Paediatric patients with a tracheostomy: a multicentre epidemiological study

RESULTS

Replies were received from 22 hospitals. Four of these stated that they did not follow up tracheostomised children during the study period. The other 18, from 11 different cities in Spain, comprised the participating centres. Of the 270 patients initially recorded, the final study included 249 as 21 were excluded: 10 because they had been decannulated before the start of the study, five because they were >18 yrs of age at the start of the study, five because they were already included in the details reported by another hospital and one who died before the start of the study. Table 2 shows the number of patients per centre. When comparing data from the different centres we found that the number of children reported in each hospital was in proportion to the number of beds in that hospital. The large cities Madrid and Barcelona, which each have several tertiary hospitals, accounted for 32% and 26%, respectively, of all the patients referred. There were 150 (60.2%) males and 99 (39.7%) females.

The age at which tracheostomy was performed varied between the first hours of life (a male with congenital laryngeal atresia) and 17 yrs (a male who received a lung transplant due to cystic fibrosis). The median age was 6 months. Given the wide age range and the variability of the underlying disorders, the patients were grouped according to age, the largest group being that of the newborns (<6 months) (n=113, 45.3%) followed by those aged 6 months–1 yr (n=37, 14.8%) (fig. 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Age at which tracheostomy was performed. Stratification by age group.

Regardless of the underlying disorder, the indications leading to the procedure varied. The main reason was prolonged ventilation (n=156, 62.6%), followed at some distance by acquired subglottic stenosis (n=34, 13.6%) and others (fig. 2). Most notable among the underlying conditions were neurological disorders (n=126, 50.6%), with 67 (53.1%) children having neuromuscular problems. Second was respiratory disorders (n=98, 39.3%), mainly bronchopulmonary dysplasia (n=32, 32.6%). Most patients, though, suffered various different processes. Of the 249 patients, 35 (14%) had been born prematurely. Given that a high percentage of the children (60.2%) were <1 yr of age at the time of the tracheostomy, figure 3 compares the conditions in these children versus the rest.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

Indication for tracheostomy. PV: prolonged ventilation; ASS: acquired subglottic stenosis; CFA: craniofacial anomaly; CAA: congenital airway anomaly; Other: acquired bilateral vocal cord paralysis (n=4), laryngeal papillomatosis (n=4), acquired tracheo-oesophageal fistula (n=1), post-operative central apnoea after cerebral tumour (n=1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3–

Most frequent disorders in children with tracheostomy. Comparison between those <1 yr and those >1 yr of age. Note: a patient may have more than one disorder. NR: neurological; Resp: respiratory; CS: congenital syndrome; Car: cardiopathy; Max: maxillofacial; HO: haematological–oncological; Dig: digestive; Other: patients with various comorbidities.

The duration of tracheostomy ranged from 1 day (due to perioperative mortality) to 19 yrs (a patient with a perinatal medullary lesion). Over the 2-yr study period, decannulation was achieved in 59 (23.7%) patients. The median duration of the tracheostomy in these 59 decannulated patients was 7 months (range 0.1–212 months). In the other 190 patients who remained tracheostomised due to their complex situation or who died while still cannulated, the median duration of the tracheostomy by the end of the study period was 34 months (range 0.03–236 months).

In addition to the tracheostomy, 92 (36.9%) patients required ventilatory support during the study period, either in ICUs (n=12, 4.8%), at home (n=79, 31.7%) or both (n=1). Five (82%) of the 249 patients required a second tracheostomy. Amongst those children ≥4 yrs of age (127 patients), 79.5% (101 patients) received some type of educational support while carrying a tracheostomy with or without ventilatory support, whether at the hospital, at home or even attending regular educational centres or other specialised centres for children with medical needs.

Concerning mortality and morbidity associated with the procedure, one or more complications were reported in 117 (46.9%) patients. Figure 4 shows those during the tracheostomy procedure itself as well as those during the perioperative period or during later care, either at home or in hospital.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4–

Complications associated with tracheostomy. SCO: severe cannula obstruction; AD: accidental decannulation; Inf: infection related to tracheostomy cannula; GT: important granulation tissue; Exit: death directly related to the tracheostomy; PTX: pneumothorax; SS: stomal stenosis; ST: suprastomal tracheomalacia; H: haemoptysis. Other: trachea-innominate artery fistula (n=1), tracheal stenosis after tracheostomy (n=1) and rupture of the cannula (n=1).

