Background
In adults, potentially life-threatening esophageal perforations and leakages can occur spontaneously as Boerhaave’s syndrome, as late manifestation of advanced esophagitis, consequence of trauma or foreign body ingestion or can be frequently encountered as complication of surgical or endoscopic interventions. Esophageal leakages are, however, not only a domain of adult medicine but can be also seen in infants with prematurely born infants affected predominantly [
1‐
8]. While there have been case reports of spontaneous esophageal perforations, they are mainly iatrogenic following procedures like resuscitation, difficult intubation or insertion of a nasogastric feeding tube. Many cases can be managed conservatively, i.e. the perforation will heal spontaneously while infants are kept on broad spectrum anti-infective agents and nothing per os (NPO) [
3]. Still, full-thickness perforations of the esophagus are a life-threatening condition and some of the infants will develop mediastinitis and require surgical therapy to close the perforation [
6,
7].
Endoscopic vacuum therapy (EVT) has become a standard treatment of perforations or anastomotic dehiscences (insufficiencies), mainly in the upper gastrointestinal (GI) tract or the rectum, since it was first described by Weidenhagen 2003 [
8‐
16]. This method is derived from vacuum-assisted closure (VAC) therapy of external wounds, as introduced by Argenta and Morykwas in 1997 [
17]. EVT within the GI tract involves a polyurethane sponge or a special open-pore film, connected to a suction catheter, which is endoscopically placed either into the lumen of the GI tract, thus covering the perforation site (intraluminal EVT), or inserted into the perforation site itself, i.e. outside of the mucosa-lined space (intracavitary EVT) [
10,
18]. The principle of EVT is to create a negatively pressurized compartment at the perforation site providing local drainage of fluids and inducing granulation of the infected wound area, eventually resulting in defect closure [
16]. The applied negative pressure (i.e. the “vacuum”) causes a collapse of the surrounding tissue thus sealing it and impeding pressure equalization. While EVT has become a widely accepted first-line therapy for esophageal perforations/leakages of different origin in adults, experience on EVT in infants is still scarce. In 2018, a first case of EVT in an 8 year-old infant was reported by Fraga et al. [
19]. Also in 2018, Manfredi et al. reported on a series of 17 children with esophageal atresia who were treated with EVT for esophageal perforations [
7]. The success rate was 88%. Less promising results have been lately reported by Ritz et al. [
20]: In all five infants, EVT alone was not sufficient and required additional treatment (surgery or insertion of a suction catheter).
In this present retrospective case series, we report on four very young infants who were successfully treated with EVT in our endoscopy unit (7, 32, 35 and 159 days old when diagnosed with perforation/leakage). In one of these infants, prematurely born in the 31st week of pregnancy and weighing 980 g only, EVT was begun on day 24 of life. To our best knowledge, this is the youngest patient up to date in whom EVT was successfully performed.
Discussion
Esophageal perforations, irrespective of their etiology, pose a vital threat to patients [
2,
3,
6]. There has been much controversy about the optimal therapeutic approach in case of perforation in infants [
7,
19]. While many cases can be managed conservatively, i.e. administration of broad-spectrum antibiotics and keeping the infant on NPO, some cases will require high-risk surgical interventions associated with high mortality rates [
2,
3,
6]. EVT on the other hand has become a standard procedure for treatment of esophageal perforations and leakages in adults and is a low-risk and easy-to-perform alternative to (thoracic) surgery [
24‐
28]. In the present case series, we confirm the feasibility of EVT in infants. The procedure was adopted from adult medicine and we found that similar EVT settings have to be applied to achieve defect closure, i.e. similar negative pressure values and EVT exchange frequencies. Thus, our data confirm previous reports from Fraga et al. and Manfredi et al. for cases of newborns and very young infants [
7,
19]. Lately, Ritz et. al reported on five children with esophageal perforations (median age 3.4 years, the youngest child being 7 months old) [
20]. In all five children, EVT helped to significantly reduce local and systemic inflammation and facilitated further treatment, although EVT alone was not sufficient to achieve complete healing of the defect: while 1 child required surgery, the other 4 children healed under subsequent therapy with a suction catheter. One possible explanation could be that Ritz et al. used longer intervals between EVT exchanges and applied lower suction values (100 cmH
2O being equal to 73.6 mmHg). In our experience (gained from infant “1”), suction values can be stepwise increased if the healing process is lacking and the EVT system can be exchanged more frequently. Also, in the report by Ritz et al. completion of defect closure was achieved by a suction catheter (100 cmH
2O) in 4 out of 5 children, which might be considered as another form of vacuum therapy. Like in the report by Ritz et al., most of the children in the reports by Fraga et al. and Manfredi et al. were older (median age 24 months, youngest child 3 months) and Manfredi et al. reported on a distinct group of children suffering from esophageal atresia. The novelty of our case series is doubtless the very young age (24, 32, 35 and 161 days of life on initiation of EVT). The youngest infant was prematurely born in the 31st week of pregnancy with a birth weight of 980 g only: on initiation of EVT, the infant was 24 days old weighing 1500 g, the esophagus being too narrow to be passed with a nasal gastroscope. For this reason, initial EVT film placement had to be done fluoroscopically by wire guidance. To our best knowledge this technical approach has not been yet described in EVT.
