Nonsurgical pneumoperitoneum detected using computed tomography: A retrospective study

Background Although several cases with pneumoperitoneum that does not require surgical intervention (nonsurgical pneumoperitoneum) have been reported, the characteristics of such cases remain unclear. The accurate diagnosis of nonsurgical pneumoperitoneum could minimize unnecessary surgery. The aim of this study was to clarify the clinical and radiological characteristics of cases with nonsurgical pneumoperitoneum detected using computed tomography. Methods This retrospective study was conducted at a single center. A total of 18513 abdominal computed tomography (CT) scans obtained between January 2010 and February 2017 were examined for pneumoperitoneum. Medical records of cases testing positive for extraluminal free air were analyzed. Results Extraluminal free air was detected in 254 examinations of 182 cases. Out of 88 examinations of 86 cases excluding iatrogenic air, colorectal perforation was the most common cause of extraluminal free air. Nonsurgical pneumoperitoneum was recognized in 25 examinations of 23 cases, and was the second most frequent. The frequency of nonsurgical pneumoperitoneum was 0.14% in all abdominal CT examinations. Most nonsurgical pneumoperitoneum cases did not exhibit severe general conditions, peritoneal signs, or leukocytosis. CT ndings of bowel wall discontinuity, segmental bowel-wall thickening, perivisceral fat stranding, and abscess were not observed. Fluid collection was present in 8 of 23 cases, and the estimated volume of uid collection was small. Pneumatosis intestinalis was simultaneously observed in 20 of 23 cases. No signicant differences in the maximum diameter of intraperitoneal free air were observed between the grades of pneumatosis intestinalis (p=0.999). Follow-up CT examination, which was performed within 7 days after the detection of nonsurgical pneumoperitoneum, showed that the pneumatosis intestinalis and/or extraluminal free air often disappeared in a short time. Conclusions Nonsurgical pneumoperitoneum was common. The cause of extraluminal free air was pneumatosis intestinalis in most cases. Well-maintained


Abstract
Background Although several cases with pneumoperitoneum that does not require surgical intervention (nonsurgical pneumoperitoneum) have been reported, the characteristics of such cases remain unclear. The accurate diagnosis of nonsurgical pneumoperitoneum could minimize unnecessary surgery. The aim of this study was to clarify the clinical and radiological characteristics of cases with nonsurgical pneumoperitoneum detected using computed tomography. Methods This retrospective study was conducted at a single center. A total of 18513 abdominal computed tomography (CT) scans obtained between January 2010 and February 2017 were examined for pneumoperitoneum. Medical records of cases testing positive for extraluminal free air were analyzed. Results Extraluminal free air was detected in 254 examinations of 182 cases. Out of 88 examinations of 86 cases excluding iatrogenic air, colorectal perforation was the most common cause of extraluminal free air. Nonsurgical pneumoperitoneum was recognized in 25 examinations of 23 cases, and was the second most frequent. The frequency of nonsurgical pneumoperitoneum was 0.14% in all abdominal CT examinations. Most nonsurgical pneumoperitoneum cases did not exhibit severe general conditions, peritoneal signs, or leukocytosis. CT ndings of bowel wall discontinuity, segmental bowel-wall thickening, perivisceral fat stranding, and abscess were not observed. Fluid collection was present in 8 of 23 cases, and the estimated volume of uid collection was small. Pneumatosis intestinalis was simultaneously observed in 20 of 23 cases. No signi cant differences in the maximum diameter of intraperitoneal free air were observed between the grades of pneumatosis intestinalis (p=0.999). Follow-up CT examination, which was performed within 7 days after the detection of nonsurgical pneumoperitoneum, showed that the pneumatosis intestinalis and/or extraluminal free air often disappeared in a short time. Conclusions Nonsurgical pneumoperitoneum was common. The cause of extraluminal free air was pneumatosis intestinalis in most cases. Well-maintained general and local conditions and normal laboratory data were the clinical characteristics. The absence of CT ndings indicative of peritonitis, little uid collection, if any, and the presence of pneumatosis intestinalis were the radiological characteristics.

