Abstract
Background
Culture-negative sepsis (CNNS) constitutes a significant proportion of neonates admitted to the NICU. However, the outcomes and factors influencing antimicrobial therapy in this group remain understudied.
Methods
We prospectively enrolled preterm neonates (<34 weeks’ gestation) with clinical features of sepsis, with or without sepsis screen positive results. Primary outcome was a composite of death, bronchopulmonary dysplasia (BPD), retinopathy of prematurity requiring treatment, Intraventricular haemorrhage ≥ 2 and periventricular leukomalacia. Details of antimicrobial therapy were also collected.
Results
Over an 18-month period, 172 neonates were enrolled. 104 had CNNS and 68 had culture-positive sepsis (CPNS). Primary outcome was observed in 19 (18.3%) neonates with CNNS and 25 (36.8%) with CPNS, with an adjusted odds ratio (aOR) of 0.50 (95% CI: 0.22–1.12, p = 0.095). Except for BPD, which was significantly lower in CNNS (aOR: 0.10; 95% CI: 0.02–0.52, p = 0.006), there was no statistically significant difference in other outcomes between groups. Multidrug-resistant organisms comprised 67.6% of the gram-negative bacterial isolates. Median (IQR) cumulative duration of antibiotic therapy was 5 (3–7) days in CNNS and 20.5 (15–24.3) days in CPNS. Prolonged cumulative antibiotic use was observed in 50 (48%) CNNS neonates (>5 days) and 50 (73.5%) CPNS neonates (>14 days). In CNNS group, 38 (36.5%) received second-line antibiotics, and 6 (5.7%) received third-line antibiotics.
Conclusion
In preterm neonates, composite outcome of mortality and major morbidities did not differ significantly between those with CNNS and CPNS. However, a considerable proportion of CNNS neonates received a prolonged course of higher antibiotics. Thus, there is a need for strategies to improve clinical outcomes and strengthen adherence to antibiotic stewardship principles.
Keywords: Meningitis, Infection, Newborn, Mortality, Antimicrobial stewardship
Introduction
Neonatal sepsis is a major contributor to neonatal mortality, with prematurity and sepsis leading the cause of neonatal mortality [1]. Preterm neonates born before 34 weeks of gestation require neonatal intensive care (NICU) admission for stabilization and transitional care. These neonates during their NICU stay are at high risk of sepsis, with up to two-thirds of them receiving antibiotics at some point based on clinical suspicion [2]. Among these, nearly 39% are treated for infections diagnosed solely on clinical grounds, without microbiological confirmation [2].
The diagnosis of neonatal sepsis, especially in preterm, presents several challenges. The preterm neonates during the transition and liable clinical phase may have signs that mimic sepsis. To add on, there is no ideal biomarker for early diagnosis. Blood culture, considered the reference standard, has limited sensitivity. The culture positivity is being influenced by the level of bacteraemia, prior antibiotic exposure, methods of isolation and volume of sample collection [3, 4]. Starting antibiotics in these cases often involves conflicting priorities. On one hand, there exists concern that inappropriate antibiotic selection (first line), in light of the significant prevalence of multidrug resistance (MDR), may increase morbidity. On the other hand, there is a compelling need to prevent unnecessary antibiotic use and adhere to the principles of antimicrobial stewardship (AMS). This challenge is pronounced during empirical treatment and in cases where clinical signs suggest sepsis but blood cultures remain negative.
Adherence to antibiotic stewardship, choosing the right drug for the right duration, remains a challenge. Neonates with culture-negative sepsis (CNNS) often receive higher antibiotics and for a prolonged duration, given the ongoing clinical instability, need for central lines or ventilation, unresolved symptoms, or fear of missing infection and lastly, reluctance to adhere to stewardship principles by treating clinicians [5–8]. The presence of culture-positive cases within the same unit can also influence empirical decisions.
The outcomes of CNNS neonates have not been well studied, and it is often assumed that their outcomes are similar to those with no infection. However, reports suggest that neonates with negative cultures have an increased risk of morbidities and neurodevelopmental outcomes compared to those with no infection [9, 10]. Thus, it is essential to focus on these neonates to improve their outcomes and understand the opportunities to improve adherence towards antibiotic stewardship.
Thus, we aimed to compare the clinically relevant outcomes between neonates with CNNS and culture-positive sepsis (CPNS), and to explore the antimicrobial choices and treatment duration in culture-negative sepsis.
Methods
Study design and settings
This single-site prospective study was conducted in a tertiary institute level III NICU, admitting both inborn and outborn neonates. The proportions of outborn neonates constitute around one third of total admissions. Admission criteria include all preterm neonates with gestation less than 34 weeks. The NICU maintains an infant-to-nurse ratio of 3–5 to 1, with personnel, including physicians, having undergone training in infection control practices. The institute also has an infection control team, consisting of a neonatologist, a microbiologist, and hospital administration, to monitor the infection control practices. With an increased proportion of neonates being referred from other hospitals, there is a predominance of MDR, with nearly two-thirds of previous isolates being MDR. The hospital has an established antibiotic stewardship program, led by a microbiologist, prior to commencement of the current study. The study was conducted in accordance with STROBE-NI standards and reported as per STROBE guidelines [11, 12]. The study was approved by the institutional ethics committee with IEC number IMS.SH/SOA/2023/566.
