World over 20 million newborns get infected each year. Of 4 million neonates dying each year, 36% are due to severe infections (<xref ref-type="bibr" rid="bib001">1</xref>). In both term and preterm infants, early signs and symptoms are often non-specific and can easily be misinterpreted as being due to non-infective causes. Infection in neonates therefore needs to be promptly diagnosed so that specific treatment (antibiotics) can be started. There is a “fear factor” in that clinicians know that to delay treatment raises the risk of mortality, but presumptive treatment with antibiotics based on subtle laboratory or clinical findings results in over-treatment. Between 11 and 23 non-infected newborns were treated with antibiotics for every one with proven sepsis (<xref ref-type="bibr" rid="bib002">2</xref>). There is no doubt that excessive antibiotic use selects for antibiotic resistant bacteria. Broad spectrum antibiotics are more potent selectors of antibiotic resistance than narrow spectrum antibiotics. Furthermore, excessive antibiotic causes neonatal fungal sepsis.Isaacs D while reviewing neonatal sepsis in India has pointed out that: (a) the organisms causing early onset sepsis are very similar to those causing late onset sepsis; (b) the commonest organisms causing early and late onset sepsis are Gram negative bacilli, particularly Klebsiella, Enterobactor and Escherichia coli. Staphylococcus aureus is the commonest Gram positive organism. Group B Streptococcus is virtually never isolated; (c) there is a high incidence of fungal infection causing late onset sepsis (<xref ref-type="bibr" rid="bib003">3</xref>). Most worrying was that there are exceedingly high rates of resistance of Gram negative bacilli to almost all antibiotics. Resistance to aminoglycosides is about 50% for amikacin, higher for netilmicin and over 75% for gentamicin. Resistance to third generation cephalosporins is 80% plus. Bacteria are less resistant (30–46%) to Piperacillin-Tazobactam. Imipenem resistance is already appearing (about 20%) (<xref ref-type="bibr" rid="bib003">3</xref>).Irrational practices which has given rise to such an alarming scenario are: not taking blood cultures before starting antibiotics, if blood cultures performed and found negative, not stopping antibiotics; adding more and more potent broad spectrum antibiotics if the baby remains “sick”; and the belief that a raised serum C-reactive protein (CRP) is proof of sepsis, even if blood cultures are negative.The goals of this article are to provide a systematic but clinically realistic approach to the diagnosis of neonatal sepsis that will (a) miss no cases and will lead to minimum overtreatment of noninfected neonates; (b) minimize duration of therapy for those infants who prove to be uninfected; (c) provide a safe observation protocol for at-risk infants who are not treated.For practical purposes clinician is confronted with diagnosis of neonatal sepsis in following three scenarios: <list list-type="order"><list-item>Neonate born with one or more risk factors for sepsis</list-item><list-item>Neonate coming from community with clinical features suggestive of sepsis</list-item><list-item>Neonate usually of low birthweight admitted in hospital and develops a non-specific sign (e.g. apnea, feed intolerance, lethargy etc.)</list-item></list>A neonate born with perinatal risk factors may be asymptomatic at birth, may develop features suggestive of sepsis within the first 72 hours of life or in very severe cases neonate may be symptomatic in-utero (fetal tachycardia, poor beat to beat variability) or within a few hours after birth.Presence of some perinatal risk factors has been associated with an increased risk of early onset sepsis. Recommendations from developed countries suggest that presence of 2 risk factors should be considered an indication for starting antibiotics (<xref ref-type="bibr" rid="bib003">3</xref>). However the main organism is group B streptococci (GBS) which is not a problem in our neonatal intensive care units. Hence, their recommendations may not be applicable to our setting. Since definitive data for our setting is lacking, an empirical approach has been recommended. Presence of the following high-risk factors has been associated with an increased risk of early onset sepsis: <list list-type="order"><list-item>Low birth weight (< 2500 grams) or preterm baby</list-item><list-item>Febrile illness in the mother within 2 weeks prior to delivery</list-item><list-item>Foul smelling and/or meconium stained liquor amnii</list-item><list-item>Prolonged rupture of membranes >24 hours.</list-item><list-item>More than 3 vaginal examinations during labor</list-item><list-item>Prolonged and difficult delivery with instrumentation</list-item><list-item>Perinatal asphyxia (Apgar score < 4 at 1 minute or age) or difficult resuscitation</list-item></list>Neonates with presence of foul smelling liquor or three of the above mentioned risk factors should be considered to have early onset sepsis and treated with antibiotics. Presence of 2 or more risk factors should be investigated with a septic screen and treated accordingly.For a newborn presenting from community or deteriorating while admitted in nursery, early signs of sepsis are often subtle and nonspecific and need a high index of suspicion for early diagnosis.</sec><sec id="sec001"><title>Role of blood culture

