Articles LancetMicrobe2024; 5: e131–41 Published Online January 11, 2024 https://doi.org/10.1016/ S2666-5247(23)00290-2 Center for Global Health, US Centers for Disease Control and Prevention, Nairobi, Kenya (J R Verani MD); Center for Global Health, US Centers for Disease Control and Prevention, Atlanta, GA, USA (D M Blau PhD, Z J Madewell PhD); Maternal and Child Health Division, International Center for Diarrhoeal Diseases Research (icddr,b), Dhaka, Bangladesh (Prof E S Gurley PhD, S El Arifeen DrPH); Departmentof Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA (Prof E S Gurley); Center for Global Health, US Centers for Disease Control and Prevention Kenya, Kisumu, Kenya (V Akelo MD); College of Health and Medical Sciences, Haramaya University, Harar, Ethiopia (Prof N Assefa PhD, M Berhane MSc, L Madrid MD, J O Oundo PhD); South African Medical Research Council Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa (V Baillie PhD, Prof S Madhi PhD, S Mahtab PhD); Child deaths caused by Klebsiella pneumoniae in sub-Saharan Africa and south Asia: a secondary analysis of Child Health and Mortality Prevention Surveillance (CHAMPS) data Jennifer R Verani, Dianna M Blau, Emily S Gurley, Victor Akelo, Nega Assefa, Vicky Baillie, Quique Bassat, Mussie Berhane, James Bunn, Anelsio C A Cossa, Shams El Arifeen, Revathi Gunturu, Martin Hale, Aggrey Igunza, Adama M Keita, Sartie Kenneh, Karen L Kotloff, Dickens Kowuor, Rita Mabunda, Zachary J Madewell, Shabir Madhi, Lola Madrid, Sana Mahtab, Judice Miguel, Florence V Murila, Ikechukwu U Ogbuanu, Julius Ojulong, Dickens Onyango, Joe O Oundo, J Anthony G Scott, Samba Sow, Milagritos Tapia, Cheick B Traore, Sithembiso Velaphi, Cynthia G Whitney, Inacio Mandomando, Robert F Breiman Summary Background Klebsiella pneumoniae is an important cause of nosocomial and community-acquired pneumonia and sepsis in children, and antibiotic-resistant K pneumoniae is a growing public health threat. We aimed to characterise child mortality associated with this pathogen in seven high-mortality settings. Methods We analysed Child Health and Mortality Prevention Surveillance (CHAMPS) data on the causes of deaths in children younger than 5 years and stillbirths in sites located in seven countries across sub-Saharan Africa (Ethiopia, Kenya, Mali, Mozambique, Sierra Leone, and South Africa) and south Asia (Bangladesh) from Dec 9, 2016, to Dec 31, 2021. CHAMPS sites conduct active surveillance for deaths in catchment populations and following reporting of an eligible death or stillbirth seek consent for minimally invasive tissue sampling followed by extensive aetiological testing (microbiological, molecular, and pathological); cases are reviewed by expert panels to assign immediate, intermediate, and underlying causes of death. We reported on susceptibility to antibiotics for which at least 30 isolates had been tested, and excluded data on antibiotics for which susceptibility testing is not recommended for Klebsiella spp due to lack of clinical activity (eg, penicillin and ampicillin). Findings Among 2352 child deaths with cause of death assigned, 497 (21%, 95% CI 20–23) had K pneumoniae in the causal chain of death; 100 (20%, 17–24) had K pneumoniae as the underlying cause. The frequency of K pneumoniae in the causal chain was highest in children aged 1–11 months (30%, 95% CI 26–34; 144 of 485 deaths) and 12–23 months (28%, 22–34; 63 of 225 deaths); frequency by site ranged from 6% (95% CI 3–11; 11 of 184 deaths) in Bangladesh to 52% (44–61; 71 of 136 deaths) in Ethiopia. K pneumoniae was in the causal chain for 450 (22%, 95% CI 20–24) of 2023 deaths that occurred in health facilities and 47 (14%, 11–19) of 329 deaths in the community. The most common clinical syndromes among deaths with K pneumoniae in the causal chain were sepsis (44%, 95% CI 40–49; 221 of 2352 deaths), sepsis in conjunction with pneumonia (19%, 16–23; 94 of 2352 deaths), and pneumonia (16%, 13–20; 80 of 2352 deaths). Among K pneumoniae isolates tested, 121 (84%) of 144 were resistant to ceftriaxone and 80 (75%) of 106 to gentamicin. Interpretation K pneumoniae substantially contributed to deaths in the first 2 years of life across multiple high-mortality settings, and resistance to antibiotics used for sepsis treatment was common. Improved strategies are needed to rapidly identify and appropriately treat children who might be infected with this pathogen. These data suggest a potential impact of developing and using effective K pneumoniae vaccines in reducing neonatal, infant, and child deaths globally. Funding Bill & Melinda Gates Foundation. Copyright Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons. org/licenses/by/4.0/). ISGlobal - Hospital Clínic, Universitat de Barcelona, Barcelona, Spain (ProfQBassatMD,RMabundaMSc); Centro de Investigação em Saúde de Manhiça (CISM), Maputo, Mozambique (Prof Q Bassat, A C A Cossa DVM, J Miguel BSc, I Mandomando PhD); Institució Catalana de Recerca I Estudis Avançats (ICREA), Barcelona, Spain (Prof QBassat); Hospital Sant Joan de Déu, Barcelona, Introduction Klebsiella pneumoniae are Gram-negative bacteria and an important nosocomial pathogen, particularly in immuno- compromised adults and neonates.1–3 Studies have also highlighted the role of K pneumoniae as a cause of community-acquired infections, including neonatal sepsis.3,4 K pneumoniae commonly colonises humans, with prevalence ranging from1% to 19% in thenasopharynx and 5% to 38% in the gastrointestinal tract.1,5 Neonates born to mothers with vaginal K pneumoniae colonisation can www.thelancet.com/microbe Vol 5 February 2024 become colonised during the perinatal period.6 K pneumo- niae can remain commensal, or infection can lead to illness with a wide range of clinical manifestations, including pneumonia, sepsis, genitourinary tract infections, intra- abdominal infections, meningitis, and soft tissue infec- tions. Risk factors for progression from colonisation to disease are poorly understood,5 and geographical variation in the burden and clinical manifestations of community- acquired infections point to gaps in understanding of K pneumoniae epidemiology.7 e131 http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ http://crossmark.crossref.org/dialog/?doi=10.1016/S2666-5247(23)00290-2&domain=pdf https://doi.org/10.1016/S2666-5247(23)00290-2 https://doi.org/10.1016/S2666-5247(23)00290-2 www.thelancet.com/microbe Spain (Prof Q Bassat); Consorcio de Investigación Biomédica en Red de Epidemiología y Salud Pública (CIBERESP), Barcelona, Spain (Prof Q Bassat); World Health Organization, Sierra Leone, Freetown, Sierra Leone (J Bunn MSc); Aga Khan University Hospital, Nairobi, Kenya (R Gunturu MD); National Health Laboratory Service, Department of Anatomical Pathology, School of Pathology, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa (Prof M Hale MD); Kenya Medical Research Institute (KEMRI), Kisumu, Kenya (A Igunza BSc); Centre pour le Développement des Vaccins (CVD-Mali), Ministère de la Santé, Bamako, Mali (A M Keita MD, S Sow MD); Ministry of Health and Sanitation, Freetown, Sierra Leone (S Kenneh MPH); Department of Pediatrics and Department of Medicine, Center for Vaccine Development and Global Health, University of Maryland School of Medicine, Baltimore, MD, USA (Prof K L Kotloff MD, Prof M Tapia MD); Crown Agents, Freetown, Sierra Leone (D Kowuor MSc, I U Ogbuanu MD); Department of Infectious Disease Epidemiology, London School of Hygiene & Tropical Medicine, London, UK (L Madrid, J O Oundo, Prof J A G Scott FMedSci); University of Nairobi, Nairobi, Kenya (F V Murila MMED); ICAP – Columbia University, Makeni, Sierra Leone (J Ojulong MSc); Kisumu County Department of Health, Kisumu, Kenya (D Onyango MD); Department of Pathological Anatomy and Cytology, University Hospital of Point G, Bamako, Mali (C B Traore MD); Department of Pediatrics, Chris Hani Baragwanath Academic Hospital, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa (Prof S Velaphi PhD); Department of Global Health, Rollins School of Public Health, Emory University, Atlanta, GA, USA (Prof C G Whitney MD, Prof R F Breiman MD); Instituto Nacional de Saúde (INS), Maputo, Mozambique Research in context Evidence before this study We searched PubMed for articles published from Jan 1, 2012, to March 31, 2022, using “Klebsiella” or “Klebsiella pneumoniae” as search terms, and focused on evidence of K pneumoniae disease in children and from low-income or middle-income settings. Numerous publications have reported on the role of Klebsiella pneumoniae as a cause of nosocomial pneumonia, particularly in immunocompromised individuals and neonates. K pneumoniae is described in published literature as an important cause of outbreaks in neonatal intensive care units. Community-acquired K pneumoniae infections are also described in the literature, although fewer studies have focused on this. Various publications reported on increasing antimicrobial resistance among K pneumoniae isolates, includingmultidrug-resistant and extremely drug-resistant strains. We also considered early findings from the Child Health and Mortality Prevention and Surveillance (CHAMPS) network, which aims to generate data on the causes of child deaths in areas with high child mortality in sub-Saharan Africa and south Asia. The first comprehensive publication of CHAMPS findings included data from five sites from December, 2016, to December, 2018, and reported that 53% of neonatal deaths and 90% of child deaths (ages 1–59months) had at least one infectious condition contributing to the death; among deaths with a contributory pathogen identified, K pneumoniae was considered contributory to 31%, making it the second most frequently implicated pathogen in neonatal deaths and the most common among child deaths. However, beyond this description of the frequency with which K pneumoniae contributes to neonatal and child deaths, data on the role of K pneumoniae in childmortality are limited. Added value of this study This study used data from CHAMPS to provide more in-depth and current data on the contribution of K pneumoniae to child mortality, reporting on data from seven countries up to December, 2021. CHAMPS generates detailed and robust data on the causes of death among stillbirths and deaths among children aged 0–59 months. Sites conduct active surveillance for deaths among defined catchment populations and seek to enrol deaths and conduct minimally invasive tissue sampling within 24 h of the death. Samples undergo extensive aetiological testing (microbiological, molecular, and pathological). Expert panels review comprehensive data from medical record review, verbal autopsy, and laboratory results; panels assign immediate, intermediate, and underlying causes of death. In this study, we build on the previous CHAMPS report on the frequency of K pneumoniae contributing to child deaths by examining the role of K pneumoniae in deaths stratified by age group and site, describing K pneumoniae-associated clinical syndromes among fatal cases, and reporting on pathogens and underlying conditions that also contributed to deaths in which K pneumoniae played a role. We also present data on the antimicrobial susceptibility patterns of K pneumoniae isolates from child deaths with K pneumoniae in the causal chain of death. Implications of all the available evidence The findings of this study further highlight the importance of K pneumoniae as a cause of death among children aged 0–59months in high-mortality settings across seven countries and provide evidence of both nosocomial and community-acquired K pneumoniae infections in these settings. Data on the age groups most commonly affected and with the highest burden of K pneumoniae-related deaths are crucial for informing prevention strategies, including potential development of vaccines against K pneumoniae. The data on antimicrobial susceptibility among K pneumoniae isolates from this study, together with existing evidence of increasing resistance among K pneumoniae, raise alarms about the ineffectiveness of commonly used antibiotics. Guidelines for empirical treatment of pneumonia and sepsis might require adaptation to optimally prevent deaths due to K pneumoniae, and infection prevention and control practices and antimicrobial stewardship efforts are crucial for curbing the spread of resistant K pneumoniae strains. The available evidence suggest that specific efforts aimed at reducing K pneumoniae-associated deaths will likely be necessary in order to meet Sustainable Development Goals for under-5 and neonatal mortality by 2030. Articles e132 Increasing antimicrobial resistance among K pneumoniae isolates and the emergence of hypervirulent strains have compounded the threat from this pathogen.8 Multidrug- resistant and extremely drug-resistant strains are increas- ingly common, greatly limiting the available antibiotic options for treatingKpneumoniae infections.9,10 Studieshave reported that carbapenem-resistant K pneumoniae strains are associated with increased case fatality.11,12 The role of K pneumoniae in child death is not well understood, particularly among children in resource-poor settings. The Child Health and Mortality Prevention Sur- veillance (CHAMPS) network tracks the burden and causes of deaths in children younger than 5 years and stillbirths in seven sites across sub-Saharan Africa and south Asia.13 Initial findings (published in 2020) from the first five sites foundKpneumoniae to be a common contributory pathogen among child deaths with an infectious cause.14 We aim to characterise key clinical and epidemiological features of K pneumoniae-associated child mortality and stillbirths in CHAMPS site settings. Methods Study design and population We conducted a secondary analysis of CHAMPS data. The CHAMPSmethods for identifying deaths and determining the causes among children younger than 5 years and still- births have been described elsewhere13,14 and are briefly summarised here. This analysis included data from CHAMPS surveillance sites located in seven countries: Baliakandi, Bangladesh; Harar, Haramaya, and Kersa, www.thelancet.com/microbe Vol 5 February 2024 https://champshealth.org/our_network/ www.thelancet.com/microbe (I Mandomando); Infectious Diseases and Oncology Research Institute, University of the Witwatersrand, Johannesburg, South Africa (Prof R F Breiman) Correspondence to: Jennifer R Verani, Center for Global Health, US Centers for Disease Control and Prevention, PO Box 606-00621, Village Market, Nairobi, Kenya jverani@cdc.gov For more on CHAMPS sites see https://champshealth.org/our_ network/ For CHAMPS protocols see https://champshealth.org/ resources/protocols Articles Ethiopia; Siaya and Kisumu, Kenya; Bamako, Mali; Quel- imane and Manhiça, Mozambique; Bombali, Sierra Leone; and Soweto, South Africa. All sites are high-mortality set- tings, and the network includes a mix of rural and urban sites.15 