Microorganisms Profile and Antimicrobial Resistance Pattern in Outborn Neonates in Northern India: A Hospital-Based Observational Study

Abstract

Background

Neonates born somewhere else (outborn) and treated in a referral centre have different microbiological profile. We report the microorganism’s profile and antimicrobial resistance (AMR) in blood culture proven sepsis in outborn neonates.

Methods

Culture positive neonatal sepsis from a neonatal unit of a referral institute catering to outborn neonates was studied over an 18 months duration. Data from the hospital information system were used to analyse the culture positivity rates, the spectrum of the microorganisms isolated and AMR pattern.

Results

Out of 5258 admitted neonates, 3687 blood samples were sent for suspect sepsis. The blood cultures were positive in 537 (14.6%) samples from 514 neonates. Gram-positive cocci (GPC) were the most common [240 (45%)] followed by gram-negative bacilli (GNB) [233 (43.4%)] and fungi [64 (11.9%)]. Coagulase negative staphylococcus (CONS) contributed to two-thirds of GPC followed by Klebsiella [93 (17.3%)] and Acinetobacter species [52 (9.7%)]. In 403 (75%) neonates, organisms grew in the samples sent at or within 24 h of admission. The case fatality rate was significantly higher in those with culture positive sepsis. The resistance to meropenem and imipenem was documented in 57.1% and 49.7%, respectively and 48% of the GNB was multidrug resistant.

Conclusions

CONS followed by Klebsiella species were the most common organisms isolated. Three-fourths of the neonates had organisms grown at or within 24 h from admission. More than half of the GNB were multidrug resistant. The case fatality rate was significantly higher in those with culture positive sepsis.

Lay Summary

Sepsis is the third most common cause of neonatal mortality globally. Outborn neonates differ in their microorganisms’ profile and antimicrobial resistance (AMR) pattern in comparison to inborn neonates. In this study, we report the microorganisms profile and their AMR pattern in blood culture proven sepsis in a large cohort of outborn (extramural) neonates admitted to the index institute. We have also presented the state-wise profile and have compared their AMR pattern. Out of the 5258 admitted neonates, 3687 blood samples were sent for culture for suspect sepsis. The blood cultures were positive in 537 (14.6%) samples from 514 neonates. Coagulase-negative staphylococcus (CONS) followed by Klebsiella species were the most common organisms isolated from this large cohort of outborn neonates. More than 75% of the neonates grew the organisms within 24 h from admission indicating that many of them harboured the organisms at admission. Case fatality rate was significantly higher in those neonates with culture positive sepsis in comparison to culture negative sepsis. Close to 50% of the gram-negative bacilli isolates were multidrug resistant and half of them were extensively drug resistant. A significant between-state difference in organism profile and their AMR patterns were observed.

outborn neonates, culture positive sepsis, organism profile, case fatality, antimicrobial resistance

INTRODUCTION

Sepsis is one of the most significant contributors to the 2.5 million global annual neonatal deaths and the third common cause of the 0.6 million annual neonatal deaths in India [1, 2]. Several reports indicate that outborn neonates (neonates born in a different place and are managed in a separate hospital) have different morbidity and mortality profile in comparison to inborn neonates [3, 4] and may differ in their microorganisms profile and antimicrobial resistance (AMR) pattern in comparison to inborn neonates due primarily to the heterogeneity in the source of infection (community vs. hospital acquired), wide variation in levels of neonatal care, a delay in referral to higher centres and treatment with multiple antimicrobials prior to referral [5, 6]. A prospective study conducted in outborn neonates in India reported that gram-negative bacteria (GNB ) contributed to half of the culture positive sepsis and also reported that more than half of the isolates were multidrug resistant (MDR) [7]. However, extrapolation of data from one particular region of the country to another region or to the country as a whole may not be appropriate due to regional variation in microorganisms profile and their AMR pattern [8]. In this study, we report the microorganisms profile and their AMR pattern in blood culture proven sepsis in a large cohort of outborn (extramural) neonates admitted to the index institute. We have also presented the state-wise profile and have compared their AMR pattern.

