The nonoperative management of blunt adult hepatic injuries, when first seriously proposed in the late 1980s, and early 1990s [1-4] was initially met with cautious restraint. Early successes with nonoperative management of blunt adult hepatic injuries,  notwithstanding, prevailing authorities in the field of trauma surgery continued to advocate the time-tested approach of early operative intervention to keep blood transfusions to a minimum and to avoid the possibility of overlooking associated intraabdominal injuries. [5,6] Despite these caveats, nonoperative management of adult blunt hepatic injuries continued to be evaluated. As of 1988, experiences with this treatment modality at 14 different level I trauma centers resulted in a mean success rate of 94%. [1-4,7-16] The number of patients managed nonoperatively varied in each series, from a low of 16 to a high of 72, as reported by Meredith et al.  A multicenter study was therefore retrospectively undertaken to accrue several hundred patients to assess whether the combined experiences at level I trauma centers could: (1) validate the currently reported high success rate, low morbidity, and virtually nonexistent mortality associated with this approach; (2) possibly assess how often patients with blunt hepatic injuries can be managed nonoperatively; (3) identify a patient population who, although having met the criteria for nonoperative management, are at greater risk for complications during the observational period; and (4) define which, if any, patient group requires closer observation in critical care areas.
MATERIALS AND METHODS
A survey of all adult blunt hepatic injuries during a 5-year period ending July 30, 1994 was conducted by 12 members of the Western Trauma Association representing 13 level I trauma centers: New York University/Bellevue Hospital Medical Center, New York, N.Y.; San Francisco General/University of California-San Francisco, San Francisco, Calif.; Huntington Memorial Hospital, Pasadena, Calif.; Cooper Hospital, Camden, N.J.; University of California-San Diego, San Diego, Calif.; Gundersen/Lutheran Medical Center, La Crosse, Wis.; Mercy Hospital, Pittsburgh, Pa.; King's County Hospital, Brooklyn, N.Y.; HCA Wesley, Wichita, Kan.; University of Vermont, Burlington, Vt.; Washington Hospital Center, Washington, D.C.; Memorial Medical Center, Savannah, Ga.; and Pennsylvania State University, Hershey, Pa. Initial hemodynamic stability or hemodynamic stability achieved with moderate fluid infusion qualified patients to be considered for inclusion in the nonoperative management group. All injuries were then stratified by grade of hepatic injury, as outlined in the hepatic organ injury scale devised by the American Association for the Surgery of Trauma (AAST)  (Table 1). Four hundred and four patients were identified and form the basis of this study. Patients were initially identified from trauma registries, but in the last 2 years of the study were prospectively entered into a specifically designed study protocol. Other organ system injuries were classified using AIS-90 to generate the Injury Severity Score (ISS). 
Criteria for Patient Inclusion
Although the criteria for patient inclusion was not previously agreed on by the participating members of the study group, all institutions adhered to guidelines extensively outlined in the literature. [3,4,10,13-15] The most crucial criteria for patient inclusion into the nonoperative management study group was the presence of hemodynamic stability. Once hemodynamic stability was ascertained, other inclusion criteria consisted of the (1) absence of peritoneal signs when the patient was neurologically intact and not under the influence of drugs or alcohol, (2) computerized axial tomography (CAT) scan (usually performed with a GE 9800 scanner and most recently with a helical scanner) delineation of the injury and AAST assessment of grade of injury, (3) absence of associated intraperitoneal and retroperitoneal injuries on CAT scan mandating urgent operative intervention, and (4) limit the need for excessive hepatic-related blood transfusions.
