(Main Manuscript) Clinical Practice Guidelines For The Screening, Diagnosis, Treatment and Prevention of Neonatal Sepsis

APPENDIX A
Draft Evidence-Based Recommendations for

Neonatal Sepsis
Clinical Practice Guidelines 2019
Developed by the Technical Working Group
This document contains the evidence guide, evidence summaries, draft recommendations and
evidence tables for questions on neonatal sepsis.
GUIDE FOR THE PANELISTS ON HOW TO USE THIS DOCUMENT
1. Each individual question was searched for evidence until October 20, 2019 and summarized.
If possible, the data from several clinical studies were pooled to come up with single estimate
or the treatment effect. This was done by conducting a meta-analysis or basing it on an existing
meta-analysis.
2. The first section are guides on how to interpret the results from evidence on diagnostic tests
and therapy and how to use the GRADE approach in grading the level of quality of evidence.
The second section contains the 11 CPG questions and their evidence summaries and
corresponding draft evidence-based recommendations.
3. The evidence for each question was presented in three (3) parts:
a. Summary of Evidence in simplified text format (including table on the interpretation of
the evidence for some questions).
b. Evidence Tables showing the assessment of the quality of evidence which may include
summary statistics
c. Appendices containing forest plots with the pooled effect
4. All the evidence summaries will be presented during the panel meeting on November 24, 2019.
We will allot ample time if you have questions and will need further clarifications. If you need
assistance prior to the meeting, you can direct your email to [email protected].
5. To the members of the panel who are not really into numbers, please just focus on the first
section: Summary of Evidence.
6. During the panel meeting, we will be assisting you the entire time in all the steps. However,
your vote for the final decision on the recommendations will be yours alone.
A2
I. QUICK GUIDE ON HOW TO INTERPRET THE EVIDENCE
A. EVIDENCE ON THERAPY
Note: This quick guide on how to read evidence on therapy will be useful in Questions 1, 2, 3, and 4.
Table 1. Ways of expressing effectiveness

Outcome Summary of result within each Comparison of results between
group two groups
Dichotomous Proportion Relative risk reduction, absolute risk
(e.g. lived or died, BP (e.g. deaths per 100 patients) reduction, relative risk
controlled or not) (see Table 2)
Rate = e.g. deaths per 100 patients Hazard ratio = rate in treatment/rate in
control group
Continuous Mean (e.g. mean blood pressure) Mean difference = mean in control –
(e.g. blood pressure in mean in treatment group
mmHg, quality of life on a
scale of 0 to 1)
Table 2. Interpreting 95% Confidence Intervals (CIs)

Instructions: When researchers express the effect of treatment using the relative risk reduction, absolute risk reduction, or
relative risk, they often give us a range of possibilities rather than a single estimate. This range of possibilities is called a ‘95%
Confidence Interval (95% CI)’ to mean ‘we are 95% sure that the true effect of a drug lies in this range’. Go through this tackle
box in rows, to discover how helpful 95% CIs are.
Interpretation of 95% Confidence Intervals (CIs)

Measure of effectiveness Superior Inferior Inconclusive Equivalent
and interpretation of (treatment surely (treatment surely (we need more (the treatments
estimates better than control) worse than control) studies) are equal)a
Relative Risk (RR)
= Rt/ Rc Both ends of 95% CI Both ends of 95% CI 95% CI wide; 95% CI narrow;
<1.0 > 1.0 straddles 1.0 straddles 1.0
<1.0 Treatment
beneficial
=1.0 Treatment no effect Example: Example: Example: Example:
>1.0 Treatment harmful RR = 0.7 RR = 2.4 RR = 1 RR = 1
[95% CI: 0.6, 0.8] [95% CI: 1.8, 3.2] [95% CI: 0.2, 5.3] [95% CI: 0.9, 1.1]
Absolute Risk Reduction

(ARR) Both ends of 95% CI Both ends of 95% CI 95% CI straddles 95% CI straddles
= Rc - Rt >0% <0% 0%; either end is 0%; either end is
(usually in percent) far from 0% close to 0%
>0% Treatment Example: Example: Example: Example:
beneficial ARR = 2% ARR = -3% ARR = 1% ARR =0.2%
=0% Treatment no effect [95% CI: 1%, 3%] [95% CI: -7%, -1%] [95% CI: -20%, [95% CI: -0.1%,
<0% Treatment harmful 32%] 0.5%]
A3
a. In both inconclusive and equivalent results, the 95% CI interval straddles the point of no effect (ARR = 0% or RR =
1.0). One end reflects the worst possible harm, while the other end reflects the best possible benefit. The only
difference is that, in equivalence, either end is close to “no effect” (i.e. any benefit is ignorable and any harm is ignorable
too). Consider the ends of the 95% CI to make sure you agree that the benefits and harms are ignorable.
b. Rc is the rate of the outcomes in the Control group; Rt is the rate of the outcome in the Treatment group
Note: The interpretations in this table only hold if the dichotomous events are expressed as adverse rather than desirable
events, e.g. death rather than survival, treatment failure rather that cure, or disease rather than disease-free. When
dichotomous outcomes are expressed as desirable events, the interpretation of benefit and harm is revered. We feel researchers
should standardize reporting and use adverse events in order to avoid confusion. Unfortunately, nobody cares what we think.
Table 3. How to interpret forest plots

Instructions: The balloons below label the most important parts of the forest plot. Go through these labels and familiarize yourself
with the anatomy of the graph. Once you feel sufficiently familiar with the anatomy, go through the notes below on what the
forest plot can signify.
A4
B. EVIDENCE ON DIAGNOSTIC TESTS
Note: This quick guide on how to read and interpret evidence on diagnostic tests will be helpful in Questions
5,6,7, and 8.
There are four conventional ways of determining how correct the test is and
would be adequate if we are comparing the results of two tests using the method
of 2x2 table.
1. Sensitivity (sn) refers to the proportion of persons with disease who correctly have
a positive test.
2. Specificity (sp) refers to the proportion of persons with no disease who correctly
have a negative test.
3. Positive predictive value (ppv) is the proportion of persons with a positive test
who correctly turn out to have disease
4. Negative predictive value (npv) is the proportion of persons with a negative test
who correctly turn out to have no disease
However, if we need to evaluate a test with multilevel results, we need a “2 x n” table

and compute for likelihood ratio (LR). LR is a measure of how much the likelihood of the
disease changes given a test result.
Table 1. Interpreting likelihood ratios (LRs)

Likelihood ratio Likelihood of Grade of likelihood
disease
LRs > 1.0 INCREASE LR<3.0 (close to 1.0) – weakly positive

LR=3.0-10.0 – moderately positive
LR10.0 is strongly positive
LRs < 1.0 DECREASE LR >0.3 (close to 1.0) – weakly negative

LR=0.3-0.1 – moderately negative
LR<0.1 strongly negative
A5
STEP 2. Determine the likelihood
ratio of the test result from the studies
reviewed. Plot this on the middle
vertical axis.
STEP 1.
Estimate the
pre-test
probability
from the
clinical
history,
physical
examination
of your STEP 3.
patient, Connect the
survey two points in
results or Step 1 and 2,
from your and extend
hospital’s the line to the
surveillance. rightmost
Plot this on vertical axis.
the left-most The point of
vertical axis. intersection is
the probability
of disease
after the test.
Figure B.1. Using Bayes nomogram for estimating post-test probability
A6
C. EVIDENCE ON SCREENING PROGRAM
Note: This quick guide on how to read evidence on diagnostic tests will be helpful in Question 9.
The criteria for evaluating screening programs are:

1. The burden of illness must be high.
2. The tests must be accurate.
3. Early treatment must be more effective than late treatment.
4. Diagnostic tests and early treatment must be safe.
5. The cost of screening strategy must be commensurate to the potential
benefit.
Figure C.1. Admissible evidence for evaluation of a screening program

Adapted from Dans Al, Dans LF and Silvestre MA. Painless Evidence – Based Medicine. 2nd edition. 2016.
A7
A.
B.
Figure C.2 Indirect (A) and direct (B) trials of the effectiveness of screening
Adapted from Dans Al, Dans LF and Silvestre MA. Painless Evidence – Based Medicine. 2nd edition. 2016.
A8
II. GRADE APPROACH IN ASSESSING THE LEVEL OF QUALITY OF EVIDENCE
(GRADE: Grading of Recommendations Assessment, Development and Evaluation; modified
from WHO Handbook in Guideline Development, 2014)
These quality ratings apply to the body of evidence assessed for the research question, not to
individual studies.
Evidence based on randomized controlled trials is initially given a high-quality rating and
evidence from observational studies is given a low-quality rating. The level is then adjusted
according to the following criteria.
Box 1. Standard criteria for grading of evidence 1
Domain Grade Characteristic

STUDY DESIGN 0 All randomized controlled trials
-1 All observational studies
0 Most of the pooled effect provided by studies, with low risk of bias (“A”)
-1 Most of the pooled effect provided by studies with moderate (“B”) or high (“C”) risk of
bias. Studies with high risk of bias weighs <40%
-2 Most of the pooled effect provided by studies with moderate (“B”) or high (“C”) risk of
bias. Studies with high risk of bias weighs ≥40%
Note:
STUDY DESIGN
Low risk of bias (no limitations or minor limitations) – “A”
LIMITATIONS
Moderate risk of bias (serious limitations or potentially very serious limitations including unclear
concealment of allocation or serious limitations, excluding limitations on randomization or
concealment of allocation) – “B”
High risk of bias (limitations for randomization, concealment of allocation, including small
blocked randomization (<10) or other very serious, crucial methodological limitations) – “C”
INCONSISTENCY 0 No severe heterogeneity (I2< 60% or X2<0.05)
-1 Severe, non-explained, heterogeneity (I2 ≥60% or X2<0.05)
If heterogeneity could be caused by publication bias or imprecision due to small studies,

downgrade only for publication bias or imprecision (i.e. the same weakness should not
be downgraded twice)
INDIRECTNESS 0 No indirectness
-1 Presence of indirect comparison, population, intervention, comparator, or outcome
0 The confidence interval is precise according to the figure below.
IMPRECISION The total cumulative study population is not very small (i.e. sample size is more than
300 participants) and the total number of events is more than 30.
A9
Box 1 (cont.). Standard criteria for grading of evidence 1
Domain Grade Characteristic
-1 One of the above-mentioned conditions is not fulfilled.

