Pneumatosis cystoides intestinalis of the ascending colon related to acarbose treatment: a case report

Introduction Pneumatosis cystoides intestinalis is characterized by the presence of multiple gas-filled cysts in the intestinal wall, the submucosa and/or subserosa of the intestine. The term pneumatosis cystoides coli is synonymous with pneumatosis cystoides intestinalis when the disorder is limited to the colon. It is a secondary finding caused by a wide variety of underlying gastrointestinal or extragastrointestinal diseases but rarely occurs in the course of treatment with an α-glucosidase inhibitor. This is the first report of pneumatosis cystoides intestinalis after 12 years of treatment with the α-glucosidase inhibitor acarbose. Case presentation A 65-year-old Caucasian German woman was referred to our hospital for hemicolectomy. She had been treated for type 2 diabetes mellitus with an α-glucosidase inhibitor (acarbose, 150 mg daily) for 12 years. Three months before referral, she had complained of left abdominal pain. 'Polyposis coli' in the ascending colon and diverticulosis were diagnosed. Colonoscopy and computed tomography scans of the abdomen were repeated and revealed pneumatosis cystoides coli located in the ascending colon, whereas diverticulosis of the sigmoid colon was confirmed. Histological examination of a biopsy specimen only showed colon mucosa. After discontinuing administration of the α-glucosidase inhibitor for 3 months and on repeated colonoscopy, the polypoid lesions had completely disappeared. Conclusion This case illustrates that pneumatosis cystoides coli can be a source of diagnostic confusion. Pneumatosis cystoides coli must be considered in the initial differential diagnosis of patients especially in the presence of multiple colonic polypoid lesions. It is important to take pneumatosis cystoides intestinalis into consideration when prescribing α-glucosidase inhibitors to patients with diabetes who have diabetic autonomic neuropathy with decreased intestinal motility, or to patients taking steroids.


Introduction
Pneumatosis cystoides intestinalis (PCI), defined as the presence of gas inside the intestinal wall, may be located in any part of the gastrointestinal tract. In PCI, gas is found in a linear or cystic form in the subserosa or submucosa [1]. The subserous cysts are most frequently found in the small bowel while the submucous localizations are predominantly seen in the colonic wall [2]. PCI is a secondary finding 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 (chronic obstructive pulmonary disease), drugs (corticosteroids, immunosuppressive therapy), and trauma (blunt abdominal trauma, endoscopy).
In most cases, PCI presents with mild gastrointestinal symptoms. Symptoms include diarrhea, mucus discharge, rectal bleeding and constipation [3]. The diagnosis is suspected by endoscopy and confirmed by computed tomography (CT) and histological examination of biopsy specimens. The endoscopic differential diagnosis of more common diseases can be difficult. In the colon, gas-filled cysts are often misdiagnosed as polyps, carcinoma, lymphoma, and colitis cystica profunda.
Patients may be treated with oxygen and/or antibiotics. Urgent surgical intervention is only required in rare cases of PCI with perforation and necrotic bowel.

