Introduction: Neonatal sepsis is associated with significant mortality and morbidity. Low-middle-income countries are disproportionately affected, but late-onset sepsis (LOS) still occurs in up to 20% of infants <28 weeks in high-income countries. Understanding site-specific data is vital to guide management. Methods: A retrospective cohort study was conducted at King Edward Memorial Hospital (KEMH), Perth. Infants admitted between January 2012 and June 2022 were included. Data were extracted from routine electronic databases. Incidence and aetiology of sepsis were determined and the association of sepsis with neonatal outcomes analysed. Results: During the study period, 23,395 newborns were admitted with a median gestation of 37 weeks and birth weight of 2,800 g. There were 370 sepsis episodes in 350 infants; 102 were early-onset sepsis (EOS) (1.6 per 1,000 live births), predominantly Streptococcus agalactiae (35, 34.3%) and Escherichia coli (27, 26.5%); 268 were LOS (0.9 per 1,000 inpatient days), predominantly coagulase-negative staphylococci (CONS) (156, 57.6%) and E. coli (30, 11.1%). The incidence of LOS declined from 2012 to 2022 (p = 0.002). Infants with EOS had increased brain injury (25.7% vs. 4.1%; p = 0.002) and mortality (18.8% vs. 1.6%; p < 0.001). Those with LOS had increased hospital stay (median 95 vs. 15 days; p < 0.001), mortality (15.3% vs. 1.6%; p = 0.018), necrotising enterocolitis (NEC) (7.4% vs. 0.5%; p < 0.001), and chronic lung disease (CLD) (58.1% vs. 5.9%; p = 0.005). Infants <28 weeks with sepsis were at increased risk of neurodevelopmental impairment compared to those without infection (43.2% vs. 30.9%, p = 0.027). Conclusions: While we observed a reduction in LOS incidence, sepsis remains associated with higher mortality, and in survivors with longer hospital stay and increased risk of brain injury, NEC, CLD, and neurodevelopmental impairment.

Neonatal sepsis is one of the main causes of neonatal mortality and accounts for between 400,000 and 900,000 deaths annually [1]. Low- and middle-income countries and specific geographical regions are disproportionately affected with a reported incidence of early-onset sepsis (EOS) and late-onset sepsis (LOS) of 1.1 and 4.9/1,000 live births, respectively, in South Africa compared with 0.3 and 1.1/1,000 in Switzerland [2‒4]. Recent estimates from Australasia report an incidence of EOS of 0.5–0.7/1,000 live births and LOS rates of up to 20% in babies born <28 weeks of gestation [5]. In addition to increased mortality, neonatal sepsis is also associated with significant morbidity in survivors, most notably longer hospital stays, substantial economic and emotional costs, and long-term neurodevelopmental impairment [6‒8].

Timely diagnosis and management of neonatal sepsis remains challenging. Presentation is often subtle, progression can be rapid, and accurate diagnosis is often complicated by low colony count bacteraemia, insufficient blood volume inoculation, relatively slow blood culture turnaround times, and absence of an ideal biomarker [9‒11]. Time to positivity (TTP) of blood cultures is an important consideration, with approximately 68% and 94% of blood cultures positive by 24 and 36 h, respectively, in EOS, compared with 54.2% and 85.1% in LOS [12]. TTP is shortest for pathogenic organisms (median TTP 21 h) and longest for contaminants (median TTP 35.4 h), coagulase-negative staphylococci (CONS) (median TTP 29 h) and fungal isolates (median TTP ∼50 h) [11‒13]. The default is empiric treatment with broad-spectrum antimicrobial agents while awaiting pathogen isolation and antimicrobial sensitivity testing to guide optimal treatment [11].

Early empiric antimicrobial treatment is justified in symptomatic, at-risk neonates to improve survival [11]. However, antimicrobial susceptibilities vary by geographical region and healthcare institution, and antibiotic choice and duration of treatment are impacted by clinician preference [11, 14]. An important consequence is the risk of antimicrobial resistance, a phenomenon increasing at an alarming rate globally and sparking action by the World Health Organisation (WHO) and other recognised agencies [15, 16]. Another is the potential for disruption of the neonatal microbiome, which has been associated with immunological and long-term health effects including altered foetal immune competency and immune priming, asthma, atopic dermatitis, diabetes, allergic diseases, obesity, cardiovascular disease, and neurological disorders [17, 18].

Balancing the benefits of early antibiotic treatment against the common antibiotic overexposure of newborns and the associated risks of antimicrobial resistance and emerging long-term effects of an altered neonatal microbiome is challenging. A minimum requirement to guide decision-making is to understand local microbial epidemiology of neonatal sepsis to determine appropriate empiric antibiotic choices and strengthen antimicrobial stewardship. This report describes the incidence and outcomes of neonatal sepsis in a large tertiary Australian NICU from 2012 to 2022.

Design

This single-centre retrospective cohort study was conducted at King Edward Memorial Hospital (KEMH) in Perth, the only tertiary perinatal centre in Western Australia (WA), which covers a vast geographical area of more than 2.5 million km2 and had a total population of 2.4 M–2.8 M during the study period. Approximately 3.3% of the WA population are Aboriginal or Torres Strait Islander, a relatively higher risk group perinatally, yet account for ∼13% of neonatal admissions. There are ∼6,000 deliveries and ∼2,300 neonatal admissions at KEMH annually, of which more than 10% require transfer in utero and ∼5% require retrieval from local, regional, and remote areas antenatally. All infants admitted to the Neonatal Directorate between January 1, 2012 and June 30, 2022, were included. Maternal, neonatal, and follow-up data were retrieved from a unit-specific electronic neonatal database using Microsoft Access (2010)®. Data were collected and defined according to the Australia and New Zealand Neonatal Network (ANZNN) guidelines [19]. The study was approved by the CAHS GEKO Triage Committee (000471).

Definitions

  • 1.

    Gestational age (GA) was assessed on maternal history and obstetric ultrasound at <20 weeks gestation or, in the absence of these, by clinical assessment as per the Dubowitz and Ballard score [20].

  • 2.

    Chronic lung disease (CLD) was defined as supplemental oxygen requirement at 36 weeks of corrected gestational age in infants born <32 weeks of gestation [19].

  • 3.

    Necrotising enterocolitis (NEC) was diagnosed at surgery or postmortem OR on clinical history plus radiological evidence (pneumatosis intestinalis, portal venous gas, or a persistent dilated loop on serial X-rays) OR clinical history plus abdominal wall cellulitis and palpable abdominal mass [19]. NEC was staged using modified Bell’s criteria [21]. NEC grade 2, 3 or NEC requiring surgery were considered significant.

  • 4.

    Neonatal sepsis was defined by positive blood or CSF culture with a bacterial or fungal pathogen AND antibiotic treatment ≥5 days [19]. Multiple episodes of sepsis were diagnosed if a subsequent blood and/or CSF culture was positive >7 days from a previous positive culture. Sepsis was further classified as:

    • -

      EOS: sepsis occurring at <48 h of life [19].

    • -

      LOS: sepsis occurring ≥48 h after birth [19].

Typically non-pathogenic organisms (contaminants) were pre-specified and included CONS, Streptococcus viridans, Streptococcus bovis, Micrococcus species, Corynebacterium, and Bacillus cereus. These were only considered pathogenic if the patient had CRP >20 mg/L within 48 h of the positive blood culture and was treated with antibiotics for ≥5 days [19]. CONS and other typical contaminants cultured from CSF were not considered pathogenic. Polymicrobial infections were counted as one episode of sepsis, but each organism was included separately in the microbiological description.

  • 5.

    Severe retinopathy of prematurity (ROP) was defined as [19]:

    • -

      Severe stage 3–5 ROP

    • -

      Aggressive ROP characterised by posterior location, prominent neovascularisation and plus disease, and rapid progression of neovascularisation

    • -

      ROP requiring treatment (laser or anti-VEGF therapy)

  • 6.

    Severe brain injury was defined as grade 3 or 4 intraventricular haemorrhage as defined by Papile or cystic periventricular leukomalacia detected on cranial ultrasound or MRI brain prior to discharge [22, 23].

  • 7.

