TITLE: Antibiotic Therapy in Moderate to Late Preterm Infants with Presumed Early Onset Sepsis is Associated with Aberrations to Feeding Tolerance and Hyperbilirubinemia

doi:10.21203/rs.3.rs-3402758/v1

Download PDF

Article

TITLE: Antibiotic Therapy in Moderate to Late Preterm Infants with Presumed Early Onset Sepsis is Associated with Aberrations to Feeding Tolerance and Hyperbilirubinemia

https://doi.org/10.21203/rs.3.rs-3402758/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Presumed early onset sepsis (PEOS) requiring antibiotics is concerning for our cohort of moderate to late preterm infants. Studies on very premature neonates have shown disruption to feeding tolerance achievement. We aimed to assess whether this concern is valid in our cohort and whether a link existed to prolonged jaundice. A single site retrospective study was performed on infants born between 32^+ 3 and 36^+ 6 weeks (2016 to 2019). PEOS infants on antibiotics (ABX) were strictly matched by gestational age (± 1 day) and birthweight (± 5%) against a reference group of non-PEOS infants (REF). Key outcomes included: enteral feeding commencement and achievement; feeding intolerance (FI); phototherapy commencement and duration; antibiotic therapy duration; maximum bilirubin (MaxBili); length of stay (LEN); net post-birth weight gain. ABX(n = 185):REF(n = 185) univariate and multivariable analysis revealed (respectively): delayed enteral feed commencement (p < 0.001, p < 0.001); no difference in FI; delayed onset of peak jaundice (p = 0.002, p = 0.001); increased duration of phototherapy (p = 0.016, p = 0.001); increased LEN (p < 0.002, p < 0.035). Univariate analysis also showed (ABX:REF): no significant difference in MaxBili; delayed full enteral feed achievement (p = 0.010). Univariant or multivariable analysis found no difference in irradiance level. However, for REF infants having 0 or 1 lights, there was a bias towards increased irradiance for ABX infants (< 0.001, 0.037 respectively). In moderate to late preterm infants, PEOS antibiotic therapy is associated with negative impacts on feeding and jaundice, adding to an infant’s burden of suffering.

Health sciences/Medical research

Health sciences/Risk factors

Early onset sepsis is a potentially serious concern amongst preterm neonates and typically occurs within 72 hours of birth [1]. Pathogenesis includes maternal gastrointestinal and genitourinary flora transmitted during labor [2].

Diagnosis is difficult, with signs and symptoms not dissimilar from other causes of neonatal distress [3]. Antibiotic therapy management is often undertaken without a confirmed infection [4] and is therefore presumed. In Australia, the most commonly used antibiotics for presumed early onset sepsis (PEOS) are a combination of benzyl penicillin (BPen) and gentamicin [5], which differs from Europe where the most common antibiotic combination is ampicillin and gentamicin [6]

Adverse outcomes accompany the use of antibiotics in neonates. These include antibiotic resistance, induced microbiome alterations and an impact on the developing immune system [7]. In premature infants, prolonged antibiotic therapy has been associated with late onset sepsis, necrotising enterocolitis and even death [8]. Understandably, such adverse outcomes have prompted antibiotic reduction strategies such as the EOS Risk Calculator [9].

In our Neonatal Unit (NNU), we care for moderate to late preterm, low birthweight infants with the key objective of optimising feeding and growing. Consequently, feeding tolerance achievement is of particular concern, but in the preterm infant early antibiotic therapy is associated with delayed feeding tolerance [10]. The mechanism is unclear [10], but feeding intolerance (FI) in the preterm infant is associated with altered gut microbiome composition [11] and antibiotic administration in premature infants affects the differential composition of gut microbiota [12].

In recent years, our review of the literature has revealed little interest regarding the impact of antibiotics on neonatal jaundice. During the 1970s and 1980s, there were concerns that antibiotics may cause bilirubin to be displaced from albumin leading to an increased risk of neonatal kernicterus. Gentamicin [13] and ampicillin [14] were investigated - gentamicin showed no displacement effects and ampicillin only a slight displacement effect. BPen was found to be a low-level displacer [15].

However, bilirubin-albumin displacement is only one facet of antibiotic-jaundice interaction. Bilirubin conjugation is via glucuronyl transferase, which is inhibited by gentamicin, but only at high concentrations [16]. Some antibiotics may compete with bilirubin for glucuronidation via UDP-glucuronosyltransferase (UGT) [17], which in the neonate occurs in both the liver and in the gut [18]. We have previously proposed that enhanced expression of intestinal UGT explains the improved bilirubin excretion we observed in breastmilk fortifier fed infants [18]. Conversely, it is possible that gut expression of UGT may be inhibited by antibiotics and bilirubin elimination compromised.

Our interest in antibiotics-jaundice association arises from the basic observation that in our infant cohort, those on antibiotics were more prone to phototherapy and the jaundice took longer to resolve. We have found only one reference to this - infants on BPen and gentamicin required more phototherapy, but these infants were more jaundiced prior to antibiotic commencement [19].