At the end of the study period, 218 (87.5%) out of the initial 249 patients remained alive. Median age for the survivors was 65 months (range 2–238 months). Only one patient was lost to follow-up. Of the 31 deaths, eight were directly related to the tracheostomy; figure 5 shows the causes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5–

General mortality and mortality directly related to the tracheostomy. Patients alive at the end of the study: 218 (87.9%) out of 249; patients dead at the end of the study 31 (12.5%) out of 249. DRUD: death related to the underlying disorder (23 out of 31, 9.7%); DRT: death directly related to the tracheostomy (eight out of 31, 3.2%), of which four were newborns during the immediate post-operative period (one case of suture dilaceration in a patient with subglottic congenital membrane, two cases of bilateral pneumothorax and one case due to complications with mechanical ventilation because of inadequate seal and cannula displacement). The other four patients died at home due to severe obstruction of the cannula and failure of resuscitation measures.

DISCUSSION

Although paediatric tracheostomy is seldom performed nowadays, it remains necessary for the management of a few “technology-dependent” children, enabling them to be discharged from the ICU [4, 12–14]. The best data regarding its true incidence stem from a study by Lewis et al. [15], which analysed the data from 2,521 US hospitals and estimated the rate to be 6.6 children per 100,000 child-yrs during 1997. Although originally this technique was developed as an emergency procedure to solve difficult acute situations, in many settings, it has now become a programmed decision, part of a multidisciplinary approach in complex patients with chronic disorders. Indeed, Corbett et al. [3] found that just 6% were performed for emergency airway management in a series of 122 children between 1987 and 2003.

Over recent years, numerous studies in different countries have reported the individual experience of various hospitals, with the number of patients varying from 36 to 362 over periods of time ranging from 2 to 37 yrs [3, 5, 6, 8, 9, 16–18]. Review of these reports reveals differences between the various centres. While the frequency of the technique has fallen and is almost absent in smaller hospitals, it has remained the same, or even risen slightly, in referral hospitals, due to its complexity and the need for multidisciplinary management of the children referred [5, 3, 9, 16].

Concerning the age at which the tracheostomy was performed, the study with the most patients published so far [15] indicates two peaks: <1 yr (32%) and 15–18 yrs (10.3%), the latter mainly resulting from cranioencephalic trauma. Others have reported figures for children <1 yr of age of between 50% and 66% [3, 8, 19], which is more like the 60.2% in our series. There has also been a growing trend for the procedure to be used in premature infants (10.8%) [15, 20] and in children with multiple chronic disorders or severe congenital diseases in neonatal or paediatric ICUs [9, 21, 22]. The mean age for tracheostomy in most series ranges from 3.2 to 7.8 months [3, 6, 8, 17], or around 3 yrs of age [2, 16]. The higher mean age in our series may be due to its multicentre nature and the inclusion of referral centres for lung transplantation, which greatly increases the upper limit for the procedure. We did not detect the second peak in the incidence of tracheostomy in adolescents seen by Lewis et al. [15], probably because the age for admission to many children's hospitals in Spain is <16 yrs, although chronic patients are usually followed-up to a later age. As in our series, most studies have found a slightly higher frequency in males (60–63.8%) [6, 15, 18], though not all (48%) [17].

The percentage of cases due to prolonged ventilation, the main indication in our series, varies (22–66%) [2, 3, 5, 6, 9, 16, 19]. The low percentages in some reference centres [5, 6] may be related to the greater use of noninvasive ventilation or the higher number of interventional procedures in chronic obstructive airway disorders [8]. Whereas in earlier decades acquired subglottic stenosis only accounted for a small proportion of tracheostomies (0% [23] to 5% [24]), it was the second most common indication in our series (13.6%), a similar percentage to that found by others [3, 8, 19], though yet others have reported much higher rates (20–36%) [5, 6, 16, 17, 25]. This increase is particularly surprising, as interventional techniques now exist to resolve this condition, often without the need for tracheostomy [5, 16]. The third indication for tracheostomy in our series, upper airway obstruction secondary to craniofacial malformations or anomalies (10%), is becoming more usual in tertiary referral hospitals (5–32.7%) [3, 5, 6, 8, 19]. Finally, specific congenital airway malformations, such as tracheal stenosis, laryngo–trachea–bronchomalacia, subglottic haemangiomas or laryngo–trachea–oesophageal cleft, represented the fourth reason for tracheostomy (9.6%). This figure is lower than that for some tertiary referral centres (19.8% [5] or 21.4% [3]), which are seeing a notable increase as a result of the new interventional possibilities. Of interest among the other indications was that conditions associated with the classical pathogens, such as C. diphtheriae or epiglottis due to H. influenzae, have disappeared from the statistics over the last 10 yrs [3], although papillomavirus persists (1.6%), as it does in other studies (1% [5] or 2.6% [3]).

The underlying disorder (fig. 3) was more varied in the patients <1 yr of age, as occurred in other studies [15]. The most usual were neurological disorders, in both the younger patients (68 out of 159, 45.3%) and those >1 yr (58 out of 99, 58.5%). The high percentage of patients with neurological involvement is also noted in other series (42%) [12]. Of note, too, in our series, was chronic respiratory involvement in children aged ≤1 yr (69 out of 150, 46%), mainly due to bronchopulmonary dysplasia (32 out of 69, 46.3%), a disorder suffered by 32 (12.8%) out of the 249 patients in this series.