Background
Pneumoperitoneum is considered to be one of the most crucial signs of severe intraabdominal diseases. doctors in the out-patient, in-patient, and emergency departments. All CT examinations were performed using a 32-slice multi-detector CT scanner (Aquilion 32, Toshiba, Japan), and a collimated slice thickness of 1 mm was used. All ndings of the CT scan images were reported in our database.

Clinical Analysis
The database was searched using "free air" as the keyword to nd cases with extraluminal free air.
Extraluminal free air consists of intraperitoneal, intramesenteric and retroperitoneal free air. In this study, intramesenteric or retroperitoneal free air was de ned as the air not continuous to the air in the bowel wall. Extraluminal free air was detected in 254 examinations of 182 cases. Medical records of the 182 cases were reviewed to con rm the cause of the free air.
A case with extraluminal free air that did not retrospectively require surgical intervention or intensive therapy was de ned as a nonsurgical pneumoperitoneum case. Nonsurgical pneumoperitoneum case was a case with pneumoperitoneum from nonsurgical cause. Nonsurgical pneumoperitoneum cases included a case receiving unnecessary surgery, but did not include a case receiving intensive therapy instead of surgery. For example, a case with perforated duodenal ulcer treated by conservative therapy without surgery was not included in nonsurgical pneumoperitoneum cases. Medical doctors who participated in this study analyzed the medical records of nonsurgical pneumoperitoneum cases in detail.

Imaging Analysis
One radiologist with 32 years of experience and one gastrointestinal surgeon with 38 years of experience reevaluated the CT examinations of nonsurgical pneumoperitoneum cases. CT ndings evaluated were bowel wall discontinuity, segmental bowel-wall thickening, perivisceral fat stranding, intraabdominal abscess, intraperitoneal uid collection, extraluminal free air, and pneumatosis intestinalis.
In each case, the maximum diameter of the largest pocket of intraperitoneal free air located either under the abdominal wall or in the perihepatic space was measured to estimate the amount of extraluminal free air. We did not measure intramesenteric or retroperitoneal free air (Fig.1), because the maximum diameter of intramesenteric or retroperitoneal air could not be correctly measured. The anteroposterior, lateral, and vertical diameters of the uid collection were measured, and the estimated volume was determined using a formula for calculating ellipsoid volume.
Pneumatosis intestinalis was de ned as the presence of air in the bowel wall identi ed using lung window settings [15]. The grade of pneumatosis intestinalis was classi ed by our criteria as follows. Bubbly, linear, or circular air ( Fig. 2 and 3) in or along less than 10 cm of the bowel wall was de ned as mild. Air in or along more than or equal to 10 cm of the bowel wall or air not only in the bowel wall but also in the mesentery (Fig.4) Table 1.

2) Causes of extraluminal free air
The causes of extraluminal free air in the 86 cases excluding iatrogenic air were studied, and are summarized in Table 2. Colorectal perforation was the most common cause of extraluminal free air. The frequency of nonsurgical pneumoperitoneum cases was 26.7% constituting the second most common nding. Perforated duodenal ulcer was the third most common cause. One case with pneumoretroperitoneum caused by pneumothorax was not included in nonsurgical pneumoperitoneum, because the case was required intensive therapy for pneumothorax.

3) Clinical features of nonsurgical pneumoperitoneum cases
Page 5/15 The clinical features of nonsurgical pneumoperitoneum cases are presented in Table 3. The ages of nonsurgical pneumoperitoneum cases were high. The chief complaints of the cases varied. All cases showed good general condition, and many cases did not show peritoneal signs or leukocytosis. One case underwent emergent laparotomy because of slight peritoneal signs and severe leukocytosis. This case showed no pathological intraabdominal ndings. Nine cases received antibiotic therapy. The antibiotics were administered for a short period of time, or used for other diseases.
Twenty-two of the 23 cases had comorbidities, such as cardiovascular, digestive, neurological, and endocrine disorders. One case had bronchial asthma, three cases received steroids, one case received αglucosidase, and one case received chemotherapy for cancer. No speci c comorbidities were shared among the nonsurgical pneumoperitoneum cases.

