Abstract
Late-onset neonatal sepsis (LONS) remains an important threat to the health of preterm neonates in the neonatal intensive care unit. Strategies to optimize care for preterm neonates with LONS are likely to improve survival and long-term neurocognitive outcomes. However, many important questions on how to improve the prevention, early detection, and therapy for LONS in preterm neonates remain unanswered. This review identifies important knowledge gaps in the management of LONS and describe possible methods and technologies that can be used to resolve these knowledge gaps. The availability of computational medicine and hypothesis-free-omics approaches give way to building bedside feedback tools to guide clinicians in personalized management of LONS. Despite advances in technology, implementation in clinical practice is largely lacking although such tools would help clinicians to optimize many aspects of the management of LONS. We outline which steps are needed to get possible research findings implemented on the neonatal intensive care unit and provide a roadmap for future research initiatives.
Impact
-
This review identifies knowledge gaps in prevention, early detection, antibiotic, and additional therapy of late-onset neonatal sepsis in preterm neonates and provides a roadmap for future research efforts.
-
Research opportunities are addressed, which could provide the means to fill knowledge gaps and the steps that need to be made before possible clinical use.
-
Methods to personalize medicine and technologies feasible for bedside clinical use are described.
Similar content being viewed by others
Introduction
Neonatal sepsis is one of the main causes of death and morbidity in preterm neonates worldwide.1,2,3 Preterm neonates are especially vulnerable for acquiring late-onset neonatal sepsis (LONS). The incidence of LONS, defined as sepsis onset >72 h after birth, is up to 20% among very low birth weight neonates (VLBW; birth weight <1500 g) with 40% of those with a birth weight <750 g having more than one LONS.2,4 Despite extensive research efforts in the past decades, many preterm neonates do not recover from LONS or survive with risk of neurodevelopmental impairment.5,6
Strategies to optimize care for preterm neonates with LONS are likely to improve survival and outcome, but many important questions on how to improve the prevention, detection, and therapy of LONS in preterm neonates remain unanswered. This review focuses on identifying knowledge gaps in the management of LONS, which could provide important information to improve the outcome of preterm neonates suffering from LONS. We recognize the importance of a clear definition of neonatal sepsis for further research efforts and easy comparison of study results, but this will not be within the scope of this review as it is currently still under debate.7,8,9
Figure 1 shows the conceptual framework for neonates at risk for LONS and the most important knowledge gaps. Below, we will focus on these knowledge gaps for the management of LONS, illustrate their importance, and suggest a roadmap for future research efforts.
This framework shows opportunities for optimizing components of LONS management and key objectives (keys) that need to be addressed. LONS late-onset neonatal sepsis, PK/PD pharmacokinetics/pharmacodynamics, PTX pentoxifylline, IVIG intravenous immune globulin, GM-CSF granulocyte–macrophage colony-stimulating factor, MSC mesenchymal stromal cells.
Prevention of LONS
In many neonatal intensive care units strategies for preventing LONS are implemented, such as hand hygiene protocols, wearing face masks, and care bundles to reduce catheter-related LONS. Below we describe supplements and drugs that could benefit prevention of LONS (summarized in Table 1).
Probiotics
Delayed bacterial colonization and an abnormal gut microflora have been observed in VLBW neonates. This might lead to interrupted integrity of the intestinal barrier, causing bacteria to translocate into the bloodstream.10 With the aim to prevent this, prophylactic use of probiotics was introduced. There is conflicting evidence on the impact of probiotic treatment on the development of LONS. A recent meta-analysis of 67 trials in 15,712 patients showed a reduced overall mortality and incidence of NEC in VLBW neonates, but the risk reduction for culture proven sepsis was non-significant. At the same time the authors acknowledge that there was large heterogeneity in the included studies.11 Another meta-analysis of 37 randomized controlled trials (RCT) showed that probiotics only decreased the risk of acquiring LONS in mother’s milk fed preterm neonates.12 Both analyses conclude that only probiotic mixtures, not single-strain products, were effective in reducing LONS incidence.11,12 This was supported by a recent network meta-analysis by Chi et al., which showed that different combinations of pre- and probiotics could be beneficial for different neonatal endpoints.13
Knowledge gaps and opportunities for future research
In the future, more focus should go to clarifying the relationship between probiotic supplementation and feeding practices in preterm neonates. Additional research should also focus on type of probiotic microorganisms, dose, starting time of and duration of administration to make the available evidence more comparable. The recent network meta-analysis could be used as a starting point for strain combinations tot test.13
Mother’s milk and donor milk
Feeding preterm neonates with their own mother’s milk has a lot of benefits including decreased rates of LONS, necrotizing enterocolitis (NEC), and improved neurodevelopmental outcomes.14,15,16,17 If mother’s milk is unavailable, donor human milk (DHM) can be used as substitute, with several limitations. Unfortunately, the process of pasteurization of DHM, when performed, destroys or significantly decreases the concentration of many of the essential protective molecules.18,19,20,21 Feeding DHM instead of formula did not significantly affect mortality or frequency of LONS.22,23 However, DHM feeding might still be more beneficial compared to formula feeding as a meta-analysis showed DHM compared to formula feeding resulted in a decreased risk of developing NEC.23 More recently, research groups have been focusing on the specific effects of colostrum which is very rich in protective immune components.24,25 However, due to its low quantity, administration via a feeding tube is difficult and extremely preterm neonates can be intolerant to feeding. Oromucosal administration of very low quantities every 3 h has been associated with higher sIgA and lactoferrin concentrations in a group of preterm infants.24 Currently, the first study on the effect of colostrum on neonatal outcome is ongoing.26
Knowledge gaps and opportunities for future research
It is important to understand which components of human milk are essential for the protective effect against LONS. Next, processes of preparation and storage of DHM need then to be adapted to retain these essential components and subsequently the adapted DHM should be evaluated in patient studies. Also, beneficial components that reduce LONS incidence may be used to improve formula feeding. Additionally, the potential benefits of early colostrum administration needs to be further studied in RCTs.
Maternal immunization
As IgA is transferred in mother’s milk, it might be also interesting to think about immunizing mothers with the pathogens found in routine skin and rectal swabs. Several studies on the protective role of breastmilk in women vaccinated during pregnancy identified a high amount of vaccine-specific sIgA in the breastmilk samples up to several weeks postpartum.27,28,29 Consequently, a lower incidence of respiratory illness with fever episodes in young infants of influenza-vaccinated mothers was reported.28,29 The potential of neonatal immunization against Gram-negative bacteria has also been described, either a maternal vaccination to pregnant woman, via breastmilk and directly to the neonate.30
Knowledge gaps and opportunities for future research
Protection offered to the newborn via secretory antibodies present in mothers milk support need for robust immunization strategies that include pregnant women. Effective protection of pregnant women may translate to dual protection to both mother and the infant. Maternal vaccination has already been shown to protect the neonate from severe infections and currently represents the best preventive option against various pathogens. Development of vaccination strategies in pregnant women with imminent preterm labor or possibly soon after preterm birth could improve prevention of LONS by enhancing the immune system of the neonate. These strategies should be evaluated in prospective, preferably randomized, studies.
Lactoferrin
Lactoferrin is a major iron-binding glycoprotein in mature human milk with direct antimicrobial and anti-inflammatory effects.31 In several small RCTs the oral supplementation of bovine lactoferrin (bLF) was associated with a decrease of LONS incidence, although studies were small with high risk of bias and possibility of publication bias was noted.32,33 Until now, however, no adverse effects of bLF in neonates have been reported.33 Larger, already ongoing studies will hopefully provide more data on the effectiveness of lactoferrin supplementation.
Knowledge gaps and opportunities for future research
Next to a potential interaction between probiotics and bLF, which might warrant further study, optimal dosing regimens, types of lactoferrin (human vs bLF), and their effects on long-term outcomes need to be determined.22
Monoclonal antibodies
The use of monoclonal antibodies to modulate the neonatal immune system to prevent LONS is interesting. Most evidence to prevent staphylococcal sepsis in preterm neonates has been reported on pagibaximab. In several smaller studies, a good safety and tolerability profile was demonstrated; however, no effect on LONS was established so far.19,34,35 Pagibaximab is a humanized mouse chimeric monoclonal antibody directed against lipoteichoic acid (LTA), a major cell wall component of Gram-positive bacteria, which promotes staphylococcal phagocytosis.36 The ability of two other antistaphylococcal immunoglobulins, Altastaph and INH A-2, to augment the neonatal immune system to prevent infections has also been studied although not recommended yet.37 A new promising humanized immunoglobulin SpAKKAA-mAb bound with high affinity, blocked IgG binding to protein A, and provided protection against S. aureus sepsis in neonatal mice.38
Knowledge gaps and opportunities for future research
Further research should focus on exploring the development and use of monoclonal antibodies to prevent LONS. Phase 1–3 trials are required for the transition to neonates. The pharmacokinetic (PK) knowledge on monoclonal antibodies and immunoglobulins in preterm neonates may highly benefit from the approach by Malik et al.39 who described the PK profile of monoclonal antibodies in preterm neonates using a physiologically based pharmacokinetic (PBPK) model, based on the reported PK of pagibaximab and palivizumab, as well as immunoglobulins. This enables a prediction of the disposition of other as well as new monoclonal antibodies and immunoglobulins in preterm infants.
