OBJECTIVES: After respiratory syncytial virus (RSV), human metapneumovirus (hMPV) was the second-ranked pathogen attributed to severe pneumonia in the PERCH study. We sought to characterize hMPV-positive cases in high burden settings, which have limited data, by comparing to RSV-positive and other cases. METHODS: Children aged 1-59 months hospitalized with suspected severe pneumonia and age/season-matched community controls in seven African and Asian countries had nasopharyngeal/oropharyngeal swabs tested by multiplex PCR for 32 respiratory pathogens, among other clinical and lab assessments at admission. Odds ratios adjusted for age and site (aOR) were calculated using logistic regression. Etiologic probability was estimated using Bayesian nested partial latent class analysis. Latent class analysis identified syndromic constellations of clinical characteristics. RESULTS: HMPV was detected more frequently among cases (267/3887, 6.9%) than controls (115/4976, 2.3%), among cases with pneumonia chest X-ray findings (8.5%) than without (5.5%), and among controls with respiratory tract illness (3.8%) than without (1.8%; all p≤0.001). HMPV-positive cases were negatively associated with the detection of other viruses (aOR=0.18), especially RSV (aOR=0.11; all p<0.0001), and positively associated with the detection of bacteria (aORs 1.77, p=0.03). No single clinical syndrome distinguished hMPV-positive from other cases. Among hMPV-positive cases, 65.2% were aged <1 year and 27.5% had pneumonia danger signs; positive predictive value was 74.5%; mortality was 3.9%, similar to RSV-positive (2.4%) and lower than other cases (9.6%). CONCLUSIONS: HMPV-associated severe pediatric pneumonia in high burden settings was predominantly in young infants and clinically indistinguishable from RSV. HMPV-positives had low case fatality, similar to that in RSV-positives.

INTRODUCTION: In 2012, the World Health Organization revised treatment guidelines for childhood pneumonia with lower chest wall indrawing (LCWI) but no 'danger signs', to recommend home-based treatment. We analysed data from children hospitalized with LCWI pneumonia in the Pneumonia Etiology Research for Child Health (PERCH) study to identify sub-groups with high odds of mortality, who might continue to benefit from hospital management but may not be admitted by staff implementing the 2012 guidelines. We compare the proportion of deaths identified using the criteria in the 2012 guidelines, and the proportion of deaths identified using an alternative set of criteria from our model. METHODS: PERCH enrolled a cohort of 2189 HIV-negative children aged 2-59 months who were admitted to hospital with LCWI pneumonia (without obvious cyanosis, inability to feed, vomiting, convulsions, lethargy or head nodding) between 2011-2014 in Kenya, Zambia, South Africa, Mali, The Gambia, Bangladesh, and Thailand. We analysed risk factors for mortality among these cases using predictive logistic regression. Malnutrition was defined as mid-upper-arm circumference <125mm or weight-for-age z-score <-2. RESULTS: Among 2189 cases, 76 (3·6%) died. Mortality was associated with oxygen saturation <92% (aOR 3·33, 1·99-5·99), HIV negative but exposed status (4·59, 1·81-11·7), moderate or severe malnutrition (6·85, 3·22-14·6) and younger age (infants compared to children 12-59 months old, OR 2·03, 95%CI 1·05-3·93). At least one of three risk factors: hypoxaemia, HIV exposure, or malnutrition identified 807 children in this population, 40% of LCWI pneumonia cases and identified 86% of the children who died in hospital (65/76). Risk factors identified using the 2012 WHO treatment guidelines identified 66% of the children who died in hospital (n = 50/76). CONCLUSIONS: Although it focuses on treatment failure in hospital, this study supports the proposal for better risk stratification of children with LCWI pneumonia. Those who have hypoxaemia, any malnutrition or those who were born to HIV positive mothers, experience poorer outcomes than other children with LCWI pneumonia. Consistent identification of these risk factors should be prioritised and children with at least one of these risk factors should not be managed in the community.

BACKGROUND: Pneumonia is the leading cause of death among children younger than 5 years. In this study, we estimated causes of pneumonia in young African and Asian children, using novel analytical methods applied to clinical and microbiological findings. METHODS: We did a multi-site, international case-control study in nine study sites in seven countries: Bangladesh, The Gambia, Kenya, Mali, South Africa, Thailand, and Zambia. All sites enrolled in the study for 24 months. Cases were children aged 1-59 months admitted to hospital with severe pneumonia. Controls were age-group-matched children randomly selected from communities surrounding study sites. Nasopharyngeal and oropharyngeal (NP-OP), urine, blood, induced sputum, lung aspirate, pleural fluid, and gastric aspirates were tested with cultures, multiplex PCR, or both. Primary analyses were restricted to cases without HIV infection and with abnormal chest x-rays and to controls without HIV infection. We applied a Bayesian, partial latent class analysis to estimate probabilities of aetiological agents at the individual and population level, incorporating case and control data. FINDINGS: Between Aug 15, 2011, and Jan 30, 2014, we enrolled 4232 cases and 5119 community controls. The primary analysis group was comprised of 1769 (41·8% of 4232) cases without HIV infection and with positive chest x-rays and 5102 (99·7% of 5119) community controls without HIV infection. Wheezing was present in 555 (31·7%) of 1752 cases (range by site 10·6-97·3%). 30-day case-fatality ratio was 6·4% (114 of 1769 cases). Blood cultures were positive in 56 (3·2%) of 1749 cases, and Streptococcus pneumoniae was the most common bacteria isolated (19 [33·9%] of 56). Almost all cases (98·9%) and controls (98·0%) had at least one pathogen detected by PCR in the NP-OP specimen. The detection of respiratory syncytial virus (RSV), parainfluenza virus, human metapneumovirus, influenza virus, S pneumoniae, Haemophilus influenzae type b (Hib), H influenzae non-type b, and Pneumocystis jirovecii in NP-OP specimens was associated with case status. The aetiology analysis estimated that viruses accounted for 61·4% (95% credible interval [CrI] 57·3-65·6) of causes, whereas bacteria accounted for 27·3% (23·3-31·6) and Mycobacterium tuberculosis for 5·9% (3·9-8·3). Viruses were less common (54·5%, 95% CrI 47·4-61·5 vs 68·0%, 62·7-72·7) and bacteria more common (33·7%, 27·2-40·8 vs 22·8%, 18·3-27·6) in very severe pneumonia cases than in severe cases. RSV had the greatest aetiological fraction (31·1%, 95% CrI 28·4-34·2) of all pathogens. Human rhinovirus, human metapneumovirus A or B, human parainfluenza virus, S pneumoniae, M tuberculosis, and H influenzae each accounted for 5% or more of the aetiological distribution. We observed differences in aetiological fraction by age for Bordetella pertussis, parainfluenza types 1 and 3, parechovirus-enterovirus, P jirovecii, RSV, rhinovirus, Staphylococcus aureus, and S pneumoniae, and differences by severity for RSV, S aureus, S pneumoniae, and parainfluenza type 3. The leading ten pathogens of each site accounted for 79% or more of the site's aetiological fraction. INTERPRETATION: In our study, a small set of pathogens accounted for most cases of pneumonia requiring hospital admission. Preventing and treating a subset of pathogens could substantially affect childhood pneumonia outcomes. FUNDING: Bill & Melinda Gates Foundation.

