency | 08 (3.0) | 02 (1.0) | 06 (7.3) | - |
| Hemoglobinopathies | 08 (3.0) | 05 (2.7) | 03 (3.6) | - |
| Malignancy | 01 (0.3) | 00 (0.0) | 01 (1.2) | - |
| Chronic liver disease | 00 (0.0) | 00 (0.0) | 00 (0.0) | - |
| Genetic disease | 20 (7.5) | 10 (5.4) | 10 (12.1) | - |
| Other conditions | 40 (15.0) | 22 (11.9) | 18 (21.9) | - |
| Clinical data on admission to PICU | ||||
| Tachypnea | 221 (83.0) | 156 (84.7) | 65 (79.2) | 0.07 |
| Increased effort of breathing | 201 (75.5) | 141 (76.6) | 60 (73.1) | 0.54 |
| SpO2 < 90% in air | 180 (67.6) | 127 (69.0) | 53 (64.6) | 0.51 |
| Altered mental status | 21 (7.8) | 16 (8.6) | 05 (6.0) | 0.47 |
| Septic shock | 64 (24.0) | 15 (8.1) | 49 (59.7) | < 0.001 |
| Laboratory/Radiological data | ||||
| C-reactive protein level ≥ 100 mg/L | 26 (9.7) | 18 (9.7) | 08 (9.7) | 0.99 |
| Leucopenia/leukocytosis (< 4 or > 15 × 109) | 118 (44.3) | 70 (38.0) | 48 (58.5) | 0.001 |
| Blood urea nitrogen (> 7.1mmol/L) | 70 (26.3) | 42 (22.8) | 28 (34.1) | 0.053 |
| Aspartate transaminase levels (> 33 IU/L) | 81 (30.4) | 47 (25.5) | 34 (41.4) | 0.009 |
| Serum sodium (< 135mEq/L) | 47 (17.6) | 28 (15.2) | 19 (23.1) | 0.12 |
| Metabolic acidosis (HCO3 < 15, pH < 7.35) | 36 (13.5) | 18 (9.7) | 18 (21.9) | 0.007 |
| Pleural effusion | 23 (8.6) | 13 (7.0) | 10 (12.4) | 0.17 |
| Multilobar involvement | 26 (9.7) | 13 (7.0) | 13 (15.8) | 0.03 |
| Respiratory support | ||||
| Invasive ventilation | 175 (65.8) | 102 (55.4) | 73 (89.1) | < 0.001 |
| Noninvasive positive pressure ventilation | 46 (17.2) | 37 (20.1) | 09 (10.9) | 0.07 |
The outcome of the children with CAP
The PICU mortality rate among children was 30.8% (n = 82). The median survival time was 24 days (IQR25 − 75: 45, 9.75), and the median length of PICU stay was 6 (IQR25 − 75: 4, 12) days (Fig. 1). We compared the total duration of hospital stay between survivors and non-survivors using the Mann-Whitney U test. The difference was not statistically significant (p = 0.56).
Fig. 1.
Kaplan–Meier estimates of cumulative survival probability of children with community-acquired pneumonia
Predictors of mortality
Logistic regression was carried out to examine the independent predictors of mortality. The model fitness was examined: X2 = 123.26, DF = 7, p < 0.001) and R2 = 0.534 and Hosmer-Lemeshow goodness of fit was not significant (p = 0.32). In the final regression model, the covariates explained 56% of the variation in PICU mortality. The variables that showed a significant statistical association were age, septic shock, invasive ventilation, multilobar involvement, WBC count, and comorbidities (Table 3). The strongest association with mortality was observed in the presence of septic shock unresponsive to fluid boluses, with an adjusted odds ratio of 13.7 (CI: 6.088–30.830, p < 0.001). Children with one or more comorbid conditions were at seven times higher risk of dying than children without comorbidities. Only 3 out of 19 children with asthma died, compared to 10 out of 19 children with chronic lung disease. It revealed that having chronic lung disease significantly increased the likelihood of mortality compared to having asthma (p = 0.03). Although respiratory rate, blood urea nitrogen, and metabolic acidosis were significantly associated with mortality in simple logistic regression, they were not significant in multiple logistic regression.