Stillbirths and deaths among children younger than 5 yearswho reside in catchment areas are promptly reported from health facilities and via networks of community reporters.Upon receivingnotification of aneligible stillbirth or death, CHAMPS staff rapidly approach family members to screen for eligibility and seek written informed consent for collection of data and specimens. Ethics committees overseeing investigators at each site and at Emory University approved overall and site-specific protocols (Emory institutional review board number 00091706). Protocols are available on theCHAMPSwebsite. Procedures Data collection (verbal autopsy using the WHO 2016 instrument, clinical record abstraction, anthropometric measurements, and photographs) was performed using standardised procedures by trained staff. Deaths for which CHAMPS staff were notified within 36 h (or ≤72 h if post- mortem refrigeration was used) were eligible for minimally invasive tissue sampling (MITS). MITS specimens were collected using biopsy needles under sterile conditions. Tissues sampled included the lungs, heart, brain, liver, bone marrow, and placenta (for stillbirths). In addition, samples of blood, cerebrospinal fluid (CSF), stool, and nasopharyngeal secretions were collected. Laboratories at each site performed culture of blood and CSF using standard microbiological procedures. For sites with automated culture systems, identification of Kpneumoniaewasperformedby automatedsystems (Vitek2 [bioMerieux, Hazelwood, MO, USA], BD Phoenix [BD, Franklin Lakes, NJ, USA] or MicroScan [Beckman Coulter, Brea, CA, USA]). For sites without access to automated systems, specimenswith growthwere inoculatedonto blood and MacConkey agars and incubated for 18–24 h; identifi- cation of pure colonies was performed by biochemical reactions and colonymorphology, and further confirmation was performed by K pneumoniae-specific PCR. Antimicrobial susceptibility testing of K pneumoniae iso- lates was not a standardised CHAMPS process; however, sites used either automated systems (Vitek 2, BD Phoenix, or MicroScan) or the Kirby–Bauer disk diffusion method. Isolates were classified as susceptible, intermediate, or resistant according to Clinical and Laboratory Standards Institute guidance.16 Specific antibiotics tested and propor- tion of K pneumoniae isolates that underwent susceptibility testing varied by site; we reported on antibiotics for which at least 30 isolates had been tested, and excluded data on antibiotics for which susceptibility testing is not recom- mended for Klebsiella spp due to lack of clinical activity (eg, penicillin and ampicillin).16 Site laboratories also testedpost- mortemblood samples forHIV andprepared and examined thin and thick smears for malaria using microscopy. Five custom TaqMan Array Cards (TAC, ThermoFisher www.thelancet.com/microbe Vol 5 February 2024 Scientific, Waltham, MA, USA) were used for molecular detection of 116 pathogens from lung tissue, blood, CSF, rectal swabs, and nasopharyngeal swabs.17 K pneumoniae targets were included on the TAC panels for lung tissue, blood, CSF, and nasopharyngeal swabs. Histopathological analyses of all tissues using routine stains was performed at site pathology laboratories and at theUSCenters forDisease Control and Prevention (CDC) in Atlanta, GA, USA. Add- itionally, when indicated by TAC, blood culture, or histo- pathology findings, additional testing including Gram stain and immunohistochemistry (IHC) was done at the CDC pathology laboratories. The select Gram-negative bacterial IHC assay uses an antibody raised againstKpneumoniae but is known to be broadly cross-reactive with other Gram- negative bacteria of relevance to CHAMPS cases, includ- ing but not limited to Escherichia coli, Pseudomonas spp, and Haemophilus influenzae. Determination of cause of death (DeCoDe) panels at each site included paediatricians, obstetricians, epidemiologists, pathologists, and microbiologists.18 The panels reviewed all available data for each case and assigned the chain of events leading to deathusing the tenth revision of the International Classification of Diseases (ICD-10) for deaths among chil- dren aged older than 28 days, and ICD-perinatal mortality (ICD-PM) for deaths in the perinatal or neonatal period. For deathswithmore thanone cause, thepanels assigned (1) the immediate or most proximal cause, (2) antecedent or comorbid causes that contributed directly to the chain, and (3) the underlying cause that initiated the chain of events that led to thedeath.Wedefined the causal chain for death to be conditions included inanyof the three categories.For any pathogendetected, theDeCoDe standards provided levels of certainty for the finding, by linking microbiological with histopathological and clinical data.19 The DeCoDe process was standardised across sites using network-wide diagnosis standards and through training.18 Outcomes and statistical analysis We conducted a descriptive analysis of deaths with K pneumoniae in the causal chain from Dec 9, 2016, to Dec 31, 2021. Because K pneumoniae was infrequently deemed to be in the causal chain of death for stillbirths, we present only the overall proportion of stillbirths with Kpneumoniae in the causal chain; subsequent analyseswere restricted to child deaths only. Deaths were classified as community or health facility deaths based on the location of the child at the timeof death; facility deathswere stratified in those occurring 48 h or less after admission, for which the source of K pneumoniae was more likely from the community, and those occurring more than 48 h after admission, for which the source of infection was more likely nosocomial. We used a multivariable logistic regres- sion model to examine the association of age, location of death, and sitewithhavingKpneumoniae in the causal chain (vs death without K pneumoniae in the causal chain). All statistical analyses were conducted using R software, version 4.1.2. e133 https://champshealth.org/resources/protocols mailto:jverani@cdc.gov https://champshealth.org/our_network/ https://champshealth.org/our_network/ https://champshealth.org/resources/protocols https://champshealth.org/resources/protocols www.thelancet.com/microbe Stillbirths Neonates 1–11 months 12–23 months 24–59 months Child deaths at all ages* n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) Deaths with MITS with COD determination Overall 1297 ⋅⋅ 1424 ⋅⋅ 485 ⋅⋅ 225 ⋅⋅ 218 ⋅⋅ 2352 ⋅⋅ Deaths with K pneumoniae in causal chain of death Overall 5 0% (0–1) 257 18% (16–20) 144 30% (26–34) 63 28% (22–34) 33 15% (11–21) 497 21% (20–23) K pneumoniae underlying COD† 4 80% (30–99) 62 24% (19–30) 28 19% (14–27) 8 13% (6–24) 2 6% (1–22) 100 20% (17–24) K pneumoniae only cause 2 40% (7–83) 21 8% (5–12) 4 3% (1–7) 0 0% (0–7) 0 0% (0–13) 25 5% (3–7) K pneumoniae underlying only 4 80% (30–99) 41 16% (12–21) 17 12% (7–18) 6 10% (4–20) 2 6% (1–22) 66 13% (10–17) K pneumoniae underlying and immediate COD but not comorbid 0 0% (0–54) 15 6% (3–10) 9 6% (3–12) 2 3% (1–12) 0 0% (0–13) 26 5% (4–8) K pneumoniae underlying and comorbid but not immediate COD 0 0% (0–54) 3 1% (0–4) 1 1% (0–4) 0 0% (0–7) 0 0% (0–13) 4 1% (0–2) K pneumoniae underlying and immediate COD and comorbid 0 0% (0–54) 3 1% (0–4) 1 1% (0–4) 0 0% (0–7) 0 0% (0–13) 4 1% (0–2) K pneumoniae in causal chain but not underlying COD 1 20% (1–70) 195 76% (70–81) 116 81% (73–86) 55 87% (76–94) 31 94% (78–99) 397 80% (76–83) K pneumoniae