EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Objectives: We evaluated the incidence, clinical presentation, laboratory findings, possible causes, and outcomes associated with pneumatosis intestinalis in the setting of lung transplant.

Materials and Methods: A departmental database showed that, between October 2002 and August 2013, our institution performed 373 lung transplants. The database was queried for all patients who had pneumatosis intestinalis and/or pneumoperitoneum after lung transplant at our institution. For this retrospective chart review, information collected included demographics, baseline variables, symptoms, signs, laboratory parameters, radiologic work-up, medications, including immunosuppressants at presentation, progress and time to resolution, length of hospital stay, and postoperative complications, as well as short-term and midterm outcomes (up to 1 year).

Results: The incidence of pneumatosis intestinalis was 2.68% (10/373 patients). Mean age of patients was 52.2 years (range, 34.9-67.9 y). Mean time for devel-opment of pneumatosis intestinalis after transplant was 352.8 days (range, 5-2495 d). Abdominal symptoms and signs were present in 6 patients (60%). The large bowel was the site of involvement in all but 1 patient, with predilection for ascending colon (80%) and transverse colon (90%) in most patients. High serum lactate levels were found in 2 patients, and both underwent bowel resection, with surgical specimens revealing evidence of ischemic changes in the gut. Mean dose of prednisone was 22 mg at the time of presentation (range, 0-60 mg). Mean hospital length of stay was 16.9 days (range, 0-40 d). Short-term survival was 100%. Midterm survival at the end of 1 year was 80%. Mean time for pneumatosis to resolve was 18 days (range, 14-35 d).

Conclusions: Pneumatosis intestinalis in bilateral lung transplant recipients can be treated with a minimal amount of imaging, and most patients can be treated conservatively. Survival outcomes are quite good, with 100% during the short term and 80% at 1 year in our series.

Key words : Bowel resection, Pneumoperitoneum, Retrospective study, Survival

Pneumatosis intestinalis (PI) is a bowel pathology defined as accumulation of gas in the submucosa or subserosa of the large or small intestine (Figure 1). It is seen in association with a number of diseases, but its pathogenesis is still obscure. It has shown an association with solid-organ and bone marrow transplant,^1-5 immunosuppressive drugs, especially steroids,^1,6,7 ischemia, intestinal obstruction, and inflammatory bowel disease. Few cases have been reported after lung transplant.^5,8Several concomitant infections have been proposed as causative agents for the development of PI, including cytomegalovirus (CMV), Clostridium difficile, human immuno-deficiency virus, and rotavirus infection.^1,2,4,7,9

Three mechanisms have been suggested to account for intramural gas: mechanical, bacterial, and pulmonary. According to the mechanical theory, a mucosal injury caused by increased intraluminal pressure causes the dissection of gas into the bowel wall (eg, endoscopic procedure, intestinal obstruction). The bacterial theory proposes that gas formed by gas-fermenting organisms infiltrates the bowel wall through mucosal injury. The pulmonary theory states that air possibly comes from the thoracic cavity through the retroperitoneum.¹⁰ The extent of PI does not necessarily correlate with the severity of the symptoms or underlying disease.¹¹ Proven PI in conjunction with a serum lactic acid level exceeding 2 mmol/L entails a high rate of fatal outcomes, of over 80%.¹² Thus, transplant physicians should be well-versed in differentiating the fatal form of PI from a benign one.

Materials and Methods

A departmental database was queried for PI and/or pneumoperitoneum after lung transplant. Between October 2002 and August 2013, we found that 373 transplants were performed. A retrospective chart review was conducted for all patients who had PI and pneumoperitoneum after lung transplant at our institution, with 10 patients identified. Systematic chart review, clinic notes, and electronic records were used to collect demographics and baseline variables (Table 1), including abdominal symptoms and signs, radiologic work-up results, concurrent infections, immunosuppressive regimens, prednisone dose at time of diagnosis, laboratory values, length of hospital stay, postoperative complications, progress, and time to resolution, as well as short-term and midterm outcomes up to 1 year. A radiologist with expertise in abdominal computed tomography (CT) scans reviewed the plain radiographic images and CT scans of all patients to confirm PI and or pneumoperitoneum.

Mean age of patients was 52.2 years (range, 34.9-67.9 y); all but 1 patient were white. Mean body mass index was 24.4 kg/m² (range, 18.2-34.8 kg/m²). Mean time for development of PI after transplant was 352.8 days (range, 5-2495 d). Mean time for PI to resolve was 18 days (range, 14-35 d). Mean dose of prednisone was 22 mg at the time of presentation (range, 0-60 mg) (Table 2).

