| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Emergency Radiology |
1 From the Division of Emergency Medicine (J.R.R.) and Department of Radiology (E.L.O., M.A.G., J.P.M.), University of California, Davis Medical Center, 2315 Stockton Blvd, Sacramento, CA 95817; and School of Medicine, Georgetown University, Washington DC (M.V.R.). Received October 17, 2003; revision requested January 12, 2004; revision received February 20; accepted March 26. Address correspondence to J.R.R. (e-mail: jrrichards@ucdavis.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: A retrospective review of results of all consecutive emergency blunt trauma US examinations performed at a level I trauma center from January 1995 to June 2002 was conducted. Data on demographics, free fluid location, and patient outcome were collected. Injuries were determined on the basis of results of computed tomography and/or laparotomy. The Student t test was used to detect differences between continuous variables, and 
2 analysis was used to evaluate differences between proportions.
RESULTS: A total of 2319 US examinations for blunt trauma were performed in girls and women between the ages of 10 and 50 years. There were 328 pregnant patients, 23 of whom had intraabdominal injury. The mean age of the pregnant patients was 24.7 years ± 6.1 (standard deviation) (age range, 14–42 years). In pregnant patients, the sensitivity of US was 61% (14 of 23 patients), the specificity was 94.4% (288 of 305 patients), and the accuracy was 92.1% (302 of 328 patients). Pregnant patients were significantly more likely to have sustained injuries from assault (odds ratio: 2.6, P < .001). The most common pattern of free fluid accumulation detected at US in pregnant patients was that of fluid in the left and right upper quadrants and pelvis (n = 4, 29%); the second most common pattern was one of isolated pelvic fluid (n = 3, 21%).
CONCLUSION: For detection of intraabdominal injury, US was less sensitive in pregnant patients than in nonpregnant patients but was highly specific in both subgroups. The sensitivity of US was highest in pregnant patients during the first trimester.
© RSNA, 2004
Index terms: Abdomen, injuries, 70.41, 80.41 • Abdomen, US, 70.1298, 80.1298 • Emergency radiology • Pregnancy, US
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Computed tomography (CT) of the abdomen accurately depicts intraabdominal injury but is relatively contraindicated in pregnancy. CT requires the administration of iodinated contrast material and radiation exposure, which may be detrimental to the developing fetus (5). In most trauma centers, patients must be transported away from the resuscitation area to the CT suite, and this may be unsuitable for unstable patients with trauma.
Ultrasonography (US) is rapid and noninvasive compared with diagnostic peritoneal lavage and CT, obviates fetal exposure to radiation, and can be performed during resuscitation. US, in addition to its long history in obstetric applications, has been described as an accurate method for detection of hemoperitoneum in the form of free fluid and canalso depict parenchymal abnormalities of solid organs that are suggestive of injury (6–9). However, few studies to date have involved examination of the use of US in pregnant patients who have sustained blunt abdominal trauma, and none have involved comparing these patients with their nonpregnant counterparts. Thus, the purpose of our study was to determine the accuracy of US for the detection of blunt intraabdominal injury in pregnant patients and to compare differences between pregnant and nonpregnant patients of childbearing age.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Final US reports were retrospectively reviewed (by either J.R.R., E.L.O., or M.V.R.) for the following characteristics: the presence of free fluid, the location of free fluid, and the presence of solid organ parenchymal abnormalities. Patients with intraabdominal injuries were identified, and demographic data, mechanism of injury, and outcome, as reported in patient charts, were recorded. Hemodynamic parameters were not recorded. Surgical reports were reviewed and correlated with US and CT reports. Pregnancy was confirmed with US and with serum and/or urine pregnancy tests. Estimated gestational age was determined with serum quantitative ß human chorionic gonadotropin levels and/or obstetric US measurements. This study was approved by the institutional review board, and exemption from the requirement of informed consent was granted.
US Examinations and Report Interpretation
US examinations were performed by registered diagnostic medical sonographers with phased-array or convex 2.5–5.0-MHz transducers. Toward the beginning of the study, examinations were performed with an XP10–128 unit (Acuson, Mountain View, Calif); later in the course of the study, examinations were performed with a Sequoia GI unit (Acuson) or a 5200S unit (Acoustic Imaging, Phoenix, Ariz). Focused abdominal US for trauma was performed as follows: The right and left upper quadrants were scanned for the presence of free fluid. The parenchyma of both the liver and the spleen was also evaluated to detect irregularities suggestive of hematomas and/or lacerations. The epigastrium was scanned for evaluation of the left lobe of the liver and the pancreas. Both right and left flanks were scanned for detection of free fluid in the right and left paracolic gutters and evaluation of the retroperitoneum. Finally, the pelvis was evaluated for the presence of free fluid.
