In her 1976 review, Dr. Frost divides the history of tracheotomy into five stages. The first stage is the time of legend, and was from 1500 BC to 1500 AD. References to incisions in the windpipe are presented in the Rig Veda, the sacred book of Hindu medicine. In the fourth century BC, Alexander the Great punctured the trachea of a soldier with the point of his sword when he saw a man choking on a piece of bone that had lodged in his throat.

The second stage occurred between the sixteenth century and 1830. During this time tracheotomy was considered an irresponsible act with only 28 successful procedures performed in almost 300 years.

However, Nicholas Habekopf of Paris did publish a 108-page book on tracheotomy, the first book devoted solely to the procedure, in which he reported four successful operations. Two of these were on children, the first reported in the literature. One of these patients was a 14-year-old boy who had swallowed a bag of gold coins to prevent their theft. The bag lodged in the esophagus producing airway obstruction. Habekopf performed a tracheotomy and manipulated the money bolus distally towards the stomach.Despite being the scandal of surgery, Dr. Elisha Cedic recommended a tracheotomy for one of his patients, but the patient's two other physicians, Dr. James Craig and Gustfus Brown, did not concur. Subsequently this patient expired on December 15, 1799. The patient was, of course, George Washington, who died of acute upper airway obstruction from a peritonsillar abscess.

The third stage in the history of tracheotomy lasted until 1932. Frances Holme, a Scottish physician, published a treatise on croup based on his observations of 12 acutely stridulous children. Nine of these patients died. Subsequent dissection revealed a membrane obstructing the trachea. Holme, of course, had described diphtheria. In 1869 Trousseau presented data on 215 cases of diphtheria treated with tracheotomy. He reported a 25% survival rate.

In 1907 Chevalier Jackson published his text on tracheobronchoscopy providing information on equipment and techniques still used today. Two years later he published his experience on tracheotomy emphasizing a long vertical incision placed low to avoid entering the thyroid cartilage and discussing the need to divide the thyroid isthmus. Jackson's teachings were largely responsible for decreasing the mortality from tracheotomy from 25% to 1%-2% and reducing the incidence of laryngeal stenosis, especially in children.

The fourth stage lasted until the mid 1960s. The polio epidemic in the 1930s brought renewed interest in tracheotomy. The procedure was thought to be a method of prophylactically treating the pulmonary infections associated with polio. Physicians realized that performing a tracheotomy reduced the volume of dead space and was helpful in treating chronic lung disease. Until patients with this illness began to be treated with artificial ventilation, the only indications for tracheotomy were to relieve upper airway obstruction or remove foreign bodies. But by the 1960s, the Salk vaccine had eliminated polio as a primary reason for tracheotomy and by that time endotracheal intubation had become a safer and quicker procedure, with a lower complication rate.

The fifth stage is the present day, and the indications for performing tracheotomy continue to evolve. The three major indications for tracheotomy in pediatrics are ventilatory support, airway obstruction, and pulmonary toilet. Respiratory failure may be due to prematurity or pulmonary reserve or central nervous system disease. Those requiring a tracheotomy for airway obstruction are generally trached at a younger age for some congenital or acquired subglottic stenosis. Children trached for pulmonary toilet usually have some component of aspiration related to underlying neurologic disease. The breakdown for these three indications is not equal and the ratios have continued to evolve over the last twenty years.

Endotracheal intubation is performed in lieu of tracheotomy with increasing frequency. Reviews from several institutions have identified this so-called change in trends. Tucker and Silverman in 1972 performed the major retrospective review. They reviewed 350 pediatric tracheotomies over a twenty-five year period from five different Philadelphia area hospitals. They found an increase in tracheotomies in the later half of the study and fifty percent of their patients were between the ages of one and five. One-third of their procedures was secondary to croup or epiglottitis. Wetmore and Penn performed two retrospective reviews over a twenty-year period. The first was from 1971 to 1982. They found a gradual decrease in the number of tracheotomies throughout the decade, represented as a percentage of hospital admissions. This can be attributed to the increasing use of prolonged intubation for croup and epiglottitis as well as some central nervous system and cardiac disorders. Other institutional reviews supported this trend.