Early detection of LONS
Timely initiation of adequate antibiotic and supportive therapy is critical to improve survival of patients suffering from LONS.40 In most cases of LONS the first clinical symptoms are subtle and non-specific (e.g. instability of feeding, temperature, or breathing), but the course can be fulminant and leading to death within a few hours.41,42
Physiological parameters
Ideally, non-invasive methods should be able to predict the onset of LONS. Up until now, several physiological parameters (physiomics) have been investigated, of which heart rate variability (HRV) is the most widely used. Multiple mathematical computations of the heart rate can be used to monitor changes in HRV.43,44 Changes in HRV are associated with the onset of LONS and HRV monitoring has been shown to reduce mortality in VLBW neonates in a large randomized controlled trial.45 However, the specificity is low, which may lead to more blood cultures taken and an increased use of antibiotics.46,47 It was already shown that adding cross-correlation of SpO2 with HRV, or respiratory rate and body temperature, increased the predictive value for occurrence of LONS and NEC, although specificity remained low.48,49 Predictive ability improved when adding physiological parameters to baseline characteristics and conventional risk factors of LONS.50 Also, adding ECG-derived estimates of neonatal movements adds information on top of HRV.51
Knowledge gaps and opportunities for future research
Despite the common knowledge that vital signs change in patients with LONS, predictive monitoring is not used widely in the NICU due to limited predictive ability (low specificity). Future research should focus on improving specificity by combining multiple physiological parameters and adding new non-invasive continuously measured physiological parameters such as skin conductance, changes in microcirculation, and peripheral pulse rate variability. The rapid progress in the development of electronics will produce other non-invasive techniques observing the infants that may outperform those currently available. Combining non-invasive signals with rapidly available (serum) biomarkers is a promising scenario to investigate.
Using today’s technology of big data mining combined with machine learning (ML) and artificial intelligence (AI) approaches52,53,54,55,56 will most likely allow to characterize those with changing physiology indicative of LONS. Data incorporated in ML/AI models include electronic health record data, vital sign analysis, administered therapy, and rapidly available biomarkers. Unfortunately, in adult ICU most ML and AI models remain in prototyping environment and are not tested in bedside settings.57 Challenges lie in data management (sample size, privacy regulations, model generalizability, standardized data formats), development of models (ensuring data quality, preprocessing of data, data labeling, measurement errors), and implementation in clinical workflow (lack of insight in model development limiting possibility of validation studies, concerns for patient safety).57,58,59 Developed predictive algorithms should be prospectively validated, ideally in randomized controlled trials, for their effect on short- and long-term outcome. To improve model performance and generalizability, initiatives should be employed to share high-frequency vital sign datasets with paired outcome data including a harmonized definition of sepsis in order to give researchers the possibility to develop predictive models with the biggest dataset possible.
Serum biomarkers
“Classical” biomarkers
Many studies investigating single biomarkers or a combination of several biomarkers to diagnose or rule out LONS have been performed without establishing an adequate solution for clinical use.60,61,62,63,64,65,66 In addition, the dynamics of many biomarkers in response to LONS as well as the different phases of a neonates immune response have not been fully understood.67 It seems unlikely that a single biomarker will have sufficient predictive value to accurately diagnose or rule out LONS. This emphasizes the need for profiles of multiple biomarkers to identify the status of infections as well as the neonatal immune response in order to improve diagnostic ability.68 A solution could be found in -omics approaches.
Systems biology; -omics approaches
Relatively novel systems biology approaches such as proteomics (simultaneous determination of large numbers of proteins), transcriptomics (RNA transcripts), and metabolomics (cellular metabolites) seem promising to detect the changing (patho)physiology of a preterm neonate during LONS.69 Mass spectrometry (MS) and liquid chromatography techniques have made high-throughput determination of -omics platforms possible.70,71,72 Several studies have investigated the use of -omics profiles in diagnosing LONS, yielding promising results.73,74,75,76,77,78,79,80,81,82,83 Several biological metabolic pathways, like glucose metabolism, oxidative stress, and the fatty acid pathway,77,78,79,80 have been shown to be involved in LONS and are promising targets to investigate. In the largest proteomic study so far including data from 258 predominantly preterm neonates,74 biomarker discovery was performed and subsequently, a composite score (ApoSAA score) was constructed. Then, prospectively validated in a single center setting, the ApoSAA score showed high specificity and sensitivity (0.84 and 0.89, respectively).74 However, the ApoSAA score remains to be evaluated in a multi-center setting. A gene-expression-based risk score (Sepsis MetaScore, comprised of 11 gene-expression levels) showed a high diagnostic yield (AUROC 0.92–0.93 in three separate centers) but remains to be prospectively validated.83 These studies illustrate that despite promising results, the route taking research results to the bedside is challenging.
Knowledge gaps and opportunities for future research
Future studies should define biomarker cut-off levels specific for (small) ranges of gestational and postnatal age as biomarker levels are affected by both.22 New biomarkers should be studied that rapidly change in a short time-period, because they are more likely to provide a high predictive value for diagnosing or excluding LONS. Because sepsis is a complex immunological process, it is likely that the predictive value of multiple pro- and anti-inflammatory biomarkers yields more information than a single biomarker, and should be assessed in age (both post-menstrual and postnatal) and weight specific groups of neonates. Large and well-defined patient populations are needed for discovery of novel -omics biomarkers as well as confirming prospective studies are needed for clinical implication.
In general, predictive models are not likely to consistently show absence of false-negative results, which has the highest clinical concern because of the potential consequences of delayed treatment of LONS. Therefore, a combination of biomarkers with a high positive predictive value could serve as alert to medical staff to consider (early) antibiotic treatment or additional testing.
From -omics research approaches to the bedside
Although hypothesis-free approaches are promising and will likely yield interesting results, the road to bedside implementation is long. Research efforts will have to be made to minimize the amount of blood needed for these -omics approaches as well as making them available 24/7 with a rapid turnover time. For both proteomics and metabolomics technical advances show promise in minimizing equipment size and blood sample volume for bedside implementation, when the appropriate proteomic and/or metabolomic targets have been identified.84,85,86,87 As -omics approaches currently are only used in research settings, biobanking of biological specimens and international collaboration will be critical for success.
Furthermore, future research should also focus on continuously measuring informative biomarkers using new techniques such as sensors and micro-dialysis devices,88 and measuring biomarkers in other, freely available body materials such as urine, saliva, or the microbiome of the skin.89 In order for these tests to have an impact on clinical decision-making and neonatal outcome such biomarkers sets should be made available as rapid onsite tests. Another option may be measuring volatile organic compounds (VOC). VOCs have been shown to change several before clinical LONS and NEC diagnosis and are specific to causative bacteria.90 An eNose in an incubator could continuously analyze exhaled air and can be an addition to early detection of LONS.91,92
Antibiotic therapy
Empirical antibiotic therapy
During the initial treatment phase, if sepsis is suspected, it is of utmost importance to assure efficacy of treatment.10,93 Usually, empirical antibiotic therapy is initiated covering the most common causative microorganisms.94 Upon determination of the causative microorganism and its antibiotic sensitivity, the antibiotic therapy can be narrowed with specific antibiotics or stopped if the blood culture remains negative and the clinical signs do not fit sepsis.10,94 Unfortunately, many blood cultures remain negative although sepsis cannot be ruled out clinically, inevitably leading to unnecessary use of antibiotics. This has been associated with an increased risk of NEC and death in patients receiving large amounts of antibiotics,95 and the emergence of multidrug-resistant Gram-negative bacteria.96
Knowledge gaps and opportunities for future research
Instantly available information on the causative microorganism would limit unnecessary use of antibiotics. This is critical for effective treatment and to prevent formation of more resistant strains. One way to go forward with this could be the use molecular diagnostics such as PCR or MALDI-TOF MS, a mass spectrometry technique on blood specimens after laser exposure, which have been shown to lead to quicker and more sensitive results, but with limited specificity.62,97 In the future the added value of these techniques in patients with negative cultures needs to be investigated to reduce the use of antibiotics in culture-negative sepsis. Also, biomarkers to detect resistant microorganisms and enabling adapted empiric antibiotic therapy are needed (Table 2).
