Table 3.

Studies reporting mortality in patients with 3GC-R BSI

Study, publication year	Study type	Population	Country	Total patients in study	Pathogens	Case-fatality rate, 3GC-R 3GC-S n (%)	Adjusted mortality estimate from 3GC-R BSI (95% CI)	Author conclusions
Blomberg17	Prospective cohort	Paediatric; 0–7 years	Tanzania	1632	Mixture of Enterobacteriaceae	15/21 (71.0)	OR 12.87 (4.95–33.48)	Inappropriate antimicrobial therapy due to 3GC resistance predicts fatal outcome
						NR	Multivariable model adjusted for: age <1 month, sex, HIV status, malaria, other underlying disease, polymicrobial blood culture
2007
		Urban referral hospital
		Children with suspected systemic infection based on IMCI
Dramowski10	Retrospective cohort	Paediatric; 0–14 years	South Africa	864	Mixture of Enterobacteriaceae (mortality data available for Klebsiella spp.)	21/122 (17.2)	Not reported by AMR type	AMR not associated with BSI mortality
		Urban referral hospital
2015
						NR
		Children with suspected sepsis or severe focal infection
Onken19	Prospective cohort	All ages; no range reported	Zanzibar	469	Mixture of Enterobacteriaceae	3/5 (60.0)	Not reported	No significantly higher case-fatality rate in 3GC-R compared with susceptible infections, but small numbers
2015		Urban referral hospital				4/11 (36.0)
		Patients with fever (≥38.3°C in adults, ≥38.5°C in children) or hypothermia (<36.0°C), tachypnoea >20/min, tachycardia >90/min or suspected systemic bacterial infection
Seboxa18	Prospective cohort	Adults; 13–98 years	Ethiopia	232	Mixture of Enterobacteriaceae	11/11 (100)	RR 9.00 (1.42–57.12)	Inappropriate antimicrobial therapy due to 3GC-R infections predicts fatal outcome
2015		Urban referral hospital				1/9 (11.1)	No multivariable analysis
		Patients with clinical suspicion of septicaemia and 2 of the 3 following criteria: axillary temperature ≥38.5°C or ≤36.5°C, pulse ≥90 beats/min and frequency of respiration ≥20/min
Buys21	Retrospective cohort	Paediatric; IQR 2–16 months	South Africa	410	Klebsiella spp.	NR	OR 1.09 (0.55–2.16)	MDR K. pneumoniae BSI is associated with high mortality in children
		Urban referral hospital					Multivariable model adjusted for: age, gender, nutrition, HIV, ESBL, patient in PICU, patient needing to go to PICU, continuous IV infusion for >3 days before the BSI, Klebsiella BSI without source, chronic underlying medical condition excluding HIV, and skin erosions
2016
		Electronic list of Klebsiella bloodstream isolates from hospital database
Eibach20	Prospective cohort	All ages; IQR 1–18 years	Ghana	7172	Mixture of Enterobacteriaceae	NR	Whole cohort: OR 3.0 (1.2–7.3) Neonates: OR 0.6 (0.1–3.7) No multivariable regression reported	3GC-R BSI is associated with higher mortality than non-3GC-R, but this is highly dependent on age
2016		Rural primary healthcare centre Patients with fever ≥38°C or history of fever within 24 h after admission or neonates with suspected neonatal sepsis
								No mortality difference from 3GC-R infections in neonates and higher overall mortality
Ndir11	Case–control	Paediatric; 0–17 years	Senegal	173	Mixture of Enterobacteriaceae	NR (54.8)	OR 2.9 (1.8–7.3)	3GC-R BSI is associated with fatal outcome in HA-BSI
2016		Urban referral hospital				NR (15.4)	Multivariable model adjusted for: age <1 month, prematurity, underlying comorbidities, admission diagnoses, invasive procedures, inappropriate antibiotics
		Cases—patients with an HA-BSI caused by Enterobacteriaceae
		Controls—patients who did not experience an infection during the study period, randomly selected from the hospital database
Marando44 2018	Prospective cohort	Neonates; IQR 4–8 days	Tanzania	304	Mixture of Enterobacteriaceae	NR (34.4) NR	HR 2.4 (1.2–4.8), Cox regression	Neonates infected with 3GC-R BSI have significantly higher mortality than EBSL negative or non-bacteraemic patients
							OR 2.71 (1.22–6.03), multivariable model adjusted for age and sex

3GC-S, 3GC susceptible; IMCI, integrated management of childhood infection.