During a 5-year period ending in July 1994, 404 adult patients sustaining blunt hepatic injuries and managed nonoperatively were accrued from 13 participating level I trauma centers. Patient ages ranged from 18 to 88 years (mean, 29.8). There were 246 (61%) male and 158 (39%) female patients. Motor vehicle crashes accounted for the majority of patients (72%) injured in this study.  (Table 2). Snowmobiling, moped, motorcycle and horseback riding, karate, as well as boating, were some of the other responsible mechanisms of injury reflecting the geographical distribution of the participating centers. The mean ISS for the entire group was 20.2 (range, 4 to 75). This score was virtually identical to the mean ISS of 21.5 for the 21 patients experiencing postobservation complications, but was markedly different from the mean ISS of 40.3 in the 27 patients who died.
AAST-CAT Scan Injury Grading
The AAST hepatic organ injury scale was the sole criteria for assessing the grade of injury at each institution. The CAT scan grading was as follows: grade I, 19% (n = 76); grade II, 31% (n = 124); grade III, 36% (n = 146); grade IV, 10% (n = 42); and grade V, 4% (n = 16).
CAT Scan Follow-Up
The CAT scan documentation of stabilization or improvement of the injury before patient discharge, although highly desirable, was inconsistently used. Many of the patients had more than one follow-up CAT scan after admission; however, only the last documented scan was considered for analysis (one follow-up scan per patient). Some of the authors felt that, when injuries were not extensive (grades I to III), follow-up CAT scan studies were superfluous. As a result, only 198 (49%) of the 404 patients managed nonoperatively underwent follow-up CAT scanning after the initial study performed on admission. Repeat CAT scanning (final CAT scan performed) varied from 1 to 238 days. Of the 198 patients scanned, 52 (26%) underwent repeat CAT scanning anywhere from 1 to 6 days from injury. Of these, only five patients (9.6%) showed any degree of resolution of injury. Eighty-nine patients (45%) underwent subsequent CAT scanning between postobservation days 7 through 10. Of these 89 patients, 69 (78%) were noted to have had improvement of the original injury. In the remaining 20 patients, CAT scan analysis could not detect any worsening of the original injury because all were considered unchanged. Forty-six patients (23%) underwent CAT scanning between postinjury days 11 and 89. The injuries were noted to be completely resolved in 93% (n = 43). Eleven patients (6%) were scanned at 90 or more days after injury. Of these 11 patients, 6 (45%) were noted on CAT scan to have small "hypodense" areas in the vicinity of the original injury.
Length of Stay
The mean length of stay for the entire group was 13.1 days (range, 0 to 162 days). The mean length of stay, however, increased to 17.7 and 26.9 days for the 27 patients who died and for the 21 patients who developed complications in the observation period, respectively. Six patients (1%) with isolated grade I hepatic injuries were discharged directly from the emergency department. An additional 186 patients were discharged within 1 week of admission (46%), and another 117 patients (29%) were discharged within 2 weeks of admission. Ninety-five (24%) remained in the hospital anywhere from 15 to 162 days. Aside from patients who expired or who developed complications, patients remaining in the hospital in excess of 21 days required treatment for associated injuries, most often in rehabilitation units.
Of the 404 patients managed nonoperatively, 361 (89%) did not receive blood, whereas 43 (11%) patients were transfused. A total of 169 U of blood were transfused (51 U to one patient), with rates that varied from 3 to 51 U (these figures were inclusive of blood given during the observation period, intraoperatively, and postoperatively). The mean transfusion rate for the entire group was less than 1/2 U (0.42 U) per patient. For the 43 patients requiring transfusion, however, the mean unit transfusion rate increased to 3.9 U. The subset of patients transfused invariably sustained multiple associated injuries, most often thoracic, orthopedic, pelvic, peripheral vascular, and neurologic.