comorbid condition only 1 20% (1–70) 31 12% (8–17) 23 16% (11–23) 13 21% (12–33) 9 27% (14–46) 76 15% (12–19) K pneumoniae immediate COD only 0 0% (0–54) 84 33% (27–39) 55 38% (30–47) 31 49% (37–62) 16 48% (31–66) 186 37% (33–42) K pneumoniae immediate and comorbid COD 0 0% (0–54) 80 31% (26–37) 38 26% (20–35) 11 17% (9–30) 6 18% (8–36) 135 27% (23–31) K pneumoniae-associated clinical syndrome in deaths with K pneumoniae in causal chain Sepsis 3 60% (17–93) 125 49% (42–55) 54 38% (30–46) 29 46% (34–59) 13 39% (23–58) 221 44% (40–49) Pneumonia 0 0% (0–54) 19 7% (5–11) 32 22% (16–30) 17 27% (17–40) 12 36% (21–55) 80 16% (13–20) Sepsis and pneumonia 0 0% (0–54) 42 16% (12–22) 36 25% (18–33) 11 17% (9–30) 5 15% (6–33) 94 19% (16–23) Meningitis 1 20% (1–70) 3 1% (0–4) 5 3% (1–8) 0 0% (0–7) 0 0% (0–13) 8 2% (1–3) Sepsis and meningitis 0 0% (0–54) 21 8% (5–12) 6 4% (2–9) 1 2% (0–10) 0 0% (0–13) 28 6% (4–8) Pneumonia and meningitis 0 0% (0–54) 4 2% (0–4) 1 1% (0–4) 0 0% (0–7) 0 0% (0–13) 5 1% (0–2) Sepsis, pneumonia, and meningitis 0 0% (0–54) 25 10% (7–14) 6 4% (2–9) 3 5% (1–14) 2 6% (1–22) 36 7% (5–10) Sepsis and other clinical syndrome‡ 0 0% (0–54) 9 4% (2–7) 1 1% (0–4) 0 0% (0–7) 0 0% (0–13) 10 2% (1–4) Sepsis, meningitis, and other clinical syndrome 0 0% (0–54) 2 1% (0–3) 0 0% (0–3) 0 0% (0–7) 0 0% (0–13) 2 0% (0–2) Other clinical syndrome 1 20% (1–70) 6 2% (1–5) 3 2% (1–6) 1 2% (0–10) 1 3% (0–18) 11 2% (1–4) Laboratory evidence of K pneumoniae among deaths with K pneumoniae in causal chain Culture§, TAC, and IHC 0 0% (0–54) 46 18% (14–23) 24 17% (11–24) 10 16% (8–28) 7 21% (10–39) 87 18% (14–21) Culture and TAC 2 40% (7–83) 135 53% (46–59) 69 48% (40–56) 30 48% (35–60) 12 36% (21–55) 248 50% (45–54) TAC and IHC 0 0% (0–54) 15 6% (3–10) 11 8% (4–14) 11 17% (9–30) 4 12% (4–29) 41 8% (6–11) Culture only 2 40% (7–83) 13 5% (3–9) 5 3% (1–8) 0 0% (0–7) 1 3% (0–18) 21 4% (3–6) TAC only 1 20% (1–70) 45 18% (13–23) 33 23% (17–31) 12 19% (11–31) 8 24% (12–43) 99 20% (17–24) IHC only 0 0% (0–54) 1 0% (0–2) 0 0% (0–3) 0 0% (0–7) 0 0% (0–13) 1 0% (0–1) No post-mortem laboratory evidence of K pneumoniae 0 0% (0–54) 2 1% (0–3) 2 1% (0–5) 0 0% (0–7) 1 3% (0–18) 5 1% (0–2) MITS=minimally invasive tissue sampling. COD=cause of death. TAC=Taqman array card, performed on blood, cerebrospinal fluid, and lung tissue. IHC=immunohistochemistry, performed on lung, liver, heart, and brain; not performed in all cases. *Excludes stillbirths. †“K pneumoniae only cause” refers to deathswith K pneumoniae as the underlying CODwith noother cause noted, while “K pneumoniae underlying only” refers to deaths with K pneumoniae as the underlying COD (but not immediate COD or comorbid condition), with or without other causes in the causal chain; these two categories are not mutually exclusive. ‡Congenital infection, other respiratory disease, and other infections. §Culture performed on blood and cerebrospinal fluid. Table 1: Frequency of deaths with Klebsiella pneumoniae in the causal chain, associated clinical syndromes, and laboratory evidence of K pneumoniae, by age group See Online for appendix Articles e134 Role of the funding source The study sponsorwas involved in discussions about design of the CHAMPS network; the sponsor had no role in col- lection, analysis or interpretation of data, writing of this manuscript, or the decision to publish. Results From Dec 9, 2016, to Dec 31, 2021, 3030 stillbirths were reported; consent was provided for 1528 (50%), MITS was performed on 1489 (97%) of 1528, and DeCoDe was com- pleted for 1297 (87%) of 1489 (appendix p 2). K pneumoniae was found to be in the causal chain for five (<1%) stillbirths (table 1), including four (2%, 95% CI 1–5) in South Africa andone (1%,0–4) inSierra Leone.Among6320 child deaths reported, consent was provided for 2797 (44%), MITS was performed on 2763 (99%) of 2797, and DeCoDe was com- pleted for 2352 (85%) of 2763 (appendix p 3). K pneumoniae was in the causal chain for 497 (21%, 95% CI 20–23) of 2352 child deaths, including 197 (19%, 17–22) of 1030 among females and 300 (23%, 21–25) of 1319 amongmales; for three deaths, the sex was indeterminate, none of which had K pneumoniae in the causal chain. Among child deaths www.thelancet.com/microbe Vol 5 February 2024 www.thelancet.com/microbe South Africa Kenya Mozambique Sierra Leone Bangladesh Mali Ethiopia n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI) Deaths with MITS with COD determination Overall 703 ⋅⋅ 432 ⋅⋅ 428 ⋅⋅ 343 ⋅⋅ 184 ⋅⋅ 126 ⋅⋅ 136 ⋅⋅ Deaths with K pneumoniae in causal chain of death Overall 181 26% (23–29) 51 12% (915) 55 13% (10–16) 107 31% (26–36) 11 6% (3–11) 21 17% (11–25) 71 52% (44–61) K pneumoniae underlying COD* 14 8% (4–13) 18 35% (23–50) 20 36% (24–50) 34 32% (23–42) 4 36% (12–68) 2 10% (2–32) 8 11% (5–22) K pneumoniae only cause 2 1% (0–4) 5 10% (4–22) 11 20% (11–33) 4 4% (1–10) 1 9% (0–43) 2 10% (2–32) 0 0% (0–6) K pneumoniae underlying only 11 6% (3–11) 13 25% (15–40) 15 27% (17–41) 20 19% (12–28) 3 27% (7–61) 2 10% (2–32) 2 3% (0–11) K pneumoniae underlying and immediate COD but not comorbid 1 1% (0–4) 5 10% (4–22) 5 9% (3–21) 13 12% (7–20) 1 9% (0–43) 0 0% (0–19) 1 1% (0–9) K pneumoniae underlying and comorbid but not immediate COD 1 1% (0–4) 0 0% (0–9) 0 0% (0–8) 1 1% (0–6) 0 0% (0–32) 0 0% (0–19) 2 3% (0–11) K pneumoniae underlying and immediate COD and comorbid 1 1% (0–4) 0 0% (0–9) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 0 0% (0–19) 3 4% (1–13) K pneumoniae in causal chain but not underlying COD 167 92% (87–96) 33 65% (50–77) 35 64% (50–76) 73 68% (58–77) 7 64% (32–88) 19 90% (68–98) 63 89% (78–95) K pneumoniae comorbid condition only 37 20% (15–27) 6 12% (5–25) 2 4% (1–14) 17 16% (10–25) 0 0% (0–32) 5 24% (9–48) 9 13% (6–23) K pneumoniae immediate COD only 64 35% (29–43) 21 41% (28–56) 28 51% (37–64) 46 43% (34–53) 5 45% (18–75) 5 24% (9–48) 17 24% (15–36) K pneumoniae immediate and comorbid COD 66 36% (30–44) 6 12% (5–25) 5 9% (3–21) 10 9% (5–17) 2 18% (3–52) 9 43% (23–66) 37 52% (40–64) K pneumoniae-associated clinical syndrome in deaths with K pneumoniae in causal chain Sepsis 66 36% (30–44) 27 53% (39–67) 27 49% (36–63) 64 60% (50–69) 8 73% (39–93) 9 43% (23–66) 20 28% (18–40) Pneumonia 32 18% (13–24) 16 31% (20–46) 14 25% (15–39) 12 11% (6–19) 0 0% (0–32) 1 5% (0–26) 5 7% (3–16) Sepsis and pneumonia 44 24% (18–31) 6 12% (5–25) 8 15% (7–27) 15 14% (8–22) 0 0% (0–32) 9 43% (23–66) 12 17% (9–28) Meningitis 4 2% (1–6) 1 2% (0–12) 0 0% (0–8) 3 3% (1–9) 0 0% (0–32) 0 0% (0–19) 0 0% (0–6) Sepsis and meningitis 7 4% (2–8) 0 0% (0–9) 0 0% (0–8) 8 7% (4–15) 2 18% (3–52) 0 0% (0–19) 11 15% (8–26) Pneumonia and meningitis 4 2% (1–6) 1 2% (0–12) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 0 0% (0–19) 0 0% (0–6) Sepsis, pneumonia, and meningitis 20 11% (7–17) 0 0% (0–9) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 2 10% (2–32) 14 20% (12–31) Sepsis and other clinical syndrome† 1 1% (0–4) 0 0% (0–9) 1 2% (0–11) 1 1% (0–6) 1 9% (0–43) 0 0% (0–19) 6 8% (3–18) Sepsis, meningitis, and other clinical syndrome 0 0% (0–3) 0 0% (0–9) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 0 0% (0–19) 2 3% (0–11) Other clinical syndrome 3 2% (0–5) 0 0% (0–9) 4 7% (2–18) 3 3% (1–9) 0 0% (0–32) 0 0% (0–19) 1 1% (0–9) Laboratory evidence of K pneumoniae among deaths with K pneumoniae in causal chain Culture‡, TAC, and IHC 30 17% (12–23) 10 20% (10–34) 14 25% (15–39) 17 16% (10–25) 3 27% (7–61) 5 24% (9–48) 8 11% (5–22) Culture