Académique Documents
Professionnel Documents
Culture Documents
SCIENCEDOMAIN international
www.sciencedomain.org
Authors contribution
The sole author designed, analyzed and interpreted and prepared the manuscript.
Article Information
DOI: 10.9734/AJMAH/2017/32627
Editor(s):
(1) Marco Vincius Chaud, Pharmaceutical Science, Laboratory of Biomaterials and Nanotechnology Sorocaba, University of
Sorocaba, Brazil.
Reviewers:
(1) P. Krishnaveni, Rajiv Gandhi University of Health Sciences, India.
(2) Dinesh Yadav, Kailash Hospital, Behror, India.
(3) Guang Hu, Soochow University, China.
(4) Simon Pius, University of Maiduguri, Nigeria.
Complete Peer review History: http://www.sciencedomain.org/review-history/18647
th
Received 6 March 2017
th
Review Article Accepted 12 April 2017
Published 15th April 2017
ABSTRACT
Introduction: The burden of neonatal sepsis continues to be significant, more so in the preterm
population. It is quite challenging to make early diagnosis because the early signs of sepsis may be
subtle, and resemble those of other non-infectious processes; furthermore, cultures take time and
the yield is quite low with wide inter-laboratory variation.
Alternative diagnostic tests are simply not accurate enough. The resulting implications of these
challenges include prolonged empirical antibiotic therapy for at-risk neonates.
An early diagnostic test that is highly accurate will be immensely beneficial in guiding clinicians in
neonatal units on commencement of antibiotics and when to stop.
This metanalysis assesses the usefulness, reliability, limitations and challenges in clinical practice
of some of the tests for the early diagnosis of neonatal sepsis.
Methods: A metanalysis of articles on the different methods of diagnosis of neonatal sepsis in
clinical practice as well as scientific papers comparing various methods was carried out. Sources
include but not limited to Pubmed, HINARI and EMBASE database.
Conclusions: Extensive work is being performed to find the ideal test for early diagnosis of
neonatal sepsis. Despite numerous studies on acute phase reactants, their use in clinical setting is
_____________________________________________________________________________________________________
limited to CRP and PCT to some extent. There is still need for further research work to find an
ideal test for early diagnosis of neonatal sepsis. However, the methodologies and study designs
are to be harmonized in order to obtain ideal cut-off acceptable values.
The accuracy of the tests available are not Hematologic scoring system (HSS) described by
satisfactory; there are cases of false negatives Rodwell, Lesilie, and Tudehope improves the
and more commonly false positives. The efficiency of the complete blood count (CBC) as
relevance of an early diagnostic test is mostly in a screening test for sepsis. It is derived from a
the evaluation of at-risk neonates and those with single test that is fast, easily performed and
subtle, non-specific signs, where delay in readily available in most hospitals and thus could
instituting appropriate antimicrobial therapy can allow a more systematic approach to decisions
lead to significant morbidity and possibly fatality regarding antibiotic therapy.
[15].
The HSS has practical advantages; it is
A number of acute-phase reactants (APR) have applicable to all infants, including those who
been evaluated with reasonably high specificity have received antibiotic therapy prior to
and sensitivity, however none has been proven evaluation and simplifies the interpretation of
to be accurately diagnostic [16,17]. This review hematologic profile [18].
draws attention to the limitations of some of the
tests for the early diagnosis of neonatal sepsis. In a Bangladeshi study by Khalada et al.
involving 100 neonates, a score of 3 was highly
2. METHODS significant (P<0.05), with sensitivity of 100%,
specificity of 21%, PPV 15%, NPV 100%. These
A metanalysis of articles on the different methods results were consistent with other studies. In the
of diagnosis of neonatal sepsis as well as study, sensitivity of score >4 was 100%,
scientific papers comparing various methods was specificity 60%, PPV 26%, NPV 100%. In
carried out. Sources include but not limited to comparison with score >3, score >4 was more
Pubmed, HINARI and EMBASE database. sensitive (P<.001) and specificity and PPV were
2
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
significantly higher as well, 60% and 26% inflammation. The most commonly used acute
respectively. But considering the high specificity phase reactant (ARP) is the C-reactive protein
and positive predictive value, the study implied (CRP). However, the CRP takes 1224 hours to
that score > 4 was more reliable as a screening increase to measurable levels; its half life is very
tool for sepsis than any of the individual long and it takes 57 days to normalize after
hematological parameter. eradication of the infectious agent. Cytokines
such as IL6, IL8, TNF-, and procalcitonin have
In the study, I/T ratio >0.2 had a sensitivity, also been studied [24,25]. Cytokines rise quickly
specificity, PPV and NPV of 100%, 04%, 13% after infection even in neonates, and are more
and 100% respectively. Specificity and positive sensitive to low concentrations of pathogens than
predictive value was low because of large CRP.