Table 3.
Predictors of mortality by using multivariable logistic regression
| AOR | 95% CI | p-value | ||
|---|---|---|---|---|
| Lower | Upper | |||
| Age | ||||
| ≥ 5 years | Ref | |||
| Infants | 5.22 | 1.158 | 23.558 | 0.03 |
| 1-5years | 3.78 | 0.782 | 18.342 | 0.09 |
| Septic shock | 13.7 | 6.088 | 30.830 | < 0.001 |
| Invasive ventilation | 3.87 | 1.564 | 9.555 | 0.003 |
| Multilobar involvement | 3.44 | 1.128 | 10.470 | 0.03 |
| Leucopenia/leukocytosis | 2.06 | 1.004 | 4.255 | 0.04 |
| Comorbidities ≥ 1 | 7.398 | 2.587 | 21.158 | < 0.001 |
AOR Adjusted odds ratio
The Kaplan-Meier survival curves (Figs. 2, 3 and 4) for comorbidities, septic shock and invasive ventilation show distinct survival patterns over time. The presence of one more comorbidity (17 vs. 30), the requirement for invasive ventilation (18 vs. 45 days) and associated septic shock (8 vs. 30 days) exhibit a lower probability of survival throughout the period.
Fig. 2.
Kaplan–Meier estimates of cumulative survival probability among CAP children with comorbidities
Fig. 3.
Kaplan–Meier estimates of cumulative survival probability among children with septic shock
Fig. 4.
Kaplan–Meier estimates of cumulative survival probability among CAP children required invasive ventilation
Discussion
This study, conducted in Sri Lanka, focused on community-acquired childhood pneumonia cases requiring intensive care, reporting a relatively high mortality rate of 31%. It also identified several factors that were significantly associated with mortality, including young age, the need for invasive ventilation, leukocytosis/leucopenia, multilobar involvement, septic shock, and the presence of comorbidities. These findings align with previous studies, highlighting the importance of these factors in evaluation and triage of critically ill children with pneumonia [21, 22].
The mortality rate reported in this study is relatively high. Lower mortality rates have been reported among children with CAP who required intensive care in other settings, for instance, 20.3% in Pakistan [23], and 12.3% and 9.3% in two separate studies from China [24, 25]. However, direct comparisons between studies should be interpreted with caution due to variations in study design, inclusion criteria, and duration of follow-up. Moreover, disparities in healthcare resources, such as the availability of paediatric intensive care beds, timely access to critical care, trained personnel, and supportive technologies, may significantly influence outcomes. Differences in immunisation coverage, as well as variations in socioeconomic conditions, health-seeking behaviours, and the burden of underlying comorbidities, may also contribute to the observed differences in mortality rates across countries.
Age emerged as a significant factor in predicting death, with infants being more vulnerable compared to older children. Organ immaturity, including an immature immune system, delay in recognising pneumonia due to non-specific presentations, and higher risk of complications are likely reasons for increased vulnerability at a young age [18]. Also, enteral antibiotic failure is one of the most vital factors associated with the increased vulnerability of infants [26].
The presence of comorbidities was also associated with an increased risk of mortality, with underweight being the most frequently observed comorbidity in our cohort. However, we acknowledge that underweight may reflect a range of underlying causes, including chronic illness or prematurity, and may not always be due to undernutrition. Similar to our findings, other studies have reported underweight, cardiovascular disease, and neurological conditions as common comorbidities in children with severe CAP [22, 25]. Notably, children with underlying cardiac disease experienced a substantially higher mortality rate compared to those without such conditions. This emphasises the need for early identification and aggressive management of children with cardiac comorbidities presenting with acute infections. It also highlights the potential need for developing tailored care protocols, such as timely PICU admission, enhanced clinical monitoring, and multidisciplinary involvement to improve outcomes in this high-risk population.