and TAC 87 48% (41–56) 14 27% (16–42) 26 47% (34–61) 60 56% (46–66) 5 45% (18–75) 12 57% (34–77) 42 59% (47–70) TAC and IHC 9 5% (2–10) 13 25% (15–40) 5 9% (3–21) 7 7% (3–13) 1 9% (0–43) 3 14% (4–37) 3 4% (1–13) Culture only 12 7% (4–12) 0 0% (0–9) 0 0% (0–8) 3 3% (1–9) 1 9% (0–43) 0 0% (0–19) 3 4% (1–13) TAC only 38 21% (15–28) 13 25% (15–40) 10 18% (10–31) 20 19% (12–28) 1 9% (0–43) 1 5% (0–26) 15 21% (13–33) IHC only 0 0% (0–3) 1 2% (0–12) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 0 0% (0–19) 0 0% (0–6) No post-mortem laboratory evidence of K pneumoniae 5 3% (1–7) 0 0% (0–9) 0 0% (0–8) 0 0% (0–4) 0 0% (0–32) 0 0% (0–19) 0 0% (0–6) MITS=minimally invasive tissue sampling. COD=cause of death. TAC=Taqman array card, performed on blood, cerebrospinal fluid, and lung tissue. IHC=immunohistochemistry, performed on lung, liver, heart, and brain; not performed in all cases. *"K pneumoniae only cause" refers to deaths with K pneumoniae as the underlying CODwith no other cause noted, while "K pneumoniae underlying only" refers to deaths with K pneumoniae as the underlying COD (but not immediate COD or comorbid condition), with or without other causes in the causal chain; these two categories are not mutually exclusive. †Congenital infection, other respiratory disease, and other infections. ‡Culture performed on blood and cerebrospinal fluid. Table 2: Frequency of child deaths with Klebsiella pneumoniae in the causal chain, associated clinical syndromes, and laboratory evidence of K pneumoniae, by site Articles withKpneumoniae in the causal chain,Kpneumoniaewas an underlying cause of 100 (20%, 95% CI 17–24), including 25 (5%, 3–7) for which K pneumoniae was the only cause listed (ie, K pneumoniae both immediate and underlying cause). K pneumoniae was considered part of the causal chain but with another cause as underlying in 397 (80%, 95% CI 76–83) child deaths, including as a comorbid con- dition in 76 (15%, 12–19), an immediate cause of death in 186 (37%, 33–42), and both a comorbid condition and an immediate cause of death in 135 (27%, 23–31). Themedian time from death toMITS for all deaths was 12 h (IQR 4–21) www.thelancet.com/microbe Vol 5 February 2024 and for deaths with K pneumoniae in the causal chain was 13 h (4–22). Among all child deaths, K pneumoniae was identified in the causal chainmore commonly among deaths in children aged 1–11 months (30%, 95% CI 26–34; 144 of 485 deaths) and 12–23 months (28%, 22–34, 63 of 225 deaths) than among deaths in other age groups. By site, the proportion of deaths withK pneumoniae in the causal chain ranged from a lowof 6%(95%CI3–11; 11of 184deaths) inBangladesh to a high of 52% (44–61; 71 of 136 deaths) in Ethiopia (table 2). For neonatal deaths, K pneumoniae was in the causal chain e135 www.thelancet.com/microbe 100 75 50 25 0 Proportion of deaths w ith K pneum oniae in causal chain (% ) 600 400 200 0 2000 Age group 18% 30% 14% 22% 20% 45% 22% 24% 15% 27%28% 15% All sitesA B South Africa Location of death Age group Location of death 1500 1000 De at hs 500 0 Neonate (n=1424) 1–11 m onths (n=485) 12–23 m onths (n=225) 24–59 m onths (n=218) Community (n=329) Health fa cili ty (n=2023) Neonate (n=463) 1–11 m onths (n=161) 12–23 m onths (n=37) 24–59 m onths (n=42) Community (n=48) Health fa cili ty (n=655) 100 75 50 25 0 Proportion of deaths w ith K pneum oniae in causal chain (% ) 9% 17% 15% 4% 14% 11% 10% 16% 30% 6% 10% 13% C D Age group Kenya Mozambique Location of death Age group Location of death 200 300 100 0 De at hs 200 400 300 100 0 Neonate (n=182) 1–11 m onths (n=140) 12–23 m onths (n=59) 24–59 m onths (n=51) Community (n=139) Health fa cili ty (n=293) Neonate (n=271) 1–11 m onths (n=64) 12–23 m onths (n=46) 24–59 m onths (n=47) Community (n=41) Health fa cili ty (n=387) 100 75 50 25 0 Proportion of deaths w ith K pneum oniae in causal chain (% )27% 39% 42% 19% 18% 33% 6% 0% 0% 0% 6% E F Age group Sierra Leone Bangladesh Location of death Age group Location of death 200 300 100 0 De at hs 150 100 50 0 Neonate (n=144) 1–11 m onths (n=76) 12–23 m onths (n=65) 24–59 m onths (n=58) Community (n=33) Health fa cili ty (n=310) Neonate (n=180) 1–11 m onths (n=3) 12–23 m onths (n=1) 24–59 m onths (n=0) Community (n=9) Health fa cili ty (n=175) (Figure 1 continues on next page) Articles e136 www.thelancet.com/microbe Vol 5 February 2024 www.thelancet.com/microbe 100 75 50 25 0 Proportion of deaths w ith K pneum oniae in causal chain (% ) 20% 10% 21% 13% 18% 53% 46% 43% 67% 30% 56% 7% G H Age group Mali Ethiopia Location of death Age group Location of death 90 60 30 0 40 80 60 20 0 De at hs Neonate (n=74) 1–11 m onths (n=28) 12–23 m onths (n=10) 24–59 m onths (n=14) Community (n=39) Health fa cili ty (n=87) Neonate (n=110) 1–11 m onths (n=13) 12–23 m onths (n=7) 24–59 m onths (n=6) Community (n=20) Health fa cili ty (n=116) Figure 1: Deaths with Klebsiella pneumoniae in the causal chain by age group, location of death, and site Theheight of each bar reflects total deaths in that category included in the analysis; the darker portion of thebar represents the subset of those deathswith K pneumoniae in the causal chain. Articles for 18% (95% CI 16–20; 257 of 1424 deaths) overall; how- ever, the proportion ranged across sites, from 9% (95% CI 5–14; 16 of 182 deaths) in Kenya to 53% (43–62; 58 of 110 deaths) in Ethiopia (figure 1). K pneumoniae was in the causal chain for 450 (22%, 95% CI 20–24) of 2023 deaths that occurred in health facilities and 47 (14%, 11–19) of 329 that occurred in the community. In Kenya, the site with the largest number of community deaths undergoing MITS, the proportion of deaths with K pneumoniae in the causal chain was higher in community (14%, 95% CI 9–21; 19 of 139 deaths) than among health facility deaths (11%, 8–15; 32 of 293 deaths). Among 47 deaths that occurred in the community with K pneumoniae in the causal chain across all sites, 30 had information on care-seeking before death and 19 (63%, 44–79) of those had been admitted to hospital during the course of illness; for the remaining 11 (37%, 21–56), no substantial exposure to health-care facilities was noted. Among the 450 health facility deaths with K pneumoniae in the causal chain fromall sites, themedian duration between hospitalisation and death was 5 days (IQR 2–11); 262 (58%, 95% CI 53–63) died more than 48 h after arrival to the facility. Results of a multivariable model comparing deaths with andwithoutK pneumoniae in the causal chainwith regard to age, site, and location of death were consistent with descriptive analyses; age 1–11 months, age 12–23 months, Ethiopia site, anddeathoccurring inhospitalmore than48h after admission all had adjusted odds ratios greater than 1 (appendix p 7). The most common clinical syndromes among the 497 deaths with K pneumoniae in the causal chain were sepsis (44%, 95% CI 40–49), sepsis in conjunction with pneumonia (19%, 16–23), and pneumonia (16%, 13–20); 2% (1–3) had meningitis without pneumonia or sepsis www.thelancet.com/microbe Vol 5 February 2024 (table 1). The predominance of sepsis and pneumonia was consistent across sites (table 2). Among children with K pneumoniae sepsis episodes contributing to deaths, 195 (50%, 95% CI 45–55) of 391 had multipathogen sepsis, and among K pneumoniae pneumonia episodes contribut- ing to deaths, 143 (67%, 60–73) of 215 were multipathogen pneumonia. Other pathogens most commonly implicated together with K pneumoniae in multipathogen sepsis