Pneumatosis intestinalis (PI), defined as gas within the bowel wall, is an uncommon radiographic sign which can represent a wide spectrum of diseases and a variety of underlying diagnoses. Because its etiology can vary greatly, management of PI ranges from surgical intervention to outpatient observation (see, Greenstein et al. (2007), Morris et al. (2008), and Peter et al. (2003)). Since PI is infrequently encountered, clinicians may be unfamiliar with its diagnosis and management; this unfamiliarity, combined with the potential necessity for urgent intervention, may place the clinician confronted with PI in a precarious medical scenario. We present a case of pneumatosis intestinalis in a patient who posed a particularly challenging diagnostic dilemma for the primary team. Furthermore, we explore the differential diagnosis prior to revealing the intervention offered to our patient; our concise yet inclusive differential and thought process for rapid evaluation may be of benefit to clinicians presented with similar clinical scenarios.
Pneumatosis intestinalis (PI) is defined simply as the radiographic finding of gas within the bowel wall. PI is an uncommon entity which has recently come to increased clinical attention due to improved radiographic identification; others have cited the radiographic incidence of PI to be present in 0.37% of patients who have abdominal CT scans [
Since PI can represent a wide range of pathology, it is in itself not diagnostic of any certain condition. This finding can represent pathologies which range from life-threatening to benign, and, for this reason, management of PI can range from urgent surgical intervention to outpatient observation. Delineating between these etiologies may be difficult but is important to establish with overall mortality estimated to be between 20% and 25% [
An 84-year-old white female with a history of atrial fibrillation, remote history of colorectal cancer treated with a partial colectomy with colostomy and postoperative radiation, and remote history of numerous small bowel obstructions necessitating resection of the terminal ileum with mild residual short gut syndrome presented with a 3-week history of diarrhea and cramping abdominal pain. The diarrhea consisted of six to eight episodes of voluminous, watery, loose, and brown stool without hematochezia or melena. Pain was mild, cramping, diffuse, and worse with food. There was no tenesmus, and the patient denied nausea or vomiting. No fever, chills, myalgias, or systemic signs of illness were present. Past surgical history was notable for remote hip arthroplasty and lower lumbar fusion. Review of systems was otherwise unremarkable.
Physical examination revealed a temperature of 36.8°C, blood pressure of 132/68 mm Hg, pulse of 72 beats per minute, 18 respirations per minute, and oxygen saturation of 99% on ambient air. The patient was a nontoxic female in no acute distress. Cardiovascular exam showed an irregularly irregular rhythm with regular rate. Pulmonary examination demonstrated lungs which were clear to auscultation; there was no wheezing, rales, or rhonchi. Her abdominal exam revealed a soft abdomen with no significant tenderness to light palpation; however there was diffuse tenderness to deep palpation. Brown, loose stool, and gas was found in the patient’s colostomy bag. There was no rebound or guarding, and bowel sounds were normal. Initial laboratory investigations were unremarkable, and pertinent negative labs included a normal white blood cell count and nonelevated lactate (Table
Laboratory data.
Variable | Reference range, adults | On admission | ||||||
---|---|---|---|---|---|---|---|---|
Hemoglobin (g/dL) | 12.0–15.5 | 13.6 | ||||||
Hematocrit (%) | 34.9–44.5 | 42.5 | ||||||
White cell count (per mm3) | 3500–10500 | 4400 | ||||||
Platelet count (per mm3) | 150,000–450,000 | 169,000 | ||||||
Sodium (mEq/dL) | 135–145 | 145 | ||||||
Chloride (mEq/dL) | 100–108 | 106 | ||||||
Potassium (mEq/dL) | 3.5–5.1 | 4.7 | ||||||
Bicarbonate (mEq/dL) | 22–29 | 28 | ||||||
BUN (mg/dL) | 12–21 | 26 | ||||||
Creatinine (mg/dL) | 0.7–1.2 | 1.4 | ||||||
Ionized calcium (mg/dL) | 4.7–5.4 | 4.6 | ||||||
Ionized magnesium (mmol/L) | 0.50–0.73 | 0.40 | ||||||
Lactate (mmol/L) | 0.9–1.7 | 0.5 | ||||||
Alkaline phosphatase (units/L) | 55–142 | 92 | ||||||
AST (units/L) | 12–31 | 20 | ||||||
ALT (units/L) | 9–29 | 12 | ||||||
Total bilirubin (mg/dL) | 0.1–1.1 | 0.3 | ||||||
Direct bilirubin (mg/dL) | 0.0–0.3 | 0.1 |
The CT scanogram demonstrated left lower quadrant bowel gas-filled loops, with one loop demonstrating an additional crescentic gas lucency (Figure
Scout film from abdominal and pelvic CT scan shows crecentic lucencies in the walls of the stool-filled cecum and ascending colon, compatible with pneumatosis.
Axial noncontrast image of the pelvis shows a curvilinear collection of gas in the dependent wall of the large bowel (arrows). The dependent location of the gas helps distinguish pneumatosis from intraluminal air.
