Neonatal Sepsis : A Profile of a Changing Spectrum

Address for correspondence: Dr. (Col) K K Venkatnarayan Associate Professor & HOD, Department of Paediatrics, Command Hospital (Southern Command), Pune, Maharastra, India. E.mail: venkatnarayan_kannan@yahoo.co.in Tel: +91 9168178855 1Dr. Kannan Venkatnarayan, MBBS, MD, DM, 2Dr. Bej PK, MBBS, DCH, 3Dr. Thapar RK, MBBS, MD. All from the Department of Paediatrics, Command Hospital (Southern Command), Pune-411040. Abstract


Introduction
W orldwide, four million babies die within fi rst 30 days of birth of an es mated 130 million babies born every year.Of these, 98% of deaths occur in developing countries and this represents 40% of all deaths of children <5 years of age 1,2 .Infec ons form the single largest cause of mortality, responsible for an es mated 35% of all neonatal deaths 3 .
The term, neonatal sepsis, has been tradi onally defi ned as bacteremia accompanied by hemodynamic compromise and systemic signs of infec on 4 .The literature on neonatal sepsis is confused by the inclusion of the en ty clinical sepsis, meaning, in developed countries, signifi cant symptoms or laboratory abnormali es, sugges ng sepsis with a nega ve blood culture.In resource-poor countries, blood cultures and adjunct laboratory tests are o en not possible and the diagnosis of neonatal sepsis is o en based solely on clinical signs 5 .
The reported incidence of neonatal sepsis varies from 7.1 to 38 per thousand live births in Asian countries 6,7,8 .Besides, there is a higher prevalence of meningi s in these babies compared to data reported from western literature 8 .
Neonatal sepsis is generally divided into earlyonset neonatal sepsis (EONS) and late-onset neonatal sepsis (LONS) with variable cutoff s for these defi ni ons, though most authors consider it at 72 h of life 5 .This designa on of EONS or LONS is important because it implies par cular routes of infec on and organisms and thus can help to guide therapy 9 .
Though blood culture remains the gold standard of diagnosis and is required to document the organism causing bacteremia, it lacks sensi vity and is o en reported as nega ve even in the presence of strong clinical evidence of sepsis 10 .The administra on of intrapartum an bio cs, insuffi cient volumes of blood drawn for culture, o en compounds this problem 11 .
Many laboratory adjuncts are available in detec ng sepsis 12 .Of these, a sep c screen consis ng fi ve items:[C-reac ve protein (CRP), total and absolute neutrophil count (ANC), immature to total ra o (ITR) and micro-erythrocyte sedimenta on rate (μ-ESR)] has been used in many studies with a fairly high sensi vity (93-100%), specifi city (83%) and nega ve predic ve value (100%) 13 .
Since, many of the lab parameters are not easily available to detect EONS in resource-poor se ngs; a large mul center study conducted to assess the usefulness of simple clinical signs iden fi ed seven such parameters that had high sensi vity (>80%) and specifi city (>70%). 14In addi on to these clinical signs to detect sepsis in community, certain maternal risk factors have been studied to be associated with increased risk of EONS. 13,15 view of these varied factors, sepsis in neonates is o en encountered with varied presenta ons such as clinical sepsis with either sep c screen posi vity or nega vity, culture posi ve sepsis with or without meningi s.In addi on, babies could receive an bio cs for presence of maternal risk factors for EONS 13,15 .Hence this study was conducted to study the profi le of neonates presen ng with two or more of the pre defi ned criteria of clinical features of sepsis with or without maternal risk factors, to assess the outcomes in terms of micro organisms cultured, sep c screen posi vity and clinical outcomes.