Consequently, we were prompted to carry out a retrospective study to look at both the impact of antibiotics on feeding tolerance in moderate to late preterm infants (most previous studies have examined more preterm infants) and to assess whether a negative antibiotics-jaundice impact existed and identify potential implications.

Study design and participants

A retrospective study was conducted on infants admitted to the NNU at The Northern Hospital (TNH) in Victoria, Australia, between January 2016 and December 2019. Cases were all infants, between 32^+ 3 and 36^+ 6 weeks gestational age who required antibiotic therapy during NNU admission for PEOS (ABX). ABX cases were strictly matched to a control infant not receiving any antibiotic therapy during NNU admission (REF), to firstly to within ± 1 day of gestational age (GA) and then within 5% of birthweight. A key aspect of selecting the REF infants admitted to NNU was to provide comparability with burden of morbidity and patient complexity. Given the constraints of PEOS and controls available from TNH Birth Register during this period of time, the final analyses included 185 matched ABX and REF pairs.

TNH serves a lower-socioeconomic, multicultural, multi-ethnic community.

Inclusion criteria were: infants requiring NNU admission at birth and born at less than 37 weeks gestation. Infant exclusion criteria were: neonatal death (within the first 28 days of life); recognised chromosomal abnormalities; transferral to tertiary hospitals shortly after birth.

Ethics approval

was by the Northern Health Office of Research (ALR No. 51.2019).

Data collection and outcomes

The study cohort was selected from a database of births at TNH provided by the Northern Health Decision Support Department.

The primary outcomes were neonatal hyperbilirubinemia and feeding tolerance – as defined below.

Definitions for the purposes of this study are as follows: PEOS – sepsis not confirmed by blood culture but presumed via clinical suspicion within 72 hours of birth; neonatal hyperbilirubinemia - requiring phototherapy for jaundice management; feeding tolerance - the preterm infant being able to safely ingest / digest prescribed enteral feeding without complications associated with aspiration, infection, and gastrointestinal dysfunction; FI - significant abdominal distention and / or emesis in the neonate resulting in a lack of feeding tolerance.

Additional hyperbilirubinemia outcomes: maximum total bilirubin concentration (µmol/L); age in hours at maximum total bilirubin concentration; maximum direct bilirubin concentration (µmol/L); maximum number of phototherapy lights (single, double or higher); total hours of phototherapy. Phototherapy was via the MediLED Mini (Medix, Jose Arias 293, Villa Lynch San Martin, 1672 Buenos Aires, Argentina). Quality control is managed routinely by the hospital’s Biomedical Engineering Department. Phototherapy requirement was determined via “Treatment for Hyperbilirubinaemia in Newborn Babies Charts” specific for gestational age and bilirubin level – developed as a guideline policy by the Royal Women’s Hospital, Melbourne, Australia.

Additional enteral feeding outcomes: admission enteral feed type (exclusively breastmilk, formula, mixed feeds); hours to commence enteral feeds; days to reach full enteral feeds (total fluid intake of ≥ 150ml/kg/day).

Secondary outcomes: NNU length of stay in days (LEN); growth as defined by the post-birth net gain in weight (discharge minus birthweight); weight Z score gain on discharge (adjusted for gestational age and gender using 2013 Fenton growth charts [20]).

Sample Size Calculations

Assuming an effect size of 0.25 (small), a power of 80% and a two-sided level of significance of 5%, a sample size of 128 matched pairs was calculated as being suitable for establishing if the mean of the paired differences is significantly different from zero - or a two-sided p-value is less than 0.05.

We aimed to keep numbers in the study above these estimates.

Statistical analysis

Data was analysed using IBM SPSS Statistics for Windows, Version 26 (IBM Corp., Armonk, N.Y., USA) and Stata Statistical Software for Windows, Release 16 (StataCorp LLC., College Station, TX., USA). Assistance was provided by the TNH research statistician. Parametric and non-parametric tests were used to determine the normality of continuous variables. For univariate analysis: Chi-square and Fisher’s Exact tests were used for categorical factors; Mann-Whitney rank-sum & Kruskal-Wallis H tests for continuous factors.

Confounding variables were identified as factors with p < 0.2 through univariate analysis plus factors considered likely to be clinically relevant. These confounders are fully listed Table 4 footnote. For all outcomes (excepting time to reach full enteral feeds), a process of backward and forward multivariable logistic regression was used (Stepwise Regression) [21]. For the time to reach full enteral feeds outcomes - Cox regression time-to-event analysis was used.

Significance was via two-tailed p-values < 0.05 in univariate and multivariable.

A total of 370 preterm infant-mother pairs were enrolled and separated into the ABX group (n = 185) or REF group (no ABX therapy) (n = 185) during NNU admission. Univariate analysis results are presented in Tables 1, 2 and 3 and multivariable analysis in Table 4.

Antibiotic Therapy

All 185 infants included in the study were given narrow-spectrum antibiotics to control EOS. The great majority (n = 169) were given a combination BPen (gram positive) and gentamicin (gram negative) only. 11 infants were given an additional narrow-spectrum antibiotic (flucloxacillin or metronidazole). Broad-spectrum antibiotics were used in ~ 3% of infants – either cefotaxime (4 infants) or ampicillin (1 infant) combined with the narrow-spectrum antibiotics of BPen and gentamicin.

