Pneumoperitoneum in the absence of intestinal injury in blunt thoracic and abdominal trauma: A case series with review of literature

INTRODUCTION

Traditionally, pneumoperitoneum following blunt torso trauma is considered a surgical emergency and is due to visceral perforation in 85 to 95% of cases.¹ In approximately 5 to 15 per cent of patients, pneumoperitoneum is not caused by perforation of a hollow viscus in the abdomen but may instead be related to a pathology involving the thorax, reproductive organs, or other abdominal conditions.² Air can enter the peritoneal cavity following blunt and penetrating chest traumas, either via normal pathways such as diaphragmatic hiatuses or through abnormal routes like congenital defects or post-traumatic diaphragmatic injuries.³ It can also be caused by the ‘Macklin effect’, where air dissects along perivascular sheaths into the mediastinum and tracks into the peritoneal cavity via diaphragmatic openings.⁴

In a retrospective review done over seven years, Kane et al. reported 18 blunt trauma patients with free intraperitoneal air on Computed Tomography (CT) images, excluding those with penetrating injuries or pre-CT peritoneal lavage.⁵ In this study, bowel injury was identified intra-operatively in only four cases (22%). Pneumothorax and mechanical ventilation were identified as alternative sources of pneumoperitoneum in the remaining patients.⁵ In another study, Nishina et al. reported on the patients who were managed following blunt thoracoabdominal trauma over 17 years.⁶ Of the 233 patients, 32 were positive for free air on radiograph; of these, 25 had a gastric perforation that was managed surgically. However, seven patients had no intra-abdominal hollow viscus injury, although all had pneumothorax in common and other intra-abdominal injuries.⁶

Historically, the criteria for performing laparotomy in patients with blunt abdominal trauma included the detection of free intraperitoneal air on imaging, evidence of intraperitoneal haemorrhage, or the presence of peritoneal signs on clinical examination. In this case series, we examine the management of a group of patients with pneumoperitoneum on radiological imaging without the presence of hollow viscus injury after blunt trauma.

CASE SERIES

Non-operative management

Case 1

A 47-year-old woman presented to the Emergency Department (ED) seven days after falling from a two-wheeler, exhibiting tachypnea, tachycardia, and subcutaneous emphysema. She was hypotensive but responsive to fluid resuscitation. An Extended Focused Assessment with Sonography for Trauma (e-FAST) revealed a left-sided hemopneumothorax, necessitating the insertion of a left intercostal drain (ICD) that immediately released air and blood. Due to ongoing respiratory distress, she was intubated and ventilated. Abdominal examination did not reveal any peritoneal signs.

A CT scan from head to pelvis revealed pulmonary contusions, residual mild pneumohemothorax, pneumomediastinum, and a suspected area of discontinuity in the anterior right diaphragm with air tracking into the abdominal wall and pneumoperitoneum, but no hemoperitoneum or bowel injury [Figure 1a and b]. Respiratory failure precluded intervention for suspected diaphragmatic rent, and she was managed non-operatively with close monitoring. She was weaned and extubated by day three, tolerated enteral feeding, and was discharged after ten days, remaining well at follow-up.

Close

Figure 1:

(a) The yellow arrow shows pneumoperitoneum surrounding small bowel loops. The yellow star shows subcutaneous emphysema, and the blue star shows air in the space of the retzius. (b) Sharpened and reformatted coronal images showing multiple rents along the right diaphragm with underlying pneumoperitoneum (Yellow arrow)

Export to PPT

Case 2

A 53-year-old man was brought to the hospital two hours after a motorcycle crash. He was gasping and was quickly intubated. Examination revealed decreased bilateral breath sounds, subcutaneous emphysema, and crepitus, leading to the insertion of bilateral chest drains. An e-FAST exam was negative; his Glasgowcoma scale (GCS) score was 6T/15 (E1VTM5), with pupils reacting to light. An abdominal exam was normal.

Imaging from head to pelvis showed a left acute subdural hematoma with associated contusions and significant mass effect. Further findings included bilateral pneumothoraces, extensive subcutaneous emphysema in the neck and retropharyngeal area, pneumomediastinum, rib fractures on the right side, and pneumoperitoneum tracking along the falciform ligament, with no intraperitoneal free fluid.

