Subglottic stenosis (SGS) is defined as narrowing of the subglottic airway to <3.5 mm in premature infants and 4.0 mm in term neonates. It can be caused by neoplasms such as hemangioma (Fig. 4), but in the premature infant, it is most often iatrogenic or congenital.

The vast majority (~95 %) of cases are acquired due to airway trauma, most commonly endotracheal intubation. SGS without a history of airway manipulation is considered congenital. The proportion of SGS that is congenital is likely underestimated because of the frequency of intubation before formal endoscopy in the setting of respiratory distress.

Congenital SGS can be due to cartilaginous or membranous stenosis. Congenital cartilaginous SGS can be due to a flattened (short anterior/posterior diameter), elliptical (shortened lateral diameter) (Fig. 5), or a “trapped” first tracheal ring that telescopes into the subglottis, narrowing the airway. Membranous stenosis results from hypertrophy of the submucosal tissues or mucous glands and appears soft on rigid endoscopy (Fig. 6). Congenital SGS is more likely to be amenable to observation and avoidance of surgical intervention. Because it does not result from an inflammatory process, it is also more likely to be successfully treated if surgical intervention becomes necessary.

The emergence and rapid advance of the specialty of neonatology since the 1960s has allowed the survival of premature infants with their associated pulmonary diseases of prematurity. Along with the increased survival of extreme premature infants, the incidence of prolonged endotracheal intubation may increase, causing ongoing prevalence despite the measures taken to minimize risks. As the understanding of the pathophysiology of subglottic/laryngotracheal stenosis has improved, techniques designed to minimize tissue damage have been developed, and the incidence of SGS associated with endotracheal intubation has decreased.

The true incidence of SGS specifically in premature infants is difficult to quantify. Available series include all intubated neonates and fail to specify whether they are premature or term. Reported incidence of SGS in intubated neonates has decreased from 4 % in the early 1980s to 0.63 % in more recent series [20].

Several factors combine to make to the level of the cricoid cartilage in the subglottis the site of the majority of instances of acquired LTS. The cricoid is the only complete cartilaginous ring in the normal airway, thus eliminating the ability of the airway to dilate to reduce pressure on the mucosa and submucosa. Additionally, the cricoid is the narrowest portion of the infant airway, creating the greatest levels of hydrostatic pressure along a fixed-diameter endotracheal tube. Hydrostatic pressure exceeding capillary pressure creates relative ischemia and leads to mucosal and submucosal damage. Reduced thickness of submucosal tissue at the level of the cricoid also leads to early exposure of the perichondrium and subsequent increase in inflammation. In postmortem anatomical examinations of low birth weight infants, exposure of the perichondrium at the level of the cricoid is seen in as few as 8 days in many specimens. Depth and size of mucosal scarring increases with the duration of intubation [21].

The same pathophysiological process of scar formation that leads to SGS frustrates attempts to surgically treat the condition. The increased scar deposition after mucosal trauma from intubation can also occur after endoscopic and open surgical treatment.

Tracheostomy is in most cases a successful treatment modality for subglottic stenosis, allowing normal feeding and providing a safe airway. It is associated with significant morbidity and is generally not undertaken in infants under 2000g, as complications increase in smaller patients. The overarching goal of other modalities is to either prevent tracheostomy in the first place or allow decannulation. The majority of studies investigating these techniques define success as allowing for extubation while avoiding tracheostomy or allowing for decannulation.

The technique of dividing the cricoid ring to allow for expansion of the subglottic airway dates to Réthi’s description in 1956, but was popularized by the description of the anterior cricoid split by Cotton and Seid [1]. This technique involves a midline division of the anterior cricoid lamina and first tracheal ring without graft placement. A stent, either an endotracheal tube in the nontracheotomized patient or an Aboulker or Rutter stent or Montgomery T-tube in a patient with previously placed tracheostomy, is left in place temporarily to allow for healing. In well-selected patients without significant comorbidities, the success of this operation is quite good, with extubation/decannulation rates of approximately 70% [22]. Patients with a longer total period of preoperative intubation and with multiple medical or neurological comorbidities fare worse with success rates around 40–50 % [23].

The development by Cotton and colleagues of laryngotracheal reconstruction (LTR) in the 1970s led to improvement in outcomes by combining division of the cricoid with placement of cartilage grafts harvested from the thyroid ala or costal cartilage. When compared with anterior cricoid split, LTR is more successful and reduces complications [24]. Success rates of up to 83 % are reported [25]. Outcomes of this procedure are worse in the setting of stenosis that extends superiorly to the glottis or supraglottis and in patients under 4 kg [26, 27].

Some evidence that single stage procedure (i.e., leaving an endotracheal tube in place postoperatively rather than a tracheostomy with a second procedure to decannulate) is more likely to result in decannulation when controlling for severity of disease and previous surgery [28].

Before the development of open airway reconstruction techniques in the 1970s by Cotton and colleagues, the mainstay of treatment for SGS was serial anterograde rigid dilation (bougienage). This proved unsatisfactory and was rapidly supplanted in many cases by the newer open techniques. Shearing trauma to the mucosa inherent to the technique of bougienage could induce inflammation and fibrosis and worsen stenosis. The adoption of balloon dilation coupled with newer endoscopic techniques has brought endoscopic treatment as a primary modality back to the forefront of treatment for SGS. Balloon dilation offers the benefit of circumferential compression that limits shearing forces and theoretically reduces mucosal damage and resulting in inflammatory response. As with previous dilation techniques, balloon dilation typically requires multiple procedures to establish a durable improvement in airway diameter as discussed below. This can be contrasted with open repair techniques which aim to establish an adequate airway with one procedure. Both techniques require close follow-up and serial bronchoscopy to ensure continued patency.

Early descriptions of the technique employed angioplasty balloon catheters under fluoroscopy [29]. Most surgeons currently performing the procedure currently employ bronchoscopic visualization for placement. More recently, noncompliant balloons able to deliver pressure directly to a stenotic segment without deformation have been developed and widely adopted. Reported success rates, as defined as avoidance of tracheostomy or decannulation, vary between 39 and 100 %. Studies with longer term follow-up report success of approximately 50–60 % [30–32].

There are currently no studies directly examining outcomes of rigid dilation versus balloon dilation. Published studies show similar success rates between the two techniques despite the theoretical advantages of balloon dilation detailed earlier [33]. As with other airway reconstruction techniques, overall patient numbers analyzed in these studies are limited enough to make meaningful statistical analysis difficult.