Three of the 43 patients who were transfused merit further comment. One patient who received a 3 U transfusion was a known cirrhotic with abnormal liver function tests, as well as an elevated prothrombin and partial thromboplastin time on admission. An additional patient who was transfused 4 U was a hemophiliac and human immunodeficiency virus-positive. The last patient who received 51 U of blood sustained a grade III hepatic injury, entered with an ISS of 29. Concomitant injuries consisted of a pelvic fracture, cardiac contusion, multiple rib fractures, bilateral pneumohemothoraces, splenic laceration, and a cerebral contusion. The patient was in an intensive care setting and was hemodynamically stable for 22 days. On the 22nd observation day, he received 2 U of blood. Two days later, he was transfused an additional 2 U. On the following day he became hemodynamically unstable and was urgently taken to the operating room, wherein both bleeding splenic and hepatic injuries were noted. Splenectomy and attempts at controlling hemorrhage from the liver laceration necessitated an additional 47-U transfusion. The patient died 3 days later from multiple organ failure.
There were 27 deaths (7%) in the series. Head injuries accounted for 59% (n = 16) of the deaths. In the remaining 11 patients, responsible mechanisms resulting in patient death in decreasing order of frequency were as follows: sepsis, 11% (n = 3); pelvic fracture, 7% (n = 2); pulmonary failure, 7% (n = 2); aortic transection, 3.5% (n = 1); and cardiac injury, 3.5% (n = 1). Only two deaths (7%) could be attributed the hepatic injury itself, representing a 0.5% cumulative mortality rate for the entire group. An analysis of these two hepaticrelated patient deaths revealed that one was a cirrhotic and actively bleeding from his liver injury. The family refused surgical intervention. The second patient (detailed in the transfusion section) became hypotensive on observation day 25 and died as a result of profuse bleeding from splenic and hepatic injuries that could not be controlled.
During the observation period, a 5% hepatic-related complication rate was documented (n = 21). The most common complication noted was hemorrhage, occurring in 14 patients (3.5%). Surgical intervention to control hemorrhage was necessary in only 3 (0.7%) of these 14 patients. Six patients were treated by transfusion alone (0.5 to 5 U), 4 underwent angioembolization, and 1 was further observed.
Other complications included two "bilomas" and three perihepatic abscesses, all of which were drained percutaneously. One of the patients with a large perihepatic abscess required surgical intervention after multiple attempts at percutaneous drainage failed. Two small intestinal injuries (one ileal and one duodenal) were initially missed 0.5%, but were detected 2 and 3 days after admission, respectively. Overall, six patients (1.4%) required operative intervention: three for bleeding, two for missed enteric injuries, and one for persistent sepsis after unsuccessful percutaneous drainage (Table 3). No delayed complications were noted after initial discharge.
Pringle,  in his 1908 classic paper on liver injuries, has inadvertently misled subsequent generations of surgeons into believing that significant lacerations of the liver would continue to bleed because the structural integrity of the liver was incapable of achieving spontaneous hemostasis. As late as 1977,  operative intervention was considered the only acceptable treatment of blunt hepatic injuries. Reports in the early 1970s detailing the successful nonoperative management of patients sustaining blunt hepatic injuries were for the most part ignored. [21,22] The technological breakthroughs in radiologic imaging techniques, specifically in the field of CAT scans, were principally responsible for the subsequent change in surgical philosophy. Advanced high-speed CAT scanners were able to not only delineate the severity of hepatic injury and quantitate the degree of hemoperitoneum, but could also provide a reliable evaluation of the gastrointestinal tract and retroperitoneal structures. The additional realization that up to 67% of all operations performed for blunt traumatic abdominal injuries have been nontherapeutic, [7,8,22-25] as well as a substantial body of literature confirming that as many as 86% of all hepatic injuries have already stopped bleeding at the time of surgery, [26-30] has led to a reassessment as to how stable blunt injuries to the adult liver should best be managed.