MATERIALS AND METHODS

This was a retrospective analysis of all the culture positive neonatal sepsis in an extramural neonatal unit of a large government tertiary care centre of North and North-western India from January 2018 to August 2019 (20 months). In brief, the Paediatric Emergency Room of the centre caters to ∼4000 neonatal admissions per annum. The admitted neonates are primarily referred from the six neighbouring states which contribute to about 15% of the population of the country. The study protocol was approved by the institutional ethics committee and a waiver for consent from the patients was obtained considering the retrospective design of the study. All neonates (<28 days postnatal age) admitted to the extramural unit with a clinical diagnosis of neonatal sepsis and in whom the blood culture grew a microorganism were included in the analysis. The admissions included first as well as readmissions provided the readmission occurred after an interval of 24 h from discharge/back-referral from the index hospital. In all eligible neonates, a blood sample was drawn for culture before starting the antibiotics. The choice of antibiotics depended on the policy of the unit, which was time and again revised based on the profile of bacterial organisms. During the study period, ciprofloxacin and amikacin were the first line empiric antibiotics of choice followed by meropenem and colistin-polymyxin as second- and third-line antibiotics, respectively. Vancomycin was added to the above antibiotics at the discretion of the clinician.

Blood culture process

During the study period, paediatric blood culture bottle BD BACTEC^TM PedsPlus^TM/F (Becton, Dickinson and Company, Sparks, MD, USA) were used in the Institute for inoculation of blood samples. These bottles had enriched soybean-casein digest broth medium (40 ml). Considering the higher relative bacterial load in neonates, a minimum of 0.5 ml was drawn for the blood cultures [9]. The standard practice in the institute is to read the BACTEC™ instrument periodically at 10 min of interval for fluorescence of the vial sensor present in each bottle. The microbiology department had also instituted a ‘critical alert’ system to inform the clinical team about a positive bacterial growth as soon as the culture system demonstrated a ‘beep’ along with a gram stain report, in order to guide the antimicrobial therapy. Subsequently, all the samples were inoculated in a suitable medium [MacConkey agar, blood-agar and Sabouraud Dextrose Agar plate (for yeast)] for organism growth. Matrix-assisted laser desorption ionization-time of flight mass spectrometry (Bruker Biotyper, Bruker Daltonics, Bremen, Germany) was used for the identification of the colonies and antibiotic susceptibility was tested by disc-diffusion and Vitek system.

The records of the blood culture during the study period were extracted from the Hospital Information System (HIS) database as well as from the database of the Department of Microbiology (both online and register based entries). From the pooled hospital HIS data, those pertaining to the study population were extracted using the keywords ‘neonate’, ‘newborn’, ‘baby of’, ‘age in days’ and the ‘staff in-charge’ and the extracted data were checked for duplicates using the bacterial lab serial number before analysis. Missing data were identified and retrieved by matching the extracted data with the computerized database of the central records department.

For this analysis, sepsis was defined as systemic inflammatory response syndrome in the presence of or possibly as a result of suspected or proven infection, modified for preterm neonates [10]. Infection was suspected (suspect sepsis) in neonates with the following clinical signs: hypothermia, hyperthermia, tachycardia, bradycardia, tachypnoea, central cyanosis, abdominal distension, increased pre-feed aspirates, apnoea, grunting, chest retractions, not looking well, refusal to feed, lethargy, seizure and sclerema [11]. Culture positive sepsis was defined as isolation of a recognized pathogen from blood in neonates suspected to have sepsis based on the clinical features (as described above) along with treatment with antibiotics for a minimum duration of 5 days. Contaminant growth was defined as per the Clinical and Laboratory Standards Institute (CLSI) guidelines document M47-A [12]. Antibiotic susceptibility of the isolates was determined as per the CLSI guidelines [12, 13] and was reported as susceptible, intermediate or resistant for each individual antibiotic [14]. MDR was defined as acquired non-susceptibility to at least one agent in three or more antimicrobial categories and extensively drug resistant was defined as non-susceptibility of the bacteria to at least one agent in all but two or lesser antimicrobial categories (i.e. bacterial isolates remain susceptible to only one or two categories) [15].

Data collection and statistical analysis

Incidence rate of culture positive sepsis was calculated by dividing number of neonates with culture positive sepsis by the total number of admissions to the extramural unit. AMR rates were calculated for major antibiotic groups and those antibiotic groups which are intrinsically resistant to certain organisms were not included in the analysis. For depicting the AMR patterns and individual AMR rates for each antibiotic, proportion of organisms sensitive, intermediately sensitive or resistant to a particular group of antibiotics were calculated. Bonferroni correction was done for multiple comparisons. Early-into-admission (growths in the cultures sent in the first 24 h into admission) were compared with late-into-admission (growths in the cultures sent after first 24 h into admission) and strengths of associations were assessed by calculating the relative risk (RR) and their 95% CI. Data entry and statistical analysis were done using Microsoft Excel™ [MS Office (2016), Microsoft, USA[ and IBM SPSS Statistics for Windows, version 26 (IBM Corp., Armonk, NY, USA).