Personalized antibiotic therapy
Ideally, we want to give the most specific, narrow, and optimally dosed antibiotic therapy, individualized for each patients with suspected LONS, and avoid empirical broad-spectrum antibiotic therapy. This would reduce unnecessary use of antibiotics and the accompanying risk of side effects and resistance. To accomplish personalized antibiotic treatment in LONS patients, rapidly available information on the causative microorganism is needed. Conventional blood culture methods are clearly too slow.98,99
Knowledge gaps and opportunities for future research
Future research should focus on the development of bedside molecular diagnostics in order to facilitate personalized antibiotic treatment. Information from classical serum biomarkers (e.g. C-reactive protein (CRP), interleukin-6 (IL-6)) provide limited information in adults, but this has not been investigated in neonatal populations.100,101,102,103 Analysis of the entire metabolome or proteome has great potential to find predictive biomarkers for the causative microorganism of LONS.104,105,106 However, it could be that only a subgroup of patients is suitable for personalized antibiotic treatments, based on their biomarker profiles predicting a causative microorganism with a high certainty. Possibly some patients do not need antibiotic therapy at all. If reliable biomarkers can be established, adapted antibiotic therapy needs to be investigated in a sufficiently powered RCT in order to assess safety and efficacy.107
PKs of antibiotics
Due to the lack of available PK data, it has long been a challenge to understand the factors driving the drug levels in septic neonates.108 Fortunately, this has drastically improved in the last decade due to more sensitive laboratory equipment that enabled quantification of drug levels in small plasma volumes and advances in PK modeling and simulation.109 Nowadays, the most commonly used antibiotics have been labeled for neonatal age and PK data are available.110 The most relevant antibiotics for the treatment of LONS are known to be predominantly eliminated renally.111 It may therefore be expected that determinants of antibiotic clearance are similar for most of these drugs. These determinants reflect prenatal (birth weight, gestational age) and postnatal maturation (postnatal age, current body weight).112,113 Nevertheless, different existing models for one antibiotic often result in slightly different dosage advice. The reasons for these difference relate to include the inclusion of different patient populations, small cohorts and different PK targets as these have not yet been clearly defined for neonates (Table 2).
Knowledge gaps and opportunities for future research
Future research should therefore focus on pooling all available data on one drug114 or performing model-based meta-analyses,115 which will result in more uniform and more reliable dosage advice. In addition, more studies are needed to identify the appropriate PK targets in neonates. In case of performing new PK trials, clinical trial simulations prior to the actual start of a study can help to choose the most informative blood sampling time scheme to ensure the study is sufficiently powered for PK analyses.116
Determining the right dose for antibiotics
Minimal inhibitory concentration
To determine the most adequate dose for the antibiotic therapy of LONS, the minimal inhibitory concentration (MIC) is often used. This concentration describes the lowest concentration at which bacterial growth is inhibited in vitro.117 However, the resulting MIC breakpoints that define whether a pathogen is susceptible are based on PK, pharmacodynamics (PD), and outcome data in adults and not in children, let alone neonates. Differences in neonatal immune response leading to altered bacterial killing patterns might lead to a need for a higher MIC to effectively kill the bacteria. This becomes even more relevant if the microorganism has a high MIC.118 Furthermore, the elimination and distribution of drugs in neonates is different from older children and adults.119 Even when reaching similar plasma concentrations in the blood by means of individualized dosing, the concentration at the site of infection might vary. This concentration, will, in turn impact the immune response, the pathogen load, and possible biomarkers for diagnosing LONS and/or monitoring treatment effect (see Fig. 2).
Pink: causative pathogen, blue: immune system, orange: organ function, green: drug, adapted from https://ascpt.onlinelibrary.wiley.com/doi/epdf/10.1002/cpt.2194.
PK/PD relationships
For different classes of antibiotics, different MIC-related PK/PD indices, relating PK parameters to efficacy, have been defined depending on the mechanism of action. These include concentration-dependent killing and time-dependent killing and a combination of the two.93 For example, vancomycin shows time- and concentration-dependent killing that is best described by the area under the curve (AUC)/MIC. In clinical practice, AUC/MIC is assessed by using the trough concentration as a surrogate for AUC.120 However, due to their significantly lower clearance in neonates and other differences in PK, a sufficient AUC/MIC ratio can be achieved aiming for lower trough concentrations.121 Moreover, the choice of dosing interval and daily dose leading to the target AUC has a great impact on the corresponding trough concentration.122 This underlines that the adult PK target cannot simply be extrapolated to neonates and that more research into suitable PK targets for neonates is needed.
Knowledge gaps and opportunities for future research
In absence of targets determined for neonates, model-informed precision dosing (MIPD) bears the potential to initiate antibiotic treatment with the optimal individual dosage. MIPD captures drug, disease, and patient characteristics in modeling approaches and can be used to perform Bayesian forecasting and dose optimization (Table 2). Despite all reported PK models, this has hardly been implemented yet.123 Depending on the antibiotic drug, this may imply to start with a loading dose for an immediate adequate exposure and quick onset of effect. Subsequently, therapeutic drug monitoring (TDM) can be used to further optimize individual exposure which should aim for the most evidence based target so far: AUC. Due to the observed large inter-individual difference in PK in preterm neonates, TDM is of additional value with respect to preventing toxicity as well as assuring treatment efficacy.124 One good example for this could be beta-lactams as recent studies have shown that in neonates with sepsis therapy failure, as little as 20% of the benzyl penicillin concentrations were within the target range.125
Monitoring sepsis course
Assessing treatment effect
LONS may compromise the cardiovascular system, leading to reduced cardiac output and/or disturbed microcirculation, but also other organ systems. Adequate biomarkers to judge effectiveness of LONS management would significantly improve clinical care; knowing you are on the right track even before the patient shows clinical recovery, potentially prevents overuse of supportive therapy and unnecessary switches to broad-spectrum antibiotics. An ideal biomarker to assess treatment effect should have a high predictive value and a fast response to changes in the disease state and easily available. Preferably, the biomarker changes before the clinical condition of the patient improves or deteriorates and is measured non-invasively.105,126 Conventional methods for assessing the hemodynamic status during sepsis, such as blood pressure, urine output, and heart rate, seem inaccurate and unreliable.127,128 More objective physiological biomarkers of the (micro) circulation could be used during disease monitoring. Repeated assessment of cardiac output measurements, obtained by (neonatologist performed) echocardiography, may also provide objective physiological biomarkers that could be used during for disease status monitoring.129,130 However, these measurements require intensive training and there is a lack knowledge regarding the targets to aim for. Furthermore, assessment is limited to central blood flow, while end-organ (micro)circulation is not measured.129,130 To this end, near-infrared spectroscopy or transcutaneous assessment of capillary flow might be used to monitor regional perfusion, but again targets are lacking and are likely to be individual.131,132,133,134
Knowledge gaps and opportunities for future research
Future research should develop methods that enable us to measure end-organ blood flow or microcirculation to improve hemodynamic assessment during sepsis, which could facilitate tailored management. Objective measuring tools are needed that generate easy interpretable results with targets available that need to be pursued. Once established, prospective studies should be performed to evaluate whether these tools improve neonatal outcome, when implemented and targeted for.