Patient selection and methods
We enrolled consecutive neonates for 18 months from July 2023 to December 2024. All preterm neonates born before 34 weeks with postnatal age more than 72 hours with clinical signs of sepsis, as per NEOKISS criteria [13], were evaluated. Neonates were included if they had clinical features of sepsis with or without a positive sepsis screen (C reactive protein > 10 mg/L, total leucocyte count < 5 × 109 /L, or absolute neutrophil count < 1.8 × 109 /L) and a blood culture sample drawn before antibiotic administration. NEOKISS criteria was met if any 2 of the following were present (without other recognized cause): fever (>38 °C) or temperature instability or hypothermia (<36.5 °C), tachycardia (>200/min), or new/ more frequent bradycardia (<80/min), capillary refill time > 2 seconds, new or more frequent apnea (>20 seconds), unexplained metabolic acidosis, new-onset hyperglycaemia (>140 mg/dl), or other signs/ symptoms of sepsis: (1 or more; skin color (pale/mottled), lethargy, poor feeding, unexplained respiratory distress, increased oxygen requirements or requiring intubation, unstable condition, and apathy) [13]. Neonates were categorized into the CNNS group if culture remained sterile and the CPNS group if culture was positive. Neonates initially categorized as CNNS who later had a positive culture during their hospital stay were reclassified as CPNS. Neonates who received antibiotics prior to admission, suspected or confirmed early onset sepsis and moribund at admission were excluded. Neonates with meningitis or arthritis, conditions requiring surgical intervention, necrotizing enterocolitis (NEC) or congenital malformations were also excluded. Neonates with meningitis, arthritis and NEC were excluded as these conditions require prolonged antibiotic therapy that is often independent of organism isolation. All neonates were followed up till discharge or death.
Data were collected prospectively, including demographic information, comorbidities, respiratory support, antibiotic therapy details (choice of antibiotics, duration and modifications), laboratory test results, and clinically relevant outcomes in a predesigned proforma, which was later transferred to Microsoft excel spreadsheet by a research assistant. All patient information was anonymized and treated with confidentiality.
Microbiological reporting
All blood cultures drawn were transferred to the in-house central laboratory. Nursing staff and doctors were trained to draw blood culture with minimum volume of 1 ml under sterile condition into the BACTEC bottles. All culture bottles were incubated in an automated system for a minimum duration of 5 days until they were labelled as sterile. The BD BACTEC FX 20 (Becton Dickinson, East Rutherford, NJ, USA) was used for blood culture, and the automated VITEK 2 system (BioMerieux, Marcy-l’Étoile, France) was used for antimicrobial susceptibility testing as per Clinical and Laboratory Standards Institute standard guidelines (2025) [14]. MDR was defined as resistance to one or more agents in at least three antibiotic classes [15].
Infection control and treatment protocol
All neonates were managed according to the unit protocol. Neonates were monitored for features of sepsis using NEOKISS criteria before starting antibiotics [13]. The sepsis evaluation included total leucocyte count, absolute neutrophil count and C-reactive protein. Other markers, including erythrocyte sedimentation rate (ESR), procalcitonin and the immature-to-total (IT) neutrophil ratio, were not routinely performed. We performed cerebrospinal fluid analysis on all positive blood cultures, clinical features of meningitis, and suspected late-onset sepsis. Enrolled neonates were initiated on piperacillin-tazobactam and amikacin as the first-line empiric antibiotics. In the CNNS group, antibiotic discontinuation was based on clinical discretion, with a unit protocol that required stopping within 5 days of initiation, provided a negative blood culture report and resolution of symptoms. Persistence of symptoms prompted a repeat blood culture to be performed after 72 hours. For the CPNS group, empirical antibiotics were modified based on antibiotic sensitivity patterns, with documented microbial clearance after 5–7 days of antibiotics, if the isolates were MDR pathogens or fungal isolates.
In neonates with negative cultures or pending culture reports, as per unit’s antibiotic policy, antibiotics were upgraded to meropenem/levofloxacin and gentamicin (second line) if clinical worsening occurred, as evidenced by the need for inotropes, bleeding, or high ventilatory settings. In the presence of a high incidence of gram-negative MDR organisms, upgradation to colistin before culture report was considered if there was a need for two inotropes or the presence of active bleeding. Antibiotics initiation, escalation and de-escalation flow chart is provided in Fig. 1. Nevertheless, the choice of initial antibiotics and upgradation to higher-level antibiotics was based on the clinician’s discretion.
Fig. 1.
Algorithm outlining initiation, escalation, and de-escalation of antibiotic therapy in neonates with suspected late-onset sepsis
Outcome assessment
We included the recommended outcome for neonatal sepsis with the primary outcome of a composite measure including all-cause hospital mortality, bronchopulmonary dysplasia (BPD) (oxygen requirement at 36 weeks postmenstrual age) [16], intraventricular haemorrhage (IVH) grade ≥ 2 [17], retinopathy of prematurity (ROP) requiring treatment [18], and periventricular leukomalacia (PVL) [19]. Secondary outcomes included individual components of composite outcomes, need for invasive and non-invasive ventilation, duration of invasive and non-invasive ventilation, need for inotropes, and duration of hospital stay.