Prior to the late 60s and early 70s, intubation of the inflamed larynx was felt to be contraindicated, but now routine use of steroid and racemic epinephrine in patients with moderate to severe laryngotracheobronchitis has contributed to this change. In addition, Wetmore and Penn found an increase in tracheotomies in those patients under two years of age paralleling an increased survival in children with bronchopulmonary dysplasia and other congenital disorders. The most frequent primary diagnoses were central nervous system disease and airway obstruction. Seven percent of their procedures were performed secondary to croup. They also noted an increase in procedures performed for airway management during elective craniofacial procedures. Their follow-up study was from 1982 to 1990 and revealed that two-thirds of the surgeries were now performed during the first year of life and 75% were in patients younger than age two. The most common diagnoses now were bronchopulmonary dysplasia and neurologic disease. Only 1% was due to croup. The number of procedures dropped slightly, from 42 to 37.5 per year.

The timing of tracheotomy versus continued intubation in the pediatric patient remains controversial. Laryngeal injury from intubation is a well-described phenomenon. The endotracheal tube can cause glottic and subglottic edema within 72 hours. An endotracheal tube will always lie and exert pressure on the posterior larynx. In the pediatric population, a point of possible significant damage is in the subglottis. Because of its relatively small caliber, children are especially vulnerable. Pressure from the endotracheal tube can cause ischemia, and pressure necrosis of mucosa and underlying cartilage. Due to potential injury to the larynx and trachea from prolonged intubation, the current standard of care in adults is that patients who require intubation for more than two weeks should have a tracheostomy tube inserted. However, this two-week grace period is extended in pediatrics through the use of uncuffed endotracheal tubes and the advancement of neonatal care. The laryngeal cartilage in infants may yield to pressure and mold more than in older patients. However, in small infants it remains difficult to provide an endotracheal tube small enough to avoid mucosal injury yet large enough to provide adequate ventilatory support and pulmonary toilet. Frequent tube changes result in additional laryngeal trauma. In these instances, placement of a larger tube also potentiates further damage. However, there remains no general consensus for the maximum safe time of endotracheal intubation. This is usually determined on an individual basis. Studies have shown that the incidence of acquired subglottic stenosis in intubated neonates ranges from 1% to 9%, with an average duration of intubation ranging from 3 to 5 weeks. There are several reasons for discrepancies between these studies. The first is the lack of a universal grading system. Also some investigators include results from autopsy specimens. It is important to note whether those patients required any surgical intervention versus those with laryngeal injury identified endoscopically. In a prospective study Nicholas and Crysdale followed 289 patients intubated for more than 48 hours. Subglottic stenosis developed in 7, or 2.5% of them. Those patients that developed subglottic stenosis were intubated for longer periods of time and more frequently. The results imply the need for tracheotomy in infants requiring prolonged intubation or repeated intubation. Obvious anatomic differences in the pediatric larynx and trachea necessitate alterations in surgical technique. The anterior/posterior diameter of the infant larynx is only 7mm compared with 19mm in the adult. In infants, the subglottic diameter is only 5mm to 7mm. The infant larynx resides higher in the neck and is often shielded by the overlying hyoid bone. The cricoid is often the most prominent structure in the extended neck. The pediatric trachea has more lateral mobility. This makes it more difficult to palpate and also easier to retract out of the surgical field. Ideally, this procedure is performed under general anesthesia. Placement of an endotracheal tube or bronchoscope makes it easier to palpate the malleable pediatric trachea. Because children have thicker and shorter necks than adults, it is often helpful to remove some subcutaneous fat surrounding the skin incision. This also allows a tracheotomy tube to rest in good position in the trachea, preventing anterior dislodgment and accidental decannulation. When the trachea is reached, the cricoid cartilage is again identified with before the incisions are made, retraction sutures are placed just lateral to the proposed incision site. These provide quick identification of the new airway should the patient become accidentally decannulated during the immediate postoperative period. Then a vertical incision is made between the second and third rings.

