blood in the belly: ct findings of hemo- peritoneum

EDUCATION EXHIBIT 109

Blood in the Belly:CT Findings of Hemo-peritoneum1

LEARNINGOBJECTIVESFOR TEST 4After reading thisarticle and takingthe test, the reader

will be able to:

� Recognize the CTmanifestations ofhemoperitoneum.

� Describe the spec-trum of disease asso-ciated with traumaticand nontraumatichemoperitoneum.

� Identify key CTfeatures that mayhelp direct manage-ment.

Meghan Lubner, MD ● Christine Menias, MD ● Creed Rucker, MDSanjeev Bhalla, MD ● Christine M. Peterson, MD ● Lisa Wang, MDBrett Gratz, MD

Hemoperitoneum may occur in various emergent conditions. In thetrauma setting, evidence of intraperitoneal blood depicted at computedtomography (CT) should lead the radiologist to conduct a carefulsearch of images for the injured visceral organ (the liver or spleen).Specific CT signs, such as a sentinel clot or extravasation of intravascu-lar contrast material, may indicate the source of bleeding and help di-rect management. In addition, the configuration of accumulated bloodmay help identify the injured organ; for example, triangular fluid col-lections are observed in the mesentery most often in the setting ofbowel or mesenteric injury. Less commonly, hemoperitoneum mayhave a nontraumatic origin. Iatrogenic hemoperitoneum may occur asa complication of surgery or other interventional procedures in the ab-dominal cavity or as a result of anticoagulation therapy. Hemoperito-neum also may be seen in the setting of blood dyscrasias such as hemo-philia and polycythemia vera. Tumor-associated hemorrhage, whichmost often occurs in hepatocellular carcinoma, hepatic adenoma, orvascular metastatic disease, also may produce hemoperitoneum. Otherpotential causes of nontraumatic hemoperitoneum are gynecologicconditions such as hemorrhage or rupture of an ovarian cyst and rup-ture of the gestational sac in ectopic pregnancy, and hepatic hematomain syndromic hemolysis with elevated liver enzymes and low plateletcount (HELLP syndrome). Vascular lesions (visceral artery aneurysmsand pseudoaneurysms) that occur in systemic vascular diseases such asEhlers-Danlos syndrome or in pancreatitis are another less commonsource of hemoperitoneum.©RSNA, 2007

Abbreviation: HELLP � hemolysis, elevated liver enzymes, and low platelet count

RadioGraphics 2007; 27:109–125 ● Published online 10.1148/rg.271065042 ● Content Codes:

1From the Mallinckrodt Institute of Radiology, Washington University School of Medicine, 510 S Kingshighway Blvd, Box 8131, St Louis, MO63110. Recipient of a Certificate of Merit award for an education exhibit at the 2003 RSNA Annual Meeting. Received March 27, 2006; revision re-quested May 3 and received June 2; accepted June 13. All authors have no financial relationships to disclose. Address correspondence to M.L.(e-mail: [email protected]).

©RSNA, 2007

CME FEATURESee accompanying

test at http://www.rsna.org

/education/rg_cme.html

See last page

TEACHING POINTS

IntroductionAlthough ultrasonography (US) may be used forthe evaluation of hemoperitoneum in the traumapatient, computed tomography (CT), because ofits speed and proximity to the trauma bay, is usedincreasingly for such evaluations. CT has highsensitivity for the detection of even small effusionsof blood in the peritoneal cavity (1). NumerousCT signs and findings of hemoperitoneum, in-cluding a sentinel clot, active arterial extravasa-tion, and mesenteric fluid, may enable the radi-ologist to locate sources of intraperitoneal hemor-rhage and help direct management. Trauma-relatedinjury to a solid organ, particularly to the liver orspleen, is by far the most common cause of hemo-peritoneum. However, the radiologist must beaware of the myriad of other possible causes ofhemoperitoneum that are unrelated to trauma,especially hemorrhage from a tumor, rupture ofan ovarian cyst or of the affected anatomy in anectopic pregnancy, and iatrogenic bleeding as aresult of surgery or other therapy. This articleprovides an overview of the key CT findings seenin patients with traumatic and nontraumatic he-moperitoneum (Table).

CT Signs of Hemoperitoneum

Peritoneal SpacesThe peritoneal cavity contains the liver, spleen,bowel, stomach, and mesentery. In the supineposition, the most dependent portion of the abdo-men is the hepatorenal fossa (Morison pouch),and the most dependent portion of the pelvis isthe pelvic cul-de-sac (pouch of Douglas). Hemo-peritoneum starts near the site of injury and flowsalong expected anatomic pathways. Hemorrhagefrom the liver typically flows in a caudal directionfrom the perihepatic spaces and hepatorenalfossa, along the right paracolic gutter, and intothe cul-de-sac, which is the rectouterine space inwomen and rectovesical space in men (Fig 1).Similarly, hemorrhage from the spleen typicallyflows in a caudal direction from the perisplenicspaces, along the left paracolic gutter, and intothe pelvis (2). A large volume of blood may col-lect in the pelvis without an obvious hematomasurrounding the source organ. Thus, the appear-ance of the cul-de-sac is crucial to the diagnosisof hemoperitoneum. Small amounts of fluid orblood in this region may be the only sign of anintraperitoneal injury and should prompt a carefulinspection of images of the viscera (3–6).