Since 1988, 495 adult patients with blunt hepatic injuries managed nonoperatively have been accrued from 14 institutions which managed 16 or more patients in this fashion.  The success rate achieved with this form of treatment in these 495 patients was a remarkable 94%, with a mean transfusion rate of 1.9 U and a mean hospital stay of 13 days for the entire group. In addition, reported complications were minimal, as evidenced by an abscess rate of 0.6%, a "biloma"-bile fistula rate of 2.8%, and a hemorrhage rate of 2.8%. Most importantly, however, no hepatic-related mortalities, or missed enteric injuries were documented. These combined studies strongly suggested that nonoperative management of stable blunt adult hepatic injuries should be considered the treatment modality of choice. The results of our multicenter study confirmed the aforementioned cited observations, in that the nonoperative management of blunt adult hepatic injuries is the preferred treatment choice in hemodynamically stable patients. Significant differences, however, do exist, despite similarities in success rates reported: 94% in the combined previous collected series and the 98.5% who successfully avoided surgical intervention in our series. In contradistinction to the previously cited data obtained from the analysis of the 495 patients gathered from the literature, we did experience two missed gastrointestinal injuries (0.5%); additionally, there were two (0.5%) hepatic-related deaths. Two of the principle investigators (HLP and SR) have, since the completion of this study, documented two additional instances wherein enteric injuries had not been appreciated on either the initial physical examination or on CAT scan analysis. In a recent prospective series, Croce et al.  noted that, of the 12 failures in a group of 112 adult patients with blunt hepatic injuries managed nonoperatively, 33% (n = 4) were caused by missing intra-abdominal injuries (two pancreatic, one renal, and one duodenal).
An increasing number of studies detailing missed intra-abdominal and retroperitoneal injuries will most likely be forthcoming, because more patients with blunt hepatic injuries are managed nonoperatively. At present, the incidence of these missed injuries varies from 0.5% reported in our present series and would have increased from 1% had the two recent patients been included, to the 3.5% reported by Croce et al.  Moreover, with prompt surgical intervention, none of the patients reported seemed to suffer from the delay in recognition of the missed injury. Despite an accuracy rate of 97 to 99% in detecting associated intra-abdominal injuries, the CAT scan is not infallible, and can by no means serve as a substitute for the constant re-evaluation of the patient, most effectively achieved by the same set of examiners.
Relationship of CAT Scan Grade Injury in Determining the Need for Surgical Intervention
In its early stages, the nonoperative management of blunt adult hepatic injuries was relegated to AAST grade I to III injuries. [10,11] Recently published data have documented that a select number of patients with the most complex hepatic injuries (grades IV and V) who were hemodynamically stable could be managed nonoperatively [9,13-15,32] (Figure 1,Figure 2,Figure 3,Figure 4 and Figure 5). Fifty-eight (14%) of the 404 patients managed nonoperatively in our study were considered to have grade IV and V injuries. The rate with which these complex injuries can be managed nonoperatively is presently unknown, but reports varying from 21 to 38% have recently been reported by Meredith et al.  and Croce et al.,  respectively. It is worth noting, however, that 66.6% of the patients in this study who required operative intervention for failure of nonoperative management were patients with either grade IV or V injury. This subgroup of patients who constitute the most severe hepatic injuries are initially at greater risk for treatment failure and therefore should be more closely monitored in a critical care unit.
It would seem that hemodynamic stability rather than CAT scan grade of hepatic injury determines whether a patient sustaining a blunt hepatic injury can be managed nonoperatively.
CAT Scan Findings Indicating the Need for Additional Adjunctive Therapy or Operative Intervention
The presence of "pooling of contrast material" within the hepatic parenchyma indicates free extravasation of blood as a result of active bleeding (Figure 6). Irrespective of CAT scan grade of injury or hemodynamic stability, prompt preparations should be made for operative intervention, because fluxes in hemodynamic stability can occur suddenly and without warning. With the recognition that active bleeding is present, and an operating room on standby secured, the preferred approach would be, if hemodynamically stability is maintained, to bring the patient to the angiography suite where angioembolization of the lesion is attempted. In the hands of an experienced angiographer, successful embolization of the offending vessel can usually be accomplished. The net effect of this maneuver, with its resultant cessation of hemorrhage, allows for the continued nonoperative management of the patient. Persistent and prolonged attempts at controlling the bleeding vessel through angiographic means, however, should be discouraged. If angioembolization fails to control bleeding from the lacerated vessel in a reasonable amount of time, prompt surgical intervention with intrahepatic hemostasis should be undertaken expeditiously and without undue delay.