    Neurodevelopmental impairment was defined as:

    • -

      Mild impairment: Bayley Scales of Infant and Toddler Development® (BSID) (3rd edition until January 2019 and 4th edition thereafter) [24] score −1 to −2 standard deviations (SD) below the mean, ambulant cerebral palsy (Gross Motor Function Classification System (GMCFS) I and II) [25], unilateral deafness, and/or any two domain fails on Ages and Stages questionnaire® (ASQ) [26].

    • -

      Moderate impairment: BSID score −2 to −3 SD below the mean, ambulant cerebral palsy with aids (GMCFS III), and/or bilateral deafness with aids

    • -

      Severe impairment: BSID score <−3SD below the mean, non-ambulant cerebral palsy (GMCFS IV and V), and/or vision <6/60 in the better eye.

Only those infants who met unit follow-up criteria underwent a neurodevelopmental assessment at 2 years of corrected age. At this non-surgical site, these criteria included gestational age <30 weeks and hypoxic-ischaemic encephalopathy requiring therapeutic hypothermia.

Statistical Analysis

Maternal and neonatal characteristics were compared between sepsis and no sepsis infants using independent t-tests or Mann-Whitney tests for parametric and non-parametric continuous data, respectively, and using χ2 or Fisher exact tests for categorical data. Durations of ventilation, PICC, and umbilical lines were summarised using Kaplan-Meier survival estimates and compared between sepsis groups using log-rank tests. Deaths were censored for survival analyses. The effect of sepsis was analysed in adjusted logistic regression models for short-term neonatal outcomes, and Cox proportional hazard models for length of stay. Potential confounders and covariates known to influence neonatal outcomes such as GA, birthweight (BW) z-scores, sex, inborn status, major congenital anomaly, and year of birth were adjusted in the modelling of short-term neonatal outcomes. The sample size for neurodevelopmental outcomes was considered too small for adjusted analyses. Infants with major congenital anomaly were excluded from neurodevelopmental assessment at 2 years of corrected age. Adjusted odds ratios (aORs) and hazard ratios (aHRs) were reported along with 95% confidence intervals (CIs). A p value <0.05 was considered significant. SPSS version 29 statistical software was used for data analysis (Armonk, NY, IBM Corp).

Patient Characteristics

There were 63,124 live births and 23,395 neonatal admissions at KEMH from January 1, 2012 to June 30, 2022. Median GA and BW were 37 weeks (IQR: 34–39 weeks) and 2,800 g (IQR: 2,010–3,450 g), respectively. A total of 16,610 blood cultures were taken during the study period of which 370 cultures from 350 infants represented episodes of neonatal sepsis. A flow diagram of infants with and without sepsis is presented in Figure 1. Neonates with sepsis differed from those with no sepsis in many pregnancy, delivery, and neonatal characteristics including increased chorioamnionitis, prematurity, exposure to TPN, ventilation, longer time to full feeds, and presence of a central venous catheter. Maternal and neonatal characteristics of the cohort are presented in Tables 1, 2, respectively.

Fig. 1.

Liveborn infants born from 2012 to 2022 with and without sepsis.

Fig. 1.

Liveborn infants born from 2012 to 2022 with and without sepsis.

Close modal
Table 1.

Maternal characteristics

All (n = 23,395)No sepsis (n = 23,045)Sepsisa (n = 350)p valueb
EOS (n = 102)LOS (n = 251)
Age, years (mean, SD) 30.4 (5.8) 30.4 (5.8) 29.7 (6.5) 29.1 (6.6) <0.001 
Ethnicity, n (%) 
 Caucasian 15,423 (65.9) 15,202 (66.0) 64 (62.7) 158 (62.9) 0.087 
 ATSI 2,196 (9.4) 2,148 (9.3) 11 (10.8) 37 (14.7)  
 Asian 3,783 (16.2) 3,732 (16.2) 19 (18.6) 33 (13.1)  
 Other 1,993 (8.5) 1,963 (8.5) 8 (7.8) 23 (9.2)  
SEIFA (<2 quintiles), n (%) 7,531 (32.2) 7,416 (32.2) 35 (34.3) 82 (32.7) 0.848 
Amnionitis, n (%) 794 (3.4) 736 (3.2) 26 (25.5) 33 (13.1) <0.001 
Preeclampsia, n (%) 2,397 (10.2) 2,358 (10.2) 1 (1.0) 44 (14.5) 0.577 
Gestational diabetes, n (%) 3,719 (15.9) 3,693 (16.0) 11 (10.8) 17 (5.6) <0.001 
Multiple birth, n (%) 3,231 (13.8) 3,176 (13.8) 7 (6.9) 48 (19.1) 0.298 
Antenatal antibiotics, n (%) 5,850 (25.0) 5,700 (24.7) 46 (45.1) 105 (41.8) <0.001 
Antenatal steroids, n (%) 
 None/<24 hc 1,822/5,533 (32.9) 1,723/5,243 (32.9) 19 (37.3) 81 (33.6) 0.653 
 At least one dose >24 hc 3,711/5,533 (67.1) 3,520/5,243 (67.1) 32 (62.7) 160 (66.4)  
 Not stated (n = 13)      
Inborn, n (%) 22,079 (94.4) 21,769 (94.5) 84 (82.4) 229 (91.2) <0.001 
ROM, h (median, IQR) 4 (0–12) 4 (0–12) 18 (1–49) 0 (0–18) 0.722 
Mode, n (%) 
 Vaginal 11,510 (49.2) 11,356 (49.2) 51 (50) 105 (39.8) <0.001 
 Caesarean section 11,865 (50.7) 11,670 (50.6) 51 (50) 145 (57.8)  
 Not stated 20 (0.1) 19 (0.1) 0 (–) 1 (0.4)  
All (n = 23,395)No sepsis (n = 23,045)Sepsisa (n = 350)p valueb
EOS (n = 102)LOS (n = 251)
Age, years (mean, SD) 30.4 (5.8) 30.4 (5.8) 29.7 (6.5) 29.1 (6.6) <0.001 
Ethnicity, n (%) 
 Caucasian 15,423 (65.9) 15,202 (66.0) 64 (62.7) 158 (62.9) 0.087 
 ATSI 2,196 (9.4) 2,148 (9.3) 11 (10.8) 37 (14.7)  
 Asian 3,783 (16.2) 3,732 (16.2) 19 (18.6) 33 (13.1)  
 Other 1,993 (8.5) 1,963 (8.5) 8 (7.8) 23 (9.2)  
SEIFA (<2 quintiles), n (%) 7,531 (32.2) 7,416 (32.2) 35 (34.3) 82 (32.7) 0.848 
Amnionitis, n (%) 794 (3.4) 736 (3.2) 26 (25.5) 33 (13.1) <0.001 
Preeclampsia, n (%) 2,397 (10.2) 2,358 (10.2) 1 (1.0) 44 (14.5) 0.577 
Gestational diabetes, n (%) 3,719 (15.9) 3,693 (16.0) 11 (10.8) 17 (5.6) <0.001 
Multiple birth, n (%) 3,231 (13.8) 3,176 (13.8) 7 (6.9) 48 (19.1) 0.298 
Antenatal antibiotics, n (%) 5,850 (25.0) 5,700 (24.7) 46 (45.1) 105 (41.8) <0.001 
Antenatal steroids, n (%) 
 None/<24 hc 1,822/5,533 (32.9) 1,723/5,243 (32.9) 19 (37.3) 81 (33.6) 0.653 
 At least one dose >24 hc 3,711/5,533 (67.1) 3,520/5,243 (67.1) 32 (62.7) 160 (66.4)  
 Not stated (n = 13)      
Inborn, n (%) 22,079 (94.4) 21,769 (94.5) 84 (82.4) 229 (91.2) <0.001 
ROM, h (median, IQR) 4 (0–12) 4 (0–12) 18 (1–49) 0 (0–18) 0.722 
Mode, n (%) 
 Vaginal 11,510 (49.2) 11,356 (49.2) 51 (50) 105 (39.8) <0.001 
 Caesarean section 11,865 (50.7) 11,670 (50.6) 51 (50) 145 (57.8)  
 Not stated 20 (0.1) 19 (0.1) 0 (–) 1 (0.4)  

EOS, early-onset sepsis; LOS, late-onset sepsis; ATSI, Aboriginal and Torres Strait Islander; SEIFA, Socioeconomic Index for Areas; ROM, rupture of membranes.

aThree cases had at least one episode of EOS and LOS.

bp value represents the comparison of any sepsis (EOS and LOS) relative to no sepsis.

cTwenty-four hours prior to delivery.

