This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Accepted Manuscript
American Journal of Perinatology
Contribution of Concurrent Comorbidities to Sepsis-Related Mortality in Pre- term Infants ≤ 32 Weeks of Gestation at an Academic Neonatal Intensive Care Network
Brian W Barnette, Benjamin T Schumacher, Richard F Armenta, James L Wynn, Andrew Richardson, John S Bradley, Sarah Lazar, Shelley M Lawrence.
Affiliations below.
DOI: 10.1055/a-1675-2899
Please cite this article as: Barnette B W, Schumacher B T, Armenta R F et al. Contribution of Concurrent Comorbidities to Sepsis-Re- lated Mortality in Preterm Infants ≤ 32 Weeks of Gestation at an Academic Neonatal Intensive Care Network . American Journal of Perinatology 2021. doi: 10.1055/a-1675-2899
Conflict of Interest: The authors declare that they have no conflict of interest.
This study was supported by Division of Intramural Research, National Institute of Allergy and Infectious Diseases (http://dx.doi.
org/10.13039/100006492), R01AI134982, National Institute of General Medical Sciences (http://dx.doi.org/10.13039/100000057), R01GM128452, Eunice Kennedy Shriver National Institute of Child Health and Human Development (http://dx.doi.
org/10.13039/100009633), R01HD089939,R01HD095547-03,R01HD097081,R01HD099250, National Institute of Biomedical Imaging and Bioengineering (http://dx.doi.org/10.13039/100000070), R43EB029863
Abstract:
Objective: The lack of a consensus definition for neonatal sepsis may complicate the accurate calculation of sepsis-related mortality in infants ≤ 32 weeks of gestation. This study evaluates whether concurrent major comorbidities influenced neonatal sepsis-related mortality in this patient population following a diagnosis of bacteremia or blood culture-negative sepsis.
Study Design: This is a retrospective chart review of infants ≤ 32 weeks of gestation, who were admitted to a single academic network of multiple neonatal intensive care units between January 1, 2012 and December 31, 2015, to determine if concur- rent co-morbidities contributed to bacteremia or blood culture-negative sepsis- related morality. Direct comparisons between early-onset sepsis (EOS; ≤ 72 hours) and late-onset sepsis (LOS; > 72 hours) were made.
Results: In our study cohort of 939 total patients ≤ 32 weeks of gestation, 182 infants were diagnosed with 198 episodes of sep- sis and 7.7% (14/182) died. Mortality rates did not significantly differ between neonates with bacteremia or blood culture-ne- gative sepsis (7/14 each group), and those diagnosed with EOS compared with LOS (6/14 vs. 8/14). Nearly 80% (11/14) of infants were transitioned to comfort care prior to their death secondary to a coinciding diagnosis of severe grade 3 or 4 intraventricu- lar hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, and/or intestinal perforation.
Conclusion: Those with sepsis-related mortality had pre-existing comorbidities that are commonly associated with extreme preterm birth. The contribution of comorbidities to sepsis-related mortality should be considered in future investigations de- signed to evaluate the efficacy of therapeutics and/or technologies that target sepsis-mediated pathways.
Corresponding Author:
Shelley M Lawrence, University of California San Diego, Pediatrics, La Jolla, United States, shelley.m.lawrence@hsc.utah.edu, babydoc- lawrence@hotmail.com
Affiliations:
Submission Date: 2021-04-27 Accepted Date: 2021-09-22 Publication Date:
Accepted Manuscript
This article is protected by copyright. All rights reserved.
2021-10-21
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Brian W Barnette, Rady Children‘s Hospital San Diego, Pediatrics, San Diego, United States
Brian W Barnette, University of California San Diego Health Sciences, Pediatrics, La Jolla, United States
Benjamin T Schumacher, University of California San Diego Health Sciences, Herbert Wertheim School of Public Health and Longevity Science, La Jolla, United States
[…]
Shelley M Lawrence, Rady Children‘s Hospital San Diego, Pediatrics, San Diego, United States
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Contribution of Concurrent Comorbidities to Sepsis-Related Mortality in Preterm Infants
≤ 32 Weeks of Gestation at an Academic Neonatal Intensive Care Network
Brian W. Barnette1, Benjamin T. Schumacher 2, Richard F. Armenta3, James L. Wynn 4,5, Andrew Richardson6, John S. Bradley7, Sarah Lazar1, and Shelley M. Lawrence1,8,*
Authors’ Information:
1 University of California, San Diego, College of Medicine, Department of Pediatrics, Division of Neonatal-Perinatal Medicine, San Diego, CA, USA
2 Herbert Wertheim School of Public Health and Longevity Science, UC San Diego School of Medicine, San Diego, CA, USA
3 California State University, San Marco, Department of Kinesiology, College of Education, Health, and Human Services, San Diego, CA, USA
4 University of Florida, College of Medicine, Department of Pediatrics, Division of Neonatal-Perinatal Medicine, Gainesville, FL, USA
5 University of Florida, Department of Pathology, Immunology, and Laboratory Medicine, Gainesville, FL, USA
6 Rady Children’s Hospital San Diego, San Diego, Clinical Research Informatics, San Diego, CA, USA
7 University of California, San Diego, College of Medicine, Department of Pediatrics, Division of Infectious Disease, San Diego, CA, USA
8 University of California, San Diego, Department of Pediatrics, Division of Host-Microbe Systems and Therapeutics, San Diego, CA, USA
* Corresponding Author: Shelley M. Lawrence, MD, MS. Associate Professor, Division of Perinatal-Neonatal Medicine, Department of Pediatrics, The University of California, San Diego;
9500 Gilman Drive, MC0760, La Jolla, California, USA 92093-0760; Phone: (858) 249-1704;
Fax: (858) 966-7483. Email: : slawrence@health.ucsd.edu Brain Barnette: bwbarnette14@gmail.com
Benjamin Schumacher: benschumacher12@gmail.com Richard Armenta: rarmenta@csusm.edu
James Wynn: james.wynn@peds.ufl.edu Andrew Richardson: acrichardson@rchsd.org John Bradley: jsbradley@health.ucsd.edu Sarah Lazar: slazar@health.ucsd.edu
Keywords: Neonatal sepsis, sepsis-related mortality; prematurity; bacteremia; blood culture negative sepsis, comorbidities; comfort care
Funding
SML: R01AI134982, R01HD099250 JSB: R01HD095547-03
JLW: R01GM128452, R01HD089939, R01HD097081, R43EB029863
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Abstract 1
2
Objective: The lack of a consensus definition for neonatal sepsis may complicate the accurate 3
calculation of sepsis-related mortality in infants ≤ 32 weeks of gestation. This study evaluates 4
whether concurrent major comorbidities influenced neonatal sepsis-related mortality in this patient 5
population following a diagnosis of bacteremia or blood culture-negative sepsis.