Statistical analysis
Descriptive statistics were used to summarize the collected data. Categorical variables were presented as frequencies and percentages, while continuous variables were reported as mean with standard deviation (SD) or median with interquartile range (IQR), depending on the distribution. Quantitative variables were compared using either Student’s t-test or the Mann–Whitney U test, as appropriate. Median differences (MD) were reported along with 95% confidence intervals (CI). Categorical variables were compared using the Chi-square test or Fisher’s exact test, as applicable. Unadjusted and adjusted odds ratios (OR) with 95% CI were calculated. A logistic regression model was used to adjust odds ratios for potential confounders. For all-cause hospital mortality, adjusted hazard ratios (HR) were estimated using multivariable Cox proportional hazards regression. Time-to-event data were analysed using Kaplan–Meier survival curves. A two-sided p-value of less than 0.05 was considered statistically significant. All analyses were conducted using Jamovi (version 2.5) (Sydney, Australia) and R Core Team (version 4.3) (Vienna, Austria) [20, 21].
Sample size
Based on hospital data from the previous year, the incidence of the composite outcome in culture-positive sepsis was approximately 45%. We assumed a lower incidence of 20% in culture-negative cases. To detect this difference with 80% power and a two-sided alpha of 0.05, and assuming a 2:1 allocation ratio, the required sample size was 173 neonates, including 102 with culture-negative and 71 with culture-positive sepsis.
Results
A total of 520 neonates with a gestational age of less than 34 weeks were admitted to the NICU. Of these, 222 were evaluated for sepsis beyond day 3 of life. Twenty-eight neonates with meningitis (cerebrospinal fluid (CSF) culture or cytology if CSF culture is negative), 3 neonates with suspected osteomyelitis, 12 neonates diagnosed to have NEC and 7 refused to consent were excluded. Thus, a total of 172 neonates were included. Among them, 104 were CNNS, and 68 neonates were labelled as CPNS. All neonates were followed till discharge or death (Fig. 2).
Fig. 2.
Study flow chart
The CNNS group was more mature, with a mean (SD) gestational age of 30.5 (2.2) weeks, compared to 29.7 (2.1) weeks in the CPNS group (p = 0.013). Other maternal and neonatal characteristics were comparable between groups, including birth weight with a mean (SD) of 1296 (356) gm in the CNNS group and 1219 (398) gm in the CPNS group (p = 0.187) (Table 1).
Table 1.
Maternal and neonatal baseline characteristics
| CNNS (n = 104) |
CPNS (n = 68) |
P value | |
|---|---|---|---|
| Gestation (week)* | 30.5 (2.2) | 29.7 (2.1) | 0.013 |
| Birth weight (grams)* | 1296.6 (356.5) | 1219.3 (398.9) | 0.187 |
| Male | 64 (61.5%) | 48 (70.6%) | 0.223 |
| Small for gestational age | 23 (22.1%) | 8 (11.8%) | 0.162 |
| Vaginal delivery | 45 (43.3%) | 33 (48.5%) | 0.498 |
| Out born | 17 (16.3%) | 14 (20.6%) | 0.479 |
| Maternal characteristics | |||
|---|---|---|---|
| Gestational hypertension | 22 (21.2%) | 17 (25%) | 0.556 |
| Gestational Diabetes | 13 (12.5%) | 10 (14.7%) | 0.678 |
| PPROM > 18 hours | 27 (26%) | 15 (22.1%) | 0.560 |
| Multiple gestation | 27 (26%) | 18 (26.5%) | 0.941 |
| Antenatal steroids | 94 (90.4%) | 62 (91.2%) | 0.861 |
| PPV at birth | 34 (32.7%) | 22 (32.4%) | 0.963 |
| Sepsis screen markers | |||
|---|---|---|---|
| C reactive Protein (mg/L) # | 0.026 (0.016–0.127) | 0.972 (0.046–4.2) | < 0.001 |
| Leucocyte Count (x 109L)# | 9.98 (6.86–14.54) | 9.74 (6.77–15.42) | 0.435 |
| Absolute Neutrophil count (x 109L) # | 4.36 (2.40–6.88) | 3.96 (2.04–7.26) | 0.459 |
Abbreviations: CNNS-culture negative neonatal sepsis, CPNS- culture positive neonatal sepsis, PPROM- preterm premature rupture of membrane, PPV- positive pressure ventilation
All values are expressed as n (%) unless specified
*Mean (standard deviation)
# Median (interquartile range)
The composite adverse outcome occurred less frequently in the CNNS group 19 (18.3%) as compared to CPNS group 25 (36.8%) (p = 0.007). However, after adjusting for gestational age, birth weight, gender, and antenatal steroid exposure, the difference was not statistically significant. (adjusted odds ratio (aOR) 0.50; 95% CI: 0.22–1.12, p = 0.095). Individual outcomes such as ROP requiring treatment, BPD, IVH and PVL were more frequent in the CPNS group, but after adjustment, only the risk for BPD remained higher in the CPNS group (aOR 0.10; 95% CI: 0.02–0.52, p = 0.006) (Table 2). Use of invasive ventilation was lower in the CNNS group 29 (27.9%) as compared with CPNS group 35 (51.5%) (OR 0.36; 95% CI: 0.19–0.69, p = 0.002). Similarly, fewer neonates in the CNNS group required non-invasive ventilation [84 (80.8%) vs. 62 (91.2%) (OR 0.40; 95% CI: 0.15–1.07, p = 0.062)]. The need for inotropic support (OR 0.57; 95% CI: 0.30–1.10, p = 0.094) and for PDA needing treatment (OR 0.90; 95% CI: 0.27–2.99, p = 0.876) did not differ significantly between groups (Table 2).