Comparison of our 23% of patients who achieved decannulation with other series, which report decannulation figures of 29–52.7% [2, 3, 6, 16, 17, 26], or even 75% in New Zealand [8], gives a notably lower percentage. Possible reasons for this include the shorter period of our study, the high percentage of patients with chronic neurological and respiratory indications for long-term ventilation, and the high percentage of patients <1 yr of age. Studies in decannulated patients indicate a mean duration of tracheostomy that can vary from 12.4 [3] to 22 months [19], depending on the underlying disorder. One of these studies found that patients <1 yr of age had the tracheostomy for much longer periods (829 days) than children >1 yr (94 days) [3]. Another study analysed the possible risk factors related to the duration of the cannulation, noting that the only determining factors were the type of disorder and the indication for tracheostomy, but not the age of the patient [27]. It is nevertheless interesting to note that despite the chronicity and complexity of the patients, 79.5% of the survivors still managed to achieve social insertion appropriate for their age via the preschool/school support system, independently of whether they had been decannulated or were receiving ventilatory support.

The frequency of complications in our series (46.9%) is high, as it is in other paediatric series (51–77%) [1, 8, 26], in comparison with the frequency in adults. In a multicentre survey, 22% of paediatric surgeons reported perioperative complications during the tracheostomy, leading to a change in practice by 58% of them [22]. A study of 54 patients with home mechanical ventilation (45 via tracheostomy and nine via nasal mask) reported that 66% of the severe emergencies were related to the tracheostomy [28]. Whereas infection is reported to be the main complication [26, 27], as much as 90% in some series [6], in our series it only accounted for 8.8% of complications. This difference may be related to both the shorter follow-up and the question of differentiating infection versus colonisation, under- or overestimating the true incidence. Different percentages have been reported for endotracheal granulation in various paediatric series, ranging from 12.3% to 56% [1, 6, 8]. We only considered granulation to be a complication when it led to symptoms and was subsidiary to intervention (8.4%). The main life-threatening complications in our study were severe obstruction of the cannula by a mucous plug (29.3%) and accidental decannulation (15.2%), this latter percentage being similar to that of other series [3, 6]. The other life-threatening complications, such as pneumothorax, subcutaneous emphysema or tracheo-innominate artery fistula, were much less usual [2, 3, 19, 26].

Mortality from the underlying condition, 9.7% in our series, varies greatly in other studies according to the main underlying disease (6.9–39.2%) [2, 3, 6, 8, 9, 19, 23]. Mortality directly related to the procedure, however, (3.2% in our series), was higher than that of other studies (0.7–3%) [2, 3, 8, 18, 26, 29]. Of note was the high mortality associated with the procedure in preterm infants (four in our series) (fig. 5) in comparison with other studies in which tracheostomy is presented as a safe technique at that age [20].

Our study does have certain limitations. It did not record a series of data that would, with hindsight, have proved interesting, such as the time between respiratory failure and the tracheostomy, given the high percentage of children with long-term ventilation as the main indication for elective tracheostomy. Unlike adult patients, this time is not clearly defined in children. Very variable intervals have been reported, even in the same centre (0–148 days) [6, 9], which explains why the indication for tracheostomy in a child is usually personalised according to the clinical status, the endoscopic findings, the experience with noninvasive ventilation and the family circumstances [1, 12, 30].

It would also have been of clinical interest to determine whether, before doing the tracheostomy, non-invasive ventilation had been attempted, how long the patient had been in the ICU and in hospital after the tracheostomy, the cost per patient and the percentage of readmissions after the tracheostomy, as it is easy to assume that the battle ends with the tracheostomy, when in fact a new one begins [4, 12, 30]. Indeed, in our series, 31.7% of the patients required assisted ventilation at home at the time of hospital discharge. This circumstance was also noted in a series from Houston, TX, USA (61%), in whom 81% had additional support devices or techniques besides the tracheostomy, such as feeding tubes or central venous access [9].

Finally, although in our study we did not collect information regarding use of devices such as speaking valves or other phonatory means, we agree that early development of language can improve the social integration of these patients. Although it has been recommended that speaking valves should not be used in very small or very ill children due to the additional work of breathing, children of all ages, including infants, may be candidates for its use [11]. A new study on the use of these devices could help define indications for their use, and the true limitations in very young, small or unstable children. New surgical techniques that have been implemented in children with airway anomalies, e.g. endoscopic anterior cricoid split and balloon dilation, laryngotracheal reconstruction and posterior costal cartilage grafting or slide thyrocricotracheoplasty in subglottic stenosis, and laser therapy or tracheobronchial stents in several airway anomalies, are not reflected in this paper either, despite the positive impact that they have had in some of our patients, who could not have achieved decannulation otherwise.