-2 The two above-mentioned are not fulfilled.
Note:
If the total number of events is less than 30 and the total cumulative sample size is
appropriatelylarge (e.g. above 3000 patients, consider not downgrading the evidence). If there
are no events in both intervention and control groups, the quality of evidence in the specific
outcome should be regarded as very low.
PUBLICATION 0 No evident asymmetry in the funnel plot or less than five studies to be plotted.
BIAS -1 Evident asymmetry in funnel plot with at least five studies
________________
1 Adapted from: Schϋnemann H, Brozek J, Oxman A, editors. GRADE handbook for grading quality of evidence and strength of recommendations. The GRADE
Working Group. Available at: http://ims.cochrane.org/revman/gradepro. (This document is contained within the “Help” section of the GRADE profiler software
version v.3.2.2.)
Table 4. Quality of evidence in GRADE

Quality Level Definition
High We are very confident that the true effect lies close to that of the estimate of the
effect.
Moderate We are moderately confident in the effect estimate: the true effect is likely to be close
to the estimate of the effect, but there is a possibility that it is substantially different.
Low Our confidence in the effect estimate is limited: the true effect may be substantially
different from the estimate of effect.
Very Low We have very little confidence in the effect estimate: the true effect is likely to be
substantially different from the estimate of effect.
References
1. Dans Al, Dans LF and Silvestre MA. Trade-off between benefit and harm is crucial in
screening recommendations. J Clin Epidem. 2010
2. Dans Al, Dans LF and Silvestre MA. Painless Evidence – Based Medicine. 2nd edition.
2016.
3. Schϋnemann H, Brozek J, Oxman A, editors. GRADE handbook for grading quality of

evidence and strength of recommendations. The GRADE Working Group. Available at:
http://ims.cochrane.org/revman/gradepro. (This document is contained within the “Help”
section of the GRADE profiler software version v.3.2.2.)
A10
TABLE OF CONTENTS
Summary of Evidence, Summary Tables and Appendices
QUESTION PAGE
Q1. Among newborns with severe sepsis, how effective and safe is double
volume exchange transfusion in addition to standard of care compared with A13
standard of care alone in decreasing morbidity (length of hospital stay,
duration of antibiotic treatment, DIC, septic shock, neurodevelopmental
outcome) and preventing mortality?
Q2 Among newborns with sclerema neonatorum, how effective and safe is A22
fresh frozen plasma transfusion in decreasing morbidity and mortality?
Q3. Among asymptomatic newborns delivered meconium-stained, how

effective and safe is giving antibiotic prophylaxis in preventing sepsis and A27
all-cause mortality?
Q4. Among newborns presenting with sepsis who improve after starting
antibiotics, how effective is a five-day course compared to standard A33
duration of antibiotics (seven days) in decreasing treatment failure and
mortality?
Q5. Among newborns with clinical sepsis, how predictive of CNS infection A39
is blood culture?
Q6. Among newborns with risk factors for neonatal infection, how accurate
A44
is each of the following CBC parameters in diagnosing sepsis? a) WBC
<5000, b) WBC >5000, c) ANC 375, d) IT ratio >0.2, e) Plt <150,000
Q7. Among asymptomatic newborns delivered to mothers with maternal

UTI within two weeks prior to delivery, how useful is a neonatal sepsis A51
calculator in predicting occurrence of sepsis and all-cause mortality?
A11
Q8. Among asymptomatic newborns, how accurate is CBC and single A53
quantitative CRP in diagnosing sepsis?
Q9. Among women near term, how effective is universal culture-based A58
screening in preventing early onset GBS sepsis?
Q10. Among newborns with no risk factors for infection presenting with A61
fever alone, what is the likelihood of them having sepsis?
Q11. Among newborns with no risk factors for infection with isolated
jaundice, what is the likelihood of sepsis? A65
A12
Q1. Among newborns with severe sepsis, how effective is double volume exchange
transfusion (DVET) as an adjunct in decreasing morbidity and mortality?
SUMMARY OF EVIDENCE
Based on high level of evidence, there was a significant decrease in the risk of
mortality among septic newborns who received double volume exchange transfusion as
an adjunct to standard of care. There was a trend towards decreased risk for organ
dysfunction and abnormal neurologic outcome but the findings were not significant.
A meta-analysis of six RCTs including 270 neonates showed significant reduction
in the risk of mortality in the DVET group compared to standard of care alone RR 0.64
[95% CI 0.50, 0.81]; (Figure Q1.1). There was no significant heterogeneity among the
included studies.1
The RCTs included had small sample sizes (ranging between 30 and 88
neonates). Both term and preterm infants were included in the RCTs.2,3,5-7 Only preterms
were included in one RCT.4 All studies included all infants regardless of birth weight.
4,5.
Exchange transfusion was performed for all septic newborns in two trials
Whereas, only infants with severe sepsis, as evidenced by disseminated intravascular
coagulation and/or organ dysfunction were included in four clinical trials.2,3,6,7 Subgroup
analysis on the effect of DVET on all-cause mortality among infants with severe sepsis
showed significant reduction in the risk of mortality in the DVET group compared to the
control group RR 0.68 [95% CI 0.52, 0.89]; (Figure Q1.2).
The use of DVET was evaluated in five studies. 2,3,5,6,7 Gunes and colleagues did
1
not specify whether single- or double-volume exchange transfusion was performed.
Sensitivity analysis was done to see the effect of the RCT by Gunes on the over-all
outcome still showed significant reduction in mortality (RR 0.66 [0.51, 0.84]; (Figure
Q1.3).1
Rates of organ dysfunction by day 14 and abnormal neurologic status at time of
discharge was reported in only one study.2 There was a trend towards decreased
incidence of organ dysfunction RR 0.78 [95% CI 0.44,1.40]; (Figure Q1.4) and abnormal
neurologic status at time of discharge RR 0.51 [95% CI 0.05, 5.43]; (Figure Q1.5) among
A13
septic neonates who received double volume exchange transfusion. However, these
results were not statistically significant.2
The reported adverse events among newborns post-DVET included
intraventricular hemorrhage (2/41; 4.8%)8; mild hypothermia (36.0 – 36.4 °C) during line
placement for DVET (12/41; 29%); transient bradycardia that spontaneously recovered
(2/41; 5%); serum potassium >6.5 mEq/L without any electrocardiography evidence
2
(2/41; 5%); and serum sodium >145 mEq/L which resolved spontaneously (2/41; 5%).
Significantly lesser neonates in the DVET group had progression of
thrombocytopenia (no DVET vs. DVET: 30/42 vs. 14/41; RR 0.48 [95% CI 0.30, 0.76];
(Figure Q1.6) and worsening of metabolic acidosis (no DVET vs. DVET: 12/42 vs. 3/41;
RR 0.26 [95% CI 0.08, 0.84]; (Figure Q1.7) following DVET.2
Table Q1. Summary of Results (please refer to Appendix to view forest plots of
combined studies)
OUTCOMES Measure of 95% CI Interpretation Basis
treatment
effect
All-cause mortality 0.64 0.50-0.81 Significant 6 RCTs
(Figure Q1.1)
Organ dysfunction by day 14 0.78 0.44-1.40 Not significant 1 RCT
(Figure Q1.4)
Abnormal neurological status at 0.51 0.05-5.43 Not significant 1 RCT
time of discharge
(Figure Q1.5)
Progression of thrombocytopenia 0.48 0.30, 0.76 Significant 1 RCT
(Figure Q1.6)
Worsening metabolic acidosis 0.26 0.08, 0.84 Significant 1 RCT
(Figure Q1.7)
The result of the meta-analysis on the outcome of all-cause mortality was driven
mainly by the four RCTs done from years 1971-1997.3,5,6,7 These RCTs showed dramatic
differences in treatment effect which may have been due to the fact that there were limited
modalities of treatment for sepsis during this period. This is evidenced by the high control
event rates in these RCTs (Gross 3/11, Mathur 7/10, Narayanan 18/20, Sadana
19/20).3,5,6,7 Subgroup analysis of these 4 RCTs showed significant reduction in all-cause
mortality in the DVET group RR 0.62 [95% CI 0.47, 0.82]; (Figure Q1.8) Whereas,
A14
2,4
subgroup analysis of the two more recent RCTs showed a trend towards decreased
mortality but the result was not significant (0.67 [0.43, 1.05]; (Figure Q1.9). This result
may be due to the low control event rates in the latter studies (Gunes 9/22, Aradhya
19/42).2,4
The RCTs in this review included infants admitted in neonatal intensive care units
of tertiary referral centers capable of providing high level of neonatal care. As such, the
results should be applied to similar populations only. Caution must be exercised in the
applicability of these results in other populations.
DRAFT RECOMMENDATION
Among newborns with severe sepsis, double volume exchange transfusion as an
adjunct treatment is RECOMMENDED.
Level of Evidence: High
Strength of Recommendation: Strong
A15
References
1. Abat-Senen K, Mantaring JV. Exchange transfusion for prevention of mortality in neonatal

sepsis: a meta-analysis. 2012 Jan. Updated 2019 Aug. (unpublished)
2. Aradhya AS, Sundaram V, Kumar P, Ganapathy SM, Jain A, Rawat A. Double volume
exchange transfusion in severe neonatal sepsis. Indian J Pediatr. 2015 Jun. DOI
10.1007/s12098-015-1841-0
3. Gross SJ, Filston HC, Anderson JC. Controlled study of treatment for disseminated
intravascular coagulation in the neonate. J Pediatr. 1982 Mar;100(3):445-8.
4. Gunes T, Koklu E, Buyukkayhan D, Kurtoglu S, Karakukcu M, Patiroglu T. Exchange
transfusion or intravenous immunoglobulin therapy as an adjunct to antibiotics for neonatal
sepsis in developing countries: a pilot study. Ann Trop Paediatr. 2006 Mar;26(1):39-42.
5. Mathur NB, Subramanian BK, Sharma VK, Puri RK. Exchange transfusion in neutropenic
septicemic neonates: effect on granulocyte functions. Acta Paediatr. 1993
Nov;82(11):939-43.
6. Narayanan I, Mitter A, Gujral VVIndian. A comparative study on the value of exchange
and blood transfusion in the management of severe neonatal septicemia with sclerema. J
Pediatr .1982 Jul-Aug;49(399):519-23.
7. Sadana S, Mathur NB, Thakur A. Exchange Transfusion in Septic Neonates with
Sclerema: Effect on Immunoglobulin and Complement Levels. Indian Pediatrics January
1997;34:20-25.
8. Pugni L, Ronchi A, Bizarri B, Consonni D, Pietrasanta C, Ghirardi B, Fumagalli M,
Ghirardello S, Mosca F. Exchange Transfusion in the Treatment of Neonatal Septic Shock:
A Ten-Year Experience in a Neonatal Intensive Care Unit. Int. J. Mol. Sci. 2016; 17 (695):
1-12.
A16
Q1 EVIDENCE TABLES
AMONG NEWBORNS WITH SEPSIS, HOW EFFECTIVE IS DOUBLE VOLUME EXCHANGE TRANSFUSION (DVET) AS AN ADJUNCT
TREATMENT IN DECREASING MORBIDITY AND MORTALITY?
Setting: Aradhya, Mathur, Narayanan, Sadana – India, Gross -US , Gunes - Turkey
Bibliography: Abat-Senen K, Mantaring JV. Exchange transfusion for prevention of mortality in neonatal sepsis: a meta-analysis. 2012 Jan. Updated 2019 Aug.
(unpublished)
Certainty assessment № of patients Effect
Certainty Importance
№ of Other Exchange no exchange Relative Absolute
Study design Risk of bias Inconsistency Indirectness Imprecision
studies considerations transfusion transfusion (95% CI) (95% CI)
Mortality
6 Randomisedtri not serious not serious not serious not serious strong 54/145 (37.2%) 75/125 RR 0.64 216 fewer ⨁⨁⨁⨁ IMPORTANT
als association (60.0%) (0.50 to per 1,000 HIGH
0.81) (from 300
fewer to
114 fewer)
Organ dysfunction by day 14
1 Randomised not serious not serious not serious not serious none 13/41 (31.7%) 17/42 (40.5%) RR 0.78 89 fewer ⨁⨁⨁⨁
trials (0.44 to per 1,000 HIGH
1.40) (from 227
fewer to
162 more)
Abnormal neurologic status at discharge amongst survivors

1 Randomised not serious not serious not serious not serious none 1/41 (2.4%) 2/42 (4.8%) RR 0.51 23 fewer ⨁⨁⨁⨁
trials (0.05 to per 1,000 HIGH
5.43) (from 45
fewer to
211 more)
Progression of thrombocytopenia
1 Randomisedtri not serious not serious not serious not serious strong -/41 30/42 (71.4%) RR 0.48 371 fewer ⨁⨁⨁⨁
als association (0.30 to per 1,000 HIGH
0.76) (from 500
fewer to
171 fewer)
A17
Certainty Importance
№ of Other Exchange no exchange Relative Absolute
Study design Risk of bias Inconsistency Indirectness Imprecision
studies considerations transfusion transfusion (95% CI) (95% CI)
Worsening metabolic acidosis

