If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
In-hospital mortality risk stratification in children aged under 5 years with pneumonia with or without pulse oximetry: A secondary analysis of the Pneumonia REsearch Partnership to Assess WHO REcommendations (PREPARE) dataset
Department of Global Public Health, Karolinska Institutet, Stockholm, Sweden and Institute for Global Health, University College London, London, United Kingdom
Global Program in Respiratory Sciences, Eudowood Division of Pediatric Respiratory Sciences, Department of Pediatrics, Johns Hopkins School of Medicine, Baltimore, United States of America and Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States of America
Division of Global Health Protection, Thailand Ministry of Public Health–US Centers for Disease Control and Prevention Collaboration, Nonthaburi, Thailand
South African Medical Research Council: Vaccines and Infectious Diseases Analytics Research Unit, School of Pathology, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation: Vaccine Preventable Diseases Unit, University of the Witwatersrand, Johannesburg, South Africa
South African Medical Research Council: Vaccines and Infectious Diseases Analytics Research Unit, School of Pathology, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation: Vaccine Preventable Diseases Unit, University of the Witwatersrand, Johannesburg, South Africa
Universidad Nacional de Asuncion, Departamento de Biología Molecular y Genética, Instituto de Investigaciones en Ciencias de la Salud, Asuncion, Paraguay
Unité d'Hygiène, Epidémiologie, Infectiovigilance et Prévention, Hospices Civils de Lyon, Lyon, France and Centre International de Recherche en Infectiologie, Institut National de la Santé et de la Recherche Médicale U1111, CNRS Unité Mixte de Recherche 5308, École Nationale Supérieure de Lyon, Université Claude Bernard Lyon 1, Lyon, France
Chinese Academy of Medical Sciences & Peking Union, Medical College Institute of Pathogen Biology, MOH Key Laboratory of Systems Biology of Pathogens and Dr Christophe Mérieux Laboratory, Beijing, China
Center for Intervention Science in Maternal and Child Health, University of Bergen, Norway and Department of Pediatrics, Tribhuvan University Institute of Medicine, Nepal
Children with pneumonia whose oxygen level was measured had a lower risk of death.
•
Hypoxemia was frequent among danger signs and chest-indrawing pneumonia cases.
•
Pulse oximeters are essential tools for hospital-based child pneumonia care.
•
Additional interventions to reduce in-hospital pneumonia deaths should be explored.
Abstract
Objectives
We determined the pulse oximetry benefit in pediatric pneumonia mortality risk stratification and chest-indrawing pneumonia in-hospital mortality risk factors.
Methods
We report the characteristics and in-hospital pneumonia-related mortality of children aged 2-59 months who were included in the Pneumonia Research Partnership to Assess WHO Recommendations dataset. We developed multivariable logistic regression models of chest-indrawing pneumonia to identify mortality risk factors.
Results
Among 285,839 children, 164,244 (57.5%) from hospital-based studies were included. Pneumonia case fatality risk (CFR) without pulse oximetry measurement was higher than with measurement (5.8%, 95% confidence interval [CI] 5.6-5.9% vs 2.1%, 95% CI 1.9-2.4%). One in five children with chest-indrawing pneumonia was hypoxemic (19.7%, 95% CI 19.0-20.4%), and the hypoxemic CFR was 10.3% (95% CI 9.1-11.5%). Other mortality risk factors were younger age (either 2-5 months [adjusted odds ratio (aOR) 9.94, 95% CI 6.67-14.84] or 6-11 months [aOR 2.67, 95% CI 1.71-4.16]), moderate malnutrition (aOR 2.41, 95% CI 1.87-3.09), and female sex (aOR 1.82, 95% CI 1.43-2.32).
Conclusion
Children with a pulse oximetry measurement had a lower CFR. Many children hospitalized with chest-indrawing pneumonia were hypoxemic and one in 10 died. Young age and moderate malnutrition were risk factors for in-hospital chest-indrawing pneumonia-related mortality. Pulse oximetry should be integrated in pneumonia hospital care for children under 5 years.
GBD 2015 LRI Collaborators Estimates of the global, regional, and national morbidity, mortality, and aetiologies of lower respiratory tract infections in 195 countries: a systematic analysis for the Global Burden of Disease Study 2015.
]. Over the last 2 decades, substantial progress has been made to reduce mortality and limit unnecessary hospitalizations. Randomized controlled trials demonstrated that children aged 2-59 months with chest-indrawing pneumonia without any general danger sign experience similar treatment failure rates with oral amoxicillin as those managed with injectable penicillin [
World Health Organization Recommendations for management of common childhood conditions : evidence for technical update of pocket book recommendations : newborn conditions, dysentery, pneumonia, oxygen use and delivery, common causes of fever, severe acute malnutrition and supportive care.
World Health Organization. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses. 2nd ed., 2013 ed. Geneva: World Health Organization, 2013.