All but 1 patient had involvement of transverse colon (90%) (Figure 2), and 8 patients (80%) had involvement of ascending colon. Descending colon, sigmoid, and rectum were not involved in any of the patients. One patient had isolated involvement of terminal ileum. Half of the patients (50%) had PI associated with air under diaphragm (pneumoperitoneum). Portal venous gas was present in only 1 patient. We observed cystic disease of the liver, kidney, and pancreas in association with 40% of the cases. Only 1 patient had CMV infection at the time of PI. Rectal swab for vancomycin-resistant enterococci (VRE) was positive in 4 patients (40%).

Abdominal symptoms were present in 6 patients (60%), whereas 4 patients were completely asymptomatic. Those with symptoms presented with diarrhea (3 patients), abdominal pain (2 patients), dehydration (2 patients), and vomiting (1 patient). Minimal abdominal signs were seen (rigidity in only 1 patient). Nine patients (90%) did not have any abdominal signs whatsoever. High serum lactate level was found in 2 patients; both of these patients required bowel resection, with surgical pathology confirming evidence of ischemic changes in the gut. White blood cell counts were mostly within normal ranges in almost all patients except 1, where abnormal levels were associated with high serum lactate concentration. Mean white cell count was 8.5 × 109/L (range, 0.4-16 × 109/L). None of the patients had documented rejection at the time of presentation of PI or pneumoperitoneum.

Mean hospital length of stay was 16.9 days (range, 0-40 d). Short-term survival was 100%. Survival at the end of 1 year was 80%. One death was the result of VRE septicemia, and another death resulted from infected pancreatic pseudocyst leading to septicemia.

Multiple factors predispose transplant patients to the development of PI, with steroid therapy, infectious colitis, graft-versus-host disease and septic shock1 being important factors. Although there have been rare case reports of PI developing in patients after thoracic organ transplant,^2,5 there is only one series published, which consisted of 7 patients presenting with pneumoperitoneum or PI after lung transplant.¹³ The clinical presentation, radiographic appearance, and outcomes of previous case reports were quite similar to those for our 10 patients with PI. To our knowledge, this is the largest case series (consisting of 10 patients) of patients presenting with PI after lung transplant. The incidence of PI in our lung transplant recipients was 2.68%. Previous published case series have reported 2% incidence of PI and/or pneumoperitoneum.

In our patients, diarrhea was the most common symptom, whereas abdominal signs seem to be rare in PI after lung transplant. In a previous study, symptoms attributed to PI (in decreasing order of frequency) were reported as diarrhea, bloody stool, abdominal pain, abdominal distention, constipation, weight loss, and tenesmus.¹⁴

Sepsis, hypotension, metabolic acidosis, thickened bowel wall, and a serum lactate concentration of ≥ 2.0 mmol/L have been shown to be associated with fatal outcomes in over 80% of patients.¹² This was substantiated in our study, where high serum lactate was found in 2 patients, with both undergoing bowel resection and with surgical specimen revealing evidence of ischemic changes in the gut.

Pneumatosis intestinalis can be associated with both benign and life-threatening conditions.15 The benign causes of PI usually result in only mild or even no abdominal symptoms. Patients usually display cardiopulmonary stability, with lactate levels within the normal limits and do not present with septic signs or respiratory acidosis.12 In these patients, there are often no CT abnormalities other than diagnosis of PI. In contrast, patients with life-threatening causes of pneumatosis intestinalis usually display peritoneal symptoms.

A retrospective analysis of a 10-year period at a single institution found that age ≥ 60 years, the presence of emesis, and a white blood cell count of > 12 000/mm³ were independently associated with surgical treatment (multivariate analysis) and that preexisting sepsis correlated with a higher risk of death.¹² In our series, 1 patient showed abnormal white blood cell count, where it was associated with high serum lactate concentration (10.5 mmol/L). This patient underwent bowel resection for ischemic changes in the gut. Therefore, a white blood cell count of 12 × 10⁹/L seems to be associated with a more life-threatening presentation and could serve as an indicator of surgical management along with high serum lactate. Mean white cell count in our patients was 8.5 × 109/L (range, 0.4-16 × 10⁹/L). None of the patients had documented rejection at the time.