The natural history of patients who present with pooling of contrast material on CAT scan in the face of hemodynamic stability is unknown. One patient in this series, who presented with pooling of contrast material on the CAT scan, spontaneously stopped bleeding, but went on to form an intrahepatic abscess that required operative intervention despite numerous attempts at percutaneous drainage (Figure 7).
Timing of Subsequent CAT Scanning
The timing of when follow-up CAT scanning should optimally be performed in patients with blunt hepatic injuries managed nonoperatively is controversial. Suggestions have varied from 48 to 72 hours, then at 3 to 6 weeks from injury,  to a series of scans performed at 24 to 48 hours, then at 5 to 7 days, and finally at 1 month,  to the total abandonment of routine follow-up imaging.  An analysis of our data revealed that only 49% (n = 198) of our patients underwent follow-up CAT scanning. Data on the 51% who did not have follow-up CAT scanning are lacking, reflecting the constraints of any multicenter retrospective study. One can only speculate that the responsible surgeons felt that repeat CAT scanning in a stable patient added little to the clinical management. Of those patients who did undergo follow-up CAT scanning, 52 (26%) were studied between observation days 1 and 6. In only 5 (9.6%) of these patients, however, was there any significant degree of change noted when compared with the original scan. When subsequent CAT scanning was performed between 7 and 10 days after injury (n = 89), 78% of scans showed improvement, whereas in the remaining 22% no worsening of the injury was noted. Of the 57 CAT scans performed beyond postinjury day 10 (11 to 238), resolution of the injury was virtually complete at 3 months (93%, totally resolved; 7%, resolving). Beyond 3 months, the only radiographic abnormalities noted were small hypodense lesions, considered as nonconsequential. Our study confirms the hypothesis proposed by Karp et al.  that the physiology of hepatic repair after blunt injury progresses in a predictable fashion that results in virtually complete restoration of hepatic integrity at the end of 3 months. In addition, as little significant changes are noted on CAT scanning performed in less than 1 week from injury, coupled with virtual total resolution noted at 3 months, there seems to be little benefit derived from repeat CAT scanning performed in these time periods. Documentation on CAT scan that the injury is improving or at least has not gotten worse, is still useful. Analysis of the scan aids in the decision to move patients with grade IV and V injuries from critical care areas to the floor, helps in determining discharge date and provides an anatomic guideline for the resumption of normal or athletic activities. The optimal time frame for follow-up CAT scan, based on the data obtained in this study, seems to be somewhere between 7 and 10 days from the original injury.
The precise number of adult patients sustaining blunt hepatic injuries who can be managed nonoperatively is still evolving. Using generally accepted criteria for patient inclusion in the nonoperative management of blunt adult hepatic injuries, almost 1,000 cases (including this study) are available for analysis. [31,32] The majority of patients (> 75%) treated in this manner have been reported during the last 2 years. At the inception of this form of treatment, it was estimated that at best only 20% of adult patients with blunt hepatic injuries could be managed nonoperatively. [5,9,10] Recent studies by Sherman et al.,  Meredith et al.,  and Croce et al.,  however, document that nonoperative management of blunt hepatic injuries constituted 50, 57, and 82%, respectively, of all blunt hepatic injuries seen at their institutions. Of 13 level I trauma centers participating in this study, nonoperative management of blunt adult hepatic injuries varied from a low of 20% to a high of 70%, with a mean nonoperative management rate for the entire group of 47%. The accrual of successful experiences with nonoperative management of adult hepatic injuries has no doubt been responsible for a greater willingness among surgeons to embark on this course of therapy, rendering it more applicable to a larger number of patients. The potential for this form of treatment, initially grossly underestimated, would seem to be applicable to at least 50% of all patients with blunt hepatic injuries and could approach a high of nearly 80%. [13,32]
Two major objections to routine nonoperative management of adult blunt hepatic injuries, the unnecessary use of excessive blood transfusions and prolonged length of stay, have never materialized. The previously cited study encompassing 14 institutions and 495 patients documented a mean transfusion rate of 1.9 U.  Likewise, a remarkably identical mean transfusion rate of 1.9 U was reported by Croce et al.  in the 112 patients they managed nonoperatively. Of the 404 patients in this study, 89% (n = 361) received no blood. A total of 169 U were used in the remaining 43 patients, with nearly one-third (51 U) going to one patient. The mean transfusion rate for those patients receiving blood was 3.9 U, but when the entire group was considered the mean transfusion rate decreased to 0.42 U per patient.