World Health Organization. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses. 2nd ed., 2013 ed. Geneva: World Health Organization, 2013.
. It recommends that children aged 2-59 months without HIV with chest indrawing but without general danger signs (unable to drink/feed; convulsions; sleepy/lethargic; vomiting everything; severe wheezing; and signs of respiratory distress, including grunting, head nodding, nasal flaring), stridor, severe malnutrition, or hypoxemia (defined as a peripheral transcutaneous oxyhemoglobin saturation [SpO2] <90%) can be treated with oral amoxicillin. Other trials in India [
A randomized controlled trial of hospital versus home based therapy with oral amoxicillin for severe pneumonia in children aged 3–59 months: the IndiaCLEN Severe Pneumonia Oral Therapy (ISPOT) Study.
Community case management of severe pneumonia with oral amoxicillin in children aged 2–59 months in Haripur district, Pakistan: a cluster randomised trial.
Effectiveness of community case management of severe pneumonia with oral amoxicillin in children aged 2–59 months in Matiari district, rural Pakistan: a cluster-randomised controlled trial.
Safety, effectiveness and feasibility of outpatient management of children with pneumonia with chest indrawing at Port Moresby general hospital, Papua New Guinea.
] demonstrated that these children could be safely treated with oral antibiotics at home. However, most studies screened for and excluded hypoxemic children, using definitions ranging from SpO2 <90% to <85%. In addition, none were powered to demonstrate the differences in mortality [
Danger signs are either according to WHO pocketbook (i.e., central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3. bpm, breaths per minute; RR, respiration rate WHO, World Health Organization.
Danger signs are either according to WHO pocketbook (i.e., central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3. bpm, breaths per minute; RR, respiration rate WHO, World Health Organization.
Not applicable
a Fast breathing for age: RR ≥50 bpm in those aged 2-11 months and RR ≥40 bpm in those aged 12-59 month;
b Danger signs are either according to WHO pocketbook (i.e., central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.bpm, breaths per minute; RR, respiration rate WHO, World Health Organization.
Chest indrawing is a cardinal feature of respiratory distress that precedes hypoxemia and respiratory failure in children. This inward movement of abdominal and chest wall soft tissue below the rib cage is due to the increased negative intrapleural pressures generated to expand lungs with poor compliance during inspiration [
]. Hypoxemia occurs most commonly when there is ventilation perfusion mismatch in the lungs from an ALRI and is most frequently measured noninvasively by a pulse oximeter device [
]. However, before COVID-19, when the data for the current analysis were collected, pulse oximetry was limited in many low-resource settings, particularly in primary and community care [
. It is recognized that in the absence of pulse oximetry, the WHO Integrated Management of Childhood Illness protocol may miss hypoxemia, leading to misclassification of patients who need oxygen and inpatient care [
Hypoxaemia in Mozambican children <5 years of age admitted to hospital with clinical severe pneumonia: clinical features and performance of predictor models.
]. A retrospective Kenyan study conducted in district hospitals without pulse oximeters found that, apart from danger signs, mild to moderate pallor, age <12 months, lower chest indrawing, respiratory rate of 70 breaths or more, admission to a hospital in a malaria-endemic region, and moderate malnutrition were independently associated with pneumonia-related mortality [
The overall goal of this study was to understand the value of pulse oximetry in evaluating hospitalized children with pneumonia. We also explored additional clinical characteristics that were risk factors for chest-indrawing pneumonia mortality and could therefore be used to identify children with a high mortality risk. Using the WHO Pneumonia Research Partnership to Assess WHO Recommendations (PREPARE) study dataset, we aimed to (i) describe and compare the clinical characteristics and case fatality risk (CFR) by pneumonia severity among children with and without a pulse oximetry reading at study enrollment and (ii) determine in-hospital mortality risk factors among children aged 2-59 months with chest-indrawing pneumonia, with and without pulse oximetry measurements.
Methods
Study sample
We conducted a secondary analysis of collated datasets from 41 studies included in the WHO PREPARE project. These studies were conducted in 31 countries, including 29 low-middle-income countries (LMICs). Figure 1 describes how the analytic dataset was selected and used for this analysis. A detailed description of the studies is shown in Table 1 [
Microorganisms associated with pneumonia in children <5 years of age in developing and emerging countries: the GABRIEL pneumonia multicenter, prospective, case-control study.
Impact of the 13-valent Pneumococcal Conjugate Vaccine on clinical and hypoxemic childhood pneumonia over three years in central Malawi: an observational study.
Influence of respiratory viruses on the evaluation of the 13-valent Pneumococcal Conjugate Vaccine effectiveness in children under 5 years old: a time-series study for the 2001–2013 period.
Incidences of vaccine-preventable Haemophilus influenzae type b pneumonia and meningitis in Indonesian children: hamlet-randomised vaccine-probe trial.
Identification of radiological alveolar pneumonia in children with high rates of hospitalized respiratory infections: comparison of WHO-defined and pediatric pulmonologist diagnosis in the clinical context.