number of false positive results. While an I/T
ratio >0.2 suggested by Rodwell, Lesilie, and 2.1.2.1 The C-reactive protein
Tudehope had a sensitivity of 96% and NPV of
99%. So this result for an elevated I/T ratio was The C-reactive protein is the most analyzed
consistent with other reports. The sensitivity of parameter for the detection of bacterial infections
I/M ratio (>0.3) was 100%, specificity 07%, PPV for years [26,27]. It is an annular pentameric
11% and NPV 100%. Rodwell et al. used I/M protein of hepatic origin that increases following
ratio of 0.3 as a predictor of infection and interleukin-6 secretion from macrophages and T
reported sensitivity, specificity, PPV and NPV of cells. It was discovered by Tillett and Francis in
93%, 81%, 32% and 99% respectively 1930 [28].
[11,18,19,21-23].
Table 1. Hematological scoring system
Ghosh et al. [22] evaluated the haematologic
parameters of 105 Indian neonates, a score of Criteria Finding Score
three was taken as the cut of point as it provided Total WBC 5000 or 25000/l 1
acceptable levels of sensitivity, specificity and count
positive predictive values. A score of 3 identified Total PMN 1800 5400 0
28 of the 30 babies with sepsis and in addition count None 2
five infants without sepsis. Each of the seven < 1800 or > 5400 1
criteria of the HSS was evaluated. An abnormal Immature PMN 600 0
1:T ratio had the highest sensitivity (93%) and count Increased 1
identified more than 90% of the infants with I:T PMN ratio 0.12 0
sepsis of probable infection. An abnormal 1:M > 0.12 1
ratio was also found to be highly sensitive (92%) I:M PMN ratio 0.3 0
in identifying infants with sepsis. These two > 0.3 1
criteria along with thrombocytopenia PMN changes Toxic granules or 1
(<150,000/cm) had a high negative predictive cytoplasmic
value over 94%. Hence the likelihood of sepsis vacuoles
being present in the absence of these findings is Platelet count 1.5 lakhs/ l 1
low. The study also found that the higher the WBC- White blood cell, PMN- Polymorphonuclear,
score the greater the certainty that sepsis was I:T- Immature to total ratio,
I:M- Immature to mature ratio.
present. The five false positive cases all had a
Reproduced from: Rodwell RL, Leslie AL, Tudehope
borderline score of 3. DI. Early diagnosis of Neonatal Sepsis using a
hematologic scoring system. J Pediatric. 1988;112:
The HSS is potentially a very useful tool in 761-767
resource poor settings and its right application Interpretation: < 2 Sepsis is very unlikely, 3 or 4
should significantly minimize the problems Sepsis is suspected
associated with prolonged empirical treatment of > 5 Sepsis or infection is very likely
at-risk babies. One of the drawbacks of the HSS
is the observer variation in the assessment of C- reactive protein is a good discriminatory
blood film for white cell properties. marker of bacterial sepsis in newborns, its half-
life of 48 hours is constant, and therefore its level
2.1.2 Biological biomarkers is determined by the rate of production. The
predictive value of CRP improves with time and
Biological biomarkers are human blood is most predictive between 24 and 48 hours after
components that increase in response to presentation with suspected infection [29,30].
