Skip to main content

Understanding antimicrobial resistance in bacteria from neonatal sepsis cases in low- and middle-income countries

Gram-positive bacteria, such as Staphylococcus aureus, are among the main causes of neonatal sepsis in low- and middle-income countries (LMICs), yet only limited information is available on their diversity and antimicrobial resistance (AMR) properties. Now, a new study led by the Ineos Oxford Institute (IOI) for Antimicrobial Research has characterised more than 100 S. aureus isolates from neonatal sepsis cases in six LMICs, revealing that these bacteria are diverse and display high levels of resistance to antibiotics.

Despite substantial progress in preventing child mortality in recent decades, every year there are still around 2.4 million children who die in the first month of life, with infections remaining a major cause of neonate death, particularly in LMICs. Although the main culprits of infections in newborns in these locations are thought to be gram-negative bacteria, such as Klebsiella pneumoniae and Escherichia coli, previous studies have also found that gram-positive bacteria, particularly S. aureus, are frequently associated with neonatal sepsis. However, owing to limited resources and the lack of simple and cost-effective methods for their identification, gram-positive bacteria isolated from blood cultures of infants in LMICs are rarely studied. As such, detailed information about the diversity of such bacteria and their characteristics – including their ability to resist antibiotics – is lacking.

To overcome this problem, Dr. Kirsty Sands and colleagues decided to characterise gram-positive bacteria isolated from neonatal blood cultures from six LMICs in Africa and South-Asia (Bangladesh, Ethiopia, Nigeria, Pakistan, Rwanda and South Africa). The clinical isolates were collected between 2015 and 2017 as part of the BARNARDS study, aimed at assessing the burden of AMR in neonates from developing nations. While BARNARDS was originally designed with an emphasis on carbapenem resistance in neonatal sepsis caused by gram-negative bacteria, the high reporting of sepsis linked to S. aureus during a network meeting prompted the researchers to retrospectively recover and systematically characterise gram-positive isolates in addition to gram-negative ones.

Whilst our primary aim with BARNARDS was to characterise the gram-negative bacteria, with particular interest in carbapenem resistance, when we were discussing our data in a network annual meeting in Kigali, Rwanda, it became clear we also needed to look into staphylococci.

In the recent study, the authors recorded 1,266 gram-positive isolates from 1,239 neonates, and identified 296 of these isolates, with the majority (271/296, or 92%) being Staphylococci. Of these 271 staphylococci, 101 (37%) were S. aureus and 170 belonged to a heterogeneous group of coagulase-negative staphylococci (CoNS) that includes Staphylococcus epidermidis, Staphylococcus haemolyticus and Staphylococcus sciuri (recently renamed as Mammaliicoccus sciuri).

To better characterise these bacteria, the authors then sequenced the genomes of multiple isolates, including 100 S. aureus and 30 CoNS isolates. Notably, although the S. aureus isolates were quite variable (belonging to 18 different sequence types, or STs), the data revealed clear epidemiological clusters in Pakistan (belonging to ST8) and South Africa (belonging to ST5). Furthermore, these clusters carried signatures of methicillin resistance (MRSA), and both were associated with higher neonate mortality compared to other S. aureus clusters. Similarly, while the CoNS isolates were also diverse across countries, the analysis revealed epidemiological clusters – particularly for M. sciuri in Pakistan – and many of the genomes of the CoNS bacteria contained signatures associated with AMR.

To better assess the AMR profile of the different isolates, the authors then performed antimicrobial susceptibility testing for 144 isolates (100 S. aureus and 44 CoNS isolates) against 14 antibiotics. These data revealed that resistance is widespread across the bacteria recovered from the blood of neonates in LMICs, with one third of the S. aureus isolates being resistant to methicillin (MRSA), which was also observed in more than a third of the CoNS isolates (11/18 (61%) of S. epidermidis isolates; 9/11 (82%) of S. haemolyticus isolates; and 2/13 (15%) of M. sciuri isolates).

Resistance is widespread across the bacteria recovered from the blood of neonates in LMICs, with one third of the S. aureus isolates being resistant to methicillin (MRSA)

Despite its limitations, notably the heterogeneity in the reporting and storage of gram-positive isolates across clinical sites, the present study provides an important characterisation of Staphylococcus spp. associated with neonatal sepsis across multiple LMICs. Collectively, these data demonstrate that these bacteria are diverse, and that AMR is common in these clinical isolates. Furthermore, the analyses suggest that certain lineages of S. aureus (such as ST5 and ST8) may be more pathogenic and problematic in neonatal sepsis in LMICs, highlighting the need for increased surveillance in these countries and additional studies elucidating how genetic differences are linked to disease severity. Increased surveillance might also reveal additional insights into the contribution of other gram-positive bacteria to neonatal sepsis, as highlighted by the description of a cluster of M. sciuri in Pakistan. Finally, such studies could also provide important clues to understand the differences between LMICs and high-income countries, where Group B Streptococci (GBS) – which was not detected in the present study – are considered the dominant cause of neonatal sepsis.

Additional studies, aimed at providing a more comprehensive picture of the contribution of gram-positive bacteria to neonatal sepsis in LMICs are already being planned. As explained by Dr. Kirsty Sands:

“Whilst our primary aim with BARNARDS was to characterise the gram-negative bacteria, with particular interest in carbapenem resistance, when we were discussing our data in a network annual meeting in Kigali, Rwanda, it became clear we also needed to look into staphylococci. S. aureus is widely recognised as a pathogen, however when it comes to coagulase negative Staphylococci (CoNS), it is tricky because these species commonly reside on the skin and in many cases are still considered contaminants. Similarly, we may have failed to recover other gram-positive bacteria in the present study, like the Streptococci, because they are trickier to grow from blood cultures. So the next question is: can we explore any association between antimicrobial resistance, virulence factors, pathogenicity and neonatal outcome with a larger designed study? Evaluating the true burden of gram-positive neonatal sepsis is very much a core aim of the next phase of BARNARDS, which we will be carrying out in collaboration with Dr. Owen Brad Spiller at Cardiff University. As part of this continuation of BARNARDS, we also aim to support increased capacity of local microbiology diagnostics in the participating countries to cover both gram-negative and gram-positive bacteria, to ensure that this next phase of the study provides an even more accurate representation of the main culprits of infections in newborns in LMICs”.

Read the full paper: Sands, K., Carvalho, M.J., Spiller, O.B. et al. Characterisation of Staphylococci species from neonatal blood cultures in low- and middle-income countries. BMC Infect Dis 22, 593 (2022).