An open-label, randomized controlled parallel intervention trial was conducted on 40 late preterm newborn infants. The study was carried out at the neonatology units of Mansoura University Children’s Hospital and Mansoura Insurance Hospital over 10 months from July 2020 to April 2021. Mansoura Faculty of Medicine institutional research board (IRB) approved the study and fully informed written consent was obtained from the neonate’s parents or guardian before enrolment in the study. The study was registered in the clinical trials database (clinicaltrials.gov, ID: NCT04474327).

The study was conducted on late preterm infants at a gestational age between 340/7 weeks and 366/7 weeks with a birth weight of more than 1.5 kg with clinical evidence of neonatal sepsis. Neonatal sepsis was defined in the presence of one of the following signs; fever (> 38 °C, rectal), hypothermia (< 36 °C, rectal), tachycardia (heart rate (HR) > 180 beats/min) or bradycardia (HR < 100 beats/min), apnea, lethargy, feeding problems, mottled skin, convulsions, hypotonia plus at least 2 of the following laboratory findings; leukopenia (white blood cells count (WBCs) < 5000 mm3), leukocytosis (WBCs > 20,000 mm3), thrombocytopenia (< 100,000 mm3), serum C-reactive protein (CRP) > 15 mg/L, fibrinogen > 150 mg/dL, or metabolic acidosis (base excess of ≤ 7 mmol/L); with (culture-confirmed) or without (culture-negative) a positive blood culture result [18, 19]. Infants presented initially with septic shock, multi-organ dysfunction syndrome (MODS), disseminated intravascular coagulopathy, or feeding intolerance requiring to be nil by mouth were excluded from the study. Besides, we did not include infants with major congenital malformations, chromosomal aberrations, or postoperative patients.

Infants were divided into the Montelukast group and the routine care group. Both groups received the same protocol for the treatment of sepsis including antibiotics and supportive measures according to the policy of neonatal units and patients’ needs. Infants in both groups were started on empirical antibiotics with Ampicillin at a dose of 50 mg/kg/12 h combined with Aminoglycoside (Gentamycin) at a dose of 4.5 mg/kg/36 h for infants with suspected early-onset sepsis while for late-onset sepsis, Cloxacillin (50 mg/kg/dose, every 8 h) was used instead of Ampicillin. After that, antibiotics were changed as appropriate according to the results of culture and antibiotic sensitivity. The use of positive pressure ventilation and its duration, inotropic drug usage and other supportive measures were guided and monitored by the treating physicians according to the policy of the neonatal units and patients' needs.

The Montelukast group received, in addition to the routine care, Montelukast sodium (Singulair, Merck Sharp and Dohme Corp.) for 10 days in a dose guided by the infant's birth weight as follows (infant’s weight 1.5 to 2 kg received 1.5 mg once daily whereas greater than 2 kg received 2 mg). The given dose was calculated according to Kim and coauthors [12]. Montelukast was given by oral route by dissolving one sachet containing four mg of the Montelukast sodium in four ml milk to get a concentration of 1 mg/1 ml.

Infants were discharged from the neonatal intensive care unit (NICU) if they fulfilled the following criteria: adequate oral feeding sufficient to support appropriate growth and gaining weight, the ability to maintain normal body temperature in a home environment and sufficient mature respiratory control that allow safe discharge in addition to acceptable bilirubin level and free from infection determined as normal WBCs count with negative blood culture results and declining CRP reaching < 10 mg/L.

For the policy of discontinuation of respiratory support; infants were weaned off mechanical ventilation if they fulfilled the following NICU guidelines: Spontaneous respiratory effort, Gag or cough with suctioning, Satisfactory blood gases (Power of the hydrogen ion (PH) more than 7.25, Partial pressure of carbon dioxide (pCO2) less than 60 mmHg, and base deficit less than 8 meq/L) on a mean airway pressure less than 8 cmH2O and frequency less than 30 breath /minute and saturation more than 88% on a fraction of inspired oxygen (FiO2) less than 30% in the preceding 24 h besides approval of the attending physician.

