Short-course antibiotic therapy for pneumonia in the neonatal intensive care unit

To determine the adherence and safety outcomes of a 5-day antibiotic course with a “time-out” for treatment of “blood culture-negative” pneumonia in the NICU. Prospective surveillance of all infants diagnosed with pneumonia at 7 NICUs from 8/2020-12/2021. Safety outcomes were defined a priori by re-initiation of antibiotic therapy within 14 days after discontinuation and overall and sepsis-related mortality. 128 infants were diagnosed with 136 episodes of pneumonia; 88% (n = 119) were treated with 5 days of definitive antibiotic therapy. Antibiotics were restarted within 14 days in 22 (16%) of the 136 pneumonia episodes. However, only 3 (3%) of the 119 episodes of pneumonia treated for 5 days had antibiotics restarted for pneumonia. Mortality was 5% (7/128); 5 of the 7 deaths were assessed as sepsis-related. Adherence to the 5-day definitive antibiotic treatment for “culture-negative” pneumonia was high and the intervention seemed safe.


INTRODUCTION
Infants in the neonatal intensive care unit (NICU) are often provided prolonged antimicrobial therapy for a diagnosis of pneumonia [1,2]. There is growing concern that prolonged or excessive antibiotic therapy may lead to perturbations in the intestinal and airway bacterial microbiome. In preterm infants, the resultant dysbiosis has been associated with both short-and long-term adverse consequences such as necrotizing enterocolitis (NEC), bronchopulmonary dysplasia, retinopathy of prematurity, late-onset sepsis, systemic candidiasis, neurodevelopmental impairment, and death [3][4][5][6][7][8][9][10]. In full-term infants, prolonged antibiotic therapy has been associated with recurrent wheezing, obesity, as well as gastrointestinal and neurologic disorders [11][12][13]. In addition, prolonged therapy with some antibiotics may increase the risk for hearing loss, renal dysfunction, and seizures [14][15][16]. After controlling for severity of illness by the Clinical Risk Index in Babies (CRIB) II score, Cantey et al. [3] showed that each additional day of antibiotic therapy provided to preterm, very-low-birth-weight (VLBW; ≤1500 g) infants in the first two weeks of age was associated with significantly increased risk of late-onset sepsis, NEC, or death. Similarly, using the Score for Neonatal Acute Physiology II (SNAP-II) as a measure of newborn disease severity, Ting et al. [5] showed that each additional day of antibiotic exposure was associated with 4.7% increased odds in the composite outcome of mortality or major morbidity among VLBW infants in Canadian NICUs. As treatment duration for neonatal pneumonia is based on sparse evidence [17], and short course antibiotic therapy appears to be effective in children and adults for community-acquired pneumonia [18,19], we sought to implement a similar strategy among NICU infants with pneumonia.
In August 2019, the Neonatal Antimicrobial Stewardship Program (NEO-ASP) at Nationwide Children's Hospital, Columbus, OH and its seven affiliated NICUs recommended a 5-day antibiotic course with a time-out for treatment of pneumonia without concomitant bacteremia or meningitis. Education was provided to all heath care professionals with prospective auditing and feedback performed. The objective of this study was to determine the adherence to the guidance and its safety outcomes among infants in the seven NICUs.