He underwent emergency decompressive craniectomy, during which a thick subdural hematoma and temporal lobe contusion were identified. Unfortunately, periprocedurally, the patient developed respiratory failure with cardiac arrest, and he succumbed to his illness.

Case 3

A 24-year-old woman, riding as a pillion passenger on a motorcycle, was admitted three hours after a road traffic incident (RTI). She experienced multiple episodes of vomiting and was intubated at a primary care centre before transfer. On primary survey, her GCS was 5T/15 (E1VTM4) with bilaterally reactive pupils and a normal abdominal exam.

Imaging revealed bilateral mild to moderate pneumothorax, subcutaneous emphysema, irregularities in the right bronchus intermedius and distal trachea, pneumomediastinum, and ild pneumoperitoneum with retroperitoneal air, without apparent bowel injury. Bilateral chest tubes were inserted; she was admitted to the trauma intensive care unit (TICU). Suspected tracheal injury was managed non-operatively. The chest tubes were removed on days 2 and 4. A tracheostomy was performed on day 4 due to prolonged intubation. Thereafter, her neurological status gradually improved, allowing her to be weaned from ventilation, and she tolerated an oral diet. She was transferred to the rehabilitation department and eventually discharged.

Case 4

A 43-year-old man was admitted three hours after an RTI. He was intubated due to a GCS score of 8 (E1V2M5), with reactive pupils. An initial e-FAST and abdominal exam were negative.

Imaging from the neck to the pelvis revealed a small right pneumothorax, bilateral lung contusions, minimal bilateral hemothorax, and a suspected right diaphragmatic injury with an elevated right hemidiaphragm. Additional findings included a right adrenal hematoma, fractures of the right ribs and clavicle, and diffuse hepatic steatosis, with no intra-abdominal free fluid. Air was observed beneath the right diaphragm and abdominal wall, consistent with pneumoperitoneum.

The patient was managed in the TICU, where a tracheostomy was performed for prolonged ventilation, and he was eventually weaned to room air. Repeat brain imaging showed no change. The bilateral pneumothorax, diaphragmatic injury, and pneumoperitoneum were managed non-operatively, as there was no respiratory compromise or bowel herniation, with the liver acting as a barrier to prevent the latter. Moreover, the management of severe head injury precluded non-emergent abdominal surgery. The patient remained stable, tolerated enteral feeding, and was discharged in good condition.

Case 5

A 19-year-old man presented two and a half hours after an RTI involving a two-wheeler collision. The primary survey was normal. GCS was E3V4M6; pupils were 2 mm and reactive to light. A right humerus fracture was present. The e-FAST revealed mild-moderate free fluid with echoes in the left hypochondrium, suggesting hemoperitoneum. No pleural/pericardial effusion or pneumothorax was seen. Abdominal exam was unremarkable.

CT revealed right middle/lower lobe lung contusions, a thin rim of right pneumothorax, a grade III liver laceration, grade II right renal injury, thin rim of pneumoperitoneum, mild-moderate hemoperitoneum, and cerebral oedema. He was admitted to the TICU for close monitoring. The right humerus fracture was immobilised. He remained stable and tolerated the diet well, but was eventually discharged against medical advice after 48 hours of care due to financial constraints.

Case 6

A 25-year-old man presented eight hours after being involved in an RTI. He was intubated and had a left ICD placed at another hospital. Primary survey revealed decreased breath sounds on the left, tachycardia, and responsive hypotension. His GCS was 2T/15 (E1VTM1) with a fixed and dilated left pupil(4 mm) and the right pupil obscured by oedema. The e-FAST exam revealed absent lung sliding on the left side. Abdominal exam was unremarkable.

CT head to pelvis revealed bilateral frontal contusions, subdural and subarachnoid haemorrhages, extensive facial fractures, left moderate pneumothorax (ICD in situ), 3rd rib fracture, moderate pneumomediastinum, suspected right posterolateral tracheal injury, possible small anterior diaphragmatic defects with subdiaphragmatic significant pneumoperitoneum and minimal pelvic fluid, grade 2 liver injury, grade 1 splenic injury [Figure 2].