1 Randomised not serious not serious not serious not serious strong 3/41 (7.3%) 12/42 (28.6%) RR 0.26 211 fewer ⨁⨁⨁⨁
trials association (0.08 to per 1,000 HIGH
0.84) (from 263
fewer to 46
fewer)
A18
APPENDIX Q1
Figure Q1.1. Exchange transfusion (DVET/ET) vs no exchange transfusion, outcome:

All-cause Mortality
Figure Q1.2. Subgroup analysis exchange transfusion vs no exchange transfusion for severe
sepsis, outcome: All-cause Mortality
Figure Q1.3. Sensitivity analysis DVET vs no exchange transfusion, outcome: All-cause Mortality.
A19
Figure Q1.4. Exchange transfusion vs no exchange transfusion, outcome:
Organ dysfunction by day 14

Abnormal neurologic status at discharge amongst survivors

Progression of thrombocytopenia

Worsening metabolic acidosis
A20
Figure Q1.8. Subgroup analysis exchange transfusion vs no exchange transfusion 4 earlier
RCTs (before year 2000), outcome: All-cause Mortality
Figure Q1.9. Subgroup analysis exchange transfusion vs no exchange transfusion 2 newer

RCTs (after year 2000), outcome: All-cause Mortality
A21
Q2. Among newborns with sclerema neonatorum, how effective and safe is fresh frozen
plasma transfusion in decreasing morbidity and mortality?
SUMMARY OF EVIDENCE
The overall quality of the evidence was downgraded due to indirectness due to the
included participants were those who are very preterm. The literature search in the different
databases was not able to yield studies specific to the effectiveness of fresh frozen plasma
(FFP) in decreasing morbidity and mortality among newborns with sclerema. The
participants in the trials included, received early treatment (fresh frozen plasma within the
first 24 hours of life).
The trials reported adequate randomization procedures and allocation concealment.
However, none reported efforts to blind caregivers and given the nature of the intervention
it is unlikely this was possible. Only one trial provided data on long term
neurodevelopment4. This was the largest trial with no losses to follow up and blinded
assessment of neurodevelopment was done at two years. Another possible limitation of the
trials would be the time period that they were conducted, thus, potentially limiting the
significance of its results to current neonatal practice.
The meta-analysis examined the potential role of early volume expansion for the
prevention of morbidity and mortality in very preterm infants.1 In the subgroup analyses,
four studies randomized infants to fresh frozen plasma or no treatment with volume
expanders. Meta-analysis of three studies reporting mortality data involving a total of 654
infants found no significant difference in mortality (RR 1.05; 95% CI 0.81, 1.36).2,3,4
Furthermore, rates of severe disability (RR 0.80; 95% CI 0.48, 1.34), death or severe
disability (RR 0.94; 95% CI 0.73, 1.22) and cerebral palsy (RR 0.79; 95% CI 0.46, 1.34)
were not significantly different but with signals of harm.
Beverley, et al 2, reported a significant reduction in periventricular/intraventricular
5
hemorrhage (P/IVH). This was in contrast with the findings of Ekblad wherein there was
no found significant difference. Meta-analysis of these two studies found a non-significant
trend to reduced P/IVH in infants receiving FFP. Among the survivors examined, a group
of researchers also reported no difference in the reduction of P/IVH.4 There was also found
no significant difference in patent ductus arteriosus (PDA) in a meta-analysis of two
A22
studies.4,5 The results of the studies also show no significant difference in the rates of
pneumothorax.
In summary, the overall rate of mortality and disability were not different between
infants who received FFP compared to no “volume expanders.” With the limited evidence
available to inform the use of FFP transfusion among neonates with sclerema, it is difficult
to make definitive recommendations. Currently, there is no supporting evidence to
recommend its use.
DRAFT RECOMMENDATIONS
The use of fresh frozen plasma among newborns with sclerema neonatorum in order
to decrease morbidity and mortality is not recommended.
Level of Evidence: Low
Strength of recommendation: Weak
References
1. Osborn DA, Evans N. Early volume expansion for prevention of morbidity and mortality in
very preterm infants. The Cochrane Database of Systematic Reviews (Complete Reviews).
2004; Issue . Art. No.: CD002055. DOI: 10.1002/14651858.CD002055.
2. Beverley DW, Pitts-Tucker TJ, Congdon PJ, Arthur RJ, Tate G. Prevention of
intraventricular haemorrhage by fresh frozen plasma. Arch Dis Child. 1985;60:710-3.
3. Gottuso MA, Williams ML, Oski FA. The role of exchange transfusion in the management
of low-birth-weight infants with and without severe respiratory distress syndrome. J Pediatr.
1976; 89:279-85.
4. Group. Lancet 1996;348:229-32.The Northern Neonatal Nursing Initiative [NNNI] Trial
Group. A randomized trial comparing the effect of prophylactic intravenous fresh frozen
plasma, gelatin or glucose on early mortality and morbidity in preterm babies. Eur J Pediatr.
1996;155:580-8
5. Ekblad H, Kero P, Shaffer SG, Korvenranta H. Extracellular volume in preterm infants:
influence of gestational age and colloids. Early Hum Dev. 1991;27:1-7
A23
Q2 EVIDENCE TABLE
Fresh frozen plasma compared to control (no treatment) for newborns with sclerema neonatorum
Patient or population: newborns with sclerema neonatorum
Setting: Beverley 1985 - Australia; Guttoso 1976 - USA; NNNI 1996 - UK ; Ekblad 1991
Intervention: fresh frozen plasma
Comparison: control (no treatment)
Anticipated absolute effects* (95% CI) Certainty

Relative effect № of participants of the
Outcomes Risk with control (no Risk with fresh frozen Comments
(95% CI) (studies) evidence
treatment) plasma (GRADE)
246 per 1,000 RR 1.05 654

Death 234 per 1,000
(0.81 to 1.36) (3 RCTs)
LOW Downgraded due to indirectness
(190 to 319)
Severe neurodevelopmental 133 per 1,000 RR 0.94 408
141 per 1,000 LOW Downgraded due to indirectness
disability (103 to 173) (0.73 to 1.22) (1 RCT)
190 per 1,000 RR 0.58 120

Reduced P/IVH 328 per 1,000
(0.30 to 1.11) (2 RCTs)
(98 to 364)
463 per 1,000 RR 1.39 54

Patent ductus arteriosus 333 per 1,000
(0.91 to 2.14) (2 RCTs)
(303 to 713)
195 per 1,000 RR 0.80 73

Pneumothorax 243 per 1,000
(0.33 to 1.92) (1 RCT)
(80 to 467)
A24
APPENDIX Q2
Figure Q8.1 Mortality
Figure Q2.1 Mortality
Figure Q2.2 Severe neurodevelopmental disability
Figure Q2.3 Reduced P/IVH
A25
Figure Q2.4 Patent Ductus Arteriosus
Figure Q2.5 Pneumothorax
A26
Q3. Among asymptomatic newborns delivered meconium-stained, how effective and safe is
giving antibiotic prophylaxis in preventing sepsis and all-cause mortality?
SUMMARY OF EVIDENCE
A meta-analysis of four (4) randomized controlled trials evaluated the effectiveness

of antibiotics in reducing infection among 695 full-term infants born as singleton cephalic
presentation through meconium-stained amniotic fluid.1 Primary outcome of the trials was
culture-proven (confirmed) sepsis. There was a low risk for bias because selection,
performance, attrition, and reporting biases were all well-accounted for. One of the
limitations would be the fact that all four trials compared antibiotics versus no treatment
and did not include a placebo control which may have resulted in biased reporting. Another
limitation would be the lack of full blinding of treating clinicians. Moreover, there was an
overall low event rate for sepsis in both the intervention and control arms of the trial.
Therefore, the evidence was downgraded due to serious imprecision related to a small
sample size and the unclear risk of bias in outcome assessment.
No difference was identified in risk of infection (confirmed sepsis) among neonates
receiving antibiotics and controls following meconium exposure. All four included trials
consistently reported no significant improvement in infection reduction with antibiotic
treatment or prophylaxis despite use of different antibiotics and treatment durations.
Only one RCT in the meta-analysis contributed data for the comparison of antibiotics
versus control (no treatment with antibiotics) in asymptomatic neonates. 2 There was no
evidence that antibiotics significantly decreased the risk of sepsis in asymptomatic
neonates exposed to meconium in amniotic fluid (RR 0.76; 95% CI 0.25 to 2.34). Results
also showed no significant difference in mortality between the group given antibiotic and
the control group of asymptomatic neonates (RR 1.07; 95% CI 0.22, 5.18).
For the morbidity outcomes, suspected sepsis was confirmed in 50% of cases in
both the antibiotic and control groups. Results showed no significant difference in the
incidence of suspected sepsis between groups. Furthermore, results also showed no
statistically significant differences in the incidence of intracranial hemorrhage, azotemia,
oliguria, diarrhea and respiratory distress between groups.
The studies included in the meta-analysis did not provide data on the cost of
antibiotic prophylaxis for newborns delivered meconium-stained. Moreover, only one study
A27
reported the outcome of “duration of hospital stay until discharge among survivors (days).”
In a study, it was noted that no statistically significant differences in the mean duration of
hospital stay between neonates who received antibiotics and those in the control group
(mean difference (MD) 0.16; 95% CI -1.15, 1.47; 146 participants.3 However, this outcome
was only reported among symptomatic newborns.
Based on the available low-quality evidence, there is no difference in the risk of
sepsis following antibiotic treatment among asymptomatic neonates born through
meconium stained amniotic fluid. There is a need to generate more studies on the efficacy
and cost-effectiveness of antibiotics as prophylaxis among asymptomatic term newborns
delivered meconium stained, before recommending antibiotic prophylaxis in routine clinical
practice.
Table Q3. Summary of Results (please refer to Appendix to view forest plots)
OUTCOMES Measure of 95% Interpretation Basis
Treatment Confidence
Effect Interval
Incidence of Confirmed sepsis in first 28 days
of life RR 0.76 0.25-2.34 Not significant 1 RCT
(Fig. Q3.1)
Mortality before discharge RR 1.07 0.22-5.18 Not significant 1 RCT
(Fig. Q3.2)
Incidence of suspected sepsis RR 0.76 0.35-1.65 Not significant 1 RCT
(Fig. Q3.3)
Incidence of intracranial hemorrhage
(Fig. Q3.4) RR 0.36 0.01-8.64 Not significant 1 RCT
Incidence of azotemia (Fig. Q3.5) RR 3.20 0.13-77.73 Not significant 1 RCT
Incidence of diarrhea (Fig. Q3.6) RR 0.12 0.01-2.18 Not significant 1 RCT
Incidence of respiratory distress
(Fig. Q3.7) RR 1.18 0.81-1.72 Not significant 1 RCT
Antibiotic prophylaxis for asymptomatic newborns delivered meconium stained is

not recommended.
Strength of Recommendation: Weak
A28
References
1. Kelly LE, Shivananda S, Murthy P, Srinivasjois R, Shah PS. Antibiotics for neonates born
through meconium stained amniotic fluid. Cochrane Database Syst Rev 2017. CD006183.
Doi: 10.1002/14651858
2. Goel A, Nangia S, Saili A, Garg A, Sharma S, Randhawa VS. Role of prophylactic antibiotics
in neonates born through meconium-stained amniotic ﬂuid (MSAF) - a randomized
controlled trial. European Journal of Pediatrics 2015; 174: 237–43.
3. Basu S, Kumar A, Bhatia BD. Role of antibiotics in meconium aspiration syndrome. Annals
of Tropical Pediatrics 2007;27:107–13.
4. Lin HC, Su BH, Tsai CH, Lin TW, Yeh TF. Role of antibiotics in management of non-
ventilated cases of meconium aspiration syndrome without risk factors for infection. Biology
of the Neonate 2005;87:51–5.
5. Shankar V, Paul VK, Deorari AK, Singh M. Do neonates with meconium aspiration
syndrome require antibiotics? Indian Journal of Pediatrics 1995;62:327–
A29
Q3 EVIDENCE TABLE
Antibiotics compared to control (no antibiotics) for asymptomatic neonates born through meconium stained amniotic fluid
Patient or population: asymptomatic neonates born through meconium stained amniotic fluid
Setting: Goel 2015 - India
Intervention: empiric antibiotics
Comparison: control (no antibiotics)
Anticipated absolute effects* (95% CI) Certainty