BACKGROUND: Pneumococcal disease is a leading cause of bacterial pneumonia and invasive bacterial disease among children globally. The reason some strains of pneumococci are more likely to cause disease, and how interventions such as vaccines and antibiotics affect pneumococcal strains is poorly understood. We aimed to identify genetic regions under selective pressure and those associated with disease through the analysis of pneumococcal whole-genome sequences. METHODS: Whole-genome sequencing was performed on pneumococcal isolates collected between January, 2005, and May, 2018, in Kathmandu, Nepal, which included programmatic ten-valent pneumococcal conjugate vaccine (PCV10) introduction in 2015. Isolates were from three distinct cohorts: nasopharyngeal swabs of healthy community-based children, nasopharyngeal swabs of children admitted to hospital with pneumonia, and sterile-site cultures from children admitted to hospital. Across these cohorts we examined serotype distribution, antibiotic resistance, strain distribution, and regions of recombination to determine genes that were undergoing diversifying selection. Genome-wide association studies comparing pneumonia and sterile-site isolates with healthy carriage were used to determine novel variants associated with disease. FINDINGS: After programmatic introduction of PCV10, there was a decline in vaccine covered serotypes; however, strains that had expressed these serotypes continued to exist in the post-PCV population. We identified GPSC9 to be a strain of concern due to its high prevalence in disease, multidrug resistance, and ability to switch to an unencapsulated phenotype via insertion of virulence factor pspC into the cps locus. Antibiotic resistance loci to co-trimoxazole were found to be prevalent (pre-PCV10 78% vs post-PCV10 81%; p=0·27) and increasingly prevalent to penicillin (pre-PCV10 15% vs post-PCV10 32%; p<0·0001). Regions with multiple recombinations were identified spanning the surface protein virulence factors pspA and pspC and antibiotic targets pbpX, folA, folC, folE, and folP. Furthermore, we identified variants in lacE2 to be strongly associated with isolates from children with pneumonia and PRIP to be strongly associated with isolates from sterile sites. INTERPRETATION: Our work highlights the effect of pneumococcal conjugate vaccines, antibiotics, and host-pathogen interaction in pneumococcal variation, and the pathogen's capability of adapting to these factors at both population-wide and strain-specific levels. Ongoing surveillance of disease-associated strains and further investigation of lacE2 and PRIP as serotype-independent targets for therapeutic interventions is required. FUNDING: Gavi, The Vaccine Alliance; WHO; Bill & Melinda Gates Foundation; Wellcome Sanger Institute; and US Centers for Disease Control and Prevention.

BACKGROUND: Pneumonia remains the leading cause of death among children <5 years of age beyond the neonatal period in Thailand. Using data from the Pneumonia Etiology Research for Child Health (PERCH) Study, we provide a detailed description of pneumonia cases and etiology in Thailand to inform local treatment and prevention strategies in this age group. METHODS: PERCH, a multi-country case-control study, evaluated the etiology of hospitalized cases of severe and very severe pneumonia among children 1-59 months of age. The Thailand site enrolled children for 24 consecutive months during January 2012-February 2014 with staggered start dates in 2 provinces. Cases were children hospitalized with pre-2013 WHO-defined severe or very severe pneumonia. Community controls were randomly selected from health services registries in each province. Analyses were restricted to HIV-negative cases and controls. We calculated adjusted odds ratios (ORs) and 95% CIs comparing organism prevalence detected by nasopharyngeal/oropharyngeal (NP/OP) polymerase chain reaction between cases and controls. The PERCH Integrated Analysis (PIA) used Bayesian latent variable analysis to estimate pathogen-specific etiologic fractions and 95% credible intervals. RESULTS: Over 96% of both cases (n = 223) and controls (n = 659) had at least 1 organism detected; multiple organisms were detected in 86% of cases and 88% of controls. Among 98 chest Radiograph positive (CXR+) cases, respiratory syncytial virus (RSV) had the highest NP/OP prevalence (22.9%) and the strongest association with case status (OR 20.5; 95% CI: 10.2, 41.3) and accounted for 34.6% of the total etiologic fraction. Tuberculosis (TB) accounted for 10% (95% CrI: 1.6-26%) of the etiologic fraction among CXR+ cases. DISCUSSION: More than one-third of hospitalized cases of severe and very severe CXR+ pneumonia among children 1-59 months of age in Thailand were attributable to RSV. TB accounted for 10% of cases, supporting evaluation for TB among children hospitalized with pneumonia in high-burden settings. Similarities in pneumonia etiology in Thailand and other PERCH sites suggest that global control strategies based on PERCH study findings are relevant to Thailand and similar settings.

