Pneumatosis cystoides intestinalis (PCI) in a patient with undiagnosed systemic sclerosis

Abstract

Pneumatosis cystoides intestinalis (PCI) refers to the presence of gas within the wall of the small or large intestine. The pathophysiology is incompletely understood and is probably multifactorial in nature. PCI is a known but rare complication of systemic scleroderma, and the aetiology of PCI in patients with scleroderma is not fully understood. We present the case of a patient who was referred to gastroenterology clinic by her general practitioner for investigation of 8 months of weight loss, urgency, diarrhoea, bloating and crampy abdominal pain. Extensive investigations were performed to exclude infective, inflammatory or malignant aetiologies for these symptoms. She was diagnosed with PCI on her colonoscopy and was subsequently screened for secondary causes. Our patient was diagnosed with the limited cutaneous (CREST) variant of systemic scleroderma. This case report illustrates that PCI could be an uncommon presentation of systemic sclerosis, therefore clinicians should be aware of the association between these conditions.

Background

This case report illustrates that pneumatosis cystoides intestinalis (PCI) can be a diagnostic dilemma especially due to the non-specific nature of a patient’s presenting symptoms. As a result, other more common differentials such as coeliac disease, inflammatory bowel disease or even an underlying malignancy must be thoroughly assessed for and excluded. This case report illustrates that PCI could be an uncommon presentation of systemic sclerosis, therefore clinicians should be aware of the association between the two conditions.

Case presentation

A 52-year-old Caucasian Australian woman was referred to gastroenterology clinic by her general practitioner (GP) for investigation of 8 months of weight loss, urgency, diarrhoea, bloating and occasional crampy abdominal pain. All of these symptoms seemed to be relieved with evacuation. She was reporting up to four bowel motions a day including overnight sometimes and occasionally had blood mixed with her stools. During this time period, she had also lost 11 kg of weight unintentionally.

Her only medical history included gastro-oesophageal reflux disease for which she was taking pantoprazole 40 mg daily. She denied taking any other medications.

She was a non-smoker and admitted to being a social drinker that did not indulge in any binge drinking episodes. She was not aware of any family history of cancers, including pancreatic or colorectal cancer.

A focused cardiovascular examination was unremarkable. She was afebrile, and her breath sounds were normal without rales or bronchial obstruction. There was no sign of chronic obstructive pulmonary disease (COPD) or asthma on the completion of a respiratory examination.

Her abdomen was distended but soft, non-tender and regular peristaltic sounds were audible. There was no evidence of organomegaly, peritonitis, guarding or rebound tenderness. A per-rectal examination was unremarkable with normal anal tone and sensation and no blood on finger withdrawal.

A rheumatological examination of any of her joints was not performed during her gastroenterology clinic review.

Investigations

Routine bloods during gastroenterology clinic review were unremarkable. Her haemoglobin was noted to be 13.2 g/dL (normal range 13–18). She had a white cell count of 6.6×10⁹/L (normal range 4.0–11.0) and her platelets were 267×10⁹/L (normal range 150–450). Her urea electrolyes and creatinine (UECs) revealed a sodium level of 138 mmol/L (normal range 135–145), potassium of 4.3 mmol/ L (normal range 3.5–5.2), urea of 3.1 mmol/L (normal range 2.5–8) and estimated glomerular filtration rate (eGFR) of >90. The patient’s liver function tests were all within the normal parameters. Her lipase was 34 units/L (normal range <34), and her albumin was 37 g/L (normal range 35–52). She also had a C reactive protein of 4.1 mg/L (normal range <5).

The gastroenterology team arranged for the patient to have a stool sample to look at the faecal calprotectin and elastase. Both tests were normal. Her stool faecal culture was also negative for any growth. She also underwent a coeliac disease screen including total IgA, IgA anti-tissue transglutaminase (anti-tTG) antibody and IgG-deamidated gliadin peptides (anti-DGP) while on a gluten-rich diet. She had an anti-tTG antibody level of 9 units (normal range <20), and an anti-DGP level of 2 units (normal range <20). She also underwent human leukocyte antigen (HLA) coeliac disease testing and tested negative for HLA DQ2, HLA DQ8 and DQA105.

She subsequently underwent an urgent outpatient gastroscopy and colonoscopy. Her gastroscopy was unremarkable except for the discovery of grade A reflux oesophagitis which did not display any evidence of dysplasia or malignancy on biopsies. Her colonoscopy revealed a crop of bubbly projections that collapsed on insufflation distal to the splenic flexure (Figure 1). Positive ‘pillow sign’ was demonstrated, and the appearance was consistent with PCI. The remainder of the colon was unremarkable including a normal terminal ileum (Figure 1). Biopsies of the terminal ileum, right, transverse, left colon and rectum were within normal limits. The surface and crypt architecture appeared normal. The lamina propria contained a normal population of inflammatory cells, and there was no acute inflammation. No granulomas or parasites were seen. The appearances of the large intestinal mucosa were not consistent with collagenous or lymphocytic colitis. There was also no evidence of dysplasia or malignancy on the biopsies from the colonoscopy.