Case presentation
A 65-year-old Caucasian German woman complained of left abdominal pain 3 months before referral. Under the suspected diagnosis of acute sigmoid diverticulitis, she had received nonspecific antibiotic therapy with ciprofloxacin for 5 days. Two weeks later, colonoscopy revealed numerous polypoid lesions located in the ascending colon. The histology of a biopsy specimen revealed normal colon mucosa. Nevertheless, she was referred to our hospital for hemicolectomy with the diagnosis still suspected to be polyposis coli.
Non-insulin-dependent diabetes mellitus had been diagnosed 12 years earlier. Since then, she had taken 150 mg acarbose every day. There were no other episodes of abdominal problems during that 12-year period.
Her medical history included hypertension, typhus abdominalis with ulcer 57 years previously, and hysterectomy and ovariectomy one year previously.
Physical examination showed normal blood pressure (120/60 mmHg), regular heart rate (76 beats/minute), and a body temperature of 36.6°C. Heart sounds were clear and the rhythm was regular; breath sounds were normal without rales or bronchial obstruction. The abdomen was not distended and regular peristaltic sounds were audible. Neurological examination revealed no pathological findings, in particular, no signs of diabetic polyneuropathy. Laboratory tests revealed the white blood cell count and C-reactive protein levels to be normal, but blood sugar (120 mg/dl) and HbA1c levels (6.3%) were raised.
We repeated a colonoscopy and it revealed multiple polypoid formations of varying sizes (1-3 cm) in the ascending colon, covered by normal mucosa with superficial vessels (Figure 1), and diverticulosis of the sigmoid colon. After biopsy, the cysts collapsed and disappeared. Furthermore, only colon mucosa was found in the biopsy specimens ( Figure 2). While X-ray film of the abdomen did not reveal any pathological findings, CT confirmed conspicuous gas bubbles in the ascending colon ( Figure 3).
At that time, we stopped the acarbose treatment. Whilst a diabetes diet was continued, neither further oral antidiabetics nor insulin were required to control diabetes mellitus. Additionally, the patient was treated with oxygen for 7 days (3 L/minute intranasally). After having discontinued the acarbose treatment for 3 months, the gas-filled cysts disappeared completely, as demonstrated by colonoscopy. Fortunately, the patient had not undergone hemicolectomy for initially suspected 'polyposis coli'. The patient has remained free of abdominal symptoms for a further 19 months.

Discussion
The pathogenesis of PCI is still unclear and several mechanisms have been postulated for its development. The mechanical theory proposes that gas diffuses into the intestinal wall from either the intestinal lumen or the pulmonary airway. The diffusion of intraluminal gas into the intestinal wall is due to increased intraluminal pressure and the presence of mucosal injury [4,5]. Additionally, gas might travel from ruptured alveoli through the mediastinum into the retroperitoneal space and find its way into the intestinal wall along perivascular spaces through the mesentery [6]. Alternatively, the bacterial theory suggests that gas-producing bacteria entering the intestinal wall through a mucosal lesion form intramural gas [7], thus forming cysts.
Acarbose, an a-glucosidase inhibitor, is a hypoglycemic agent that can suppress postprandial hyperglycemia by delaying absorption of carbohydrates in the small intestine through antagonistic as well as dose-dependent suppression of a-glucosidase (a-GI). Well-known side effects of a-GIs include flatulence and abdominal distension resulting from fermentation by intestinal bacteria that produce carbon dioxide, methane and hydrogen from unabsorbed carbohydrates [8].
Our patient did not experience abdominal distension or flatulence. Nevertheless, she probably had elevated intraluminal pressure because she had diverticulosis of the sigmoid colon, possibly as a result of elevated intraluminal pressure and reduced power of resistance of the intestinal wall. We found one report regarding pneumatosis cystoides coli (PCC) located in the sigmoid colon due to a solitary sigmoid diverticulum-the patient was treated with corticosteroids for periarteritis nodosa [9]. In our patient, diverticulosis of the sigmoid colon was most likely not responsible for PCC in the ascending colon.
The term pneumatosis cystoides coli (PCC) is synonymous with pneumatosis cystoides intestinalis (PCI) when the disorder is limited to the colon. In our patient, we suppose that PCC was basically caused by acarbose. Two months after her initial presentation, no regression of PCC was found via colonoscopy during continuation of her treatment with acarbose. However, after discontinuing treatment with acarbose for 3 months, the PCC located in the ascending colon had completely disappeared. It is assumed that, in our patient, PCC developed from a combination of the a-GI leading to elevated intraluminal pressure from increased gas volume due to bacterial overgrowth as well as mucosal damage in the ascending colon due to elevated intraluminal pressure. We can only speculate that the susceptibility to the formation of PCC was due to colon mucosal vulnerability after the patient's typhus abdominalis with ulcer about 57 years previously. There were no symptoms indicating previous episodes of PCI in this patient. We believe that ageing and its related changes to the colonic wall reached a critical point after 12 years, allowing gas invasion by intraluminal pressure. Patients with PCI after acarbose treatment are generally older, the youngest one being 53 years of age (Table 1).
In our patient, we used oxygen treatment applied by a mask for 8 hours daily for 1 week. We did not use any antibiotics. However, we do not definitely believe that oxygen therapy basically caused regression of PCC, because the patient was discharged home without oxygen therapy.