Hyperinflammation
The preterm neonates immune system is immature and relies on a less well developed innate immune system, making preterm neonates more vulnerable to infections.135,136,137,138,139 Studies in neonatal cord blood show that the neonatal immune system responds differently to bacterial antigens at birth, compared to adults, which results in different cytokine responses, e.g. lower levels of tumor necrosis factor-α (TNF-α) and higher levels of IL-6.137,140 These differences can result in a dysregulated pro-inflammatory response to infection, sometimes also called “cytokine-storm”.141 Such an inflammatory response can cause damage to different organ systems126 which can, in turn, affect the PKs of drugs (see Fig. 2). Depending on the causative pathogen, neonatal immune response will vary. Cell models comparing the response to different aerobe and anaerobe Gram-positive and Gram-negative bacteria in umbilical cord cells and adults cells indicate specific differences in the immune response for different pathogens.142
Knowledge gaps and opportunities for future research
Means to identify the different stages of inflammation in preterm neonatal sepsis are of paramount importance to increase clinical understanding and potentially guide therapy. It is likely that the level of biomarkers reflecting the immune response are affected by complex host–pathogen–drug interactions.143 Many factors can determine the level of such a biomarker: (1) the developing immune system of the neonate, (2) the inflammatory state, (3) the causative pathogen, and (4) administered drugs (see Fig. 2). Research on immune response biomarkers increases the understanding of the neonatal immune response and its dynamics, and should aim to define (immune) biomarkers that can be clinically used for treatment response and for PK/PD of drug treatment. The neonatal immune response could relatively easily be studied in in vitro or high-throughput animal models of sepsis.144 Up until now animal models of sepsis rarely yielded new biomarkers or therapeutic agents that have been successfully used in clinical practice.145 A promising new candidate model is zebrafish larvae. The immune system of the zebrafish closely resembles the human immune system.146 In zebrafish models, larvae can be infected under controlled conditions and the immune response can be observed for different pathogens and maturational statuses. Zebrafish already have been successfully exposed to pathogens relevant for neonatal sepsis, which has already resulted in the identification of factors necessary for the establishment and spread of the infection.144 Furthermore, the emergence of -omics approaches enables researchers to map and understand physiological changes during sepsis. Rapid onsite omics approaches could be used to monitor the inflammatory response during the course of sepsis, which would enable individualized management according to the patient’s functional immune status.147
Prediction of severity of sepsis
During the sepsis course, some patients will develop severe sepsis symptoms with need for inotropes or mechanical ventilation and risk of mortality, while others may not. Identification of patients with oncoming severe sepsis could enable personalized monitoring and possible additional interventions to prevent subsequent clinical deterioration. A recent study showed that commonly used chemical biomarkers, such as IL-6, PCT, and CRP, may provide information on who will have a severe LONS.148 Cut-off values for the biomarkers were calculated to be used in clinical practice as a warning signal for subsequent clinical deterioration. As one may recognize, the severity of illness is also related to the underlying causative microorganism (Gram-negative bacteria); thus, biomarkers establishing the causative organism rapidly after LONS diagnosis are likely to show overlap with predicting severe LONS. In adult sepsis it has been shown that PCT levels differ according to specific causative organism101 and therefore could potentially be used to predict the microorganism even before the result of the blood culture. However, the discriminatory power of PCT as a sole biomarker is too low to guide therapeutic decisions.101
Knowledge gaps and opportunities for future research
Future research should focus on identifying patients with severe sepsis and a developing hyper-inflammatory syndrome in order to facilitate personalized monitoring (e.g. inserting peripheral arterial line) and additional therapy (e.g. immune-modulating treatment) in the early stages of sepsis, thereby going from reactive management towards forward-looking management. Early prediction of the causative organism by (set of) biomarkers might be one way to facilitate this, where the focus should lie on identifying a set of biomarkers rather than a single biomarker. Another potential biomarker that could provide information on sepsis severity could be a biomarker reflecting bacterial load. Prospective studies are needed to investigate which biomarkers could be of value in clinical practice.
Additional therapy
Next to antibiotic therapy, several studies in neonates have focused on additional therapies consisting of a variety of immune-modulating intervention, up until now evidence remains sparse (knowledge gaps and opportunities are summarized in Table 1).149 Evidence for the most interventions, such as IVIG, lactoferrin, and granulocyte–macrophage colony-stimulating factor, remains inconclusive.10,33,94,150,151,152,153 When considering additional therapies, it is important to adapt an individual approach, since the immune response varies over time and the timing of immunomodulatory therapy is critical.154
Pentoxifylline (PTX)
PTX is a phosphodiesterase inhibitor, originally registered for intermittent claudication in adults,155 which suppresses the production of TNFα and other inflammatory cytokines and prevents their subsequent effects.156,157,158,159 Furthermore, PTX has beneficial effects on endothelial cell function, coagulation, and microcirculation in preterm neonates with LONS.160,161,162,163 A recent meta-analysis of six studies showed that PTX is associated with a decreased mortality during hospitalization,164 which makes PTX a promising candidate for the adjunctive treatment of LONS. A larger study to confirm these effects and study long-term outcome is currently ongoing (ACTRN12616000405415). Unfortunately, available data on PKs of PTX and its active metabolites in neonates are limited. Most clinical studies of PTX in neonates used the same dosage of 5 mg/kg/h during 6 h.157,158 PTX dosages of 5 mg/kg/h during 12 h led to increased plasma levels and were well tolerated in seven preterm neonates.158,165 However, the most effective dosage and target concentration as well as timing in preterm neonates still need to be determined (dose per kilogram per day, duration of infusion and treatment). In vivo animal models might help define optimal timing and dosing, which subsequently could be validated in a clinical trial.
Mesenchymal stromal cells
Mesenchymal stromal cells (MSCs) are non-hematopoietic multi potent stromal cell that can be isolated from various tissues, such as umbilical cord tissue. MSCs may reduce multi-organ dysfunction and improve survival in sepsis by several mechanisms, which include reducing the inflammatory response, vascular injury, and bacterial load.166,167,168,169 A phase-I trial in nine adults with sepsis showed that infusion of MSCs seemed safe.170 Although MSC-treatment seems promising in neonatal animal models,171 information on safety, correct dosage, underlying working mechanism, and possible patient selection is missing in neonatal sepsis. Clinical trials are needed to assess safety, feasibility, effectiveness, and to determine the effective dosage.
Timing of additional therapy
The host immune response in sepsis varies in the same patient over time172 and gestational age and postnatal age, but also sex, affect the immune responses.173,174,175 Therefore, mapping the immune status for each individual patient during the course of sepsis would enable to select patients that might benefit from immunomodulatory treatment at a certain time point during the disease process (see “Hyperinflammation”).176
Knowledge gaps and opportunities for future research
Depending on the immune state of the patient, immune-modulating drugs may benefit the patient, while other patients may suffer from adverse effects if the drug is administered at an inappropriate time point because of its potentially harmful interference with the developing immune system.176 For example, giving pentoxifylline in a late stage of severe sepsis has adverse consequences in a swine model of sepsis.177 However, knowledge regarding sepsis phases in preterm human neonates is still largely lacking.22 Future research should aim to understand immune response mechanisms and study how and when the immune response should be modulated to improve LONS outcome. As this is complicated to study in preterm neonates, in vitro experiments and/or animal models146 should be explored first before clinical application is feasible. As a final step, an individualized management based on immune status of the status should be prospectively compared with a group-based approach ideally in a randomized controlled trial.
Personalized medicine in clinical practice
Ideally all information from a patient comes together in real-time bedside “dashboard” of the patients clinical status. This dashboard can incorporate physiological data, patient characteristics (e.g. GA, PNA, sex), clinical data (e.g. previous diagnoses and therapies), biomarkers, and metabolomic and proteomic profiles. Using supervised ML models, all these characteristics could be able to predict a causative organism, stratify the patient according to risk, provide information on immune status, and detect if the patient is responding to the initiated therapy.178 An optimal (additional) treatment and monitoring strategy tailored to the individual patient could be recommended. By continuously extracting patient (physiological) data and additionally determined biomarkers, the data driven “dashboard” model could update treatment and monitoring strategies and provide additional suggestions for LONS management if needed.69,179 Development of predictive modeling dashboards will require harmonized and large-scale databases, fed by multicentre collaborations, if possible including proteomics and/or metabolomics data.180
A roadmap for future research
In conclusion, the research field of LONS, is entering an exciting new era. The availability of computational power, computational medicine algorithms, and hypothesis-free -omics approaches give way to building bedside feedback tools to guide clinicians in personalized management of LONS. However, important knowledge gaps still need to be addressed. First, we need to identify biomarkers, which ideally can be measured continuously and non-invasively or in low fluid volumes, to identify preterm neonates with LONS as early as possible and provide information on the severity of LONS. This will enable personalized therapy and monitoring of treatment effect. Second, means of assessing the immune state of a preterm neonate are likely to harness important information to provide optimal immune-modulating therapy at the right time. Third, we need to improve our understanding and utilization of neonatal pharmacology (pharmacokinetics and pharmacodynamics), which may provide means to immediately optimize treatment and supportive strategies. Last, we need to develop clear and informative bedside dashboards, incorporating biomarker information, ML, and AI and monitoring to help clinicians optimize every step in the management of LONS in preterm neonates and improve outcome.