Close

Figure 2:

Diaphragmatic defect (yellow arrows) with pneumoperitoneum (yellow stars) below the domes. The blue stars indicate pneumomediastinum and left pneumothorax

Export to PPT

An additional ICD was placed on the left side for the persistent hemopneumothorax. Pneumomediastinum and pneumoperitoneum were attributed to the extension of pneumothorax and the tracking of air via the small anterior diaphragmatic defects. In view of the radiological findings strongly suggesting air tracking from the chest, the absence of obvious features of bowel injury, and the lack of abdominal signs, the patient was managed nonoperatively. He remained stable, but was discharged against medical advice 48 hours after injury due to financial constraints.

Operative management

Case 7

A 59-year-old man presented eight hours after a two-wheeler RTI. He was intubated for low sensorium and had a right hemothorax (ICD in situ). The primary survey revealed a transient response to blood and a GCS of 2T/15 (E1VTM1) with anisocoria.

CT head to pelvis reported bilateral parietal subarachnoid haemorrhage, minimal right parietal contusion, right lateral tracheal rent (Thoracic D2–D3 level), pneumomediastinum, bilateral mild pneumothorax, multiple rib fractures, mild-moderate pneumoperitoneum extending to upper pelvis, mesenteric and retroperitoneal air, no hollow viscus or intra/extraperitoneal fluid.

In view of transient responder status, he underwent emergency exploratory laparotomy and drain placement. Intraoperatively, pneumoperitoneum, pneumoretroperitoneum, mesenteric tear with retroperitoneal hematoma at the rectosigmoid, and small mesenteric hematomas were noted with no evidence of bowel injury. Postoperative course in the ICU was uneventful, and he was discharged for further neurorehabilitation to another centre.

Case 8

A 39-year-old man presented 4 hours following a fall from a two-wheeler. Primary survey revealed respiratory distress requiring intubation, subcutaneous emphysema warranting bilateral chest tubes, a GCS of 14/15, and anisocoria. He had an open cranial fracture with cerebrospinal fluid (CSF) leak and exposed brain tissue.

CT head to pelvis showed a left frontal hemorrhagic contusion, depressed frontal fracture, temporal and cribriform plate fractures, extensive subcutaneous emphysema, bilateral mild pneumothorax (ICD in situ), pneumomediastinum, pneumoperitoneum, air tracking along the mesenteric root, and grade 1 liver contusion. [Figure 3a and b]

Close

Figure 3:

(a) Pneumoperitoneum (yellow stars) under the right dome of the diaphragm, pneumoretroperitoneum tracking along the diaphragmatic crura (yellow arrows), (b) Sagittal view

Export to PPT

Emergency frontal craniectomy and lumbar drain were performed. Despite the presence of pneumoperitoneum and pneumomediastinum, the abdomen was soft, and imaging revealed no intra-abdominal fluid; the patient was initially managed non-operatively. On postoperative day 1, fever prompted a suspicion of meningitis (a lumbar puncture was normal). Persistent fever and pneumoperitoneum led to exploratory laparotomy: findings included pneumoperitoneum, lesser sac and Morrison’s pouch seropurulent fluid and pus flakes, but no bowel injury. He recovered and was discharged after initiation of a regular diet. The management and outcomes of all eight cases are summarised in Table 1.

Table 1: Summary of management and outcomes in patients with traumatic pneumoperitoneum

