Antimicrobial resistance in neonates with suspected sepsis

SETTING: Nobel Medical College and Teaching Hospital, Biratnagar, Nepal. OBJECTIVE: To determine the pattern of antimicrobial resistance and hospital exit outcomes in neonates with suspected sepsis in a tertiary neonatal intensive care unit (NICU). DESIGN: This hospital-based cohort study was conducted to follow patients from January to December 2019. All identified cases of suspected sepsis were enlisted from hospital records. RESULTS: Sepsis was suspected in 177 (88%) of the 200 cases admitted in the NICU; 52 (29%) were culture-positive. Pseudomonas was the predominant organism isolated (n = 40; 78%), followed by coagulase negative staphylococcus (n = 12, 23%). Nine (17%) of the 52 isolates were resistant to the Access and Watch group of antibiotics, including some resistance to Reserve group drugs such as imipenem and linezolid. Most treated cases (n = 170, 96%) improved, although 7 (4%) left against medical advice. CONCLUSION: Most of the pathogens were resistant to WHO Access and Watch antibiotics and occasional resistance was observed to Reserve group drugs. Most sepsis was caused by Gram-negative bacilli. Improving turnaround times for antibiotic sensitivity testing using point-of-care testing, and a greater yield of culture-positive results are needed to enhance the management of neonatal sepsis.

N eonatal sepsis refers to generalised bacterial infection with systemic signs, and is diagnosed using a positive blood culture in the first 4 weeks of life. 1,2 It is classified by time of occurrence of infection as early onset sepsis, i.e., bacteraemia in the first 72 h of life, and thereafter, as late onset sepsis. 3 It is a life-threating illness and a major cause of neonatal morbidity and mortality worldwide. 3 Global incidence is estimated to be 1.3 million cases annually, resulting in about 203,000 attributable deaths. Estimates that 2,202 neonates per 100,000 live births develop neonatal sepsis globally, with a mortality rate of 11-19% (242-418 deaths/100,000 live births), are derived from high-income country data. 4 However, lowand middle-income countries, where data are limited, are most affected due to the frequency of infections and limited access to good healthcare.
Due to the non-specific nature of the signs and symptoms in the new-born, most cases of sepsis are clinically diagnosed, then treated with empirical broad-spectrum antibiotics. This use of antibiotics promotes antimicrobial resistance (AMR). Bedside tests are not available and empirical antibiotic treatment contributes to the increased prevalence of antibiotic-resistant organisms. Examples include methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus and multidrug-resistant Gram-negative rods. 5,6 At least 700,000 deaths occur annually due to drug-resistant diseases, which is projected to rise to more than 10 million deaths globally by 2050 if action is not taken. 7 The spectrum of multidrug-resistant bacteria varies. In neonates, sepsis 48 h after birth is most often caused by Group B streptococcus (43%) and Escherichia coli (16-29%). 8,9 For very low birth weight infants with early onset sepsis, the E. coli rate exceeds that of Group B streptococcus (5.1 vs. 2.1/1,000 live births). 9 Late onset sepsis is mainly caused by Gram-positive bacteria (49%), most often coagulase-negative Staphylococcus (45%). Gram-negative late onset sepsis is less common (23%), but is associated with greater mortality in neonatal intensive care units (NICUs) (19-36%). 9 Both forms of sepsis are related to birth weight and the location where the infection was acquired. 10 The neonatal mortality rate in Nepal is 21/1,000 live births. 11 Neonatal infection is the third most common cause of neonatal mortality (17%; 3.6/1,000 live births), after preterm birth complications (33%) and intrapartum-related events (23%). 11 In Nepal, 10-42% of all the patients are prescribed antibiotics for therapeutic or prophylactic purposes without bacterial confirmation or susceptibility testing. Further data on AMR in Nepal are limited, possibly due to a lack of prioritisation and the absence of routine surveillance. 12 An appropriate antimicrobial agent should be used for sepsis to prevent the development of multidrug resistance and reduce morbidity and mortality. 13 Nepal has significantly more multidrug-resistant bacterial strains closer to health care sites. 12 In one tertiary hospital, coagulase-negative Staphylococcus was the predominant Gram-positive organism isolated from both early and late onset sepsis cases (46.6%), followed by S. aureus (14.6%). Other Gram-negative multidrug-resistant bacteria were associated with early onset sepsis: Acinetobacter species (9.5%) and Klebsiella pneumoniae (7.7%). 1 Antibiotics used in Nepali hospitals were found to be effective in curing only 50% of infections. 14 Furthermore, AMR contributes to increased treatment costs, hospital stay, morbidity and mortality. 15