References
Stoll, B. J. & Hansen, N. Infections in VLBW infants: studies from the NICHD Neonatal Research Network. Semin. Perinatol. 27, 293–301 (2003).
Stoll, B. J. et al. Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network. Pediatrics 110, 285–291 (2002).
Lawn, J. E., Wilczynska-Ketende, K. & Cousens, S. N. Estimating the causes of 4 million neonatal deaths in the year 2000. Int. J. Epidemiol. 35, 706–718 (2006).
Shane, A. L., Sanchez, P. J. & Stoll, B. J. Neonatal sepsis. Lancet 390, 1770–1780 (2017).
Stoll, B. J. et al. Neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection. JAMA 292, 2357–2365 (2004).
Schlapbach, L. J. et al. Impact of sepsis on neurodevelopmental outcome in a Swiss National Cohort of extremely premature infants. Pediatrics 128, e348–e357 (2011).
Fleiss, N. et al. Evaluation of the neonatal sequential organ failure assessment and mortality risk in preterm infants with late-onset infection. JAMA Netw. Open 4, e2036518 (2021).
Molloy, E. J. et al. Neonatal sepsis: need for consensus definition, collaboration and core outcomes. Pediatr. Res. 88, 2–4 (2020).
Wynn, J. L. et al. Time for a neonatal-specific consensus definition for sepsis. Pediatr. Crit. Care Med. 15, 523–528 (2014).
Dong, Y. & Speer, C. P. Late-onset neonatal sepsis: recent developments. Arch. Dis. Child Fetal Neonatal Ed. 100, F257–F263 (2015).
Morgan, R. L. et al. Probiotics reduce mortality and morbidity in preterm, low-birth-weight infants: a systematic review and network meta-analysis of randomized trials. Gastroenterology 159, 467–480 (2020).
Rao, S. C. et al. Probiotic supplementation and late-onset sepsis in preterm infants: a meta-analysis. Pediatrics 137, e20153684 (2016).
Chi, C. et al. Effects of probiotics in preterm infants: a network meta-analysis. Pediatrics 147, e20200706 (2021).
Schanler, R. J., Shulman, R. J. & Lau, C. Feeding strategies for premature infants: beneficial outcomes of feeding fortified human milk versus preterm formula. Pediatrics 103, 1150–1157 (1999).
de Silva, A., Jones, P. W. & Spencer, S. A. Does human milk reduce infection rates in preterm infants? A systematic review. Arch. Dis. Child Fetal Neonatal Ed. 89, F509–F513 (2004).
Cacho, N. T. & Lawrence, R. M. Innate immunity and breast milk. Front. Immunol. 8, 584 (2017).
Underwood, M. A. Human milk for the premature infant. Pediatr. Clin. North Am. 60, 189–207 (2013).
Cacho, N. T., Parker, L. A. & Neu, J. Necrotizing enterocolitis and human milk feeding: a systematic review. Clin. Perinatol. 44, 49–67 (2017).
Meier, P., Patel, A. & Esquerra-Zwiers, A. Donor human milk update: evidence, mechanisms, and priorities for research and practice. J. Pediatr. 180, 15–21 (2017).
Peila, C. et al. The effect of holder pasteurization on nutrients and biologically-active components in donor human milk: a review. Nutrients 8, 477 (2016).
Daniels, B. et al. The effect of simulated flash-heat pasteurization on immune components of human milk. Nutrients 9, 178 (2017).
Schuller, S. S. et al. Immunomodulation to prevent or treat neonatal sepsis: past, present, and future. Front. Pediatr. 6, 199 (2018).
Quigley, M., Embleton, N. D. & McGuire, W. Formula versus donor breast milk for feeding preterm or low birth weight infants. Cochrane Database Syst. Rev. 7, CD002971 (2019).
Maffei, D. et al. Early oral colostrum administration in preterm infants. J. Perinatol. 40, 284–287 (2020).
Rodriguez, N. A., Meier, P. P., Groer, M. W. & Zeller, J. M. Oropharyngeal administration of colostrum to extremely low birth weight infants: theoretical perspectives. J. Perinatol. 29, 1–7 (2009).
da Cruz Martins, C. et al. Colostrum oropharyngeal immunotherapy for very low birth weight preterm infants: protocol of an intervention study. BMC Pediatr. 20, 371 (2020).
Maertens, K. et al. Breastfeeding after maternal immunisation during pregnancy: providing immunological protection to the newborn: a review. Vaccine 32, 1786–1792 (2014).
Schlaudecker, E. P. et al. IgA and neutralizing antibodies to influenza a virus in human milk: a randomized trial of antenatal influenza immunization. PLoS ONE 8, e70867 (2013).
Cinicola, B. et al. The protective role of maternal immunization in early life. Front. Pediatr. 9, 638871 (2021).
Broad, J. et al. The current state of immunization against Gram-negative bacteria in children: a review of the literature. Curr. Opin. Infect. Dis. 33, 517–529 (2020).
Jakaitis, B. M. & Denning, P. W. Human breast milk and the gastrointestinal innate immune system. Clin. Perinatol. 41, 423–435 (2014).
Manzoni, P. et al. Bovine lactoferrin supplementation for prevention of late-onset sepsis in very low-birth-weight neonates: a randomized trial. JAMA 302, 1421–1428 (2009).
Pammi, M. & Suresh, G. Enteral lactoferrin supplementation for prevention of sepsis and necrotizing enterocolitis in preterm infants. Cochrane Database Syst. Rev. 3, CD007137 (2020).
Patel, M. & Kaufman, D. A. Anti-lipoteichoic acid monoclonal antibody (pagibaximab) studies for the prevention of staphylococcal bloodstream infections in preterm infants. Expert Opin. Biol. Ther. 15, 595–600 (2015).
Weisman, L. E. et al. A randomized study of a monoclonal antibody (pagibaximab) to prevent staphylococcal sepsis. Pediatrics 128, 271–279 (2011).
Theilacker, C. et al. Protection against Staphylococcus aureus by antibody to the polyglycerolphosphate backbone of heterologous lipoteichoic acid. J. Infect. Dis. 205, 1076–1085 (2012).
Shah, P. S. & Kaufman, D. A. Antistaphylococcal immunoglobulins to prevent staphylococcal infection in very low birth weight infants. Cochrane Database Syst. Rev. CD006449 (2009).
Thammavongsa, V., Kim, H. K., Missiakas, D. & Schneewind, O. Staphylococcal manipulation of host immune responses. Nat. Rev. Microbiol 13, 529–543 (2015).
Malik, P. R. V. & Edginton, A. N. Physiologically-based pharmacokinetic modeling vs. allometric scaling for the prediction of infliximab pharmacokinetics in pediatric patients. CPT Pharmacomet. Syst. Pharm. 8, 835–844 (2019).
Kumar, A. et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit. Care Med. 34, 1589–1596 (2006).
Ng, P. C. Diagnostic markers of infection in neonates. Arch. Dis. Child Fetal Neonatal Ed. 89, F229–F235 (2004).
Ng, P. C. et al. Diagnosis of late onset neonatal sepsis with cytokines, adhesion molecule, and C-reactive protein in preterm very low birthweight infants. Arch. Dis. Child Fetal Neonatal Ed. 77, F221–F227 (1997).
Chiera, M. et al. Heart rate variability in the perinatal period: a critical and conceptual review. Front. Neurosci. 14, 561186 (2020).
Leon, C., Carrault, G., Pladys, P. & Beuchee, A. Early detection of late onset sepsis in premature infants using visibility graph analysis of heart rate variability. IEEE J. Biomed. Health Inf. 25, 1006–1017 (2021).
Moorman, J. R. et al. Mortality reduction by heart rate characteristic monitoring in very low birth weight neonates: a randomized trial. J. Pediatr. 159, 900–906 e901 (2011).
Fairchild, K. D. et al. Septicemia mortality reduction in neonates in a heart rate characteristics monitoring trial. Pediatr. Res. 74, 570–575 (2013).
Coggins, S. A. et al. Heart rate characteristic index monitoring for bloodstream infection in an NICU: a 3-year experience. Arch. Dis. Child Fetal Neonatal Ed. 101, F329–F332 (2016).
Fairchild, K. D. et al. Vital signs and their cross-correlation in sepsis and NEC: a study of 1,065 very-low-birth-weight infants in two NICUs. Pediatr. Res. 81, 315–321 (2017).