Table 2.

Neonatal characteristics

All (n = 23,395)No sepsis (n = 23,045)Sepsisa (n = 350)p valueb
EOS (n = 102)LOS (n = 251)
Gestational age, weeks (median, IQR) 37 (34–39) 37 (34–39) 34 (27–38) 26 (25–29) <0.001 
 <28, n (%) 1,133 (4.8) 947 (4.1) 27 (26.5) 160 (63.7) <0.001 
 28–31+6, n (%) 2,177 (9.3) 2,099 (9.1) 10 (9.8) 69 (27.5)  
 32–36+6, n (%) 8,103 (34.6) 8,059 (35.0) 29 (28.4) 15 (6.0)  
 ≥37, n (%) 11,982 (51.2) 11,940 (51.8) 36 (35.3) 7 (2.8)  
Birth weight, g (median, IQR) 2,800 (2,010–3,450) 2,820 (2,040–3,455) 2,205 (1,071–3,266) 868 (655–1,168) <0.001 
Birth weight z-score (median, IQR) −0.06 (−0.81 to 0.69) −0.06 (−0.81 to 0.69) 0.07 (−0.46 to 0.66) −0.02 (−0.72 to 0.75) 0.236 
Male sex, n (%) 12,964 (55.4) 12,739 (55.3) 57 (55.9) 169 (67.3) <0.001 
Apgar <7 at 1 min, n (%) 7,226 (30.9) 6,989 (30.3) 56 (54.9) 183 (72.9) <0.001 
Apgar <7 at 5 min, n (%) 1,909 (8.2) 1,808 (7.8) 25 (24.5) 78 (31.1) <0.001 
Age at onset of sepsis, days 8 (1–14) NA 0 (range: 0–2) 11 (8–17) NA 
TPN days (median, IQR) 5 (3–8) 5 (3–8) 8 (3–12) 16 (11–23) <0.001 
Age at full feeds, days, median, IQR 1 (0–3) 1 (0–3) 4 (1–9) 14 (9–20) <0.001 
Mode of feeding (at discharge)c, n (%) 
 Breastmilk 9,846 (43.9) 9,684 (43.7) 41 (50.0) 122 (59.5) <0.001 
 Formula 3,112 (13.9) 3,032 (13.7) 21 (25.6) 60 (23.9)  
 Mixed 9,463 (42.2) 9,421 (42.6) 20 (24.4) 23 (11.2)  
Ventilation, n (%) 3,377 (14.4) 3,064 (13.3) 49 (48.0) 219 (87.3) <0.001 
 Duration, hd 18 (7–89) 16 (7–64) 176 (31–673) 346 (78–790) <0.001 
CPAP, n (%) 8,922 (38.1) 8,592 (37.4) 64 (62.7) 216 (86.1) <0.001 
 Duration, hd 25 (8–112) 24 (7–99) 53 (20–749) 925 (443–1,250) <0.001 
Umbilical line, n (%) 1,709 (7.3) 1,503 (6.5) 38 (37.3) 169 (67.3) <0.001 
 Duration, daysd 86 (47–124) 83 (42–120) 120 (89–149) 120 (80–150) <0.001 
PICC line, n (%) 544 (2.3) 407 (1.8) 23 (22.5) 115 (45.8) <0.001 
 Duration, daysd 8 (6–12) 8 (5–11) 10 (7–17) 11 (7–18) <0.001 
Major anomaly, n (%) 525 (2.2) 509 (2.2) 4 (3.9) 12 (4.8) 0.003 
Antibiotics given, n (%) 12,539 (53.6) 12,189 (52.9) 102 (100) 251 (100) <0.001 
Antibiotic days per sepsis episode (median, IQR) 8 (6–12) (n = 350) NA 8 (6–14) 8 (6–12) NA 
All (n = 23,395)No sepsis (n = 23,045)Sepsisa (n = 350)p valueb
EOS (n = 102)LOS (n = 251)
Gestational age, weeks (median, IQR) 37 (34–39) 37 (34–39) 34 (27–38) 26 (25–29) <0.001 
 <28, n (%) 1,133 (4.8) 947 (4.1) 27 (26.5) 160 (63.7) <0.001 
 28–31+6, n (%) 2,177 (9.3) 2,099 (9.1) 10 (9.8) 69 (27.5)  
 32–36+6, n (%) 8,103 (34.6) 8,059 (35.0) 29 (28.4) 15 (6.0)  
 ≥37, n (%) 11,982 (51.2) 11,940 (51.8) 36 (35.3) 7 (2.8)  
Birth weight, g (median, IQR) 2,800 (2,010–3,450) 2,820 (2,040–3,455) 2,205 (1,071–3,266) 868 (655–1,168) <0.001 
Birth weight z-score (median, IQR) −0.06 (−0.81 to 0.69) −0.06 (−0.81 to 0.69) 0.07 (−0.46 to 0.66) −0.02 (−0.72 to 0.75) 0.236 
Male sex, n (%) 12,964 (55.4) 12,739 (55.3) 57 (55.9) 169 (67.3) <0.001 
Apgar <7 at 1 min, n (%) 7,226 (30.9) 6,989 (30.3) 56 (54.9) 183 (72.9) <0.001 
Apgar <7 at 5 min, n (%) 1,909 (8.2) 1,808 (7.8) 25 (24.5) 78 (31.1) <0.001 
Age at onset of sepsis, days 8 (1–14) NA 0 (range: 0–2) 11 (8–17) NA 
TPN days (median, IQR) 5 (3–8) 5 (3–8) 8 (3–12) 16 (11–23) <0.001 
Age at full feeds, days, median, IQR 1 (0–3) 1 (0–3) 4 (1–9) 14 (9–20) <0.001 
Mode of feeding (at discharge)c, n (%) 
 Breastmilk 9,846 (43.9) 9,684 (43.7) 41 (50.0) 122 (59.5) <0.001 
 Formula 3,112 (13.9) 3,032 (13.7) 21 (25.6) 60 (23.9)  
 Mixed 9,463 (42.2) 9,421 (42.6) 20 (24.4) 23 (11.2)  
Ventilation, n (%) 3,377 (14.4) 3,064 (13.3) 49 (48.0) 219 (87.3) <0.001 
 Duration, hd 18 (7–89) 16 (7–64) 176 (31–673) 346 (78–790) <0.001 
CPAP, n (%) 8,922 (38.1) 8,592 (37.4) 64 (62.7) 216 (86.1) <0.001 
 Duration, hd 25 (8–112) 24 (7–99) 53 (20–749) 925 (443–1,250) <0.001 
Umbilical line, n (%) 1,709 (7.3) 1,503 (6.5) 38 (37.3) 169 (67.3) <0.001 
 Duration, daysd 86 (47–124) 83 (42–120) 120 (89–149) 120 (80–150) <0.001 
PICC line, n (%) 544 (2.3) 407 (1.8) 23 (22.5) 115 (45.8) <0.001 
 Duration, daysd 8 (6–12) 8 (5–11) 10 (7–17) 11 (7–18) <0.001 
Major anomaly, n (%) 525 (2.2) 509 (2.2) 4 (3.9) 12 (4.8) 0.003 
Antibiotics given, n (%) 12,539 (53.6) 12,189 (52.9) 102 (100) 251 (100) <0.001 
Antibiotic days per sepsis episode (median, IQR) 8 (6–12) (n = 350) NA 8 (6–14) 8 (6–12) NA 

EOS, early-onset sepsis; LOS, late-onset sepsis; NA, not applicable; CPAP, continuous positive airway pressure; PICC, line: peripherally inserted central line.

aThree cases had at least one episode of EOS and LOS.

bp value represents the comparison of any sepsis (EOS and LOS) relative to no sepsis.

cExcludes 974 infants who died before discharge (n = 203), never fed/NBM (n = 582), or were unknown (n = 189).

dKaplan-Meier survival estimates used to summarise median (25th–75th percentile) durations of respiratory, central venous access, and antibiotic exposure.