Based on the radiographic finding of air within the bowel wall seen on CT, the patient was diagnosed with pneumatosis intestinalis (PI). The pathogenesis of pneumatosis is currently thought to be the result of many contributing factors [
The composition of the gas which results in PI has been well studied and gives insight into PI’s formation and etiology. The gas-filled colonic cysts in patients with chronic pneumatosis intestinalis have previously been identified as containing nitrogen, hydrogen, and carbon dioxide, but interestingly not the typical colonic gas methane [
This is where the “counterperfusion” aspect comes into play and provides a model whereby supersaturation can occur without increases in pressure. In this model, two inert gases attempt to travel across their normal diffusion gradients, but this exchange of gases is limited by two resistances. In the human model, the two inert gases are hydrogen, which is typically found in the lumen of the gut due to production by bacteria, and nitrogen, which is found in the blood as a result of pulmonary gas exchange. It had been identified that many patients with idiopathic PI are H2 superproducers [
Furthermore, other authors have examined the relationship between alkyl halides and pneumatosis intestinalis [
Prior authors have devised multiple methods to classify the expansive differential diagnosis which encompasses PI. A more inclusive list of conditions associated with PI is discussed elsewhere [
Prominent causes of pneumatosis intestinalis.
Nonurgent | Life-threatening |
---|---|
Traumatic | (i) Ischemia |
(i) Surgical anastamosis | (ii) Infarction |
(ii) Endoscopy | |
Infectious | Traumatic/mechanical |
(i) Clostridium difficile | (i) Volvulus |
(ii) HIV and AIDS | (ii) Malrotation |
(iii) Cryptosporidium | (iii) Intussusception |
(iv) Cytomegalovirus | (iv) Obstruction/strangulation |
(v) Pneumocystis carinii | (v) Blunt abdominal trauma |
(vi) Rotavirus | |
(vii) Adenovirus | |
Inflammatory/Autoimmune | |
(i) Crohn's Disease | |
(ii) Ulcerative Colitis | |
Other | |
(i) Graft versus host | |
(ii) Pseudo-obstruction | |
(iii) Immunosuppression | |
(iv) Iatrogenic | |
Pulmonary | |
(i) Asthma | |
(ii) COPD | |
(iii) Cystic fibrosis |
When considering ischemia or infarction, rapid recognition of the overall clinical picture is imperative. There are several key features of the current clinical scenario, patient past medical history, physical exam, and laboratory data, which can heighten suspicion for life-threatening PI (Table
Findings concerning for mesenteric ischemia and infarction within pneumatosis intestinalis.
Clinical Scenario |
---|
(i) Low-flow states |
(a) CHF |
(b) Sepsis |
(c) IV pressors |
(d) Hypotension |
(ii) Arrhythmias |
Past medical history |
(i) Vascular disease |
(ii) Risk factors for vascular disease: CAD, HTN, hyperlipidemia, diabetes, smoking |
Physical examination |
(i) “Pain out of proportion to exam” |
(ii) Peritonitis |
Laboratory data |
(i) Elevated lactate/acidemia |
Radiographic details |
(i) Gas within vasculature |
(ii) Linear/crescentic gas pattern |
(iii) Small bowel gas |
Furthermore, details of the radiographic findings can help define the nature of PI. The finding of additional gas in the vasculature, particularly portal venous gas, can be an ominous sign and correlates to transmural bowel necrosis [
After the more serious etiologies have been ruled out, nonurgent pathologies should be considered. Mechanical trauma such as recent surgical anastomosis and endoscopy can impair the normal mucosal barrier and may represent the underlying cause [
The cause of pneumatosis intestinalis is not exclusively limited to GI tract pathology. While debatable, it has been suggested that cystic fibrosis, asthma, and other obstructive bronchopathologies such as COPD can cause PI [
Interestingly, location of the pneumatosis intestinalis which can be anywhere from the stomach to the rectum may guide the clinician toward the cause. More proximal disease can represent rarer causes of PI such as pyloric stenosis, gastric ulcer, or gastric malignancy while distal lesions may stem from pathology such as appendicitis or diverticulitis.
Taking a “worst case scenario” approach may be of particular benefit towards patients with PI. This tactic will help the clinician in expeditiously ruling out urgent pathology while ultimately leading to a diagnosis. In this particular instance, the clinical context and radiographic findings did not suggest urgent pathology. Nevertheless, a surgical consultation was ordered as the patient had “pain out of proportion to the physical exam” and therefore could have been suffering from subacute bowel infarction. After evaluation by surgery, it was noted that the lactate was normal and urgent surgical exploration was not deemed necessary. With this and other serious pathologies also ruled out, pneumatosis intestinalis secondary to a less emergent cause was considered. While the patient did have a history of COPD, absence of signs and symptoms of COPD exacerbation and history of prior CTs without evidence of PI make COPD an unlikely cause. Further workup revealed a negative stool WBC and Clostridium difficile toxin. With no other plausible cause of PI, the patient’s presentation was ultimately thought to represent a mild bacterial infection yielding minimal clues to the definitive diagnosis aside from radiographic data. The patient was treated with metronidazole and her diarrhea resolved without sequelae, further reinforcing the diagnosis.
It is important to recognize that pneumatosis intestinalis is a clinical sign and is in itself not a diagnosis. Because it represents such a wide spectrum of diseases, rapid evaluation of PI may be difficult. An efficient recognition of the clinical scenario, encompassing the current clinical context, comorbid conditions, physical examination findings, laboratory data, and radiographic details, assists the clinician in reaching the correct diagnosis and offering appropriate treatment.