It is important to bear in mind that, in traumacases, significant abdominal injuries may be

Causes and CT Findings of Hemoperitoneum

Cause Findings

TraumaticSolid organ injury Sentinel clot sign with accumulations of high-attenuation

fluid near or around the injured organ, in the paracolicgutters, and in the pelvis; serpiginous border of high-attenuation area suggestive of active extravasation ofcontrast material

Mesenteric or bowel injury Triangular high-attenuation interloop mesenteric fluidcollections; bowel wall thickening; high-attenuationfree fluid indicative of active extravasation of intrave-nous contrast material from mesenteric tear or of oralcontrast material from site of bowel injury

NontraumaticIatrogenic injury (complication of surgery,

interventional procedure, or anticoagula-tion therapy)

High-attenuation fluid in surgical bed, at site of interven-tion, or around liver and spleen in patient undergoinganticoagulation therapy

Tumor-associated hemorrhage Tumor site in liver, spleen, or peritoneum, with sur-rounding high-attenuation fluid; sentinel clot sign

Gynecologic condition (ruptured ovarian cyst,ectopic pregnancy, HELLP syndrome)

Fluid with internal echogenicity or high attenuationaround the uterus and adnexa; associated mixed-attenuation adnexal mass with a high-attenuationcomponent or fluid-fluid level

Vascular lesion (aneurysm or pseudoaneu-rysm of visceral artery)

Accumulation of contrast material within an apparentaneurysmal sac surrounded by high-attenuation freefluid; clinical history of systemic vascular disease (eg,Ehlers-Danlos syndrome) or pancreatitis

110 January–February 2007 RG f Volume 27 ● Number 1

present without apparent hemoperitoneum. Forexample, intraparenchymal lacerations or contu-sions may be present without penetration of theorgan capsule. Knowledge of the relationship ofabdominal organs to the surrounding peritonealspaces is crucial. For example, because the bare areaof the liver is contiguous with the retroperitoneum,retroperitoneal fluid rarely may be the only sign ofhepatic injury, although intraperitoneal fluid is morecommonly seen in cases of solid organ injury. In apatient with abdominal trauma, a CT finding of freeintraperitoneal fluid without solid organ injuryshould arouse suspicion about the possible presenceof a bowel laceration or perforation (7,8).

High Attenuation of BloodThe utility of CT for the detection of hemoperito-neum lies in its ability to depict the difference be-tween low-attenuation fluid and blood, which hasa higher attenuation. Attenuation measurementsat CT can help differentiate among simple as-

cites, unclotted blood from recent bleeding, he-matoma, bile, urine, chyle, and active bleeding(9). The attenuation values of fluids that haveapproximately the same density as water (eg, bile,urine, and the intestinal contents) range from 0 to15 HU. The recognition of such fluids on CTimages can be aided by visual comparison withadjacent fluid-filled structures such as the urinarybladder and gallbladder. Attenuation should bedirectly measured in cases in which the differ-ences are subtle, because the appearance may bemisleading. The administration of both intrave-nous and water-soluble oral contrast material maybe helpful; the presence of free fluid from a bowelperforation can be confirmed by showing free in-traperitoneal water-soluble oral contrast material,and urinary ascites can be distinguished from he-moperitoneum with delayed phase (urographicphase) CT or CT cystography (Fig 2).

Figure 1. Typical flow pattern of hemorrhage from a liver laceration. (a) Contrast material–enhanced CT image showshigh-attenuation fluid (arrows) in perihepatic and perisplenic regions, the hepatorenal fossa, and the lesser sac. (b) Con-trast-enhanced CT image at a lower level than a shows blood tracking laterally along the paracolic gutters (arrows).

Figure 2. Free intraperitoneal fluid after trauma. (a) Contrast-enhanced CT image shows a large volume ofhigh-attenuation free intraperitoneal fluid. A subsequent CT cystogram revealed that most of the fluid wasurine leaking from an injured bladder. (b) Contrast-enhanced CT image shows leakage of urine from a bladderrupture at the 1-o’clock position (arrow).

RG f Volume 27 ● Number 1 Lubner et al 111

Blood usually has higher measured attenuationthan other body fluids; however, its attenuationmay vary, and its CT appearance often dependson the age, extent, and location of the hemor-rhage. Because of its high protein content, unclot-

ted extravascular blood usually has a measuredattenuation of 30–45 HU; however, its attenua-tion must be assessed on the basis of the indi-vidual patient. For example, hemoperitoneummay have a measured attenuation of less than30 HU in a patient with a decreased serum he-matocrit level (2) (Fig 3) or in a patient with ahemorrhage that is more than 48 hours old(10).

Figure 3. Low-attenuation hemoperitoneum in a hemophiliac patient with a splenic rupture. (a, b) Contrast-en-hanced CT images (b at a lower level than a) show much lower attenuation in the intraperitoneal hemorrhage than inthe adjacent muscle. The patient’s hematocrit level was 0.24 at the time of CT.

Figure 4. Sentinel clot sign. (a) Contrast-enhanced CT image shows a sentinel clot secondary to laceration alongthe fissure for the ligamentum teres (arrow), in the perihepatic space and lesser sac. (b) Contrast-enhanced CT im-age obtained in a patient who was undergoing anticoagulation therapy for a protein C deficiency shows a sentinel clot(arrow) surrounding the spleen.


TeachingPoint

Teaching Point Blood usually has higher measured attenuation than other body fluids; however, its attenuation may vary, and its CT appearance often depends on the age, extent, and location of the hemorrhage.

The measured attenuation of clotted blood is45–70 HU. The body attempts hemostasis byforming a blood clot at the site of injury. Thus, onCT images, the highest-attenuation hematoma,or sentinel clot, is that closest to the site of bleed-ing, whereas lower-attenuation unclotted blood islocated farther from the source (11). The sentinelclot sign is useful for identifying the dominant

source of hemoperitoneum in patients with mul-tiple injuries from trauma (Fig 4). The detectionof a high-attenuation clot may be facilitated byusing narrow window settings. The sentinel clotsign also may be valuable in cases of nontrau-matic hemoperitoneum (Fig 5). Occasionally, theCT appearance of fresh blood in the peritonealcavity is altered by the hematocrit effect, in whichsedimented red blood cells produce a dependentlayer of high attenuation within the hemoperito-neum (Fig 6).

Active arterial extravasation of contrast mate-rial is depicted at CT with a measured attenua-tion value higher than that of free or clottedblood, a feature indicative of the presence of in-travenous contrast material in the blood. In a se-ries of trauma patients, Shanmuganathan et alfound attenuation values of extravasated contrastmaterial that ranged from 85 to 370 HU, with amean of 132 HU (12). Although in many caseshemoperitoneum can be managed nonsurgicallywith excellent results (13–15), a CT finding ofactive bleeding is indicative of a need for emer-gent embolization or surgical treatment (12,16,17). The visually apparent attenuation in an areaof active bleeding resembles that in the aorta or

Figure 5. Bleeding hepatoma. (a) Unenhanced CT image shows perihepatic and perisplenic fluid, with an area ofslightly higher attenuation (arrow) indicative of a clot along the left hepatic lobe. (b) Contrast-enhanced CT imageshows the origin of the clot, a ruptured mass (arrow) that subsequently was proved to be a hepatocellular carcinoma,as well as active extravasation (arrowhead).