Baseline Characteristics - Table 1

The key aspects of maternal age, neonatal birthweight, gestational age and gender were all matched well. In the ABX group, there was a higher frequency of maternal intrapartum fever, neonatal vaginal birth, instrument delivery, birth trauma, resuscitation and respiratory distress (it is noted that a number of these factors are prompts for the suspicion of PEOS).

PEOS (due to blood culture negative status) accounted for 100% of the ABX group.

Table 1

Maternal and Infant Cohort Comparisons – no antibiotics vs antibiotic therapy
	No ABX (n = 185)	ABX (n = 185)	p value
Maternal Characteristics
Age median (IQR) (years) (paired)	30 (7)	30 (7)	0.288
Region of Birth, %
Australia, New Zealand vs Other (paired)	51.9%	54.1%	0.755
Gravida, median (IQR) (paired)	2 (2)	2 (2)	0.658
Para, median (IQR)	1 (1)	1 (1)	0.790
GBS-positive status	22.0%	25.2%	0.645
Maternal intrapartum fever, %	0%	5.9%	0.001
Maternal intrapartum antibiotic, %	14.1%	20.5%	0.130
Maternal CRP, mg/L median (IQR)	7.0 (7.3)	8.4 (9.5)	0.060
Neonatal Characteristics
Gestational age, median (IQR) (weeks) (paired)	35.5 (1.3)	35.4 (1.3)	0.194
Birthweight, mean (SD) (grams) (paired)	2364 (454)	2365 (451)	0.958
Male, %	53.5%	58.4%	0.402
Apgar 5-minute, median (IQR) (unpaired)	9 (0)	9 (0)	0.030
Fetal distress, %	35.7%	35.7%	1.000
Delivery mode, %
Normal Vaginal Delivery	34.1%	63.8%	< 0.001
Emergency Caesarean	43.8%	30.8%	0.013
Elective Caesarean	22.2%	5.4%	< 0.001
Instrumentation at delivery, %	11.4%	14.1%	0.533
Neonatal birth trauma, %	2.7%	14.1%	< 0.001
Neonatal resuscitation, %	28.6%	42.2%	0.009
Neonatal respiratory distress, %	29.7%	53.5%	< 0.001
Neonatal hypoglycaemia, %	35.7%	31.4%	0.441
Haemolysis present, %	3.2%	2.3%	0.716
Birth Temperature, mean (SD) °C (unpaired)	36.5 (0.37)	36.6 (0.47)	0.062
Neonatal blood culture positive	NIL	NIL	-------
Antibiotic Duration, median (IQR), days	------	1 (1)	-------

Chi-square test for categorical variables, Students t-test or Mann-Whitney U test.

Jaundice Outcomes - Table 2

There was no evidence of a difference in the maximum total direct bilirubin. Peak jaundice level occurred ~ 20 hours later in the ABX group.

The duration of phototherapy was almost 10 hours longer for the ABX group.

When the REF group was used as a baseline for those infants having no jaundice or jaundice requiring only one phototherapy light, there was a bias towards increased irradiance (light number exposure) in the ABX group.

Table 2

Neonatal Jaundice
	Use of Antibiotic Therapy
	No	Yes	p-value
Neonatal Jaundice Outcomes
Any Phototherapy	75.2%	76.5%	0.806
Maximum total bilirubin, mean (SD) (µmol/L)	213 (45)	214 (46)	0.840
Maximum direct bilirubin, mean (SD) (µmol/L)	12 (4)	11(4)	0.398
Age at maximum total bilirubin, median (IQR) (hours of life)	72 (51)	93 (55)	0.002
Phototherapy treatment duration, median (IQR) (hours)	28 (22)	38(26)	0.016
Phototherapy Light Dosage	Number of Lights
Median (IQR) [paired, n = 20]	0(0)	1.5(1.25)	< 0.001
Median (IQR) [paired, n = 18]	1(0)	2(1.75)	0.037
Median (IQR) [paired, n = 27]	2(0)	2(3)	0.085
Maximum total bilirubin (paired, n = 86); Maximum direct bilirubin (paired n = 87) ; Age at maximum total bilirubin (unpaired, non-ABX n = 125, ABX n = 132); Phototherapy treatment duration (unpaired, non-ABX n = 93, ABX n = 100)

Table 3 - Feeding Tolerance and Secondary results

There was no difference in the FI of both groups.

For the ABX group, median enteral feed commencement was delayed by 2 hours. Median full enteral feed achievement was delayed by 2 days.

Median LEN was approximately two days longer for ABX group compared to REF.