The Pneumonia Etiology Research for Child Health (PERCH) study evaluated the etiology of severe and very severe pneumonia in children hospitalized in 7 African and Asian countries. Here, we summarize the highlights of in-depth site-specific etiology analyses published separately in this issue, including how etiology varies by age, mortality status, malnutrition, severity, HIV status, and more. These site-specific results impart important lessons that can inform disease control policy implications.

Rhinovirus (RV) is commonly detected in asymptomatic children; hence, its pathogenicity during childhood pneumonia remains controversial. We evaluated RV epidemiology in HIV-uninfected children hospitalized with clinical pneumonia and among community controls. PERCH was a case-control study that enrolled children (1-59 months) hospitalized with severe and very severe pneumonia per World Health Organization clinical criteria and age-frequency-matched community controls in seven countries. Nasopharyngeal/oropharyngeal swabs were collected for all participants, combined, and tested for RV and 18 other respiratory viruses using the Fast Track multiplex real-time PCR assay. RV detection was more common among cases (24%) than controls (21%) (aOR = 1.5, 95%CI:1.3-1.6). This association was driven by the children aged 12-59 months, where 28% of cases vs. 18% of controls were RV-positive (aOR = 2.1, 95%CI:1.8-2.5). Wheezing was 1.8-fold (aOR 95%CI:1.4-2.2) more prevalent among pneumonia cases who were RV-positive vs. RV-negative. Of the RV-positive cases, 13% had a higher probability (>75%) that RV was the cause of their pneumonia based on the PERCH integrated etiology analysis; 99% of these cases occurred in children over 12 months in Bangladesh. RV was commonly identified in both cases and controls and was significantly associated with severe pneumonia status among children over 12 months of age, particularly those in Bangladesh. RV-positive pneumonia was associated with wheezing.

BACKGROUND: Severity of viral respiratory illnesses can be increased with bacterial coinfection and can vary by sex, but influence of coinfection and sex on human endemic coronavirus (CoV) species, which generally cause mild to moderate respiratory illness, is unknown. We evaluated CoV and pneumococcal co-detection by sex in childhood pneumonia. METHODS: In the 2011-2014 Pneumonia Etiology Research for Child Health study, nasopharyngeal and oropharyngeal (NP/OP) swabs and other samples were collected from 3981 children <5 years hospitalized with severe or very severe pneumonia in 7 countries. Severity by NP/OP detection status of CoV (NL63, 229E, OC43 or HKU1) and high-density (≥6.9 log10 copies/mL) pneumococcus (HDSpn) by real-time polymerase chain reaction was assessed by sex using logistic regression adjusted for age and site. RESULTS: There were 43 (1.1%) CoV+/HDSpn+, 247 CoV+/HDSpn-, 449 CoV-/HDSpn+ and 3149 CoV-/HDSpn- cases with no significant difference in co-detection frequency by sex (range 51.2%-64.0% male, P = 0.06). More CoV+/HDSpn+ pneumonia was very severe compared with other groups for both males (13/22, 59.1% versus range 29.1%-34.7%, P = 0.04) and females (10/21, 47.6% versus 32.5%-43.5%, P = 0.009), but only male CoV+/HDSpn+ required supplemental oxygen more frequently (45.0% versus 20.6%-28.6%, P < 0.001) and had higher mortality (35.0% versus 5.3%-7.1%, P = 0.004) than other groups. For females with CoV+/HDSpn+, supplemental oxygen was 25.0% versus 24.8%-33.3% (P = 0.58) and mortality was 10.0% versus 9.2%-12.9% (P = 0.69). CONCLUSIONS: Co-detection of endemic CoV and HDSpn was rare in children hospitalized with pneumonia, but associated with higher severity and mortality in males. Findings may warrant investigation of differences in severity by sex with co-detection of HDSpn and SARS-CoV-2.

BACKGROUND: To estimate the infectious period of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in older adults with underlying conditions, we assessed duration of coronavirus disease 2019 (COVID-19) symptoms, reverse-transcription polymerase chain reaction (RT-PCR) positivity, and culture positivity among nursing home residents. METHODS: We enrolled residents within 15 days of their first positive SARS-CoV-2 test (diagnosis) at an Arkansas facility from July 7 to 15, 2020 and instead them for 42 days. Every 3 days for 21 days and then weekly, we assessed COVID-19 symptoms, collected specimens (oropharyngeal, anterior nares, and saliva), and reviewed medical charts. Blood for serology was collected on days 0, 6, 12, 21, and 42. Infectivity was defined by positive culture. Duration of culture positivity was compared with duration of COVID-19 symptoms and RT-PCR positivity. Data were summarized using measures of central tendency, frequencies, and proportions. RESULTS: We enrolled 17 of 39 (44%) eligible residents. Median participant age was 82 years (range, 58-97 years). All had ≥3 underlying conditions. Median duration of RT-PCR positivity was 22 days (interquartile range [IQR], 8-31 days) from diagnosis; median duration of symptoms was 42 days (IQR, 28-49 days). Of 9 (53%) participants with any culture-positive specimens, 1 (11%) severely immunocompromised participant remained culture-positive 19 days from diagnosis; 8 of 9 (89%) were culture-positive ≤8 days from diagnosis. Seroconversion occurred in 12 of 12 (100%) surviving participants with ≥1 blood specimen; all participants were culture-negative before seroconversion. CONCLUSIONS: Duration of infectivity was considerably shorter than duration of symptoms and RT-PCR positivity. Severe immunocompromise may prolong SARS-CoV-2 infectivity. Seroconversion indicated noninfectivity in this cohort.