Figure 6. Hematocrit effect. Contrast-enhanced CTimage shows an obvious layer of high attenuation pro-duced by sedimented blood cells (arrow) in a pelvichemoperitoneum.


TeachingPoint

TeachingPoint

Teaching Point On CT images, the highest-attenuation hematoma, or sentinel clot, is that closest to the site of bleeding, whereas lower-attenuation unclotted blood is located farther from the source.

Teaching Point Although in many cases hemoperitoneum can be managed nonsurgically with excellent results, a CT finding of active bleeding is indicative of a need for emergent embolization or surgical treatment.

adjacent arteries. A focus of active bleeding mayappear as a serpiginous or amorphous high-attenuation area intermixed with or surroundedby a large hematoma (Fig 7). On images obtainedwith soft-tissue window settings, active extravasa-tion may have attenuation that allows it to be mis-taken for adjacent bone or a bone fracture. Theuse of bone window settings in the evaluation ofimages of the region should prevent such confu-sion. Free oral contrast material from a perforatedbowel also may appear as a focal area of high at-tenuation, but there is no surrounding hematomalike that seen in arterial extravasation.

Mesenteric FluidHemorrhage from the bowel or mesentery typi-cally flows into areas between bowel loops,termed interloop spaces. Mesenteric fluid has acharacteristic triangular shape due to the mesen-teric leaves that surround the bowel and entrapfluid (Fig 8). Liver or splenic hemorrhage moretypically descends peripherally along the paracolicgutters into the pelvis and is not entrapped in in-terloop spaces. Thus, centrally located, triangularareas of high-attenuation abdominal fluid shouldprompt a search for intraperitoneal bowel or mes-enteric injury (7,8).

Traumatic Hemoperitoneum

Solid Organ InjuryThe spleen is the most frequently injured organ inblunt abdominal trauma; splenic injuries accountfor 40% of abdominal organ injuries (18). Injuries

Figure 7. Active arterial extravasation. Note the sentinel clot sign produced by serpiginous areas of high attenua-tion surrounded by lower-attenuation areas of hematoma. (a) Contrast-enhanced CT image shows active bleedingfrom a branch of the internal iliac artery (arrow) in a patient with pelvic fractures. (b) Contrast-enhanced CT imageshows active bleeding from a splenic rupture due to blunt trauma.

Figure 8. Mesenteric tear. Contrast-enhanced CTimage depicts high-attenuation triangular fluid collec-tions (arrows) within the mesentery of the distal smallbowel in a patient with a surgically proved ileal mesen-teric tear.


to the liver account for 20% of solid organ injuriesrelated to blunt abdominal trauma (19). How-ever, the liver is the single most commonly in-jured organ when blunt and penetrating traumaare combined. There are numerous systems forclassifying visceral injuries, all of which includethe following categories, in some form: intrapa-renchymal contusion or hematoma, subcapsularhematoma, laceration, fracture, and vascularpedicle injury (Fig 9) (20–22).

In the trauma set-ting, CT can help accurately diagnose the typeand extent of the primary injury, determine thepresence and amount of hemoperitoneum, anddetect associated injuries (19,23).

Traumatic injury to the right hepatic lobe, es-pecially the posterior segment, is much morecommon than is left lobe trauma, whereas injuryto the caudate lobe is rare. The presence of small

Figure 9. Spectrum of hepatic and splenic injuries due to trauma. (a) Contrast-enhanced CT image shows a largesubcapsular hepatic hematoma with laceration and periportal edema and with a typical indentation in the fibrouscapsule of the liver. (b) Contrast-enhanced CT image shows multiple irregular low-attenuation contusions in bothhepatic lobes (arrows). (c) Contrast-enhanced CT image depicts a single hepatic laceration that extends alongsidethe fissure of the ligamentum teres (arrow). Lacerations commonly occur along the natural planes of hepatic vesselsand fissures. (d) Contrast-enhanced CT image shows a single jagged laceration in the posterior tip of the spleen (ar-row).


TeachingPoint

Teaching Point In the trauma setting, CT can help accurately diagnose the type and extent of the primary injury, determine the presence and amount of hemoperitoneum, and detect associated injuries.

amounts of contrast material in the liver or spleenwith adjacent hemoperitoneum may be indicativeof arterial injury or pseudoaneurysm formationand should prompt surgical or angiographic inter-vention. Poletti et al (24) reported that major vas-cular injury is more likely in the presence of high-grade liver lacerations that involve branches of thehepatic or portal veins or from which there is ac-tive bleeding, and these findings should promptangiography. There is also a high likelihood ofvenous disruption and biliary injury in the pres-ence of lacerations that extend into the region ofthe inferior vena cava or the major hepatic veins(25) (Fig 10). Fang et al (16,17) showed a signifi-cant correlation between the need for surgicalintervention and the presence of hepatic and in-traperitoneal bleeding or parenchymal contrastmaterial extravasation in association with hemo-peritoneum (Fig 11).

Bowel and Mesenteric InjuriesBowel and mesenteric injuries are found in ap-proximately 5% of patients who undergo surgeryfor blunt abdominal trauma (26). Bowel and mes-enteric injuries usually occur together but also

may occur separately, with purely mesentericavulsions sometimes seen in the setting of seat-belt-related injuries. Bowel injuries can be diffi-cult to detect clinically; abdominal rigidity, pain,and decreased bowel sounds are found during thephysical examination of only one-third of patientsat initial presentation (27). Undiagnosed bowelinjuries are associated with significant morbidityand mortality from peritonitis, sepsis, and hemor-rhage (6).