Table 3

Feeding and secondary outcomes
	Antibiotic Therapy
	No	Yes	p-value
Feeding Intolerance %	10.8%	10.8%	1.000
Enteral feed during admission
Breast milk-Exclusive %	4.3%	8.6%	0.138
Mixed %	87.0%	82.2%	0.249
Enteral Feeds Commencement, median (IQR) (hours)	1.0 (0.6)	2.0 (8.45)	< 0.001
Full Enteral Feed Achievement, median (IQR) (days)	6 (1.0)	8(12)	0.010
Secondary Outcomes
LEN, median (IQR) (days)	8 (12)	10 (11)	0.002
Readmission %	13.0%	13.0%	1.000
Weight gain from birth at discharge g, mean (SD)	81(279)	116 (271)	0.031
Enteral Feed Admission (n = 185); LEN (paired n = 184); Enteral Feed Commencement (unpaired n = 185); Feeding Intolerance (n = 185); Full Enteral Feeds (unpaired No n = 130, Yes n = 185); Weight Gain (paired n = 185);

Ref: No – all groups referenced to the non-antibiotic use (Reference) group. LEN - length of stay in NNU.

Table 4 - Multivariable Analysis. Many of the findings for univariate analysis were also found to be significant with multivariable analysis. These included for the ABX compared to REF: delayed onset of jaundice and greater duration of phototherapy, delayed time to commence enteral feeds and longer time achieve full enteral feeds; longer LEN.

Table 4

Multivariable regression analysis for infants who received antibiotic therapy compared to the control group (n = 370)
Outcome	Odds or Hazards Ratio	95% Confidence Interval	P-value
Neonatal hyperbilirubinaemia/any phototherapy*
Use of antibiotic therapy	1.04	0.77–1.4	0.791
Triple phototherapy*
Use of antibiotic therapy	1.54	0.65–3.65	0.325
Double phototherapy*
Use of antibiotic therapy	0.92	0.39–2.19	0.853
Duration of phototherapy*
Use of antibiotic therapy	1.24	1.10–1.41	0.001
Age at maximum bilirubin*
Use of antibiotic therapy	1.24	1.12–1.37	0.001
Feeding intolerance †
Use of antibiotic therapy	1.00	0.52–1.93	1.000
Time to commence enteral feeds †
Use of antibiotic therapy	2.75	2.32–3.27	< 0.001
Time to reach full enteral feeds †
Use of antibiotic therapy	1.10	0.96–1.25	0.173
Length of stay ‡
Use of antibiotic therapy	1.31	1.02–1.67	0.035
Readmission ‡
Use of antibiotic therapy	1.65	0.65–4.16	0.290
Weight gain on discharge^
Use of antibiotic therapy	1.10	0.87–1.38	0.433
Weight Z score gain on discharge^
Use of antibiotic therapy	0.96	0.89–1.04	0.303

In all cases, statistical significance (< 0.05) indicates a compromised outcome for the ABX group.

For the time to reach full enteral feeds, a time-to-event analysis using Cox regression was used.

These results are presented as hazard ratios with 95% confidence intervals. All other results are presented as odds ratios with 95% confidence intervals.

* Variables included – use of maternal intrapartum antibiotics, mode of delivery, neonatal hypoglycaemia, birth trauma, maternal blood group, neonatal blood group, gestational age, birthweight, maternal WCC count, neonatal temperature.

† Variables included – gestational age, birthweight, maternal intrapartum fever, use of maternal intrapartum antibiotics, use of maternal analgesia, fetal distress, birth trauma, neonatal resuscitation, respiratory distress, neonatal hypoglycaemia, neonatal temperature.

‡ Variables included – maternal ethnicity, maternal blood group, maternal analgesia, neonatal hypoglycaemia, neonatal blood group, fetal distress, mode of delivery, use of instrumentation during birth, birth trauma, neonatal resuscitation, respiratory distress, gestational age, maternal age, birthweight, APGAR 1 and 5 minutes.

^ Variables included – maternal fever, neonatal hypoglycaemia, birth trauma, neonatal temperature, fetal distress, use of maternal intrapartum antibiotics, mode of delivery, respiratory distress, gestational age.

The key findings of our study relate to both feeding tolerance and neonatal jaundice in moderate to late preterm infants. In summary, univariate and multivariable analysis revealed that: antibiotic therapy was associated with a delay in enteral feed commencement and delay in reaching full enteral feeds; antibiotic therapy was associated with delayed onset of peak jaundice by around 20 hours and increased the duration of phototherapy. To our knowledge the findings relating to phototherapy have not been reported previously.

Given the importance of nutrition, there have been a number of studies on antibiotic use and feeding implications in very low birthweight infants (VLBW - <1500g birthweight). Ampicillin and gentamicin delayed feeding initiation and total milk intake [8]. Antibiotic therapy prolonged total parental nutrition [10]. Early antibiotics delayed the achievement of milk feeds, but antibiotics commenced 9 days after birth reduced this delay [22] – suggesting that early antibiotics interfere with gastrointestinal (GI) tract maturation. This study’s findings for the moderate to late preterm infant are similar to those for the VLBW infant.