or pneumonia were Acinetobacter baumannii (n=87; 32%, 95% CI 26–38), E coli (n=57; 21%, 16–26), Streptococcus pneumoniae (n=45; 16%, 12–22), and Pseudomonas aerugi- nosa (n=30; 11%, 8–15; appendix p 4). The distribution of other pathogens involved in multipathogen sepsis or pneumonia varied by location and timing of death; A baumannii was more common in deaths that occurred more than 48 h after admission (50%, 95% CI 42–58; 76 of 153 deaths), whereas S pneumoniae and E coli were more common among deaths that occurred in the community (45%, 30–64, 13 of 29 deaths; and 31%, 16–51, nine of 29 deaths, respectively) and within 48 h of admission to hospital (29%, 20–39, 26 of 91 deaths; and 26%, 18–37, 24 of 91 deaths, respectively). Meningitis was multipathogen in 27 (34%, 95% CI 24–46) of 79 cases. Among deaths in which K pneumoniae was either an immediate cause of death or a comorbid condition, themost common underlying causes varied by age group (figure 2A; appendix p 5). Among neonates with K pneumoniae as an immediate or comorbid condition, the most common underlying causes were preterm birth complications (56%, 95% CI 49–63), perinatal asphyxia or hypoxia (21%, 15–27), and congenital birth defects (12%, 8–18). Among infants aged 1–11 months, malnutrition (28%, 95% CI 20–37), preterm birth complications (21%, 14–29), congenital birth defects (17%, 11–26), and HIV (11%, 6–19) were the most frequent underlying causes. Among children aged e137 www.thelancet.com/microbe Neonate (n=195) 1–11 m onths (n=116) 12–59 m onths (n=86) Bangladesh (n=7) Ethiopia (n=63) Kenya (n=33) Mali (n=19) Mozambique (n=35) Sierra Leone (n=73) South Afric a (n=167) 100 75 Pr op or tio n of d ea th s ( % ) Age groupA B Site 50 25 0 Birth trauma Cancer Other infections Other respiratory disease Other skin and subcutaneous diseases Liver disease Measles Other Neonatal encephalopathy Syphilis Other neonatal disorders Other neurological disorders Sepsis Lower respiratory infections Neonatal aspiration syndromes Malaria Congenital infection Injury Diarrhoeal diseases HIV Perinatal asphyxia or hypoxia Congenital birth defects Malnutrition Neonatal preterm birth complications Figure 2: Underlying causes of death among deaths with Klebsiella pneumoniae as immediate cause of death or comorbid condition among all sites, by age group (A) and site (B) Articles e138 12–59 months, malnutrition (35%, 95% CI 25–46) was the most common underlying cause, followed by injury (16%, 10–26) and HIV (16%, 10–26). Variations in under- lying causes of death across sites (figure 2B; appendix p 6) reflected differences in age distribution. Among the 2352 child deaths that underwent MITS and had DeCoDe results during the study period, K pneumoniae was detected by TAC in at least one specimen of 1297 (55%, 95% CI 53–57); 36% (34–39) of those with K pneumoniae detected by TAC were found to have K pneumoniae in the causal pathway, thus 64% (61–66) of these detectionswere judged to be incidental andnot related to death. Among 349 deaths in which K pneumoniae was detected solely through TAC testing of nasopharyngeal swabs, 22 (6%, 95%CI4–10)hadKpneumoniae in the causal chain, 21 of which were deaths that occurred in health facilities. Among 948 deaths in which K pneumoniae was detected by TAC on either blood, lung, or CSF, 450 (47%, 95%CI 44–51) hadK pneumoniae in the causal chain. Among all deaths with K pneumoniae in the causal pathway, 248 (50%, 95% CI 45–54) had evidence of K pneumoniae by culture (blood or CSF, or both) and by TAC (on blood, CSF, or lung tissue, or a combination of these samples), and 87 (18%, 14–21) had evidence of K pneumoniae by culture (blood or CSF, or both), by TAC (on blood, CSF, lung tissue, or a combination), and by IHC (lung, liver, heart, or brain, or a combination); 99 (20%, 17–24) had K pneumoniae detected by TAC only (table 1). Five deaths had no post- mortem laboratory evidence of K pneumoniae but had K pneumoniae isolated from antemortem blood cultures. Of 497 child deaths withK pneumoniae in the causal chain across all sites, 157 (32%, 95% CI 28–36) had K pneumoniae isolates that underwent antimicrobial susceptibility testing for relevant antibiotics. The number of isolates tested for each antibiotic ranged from 42 (chloramphenicol) to 144 (ceftriaxone; figure 3A). Resistance to ceftriaxone andgentamicinwas observed in121 (84%, 95%CI77–89)of 144 and 80 (75%, 66–83) of 106 isolates tested, respectively. The frequency of resistance to imipenem and meropenem was 30% (95%CI 20–43; 19 of 63 isolates) and 21% (14–31; 21 of 100 isolates), respectively. Resistance to most anti- biotics tested was more common among deaths that occurred more than 48 h after admission (figure 3B). Discussion This analysis of comprehensive child mortality data from seven countries highlights the importance of K pneumoniae as a causeof death among childrenunder 5 years in resource- poor settings.Kpneumoniaewas implicated inmore thanone in four deaths among children aged 1–23 months. While the proportion of neonatal deaths with K pneumoniae in the causal chain was lower (18%) than in older age groups, the high mortality rate among neonates, especially in sub- Saharan Africa and Asia20 yields a substantial global burden of neonatal deaths in which K pneumoniae appears to play a causal role. Our results also illustrate the complexity of factors contributing to child deaths;21 K pneumoniae was frequently identified in conjunction with other pathogens and in children with prematurity, birth complications, HIV, and malnutrition. If Sustainable Development Goals for under-5 and neonatal mortality are to be met by 2030,22 specific efforts aimed at reducing K pneumoniae-associated deaths, including multifaceted interventions addressing multiple interacting conditions contributing tomortality will likely be necessary.21 We found K pneumoniae to play an important role in deaths occurring both in the community and in health facilities. The proportion of community deaths and health facility deaths with K pneumoniae in the causal chain varied across sites; however, interpretation is limited by small numbers of community deaths at several sites. Of note, the location of death is not necessarily reflective of where the exposure to K pneumoniae occurred. For many of the in-hospital deaths with K pneumoniae in the causal chain, www.thelancet.com/microbe Vol 5 February 2024 www.thelancet.com/microbe An tim icr ob ia l Amikacin Amoxicillin–clavulanate Cefepime Cefotaxime Ceftazidime Ceftriaxone Cefuroxime Chloramphenicol Ciprofloxacin Co-trimoxazole Gentamicin Imipenem Meropenem Piperacillin–tazobactam n=44 n=127 n=101 n=85 n=71 n=144 n=75 n=42 n=82 n=62 n=106 n=63 n=100 n=52 SusceptibleIntermediateResistant A Amikacin ≤48 h or in the communityB Amoxicillin–clavulanate Cefepime Cefotaxime Ceftazidime Ceftriaxone Cefuroxime Chloramphenicol Ciprofloxacin Co-trimoxazole Gentamicin Imipenem Meropenem Piperacillin–tazobactam An tim icr ob ia l n=46 n=16 n=30 n=56 n=15 n=25 n=21 n=17 n=36 n=37 n=24 n=18 n=28 n=36 0 25 50 75 100 Proportion of isolates (%) >48 h Amikacin Amoxicillin–clavulanate Cefepime Cefotaxime Ceftazidime Ceftriaxone Cefuroxime Chloramphenicol Ciprofloxacin Co-trimoxazole Gentamicin Imipenem Meropenem Piperacillin–tazobactam An tim icr ob ia l n=60 n=55 n=71 n=88 n=29 n=50 n=64 n=35 n=46 n=90 n=18 n=45 n=72 n=26 Figure 3: Antimicrobial susceptibility of Klebsiella pneumoniae isolates from deaths with K