BACKGROUND: Whether digitally recorded lung sounds are associated with radiographic pneumonia or clinical outcomes among children in low-income and middle-income countries is unknown. We sought to address these knowledge gaps. METHODS: We enrolled 1-59 month old children hospitalized with pneumonia at eight African and Asian Pneumonia Etiology Research for Child Health sites in six countries, recorded digital stethoscope lung sounds, obtained chest radiographs, and collected clinical outcomes. Recordings were processed and reclassified into binary categories positive or negative for adventitial lung sounds. Listening and reading panels classified recordings and radiographs. Recording classification associations with chest radiographs with World Health Organization (WHO)-defined primary endpoint pneumonia (radiographic pneumonia) or mortality were evaluated. We also examined case fatality among risk strata. RESULTS: Among children without WHO danger signs, wheezing (without crackles) had a lower adjusted odds ratio (aOR) for radiographic pneumonia (0.35, 95% confidence interval (CI) 0.15, 0.82), compared to children with normal recordings. Neither crackle only (no wheeze) (aOR 2.13, 95%CI 0.91, 4.96) or any wheeze (with or without crackle) (aOR 0.63, 95%CI 0.34, 1.15) were associated with radiographic pneumonia. Among children with WHO danger signs no lung recording classification was independently associated with radiographic pneumonia, although trends towards greater odds of radiographic pneumonia were observed among children classified with crackle only (no wheeze) or any wheeze (with or without crackle). Among children without WHO danger signs, those with recorded wheezing had a lower case fatality than those without wheezing (3.8% vs 9.1%, p=0.03). CONCLUSIONS: Among lower risk children without WHO danger signs digitally recorded wheezing is associated with a lower odds for radiographic pneumonia and with lower mortality. Although further research is needed, these data indicate that with further development digital auscultation may eventually contribute to child pneumonia care. This article is protected by copyright. All rights reserved.

In this ecological study, we attempt to quantify the extent to which differences in homicide and suicide death rates between three countries, and among states/provinces within those countries, may be explained by differences in their social, economic, and structural characteristics. We examine the relationship between state/province level measures of societal risk factors and state/province level rates of violent death (homicide and suicide) across Australia, Canada, and the United States. Census and mortality data from each of these three countries were used. Rates of societal level characteristics were assessed and included residential instability, self-employment, income inequality, gender economic inequity, economic stress, alcohol outlet density, and employment opportunities). Residential instability, self-employment, and income inequality were associated with rates of both homicide and suicide and gender economic inequity was associated with rates of suicide only. This study opens lines of inquiry around what contributes to the overall burden of violence-related injuries in societies and provides preliminary findings on potential societal characteristics that are associated with differences in injury and violence rates across populations.

BACKGROUND: In 2015, pneumonia remained the leading cause of mortality in children between 1-59 months old. METHODS: Data from 1802 HIV-negative children between 1-59 months old enrolled in the Pneumonia Etiology Research for Child Health (PERCH) study with severe or very severe pneumonia during 2011-14 were used to build a parsimonious multivariable model predicting mortality using backwards stepwise logistic regression. The PERCH severity score, derived from model coefficients, was validated on a second, temporally discrete dataset of a further 1819 cases and compared to other available scores using the c-statistic. RESULTS: Predictors of mortality, across seven low and middle-income countries, were: age <1 year, female sex, 3 or more days of illness prior to presentation to hospital, low weight-for-height, unresponsiveness, deep breathing, hypoxemia, grunting and the absence of cough. The model discriminated well between those who died and those who survived (c-statistic: 0.84), but the predictive capacity of the PERCH 5-stratum score derived from the coefficients was moderate (c=0.76). The performance of the Respiratory Index of Severity in Children (RISC) score was similar (c=0.76). The number of WHO danger signs demonstrated the highest discrimination (c=0.82; 1.5% died if no danger signs, 10% if 1 danger sign and 33% if 2 or more danger signs). CONCLUSIONS: The PERCH severity score could be used to interpret geographic variations in pneumonia mortality and etiology. The number of WHO danger signs on presentation to hospital could be the most useful, of the currently available tools, to aid clinical management of pneumonia.

OBJECTIVES: The aim of this study was to highlight the differences in injury rates between populations through a descriptive epidemiological study of population-level trends in injury mortality for the high-income countries of Australia, Canada and the United States. METHODS: Mortality data were available for the US from 2000 to 2014, and for Canada and Australia from 2000 to 2012. Injury causes were defined using the International Classification of Diseases, Tenth Revision external cause codes, and were grouped into major causes. Rates were direct-method age-adjusted using the US 2000 projected population as the standard age distribution. RESULTS: US motor vehicle injury mortality rates declined from 2000 to 2014 but remained markedly higher than those of Australia or Canada. In all three countries, fall injury mortality rates increased from 2000 to 2014. US homicide mortality rates declined, but remained higher than those of Australia and Canada. While the US had the lowest suicide rate in 2000, it increased by 24% during 2000-2014, and by 2012 was about 14% higher than that in Australia and Canada. The poisoning mortality rate in the US increased dramatically from 2000 to 2014. CONCLUSION: Results show marked differences and striking similarities in injury mortality between the countries and within countries over time. The observed trends differed by injury cause category. The substantial differences in injury rates between similarly resourced populations raises important questions about the role of societal-level factors as underlying causes of the differential distribution of injury in our communities.

Background.: It is standard practice for laboratories to assess the cellular quality of expectorated sputum specimens to check that they originated from the lower respiratory tract. The presence of low numbers of squamous epithelial cells (SECs) and high numbers of polymorphonuclear (PMN) cells are regarded as indicative of a lower respiratory tract specimen. However, these quality ratings have never been evaluated for induced sputum specimens from children with suspected pneumonia. Methods.: We evaluated induced sputum Gram stain smears and cultures from hospitalized children aged 1-59 months enrolled in a large study of community-acquired pneumonia. We hypothesized that a specimen representative of the lower respiratory tract will contain smaller quantities of oropharyngeal flora and be more likely to have a predominance of potential pathogens compared to a specimen containing mainly saliva. The prevalence of potential pathogens cultured from induced sputum specimens and quantity of oropharyngeal flora were compared for different quantities of SECs and PMNs. Results.: Of 3772 induced sputum specimens, 2608 (69%) had <10 SECs per low-power field (LPF) and 2350 (62%) had >25 PMNs per LPF, measures traditionally associated with specimens from the lower respiratory tract in adults. Using isolation of low quantities of oropharyngeal flora and higher prevalence of potential pathogens as markers of higher quality, <10 SECs per LPF (but not >25 PMNs per LPF) was the microscopic variable most associated with high quality of induced sputum. Conclusions.: Quantity of SECs may be a useful quality measure of induced sputum from young children with pneumonia.