The results of our multicenter study not only corroborate the data from previous studies,  but firmly establish that hemodynamically stable adult patients with blunt hepatic injuries can be managed nonoperatively without resorting to excessive transfusions. In the patient with associated injuries, a falling hematocrit is not uncommon, but can often present the surgeon with a difficult dilemma as to whether the liver lesion is the cause of the dropping hematocrit. Under these circumstances, a repeat CAT should be undertaken immediately. If the hepatic lesion is unchanged, one can rest assured that the need for transfusion is not hepatic related. On the other hand, if the hepatic lesion has become more extensive, especially in light of an increase in perihepatic or intraperitoneal blood, then additional treatment is required rather than relying on transfusions alone. If hemodynamically stable, the patient should undergo angiography for the express purpose of embolizing the bleeding vessel. It should also be stressed that hemodynamic instability at any given time is indicative of significant ongoing hepatic hemorrhage and mandates prompt operative intervention to arrest the bleeding.
Similarly, the length of stay for patients treated nonoperatively did not prove to be excessive. Of the 404 patients managed nonoperatively, 47% (n = 92) were discharged within 1 week of admission, and an additional 29% (n = 117) were discharged within 2 weeks. The mean number of days spent in the hospital was 13.1, varying from 0 to 162 days. The mean length of stay of 13.1 days reported in this study is remarkably similar to the mean hospitalization of 13 days noted in the 495 patients with blunt hepatic injuries who were managed nonoperatively at 14 different institutions.  In instances wherein comparable groups of patients underwent operative intervention, three authors recently noted that intensive care unit stay and mean length of hospitalization for these patients were either comparable or exceeded that of patients managed nonoperatively. [13-15]
For the nonoperative management of adult blunt hepatic injuries to be accepted on a routine basis, the complication rate reported with its use must be kept to a minimum. Foremost among potential complications associated with the nonoperative management of blunt hepatic injuries is the valid concern of sudden hepatic related hemorrhage during the observation period, as well as the possibility of delayed hemorrhage after patient discharge. The validity of this concern has only been heightened by a study from the pediatric literature documenting an instance of delayed hemorrhage with subsequent death in a patient with a hepatic injury who was managed nonoperatively.  A recent retrospective survey of the literature by Gates,  documented 13 hemorrhagic complications in patients with blunt hepatic injuries managed nonoperatively.  Ten of the 13 patients (77%), however, did not require surgical intervention. Of the 21 documented complications in our study, hemorrhage was the most common, accounting for 66% (n = 14) of all complications noted. The overall bleeding rate for the 404 patients, however, was 3.5%. Six of the 14 patients who remained hemodynamically stable were treated by transfusions alone (0.5 to 5 U). Four other patients were managed by angioembolization, whereas one patient was further observed without either transfusion or adjunctive therapy.
