Indian J Gastroenterol DOI 10.1007/s12664-014-0449-z
ORIGINAL ARTICLE
Gastrointestinal intramural hematoma—Analysis of clinical and radiological features for early differentiation from mesenteric ischemia R. Subhash & G. Unnikrishnan & Dinesh Balakrishnan & O. V. Sudheer & Puneet Dhar & S. Sudhindran
Received: 8 November 2013 / Accepted: 13 February 2014 # Indian Society of Gastroenterology 2014
Abstract Introduction Long-term anticoagulation is associated with hemorrhage at various sites. Gastrointestinal intramural bleeds and hematomas (IMH) often mimic mesenteric ischemia (MI) due to similar clinical settings and imaging features, making early differentiation difficult. Aim To compare the demography, clinical features and imaging characteristics of patients presenting with IMH with those of MI, so as to help in evolving clinical and imaging guidelines to differentiate both early in the course of the disease. Methods All radiologically (contrast-enhanced computed tomogram [CT]) diagnosed cases of gastrointestinal IMH from the hospital database during the period between 2006 and 2012 were retrospectively analyzed. This data was compared with the clinical and imaging features of a group of surgically confirmed MI during the same period. Patients not on anticoagulation therapy at the time of presentation and those with incomplete clinical or radiological data were excluded from the study. Results There were 16 patients in IMH group and 54 patients in MI group. Clinical features like overt rectal bleeding or melena, and prolonged prothrombin time-international normalized ratio (PT-INR) more than three, and CT features like proximal location in the bowel, increased bowel wall thickness, hyperdensity on plain scan (>40 Hounsfield units (HU)), and short segment bowel involvement were significantly associated with IMH. Visualization of embolus and absent
R. Subhash and G Unnikrishnan contributed equally to this work. R. Subhash (*) : G. Unnikrishnan : D. Balakrishnan : O. V. Sudheer : P. Dhar : S. Sudhindran Department of Gastrointestinal Surgery and Liver Transplantation, Amrita Institute of Medical Sciences and Research Centre, Aims Ponekkara PO, Kochi 682 041, India e-mail: [email protected]
mesenteric vasculature to a segment of intestine in CT was significantly associated with MI. Conclusion Attention to clinical features and early CT scan can aid in early differentiation of IMH from MI, facilitating appropriate intervention early in the course of disease. Keywords Anticoagulation . Gastrointestinal tract . Hemorrhage
Introduction Long-term oral anticoagulation is recommended for patients with mechanical heart valves, atrial fibrillation, cardiomyopathy, and similar cardiac conditions to prevent thromboembolic complications. When anticoagulation is inadequate in these subjects, arterial embolization can occur to various sites in the body including mesenteric arteries, resulting in mesenteric ischemia (MI). Conversely, when anticoagulation is excessive, it may lead to hemorrhagic complications [1]. Anticoagulation-related gastrointestinal (GI) bleeds may be luminal or intramural [2]. Identification of luminal GI bleeding is clinically evident and straightforward. However, the detection of intramural hematoma (IMH) or bleed is difficult and can often be confused with MI secondary to embolization [3]. Usual clinical presentation for both the conditions is typically a patient on anticoagulant therapy presenting with acute severe abdominal pain, hypotension, and tachycardia. Whilst mesenteric embolization may necessitate urgent radiologic or surgical intervention to remove the embolus from the ischemic site to restore blood flow, intramural bleeding most often merely requires correction of the abnormal anticoagulation by medical measures. Differentiating between
Indian J Gastroenterol
MI and IMH early at the time of presentation is therefore crucial to the management strategy in such patients. The aim of this study was to compare the demography, clinical features, and imaging characteristics of patients presenting with IMH with those of MI, among patients on anticoagulation therapy.