METHODS
This prospective surveillance study included all infants who received antibiotic therapy for pneumonia at seven Nationwide Children's Hospital NICUs (1, level 4, outborn; 5, level 3 and 1, level 2, inborn; 263 total beds) from August 2020 to December 2021 [20]. Following the NEO-ASP recommendation for a 5-day antibiotic treatment course for pneumonia, NICU pharmacists monitored its implementation on daily patient rounds. The antibiotic order for pneumonia in the electronic health record was set with an automatic stop at 5 days with a "time-out." The antibiotic time-out prompted a reassessment of the continuing need and choice of antibiotics. If the health care provider decided to extend antibiotic therapy beyond 5 days, either the previous order was updated or a new order was created with a new stop date determined by the treatment team. In addition, the study pharmacists (PP, JKM) with a neonatologist and pediatric infectious diseases specialist (PJS) conducted NEO-ASP rounds every Monday/ Thursday and recorded information on all infants who received antibiotic therapy for pneumonia and discussed the NEO-ASP recommendation with the infant's treatment team.
Infants were diagnosed with pneumonia by the attending neonatologist based on clinical and radiographic abnormalities. These included escalation of respiratory support, lack of improvement after surfactant treatment, increased tracheal secretions, apnea, bradycardia, frequent desaturation episodes, temperature instability, abnormal white blood cell count, elevated c-reactive protein or procalcitonin levels, or an abnormal chest radiograph. Bacterial culture of endotracheal aspirate among ventilated infants was not routinely recommended, performed, or required for a diagnosis of pneumonia. Infants with positive bacterial cultures of blood, urine, or cerebrospinal fluid were excluded. Infants also were excluded if they tested positive for a respiratory virus by polymerase chain reaction. In general, empirical antibiotic therapy for possible early-onset sepsis (<72 h of age) was ampicillin and gentamicin. For empirical therapy for possible late-onset sepsis (>72 h of age), infants received gentamicin and either nafcillin or vancomycin based on previously published guidance [21]. Infants without a history of colonization or infection with methicillinresistant Staphylococcus aureus (MRSA) received empiric treatment with nafcillin rather than vancomycin irrespective of clinical or laboratory parameters such as the presence of a central venous catheter, hypotension, or abnormal acute phase reactants. Infants who were known to be colonized with MRSA at the time of the sepsis evaluation received empirical vancomycin treatment. Subsequent antibiotic therapy was based on results of bacterial cultures (e.g., tracheal aspirate) if obtained as part of the sepsis evaluation or clinical response to initial antibiotic coverage. Pertinent demographic, clinical, laboratory, and microbiologic data were obtained from the electronic health records and entered into REDCap for subsequent analysis. Information included chronological age, gestational age, birth weight, sex, bacterial culture results, duration and dates of antibiotic therapy, and mortality. Chorioamnionitis was present if documented on the maternal or neonatal medical record. Bronchopulmonary dysplasia was defined as any grade of respiratory support provided to a preterm infant (<32 weeks' gestational age) at 36 weeks' postmenstrual age as defined in the executive summary of a workshop sponsored by the Eunice Kennedy Shriver National Institute of Child Health and Human Development [22,23]. Presence of tracheostomy was recorded as was the respiratory support at the time of the pneumonia diagnosis.
The start and stop dates for all antibiotics provided to infants with pneumonia were recorded, including if provided intravenously or by the oral route. The length of therapy (LOT) was defined as any calendar day that the infant received ≥1 antibiotic. Definitive antibiotic therapy was defined as the actual treatment course that was decided upon by the treatment team based upon clinical course and tracheal aspirate cultures, if performed. Safety outcomes were defined a priori by re-initiation of antibiotic therapy within 14 days after discontinuation of the definitive treatment course and overall and sepsis-related mortality while in the NICU. The clinical and outcome characteristics of infants who received five versus a longer course of antibiotic therapy were compared. The study was approved by the Institutional Review Board of Nationwide Children's Hospital with waiver of informed consent.

Statistical analyses
Descriptive data were analyzed as median with interquartile range (IQR) with Chi-square test used to compare safety outcomes. Comparison of median duration of definitive antibiotic therapy was performed using the Mann-Whitney U test. Data management and descriptive statistics were conducted using Microsoft Excel and R (R Core Team 2019). Analyses were performed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY), R (R Core Team 2019), and Stata 16.
There was no difference in the median birth weight and gestational age of infants who received who received ≤5 days vs. >5 days of antibiotics for early-onset pneumonia (3250 g vs. The safety outcomes were defined as antibiotic restart within 14 days of discontinuation of definitive antibiotic treatment as well as mortality ( Table 2). Of the 136 episodes of pneumonia, 22 (16%) episodes had antibiotics restarted within 14 days. In early onset pneumonia, antibiotic therapy was restarted in 4% (3/75) of episodes and all 3 were among infants who received 5 days of therapy and none among the 7 episodes treated with 7 or 8 days of antibiotics. Among late-onset episodes, 19 were restarted within 14 days, with the majority (n = 16) restarted after 5 days of definitive antibiotic therapy. Overall, pneumonia was the reason for antibiotic restart in 3 (14%) of the 22 episodes with NEC being the most common reason (18%, 4/22). The overall mortality was 5% (7/128), with 5 of the 7 deaths assessed as sepsis-related (Table 3; patients #1, 2, 3, 4, 7). Three of the deaths occurred within 14 days of antibiotic discontinuation and all were due to blood culture-proven sepsis due to gentamicin-susceptible Pseudomonas aeruginosa. One of the infants (Table 3, patient #1) had received 7 days of antibiotic therapy for previous pneumonia when the endotracheal aspirate had grown Pseudomonas aeruginosa. The endotracheal aspirate from patient #3 (Table 3) had grown Staphylococcus epidermidis with the previous diagnosis of pneumonia. Patient #2 (Table 3) did not have a culture performed of the endotracheal aspirate. There was no significant difference in the combined safety outcomes of antibiotic restart or mortality between ≤5 (18%; 26/146) and 6-20 days (16%; 3/19) of definitive antibiotic therapy (p = 0.83).