Relative effect № of participants of the
Outcomes Risk with control (no Risk with empiric Comments
(95% CI) (studies) evidence
antibiotics) antibiotics (GRADE)
Incidence of confirmed sepsis in first 28 0 per 1,000 250

54 per 1,000 RR: 0.76 (0.25 - 2.34) LOW Downgraded due to serious imprecision
days (0 to 0) (1 RCT)
25 per 1,000 250

Mortality ( before discharge) 23 per 1,000 RR: 1.07 (0.22 to 5.18)
(1 RCT)
LOW Downgraded due to serious imprecision
(5 to 120)
0 per 1,000 250

Incidence of suspected sepsis 109 per 1,000 RR: 0.76 (0.35 - 1.65)
(1 RCT)
LOW Downgraded due to serious imprecision
(0 to 0)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval
GRADE Working Group grades of evidence

High certainty: We are very confident that the true effect lies close to that of the estimate of the effect
Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect
Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
A30
APPENDIX Q3
Figure Q3.1 Incidence of confirmed sepsis in first 28 days of life
Figure Q3.2 Mortality before discharge
Figure Q3.3 Incidence of suspected sepsis
Figure Q3.4 Incidence of intracranial hemorrhage
A31
Figure Q3.5 Incidence of azotemia
Figure Q3.6 Incidence of diarrhea
Figure Q3.7 Incidence of respiratory distress
A32
Q4. Among newborns with sepsis who improved after starting antibiotics, how effective is a
shorter course (5-day) of intravenous antibiotics compared to seven days in decreasing
treatment failure and mortality?
SUMMARY OF EVIDENCE
Treatment failure is defined a priori as recurrence of symptoms after completing a
certain duration of intravenous antibiotics.
There were no studies available comparing the original antibiotic duration (E) from
the question despite the systematic search of major electronic databases (MEDLINE via
Pubmed, CENTRAL, ClinicalTrials.gov) and hand-search of 22 articles. There were also
no studies available in local databases included in this search. However, three randomized
controlled trials were deemed indirect but relevant to our clinical question, thus, were used
as available evidence.
These randomized controlled trials were done in low-income countries1-3. All studies
included newborns in neonatal intensive care units with probable sepsis or those with
clinical signs of sepsis but had negative results of blood culture after having empirical
intravenous antibiotics.. In one study, authors included infants with birth weight over 1500g,
gestational age over 34 weeks and were observed within 7 days postnatal age.1 These
infants were randomized after 72 hours of being asymptomatic. Investigators in Iran
compared 3-day and 5-day IV antibiotics in terms of preventing treatment failure. In the
second RCT, newborns included were those with more than 1000g and with gestational
age of more than 30 weeks.2 They also started antibiotics for probable sepsis supported
by a positive C-reactive protein (CRP) test. The infants randomized in the short-course
group (between 48 hours to 96 hours) did not receive antibiotics after having negative blood
culture results. The control group received 7 days of antibiotics. Outcome in both studies
was treatment failure defined as reappearance of signs and symptoms of neonatal sepsis
diagnosed by an expert physician and supported by laboratory findings.2,3 This outcome
was observed within two weeks (14-15 days).2,3 In addition, mortality was not mentioned
as the primary outcome but was reported in the study results (Table 1). The last clinical
trial included in this review was still indirect in terms of intervention (3-day and 7-day
antibiotics). Neonates with probable sepsis identified based on clinical signs of sepsis and
positive septic screen and who were on antibiotic therapy and had sterile blood culture at
the third day of admission were observed in NICU. The investigators determined outcomes
such as successful therapeutic outcome including absence of apparent signs of sepsis
A33
ascertained by clinical examination and absence of rehospitalisation caused by sepsis.
Mortality was not also reported as outcome in this trial.
The study analysing the effects of shorter course versus 7-day IV antibiotics has
lower risk of bias than the other RCT investigating similar duration. However, combining
these studies still result to very serious methodological limitations. One study comparing
the two shorter courses of intervention also has poor methodological quality. (Refer to Q4
Evidence Table)
Table Q4. Summary of results
Outcomes Measure of 95% Confidence Interpretation Basis
Treatment Effect Interval
Treatment failure RR 0.14 0.01 - 2.63 Not significant 2 RCTs2,3

(3 days vs 7 days)
Treatment failure RR 3.00 0.13 - 70.83 Not significant 1 RCT1

(3 days vs 5 days)
Mortality (3 days RR 0.33 0.01- 7.81 Not significant 1 RCT2

vs 7 days)
Although the point estimate shows 86% reduction in treatment failure with the
shorter course, the confidence interval is wide with the worst case scenario raising a signal
of harm (Table Q4). It means that those who receive a 3-day intervention can be three
times as likely to be at risk of having treatment failure as those who receive a 7-day IV
antibiotics. Similarly, there is no sufficient evidence to say that there is a decrease by 67%
in the risk of mortality among newborns who receive IV antibiotics in an average of 3 days1.
The shorter course when compared with a five-day intervention can also be harmful but it
also has reported 87% reduction in the risk of treatment failure.1 This study has poor
methodological quality aside from its inconclusive results. There were no reported cases
of neonatal deaths in the two groups of treatment regimen.1
Given these indirect and inconclusive findings, the available evidence is not
adequate to determine the effectiveness of a shorter course of intravenous antibiotics
compared with a 7-day course among newborns with sepsis who improve after starting
antibiotics. However, this current available evidence on shorter course of antibiotics among
these newborns still showed an apparent risk of harm.
A34
Recommendations of Others
In UK guideline, the usual length of antibiotic treatment for newborns with a positive
blood culture and for those with strong suspicion of sepsis even with a negative blood
culture is 7 days. Similarly, WHO supported the Integrated Management of Childhood
Illness (IMCI) which recommends that IV antibiotics should be at least 7-10 days in infants
aged less than two months with serious bacterial infection. In US, AAP recommend when
blood cultures are sterile, antibiotics should be stopped by 36 to 48 hours of incubation
unless there is proven infection on a specific site for neonates born ≥35 weeks’ and <34
weeks’ gestation. However, persistently unstable cardiopulmonary status is not unusual
among premature infants with VLBW; it is not a sole indication for prolonged empirical
antibiotic therapy. Moreover, using abnormal laboratory tests only cannot always justify
prolonged antibiotics, particularly among preterm infants with lower risk for early-onset
sepsis (e.g. those with maternal delivery conditions affecting hematopoiesis).
A shortened 5-day course of antibiotics for newborns who improve after initial
antibiotic coverage is NOT RECOMMENDED.
Level of Evidence: Low to Very Low
References
1. Pasha YZ, Ahmadpour-Kacho M, Behmadi R and Jahangir T. 3-day versus 5-day course of
intravenous antibiotics for suspected early onset neonatal sepsis: a randomized controlled
trial. Iran J Pediatr. 2014; 24(6):673-678.
A35
2. Saini, S. S., Dutta, S., Ray, P., & Narang, A. (2010). Short course versus 7-day course of
intravenous antibiotics for probable neonatal septicemia: A pilot, open-label, randomized
controlled trial. Indian Pediatrics. 2010; 48(1), 19–24. doi:10.1007/s13312-011-0019-8
3. Prashanth SN, Rashmi N, Sandeep Patil. 3 days vs 7 days course of intravenous
antibiotics for probable neonatal sepsis. Pediatric Review: International Journal of
Pediatric Research. 2017; 4(2):xx-xx.
4. National Institute for Health and Care Excellence (NICE). Neonatal infection (early onset):
antibiotics for prevention and treatment. 2012
5. Puopolo KM, Benitz WE, Zaoutis TE, AAP Committee on Fetus and Newborn, AAP
Committee on Infectious Diseases. Management of neonates born at ≥35 0/7 weeks’
gestation with suspected or proven early-onset bacterial sepsis. Pediatrics 2018;
142(6):e20182894
6. Puopolo KM, Benitz WE, Zaoutis TE, AAP Committee on Fetus and Newborn, AAP
Committee on Infectious Diseases. Management of neonates born at <34 6/7 weeks’
gestation with suspected or proven early-onset bacterial sepsis. Pediatrics 2018;
142(6):e20182894
7. Fuchs A, Bielicki J, Mathur S, Sharland M, Van Den Anker JN. Antibiotic use for sepsis in
neonates and children: 2016 Evidence update. WHO Reviews 2016. X:1-22
A36
Q4 EVIDENCE TABLE
Date: October 20, 2019
Question: Newborns presenting with sepsis and improve after starting antibiotics receive shorter course compared to 7-day or 5-day antibiotic regimen for prevention of treatment failure and mortality
Setting: Iran1 and India2
Bibliography:
1. Pasha YZ, Ahmadpour-Kacho M, Behmadi R and Jahangir T. (2014). 3-day versus 5-day course of intravenous antibiotics for suspected early onset neonatal sepsis: a randomized controlled trial.
Iran J Pediatr; 24(6):673-678.
2. Saini, S. S., Dutta, S., Ray, P., & Narang, A. (2010). Short course versus 7-day course of intravenous antibiotics for probable neonatal septicemia: A pilot, open-label, randomized controlled trial.
Indian Pediatrics, 48(1), 19–24. doi:10.1007/s13312-011-0019-8
newborns
presenting
with sepsis
7-day or 5-day Certainty Importance
№ of Risk of Other and improve Relative Absolute
Study design Inconsistency Indirectness Imprecision antibiotic
studies bias considerations after starting (95% CI) (95% CI)
regimen
antibiotics
receive
shorter course
Treatment failure (3-day vs 7-day IV antibiotics) (assessed with: Risk ratio)
2 randomised serious a not serious not serious serious b none 0/25 (0.0%) 3/25 (12.0%) RR 0.14 103 fewer ⨁⨁◯◯
trials (0.01 to per 1,000 LOW
2.63) (from 119
fewer to 196
more)
Treatment failure (3-day vs 5-day IV antibiotics) (assessed with: Risk ratio)
1 randomised very not serious not serious serious b none 1/30 (3.3%) 0/30 (0.0%) RR 3.00 0 fewer per ⨁◯◯◯
trials serious c (0.13 to 1,000 VERY LOW
70.83) (from 0 fewer
to 0 fewer)
Mortality (3-day vs 7-day IV antibiotics)
A37
newborns
presenting
with sepsis
7-day or 5-day Certainty Importance
№ of Risk of Other and improve Relative Absolute
Study design Inconsistency Indirectness Imprecision antibiotic
studies bias considerations after starting (95% CI) (95% CI)
regimen
antibiotics
receive
shorter course
1 randomised serious a not serious not serious serious b none 0/25 (0.0%) 1/25 (4.0%) RR 0.33 27 fewer per ⨁⨁◯◯
trials (0.01 to 1,000 LOW
7.81) (from 40
fewer to 272
more)
CI: Confidence interval; RR: Risk ratio
Explanations
a. unclear blinding of personnel and attrition bias
b. wide confidence interval
c. no allocation concealment, no blinding of participants and personnel, unclear blinding of outco
A38
Q5. Among newborns with clinical sepsis, how predictive of CNS infection is blood
culture?
SUMMARY OF EVIDENCE
Neonatal meningitis is a major concern in the neonatal intensive care unit and early
detection among high risk neonates remains a priority. Most of the studies performed
lumbar puncture along with other body fluid cultures before initiating antibiotic
therapy.2,3,4 The studies were grouped into three, based on the onset of sepsis as being
early of onset, late-onset and those that looked at sepsis from delivery until 30 to 150 days
of life.
Septicemia was defined as having a positive blood culture for a bacterial pathogen.
Early-onset sepsis was defined as positive blood or CSF culture in neonates whose
specimen were obtained within 72 hours of delivery, and late-onset sepsis as occuring after
day 3 of delivery.1,3,5,6 Two studies, Hervas (1993) and Johnson (1997) used 5 and 7 days
as their cut-off for early-onset sepsis, respectively.2,9
A. Early-onset Sepsis
Based on five retrospective and prospective studies spanning 6 months to 7 years,