Materials and Methods
The study was conducted at Command Hospital (Eastern Command), a ter ary care center, from Jan 2011 to Jul 2012.Fi y consecu ve newborn babies presen ng with more than one of the following syndrome of sepsis were included in the study: Refusal to suck, lethargy, convulsions, Low/high temp, diarrhea, vomi ng, abdominal distension, Rapid breathing, apnea, Irritability as assessed by two care-takers, Fever, Jaundice, sclerema and superfi cial infec ons e.g.umbilical sepsis, pyoderma.The included babies were also assessed for associated maternal risk factors (PROM>18 hours, foul smelling liquor, maternal fever in 14 days, confi rmed maternal urinary tract infec on, repeated PV examina ons (clean>3, unclean≥1) during labor).The following were excluded: Birth asphyxia requiring resuscita on, major congenital malforma ons, and respiratory distress caused by pneumothorax, hyaline membrane disease, transient tachypnea of newborn.
Outcomes and their measurements: The included babies underwent gesta onal assessment and were calculated from the fi rst day of last menstrual period if the mother was sure of dates; if not, it was calculated based on the fi rst trimester ultrasound assessment or by Expanded New Ballard Score.The babies were classifi ed as Early Onset Neonatal Sepsis (EONS) or Late Onset Neonatal Sepsis (LONS) depending on the me of ini al presenta on (<72h or >72h, respec vely).
A sep c screen and blood culture were carried out in all babies.The sep c screen was carried out at the me of presenta on and consisted of C-reac ve protein (CRP), total leukocyte and absolute neutrophil count (TLC & ANC), immature to total ra o (ITR) and micro-erythrocyte sedimenta on rate (μ-ESR).The following were considered signifi cant: TLC<5000/ mm 3 , ANC as per Manroe/ Mouzinho's chart (for term/ VLBW babies respec vely), ITR>0.2, μ-ESR>15mm in 1 st h, CRP>1mg/dL.Presence of two or more signifi cant parameters was considered as sep c screen posi ve.
Blood culture was taken with strict asep c precau ons as per laid down guidelines and standard techniques. 17Cerebro Spinal Fluid (CSF) study for meningi s was done in all babies with LONS and in babies with EONS, when the sep c screen was posi ve.Meningi s was diagnosed based on the CSF reports, as published. 13Urine culture was done in all babies presen ng with LONS.
Mothers presen ng with premature rupture of membranes were rou nely given an bio cs consis ng of Ampicillin and Azithromycin, as per the policies of the Obstetric team.The babies with suspected sepsis were treated with fi rst-line an bio cs (Cefotaxime and Amikacin) as per standard guidelines and upgraded to the second-line an bio cs (Vancomycin and Piperacillin-Tazobactum) in case of inadequate clinical response by 96h of ini a on of therapy or clinical deteriora on.The clinical outcome was assessed based on the response and the clinical course (responded to fi rst line an bio cs, needed a change of an bio cs, requirement of inotropes, death).

Profi le of subjects
A total of 50 consecu ve newborns were studied with clinical features of sepsis.Male comprised of 32 babies (64%) in total and accounted for 66% of culture proven sepsis.Four babies were born extramurally and were clinically diagnosed to have clinical sepsis, without screen or culture posi vity.Twenty babies (40%) were term.In total, 38 babies (76%) were EONS and the rest (24%), LONS.Fi een babies (30%) were blood culture posi ve and it accounted for 29% of EONS and 33% of LONS.Sixteen babies (32%) had posi ve lab parameters in addi on to clinical features but did not grow any organism on blood culture (Table 1 and Fig 1).

Maternal Risk Factors
The babies were assessed for presence of maternal risk factors of sepsis and 62% (31) babies had at least one risk factor.The details of risk factors are as depicted in Table 2.

Manifesta ons of clinical syndrome of sepsis
Eleven clinical features of neonatal sepsis were used for screening babies and these included the seven parameters of the Young Infant Study group 14 .Of the1197 babies assessed for eligibility for the study, 26 were excluded based on predefi ned criteria.Fi y consecu ve babies were screened for the eleven clinical features of neonatal sepsis syndrome.Feeding problems were seen in all the babies of the study, irrespec ve of EONS or LONS.Temperature instability was no ced only in EONS group.Jaundice was taken as a parameter for screening of the babies only when it was associated with another syndrome of sepsis (Table 3).

Sep c screen posi vity
All the pa ents underwent sep c screen as specifi ed earlier.A total of twenty seven babies were screen posi ve.When assessed by the clinical presenta on, the screen was posi ve in 17 babies (45 %) with EONS and in 10 babies (83%) with LONS.Of these babies, sixteen babies did not grow any organism on blood culture (screen posi ve sepsis).Nineteen babies were nega ve on both sep c screen and blood culture (clinical sepsis) (Table1, Figure 1).When compared with blood culture posi vity, which was taken as gold standard, the sensi vity was 73% and specifi city 54%, with a posi ve predic ve value of 41% and nega ve predic ve value of 83%.Four babies were nega ve on sep c screen but grew organisms on blood culture (False Nega ve) and all of them belonged to EONS group.