Methods We performed a retrospective analysis of all radiologically (contrast-enhanced computed tomogram [CT]) diagnosed cases of gastrointestinal IMH from the hospital database of our institution during the period between 2006 and 2012 (n=19). This data was compared with the clinical and imaging features of a group of surgically confirmed MI during the same period (n= 59). Patients not on anticoagulation at the time of presentation and those with incomplete clinical or radiological data were excluded from the study (n=8, three from IMH group and five from MI group). All patients had undergone CT abdomen with plain and contrast phases (multidetector spiral CT scanner, Siemens Somatom Sensation 64, Siemens AG, Munich, Germany). All the imaging features were reevaluated by two senior abdominal radiologists with blinded diagnosis (images archived and reviewed on a workstation named Amrita medvision version 4.2.2686, Amrita Technologies, India). The location (proximal bowel [duodenum and jejunum] vs. distal bowel [ileum and colon]), length of involvement (short vs. long segment, arbitrarily defined by a cutoff value of 15 cm) [4], extent of mural thickening, fat stranding, dilatation of the bowel, presence of hemoperitoneum or ascites, pattern of attenuation, and other obvious findings if present were studied in both the groups. Early clinical symptoms and international normalized ratio (INR) for prothrombin time were correlated with these radiological findings and were compared between the two groups. Clinical signs were indistinguishable early in the course of both the diseases. Also, features of systemic toxicity and abdominal signs were very evident with disease progression late in the course of MI. So, the clinical signs were not considered in our study, as our aim was to differentiate both early in the course of disease. Standard statistical methods with chi-square test (for categorical variables) and nonparametric Mann-Whitney U test (for continuous variables) were used to compare each variable. All analyses were done using SPSS.v.18.
Results Initially, we identified 19 patients with IMH and 59 patients with MI. Of these, three patients with IMH and five patients with MI were excluded from the study, leaving 16 patients in IMH group (group I) and 54 patients in the MI group (group II), based on set
inclusion and exclusion criteria. Age and sex distribution did not differ significantly between the two groups (Table 1). Symptoms like abdominal pain, vomiting, and distension were present in both the groups, and the difference was statistically not significant. However, melena or bleeding per rectum as a symptom was more common in IMH group, and the difference was statistically significant (p-value-0.033) when comparing with MI. The prothrombin time-international normalized ratio (PTINR) was prolonged and more than three in all but one patient (93.7 %) in IMH group, while it was less than three in 55.5 % patients with MI. The p-value was significant (p=0.001) when this cutoff value of three for PT-INR was considered. On evaluation of CT findings (Table 2), IMH was mainly identified in proximal bowel segments in comparison to MI. The site distribution of IMH was esophagus 6.2 %, duodenum 6.2 %, jejunum 62.5 %, ileum 25 % and that of MI was jejunum 13 %, ileum 33 %, jejunum and ileum 26 %, colon 17 %, ileum and colon 11 %. On comparing the two groups of patients (IMH and MI) based on the site of involvement (considering duodenum and jejunum as proximal and ileum and colon as distal), the difference was statistically significant (p-value 0.001). The length of the involved bowel segment was arbitrarily defined as short segment and long segment involvement based on a cutoff value of 15 cm [4]. Of patients with IMH, 87.5 % fell under the category of short segment, whereas the majority of MI (74 %) had long segment bowel involvement. The difference was statistically significant with a p-value of 0.006. The median thickness of the involved bowel was 17.6 mm in IMH group and 8.2 mm in MI group, showing the bowel wall thickening to be significantly associated with IMH (p-value 0.001). The attenuation value of the mural thickening calculated in plain CT had a median value of 65.2 Hounsfield units (HU) for IMH and 34.7 for MI showing a statistically significant difference (p-value 0.001). All the patients with IMH were showing an attenuation value above 40 HU.