Neonatal Sepsis

EOS Epidemiology

There were 102 episodes of EOS with an incidence of 1.6 per 1,000 live births for all GA groups combined and 0.6 per 1,000 live births >35 weeks GA. The predominant blood culture isolates in EOS were Streptococcus agalactiae (35, 34.3%), Escherichia coli (27, 26.5%), H. influenzae (7, 6.9%), Streptococcus pneumoniae (6, 5.9%), and S. aureus (5, 4.9%) (Fig. 2). E. coli was more common in preterm infants, particularly those born <28 weeks (n = 10/27, 37.0%) compared with those >35 weeks (n = 5/27, 18.5%), while S. agalactiae was more common in mature infants >35 weeks (n = 16) compared with those <28 weeks (n = 3). There was no significant change in pathogen pattern over time.

Fig. 2.

Blood culture isolates for early- and late-onset neonatal sepsis.

Fig. 2.

Blood culture isolates for early- and late-onset neonatal sepsis.

Close modal

LOS Epidemiology

There were 268 episodes of LOS in 251 infants, with an overall incidence of 4/1,000 live births and 0.7/1,000 inpatient days. LOS was almost exclusively observed in preterm infants in this non-surgical cohort and occurred in 160/1,088 (14.7%) babies <28 weeks, 69/2,165 (3.2%) at 28–31+6 weeks, 15/20,051 (0.2%) 32–36+6 weeks, and 7/7,000 (0.1%) ≥37 weeks gestation. The predominant LOS blood culture isolates were CONS (156, 57.6%), E. coli (30, 11.1%), S. aureus (27, 10.0%), K. pneumoniae (10, 3.7%), and S. agalactiae (9, 3.3%) (Fig. 2). There was a significant decrease in LOS episodes due to CONS isolates in infants <32 weeks gestation from 112/154 (72.7%) in 2012–2016 to 33/66 (50.0%) in 2017–2022 (p < 0.001). Three infants (0.9%) had episodes of both EOS and LOS (Table 3).

Table 3.

Culture isolates by gestational age and time of onset

2012–20162017–2022p value
EOS, n (%) 
 <28 weeks, n 
  S. agalactiae 2 (0.4) 1 (0.2) 0.609 
  E. coli 6 (1.1) 4 (0.7) 0.533 
  H. influenzae 4 (0.7) 0 (–) 0.052 
  Total, n 541 591  
 28–34 weeks, n 
  S. agalactiae 9 (0.3) 1 (0.03) 0.024 
  E. coli 7 (0.2) 5 (0.2) 0.779 
  H. influenzae 1 (0.03) 1 (0.03) 0.999 
  Total, n 3,304 2,919  
 35+ weeks, n 
  S. agalactiae 4 (0.02) 12 (0.04) 0.082 
  E. coli 3 (0.01) 2 (0.01) 0.674 
  H. influenzae 0 (–) 0 (–)  
  Total, n 25,045 27,619  
 All 
  S. agalactiae 15 (0.05) 14 (0.04) 0.714 
  E. coli 16 (0.06) 11 (0.04) 0.256 
  H. influenzae 5 (0.02) 1 (0.003) 0.112 
  Total, n 28,890 31,129  
LOS 
 <32 weeks 
  CONS 112 (7.2) 33 (2.1) <0.001 
  E. coli 13 (0.8) 11 (0.7) 0.686 
  S. aureus 12 (0.8) 8 (0.5) 0.376 
  K. pneumoniae 6 (0.4) 3 (0.2) 0.337 
  S. agalactiae 3 (0.06) 5 (0.3) 0.726 
  Total, n 1,549 1,590  
2012–20162017–2022p value
EOS, n (%) 
 <28 weeks, n 
  S. agalactiae 2 (0.4) 1 (0.2) 0.609 
  E. coli 6 (1.1) 4 (0.7) 0.533 
  H. influenzae 4 (0.7) 0 (–) 0.052 
  Total, n 541 591  
 28–34 weeks, n 
  S. agalactiae 9 (0.3) 1 (0.03) 0.024 
  E. coli 7 (0.2) 5 (0.2) 0.779 
  H. influenzae 1 (0.03) 1 (0.03) 0.999 
  Total, n 3,304 2,919  
 35+ weeks, n 
  S. agalactiae 4 (0.02) 12 (0.04) 0.082 
  E. coli 3 (0.01) 2 (0.01) 0.674 
  H. influenzae 0 (–) 0 (–)  
  Total, n 25,045 27,619  
 All 
  S. agalactiae 15 (0.05) 14 (0.04) 0.714 
  E. coli 16 (0.06) 11 (0.04) 0.256 
  H. influenzae 5 (0.02) 1 (0.003) 0.112 
  Total, n 28,890 31,129  
LOS 
 <32 weeks 
  CONS 112 (7.2) 33 (2.1) <0.001 
  E. coli 13 (0.8) 11 (0.7) 0.686 
  S. aureus 12 (0.8) 8 (0.5) 0.376 
  K. pneumoniae 6 (0.4) 3 (0.2) 0.337 
  S. agalactiae 3 (0.06) 5 (0.3) 0.726 
  Total, n 1,549 1,590  

Temporal Trend

There was no significant change in EOS incidence in any GA group during the study period, and the trend toward lower EOS incidence in those <28 weeks of GA was not significant (p = 0.140). However, there was a decrease in S. agalactiae in infants 28–34 weeks of gestation which decreased from 9/3,304 (0.3%) to 1/2,919 (0.03%) between periods 2012–2016 and 2017–2022, respectively (p = 0.024) (Table 3).

There was a decrease in LOS across all GA groups from 1.7% of admissions in 2012/2013 to 0.6% of admissions in 2020/2021 (p < 0.001), but most significantly in the highest risk group <28 weeks of GA from 21.2% in 2012/2013 to 9.9% in 2020/2021 (p = 0.002) (Fig. 3). There appeared to be an increase in both EOS and LOS between 2014 and 2018 (Fig. 3). However, comparison of the incidence of EOS in 2014–2018 with the preceding (2012–2013) and following (2019–2021) periods showed no statistically significant difference, and the incidence of LOS was higher in the preceding (2012–2013) and lower in the following (2019–2021) period compared with 2014–2018, consistent with a significant reduction in LOS over time.

Fig. 3.

Incidence of EOS and LOS by gestational age and year.

Fig. 3.

Incidence of EOS and LOS by gestational age and year.

Close modal

Sepsis with Meningitis

There were 18 CSF culture-positive cases, all of which corresponded with a positive blood culture with the same organism. Meningitis occurred in 10/102 (9.8%) cases of EOS and 8/251 (3.2%) cases of LOS. Of the 10 EOS meningitis cases, CSF cultures grew Streptococcus agalactiae (n = 4), Streptococcus pneumoniae (n = 1), and Escherichia coli (n = 5), while in the 8 infants with LOS meningitis, CSF cultures were positive with Streptococcus agalactiae (n = 2), Enterococcus faecalis (n = 1), Klebsiella pneumonia (n = 1), and Escherichia coli (n = 4).

Infants with sepsis and meningitis had a greater median gestational age (39 vs. 33 weeks, p = 0.015) and birth weight (3,005 g vs. 1,935 g, p = 0.02) compared to infants with sepsis without positive CSF culture. Antibiotic duration in CSF-positive cases was a median of 20 days (IQR: 14–23) compared with a median of 8 days (IQR: 6–12) for neonatal sepsis without meningitis (online suppl. File; for all online suppl. material, see https://doi.org/10.1159/000539174).

Neonatal Sepsis and Outcomes until Hospital Discharge

Term infants with EOS had longer hospital stay (median 8 [IQR: 7–15] versus 3 [IQR: 2–5] days; aHR: 2.42, 95% CI: 1.73–3.39, p < 0.001) and higher mortality (5.7% vs. 0.3%, p < 0.001) (sample size insufficient for adjusted models) than those without EOS. Other neonatal outcomes were not different between infants with and without sepsis (Table 4).

Table 4.