Chloramphenicol versus ampicillin plus gentamicin for community acquired very severe pneumonia among children aged 2–59 months in low resource settings: multicentre randomised controlled trial (SPEAR study).
Efficacy of zinc given as an adjunct in the treatment of severe and very severe pneumonia in hospitalized children 2–24 mo of age: a randomized, double-blind, placebo-controlled trial.
Efficacy of nine-valent Pneumococcal Conjugate Vaccine against pneumonia and invasive pneumococcal disease in the Gambia: randomised, double-blind, placebo-controlled trial.
. The primary data collection occurred between 1994 and 2014, and countries were at varying stages of Pneumococcus and Haemophilus influenzae type B (Hib) vaccine implementation.
Figure 1WHO signs: Fast breathing (respiratory rate 50 or more per minute in 2-11 months old and 40 or more per minute in 12-59 months old), lower chest indrawing, or danger signs (defined below)†Of 26 hospital-based studies in the PREPARE dataset, 17 studies reported information on the presence of chest indrawing in children 2-59 months of age.‡Danger signs are either according to WHO pocketbook (i.e., central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.
The patient records of hospitalized children aged 2-59 months with WHO-defined pneumonia and had survival outcome were included in the current analyses. We excluded patient records from community-based studies, those without WHO signs for pneumonia classification, or those who had no survival outcomes (Figure 1). Children received hospital-based care, including antibiotics and supplemental oxygen when indicated and available according to local norms.
Definitions and variables
The WHO pneumonia severity was defined as fast breathing (respiratory rate above the age-specific cut-off), chest indrawing, or danger sign (Box 1). Before the 2012 WHO guidance revision [
World Health Organization Recommendations for management of common childhood conditions : evidence for technical update of pocket book recommendations : newborn conditions, dysentery, pneumonia, oxygen use and delivery, common causes of fever, severe acute malnutrition and supportive care.
], it was recommended that all children with chest indrawing, even those without danger signs, should be hospitalized for injectable antibiotics and supportive care. Variables were chosen a priori due to clinical significance and potential association with mortality based on previous studies. These variables include: age, sex, weight, weight-for-age z-score, temperature (normothermia [35.5-37.9°C], fever [≥38°C], and hypothermia [<35.5°C]), age-adjusted tachypnea, severe tachypnea (defined as respiratory rate ≥70 breaths per minute), signs of severe respiratory distress (i.e., grunting, head nodding, or nasal flaring), and SpO2 (if reported) [
World Health Organization. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses. 2nd ed., 2013 ed. Geneva: World Health Organization, 2013.
], these data were not routinely included in our dataset. Some of the studies that contributed to the dataset were multicountry and included malaria hyperendemic regions and nonhyperendemic regions, but the dataset identifies cases by study and does not include if the case was from a specific country. Our primary outcome of interest was in-hospital pneumonia-related mortality.
Statistical analysis
To address the first objective, we described and compared the frequency, proportion with 95% confidence interval (CI), mean, median, and missingness of data on the clinical characteristics at the time of admission by pneumonia severity and if pulse oximetry was measured. Then, we reported the CFR by WHO pneumonia severity and compared these CFRs by the presence of pulse oximetry measurements. Pulse oximetry was categorized into no SpO2 measurement and any SpO2 measured at presentation. The measured group was further stratified into SpO2 <90%, 90-92%, and 93-100% to explore the impact of these categories on CFR [
Predicting hospitalised paediatric pneumonia mortality risk: an external validation of RISC and mRISC, and local tool development (RISC-Malawi) from Malawi.
Hypoxaemia as a mortality risk factor in acute lower respiratory infections in children in low and middle-income countries: systematic review and meta-analysis.
For the second objective, we exclusively used the data from the chest-indrawing pneumonia cases subset to fit two mixed-effects logistic regression models to explore the associations between demographic characteristics, nutritional status, and clinical signs at initial presentation with mortality. Mixed-effects modeling was chosen because some parameters had clear fixed effects on the outcome (tachypnea vs no tachypnea), whereas we assumed other parameters had a variable or unknown effect on the outcome (male vs female). In addition, we included variables to reflect the study type (observational vs randomized controlled trial) and pneumococcal vaccine (PCV) implementation at time of data collection because these factors had clear effects on mortality. Heterogeneity was accounted for at each study level. The first model included chest-indrawing cases with pulse oximetry measurements at hospitalization, and the second included those without pulse oximetry measurements. In the model that included pulse oximetry, SpO2 categories (<90%, 90-92%, and >93%) were treated as ordinal co-variables. To assess for bias, we described the variable missingness. We then conducted a bivariate analysis with complete cases. Variables with >15% missingness were excluded from the multivariable model. We reported the adjusted odds ratio (aOR) with 95% CI.