Background.: Antibiotic exposure and specimen volume are known to affect pathogen detection by culture. Here we assess their effects on bacterial pathogen detection by both culture and polymerase chain reaction (PCR) in children. Methods.: PERCH (Pneumonia Etiology Research for Child Health) is a case-control study of pneumonia in children aged 1-59 months investigating pathogens in blood, nasopharyngeal/oropharyngeal (NP/OP) swabs, and induced sputum by culture and PCR. Antibiotic exposure was ascertained by serum bioassay, and for cases, by a record of antibiotic treatment prior to specimen collection. Inoculated blood culture bottles were weighed to estimate volume. Results.: Antibiotic exposure ranged by specimen type from 43.5% to 81.7% in 4223 cases and was detected in 2.3% of 4863 controls. Antibiotics were associated with a 45% reduction in blood culture yield and approximately 20% reduction in yield from induced sputum culture. Reduction in yield of Streptococcus pneumoniae from NP culture was approximately 30% in cases and approximately 32% in controls. Several bacteria had significant but marginal reductions (by 5%-7%) in detection by PCR in NP/OP swabs from both cases and controls, with the exception of S. pneumoniae in exposed controls, which was detected 25% less frequently compared to nonexposed controls. Bacterial detection in induced sputum by PCR decreased 7% for exposed compared to nonexposed cases. For every additional 1 mL of blood culture specimen collected, microbial yield increased 0.51% (95% confidence interval, 0.47%-0.54%), from 2% when volume was ≤1 mL to approximately 6% for ≥3 mL. Conclusions.: Antibiotic exposure and blood culture volume affect detection of bacterial pathogens in children with pneumonia and should be accounted for in studies of etiology and in clinical management.

Pneumonia kills more children each year worldwide than any other disease. Nonetheless, accurately determining the causes of childhood pneumonia has remained elusive. Over the past century, the focus of pneumonia etiology research has shifted from studies of lung aspirates and postmortem specimens intent on identifying pneumococcal disease to studies of multiple specimen types distant from the lung that are tested for multiple pathogens. Some major challenges facing modern pneumonia etiology studies include the use of nonspecific and variable case definitions, poor access to pathologic lung tissue and to specimens from fatal cases, poor diagnostic accuracy of assays (especially when testing nonpulmonary specimens), and the interpretation of results when multiple pathogens are detected in a given individual. The future of childhood pneumonia etiology research will likely require integrating data from complementary approaches, including applications of advanced molecular diagnostics and vaccine probe studies, as well as a renewed emphasis on lung aspirates from radiologically confirmed pneumonia and postmortem examinations.

Background.: Mycobacterium tuberculosis (Mtb) contributes to the pathogenesis of childhood acute community-acquired pneumonia in settings with a high tuberculosis burden. The incremental value of a repeated induced sputum (IS) sample, compared with a single IS or gastric aspirate (GA) sample, is not well known. Methods.: Two IS samples were obtained for Mtb culture from children enrolled as cases in the Pneumonia Etiology Research for Child Health (PERCH) study in South Africa. Nonstudy attending physicians requested GA if pulmonary tuberculosis was clinically suspected. We compared the Mtb yield of 2 IS samples to that of 1 IS sample and GA samples. Results: . Twenty-seven (3.0%) culture-confirmed pulmonary tuberculosis cases were identified among 906 children investigated with IS and GA samples for Mtb. Results from 2 IS samples were available for 719 children (79.4%). Of 12 culture-confirmed pulmonary tuberculosis cases identified among children with ≥2 IS samples, 4 (33.3%) were negative at the first IS sample. In head-to-head comparisons among children with both GA and IS samples collected, the yield of 1 GA sample (8 of 427; 1.9%) was similar to that of 1 IS sample (5 of 427, 1.2%), and the yield of 2 GA samples (10 of 300; 3.3%) was similar to that of 2 IS samples (5 of 300; 1.7%). IS samples identified 8 (42.1%) of the 19 culture-confirmed pulmonary tuberculosis cases that were identified through submission of IS and GA samples. Conclusions.: A single IS sample underestimated the presence of Mtb in children hospitalized with severe or very severe pneumonia. Detection of Mtb is enhanced by combining 2 IS with GA sample collections in young children with acute severe pneumonia.

Over the last 20–30 years, enormous reductions have occurred in the absolute and relative burden of pneumonia mortality in young children around the world. Only 20 years ago, when the population of young children was approximately 625 million, approximately 1.7 million young children died from pneumonia before their 5th birthday (Figure 1) [1–4]. Mortality from pneumonia among children aged <5 years fell to 921 000 in 2015, whereas the population of young children rose to >670 million [1, 2, 5]. This remarkable improvement in child survival and health has resulted from advances in social conditions and economic development [6] but has also been influenced by at least 4 pivotal innovations: (1) the development of a global vaccination program, the World Health Organization’s Expanded Program on Immunizations (begun in 1974), which created the architecture around which country investments, donor funding, program strategies, and outcome measurements could be envisioned and implemented; (2) the global consensus to focus funding, programs, and momentum on 6 development targets articulated by the United Nations General Assembly through the Millennium Development Goals (MDGs, agreed upon in 2000) with MDG4 targeting child survival; (3) the advent of large, health-focused nongovernmental organizations; and (4) the founding of the Global Alliance for Vaccines and Immunization (the Gavi Alliance, formally launched at the World Economic Forum in January 2000), a multilateral funding organization that has allowed for an unprecedented pace of introduction and expanded use of life-saving vaccines in low-income countries. In part, as a result of this multidimensional, multisectoral consensus approach enacted through critical large-scale investments in prevention, protection, and treatment, pneumonia mortality has fallen substantially in many parts of the world because the most fatal of the pathogens and the underlying conditions that put children at risk are being targeted.