Mithal, L. B. et al. Vital signs analysis algorithm detects inflammatory response in premature infants with late onset sepsis and necrotizing enterocolitis. Early Hum. Dev. 117, 83–89 (2018).
Sullivan, B. A. et al. Clinical and vital sign changes associated with late-onset sepsis in very low birth weight infants at 3 NICUs. J. Neonatal Perinatal Med. https://doi.org/10.3233/NPM-200578 (2021).
Joshi, R. et al. Predicting neonatal sepsis using features of heart rate variability, respiratory characteristics, and ECG-derived estimates of infant motion. IEEE J. Biomed. Health Inf. 24, 681–692 (2020).
Kumar, N. et al. Continuous vital sign analysis for predicting and preventing neonatal diseases in the twenty-first century: big data to the forefront. Pediatr. Res. 87, 210–220 (2020).
Beam, A. L. & Kohane, I. S. Big data and machine learning in health care. JAMA 319, 1317–1318 (2018).
Rajkomar, A., Dean, J. & Kohane, I. Machine learning in medicine. N. Engl. J. Med. 380, 1347–1358 (2019).
Masino, A. J. et al. Machine learning models for early sepsis recognition in the neonatal intensive care unit using readily available electronic health record data. PLoS ONE 14, e0212665 (2019).
Helguera-Repetto, A. C. et al. Neonatal sepsis diagnosis decision-making based on artificial neural networks. Front. Pediatr. 8, 525 (2020).
van de Sande, D. et al. Moving from bytes to bedside: a systematic review on the use of artificial intelligence in the intensive care unit. Intensive Care Med. 47, 750–760 (2021).
Van Laere, D. et al. Machine learning to support hemodynamic intervention in the neonatal intensive care unit. Clin. Perinatol. 47, 435–448 (2020).
Bellini, V., Montomoli, J. & Bignami, E. Poor quality data, privacy, lack of certifications: the lethal triad of new technologies in intensive care. Intensive Care Med, https://doi.org/10.1007/s00134-021-06473-4 (2021).
Delanghe, J. R. & Speeckaert, M. M. Translational research and biomarkers in neonatal sepsis. Clin. Chim. Acta 451, 46–64 (2015).
Sharma, D., Farahbakhsh, N., Shastri, S. & Sharma, P. Biomarkers for diagnosis of neonatal sepsis: a literature review. J. Matern Fetal Neonatal Med. 31, 1646–1659 (2018).
Iroh Tam, P. Y. & Bendel, C. M. Diagnostics for neonatal sepsis: current approaches and future directions. Pediatr. Res. 82, 574–583 (2017).
Kuster, H. et al. Interleukin-1 receptor antagonist and interleukin-6 for early diagnosis of neonatal sepsis 2 days before clinical manifestation. Lancet 352, 1271–1277 (1998).
Arnon, S. et al. Serum amyloid A: an early and accurate marker of neonatal early-onset sepsis. J. Perinatol. 27, 297–302 (2007).
Ho, J. et al. Pathological role and diagnostic value of endogenous host defense peptides in adult and neonatal sepsis: a systematic review. Shock 47, 673–679 (2017).
Bellos, I. et al. The diagnostic accuracy of presepsin in neonatal sepsis: a meta-analysis. Eur. J. Pediatr. 177, 625–632 (2018).
Bhandari, V. Effective biomarkers for diagnosis of neonatal sepsis. J. Pediatr. Infect. Dis. Soc. 3, 234–245 (2014).
Ludwig, K. R. & Hummon, A. B. Mass spectrometry for the discovery of biomarkers of sepsis. Mol. Biosyst. 13, 648–664 (2017).
Ng, S. et al. Precision medicine for neonatal sepsis. Front. Mol. Biosci. 5, 70 (2018).
Cao, Z. & Robinson, R. A. The role of proteomics in understanding biological mechanisms of sepsis. Proteom. Clin. Appl. 8, 35–52 (2014).
Schmerler, D. et al. Targeted metabolomics for discrimination of systemic inflammatory disorders in critically ill patients. J. Lipid Res. 53, 1369–1375 (2012).
Stewart, C. J. et al. Metabolomic and proteomic analysis of serum from preterm infants with necrotising entercolitis and late-onset sepsis. Pediatr. Res. 79, 425 (2015).
Kim, D. et al. Pilot application of magnetic nanoparticle-based biosensor for necrotizing enterocolitis. J Proteomics Bioinformatics, Suppl 5:002 (2015).
Ng, P. C. et al. Host-response biomarkers for diagnosis of late-onset septicemia and necrotizing enterocolitis in preterm infants. J. Clin. Invest. 120, 2989–3000 (2010).
Buhimschi, C. S. et al. Proteomics mapping of cord blood identifies haptoglobin “switch-on” pattern as biomarker of early-onset neonatal sepsis in preterm newborns. PLoS ONE 6, e26111 (2011).
Mardegan, V. et al. Untargeted and targeted metabolomic profiling of preterm newborns with early onset sepsis: a case-control study. Metabolites 11, 115 (2021).
Dessi, A. et al. Monitoring neonatal fungal infection with metabolomics. J. Matern. Fetal Neonatal Med. 27(Suppl. 2), 34–38 (2014).
Sarafidis, K. et al. Urine metabolomics in neonates with late-onset sepsis in a case-control study. Sci. Rep. 7, 1–9 (2017).
Fanos, V. et al. Urinary 1H-NMR and GC-MS metabolomics predicts early and late onset neonatal sepsis. Early Hum. Dev. 90, S78–S83 (2014).
Mickiewicz, B., Vogel, H. J., Wong, H. R. & Winston, B. W. Metabolomics as a novel approach for early diagnosis of pediatric septic shock and its mortality. Am. J. Respir. Crit. Care Med. 187, 967–976 (2013).
Ou, H. et al. Expression of microRNA-23b in patients with sepsis and its effect on leukocytes and the expression of E-selectin and ICAM-1. Exp. Ther. Med. 16, 4707–4711 (2018).
Fatmi, A. et al. miRNA-23b as a biomarker of culture-positive neonatal sepsis. Mol. Med. 26, 94 (2020).
Sweeney, T. E. et al. Validation of the sepsis metascore for diagnosis of neonatal sepsis. J. Pediatr. Infect. Dis. Soc. 7, 129–135 (2018).
Tebani, A., Afonso, C., Marret, S. & Bekri, S. Omics-based strategies in precision medicine: toward a paradigm shift in inborn errors of metabolism investigations. Int. J. Mol. Sci. 17, 1555 (2016).
Nagana Gowda, G. A. & Raftery, D. Biomarker discovery and translation in metabolomics. Curr. Metabolomics 1, 227–240 (2013).
Buchegger, P., Sauer, U., Toth-Szekely, H. & Preininger, C. Miniaturized protein microarray with internal calibration as point-of-care device for diagnosis of neonatal sepsis. Sensors (Basel) 12, 1494–1508 (2012).
Tighe, P. J., Ryder, R. R., Todd, I. & Fairclough, L. C. ELISA in the multiplex era: potentials and pitfalls. Proteomics Clin. Appl. 9, 406–422 (2015).
Waelgaard, L., Pharo, A., Tonnessen, T. I. & Mollnes, T. E. Microdialysis for monitoring inflammation: efficient recovery of cytokines and anaphylotoxins provided optimal catheter pore size and fluid velocity conditions. Scand. J. Immunol. 64, 345–352 (2006).
Pammi, M. et al. Microbiome signatures in neonatal central line associated bloodstream infections. PLoS ONE 15, e0227967 (2020).
Berkhout, D. J. C. et al. Late-onset sepsis in preterm infants can be detected preclinically by fecal volatile organic compound analysis: a prospective, multicenter cohort study. Clin. Infect. Dis. 68, 70–77 (2019).
Gofin, Y. et al. Salivary C-reactive protein—a possible predictor of serum levels in pediatric acute respiratory illness. Eur. J. Pediatr, 180, 2465–2472 (2021).
El Manouni El Hassani, S. et al. Profound pathogen-specific alterations in intestinal microbiota composition precede late onset sepsis in preterm infants: a longitudinal multicenter case-control study. Clin. Infect. Dis. 73, e224–e232(2021).
van Donge, T., Bielicki, J. A., van den Anker, J. & Pfister, M. Key components for antibiotic dose optimization of sepsis in neonates and infants. Front Pediatr. 6, 325 (2018).