6
Study Design: This is a retrospective chart review of infants ≤ 32 weeks of gestation, who were 7
admitted to a single academic network of multiple neonatal intensive care units between January 8
1, 2012 and December 31, 2015, to determine if concurrent co-morbidities contributed to 9
bacteremia or blood culture-negative sepsis- related morality. Direct comparisons between early- 10
onset sepsis (EOS; ≤ 72 hours) and late-onset sepsis (LOS; > 72 hours) were made.
11
Results: In our study cohort of 939 total patients ≤ 32 weeks of gestation, 182 infants were 12
diagnosed with 198 episodes of sepsis and 7.7% (14/182) died. Mortality rates did not significantly 13
differ between neonates with bacteremia or blood culture-negative sepsis (7/14 each group), and 14
those diagnosed with EOS compared with LOS (6/14 vs. 8/14). Nearly 80% (11/14) of infants 15
were transitioned to comfort care prior to their death secondary to a coinciding diagnosis of severe 16
grade 3 or 4 intraventricular hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, 17
and/or intestinal perforation.
18
Conclusion: Those with sepsis-related mortality had pre-existing comorbidities that are commonly 19
associated with extreme preterm birth. The contribution of comorbidities to sepsis-related 20
mortality should be considered in future investigations designed to evaluate the efficacy of 21
therapeutics and/or technologies that target sepsis-mediated pathways.
22
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Key Points:
23
A. Concurrent co-morbidities significantly contribute to, and may artificially inflate, sepsis- 24
related mortality in preterm infants.
25
B. The absence of a consensus definition for neonatal sepsis complicates the investigation of 26
infection and the precise determination of sepsis-related mortality.
27
C. Accurate assessment of the incidence of sepsis in very low birth weight infants is vital for 28
future investigations of therapeutics and/or technologies that target sepsis-mediated 29
pathways.
30
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Introduction 31
32
Neonates are at considerable risk for developing life-threatening infections with 33
aggressive, virulent organisms due to gestational age-related functional deficiencies of their innate 34
and adaptive immune responses1-3. Worldwide, an estimated 13-15% of all neonatal deaths are 35
attributed to sepsis and 42% of losses are experienced within the first week of life due to 36
complications associated with early-onset sepsis (EOS; sepsis occurring ≤ 72 h of life)4, 5. In the 37
United States, sepsis is the fifth leading cause of neonatal mortality and is surpassed only by 38
complications related to pregnancy, preterm birth, and congenital anomalies6. 39
To date, no consensus definition for neonatal sepsis has been established. Historically, 40
bacteremia has been considered the “Gold Standard” for the diagnosis of neonatal sepsis (sepsis 41
diagnosed ≤28 days of life), but up to 14% of newborns with unequivocal infection documented 42
at autopsy reported negative premortem blood cultures7. Blood culture-negative neonatal sepsis 43
leading to life-threatening organ dysfunction may occur when the primary site of infection includes 44
meningitis, soft-tissue (omphalitis), osteomyelitis, pneumonia, and perforated bowel [e.g., 45
spontaneous intestinal perforation (SIP) or necrotizing enterocolitis (NEC)]8-11. In addition, blood 46
culture-negative sepsis is diagnosed in more than half of adult and pediatric patients presenting 47
with clinical signs and symptoms of septic shock11-13. 48
While the incidence of neonatal sepsis in the US is reported to be 0.9-1.5 per 1000 live 49
births for EOS and 3.0-3.7 per 1000 live births for late-onset sepsis (LOS; > 72 hours of life)6, the 50
diagnosis of infection is inversely proportional to gestational age and weight at birth1. Thus, the 51
least mature, smallest, and most critically-ill infants have the highest incidence of neonatal sepsis1. 52
Infection risks are heightened in very low birth weight (VLBW; birth weight < 1500 grams) infants 53
Accepted Manuscript
This article is protected by copyright. All rights reserved.
because of the need for specialized intensive care supportive measures. Common interventions and 54
procedures that are vital for sustaining developmentally immature organ systems, including 55
prolonged placement of indwelling central lines, frequent laboratory tests, and non-invasive 56
respiratory support or mechanical ventilation, may also promote sepsis-related morbidity and lead 57
to protracted hospital stays14-17. Consequently, the mortality rate for VLBW infants compared with 58
their term counterparts is significantly greater (30%-35% for EOS and 18% for LOS compared to 59
an overall upper limit of 3% mortality for term infants)1, 18-20. 60
The development of neonatal co-morbidities, such as severe intraventricular hemorrhage 61
(IVH), periventricular leukomalacia (PVL), retinopathy of prematurity (ROP), patent ductus 62
arteriosus (PDA), and neonatal death, have been reported as complications resulting from placental 63
chorioamnionitis and funisitis in multiple studies21-25. The direct association between increased 64
sepsis-related mortality and decreased gestational age is also well documented, but the contribution 65
of coexisting comorbidities that commonly accompany extreme preterm birth has not been well 66
described. Ascertaining this information through data extraction (without concurrent chart review) 67
may be challenging in large, multicenter studies due to differences in electronic medical record 68
networks between study sites and limitations of individual hospital computer systems. The purpose 69
of this study, therefore, is to determine how concurrent comorbidities may contribute to death in 70
infants born ≤ 32 weeks of gestation and admitted to the University California, San Diego, 71
(UCSD) or the Rady Children’s Hospital of San Diego’s (RCHSD) network of regional neonatal 72
intensive care units (NICUs), following a diagnosis of bacteremia or culture negative sepsis. We 73
additionally characterize pathogens associated with positive blood culture results in our patient 74
cohort.