Table 2.
Primary and secondary outcomes of neonates in culture positive and culture negative group
| Total (n = 172) |
CNNS (n = 104) |
CPNS (n = 68) |
Unadjusted OR/MD (95% CI) |
p value |
Adjusted OR† (95% CI) |
P value | |
|---|---|---|---|---|---|---|---|
| Composite outcome* | 44 (25.6%) | 19 (18.3%) | 25 (36.8%) |
0.38 (0.19, 0.77) |
0.007 |
0.50 (0.22–1.12) |
0.095 |
| Invasive Ventilation | 64 (61.5%) | 29 (27.9%) | 35 (51.5%) |
0.36 (0.19, 0.69) |
0.002 | - | - |
| Duration of invasive ventilation# | 0 (0–2) | 0 (0–1) | 1 (0–4) | 0 (−1, 0) | 0.001 | - | - |
| Non-Invasive Ventilation | 146 (84.9%) | 84 (80.8%) | 62 (91.2%) |
0.40 (0.15, 1.07) |
0.062 | - | - |
| • CPAP | 135 (78.5%) | 74 (71.2%) | 61 (89.7%) |
0.28 (0.11, 0.68) |
0.004 | - | - |
| • NIMV | 69 (40.1%) | 28 (26.9%) | 41 (60.3%) |
0.24 (0.12, 0.46) |
<0.001 | - | - |
| • HHFNC | 90 (52.3%) | 46 (44.2%) | 44 (64.7%) |
0.43 (0.23, 0.81) |
0.009 | - | - |
| Duration of non-invasive ventilation # | 7 (2–25) | 4 (1–11) | 14.5 (5–48) |
− 8.9 (−15, − 4.9) |
<0.001 | - | - |
| Need for Inotropes | 58 (33.7%) | 30 (28.8%) | 28 (41.2%) |
0.57 (0.30, 1.10) |
0.094 | - | - |
| Patent Ductus Arteriosus | 12 (7%) | 7 (6.7%) | 5 (7.4%) |
0.909 (0.276–2.99) |
0.876 | - | - |
| Retinopathy of prematurity | 13 (7.6%) | 5 (4.8%) | 8 (11.8%) |
0.37 (0.11, 1.21) |
0.091 |
0.47 (0.13, 1.65) |
0.243 |
| Bronchopulmonary Dysplasia | 15 (8.7%) | 3 (2.9%) | 12 (17.6%) |
0.13 (0.03, 0.51) |
0.001 |
0.10 (0.02, 0.52) |
0.006 |
| Intraventricular haemorrhage > 2 | 8 (4.7%) | 4 (3.8%) | 4 (5.9%) |
0.64 (0.15, 2.65) |
0.714 |
0.86 (0.20, 3.77) |
0.893 |
| Periventricular Leukomalacia | 2 (1.2%) | 0 (0%) | 2 (2.9%) |
0.12 (0.006, 2.69) |
0.155 | - | - |
| Duration of Hospital stay# | 34 (20–58.8) | 29 (16–47) | 51.5 (26.3–72) |
−17 (−26, − 8.9) |
<0.001 | - | - |
| Death | 18 (10.5%) | 12 (11.5%) | 6 (8.8%) |
1.35 (0.48, 3.78) |
0.570 |
1.57 (0.52, 4.74) |
0.383 |
| • < 28 weeks | 6 (3.4%) | 4 (3.8%) | 2 (2.9%) |
2.57 (0.41, 16.1) |
0.395 | - | - |
| • 28 - 32 weeks | 8 (4.6%) | 5 (4.8%) | 3 (4.4%) |
1.22 (0.27, 5.46) |
1.0 | - | - |
| • > 32 weeks | 4 (2.3%) | 3 (2.8%) | 1 (1.5%) | 0.82(0.07, 9.07) | 1.0 | - | - |
Abbreviations: CNNS-culture negative neonatal sepsis, CPNS- culture positive neonatal sepsis, OR odds ratio, MD- median difference, CI- confidence interval, CPAP- continuous positive airway pressure, NIMV -non-invasive mechanical ventilation, HHFNC - humidified high flow nasal cannula
*Composite outcome: Death, bronchopulmonary dysplasia, intraventricular haemorrhage ≥ 2, retinopathy of prematurity requiring treatment, periventricular leukomalacia
† adjusted for gestational age, birth weight, gender, and antenatal steroids
All values are expressed in n (%) until otherwise specified
# Median (interquartile range)
In the CPNS group, 73 bacterial isolates were identified from 63 neonates, the most common being Klebsiella pneumoniae 33 (45.3%), Acinetobacter species 11 (14.7%), and Escherichia coli 7 (9.3%). MDR organisms were detected in 67.6% (22/68) of gram-negative bacterial isolates. Among the common pathogens, MDR isolates for Klebsiella pneumoniae were 23/33 (69.6%), Acinetobacter species 4/11 (36.3%), and Escherichia coli 6/7 (85.7%). (Fig. 3) In 17 (25%) CPNS cases, new onset culture-positive neonatal sepsis was documented. Candida species were isolated in 10 neonates (Fig. 3).
