Early neonatal pyloric stenosis after exposure to maternal macrolide therapy

Introduction

Infantile hypertrophic pyloric stenosis (IHPS) is due to hypertrophy of the muscular layer of the pylorus causing gastric outlet obstruction [1]. Generally, clinical signs of IHPS present between the first 3 to 5 weeks of life [1]. Documented reports of IHPS in the first days of life are rare [2]. A link between the use of macrolide antibiotics and IHPS has been described [1]. We report here a newborn who developed vomiting on the second day of life (DOL), subsequently diagnosed as IHPS whose mother received azithromycin during the second trimester of pregnancy.

Case report

AG was born at full-term, weighed 3724 g and was the firstborn male of a 23 year-old mono-gravid Jewish Sephardic mother. The pregnancy was unremarkable, except for a maternal upper respiratory infection at 17 weeks’ gestation, which was treated with a course of azithromycin (500 mg, once daily, for 3 days). A regular maternal ultrasound performed during the second trimester was interpreted as normal. The baby was born by cesarean section because of pre labor rupture of membranes accompanied by maternal fever. Intravenous penicillin G was administered perinatally to the mother because of suspected chorioamnionitis.

At 6 h of age the baby developed respiratory distress and was transferred to the neonatal intensive care unit (NICU). He was treated immediately with continuous positive airway pressure (CPAP) ventilation with oxygen. After a septic screen was performed, intravenous ampicillin and gentamycin therapy were started and his respiratory status improved quickly. On the second DOL, enteral feeding was initiated and was followed by intermittent non-bilious vomiting. Abdominal X-ray was not diagnostic. On DOL 6 a barium examination of the upper gastrointestinal tract and an abdominal ultrasound were performed indicating pyloric stenosis (Figure 1). Successful pyloromyotomy was performed with no surgical complications and the baby was gradually able to tolerate full enteral feeding. Repeated chest radiograph demonstrated persistent air filled lucencies in the right and left lower lobes. A chest CT demonstrated multiple cysts with air-fluid levels, largely isolated to the lower lobes (Figure 2). A polymerase chain reaction (PCR) examination of fluid from an ultrasound-guided aspiration demonstrated Escherichia coli infection. Antibiotic treatment was switched to IV amoxicillin/clavulanic acid. Interestingly, the culture of the placenta also grew E. coli. The infant was discharged on DOL 32 in good condition without fever, vomiting or respiratory distress. An additional 2 weeks of oral amoxicillin/clavulanic acid therapy was recommended at home. At the age of 6 months, the child continues to feel well and a follow-up CT showed normal lungs without remnants of the cysts (Figure 3).

Figure 1:

Abdominal ultrasound demonstrating pyloric stenosis.

Figure 2:

Chest CT demonstrating multiple cysts with air-fluid levels, largely isolated to the lower lobes.

Figure 3:

Chest CT at the age of 6 months demonstrating resolution of all the lungs cysts.

Discussion

We report here a child who developed symptoms of IHPS on the 2^nd DOL. This unique case suggests an association between maternal use of azithromycin as early as the second trimester and intrauterine hypertrophy of the pyloric muscle.

In the above case, the diagnosis was achieved on the 6^th DOL and definitive surgery was performed the next day. IHPS is rarely diagnosed prior to 3 weeks of age (8.4%) [2] with a reported mean age at diagnosis of 40.47±1.25 days [2]. Even in the sub group of those diagnosed prior to 21 days, the mean age at diagnosis was 17.22±1.14 days [2]. The unusually early and severe nature of the IHPS in our case suggests an intrauterine etiology.

IHPS has an increased incidence among males particularly if first born to a young mother [3], risk factors that were all present in our case. However, he had no first degree relatives who developed IHPS – a factor known to increase the risk of IHPS many fold [3]. Regarding ethnicity, the child presented above was of Middle Eastern and North African mixed origin. There was no evidence found for increased incidence from this population but the White:Black ratio in the US has been reported to be 2:1 [3]. The child we presented had a limited number of known risk factors however, these did not fully explain the very early presentation.

Neonatal erythromycin therapy during the first weeks of life has been reported to be associated with the development of IHPS [4]. In a 12-year retrospective study of children of members of the US armed forces, this association was found to be stronger if therapy was administered during the first 2 weeks and was found to be true for erythromycin and azythromycin [5]. Additionally, a potential association has been shown with antenatal third trimester exposure to erythromycin and non-erythromycin macrolides [4]. Cooper et al. found the association between with IHPS to be potentially true only for non-erythromycin macrolides [6]. However, in a retrospective study of infants with a diagnosis of IHPS, Lin et al. found no association with macrolide therapy during the first or third trimester of pregnancy [7]. The prevailing theory of the pathogenesis of IHPS, which also explains its development over weeks, is increased peristalsis of the stomach via various mediators [8]. Erythromycin has motilinomimetic effects on antral smooth muscle function, including the migrating motor complexes causing excessive peristalsis and hypertrophy of the pylorus and this has recently been shown to apply to azithromycin as well [9]. Fetal gastric peristalsis is present as early as between the 12^th and the 14^th week of gestation. Motillin receptors have been shown to exist at 20 weeks gestation in a similar distribution to adults.

There is evidence of transplacental transfer of azithromycin to the fetus in animal models. Multiple dose regimes in monkeys effectively treated ureaplasma infection in amniotic fluid [10] and in sheep, significant tissue deposition was demonstrated after multiple doses were administered [11]. In a trial of a single dose infused into human placentae post partum, there was a low transfer rate of azenil to the placenta [12]. This was found with the animal models and only increased after multiple doses were administered.

Hence, we speculate that excessive intrauterine motility may cause hypertrophy of the pylorus early in utero in a similar manner to postnatal exposure. Furthermore, we postulate that the respiratory distress of the child few hours after his birth was probably linked to a pulmonary aspiration predisposed by the IHPS. The E. coli found in the amniotic fluid and documented by the culture from the pulmonary cyst supports our hypothesis. The complete disappearance of the lung findings in the chest CT at age of 6 months further supports the hypothesis of an infectious process in the lungs which recovered.

It is improbable that the clinical picture could be attributed to the maternal penicillin or the antibiotics given after the septic screen. Firstly, no association between non-macrolide antibiotics and IHPS has been reported [1] and secondly there was insufficient time after the penicillin was initiated for development of the significant pyloric hypertrophy required to obstruct gastric outflow.

To the best of our knowledge this is the first time that a possible association between maternal macrolide therapy in the second trimester and IHPS has been reported. When assessing non-bilious vomiting in newborns, a history of exposure to macrolides from as early as the second trimester should alert clinicians to the possibility of IHPS.