CT has high accuracy for the detection ofbowel and mesenteric injuries (28). Various in-vestigators in previous studies reported sensitivityof 84%–94% and accuracy of 84%–99% for thediagnosis of bowel injury at CT (29,30). Thepresence of free air is a less common sign of bowelperforation than is the presence of free fluid, espe-cially in cases of small-bowel injury; the intralu-minal contents of the small bowel consist mostly

Figure 10. (a–c) Contrast-enhanced CT images atprogressively lower levels show a liver laceration extend-ing to the level of the inferior vena cava, with evidence ofactive extravasation and hemoperitoneum, in a patientafter a motor vehicle collision. At surgery, the liver wasfound to be totally avulsed from the inferior vena cava,leaving the latter open at the level of the diaphragm.


of fluid rather than gas. More reliable signs ofbowel or mesenteric injury are interloop fluid,mesenteric hematoma or infiltration, wall thick-ening with associated enhancement, and bowelwall discontinuity (31,32). Oral contrast materialextravasation, though rare, is the most specificsign of bowel rupture. Active extravasation of in-travenous contrast material can be seen in thesetting of mesenteric vessel injury (Fig 12).

The intraperitoneal bowel segments that aremost commonly affected by trauma are the por-tion of the jejunum near the ligament of Treitzand the portion of the ileum near the ileocecalvalve, presumably because these segments aremore fixed than the rest of the small bowel. Injuryoccurs more commonly on the antimesentericside than on the mesenteric side.

Nontraumatic HemoperitoneumNontraumatic hemoperitoneum may occur spon-taneously or may be iatrogenic. Spontaneous he-moperitoneum is an uncommon and often unsus-pected condition that may be catastrophic if it isnot promptly diagnosed and treated. It may beassociated with severe abdominal pain and disten-tion, a decreased hematocrit level, or, more rarely,hypovolemic shock (33). However, these signsmay not be present or may be equivocal, and thehematocrit level in a serum sample obtained atinitial presentation may not reflect the extent of ahemorrhage (34). In such cases, the diagnosisfrequently is made on the basis of radiologic find-ings (35,36). Nontraumatic hemoperitoneum hasvarious possible causes, including hemorrhagefrom a highly vascular neoplasm (tumor-associ-ated hemorrhage), hemorrhage or rupture of anovarian cyst, rupture of the gestational sac or

Figure 11. Liver laceration with pseudoaneurysm formation. (a) Contrast-enhanced CT image shows a large liverlaceration with internal pooling of high-attenuation material (arrow), a finding suggestive of an aneurysm or pseudo-aneurysm. (b) Angiogram shows a pseudoaneurysm (arrow).

Figure 12. Mesenteric tear. Contrast-enhanced CTimage shows active extravasation from jejunal mesen-teric vessels with associated jejunal injury and mesen-teric hemorrhage indicated by a triangular region ofhemoperitoneum (arrow).


other affected anatomic part in an ectopic preg-nancy, and bleeding from a vascular lesion suchas an arterial aneurysm. Nontraumatic hemoperi-toneum also may result from anticoagulationtherapy, surgery, or an invasive procedure.

In nontraumatic hemoperitoneum, as in trau-matic hemoperitoneum, it is critical to identifyany active bleeding. A finding of active bleedingshould lead to the abandonment of a conservativemanagement strategy in favor of surgical or endo-vascular intervention. After active bleeding hasbeen excluded, a careful search should be under-taken for the source of hemoperitoneum (35).

Iatrogenic CausesAny surgical procedure performed within theperitoneal cavity may be complicated by hemo-peritoneum (Fig 13). Even minimally invasive per-cutaneous or endovascular procedures occasionallylead to intraperitoneal hemorrhage (Figs 14, 15).

Bleeding diatheses are common in patientswho are undergoing anticoagulation therapy withheparin or coumadin. In most cases, the risk ofbleeding is directly related to the intensity of theanticoagulant effect (37). Anticoagulation therapymost commonly causes hemorrhage into thepsoas or rectus muscles, but it also occasionallyresults in hemoperitoneum. Hemoperitoneumdue to rupture of the gallbladder or spleen hasbeen reported in patients undergoing anticoagula-tion therapy (38,39). The use of CT to evaluatesuch patients for abdominal hemorrhage had adirect effect on clinical management for aboutone-half of the patients in one study, and positiveand negative findings at CT were equally likely toaffect management (40).

Like anticoagulation therapy, blood dyscrasiasmay alter coagulation status and lead to hemo-peritoneum (Fig 16).

Tumor-associated HemorrhageSpontaneous hemoperitoneum rarely occurs inthe absence of trauma, a surgical or interventionalprocedure, or anticoagulation therapy (41). Insuch cases, the possibility of the rupture of an un-identified neoplasm must be excluded. Althoughthe occurrence is uncommon, any primary or

metastatic tumor can rupture and bleed into theperitoneal cavity.

The most common cause of nontraumatic he-patic hemorrhage is a hypervascular neoplasmsuch as hepatocellular carcinoma or hepatocellu-lar adenoma (36). Hepatocellular carcinoma iscommonly seen in the setting of cirrhosis orchronic hepatic inflammation and is the mostcommon primary malignancy of a solid organ inthe global population (42). Rupture has been re-ported to occur in 6.9%–14% of cases of hepato-cellular carcinoma in Asia and Africa, where theprevalence of this type of cancer is high, and it isthe most common cause of nontraumatic hemo-peritoneum in male patients of all ages; however,ruptured hepatocellular carcinoma is less com-mon in Europe and the United States (43,44).Patients with large or peripherally located tumorsdevoid of normal overlying tissue are at a higherrisk for rupture (45). There also seems to be ahigher likelihood of rupture in patients with bothcirrhosis and hepatocellular carcinoma (36,46).In these cases, a wide spectrum of imaging find-ings can be seen, ranging from minor intrahepaticbleeding to subcapsular hemorrhage and rupturethrough the hepatic capsule into the peritoneum.Blood-tinged ascites is present in almost all cases(47). Necrosis and hemorrhage also may occur inless common hypervascular hepatic tumors, such

Figure 13. Hematoma after a pancreatoduodenec-tomy (Whipple operation). Contrast-enhanced CT im-age shows a large abdominal hematoma that involvesthe left perihepatic and perisplenic space, with evidenceof active extravasation (arrow).


TeachingPoint

Teaching Point The most common cause of nontraumatic hepatic hemorrhage is a hypervascular neoplasm such as hepatocellular carcinoma or hepatocellular adenoma.

as hepatic angiosarcoma (Fig 17) (48). Hemangi-omas, though common, are not associated withnontraumatic hemoperitoneum. A lesion that re-sembles a hemangioma on CT images obtained in

a patient with nontraumatic hemoperitoneummay actually be an atypical angiomatous hepaticlesion such as an angiosarcoma.