Fig. 3.
Organisms isolated from the blood culture. A. Bacterial isolates. B. Fungal isolates
All neonates in both groups received antibiotics. The median (IQR) cumulative duration of antibiotic therapy, defined as the total number of calendar days during which the neonate received at least one antibiotic, was 5 (3–7) days in the CNNS group and 20.5 (15–24.3) days in the CPNS group. In the CNNS group, the number of neonates receiving second and third line antibiotics, combination antibiotics has been depicted in Table 3. Antibiotic therapy was restarted in 7 (6.7%) CNNS cases and 24 (35.3%) CPNS cases. Nearly 50 (48%) of neonates in the CNNS group received cumulative antibiotics for more than 5 days, while 50 (73.5%) in the CPNS group received cumulative antibiotics for more than 14 days. (Table 3) On multivariable Cox regression analysis adjusted for gestational age, birth weight, sex, and antenatal steroid exposure, there was no significant difference in mortality risk between the groups (HR: 2.07; 95% CI: 0.70–6.08, p = 0.187) (Fig. 4).
Table 3.
Microbial isolates and antimicrobial use in culture negative and culture positive group
| CNNS (n = 104) |
CPNS (n = 68) |
|
|---|---|---|
| MDR isolates (Gram negative pathogens) | - | 46 (67.6%) |
| ≥ 2 Organisms isolated | - | 17 (25%) |
| Antibiotics | ||
|---|---|---|
| Piperacillin and Tazobactam | 95 (91.3%) | 55 (80.9%) |
| Amikacin | 86 (82.7%) | 44 (64.7%) |
| Gentamicin | 37 (35.6%) | 52 (76.5%) |
| Meropenem | 16 (15.4%) | 46 (67.6%) |
| Levofloxacin | 16 (15.4%) | 24 (35.3%) |
| Colistin | 9 (8.7%) | 23 (33.8%) |
| Tigecycline | 1 (1.0%) | 13 (19.1%) |
| Vancomycin | 1 (1.0%) | 7 (10.3%) |
| Ciprofloxacin | 1 (1.0%) | 3 (4.4%) |
| Piperacillin and tazobactam + Amikacin | 83 (79.8%) | 40 (58.8%) |
| Piperacillin and tazobactam + Gentamicin | 1 (1%) | 6 (8.8%) |
| Meropenem + Amikacin | 13 (12.5%) | 30 (44.1%) |
| Meropenem + Gentamicin | 13 (12.5%) | 30 (44.1%) |
| Meropenem + Colistin | 3 (2.9%) | 16 (23.5%) |
| Colistin + Gentamicin | 3 (2.9%) | 13 (19.1%) |
| Levofloxacin + Gentamicin | 12 (11.5%) | 22 (32.4%) |
| Colistin + Tigecycline | 0 (0%) | 10 (14.7%) |
| First line antibiotics | 83 (79.8%) | 40 (58.8%) |
| Second line antibiotics | 38 (36.5%) | 52 (76.4%) |
| Third line antibiotics | 6 (5.7%) | 36 (52.9%) |
| Restarting of Antibiotics | 7 (6.7%) | 24 (35.3%) |
| Median cumulative antibiotics days | 5 (3–7) | 20.5 (15–24.3) |
| > 5 days of antibiotics | 50 (48%) | - |
| >14 days of antibiotics | - | 50 (73.5%) |
| Amphotericin B | 0 (0%) | 14 (20.6%) |
| Caspofungin | 0 (0%) | 2 (2.9%) |
| Antifungal Duration | - | 14 (5.5–21) |
All values are expressed as n (%) unless specified
# Median (interquartile range)
Abbreviations: CNNS-culture negative neonatal sepsis, CPNS- culture positive neonatal sepsis, MDR multidrug resistance
Fig. 4.
Kaplan-Meier for all-cause mortality in culture negative and culture positive group
Discussion
Our single-center prospective study evaluated clinically relevant outcomes among preterm neonates born before 34 weeks of gestation, comparing culture negative sepsis to those with culture positive sepsis. As anticipated, the CPNS group had a higher frequency of composite adverse outcomes. However, after adjusting for gestational age and other confounding factors, the risk of composite outcomes did not differ significantly between groups, suggesting that neonates with CNNS are also at increased risk for the adverse outcomes. Among the secondary outcomes, the BPD rates were higher in the CPNS group. However, the incidence of mortality, IVH ≥ grade 2, PVL, and ROP requiring treatment showed no significant differences between the groups.
Before the initiation of the study, we hypothesised that neonates may experience more adverse outcomes, although fewer than those with culture-positive sepsis. We anticipated that adherence to antibiotic stewardship would be poorer in the CPNS group but not in the CNNS group. We observed an increased proportion of NICU admissions (20%) with gestation less than 34 weeks were labelled as CNNS. This increased proportion of CNNS is likely due to increased admissions of outborn neonates and failure to isolate high-risk cases. Although factors such as prior antibiotic exposure and sample transport were controlled, the blood volume (1 mL) may still have influenced culture positivity. Thus, sample volume could have contributed to false-negative results, potentially inflating cases classified as CNNS. Lastly, our evaluation did not include viral assays, which may mimic bacterial sepsis. Viral infections occur in about 1% of NICU admissions, mostly community-acquired [22]. However, our study reflects a resource limited setting where viral assays by culture or PCR are seldom practiced.