Public Health Action
In Nobel Medical College and Teaching Hospital (NMCTH), Biratnagar, Nepal, a public-private partnership tertiary care referral centre, sepsis is the third most common cause of admission in the NICU, after asphyxia and meconium aspiration. 16 However, there is no information about the AMR pattern, particularly in relation to patient outcomes. In the present study, we described the bacterial profile and determined the antimicrobial resistance pattern and hospital exit outcomes in neonates with suspected sepsis in the NMCTH NICU.

Study design
This was a hospital-based, cohort study conducted from January to December 2019. Nepal, in South Asia, is mainly located in the Himalayas, with an estimated population of 30.2 million. 17 NMCTH is in Biratnagar, the second most densely populated and third most populous city of Nepal, with a population of 242,548 (2011 census; https://unstats.un.org/unsD/demographic/sources/ census/wphc/Nepal/Nepal-Census-2011-Vol1.pdf).

General setting
Nepal is moving towards an integrated response to AMR. The National Action Plan, led by the AMR Steering Committee at the Curative Division of Ministry of Health and Population, recognises the National Public Health Laboratory (under the Department of Health Services) as the focal point for AMR activities.

Specific setting
NMCTH, affiliated to Kathmandu University, has 911 beds, 40,000 admissions per year and a bed occupancy rate of ~80%. The average outpatient flow is ~1000/day. The NICU is well equipped, with 17 beds, a nursing staff: patient ratio of 1:2-3 and the highest volume referral for tertiary care in eastern Nepal.

Laboratory testing
Blood samples from the neonatal ward/emergency department were collected in brain heart infusion (BHI) broth (1 ml blood in 5 ml of BHI broth). After 24 h culture, the sample was sub-cultured onto blood and MacConkey agar (aerobic, 35°C) and chocolate agar (anaerobic). Any organism grown was identified using standard biochemical tests, followed by antimicrobial susceptibility testing (Clinical and Laboratory Standard Institute guidelines). 18 The timing of the reporting of culture results was noted (standard: within 24-48 h), as was timing of the reporting of the resistance pattern (standard: 72 h).

Study population/participants
All neonates admitted to the NICU of NMCTH, Biratnagar, Nepal, between 1 January and 31 December 2019 were ascertained from hospital records. We identified cases of suspected sepsis using criteria adapted from All India Institute of Medical Sciences Protocols in Neonatology (Table 1). 19

Sources of data and data collection
From the clinical and laboratory records and discharge summary, we collected sociodemographic characteristics, risk factors associated with sepsis, microorganism growth and antibiotics susceptibility patterns, and hospital exit outcome in relation to septicaemia treatment using a structured pro forma.

Data validation
All data were double-entered and validated using Epi-Data database v3.1 for entry (EpiData Association, Odense, Denmark). Data were analysed using SPSS v21 (IBM SPSS, Armonk, NY, USA).

Analysis and statistics
The number and proportion of neonates admitted for suspected septicaemia and with samples sent for blood culture were calculated, along with median and interquartile ranges for the time taken for laboratory procedure, timing of changes in regimen from the issuing of the laboratory report and duration of empirical treatment. Poisson regression modelled any association between the demographic and clinical characteristics, hospital exit outcome, potential confounders and antibiotic resistance. The level of significance was set at P  0.05.

Ethics approval
Permission for the study was obtained from Institutional Review Committee (IRC), Nobel Medical College and Teaching Hospital (NMCTH) (Ref: IRC-NMCTH 162/076 of 29/01/2020). The study was also approved by the Ethics Advisory Group of the International Union Against Tuberculosis and Lung Disease, Paris, France (EAG number: 08/20, date 2020-02-05). As this was a record review with no patient identifiers, the issue of informed patient consent did not apply.