Background.: The etiologic inference of identifying a pathogen in the upper respiratory tract (URT) of children with pneumonia is unclear. To determine if viral load could provide evidence of causality of pneumonia, we compared viral load in the URT of children with World Health Organization-defined severe and very severe pneumonia and age-matched community controls. Methods.: In the 9 developing country sites, nasopharyngeal/oropharyngeal swabs from children with and without pneumonia were tested using quantitative real-time polymerase chain reaction for 17 viruses. The association of viral load with case status was evaluated using logistic regression. Receiver operating characteristic (ROC) curves were constructed to determine optimal discriminatory viral load cutoffs. Viral load density distributions were plotted. Results.: The mean viral load was higher in cases than controls for 7 viruses. However, there was substantial overlap in viral load distribution of cases and controls for all viruses. ROC curves to determine the optimal viral load cutoff produced an area under the curve of <0.80 for all viruses, suggesting poor to fair discrimination between cases and controls. Fatal and very severe pneumonia cases did not have higher viral load than less severe cases for most viruses. Conclusions.: Although we found higher viral loads among pneumonia cases than controls for some viruses, the utility in using viral load of URT specimens to define viral pneumonia was equivocal. Our analysis was limited by lack of a gold standard for viral pneumonia.

Despite tremendous advances in diagnostic laboratory technology, identifying the pathogen(s) causing pneumonia remains challenging because the infected lung tissue cannot usually be sampled for testing. Consequently, to obtain information about pneumonia etiology, clinicians and researchers test specimens distant to the site of infection. These tests may lack sensitivity (eg, blood culture, which is only positive in a small proportion of children with pneumonia) and/or specificity (eg, detection of pathogens in upper respiratory tract specimens, which may indicate asymptomatic carriage or a less severe syndrome, such as upper respiratory infection). While highly sensitive nucleic acid detection methods and testing of multiple specimens improve sensitivity, multiple pathogens are often detected and this adds complexity to the interpretation as the etiologic significance of results may be unclear (ie, the pneumonia may be caused by none, one, some, or all of the pathogens detected). Some of these challenges can be addressed by adjusting positivity rates to account for poor sensitivity or incorporating test results from controls without pneumonia to account for poor specificity. However, no classical analytic methods can account for measurement error (ie, sensitivity and specificity) for multiple specimen types and integrate the results of measurements for multiple pathogens to produce an accurate understanding of etiology. We describe the major analytic challenges in determining pneumonia etiology and review how the common analytical approaches (eg, descriptive, case-control, attributable fraction, latent class analysis) address some but not all challenges. We demonstrate how these limitations necessitate a new, integrated analytical approach to pneumonia etiology data.

Background.: Lack of a gold standard for identifying bacterial and viral etiologies of pneumonia has limited evaluation of C-reactive protein (CRP) for identifying bacterial pneumonia. We evaluated the sensitivity and specificity of CRP for identifying bacterial vs respiratory syncytial virus (RSV) pneumonia in the Pneumonia Etiology Research for Child Health (PERCH) multicenter case-control study. Methods.: We measured serum CRP levels in cases with World Health Organization-defined severe or very severe pneumonia and a subset of community controls. We evaluated the sensitivity and specificity of elevated CRP for "confirmed" bacterial pneumonia (positive blood culture or positive lung aspirate or pleural fluid culture or polymerase chain reaction [PCR]) compared to "RSV pneumonia" (nasopharyngeal/oropharyngeal or induced sputum PCR-positive without confirmed/suspected bacterial pneumonia). Receiver operating characteristic (ROC) curves were constructed to assess the performance of elevated CRP in distinguishing these cases. Results.: Among 601 human immunodeficiency virus (HIV)-negative tested controls, 3% had CRP ≥40 mg/L. Among 119 HIV-negative cases with confirmed bacterial pneumonia, 77% had CRP ≥40 mg/L compared with 17% of 556 RSV pneumonia cases. The ROC analysis produced an area under the curve of 0.87, indicating very good discrimination; a cut-point of 37.1 mg/L best discriminated confirmed bacterial pneumonia (sensitivity 77%) from RSV pneumonia (specificity 82%). CRP ≥100 mg/L substantially improved specificity over CRP ≥40 mg/L, though at a loss to sensitivity. Conclusions.: Elevated CRP was positively associated with confirmed bacterial pneumonia and negatively associated with RSV pneumonia in PERCH. CRP may be useful for distinguishing bacterial from RSV-associated pneumonia, although its role in discriminating against other respiratory viral-associated pneumonia needs further study.

In pneumonia, specimens are rarely obtained directly from the infection site, the lung, so the pathogen causing infection is determined indirectly from multiple tests on peripheral clinical specimens, which may have imperfect and uncertain sensitivity and specificity, so inference about the cause is complex. Analytic approaches have included expert review of case-only results, case-control logistic regression, latent class analysis, and attributable fraction, but each has serious limitations and none naturally integrate multiple test results. The Pneumonia Etiology Research for Child Health (PERCH) study required an analytic solution appropriate for a case-control design that could incorporate evidence from multiple specimens from cases and controls and that accounted for measurement error. We describe a Bayesian integrated approach we developed that combined and extended elements of attributable fraction and latent class analyses to meet some of these challenges and illustrate the advantage it confers regarding the challenges identified for other methods.

