Azithromycin therapy for prevention of chronic lung disease of prematurity (AZTEC): a randomised placebo-controlled trial

Systematic reviews have reported conflicting evidence to confirm if macrolides reduce rates of chronic lung disease of prematurity in at-risk preterm-born infants, including in those colonised with pulmonary Ureaplasma spp. Since an adequately powered trial has been lacking, we conducted a double-blind, randomised, placebo-controlled trial to assess if the macrolide azithromycin improved survival without the development of physiologically defined moderate or severe chronic lung disease of prematurity in infants born at < 30 weeks' gestation.

Infants recruited from 30 neonatal units (median gestational age 27.0 weeks, interquartile range 25.3-28.6) requiring respiratory support within 72 hours of birth were randomised to intravenous azithromycin 20 mg/kg/day for 3 days followed by 10 mg/kg for 7 days or to placebo. Primary outcome was survival without development of physiologically defined moderate/severe chronic lung disease of prematurity at 36 weeks' postmenstrual age. A total of 796 infants were required to detect 12% improvement in survival without development of moderate or severe chronic lung disease of prematurity, including 10% dropout, with two-sided α-level of 5% and 90% power. The primary outcome was analysed using three-level logistic regression to account for clustering of multiple births and participants within centres and was adjusted for gestational age as a fixed effect. Secondary outcomes included death, chronic lung disease of prematurity severity, treatment interaction with Ureaplasma spp. colonisation, days of invasive and days of non-invasive respiratory support, treatment for nosocomial infections, treated patent ductus arteriosus, severe intraventricular haemorrhage, necrotising enterocolitis, treated retinopathy of prematurity and emergence of azithromycin resistance in stool and respiratory samples. Quantitative polymerase chain reaction identified respiratory Ureaplasma spp. and antibiotic resistance genes. Safety was also monitored.

After three withdrawals, 796 randomised infants were included in the final analyses. Survivors without physiologically defined moderate/severe chronic lung disease of prematurity were: 166/394 (42.1%) and 179/402 (44.5%) in the intervention and placebo groups, respectively (adjusted odds ratio 0.84; 95% confidence interval 0.55 to 1.29; p = 0.43). Secondary outcomes were not significantly different between the treatment groups, except for treated retinopathy of prematurity in survivors (3.5% vs. 7.4%, azithromycin vs. placebo; odds ratio: 0.42, 95% confidence interval 0.18 to 0.98). Ureaplasma spp. colonisation did not influence treatment effect. No significant serious adverse effects were reported. From 1108 (n = 541 azithromycin, n = 567 placebo) respiratory aspirates and 709 stool samples from 348 infants, erm(C) and msr(A) were the most prevalent macrolide-resistance genes, but erm(C) increased with azithromycin treatment in both sample types (11% at baseline, 16% at day 14 in respiratory samples; 0% at baseline, 69% at day 14 in stool samples).

Prophylactic use of azithromycin did not improve survival without development of physiologically defined chronic lung disease of prematurity regardless of Ureaplasma spp. colonisation. Thus, it cannot be recommended in clinical practice. Since preterm-born infants are exposed to a range of antibiotics, in addition to the trial azithromycin, judicious use of antibiotics is required, given the emergence of multiresistant bacteria in this vulnerable group of infants.

Follow-up at ages 1 and 2 years will assess the medium-term effects. Investigating whether treatment modulated proinflammatory cytokine concentrations, including whether this was more prevalent in the Ureaplasma spp. colonised or non-colonised group, will be crucial to providing further assurances to the clinical community.

Limitations include the (limited) missed oxygen reduction tests, inadequate collection of respiratory support data and lower-than-anticipated baseline sampling.

This synopsis presents independent research funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme as award number 16/111/106.