only two percent (2%) of clinically septic neonates diagnosed within 72 hours with positive
blood culture were positive for cerebrospinal fluid (CSF) culture.1-5 The largest study with
5,135 symptomatic and asymptomatic term infants had an incidence of 0.6% (11/1,712). 1
All asymptomatic neonates were negative for CSF culture. Two studies, which evaluated
maternal risk factors and included 431 preterm and term infants, had a prevalence rate of
1.4%.3,4 Overall specificity and negative predictive values were both 100% and the positive
likelihood ratio (+ LR) was 9.98 (95% CI 9.08-10.97). A positive CSF and blood culture is
strongly indicative of CNS infection. The authors believe that an early lumbar puncture may
lead to false negative result and puts the infant at risk for further infection. 3
B. Late-onset Sepsis
Based on three retrospective and prospective studies spanning 1 to 3.3 years and
a total of 3,141 neonatal records, the prevalence rate of CNS infection was 5% with only
3.1% (97/3,141) positive for both CSF and blood culture. 1,3,6 The negative predictive value
A39
remained high at 96%, suggesting that a negative blood culture result rules out a CNS
infection. The sensitivity across the studies ranged from 26% to 98% (overall 62%).
However, the largest study in the series was done on very low birth weight (VLBW) babies
and this may have overestimated the risks, thus affecting the Specificity, Sensitivity and
likelihood ratios.3,6 In two of the studies, lumbar puncture was routinely done on septic
neonates before initiating antibiotic therapy. 1,3 In a study, infants <7 days old were
classified under EOS, possibly underestimating the rate of meningitis for LOS.1 There was
no difference in the risk of death between septic infants with a positive and those with
negative lumbar puncture at 10% each (284/2877 vs 661/6764, OR = 1.01; 95% CI 0.87,
1.17).6
C. Combined Early and Late Onset Sepsis
There were three large retrospective studies done from 7-15 years with 12,915
neonatal records combined.7,8,9 The prevalence rate of CSF positive among blood culture-
positive infants less than 31 days old was 1.8%. The largest study with 8,596 subjects had
a prevalence of 1.07%.8 The Sn, Sp and NPV were 68%, 92% and 99%, respectively,
similar to the individual findings for early and late onset sepsis. A high LR+ (8.73; 95% CI
8.11, 9.39) showed a moderately positive likelihood of CNS infection when the blood culture
is positive in septic neonates. The large effect of the population with consistent findings
supports doing a lumbar puncture in a stable neonate with clinical sepsis to diagnose
meningitis.
A40
Recommendation Level of Evidence Strength of Recommendation
5a) Among culture-positive

newborns with early onset
sepsis, there is a 2% prevalence
of CSF culture positivity. A
Low Strong
negative blood culture does not
warrant a CSF culture.
5b) Among culture-positive

neonates with late onset sepsis,
there is 5% prevalence of CSF
culture positivity and a lumbar
Very Low Weak to Moderate
tap is recommended.
5c) Among culture-positive

infants less than 31 days old
with clinical sepsis, there is an
overall prevalence of 1.8%, and Low Strong
a lumbar puncture is
recommended if the neonate is
stable.
References
1. Schwersenski J, McIntyre L, Bauer CR. Lumbar puncture frequency and cerebrospinal fluid
analysis in the neonate. Am J Dis Child. 1991;145(1):54–58
2. Johnson CE, Whitwell JK, Pethe K, Saxena K, Super DM. Term newborns who are at risk
for sepsis: are lumbar punctures necessary? Pediatrics. 1997;99(4):E10
3. Visser VE,Hall RT. Lumbar puncture in the evaluation of suspected neonatal sepsis. J
Pediatr. 1980;96(6):1063–1067
4. Kumar P, Sarkar S, Narang A. Role of routine lumbar puncture in neonatal infection. J
Paediatr Child Health 1995;31:8–10.
5. May M, Daley AJ, Donath S, et al. Australasian study group for neonatal infections. Early
onset neonatal meningitis in Australia and New Zealand, 1992-2002. Arch Dis Child Fetal
Neonatal Ed.2005; 90:F324–7.
6. Stoll BJ, Hansen N, Fanaroff AA, et al. To tap or not to tap: high likelihood of meningitis
without sepsis among very low birth weight infants. Pediatrics. 2004;113(5):1181–1186
7. Smith PB, Garges HP, Cotton CM, Walsh TJ, Clark RH, Benjamin DK Jr. Meningitis in
preterm neonates: importance of cerebrospinal fluid parameters. Am J Perinatol. 2008;
25(7):421–426
8. Garges HP, Moody A, Cotten CM, et al. Neonatal meningitis: what is the correlation among
cerebrospinal fluid cultures, blood cultures, and cerebrospinal fluid parameters? Pediatrics.
2006;117(4):1094 –1100
9. Hervas JA, Alomar A, Salva F, Reina J, Benedi VJ. Neonatal sepsis and menmingitis in
Mallorca, Spain, 1977-1991. Clin Infect Dis. 1993; 16:719-24.
A41
A42
Q5 EVIDENCE TABLES
Table Q5.1 GRADE Evidence Profile
Table Q5.2 Summary of Results
A43
Q6. Among newborns with risk factors for neonatal infection, how accurate is each of the
following CBC parameters in diagnosing sepsis?
a. White Blood Cells (WBC) <5000
b. White Blood Cells (WBC) >30,000
c. Absolute Neutrophil Count (ANC) <1500
d. Immature to Total Neutrophil Count Ratio (IT Ratio) >0.2
e. Platelet <150,000
SUMMARY OF EVIDENCE
Based on high level of evidence, the likelihood of sepsis in a neonate with risk factors
fulfilling the category of an abnormal complete blood count is high.
A series of 13 prospective and retrospective cohort cross-sectional studies done
over a span of 39 years that included 4854 neonates less than 31 days old who fulfilled the
criteria of an abnormal complete blood count has an over-all prevalence of 10.4%
(505/4854), incidence risk of 104/1000 patients 95% CI 95.8-112.9. The prevalence rates
across all studies ranged from 0.93% to 61.4%. The positive and negative likelihood ratios
of each of the CBC parameters are summarized in Table 6.1.
In general, using a single CBC parameter is not helpful in diagnosing sepsis. Each
of the parameter has low positive and negative likelihood ratios (positive LR < 3.0; negative
LR > 0.3) which indicate that the likelihood of having sepsis will not change using a CBC
parameter.
Across the five parameters the PPVs in general are also low. The NPVs are better
across the 5 and show that the ANC <1500 has the highest NPV.
Although the quality of evidence is high and there is consistency among the findings
from 39 years of accumulated data, the predictive values of tests are not very useful across
populations with different prevalences of sepsis.
Among newborns with risk factors for neonatal infection, a single complete blood
cell count done within the 6th-24th hour of life showing at least one abnormal CBC
parameter (WBC <5000, WBC >30,000, ANC <1500, IT ratio >0.2 or platelet count
<150,000) should not be used in diagnosing sepsis.
Level of Evidence: High

Strength of Recommendation: Strong
A44
References
1. Philip AGS, Hewitt JR. Early diagnosis of neonatal sepsis. Pediatrics. 1980;65(5):1036–41.
2. Berger C, Uehlinger J, Ghelfi D, Blau N, Fanconi S. Comparison of C-reactive protein and
white blood cell count with differential in neonates at risk for septicaemia. Eur J Pediatr.
1995;154(2):138–44.
3. Rodwell RL. Early diagnosis of neonatal sepsis using a hematologic scoring system. Fetal
and Neonatal Medicine. 1983;(November).
4. Escobar GJ, Li DK, Armstrong MA, Gardner MN, Folck BF, Verdi JE, et al. Neonatal sepsis
workups in infants ≥2000 grams at birth: A population-based study. Pediatrics. 2000; 106(2
I):256–63.
5. Blommendahl J, Janas M, Laine S, Miettinen A, Ashorn P. Comparison of procalcitonin with
CRP and differential white blood cell count for diagnosis of culture-proven neonatal sepsis.
Scand J Infect Dis. 2002;34(8):620–2.
6. Manucha V, Rusia U, Sikka M, Faridi MMA, and Madan N. Utility of hematologic parameters
and C-reactive protein in the detection of neonatal sepsis. J Pediatr. Child Health. 2002;
38, 459-464.
7. Ottolini MC, Lundgren K, Mirkinson LJ, Cason S, Ottolini MG. Utility of complete blood count
and blood culture screening to diagnose neonatal sepsis in the asymptomatic at risk
newborn. Pediatr Infect Dis J. 2003;22(5):430–4.
8. Narasimha A, Harendra Kumar ML. Significance of hematological scoring system (HSS) in
early diagnosis of neonatal sepsis. Indian J Hematol Blood Transfus. 2011;27(1):14–7.
9. Mondal SK, Nag DR, Bandyopadhyay R, Chakraborty D, Sinha SK. Neonatal sepsis: Role
of a battery of immunohematological tests in early diagnosis. Int J App Basic Med Res
2012;2:43-7.
10. Makkar M, Pathak R, Garg S, Gupta C, Mahajan N. Performance evaluation of hematologic
scoring system in early diagnosis of neonatal sepsis. J Clin Neonatol. 2013;2(1):25.
11. Chacha F, Mirambo MM, Mushi MF, Kayange N, Zuechner A, Kidenya BR, et al. Utility of
qualitative C- reactive protein assay and white blood cells counts in the diagnosis of
neonatal septicaemia at Bugando Medical Centre, Tanzania. BMC Pediatr. 2014;14(1):1–
8.
12. Saied DA. Can we rely on the neutrophil left shift for the diagnosis of neonatal sepsis? Need
for re-evaluation. Egypt Pediatr Assoc Gaz. 2018;66(1):22–7. Available from:
https://doi.org/10.1016/j.epag.2017.12.002
13. Saboohi E, Saeed F, Khan RN, Khan MA. Immature to total neutrophil ratio as an early
indicator of early neonatal sepsis. Pakistan J Med Sci. 2019;35(1):241–6.
A45
Q6 EVIDENCE TABLES
TABLE Q6.1. Summary of Results
No of Studies Total TP FP FN TN Sn Sp PPV NPV PLR NLR Accuracy