Results
Included studies
Among the 285,839 children from 41 studies in the PREPARE dataset, 164,244 (57.5%) from 26 of the 41 included studies (conducted in 29 countries) met the inclusion criteria for analysis (Figure 1 and Table 1). All cases that met the inclusion criteria were enrolled in-hospital-based studies. Pulse oximetry measurements were reported in 27.8% (n = 45,675) of cases. Among the 164,244 children included in the analyses, there were 7921 deaths (CFR 4.8%). Of included cases, 12.6% (n = 20,672) had only fast-breathing pneumonia, 39.1% (n = 64,256) had chest indrawing with or without fast breathing, and 48.3% (n = 79,316) had any danger sign at the time of admission.
The majority of cases with a pulse oximetry measurement (62.5%, 28,554/45,675) were from three studies conducted in Malawi [
Impact of the 13-valent Pneumococcal Conjugate Vaccine on clinical and hypoxemic childhood pneumonia over three years in central Malawi: an observational study.
Impact of the 13-valent Pneumococcal Conjugate Vaccine on clinical and hypoxemic childhood pneumonia over three years in central Malawi: an observational study.
Incidences of vaccine-preventable Haemophilus influenzae type b pneumonia and meningitis in Indonesian children: hamlet-randomised vaccine-probe trial.
A countrywide pneumonia surveillance study conducted in Malawi from 2001 to 2012 provided 85.3% (n = 101,182/118,569) of cases without pulse oximetry measurement [
]. This study contributed 90.0% (n = 46,160/51,320) of the cases with chest-indrawing pneumonia without pulse oximetry measurement. Four studies with overall CFRs ≥6.0% were from India [
Incidences of vaccine-preventable Haemophilus influenzae type b pneumonia and meningitis in Indonesian children: hamlet-randomised vaccine-probe trial.
Efficacy of nine-valent Pneumococcal Conjugate Vaccine against pneumonia and invasive pneumococcal disease in the Gambia: randomised, double-blind, placebo-controlled trial.
The overall hypoxemia (SpO2 <90%) prevalence was 17.7% (95% CI 17.3-18.0%). A nearly similar prevalence of hypoxemia was observed in patients with chest-indrawing pneumonia (19.7%; 95% CI 19.0-20.4%) and patients with danger sign pneumonia (20.7%; 95% CI 20.2-21.2%). Of the 164,244 cases in the dataset, 6.1% reported data on signs of severe respiratory distress. Children with chest-indrawing pneumonia with and without pulse oximetry measurements had similar characteristics except for differences in the prevalence of temperature ≥38°C (35.9% with vs 40.5% without SpO2) and severe tachypnea (10.7% with vs 13.9% without SpO2), which might reflect the differences in frequency of missing data among cases without an SpO2 measurement (Table 2). Among fast-breathing pneumonia cases, a larger proportion of children with versus without pulse oximetry measurements were aged 2-5 months (14.1% with vs 9.7% without SpO2). Otherwise, the demographic and clinical characteristics by pneumonia severity were similar between the SpO2 measured and not measured cohorts.
Table 2Baseline characteristics of hospitalized children aged 2-59 months with pneumonia by pneumonia classification and pulse oximetry assessment (N = 164,244).
Danger signs are either according to WHO pocketbook [6]; i.e.; central cyanosis, apnea, gasping, grunting, nasal flaring, audible wheeze, head nodding) or according to IMCI [7]; i.e.; general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.
(n = 79,316)
Pulse oximetry not measured (n = 11,124)
Pulse oximetry measured (n = 9548)
Pulse oximetry not measured (n = 51,320)
Pulse oximetry measured (n = 12,936)
Pulse oximetry not measured (n = 56,125)
Pulse oximetry measured (n = 23,191)
Study design Clinical trial Observational study
56 (0.5, 0.4, 0.6) 11,068 (99.5, 99.3, 99.6)
1259 (13.2, 12.5, 13.9) 8289 (86.8, 86.1, 87.5)
1949 (3.8, 3.6, 4.0) 49,371 (96.2, 96.0, 96.4)
2717 (21.0, 20.3, 21.7) 10,219 (79.0, 78.3, 79.7)
454 (0.8, 0.7, 0.9) 55,671 (99.2, 99.1, 99.3)
9022 (8.9, 38.3, 39.5) 14,169 (61.1, 60.5, 61.7)
Pneumococcal vaccine rollout Yes No
29 (0.3, 0.2, 0.4) 11,095 (99.7, 99.6, 99.8)
903 (9.5, 8.9, 10.1) 8645 (90.5, 89.9, 91.1)
546 (1.1, 1.0, 1.2) 50,774 (98.9, 98.8, 99.0)
2524 (19.5, 18.8, 20.2) 10,412 (80.5, 79.8, 81.2)
1293 (2.3, 2.2, 2.4) 54,832 (97.7, 97.6, 97.8)
9336 (40.3, 39.6, 40.9) 13,855 (59.7, 59.1, 60.4)
Age in months median value, (IQR) Age category 2-5 months old, n (%, 95% CI) 6-11 months old, n (%, 95% CI) 12-59 months old, n (%, 95% CI)
World Health Organization. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses. 2nd ed., 2013 ed. Geneva: World Health Organization, 2013.