3
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
Several hours are needed for CRP levels to case NPV and PPV, generally low threshold
increase in serum (including activation of values for diagnosis gives a higher PPV and
neutrophils, elaboration of interleukin-6, and lower NPV, the converse applies to high cut-off
induction of hepatic synthesis of CRP) therefore values.
limiting the sensitivity of this test in diagnosing
neonatal sepsis (NNS). Its levels are consistently In an Australian study, CRP values greater than
elevated 24 to 48 hours after the onset of 15 mg/l was said to confer a 7% risk of death and
infection; therefore serial normal levels done at up to 21% when levels are greater than 30 mg/l.
18 to 24 hour intervals may be useful for With regards to the utility of CRP as a prognostic
identification of infants who do not have bacterial marker, the majority of studies failed to
infection [31,32]. It is not a highly specific marker, demonstrate any correlation between CRP
as false positive elevation of CRP is found in viral concentration and mortality [36-38].
infections, meconium aspiration syndrome,
cephalhaematoma, bruising and surgery [33]. Serial measurements of CRP is required to make
decisions in clinical practice, this can be
The wide variation in sensitivity, specificity, expensive in resource limited settings where
positive predictive value (PPV) and negative health insurance is grossly suboptimal.
predictive values (NPV) reported in various
studies are partly due to different cut-off values 2.1.2.2 Procalcitonin (PCT)
ranging from 5 to 20 mg/l, as well as different
testing methods. Han et al. reported the Procalcitonin (PCT) has been reported as a
sensitivity, specificity, PPV and NPV of serial measurable laboratory marker in the
CRP as 82%, 78%, 15% and 98.9% respectively. inflammatory response to infection in some
Ayazi et al. reported 55%, 80%, 24% and 90% studies.
respectively, while Borma reported 79%, 85%,
36% and 97% respectively. Nuntnarumit et al. Procalcitonin is a 116-amino acid protein, a
reported a sensitivity of 100%, specificity of 94%, precursor of calcitonin which is produced by the
PPV and NPV of 91.6% and 100% respectively thyroid. In sepsis, macrophages and the
of CRP for detecting proven sepsis and localized monocytic cells of the liver are involved in the
infection at cut off point 5 mg/l. Himayun et al. synthesis of PCT. A number of studies have
found 70%, 72.3%,28% and 94% respectively at reported on the usefulness of the quantitative
a cut-off of 8 mg/l [2,31,34,35]. measurement of PCT for an early diagnosis of
sepsis in newborns [39].
The NPV of CRP is consistently high and it is
most useful in indicating an absence of infection. In healthy persons, PCT levels are undetectably
A negative CRP value is more important than a low, however in severe bacterial, fungal, parasitic
positive CRP value in that it excludes infection infections with systemic manifestations, a
with a high certainty and can guide the decision significant rise is seen. In this condition the
to discontinue antibiotics. production site is the extra thyroid tissues [40].
The challenge with interpretation and application In healthy individuals, production of PCT and
of the studies is defining the reliability, in this subsequently calcitonin is restricted to the thyroid
4
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
C-cells. Bacterial infections selectively induce an categorized as proven sepsis, suspected sepsis
increase in the concentration of PCT; because and clinical sepsis respectively. Procalcitonin
both endotoxins released from the bacterial cell was positive in 100% compared to the CRP
wall as well as the host responses to infection positivity in 63.6% of the proven sepsis cases.
activate the production of PCT mainly in Chiesa et al. determined reference ranges for
parenchymal tissues [41,42]. PCT across the range of postnatal hours from 0
to 48 h and, in a further study, these authors
In1993, PCT was first described as a marker of showed that PCT sensitivity and specificity were
the extent and course of systemic inflammatory greater than those of CRP or interleukin 6 if
response to bacterial and fungal infections. The different cutoff points at birth and at 24 h and 48
advantage of PCT as compared to Creactive h of life were used [29,36].
