Spontaneous pneumomediastinum is free air in the mediastinum with no traumatic or iatrogenic cause. It occurs most often in young adults after coughing, retching, exertion or a Valsalva manoeuvre, and it is usually self-limiting. The clinical work sits in what has to be excluded first — oesophageal perforation, tracheobronchial injury, and tension physiology. In the largest published series, 10 per cent of patients presenting with mediastinal air were found to have an oesophageal perforation. Mr Lawrence Okiror is a Consultant Thoracic and Robotic Surgeon at London Bridge Hospital and The Lister Hospital Chelsea, and receives these referrals through the thoracic on-call at Guy’s and St Thomas’. Private appointments within 2–3 working days. Self-referrals welcome.
Last reviewed: September 2026 · Mr Lawrence Okiror FRCS(CTh) FRCSEd(CTh) · GMC 6150382
Alveolar rupture, with air dissecting along the bronchovascular sheath to the hilum and into the mediastinum — the Macklin effect, described in 1939. The visceral pleura stays intact, which is why most patients have no pneumothorax
In a series of 249 patients presenting with mediastinal air, 24 (10%) had an oesophageal perforation. Older age, a pleural effusion and a raised white cell count were the independent predictors
Observation, analgesia and treatment of the precipitant, with repeat imaging at 24–48 hours to confirm a settling course. A concomitant pneumothorax is the finding that brings a chest drain into the plan
Mediastinal air on a CT report in a young adult is usually benign and self-limiting. The word spontaneous is applied once oesophageal perforation, tracheobronchial injury and tension physiology have been accounted for. In the largest published series, 10 per cent of patients presenting this way had an oesophageal perforation, and the features that identified them were older age, a pleural effusion, and a raised white cell count.
Most of those exclusions are made on the history, the observations and the blood results rather than on further imaging. The decision that matters at the point of referral is which patients need a contrast study and which can be observed. Request a consultation at London Bridge Hospital within 2–3 working days →
A sudden rise in intra-alveolar pressure ruptures alveoli at their junction with the bronchovascular sheath. The escaped air then dissects centripetally along that sheath — towards the hilum rather than out through the pleura — and enters the mediastinum. From there it tracks upward into the deep cervical fascial planes and outward into the chest wall, which is what produces the subcutaneous emphysema patients notice as crackling under the skin of the neck.
Macklin described this in an animal model in 1939, and Macklin and Macklin set out the full pathophysiology in 1944. The sequence has three steps: alveolar rupture, air dissection along the bronchovascular sheaths as pulmonary interstitial emphysema, and spread of that air into the mediastinum. On CT it appears as linear collections of air running alongside the bronchovascular bundles, and subpleural pulmonary interstitial emphysema is present in around half of cases.
The practical consequence of the route the air takes is that the visceral pleura is not breached. That is why the majority of patients have mediastinal and subcutaneous air with no pneumothorax at all, and why the appearance on the scan is often more alarming than the physiology. Where a pneumothorax is present as well, it changes the management, and that situation is dealt with below.
Pneumomediastinum, mediastinal emphysema and Hamman syndrome all describe the same finding. Hamman published the first case series in 1939. The Macklin effect refers to the mechanism, and on a CT report it describes the linear peribronchovascular air that demonstrates a pulmonary rather than an oesophageal source.
The dominant symptom is retrosternal chest pain, often radiating to the neck or the back and frequently worse on deep inspiration or swallowing. Breathlessness, odynophagia and a change in the voice are common. The physical findings are subcutaneous emphysema in the supraclavicular fossae and neck, and occasionally Hamman’s sign — a crunching sound synchronous with the heartbeat on auscultation over the praecordium.
The differential at first contact is broad, because a young adult with sudden retrosternal pain is assessed for acute coronary syndrome, pulmonary embolism, pericarditis, pneumothorax and musculoskeletal chest pain before mediastinal air is considered. Where the pain is not explained and the chest radiograph is reported as normal, a lateral film is worth having: retrosternal air is often visible on the lateral projection when the frontal view looks unremarkable. Patients whose chest pain has no clear cause on first assessment are dealt with more broadly on the unexplained chest pain page.