In this study, septic shock on admission emerged as the most significant predictor for mortality in children with severe community-acquired pneumonia. This finding is consistent with previous studies highlighting the prognostic significance of septic shock in pneumonia outcomes [27, 28]. Notably, septic shock is considered a major criterion for intensive care unit (ICU) admission in the PIDS/IDSA guidelines for managing CAP [7]. This signifies the importance of early detection of septic shock and early intensive care referral directly from the emergency department or from paediatric wards.
Furthermore, the need for invasive ventilation is included in most prognostic scores for paediatric and adult pneumonia patients [7, 28]. This underscores the importance of recognising the early signs of respiratory compromise and promptly initiating lung-protective invasive ventilation strategies, including optimal lung recruitment. Paediatric intensivists should consider these findings in their approach to managing severe CAP, advocating for more aggressive therapeutic interventions to improve patient outcomes.
Leukocytosis is often considered a predictor of the severity of childhood pneumonia [18, 27]. Conversely, leukopenia (white blood cell count of less than 4 × 109/L has been reported as an indicator of high mortality [18]. In the present study, the subjects with leukopenia or leukocytosis had twice the risk of mortality compared to those with a normal white blood cell count. Furthermore, the presence of bilateral multilobar pneumonia is identified as an independent predictor of mortality, highlighting the importance of vigilant monitoring and timely intervention in affected children [29].
Another important observation was that higher proportions of survivors had tachypnoea, increased effort of breathing, spO2 < 90 in air and altered mental status at the time of admission compared to patients who did not survive. One possible explanation is that despite survivors being presented with severe clinical signs, they may have responded better to intensive care management. In contrast, patients who died might have had predominant circulatory instability or infections with resistant organisms, contributing to a poorer prognosis despite having less severe signs at initial presentation.
This is the first and the largest study exploring the outcomes of children with community-acquired pneumonia admitted to paediatric intensive care units in Sri Lanka. The findings offer important insights for policymakers and healthcare providers by highlighting not only the clinical severity of CAP but also its broader implications for the healthcare system. The findings, including the length of ICU stay, underscore the significant burden on already limited PICU resources. This reflects the pressing need for effective strategies to prevent progression to critical illness.
Strengthening of public health interventions, particularly the expansion of immunisation programmes in Sri Lanka to include key respiratory pathogens such as Streptococcus pneumoniae, is vital. Many countries have observed a marked decline in pneumonia-related morbidity and mortality following the introduction of pneumococcal vaccination [30]. Sri Lanka is currently in the process of evaluating the inclusion of the pneumococcal vaccine in the national immunisation program, which could offer substantial benefits in preventing severe CAP and alleviating pressure on critical care services. Additionally, the potential for community-level preventive interventions, such as early identification and outpatient management of CAP, improved nutrition, and caregiver education, should be explored. These measures could help reduce the number of cases requiring hospitalisation or intensive care, preserving scarce PICU beds for the most critically ill children.
The findings of this study should be interpreted with several limitations in mind. Although we initially planned to assess the IDSA-PIDS guidelines and the Paediatric Early Warning Score, the extent of missing data made this analysis unfeasible. The inclusion of pathogen-specific data at admission would have significantly enriched our findings; however, resource constraints, such as the lack of viral diagnostics and the absence of blood or sputum cultures before initiating antibiotics, limited our ability to accurately identify causative organisms. Inconsistencies in medical records further hindered the analysis of microbiological data and antibiotic prescribing patterns. The differentiation of community-acquired pneumonia (CAP) from other respiratory illnesses, such as bronchiolitis, was based on available documentation, introducing potential misclassification, although a panel of experienced clinicians reviewed unclear cases to reach a consensus. We were also unable to evaluate the impact of intravenous fluid administration, which has been associated with increased mortality [31] in resource-limited settings, due to the absence of relevant data. Similarly, we did not examine the exact causes of death, as it was beyond the scope of our objectives. While some patients in our cohort had prolonged PICU stays, detailed data on specific ICU interventions and ventilation duration were unavailable, limiting our ability to assess resource utilisation. This gap highlights the need for prospective research to better understand ICU demands and long-term outcomes. Furthermore, important factors affecting the outcome such as immunisation status, institutional admission criteria, and pre-hospital delays were not assessed. Lastly, the absence of a centralised admission database meant that we could not include children treated in adult ICUs or those who died of CAP without access to intensive care services.