75 76
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Methods:
77 78
Study Population 79
A retrospective chart review of infants born ≤ 32 weeks of gestation between January 1, 80
2012 and December 31, 2015 and admitted to the following hospitals were included: UCSD 81
Hillcrest Medical Center, Rady Children’s Hospital of San Diego, and Rady Children’s Hospital 82
satellite NICUs at Scripps Memorial Hospital (La Jolla), Scripps Mercy Hospital (Hillcrest), 83
Scripps Chula Vista, Scripps Memorial Hospital (Encinitas), Palomar Medical Center 84
(Escondido), and Rancho Springs Medical Center. Infants were included for analysis if they had a 85
positive blood culture(s), ICD codes corresponding to culture negative sepsis (A41.89 – other 86
specified sepsis, A41.9 – sepsis, unspecified organism, P36.8 – other bacterial sepsis of newborn, 87
P36.9 – Bacterial sepsis of newborn, unspecified, R65.20 – severe sepsis without septic shock, 88
R65.21 - severe sepsis with septic shock, R68.89 – other general symptoms and signs, Z05.1 – 89
observation and evaluation of newborn for suspected infectious condition ruled out, and Z91.89 – 90
other specified personal risk factor, not elsewhere classified), or treatment with antibiotics ≥ 5 days 91
(or death prior to 5 days with intention to treat).
92 93
Patient Protection 94
Our research plan (#170992) was approved by the Rady/UCSD Institutional Review Board 95
and determined to present no greater than minimal risk, in accordance with research guidelines 96
involving children (45 CFR 46.404). A waiver of assent was also approved, as set forth in HHS 97
regulations at 45 CFR 46.408 and specified in 45 CFR 46.116(d).
98 99
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Definition of Sepsis, Blood Culture Negative Sepsis, and Sepsis-Related Mortality 100
There is currently no consensus definition for neonatal sepsis26-29. We defined sepsis as 101
either the laboratory finding of bacteremia (positive blood culture) or as blood “culture negative 102
sepsis”, if the blood culture result remained negative but the physician believed the patient to have 103
bacterial sepsis with antibiotic administration for >5 days and a diagnosis of culture negative sepsis 104
reported in the patient’s electronic medical record. Sepsis was further defined as early, if onset 105
occurred within the first 72 hours of life, or late if diagnosed > 72 hours. Determination of 106
contaminating blood isolates were made at the discretion of the attending neonatologist, if the 107
organism was known to be a possible contaminant [i.e., coagulase-negative Staphylococcus 108
(CoNS) or S. epidermidis] or if the infant received < 5 days of antimicrobial therapy. Infants 109
routinely completed blood culture analysis using a single 1.0 mL peripheral blood volume sample, 110
unless a central line was in place for which a 1.0 mL peripheral and central blood volume sample 111
were simultaneously tested.
112
While all patient deaths were recorded and correlated to their proximity to their sepsis 113
episode, sepsis-related mortality was defined as mortality within 14 days of a sepsis diagnosis 114
(either blood culture positive or negative), death beyond 14 days of a sepsis diagnosis if the infant 115
remained on a prolonged antibiotic course, and documentation that the sepsis event was clinically 116
relevant. Acute clinical deterioration, including increased apnea/bradycardia/desaturation events, 117
respiratory decompensation requiring increased support or intubation with mechanical ventilation, 118
or demonstration of acute onset end-organ damage or failure, may lead a clinician to determine the 119
event as clinically relevant. Institution prescribing guidelines for antimicrobial therapeutics were 120
similar among hospital systems. Decisions to prolong antibiotics beyond fourteen days were 121
attending neonatologist-dependent and typically based on the source of the primary infection 122
Accepted Manuscript
This article is protected by copyright. All rights reserved.
and/or clinical course in conjunction with a pediatric infectious disease consultation. Transition to 123
comfort care and documented contributors to death were also noted.
124 125
Outcome Measures 126
The electronic health records of these infants were reviewed to ascertain and document: (a) 127
date and time of the infectious workup, including blood culture collection, (b) date and time of 128
blood culture positivity, if applicable, with pathogen identification, and (c) antibiotics use with 129
their duration. If the patient died, then the date and time, interval from evaluation, and 130
circumstances of the patient’s death were recorded and analyzed, including transition to comfort 131
measures. Additionally, nSOFA (neonatal sequential organ failure assessment) scores were 132
calculated around the neonate’s sepsis workup and time of death30, 31. 133
134
Statistical Analysis 135
A descriptive analysis was conducted using R Studio 3.6.3 (RStudio, Inc., Boston, MA).
136
Infants who had unique cases of both early-onset and late-onset sepsis were included in early-onset 137
and late-onset groups based on the timing for each episode. Data are presented as means and 138
standard deviations (SD) for numeric data and as frequencies and percentages for nominal data.
139
Chi-square tests were used to compare counts across strata if all expected cell sizes were ≥ 5, 140
otherwise Fisher’s exact test was used. Additionally, after visual confirmation of normality from 141
Q-Q plots for all relevant covariates, t-tests were used to compare means across strata. A p-value 142
less than 0.05 was considered statistically significant.
143 144
Results:
145
Accepted Manuscript
This article is protected by copyright. All rights reserved.
146
Patient Demographics 147
Between January 1, 2012 and December 31, 2015, a total of 939 infants ≤ 32 weeks of 148
gestation were admitted to a NICU within our hospital network, of which 183 patients were found 149
to have either ≥ 1 positive blood culture result(s) or medical coding concerning for culture negative 150
sepsis for a total of 198 sepsis-related episodes (Table 1). One patient coded for culture-negative 151
sepsis was excluded, as this patient had a single documented negative culture, no indication of 152
physician concerns regarding sepsis, and discontinuation of antibiotics at 48 hours after birth with 153
no further antibiotic exposure during their hospital course. A total of 182 patients with a total of 154
198 sepsis episodes were, therefore, used for analysis, including 109 positive blood culture results 155
and 89 diagnoses of blood culture-negative sepsis. Four patients had separate episodes of early- 156
onset and late-onset sepsis, which were classified accordingly. No statistical differences were 157
observed in regard to sex, ethnicity, or mode of delivery between bacteremic and blood culture- 158
negative infants. As expected, neonates born at lower birth weights and gestational ages exhibited 159
a higher incidence of LOS than EOS (946 gm vs. 1241 gm and 27 weeks vs. 29 weeks of gestation;
160
p<0.001). The diagnosis of EOS was 5.8-times more likely to be classified as blood culture- 161
negative than -positive sepsis, and bacteremia was nearly 4-times more likely to be identified in 162
infants diagnosed with LOS.
