Pneumomediastinum (PM) was first described by Laennec in 1819 as a consequence of traumatic injury. Spontaneous pneumomediastinum (SPM) was reported in 1939 by Hamman, for whom the Hamman sign (see Physical) is named. Pneumomediastinum is an uncommon entity in pediatric practice. It is defined as free air or gas contained within the mediastinum, which almost invariably originates from the alveolar space or the conducting airways. The etiology of pneumomediastinum is multifactorial. 
Many authors distinguish spontaneous pneumomediastinum as a form of pneumomediastinum that is not associated with blunt force or penetrating chest trauma, endobronchial or esophageal procedures, neonatal lung disease, mechanical ventilation, or chest surgery or other invasive procedures. Because of the infrequent occurrence of pneumomediastinum, the literature relating to pneumomediastinum involves individual case reports or small case series and is retrospective in nature.
The following images demonstrate radiographic findings in patients with pneumomediastinums.
Pneumomediastinum rarely leads to clinically significant complications. More commonly, the associated or precipitating condition underlying pneumomediastinum may be the cause of significant illness. Rarely, tension pneumomediastinum has been reported in which elevated mediastinal pressure leads to diminished cardiac output because of direct cardiac compression or reduced venous return. When extensive subcutaneous and mediastinal gas is present, airway compression may also occur. Jennings et al (2013) report this as a complication following blunt thoracic trauma. 
The generally accepted explanation for the development of pneumomediastinum is that free air tracks from ruptured alveoli along peribronchial vascular sheaths toward the hilum of the lung. From there, it extends proximally within the mediastinum.
The Macklin effect, first described in 1939, highlights the sequence of events in the development of pneumomediastinum as follows: (1) alveolar rupture, (2) air dissection along the bronchovascular sheath, and (3) free air reaching the mediastinum.
The dissection of free air may not be confined solely to the mediastinum. Zylak et al note that the mediastinum communicates with the submandibular space, the retropharyngeal space, and vascular sheaths within the neck.  In addition, 2 routes of communication with the retroperitoneum have been noted: via a tissue plane extending through the sternocostal attachment to the diaphragm, as well as periaortic and periesophageal fascial planes. As a result, air present within the mediastinum may dissect through these tissue planes, causing pneumopericardium, pneumothorax, subcutaneous emphysema, pneumoperitoneum, or pneumoretroperitoneum.
Carolan et al report a case of spontaneous pneumomediastinum associated with the presence of air within the cervical and thoracic spinal canal indicative of air dissection from the mediastinum, neck, and back through soft-tissue planes and the neural foramina into the spinal canal.  They suggest use of the term “spontaneous pneumorrhachis” for patients with this finding.
In a study of 87 patients by Wong et al, the common causes of secondary spontaneous pneumomediastinum included asthmatic exacerbation, infections (eg, pneumonia, lower respiratory tract infections), and choking. 
The epidemiology of pneumomediastinum reflects that of the associated disease states, when present.
Spontaneous pneumomediastinum is a rare condition. A review by Chalumeau et al summarized the available literature.  Based on previous studies, they determined a prevalence of spontaneous pneumomediastinum ranging from 1 per 800 to 1 per 42,000 pediatric patients presenting to a hospital emergency department. Stack et al reported a 0.3% incidence of pneumomediastinum in association with asthma presenting to their institution over a 10-year period.  The mean age of affected patients was 11 years. No sex differences were observed in this cohort.
A study from Nashville, Tennessee, reported the frequency of extra-abdominal gas in a series of patients undergoing laparoscopic esophageal surgery. Forty-seven percent of patients (N = 45) had evidence of extra-abdominal air on chest radiography. Of these, 86% had a pneumomediastinum. Pneumomediastinum persisted at least 1 postoperative day in two thirds of these cases. However, no mortality or morbidity was attributable to the presence of pneumomediastinum.
In a series of patients with sepsis-induced acute respiratory distress syndrome (ARDS), air leaks of any type, excluding pneumothorax, occurred in 3.7% of patients.  Ventilator pressures and volumes delivered were not correlated with the development of air leak.
In a series of adult patients presenting with blunt chest trauma, as many as 10% had evidence of pneumomediastinum.
A study by Briassoulis et al from Athens, Greece, evaluating the frequency of air leaks in children receiving mechanical ventilation reported a prevalence of 27%.  However, they did not report the prevalence of specific types of air leak.
Esayag et al reported an Israeli series of 13 patients with spontaneous pneumomediastinum.  This group represented 1 in 41,600 referrals to the emergency room and 1 in 15,500 hospitalizations. The median age of the patients was 19 years (range 2–72 y). Males comprised 77% of this group.
A case series from the Children's Medical Center, China Medical University in Taiwan reported by Lee et al defined an incidence of spontaneous pneumomediastinum in children of 1:8,302 patient visits to the pediatric emergency department.  They observed a bimodal distribution, with cases occurring in children younger than 4 years old and in adolescents aged 15-18 years. Males outnumbered females by a ratio of 4:1.
The mortality and morbidity associated with pneumomediastinum are generally attributable to underlying disease states. Spontaneous pneumomediastinum is usually a self-limited condition that rarely produces significant or life-threatening symptoms.
The mortality rate associated with pneumomediastinum may be as high as 50-70% as seen in Boerhaave syndrome (esophageal rupture following vomiting). The development of air leak, according to Weg et al, is not associated with an increased mortality rate in patients with sepsis-induced ARDS.  Other predisposing conditions associated with high mortality rates include trauma (blunt and penetrating, especially high velocity injury), asthma, and tracheobronchial perforation.
The most common morbidities attributable to pneumomediastinum are symptoms such as chest pain, voice change, and cough. Rarely, tension pneumomediastinum may result in decreased cardiac output. Laryngeal compression leading to stridor has been reported. Gas embolism has rarely been reported.
In a series of pneumomediastinum occurring in persons with asthma, there was a very slight male predominance in the prevalence of spontaneous pneumomediastinum. Other series confirm this excess of male cases. Damore reported 29 cases of pneumomediastinum over a 10-year period unrelated to trauma, intubation, or surgical procedures; 69% of patients were male.  Some have suggested that a body habitus favoring a tall thin build is an additional risk factor for the development of spontaneous pneumomediastinum. The mechanisms underlying this association are unclear. Traumatic pneumomediastinum is more common in males, reflecting the male predominance among those who experience trauma and accidents.
The peak prevalence of spontaneous pneumomediastinum is seen in the second to fourth decades of life. This presumably reflects involvement in activities that increase the risk for developing pneumomediastinum, such as diving or marked physical exertion (eg, athletic activities, weight lifting). Moreover, the force of an individual's cough, vomit, and Valsalva maneuvers (all of which may lead to pneumomediastinum) attenuates with age, accounting for the decline in the prevalence of pneumomediastinum with age.
The age distribution for pneumomediastinum occurring in conjunction with specific disease processes reflects the age profile of the particular disease.
What would you like to print?