; i.e.; general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.
In Table 3, we compare the CFR of cases with and without pulse oximetry integrated into their care. Column 1 reflects the data from four studies with 100% missing SpO2 values. Pulse oximetry was not documented in these studies because it was not integrated into the overall study design. Column 2 reflects the data from studies with <100% missing values. Except for two retrospective studies by Lu (55.5%) and Wulandari (26.0%), all studies had <15% missing SpO2 values. We chose to include the Lu and Wulandari studies in column 2 because the subanalyses suggested that pulse oximetry measurements were missing at random because there was no difference in the CFR among children with and without a documented SpO2 measurement. Most missing measurements were in children with fast-breathing pneumonia, and pulse oximetry is routinely not used in these cases.
Table 3CFR of children aged 2-59 months hospitalized with pneumonia by pulse oximetry assessment and pneumonia classification (n = 164,244).
Danger signs are either according to World Health Organization pocketbook; i.e.; central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI; i.e.; general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.
a Danger signs are either according to World Health Organization pocketbook; i.e.; central cyanosis, apnea, gasping, grunting, nasal flaring, severe wheezing, head nodding) or according to IMCI; i.e.; general danger sign (inability to drink, lethargy or unconscious, convulsions, vomit everything), stridor in a calm child or weight-for-age z-score <-3.
The CFR of patients without pulse oximetry measurement integrated as part of their care was significantly higher than those with a recorded measurement (5.8% vs 2.1%) (Table 3). This was particularly notable in children with any danger sign because those without a pulse oximetry measurement had a CFR of 8.5% (95% CI 8.2-8.7%) versus 4.6% (95% CI 4.3-4.9%) among those with a measurement. In patients with SpO2 measurement, independent of hypoxemia, the overall CFR of chest-indrawing pneumonia was 3.5% (95% CI 3.2-3.8%) versus 4.6% (95% CI 4.3-4.9%) in danger sign cases. Hypoxemic pneumonia cases of SpO2 <90% with a danger sign or chest indrawing had high CFRs of 11.8% and 10.3%, respectively. The CFR of patients with chest indrawing with SpO2 <90% (10.3%, 9.1-11.5%) was four times higher than those with an SpO2 of 90-92% (2.6%; 95% CI 2.0-3.3%) and six times higher than those with an SpO2 of 93-100% (1.6%; 95% CI 1.3-1.9%) (Table 3).
Mortality risk factors among chest-indrawing cases
We report the models of risks for chest-indrawing pneumonia-associated mortality with (Table 4) and without (Supplementary Tables 1 and 2) pulse oximetry measurements. Of cases with a measured SpO2, age bands 2-5 months (aOR 9.94, 95% CI 6.67-14.84) and 6-11 months (aOR 2.01, 95% CI 1.27-3.18), SpO2 <90% (aOR 3.47, 95% CI 2.66-4.52), -3 <weight-for-age z-score <-2 (aOR 2.67, 95% CI 1.71-4.16), and female sex (aOR 1.82, 95% CI 1.43-2.32) were associated with in-hospital mortality (Table 4). Notably, 41.8% (95% CI 37.2-46.5%; 188/450) of deaths occurred in children with an SpO2 >90%. Children with chest-indrawing pneumonia without a pulse oximetry measurement had similar pneumonia-related mortality risk factors (Supplementary Table 1). Because 90.0% of these cases came from Malawi, we conducted a sensitivity analysis by excluding Malawi cases and found that only moderate malnutrition remained a mortality risk factor (Supplementary Table 2).
Table 4Clinical characteristics associated with death of children aged 2-59 months hospitalized with chest-indrawing pneumonia with pulse oximetry assessment (n = 12,936).
In this study, the CFR of cases with an SpO2 measurement was lower than those without. Hypoxemia of SpO2 <90% was highly prevalent among children with chest-indrawing or danger sign pneumonia. Patients with chest-indrawing and danger sign pneumonia with an SpO2 <90% had a CFR of 10.3% and 11.8%, respectively. Age bands 2-5 months and 6-11 months, SpO2 <90%, moderate malnutrition, and female sex were independently associated with chest-indrawing pneumonia-related in-hospital death. We used a large multicountry dataset of hospitalized patients with pneumonia to explore the clinical outcomes in child pneumonia cases with and without SpO2 measurement and focused on cases with chest-indrawing pneumonia. Given the size of the dataset, our post hoc power estimate was greater than 95%. All data-contributing studies were conducted before the implementation of the 2012 WHO pneumonia management guidance recommending that children with chest-indrawing pneumonia without danger signs or hypoxemia (if pulse oximetry is available) could be safely managed in outpatient settings with oral amoxicillin [
World Health Organization Recommendations for management of common childhood conditions : evidence for technical update of pocket book recommendations : newborn conditions, dysentery, pneumonia, oxygen use and delivery, common causes of fever, severe acute malnutrition and supportive care.