RESULTS

Amongst the 3687 blood samples sent for suspect sepsis, 537 (14.6%) blood cultures were positive from 514 neonates (Figs 1, 2 and Table 1). Gram-positive cocci (GPC) were the most common isolates [240 (45%)] followed by GNB [233 (43.4%)] and fungi [64 (11.9%)] (Fig. 1). CONS contributed to two-thirds of GPC. Klebsiella species [93 (17.3%)] and Acinetobacter species [52 (9.7%)] were the most common GNB isolates.

Fig. 1.

Flow of the study.

Fig. 2.

Profile of the organisms.

Table 1

Description of the study population (n = 514)

	Characteristic	Value
1	Gestational age (weeks); mean (SD)	34.4 (3.2)
2	Prematurity (<37 completed weeks of gestation); n (%)	358 (70)
	<28 weeks	31 (6)
	28–31 weeks	152 (30)
	32–35 weeks	73 (14)
	36–37 weeks	196 (38)
3	Birth or admission weight (g); mean (SD)	2080 (749)
4	Male sex; n (%)	374 (73)
5	Duration of stay (days): median (IQR)	4 (2, 8)
6	State of referrala; n (%)
	Haryana	239/1300 (18)
	Punjab	151/1053 (14)
	Himachal Pradesh	54/320 (16)
	Uttar Pradesh	47/147 (32)
	Chandigarh	9/94 (9)
	Others	14/80 (17)
7	Case fatality rate of culture positive neonates; n (%)	105 (20.4)

a

Denominators are the number of referred neonates from respective states in whom blood cultures were sent.

Out of the culture positive sepsis subjects, 403 (75%) isolates grew in the samples sent within 24 h of admission (early-into-admission). However, the culture positivity rates were similar between early-into-admission vs. late-into-admission cultures (after 24 h from admission) [403 (15.3%) vs. 134 (12.8%), RR (95% CI): 1.19 (0.99, 1.43), p = 0.05]. On comparing the organisms in similar fashion, CONS and Klebsiella species were isolated significantly more commonly in early-into-admission cultures whereas Acinetobacter species were isolated significantly more often in the late-into-admission cultures (Table 2). Significantly higher mortality was observed in subjects with culture positive sepsis [105 (20.4%) vs. 376 (15.2%); RR (95% CI): 1.35 (1.11–1.64); p = 0.003].

Table 2

Comparison of early vs. late-into-admission isolates

	Organism	Early-into-admission (within 24 h from admission)	Late-into-admission (after 24 h from admission)	RR (95% CI)	p
1	Blood culture positive; n/N (%)	403/2640 (15.3%)a	134/1047 (12.8%)a	1.2 (0.99, 1.4)	0.06
2	Acinetobacter species (n = 52)	26 (1%)	26 (2.5%)	0.4 (0.2, 0.7)	0.001
3	Burkholderia species (n = 21)	17 (0.6%)	4 (0.4%)	1.7 (0.6, 5)	0.3
4	Enterobacter species (n = 12)	12 (0.5%)	0 (0%)	4.8 (0.6, 37)	0.09
5	Enterococcus species (n = 19)	12 (0.5%)	7 (0.7%)	0.7 (0.3, 1.7)	0.4
6	Escherichia coli (n = 19)	14 (0.5%)	5 (0.5%)	1.1 (0.4, 3.1)	0.8
7	Klebsiella pneumoniae (n = 93)	84 (3.2%)	9 (0.9%)	3.7 (1.9, 7.3)	<0.001
8	Pseudomonas species (n = 18)	13 (0.5%)	5 (0.5%)	1.03 (0.4, 2.9)	0.9
9	S. aureus (n = 21)	14 (0.5%)	7 (0.7%)	0.8 (0.3, 1.9)	0.6
10	CONS (n = 184)	148 (5.6%)	36 (3.4%)	1.6 (1.1, 2.3)	0.006
11	Streptococcus species (n = 8)	7 (0.3%)	1 (0.1%)	2.8 (0.3, 22.5)	0.3
12	Candida species (n = 33)	21 (0.8%)	12 (1.1%)	0.7 (0.3, 1.4)	0.3
13	Wickerhemomyces (n = 27)	17 (0.6%)	10 (1%)	0.7 (0.3, 1.5)	0.3

Notes: CONS, coagulase-negative staphylococcus; RR, relative risk.

a

All the denominators for calculation of proportions are total number of blood cultures sent.