4) CT ndings of nonsurgical pneumoperitoneum cases
The CT ndings of 23 nonsurgical pneumoperitoneum cases are summarized in Table 4. Bowel wall discontinuity, segmental bowel-wall thickening, perivisceral fat stranding and abscess were not recognized. Fluid collection was not observed in 15 of 23 cases, and the estimated total volume of uid collection was small in other cases. Pneumatosis intestinalis was simultaneously detected in 20 cases, and the average age of the 20 cases was 79.8 ±9.8 (mean±SD) years. Intraperitoneal free air was detected in 21 cases, and the maximum diameter of intraperitoneal free air varied.
The maximum diameter of intraperitoneal free air was compared between mild and severe grades of pneumatosis intestinalis, as shown in Figure 5. No signi cant differences in the maximum diameter of intraperitoneal free air were observed between the grades (p=0.999).

5) Follow-up of CT ndings in nonsurgical pneumoperitoneum cases
The follow-up CT examination, which was performed to observe the clinical course within 7 days after the detection of nonsurgical pneumoperitoneum, was performed in eight cases. The CT ndings when nonsurgical pneumoperitoneum was detected and the follow up CT ndings in the 8 cases were compared, and presented in Table 5. When the follow-up CT was performed, pneumatosis intestinalis alone disappeared in 2cases, extraluminal free air alone disappeared in one case, and both pneumatosis intestinalis and extraluminal free air disappeared in three cases.