The CFR was higher among child pneumonia cases without a documented SpO2 measurement. The reduced CFR with pulse oximeter use may reflect the impact of pulse oximetry on hospital outcomes or effects of a more functional health system [
Pulse oximeter with integrated management of childhood illness for diagnosis of severe childhood pneumonia at rural health institutions in Southern Ethiopia: results from a cluster-randomised controlled trial.
. In contrast, clinicians may use pulse oximetry as an objective measurement to improve their assessments, which, in some cases, may result in the de-escalation of unnecessary care, presumably freeing up resources for children who could benefit from them [
The impact of pulse oximetry on diagnosis, management and outcomes of acute febrile illness in low-income and middle-income countries: a systematic review.
Our hypoxemia prevalence and some of our CFR findings differ from other published reports. Our estimates are slightly higher than that of a 2009 metanalysis [
]. This study included a small amount of data (<10%) from studies that enrolled children aged up to 12 years. Hypoxemic pneumonia is less frequent in older children, which may explain their slightly lower estimates. In contrast, our findings are much lower than that reported by two meta-analyses [
Pneumonia Etiology Research for Child Health (PERCH) Study Group Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia: the PERCH multi-country case-control study.
] (47-35.8%). Misclassification bias could explain these differences. In the PERCH study, at most study sites, hypoxemia was defined as an SpO2 <92%. In Rahman et al.’s metanalysis, they were unable to disaggregate data, and 17 of the 57 included studies defined hypoxemia as an SpO2 <92-95% [
Predicting hospitalised paediatric pneumonia mortality risk: an external validation of RISC and mRISC, and local tool development (RISC-Malawi) from Malawi.
Hypoxaemia as a mortality risk factor in acute lower respiratory infections in children in low and middle-income countries: systematic review and meta-analysis.
The predictive performance of a pneumonia severity score in human immunodeficiency virus-negative children presenting to hospital in 7 low- and middle-income countries.
], an SpO2 of 90-92% was proposed as an ALRI-associated mortality risk factor, independent of clinical severity, but this was not associated with chest-indrawing pneumonia mortality in our data. In our study, hypoxemia (SpO2 <90%) prevalence was high and it put children, particularly those with chest indrawing or danger signs, at risk for death. Despite the WHO's recommendation to use pulse oximetry to assess hypoxemia and clear evidence that pulse oximeters are essential medical devices, many outpatient facilities and hospitals in LMICs do not have or do not use pulse oximeters in routine care [
Use of pulse oximetry during initial assessments of children under five with pneumonia: a retrospective cross-sectional study from 14 hospitals in Ethiopia.
. Unfortunately, clinical signs alone are not reliable predictors of hypoxemia, resulting in both false-positive and false-negative classifications, leading to many hypoxemic children not receiving oxygen and potentially contributing to pneumonia-related deaths [
Pulse oximeter with integrated management of childhood illness for diagnosis of severe childhood pneumonia at rural health institutions in Southern Ethiopia: results from a cluster-randomised controlled trial.
. Global uptake of pulse oximeters at the health system level will take time due to funding and implementation challenges, such as procurement, training, promotion of use, and ongoing monitoring and feedback [
. Nonetheless, pulse oximetry implementation is a necessary investment.
The value of pulse oximetry is clear; however, other factors may play an important role in reducing chest-indrawing pneumonia deaths. Unlike children with chest-indrawing pneumonia receiving outpatient care, the children in this study had access to supplemental oxygen and hospital-based care and received injectable antibiotics yet they still died. Similar to ours, other studies identified age 2-11 months, moderate malnutrition, severe tachypnoea, and female sex as mortality risk factors among children with pneumonia [
Predicting hospitalised paediatric pneumonia mortality risk: an external validation of RISC and mRISC, and local tool development (RISC-Malawi) from Malawi.
. These risk factors are plausible. Sex-based health disparities, including delayed care seeking, have been demonstrated in Africa and Asia and may explain some of these findings [
Predicting hospitalised paediatric pneumonia mortality risk: an external validation of RISC and mRISC, and local tool development (RISC-Malawi) from Malawi.
Risk factors for mortality from acute lower respiratory infections (ALRI) in children under five years of age in low and middle-income countries: a systematic review and meta-analysis of observational studies.
Gender inequalities in health and wellbeing across the first two decades of life: an analysis of 40 low-income and middle-income countries in the Asia-Pacific region.