The chest radiograph confirms the diagnosis in most cases. CT of the chest defines the extent of the air, identifies any associated pneumothorax, demonstrates the peribronchovascular air that supports a pulmonary source, and allows the mediastinum and oesophagus to be interrogated directly. Where there is a reason to suspect oesophageal injury, CT is performed with oral contrast rather than as a plain study. What each chest scan is for →
One point deserves emphasis for young patients. Mediastinal air with a cough or haemoptysis that the episode does not fully explain should not close the assessment. A normal chest radiograph does not exclude a central airway lesion, and persistent respiratory symptoms in a young adult warrant their own work-up. When cough or haemoptysis in a young adult needs a CT →
Air reaches the mediastinum from the lung, from the airway, or from the oesophagus. Establishing which of those applies is the whole of the early assessment, and the published data give a clear sense of the stakes. In a retrospective review of 249 patients presenting with pneumomediastinum over ten years, 24 — ten per cent — were ultimately found to have an oesophageal perforation, diagnosed at endoscopy, on contrast study, or at emergency surgery. On multivariate analysis only three variables independently predicted perforation: older age, a pleural effusion, and a raised white cell count.
The same series concluded that where those features are absent, most patients can be observed without further diagnostic testing. That is the useful message for a referring team at two in the morning: the discriminating information is already in the history, the observation chart and the blood results.
Exclusion 1
Boerhaave syndrome classically follows forceful retching or vomiting. The patient is usually systemically unwell, with pain out of proportion to the findings, and often a left pleural effusion. Iatrogenic perforation after endoscopy or transoesophageal echocardiography is commoner than Boerhaave and belongs in the same differential.
Where suspected: a water-soluble contrast swallow, or CT with oral contrast. Barium is avoided. A negative swallow does not settle the question in a patient who remains clinically suspicious.
Exclusion 2
Largely a question of context. Trauma to the neck or chest, recent intubation or airway instrumentation, a rapidly enlarging volume of air, progressive subcutaneous emphysema, or a pneumothorax with a large or persistent air leak all move this up the list.
Where suspected: bronchoscopy, arranged after a thoracic surgical opinion. Central airway interventions →
Exclusion 3
Rare, and the one presentation that is time-critical. Accumulating mediastinal air compresses the heart and the great veins, reducing venous return and producing hypotension, tachycardia and distended neck veins.
Where suspected: immediate decompression and an urgent thoracic surgical opinion. Haemodynamic compromise in a patient with mediastinal air is never attributed to the air being uncomfortable.
Once the three exclusions above are satisfied, treatment is supportive. The mediastinal air is reabsorbed without intervention, and the components of management are analgesia, rest, avoidance of straining and Valsalva manoeuvres, and proper treatment of whatever precipitated the episode. An asthma exacerbation that has not been optimised is the most important example, because it is both the cause of the current episode and the risk factor most consistently associated with a further one.
Mr Okiror’s practice for patients referred to the thoracic on-call is a short admission, usually one to two days, with repeat imaging at 24 to 48 hours to confirm that the volume of air is static or reducing and that a pneumothorax has not developed in the interval. The purpose of the second study is to document a settling course before the patient is discharged with reassurance, and to catch the small number in whom the picture is progressing. A patient whose symptoms are worsening is re-imaged regardless of the interval.
Two interventions are commonly applied and are not required. Prophylactic antibiotics have no role once an oesophageal or airway source has been excluded; they are appropriate only where perforation is confirmed or still being actively excluded. Restriction of oral intake is similarly unnecessary in that setting. High-flow oxygen is sometimes given on the reasoning that nitrogen washout speeds reabsorption of the trapped air, but that evidence is extrapolated from pneumothorax rather than demonstrated in pneumomediastinum, and it is not a routine requirement in a patient who is not hypoxic.
There is no randomised evidence governing length of stay. Published series report admissions ranging from one to five days, and comparisons of admitted patients against those discharged directly from the emergency department have not demonstrated a difference in complications or recurrence. Patients with ongoing pain, nausea, an untreated precipitant, or an incomplete work-up for oesophageal injury are the ones who benefit from an inpatient bed.