There are, however, some differences to EVT in adults. Firstly, the EVT catheter can be placed perorally and not transnasally, which in our opinion facilitates EVT placement and EVT exchanges and prevents contact lesions within the nose. In adults, peroral EVT catheter placement will usually not be tolerated by the patient. Secondly, the infant’s esophagus has a much smaller diameter, so EVT placement can be more cumbersome and requires the use of a nasal gastroscope or even fluoroscopic wire guidance (as described in case “1”). In all four cases, we used thin 8 Fr suction catheters and began therapy with EVT films (instead of sponges) because EVT films can be shaped much thinner. Thus, owing to the availability of thin open-pore drainage films, the use of which to our best knowledge has not been yet described in infants, EVT becomes technically feasible even in very young infants with small esophageal diameters. However, from our own and other authors’ experience in adults, EVT sponges adhere more strongly to the tissue and for this reason exert higher debriding properties than EVT films, inducing more granulation and eliminating necroses more efficiently [
29]. In older wounds containing a lot of debris (like in cases “1” and “2” described here) it is therefore desirable to switch to EVT sponges if technically possible. In adults, EVT placement and exchanges are performed under conscious sedation, but patients usually stay awake during ongoing EVT. Similarly, EVT in infants does not seem to require general anesthesia on principle, and lowest sedation levels possible should be rather sought. Most infants in our cases (“1”, “3” and “4”) were kept on general anesthesia for the time of ongoing EVT, however, in infants “3” and “4” general anesthesia was maintained because of their complex diseases including chronic respiratory insufficiency. In infant “1” the combination of extreme prematurity, little birth weight, status post thoracic surgery and the imminent dislocation risk of the EVT catheter added up to the decision to maintain general anesthesia. Finally, enteral nutrition is an important issue in ongoing EVT, especially given that EVT can last up to several weeks. While patients must be NPO because oral nutrition would contaminate both the perforation site and the film/sponge, thus making EVT ineffective, adults in some cases can be nourished via a nasoduodenal or nasojejunale tube. However, placing a second line for nutritional purposes will be very difficult in an infantile esophagus due to its narrow size or might interfere with the EVT system. Double-lumen feeding tubes on the other hand could help to shape individual EVT film systems that could allow for simultaneous enteral nutrition in future cases, as described by Loske et al. [
30,
31].
Owing to its retrospective nature, this report of four more or less independent cases leaves some questions unanswered and makes it difficult to generalize our promising results. Existing data suggests that most infants recover spontaneously after esophageal perforation [
3]. Thus, it can be speculated that some of our cases presented here might have also recovered without EVT. On the other hand, cases “1” and “2” describe post-operative situations where an esophageal leakage remained despite or due to surgical therapy, which is why recovery without further specific therapy seemed very unlikely. In cases “3” and “4” the perforation site was very extensive in size (6 and 4 cm) but owing probably to the rapid intervention immediately after perforation, complete closure could be achieved after 7 and 11 days of continuous EVT. Thus, with availability of EVT as a minor invasive and easy-to-perform procedure with a low-risk profile, a watch-and-wait approach after esophageal perforation might be questioned. Though infant “2” developed an anastomotic stenosis after EVT, it can be argued that the stenosis was a consequence of ischemia and local inflammation at the insufficient anastomosis rather than EVT itself. In infant “3”, an ulcer developed in the gastric antrum during ongoing EVT, most probably caused by contact with the PEG tube. Such contact could have been aggravated due to continuous deflation of the stomach, so a causative association with EVT cannot be excluded. Apart from that, no other EVT-associated adverse events were recorded.
Finally, we did not compare EVT with other endoscopic closure techniques, like clip or stent application or insertion of a simple suction catheter. However, several meta-analyses found that EVT in comparison to the use of self-expandable metal stents (SEMS) for esophageal leaks had higher closure rates, shorter treatment times and lower mortality rates [
27,
28]. Manfredi et al. in their study also compared EVT with stent placement in infants for esophageal perforation due to atresia therapy: they found EVT to be more beneficial than stent placement [
7].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.