Fewer studies have been done on antibiotic-feeding concerns in infants born towards term. A recent study [23] reported that term infants commenced on early ampicillin and gentamicin, four weeks after ceasing antibiotics had increased levels of potentially pathogenic Proteobacteria and decreased beneficial Actinobacteria, Bifidobacteria and Lactobacillus. These microbiota changes persisted, for at 8 weeks post antibiotics, Proteobacteria was still elevated and Bifidobacterium diversity was reduced.

Changes in the GI microbiota of antibiotic exposed newborn infants perhaps explain the feeding dysfunctions reported in VLBW infants and the similar findings in our moderate to late preterm infants. Animal studies show that postnatal maturation of the intestine is partially modulated by bacterial colonisation aiding in the establishment of an efficient barrier to luminal antigens [24]. Gut immunological, sensory and motor functions in the immature GI tract are developed and maintained by commensal bacteria, with phylum including Bifidobacterium and Lactobacillus enhancing intestinal epithelia survival [25]. Consequently, antibiotic altered GI microbiota may disrupt development of GI immune and neural patterns contributing to the aetiology of feeding dysfunction [25].

While enteral feeding delay in the ABX group was not unexpected, we were surprised that peak jaundice was also delayed. Reassuringly, there was no difference in peak bilirubin level between the groups – no evidence that antibiotics increased kernicterus risk. The findings of increased duration of ABX phototherapy, and a bias towards increased irradiation (more phototherapy lights) when compared to the REF group as a baseline, concorded with our previous basic observations. There are research findings that may explain these results.

The first step towards jaundice in the neonate is the breakdown of red blood cells, followed by heme degradation, via heme oxygenase to biliverdin and then bilirubin [26]. The steep rise in bilirubin production is due to a shortened life span of erythrocytes in full term and preterm infants. Eighty to ninety percent of bilirubin is formed from haemolysed or senescent erythrocytes [26]. We suspect the observed delay in ABX group peak bilirubin onset is due to a perturbation of this first step. Antibiotics strengthen the erythrocyte membrane leading to a reduction in haemolysis. Penicillin and gentamicin have both been shown to decrease haemolysis in a dose dependant manner [27]. In the case of BPen, increased cellular potassium levels are implicated in this strengthening of the erythrocyte membrane [28]. Interestingly, this antibiotic strengthening of erythrocyte membranes has allowed gentamicin encapsulated red blood cells to be proposed as an effective slow-release biodegradable drug carrier with a half-life greater than other known carriers [29]. Consequently, it is plausible that the ABX vs REF peak bilirubin delay is a result of delayed levels of initial erythrocyte haemolysis due to an antibiotic related effect of cell membrane strengthening.

Bilirubin elimination is a complex process, requiring conjugation and gut elimination. UDP-glucuronosyltransferase 1A1 (UGT1A1) is solely responsible for bilirubin conjugation and is expressed in both the liver and intestine [18]. In the gut, UGT1A1 is induced by enteral feeding and especially glucose (eg starch maltodextrin as present in milk fortifiers) and has the potential to be very effective as it can re-conjugate bilirubin which has been de-conjugated by β-glucuronidase present in breast milk [30] and in gut bacteria [31], enhancing bilirubin elimination and enterohepatic re-circulation [18]. We hypothesize that the increase in ABX group duration of phototherapy is potentially related to: antibiotic gut dysfunction that delays enteral feeding, thus downregulating intestinal UGT1A1 expression and burdening liver UGT1A1; antibiotic induced cell membrane strengthening which result in longer erythrocyte half-life and a prolonged degradation of heme.

Our study showed an increased LEN for the ABX group. While perhaps not unexpected in an infant coping with the morbidity of an infection, these results are also consistent with an infant challenged by the morbidities of exacerbated hyperbilirubinemia and compromised enteral feeding. Given the limitations of a retrospective study, it is not possible to differentiate these findings further. However, it does highlight concerns of antibiotic overtreatment of EOS and moderation of antibiotic use via alternative EOS assessment pathways such as the EOS Risk Calculator [9].

A key limitation of this study is that it is retrospective and consequently has inferior evidence when compared to a prospective study. It is also a single centre study, which may limit the applicability of the results. However, we follow national guidelines for antibiotic management of EOS and our usage of antibiotics reflects national patterns [5]. A strength of the study is the strict matching of the ABX and REF group. As with any retrospective study, the key findings cannot be seen as causative – we only claim an association with our findings.

A not uncommon problem with retrospective studies on neonatal populations is that while case-control helps to minimise differences between the compared cohorts, there are usually some differences remaining. In this study, the key results relate to jaundice and feeding. With regard to jaundice, studies have clearly shown that not only gestational age, but also birthweight is a confounding factor – jaundice frequency and levels are negatively correlated to both gestational age and weight in preterm infants [32, 33]. While gestational age has a role in feeding maturation in the preterm infant, birthweight is positively associated with feeding strength and intensity [34]. Consequently, the strict matching of birthweight, GA and gender in this study goes some way towards addressing these confounding aspects. However, the key technique in adjusting for such differences comes from multivariable analysis. When there are potential confounders, multivariable analysis provides the only solution of control [35]. We have rigorously applied Stepwise Regression and provided a detailed listing of variables retained or removed as a footnote to Table 4.