Conclusions
The mortality rate among children with community-acquired pneumonia was high in this study population. Younger age, the need for invasive ventilation, septic shock, multilobar involvement and underlying comorbidities were identified as statistically significant independent predictors of mortality. However, the high mortality rate observed may also reflect other contributing factors not captured within the scope of this study, highlighting the need for further research to explore these influences.
Supplementary Information
Authors’ contributions
G.L. Conceptualisation, Methodology, Data Acquisition, Formal analysis, Writing - original draft, Writing - reviewing & editing, Supervision. R.J. Methodology, Data Acquisition, Writing - original draft, Writing - reviewing & editing, Supervision. D.A. Methodology, Data Acquisition, Writing-reviewing & editing. K.P. Methodology, Data Acquisition, Writing-reviewing & editing. P.J. Methodology, Data Acquisition, Writing-reviewing & editing. S.A. Methodology, Data Acquisition, Writing-reviewing & editing. T.K. Methodology, Data Acquisition, Writing - original draft, Writing - reviewing & editing.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Data availability
The data is available from the corresponding author, GL, and will be provided at a reasonable request.
Declarations
Ethics approval and consent to participate
Ethical clearance was obtained from the Ethical Review Committee of the Sri Lanka College of Paediatricians (SLCP/ ERC/2022/24). Consent to participation was not required as this study involved a retrospective analysis of patient records. The data were de-identified prior to analysis, ensuring patient confidentiality, and the study was conducted in accordance with relevant ethical guidelines for retrospective research.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.World Health Organization, Geneva. Pneumonia in children. Switzerland: World Health Organization; 2002. https://www.who.int/news-room/fact-sheets/detail/pneumonia. Accessed on 20 June 2024. [Google Scholar]
- 2.Department of Census and Statistics. The department of census and statistics. Vital statistics-2019. Sri Lanka; 2019. http://www.statistics.gov.lk/Population/Vital_Statistics-2019#gsc.tab=0. Accessed on 20 June 2024.
- 3.Sharma PK, Vinayak N, Aggarwal GK, Srivastava RD, Aggarwal PK, Kushwaha AS. Severe necrotising pneumonia in children: A challenge to intensive care specialist. J Trop Pediatr. 2020;66(6):637–44. 10.1093/tropej/fmaa027. [DOI] [PubMed] [Google Scholar]
- 4.Shi T, Chen C, Huang L, Fan H, Lu G, Yang D, et al. Risk factors for mortality from severe community-acquired pneumonia in hospitalised children transferred to the pediatric intensive care unit. Pediatr Neonatol. 2020;61(6):577–83. 10.1016/j.pedneo.2020.06.005. [DOI] [PubMed] [Google Scholar]
- 5.Koh JWJC, Wong JJ, Sultana R, Wong PPC, Mok YH, Lee JH. Risk factors for mortality in children with pneumonia admitted to the pediatric intensive care unit. Pediatr Pulmonol. 2017;52(8):1076–84. 10.1002/ppul.23702. [DOI] [PubMed] [Google Scholar]