Infants were weaned off continuous positive airway pressure (CPAP) if they fulfilled the following NICU guidelines: Pressure of 3 to 6 cm H2O for 24 h, FiO2 less than 30% to keep target saturation in the preceding 24 h, respiratory rate less than 60 breath/minute, no single apnea requiring bagging in the preceding 24 h, no more than 6 apneas requiring stimulation in the preceding 24 h, satisfactory blood gases (PH > 7.25, pCO2 < 60 mmHg, and base deficit < 8 meq/L) and infant tolerates time off CPAP during nursing care in addition to approval of attending physician.

Infants were weaned off oxygen therapy if they fulfilled the following NICU guidelines: Infant's saturation remained above 91% in less than 30% FiO2 for 24 to 48 h or the infant could tolerate a trial of discontinuing oxygen therapy to 21% FiO2 for 1 h.

The primary outcome was serum TNF Alpha level, an inflammatory marker, at day 10 after receiving therapy. Serum TNF alpha level was measured by the double-antibody sandwich Enzyme-linked immunosorbent assay (ELISA) technique (TNFa ELISA kits, Sunred PeloBiotech GmbH, Germany) [20]. Secondary outcome measures included; needs and duration for invasive mechanical ventilation, needs, and duration of non-invasive ventilation, needs, and duration of inotropic support, the total duration of NICU admission and serum CRP level at day 10 after receiving therapy. Pediatric Logistic Organ Dysfunction (PELOD) score was performed by the treating physician on admission and on day 10 in order to follow the improvement or deterioration of the studied patients in both groups. Furthermore, delta PELOD (score at day 10–score on admission) was calculated. PELOD score is used in neonates to calculate sepsis-induced organ dysfunction [21]. The score assesses cardiovascular function (heart rate and systolic blood pressure), neurologic function (Glasgow coma scale and pupillary reaction), hepatic function (aspartate aminotransferase (AST) and international normalized ratio (INR)), pulmonary function (partial pressure of oxygen in arterial blood (PaO2)/FiO2, PCO2 and whether the patient is on mechanical ventilation), hematologic function (WBCs and platelet count), and renal function (serum creatinine). Patient survival was recorded during the duration of NICU admission. In our research, adverse effects of Montelukast therapy such as diarrhea, vomiting, fever, cough, conjunctivitis, and skin signs (rash, eczema, bruises and erythematous lesions) [22] were observed during hospital admission.

Clinical data such as demographic data, cause of admission, site of infection (bacteremia, pneumonia, meningitis or septic arthritis), vital signs, activity, feeding tolerance, respiratory symptoms, duration of respiratory support, and use of inotropes were registered. A venous sample was withdrawn for a complete blood picture, serum CRP, TNF alpha, creatinine, liver enzymes, INR, and blood culture. Blood sampling in the studied groups was performed twice: First, at the start of the study, once sepsis was suspected, and the second sample was after 10 days of treatment. A lumbar puncture was performed when meningitis was suspected.

Infants were assigned randomly to treatment groups using an internet-based random table technique with a block size of four. Cards in sequentially numbered, opaque, sealed envelopes were kept in NICU. A designated nurse who was not involved in the study was responsible for the randomization of selected infants.

Sample size calculation was based on the mean TNF level between the studied groups (routine care group and Montelukast group). A pilot study was carried out on 10 cases (5 in each group) to calculate sample size. During this stage of the clinical trial, none of the participants was withdrawn; moreover, none had intolerable adverse events. Using the G*power calculator to calculate the difference between 2 means using t test with an effect size of 0.894, 2-tailed, with α error = 0.05 and a power = 80%, mean ± standard deviation (SD) of mean TNF level (110 ± 30 and 90.73 ± 10), the total calculated sample size was 40 infants (20 in each group).

Statistical Package for the Social Sciences (SPSS) for Windows (SPSS, Inc, an IBM Company, Chicago, IL, USA), version 20 was used for statistical analysis of the collected data. Shapiro–Wilk test was used to check the normality of data distribution. Quantitative variables were expressed as mean ± SD or median (minimum–maximum) as appropriate while categorical variables were expressed as frequency and percentage. Independent sample t test and Mann–Whitney U tests were used for inter-group comparison of parametric and non-parametric continuous data. Chi-square /Fisher´s Exact tests were used for inter-group comparison of nominal data using the crosstabs function. Probability (p value) < 0.05 was considered statistically significant.