. The excess mortality burden in infancy may reflect incomplete vaccination or a higher risk of occult untreated serious bacterial infection other than pneumonia, such as bacteremia, urinary tract infection, malaria, and meningitis [
. Exploratory studies are necessary to identify how these risk factors could inform medical decision making in the triage, follow-up, and hospital care of children with chest-indrawing pneumonia. Although known to contribute to pneumonia-related [
] and all-cause mortality, there is no formal disease-specific guidance on the care of children who are moderately malnourished. Targeted interventions could reduce pneumonia-related mortality in this group. For instance, the association of enteral protein intake during hospitalization with reduced 60-day mortality is well documented in critically ill children who are mechanically ventilated, independent of baseline nutrition status [
]. It is plausible that protein supplementation could reduce pneumonia-related mortality, particularly in children who are moderately malnourished. An ongoing phase II randomized controlled trial in Kenya and Uganda addresses this issue in children with severe (danger sign) pneumonia [
]. Other potential studies could evaluate if close outpatient follow-up, earlier hospital referral, or more intensive in-hospital monitoring of select groups, such as young infants aged 2-5 months, may reduce hospitalized pneumonia deaths.
Limitations
This study had some limitations. First, most of these data are derived from studies conducted before or during the widespread PCV and Hib vaccine implementation. Accordingly, we may be overestimating the prevalence of severe and hypoxemic pneumonia and its associated mortality because 73% of infants now receive Hib vaccine, and around 45% receive PCV [
]. However, suggesting otherwise, our findings have a similar CFR as that of the PERCH study (6.7%), which examined the etiology of severe pneumonia in the postpneumococcal vaccination era [
Pneumonia Etiology Research for Child Health (PERCH) Study Group Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia: the PERCH multi-country case-control study.
]. Second, because the studies included in our dataset occurred before the WHO recommendation that chest-indrawing cases be managed with oral amoxicillin, the majority of these cases received injectable antibiotics. Third, HIV co-morbidity data were not commonly documented in the dataset. Pneumonia-associated hypoxemia and mortality are higher among children who are HIV-positive or -exposed [
]. Therefore, our findings are not generalizable to this high-risk patient group. In addition, other co-infections, such as malaria, were not accounted for [
]. Although, this may also reflect real-world conditions in settings without reliable HIV and malaria testing resources. Fourth, there are inherent differences in pulse oximeter devices, training, and supervision, which could affect the accuracy of SpO2 measurements and our findings. Fifth, only 6.1% of this dataset included information on signs of respiratory distress. Given that these are considered danger signs that warrant hospitalization, we cannot be certain that some of the cases with chest indrawing did not also have signs of respiratory distress. Sixth, we collated data from a diverse range of settings and a large proportion of unmeasured pulse oximetry cases came from a single 10-year study in Malawi. To address this, we conducted a sensitivity analysis and found that moderate malnutrition remained a mortality risk factor even when the Malawian cases were excluded. Seventh, we were unable to assess for study-level variance in the duration of illness before hospitalization, length of hospitalization, and time to inpatient death, which may reflect the differences in care-seeking behaviors and clinician judgment to hospitalize a patient. There are certainly other unmeasured factors affecting child pneumonia in-hospital mortality, which we are unable to account for. Finally, we aggregated data from a wide variety of studies conducted in many different countries, which may limit the applicability of our findings to local contexts. Some of the data are from clinical trials or prospective studies with dedicated study staff, whereas others are from routine care settings; as such, there is variability in the quality of the reported data. We are unable to account for how management differences affected the patient outcomes. From an implementation perspective, the use of data collected outside of a funded, well-staffed, and well-supplied clinical trial may be a strength because these data more accurately reflect the real-world conditions of healthcare delivery in LMICs.
Conclusion
Pulse oximetry use is critical to providing effective pneumonia care. Given that many LMIC ALRI care settings do not have or use pulse oximeters and that danger signs and chest-indrawing cases had a high prevalence of hypoxemia and associated CFR, we can conclude that many children who could benefit from supplemental oxygen are going unrecognized. This represents a missed opportunity to reduce child pneumonia deaths. A substantial proportion of chest-indrawing pneumonia deaths were not hypoxemic. Exploratory research is needed to understand how mortality risk factors, such as moderate malnutrition and young age, could be used to guide care to reduce mortality. Our findings suggest that pulse oximetry should be integrated in the clinical evaluation of children aged 5 years who are hospitalized with ALRI, particularly for children with chest-indrawing pneumonia.
Declaration of competing interest
The authors have no competing interests to declare. YBN is a staff member of the WHO. The expressed views and opinions do not necessarily represent the policies of the WHO.
Funding
The Bill and Melinda Gates Foundation, Seattle, WA, USA (#INV-007927).
Ethical approval
All studies included in this deidentified data set were previously granted clearance by ethical review boards at each participating site.
Acknowledgments
The authors would like to thank the children and their families who participated in the included studies.