Figures 14, 15. Hematoma after interventional procedures. (14) Contrast-enhanced CT image shows a hematomawithin the lesser sac after a laparoscopic Nissen fundoplication. (15) Contrast-enhanced CT image in another patientdepicts a large perigastric hematoma (arrow) after a percutaneous gastrostomy tube placement.

Figure 16. Splenic rupture. (a) Contrast-enhanced CT image shows a splenic rupture and pooling of contrast ma-terial in a patient with splenomegaly due to polycythemia vera. (b) Angiogram reveals multiple splenic pseudoaneu-rysms that were subsequently treated with embolization.


The spontaneous rupture of a metastatic lesionin a solid organ is rare but usually results in mas-sive hemoperitoneum (47). Lung carcinoma, re-nal cell carcinoma, and melanoma are the meta-static lesions that most often cause hemoperito-neum (Figs 18, 19) (41).

Hemoperitoneum associated with a benignneoplasm most often occurs with hepatic ad-enoma. These tumors are strongly associated withoral contraceptive use and estrogen steroid therapy;they most often are found in women of childbear-ing age who have a history of prolonged oral con-traceptive use (49). The identification of a hepaticadenoma is essential, given the associated risk oflife-threatening hemorrhage (50). Focal nodular

Figure 17. Ruptured hepatic angiosarcoma. (a, b) Contrast-enhanced CT images (a at a higher level than b) showa heterogeneous low-attenuation lesion in the left hepatic lobe, with evidence of increased vascularity, active extrava-sation, and hemoperitoneum. At pathologic analysis, the lesion was found to be a hepatic angiosarcoma. The high-attenuation material adjacent to the lesion represents perihepatic blood secondary to a capsular rupture.

Figures 18, 19. Hemorrhagic metastases. (18) Contrast-enhanced CT image shows marked hemoperitoneum withactive extravasation in a patient with metastatic peritoneal melanoma. Note the sentinel clot that surrounds the tu-mor implant, and the subcutaneous melanotic nodule (arrowhead) in the left posterior region. (19) Contrast-en-hanced CT image depicts multiple hemorrhagic hepatic lesions with central high attenuation in a patient with knownmetastatic testicular choriocarcinoma. Hemoperitoneum (arrow) is seen adjacent to the largest lesion.


hyperplasia, in contrast, is not usually accompa-nied by spontaneous hemorrhage. Large adeno-mas are prone to bleed (51), and because the tu-mor capsule is incomplete, a hemorrhage mayspread into the liver or the abdominal cavity (Fig20) (52). For these reasons, although a very lowrate of malignant degeneration has been reported,hepatic adenomas are surgically resected (53).

Pathologic splenic rupture may occur as acomplication of a viral infection, including infec-

tion by cytomegalovirus, malaria, or Epstein-Barrvirus (54); a congenital disease; a metabolic ab-normality such as Gaucher disease or amyloido-sis (55); and, rarely, a neoplastic process such ashemangiomatosis, angiosarcoma, leukemia, orlymphoma (Figs 21, 22) (56). The prevalenceof bleeding has been correlated directly with

Figure 20. Ruptured hepatocellular adenoma. (a) Contrast-enhanced CT image shows a large heterogeneous low-attenuation mass (arrow) in the right hepatic lobe, and a loss of liver capsule integrity anterolaterally. (b) Coronalvolume-rendered CT image shows the peripherally enhancing mass, ruptured liver capsule, and perihepatic hemor-rhage (arrow).

Figures 21, 22. Splenic rupture. (21) Unenhanced CT image shows an enlarged spleen with a perisplenic sentinelclot in a patient with lymphoma and thrombocytopenia who was evaluated for flank pain. (22) Contrast-enhancedCT image shows a hemoperitoneum that surrounds an enlarged ruptured spleen in a patient with malaria. (Fig 22courtesy of Howard J. Harvin, MD, Scottsdale, Ariz.)


increased splenic weight (56). At times, evenseemingly minor trauma, such as that caused bylying prone, can induce hemorrhage in patientswith splenomegaly.

Bleeding Due toGynecologic ConditionsThe primary imaging modality used for the as-sessment of gynecologic conditions is US. How-ever, CT is often requested if the clinical findingsare nonspecific. The reproductive tract is themost common source of spontaneous hemoperi-toneum in women of childbearing age. The mostcommon gynecologic sources of bleeding are ec-topic pregnancy and ruptured ovarian cyst (57,58). Hemoperitoneum also has been seen in thesetting of endometriosis and uterine rupture, butsuch occurrences are uncommon (36,59).

A hemorrhagic ovarian cyst (usually, a corpusluteal or follicular cyst) is a frequent cause ofacute pelvic pain in women of childbearing age(57). Accurate diagnosis of a hemorrhagic ovariancyst is important because, although it rarely causesa clinically significant blood loss, the clinicalmanifestations may be confused with those ofconditions that require immediate surgery, suchas appendicitis (60,61). Hemorrhage into anovarian cyst is a relatively common occurrence,but few hemorrhagic ovarian cysts rupture andcause hemoperitoneum (61). Fluid with internalechogenicity or high attenuation can be seen sur-rounding the uterus and adnexa in the presence ofrecent bleeding, and an associated mixed-attenu-

ation adnexal mass with a high-attenuation com-ponent and at times a fluid-fluid level may bepresent (Fig 23) (57).

Ectopic pregnancy can cause life-threateningbleeding and therefore must be considered in ev-ery woman of childbearing age who presents withabdominal or pelvic pain. Ectopic pregnancy ac-counts for up to 1% of pregnancies, with 97% ofoccurrences located in either the ampullary (mostcommon) or the isthmic portion of the fallopiantube (62). Risk factors include previous ectopicpregnancy, pelvic inflammatory disease, in vitrofertilization, use of an intrauterine device, andtubal surgery (63,64). In the context of a positivehuman chorionic gonadotropin level of more than2000 IU/L and no intrauterine pregnancy, thefinding of an extraovarian mass is highly sugges-tive of an ectopic pregnancy (Fig 24) (58). He-moperitoneum in a patient with an ectopic preg-nancy is not necessarily indicative of tubal rup-ture; however, the larger the amount of fluid, thehigher the likelihood of tubal rupture (36).