Stockmann, C. et al. Considerations in the pharmacologic treatment and prevention of neonatal sepsis. Paediatr. Drugs 16, 67–81 (2014).
Kuppala, V. S., Meinzen-Derr, J., Morrow, A. L. & Schibler, K. R. Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants. J. Pediatr. 159, 720–725 (2011).
Tsai, M. H. et al. Risk factors and outcomes for multidrug-resistant Gram-negative bacteremia in the NICU. Pediatrics 133, e322–e329 (2014).
Verroken, A. et al. Clinical impact of MALDI-TOF MS identification and rapid susceptibility testing on adequate antimicrobial treatment in sepsis with positive blood cultures. PLoS ONE 11, e0156299 (2016).
Connell, T. G. et al. How reliable is a negative blood culture result? Volume of blood submitted for culture in routine practice in a children’s hospital. Pediatrics 119, 891–896 (2007).
Jardine, L., Davies, M. W. & Faoagali, J. Incubation time required for neonatal blood cultures to become positive. J. Paediatr. Child Health 42, 797–802 (2006).
Brodská, H. et al. Significantly higher procalcitonin levels could differentiate Gram-negative sepsis from Gram-positive and fungal sepsis. Clin. Exp. Med. 13, 165–170 (2013).
Thomas-Rüddel, D. O. et al. Influence of pathogen and focus of infection on procalcitonin values in sepsis patients with bacteremia or candidemia. Crit. Care 22, 128 (2018).
Bassetti, M. et al. Role of procalcitonin in bacteremic patients and its potential use in predicting infection etiology. Expert Rev. Anti Infective Ther. 17, 99–105 (2019).
Nakajima, A. et al. Clinical utility of procalcitonin as a marker of sepsis: a potential predictor of causative pathogens. Intern. Med. 53, 1497–1503 (2014).
Clish, C. B. Metabolomics: an emerging but powerful tool for precision medicine. Mol. Case Stud. 1, a000588 (2015).
Dessì, A. et al. New diagnostic possibilities in systemic neonatal infections: metabolomics. Early Hum. Dev. 90, S19–S21 (2014).
Eckerle, M. et al. Metabolomics as a driver in advancing precision medicine in sepsis. Pharmacotherapy 37, 1023–1032 (2017).
McShane, L. M. et al. Criteria for the use of omics-based predictors in clinical trials. Nature 502, 317–320 (2013).
Laughon, M. M. et al. Innovative clinical trial design for pediatric therapeutics. Expert Rev. Clin. Pharmacol. 4, 643–652 (2011).
Rivera-Chaparro, N. D., Cohen-Wolkowiez, M. & Greenberg, R. G. Dosing antibiotics in neonates: review of the pharmacokinetic data. Future Microbiol. 12, 1001–1016 (2017).
Flint, R. B. et al. Large differences in neonatal drug use between NICUs are common practice: time for consensus? Br. J. Clin. Pharmacol. 84, 1313–1323 (2018).
Blot, S. I., Pea, F. & Lipman, J. The effect of pathophysiology on pharmacokinetics in the critically ill patient—concepts appraised by the example of antimicrobial agents. Adv. Drug Deliv. Rev. 77, 3–11 (2014).
Tremoulet, A. et al. Characterization of the population pharmacokinetics of ampicillin in neonates using an opportunistic study design. Antimicrob. Agents Chemother. 58, 3013–3020 (2014).
Allegaert, K., Mian, P. & van den Anker, J. N. Developmental pharmacokinetics in neonates: maturational changes and beyond. Curr. Pharmacol. Des. 23, 5769–5778 (2017).
Voller, S. et al. Model-based clinical dose optimization for phenobarbital in neonates: an illustration of the importance of data sharing and external validation. Eur. J. Pharmacol. Sci. 109S, S90–S97 (2017).
Mould, D. R. Model-based meta-analysis: an important tool for making quantitative decisions during drug development. Clin. Pharmacol. Ther. 92, 283–286 (2012).
Holford, N. H., Kimko, H. C., Monteleone, J. P. & Peck, C. C. Simulation of clinical trials. Annu. Rev. Pharmacol. Toxicol. 40, 209–234 (2000).
Mueller, M., de la Pena, A. & Derendorf, H. Issues in pharmacokinetics and pharmacodynamics of anti-infective agents: kill curves versus MIC. Antimicrob. Agents Chemother. 48, 369–377 (2004).
Nielsen, L. E. et al. One size fits all? Application of susceptible-dose-dependent breakpoints to pediatric patients and laboratory reporting. J. Clin. Microbiol. 58, e01446–19 (2019).
Kearns, G. L. et al. Developmental pharmacology–drug disposition, action, and therapy in infants and children. N. Engl. J. Med. 349, 1157–1167 (2003).
Liu, C. et al. Clinical practice guidelines by the infectious diseases society of america for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children: executive summary. Clin. Infect. Dis. 52, 285–292 (2011).
Stockmann, C. et al. Predictive performance of a vancomycin population pharmacokinetic model in neonates. Infect. Dis. Ther. 4, 187–198 (2015).
Allegaert, K., Flint, R. & Smits, A. Pharmacokinetic modelling and Bayesian estimation-assisted decision tools to optimize vancomycin dosage in neonates: only one piece of the puzzle. Expert Opin. Drug Metab. Toxicol. 15, 735–749 (2019).
Abdulla, A. et al. Model-informed precision dosing of antibiotics in pediatric patients: a narrative review. Front Pediatr. 9, 624639 (2021).
Mian, P. et al. Therapeutic drug monitoring in neonates: what makes them unique? Curr. Pharmacol. Des. 23, 5790–5800 (2017).
Jansen, S. J. et al. Benzylpenicillin serum concentrations in neonates with group B Streptococci sepsis or meningitis: a descriptive cohort study. Pediatr. Infect. Dis. J. 40, 434–439 (2021).
Ng, P. C., Ma, T. P. & Lam, H. S. The use of laboratory biomarkers for surveillance, diagnosis and prediction of clinical outcomes in neonatal sepsis and necrotising enterocolitis. Arch. Dis. Child Fetal Neonatal Ed. 100, F448–F452 (2015).
de Boode, W.-P. Clinical monitoring of systemic hemodynamics in critically ill newborns. Early Hum. Dev. 86, 137–141 (2010).
Tibby, S. M., Hatherill, M., Marsh, M. J. & Murdoch, I. A. Clinicians’ abilities to estimate cardiac index in ventilated children and infants. Arch. Dis. Child. 77, 516–518 (1997).
de Boode, W. P. et al. The role of neonatologist performed echocardiography in the assessment and management of neonatal shock. Pediatr. Res. 84, 57–67 (2018).
De Boode, W. P. et al. Recommendations for neonatologist performed echocardiography in Europe: consensus statement endorsed by European Society for Paediatric Research (ESPR) and European Society for Neonatology (ESN). Pediatr. Res. 80, 465–471 (2016).
da Costa, C. S., Greisen, G. & Austin, T. Is near-infrared spectroscopy clinically useful in the preterm infant? Arch. Dis. Child.-Fetal Neonatal Ed. 100, F558–F561 (2015).
Liem, K. D. & Greisen, G. Monitoring of cerebral haemodynamics in newborn infants. Early Hum. Dev. 86, 155–158 (2010).
Sood, B. G., McLaughlin, K. & Cortez, J. Near-infrared spectroscopy: applications in neonates. Semin. Fetal Neonatal Med. 164–172 (2015).
van der Laan, M. E. et al. Multisite tissue oxygenation monitoring indicates organ-specific flow distribution and oxygen delivery related to low cardiac output in preterm infants with clinical sepsis. Pediatr. Crit. Care Med. 17, 764–771 (2016).
Wynn, J. L., Neu, J., Moldawer, L. L. & Levy, O. Potential of immunomodulatory agents for prevention and treatment of neonatal sepsis. J. Perinatol. 29, 79–88 (2009).
Dowling, D. J. & Levy, O. Ontogeny of early life immunity. Trends Immunol. 35, 299–310 (2014).
Olin, A. et al. Stereotypic immune system development in newborn children. Cell 174, 1277–1292 e1214 (2018).
Melville, J. M. & Moss, T. J. The immune consequences of preterm birth. Front. Neurosci. 7, 79 (2013).
Kollmann, T. R., Levy, O., Montgomery, R. R. & Goriely, S. Innate immune function by Toll-like receptors: distinct responses in newborns and the elderly. Immunity 37, 771–783 (2012).