DISCUSSION
In response to concerns for potential adverse effects of prolonged antibiotic therapy provided to infants in the NICU, our NEO-ASP Program recommended a 5-day antibiotic course with a time-out for treatment of "blood culture-negative" pneumonia. The adherence to the guidance was high as 88% (119/136) of pneumonia episodes were treated for 5 days. Moreover, adherence was not associated with either lower median birth weight or gestational age for both early-and-late onset pneumonia groups. Importantly, there was no apparent safety signal for the 5-day treatment regimen with respect to re-initiation of antibiotic therapy within 14 days of its discontinuation or sepsisrelated death during the hospitalization. Only 3 (3%) of the 119 episodes of pneumonia treated for 5 days had antibiotics restarted for pneumonia. In addition, all 3 infants who died within 14 days of discontinuation of antibiotic therapy had P. aeruginosa sepsis, one of whom had received 7 days of antibiotic treatment (Tables 1  and 2). The fourth sepsis-related death occurred 17 days after antibiotic discontinuation in a preterm infant who had Klebsiella pneumoniae sepsis as a complication of NEC (Table 2, infant #4).
Treatment duration for pneumonia in pediatric patients as well as in adults has been decreasing in recent years [19,24]. For community-acquired pneumonia in non-hospitalized pediatric patients aged 6 months to 10 years, Pernica et al. [25] showed that 5 days of high dose amoxicillin therapy was comparable to 10 days in previously healthy children. In a systematic review of antibiotic duration for community-acquired pneumonia in outpatient children ≥6 months of age in high-income countries, a short antibiotic treatment duration of 3-5 days was equally effective and safe compared to 7-10 days [19]. The need for retreatment was 8.3% in the short-course and 7% in the long-course group. Similar results have been seen in adults with antibiotic courses as short as 5 days recommended if the patient achieves stability [26]. However, if the causative organism is suspected or proven methicillin-resistant S. aureus or P. aeruginosa, antibiotic duration is 7 days as per recent hospital-acquired pneumonia and ventilator-associated pneumonia guidelines [18,27,28]. Similar guidance has not been extended to high-risk infants in the NICU although the occurrence of fulminant P. aeruginosa sepsis without pneumonia within two weeks of antibiotic discontinuation in three preterm infants may warrant a similar strategy. In our cohort of NICU infants who received a 5-day course of therapy, re-initiation of antibiotic therapy occurred in 4% (3/68) and 31% (16/51) of pneumonia episodes in early and late-onset episodes, respectively. However, re-initiation of antibiotics for a new diagnosis or relapse of pneumonia was lower at 1% (1/68) and 4% (2/51), respectively, suggesting that a short course may be adequate for most cases of blood culture-negative pneumonia.
In late preterm and term newborns in the normal newborn nursery, Engle et al. [29,30] showed that 4 days of antibiotic Klebsiella spp.    therapy provided to 35 infants was comparable to 7 days (n = 38) for treatment of pneumonia if the infants were asymptomatic after 48 h of antibiotic therapy. All of the infants had sterile blood cultures and had not required supplemental oxygen for >8 h. None had received nasal CPAP or mechanical ventilation. Two infants in the 4-day group developed tachypnea in the 24 h after antibiotic discontinuation and antibiotic therapy was re-initiated for 5 days in one infant for a failure rate of 3% (1/35). Pneumonia is a frequent reason for prolonged (>5 days) antibiotic administration despite sterile cultures in the NICU [1]. In response to the marked variability of anywhere from 5 to 14 days of antibiotic therapy for pneumonia, Cantey et al. [17] implemented a 5-day antibiotic treatment course among infants in a Level 3 NICU and, utilizing prospective surveillance, found that no infant required additional antibiotics for the same indication within 14 days of completing treatment. Data from this study formed the basis for the NEO-ASP recommendation of a short course of antibiotic therapy for culture-negative pneumonia.
In our study, we were able to only evaluate safety outcomes as the majority of infants received a 5-day course of antibiotic treatment. Overall, 16% of the pneumonia episodes had antibiotics restarted within 14 days but only 3 (3%) of the 119 episodes were restarted for a diagnosis of pneumonia suggesting that the time-out evaluation at 5 days was for the most part successful and in line with past studies. Overall, 7 infants died during the NICU hospitalization, and all except one had received a 5-day course of antibiotics for pneumonia ( Table 2). All 7 had underlying co-morbidities that necessitated NICU hospitalization, and one of the 3 infants who had fulminant P. aeruginosa sepsis had received 7 days of antibiotics. Future studies are needed to ascertain optimal treatment duration for pneumonia due to Pseudomonas species.
Limitations of the study include its observational nature that documented possible adverse outcomes following implementation of a NEO-ASP guidance of treating blood culture-negative pneumonia with 5 days of antibiotic therapy with a time-out. The small number of infants who received more than 5 days of definitive antibiotic therapy made direct correlations difficult to ascertain. Certainly, the input from the NEO-ASP during the study period informed practice and was a driving force responsible for its implementation and adherence. The diagnosis of pneumonia, however, remains difficult to establish so it is possible that some infants may have had other noninfectious reasons for respiratory deterioration [31]. Specific criteria for establishing a diagnosis of pneumonia, including performance of endotracheal aspirate cultures, was not mandated, but the current design does give the study "real world" applicability. In addition, a systematic review of all autopsies performed during the study period to check on any missed diagnosis of pneumonia as the cause of death was not performed. However, all autopsies performed on infants who received antibiotic therapy for pneumonia were reviewed and are included in the study analyses. Respiratory viral testing was not performed routinely and mitigation strategies involving the COVID-19 pandemic may have influenced the occurrence of pneumonia. In addition, the majority of infants had a birth weight >2500 g (62%) and gestational age ≥34 weeks (67% ; Table 1). Finally, all of the seven NICUs that participated in the study belonged to one NCH neonatal network and, therefore, the safety of the shortened course of antibiotic therapy should be studied in randomized clinical trials involving diverse demographic and geographic regions.