Another issue which is pertinent is to what extent infection itself, rather than the use of antibiotics have influenced the results. Interestingly, there is an association between infection and jaundice. UTI has been found in 12.5% of asymptomatic jaundiced neonates in the first week of life [36]. However, the action of the bilirubin molecule in response to infection is equivocal, militating against a consistent underlying physiological protective response. While bilirubin itself has been shown to have antibacterial properties against Group B Streptococcus and some gram-positive organisms, it can act as a protectant to E. Coli strains – an organism which accounts for around 25% of early onset sepsis [37]. Regarding the effects of sepsis on the function of the preterm gut itself, evidence is not extensive. What is known is that sepsis induces inflammation in the gut mucosa which influences microbiota changes and extends inflammation associated damage to other organs. Sepsis related oxidative stress has the potential to reduce beneficial anaerobes in the gut lumen [38]. On the other hand, there is evidence to suggest that antibiotics themselves are associated with an increase in morbidity in a dose dependant manner [39]. As with jaundice and feeding, we have relied on multivariable analysis to highlight associations. As a final comment on limitations in regard to infection, given that the treatment of early onset sepsis with antibiotics is empirical, our cohort will have been expected to include a large proportion of infants without infection [40]. This is supported by our data in that there were no positive blood cultures detected in the ABX group.

The management of the newborn infant with proven or presumed EOS is a dilemma for the neonatologist. The use of antibiotics is designed to combat perinatal infection, which accounts for up to 10% of neonatal deaths in Australia [5]. However, the use of antibiotics has the potential to increase morbidity and add to the burden of suffering of an infant. The results of this study add valuable information for the clinician assessing the management options open to an infant with EOS. This study raises concerns that early neonatal antibiotic use increases the risk of developing feeding dysfunction in the moderate to late preterm infant. Importantly, it further extends the knowledge of morbidity risk to the very common neonatal concern of hyperbilirubinemia – the moderate to late preterm infant will likely have delayed jaundice that is more intractable. This latter outcome has immediate phototherapy management ramifications including insensible water loss, increased peripheral blood flow and potential childhood sequelae such as type 1 diabetes and asthma [41].

Data Availability The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments:

Thanks to Mark Tacey, statistician, for his invaluable advice and input into data analysis.

Ethical Approval This study was performed in line with the principles of the Declaration of Helsinki. Ethics approval was by the Northern Health Office of Research (ALR No. 51.2019). As this is a retrospective study with de-identified patient data, the Northern Health Office of Research considered informed consent from participants as not applicable.

Funding This study received no funding.

Conflict of interest The authors have no relevant financial or non-financial interests to disclose.