Blood culture reports
Blood culture was done in all babies.Overall, 15 babies (30%) grew organisms on blood culture.All these babies were born intramurally.They represented 29% of babies with EONS and 33% of babies with LONS.The most common organism was Staphylococcus aureus in both EONS and LONS group.GBS and pseudomonas were encountered only in EONS, while CONS and E Coli were seen in both the groups (Table 4, Figure 2)

An bio c sensi vity pa erns
An bio c sensi vity pa erns for the cultured organisms were analyzed with commonly used drugs, so as to evolve a protocol for the unit for the fi rst line therapy in case of suspected sepsis syndrome.The an bio c sensi vity pa ern for each organism per an bio c is depicted in Figure 3. Amongst Staphylococcus aureus, 66.6% (4) were sensi ve to Ampicillin and 85.7% (6) to Vancomycin.Almost 71.4% were resistant to Cefotaxime and Amikacin.66.7% of E Coli was sensi ve to Ampicillin, Cefotaxime and Gentamicin independently.And all (100%) of the E Coli culture were sensi ve to Amikacin.Coagulase Nega ve Staphylococcus showed 100% sensi vity to Vancomycin and 50% sensi vity to Ampicillin, Cefotaxime and Amikacin.Pseudomonas showed 100% sensi vity to Piperacillin, Cabenicillin and 50% sensi vity to Amikacin and Ampicillin.All were resistant to Gentamicin.Group B Streptococci showed 100% sensi vity to Ampcillin and Vancomycin and 50% resistance to Amikacin.All the GBS were resistant to Cefotaxime.

Outcome to treatment
On inclusion in the study, all the babies were administered fi rst line an bio c (Cefotaxime and Amikacin), at the prescribed doses, as per unit policy.Babies who were sep c screen nega ve but with clinical features, were given an bio cs for seven days.The dura on of treatment for those who were ini ated on treatment solely based on >2 maternal risk factors was three days.Babies who were sep c screen posi ve but culture nega ve were given an bio cs for ten days.On culture posi vity, the dura on was increased to fourteen days.Babies with CSF study sugges ve of meningi s were administered treatment for twenty one days.74% of babies responded to fi rst line an bio cs, 22% required change of an bio cs and 2% required inotropes.The case fatality of sepsis in our unit during the study period was 4%.