The exact prevalence of bleeding in patients with blunt hepatic injuries managed nonoperatively is presently unknown. However, based on a 3.5% incidence in the 404 patients examined in our study and the 2.8% incidence documented in the 495 patients accrued from 14 different institutions,  it would seem that the overall occurrence rate of bleeding is relatively low. More importantly, of the 13 patients who underwent surgical exploration for hemorrhagic complications in the collected series of 495 patients, operative intervention was deemed nontherapeutic in 46% (n = 6).  Thus, only 1.7% (n = 7) of the 495 patients actually required operative intervention for ongoing hemorrhage. Similarly, only 3 (0.7%) of the 14 patients in our series who experienced hemorrhage during the observational period required operative intervention. Previous studies regarding the nonoperative management of adult blunt hepatic injuries have failed to document instances of hemorrhagic-related death. Two (0.5%) of the patients in our series, however, died as a result of hemorrhagic hepatic related deaths. The first patient was a cirrhotic whose family refused further treatment. The second patient was the only true failure of nonoperative management who experienced delayed bleeding 3 weeks after the initial injury that eventually led to his demise. In this instance, 51 U of blood transfusion was required to deal with the combination of ongoing hemorrhage from both the liver and spleen.
The second most common complication encountered with the nonoperative management of adult blunt hepatic injuries is either the development of a collection of bile (biloma) or the formation of a biliary fistula. These complications were noted to occur in 14 patients (2.8%) in the series of 495 patients accrued from the literature.  Two (0.5%) significant bile collections were documented in our series, both of which were successfully treated by percutaneous drainage. No instances of biliary fistulas were documented. Similarly, the three patients who developed intrahepatic abscesses were initially managed by percutaneous catheter drainage. This approach was successful in two of the three patients. The third patient required operative intervention after multiple attempts at percutaneous catheter drainage failed to control ongoing sepsis.
Nonoperative management of blunt adult hepatic injuries is clearly the treatment modality of choice in hemodynamically stable patients, irrespective of grade of injury. Current data would suggest that 50 to 80% of all adult patients with blunt hepatic injuries are candidates for this form of therapy. [13,15,16] Of 260 adult patients managed in this fashion since 1994, an extraordinarily high mean success rate of 94% was achieved. [13-15,32] That 98.5% of the 404 patients in this cooperative study were able to avoid surgical intervention merely validates the previously cited data. [15,16,32] Moreover, the inordinately high success rate achieved in this study was accompanied by an exceedingly low overall morbidity of 5% and an even lower mortality of 0.5%. In addition, nonoperative management of blunt hepatic injuries has not resulted in prolonged length of stay, unnecessary blood transfusions, or significant instances of either hemorrhagic complications or missed associated intra-abdominal injuries.
1. Farnell MB, Spencer MP, Thompson E, et al: Nonoperative management of blunt hepatic trauma in adults. Surgery 104:748, 1988
2. Mirvis SE, Whitely NO, Vainright JR, et al: Blunt hepatic trauma in adults: CT based classification and correlation with prognosis and treatment. Radiology 171:27, 1989
3. Hiatt JR, Harrier HD, Koenig BV, et al: Nonoperative management of major blunt liver trauma with hemoperitoneum. Arch Surg 125:101, 1990
4. Federico JA, Horner WR, Clark DE, et al: Blunt hepatic trauma: Nonoperative management in adults. Arch Surg 120:103, 1990
5. Walt AJ: Nonoperative management of major blunt injury with hemoperitoneum: Invited commentary. Arch Surg 125:909, 1990