HELLP syndrome is a peripartum conditionthat comprises the clinical triad of hemolysis, el-evated liver enzymes, and low platelet count(65,66). Disseminated intravascular coagulationoccurs in 20%–40% of patients with HELLP syn-drome, and other complications, including he-patic infarction, hematoma, hepatic rupture, andplacental abruption, may occur (67). CT may bethe modality of choice for the demonstration ofsubcapsular hematoma, hepatic rupture, and he-moperitoneum (Fig 25) (36,68). Hepatic infarc-tion appears on CT images as a wedge-shapedperipheral area with low attenuation and withouta mass effect (68).

Figure 23. Ruptured hemorrhagic ovarian cyst. (a) Contrast-enhanced CT image shows an adnexal cyst withan internal hematocrit level (arrow), surrounded by a pelvic hemorrhage. A curvilinear focus of extravasation(arrowhead) visible posterior to the cyst was found at surgery to represent active bleeding from an internal iliacartery branch. (b) Contrast-enhanced CT image from a different patient shows an ovarian cyst (arrow) andsurrounding hemorrhage with a hematocrit effect.


Vascular SourcesIntraperitoneal bleeding from vascular lesions (eg,ruptured abdominal aortic aneurysms) is less com-mon than retroperitoneal hemorrhage from suchlesions. However, abdominal aortic aneurysms

with very large leaks occasionally result in bleed-ing that extends into the peritoneal cavity (69).

Splenic artery aneurysms account for 60% ofvisceral artery aneurysms (70). Splenic artery an-eurysms are four times as common in women asthey are in men, and the rate of rupture is particu-larly increased during pregnancy (71). Spontane-ous rupture occurs in 3%–10% of splenic arteryaneurysms and produces a high mortality rate,estimated to be approximately 36% (72). Hepaticartery aneurysms are the second most commontype (20%) of visceral artery aneurysms (73). Inyoung patients, splanchnic artery aneurysmsshould precipitate a thorough search for systemicvascular disease, most notably type IV Ehlers-Danlos syndrome. These patients may presentwith a spontaneous aneurysm rupture, and an-giography may be contraindicated because of therisk of aneurysm formation at the site of punc-ture. Pseudoaneurysms of the hepatic, splenic,and gastroduodenal arteries are complicationsthat must be excluded during an imaging evalua-tion for pancreatitis (Fig 26) (74).

SummaryThere are numerous causes of hemoperitoneumin both traumatic and nontraumatic settings. CTis the imaging modality of choice because of itsability to help distinguish blood from other fluids.The crucial initial steps for the radiologist are touse the various CT signs to detect the presence ofintraperitoneal blood, locate the source of hemor-rhage, and determine whether emergent interven-tion is indicated. After these steps are completed,a differential diagnosis may be generated on thebasis of the clinical history and radiologic findings.

Figure 24. Adnexal rupture due to an ectopic preg-nancy. CT image shows a hemorrhagic right adnexalmass that corresponds to an ectopic gestational sacwith associated active extravasation (arrow) and bloodin the cul-de-sac (arrowhead). Ectopic pregnancy andrupture were proved at surgery. (Courtesy of AlvaroHuete, MD, Santiago, Chile.)

Figure 25. HELLP syndrome. Coronal CT imagedemonstrates a subcapsular hematoma (arrow) in arecent postpartum patient with a history of HELLPsyndrome. Note the hypervascularity of the postpartumuterus.

Figure 26. Hemorrhagic pancreatitis. Contrast-en-hanced CT image shows a hematoma in the gastrocolicligament, with a focus of active extravasation (arrow).At angiography, the bleeding vessel was found to be apancreatic branch of the splenic artery. Subsequentembolization was successful.


References1. Federle MP, Goldberg HI, Kaiser JA, et al. Evalu-

ation of trauma by computed tomography. Radiol-ogy 1981;138:637–644.

2. Shuman WP. CT of blunt abdominal trauma inadults. Radiology 1997;205:297–306.

3. Roberts JL. CT of abdominal and pelvic trauma.Semin Ultrasound CT MR 1996;17:142–169.

4. Raptopoulos V. Abdominal trauma: emphasis oncomputed tomography. Radiol Clin North Am1994;32:969–987.

5. Wolfman NT, Cechtold RE, Scharling ED,Meredith JH. Blunt upper abdominal trauma:evaluation by CT. AJR Am J Roentgenol 1992;158:493–501.

6. Shanmuganathan K. Multi-detector row CT im-aging of blunt abdominal trauma. Semin Ultra-sound CT MR 2004;25:180–204.

7. Harris HW, Morabito DJ, Mackersie RC, et al.Leukocytosis and free fluid are important indica-tors of isolated intestinal injury after blunt trauma.J Trauma 1999;46:656–659.

8. Brasel KJ, Olson CJ, Stafford RE, et al. Incidenceand significance of free fluid on abdominal com-puted tomographic scan in blunt trauma.J Trauma 1998;44:889–892.

9. Mirvis SE, Rodriguez A. Abdominal and pelvictrauma. In: Mirvis SE, Young JWR, eds. Imagingin trauma and critical care. Baltimore, Md: Wil-liams & Wilkins, 1992; 145–242.

10. Federle MP, Jeffrey RB Jr. Hemoperitoneum stud-ied by computed tomography. Radiology 1983;148:187–192.

11. Orwig D, Federle MP. Localized clotted blood asevidence of visceral trauma on CT: the sentinelclot sign. AJR Am J Roentgenol 1989;153:747–750.

12. Shanmuganathan K, Mirvis SE, Reaney SM. Pic-torial review: CT appearances of contrast mediumextravasations associated with injury sustainedfrom blunt abdominal trauma. Clin Radiol 1995;50:182–187.

13. Pachter HL, Knudson MM, Esrig B, et al. Statusof nonoperative management of blunt hepatic in-juries in 1995: a multicenter experience with 404patients. J Trauma 1996;40:31–38.

14. Croce MA, Fabian TC, Menke PG, et al. Nonop-erative management of blunt hepatic trauma is thetreatment of choice for hemodynamically stablepatients: results of a prospective trial. Ann Surg1995;221:744–755.