163 164
Cause of Death 165
Of the 182 patients reviewed, a total of fourteen patients died (Table 2 and 3). Mortality 166
was not significantly different between EOS and LOS groups [24.4% (6/14) vs. 17.1% (8/14)], nor 167
for blood culture -positive vs. negative sepsis [50% (7/14) each group]. The majority of infants 168
Accepted Manuscript
This article is protected by copyright. All rights reserved.
(80% or 11/14) in this cohort died at ≤ 14 days of life. Two infants, born at 24 weeks of gestation, 169
died within 24 hours of life. The first (Patient 8) died following a diagnosis of blood culture 170
negative sepsis, spontaneous intestinal perforation, and severe metabolic acidosis. The second, 171
diagnosed with Escherichia coli bacteremia, had a clinical course complicated by grade 4 IVH 172
(intraventricular hemorrhage) and bilateral pneumothoraces (Patient 14).
173
In our cohort (Table 3), seven patients very likely died as a direct result of infection, 174
irrespective of transition to comfort care: (a) Patient 5 was newly diagnosed with high-stage NEC 175
and severe metabolic acidosis, (b) Patient 10 with S. aureus bacteremia, CDH, and intestinal 176
perforation on DOL 14, (c) Patient 14 discussed above, and (d) four infants who succumbed 177
following acute intestinal perforation, with or without fungemia/bacteremia and/or severe IVH 178
(Patients 8, 11, 12, and 13). In our cohort, all infants except for one had a diagnosis of NEC or SIP 179
that preceded their diagnosis of bacteremia. Patients 10, 12, and 13 had their blood cultures 180
obtained after the intestinal perforation was diagnosed (positive for S. aureus, Candida albicans, 181
Enterobacter cloacae, respectively), while Patient 11 was convalescing from E. coli EOS when 182
they developed a SIP six days into antimicrobial therapy, with a negative blood culture obtained 183
at the time of their SIP diagnosis.
184
Two other infants died following a positive blood culture for CoNS while receiving 185
antibiotics. The first infant (Patient 2) was born at 25 weeks of gestation with complex congenital 186
anomalies, including congenital heart disease and CDH, experienced progressive deterioration in 187
their clinical course with death occurring on DOL 117 from pulseless ventricular tachycardia. The 188
second infant (Patient 9) died of severe pulmonary hemorrhage and bilateral grade 3 IVH on the 189
third DOL after preterm birth at 24 weeks of gestation. Although blood cultures isolated CoNS in 190
both of these cases, the attending physicians’ documentation clearly defined each as a contaminant 191
Accepted Manuscript
This article is protected by copyright. All rights reserved.
and death was attributed to noninfectious-related causes. The accurate determination of blood 192
culture contamination, however, can be challenging in this patient population if multiple cultures 193
are not obtained and mortality occurred while the patient was still receiving antibiotics.
194
Five patients diagnosed with blood culture-negative sepsis highlight the importance of 195
concurrent comorbidities in relation to sepsis mortality in this patient cohort and the need to 196
develop a consensus definition for neonatal sepsis. The first (Patient 3) died following failure of 197
full resuscitative measures from a pulmonary hemorrhage associated with severe coagulopathy 198
following a traumatic hepatic injury sustained at the time of birth. The other four (Patients 1, 4, 6, 199
and 7) were transitioned to comfort care measures due to severe IVH and/or periventricular 200
leukomalacia (PVL), with two infants each at 31 weeks’ gestational age also having concurrent 201
congenital anomalies, including hypoplastic left heart syndrome (Patient 6) and TEF (Patient 7).
202
While each of these patients fulfilled our a priori definition of sepsis-related death and had culture- 203
negative sepsis listed as a contributing factor to mortality in their medical records, it remains most 204
plausible that non-infectious comorbidities and congenital anomalies were primarily involved in 205
their deaths.
206
Two deaths were classified as “not sepsis related” after careful review of their medical 207
records and failure to meet our a priori definition. Patient 1, born at 23 weeks of gestation, died 208
after initiation of comfort care measures on day of life (DOL) 103 due to worsening seizure activity 209
associated with post-hemorrhagic IVH and evolving PVL with previous history of intestinal 210
perforation. The other infant, Patient 2 discussed above, was born extremely preterm at 25 weeks’
211
CGA and succumbed after full resuscitative measures failed on DOL 117 from complications 212
related to congenital heart disease (coarctation with Ebstein’s anomaly) and congenital 213
diaphragmatic hernia (CDH).
214
Accepted Manuscript
This article is protected by copyright. All rights reserved.
In our patient cohort the transition to comfort care procedures was associated with the 215
majority of sepsis-related deaths. We determined that 80% (11/14) of infants with either 216
bacteremia or blood culture negative sepsis died following transition to comfort measures. In this 217
subgroup, nearly three-fourths (8/11) of patients had a corresponding diagnosis of severe grade 218
3/4 IVH or PVL, an equal number (8/11) had developed NEC or suffered from an intestinal 219
perforation (IP), and almost half (5/11) had a concurrent diagnosis of IVH/PVL and NEC/IP.
220 221
nSOFA Scores of Infants Who Died 222
The nSOFA score incorporates a patient’s respiratory dysfunction (need for ventilatory 223
support, supplemental oxygen provided in the context of transcutaneous oxygen saturations), 224
cardiovascular dysfunction (pharmacologic blood pressure support including inotropes, 225
vasopressors, and steroids), and hematologic dysfunction (platelet counts) to determine an overall 226
risk of mortality following a diagnosis of neonatal sepsis30, 31. The score ranges from a minimum 227
of 0 (best score) to a maximum of 15 (worst score). Patients with late-onset infection with a nSOFA 228
score of >4 at evaluation were five-times more likely to die compared with those who score ≤430. 229
Among the patients that experienced mortality, the mean nSOFA at evaluation was 4.5 [standard 230
deviation (SD) 3.4], and the mean maximum nSOFA during the episode of 8.8 (SD 4.0) (Table 231
3). Five infants in our study had nSOFA scores at evaluation <4, including three with concurrent 232
congenital anomalies (hypoplastic left heart syndrome, TEF, and CDH with Epstein’s anomaly 233
and coarctation), one with evolving PVL and severely abnormal electroencephalogram, and one 234
died of an acute pulmonary hemorrhage with bilateral grade 3 IVH. Three of these five infants 235
died following transition to comfort care protocols.