All US examinations were performed before CT, diagnostic peritoneal lavage, or laparotomy. US results were considered positive when any abnormality that could have resulted from trauma was detected: Free fluid was assumed to represent hemoperitoneum, and irregularities within the parenchyma of solid organs or subcapsular collections of fluid were assumed to represent lacerations or hematomas. US was performed as soon as possible after the patient arrived in the emergency department. Not all patients who had sustained blunt trauma underwent US during the entire study period, and selection of imaging modality was left to the discretion of the attending physician who was caring for the patient.
The initial interpretation of the US results was performed by the on-call faculty, fellow, or resident radiologist. All images were recorded on film hard copy or at a central US workstation (Kinetics, Siemens-Acuson, Mountain View, Calif) and were reviewed immediately. Final interpretation by faculty radiologists was performed soon thereafter, and the report of the final interpretation was used in this study. There were seven faculty radiologists, who had between 2 and 24 years of experience in US image interpretation. Information was obtained from these final interpretation reports and not at an actual image review. If multiple US examinations were performed, only the results of the initial US examination performed at presentation after the accident were used. US results were considered true-positive, false-positive, or false-negative in comparison to results of CT and/or laparotomy, which were used as the reference-standard examinations. Patients with true-negative US results were considered as such on the basis of results of clinical observation for abdominal injury or results of CT and/or laparotomy.
CT Examinations
Toward the beginning of this study, CT was performed with a 900SX scanner (Toshiba, Tustin, Calif); later, a CTi or a LightSpeed scanner (GE Medical Systems, Milwaukee, Wis) was used. All CT examinations were performed by using rapid injection of iohexol (Omnipaque 330; Amersham Health, Princeton, NJ) at an injection rate of 2 mL/sec. Images were reconstructed with a section thickness of 5 mm. Oral contrast material was not administered.
Statistical Analysis
Unless otherwise stated, data are reported as means ± standard deviations. Odds ratios and 95% confidence intervals are reported when appropriate. Comparisons were performed and their results were analyzed for pregnant and nonpregnant patients. Differences between continuous variables and proportions were examined by using the Student t test and the 2 test, respectively. P < .05 was considered to indicate a statistically significant difference. Statistical analyses were performed by using Stata 8 software (Stata, College Station, Tex).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Pregnant Patients
There were 328 pregnant patients who had sustained blunt trauma, and 23 (7.0%) of them had intraabdominal injury. The mean age of the pregnant patients was 24.7 years ± 6.1 (age range, 14–42 years). Eighty-one (24.7%) of the patients were in their first trimester of pregnancy, 134 (41.0%) were in their second trimester, and 113 (34.5%) were in their third trimester. Free fluid was detected with US in 29 pregnant patients, 15 of whom had false-positive results for abdominal injury. There were two patients with false-positive US results whose examinations did not reveal free fluid but did reveal suspicious parenchymal abnormalities of the liver (in one patient) or the spleen (in the other). The most common false-positive finding was isolated pelvic free fluid (n = 7). Fourteen patients had true-positive US results, and in one of these patients, US revealed a splenic abnormality without free fluid that was confirmed to be a laceration during exploratory laparotomy. With regard to gestational age and true-positive US findings, pregnant patients in the first (n = 9) and second (n = 1) trimesters all had true-positive results for free fluid in the pelvis, whereas those in the third trimester (n = 4) had no true-positive results for free fluid in the pelvis.
Nine patients had false-negative US results, and 288 patients had true-negative US results. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of focused abdominal US for trauma, as compared among pregnant patients according to age of gestation, are reported in Table 1. US had the highest sensitivity but the lowest specificity and accuracy for patients in the first trimester.
|
Twenty-seven pregnant patients underwent CT scanning, at which nine patients were diagnosed with intraabdominal injury, after US. The likelihood of CT scanning was significantly higher in pregnant patients with actual injury than in those without injury (nine [40%] of 23 patients with injury vs 18 [5.9%] of 305 patients without injury, odds ratio = 10.3, P < .001). Examples of intraabdominal injuries in pregnant patients that were detected at US are provided in Figures 1–3 . Figure 4 shows an example of false-positive US and CT results.