The overall mortality rate (9.5%), BPD (8.7%), ROP requiring treatment (7.5%), and IVH ≥ grade 2 (4.6%) observed in our cohort were much lower than those reported in similar settings [23, 24]. In a retrospective study by Jiang et al., the composite outcome rates were 24.1% in the CNNS group and 30.1% in the CPNS group. Similarly, our study found lower composite outcomes in the CNNS group (18.3%) but higher rates in the CPNS group (36.8%). However, Jiang et al. reported significantly higher adverse outcomes in CPNS even after adjustment, whereas we found no significant differences. No differences observed in our study may be attributed to our exclusion of neonates with NEC and meningitis in the CPNS group. Although most adverse outcomes were more frequent in the CPNS group in our study, only BPD remained significantly higher after adjustment [9]. Similar outcomes were also observed in high resource settings, but with a lower gestation below 28 weeks, where it was found that confirmed sepsis had higher mortality, whereas other outcomes like BPD, ROP and PVL were similar in culture positive and culture negative group [10].
Adherence to AMS remains a significant challenge in the antibiotic therapy for preterm neonates. The common belief that empiric antibiotic therapy benefits all neonates with suspected infection, even in the absence of positive culture, has been questioned [8]. Unnecessary antibiotic exposure and prolonged antibiotic (>5 days) courses can have serious consequences, including the emergence of MDR, fungal infections, NEC, and increased risk of adverse clinical outcomes [25–27]. Current guidelines for CNNS vary widely, recommending durations of antibiotic therapy ranging from 36 hours to 7 days, with stratification of neonates based on symptom resolution, initial clinical risk, and negative culture results. However, due to the limited number of studies done on optimal antibiotic duration in this population, these recommendations remain inconsistent [28]. Previous studies have defined antibiotic therapy exceeding 5 days as prolonged, while a 5-day course is generally regarded as appropriate for promoting antimicrobial stewardship [29–31]. Our study findings indicate that although the median duration of antibiotics in CNNS was indeed 5 days, nearly 48% of neonates received therapy beyond this period. The concept of automatic cessation of antibiotics after a fixed duration may not be suitable for preterm neonates. Prolonged antibiotic use in our cohort was primarily driven by persistent clinical symptoms, ongoing systemic illness, laboratory abnormalities, and an overall clinical picture strongly suggestive of sepsis. Additionally, clinician hesitation in discontinuing antibiotics likely contributed to the extended duration. Nevertheless, a similar challenge was also faced in achieving agreement among providers to shorten the duration of antibiotic therapy, with nearly 38% still continued to receive a prolonged antibiotic course (>5 days) [30].
Another area of concern is escalation of antibiotics. Ideally, neonates with CNNS should receive first-line empirical antibiotics as per unit protocol. However, in our cohort, antibiotics were escalated in 42% of neonates, with second-line in 36.5% and third-line in 5.7%. Escalation typically occurred within the first 48 hours of antibiotic initiation, prompted by clinical deterioration or the presence of other culture-positive cases in the unit. Similarly, a significant proportion of culture-positive neonates (73.5%) received antibiotics for more than 14 days, often due to escalation from first-line regimens or recurrent sepsis episodes requiring re-initiation of therapy, which was observed in one third of the CPNS cohort. These findings reflect those of a multicentric study that reported deviations from standard protocols (including WHO recommendations) and frequent antibiotic switching [32].
Several factors contribute to the nonadherence to the AMS principles in preterm neonates. Firstly, the absence of reliable biomarkers to clearly distinguish between sepsis and non-sepsis cases, and the lack of consensus on the definition of culture negative sepsis, pose major challenges. Furthermore, there is limited evidence to guide biomarker based antibiotic duration and the role of discontinuation algorithms, including automated stop orders [33, 34]. Nonetheless, key AMS strategies include diagnostic stewardship, which ensures aseptic sample collection, adequate blood volume, and timely transport to the laboratory, and therapeutic stewardship, which involves structured AMS interventions such as 48 to 72 hour antibiotic reviews, regular audits, adherence to updated unit protocols, continuous clinician education, establishment of dedicated neonatal AMS teams, and ongoing surveillance integrated with quality improvement initiatives [34–36].
Our study has several strengths. As a prospective study with clearly defined clinical criteria to identify neonates with sepsis, and by excluding neonates who had received prior antibiotic therapy, we improved patient selection. We adhered to the recommended core outcomes for neonatal sepsis, including microbiological data, resource utilization, and details of antibiotic therapy, all of which have substantial clinical relevance. We adhered to standard reporting guidelines for neonatal infection. Nevertheless, the study has several limitations. Being a single-center study conducted in a setting with a high prevalence of MDR organisms and exclusion of neonates with suspected early-onset sepsis, outborn neonates who received prior antibiotics, NEC and meningitis may limit the representativeness of the culture-negative sepsis cohort and introduce inclusion bias and limit the generalizability of the findings. Although we had an established unit antibiotic protocol, the treatment of neonates was still based on clinician discretion, resulting in variability in treatment. Nonetheless, this reflects real-world, pragmatic clinical practices. As an observational study, our findings cannot establish causality between adverse outcomes and either CNNS or antibiotic duration. In addition, we did not achieve our initially estimated sample size, which may have affected the statistical power. The limited sample size also precluded analysis of factors associated with prolonged antibiotic duration in CNNS. Finally, our sepsis evaluation did not include viral testing, which might have identified non-bacterial causes of sepsis presentations.