AMR to carbapenems, third-generation cephalosporins, aminoglycosides, fluoroquinolones and the colistin/polymyxin group of antibiotics was observed in 117/220 (53.2%), 174/230 (76%), 93/186 (50%), 212/422 (50%) and 6/64 (8.1%) isolates, respectively. Resistance to any of the carbapenems was observed in 41/52 (79%), 52/93 (56%) and 6/19 (32%) of Acinetobacter, Klebsiella and Escherichia species, respectively. Methicillin resistance was documented in 132/184 (72%) of CONS and 14/21 (67%) of Staphylococcus aureus. Close to 95% (n = 222) of the GPC that were tested were sensitive to vancomycin.

In the carbapenems group, the resistance to meropenem and imipenem was documented in 57.1% and 49.7%, respectively, whereas amongst the cephalosporins 84%, 77%, 76% and 56% of the isolates were resistant to cefotaxime, ceftriaxone, ceftazidime and cefoperazone-sulbactam, respectively. Bacterial isolates from the referrals from the state of Punjab showed a significantly higher resistance rates for carbapenems and colistin/polymyxins, whereas the state of Haryana showed a significantly higher resistance for carbapenems in comparison to other neighbouring states (Table 3). The resistance to few other antibiotics approached 100% in certain states, but the sample size was too less to draw meaningful interpretations.

Table 3

Comparison of AMR rates amongst major referral states

	State of referral and antimicrobials	Resistance rates [n (%)]	RR (95% CI)	p
1	State of Punjab; n (%)
	Carbapenems	42 (67) vs.75 (48)	1.3 (1.1, 1.7)	0.006
	Third-generation cephalosporins	13 (62) vs. 161 (77)	0.8 (0.6, 1.1)	0.13
	Colistin/Polymyxin	2 (33) vs. 4 (6)	5.7 (1.3, 25)	0.012
	Vancomycin	3 (9) vs. 8 (4)	2.4 (0.7, 8.6)	0.20
2	State of Haryana; n (%)
	Carbapenems	47 (44) vs. 70 (61)	0.7 (0.6, 0.9)	0.01
	Third-generation cephalosporins	78 (72) vs. 96 (79)	0.9 (0.8, 1.1)	0.20
	Colistin/Polymyxin	4 (13) vs. 2 (5)	2.6 (0.5, 13.5)	0.15
	Vancomycin	6 (6) vs. 5 (4)	1.6 (0.5, 5.2)	0.40
3	State of Himachal Pradesh; n (%)
	Carbapenems	7 (33) vs. 110 (55)	0.6 (0.33, 1.12)	0.05
	Third-generation cephalosporins	53 (79) vs. 121 (74)	1.07 (0.91, 1.24)	0.61
	Colistin/Polymyxin	0 (0) vs. 6 (11)	1.07 (0.94, 1.23)	0.33
	Vancomycin	1 (1) vs. 10 (6)	0.26 (0.03, 1.98)	0.15
4	State of Uttar Pradesh; n (%)
	Carbapenems	14 (64) vs. 103 (52)	1.22 (0.87, 1.72)	0.30
	Third-generation cephalosporins	20 (83) vs. 154 (75)	1.11 (0.92, 1.36)	0.40
	Colistin/Polymyxin	0 (0) vs. 6 (10)	1.01 (0.84, 1.22)	0.43
	Vancomycin	1 (6) vs. 10 (5)	1.34 (0.18, 9.87)	0.77

Notes: Punjab, Haryana, Himachal Pradesh and Uttar Pradesh were compared, individually against the combined category of all other states/Union Territories in a binomial categorical fashion after creating a 2 × 2 contingency table. States/Union Territories that were included were Punjab, Haryana, Himachal Pradesh, Uttar Pradesh, Chandigarh, Uttarakhand and Bihar. Last three states were not compared with the rest due to their small numbers.

DISCUSSION

This report is from one of the largest hospital-based cohort of outborn culture positive septic neonates referred from six North and North-western Indian states. The key observations from this study are (i) a high blood culture positivity rate [14.6% (amongst suspected sepsis) and 9.8% (amongst admissions)], similar to that of the Delhi Neonatal Infection Study (DeNIS) outborn cohort and the Young Infant Study Group (YISG) observation reported few years ago [7, 16] but lower than an older report from the National Neonatal-Perinatal Database—2002 (32.2% in outborn neonates) [17]; (ii) a large majority of the neonates were septic at admission; (iii) a significant proportion grew gram-positive cocci dominated by CONS; (iv) a high case fatality rate amongst culture positive sepsis neonates; and (v) an alarmingly high AMR to broad spectrum antibiotics with significant difference between states

Close to 75% of the infections/isolates were identified within 24 h from admission, suggesting that these neonates were already infected at the time of admission. Moreover, a higher GPC isolates in this study is also in stark contrast to the report of the DeNIS outborn cohort report [7]. Biologically plausible reasons for our observations could be unscrupulous use of broad-spectrum antibiotics and poor housekeeping and infection control practices (ICPs) in the referral states, which if proven true, could lead further dangerous increase in AMR. Many large network studies across the world have also reported CONS as a common isolate [18–20]. Similarly, studies from India as well as other LMIC’s have reported GPC as common bacterial isolates [5, 16, 21, 22].