Number
of
patients
WBC <50008 1208 51 135 129 863 28.33% 86.47% 27.42% 87% 2.09 0.83 77.59%
(21.88- (84.19-88.53) (22.2-33.3) (85.88- (1.58-2.77) (0.75-0.91) (75.1-79.94)
35.52) 88.03)
WBC >30,0005 786 82 164 75 463 52.23% 73.84% 33.33% 86.06% 2.0 0.65 69.52%
(44.1-60.25) (70.22-77.24) (29.06-37.09) (83.89- (1.64-2.44) (0.55-0.77) (66.16-72.72)
87.89)
ANC <15002 350 10 75 30 235 25% 75.81% 11.76% 88.68% 1.03 0.99 70.00%
(12.69-41.2) (70.64-80.47) (7.0-19.11) (86.63- (0.58-1.83) (0.82-1.2) (64.9-74.76)
90.45)
IT Ratio >0.29 1527 178 265 181 902 49.58% 77.29% 40.18% 83.29% 2.18 0.65 70.77%
(44.29- (74.78-79.67) (36.67-43.80) (81.74- (1.88-2.53) (0.59-0.73) (68.42-73.05)
54.88) 84.73)
Platelet 410 75 89 63 181 54.35% 67.04% 45.73% 74.18% 1.65 0.68 62.75%
<150,0005 (45.66- (61.08-72.61) (40.13-51.44) (70.16- (1.31-2.07) (0.56-0.83) (57.85-67.45)
62.85) 77.83)
A46
Table Q6.2 GRADE Evidence Profile (WBC <5000)
Table Q6.3 GRADE Evidence Profile (WBC >30,000)
A47
Table Q6.4 GRADE Evidence Profile (ANC <1500)
A48
Table Q6.5 GRADE Evidence Profile (IT >0.2)
A49
TABLE Q6.6 GRADE Evidence Profile (Platelet count <150,000)
A50
Q7. Among asymptomatic newborns delivered to mothers with maternal UTI within two weeks
prior to delivery, how useful is a neonatal sepsis calculator in predicting occurrence of sepsis
and all-cause mortality?
SUMMARY OF EVIDENCE
There were no studies found to directly answer the question and the specified
outcomes (i.e. predicting sepsis occurrence and all-cause mortality) despite systematic
search in major databases using MeSH terms and free text. The search yielded two (2)
systematic reviews and meta-analyses that looked into the comparison of the use of an
EOS calculator versus conventional/standard management strategies among newborns
(i.e. >34 weeks age of gestation). However, these did not specify maternal UTI as a risk
factor and used reduction of antibiotic therapy as the specified outcome. The variables (risk
factors) used in deriving the multivariate model for the EOS calculator were gestational
age, group B Streptococcus (GBS) status, rupture of membrane (ROM) time, highest
intrapartum temperature and intrapartum antibiotic prophylaxis.
In the first meta-analysis, Achten and colleagues1 reported a relative risk of antibiotic
use of 52% (95% CI 53%, 59%) among newborns with the use of the EOS calculator. In
this meta-analysis, the proportions of missed cases of EOS were comparable between the
EOS-guided group (28%) versus the conventional management strategies group (29%)
(pooled OR, 0.96; [95% CI 0.26, 3.52]; P = .95). In the second meta-analysis, Deshmukh
also demonstrated a reduction in antibiotic use among newborns > 34 weeks with an OR
of 0.22 (95% CI 0.14, 0.36); p<0.00001) in favor of EOS calculator versus standard
management strategies2.
Currently, there is no data to support the use of an early-onset neonatal sepsis
calculator in women with UTI within 2 weeks of delivery to predict occurrence of sepsis and
all-cause mortality. This is an identified research gap (maternal UTI as risk factor for EOS)
and the applicability/adaptability of the EOS calculator to our Filipino mothers and
neonates.
No recommendation can be made at this time for use of the early onset neonatal
sepsis calculator for neonates delivered to mothers with UTI within two weeks prior to
delivery.
A51
Maternal UTI was not included as a risk factor for EOS in multivariate models used
in EOS calculators reviewed thus there is no direct evidence currently available to support
use of an early-onset neonatal sepsis calculator, compared to conventional management
therapies, to predict occurrence of sepsis and all-cause mortality in this population of
neonates
Level of evidence: Very Low
References
1. Achten N, Klingenberg C, Benitz W, Stocker M, Schlapbach L, Giannoni E et al. Association
of use of the neonatal early-onset sepsis calculator with reduction in antibiotic therapy and
safety. JAMA Pediatrics. 2019173(11):1032-1040
2. Deshmukh M, Mehta S, Patole S. Sepsis calculator for neonatal early onset sepsis – a
systematic review and meta-analysis. The Journal of Maternal-Fetal & Neonatal Medicine.
2019:1-9.
A52
Q8. Among asymptomatic newborns, what is the accuracy of CBC and a single quantitative CRP
in the diagnosis of sepsis?
SUMMARY OF EVIDENCE
Searching literature in electronic databases (e.g. MEDLINE, clinicaltrials.gov,
Cochrane database, TRIP database) using MeSH terms and free texts and hand-searching
relevant published clinical practice guidelines, there is one (1) study found to directly
answer the question and the specified outcome.
A cross-sectional study1 showed that any CRP done within three (3) days of birth in
combination with WBC count showed better diagnostic accuracy in neonates with clinical
suspicion of neonatal sepsis. In this research, Chacha and colleagues included all
neonates with clinical suspicion of neonatal sepsis according to the WHO criteria: (1)
admitted at NICU and premature units. Combination of WBC and CRP, the sensitivity,
specificity, PPV and NPV is 90.3%, 50.2%, 31.6% and 95.3% respectively. Given these
estimates, the positive and negative likelihood ratios (LRs) are 1.813 and 0.053,
respectively.
The search also yielded two studies2,3 that looked into the diagnostic accuracy of
CRP and CBC among newborns. However, these studies included neonates with
diagnosed sepsis and/or with late-onset sepsis (P) and qualitative determination of CRP
(I), thus, indirectly addressing the question at hand.
The study by Beltempo and colleagues2 showed that among very low birth weight
neonates with late-onset sepsis, the combination of CBC at T0 and qualitative CRP at T24
offered better diagnostic accuracy in neonatal sepsis. Combination of WBC at T0 and CRP
at T24, the sensitivity, specificity, PPV and NPV is 88.0%, 60.0%, 46.0% and 93.0%
respectively. Using these estimates, the positive and negative likelihood ratios are 2.20
and 0.20, respectively.
The other study3 showed that qualitative CRP alone or in combination with WBC
count resulted to better diagnostic accuracy in neonatal sepsis. Combination of WBC and
qualitative CRP, the sensitivity, specificity, PPV and NPV is 78.5%, 83%, 60% and 93%,
respectively. Moreover, the positive and negative LRs are 4.618 and 0.259, respectively.
A53
In well-appearing infants with risk factors for sepsis, management options include a
combination of monitoring of clinical signs, abnormal laboratory parameters, blood culture
determination and empiric antibiotics. The essential examinations done in a sepsis workup
includes a complete blood count (CBC) and blood culture and sensitivity (CS). Their results
are augmented by the additional determination of concentrations of acute phase reactants,
of which C-reactive protein (CRP) and procalcitonin have been the most widely studied.
CRP concentration is low at birth, and may be normal during the initial stages of an
infection. CRP levels may rise at least 12 hours after the onset of an infection, and its
sensitivity increases with serial determinations 24-48 hours after the onset of symptoms.4
Thus, CRP determination at birth or at initial presentation of infection is not recommended.5
Most CPGs from the American Association of Pediatrics (AAP) and the National Institute
for Health and Care Excellence (NICE) recommend serial determination, as opposed to a
single CRP determination, for monitoring the response to treatment and for ruling out an
infection because the diagnostic accuracy of a single CRP at the time of initial investigation
is poor.4, 6-7 However, its reliability increases in the 24-48 hours after birth or after the onset
of infection. At this time, a single CRP has 93% sensitivity for probable sepsis5, and this
increases further with a repeat determination at 48 hours. These guidelines likewise stated
that serial CRP in the first 48 hours was the marker of choice in the absence of clinical
symptoms since it has negative predictive accuracy of 99.7% and a negative likelihood
ratio of 0.15 for proven neonatal sepsis.4 Systematic reviews and other studies on the
likelihood ratios for leucocyte indexes and CRP to predict sepsis have too much
heterogeneity among studies to arrive at recommendations on an ideal test or combination
of tests.8
Combination of CBC and any single quantitative CRP may be performed in

diagnosing sepsis among asymptomatic newborns.
Level of evidence: Low

Strength of recommendation: Weak
A54
References
1. Chacha F, Mirambo M, Mushi M, Kayange N, Zuechner A, Kidenya B et al. Utility of
qualitative C- reactive protein assay and white blood cells counts in the diagnosis of
neonatal septicaemia at Bugando Medical Centre, Tanzania. BMC Pediatrics. 2014;14(1).
2. Beltempo M, Viel-Thériault I, Thibeault R, Julien A, Piedboeuf B. C-reactive protein for late-
onset sepsis diagnosis in very low birth weight infants. BMC Pediatrics. 2018;18(1).
3. Sorsa, A. (2018). Diagnostic Significance of White Blood Cell Count and C-Reactive Protein
in Neonatal Sepsis; Asella Referral Hospital, South East Ethiopia. The Open Microbiology
Journal, 12(1), pp.209-217.
4. Richard A. Polin, MD and the Committee on Fetus and Newborn. Clinical report:
Management of neonates with suspected or proven early-onset bacterial sepsis. Pediatrics
2012;129;1006.
5. Auckland District Health Board Newborn Services Clinical Guideline. C-Reactive Protein
(CRP) Indications and Interpretation. www.adhb.govt.nz
6. Brown JVE, Meader N, Cleminson J, McGuire W. C-reactive protein for diagnosing late-
onset infection in newborn infants. Cochrane Database Syst Rev. 2019;1:CD012126. Epub
2019 Jan 14
7. National Institute for Health and Care Excellence (NICE) . Neonatal infection (early onset):
antibiotics for prevention and treatment. Published 22 August 2012.
www.nice.org.uk/guidance/cg149
8. Bedford Russell AR, Kumar R. Early onset neonatal sepsis: diagnostic dilemmas and
practical management. Arch Dis Child Fetal Neonatal Ed. 2015;100:F350–F354.
A55
Q8 EVIDENCE TABLE
Reference Study Design, Location Participants (N) Synopsis Major outcomes
1. Chacha 2013 Analytical cross sectional N = 305 Immunopak done at at 24 hours, 48 The positive CRP assay was observed in 67 (22.0%), 80 (26.2%)
study hours and 88 (28.9%) of neonates on day 1, 2 and 3 respectively; with
Neonates with clinical and 72 hours after admission any CRP positive occurred in 104 (34.1%) of neonates. The
Tanzania suspicion of neonatal sensitivities of CRP assay in the diagnosis of septicaemia
sepsis Blood culture done on admission using culture as gold standard on day 1, 2, 3 and any positive
according to WHO were 40.4%, 53.2%, 54.8% and 62.9% respectively.
criteria admitted at CBC done on admission While specificities were 82.7%, 80.7%, 77.8% and 73.3%
NICU and respectively. Higher sensitivity of 75% was observed when
premature Unit CRP was used to diagnose gram negative septicaemia
compared to 50% that was observed in the diagnosis of
gram positive septicaemia. WBC count of ≥13 × 109 /L had
sensitivity and specificity of 64.5% and 66.7% respectively with
area under the curve of 0.694. When the any positive CRP and
WBC of ≥13 × 109 /L were used the sensitivity increased to
90.3% with specificity of 50%. Neonates with septicaemia due
to gram negative bacteria were significantly found to have
higher rates of positive CRP than neonates with gram positive
septicaemia and with negative culture (p < 0.001, OR 8.2, 95 CI;
2.9-26).
2. Beltempo 5-year retrospective N = 416 VLBW infants CRP and CBC were drawn at time of At T0, combining the CBC and the CRP had the highest
2018 cohort with late onset sepsis initial blood culture (T0), at 16–24 h sensitivity of 66% (95% confidence interval [CI], 58–73)
(>3 days old at initial (T24) and 40–48 h (T48) after. compared to both individual tests for predicting late onset
sepsis evaluation) sepsis. At T24, CRP’s sensitivity was 84% (95% CI, 78–89) and
Symptomatic patient was statistically higher than the CBC’s 59% (95% CI, 51–67).
The combination of CBC at T0 and CRP at T24 offered the
greatest sensitivity of 88% (95% CI, 82–92) and negative
predictive value 93% (95% CI, 89–96), with fewer samples,
compared to any other combination of tests. The area under
the curve for the change in the white blood cell count from T0
to T24 was 0.82.
At initial sepsis evaluation (T0), both CBC and CRP should be
performed to increase sensitivity. A highly negative predictive
value is reachable with only two tests: a CBC at T0 and a CRP a
T24.
A56
3. Sorsa 2018 Prospective cross- N = 303 neonates with WBC done The Sensitivity and Specificity of CRP were 65.6% and 78% with
sectional study risk factors AND signs Quali CRP done after 24th hour of life positive predictive value (PPV) and negative predictive value
Ethiopia of clinical sepsis Blood CS done (NPV) of 42% and 91%, respectively. On the other hand; the
Sensitivity and Specificity of WBC were 59.5% and 79.6% with
PPV and NPV of 52% and 64.5% respectively, and area under
the curve of 0.686 (Fig. 1). When
CRP is combined with WBC count the sensitivity and specificity
became much better (Table 2).
From this study finding, it can be concluded that CRP alone or
in combination with WBC count demonstrated a better
screening output in neonatal sepsis evaluation for timely
management decision.
A57
Q9. Among women near term, how effective is universal culture-based screening in
preventing early onset GBS sepsis?
SUMMARY OF EVIDENCE
There are no randomized controlled trials found in the systematic literature search
to answer this question. However, two observational studies1,2 which were included in a
meta-analysis3 provided a very low level of evidence.
In an observational study done in Japan,1 only six neonates from the 710 mothers
screened developed early-onset GBS diseasewhile two newborns from the 20 unscreened
mother had the disease. The computed odds ratio (OR) from this retrospective cohort study
was 0.07 (95% CI 0.01, 0.41; p-value = 0.041) indicating benefit or a decrease in odds of
developing early-onset GBS sepsis by 93%. However, there were no clear statement on
method of culture and timing of screening among these pregnant women. In another
retrospective cohort (United Kingdom) 2, in the prescreening period, early onset GBS rate
was 0.99/1000 live births (25/25276). In the screening period, the rate was 0.33/1000 and
in the subset of mothers actually screened, the rate was 0.16/1000 live births (1/6309).
Despite the low point estimate (OR =0.22), there was no significant benefit suggested by a
wide confidence interval with a signal of harm (95% CI 0.02, 2.44).
Based on these results, further research is needed to determine the effectiveness
of universal culture-based screening among women near term in preventing early-onset
GBS sepsis.
Therefore, based on the very low level of evidence and inconclusive results,
universal culture-based screening among women near term is not recommended in
preventing early-onset GBS sepsis.
In 2015, the incidence of EOGBS in United Kingdomand Ireland was 0.57/1000
births (517 cases), a significant increase from the previous surveillance undertaken in 2000
where an incidence of 0.48/1000 was recorded. Despite this increase, the Royal College
of Obstetricians and Gynaecologists (RCOG) does not recommend universal bacteriologic
screening of pregnant women. 4
Unlike the RCOG, The American College of Obstetrics and Gynecologists
recommends performing universal GBS screening between 36 0/7 and 37 6/7 weeks of
A58
gestation. (2017) This new recommended timing for screening provides a 5-week window
for valid culture results that includes births that occur up to a gestational age of at least 41
0/7 weeks. Key component of the program is targeted IV intrapartum antibiotic prophylaxis
in neonates born to women with positive antepartum GBS cultures and women with other
risk factors for intrapartum GBS colonization.5
Universal culture-based screening of women near term should not be performed to