Conclusion
In culture negative sepsis among preterm neonates with gestation less than 34 weeks, CNNS was associated with adverse composite outcomes that were not statistically different from those observed in CPNS. These findings highlight a substantial healthcare burden posed by CNNS, particularly in resource-limited settings, and emphasize the need for targeted antimicrobial stewardship strategies and improved diagnostic precision to strengthen adherence to stewardship.
Acknowledgment
We thank Priyabrata Singh, Research Assistant, for providing assistance with data collection and entry.
Authors contributions
GR conceptualized and designed the study, collected data, conducted data analysis, and drafted the initial manuscript. RS conceptualized and designed the study, supervised the study, and critically revised the manuscript. RD , PJ, and DN provided clinical input, supported data interpretation, and reviewed the manuscript. VK conceptualized and designed the study, collected data, supervised the study, conducted data analysis, and revised the initial draft. All authors reviewed the manuscript.
Funding
Open access funding provided by Siksha 'O' Anusandhan (Deemed To Be University)
Data availability
Data will be made available on suitable request.
Declarations
Ethics approval
The study was performed in line with the principles of the Declaration of Helsinki and was approved by the Institutional Ethics Committee (Institute of Medical Sciences (IMS) and SUM hospital) (reference number 2023/566).
Consent to participate
Informed consent was obtained from all patient caregivers before inclusion in the study and after assuring confidentiality.
Consent for publication
Not applicable
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Reference
- 1.Milton R, Gillespie D, Dyer C, Taiyari K, Carvalho MJ, Thomson K, Sands K, Portal EAR, Hood K, Ferreira A, et al. Neonatal sepsis and mortality in low-income and middle-income countries from a facility-based birth cohort: an international multisite prospective observational study. Lancet Glob Health 2022;10:e661–e672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Stoll BJ, Hansen NI, Adams-Chapman I, Fanaroff AA, Hintz SR, Vohr B, Higgins RD.National institute of child health and human development neonatal research network ft: neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection. JAMA. 2004;292(19):2357–65. [DOI] [PubMed] [Google Scholar]
- 3.Piantino JH, Schreiber MD, Alexander K, Hageman J.Culture negative sepsis and systemic inflammatory response syndrome in neonates. NeoReviews. 2013;14(6):e294–e305. [Google Scholar]
- 4.Mukhopadhyay S, Briker SM, Flannery DD, Dhudasia MB, Coggins SA, Woodford E, Walsh EM, Li S, Puopolo KM, Kuzniewicz MW.Time to positivity of blood cultures in neonatal late-onset bacteraemia. Arch Dis Child Fetal Neonatal Ed. 2022;107(6):583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Rueda MS, Calderon-Anyosa R, Gonzales J, Turin CG, Zea-Vera A, Zegarra J, Bellomo S, Cam L, Castaneda A, Ochoa TJ.Antibiotic overuse in premature low birth weight infants in a developing country. Pediatr Infect Dis J. 2019;38(3):302–07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Huncikova Z, Stensvold HJ, Øymar KAA, Vatne A, Lang AM, Støen R, Brigtsen AK, Moster D, Eriksen BH, Selberg T, et al. Variation in antibiotic consumption in very preterm infants—a 10 year population-based study. J Antimicrob Chemother 2023;79:143–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Thampi N, Shah PS, Nelson S, Agarwal A, Steinberg M, Diambomba Y, Morris AM.Prospective audit and feedback on antibiotic use in neonatal intensive care: a retrospective cohort study. BMC Pediatr. 2019;19(1):105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mukhopadhyay S, Sengupta S, Puopolo KM.Challenges and opportunities for antibiotic stewardship among preterm infants. Arch Dis Child Fetal Neonatal Ed. 2019;104(3):F327–f332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Jiang S, Yang Z, Shan R, Zhang Y, Yan W, Yang Y, Shah PS, Lee SK, Cao Y.Neonatal outcomes following culture-negative late-onset sepsis among preterm infants. Pediatr Infect Dis J. 2020;39(3):232–38. [DOI] [PubMed] [Google Scholar]
- 10.Mukhopadhyay S, Puopolo KM, Hansen NI, Lorch SA, DeMauro SB, Greenberg RG, Cotten CM, Sanchez PJ, Bell EF, Eichenwald EC, et al. Neurodevelopmental outcomes following neonatal late-onset sepsis and blood culture-negative conditions. Arch Dis Child Fetal Neonatal Ed 2021;106:467–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fitchett EJA, Seale AC, Vergnano S, Sharland M, Heath PT, Saha SK, Agarwal R, Ayede AI, Bhutta ZA, Black R, et al. Strengthening the reporting of observational studies in epidemiology for newborn infection (STROBE-NI): an extension of the STROBE statement for neonatal infection research. Lancet Infect Dis 2016;16:e202–e213. [DOI] [PubMed] [Google Scholar]
- 12.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP.The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007;147(8):573–77. [DOI] [PubMed] [Google Scholar]
- 13.Gastmeier P, Geffers C, Schwab F, Fitzner J, Obladen M, Rüden H.Development of a surveillance system for nosocomial infections: the component for neonatal intensive care units in Germany. J Hosp Infect. 2004;57(2):126–31. [DOI] [PubMed] [Google Scholar]
- 14.Lewis JS II, Bobenchik AM, Bryson AL, Campeau S, Cullen SK, Dingle T, Esparza G, Humphries RM. Performance standards for antimicrobial susceptibility testing (CLSI M100). 22nd ed. Wayne (PA): Clinical and Laboratory Standards Institute; 2022.