CONCLUSION
In conclusion, in our NICU network, there was excellent adherence (88%) to the NEO-ASP recommendation of a 5-day definitive antibiotic treatment course for "blood culture-negative" pneumonia. The intervention seemed safe as only 3% (3/119) of pneumonia episodes relapsed in the 14 days after a 5 day definitive antibiotic treatment course. However, additional studies are needed across other demographic and geographic centers [32,33]. Ongoing education, prospective auditing and feedback, and collaboration among neonatal subspecialties helped inform antimicrobial prescribing in the NICU and allowed the successful incorporation of a short antibiotic course with "time-out" into our NEO-ASP bundle.

DATA AVAILABILITY
The de-identified dataset generated from the study is available from the corresponding author on reasonable request and following approval by the Institutional Review Board of Nationwide Children's Hospital.

AUTHOR CONTRIBUTIONS
ZSL collected and analyzed the data, assisted with the analyses, wrote the first draft of the manuscript, and reviewed and revised the manuscript. PP conceptualized and designed the study, designed the data collection instruments, collected data, assisted with the analyses, and reviewed and revised the manuscript. JKM conceptualized and designed the study, collected data, and reviewed and revised the manuscript. MJK assisted with the analyses, reviewed and revised the manuscript. CdAR collected and helped to analyze the data, reviewed and revised the manuscript, and approved the final manuscript as submitted. NOW conceptualized and designed the study, reviewed and revised the manuscript. RRM conceptualized and designed the study, reviewed and revised the manuscript. RM conceptualized and designed the study, reviewed and revised the manuscript. ART conceptualized and designed the study, reviewed and revised the manuscript. PJS conceptualized and designed the study, coordinated and supervised data collection, analyzed the data, and critically reviewed and revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING
ZSL received a Pediatric Infectious Diseases Society Summer Research Scholar Award (SUMMERS) for his work on this study.