prevent early-onset GBS disease.
Level of Evidence: Very low
References
1. Toyofuku M, Morozumi M, Hida M, Satoh Y, Sakata H., Shiro H, Iwata S. Effects of
intrapartum antibiotic prophylaxis on neonatal acquisition of Group B Streptococci. The
Journal of Pediatrics. 2017; 190, 169–173.e1
2. Gopal Rao G, Nartley G, McAree T, et al. Outcome of a screening programme for the
prevention of neonatal invasive early-onset group B streptococcus infection in a UK
maternity unit: an observational study. BMJ Open. 2017;7:e014634
3. Da Silva HD, & Kretli Winkelströter L. Universal gestational screening for Streptococcus
agalactiae colonization and neonatal infection — A systematic review and meta-analysis.
Journal of Infection and Public Health. 2019; 12:479-481.
4. Hughes RG, Brocklehurst P, Steer PJ, Heath P, Stenson BM on behalf of the Royal College
of Obstetricians and Gynaecologists. Prevention of early-onset neonatal group B
streptococcal disease. Green-top Guideline No. 36. BJOG. 2017; 124:e280–e305.
5. American College of Obstetricians and Gynecologists. Committee Opinion: Prevention of
Group B Streptococcal Early Onset Disease in Newborns. American College of Obstetrics
and Gynecology.2019;134 (1):xx
A59
Q9 EVIDENCE TABLE
Universal screening among pregnant women near term compared to no universal screening between 35 to 37 weeks' gestation for prevention
of early-onset Group B streptococcus disease in their newborns
Setting: UK1, Japan2
Bibliography:
1. Gopal Rao G, Nartley G, McAree T et al. (2017). Outcome of a screening programme for the prevention of neonatal invasive early-onset group B
streptococcus infection in a UK maternity unit:an observational study. BMJ Open 2017;7:e014634.
2. Toyofuku M, Morozumi M, Hida M, Satoh Y, Sakata H., Shiro H, Iwata S. (2017). Effects of Intrapartum Antibiotic Prophylaxis on Neonatal Acquisition of
Group B Streptococci. The Journal of Pediatrics, 190, 169–173.e1. doi:10.1016/j.jpeds.2017.07.039
no
universal
universal
screening
screening Certainty Importance
№ of Risk of Other at Relative Absolute
Study design Inconsistency Indirectness Imprecision among
studies bias considerations
pregnant
between (95% CI) (95% CI)
35 to 37
women
weeks'
gestation
Early-onset Group B streptococcus (GBS) disease
2 observational very serious b not serious serious c none 7/7019 4/2809 not ⨁◯◯◯
studies serious (0.1%) (0.1%) estimable VERY
a LOW
CI: Confidence interval
Explanationsa. No statistical adjustments were made for the prognostic factors. In the study done by Gopal Rao and colleagues in 2017, follow-up was unclear and outcome measurement
of those who were not screened were also not stated even if events were reported.
b. Outcomes were inconsistent in Gopal Rao's work. They primarily intended to determine the effectiveness of the screening program. Focusing the study on the period 2014 and 2015 from a cohort of
pregnant women that agreed and refused to be screened would have been sufficient. However, their study focused on comparison of two incidence rates of two different periods: a period of no screening
(2009-2013) and a period when screening is an option (2014 and 2015).
c. A cohort study done in UK resulted to imprecise results computed in odds ratio. Although the point estimate declared a benefit of reduction in early-onset GBS disease by almost 80%, we cannot deny
the risk of harm because the interval ranged from 0.02 to 2.44. The other study included in this review did not also provide evidence of decrease in risk of having EOD among those neonates whose
mothers had underwent universal screening.
A60
Q10. Among newborns with no risk factors for infection presenting with fever alone, what is
the likelihood of them having sepsis?
SUMMARY OF EVIDENCE
Based on low to moderate level of evidence, the likelihood of sepsis in an otherwise

healthy-looking infant who presented with isolated fever was low, especially when the
neonate was: 1) previously healthy with no prenatal or natal antibiotic treatment; 2) no
chronic illness; 3) born term; 4) did not have any evidence of skin, bone, joint, or ear
infection, and; 5) laboratory values and blood culture were normal. 1,3,4
Seven prospective and retrospective cohort studies over a span of 1.5 to 10 years
included 1,894 neonates less than 31 days old who presented to the emergency room with
a rectal temperature of ≥38°C. The standard laboratory screening tests for sepsis included
white blood cell count (WBC), absolute neutrophil count (ANC), quantitative C-reactive
protein determination, urine analysis and culture of blood, CSF, urine and other body fluids.
Sepsis was considered present if the WBC <5000/mm or >15000/mm and/or ANC
>10000/mm and /or CRP >20mg/L and/or positive urinalysis.2,5-7 Systemic bacterial
infection (SBI) was defined when a bacterial pathogen was isolated from a culture of the
CSF, blood, urine, joint fluid, stool, pus or other body fluid. 1,3,4,5,6,7 Three studies included
radiographic studies in their criteria for SBI.2,6,7 These cohort studies reported an overall
prevalence of sepsis of 20% (374/1894), 95%CI 18-22%. The prevalence rates across all
studies were almost similar, ranging from 12.6% to 25.3%.
Only one study specifically used blood culture to diagnose sepsis among the
infants1. This 36-month consecutive cohort (Pennsylvania) included 254 febrile, 3-to-28-
day old neonates who presented to the ER with rectal temperatures of 38°C. Of the 254
febrile neonates, 109 (42.9%) were identified as at low risk for SBI. Of these 109, five low-
risk neonates (4.6%) had bacterial diseases while the 104 (95.4%) had either a viral
syndrome (n=87), gastroenteritis (n=14), bronchiolitis (n=1) and viral encephalitis (n=2).
Thirty-two neonates out of those 254 (12.6%) had SBI (27 were positive for the screening
test and 5 low-risk infants). The remaining 222 neonates (87.4%) were culture-negative.
The study identified five low-risk neonates who turned out to be culture-positive, hence, the
recommendation to do a complete evaluation with empirical antibiotic treatment. 1
A61
The negative predictive value (NPV) was high at 95.4% (95% CI 90-98%) while the
sensitivity (Sn), specificity (Sp) and positive predictive value (PPV) were 84.4%, 46.8%,
and 18.6%, respectively.
Six studies had almost similar protocols, which assessed the reliability and
usefulness of the low risk criteria (LRC) in identifying the febrile neonates unlikely to have
sepsis. LRC was based on the parameters mentioned above plus laboratory findings such
as normal WBC (5,000-15,000 with < 1,500 bands), ESR <30mm/hr, urinalysis with < 1
WBC/hpf, CRP <20mg/L, and normal stools (if no pus cells). Sepsis was generally defined
as having a bacterial pathogen isolated from culture of CSF, blood, urine, joint, stool or
other body fluids with ≥105 colonies/ml of a single pathogen.
The studies ranged from 1.5 years to 10 years of accumulated data on 1,640
neonates. The prevalence rate or pre-test probability was 21% (344/1,640) ranging from
16.4% to 25.3%. The studies showed similar findings regarding the accuracy of the LRC in
febrile neonates. The overall sensitivity (sn) was 93% (319/344, CI 89-95%) and an NPV
of 97% (744/769, CI 95-98%). The negative likelihood ratio (LR) was 0.13 (95%CI 0.1-0.16)
while the positive LR was 2.18 (95%CI 2.0-2.37). The studies, done from 1992 to 2009,
had consistently high NPV and Sn.2-7
One study had a higher incidence of sepsis, particularly among males due to the
practice of ritual circumcision at day 8 which may have led to more cases of UTI.6 The
investigator observed that UTI was the most commonly missed source of systemic bacterial
infection among neonates who met the LRC.5 The relatively high Sn and low negative LR
supported that otherwise healthy-looking but febrile neonates are at low risk for bacterial
infection and therefore can be managed safely without antibiotics. There were no identified
local studies on the use of low risk criteria as a tool for predicting sepsis among healthy-
looking newborns presenting with fever. Moreover, the practice of doing a sepsis panel
(culture of all possible fluid sources) poses economical and pragmatic concerns in the local
setting.
No local studies were available.
While the quality of available evidence is low given that most of these were
observational and not controlled, there is consistency among the findings from 17 years of
accumulated data. The variability lie in the modified low risk criteria and laboratories
performed on the neonates and the concomitant action done at the emergency room during
A62
the initial presentation of the febrile infant. Given these findings, an otherwise healthy
neonate presenting with isolated fever and fulfilling the low risk criteria (LRC), has 20%
chance of developing sepsis.
Among healthy-looking newborns presenting with fever alone, 20% of them will
develop sepsis. If the low risk criteria (LRC) has been satisfied, a negative blood culture
suggests that it is safe to discontinue antibiotics. If the blood culture is positive, antibiotics
should be continued.
References
1. Baker MD, & Bell LM. Unpredictability of serious bacterial illness in febrile infants from birth
to 1 month of age. Archives of Pediatrics & Adolescent Medicine. 1999; 153(5), 508- 511.
2. Bressan S, Andreola B, Cattelan F, Zangardi T, Perilongo G, & Da Dalt L. Predicting severe
bacterial infections in well-appearing febrile neonates: laboratory markers accuracy and
duration of fever. The Pediatric Infectious Disease Journal. 2010; 29(3), 227-232.
3. Chiu CH, Lin TY, & Bullard MJ. Application of criteria identifying febrile outpatient neonates
at low risk for bacterial infections. The Pediatric Infectious Disease Journal. 1994; 13(11),
946-949.
4. Chui CH, Lin TY, & Bullard MJ. Identification of febrile neonates unlikely to have bacterial
infections. Pediatr Infect Dis J. 1997; 16(1):59-63.
5. Marom R, Sakran W, Antonelli J, Horovitz Y, Zarfin Y, Koren A, & Miron D. Quick
identification of febrile neonates with low risk for serious bacterial infection: an observational
study. Archives of Disease in Childhood-Fetal and Neonatal Edition. 2007; 92(1), F15-F18.
6. Schwartz S, Raveh D, Toker O, Segal G, Godovitch N, & Schlesinger Y. A week by-week
analysis of the low-risk criteria for serious bacterial infection in febrile neonates. Archives of
Disease in Childhood. 2009; 94(4), 287-292.
7. Zarkesh M, Hashemian H, Momtazbakhsh M, & Rostami T. Assessment of febrile neonates
according to low risk criteria for serious bacterial infection. Iranian Journal of Paediatrics.
2011; 21(4), 436.
8. Esposito S, Rinaldi VE. et. al. Approach to neonates and young infants with fever without a
source who are at risk for severe bacterial infection. Mediators of Inflammation.2018; 11
pages. Doi.org/10.1155/2018/4869329.
A63
Q10 EVIDENCE TABLES
Table Q10.1. GRADE Evidence Profile
Table Q10.2. Summary of Results
LRC – low risk criteria