15. Delgado Millan MA, Deballon PO. Computedtomography, angiography, and endoscopic retro-grade cholangiopancreatography in the nonopera-tive management of hepatic and splenic trauma.World J Surg 2001;25:1397–1402.

16. Fang JF, Chen RJ, Wong YC, et al. Pooling ofcontrast material on computed tomography man-dates aggressive management of blunt hepatic in-jury. Am J Surg 1998;176:315–319.

17. Fang JF, Chen RJ, Wong YC, et al. Classificationand treatment of pooling of contrast material oncomputed tomographic scan of blunt hepatictrauma. J Trauma 2000;49:1083–1088.

18. Novelline RA, Rhea JT, Bell T. Helical CT of ab-dominal trauma. Radiol Clin North Am 1999;37:591–612.

19. Shanmuganathan K, Mirvis SE. CT evaluation ofthe liver with acute blunt trauma. Crit Rev DiagnImaging 1995;36:73–113.

20. Weishaupt D, Grozaj AM, Willman JK, et al.Traumatic injuries: imaging of abdominal and pel-vic injuries. Eur Radiol 2002;12:1295–1311.

21. Moore EE, Shackford SR, Pachter HL, et al. Or-gan injury scaling: spleen, liver, and kidney.J Trauma 1989;29:1664–1666.

22. Moore EE, Cogbill TH, Jurkovich GJ, ShackfordSR, Malangoni MA, Champion HR. Organ injuryscaling: spleen and liver (1994 revision). J Trauma1995;38:323–324.

23. Ashlock SJ, Harris JH, Kawashima A. Computedtomography of splenic trauma. Emerg Radiol1998;5:192–202.

24. Poletti PA, Mirvis SE, Shanmuganathan K, et al.CT criteria for management of blunt liver trauma:correlation with angiographic and surgical find-ings. Radiology 2000;216:418–427.

25. Savolaine ER, Grecos GP, Howard J, White P.Evolution of CT findings in hepatic hematoma.J Comput Assist Tomogr 1985;9:1090–1096.

26. Rizzo MJ, Federle MP, Griffiths BG. Bowel andmesenteric injury following blunt abdominaltrauma: evaluation with CT. Radiology 1989;173:143–148.

27. Donohue JH, Crass RA, Trunkey DD. Manage-ment of duodenal and small intestinal injury.World J Surg 1985;9:904–913.

28. Nghiem HV, Jeffrey RB Jr, Mindelzun RE. CT ofblunt trauma to the bowel and mesentery. AJRAm J Roentgenol 1993;160:53–58.

29. Malhotra AK, Fabian TC, Katsis SB, et al. Bluntbowel and mesenteric injuries: the role of screen-ing computed tomography. J Trauma 2000;48:991–998.

30. Killeen KL, Shanmuganathan K, Poletti PA, et al.Helical computed tomography of bowel and mes-enteric injuries. J Trauma 2001;51:26–36.

31. Breen DJ, Janzen DL, Swirewich CV, Nagy AG.Blunt bowel and mesenteric injury: diagnostic per-formance of CT signs. J Comput Assist Tomogr1997;21:706–712.

32. Levine CD, Gonzales RN, Wachsberg RH, Gha-nekar D. CT findings in bowel and mesentericinjury. J Comput Assist Tomogr 1997;21:974–979.

33. McCort JJ. Intraperitoneal and retroperitonealhemorrhage. Radiol Clin North Am 1976;14:391–405.

34. Scott WW, Fishman EK, Siegelmann SS. Antico-agulants and abdominal pain: the role of com-puted tomography. JAMA 1984;252:2053–2056.

35. Lucey BC, Varghese JC, Soto JA. Spontaneoushemoperitoneum: causes and significance. CurrProbl Diagn Radiol 2005;34:182–195.

36. Mortele KJ, Cantisani V, Brown DL, Ros PR.Spontaneous intraperitoneal hemorrhage: imagingfeatures. Radiol Clin North Am 2003;41:1183–1201.

37. Coon WW, Wills PW 3rd. Hemorrhagic complica-tions of anticoagulant therapy. Arch Intern Med1974;133:386–392.

38. Pissas A, Danel V, Finet P. A rare complication ofanticoagulants: hemoperitoneum caused by gall-bladder rupture: review of the literature a proposof a case [in French]. J Chir (Paris) 1985;122:235–245.


39. Ghobrial MW, Karim M, Mannam S. Spontane-ous splenic rupture following the administration ofintravenous heparin: case report and retrospectivecase review. Am J Hematol 2002;71:314–317.

40. Nazarian LN, Lev-Toaff AS, Spettell CM, Wech-sler RJ. CT assessment of abdominal hemorrhagein coagulopathic patients: impact on clinical man-agement. Abdom Imaging 1999;24:246–249.

41. Casillas VJ, Amendola MA, Caascue A, Pinnar N,Levi JU, Perez JM. Imaging of nontraumatic hem-orrhagic hepatic lesions. RadioGraphics 2000;20:367–388.

42. Levy AD. Malignant liver tumors. Clin Liver Dis2002;6:147–164.

43. Meyers WC. The liver. In: Sabiston DC Jr, LyerlyHK, eds. Textbook of surgery. 15th ed. Philadel-phia, Pa: Saunders, 1997; 1045–1116.

44. Fernandez MP, Redvanly RD. Primary hepaticmalignant neoplasms. Radiol Clin North Am1998;36:333–348.

45. Kanematsu M, Imaeda T, Yamawaki Y, et al.Rupture of hepatocellular carcinoma: predictivevalue of CT findings. AJR Am J Roentgenol 1992;158:1247–1250.

46. Hirai K, Kawazoe Y, Yamashita K, et al. Trans-catheter arterial embolization for spontaneous rup-ture of hepatocellular carcinoma. Am J Gastroen-terol 1986;81:275–279.

47. Miyamoto M, Sudo T, Kuyama T. Spontaneousrupture of hepatocellular carcinoma: a review of172 Japanese cases. Am J Gastroenterol 1991;86:67–71.

48. Molina E, Hernandez A. Clinical manifestationsof primary hepatic angiosarcoma. Dig Dis Sci2003;48:677–682.

49. Rabe T, Feldmann K, Grunwald K, RunnebaumB. Liver tumours in women on oral contracep-tives. Lancet 1994;344:1568–1569.