236 237
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Distribution of Bacterial Pathogens Isolated from Blood Culture Analysis 238
A total of 162 positive blood culture results were identified (Table 4). E.coli was prominent 239
in infants with EOS (46.2% or 6/13 blood cultures), while 80% of LOS infections were caused by 240
gram-positive organisms including CoNS, Enterococcus cloacae, and S. aureus. Similar to 241
published studies concerning LOS, CoNS was the primary organism isolated in blood culture tests 242
in more than half of cases (56%). Fungal pathogens were exclusively found in the evaluation and 243
diagnosis of LOS.
244 245
Discussion:
246 247
Infection remains an important clinical entity in neonatology. In the United States, an 248
estimated six of ten VLBW infants admitted to neonatal intensive care units will be diagnosed with 249
bacteremia or blood culture negative sepsis, and nearly one-third may die from complications 250
related to their infection20, 32. Although multiple factors contribute to the heightened risk for 251
infection in VLBW infants, multi-organ immaturity and other natural, gestational age-appropriate 252
physiologic processes may closely resemble infectious processes. A primary fear of missing or 253
misdiagnosing an infection may, therefore, lead clinicians to administer prolonged antibiotic 254
courses, even if presented with a negative blood culture result6. 255
Case-fatality within a given time interval is typically employed in clinical studies to detail 256
neonatal sepsis-related mortality. The time interval, generally defined as the period between 257
culture procurement and time of death, normally transpires within 72 hours, between 4-7 days, >
258
7 days, within 30 days, or during the infant’s initial hospital course14, 18, 20, 33-35. While half of all 259
EOS deaths will occur within the first three days of life20, an estimated 40% of neonates with LOS 260
Accepted Manuscript
This article is protected by copyright. All rights reserved.
died more than 7 days from their last blood culture14. As expected, death associated with LOS was 261
primarily attributable to infection when mortality occurred within three days of a positive culture, 262
while death ≥ 4 days was more likely to be caused by deleterious, non-infectious, but chronic 263
etiologies14. Similar results were obtained in this study, where the classical pro-inflammatory 264
sepsis pathways and multi-organ dysfunction did not directly lead to mortality of two patients who 265
died >14 days from their last sepsis evaluation (Patients 1 and 2), but may have contributed to 266
neurologic or cardiac injury and dysfunction that contributed to their deaths.
267
Apart from physiologic differences, the risk of developing sepsis also varies greatly 268
between preterm versus term infants, as recently reported in a study regarding culture-positive 269
EOS by Stoll and colleagues20. Using the Eunice Kennedy Shriver National Institute of Child 270
Health and Human Development Neonatal Research Network, these authors determined the rate 271
of EOS to be 30-fold higher in preterm infants born between 22 to 28 weeks of gestation compared 272
to their term counterparts. While all term infants survived their infection, nearly three in ten 273
bacteremic preterm infants born between 22 to 36 weeks of gestation died from sepsis-related 274
complications. The median gestational age of infants who died was 25.5 (IQR, 24-28) weeks with 275
a median birth weight of 850 (IQR, 680-1370) grams, with half of these deaths occurring within 276
the first 3 days of life. While concurrent comorbidities and transition to comfort care measures 277
where not reported in this study, all infants ≤32 weeks of gestation demonstrated clinical symptoms 278
within 72 hours of life, including respiratory compromise and hypotension, which the authors 279
suggests may be compatible with sepsis but are also common in preterm neonates who are not 280
infected.
281
Our findings are similar to those reported by Jacobs and colleagues36, who employed a 282
large Mednax database to determine the primary cause of death of 641 preterm and term infants.
283
Accepted Manuscript
This article is protected by copyright. All rights reserved.
In this prospectively-defined patient population, the leading causes of mortality were attributable 284
to complications associated with extreme birth, including IVH, NEC, sepsis, and respiratory failure 285
resulting from progressive respiratory distress. The etiology of mortality, however, shifted with 286
increasing gestational age to include hypoxic-ischemic encephalopathy and genetic and structural 287
anomalies in infants born closer to term gestation.
288
In our study, sepsis-related mortality largely resulted from the transition of intensive care 289
clinical management to comfort care protocols. This conversion was primarily guided by pre- 290
existing serious comorbidities, including severe grade 3-4 IVH, respiratory failure or lung 291
hypoplasia, and/or congenital anomalies, suggesting a baseline increased risk for poor survivability 292
and/or substantial long-term neurodevelopmental impairments. Because infection-associated pro- 293
inflammatory immune responses may exacerbate or worsen baseline neonatal outcomes32, 37, 38, the 294
additional diagnosis of sepsis may have further facilitated discussions of initiation of comfort 295
protocols with the patient’s parents and family.
296
In this study, mortality data demonstrates a lower rate of death than historically described, 297
with an overall case-fatality rate of 7.7%. The inclusion of both bacteremia and blood culture- 298
negative sepsis in this study cohort may have contributed to this discrepancy, as many neonatal 299
studies only include bacteremic patients. No significant differences in mortality caused by Gram - 300
positive compared with -negative sepsis were demonstrated, which could be related to low sample 301
size. Other limitations of this study include its retrospective nature and decision to define neonatal 302
sepsis based on blood culture results, excluding other important sources of infection such as 303
cerebrospinal fluid or urine. The analysis of a small sample size and failure to capture infants with 304
positive viral studies (culture or quantitative nucleic acid testing) are additional weaknesses. Even 305
though this is a multicenter study, it encompasses a single academic neonatal-perinatal medicine 306
Accepted Manuscript
This article is protected by copyright. All rights reserved.
practice group with approximately twenty-two neonatologists, so findings may not be applicable 307
to other neonatal centers. Transitions to comfort care, as embraced by our institution, may also not 308
be acceptable to other neonatal providers.