protein is that the increase of the former in
bacterial infection and its restoration to normal is In these studies, it was reported that serum
more rapid. It can therefore help in deciding the levels had also increased in non-infected
duration of antibiotics administration [43,44]. neonates with maternal chorioamnionitis, GBS
Furthermore, the finding that PCT is released colonization, perinatal asphyxia, intracranial
into the circulation within 3 h after endotoxin hemorrhage, pneumothorax, or after
injection, plateaus at 6 h, and remains elevated resuscitation, and these conditions had
for 24 h, makes PCT a promising new agent for negatively affected the specificity of PCT. Janota
early and sensitive identification of infected et al. reported a significant increase in serum
patients [45]. PCT concentrations within 72 h of age in preterm
infected and uninfected newborns born to
The results of recent studies on the usefulness of mothers with chorioamnionitis. Gendrel and
PCT for early diagnosis of neonatal sepsis have Bohuon suggested that hypoxemia may be
been inconclusive [27,46-53]. Some authors responsible for increased PCT values in
have reported markedly increased neonates. They also found that neonates born to
concentrations during the first 48 h of life in mothers with preeclampsia had higher PCT
newborn infants without bacterial infection [53- concentrations at both 24 and 48 h of life [58,59].
55]. In these studies, PCT sensitivity in the early
diagnosis of neonatal sepsis was found to be 83- Although PCT may cross the placental barrier,
100% while the specificity was 70-100% the findings of higher PCT concentrations in cord
[54,56,57]. sera compared with maternal samples, with even
larger differences at 24 and 48 h of age, cannot
The sensitivity of PCT in initial determinations for be explained on the grounds of maternal transfer
the diagnosis of early-onset neonatal sepsis has alone; therefore, the postnatal surge of PCT may
been reported to vary from 61% to 85%, represent endogenous synthesis [27]. This
increasing to 72100% within the subsequent phenomenon might be attributed to direct stress
24 h [51,53,54]. on the baby during the perinatal period or to the
adaptation to the extrauterine environment
Sastre et al. [12] evaluated 317 newborns in a
[60,61].
multicentre Spanish study (confirmed vertical
sepsis: 31, vertical clinical sepsis: 38, non-
Higher concentrations of PCT have been
infectious diseases: 79). In asymptomatic
observed in uninfected infants with respiratory
neonates, PCT values at 1224 h were
disorders (mostly respiratory distress syndrome
significantly higher than at birth and at 3648 h of
(RDS) compared with asymptomatic infants.
life. Neonates with confirmed vertical sepsis
Because no significant effect related to the
showed significantly higher PCT values than
severity of RDS could be detected, Monneret et
those with clinical sepsis. PCT thresholds for the
al. suggested that hypoxemia (which is transient
diagnosis of sepsis were 0.55 ng/mL at birth
during delivery) could be responsible for these
(sensitivity 75.4%, specificity 72.3%); 4.7 ng/mL
increased PCT values, providing further support
within 1224 h of life (sensitivity 73.8%,
for the hypothesis of pulmonary PCT synthesis
specificity 80.8%); and 1.7 ng/mL within 3648 h
[53,62].
of life (sensitivity 77.6%, specificity 79.2%).
Active resuscitation at birth was independently
The wide disparity in cut-off points used in these
associated with high PCT values.
studies (0.55 ng/mL) may account for the
Ali et al. in a study of 70 Neonates admitted differences in results. The inability of PCT to
in NICU, 31.4%, 42.9% and 25.7% were differentiate bacterial infections from other
5
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
6
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
sepsis of vertical transmission. BMC 25. Franz AR, Steinbach G, Kron M, Pohlandt
Pediatrics. 2007;7:9. F. Reduction of unnecessary antibiotic
DOI: 10.1186/1471-2431-7-9 therapy in newborn infants using
13. Hodge G, Hodge S, Haslam R, McPhee A, interleukin-8 and C-reactive protein as
Sepulveda H, Morgan E, Nicholson I, Zola markers of bacterial infections. Pediatrics.
H. Rapid simultaneous measurement of 1999;104:447453.
multiple cytokines using 100 microl sample 26. Monneret G, Labaune J, Isaac C,
volumes--association with neonatal sepsis. Bienvenu F, Putet G. Procalcitonin and C-
Clin Exp Immunol. 2004;137:402-407. reactive protein levels in neonatal
14. Nemsadze K. Early detection and infections. Acta Paediatrica. 1997;86:209-
prevention of neonatal sepsis. 12.