Neonatal outcomes at time of hospital discharge in preterm patients (gestational age <37 weeks) with EOSa

EOS (N = 64)No EOS (N = 11,105)aOR/aHR (95% CI)p value
Length of stay, daysb (median, IQR) 31 (10–85) 12 (3–29) 1.26 (0.96–1.66) 0.095 
Mortality, n (%) 12 (18.8) 178 (1.6) 3.87 (1.78–8.42) <0.001 
NEC ≥grade 2, n (%) 1 (1.6) 53 (0.5) 0.92 (0.12–6.98) 0.932 
CLDc, n (%) 15/52 (28.8) 642/10,950 (5.9) 2.14 (0.69–6.60) 0.186 
ROP≥stage 3d, n (%) 5/34 (14.7) 118/2,440 (4.8) 1.21 (0.42–3.45) 0.723 
Significant brain injurye, n (%) 9/35 (25.7) 124/3,046 (4.1) 3.96 (1.68–9.36) 0.002 
EOS (N = 64)No EOS (N = 11,105)aOR/aHR (95% CI)p value
Length of stay, daysb (median, IQR) 31 (10–85) 12 (3–29) 1.26 (0.96–1.66) 0.095 
Mortality, n (%) 12 (18.8) 178 (1.6) 3.87 (1.78–8.42) <0.001 
NEC ≥grade 2, n (%) 1 (1.6) 53 (0.5) 0.92 (0.12–6.98) 0.932 
CLDc, n (%) 15/52 (28.8) 642/10,950 (5.9) 2.14 (0.69–6.60) 0.186 
ROP≥stage 3d, n (%) 5/34 (14.7) 118/2,440 (4.8) 1.21 (0.42–3.45) 0.723 
Significant brain injurye, n (%) 9/35 (25.7) 124/3,046 (4.1) 3.96 (1.68–9.36) 0.002 

EOS, early-onset sepsis; LOS, late-onset sepsis; aHR, adjusted hazard ratio; aOR, adjusted odds ratio; CI, confidence interval; NEC, necrotising enterocolitis; CLD, chronic lung disease; ROP, retinopathy of prematurity; significant brain injury, IVH ≥ grade 3 and/or cystic PVL.

aTerm EOS infants (not included in table) had higher odds of longer hospital stay (median 3 vs. 1 days; aHR: 1.71, 95% CI: 1.21–2.40, p = 0.002) and mortality (5.7% vs. 0.3%, p < 0.001) (insufficient numbers to analyse in adjusted models). Other neonatal outcomes were not relevant to this group.

bDeaths excluded, missing length of stay days (n = 8).

cExcludes deaths before 36 weeks corrected gestational age.

dAssessed among infants of gestational age <31 weeks and/or BW <1,250 g.

eAssessed among infants of gestational age <32 weeks.

Preterm infants with EOS had increased odds of brain injury (25.7% vs. 4.1%; aOR: 3.96, 95% CI: 1.68–9.36, p = 0.002) and mortality (18.8% vs. 1.6%; aOR: 3.87, 95% CI: 1.78–8.42, p < 0.001) compared with preterm infants without sepsis. There were no significant differences in length of stay, rate of NEC ≥ grade 2, CLD, or ROP ≥ stage 3 (Table 4).

Overall, preterm infants with LOS had longer hospital stay (median 95 vs. 15 days; aHR: 1.57, 95% CI: 1.36–1.81, p < 0.001), and increased odds of mortality (15.3% vs. 1.6%; aOR: 1.68, 95% CI: 1.09–2.58, p = 0.018), NEC ≥ grade 2 (7.4% vs. 0.5%; aOR: 3.05, 95% CI: 1.64–5.66, p < 0.001), and CLD (58.1% vs. 5.9%; aOR: 1.78, 95% CI: 1.19–2.65, p = 0.005). There were no significant differences in brain injury or ROP ≥ stage 3 (Table 5). These findings were consistent across gestational age categories <28 weeks, 28–31+6 weeks and 32–34+6 weeks, with the exception of ROP and brain injury which are not routinely assessed in infants ≥32 weeks of gestation (Table 5).

Table 5.

Neonatal outcomes at time of hospital discharge by gestational age in patients with LOS

Gestational age
<28 weeks28–31+6 weeks32–34+6 weeks≥35 weeks
LOS (N = 160)No LOS (N = 947)p valueLOS (N = 69)no LOS (N = 2,099)p valueLOS (N = 12)no LOS (N = 4,012)p valueLOS (N = 12)no LOS (N = 4,012)p value
Length of stay, days, median (IQR)a 101 (74–120) 87 (67–108) <0.001 49 (33–65) 39 (25–57) <0.001 45 (20–71) 13 (7–21) 0.001 15 (8–23) 4 (2–6) <0.001 
Mortality, n (%) 30 (18.8) 111 (11.7) 0.014 4 (5.8) 30 (1.4) 0.021b 3 (25.0) 23 (0.6) <0.001b 0 (–) 57 (0.4) 1.000b 
NEC ≥grade 2, n (%) 13 (8.1) 27 (2.9) <0.001 3 (4.3) 15 (0.7) 0.018b 2 (16.7) 8 (0.2) <0.001b 0 (–) 5 (0.03) 1.000b 
CLDc, n (%) 104 (78.2) 460 (54.9) <0.001 17 (25.4) 171 (8.2) <0.001 2 (22.2) 10 (0.3) <0.001b 0 (–) 1 (0.006) 1.000b 
ROP≥stage 3d, n (%) 22 (13.8) 109 (11.5) 0.417 1 (1.6) 9 (0.6) 0.360b 0 (–) 0 (–) NA NA NA 
Significant brain injurye, n (%) 16 (10.0) 85 (9.0) 0.677 3 (4.3) 39 (1.9) 0.147b NA NA NA NA NA 
Gestational age
<28 weeks28–31+6 weeks32–34+6 weeks≥35 weeks
LOS (N = 160)No LOS (N = 947)p valueLOS (N = 69)no LOS (N = 2,099)p valueLOS (N = 12)no LOS (N = 4,012)p valueLOS (N = 12)no LOS (N = 4,012)p value
Length of stay, days, median (IQR)a 101 (74–120) 87 (67–108) <0.001 49 (33–65) 39 (25–57) <0.001 45 (20–71) 13 (7–21) 0.001 15 (8–23) 4 (2–6) <0.001 
Mortality, n (%) 30 (18.8) 111 (11.7) 0.014 4 (5.8) 30 (1.4) 0.021b 3 (25.0) 23 (0.6) <0.001b 0 (–) 57 (0.4) 1.000b 
NEC ≥grade 2, n (%) 13 (8.1) 27 (2.9) <0.001 3 (4.3) 15 (0.7) 0.018b 2 (16.7) 8 (0.2) <0.001b 0 (–) 5 (0.03) 1.000b 
CLDc, n (%) 104 (78.2) 460 (54.9) <0.001 17 (25.4) 171 (8.2) <0.001 2 (22.2) 10 (0.3) <0.001b 0 (–) 1 (0.006) 1.000b 
ROP≥stage 3d, n (%) 22 (13.8) 109 (11.5) 0.417 1 (1.6) 9 (0.6) 0.360b 0 (–) 0 (–) NA NA NA 
Significant brain injurye, n (%) 16 (10.0) 85 (9.0) 0.677 3 (4.3) 39 (1.9) 0.147b NA NA NA NA NA 

LOS, late-onset sepsis; NEC, necrotising enterocolitis; CLD, chronic lung disease; ROP, retinopathy of prematurity; Significant brain injury, IVH ≥ grade 3 and/or cystic PVL, NA, not applicable.

aDeaths excluded.

bFisher Exact test used.

cExcludes deaths before 36 weeks corrected gestational age.

dAssessed among infants of gestational age <31 weeks and/or BW <1,250 g.

eAssessed among infants of gestational age <32 weeks.

Certain LOS outcomes were associated with specific organism types, for example, mortality was significantly higher in LOS caused by gram-negative (15, 25.4%) compared to gram-positive organisms (3, 7.1%) or CONS (16, 11.0%) (p = 0.01). In preterm infants, NEC ≥ grade 2 was associated with gram-negative organisms (9, 15.3%) but not gram-positive organisms (0) or CONS (9, 6.3%) (p = 0.011) (Table 6).

Table 6.