Author contributions
SAQ secured the funding. SH, CKin, YBN, and SAQ conceptualized and designed the study, interpreted the data, and wrote the first draft of the manuscript. YBN verified the data and conducted statistical analyses. CKin, EDM, TC, NL, CM, CG, ST, CC, SM, MN, BG, TH, JM, EA, NC, MH, PH, PJ, JL, WM, AP, DT, NN, SZ, RR, ML, CK, CT, RA, SB, II, IM, GM, SKS, MS, SS, SAwa, AB, MC, PN, JP, VR, GR, MSyl, PV, JW, SBha, TS, MN, LA, ME, SBas, NW, RL, SA, AG, MC, SHir, KO, AC, CR, HC, HN, JF, LW, and MH oversaw the data collection, verified the data, assisted with interpretation of the data, and reviewed and edited the final manuscript. SH and YBN had final responsibility for the decision to submit for publication.
Data sharing
Data used for this study will be available upon request. Deidentified participant data and a data dictionary will be made available after all planned analyses are completed. Data will be shared after approval of request. All data requests should be directed to the corresponding author, Dr. Yasir Bin Nisar.
Estimates of the global, regional, and national morbidity, mortality, and aetiologies of lower respiratory tract infections in 195 countries: a systematic analysis for the Global Burden of Disease Study 2015.
Recommendations for management of common childhood conditions : evidence for technical update of pocket book recommendations : newborn conditions, dysentery, pneumonia, oxygen use and delivery, common causes of fever, severe acute malnutrition and supportive care.
World Health Organization. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses. 2nd ed., 2013 ed. Geneva: World Health Organization, 2013.
A randomized controlled trial of hospital versus home based therapy with oral amoxicillin for severe pneumonia in children aged 3–59 months: the IndiaCLEN Severe Pneumonia Oral Therapy (ISPOT) Study.
Community case management of severe pneumonia with oral amoxicillin in children aged 2–59 months in Haripur district, Pakistan: a cluster randomised trial.
Effectiveness of community case management of severe pneumonia with oral amoxicillin in children aged 2–59 months in Matiari district, rural Pakistan: a cluster-randomised controlled trial.
Safety, effectiveness and feasibility of outpatient management of children with pneumonia with chest indrawing at Port Moresby general hospital, Papua New Guinea.
Hypoxaemia in Mozambican children <5 years of age admitted to hospital with clinical severe pneumonia: clinical features and performance of predictor models.
Microorganisms associated with pneumonia in children <5 years of age in developing and emerging countries: the GABRIEL pneumonia multicenter, prospective, case-control study.
Impact of the 13-valent Pneumococcal Conjugate Vaccine on clinical and hypoxemic childhood pneumonia over three years in central Malawi: an observational study.
Influence of respiratory viruses on the evaluation of the 13-valent Pneumococcal Conjugate Vaccine effectiveness in children under 5 years old: a time-series study for the 2001–2013 period.
Incidences of vaccine-preventable Haemophilus influenzae type b pneumonia and meningitis in Indonesian children: hamlet-randomised vaccine-probe trial.
Identification of radiological alveolar pneumonia in children with high rates of hospitalized respiratory infections: comparison of WHO-defined and pediatric pulmonologist diagnosis in the clinical context.
Chloramphenicol versus ampicillin plus gentamicin for community acquired very severe pneumonia among children aged 2–59 months in low resource settings: multicentre randomised controlled trial (SPEAR study).
Efficacy of zinc given as an adjunct in the treatment of severe and very severe pneumonia in hospitalized children 2–24 mo of age: a randomized, double-blind, placebo-controlled trial.
Efficacy of nine-valent Pneumococcal Conjugate Vaccine against pneumonia and invasive pneumococcal disease in the Gambia: randomised, double-blind, placebo-controlled trial.
Predicting hospitalised paediatric pneumonia mortality risk: an external validation of RISC and mRISC, and local tool development (RISC-Malawi) from Malawi.
Hypoxaemia as a mortality risk factor in acute lower respiratory infections in children in low and middle-income countries: systematic review and meta-analysis.
Pulse oximeter with integrated management of childhood illness for diagnosis of severe childhood pneumonia at rural health institutions in Southern Ethiopia: results from a cluster-randomised controlled trial.
The impact of pulse oximetry on diagnosis, management and outcomes of acute febrile illness in low-income and middle-income countries: a systematic review.
Pneumonia Etiology Research for Child Health (PERCH) Study Group
Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia: the PERCH multi-country case-control study.
The predictive performance of a pneumonia severity score in human immunodeficiency virus-negative children presenting to hospital in 7 low- and middle-income countries.
Use of pulse oximetry during initial assessments of children under five with pneumonia: a retrospective cross-sectional study from 14 hospitals in Ethiopia.
Risk factors for mortality from acute lower respiratory infections (ALRI) in children under five years of age in low and middle-income countries: a systematic review and meta-analysis of observational studies.
Gender inequalities in health and wellbeing across the first two decades of life: an analysis of 40 low-income and middle-income countries in the Asia-Pacific region.
00050-4&id=gr1.jpg){kind=link}