It is the gold standard for the diagnosis of septicemia and should be done in all cases of suspected sepsis prior to starting antibiotics. Sterile acquisition by venepuncture or arterial puncture of 0.5 mL to 1.0 mL of blood, placed in a standard soy broth culture bottle, is a standard microbiologic technique available in all hospital laboratories. Blood cultures should be collected from a fresh venepuncture site because samples collected from indwelling lines and catheters are likely to be contaminated. The culture is incubated for up to 5 days, although the modern culture systems (BACTEC/BactAlert) will identify almost all early-onset neonatal pathogens within 48 hours. It is wrong not to send blood culture thinking that it has low sensitivity. A positive blood culture with a pathogenic organism is not only diagnostic of neonatal sepsis but also provides valuable information regarding prevailing organisms in a particular setting and their antibiotic sensitivity pattern. If taken properly before starting antibiotics, transported rapidly to a standardized laboratory blood culture has shown a sensitivity reaching 85%.

Role of septic screen

All newborns suspected to have neonatal sepsis should have a septic screen to corroborate the diagnosis of sepsis. A positive sepsis screen does not confirm sepsis but a negative screen strongly rules out sepsis. The various components of the septic screen include total leukocyte count, absolute neutrophil count, immature to total neutrophil ratio, micro-erythrocyte sedimentation rate and C reactive protein (Table 1). The absolute neutrophil count varies considerably in the immediate neonatal period and normal reference ranges are available in Manroe's charts. The lower limit for normal total neutrophil counts in the newborn begins at 1800/cmm, rises to 7200/cmm at 12 hours of age and then declines and persists at immature to total neutrophils (I/T ratio) is <0.16 at 1800/cmm after 72 hours of age. The ratio of birth and declines to a peak value of 0.12 after 72 hours of age. If two parameters are abnormal, it should be considered as a positive septic screen and it is reasonable to start antibiotic therapy. Presence of two abnormal parameters in a screen is associated with a sensitivity of 93–100%, specificity of 83%, positive and negative predictive values of 27% and 100% respectively in detecting sepsis. If a septic screen is negative in the presence of strong clinical suspicion, it should be repeated within 12 hours. If the screen is still negative, sepsis can be excluded with reasonable certainty. Table 1

Sepsis Screen Panel

Components	Abnormal Value

Total leukocyte count	<5000 mm³
Absolute neutrophil count	As per Manroe's Chart
Immature/Total neutrophil	>0.2
Micro-ESR	>15 mm in 1^st hour
C-reactive protein (CRP)	>1 mg/dl

Figure 1 outlines a suggested protocol for managing a newborn with suspected sepsis. Figure 1A suggested protocol for management of suspected sepsis (4) Protocol for Sepsis

Points to remember

While diagnosing neonatal sepsis always take blood culture before starting antibiotics. Rely on the blood culture results. Negative sepsis screen is very helpful in ruling out sepsis. If blood cultures are negative at 2–3 days, it is almost always safe and appropriate to stop antibiotics.

While treating neonatal sepsis use the narrowest spectrum antibiotics possible, almost always a penicillin and an aminoglycoside. Try not to use antibiotics for long periods. Do not start treatment, as a general rule, with a third generation cephalosporin.

Rational approach to diagnosis of neonatal sepsis