Sepsis
Sepsis was clinically suspected in 177 (88.5%) cases. More than two thirds of them presented with features of early onset sepsis, while many also had a range of comorbidities. Most of the sepsis cases were inborn, Training Initiative (SORT IT) (Table 2). Among inborn neonates, early onset sepsis (n = 100, 88.7%) was more common, whereas late onset sepsis (n = 38, 71.7%) was more common in outborn neonates.
Out of the total samples tested (n = 177), 52 (29.3%) were culture-positive, of which in 36 (69.2%) the empirical antibiotic regimen was revised as per blood culture sensitivity pattern. Almost all cases admitted with sepsis improved with the given regimen ( Table 2). Gram-negative bacterial infection was the main cause of septicaemia among neonates aged 1 week at admission, predominantly affecting males, inborn babies and those with low birth weight. Gram-negative sepsis was more common in babies of all gestational ages, and was most common in both early and late onset neonatal sepsis (Table 3).

Antimicrobial resistance
Pseudomonas predominated, followed by coagulase-negative Staphylococcus (Table 4). Overall, 9 (17.0%) culture-positive cases of septicaemia were resistant to the AWaRe (Access, Warch, Reserve) group of antibiotics. The most common isolate was sensitive to Access and Watch group antibiotics, whereas occasional resistance was observed with the Reserve group of drugs (Table 5).

Turnaround times
The median turnaround time for laboratory culture was 3 days (range 2.0-4.0). The median time to change empirical therapy after a positive culture and sensitivity report was 3 days (range 3.0-6.0).

Associations with culture-proven sepsis
Of culture-positive cases, two characteristics showed a statistically significant effect on positive culture results: presence of comorbidities reduced the risk of a positive culture (adjusted relative risk [aRR] 0.23, 95% confidence interval [CI] 0.09-0.57), while being post-term increased the risk (aRR 1.33, 95% CI 1.05-1.69) ( Table  6). None of the factors which affected the risk of a positive culture were significantly associated with single drug resistance.

Demographic and clinical characteristics
In a tertiary NICU in Nepal, over 88% of cases were diagnosed with clinically suspected sepsis, of which two thirds presented with features of early onset sepsis. However, under one third had culture-proven septicaemia, mainly caused by Gram-negative bacteria. Risk factors for sepsis were not identified; AMR was noted but not statistically associated with risk factors or exit outcome of the neonate from NICU. The presence of comorbidities was associated with a reduced risk of being culture-positive.

Neonatal sepsis
Clinical sepsis was confirmed in under one third of neonates. Our confirmation rate is in the middle of other studies in Nepal and elsewhere. [20][21][22][23][24][25] Empirical treatment may have started before blood was taken for culture or before admission to the NICU. Early use of antibiotics before taking blood for culture is known to lead to sterile culture reports. This needs further investigation, possible remedial action within the NICU and obstetric unit and consideration if the integrated preventive and protective measures against infections have failed, and why, or if there are other factors influencing the high rate of unconfirmed sepsis.
Among suspected sepsis cases, more than two thirds presented with features of early onset sepsis, which is higher than other nearby reports. [24][25][26] Probable explanations include 1) most of our new-borns were inborn with subsequent early identification of clinical sepsis and prompt admission to NICU; and 2) the causative agents for early onset sepsis were acquired from the mother during birth. 1 Further investigation into the possible acquisition of sepsis at birth is needed to protect vulnerable new-borns.

Public Health Action
The predominance of male new-borns has been reported elsewhere. 2,24,27,28 This may be due to the prevalent societal preference given to males in seeking medical care. 26 However, males may also be vulnerable to neonatal sepsis due to an X-linked immune regulatory gene factor. 29 With improved confirmation of sepsis, it would be possible to investigate this sex difference and begin to counter any bias against female infants.