309
In conclusion, inherent difficulties remain in our ability to accurately assess sepsis-related 310
mortality in preterm neonates. Because current published definitions of neonatal sepsis are 311
heterogeneous (bacteremia +/- culture negative sepsis) and without considerations for new onset 312
organ dysfunction, the establishment of a consensus definition is critical. The primary use of 313
bacteremia to diagnose neonatal sepsis falsely assumes only bacteremic patients can experience 314
life-threatening organ dysfunction resulting in sepsis-related mortality. Development of a 315
consensus definition for neonatal sepsis could improve clarity and promote opportunities to 316
consider how sepsis-related deaths should be regarded in terms of neonatal statistics, research, and 317
education, thereby aligning clinicians, researchers, and epidemiologists to improve patient 318
outcomes26-29. 319
A higher number of severe comorbidities were observed in our VLBW infant cohort who 320
died following a diagnosis of sepsis compared to our historic baseline NICU data of approximately 321
40%. Future research should examine the number and type of comorbidities and their role in sepsis 322
related mortality, as these common conditions may artificially inflate sepsis-related case-fatality 323
rates26. As emerging therapeutics and technologies are engineered to remedy sepsis-mediated 324
pathophysiologic processes, an accurate assessment of the incidence of sepsis will be vital to 325
determine study feasibility, calculate sample size, and validate outcome measures. Future studies 326
should incorporate data regarding confounding factors and/or comorbidities that contribute to 327
neonatal sepsis-related mortality, including the transition to comfort care protocols.
328
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Table 1. Patient Demographics and Infection Evaluations, 2012-2015
Characteristic† Total Early Onset Late Onset P-valuea
Gestational Age at Birth (Weeks),
Mean (SD)* 28 (3) 29 (3) 27 (3) <0.001
Birth Weight (Grams), Mean (SD)* 1066 (452) 1241 (459) 946 (407) <0.001
Race/Ethnicity, n (%)* 0.29
Hispanic 101 (54.0) 37 (48.7) 64 (57.7)
Non-Hispanic 86 (46.0) 39 (51.3) 47 (42.3)
Sex, n (%)* 0.40
Male 121 (64.7) 46 (60.5) 75 (67.6)
Female 66 (35.3) 30 (39.5) 36 (32.4)
Mode of Delivery, n (%)* 0.71
Caesarean Section 132 (70.6) 52 (68.4) 80 (72.1)
Vaginal 55 (29.4) 24 (31.6) 31 (27.9)
Culture Status, n (%)** <0.001
Culture Positive 109 (55.1) 11 (14.7) 98 (79.7)
Culture Negative 89 (44.9) 64 (85.3) 25 (20.3)
†Out of 182 unique patients, unless otherwise noted. (n = 198 sepsis episodes)
aP-value from the t-test for continuous variables and Chi-Sq. goodness of fit test for categorical variables across sepsis onset categories.
*Four (4) patients had separate episodes of early and late onset sepsis were classified as both early and late onset;
thus, for time-fixed variables, the sample size will be 4 patients larger than the reported sample size.
**Out of 198 sepsis episodes; the 4 patients with early and late onset (8 episodes) were not recategorized at all.
Bold indicates significance at the P < 0.05 level.
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Table 2. Survival and Cause of Death by Sepsis Onset, 2012-2015
Culture Positive Culture Negative
Characteristic†
Total Early Onset Late Onset Early Onset Late Onset
P-value*
(n = 198) (n = 11) (n = 98) (n = 64) (n = 25) Survival to Discharge, n
(%) 0.20
Yes 184
(92.9) 9 (81.8) 93 (94.9) 60 (93.8) 22 (88.0)
No 14 (7.1) 2 (18.2) 5 (5.1) 4 (6.2) 3 (12.0)
Sepsis-Related Death** 0.78
No 2 (14.3) 0 (0.0) 1 (20.0) 0 (0.0) 1 (33.3)
Yes 12 (85.7) 2 (100.0) 4 (80.0) 4 (100.0) 2 (66.7)
Transition of Care 10 (71.4) 2 (100.0) 3 (60.0) 3 (75.0) 2 (66.7)
†Out of 198 episodes, unless otherwise noted. (n = 182 patients)
*P-value for Fisher’s exact test across the four categories of sepsis and its onset.
**Out of 14 deaths.
Sepsis-related death is defined as death ≤ 14 days of sepsis onset or death while on antibiotic treatment.
Bold indicates significance at the P < 0.05 level.
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Table 3: Patient-Specific Clinical Conditions Attributable to Cause of Death
EEG: electroencephalogram; CHD: congenital heart disease; TEF: tracheoesophageal fistula; CDH: congenital diaphragmatic hernia; IVH: intraventricular hemorrhage; NEC: necrotizing enterocolitis; PVL: periventricular leukomalacia
Patient GA at Birth
Birth Weight
(gm)
Age at Last Blood
Culture (days)
Age at Death (days)
Max nSOFA
score
Organism Sepsis- Related
Death
Transition to Comfort
Care
Associated Clinical Conditions
1 23 575 44 103 7 Culture Negative Sepsis No Yes Worsening post-hemorrhagic changes from grade 4 IVH with
severely abnormal EEG; previous h/o intestinal perforation
2 25 610 52 117 1 Staphylococcus epidermitis No No Pulseless ventricular tachycardia associated with CDH and CHD
3 26 940 0 2 12 Culture Negative Sepsis Yes No Hepatic hemorrhage with intrabdominal bleeding; metabolic
acidosis; coagulopathy; acute severe pulmonary hemorrhage
4 26 790 6 8 7 Culture Negative Sepsis Yes Yes New onset right grade 3 and left grade 4
5 24 605 10 10 12 Culture Negative Sepsis Yes Yes NEC; severe metabolic acidosis
6 31 1528 30 33 14 Culture Negative Sepsis Yes Yes Hypoplastic left heart syndrome; NEC s/p laparotomy; IVH with evolving PVL
7 31 1800 2 9 6 Culture Negative Sepsis Yes Yes Acute onset bilateral grade 4 IVH, TEF with esophageal atresia, duodenal atresia, and double outlet right ventricle