Case	Age/ sex	Mechanism	Associated injuries/ findings	Etiology of pneumoperitoneum, proposed mechanism	Management	Outcome
1	47/F	Fall from a 2-wheeler	Pulmonary contusions, left hemopneumothorax, pneumomediastinum, pneumoperitoneum; diaphragmatic rent; no bowel injury	Thoracic injury, Macklin effect, Diaphragmatic rent	Observation, ICD, ventilation	Recovered, discharged
2	53/M	RTI, 2-wheeler	Subdural hematoma, temporal/frontal contusions, bilateral pneumothorax, pneumomediastinum, pneumoperitoneum, rib fractures, no free fluid	Thoracic injury, Macklin effect	Decompressive craniectomy for head injury	Died intra-op (non-abdominal cause, respiratory failure, severe head injury)
3	24/F	RTI, pillion	Bilateral pneumothorax, tracheal/bronchial injury, pneumomediastinum, pneumoperitoneum; no bowel injury	Airway injury, Macklin effect	ICDs, tracheostomy, observation	Recovered, discharged
4	43/M	RTI	Lung contusions, minimal pneumothorax/hemothorax, suspected diaphragmatic hernia, adrenal hematoma, hepatic steatosis, pneumoperitoneum.	Thoracic/diaphragm injury, Possible Diaphragmatic defect	Tracheostomy, observation	Recovered, discharged
5	19/M	RTI, 2-wheeler	Lung contusions, hemoperitoneum, liver (Gr III) & renal injury, right humerus fracture, pneumoperitoneum	Thoracic injury, Macklin effect	Observation, immobilization	Discharged against medical advice, Lost to follow up
6	25/M	RTI	Frontal contusions, facial fractures, pneumothorax, pneumomediastinum, tracheal injury, diaphragmatic defects, liver (Gr II), spleen (Gr I) injury, pneumoperitoneum	Thoracic injury, Macklin effect, Possible Diaphragmatic defect	ICDs, wound wash, observation	Discharged against medical advice, Lost to follow up
7	59/M	RTI, 2-wheeler	Subarachnoid hemorrhage, tracheal rent, rib fractures, pneumomediastinum, pneumoperitoneum, mesenteric tear, retroperitoneal air and hematoma	Airway injury; Macklin effect with air tracking into the retroperitoneum and thereafter into the peritoneum through a retroperitoneal defect in the mesentery	Exploratory laparotomy	Discharged against medical advice, lost to follow-up
8	39/M	RTI, 2-wheeler	Frontal contusion, depressed skull fracture, CSF leak, subcutaneous emphysema, mild liver contusion, pneumoperitoneum	Thoracic injury, Macklin effect	Craniectomy, lumbar drain, laparotomy	Recovered, discharged

DISCUSSION

In as many as 5 to 15 per cent of patients, pneumoperitoneum on imaging may not be the result of perforation of a hollow viscus.² We conducted a literature review of studies published in the last 25 years, using the following keywords: ‘Macklin effect’, ‘pneumoperitoneum’, ‘pneumothorax’, and ‘trauma’ in MEDLINE and Google Scholar, and identified seven articles that were similar to our own. The key findings from these studies are summarized in Table 2.^7-13

In our case series, the finding of air on imaging in the peritoneal cavity was not due to underlying abdominal injury. A recent retrospective screening of 492 CT scans of the chest, abdomen, and pelvis by Currin et al. found that pneumoperitoneum in cases of blunt trauma, when no other causes are evident, could be a benign posttraumatic finding occurring in about 5% of patients and is seen in high-velocity injuries.⁷ Similarly, Marek et al. (2014) evaluated 78 cases and concluded that CT scans may detect free air that is not always clinically significant. In their study, as high as 61% of the patients had intra-abdominal free air that was attributable to causes other than substantial intra-abdominal injuries.¹⁴ The study suggested that the coexistence of free fluid, seatbelt sign, or signs of bowel trauma alongside pneumoperitoneum may strongly suggest an underlying injury.¹⁴

In our series, the degree of pneumoperitoneum was described as mild to moderate in all eight patients. As seen in existing literature, most of our patients had accompanying thoracic injuries-pneumothorax, pneumomediastinum, rib fractures, pulmonary contusions, or subcutaneous emphysema. Goyal et al. reported a young male who underwent exploratory laparotomy for pneumoperitoneum following a high-velocity motor vehicle crash, with no evidence of injury to the gastrointestinal tract or diaphragm.¹⁵ They highlighted the association of panfacial fractures and extensive subcutaneous emphysema in the neck area with the presence of pneumoperitoneum.¹⁵ Similarly, in 2014, Hakim et al. described a case of blunt neck trauma from a motorcycle rider crashing into a stationary object, which caused severe laryngotracheal injury.⁸ The patient later developed bilateral pneumothoraces and pneumoperitoneum, despite no injury to the internal viscera.⁸