Sn – sensitivity
Sp – specificity
NPV – negative predictive value
PPV – positive predictive value
LR – likelihood ratio
A64
Q11. Among newborns with no risk factors presenting with isolated jaundice, what is their
likelihood of developing sepsis?
SUMMARY OF EVIDENCE
Based on low level of evidence, the likelihood of developing sepsis in an otherwise

healthy-looking newborns presenting with isolated jaundice is low, especially when the
neonate: 1) has no other risk factors for developing sepsis, 2) has negative blood culture
results, 3) has negative urine cultures or 4) has negative CSF cultures
1,5, 2,3
A series of five studies (two prospective two retrospective cohort studies and
a case series 4) included 774 neonates presenting with jaundice with 122 healthy patients
as controls. Neonates less than 31 days old who were admitted to the ICU and
readmissions presented with jaundice with no other risk factors nor other presenting signs
and symptoms, but with positive blood culture, urine culture and CSF studies. The over-all
prevalence rate of jaundice proven by culture studies was 57.60% (516 jaundiced out of
896 total patients included in the studies) with an odds ratio of 27.65 (95% CI 3.77, 202.76).
The prevalence rate of developing sepsis is 3.9% (35 true positives/896 total patients).
Investigators in three studies used blood, urine and CSF cultures.1-3 On the other
hand, either blood and urine culture only or urine cultures alone in the remaining two
studies.4,5 The case series yielded a prevalence rate of 80% (4/5) from the urine culture.4
1-3
Records review and observations were used in all the three older studies in
analyzing the patient’s characteristics and factors while a comparison of two groups of
4,5
healthy-looking neonates in terms of presence of UTI in more recent studies to
determine the diagnosis of jaundice. There were no deaths reported in the studies.
Antibiotics were given to those who had positive urine and blood cultures, and who
presented with symptoms. Notably, UTI was the most common infection associated with
idiopathic jaundice.
While the quality of evidence is low given that most of the studies were observational
and not controlled, there is consistency among the findings from 45 years of accumulated
data. Thus given these findings, an otherwise healthy neonate presenting with isolated
jaundice has 3.9% probability of developing sepsis.
A65
Among newborns with no risk factors for infection presenting with isolated jaundice,
the risk of sepsis is 3.9%.

References
1. Chavalitdhamrong PO, Escobedo MB, Barton LL, Zarkowsky H and Marshall RE.
Hyperbilirubinemia and bacterial infection in the newborn, a prospective study. Arch
Dis Child. 1975; 50: 652-654.
2. Escobedo MB, Barton LL, Zarkowsky H and Marshall RE. The frequency of jaundice
in neonatal bacterial infection. Clinical Pediatrics. 1974; 13 (8): 656-657.
3. Maisels MJ and Kring E. Risk of sepsis in newborns with severe hyperbilirubinemia.
Pediatrics. 1992;90(5): 741-743.
4. Afzal N, Qadir M, Qureshi S, Ali R, Ahmel S, and Ahmed K. Urinary tract infection
presenting as jaundice in neonates. J Pak Med Assoc. 2012; 62(7): 735-737.
5. Shahian M, Rashtian P, and Kalani M. Unexplained neonatal jaundice as an early
diagnostic sign of urinary tract infection. International Journal of Infectious Diseases.
2011; 6:e487-e490.
A66
Q11 EVIDENCE TABLE
Author Study Population Criteria for infection Outcome Prevalence rate

Type
Chavalitdhamrong Prospective 69 term and Bacterial isolate for blood, suprapubic  3% (2 of 69) had bacterial infection
1975 preterm (ICU) aspirated urine, CSF, >105 org/mm3 with hyperbilirubinemia 0%
clean voided specimen
 0 of 69 blood or CSF cultures (+)
bacterial isolate
Escobedo 1974 Retrospective 274 term and (+) blood culture drawn aseptically  6% (16 of 274) 5.84%
preterm (ICU) Supraprubic aspirated or clean voided Had proven bacterial infections which
urine >105 org/mm3, lead to unexplained
(+) CSF viable bacteria hyperbilirubinemia
 6 of 16 (Asymptomatic) jaundice as
single sign of infection
 10/16 (symptomatic)
Maisels 1992 Retrospective 306 readmission Blood culture  19% (59/306) performed blood 0%
within 21 days Urine culture culture (result (-))
old diagnosed CSF culture  41.2% (126/306) performed urine
with Indirect culture (result (-))
HB*  4.9% (15/306) performed CSF
culture(result (-))
 ALL CULTURES ARE NEGATIVE
 2 (symptomatic)
Afzal 2012 Case series 5 patients Urine Culture  5 patients (+) UTI 80%
Blood Culture  4 (+) Urine culture
 1 (-) Urine culture but TLC 16/hpf
 All 5 patients (-) blood cultures
A67
Shahian 2011 Prospective 120 patients Urine culture  15 asymptomatic (+) urine cultures 12.5%
Cross out of the 120
sectional
 1 (+) bacteremia and UTI out of 15
 3 hydronephrosis but no
abnormalities in cystourethrogram
out of 15
 5 (+) pyuria from the 15.
A68
A69

(Main Manuscript) Clinical Practice Guidelines For The Screening, Diagnosis, Treatment and Prevention of Neonatal Sepsis

Uploaded by

Copyright:

Available Formats

(Main Manuscript) Clinical Practice Guidelines For The Screening, Diagnosis, Treatment and Prevention of Neonatal Sepsis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Main Manuscript) Clinical Practice Guidelines For The Screening, Diagnosis, Treatment and Prevention of Neonatal Sepsis

Uploaded by

Copyright:

Available Formats

APPENDIX A

Draft Evidence-Based Recommendations for

Developed by the Technical Working Group

Table 1. Ways of expressing effectiveness

Table 2. Interpreting 95% Confidence Intervals (CIs)

Interpretation of 95% Confidence Intervals (CIs)

Absolute Risk Reduction

Table 3. How to interpret forest plots

However, if we need to evaluate a test with multilevel results, we need a “2 x n” table

Table 1. Interpreting likelihood ratios (LRs)

LRs > 1.0 INCREASE LR<3.0 (close to 1.0) – weakly positive

LRs < 1.0 DECREASE LR >0.3 (close to 1.0) – weakly negative

Figure B.1. Using Bayes nomogram for estimating post-test probability

The criteria for evaluating screening programs are:

Figure C.1. Admissible evidence for evaluation of a screening program

Box 1. Standard criteria for grading of evidence 1

Domain Grade Characteristic

If heterogeneity could be caused by publication bias or imprecision due to small studies,

Domain Grade Characteristic

-1 One of the above-mentioned conditions is not fulfilled.

Table 4. Quality of evidence in GRADE

3. Schϋnemann H, Brozek J, Oxman A, editors. GRADE handbook for grading quality of

Q3. Among asymptomatic newborns delivered meconium-stained, how

Q7. Among asymptomatic newborns delivered to mothers with maternal

1. Abat-Senen K, Mantaring JV. Exchange transfusion for prevention of mortality in neonatal

Organ dysfunction by day 14

Abnormal neurologic status at discharge amongst survivors

Worsening metabolic acidosis

Figure Q1.1. Exchange transfusion (DVET/ET) vs no exchange transfusion, outcome:

Figure Q1.5. Exchange transfusion vs no exchange transfusion, outcome:

Figure Q1.6. Exchange transfusion vs no exchange transfusion, outcome:

Figure Q1.7. Exchange transfusion vs no exchange transfusion, outcome:

Figure Q1.9. Subgroup analysis exchange transfusion vs no exchange transfusion 2 newer

Anticipated absolute effects* (95% CI) Certainty

246 per 1,000 RR 1.05 654

190 per 1,000 RR 0.58 120

463 per 1,000 RR 1.39 54

195 per 1,000 RR 0.80 73

Figure Q8.1 Mortality

Figure Q2.1 Mortality

Figure Q2.2 Severe neurodevelopmental disability

Figure Q2.3 Reduced P/IVH

Figure Q2.5 Pneumothorax

A meta-analysis of four (4) randomized controlled trials evaluated the effectiveness

Antibiotic prophylaxis for asymptomatic newborns delivered meconium stained is

Anticipated absolute effects* (95% CI) Certainty

Incidence of confirmed sepsis in first 28 0 per 1,000 250

25 per 1,000 250

0 per 1,000 250

CI: Confidence interval

GRADE Working Group grades of evidence

Figure Q3.1 Incidence of confirmed sepsis in first 28 days of life

Figure Q3.2 Mortality before discharge

Figure Q3.3 Incidence of suspected sepsis

Figure Q3.4 Incidence of intracranial hemorrhage

Figure Q3.6 Incidence of diarrhea

Figure Q3.7 Incidence of respiratory distress

Treatment failure RR 0.14 0.01 - 2.63 Not significant 2 RCTs2,3

Treatment failure RR 3.00 0.13 - 70.83 Not significant 1 RCT1