Background: Chest radiographs (CXRs) are frequently used to assess pneumonia cases. Variations in CXR appearances between epidemiological settings and their correlation with clinical signs are not well documented. Methods: The Pneumonia Etiology Research for Child Health project enrolled 4232 cases of hospitalized World Health Organization (WHO)-defined severe and very severe pneumonia from 9 sites in 7 countries (Bangladesh, the Gambia, Kenya, Mali, South Africa, Thailand, and Zambia). At admission, each case underwent a standardized assessment of clinical signs and pneumonia risk factors by trained health personnel, and a CXR was taken that was interpreted using the standardized WHO methodology. CXRs were categorized as abnormal (consolidation and/or other infiltrate), normal, or uninterpretable. Results: CXRs were interpretable in 3587 (85%) cases, of which 1935 (54%) were abnormal (site range, 35%-64%). Cases with abnormal CXRs were more likely than those with normal CXRs to have hypoxemia (45% vs 26%), crackles (69% vs 62%), tachypnea (85% vs 80%), or fever (20% vs 16%) and less likely to have wheeze (30% vs 38%; all P < .05). CXR consolidation was associated with a higher case fatality ratio at 30-day follow-up (13.5%) compared to other infiltrate (4.7%) or normal (4.9%) CXRs. Conclusions: Clinically diagnosed pneumonia cases with abnormal CXRs were more likely to have signs typically associated with pneumonia. However, CXR-normal cases were common, and clinical signs considered indicative of pneumonia were present in substantial proportions of these cases. CXR-consolidation cases represent a group with an increased likelihood of death at 30 days post-discharge.

The Pneumonia Etiology Research for Child Health (PERCH) study is the largest multicountry etiology study of pediatric pneumonia undertaken in the past 3 decades. The study enrolled 4232 hospitalized cases and 5325 controls over 2 years across 9 research sites in 7 countries in Africa and Asia. The volume and complexity of data collection in PERCH presented considerable logistical and technical challenges. The project chose an internet-based data entry system to allow real-time access to the data, enabling the project to monitor and clean incoming data and perform preliminary analyses throughout the study. To ensure high-quality data, the project developed comprehensive quality indicator, data query, and monitoring reports. Among the approximately 9000 cases and controls, analyzable laboratory results were available for ≥96% of core specimens collected. Selected approaches to data management in PERCH may be extended to the planning and organization of international studies of similar scope and complexity.

Background.: Sputum microscopy and culture are commonly used for diagnosing the cause of pneumonia in adults but are rarely performed in children due to difficulties in obtaining specimens. Induced sputum is occasionally used to investigate lower respiratory infections in children but has not been widely used in pneumonia etiology studies. Methods.: We evaluated the diagnostic utility of induced sputum microscopy and culture in patients enrolled in the Pneumonia Etiology Research for Child Health (PERCH) study, a large study of community-acquired pneumonia in children aged 1-59 months. Comparisons were made between induced sputum samples from hospitalized children with radiographically confirmed pneumonia and children categorized as nonpneumonia (due to the absence of prespecified clinical and laboratory signs and absence of infiltrate on chest radiograph). Results.: One induced sputum sample was available for analysis from 3772 (89.1%) of 4232 suspected pneumonia cases enrolled in PERCH. Of these, sputum from 2608 (69.1%) met the quality criterion of <10 squamous epithelial cells per low-power field, and 1162 (44.6%) had radiographic pneumonia. Induced sputum microscopy and culture results were not associated with radiographic pneumonia, regardless of prior antibiotic use, stratification by specific bacteria, or interpretative criteria used. Conclusions.: The findings of this study do not support the culture of induced sputum specimens as a diagnostic tool for pneumonia in young children as part of routine clinical practice.

Background.: We investigated the performance of polymerase chain reaction (PCR) on blood in the diagnosis of pneumococcal pneumonia among children from 7 low- and middle-income countries. Methods.: We tested blood by PCR for the pneumococcal autolysin gene in children aged 1-59 months in the Pneumonia Etiology Research for Child Health (PERCH) study. Children had World Health Organization-defined severe or very severe pneumonia or were age-frequency-matched community controls. Additionally, we tested blood from general pediatric admissions in Kilifi, Kenya, a PERCH site. The proportion PCR-positive was compared among cases with microbiologically confirmed pneumococcal pneumonia (MCPP), cases without a confirmed bacterial infection (nonconfirmed), cases confirmed for nonpneumococcal bacteria, and controls. Results.: In PERCH, 7.3% (n = 291/3995) of cases and 5.5% (n = 273/4987) of controls were blood pneumococcal PCR-positive (P < .001), compared with 64.3% (n = 36/56) of MCPP cases and 6.3% (n = 243/3832) of nonconfirmed cases (P < .001). Blood pneumococcal PCR positivity was higher in children from the 5 African countries (5.5%-11.5% among cases and 5.3%-10.2% among controls) than from the 2 Asian countries (1.3% and 1.0% among cases and 0.8% and 0.8% among controls). Among Kilifi general pediatric admissions, 3.9% (n = 274/6968) were PCR-positive, including 61.7% (n = 37/60) of those with positive blood cultures for pneumococcus. Discussion.: The utility of pneumococcal PCR on blood for diagnosing childhood pneumococcal pneumonia in the 7 low- and middle-income countries studied is limited by poor specificity and by poor sensitivity among MCPP cases.

Background.: Detection of pneumococcus by lytA polymerase chain reaction (PCR) in blood had poor diagnostic accuracy for diagnosing pneumococcal pneumonia in children in 9 African and Asian sites. We assessed the value of blood lytA quantification in diagnosing pneumococcal pneumonia. Methods.: The Pneumonia Etiology Research for Child Health (PERCH) case-control study tested whole blood by PCR for pneumococcus in children aged 1-59 months hospitalized with signs of pneumonia and in age-frequency matched community controls. The distribution of load among PCR-positive participants was compared between microbiologically confirmed pneumococcal pneumonia (MCPP) cases, cases confirmed for nonpneumococcal pathogens, nonconfirmed cases, and controls. Receiver operating characteristic analyses determined the "optimal threshold" that distinguished MCPP cases from controls. Results.: Load was available for 290 of 291 cases with pneumococcal PCR detected in blood and 273 of 273 controls. Load was higher in MCPP cases than controls (median, 4.0 x 103 vs 0.19 x 103 copies/mL), but overlapped substantially (range, 0.16-989.9 x 103 copies/mL and 0.01-551.9 x 103 copies/mL, respectively). The proportion with high load (≥2.2 log10 copies/mL) was 62.5% among MCPP cases, 4.3% among nonconfirmed cases, 9.3% among cases confirmed for a nonpneumococcal pathogen, and 3.1% among controls. Pneumococcal load in blood was not associated with respiratory tract illness in controls (P = .32). High blood pneumococcal load was associated with alveolar consolidation on chest radiograph in nonconfirmed cases, and with high (>6.9 log10 copies/mL) nasopharyngeal/oropharyngeal load and C-reactive protein ≥40 mg/L (both P < .01) in nonconfirmed cases but not controls. Conclusions.: Quantitative pneumococcal PCR in blood has limited diagnostic utility for identifying pneumococcal pneumonia in individual children, but may be informative in epidemiological studies.