8 24 620 0 0 8 Culture Negative Sepsis Yes Yes Intestinal perforation, profound metabolic acidosis, bleeding, and hypotension
9 24 610 4 4 11 Staphylococcus epidermitis Yes No Severe pulmonary hemorrhage, bilateral grade 3 IVH
10 31 1380 11 14 15 Staphylococcus aureus Yes Yes Left-sided CDH, severe pulmonary hypertension, and intestinal perforation
11 25 791 2 8 8 Escherichia coli Yes Yes Recovering from E. coli sepsis (6 days into treatment) with acute
intestinal perforation; pre-existing right grade 3; left grade 4 IVH
12 23 646 6 6 4 Candida albicans Yes Yes Intestinal perforation with right grade 4; left grade 3 IVH
13 23 561 7 7 6 Enterobacter cloacae Yes Yes Intestinal perforation with right grade 4; left grade 3 IVH
14 24 709 0 0 12 Escherichia coli Yes Yes Severe metabolic acidosis; coagulopathy, anemia, poor cardiac
function; bilateral pneumothoraces; bilateral IVH
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Table 4: Blood Culture Pathogens, 2012-2015
Characteristic
Total Cultures Isolates
P-value*
Early Late
(n = 162) (n = 13) (n = 149)
Pathogen, n (%) <0.001
Gram-Positive
Coagulase-negative Staphylococcus 86 (53.1) 2 (15.4) 84 (56.4)
Enterococcus spp. 18 (11.1) 1 (7.7) 17 (11.4)
Staphylococcus aureus 17 (10.5) 0 (0.0) 17 (11.4)
Group B Streptococcus 5 (3.1) 1 (7.7) 4 (2.7)
Enterobacter spp. 4 (2.5) 0 (0.0) 4 (2.7)
Other Streptococcus spp. 1 (0.6) 0 (0.0) 1 (0.7)
Bacillus spp. 1 (0.6) 0 (0.0) 1 (0.7)
Micrococcus spp. 1 (0.6) 0 (0.0) 1 (0.7)
Gram-Negative
Escherichia coli 9 (5.6) 6 (46.2) 3 (2.0)
Klebsiella pneumoniae 5 (3.1) 1 (7.7) 4 (2.7)
Pseudomonas spp. 3 (1.9) 1 (7.7) 2 (1.3)
Acinetobacter spp. 1 (0.6) 0 (0.0) 1 (0.7)
Citrobacter koseri 1 (0.6) 1 (7.7) 0 (0.0)
Salmonella spp. 1 (0.6) 0 (0.0) 1 (0.7)
Serratia spp. 1 (0.6) 0 (0.0) 1 (0.7)
Fungal
Candida spp. 7 (4.3) 0 (0.0) 7 (4.7)
Other
Paenibacillus spp. 1 (0.6) 0 (0.0) 1 (0.7)
n = 162 true positive blood culture results
*P-value for Fisher’s exact test across the categories of sepsis onset.
Bold indicates significance at the P < 0.05 level.
Accepted Manuscript
This article is protected by copyright. All rights reserved.
Acknowledgements:
329
We would like to thank the members of the Altman Clinical & Translational Research Institute, 330
UC San Diego, for UCSD data acquisition. We would also like to thank Ms. Sarah Lazar and the 331
UCSD Neonatology Research Team for coordinating data acquisition from UCSD and RCHSD.
332 333
Author’s Contribution:
334
Study Concept and design: Barnette, Wynn, and Lawrence 335
Acquisition of data: Barnette, Lazar, and Richardson.
336
Analysis and interpretation of data: Barnette, Schumacher, Armenta, and Lawrence.
337
Drafting of Manuscript: Barnette and Lawrence 338
Critical Revision: All authors 339
Accepted Manuscript
This article is protected by copyright. All rights reserved.
References:
1. Shane AL, Sanchez PJ, Stoll BJ. Neonatal sepsis. Lancet. Oct 14 2017;390(10104):1770- 1780. doi:10.1016/S0140-6736(17)31002-4
2. Puopolo KM, Benitz WE, Zaoutis TE. Management of neonates born at </=34 6/7 weeks' gestation with suspected or proven early-onset bacterial sepsis. Pediatrics. Dec
2018;142(6)doi:10.1542/peds.2018-2896
3. Anand KJ, Aranda JV, Berde CB, et al. Summary proceedings from the neonatal pain- control group. Pediatrics. Mar 2006;117(3 Pt 2):S9-S22. doi:10.1542/peds.2005-0620C 4. Zea-Vera A, Ochoa TJ. Challenges in the diagnosis and management of neonatal sepsis. J
Trop Pediatr. Feb 2015;61(1):1-13. doi:10.1093/tropej/fmu079
5. Khan AM, Morris SK, Bhutta ZA. Neonatal and perinatal infections. Pediatr Clin North Am. Aug 2017;64(4):785-798. doi:10.1016/j.pcl.2017.03.008
6. Ely DM, Driscoll AK. Infant mortality in the United States, 2017: Data from the period linked birth/infant death file. National Vital Statistics Reports. 2019;68(10):1-20.
7. Squire E, Favara B, Todd J. Diagnosis of neonatal bacterial infection: hematologic and pathologic findings in fatal and nonfatal cases. Pediatrics. Jul 1979;64(1):60-4.
8. Gupta S, Sakhuja A, Kumar G, McGrath E, Nanchal RS, Kashani KB. Culture-negative severe sepsis: nationwide trends and outcomes. Chest. Dec 2016;150(6):1251-1259.
doi:10.1016/j.chest.2016.08.1460
9. Matics TJ, Sanchez-Pinto LN. Adaptation and validation of a pediatric sequential organ failure assessment score and evaluation of the Sepsis-3 Definitions in critically ill children. JAMA Pediatr. Oct 2 2017;171(10):e172352.
doi:10.1001/jamapediatrics.2017.2352
10. Weiss SL, Fitzgerald JC, Pappachan J, et al. Global epidemiology of pediatric severe sepsis: the sepsis prevalence, outcomes, and therapies study. Am J Respir Crit Care Med.
May 15 2015;191(10):1147-57. doi:10.1164/rccm.201412-2323OC
11. Wynn JL, Cvijanovich NZ, Allen GL, et al. The influence of developmental age on the early transcriptomic response of children with septic shock. Mol Med. 2011;17(11- 12):1146-56. doi:10.2119/molmed.2011.00169
12. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. Apr 17 2003;348(16):1546-54.
doi:10.1056/NEJMoa022139
13. Seymour CW, Iwashyna TJ, Cooke CR, Hough CL, Martin GS. Marital status and the epidemiology and outcomes of sepsis. Chest. Jun 2010;137(6):1289-96.
doi:10.1378/chest.09-2661
Accepted Manuscript
This article is protected by copyright. All rights reserved.
14. Stoll BJ, Hansen N, Fanaroff AA, et al. Late-onset sepsis in very low birth weight neonates: The experience of the NICHD Neonatal Research Network. Pediatrics.