|
|
|
|
|
|
|
|
|
|
|
US in nonpregnant patients had a sensitivity of 71.2% (148 of 208 patients; 95% confidence interval: 64, 77), a specificity of 97.4% (1737 of 1783 patients; 95% confidence interval: 97, 98), a positive predictive value of 76.2% (148 of 194 patients; 95% confidence interval: 69, 82), a negative predictive value of 96.7% (1737 of 1797 patients; 95% confidence interval: 96, 97), and an accuracy of 94.6% (95% confidence interval: 94, 96). The most common intraabdominal injury in nonpregnant patients was liver injury (n = 99), followed by splenic injury (n = 96), bowel injury (n = 69), kidney injury (n = 28), adrenal gland injury (n = 3), and a ruptured bladder (n = 1). Sixty-two patients had multiple organ injuries.
There was no significant difference in mean age between the pregnant and the nonpregnant group. Differences in mechanisms of injury between pregnant and nonpregnant patients are shown in Table 2. Pregnant patients were significantly more likely to have been assaulted than nonpregnant patients (odds ratio: 2.6, P < .001). The rate of injury was slightly lower for pregnant than for nonpregnant patients (7.0% vs 10.4%, odds ratio: 0.6, P = .06).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
One explanation for the marked number of false-negative US results in our study and other studies is that focused abdominal US for trauma is performed relatively early in the resuscitation process, at a time when hemoperitoneum may not have accumulated to a detectable amount. In our study, a marked number of false-negative US results (19 of 69) were observed in patients with bowel and mesenteric injuries; this finding is consistent with findings of previous studies involving both US and CT (12).
Miller and colleagues (10) suggested that all hemodynamically stable patients who have sustained blunt abdominal trauma should undergo CT scanning rather than focused abdominal US for trauma to prevent the possible underdiagnosis of intraabdominal injury. This may be a difficult concept to apply in the assessment of pregnant patients because the risk of teratogenesis from CT radiation exposure, however minor, may be unacceptable. Further consideration should be given to the gestational age of the fetus, because radiation exposure in early pregnancy would have a more deleterious effect than radiation exposure in later pregnancy (13).
In one series at a large trauma center (14), only 7% of pregnant patients with trauma underwent CT scanning; this is comparable to the 8.2% of patients who underwent it in our study. Results of both animal and human studies have demonstrated no increase in gross congenital aberrations resulting from exposures of between 5 and 10 rad (50 and 100 mGy) (15). It appears that the decision to perform a CT examination is arbitrary and may depend on the presence of equivocal objective findings that might result in an unnecessary laparotomy.
Although not as sensitive as CT, focused abdominal US for trauma has a distinct advantage over CT in the rapid triage of unstable patients with blunt trauma who cannot safely travel to the CT suite. Detection of free fluid and/or parenchymal abnormality in this setting results in safer and faster disposition to the operating room than could be accomplished with either CT or diagnostic peritoneal lavage. If no free fluid is detected, the patient may be transferred to the labor and delivery area for fetal monitoring and potential delivery. With this method, the possibility of teratogenesis from the ionizing radiation at CT and the risk of an allergic reaction to intravenous CT contrast material are avoided. In the future, US performed with contrast material may be a viable alternative to CT. In one small, blinded study with 15 patients with abdominal trauma that involved comparing CT with power Doppler US, there were no false-positive or false-negative examination results in the US arm (16).
In the present study, the most common pattern of free fluid accumulation in pregnant patients with fetuses of all gestational ages was a pattern of accumulation in the left and right upper quadrants and the pelvis. However, US depicted pelvic free fluid in no pregnant patients in the third trimester, and the sensitivity of focused abdominal US for trauma was highest for patients who were in the first trimester of pregnancy. One possible explanation for this may be that the compression of intraabdominal structures, specifically the paracolic gutters, by the expanding uterus may make it more difficult to detect free fluid in the paracolic gutters and pelvis (17).
Isolated pelvic free fluid was the second most common true-positive fluid accumulation pattern observed in our study. The ability to distinguish between physiologic free fluid and free fluid resulting from injury has been addressed in a previous study: Sirlin and colleagues (18) reported isolated free fluid in the cul-de-sac in 56 patients, and only two had injuries, but they made no distinction between pregnant and nonpregnant patients. Their conclusion was that isolated free fluid in the pelvis was likely to be physiologic and not due to injury. In our study, 54% of the false-positive US examinations in pregnant and nonpregnant patients combined revealed isolated free fluid in the pelvis.