Background.: There is limited information on the association between colonization density of upper respiratory tract colonizers and pathogen-specific pneumonia. We assessed this association for Haemophilus influenzae, Moraxella catarrhalis, Staphylococcus aureus, and Pneumocystis jirovecii. Methods.: In 7 low- and middle-income countries, nasopharyngeal/oropharyngeal swabs from children with severe pneumonia and age-frequency matched community controls were tested using quantitative polymerase chain reaction (PCR). Differences in median colonization density were evaluated using the Wilcoxon rank-sum test. Density cutoffs were determined using receiver operating characteristic curves. Cases with a pathogen identified from lung aspirate culture or PCR, pleural fluid culture or PCR, blood culture, and immunofluorescence for P. jirovecii defined microbiologically confirmed cases for the given pathogens. Results.: Higher densities of H. influenzae were observed in both microbiologically confirmed cases and chest radiograph (CXR)-positive cases compared to controls. Staphylococcus aureus and P. jirovecii had higher densities in CXR-positive cases vs controls. A 5.9 log10 copies/mL density cutoff for H. influenzae yielded 86% sensitivity and 77% specificity for detecting microbiologically confirmed cases; however, densities overlapped between cases and controls and positive predictive values were poor (<3%). Informative density cutoffs were not found for S. aureus and M. catarrhalis, and a lack of confirmed case data limited the cutoff identification for P. jirovecii. Conclusions.: There is evidence for an association between H. influenzae colonization density and H. influenzae-confirmed pneumonia in children; the association may be particularly informative in epidemiologic studies. Colonization densities of M. catarrhalis, S. aureus, and P. jirovecii are unlikely to be of diagnostic value in clinical settings.

Background.: Sputum examination can be useful in diagnosing the cause of pneumonia in adults but is less well established in children. We sought to assess the diagnostic utility of polymerase chain reaction (PCR) for detection of respiratory viruses and bacteria in induced sputum (IS) specimens from children hospitalized with severe or very severe pneumonia. Methods.: Among children aged 1-59 months, we compared organism detection by multiplex PCR in IS and nasopharyngeal/oropharyngeal (NP/OP) specimens. To assess whether organism presence or density in IS specimens was associated with chest radiographic evidence of pneumonia (radiographic pneumonia), we compared prevalence and density in IS specimens from children with radiographic pneumonia and children with suspected pneumonia but without chest radiographic changes or clinical or laboratory findings suggestive of pneumonia (nonpneumonia group). Results.: Among 4232 cases with World Health Organization-defined severe or very severe pneumonia, we identified 1935 (45.7%) with radiographic pneumonia and 573 (13.5%) with nonpneumonia. The organism detection yield was marginally improved with IS specimens (96.2% vs 92.4% for NP/OP specimens for all viruses combined [P = .41]; 96.9% vs 93.3% for all bacteria combined [P = .01]). After accounting for presence in NP/OP specimens, no organism was detected more frequently in the IS specimens from the radiographic pneumonia compared with the nonpneumonia cases. Among high-quality IS specimens, there were no statistically significant differences in organism density, except with cytomegalovirus, for which there was a higher quantity in the IS specimens from cases with radiographic pneumonia compared with the nonpneumonia cases (median cycle threshold value, 27.9 vs 28.5, respectively; P = .01). Conclusions.: Using advanced molecular methods with IS specimens provided little additional diagnostic information beyond that obtained with NP/OP swab specimens.

Background.: Previous studies suggested an association between upper airway pneumococcal colonization density and pneumococcal pneumonia, but data in children are limited. Using data from the Pneumonia Etiology Research for Child Health (PERCH) study, we assessed this potential association. Methods.: PERCH is a case-control study in 7 countries: Bangladesh, The Gambia, Kenya, Mali, South Africa, Thailand, and Zambia. Cases were children aged 1-59 months hospitalized with World Health Organization-defined severe or very severe pneumonia. Controls were randomly selected from the community. Microbiologically confirmed pneumococcal pneumonia (MCPP) was confirmed by detection of pneumococcus in a relevant normally sterile body fluid. Colonization density was calculated with quantitative polymerase chain reaction analysis of nasopharyngeal/oropharyngeal specimens. Results.: Median colonization density among 56 cases with MCPP (MCPP cases; 17.28 x 106 copies/mL) exceeded that of cases without MCPP (non-MCPP cases; 0.75 x 106) and controls (0.60 x 106) (each P < .001). The optimal density for discriminating MCPP cases from controls using the Youden index was >6.9 log10 copies/mL; overall, the sensitivity was 64% and the specificity 92%, with variable performance by site. The threshold was lower (≥4.4 log10 copies/mL) when MCPP cases were distinguished from controls who received antibiotics before specimen collection. Among the 4035 non-MCPP cases, 500 (12%) had pneumococcal colonization density >6.9 log10 copies/mL; above this cutoff was associated with alveolar consolidation at chest radiography, very severe pneumonia, oxygen saturation <92%, C-reactive protein ≥40 mg/L, and lack of antibiotic pretreatment (all P< .001). Conclusions.: Pneumococcal colonization density >6.9 log10 copies/mL was strongly associated with MCPP and could be used to improve estimates of pneumococcal pneumonia prevalence in childhood pneumonia studies. Our findings do not support its use for individual diagnosis in a clinical setting.