6. Trunkey D: Blunt hepatic trauma: Nonoperative management in adults: Invited commentary. Arch Surg 125:909, 1990
7. Hammond JC, Canal DF, Broadie TA: Nonoperative management of adult blunt hepatic trauma in a municipal trauma center. Am Surg 58:551, 1992
8. Hollands MJ, Little JM: Nonoperative management of blunt liver injuries. Br J Surg 78:968, 1992
9. Knudson MM, Lim RC, Oakes DD, et al: Nonoperative management of blunt liver injuries: The need for continued surveillance. J Trauma 30:1454, 1990
10. Pachter HL, Spencer FC, Hofstetter SR, et al: Significant trends in the treatment of hepatic trauma: Experience with 411 injuries. Ann Surg 215:492, 1992
11. Durham RM, Buckley J, Keegan M, et al: Management of blunt hepatic injuries. Am J Surg 164:477, 1992
12. Bynoe RP, Bell RM, Miles WS, et al: Complications of nonoperative management of blunt hepatic injuries. J Trauma 32:308, 1992
13. Boone DC, Federle MP, Billiar TR, et al: Evolution of nonoperative of major hepatic trauma: Identification of patterns of injury. Presented at the Eastern Association for the Surgery of Trauma. Freeport, Bahamas Jan. 14, 1994
14. Sherman HF, Savage BA, Jones MA, et al: Nonoperative management of blunt hepatic injuries: Safe at any grade? J Trauma 37:616, 1994
15. Meredith JW, Young JS, Bowling J, et al: Nonoperative management of blunt hepatic trauma: The exception or the rule. J Trauma 36:529, 1994
16. Sugimoto K, Asari Y, Sakaguchi T, et al: Endoscopic retrograde cholangiography in the nonsurgical management of blunt liver injury. J Trauma 35:192, 1993
17. Moore EE, Shackford SR, Pachter HL, et al: Organ injury scaling: Spleen, liver and kidney. J Trauma 29:1664, 1989
18. Association for the Advancement of Automotive Medicine: The Abbreviated Injury Scale--1990 Revision. Des Plaines, Ill., AAAM, 1990
19. Pringle JH: Notes on the arrest of hepatic hemorrhage due to trauma. Ann Surg 48:541, 1908
20. Flint L, Mays ET, Aaron WS, et al: Selectivity in the management of hepatic trauma. Ann Surg 185:613, 1977
21. Richie JP, Fonkalsrud EW: Subcapsular hematoma of the liver: Nonoperative management. Arch Surg 104:781, 1972
22. Popovsky J, Wiener SN, Felder PA, et al: Liver trauma: Conservative management and the liver scan. Arch Surg 108:184, 1974
23. Karp MP, Cooney DR, Pros GA, et al: The nonoperative management of pediatric hepatic trauma. J Pediatr Surg 18:512, 1983
24. Carmona RH, Lim RC Jr, Clark GC: Morbidity and mortality in hepatic trauma: A 5 year study. Am J Surg 144:88, 1982
25. Jones S, Walsh JW, Maull KI: Diagnostic imaging in blunt trauma of the abdomen. Surg Gynecol Obstet 157:389, 1983
26. Cywes BS, Rode H, Millar AJW: Blunt liver trauma in children: Nonoperative management. J Pediatr Surg 20:14, 1985
27. Scalfani SJA, Shaftan GW, McAuley J, et al: Interventional radiology in the management of hepatic trauma. J Trauma 24:256, 1984
28. Moore EE: Critical decisions in the management of hepatic trauma. Am J Surg 145:712, 1984
29. Anderson R, Alwmark A, Gullstrand P, et al: Nonoperative treatment of blunt trauma to the liver and spleen. Acta Chir Scand 152:739, 1986
30. Meredith JW, Ditesheim JA, Stonehouse S, et al: Computed tomography and diagnostic peritoneal lavage--complimentary roles in blunt trauma. Am Surg 58:44, 1992
31. Pachter HL, Hofstetter SR: The current status of nonoperative management of adult blunt hepatic injuries. Am J Surg 169:442, 1995
32. Croce MA, Fabian TC, Menke PG, et al: Nonoperative management of blunt hepatic trauma is the treatment of choice for hemodynamically stable patients: Results of a prospective trail. Ann Surg 221:744, 1995
33. Berman SS, Monney EK, Weireter LJ Jr: Late fatal hemorrhage in pediatric liver trauma. J Pediatr Surg 27:1546, 1992
© Williams & Wilkins 1996. All Rights Reserved.
34. Gates J: Delayed hemorrhage with free rupture complicating the nonsurgical management of blunt hepatic trauma: Case report and review of the literature. J Trauma 36:572, 1994