Simonsen, K.A., Anderson-Berry, A.L., Delair, S.F., & Davies, H.D. Early-onset neonatal sepsis. Clin. Microbiol. Rev. 27, 21–47. doi:10.1128/CMR.00031-13 (2014).
Puopolo, K.M., Benitz, W.E., & Zaoutis, T.E. COMMITTEE ON FETUS AND NEWBORN; COMMITTEE ON INFECTIOUS DISEASES. Management of Neonates Born at ≥ 35 0/7 Weeks' Gestation With Suspected or Proven Early-Onset Bacterial Sepsis. Pediatrics. 142: e20182894. doi:10.1542/peds.2018-2894 (2018).
Gerdes, J.S. Diagnosis and management of bacterial infections in the neonate. Pediatr. Clin. North Am. 51, 939-ix. doi:10.1016/j.pcl.2004.03.009 (2004).
Schulman, J., Profit, J., Lee, H.C., Dueñas, G., Bennett, M.V., Parucha, J., Jocson, M.A.L., & Gould, J.B. Variations in Neonatal Antibiotic Use. Pediatrics. 142, e20180115. doi:10.1542/peds.2018-0115 (2018).
McMullan, B., Cooper, C., Spotswood, N., James, R., Jones, C., Conecny, P., Blyth, C., & Karen, T. Antibiotic prescribing in neonatal sepsis: an Australian nationwide survey. BMJ Paediatrics Open. 4, e000643. doi: 10.1136/bmjpo-2020-000643 (2020).
Garrido, F., Allegaert, K., Arribas, C., Villamor, E., Raffaeli, G., Paniagua, M., & Cavallaro, G. On behalf of European Antibiotics Study Group (EASG). Variations in Antibiotic Use and Sepsis Management in Neonatal Intensive Care Units: A European Survey. Antibiotics. 10, 1046. https://doi.org/10.3390/antibiotics10091046 (2021).
Cotton, C.M. Adverse consequences of neonatal antibiotic exposure Curr. Opin. Pediatr. 28, 141–149. doi:10.1097/MOP.0000000000000338 (2016).
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. doi:10.1016/j.jpeds.2011.05.033 (2011).
Achten, N.B., Klingenberg, C., Benitz, W.E., Stocker, M., Schlapbach, L.J., Giannoni, E., Bokelaar, R., Driessen, G.J.A., Brodin, P., Uthaya, S., van Rossum, A.M.C., & Plötz, F.B. Association of Use of the Neonatal Early-Onset Sepsis Calculator With Reduction in Antibiotic Therapy and Safety: A Systematic Review and Meta-analysis. JAMA Pediatr. 173, 1032–1040. doi:10.1001/jamapediatrics.2019.2825 (2019).
Martinez, F.E., Ferri, W.A.G., Leone, C.R., de Almeida, M.F.B., Guinsburg,R., Meneses, J.D.A., Vale, M.S.D., Marba, S.T.M., de Carvalho, W.B., de Souza Rugolo, L.M.S. et al Early Empiric Antibiotic Use Is Associated With Delayed Feeding Tolerance in Preterm Infants: A Retrospective Analysis. J. Pediatr. Gastroenterol. Nutr. 65, 107–110. doi:10.1097/MPG.0000000000001490 (2017).
Yuan, Z., Yan, J., Wen, H., Deng, X., Li, X., & Su, S. Feeding intolerance alters the gut microbiota of preterm infants. PLoS One. 14, e0210609. doi:10.1371/journal.pone.0210609 (2019).
Arboleya, S., Sánchez, B,. Milani, C., Duranti, S., Solís, G., Fernández, N., de los Reyes-Gavilán, C.G., Ventura, M., Margolles, A., & Gueimonde, M. Intestinal microbiota development in preterm neonates and effect of perinatal antibiotics. J. Pediatr. 166, 538–544. doi:10.1016/j.jpeds.2014.09.041 (2015).
Wennberg, R.P., & Rasmussen, L.F. Effects of gentamicin on albumin binding of bilirubin J. Pediatr. 86, 611–613. doi:10.1016/s0022-3476(75)80163-6 (1975).
Brodersen, R., & Ebbesen, F. Bilirubin-displacing effect of ampicillin, indomethacin, chlorpromazine, gentamicin, and parabens in vitro and in newborn infants. J. Pharm. Sci. 72, 248–253. doi:10.1002/jps.2600720312 (1983).
Wadsworth, S.J., & Suh, B. In vitro displacement of bilirubin by antibiotics and 2-hydroxybenzoylglycine in newborns. Antimicrob. Agents Chemother. 32, 1571–1575. doi:10.1128/AAC.32.10.1571 (1988)
Malaka-Zafriu, K., Tsiouris, I., & Cassimos, C. The effect of gentamicin on liver glucuronyl transferase. J. Pediatr. 82, 118–120. doi:10.1016/s0022-3476(73)80028-9 (1973).
Fujiwara, R., Sumida, K., Kutsuno, Y., Sakamoto, M., & Itoh, T. UDP-glucuronosyltransferase (UGT) 1A1 mainly contributes to the glucuronidation of trovafloxacin. Drug Metab. Pharmacokinet. 30, 82–88. doi:10.1016/j.dmpk.2014.09.003 (2015).
Ma, X.W., Fan, W.Q. Earlier Nutrient Fortification of Breastmilk Fed LBW Infants Improves Jaundice Related Outcomes. Nutrients. 12: 2116. https://doi.org/10.3390/nu12072116 (2020).
Molyneux, E.M., Dube, Q., Banda, F.M., Chiume, M., Singini, I., Mallewa, M., Schwalbe, E.C., & Heyderman, R.S. The Treatment of Possible Severe Infection in Infants: An Open Randomized Safety Trial of Parenteral Benzylpenicillin and Gentamicin Versus Ceftriaxone in Infants < 60 days of Age in Malawi. Pediatr. Infect. Dis. J. 36, e328-e333. doi:10.1097/INF.0000000000001576 (2017).
Fenton, T.R., & Kim, J.H. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatr. 13, 59 https://doi.org/10.1186/1471-2431-13-59 (2013).
Chowdhury, M.Z.I., & Turin, T.C. Variable selection strategies and its importance in clinical prediction modelling. Fam. Med. Community Health. 8, e000262. doi:10.1136/fmch-2019-000262 (2020).