Discussion
With changes in an bio c usage, mode of diagnosis and management of the condi on, the bacteriological profi le of sepsis in newborn is constantly changing.Besides, the spectrum of causa ve organisms in the developing countries is diff erent from those from the developed countries 18,19 .The reported data of sepsis from our country has been predominately from Level 3 NICUs from ter ary care units 8 .The clinical spectrum of neonatal sepsis also varies between the me of presenta on (early, <72h or late, >72h), as it has rela on to the mode of infec on 9 .Though, some studies highligh ng hospital acquired infec ons in ins tu onal deliveries have shown this diff erence to be of less importance 20 .Since there is no single test that can be confi dently relied upon for diagnosis and as the clinical manifesta ons of sepsis are protean, its diagnosis relies mainly on assessing the risk factor predisposi on 15 , clinical syndrome 14 and ba ery of tests 13 .We used a combina on of all these to include babies in our study, keeping blood culture posi vity as gold standard.
The rate of confi rmed sepsis in our study was 12.5 per 1000 births and those of meningi s were 1.7 per 1000 live births.The reported incidence of neonatal sepsis varies from 7.1 to 38 per 1000 live births in Asia 10,20 .The Na onal Neonatal Perinatal Database reported an incidence of sepsis to vary from 10 to 45 per 1000 live births from eighteen hospitals in 2003 and 30 per 1000 live births in 2006 8 .In Indian nurseries, the incidence of meningi s was es mated as 3 per 1000 live births, 16 compared to incidence in the west of 0.7 to 1 per 1000 live births 21 .Depending on the me of presenta on, the incidence of EONS in our study if defi ned by clinical indices alone, was approximately 32 per 1000 live births and that of blood culture posi ve sepsis was 9 per 1000 live births.There have been rela vely few studies from India repor ng specifi c rates of EONS.A study from a home based surveillance program from Bangladesh reported the clinical EONS to be 50 per 1000 live births and that of blood cultureconfi rmed EONS, 2.9 per 1000 live births 22 .
A clinical diagnosis of sepsis was considered based on a set of pre-defi ned clinical features, which had been validated 14 .The most common clinical presenta on was related to feeding diffi cul es, followed by lethargy and respiratory distress.These were similar to the fi ndings from two large studies from India 23,24 .However, our study being a hospital based rather than community based, we also included more objec ve criteria such as presence of sclerema and apneic spells.Jaundice was taken as a sign based on previous studies 25,26 and in view of its common presenta on in neonates; it was considered signifi cant when only associated with another defi ned clinical condi on.Sep c screen was done in all babies to decide the dura on of treatment of an bio cs and requirement of a CSF study 13 .The trends of specifi city, sensi vity, nega ve and posi ve predic ve value of our study is similar to those of other reported studies 13,27 .The blood culture posi vity yield of our study were 30%, higher than those reported from other studies from India 19,28,29 .
Our study shows a preponderance of Gram posi ve sepsis (67%) with Staphylococcus aureus as the predominant organism (14%) in both the EONS (13%) and LONS (17%).The reported studies on blood culture from India and other countries from Asia have shown predominance of Gram nega ve bacteria with Klebsiella pneumoniae as the predominant organism 10,13,16,19,20,28,29   .Staphylococcus aureus has been found to be the predominant organism in some of the cohorts of neonatal sepsis from the community 29 or in the LONS group 30 .Some centers have reported it as the major causa ve organism accoun ng for 57-87% of blood culture posi ve sepsis, repeatedly over the years 31 .The probable reasons for the predominance of Staphylococcus aureus with a lack of Klebsiella in our study popula on could be a changing pa ern of microorganisms causing neonatal sepsis in our se ng.The other over-riding reasons could be: eff ec ve selec ve antenatal an bio c cover for Gram nega ve organisms and repeated point-source Staphylococcal infec ons in the immediate post natal period 21 .Similar condi ons may be prevailing in other units 20 .The an biogram of the micro-organisms cultured in our study shows some interes ng trends.Overall, Ampicillin seems to fare be er than Cefotaxime.Vancomycin retained sensi vity across the Gram posi ve group and Amikacin was superior to Gentamicin.
The strengths of our study are its prospec ve nature and the study being done in a standardized manner in a teaching ins tu on using established techniques.The main limita on of the study was a lack of control group and a limited sample size.

Conclusion
To conclude our study has shown a changing profi le of organisms causing neonatal sepsis, with a preponderance of Gram -posi ve organisms par cularly Staphylococccus aureus.The an bio c sensi vity spectrum revealed a prevailing sensi vity to a combina on of Ampicillin and Amikacin, which fared be er than Cefotaxime.

Fig 2 :
Fig 2: Micro organisms cultured in EONS and LONS

Fig 3 :
Fig 3: An bio c sensi vity pa ern of organisms cultured in the study.The lightly shaded bar chart on the le depict number of cultures resistant and the darkly shaded bar chart on the right depict sensi vity, per an bio c Staph aureus (Staphylococcus aureus), CONS (Coagulase Nega ve Staphylococcus aureus), GBS (Group-B Streptococci), E Coli (Escherisha Coli), Pip-Taz (Piperacillin Tazobactum)

Table 1 :
Profi le of pa ents studied Babies having only clinical features of sepsis but are nega ve for sepsis screen and blood culture β Babies in addi on to clinical features are posi ve on sep c screen but nega ve on blood culture μ Babies in addi on to clinical features are posi ve on sep c screen and blood culture α

Table 2 :
Details of maternal risk factors for neonatal sepsis

Table 3 :
Clinical syndrome of neonatal sepsis for screening the babies Resp distress: Resp rate>60/min or Grun ng or chest in-drawing (cases of diagnosed pneumothorax TTNB, RDS excluded) # Along with one more indicator of clinical syndrome of sepsis β Cessa on of respira on>20 sec or associated with bradycardia or cyanosis or requiring bag and mask resuscita on α

Table 4 :
Spectrum of organisms grown on blood culture