Over the years there has been much discussion concerning the proper incision into the trachea. A common complication of pediatric tracheotomy is anterior wall collapse. Fry and colleagues at North Carolina studied different incisions in a pediatric animal model. They compared three commonly used incisions: the inferiorly based trap door, the vertical slit, and the horizontal H. These procedures were performed on male ferrets 8 to 9 weeks old that weighed 550-650 grams. They were decannulated after eight days and then sacrificed 1 week later. With the trap door technique, there was a significant anterior shelf. With both the vertical slit and the horizontal H, there was an hour glass narrowing. Tracheal compromise subsequent to healing after decannulation was greater endoscopically and radiologically in the inferior based trap door and horizontal H incision groups. Computer analysis of the area of maximal stenosis compared to the normal lumen showed a statistically significant increase in stenosis in the inferior based trap door incision. Furthermore, airflow resistance was significantly less in the vertically incised trachea.

Insertion of the tracheostomy tube produces physiologic changes because the nose, pharynx, and larynx are bypassed. Laryngeal bypass results in ineffective cough, loss of voice, ineffective glottic closure and aspiration. Nasal bypass results in the loss of the nasal effect on inspired air including decreased olfaction, humidification, cleansing, and warming. Tracheotomy also produces histologic changes. The trachea is normally compromised of pseudostratified columnar epithelium. At 37 degrees Celsius, the cilia beat about 1000 times per minute. Their coordinated activity moves the overlying mucous blanket at about 1.5cm per minute. The mucosa contains goblet cells and tubuloacinar glands that produce secretions which, together with the normal transudation of fluid, amount to 10cc per day. Normal elimination of these secretions depends on excretion through cilia transport and cough. With a 50% decrease in humidity, as seen in the tracheotomy, cilia motion ceases, resulting in thickening of secretions, crusting, and mucosal inflammation. Furthermore, conversion of ciliated respiratory epithelium and the stratified epithelium have been documented after both intubation and tracheotomy. The appropriate size of the tracheostomy tube depends on the clinical indications for the procedure and the size of the airway. Generally, the smallest tube that is capable of giving adequate air exchange is chosen. It is possible to estimate the age appropriate endotracheal tube and then estimate the appropriate size tracheostomy tube. Some tracheostomy tubes are numbered according to inner diameter, as are most endotracheal tubes. A healthy newborn should be able to accept a 3.5mm inner diameter tube. Usually a neonate may only accept a 3.0mm inner diameter tube. The appropriate length must also be estimated. Most tubes are manufactured in standard neonatal, pediatric, and pediatric long lengths. Generally children up to 7kg accept the neonatal length while above this weight the pediatric length is usually adequate. However, some children with tracheomalacia may need longer tubes to stent the airway open. The only way to accurately determine the correct length is to assess the position of the tracheostomy tube tip. The tip should be 6mm to 20mm above the carina. This can be done with a flexible telescope, a rigid bronchoscope, or a chest radiograph. Patients are monitored closely in an intensive care unit setting to avoid accidental decannulation. After 5 to 7days, the stoma is usually mature and the first trach change can be performed.

Unlike adults, children continue to grow, changing the length, caliber, and orientation of the trachea. It is important to evaluate the airway periodically to ensure the tube is in correct position, is the correct size, and that no potential complications are developing. The frequency of interval examination must be tailored to the child. Younger children who are growing more rapidly should be followed more closely.

Complications are best prevented by attention to detail intraoperatively and careful long-term care. Nonetheless, complications still occur. They are usually classified as early, which means occurring preoperatively or within the first week, or late. Some, such as accidental decannulation or tube obstruction can occur in both periods. In general, the complication rate for children is higher than for adults, with the overall complication rate ranging from 33% to 49%. This higher complication rate can be attributed to the increased incidence of air tracking problems, such as pneumothorax or subcutaneous emphysema, and to accidental decannulation. Some reviews include pneumonia as a postoperative complication. However, it is often difficult to determine if the underlying disease process was not present preoperatively. Likewise, others have recognized delayed decannulation as a complication. However, this is more likely due to the underlying disease process more than the tracheotomy itself. We will start by taking a look at some early complications. Major hemorrhage during tracheotomy is rare. The incidence of hemorrhage following tracheotomy ranges from 1% to 37%. The surgeon, though, should be aware of vascular variance including a high-riding innominate artery. But bleeding during surgery can be prevented with meticulous technique. The tracheostomy tube can cause airway irritation resulting in coughing, which raises venous pressure predisposing patients to bleeding. Persistent oozing is more often seen in coagulopathic patients, and can be controlled with loose, absorbable packing. Packing that is too tight can cause a ball valve effect and result in subcutaneous emphysema and pneumomediastinum. Bleeding that is more brisk usually mandates wound exploration.