Zhou, W. Association Between Broad-Spectrum Antibiotic Use at Different Stage and Neonatal Enteral Nutrition and Prognosis in Very Low-Birth-Weight Infants with Culture-Negative Sepsis. J. Pediatr. & Child Health Care. 5, 1033. (2020).
Fouhy, F., Guinane, C.M., Hussey, S., Wall, R., Ryan, C.A., Dempsey, E.M., Murphy, B., Ross, R.P., Fitzgerald, G.F., Stanton, C., & Cotter P.D. High-throughput sequencing reveals the incomplete, short-term recovery of infant gut microbiota following parenteral antibiotic treatment with ampicillin and gentamicin. Antimicrob. Agents Chemother. 56, 5811–5820. doi:10.1128/AAC.00789-12 (2012).
Schumann, A., Nutten, S., Donnicola, D., Comelli, E.M., Mansourian, R., Cherbut, C., Corthesy-Theulaz,I., & Garcia-Rodenas, C. Neonatal antibiotic treatment alters gastrointestinal tract developmental gene expression and intestinal barrier transcriptome. Physiol. Genomics. 23, 235–245. doi: 10.1152/physiolgenomics.00057.2005 (2005).
Di Mauro, A., Neu, J., Riezzo, G., Raimondi, F., Martinelli, D., Francavilla, R., & Indrio, F. Gastrointestinal function development and microbiota. Ital. J. Pediatr. 39, 15. doi:10.1186/1824-7288-39-15 (2013).
Stevenson, D.K., Dennery, P.A., & Hintz, S.R. Understanding newborn jaundice. J. Perinatol. 21, Suppl 1, S21-S39. doi:10.1038/sj.jp.7210628 (2001).
Lijana, R.C., & Williams, M.C. The effects of antibiotics on hemolytic behavior of red cells. Cell Biophys. 8, 223–242. doi:10.1007/BF02788514 (1986).
Vereshchagina, V.M. Deĭstvie penitsillina na membranu éritrotsitov [Action of penicillin on the erythrocyte membrane]. Antibiotiki. 1984; 29: 40–42. Russian. PMID: 6320715. (1984).
Eichler, H.G., Rameis, H., Bauer, K., Korn, A., Bacher, S., & Gasić, S. Survival of gentamicin-loaded carrier erythrocytes in healthy human volunteers. Eur. J. Clin. Invest. 16, 39–42. doi:10.1111/j.1365-2362.1986.tb01305.x (1986).
Gourley, G.R., Kreamer, B., Cohnen, M., & Kosorok, M.R. Neonatal jaundice and diet. Arch. Pediatr. Adolesc. Med. 153, 184–188. doi:10.1001/archpedi.153.2.184 (1999).
Tiribelli, C., Ostrow, J.D. Intestinal flora and bilirubin. J. Hepatol. 42, 170–172. doi:10.1016/j.jhep.2004.12.002 (2005).
Wood, B., Culley, P., Roginski, C., Powell, J., & Waterhouse, J. Factors affecting neonatal jaundice. Arch. Dis. Child. 54,111–5. doi: 10.1136/adc.54.2.111. PMID: 434885; PMCID: PMC1545363. (1979).
Billing, B.H., Cole, P.G., & Lathe, G.H. Increased plasma bilirubin in newborn infants in relation to birth weight. Br. Med. J. 2, 1263–1265. doi:10.1136/bmj.2.4899.1263 (1954).
Wrotniak, B.H., Stettler, N., & Medoff-Cooper, B. The relationship between birth weight and feeding maturation in preterm infants. Acta. Paediatr. 98, 286–290. doi:10.1111/j.1651-2227.2008.01111.x (2009).
Pourhoseingholi, M.A., Baghestani, A.R., & Vahedi, M. How to control confounding effects by statistical analysis. Gastroenterol. Hepatol. Bed Bench. 5, 79–83. PMID: 24834204; PMCID: PMC4017459. (2012).
Shahian, M., Rashtian, P., & Kalani, M. Unexplained neonatal jaundice as an early diagnostic sign of urinary tract infection. Int. J. Infect. Dis. 16, e487-e490. doi:10.1016/j.ijid.2012.02.011 (2012).
Hansen, R., Gibson, S., de Paiva, E., Goddard, M., MacLaren, A., Karcher, A., Berry, S., Collie-Duguid, E., El-Omar, E., Munro, M. et al Adaptive response of neonatal sepsis-derived Group B Streptococcus to bilirubin. Sci. Rep.. 8,1–10. https://doi.org/10.1038/s41598-018-24811-3 (2018).
Cernada, M., Bäuerl, C., Serna, E., Collado, M.C., Martínez, G.P., & Vento, M. Sepsis in preterm infants causes alterations in mucosal gene expression and microbiota profiles compared to non-septic twins. Sci. Rep. 6, 25497. doi:10.1038/srep25497 (2016).
Fajardo, C., Alshaikh, B., & Harabor, A. Prolonged use of antibiotics after birth is associated with increased morbidity in preterm infants with negative cultures..J Matern. Fetal Neonatal. Med. 32,4060–4066. doi:10.1080/14767058.2018.1481042 (2019).
Fjalstad, J.W., Stensvold, H.J., Bergseng, H., Simonsen, G.S., Salvesen, B., Rønnestad, A.E., Klingenberg, C. Early-onset Sepsis and Antibiotic Exposure in Term Infants: A Nationwide Population-based Study in Norway. Pediatr. Infect. Dis. J. 35, 1–6. doi:10.1097/INF.0000000000000906 (2016).
Maisels, M.J., McDonagh, A.F. Phototherapy for neonatal jaundice. N. Engl. J. Med. 358, 920–928. doi:10.1056/NEJMct0708376 (2008).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

TITLE: Antibiotic Therapy in Moderate to Late Preterm Infants with Presumed Early Onset Sepsis is Associated with Aberrations to Feeding Tolerance and Hyperbilirubinemia

Status:

Version 1

Abstract

Introduction

Methods

Study design and participants

Data collection and outcomes

Sample Size Calculations

Statistical analysis

Results

Antibiotic Therapy

Baseline Characteristics - Table 1

Jaundice Outcomes - Table 2

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1