- 6.Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al. Infectious diseases society of America; American thoracic society. Infectious diseases society of America/American thoracic society consensus guidelines on managing community-acquired pneumonia in adults. Clin Infect Dis. 2007;44(Suppl 2):S27–72. 10.1086/511159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Florin TA, Brokamp C, Mantyla R, DePaoli B, Ruddy R, Shah SS, Ambroggio L. Validation of the pediatric infectious diseases society-Infectious diseases society of America severity criteria in children with Community-Acquired pneumonia. Clin Infect Dis. 2018;67(1):112–9. 10.1093/cid/ciy031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sharma R, Sandrock CE, Meehan J, Theriault N. Community-Acquired Bacterial Pneumonia-Changing Epidemiology, Resistance Patterns, and Newer Antibiotics: Spotlight on Delafloxacin. Clin Drug Investig. 2020;40(10):947–960. 10.1007/s40261-020-00953-z. Erratum in: Clin Drug Investig. 2021;41(2):197. doi: 10.1007/s40261-020-00998-0. [DOI] [PMC free article] [PubMed]
- 9.Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology: A basic science for clinical medicine. Boston, MA: Little, Brown & Company; 1991. [Google Scholar]
- 10.Sirvent JM, Carmen de la Torre M, Lorencio C, Taché A, Ferri C, Garcia-Gil J, et al. Predictive factors of mortality in severe community-acquired pneumonia: A model with data on the first 24 h of ICU admission. Med Intensiva. 2013;37(5):308–15. 10.1016/j.medin.2013.03.003. [DOI] [PubMed]
- 11.Dean P, Florin TA. Factors associated with pneumonia severity in children: A systematic review. J Pediatr Infect Dis Soc. 2018;7(4):323–34. 10.1093/jpids/piy046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Harris M, Clark J, Coote N, et al. British thoracic society guidelines for the management of community-acquired pneumonia in children: update 2011. Thorax. 2011;66(Suppl 2):ii1–23. [DOI] [PubMed] [Google Scholar]
- 13.Sri Lanka College of Paediatricians & Ministry of Health and Indigenous Medicines, Sri Lanka. National Guidelines on Paediatric Respiratory Disorders. Sri Lanka College of Paediatricians. 2019. Available at: https://slcp.lk/wp-content/uploads/National-Guidelines-on-Paediatric-Respiratory-Disorders-Respiratory-Tract-Inflections-and-Chronic-Lung-Diseases.pdf).
- 14.Masarone D, Valente F, Rubino M, Vastarella R, Gravino R, Rea A, Russo MG, Pacileo G, Limongelli G. Pediatric heart failure: A practical guide to diagnosis and management. Pediatr Neonatol. 2017;58(4):303–12. 10.1016/j.pedneo.2017.01.001. [DOI] [PubMed] [Google Scholar]
- 15.World Health Organization. Revised WHO Classification and Treatment of Pneumonia in Children at Health Facilities: Evidence Summaries. 2014. Geneva, Switzerland: World Health Organization; 2014. https://www.ncbi.nlm.nih.gov/books/NBK264162/. Accessed on 20 June 2024. [PubMed]
- 16.2024. World Health Organization. Malnutrition in children, Geneva. Switzerland: World Health Organization; 2024. https://www.who.int/data/nutrition/nlis/info/malnutrition-in-children. Accessed on 03 July 2024.
- 17.Elemraid MA, Rushton SP, Thomas MF, Spencer DA, Gennery AR, Clark JE. Utility of inflammatory markers in predicting the aetiology of pneumonia in children. Diagn Microbiol Infect Dis. 2014;79(4):458–62. 10.1016/j.diagmicrobio.2014.04.006. [DOI] [PubMed] [Google Scholar]
- 18.Araya S, Lovera D, Zarate C, Apodaca S, Acuña J, Sanabria G, et al. Application of a prognostic scale to estimate the mortality of children hospitalised with Community-acquired pneumonia. Pediatr Infect Dis J. 2016;35(4):369–73. 10.1097/INF.0000000000001018. [DOI] [PubMed] [Google Scholar]
- 19.Ellis SM, Flower CDR. The WHO manual of diagnostic imaging: radiographic anatomy and interpretation of the chest and the pulmonary system; EditorsOstensen H, Pettersson H, In : 2006. https://iris.who.int/bitstream/handle/10665/43293/9241546778_eng.pdf?sequence=1.