A tracheotomy is considered a clean contaminated wound. In 24 hours the tracheostomy stoma always becomes colonized. However, wound infections are rare but can cause problems with skin breakdown in the perioperative period. This complication can be prevented with meticulous wound care. True infections respond well to local care plus antibiotics for any cellulitis of the surrounding soft tissues. Accidental decannulation can be a life-threatening complication especially in a fresh tracheotomy before a mature tract is formed. This is a primary reason why patients should be observed in an intensive care unit setting until the first tube change. Signs of tube displacement include an obstructed airway or inability to pass a suction catheter.

Again, prevention is the best treatment for this problem. This means secure placement of trach ties and stay sutures. Suturing the tracheostomy tube phalange to the skin is an additional safety feature. This can also be avoided by selecting a tube that is not too short.

If this does happen and the opening is easily visualized, the tube can be directly inserted into the lumen. If the tracheostomy opening lies deep to the soft tissues, a suction catheter or flexible fiberoptic endoscope can be used as a guide. If cannulation of the tracheostomy still appears potentially difficult, the airway can be stabilized by intubation in the absence of upper airway obstruction. What you want to avoid is the unsuccessful reinsertion that results in a false passage.

Tube obstruction can prevented with appropriate tracheostomy care: frequent suctioning to prevent build-up and plugging of the single lumen tubes. Also the posterior or anterior wall of the trachea may obstruct a tube that is too long or too short respectively. In the immediate postoperative period, a tracheoesophageal fistula is usually the result of an accidental laceration through the posterior membranous tracheal wall into the esophagus. This can be prevented by performing the incision against an endotracheal tube or a bronchoscope. When a fistula is formed, it should be repaired at the time of the initial surgery. Left untreated, this can progress to pneumothorax or mediastinitis.

The suction of air into the peritracheal tissues is the cause of subcutaneous emphysema, pneumothorax or pneumomediastinum. These are the so-called air tracking complications. Some degree of mild postoperative subcutaneous emphysema is expected. This can become more extensive, however, with airway trapping within soft tissue planes. It is usually caused by positive pressure assisted ventilation against a suture or packed wound. Consequently, it is important not to suture the skin incision and not to pack the wound too tightly for control of bleeding. Confining the dissection to the midline also minimizes this complication.

As we all know, subcutaneous emphysema is noted by crepitus over the neck or chest. This is usually a self-limited process and the body reabsorbs the trapped air in a few days. If subcutaneous emphysema becomes extensive, treatment is by opening the neck wound and placement of a drain. A chest x-ray should also be performed to rule out further complications. If left untreated, significant subcutaneous emphysema can lead to pneumomediastinum, or pneumothorax. Large amounts of air in the mediastinum can lead to hemodynamic instability. Treatment involves enlargement of the neck incision and placement of a drain to offer escape for the mediastinal air.

A pneumothorax results from three basic mechanisms. The first is from direct damage to the pleura. This occurs more commonly in children because the dome of the pleura extends above the clavicles. Dissection should thus be limited to the midline. This anatomical difference is thought to be the primary reason behind the pneumothorax rate in children, which is 10% to 17%, versus adults where it is reported as 0% to 4%. Mediastinal air can also rupture into the pleural space resulting in a pneumothorax. Patients breathing against an obstructed airway generate a very negative intrathoracic pressure making them susceptible to the suction of air along pleural planes. This risk of pneumothorax can be decreased in these patients by insertion of an endotracheal tube or bronchoscope prior to the procedure. False passage of the tracheostomy tube anterior to the trachea followed by positive pressure ventilation leads to the same results. The third mechanism is rupture of an alveolar bleb. Pneumothorax, unless small, is usually treated with a chest tube. Traditionally a postoperative chest x-ray is ordered to evaluate tube location and rule out the presence of a pneumothorax.