Portal venous gas was present in 1 patient. Interestingly, this patient also had high serum lactate level (10.5 mmol/L) along with significant abdominal distention and underwent right hemicolectomy with end ileostomy. The surgical specimen revealed evidence of ischemic changes in the gut. Thus, presence of portal venous gas seems to be suggestive of bowel ischemia This is in agreement with suggestions made by some authors that the combination of PI and portal venous gas is associated with bowel ischemia in about 70% of cases.^16-18 It is necessary to identify factors responsible for poor prognosis and high mortality and to make a decision regarding surgical intervention, if needed, early to avoid life-threatening complications resulting from bowel perforation and resulting peritonitis. It is imperative that imaging findings of PI are carefully correlated with clinical history, abdominal signs, and blood work indicating any metabolic derangement to differentiate patients who need surgical intervention from those who can be treated conservatively. Of note, we observed cystic disease of the liver, kidney, and pancreas in association with 40% of the cases. This observation needs further exploration regarding whether there is an association between PI and cystic disease involving abdominal organs.

According to the results shown in our patients, most cases of PI after lung transplant have a predilection for the large bowel, mostly with involve-ment of the ascending and transverse colon. This agrees with an earlier case series in which all cases were noted to involve the large bowel.13 We found that 50% of our patients had PI associated with air under the diaphragm (pneumoperitoneum). Presence of pneumoperitoneum with PI was not associated with poor outcomes (1 death was due to late lung infection), and only 1 patient required surgical exploration with documented bowel ischemia.

Some investigators have suggested the role of steroids in the development of pneumoperitoneum.^1,7,9 It has been proposed that steroids cause atrophy of the lymphoid tissue (Peyer’s patches) in the gastro-intestinal tract, resulting in mucosal degradation and thus allowing dissection of intraluminal air into the intestinal wall. In our case series, although 9 of the 10 patients were on prednisone, we did not find higher doses given at the time of presentation of PI. Mean dose of prednisone was 22 mg at the time of presentation (range, 0-60 mg).

Development of PI in the setting of transplant has also been linked to rotavirus1,6 and Clostridium difficile.^1,5 In 1 case report, Clostridium difficile colitis was thought to be the cause of the PI after lung transplant.⁵ Inflammation associated with these infections may cause a break in the integrity of the intestinal mucosa, thus allowing intraluminal air to dissect into the bowel wall. Active CMV infection has been widely described in association with the development of PI in the setting of transplant.^2,19 We did not encounter this association in our patients. Only 1 patient had CMV infection at the time of presentation, and none of our patients had associated Clostridium difficile infection. On the other hand, rectal swab for VRE was positive in 4 patients (40%).

The mean time for development of PI after transplant was 352.8 days (range, 5-2495 d). Most of our patients (80%) presented with PI after primary hospital discharge. Only 2 patients presented with PI during primary admission for lung transplant.

Mean hospital length of stay was 16.9 days (range, 0-40 d). In addition, 2 of our patients who presented to an outpatient department for routine evaluation were found to have PI after routine chest radiography. Because thorough examination showed no symptoms, signs, or abnormal laboratory results, the patients were treated as outpatients. Both patients had radiologic resolution of PI within 2 weeks without any complication.

Mean time for PI to resolve (radiologic resolution) was 18 days. Short-term survival was 100%, and survival at the end of 1 year was 80%. No deaths were attributable to the development of PI or pneumoperitoneum. One death was the result of VRE septicemia on a separate hospital admission, and another death resulted from infected pancreatic pseudocyst leading to septicemia. It appears that colonic PI in bilateral lung transplant recipients can be treated with a minimal amount of imaging, and most cases can be treated conservatively. The radiologist should alert the clinician to perform the appropriate clinical and laboratory work-up to rule out a less common case scenario, such as PI associated with bowel ischemia. A careful correlation of imaging findings of PI with clinical history, physical examination, and laboratory work indicating that any metabolic derangement can be used to differentiate patients who need surgical intervention from those who can be treated conservatively.

Pneumatosis intestinalis in bilateral lung transplant recipients can be treated with a minimal amount of imaging, and most cases can be treated conservatively. It seems to present with few symptoms (60% of patients in this series) and minimal abdominal signs (only 1 patient in our series). The benign form generally resolves over 2 to 5 weeks (mean of 18 days in this study). This study confirms that most cases of PI after lung transplant have a predilection for the large bowel, with most cases involving the ascending and transverse colon. Although 9 of the 10 patients were on prednisone, most our patients were not on high doses of steroids (mean dose of 22 mg) at the time of presentation of PI. None of our patients had associated Clostridium difficile infection, and only 1 patient had CMV infection at the time of PI. On the other hand, rectal swab for VRE was positive in 4 patients (40%). Presence of abdominal symptoms and signs along with high white blood cell count, high serum lactate, and portal venous gas in the CT scan seems to be associated with a more life-threatening presentation and could serve as an indicator of need for surgical management. Survival outcomes are quite good, with 100% short-term survival and 80% 1-year survival in our patients.

Volume: 16 Issue: 1 February 2018

FULL TEXT