Table 1 Comparison of demography and clinical findings
Sex, male/female Age, years Pain abdomen Vomiting Rectal bleeding/melena PT-INR >3
IMH (n=16)
MI (n=54)
p-value
12/4 54 (51–60)# 16/16 (100) 14/16 (87.5) 6/16 (37.5) 15/16 (93.75)
36/18 49 (45–61)# 46/54 (85.18) 34/54 (62.96) 4/54 (7.4) 24/54 (44.44)
0.508 0.355 0.247 0.189 0.033* 0.001**
All values in parentheses, unless marked otherwise, represent percentages IMH intramural bleeds and hematomas, MI mesenteric ischemia, PT-INR prothrombin time-international normalized ratio # Expressed value—median and interquartile range; *significant at 5 % level (p
ORIGINAL ARTICLE
Gastrointestinal intramural hematoma—Analysis of clinical and radiological features for early differentiation from mesenteric ischemia R. Subhash & G. Unnikrishnan & Dinesh Balakrishnan & O. V. Sudheer & Puneet Dhar & S. Sudhindran
Received: 8 November 2013 / Accepted: 13 February 2014 # Indian Society of Gastroenterology 2014
Abstract Introduction Long-term anticoagulation is associated with hemorrhage at various sites. Gastrointestinal intramural bleeds and hematomas (IMH) often mimic mesenteric ischemia (MI) due to similar clinical settings and imaging features, making early differentiation difficult. Aim To compare the demography, clinical features and imaging characteristics of patients presenting with IMH with those of MI, so as to help in evolving clinical and imaging guidelines to differentiate both early in the course of the disease. Methods All radiologically (contrast-enhanced computed tomogram [CT]) diagnosed cases of gastrointestinal IMH from the hospital database during the period between 2006 and 2012 were retrospectively analyzed. This data was compared with the clinical and imaging features of a group of surgically confirmed MI during the same period. Patients not on anticoagulation therapy at the time of presentation and those with incomplete clinical or radiological data were excluded from the study. Results There were 16 patients in IMH group and 54 patients in MI group. Clinical features like overt rectal bleeding or melena, and prolonged prothrombin time-international normalized ratio (PT-INR) more than three, and CT features like proximal location in the bowel, increased bowel wall thickness, hyperdensity on plain scan (>40 Hounsfield units (HU)), and short segment bowel involvement were significantly associated with IMH. Visualization of embolus and absent
R. Subhash and G Unnikrishnan contributed equally to this work. R. Subhash (*) : G. Unnikrishnan : D. Balakrishnan : O. V. Sudheer : P. Dhar : S. Sudhindran Department of Gastrointestinal Surgery and Liver Transplantation, Amrita Institute of Medical Sciences and Research Centre, Aims Ponekkara PO, Kochi 682 041, India e-mail: [email protected]
mesenteric vasculature to a segment of intestine in CT was significantly associated with MI. Conclusion Attention to clinical features and early CT scan can aid in early differentiation of IMH from MI, facilitating appropriate intervention early in the course of disease. Keywords Anticoagulation . Gastrointestinal tract . Hemorrhage
Introduction Long-term oral anticoagulation is recommended for patients with mechanical heart valves, atrial fibrillation, cardiomyopathy, and similar cardiac conditions to prevent thromboembolic complications. When anticoagulation is inadequate in these subjects, arterial embolization can occur to various sites in the body including mesenteric arteries, resulting in mesenteric ischemia (MI). Conversely, when anticoagulation is excessive, it may lead to hemorrhagic complications [1]. Anticoagulation-related gastrointestinal (GI) bleeds may be luminal or intramural [2]. Identification of luminal GI bleeding is clinically evident and straightforward. However, the detection of intramural hematoma (IMH) or bleed is difficult and can often be confused with MI secondary to embolization [3]. Usual clinical presentation for both the conditions is typically a patient on anticoagulant therapy presenting with acute severe abdominal pain, hypotension, and tachycardia. Whilst mesenteric embolization may necessitate urgent radiologic or surgical intervention to remove the embolus from the ischemic site to restore blood flow, intramural bleeding most often merely requires correction of the abnormal anticoagulation by medical measures. Differentiating between
Indian J Gastroenterol
MI and IMH early at the time of presentation is therefore crucial to the management strategy in such patients. The aim of this study was to compare the demography, clinical features, and imaging characteristics of patients presenting with IMH with those of MI, among patients on anticoagulation therapy.