The proposed mechanisms for the accumulation of free air are many. Ectopic gas was thought to accumulate near the costochondral junction due to a vacuum phenomenon caused by traumatic impact.⁷ Macklin, in 1937, described the phenomenon where alveolar rupture caused by trauma allows air to dissect along the bronchovascular bundles, resulting in the spread of pulmonary interstitial emphysema and, subsequently, pneumomediastinum, now known as the Macklin effect.⁴ This air can then track through existing diaphragmatic hiatuses or trauma-related small defects into the intraperitoneal and retroperitoneal compartments [Figure 4]. Wintermark and Schnyder reported that 39% of severe blunt traumatic pneumomediastinum cases detected by CT exhibit the Macklin effect.¹⁶ Grosfeld et al. demonstrated that in cases of pneumothorax and pneumomediastinum, the pressure inside the chest exceeds that in the abdomen, unlike normal circumstances, where the opposite is true.¹⁷ This reversal of pressure causes air to shift from the chest cavity into the abdominal space, leading to pneumoperitoneum.¹⁷ In some cases, the Macklin effect may lead to malignant pneumomediastinum, a potentially fatal condition requiring prompt intervention to prevent pulmonary oedema and circulatory failure. The spread of air from the mediastinum into the retroperitoneal space and then into the peritoneal cavity can even cause intra-abdominal hypertension.¹⁸

Close

Figure 4:

Alveolar rupture with air leak tracking along the bronchovascular sheaths (Macklin Effect), possible routes via which air enters the peritoneal cavity.

Export to PPT

All eight patients who presented to our centre had significant thoracic injuries, which suggests that the Macklin effect could be the primary explanation for the development of pneumoperitoneum in the majority of them. Two of these patients had additional injuries to the airway, causing massive leakage of air, while a few others had diaphragmatic rents, which may have facilitated the phenomenon. The diaphragmatic tears in these patients may not have been large, but just enough to allow air to escape.

Two patients required an exploratory laparotomy, as there were features of hemodynamic compromise (case 7) or persistent sepsis (case 8), although no intra-abdominal pathology could be identified to explain the pneumoperitoneum. This further emphasises the need for close monitoring of these patients and an individualised approach to management. Careful assessment to look for any signs of lingering sepsis, peritonitis, or any deviation from the normal course should alert the managing team to rule out hollow viscus injury.

A recent review of studies from 1990 to 2015, by AbdelAziz et al, analysed the largest patient group to date (11,924 individuals) and found that free peritoneal fluid was the most sensitive indicator of bowel injury on CT scan.¹⁹ When there was free fluid without solid organ injury, sensitivity was lower (53%), but specificity was higher (81%). Bowel wall thickening and mesenteric stranding were also significant. Highly specific signs of bowel injury included bowel wall hematoma, oral contrast leakage, and free intraperitoneal air without pneumothorax.²⁰ However, the interpretation of both clinical and radiological signs must be tailored to each case for successful management.

Hoover et al.¹⁹ suggested that the rate of negative laparotomies associated with pneumoperitoneum could be reduced without missing significant pathology by following a few key steps: (a) identifying specific conditions known to be linked with nonsurgical spontaneous pneumoperitoneum; (b) integrating patient history, physical examination, laboratory findings, and radiographic imaging to assess the likelihood of a non-surgical cause; and (c) using careful observation, including regular clinical assessments by experienced healthcare providers, along with diagnostic procedures such as peritoneal taps and lavage, especially in cases where the diagnosis is uncertain.¹⁹ In our series, of the six patients managed non-operatively, five were discharged (three directly, two discharged against medical advice (AMA) – lost to follow-up). Only one patient died, but the death was attributed to respiratory failure and severe head injury, and not to abdominal complications. None of the patients in the non-operative group required surgical intervention at a later date.

We acknowledge that this case series has several limitations, including its retrospective nature, a relatively small number of cases, incomplete follow-up for patients who left against medical advice, and the subjective grading of pneumoperitoneum. We appreciate the challenges involved in conducting such studies and recognise the importance of these findings while also emphasising the need for further research.

CONCLUSION

Pneumoperitoneum can be present in the absence of hollow viscus injury in blunt abdominal or thoracic trauma. Carefully selecting patients in whom there is a lack of clinical or imaging evidence of bowel injury alongside an identifiable overwhelming thoracic injury allows for successful non- operative management of pneumoperitoneum, resulting in favourable outcomes.