- 20.Duncan H, Hutchison J, Parshuram CS. The pediatric early warning system score: a severity of illness score to predict urgent medical need in hospitalised children. J Crit Care. 2006;21(3):271–8. 10.1016/j.jcrc.2006.06.007. [DOI] [PubMed] [Google Scholar]
- 21.Patel A, Mamtani M, Hibberd PL, Tuan TA, Jeena P, Chisaka N, et al. Value of chest radiography in predicting treatment response in children aged 3–59 months with severe pneumonia. Int J Tuberc Lung Dis. 2008;12(11):1320–6. [PubMed] [Google Scholar]
- 22.Caggiano S, Ullmann N, De Vitis E, Trivelli M, Mariani C, Podagrosi M, et al. Factors that negatively affect the prognosis of pediatric Community-Acquired pneumonia in district hospital in Tanzania. Int J Mol Sci. 2017;18(3):623. 10.3390/ijms18030623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zeeshan A, Abbas Q, Siddiqui A, Khalid F, Jehan F. Critical illness related to community acquired pneumonia, its epidemiology and outcomes in a pediatric intensive care unit of Pakistan. Pediatr Pulmonol. 2021;56(12):3916–23. 10.1002/ppul.25668. [DOI] [PubMed] [Google Scholar]
- 24.Shi T, Chen C, Huang L, Fan H, Lu G, Yang D, Zhao C, Zhang D. Risk factors for mortality from severe community-acquired pneumonia in hospitalized children transferred to the pediatric intensive care unit. Pediatr Neonatol. 2020;61(6):577–83. 10.1016/j.pedneo.2020.06.005. [DOI] [PubMed] [Google Scholar]
- 25.Cao L, Ji Z, Zhang P, Wang J. Epidemiology and mortality predictors for severe childhood community-acquired pneumonia in icus: A retrospective observational study. Front Pediatr. 2023;11:1031423. 10.3389/fped.2023.1031423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Mamtani M, Patel A, Hibberd PL, Tuan TA, Jeena P, Chisaka N, et al. A clinical tool to predict failed response to therapy in children with severe pneumonia. Pediatr Pulmonol. 2009;44(4):379–86. 10.1002/ppul.21014. [DOI] [PubMed] [Google Scholar]
- 27.Williams DJ, Zhu Y, Grijalva CG, Self WH, Harrell FE Jr, Reed C, et al. Predicting severe pneumonia outcomes in children. Pediatrics. 2016;138(4):e20161019. 10.1542/peds.2016-1019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Marshall JC. The PIRO (predisposition, insult, response, organ dysfunction) model: toward a staging system for acute illness. Virulence. 2014;5(1):27–35. 10.4161/viru.26908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Liapikou A, Cillóniz C, Gabarrús A, Amaro R, De la Bellacasa JP, Mensa J, Sánchez M, Niederman M, Torres A. Multilobar bilateral and unilateral chest radiograph involvement: implications for prognosis in hospitalised community-acquired pneumonia. Eur Respir J. 2016;48(1):257–61. 10.1183/13993003.00191-2016. [DOI] [PubMed] [Google Scholar]
- 30.de Oliveira LH, Shioda K, Valenzuela MT, Janusz CB, Rearte A, Sbarra AN, et al. Multinational study for PCV impact in mortality study Team. Declines in pneumonia mortality following the introduction of Pneumococcal conjugate vaccines in Latin American and Caribbean countries. Clin Infect Dis. 2021;73(2):306–13. 10.1093/cid/ciaa614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Brown RM, Semler MW. Fluid management in sepsis. J Intensive Care Med. 2019;34(5):364–73. 10.1177/0885066618784861. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data is available from the corresponding author, GL, and will be provided at a reasonable request.