Methods We performed a retrospective analysis of all radiologically (contrast-enhanced computed tomogram [CT]) diagnosed cases of gastrointestinal IMH from the hospital database of our institution during the period between 2006 and 2012 (n=19). This data was compared with the clinical and imaging features of a group of surgically confirmed MI during the same period (n= 59). Patients not on anticoagulation at the time of presentation and those with incomplete clinical or radiological data were excluded from the study (n=8, three from IMH group and five from MI group). All patients had undergone CT abdomen with plain and contrast phases (multidetector spiral CT scanner, Siemens Somatom Sensation 64, Siemens AG, Munich, Germany). All the imaging features were reevaluated by two senior abdominal radiologists with blinded diagnosis (images archived and reviewed on a workstation named Amrita medvision version 4.2.2686, Amrita Technologies, India). The location (proximal bowel [duodenum and jejunum] vs. distal bowel [ileum and colon]), length of involvement (short vs. long segment, arbitrarily defined by a cutoff value of 15 cm) [4], extent of mural thickening, fat stranding, dilatation of the bowel, presence of hemoperitoneum or ascites, pattern of attenuation, and other obvious findings if present were studied in both the groups. Early clinical symptoms and international normalized ratio (INR) for prothrombin time were correlated with these radiological findings and were compared between the two groups. Clinical signs were indistinguishable early in the course of both the diseases. Also, features of systemic toxicity and abdominal signs were very evident with disease progression late in the course of MI. So, the clinical signs were not considered in our study, as our aim was to differentiate both early in the course of disease. Standard statistical methods with chi-square test (for categorical variables) and nonparametric Mann-Whitney U test (for continuous variables) were used to compare each variable. All analyses were done using SPSS.v.18.
Results Initially, we identified 19 patients with IMH and 59 patients with MI. Of these, three patients with IMH and five patients with MI were excluded from the study, leaving 16 patients in IMH group (group I) and 54 patients in the MI group (group II), based on set
inclusion and exclusion criteria. Age and sex distribution did not differ significantly between the two groups (Table 1). Symptoms like abdominal pain, vomiting, and distension were present in both the groups, and the difference was statistically not significant. However, melena or bleeding per rectum as a symptom was more common in IMH group, and the difference was statistically significant (p-value-0.033) when comparing with MI. The prothrombin time-international normalized ratio (PTINR) was prolonged and more than three in all but one patient (93.7 %) in IMH group, while it was less than three in 55.5 % patients with MI. The p-value was significant (p=0.001) when this cutoff value of three for PT-INR was considered. On evaluation of CT findings (Table 2), IMH was mainly identified in proximal bowel segments in comparison to MI. The site distribution of IMH was esophagus 6.2 %, duodenum 6.2 %, jejunum 62.5 %, ileum 25 % and that of MI was jejunum 13 %, ileum 33 %, jejunum and ileum 26 %, colon 17 %, ileum and colon 11 %. On comparing the two groups of patients (IMH and MI) based on the site of involvement (considering duodenum and jejunum as proximal and ileum and colon as distal), the difference was statistically significant (p-value 0.001). The length of the involved bowel segment was arbitrarily defined as short segment and long segment involvement based on a cutoff value of 15 cm [4]. Of patients with IMH, 87.5 % fell under the category of short segment, whereas the majority of MI (74 %) had long segment bowel involvement. The difference was statistically significant with a p-value of 0.006. The median thickness of the involved bowel was 17.6 mm in IMH group and 8.2 mm in MI group, showing the bowel wall thickening to be significantly associated with IMH (p-value 0.001). The attenuation value of the mural thickening calculated in plain CT had a median value of 65.2 Hounsfield units (HU) for IMH and 34.7 for MI showing a statistically significant difference (p-value 0.001). All the patients with IMH were showing an attenuation value above 40 HU.
Table 1 Comparison of demography and clinical findings
Sex, male/female Age, years Pain abdomen Vomiting Rectal bleeding/melena PT-INR >3
IMH (n=16)
MI (n=54)
p-value
12/4 54 (51–60)# 16/16 (100) 14/16 (87.5) 6/16 (37.5) 15/16 (93.75)
36/18 49 (45–61)# 46/54 (85.18) 34/54 (62.96) 4/54 (7.4) 24/54 (44.44)
0.508 0.355 0.247 0.189 0.033* 0.001**
All values in parentheses, unless marked otherwise, represent percentages IMH intramural bleeds and hematomas, MI mesenteric ischemia, PT-INR prothrombin time-international normalized ratio # Expressed value—median and interquartile range; *significant at 5 % level (p