A systematic review from the community acquired neonatal sepsis data from developing countries from Asia and Africa has reported gram-positive bacteria as the most common isolates followed by Klebsiella and Escherichia coli [23]. YISG reported associated focal infections in addition to systemic infections following gram-positive infections [16]. The population characteristics of the above studies were similar to the current report even though neonates in the DeNIS network report were larger in weight and were less premature [24]. More importantly, we observed a statistically significant higher number of CONS and Klebsiella species isolates in the early-into-admission cultures, signifying a strong possibility of these organisms being acquired from the referring hospitals, synonymous to early-onset neonatal sepsis (Table 3).

We observed a lower, albeit significant, fungal sepsis rate in comparison to the DeNIS outborn cohort data (11.5% vs. 22.6%) [7]. This difference could possibly be due to the variability in broad-spectrum antibiotics usage and other risk factors that have been specifically described in relation to fungal infections such as use of central venous lines, use of parenteral nutrition and certain housekeeping practices. However, this aspect needs further exploration.

A high prevalence of multidrug (48%) and extreme drug resistance amongst GNB’s, high prevalence of GNB resistance to carbapenems (53%) and third-generation cephalosporins (76%), a higher colistin resistance (2%) in comparison to the DeNIS outborn cohort (1%) and a significant inter-state difference in the bacterial isolates and their AMR pattern are the stand outs in this report. Similar reports of drug resistance have been published from eastern India [25], from a separate point prevalence study of 41 countries [26] and in another report from five South-East Asia Region countries [27]. Emergence of colistin resistance has been reported in Klebsiella, E. coli, Acinetobacter and Pseudomonas species of bacteria, possibly due to wider use of colistin in carbapenem resistant organisms [28]. The state-wise organism profile and AMR pattern reported by us would serve as a guidance document to set up and strengthen the measures to reduce health care associated infections and improve ICP activities in both public and private sector hospitals and rationalize the use of antimicrobials by implementation of stringent policies, similar to the efforts made by the World Health Organization (Global Antimicrobial resistance Surveillance System) and the Global Antibiotic Resistance Partnership [29, 30].

This study has few important limitations. First, being a retrospective data analysis, this study had limited availability of patient level data for analysis. Second, the current data as well as that of the older studies do not fully reflect the sepsis rates of the peripheral and district hospitals as the cohort that reach the referral hospitals tend to be sicker and are on multiple broad-spectrum antibiotics by the time, they reach the hospital. Even though not a limitation, we did not classify the culture positive subjects as community acquired and hospital acquired, as was done in the DeNIS outborn cohort, due to the arbitrariness involved in such classification [7]. Moreover, based on our anecdotal experience, we chose to use 24 h as the cut-off point to classify the organisms as outside hospital acquired vs. study hospital acquired instead of 48 or 72 h from admission [7]. Despite all these limitations and variations, we believe that the current report is one of the largest studies done in the country in an effort to understand the microorganisms profile and their AMR pattern in a cohort of outborn referral neonatal population.

To conclude, a 14% culture positive sepsis rate was observed in the outborn neonates referred to the index hospital. Majority of the isolates grew in the first 24 h of admission implying these neonates were already septic at admission. CONS, Klebsiella species and Acinetobacter species were the top three common isolates. Close to 50% of the GNB isolates were multidrug resistant and half of them were also extreme drug resistant. A significant state wise difference in organism profile and their AMR patterns were observed implying the need for a more detailed state wise microbiological data analysis and state-specific strategies for Antibiotic Stewardship Program.

ACKNOWLEDGEMENTS

We acknowledge the contribution of the nursing staff and the residents of the Department of Pediatrics, who helped in completing this study.

Author contributions

Dhir SK and Sundaram V conceptualised the study and the design, analysed the data, wrote the first draft of the manuscript as well as critically edited the manuscript. Gautam V, Munda VS and Tiewsoh JBA did the microbiological analysis, analysed the data of the microorganisms for antibiotic susceptibility and critically edited the manuscript. Angurana SK, Kumar J, Saini SS, Dutta S and Kumar P helped in study designing and implementation, provided critical inputs in data analysis and critically edited the manuscript. All the authors give approval to the final version of the draft and would remain accountable for all aspects of the work associated with this paper.