Cardiopulmonary complications also occur after tracheotomy. Manipulation of the tracheobronchial tree can lead to bradycardia and hypotension mediated by vagal reflexes. This usually ceases when the manipulation is stopped. Also patients with long-standing obstruction are driven to ventilate by their hypoxia. Sudden relief of the obstruction combined with oxygen administration may correct the hypoxia but cause loss of the respiratory drive and respiratory arrest. In a similar fashion, post-obstructive pulmonary edema can occur in a child with chronic airway obstruction. The high negative intrathoracic inspiratory pressure increases pulmonary interstitial fluid. The negative intrathoracic pressure increases venous return, which increases right cardiac pre-load and increases the alveolar capillary hydrostatic pressure. Hypoxia leads to additional sympathetic discharge and vasoconstriction, shifting more fluid in the pulmonary space, but this pathophysiology is counteracted when the patient exhales against a fixed obstruction creating a positive alveolar pressure. Relief of the obstruction disrupts this balance. Pulmonary edema usually responds to positive pressure ventilation. Misplacement of the tube into the soft tissues or main stem bronchi can lead to cardiopulmonary complications as well.

The first "later" complication we will consider is subglottic stenosis. The incidence of this complication has certainly decreased since Jackson emphasized the danger of high tracheotomy 75 years ago. Other factors that may contribute include subglottic trauma, generally from endotracheal intubation, and low-grade inflammation from local infection originating from the tracheostomy stoma. Stenosis is also associated with emergency surgery in a child with an unsecured airway.

The most common complication is a tracheogranuloma. Some otolaryngologists have argued that since tracheogranulomas are so common they should not even be considered a complication. These occur most frequently just above the tracheal stoma. They can be pedunculated or can occur as an epithelealized tract that grows over the tracheostomy tube. They are believed to be caused by chronic low-grade inflammation resulting from multiple causes including secretions pooling just above the cannula as it enters the trachea, foreign body reaction to the cannula, or frictional trauma from the tube. Treatment depends on the size. Those that are not obstructive can be watched. Larger ones should be removed, since they obstruct the stoma during tube changes. Removal can be done through a bronchoscope or through the stoma. More distal granulomas can cause tube obstruction. A poorly positioned tube should be repositioned away from the site of granulation. Changing the tube length and material may also be helpful. Longer tubes can be inserted past the granulation, but this risks creating more granulation distally. Longer tubes are also limited by the carina.

Pressure on the first and second rings from the tracheostomy tube can cause local chondritis, weakening of the upper tracheal rings, with the result in super-stomal collapse. Upper pressure from an undivided thyroid isthmus can also cause this complication. As I mentioned earlier, transverse incisions, as well as removal of any cartilage at the time of tracheotomy, can lead to this complication. Super-stomal collapse can cause significant obstruction of the airway making future decannulation impossible. However, prevention of super-stomal collapse is not always possible. It is more common in younger patients who require prolonged tracheotomies. Relief of the obstruction involves surgical intervention, which ranges from stomal revision to tracheal resections with primary anastomosis.

A tracheoesophageal fistula is due to erosion of the posterior tracheal wall causing a fistula to form between the trachea and the esophagus. Pressure on the posterior tracheal wall from the distal tip of the tube is the most common cause. The posterior wall may be in contact with the distal tip if the tube is too short or if there is an anatomic abnormality of the trachea, such as in children with severe scoliosis. A delayed tracheoesophageal fistula can also occur in patients with a nasogastric tube. The tissue between the NG tube and the tracheostomy tube may undergo pressure necrosis. Smaller fistulas can be managed with removal of the offending tubes and drainage of the wound, while larger fistulas require surgical repair with inner position of muscle flaps between the trachea and esophagus.

Tracheal innominate artery fistula is the most-feared, late postoperative complication. The innominate artery arises from the aortic arch to cross in front of the trachea to give rise to the right common carotid and subclavian arteries. The tracheostomy tube may erode the artery when the tube contacts the artery through the tracheotomy incision being placed too inferiorly or an when an artery is aberrant or unusually high. It can be due to an excessively long or poorly fitting tube, which may erode through the anterior tracheal wall.