50. Meissner K. Hemorrhage caused by ruptured livercell adenoma following long term oral contracep-tives: a case report. Hepatogastroenterology 1998;45:224–225.

51. Leese T, Farges O, Bismuth H. Liver cell adeno-mas: a 12-year surgical experience from a special-ist hepato-biliary unit. Ann Surg 1988;208:558–564.

52. Molina EG, Schiff ER. Benign solid lesions of theliver. In: Schiff ER, Sorrell MF, Maddrey WC,eds. Schiff’s diseases of the liver. 8th ed. Philadel-phia, Pa: Lippincott-Raven, 1999; 1245–1267.

53. Flowers BF, McBurney RP, Vero SR. Rupturedhepatic adenoma: a spectrum of presentation andtreatment. Am Surg 1990;56:380–384.

54. Bellaiche G, Habib E, Baledent F, et al. Hemo-peritoneum due to spontaneous splenic rupture: arare complication of primary cytomegalovirus in-fection. Gastroenterol Clin Biol 1998;22:107–108.

55. Khan AZ, Escofet X, Roberts KM, et al. Sponta-neous splenic rupture—a rare complication ofamyloidosis. Swiss Surg 2003;9:92–94.

56. Giagounidis AA, Burk M, Meckenstock G, KochAJ, Schneider W. Pathologic rupture of the spleenin hematologic malignancies: two additional cases.Ann Hematol 1996;73:297–302.

57. Hertzberg BS, Kliewer MA, Paulson EK. Ovariancyst rupture causing hemoperitoneum: imagingfeatures and potential for misdiagnosis. AbdomImaging 1999;24:304–308.

58. Hertzberg BS, Kliewer MA, Bowie JD. Adnexalring sign and hemoperitoneum caused by hemor-rhagic ovarian cyst: pitfall in the sonographic diag-nosis of ectopic pregnancy. AJR Am J Roentgenol1999;173:1301–1302.

59. Dibbs KI, Ball RH, Huetner PC. Spontaneousuterine rupture and hemoperitoneum in the firsttrimester. Am J Perinatol 1995;12:439–441.

60. Yoffe N, Bronshtein M, Brandes J, et al. Hemor-rhagic ovarian cyst detection by transvaginalsonography: the great imitator. Gynecol Endocri-nol 1991;5:123–129.

61. Hallatt JG, Steele CH, Snyder M. Ruptured cor-pus luteum with hemoperitoneum: a study of 173surgical cases. Am J Obstet Gynecol 1984;149:5–9.

62. Della-Giustina D, Denny M. Ectopic pregnancy.Emerg Med Clin North Am 2003;21:565–584.

63. Ankum WM, Mol BW, Van der Veen F, BossuytPM. Risk factors for ectopic pregnancy: a meta-analysis. Fertil Steril 1996;65:1093–1099.

64. Lozeau AM, Potter B. Diagnosis and managementof ectopic pregnancy. Am Fam Physician 2005;72:1707–1714.

65. Pritchard JA, Weisman R, Ratnoff OD, VosburghGJ. Intravascular hemolysis, thrombocytopenia,and other hematologic abnormalities associatedwith severe toxemia of pregnancy. N Engl J Med1954;250:89–98.

66. Weinstein L. Syndrome of hemolysis, elevatedliver enzymes, low platelet count: a severe conse-quence of hypertension in pregnancy. Am J ObstetGynecol 1982;142:159–165.

67. Byrd DE, Riely CA. Liver disease in preeclampsia.Gastroenterologist 1996;4:65–69.

68. Nunes JO, Turner MA, Fulcher AS. Abdominalimaging feature of HELLP syndrome: a 10-yearretrospective review. AJR Am J Roentgenol 2005;185:1205–1210.

69. Golledge J, Abrokwah J, Shenoy KN, Armour RH.Morphology of ruptured abdominal aortic aneu-rysms. Eur J Vasc Endovasc Surg 1999;18:96–104.

70. Borioni R, Garofalo M, Innocenti P, et al. Hemo-peritoneum due to spontaneous rupture of an an-eurysm of the left gastroepiploic artery. J Cardio-vasc Surg (Torino) 1999;40:63–64.

71. Stanley JC, Thompson NW, Fry WJ. Splanchnicartery aneurysms. In: Rutherford RB, ed. Vascularsurgery. 4th ed. Philadelphia, Pa: Saunders, 1995;1124–1139.

72. Shanley CJ, Shah NL, Messina LM. Commonsplanchnic artery aneurysms: splenic, hepatic andceliac. Ann Vasc Surg 1996;10:315–322.

73. Berceli SA. Hepatic and splenic artery aneurysms.Semin Vasc Surg 2005;18:196–201.

74. Gadacz TR, Trunkey D, Kieffer RF Jr. Visceralvessel erosion associated with pancreatitis. ArchSurg 1978;113:1438–1440.


This article meets the criteria for 1.0 credit hour in category 1 of the AMA Physician’s Recognition Award. To obtaincredit, see accompanying test at http://www.rsna.org/education/rg_cme.html.

RG Volume 27 • Volume 1 • January-February 2007 Lubner et al

Blood in the Belly: CT Findings of Hemoperitoneum Meghan Lubner, MD et al

Page 112 Blood usually has higher measured attenuation than other body fluids; however, its attenuation may vary, and its CT appearance often depends on the age, extent, and location of the hemorrhage. Page 113 On CT images, the highest-attenuation hematoma, or sentinel clot, is that closest to the site of bleeding, whereas lower-attenuation unclotted blood is located farther from the source. Page 113 Although in many cases hemoperitoneum can be managed nonsurgically with excellent results, a CT finding of active bleeding is indicative of a need for emergent embolization or surgical treatment. Page 115 In the trauma setting, CT can help accurately diagnose the type and extent of the primary injury, determine the presence and amount of hemoperitoneum, and detect associated injuries. Page 118 The most common cause of nontraumatic hepatic hemorrhage is a hypervascular neoplasm such as hepatocellular carcinoma or hepatocellular adenoma.

RadioGraphics 2007; 27:109–125 ● Published online 10.1148/rg.271065042 ● Content Codes:

blood in the belly: ct findings of hemo- peritoneum

Documents