DOI: 10.5772/56121 27. Chiesa C, Signore F, Assumma M, Buffon
15. Chiesa C, Panero A, Osborn J, Simonetti E. Serial measurements of the C-reactive
A, Pacifico L. Diagnosis of Neonatal protein and interleukin 6 in the immediate
Sepsis: A clinical and laboratory challenge. postnatal period: The reference intervals
Clin Chem. 2004;50:279-287. and the analysis of the maternal and
perinatal confounders. Clinchem. 2001;47:
DOI: 10.1373/clinchem.2003.025171
1016-22.
16. Hodge G, Hodge S, Haslam R, McPhee A,
28. Neonatal sepsis. Wikipedia accessed on
Sepulveda H, Morgan E, Nicholson I, Zola
Dec 12, 2015.
H. Rapid simultaneous measurement of
29. Ayazi P, Daneshi M, Hashemi H. The role
multiple cytokines using 100 microl sample
of serial serum C-reactive protein level in
volumes--association with neonatal
the diagnosis of neonatal infection. Iranian
sepsis.Clin Exp Immunol. 2004;137:402-
Journal of Pediatrics Society. 2007;1(1):
407.
47-51.
17. Hodge G, Hodge S, Han P, Haslam R: 30. Bomela HN, Ballot DE, Cory BJ, Cooper
Multiple leucocyte activation markers to PA. Use of C-reactive protein to guide
detect neonatal infection. Clin Exp duration of empiric antibiotic therapy in
Immunol. 2004;135:125-129. suspected early neonatal sepsis. Pediatr
18. Rodwell RL, Leslie AL ,Tudehope DI. Early Infect Dis J. 2000;19(6):531-5.
diagnosis of Neonatal Sepsis using a 31. Himayun M, Ahmad S, Rasool A. Role of
hematologic scoring system. J Pediatric. C-reactive protein in early onset neonatal
1988;112:761-767. sepsis. The Internet Journal of Pediatrics
19. Philip AGS, Hewitt JR. Early diagnosis of and Neonatology. 2009;11:2.
Neonatal Sepsis. Pediatrics. 1980;65(5): 32. Ehl S, Gering B, Pohlandt F. A detailed
1036-1041. analysis of changes in serum C-reactive
20. Engle WD, Rosenfeld CR. Neutropenia in protein levels in neonates treated for
high risk neonates. J Pediatric; 1984. bacterial infection. Eur J Pediatr. 1999;
21. Manroe BL, Weinberg AG, Rosenfeld CR, 158:238-42.
Browne R. The neonatal blood count in 33. Pourcyrous M, Bada HS, Korones SB,
health and diaease, i,e. Reference values Baselski V, Wong SP. Significance of
for neutrophilic cells. J Pediatrics. 1979; serial C-reactive protein responses in
95(1):89-98. neonatal infection and other disorders.
22. Ghosh S, Mittal M, Jaganathan G. Early Pediatrics. 1993;92(3):431-5.
diagnosis of neonatal sepsis using a 34. Han Michael Y, et al. Serial measurements
hematologic scoring system. Indian J Med of serum C-reactive protein in the
Sci. 2001;55(9):495-500. diagnosis of neonatal infection. Pediatrics.
23. Manucha V, Rusia U, Sikka M, Faridi 1997;100(35):493-4.
MMA, Madan N. Utility of hematological 35. Nuntnarumit P, Pinkaew O, Kitiwanwanich
parameters & c-reactive protein in the S. Predictive values of serial C- Reactive
detection of neonatal sepsis. J Paediatr & protein in neonatal sepsis. J Med Assoc
Child Health. 2002;38:459-464. Thai. 2002;S1151-8.
24. Verboon-Maciolek MA, Thijsen SF, Hemels 36. Ali AM, Elkhatib WF, Abdelaziz SS.
MA, Menses M, van Loon AM, et al. Procalcitonin versus C-reactive protein in
Inflammatory mediators for the diagnosis neonatal sepsis. J Immunol Infect Dis.
and treatment of sepsis in early infancy. 2014;1(1):103.
Pediatr Res. 2006;59:457461. DOI: 10.15744/2394-6512.1.103
7
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
8
Kayode-Adedeji; AJMAH, 4(2): 1-9, 2017; Article no.AJMAH.32627
Peer-review history:
The peer review history for this paper can be accessed here:
http://sciencedomain.org/review-history/18647