An overinflated cuff may cause ischemic necrosis of the trachea. Tracheal infection can also contribute. Granted, most bleeding will come from stomal granulation tissue or irritation from deep suctioning, but any time blood is reported coming from the tracheostomy tube, the possibility of a tracheoinnominate fistula must be entertained and certainly a high index of suspicion is necessary for prompt diagnosis and successful management. Erosion of the innominate artery is usually preceded by the famous sentinel bleed during which the patient coughs up a small amount of bright blood prior to the actual massive hemorrhage. When such a bleed is recognized, a cuffed tube should be inserted into the trachea and suprasternal pressure applied. Definitive management requires a sternotomy and division of the innominate artery.

How often do these complications occur? If we look back at Tucker and Silverman's initial study, they found an overall complication rate of 46.5%. This includes 163 complications in 350 tracheotomies. Interestingly, however, their most common complication was delayed decannulation, which they define as a tracheostomy tube in place more than thirty days after the procedure. They had a total of fifteen pneumothoraces. There were no deaths associated with a pneumothorax, and they had no major vascular erosions. They had five total mortalities, commonly from tube obstruction or accidental decannulation. Their review also included an analysis of the world literature at that time, which included 4,000 pediatric tracheotomies. They identified 201 pneumothoraces, which has the most common fatal complication with 27 deaths. The majority of these occurred before 1960.

If we move to more recent studies on complications from larger series since 1980, the number of patients ranges from 164 to the two Wetmore studies, which I reviewed earlier, which had 420 and 373 patients. The complication rates weren't readily accessible from the article, but looking at the air tracking complications, their pneumothorax rates were 2.6% or less. Their most frequent complications were accidental decannulation or tube obstruction. If these series are reviewed for their later complications, the percentages tend to range from 23% to 58%.

By far the most common are tracheal and stomal granulations, with an incidence of 10% to 50%. Tube obstruction and accidental decannulation are usually the other common late complications. Accidental decannulation probably occurs more frequently than is reported and most of those remain unreported unless significant morbidity occurs. Wetmore's reviews also include tracheocutaneous fistula rates of 20%, which increased their overall complication rates. They found that tracheocutaneous fistulas were present in those patients with the tube in place for longer periods of time. Again, there is debate whether this should be included as an actual complication. Overall, out of all 1276 cases, they had two incidences of fatal tracheal innominate fistulas. In terms of mortality, the tracheotomy related mortality ranges from 0% to 4%. In contrast, looking at the mortality for children with a tracheostomy tube in place, that is much higher at 14% to 33%. This, of course, reflects significant comorbidities in these patients.

As to the causes of tracheotomy-related mortality reported in the literature, most are due to tube obstruction or accidental decannulation. Most of the hemorrhages were due to TI fistulas. Which patients are more prone to complications? Researchers have looked back and have shown that younger patients, less than one year old, are prone to more complications; those patients with bronchopulmonary dysplasia also tend to get more complications; and, the longer the patient has the tracheostomy tube, the greater the opportunity to develop a complication. They also all recognize the hazards of emergency surgery and certainly this increases the risk of air tracking complications such as subcutaneous emphysema or pneumothorax. They all recognize that complications are minimized with a controlled airway. Some studies have focused specifically on those patients younger than one year of age. Gianoli retrospectively reviewed 60 patients under one year of age who underwent tracheotomy over an 8-year time period. Exactly half of these patients were premature and half were full term. They had quite a low intraoperative complication rate, a little higher early complication rate, and an even higher late complication rate. Complication rates were twice as high in the pre-term patients as in the term patients. A higher complication rate was found in those patients with upper airway obstruction, which is not surprising, since these patients had less comorbidity and tended to live longer. They found a direct linear correlation between the survival time and the amount of complications. They did not find a correlation between more complications and higher comorbidity score. Quite simply, those patients didn't live as long. They also found higher complications with emergency surgery. They had one emergency case that accounted for 66% of their intraoperative complications. They had one tracheotomy-related mortality that correlated to 1.6% and 42% of their patients died with a tracheostomy tube in place.

Case Presentation