Basha, S., Surendran, N. & Pichichero, M. Immune responses in neonates. Expert Rev. Clin. Immunol. 10, 1171–1184 (2014).
Hibbert, J. E., Currie, A. & Strunk, T. Sepsis-induced immunosuppression in neonates. Front. Pediatr. 6, 357 (2018).
Karlsson, H., Hessle, C. & Rudin, A. Innate immune responses of human neonatal cells to bacteria from the normal gastrointestinal flora. Infect. Immun. 70, 6688–6696 (2002).
Aulin, L. B. S. et al. Biomarker-guided individualization of antibiotic therapy. Clin. Pharmacol. Ther. 110, 346–360 (2021).
Keij, F. M. et al. Zebrafish larvae as experimental model to expedite the search for new biomarkers and treatments for neonatal sepsis. J. Clin. Transl. Sci. 18, e140 (2021).
Buras, J. A., Holzmann, B. & Sitkovsky, M. Animal models of sepsis: setting the stage. Nat. Rev. Drug Discov. 4, 854–865 (2005).
Meeker, N. D. & Trede, N. S. Immunology and zebrafish: spawning new models of human disease. Dev. Comp. Immunol. 32, 745–757 (2008).
Bauer, M. et al. A transcriptomic biomarker to quantify systemic inflammation in sepsis—a prospective multicenter phase II diagnostic study. EBioMedicine 6, 114–125 (2016).
Kurul, S. et al. Association of inflammatory biomarkers with subsequent clinical course in suspected late onset sepsis in preterm neonates. Crit. Care 25, 12 (2021).
Pammi, M. & Weisman, L. E. Late-onset sepsis in preterm infants: update on strategies for therapy and prevention. Expert Rev. Anti Infect. Ther. 13, 487–504 (2015).
Ohlsson, A. & Lacy, J. B. Intravenous immunoglobulin for suspected or proven infection in neonates. Cochrane Database Syst. Rev. 1, CD001239 (2020).
Haque, K. N., Remo, C. & Bahakim, H. Comparison of two types of intravenous immunoglobulins in the treatment of neonatal sepsis. Clin. Exp. Immunol. 101, 328–333 (1995).
Carr, R., Modi, N. & Dore, C. G-CSF and GM-CSF for treating or preventing neonatal infections. Cochrane Database Syst. Rev. CD003066 (2003).
Castagnola, E. & Dufour, C. Role of G-CSF GM-CSF in the management of infections in preterm newborns: an update. Early Hum. Dev. 90(Suppl 2), S15–S17 (2014).
Opal, S. M. New perspectives on immunomodulatory therapy for bacteraemia and sepsis. Int J. Antimicrob. Agents 36(Suppl. 2), S70–S73 (2010).
Ward, A. & Clissold, S. P. Pentoxifylline. A review of its pharmacodynamic and pharmacokinetic properties, and its therapeutic efficacy. Drugs 34, 50–97 (1987).
Neuner, P. et al. Pentoxifylline in vivo down-regulates the release of IL-1 beta, IL-6, IL-8 and tumour necrosis factor-alpha by human peripheral blood mononuclear cells. Immunology 83, 262–267 (1994).
Lauterbach, R. & Zembala, M. Pentoxifylline reduces plasma tumour necrosis factor-alpha concentration in premature infants with sepsis. Eur. J. Pediatr. 155, 404–409 (1996).
Szymura-Oleksiak, J., Bury, J., Lauterbach, R. & Pawloski, M. Serum concentrations of pentoxifylline and its metabolites in premature infants with sepsis when administered by continuous intravenous infusion. Pharmaceut. Sci. 3, 367–371 (1997).
Mandell, G. L. Cytokines, phagocytes, and pentoxifylline. J. Cardiovasc. Pharmacol. 25(Suppl 2), S20–S22 (1995).
Windmeier, C. & Gressner, A. M. Pharmacological aspects of pentoxifylline with emphasis on its inhibitory actions on hepatic fibrogenesis. Gen. Pharmacol. 29, 181–196 (1997).
Schroer, R. H. Antithrombotic potential of pentoxifylline. A hemorheologically active drug. Angiology 36, 387–398 (1985).
Yang, S. et al. Gut-derived norepinephrine plays a critical role in producing hepatocellular dysfunction during early sepsis. Am. J. Physiol. Gastrointest. Liver Physiol. 279, G1274–G1281 (2000).
Danjo, W., Fujimura, N. & Ujike, Y. Effect of pentoxifylline on diaphragmatic contractility in septic rats. Acta Med Okayama 62, 101–107 (2008).
Pammi, M. & Haque, K. N. Pentoxifylline for treatment of sepsis and necrotizing enterocolitis in neonates. Cochrane Database Syst. Rev. CD004205 (2015).
Salman, S. et al. Effects of maturation and size on population pharmacokinetics of pentoxifylline and its metabolites in very preterm infants with suspected late-onset sepsis or necrotizing enterocolitis: a pilot study incorporating clinical outcomes. Br. J. Clin. Pharmacol. 85, 147–159 (2019).
Lalu, M. M. et al. Evaluating mesenchymal stem cell therapy for sepsis with preclinical meta-analyses prior to initiating a first-in-human trial. Elife 5, e17850 (2016).
Johnson, C. L., Soeder, Y. & Dahlke, M. H. Concise review: mesenchymal stromal cell-based approaches for the treatment of acute respiratory distress and sepsis syndromes. Stem Cells Transl. Med. 6, 1141–1151 (2017).
Laroye, C., Gibot, S., Reppel, L. & Bensoussan, D. Concise review: mesenchymal stromal/stem cells: a new treatment for sepsis and septic shock? Stem Cells 35, 2331–2339 (2017).
Liau, L. L. et al. The potential of mesenchymal stromal cell as therapy in neonatal diseases. Front Pediatr. 8, 591693 (2020).
McIntyre, L. A. et al. Cellular immunotherapy for septic shock. A phase I clinical trial. Am. J. Respir. Crit. Care Med. 197, 337–347 (2018).
Sato, Y. et al. Prophylactic therapy with human amniotic fluid stem cells improved survival in a rat model of lipopolysaccharide-induced neonatal sepsis through immunomodulation via aggregates with peritoneal macrophages. Stem Cell Res. Ther. 11, 300 (2020).
Hotchkiss, R. S. & Karl, I. E. The pathophysiology and treatment of sepsis. N. Engl. J. Med. 348, 138–150 (2003).
Zhao, J. et al. Hyper innate responses in neonates lead to increased morbidity and mortality after infection. Proc. Natl Acad. Sci. USA 105, 7528–7533 (2008).
Wynn, J. L. et al. The influence of developmental age on the early transcriptomic response of children with septic shock. Mol. Med. 17, 1146–1156 (2011).
Wynn, J. L. et al. Postnatal age is a critical determinant of the neonatal host response to sepsis. Mol. Med. 21, 496–504 (2015).
Schüller, S. S. et al. Immunomodulation to prevent or treat neonatal sepsis: past, present, and future. Front. Pediatr. 6, 199 (2018).
Ridings, P. C. et al. Beneficial cardiopulmonary effects of pentoxifylline in experimental sepsis are lost once septic shock is established. Arch. Surg. 129, 1144–1152 (1994).
Itenov, T. S., Murray, D. D. & Jensen, J. U. S. Sepsis: personalized medicine utilizing ‘omic’ technologies—a paradigm shift? Healthcare (Basel) 6, 111 (2018).
van Karnebeek, C. D. M. et al. The role of the clinician in the multi-omics era: are you ready? J. Inherit. Metab. Dis. 41, 571–582 (2018).
Agakidou, E., Agakidis, C., Gika, H. & Sarafidis, K. Emerging biomarkers for prediction and early diagnosis of necrotizing enterocolitis in the era of metabolomics and proteomics. Front. Pediatr. 8, 602255 (2020).
Author information
Authors and Affiliations
Contributions
H.R.T. and S.V. designed and directed the manuscript. H.T., S.V., S.K., K.F. and R.F. performed literature searches and drafted the document. I.K.M.R., S. H.P.S. and H.K. reviewed and revised it critically for important intellectual content. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Consent statement
No patient consent was required.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Kurul, S., Fiebig, K., Flint, R.B. et al. Knowledge gaps in late-onset neonatal sepsis in preterm neonates: a roadmap for future research. Pediatr Res 91, 368–379 (2022). https://doi.org/10.1038/s41390-021-01721-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41390-021-01721-1
This article is cited by
-
Paediatric and neonatal sepsis and inflammation
Pediatric Research (2022)