Neonatal characteristics and outcomes by culture type for the first episode of LOS

Gram-positive (N = 42)CONS (N = 145)Gram-negative (N = 59)p value
Neonatal characteristics 
 Gestational age, weeks (median, IQR) 26.4 (24.7–28.7) 26.6 (25.0–29.1) 25.9 (24.3–28.7) 0.297 
 BW (median, IQR) 827 (643–1,152) 900 (685–1,212) 850 (620–1,215) 0.310 
 BW z-score (median, IQR) 0.17 (−0.87 to 1.08) −0.06 (−0.83 to 0.74) 0.07 (−0.68 to 0.62) 0.881 
 Male, n (%) 23 (54.8) 98 (67.6) 44 (74.6) 0.111 
Neonatal outcomes 
 Length of stay, daysa 79 (46–111) 71 (44–110) 72 (35–102) 0.569 
 Mortality, n (%) 3 (7.1) 16 (11.0) 15 (25.4) 0.010 
 NEC ≥grade 2, n (%) 0 (–) 9 (6.3) 9 (15.3) 0.011 
 CLDb, n (%) 26 (65.0) 69 (52.3) 26 (57.8) 0.348 
 ROP ≥stage 3c, n (%) 7 (17.9) 11 (8.5) 4 (7.7) 0.206 
 IVH ≥grade 3d, n (%) 4 (10.5) 7 (5.2) 5 (9.8) 0.348 
 Cystic PVLd, n (%) 0 (–) 2 (1.5) 2 (3.9) 0.349 
 Brain injuryd, n (%) 4 (10.5) 7 (5.2) 7 (13.7) 0.115 
Gram-positive (N = 42)CONS (N = 145)Gram-negative (N = 59)p value
Neonatal characteristics 
 Gestational age, weeks (median, IQR) 26.4 (24.7–28.7) 26.6 (25.0–29.1) 25.9 (24.3–28.7) 0.297 
 BW (median, IQR) 827 (643–1,152) 900 (685–1,212) 850 (620–1,215) 0.310 
 BW z-score (median, IQR) 0.17 (−0.87 to 1.08) −0.06 (−0.83 to 0.74) 0.07 (−0.68 to 0.62) 0.881 
 Male, n (%) 23 (54.8) 98 (67.6) 44 (74.6) 0.111 
Neonatal outcomes 
 Length of stay, daysa 79 (46–111) 71 (44–110) 72 (35–102) 0.569 
 Mortality, n (%) 3 (7.1) 16 (11.0) 15 (25.4) 0.010 
 NEC ≥grade 2, n (%) 0 (–) 9 (6.3) 9 (15.3) 0.011 
 CLDb, n (%) 26 (65.0) 69 (52.3) 26 (57.8) 0.348 
 ROP ≥stage 3c, n (%) 7 (17.9) 11 (8.5) 4 (7.7) 0.206 
 IVH ≥grade 3d, n (%) 4 (10.5) 7 (5.2) 5 (9.8) 0.348 
 Cystic PVLd, n (%) 0 (–) 2 (1.5) 2 (3.9) 0.349 
 Brain injuryd, n (%) 4 (10.5) 7 (5.2) 7 (13.7) 0.115 

Excludes LOS cases with fungal pathogens (n = 3) and polysepsis with gram-positive and gram-negative pathogens (n = 2). p value represents the overall comparison between LOS groups by the independent samples Kruskal-Wallis test for continuous data and the χ2 or Fisher-Freeman-Halton Exact test for categorical data. All percentages are the percentage of those tested.

LOS, late-onset sepsis; NEC, necrotising enterocolitis; CLD, chronic lung disease; ROP, retinopathy of prematurity; significant brain injury, IVH ≥grade 3 and/or cystic PVL; IVH, intraventricular haemorrhage; PVL, periventricular cysts.

aExcludes deaths before discharge.

bExcludes deaths before 36 weeks corrected gestational age.

cAssessed among infants of gestational age <31 weeks and/or BW <1,250 g.

dAssessed among infants of gestational age <32 weeks.

Neurodevelopmental Outcomes

Developmental outcome data from the routinely offered follow-up program were available for 1,029 infants. BSID assessments were conducted at a median corrected age of 24 months (IQR: 24–25; range: 19–42 months). There were 15 exclusions due to major congenital anomalies that may affect developmental progress, of whom only one infant had sepsis. Therefore, developmental outcome data were analysed in 1,014 infants. The prevalence of any degree of impairment was higher in sepsis infants born <28 weeks of gestation on univariate analysis (43.2% vs. 30.9%, p = 0.027) (Table 7). As expected, infants who did not attend neurodevelopmental assessments in early childhood were more likely to be outborn, offspring of younger, non-Caucasian mothers of lower socioeconomic status, and with higher BW z-scores (online suppl. Material).

Table 7.

Univariate analysis of neurodevelopmental impairment in survivors at 2 years of corrected age by gestational age and sepsis category (n = 1,014)

AllNo sepsisSepsisp valuea
Impairment level 
All n = 1,014 n = 887 n = 127  
 None 624 (71.4) 642 (72.4) 82 (64.6) 0.068 
 Mild-severe 290 (28.6) 245 (27.6) 45 (35.4)  
<28 weeks n = 464 n = 376 n = 87  
 None 310 (66.8) 260 (69.1) 50 (56.8) 0.027 
 Mild-severe 154 (33.2) 116 (30.9) 38 (43.2)  
≥28 weeks n = 550 n = 511 n = 39  
 None 414 (75.3) 382 (74.8) 32 (82.1) 0.309 
 Mild-severe 136 (24.7) 129 (25.2) 7 (17.9)  
AllNo sepsisSepsisp valuea
Impairment level 
All n = 1,014 n = 887 n = 127  
 None 624 (71.4) 642 (72.4) 82 (64.6) 0.068 
 Mild-severe 290 (28.6) 245 (27.6) 45 (35.4)  
<28 weeks n = 464 n = 376 n = 87  
 None 310 (66.8) 260 (69.1) 50 (56.8) 0.027 
 Mild-severe 154 (33.2) 116 (30.9) 38 (43.2)  
≥28 weeks n = 550 n = 511 n = 39  
 None 414 (75.3) 382 (74.8) 32 (82.1) 0.309 
 Mild-severe 136 (24.7) 129 (25.2) 7 (17.9)  

ap value represents the χ2 test comparison in disability level between sepsis categories.

This large, retrospective cohort study including over 23,000 neonates showed a decline in LOS but not EOS incidence from 2012 to 2022. While gram-positive organisms were most commonly isolated from blood cultures, S. agalactiae in EOS and CONS in LOS, gram-negative organisms were associated with higher mortality. In the cohort overall, any neonatal sepsis was associated with longer hospital stay, higher mortality, increased NEC ≥ grade 2, and increased CLD. In preterm infants, EOS was associated with increased mortality and significant brain injury, whereas LOS was associated with longer hospital stays, higher mortality, increased rate of NEC ≥ grade 2, and increased CLD.

The observed EOS incidence of 0.6/1,000 live births is similar to that in other areas of Australia and other high-income countries [27, 28]. During the study period, the EOS incidence did not change significantly; however, there was a decrease in the number of S. agalactiae isolates in infants 28–34 weeks of gestation. The introduction of intrapartum antibiotic prophylaxis in the early 1990s resulted in a decline in EOS as reported in several studies from Australasia between 1992 and 2006, which predated the current project and correlated with the steepest decline in EOS [29, 30]. Ongoing efforts to reduce EOS incidence, such as enhanced screening, point-of-care testing, and development of an effective vaccine, are warranted [31, 32].

The LOS rates of 14.7%, 3.2%, and 0.1% in neonates <28 weeks, 28–31+6, and ≥32 weeks of gestation respectively are in line with or lower than those from other high-income countries such as the USA, Canada, and Israel [33]. The observed decline in LOS, predominantly due to a decrease in CONS, is likely associated with the implementation of changes to routine management of the most vulnerable infants, such as probiotic supplementation, earlier introduction and faster advancement of enteral feeds, use of topical coconut oil for skin integrity, improved insertion technique and enhanced surveillance of intravenous cannulas, and attention to hand hygiene [33, 34]. Although suggestive, these practices were introduced as changes to routine patient care, not as part of clinical trials, and therefore firm conclusions regarding the efficacy of individual initiatives cannot be reached. Evidence-based strategies to prevent LOS are limited and largely based on moderate quality and/or observational research as opposed to randomised control trials. A multifaceted approach will likely have the greatest impact, including early skin-to-skin contact, human milk feeding, probiotic supplementation, maintaining skin integrity, judicious use of vascular access devices, evidence-based bundles for central line management, isolation and cohorting practices, hand hygiene, and environmental decontamination [35, 36].