Public Health Action
We found that comorbidities (meconium aspiration syndrome, bowel obstruction, etc.) significantly reduce the risk of having a positive culture. However, there was a lot of missing data, possibly biasing this unexpected finding. Comorbidities could be expected to predispose to the overuse of broad-spectrum antibiotics, leading to sterile cultures. In addition, NMCTH receives referral cases, many with comorbidities and previous treatment, from different parts of Nepal.
There are at least two approaches to sterile cultures. First, any use of empirical antibiotics needs to be kept to a minimum by more rapid turnaround times from the laboratory, or (preferably) quick and accurate bedside point-of-use diagnostic tests for sepsis, thereby reducing the need for empirical treatment, as in other infections. 30,31 Second, enhanced professional and continuing education of all health care providers, wherever situated, would ensure proper understanding and awareness regarding antimicrobials, including the collection of blood samples for culture before starting empirical therapy.

Antimicrobial resistance profile
Gram-negative organisms are the most common agent causing neonatal sepsis, 2,25,32 as we were able to confirm (Table 4). However, few studies have reported Gram-positive bacterial septicaemia among neonates in the hospital setting. Among these, methicillin-resistant S. aureus was a critically important pathogen in the NICU population, associated with both endemic and epidemic infections. 20,28 Host and bacterial properties unite to cause sepsis in newborns. The organism varies according to the environmental conditions of the hospital, e.g., contamination during sampling, care of both in-born and out-born cases, and surgical cases mixed with medical cases. 26 Contamination during sampling or during birth from the mother's birth canal could account for coagulase-negative Staphylococcus. The latter is more life-threatening, and further investigation to distinguish between the two sources is important. In our study, the predominance of Pseudomonas suggests transmission of infection from healthcare workers to patients via contaminated hands. Accordingly, re-enforcement of infection control guidelines and education of healthcare staff are important. Our study further prompts the need for the implementation of the WHO AMR action plan in Nepal.
The most common antibiotic resistance was in Pseudomonas, unlike a previous report from Nepal, 26 which showed that Gram-negative organisms were resistant to ampicillin. Other isolates in our study were, not surprisingly, resistant to first-line and second-line therapy. A potential way to curb this would be increased and timely surveillance of antibiotic sensitivity patterns and revision of policies regarding empirical therapy based on AMR, along with the implementation of the aforementioned point-of-use diagnostic testing for sepsis.

Turnaround time for samples and empirical therapy
The longer the administration of broad-spectrum antibiotics, the greater the progression of severe sepsis to septic shock, with in- Public Health Action creased mortality. 33 The median turnaround time taken for laboratory culture plus the median time to changing to an appropriate antibiotic regime means that it is possible for a septic neonate to wait 6 days for a necessary change to appropriate treatment. This again indicates the need for rapid diagnostic tests.

Strengths
The study is one of the few which report AMR and the turnaround time for the reporting of culture results and consequent change in antibiotic management.

Limitations
The small numbers of positive blood cultures meant that none of the risks likely to be associated with resistance to at least one antibiotic or to a hospital exit outcome were fully assessed. Second, as critically ill patients who left against medical advice were likely to have died, our estimate of the mortality rate may have been an underestimate.

CONCLUSION
Pseudomonas was the most common isolate in culture-proven sepsis. Gram-negative bacilli were susceptible to Access and Watch group antibiotics, whereas, occasional resistance was observed with last resort (Reserve) antibiotics. Rapid diagnostic tools for sepsis, which would also identify AMR and conform to accepted standards and guidelines, are urgently needed in NICUs to enable the antibiotic regime to be appropriate from the earliest time possible. As sepsis is a major health risk for neonates, a standardised national and global surveillance programme collecting neonatal AMR data, including risk factors and clinical outcomes, is critical. This would enable benchmarking, the identification of remedial risks, the design and implementation of successful interventions, and allow review of NICU programmes.
Public Health Action (PHA) welcomes the submission of articles on all aspects of operational research, including quality improvements, costbenefit analysis, ethics, equity, access to services and capacity building, with a focus on relevant areas of public health (e.g. infection control, nutrition, TB, HIV, vaccines, smoking, COVID-19, microbial resistance, outbreaks etc). This is an Open Access article distributed under the terms of the Creative Commons Attribution License CC-BY 4.0 published by The Union (www.theunion.org). Contact: pha@theunion.org Information on PHA: http://www.theunion.org/what-we-do/journals/pha