- 15.Magiorakos AP, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CG, Harbarth S, Hindler JF, Kahlmeter G, Olsson-Liljequist B, et al.. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect 2012;18:268–81. [DOI] [PubMed] [Google Scholar]
- 16.Jensen EA, Dysart K, Gantz MG, McDonald S, Bamat NA, Keszler M, Kirpalani H, Laughon MM, Poindexter BB, Duncan AF, et al. The diagnosis of bronchopulmonary dysplasia in very preterm infants. An evidence-based approach. Am J Respir Crit Care Med 2019;200:751–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.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. [DOI] [PubMed] [Google Scholar]
- 18.Shukla R, Murthy GVS, Gilbert C, Vidyadhar B, Mukpalkar S.Operational guidelines for ROP in India: a summary. Indian J Ophthalmol. 2020;68(Suppl 1):S108–s114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.de Vries LS, Eken P, Dubowitz LM.The spectrum of leukomalacia using cranial ultrasound. Behav Brain Res. 1992;49(1):1–6. [DOI] [PubMed] [Google Scholar]
- 20.The jamovi project. In: jamovi. 2.5 edn; 2024.
- 21.R Core Team. R: a language and environment for statistical computing_. R foundation for statistical computing. In 2023.
- 22.Verboon-Maciolek MA, Krediet TG, Gerards LJ, Fleer A, van Loon TM.Clinical and epidemiologic characteristics of viral infections in a neonatal intensive care unit during a 12-year period. Pediatr Infect Dis J. 2005;24(10):901–04. [DOI] [PubMed] [Google Scholar]
- 23.Singh J, Dalal P, Gathwala G.Clinical profile and predictors of mortality among the referred neonates at a tertiary care centre in north India: a prospective observational study. Trop Doct. 2020;50(3):221–27. [DOI] [PubMed] [Google Scholar]
- 24.Ramaswamy VV, Abiramalatha T, Bandyopadhyay T, Shaik NB, Bandiya P, Nanda D, Pullattayil SA, Murki S, Roehr CC. ELBW and ELGAN outcomes in developing nations: systematic review and meta-analysis. PLoS One. 2021;16(8):e0255352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kuppala VS, Meinzen-Derr J, Morrow AL, Schibler KR.Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants. J Pediatr. 2011;159(5):720–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ting JY, Roberts A, Sherlock R, Ojah C, Cieslak Z, Dunn M, Barrington K, Yoon EW, Shah PS. Duration of initial empirical antibiotic therapy and outcomes in very low birth weight infants. Pediatrics. 2019;143(3). [DOI] [PubMed]
- 27.Cotten CM.Adverse consequences of neonatal antibiotic exposure. Curr Opin Pediatr. 2016;28(2):141–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Neonatal infection: antibiotics for prevention and treatment.
- 29.Cotten CM, Taylor S, Stoll B, Goldberg RN, Hansen NI, Sánchez PJ, Ambalavanan N, Benjamin Jr DK.Prolonged duration of initial empirical antibiotic treatment is associated with increased rates of necrotizing enterocolitis and death for extremely low birth weight infants. Pediatrics. 2009;123(1):58–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Cantey JB, Wozniak PS, Pruszynski JE, Sánchez PJ.Reducing unnecessary antibiotic use in the neonatal intensive care unit (SCOUT): a prospective interrupted time-series study. Lancet Infect Dis. 2016;16(10):1178–84. [DOI] [PubMed] [Google Scholar]
- 31.Lars G, Simon L, Anders E, Elisabet H, Marie S.Reduced antibiotic use in extremely preterm infants with an antimicrobial stewardship intervention. BMJ Paediatr Open. 2020;4(1):e000872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Russell NJ, Stöhr W, Plakkal N, Cook A, Berkley JA, Adhisivam B, Agarwal R, Ahmed NU, Balasegaram M, Ballot D, 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:e1004179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Van Wyk L, Lloyd LG. Biomarker-guided antibiotic stewardship in neonatology: time for implementation? Pediatr Res. 2025;97(6). [DOI] [PMC free article] [PubMed]
- 34.Cantey JB, Prusakov P. A proposed framework for the clinical management of neonatal “Culture-Negative” sepsis. J Pediatr. 2022;244:203–11. [DOI] [PubMed] [Google Scholar]
- 35.Jain M, Bang A, Meshram P, Gawande P, Kawhale K, Kamble P, Deotale V, Datta V, Dhanireddy R. Institution of an antibiotic stewardship programme for rationalising antibiotic usage: a quality improvement project in the NICU of a public teaching hospital in rural central India. BMJ Open Qual. 2021;10(Suppl 1). [DOI] [PMC free article] [PubMed]
- 36.Mascarenhas D, Ho MSP, Ting J, Shah PS. Antimicrobial stewardship programs in neonates: a meta-analysis. Pediatrics. 2024;153(6). [DOI] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data will be made available on suitable request.