Figure 1.

Colonscopy images from the patient consistent with a diagnosis of pneumatosis cystoides intestinalis (crop of bubbly projections in the descending colon). The visualised caecum and terminal ileum are unremarkable.

In the setting of her diagnosis of PCI, she had bloods collected as part of an autoimmune screen immediately after her scopes to exclude a connective tissue disorder as a secondary cause for her PCI. Her GP was advised to follow up the results of her autoimmune screen. Our patient was subsequently referred to rheumatology outpatient clinic on the basis of her screening results by her GP (see the Differential diagnosis section).

Differential diagnosis

While awaiting the results of the above investigations and histopathology results of her gastroscopy and colonoscopy, the patient developed an episode of mild crampy abdominal pain. Owing to the fact that she lived in a rural area and was isolated without any social supports, she decided to call an ambulance. While in the emergency department, she had a CT scan of her abdomen which revealed multiple rounded well-defined gas lucencies within the wall of the distal transverse colon and splenic flexure consistent with PCI (Figure 2). Her abdominal pain subsided with some analgesia. She was also reporting some constipation and was given aperients in the emergency department with good effect. Her abdominal pain completely resolved in the department. As a result, she was conservatively managed and discharged back to the care of our gastroenterology team with a view to an outpatient clinic follow-up.

Figure 2.

Transverse and coronal views of the patient’s CT abdomen/pelvis highlighting gas lucencies within the wall of the distal transverse colon and splenic flexure consistent with pneumatosis cystoides intestinalis (red arrows).

Pulmonary disorders are often associated with the development of PCI but have been excluded in our patient.¹ Her haemoglobin A1c was also within normal range excluding diabetes. Other aetiologies that could have potentially caused her symptoms including colorectal cancer and inflammatory bowel disease were excluded on the basic of her scopes, and histopathology from her biopsies.

As part of her work-up, she also underwent an autoimmune screen to exclude a connective disorder as a secondary cause for her PCI. Interestingly, she tested positive for antinuclear antibodies with a centromere antibody titre of more than 640 (normal range <80). Her rheumatoid factor was 15 IU/mL (normal range <14) with antidouble stranded DNA antibody titre of 4 IU/mL (normal range 0–7). Her anticitrullinated protein antibody titre was <7 U/mL (normal range <17.1), and her C-anti neutrophil cytoplasmic antibodies (C-ANCA) titre was 55 U/mL (normal range <20). She tested negative for perinuclear anti neutrophil cytoplasmic antibodies (P-ANCA). Her complement studies were within normal parameters. As a result of her autoimmune screen, our patient was referred to the mixed connective tissue disorders clinic by her GP.

Treatment

She was started on metronidazole 500 mg three times a day to be taken for the next 3 months to assist with the clinical and radiographic resolution of PCI.² She was advised of the potential adverse reactions associated with metronidazole and advised to avoid alcohol while on the antibiotic.

She was also reviewed by the dietitians at gastroenterology clinic and was started on an oral elemental diet for 2 weeks.

We did not employ any oxygen therapy in our patient.

Outcome and follow-up

Centromere antibodies are most commonly associated with limited forms of scleroderma of the CREST variant and can be detected years in advance of symptoms specific to scleroderma.^3–5 As a result of her autoimmune screen, our patient was referred to the mixed connective tissue disorders clinic.

In rheumatology clinic, her history was further explored, and she was re-examined. While she did not display several of the cutaneous changes characteristically associated with CREST syndrome such as skin thickening and tightening, puffy fingers, calcinosis, digital pitting and tendon friction rubs, she did report Raynaud’s phenomenon especially during the cold weather and was noted to have characteristic telangiectasia and capillary changes on nailfold capillaroscopy. She also reported having to cut up pieces of meat (and other solids) into smaller portions as she was beginning to experience dysphagia and occasional chest pain after having a large meal.

On the basis of her clinic findings and testing positive for scleroderma-related antibodies, our patient met the 2013 American College of Rheumatology/European League Against Rheumatism criteria for the diagnosis of systemic sclerosis. The rheumatology team anticipated that she would eventually develop other typical cutaneous changes as the disease inevitably progressed. Her erythrocyte sedimentation rate was 9 (normal range 5–12), and her creatine kinase was not elevated during her rheumatology review.

She will continue to be monitored for cutaneous and extracutaneous organ involvement by the rheumatology team into the future.

She will also continue to be followed up by the gastroenterology team. In the setting of her symptoms of potential oesophageal hypomotility, she is also being planned for gastric emptying study in the outpatient setting.