2002;110(2):285-291. doi:10.1542/peds.110.2.285
15. Ramasethu J. Prevention and treatment of neonatal nosocomial infections. Matern Health Neonatol Perinatol. 2017;3:5. doi:10.1186/s40748-017-0043-3
16. Kiser C, Nawab U, McKenna K, Aghai ZH. Role of guidelines on length of therapy in chorioamnionitis and neonatal sepsis. Pediatrics. Jun 2014;133(6):992-8.
doi:10.1542/peds.2013-2927
17. Greenwood C, Morrow AL, Lagomarcino AJ, et al. Early empiric antibiotic use in preterm infants is associated with lower bacterial diversity and higher relative abundance of Enterobacter. J Pediatr. Jul 2014;165(1):23-9. doi:10.1016/j.jpeds.2014.01.010
18. Stoll BJ, Hansen NI, Higgins RD, et al. Very low birth weight preterm infants with early onset neonatal sepsis: the predominance of gram-negative infections continues in the National Institute of Child Health and Human Development Neonatal Research Network, 2002-2003. Pediatr Infect Dis J. Jul 2005;24(7):635-9.
doi:10.1097/01.inf.0000168749.82105.64
19. Schrag SJ, Cutland CL, Zell ER, et al. Risk factors for neonatal sepsis and perinatal death among infants enrolled in the prevention of perinatal sepsis trial, Soweto, South Africa.
Pediatr Infect Dis J. Aug 2012;31(8):821-6. doi:10.1097/INF.0b013e31825c4b5a 20. Stoll BJ, Puopolo KM, Hansen NI, et al. Early-onset neonatal sepsis 2015 to 2017, the
rise of Escherichia coli, and the need for novel prevention strategies. JAMA Pediatr. Jul 1 2020;174(7):e200593. doi:10.1001/jamapediatrics.2020.0593
21. Moscuzza F, Belcari F, Nardini V, et al. Correlation between placental histopathology and fetal/neonatal outcome: chorioamnionitis and funisitis are associated to
intraventricular haemorrage and retinopathy of prematurity in preterm newborns. Gynecol Endocrinol. May 2011;27(5):319-23. doi:10.3109/09513590.2010.487619
22. Tang Q, Zhang L, Li H, Shao Y. The fetal inflammation response syndrome and adverse neonatal outcomes: a meta-analysis. J Matern Fetal Neonatal Med. Dec 18 2019:1-13.
doi:10.1080/14767058.2019.1702942
23. Ericson JE, Laughon MM. Chorioamnionitis: implications for the neonate. Clin Perinatol. Mar 2015;42(1):155-65, ix. doi:10.1016/j.clp.2014.10.011
24. Jung E, Romero R, Yeo L, et al. The fetal inflammatory response syndrome: the origins of a concept, pathophysiology, diagnosis, and obstetrical implications. Semin Fetal Neonatal Med. Aug 2020;25(4):101146. doi:10.1016/j.siny.2020.101146
25. Thomas W, Speer CP. Chorioamnionitis: important risk factor or innocent bystander for neonatal outcome? Neonatology. 2011;99(3):177-87. doi:10.1159/000320170
Accepted Manuscript
This article is protected by copyright. All rights reserved.
26. McGovern M, Giannoni E, Kuester H, et al. Challenges in developing a consensus definition of neonatal sepsis. Pediatr Res. Jul 2020;88(1):14-26. doi:10.1038/s41390- 020-0785-x
27. Molloy EJ, Wynn JL, Bliss J, et al. Neonatal sepsis: need for consensus definition, collaboration and core outcomes. Pediatr Res. Jul 2020;88(1):2-4. doi:10.1038/s41390- 020-0850-5
28. Wynn JL. Defining neonatal sepsis. Curr Opin Pediatr. Apr 2016;28(2):135-40.
doi:10.1097/mop.0000000000000315
29. Wynn JL, Wong HR, Shanley TP, Bizzarro MJ, Saiman L, Polin RA. Time for a neonatal-specific consensus definition for sepsis. Pediatr Crit Care Med. Jul 2014;15(6):523-8. doi:10.1097/pcc.0000000000000157
30. Wynn JL, Polin RA. A neonatal sequential organ failure assessment score predicts mortality to late-onset sepsis in preterm very low birth weight infants. Pediatr Res. Jul 2020;88(1):85-90. doi:10.1038/s41390-019-0517-2
31. Fleiss N, Coggins SA, Lewis AN, et al. Evaluation of the neonatal sequential organ failure assessment and mortality risk in preterm infants with late-onset infection. JAMA Netw Open. Feb 1 2021;4(2):e2036518. doi:10.1001/jamanetworkopen.2020.36518 32. Stoll BJ, Hansen NI, Adams-Chapman I, et al. Neurodevelopmental and growth
impairment among extremely low-birth-weight infants with neonatal infection. Jama.
Nov 17 2004;292(19):2357-65. doi:10.1001/jama.292.19.2357
33. Bizzarro MJ, Dembry LM, Baltimore RS, Gallagher PG. Changing patterns in neonatal Escherichia coli sepsis and ampicillin resistance in the era of intrapartum antibiotic prophylaxis. Pediatrics. Apr 2008;121(4):689-96. doi:10.1542/peds.2007-2171 34. Stoll BJ, Hansen N. Infections in VLBW infants: studies from the NICHD Neonatal
Research Network. Semin Perinatol. Aug 2003;27(4):293-301. doi:10.1016/s0146- 0005(03)00046-6
35. Stoll BJ, Hansen NI, Sánchez PJ, et al. Early onset neonatal sepsis: the burden of group B Streptococcal and E. coli disease continues. Pediatrics. May 2011;127(5):817-26.
doi:10.1542/peds.2010-2217
36. Jacob J, Kamitsuka M, Clark RH, Kelleher AS, Spitzer AR. Etiologies of NICU deaths.
Pediatrics. Jan 2015;135(1):e59-65. doi:10.1542/peds.2014-2967
37. Stoll BJ, Hansen NI, Bell EF, et al. Neonatal outcomes of extremely preterm infants from the NICHD Neonatal Research Network. Pediatrics. Sep 2010;126(3):443-56.
doi:10.1542/peds.2009-2959
Accepted Manuscript
This article is protected by copyright. All rights reserved.
38. Sung TJ, Sohn JA, Oh S, Lee JA. The influence of the variation in sepsis rate between neonatal intensive care units on neonatal outcomes in very-low-birth-weight infants. Sci Rep. Apr 21 2020;10(1):6687. doi:10.1038/s41598-020-63762-6