Also of interest in the present study are the three patients with undiagnosed ectopic pregnancies that ruptured after their traumatic events. The incidence of ectopic pregnancy has steadily increased over the past 3 decades, and these patients are seen more frequently in the acute care setting (19). Transvaginal US has been the imaging study of choice for detecting ectopic pregnancy, but less is known regarding transabdominal US for this indication (20). US assessment of free fluid in the three patients in our study revealed isolated pelvic free fluid in two patients and free fluid in the Morison pouch and the pelvis in one patient. On the basis of our findings, isolated free fluid in the pelvis cannot necessarily be discounted as being physiologic in the pregnant patient.
Sirlin et al (21) analyzed the results of 2693 US examinations for blunt abdominal trauma and determined that free fluid present in the left upper quadrant, in both upper quadrants, or diffusely was significantly associated with splenic injuries. In our study, all pregnant patients with splenic injuries had free fluid in the left upper quadrant. The dynamics of flow in the abdomen are of interest in that free fluid tended to flow from the left to the right upper quadrant rather than down the left paracolic gutter into the pelvis. One explanation for this may be that hemorrhage from the spleen first accumulates in the left and then progresses to the right upper quadrant because the phrenocolic ligament acts as a relative barrier to the movement of fluid to the left gutter (22). It also appears that fluid from the right upper quadrant flowed down the right paracolic gutter rather than toward the left upper quadrant, perhaps because of the gravity dependence of the right paracolic gutter and pelvis.
It has been suggested in the past that the fetus is well protected against injury from blunt trauma because it is encased in a fluid-filled structure (1). In our study, intraabdominal injuries in pregnant patients were most commonly to the spleen (n = 7) or placenta (n = 7), necessitating precipitous delivery or resulting in fetal demise. Thus, the shear forces present in even low-force injuries such as falls cannot be discounted, and we recommend that US, as well as fetal monitoring for patients whose fetuses are past 20 weeks gestation, routinely be used in pregnant patients with trauma. Results of one study indicated that continuous fetal monitoring is more sensitive but is less specific than US for the detection of placental abruption (23).
There were several limitations in this study. First, it was a retrospective review, and several faculty radiologists were involved in the interpretation of the US results. This may have resulted in variable consistency with regard to the detection of subtle findings such as parenchymal lesions. The focused abdominal US for trauma technique itself is limited to the detection of small amounts of pelvic free fluid present in nonpathologic states in the female patient. A large portion of the patient subgroup with true-negative findings in this study was followed up clinically after initial US, and no further imaging was performed. Smaller, although not life-threatening, injuries may have been missed as a result. Although we did not analyze hemodynamic parameters in the study population, it would be safe to assume that any patients in the true-negative subgroup would have been triaged to additional imaging examinations and/or procedures if they developed hemodynamic instability during their hospital stay. Last, not all consecutive patients with blunt trauma underwent US, and this also may have affected the results.
In conclusion, US is less sensitive for detection of intraabdominal injury in pregnant patients than in nonpregnant patients but is highly specific in both subgroups. Patterns of free fluid accumulation differ between pregnant and nonpregnant patients, with a pattern of free fluid in the right and left upper quadrants and pelvis being the most common pattern of accumulation in pregnant patients. Isolated pelvic free fluid does not appear to have a greater association with intraabdominal injury in the pregnant patient than in the nonpregnant patient. The sensitivity of US is highest in pregnant patients who are in the first trimester.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, all authors; study concepts and design, all authors; literature research, J.R.R., E.L.O., M.V.R., J.P.M.; clinical studies, J.R.R., E.L.O., M.V.R., J.P.M.; data acquisition, J.R.R., E.L.O., M.V.R.; data analysis/interpretation, all authors; statistical analysis, J.R.R., E.L.O.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. R Cunningham FAST Scan: Ultrasound's Role in Trauma Radiol. Technol., May 1, 2008; 79(5): 455 - 458. [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Patel, D. L. Reede, D. S. Katz, R. Subramaniam, and J. K. Amorosa Imaging the Pregnant Patient for Nonobstetric Conditions: Algorithms and Radiation Dose Considerations RadioGraphics, November 1, 2007; 27(6): 1705 - 1722. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. G. ODLE Blunt Pelvic Trauma Radiol. Technol., January 1, 2006; 77(3): 200 - 219. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