pneumoniae in the causal chain Overall (A) and stratified bydeaths occurring in the community orwithin 48h of admission and those occurringmore than 48 h after admission (B). Articles the duration of admission was long enough (>48 h) for K pneumoniae to have been acquired during hospitalisation. The epidemiology of K pneumoniae as a nosocomial patho- gen is well characterised, and nosocomial outbreaks of K pneumoniae in neonatal intensive care contribute importantly to hospital-acquired K pneumoniae infections in children.1,2 During the study period, documented K pneumoniae outbreaks occurred in the neonatal intensive care wards at CHAMPS enrolment sites in South Africa23 and Sierra Leone. However, 42% of overall in-hospital deaths with K pneumoniae in the causal chain occurred within 48 h of admission, suggesting that K pneumoniae might have been acquired before admission. We also observed deaths with K pneumoniae in the causal chain that occurred in the community among children with limited or no contact with health-care facilities during the illness that led to their death. Data on community-acquired K pneumo- niae infections in children are sparse.24,25 Reducing deaths due to K pneumoniae will require a better understanding of sources of infection in the community as well as risk factors for developing K pneumoniae disease given colonisation.26 Distinguishing the exact role of K pneumoniae in a child’s death based on post-mortem specimens is difficult because K pneumoniae commonly colonises the gastrointestinal tract and nasopharynx,5 so detection of the pathogen postmortem could potentially reflect agonal spread, post-mortem translocation, post-mortem overgrowth, or contamination. Detection of K pneumoniae was common among deaths in CHAMPS during the study period, with over half of all deaths havingK pneumoniae detected by TAC on at least one specimen (including nasopharyngeal swabs); because K pneumoniae was not tested for in the stool, this likely is an underestimate of the frequency ofK pneumoniae colonisation among child deaths. However, the compre- hensive testing in CHAMPS, which includesmicrobiology, molecular assays, histopathology, and IHC linked to clinical presentation, provides important insight into the patho- genic role of K pneumoniae. Histopathology and IHC aid in distinguishing between post-mortem contamination and antemortem infection because the pathogen can be observed in affected tissues. Also, CHAMPS MITS collec- tion methods reduce the risk of contamination and post- mortem translocation by using sterile procedures and minimising the time between death and sample collec- tion.27,28 Among cases deemed to have K pneumoniae in the causal pathway, 68% had K pneumoniae detected by both culture and TAC from normally sterile specimens, and many had additional supporting evidence from IHC. Nonetheless, it is possible that some K pneumoniae identi- fied post mortem could reflect invasion at the time of dying rather than the true cause of death. Surveillance for severe disease with extensive laboratory testing conducted at the CHAMPS sites could provide complementary data to help better characterise the role of K pneumoniae in child illness and deathby examining theburden ofKpneumoniaedisease among living children and the role ofK pneumoniae in child death in the same settings. www.thelancet.com/microbe Vol 5 February 2024 Among the K pneumoniae isolates, resistance was very common to antibiotics recommended by WHO for the treatment of paediatric pneumonia and sepsis, including gentamicin (first-line recommendation together with ampicillin, which is not clinically active against Klebsiella spp) and ceftriaxone (second-line recommendation).29 Fur- thermore, among isolates tested for carbapenem resistance (imipenem and meropenem), 21–30% were resistant, and carbapenem-resistant K pneumoniae infections are e139 www.thelancet.com/microbe Articles e140 associated with increased mortality, based on data from primarily high-income settings.11,12 Risk factors for carbapenem-resistant K pneumoniae infections have been described for predominantly adult populations,30 but data for children are limited. The findings of our study highlight the need for clinical predictors of K pneumoniae infections and rapid diagnostics in order to identify childrenwho need antibiotics that would cover K pneumoniae, as well as mon- itoring of K pneumoniae resistance patterns. Furthermore, guidelines for empirical treatment of pneumonia and sepsis might require adaptation to optimally prevent deaths due to K pneumoniae. Infection prevention and control practices and antimicrobial stewardship efforts are also important for curbing the spread of resistant K pneumoniae strains. The non-standardisation of antimicrobial susceptibility testing limits the interpretation and generalisability of the antimicrobial resistance findings. We did not perform genomic characterisation, so it is unknown whether any of thedeathswithKpneumoniae in the causal pathway involved the emerging hypervirulent pathotype;8 however, this ana- lysis is planned. Small numbers from several sites limited our ability to compare and contrast across sites. As noted above, it is possible that for some deaths found to have K pneumoniae in the causal chain, K pneumoniae infection represented agonal spread or contamination. Cases with less robust evidence of K pneumoniae involvement might have been misclassified as having K pneumoniae in the causal chain, potentially leading to an overestimation of the role of K pneumoniae in child deaths. Despite efforts to comprehensively capture all eligible deaths in theCHAMPS catchment areas, the sample of deaths included in this analysis might not be entirely representative of all deaths occurring in the surveillance area. We present evidence that K pneumoniae substantially contributed to community and facility deaths in the first 2 years of life across multiple high-mortality settings. Fur- ther research is needed to better understand the source of infection and risk factors forK pneumoniaemortality. Given the burden of K pneumoniae and the frequent resistance to first-line antibiotics used for pneumonia, sepsis, and men- ingitis, improved strategies are needed to rapidly identify childrenwhomightbe infectedwithKpneumoniae andneed more appropriate antibiotic treatment. Our data also sug- gest a potential impact of developing and using effective K pneumoniae vaccines in reducing neonatal, infant, and child deaths globally.31 Contributors JRV, DMB, ESG, CGW, IM, RFB, QB, VA, NA, SEA, KLK, SMad, IUO, DO, JAGS, and SS conceptualised the study. DMB and ZJM were responsible for data curation, visualisation, and formal analysis and accessed and verified all data. CGW and RFB acquired funding. VA, NA, VB, QB, MB, RFB, DMB, JB, ACAC, SEA, RG, ESG, MH, AI, AMK, SK, KLK, DK, RM, ZJM, SMad, LM, SMah, IM, JM, FVM, IUO, JO, DO, JOO, JAGS, SS, MT, CBT, SV, and CGW contributed to investigation and project administration. JRV, DMB, and RFB contributed to methodology. CGW, IM, RFB, QB, VA, NA, SEA, ESG, KLK, SMad, IUO, DO, JAGS, SS, and DMB provided supervision. JRV wrote the original draft of the paper and VA, NA, VB, QB, MB, RFB, DMB, JB, ACAC, SEA, RG, ESG, MH, AI, AMK, SK, KLK, DK, RM, ZJM, SMad, LM, SMah, IM, JM, FVM, IUO, JO, DO, JOO, JAGS, SS, MT, CBT, SV, and CGW contributed to review and editing. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication. Declaration of interests SEA, KLK, and MT report receiving funding from Emory University. JAGS reports receiving funding from Emory University, the Bill & Melinda Gates Foundation, Wellcome Trust, UK Foreign Commonwealth & Development Office, European Union, and National Institute for Health Research. SMad reports receiving funds from the Bill & Melinda Gates Foundation, Pfizer, Minervax, GSK, South African Medical Research Council, Merck, PATH, and Center for the AIDS Programme of Research in South Africa. All other authors declare no competing interests. Data sharing CHAMPS data are available online at https://champshealth.org/. Acknowledgments CHAMPS would like to extend sincere appreciation to all the families who agreed to participate in this study during a time of grief and loss. We would also like to thank the many teams that support different components of CHAMPS, including demographic surveillance, social sciences, different laboratory components, data curation and informatics technology, clinical activities, data to action, and reporting to families of study results. This work was supported by the Bill & Melinda Gates Foundation (OPP1126780). The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the US Centers for Disease Control and Prevention. References 1 Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11: 589–603. 2 Gupta A. Hospital-acquired infections in the neonatal intensive care unit—Klebsiella pneumoniae. Semin Perinatol 2002; 26: 340–45. 3 Wen SCH, Ezure Y, Rolley L, et al. Gram-negative neonatal sepsis in low- and lower-middle-income countries and WHO empirical antibiotic recommendations: a systematic review and meta-analysis. PLoS Med 2021; 18: e1003787. 4 van Duin D, Paterson DL. Multidrug-resistant bacteria in the community: trends and lessons learned. Infect Dis Clin North Am 2016; 30: 377–90. 5 Martin RM, Bachman MA. Colonization, infection, and the accessory genome of Klebsiella pneumoniae. Front Cell Infect Microbiol 2018; 8: 4. 6 Cutland CL, Schrag SJ, Zell ER, et al. Maternal HIV infection and vertical transmission of pathogenic bacteria. Pediatrics 2012; 130: e581–90. 7 Ko WC, Paterson DL, Sagnimeni AJ, et al. Community-acquired Klebsiella pneumoniae bacteremia: global differences in clinical patterns. Emerg Infect Dis 2002; 8: 160–66. 8 Russo TA, Marr CM. Hypervirulent Klebsiella pneumoniae. Clin Microbiol Rev 2019; 32: e00001-19. 9 Munoz-Price LS, Poirel L, Bonomo RA, et al. Clinical epidemiology of the global expansion of Klebsiella pneumoniae carbapenemases. Lancet Infect Dis 2013; 13: 785–96. 10 Moradigaravand D, Martin V, Peacock SJ, Parkhill J. Evolution and epidemiology of multidrug-resistant Klebsiella pneumoniae in the United Kingdom and Ireland. MBio 2017; 8: e01976-16. 11 Kohler PP, Volling C, Green K, Uleryk EM, Shah PS, McGeer A. Carbapenem resistance, initial antibiotic therapy, and mortality in Klebsiella pneumoniae bacteremia: a systematic review and meta- analysis. Infect Control Hosp Epidemiol 2017; 38: 1319–28. 12 Xu L, Sun X, Ma X. Systematic review and meta-analysis of mortality of patients infected with carbapenem-resistant Klebsiella pneumoniae. Ann Clin Microbiol Antimicrob 2017; 16: 18. 13 Salzberg NT, Sivalogan K, Bassat Q, et al. Mortality surveillance methods to identify and characterize deaths in child health and mortality prevention surveillance network sites. Clin Infect Dis 2019; 69 (suppl 4): S262–73. www.thelancet.com/microbe Vol 5 February 2024 https://champshealth.org/ www.thelancet.com/microbe Articles 14 Taylor AW, Blau DM, Bassat Q, et al. Initial findings from a novel population-based child mortality surveillance approach: a descriptive study. Lancet Glob Health 2020; 8: e909–9. 15 Cunningham SA, Shaikh NI, Nhacolo A, et al. Health and demographic surveillance systems within the Child Health and Mortality Prevention Surveillance Network. Clin Infect Dis 2019; 69 (suppl 4): S274–79. 16 Clinical and Laboratory Standards Institute. M100: performance standards for antimicrobial susceptibility testing. 30th edn. Malvern, PA: Clinical and Laboratory Standards Institute, 2020. 17 Diaz MH, Waller JL, Theodore MJ, et al. Development and implementation of multiplex TaqMan array cards for specimen testing at Child Health and Mortality Prevention Surveillance site laboratories. Clin Infect Dis 2019; 69 (suppl 4): S311–21. 18 Blau DM, Caneer JP, Philipsborn RP, et al. Overview and development of the Child Health and Mortality Prevention Surveillance determination of cause of death (DeCoDe) process and DeCoDe diagnosis standards. Clin Infect Dis 2019; 69 (suppl 4): S333–41. 19 CHAMPS. Diagnosis standards (DS) for determination of cause of death (DeCoDe). Atlanta, GA: Child Health and Mortality Prevention Surveillance, 2019. 20 Hug L, Alexander M, You D, Alkema L. National, regional, and global levels and trends in neonatal mortality between 1990 and 2017, with scenario-based projections to 2030: a systematic analysis. Lancet Glob Health 2019; 7: e710–20. 21 Breiman RF, Blau DM, Mutevedzi P, et al. Postmortem investigations and identification of multiple causes of child deaths: an analysis of findings from the Child Health and Mortality Prevention Surveillance (CHAMPS) network. PLoS Med 2021; 18: e1003814. www.thelancet.com/microbe Vol 5 February 2024 22 Boerma T, Requejo J, Victora CG, et al. Countdown to 2030: tracking progress towards universal coverage for reproductive, maternal, newborn, and child health. Lancet 2018; 391: 1538–48. 23 Madhi SA, Pathirana J, Baillie V, et al. Unraveling specific causes of neonatal mortality using minimally invasive tissue sampling: an observational study. Clin Infect Dis 2019; 69 (suppl 4): S351–60. 24 Tsay RW, Siu LK, Fung CP, Chang FY. Characteristics of bacteremia between community-acquired and nosocomial Klebsiella pneumoniae infection: risk factor for mortality and the impact of capsular serotypes as a herald for community-acquired infection. Arch Intern Med 2002; 162: 1021–27. 25 Hamer DH, Darmstadt GL, Carlin JB, et al. Etiology of bacteremia in young infants in six countries. Pediatr Infect Dis J 2015; 34: e1–8. 26 Broberg CA, Palacios M, Miller VL. Klebsiella: a long way to go towards understanding this enigmatic jet-setter. F1000Prime Rep 2014; 6: 64. 27 Riedel S. The value of postmortem microbiology cultures. J Clin Microbiol 2014; 52: 1028–33. 28 Morris JA, Harrison LM, Partridge SM. Practical and theoretical aspects of postmortem bacteriology. Curr Diagn Pathol 2007; 13: 65–74. 29 WHO. Pocket book of hospital care for children: guidelines for the management of common childhood illnesses, 2nd edn. Geneva: World Health Organization, 2013. 30 Zhu W-m, Yuan Z, Zhou H-y. Risk factors for carbapenem-resistant Klebsiella pneumoniae infection relative to two types of control patients: a systematic review and meta-analysis. Antimicrob Resist Infect Control 2020; 9: 23. 31 Arato V, Raso MM, Gasperini G, Berlanda Scorza F, Micoli F. Prophylaxis and treatment against Klebsiella pneumoniae: current insights on this emerging anti-microbial resistant global threat. Int J Mol Sci 2021; 22: 4042. e141 www.thelancet.com/microbe Child deaths caused by Klebsiella pneumoniae in sub-Saharan Africa and south Asia: a secondary analysis of Child Health and ... Introduction Methods Study design and population Procedures Outcomes and statistical analysis Role of the funding source Results Discussion Contributors Declaration of interests Data sharing Acknowledgments References