The pattern of EOS pathogens was consistent with other high-income countries, with S. agalactiae and E. coli most commonly identified. Similar to findings from other centres, E. coli was the predominant cause of EOS in preterm infants, particularly those <28 weeks of gestation, and S. agalactiae most commonly isolated in infants >35 weeks [37]. Several other centres have found a significant decrease in EOS due to S. agalactiae, this associated with static or increasing EOS with E. coli, across all gestational age groups [38, 39]. In our cohort, total numbers per organism were small, limiting meaningful conclusions.

Consistent with reports from other high-income settings, CONS, E. coli, and S. aureus were the most common LOS pathogens [33, 40]. Although gram-positive organisms were more commonly isolated, gram-negative organisms were associated with higher mortality. The recent NeoOBS study similarly reports 21.3% mortality in gram-negative compared with 8.5% in gram-positive sepsis [14]. In addition, gram-negative sepsis was responsible for >75% of all culture-confirmed sepsis deaths in that cohort [14].

The prevalence of meningitis complicating EOS and LOS was similar to previously reported findings from Australia and New Zealand [40, 41]. In this study, all positive CSF cultures corresponded with a positive blood culture. However, 15–38% discordance between blood and CSF cultures has been reported, and performing an LP only in case of a positive blood culture will miss a substantial proportion of infants with meningitis [42]. Conversely, performing an LP in all infants with suspected EOS, especially based on maternal risk factors or biomarkers instead of clinical signs, will have a low yield and thus an unfavourable risk-to-benefit ratio [43]. Considering the life-time risk of meningitis is highest during the neonatal period, particularly in preterm infants, and there is currently no ancillary marker for meningitis risk, the decision to perform an LP should be based on clinical suspicion in EOS, whereas a more routine approach may be appropriate in LOS in the extremely preterm infant [43]. If CSF is obtained, it should occur as early as possible in the course of the illness, ideally before antibiotic treatment is commenced, to optimise pathogen detection and determine choice of antimicrobial agent and treatment duration [43].

Any neonatal sepsis results in significantly poorer outcomes, even after adjusting for potential confounding factors. In this study, and consistent with published findings, sepsis was associated with increased mortality, higher rates of NEC, increased CLD, longer hospital stay, and risk neurodevelopmental impairment in infants <28 weeks gestation [6, 7]. It is possible that the relatively small sample was insufficiently powered to detect a difference in neurodevelopmental impairment in those >28 weeks of gestation. The retrospective study design limited availability of routine developmental follow-up data to infants meeting preset criteria. It also limited capacity for tracing non-attenders, resulting in suboptimal follow-up rates and potential exclusion of high-risk patients. In addition, specific organisms may have greater association with neurodevelopmental impairment than others. The predominant blood culture isolate in this study was CONS which has been associated with cognitive impairment but not major disability [44].

The study has several strengths including a large sample size for a vast geographical area over a significant time period. It also provides important information regarding incidence of sepsis and culture isolates to benchmark against other neonatal units and to guide empiric, site-specific antimicrobial choices. However, the study also has limitations. Firstly, exclusion of culture-negative sepsis may result in under reporting of true sepsis. Diagnosing true sepsis due to common skin commensals is challenging and, despite our best efforts, may have resulted in over-representation of CONS as a pathogenic isolate. In addition, the data analysed was from a perinatal NICU and will not reflect newborn and microbial epidemiology or a surgical NICU. Finally, the retrospective nature of the study limited data verification and participant retention in follow-up. However, these challenges are not unique to the current study, and best attempts were made to mitigate their impact.

In conclusion, we report a decreasing trend in LOS in the study centre. However, neonatal sepsis remains an important cause of morbidity and mortality, and all possible action should be taken to prevent it. Current site-specific data should be used to guide appropriate empiric antibiotic choices while simultaneously optimising infection prevention and control practices, improving accurate and timely diagnosis, and considering treatment adjuncts to improve outcome.

This study protocol was reviewed and approved by the CAHS GEKO Triage Committee, approval number 000471. The need for informed consent was waived by the CAHS GEKO Triage Committee.

The authors have no conflicts of interest to declare.

There is no funding to declare.

Cheryl Anne Mackay: investigation, data curation, resources, and writing – original draft and editing; Elizabeth A Nathan: data curation, formal analysis, resources, software, and writing – review and editing; Michelle Claire Porter: conceptualisation and writing – review and editing; Damber Shrestha: data curation, resources, software, and writing – review and editing; Rolland Kohan: data curation, resources, and writing – review and editing; Tobias Strunk: conceptualisation, investigation, data curation, supervision, and writing – review and editing.

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.