Discussion

PCI refers to the presence of gas within the wall of the small or large intestine. The pathophysiology is incompletely understood and is probably multifactorial in nature. PCI is characterised by the presence of numerous gas-containing cysts within the subserosal or submucosal layer of the intestinal wall. The subserous cysts are most frequently found in the small bowel while the submucous localisations are predominantly seen in the colonic wall.^6–8

PCI is idiopathic (in 15%) or a secondary finding (in 85%) caused by a wide variety of underlying gastrointestinal or extragastrointestinal diseases such as autoimmune (scleroderma, dermatomyositis), inflammatory (inflammatory bowel disease) or infectious diseases (Clostridium difficile, HIV), pulmonary disease (asthma, COPD, cystic fibrosis), drugs (corticosteroids, immunosuppressive therapy) and trauma (blunt abdominal trauma, endoscopy).⁶

Symptoms of PCI can include but are not limited to diarrhoea, constipation, abdominal pain, rectal bleeding and mucous discharge. However, the majority of patients with PCI are asymptomatic and never come to clinical attention. PCI may be discovered incidentally on screening colonoscopy or sigmoidoscopy and is confirmed by abdominal CT.⁹

Various theories have been proposed to explain the pathogenesis of PCI, including biochemical, mechanical and bacterial causes. In spite of the fact that these proposed theories are diverse, they are not necessarily mutually exclusive, and it is highly likely that multiple pathogenic mechanisms play a role in the development of PCI.^{2 10–15}

The biochemical theory proposes that a patient’s luminal bacterial flora can produce excessive amounts of hydrogen gas through fermentation of carbohydrates and other digested food items. This process increases the pressure of the gas within the intestinal lumen forcing the gas through the mucosa and causing it to become trapped within the submucosa.^10–12 The mechanical theory hypothesises that increased abdominal pressure can predispose to gas dissecting into the bowel wall. Studies have shown that patients with previous bowel obstructions and conditions such as COPD, chronic constipation, inflammatory bowel disease, which are known to increase abdominal pressure, are at increased risk of PCI.^13–15 Last but not the least, the bacterial theory states that gas-producing bacteria, some of which include Escherichia Coli, C. difficile and C, perfringens, are able to breach the mucosal wall through a site of injury and form copious amounts of intramural gas. Over time, the intramural gas builds up and leads to the development of PCI.^{2 16}

PCI is a rare complication of scleroderma, and the aetiology of PCI in patients with scleroderma is not fully understood.^17–19 Gastrointestinal manifestations of the CREST variant of scleroderma include but are not limited to oesophageal dysmotility, gastroparesis, decreased oesophageal sphincter function and Barrett’s oesophagus.²⁰ The development of PCI is also dependent on the severity of the gastrointestinal tract involvement from scleroderma rather than overall disease duration which could explain the general lack of cutaneous and extracutaneous organ involvement in our patient.¹⁷ Furthermore, as mentioned previously, centromere antibodies are most commonly associated with limited forms of scleroderma of the CREST variant and can be detected years in advance of symptoms specific to scleroderma highlight the importance of screening patients for secondary causes of PCI.^3–5

The management of PCI is based on the severity of symptoms. The underlying cause of PCI should be treated in all patients, regardless of the presence of symptoms. Asymptomatic patients do not require any additional therapy.²¹ For patients with mild symptoms who can be managed as outpatients, a combination of antibiotics and an elemental diet is advised.^{22 23} In patients with moderate to severe symptoms who require hospitalisation, a combination of antibiotics, an elemental diet and oxygen therapy are recommended.^{24 25}

Surgery is generally reserved for patients with PCI who remain symptomatic despite medical therapy or who develop complications from PCI such as bowel obstruction, perforation, peritonitis and necrotic bowel.²⁶

Various case reports have demonstrated a link between the development of PCI in the setting of a connective tissue disorder particularly systemic scleroderma.^16–18 The case report by Kaneko and colleagues also included a literature review of a total of 39 patients with scleroderma of which, 23 had CREST syndrome and revealed that the median duration from the onset of scleroderma to the development of PCI was 6 years.¹⁷ However, these case reports followed up patients with a known diagnosis of a connective tissue disorder, and our case report is unique in the sense that we explored secondary causes for our patient’s PCI and diagnosed her with CREST syndrome which was previously undiagnosed.

Patient’s perspective.

I consent to my medical information being used for the purpose of this research article. I wish for my medical condition to help other medical practitioners in their research so that I can make a difference to other patients and families who are in a similar situation to me. I am grateful for the care that I received from the medical and nursing staff at Austin Hospital.

Learning points.

• Pneumatosis cystoides intestinalis (PCI) is idiopathic (15%) or secondary (85%) to a wide variety of gastrointestinal and non-gastrointestinal illnesses.

• The pathogenesis of PCI is likely multifactorial due to mechanical, bacterial and biochemical causes.

• The underlying cause of secondary PCI should be treated in all patients, regardless of the presence of symptoms.

• The CREST variant of scleroderma is associated with PCI, and antibodies can be detected years in advance of symptoms specific to the rheumatological condition.

• This case highlights the importance of screening patients for a secondary cause of PCI and the recruitment of multidisciplinary teams to manage such cases.