Discussion
In this study, we demonstrated that the frequency of nonsurgical pneumoperitoneum was 0.14% among all abdominal CT examinations. When the causes of extraluminal free air in the cases excluding those with iatrogenic air were studied, nonsurgical pneumoperitoneum was observed in 26.7% of the cases.
This frequency was the second highest and next to the frequency of colorectal perforation. These results indicate that nonsurgical pneumoperitoneum cases are common. Mularski et al [16] reported in their review article that pneumoperitoneum was caused by visceral perforation in 85% to 95% of all occurrences and, in 5% to 15% of cases, pneumoperitoneum resulted from another source that did not require emergency surgery. Their 2000 study was based on many reports using plain radiography rather than CT examination; since 2000, CT has improved and been increasingly adopted by many medical institutions worldwide. The present study is the rst to clarify the frequency of nonsurgical pneumoperitoneum cases detected using CT.
The main causes of nonsurgical pneumoperitoneum were reported as being intrathoracic, gynecologic, abdominal, and idiopathic [9,16]. Pneumothorax, pneumomediastinum, cardiopulmonary resuscitation, positive pressure ventilation, etc. have been reported as intrathoracic causes. Vaginal douching, postpartum exercises, oral-genital insu ation, and coitus have been reported as gynecologic causes. Abdominal causes were classi ed as iatrogenesis and pneumatosis intestinalis. The common iatrogenic causes of pneumoperitoneum were following abdominal surgery and following gastrointestinal endoscopy [9,16]. In the present study, the cases with iatrogenic pneumoperitoneum were not analyzed as nonsurgical pneumoperitoneum cases, because iatrogenic pneumoperitoneum could be easily diagnosed as iatrogenic with an accurate medical history.
Benign pneumatosis intestinalis can result in pneumoperitoneum caused by rupture of subserosal cysts [17]. Jamart [18] reported that pneumoperitoneum was present in 9%, and Saito et al [19] reported that intraperitoneal free air was detected in 26.4% of cases with pneumatosis intestinalis. In the present study, simultaneous pneumatosis intestinalis was detected in 20 (87.0%) of the nonsurgical pneumoperitoneum cases. Pneumatosis intestinalis was considered to be a cause of extraluminal free air in the 20 cases, and the remaining 3 cases without pneumatosis intestinalis were classi ed as idiopathic. Our follow-up CT ndings showed that the pneumatosis intestinalis and/or extraluminal free air often disappeared in a short time. These ndings indicated that extraluminal free air alone could be detected in nonsurgical pneumoperitoneum cases caused by pneumatosis intestinalis when CT examination was delayed. These results suggest that pneumatosis intestinalis may be one of the causes of extraluminal free air in idiopathic nonsurgical pneumoperitoneum cases.
The average age of patients with both nonsurgical pneumoperitoneum and pneumatosis intestinalis was 79.8±9.8 (mean±SD) years. The mean ages of the patients with pneumatosis intestinalis reported by Wu et al [20], DuBose et al [21], and Saito et al [19] were 45.3, 53.3, and 64.7 years, respectively. The ages of our cases were particularly high compared with those of the patients enrolled in the aforementioned studies. Our study may indicate that pneumatosis intestinalis readily leads to ruptures in older patients, resulting in nonsurgical pneumoperitoneum. Mularski et al [8] reported that nonsurgical causes of pneumoperitoneum should be considered when abdominal pain and distension are minimal and peritoneal signs, fever, and leukocytosis are absent. Tang et al [14] reported that nonsurgical pneumoperitoneum cases often present with minimal pain and typically have no fever, leukocytosis, peritoneal signs, or metabolic acidosis. The results for many of the cases in the present study resembled those of the aforementioned studies. Based on these reports, well-maintained general and local conditions and normal laboratory data were the clinical characteristics of nonsurgical pneumoperitoneum cases.
Radiological ndings of nonsurgical pneumoperitoneum cases were described by Williams et al [9].
However, the CT ndings of nonsurgical pneumoperitoneum cases have not been studied. From this analysis of CT ndings, bowel wall discontinuity, segmental bowel-wall thickening, perivisceral fat stranding, and abscess were not observed in nonsurgical pneumoperitoneum cases. These CT ndings were indicative of acute bowel disease and gastrointestinal perforation [22,23]. Fluid collection was observed in a few cases, and the estimated volume of uid collection was small. Pneumatosis intestinalis was simultaneously observed in most nonsurgical pneumoperitoneum cases. From these results, we demonstrated that the absence of CT ndings indicative of peritonitis, little uid collection, if any, and the simultaneous presence of pneumatosis intestinalis were the radiological characteristics of nonsurgical pneumoperitoneum cases. It was noteworthy that there were no differences in the maximum diameter of intraperitoneal free air between the grades of pneumatosis intestinalis. Therefore, it is crucial to consider nonsurgical pneumoperitoneum when pneumoperitoneum cases exhibit simultaneous pneumatosis intestinalis, even if the extraluminal free air is massive and the pneumatosis intestinalis is minimal.
A 1992 study by Hoover et al [24] indicated that meticulous study of medical history, complete physical examinations, laboratory data with radiographic studies, and frequent reevaluation of these ndings are necessary for the accurate diagnosis of nonsurgical pneumoperitoneum. Although their recommendations remain relevant, we should elaborate on the ndings characteristic to nonsurgical pneumoperitoneum to make an accurate diagnosis and minimize unnecessary surgery. In this study, we demonstrated some CT ndings often observed in nonsurgical pneumoperitoneum cases. CT images should be carefully evaluated using lung window settings, if some clinical ndings to suspect nonsurgical pneumoperitoneum are present.

Conclusion
Nonsurgical pneumoperitoneum detected using computed tomography were common. The most common cause of free air was pneumatosis intestinalis, which may lead to ruptures more readily in elderly patients. Well-maintained general and local conditions and normal laboratory data were the clinical characteristics. The absence of CT ndings indicative of peritonitis, little uid collection, if any, and the presence of pneumatosis intestinalis were the radiological characteristics. The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.