1.
Rosa-Mangeret
F
,
Benski
AC
,
Golaz
A
,
Zala
PZ
,
Kyokan
M
,
Wagner
N
, et al
.
2.5 million annual deaths-are neonates in low- and middle-income countries too small to be seen? A bottom-up overview on neonatal morbi-mortality
.
Trop Med Infect Dis
.
2022
;
7
(
5
):
64
.
2.
Giannoni
E
,
Agyeman
PKA
,
Stocker
M
,
Posfay-Barbe
KM
,
Heininger
U
,
Spycher
BD
, et al
.
Neonatal sepsis of early onset, and hospital-acquired and community-acquired late onset: a prospective population-based cohort study
.
J Pediatr
.
2018
;
201
:
106
14.e4
.
3.
Li
J
,
Shen
L
,
Qian
K
.
Global, regional, and national incidence and mortality of neonatal sepsis and other neonatal infections, 1990-2019
.
Front Public Health
.
2023
;
11
:
1139832
.
4.
Mashau
RC
,
Meiring
ST
,
Dramowski
A
,
Magobo
RE
,
Quan
VC
,
Perovic
O
, et al
.
Culture-confirmed neonatal bloodstream infections and meningitis in South Africa, 2014-19: a cross-sectional study
.
Lancet Glob Health
.
2022
;
10
(
8
):
e1170
8
.
5.
Chow
SSW
,
Holberton
JR
,
Chambers
GM
,
Lui
K
.
Report of the Australian and New Zealand Neonatal Network 2021
.
ANZNN: Sydney
:
Australian and New Zealand Neonatal Network 2021
;
2023
.
6.
Cai
S
,
Thompson
DK
,
Anderson
PJ
,
Yang
JYM
.
Short- and long-term neurodevelopmental outcomes of very preterm infants with neonatal sepsis: a systematic review and meta-analysis
.
Children
.
2019
;
6
(
12
):
131
.
7.
Coggins
SA
,
Glaser
K
.
Updates in late-onset sepsis: risk assessment, therapy, and outcomes
.
Neoreviews
.
2022
;
23
(
11
):
738
55
.
8.
Salman
O
,
Procter
SR
,
McGregor
C
,
Paul
P
,
Hutubessy
R
,
Lawn
JE
, et al
.
Systematic review on the acute cost-of-illness of sepsis and meningitis in neonates and infants
.
Pediatr Infect Dis J
.
2020
;
39
(
1
):
35
40
.
9.
Connell
TG
,
Rele
M
,
Cowley
D
,
Buttery
JP
,
Curtis
N
.
How reliable is a negative blood culture result? Volume of blood submitted for culture in routine practice in a children’s hospital
.
Pediatrics
.
2007
;
119
(
5
):
891
6
.
10.
Eichberger
J
,
Resch
E
,
Resch
B
.
Diagnosis of neonatal sepsis: the role of inflammatory markers
.
Front Pediatr
.
2022
;
10
:
840288
.
11.
Shane
AL
,
Sánchez
PJ
,
Stoll
BJ
.
Neonatal sepsis
.
Lancet
.
2017
;
390
(
10104
):
1770
80
.
12.
Kuzniewicz
MW
,
Mukhopadhyay
S
,
Li
S
,
Walsh
EM
,
Puopolo
KM
.
Time to positivity of neonatal blood cultures for early-onset sepsis
.
Pediatr Infect Dis J
.
2020
;
39
(
7
):
634
40
.
13.
Mukhopadhyay
S
,
Briker
SM
,
Flannery
DD
,
Dhudasia
MB
,
Coggins
SA
,
Woodford
E
, et al
.
Time to positivity of blood cultures in neonatal late-onset bacteraemia
.
Arch Dis Child Fetal Neonatal Ed
.
2022
;
107
(
6
):
583
8
.
14.
Russell
NJ
,
Stöhr
W
,
Plakkal
N
,
Cook
A
,
Berkley
JA
,
Adhisivam
B
, et al
.
Patterns of antibiotic use, pathogens, and prediction of mortality in hospitalized neonates and young infants with sepsis: a global neonatal sepsis observational cohort study (NeoOBS)
.
Plos Med
.
2023
;
20
(
6
):
e1004179
.
15.
Antimicrobial Resistance Collaborators
.
Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis
.
Lancet
.
2022
;
399
(
10325
):
629
55
.
16.
World Health Organisation
.
Global action plan on antimicrobial resistance
.
Geneva, Switzerland
:
WHO
;
2015
. p.
45
.
17.
Mishra
A
,
Lai
GC
,
Yao
LJ
,
Aung
TT
,
Shental
N
,
Rotter-Maskowitz
A
, et al
.
Microbial exposure during early human development primes fetal immune cells
.
Cell
.
2021
;
184
(
13
):
3394
409.e20
.
18.
Sarkar
A
,
Yoo
JY
,
Valeria Ozorio Dutra
S
,
Morgan
KH
,
Groer
M
.
The association between early-life gut microbiota and long-term health and diseases
.
J Clin Med
.
2021
;
10
(
3
):
459
.
19.
Australia and New Zealand Neonatal Network
.
ANZNN data dictionary 2023
.
Sydney, Australia
;
2022
.
20.
Ballard
JL
,
Khoury
JC
,
Wedig
K
,
Wang
L
,
Eilers-Walsman
BL
,
Lipp
R
.
New Ballard Score, expanded to include extremely premature infants
.
J Pediatr
.
1991
;
119
(
3
):
417
23
.
21.
Kliegman
RM
,
Walsh
MC
.
Neonatal necrotizing enterocolitis: pathogenesis, classification, and spectrum of illness
.
Curr Probl Pediatr
.
1987
;
17
(
4
):
213
88
.
22.
de Vries
LS
,
Eken
P
,
Dubowitz
LM
.
The spectrum of leukomalacia using cranial ultrasound
.
Behav Brain Res
.
1992
;
49
(
1
):
1
6
.
23.
Papile
LA
,
Burstein
J
,
Burstein
R
,
Koffler
H
.
Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm
.
J Pediatr
.
1978
;
92
(
4
):
529
34
.
24.
Bayley
N
.
Bayley Scales of infant and toddler development
. 3rd ed.
San Antonio
:
Harcourt Assessment Inc
;
2006
.
25.
Park
EY
.
Stability of the gross motor function classification system in children with cerebral palsy for two years
.
BMC Neurol
.
2020
;
20
(
1
):
172
.
26.
Lamsal
R
,
Dutton
DJ
,
Zwicker
JD
.
Using the ages and stages questionnaire in the general population as a measure for identifying children not at risk of a neurodevelopmental disorder
.
BMC Pediatr
.
2018
;
18
(
1
):
122
.
27.
Braye
K
,
Foureur
M
,
de Waal
K
,
Jones
M
,
Putt
E
,
Ferguson
J
.
Epidemiology of neonatal early-onset sepsis in a geographically diverse Australian health district 2006-2016
.
PLoS One
.
2019
;
14
(
4
):
e0214298
.
28.
Giannoni
E
,
Dimopoulou
V
,
Klingenberg
C
,
Navér
L
,
Nordberg
V
,
Berardi
A
, et al
.
Analysis of antibiotic exposure and early-onset neonatal sepsis in europe, north America, and Australia
.
JAMA Netw Open
.
2022
;
5
(
11
):
e2243691
.
29.
Daley
AJ
,
Isaacs
D
;
Australasian Study Group for Neonatal Infections
.
Ten-year study on the effect of intrapartum antibiotic prophylaxis on early onset group B streptococcal and Escherichia coli neonatal sepsis in Australasia
.
Pediatr Infect Dis J
.
2004
;
23
(
7
):
630
4
.
30.
Schrag
SJ
,
Zywicki
S
,
Farley
MM
,
Reingold
AL
,
Harrison
LH
,
Lefkowitz
LB
, et al
.
Group B streptococcal disease in the era of intrapartum antibiotic prophylaxis
.
N Engl J Med
.
2000
;
342
(
1
):
15
20
.
31.
Hartvigsen
CM
,
Nielsen
SY
,
Møller
JK
,
Khalil
MR
.
Reduction of intrapartum antibiotic prophylaxis by combining risk factor assessment with a rapid bedside intrapartum polymerase chain reaction testing for group B streptococci
.
Eur J Obstet Gynecol Reprod Biol
.
2022
;
272
:
173
6
.
32.
Lawn
JE
,
Chandna
J
,
Paul
P
,
Jit
M
,
Trotter
C
,
Lambach
P
, et al
.
Every country, every family: time to act for group B streptococcal disease worldwide
.
Clin Infect Dis
.
2022
;
74
(
Suppl l_1
):
S1
4
.
33.
Sass
L
,
Karlowicz
MG
.
Healthcare-associated infections in the neonate
. In:
Principles and practice of pediatric infectious diseases
;
2018
; p.
560
6.e3
.
34.
Brachio
SS
,
Gu
W
,
Saiman
L
.
Next steps for health care-associated infections in the neonatal intensive care unit
.
Clin Perinatol
.
2023
;
50
(
2
):
381
97
.
35.
Fleiss
N
,
Tarun
S
,
Polin
RA
.
Infection prevention for extremely low birth weight infants in the NICU
.
Semin Fetal Neonatal Med
.
2022
;
27
(
3
):
101345
.
36.
Sivanandan
S
,
Sankar
MJ
.
Kangaroo mother care for preterm or low birth weight infants: a systematic review and meta-analysis
.
BMJ Glob Health
.
2023
;
8
(
6
):
e010728
.
37.
Stoll
BJ
,
Puopolo
KM
,
Hansen
NI
,
Sánchez
PJ
,
Bell
EF
,
Carlo
WA
, et al
.
Early-onset neonatal sepsis 2015 to 2017, the rise of Escherichia coli, and the need for novel prevention strategies
.
JAMA Pediatr
.
2020
;
174
(
7
):
e200593
.
38.
Duggan
HL
,
Chow
SSW
,
Austin
NC
,
Shah
PS
,
Lui
K
,
Tan
K
, et al
.
Early-onset sepsis in very preterm neonates in Australia and New Zealand, 2007-2018
.
Arch Dis Child Fetal Neonatal Ed
.
2023
;
108
(
1
):
31
7
.
39.
Schrag
SJ
,
Farley
MM
,
Petit
S
,
Reingold
A
,
Weston
EJ
,
Pondo
T
, et al
.
Epidemiology of invasive early-onset neonatal sepsis, 2005 to 2014
.
Pediatrics
.
2016
;
138
(
6
):
e20162013
.
40.
Gowda
H
,
Norton
R
,
White
A
,
Kandasamy
Y
.
Late-onset neonatal sepsis-A 10-year review from north queensland, Australia
.
Pediatr Infect Dis J
.
2017
;
36
(
9
):
883
8
.
41.
May
M
,
Daley
AJ
,
Donath
S
,
Isaacs
D
;
Australasian Study Group for Neonatal Infections
.
Early onset neonatal meningitis in Australia and New Zealand, 1992–2002
.
Arch Dis Child Fetal Neonatal Ed
.
2005
;
90
(
4
):
F324
7
.
42.
Ku
LC
,
Boggess
KA
,
Cohen-Wolkowiez
M
.
Bacterial meningitis in infants
.
Clin Perinatol
.
2015
;
42
(
1
):
29
45
. vii-viii.
43.
Aleem
S
,
Greenberg
RG
.
When to include a lumbar puncture in the evaluation for neonatal sepsis
.
Neoreviews
.
2019
;
20
(
3
):
e124
34
.
44.
Alshaikh
B
,
Yee
W
,
Lodha
A
,
Henderson
E
,
Yusuf
K
,
Sauve
R
.
Coagulase-negative staphylococcus sepsis in preterm infants and long-term neurodevelopmental outcome
.
J Perinatol
.
2014
;
34
(
2
):
125
9
.