The finding that changes the plan is a pneumothorax alongside the mediastinal air. Where one is present, the decision to insert an intercostal drain follows the same reasoning as in any other pneumothorax: size, symptoms and physiological effect, rather than the presence of air alone. A small apical pneumothorax in a comfortable patient can be observed alongside the pneumomediastinum and followed on the repeat study. A symptomatic or enlarging pneumothorax, or one producing physiological compromise, is drained. How pneumothorax is treated →
Where a drain is inserted and the air leak does not settle, the situation is a persistent air leak and is managed on its own terms — suction, an endobronchial valve, or a blood patch. Mr Okiror is Co-PI at Guy’s and St Thomas’ for the NIHR-funded PRO-SEAL trial (ISRCTN15099654), which compares those options directly. How a persistent air leak is managed →
Surgery has no role in spontaneous pneumomediastinum itself. Operative intervention in this setting is directed at the underlying cause — repair of an oesophageal perforation, repair of a tracheobronchial injury, or decompression of tension physiology — and at a pneumothorax that behaves in a way that warrants it. Where a patient has both mediastinal air and a background of bullous lung disease, the bullous disease is what drives any surgical discussion. Bullectomy and giant bulla surgery →
Mediastinal air is reabsorbed over days to weeks, and occasionally remains visible on imaging for longer without clinical significance. Pain usually settles within the first few days. Patients are advised to avoid heavy lifting, straining and Valsalva manoeuvres until symptoms have resolved, and to return if pain worsens, breathlessness develops, a fever appears, or the swelling in the neck increases.
Recurrence is uncommon. Most published series report rates of around one per cent, although a 2023 cohort reported 17 per cent, and asthma is the risk factor identified most consistently. Long-term follow-up is not needed for every patient; it is worth arranging where the precipitant is unresolved, where there were respiratory symptoms the episode did not explain, or where the patient has had more than one episode.
Two activities need specific advice. Flying is reasonable once the air has resolved and the patient is asymptomatic; the reasoning behind air-travel advice after intrathoracic air is set out on the flying after a collapsed lung page. Diving is a different matter. A history of pneumomediastinum is generally treated as a contraindication to scuba diving, on the same reasoning that applies to pneumothorax, and anyone who dives should have fitness to dive assessed formally before returning to the water rather than relying on symptom resolution alone.
Most patients with spontaneous pneumomediastinum are managed by the admitting team without a thoracic surgeon becoming involved, and that is appropriate. The situations where an opinion changes something are set out below. Referrals are accepted from emergency departments, acute medicine, respiratory and general practice, and self-referrals are welcome.
For the private pathway, referral routes and practical information for referring clinicians, see For GPs and Referring Clinicians. Patients who have been discharged after an episode and want an independent review of the imaging and the advice they were given can be seen at London Bridge Hospital or The Lister Hospital Chelsea within 2–3 working days. Specialist second opinion →
References
Questions from referring teams and from patients who have been told there is air in the chest. For a collapsed lung rather than mediastinal air, see the pneumothorax page →
Book a Consultation →Or call Jo Mitchelson:
020 7952 2882
Bringing the imaging to the consultation allows the scan to be reviewed directly rather than summarised from the written report. Appointments at London Bridge Hospital and The Lister Hospital Chelsea within 2–3 working days. Self-referrals welcome.
Jo Mitchelson, PA · 020 7952 2882 · pa@lungsurgeon.co.uk
St Thomas’ Hospital #1 UK · Guy’s Hospital #2 UK · London Bridge Hospital #10 UK · Newsweek World’s Best Hospitals 2026
When air is in the pleural space rather than the mediastinum — drainage, and surgery to prevent recurrence
Flying After a Collapsed LungAir-travel advice after intrathoracic air, and the reasoning behind the intervals
Respiratory Symptoms in Young AdultsWhen cough or haemoptysis in a young patient needs a CT, and why a normal chest X-ray does not settle it
Unexplained Chest PainChest pain with no cardiac cause — the chest wall, rib and mediastinal causes that get missed
Persistent CoughA cough lasting beyond three weeks — the commonest precipitant, and when it needs investigating in its own right
Chest ImagingWhat each scan is asked to answer — chest X-ray, CT with and without contrast, and when CT adds something
Central Airway InterventionsBronchoscopy and airway surgery — the route when tracheobronchial injury is the question
Pleural Surgery in 2026 — Clinical ReferenceBTS 2023 pneumothorax management, empyema and pleural malignancy. For GPs and specialists
Pleural Drainage — Clinical ReferenceHow chest drains work, air-leak interpretation, and the management of a persistent air leak
Genetic Causes of a Collapsed LungFamilial pneumothorax and Birt-Hogg-Dubé syndrome — when recurrent intrathoracic air has an inherited cause
For GPs & Referring CliniciansReferral routes, acknowledgement times, insurance recognition and outpatient locations
Specialist Second OpinionIndependent review of the imaging and the advice given, within 2–3 working days