- Email: [email protected]
Trend toward nonoperative management of splenic injuries
Trend Toward Nonoperative Management of Splenic Injuries Mark D. Williams, MD, David H. Young, MD, William R. Schiller, MD, Phoenix, Arizona
Treatment of splenic injuries has evolved over the past decade to reflect more effort to conserve function of the spleen. Records of 169 patients admitted over a 6-year period were identified as documenting the treatment of splenic injuries. We collected data regarding patient age, gender, degree of hemodynamic stability, number of units of blood required, severity of splenic injury, Injury Severity Score, and results of treatment. There were 143 adults (age greater than 16 years) and 26 pediatric patients (age less than 17 years), with mean age in the 2 groups of 31.6 and 11.4 years, respectively. Males comprised 72% of the group, and blunt injury occurred in 154 of the 169 patients. In the adults, spleneetomy, splenorrhaphy, laparotomy without operative treatment of the spleen, and nonoperative management were observed 48%, 30%, 14%, and 8% of the time and in the pediatric group 31%, 27%, 19%, and 23% of the time, respectively. By using operative splenic repair techniques and increased use of nonoperative management, the splenic salvage rate has increased in the last 6 years from 41% to 61% without an increase in morbidity and mortality. Incidence of spleen salvage correlated with severity of spleen and overall injury and cardiovascular stability.
he evaluation and treatment of splenic injuries have T changed over the years, partially due to the documented incidence of serious infection after removal [1]. Since postsplenectomy sepsis is a significant risk in children, nonoperative management is now the usual practice for pediatric injuries involving the spleen [2]. At the same time, splenorrhaphy has become a standard alternative method of managing adult splenic trauma [3,4]. Recently, successful nonoperative management of blunt adult splenic injury has been reported and selectively advocated [5-11]. However, in most institutions, controversy exists over the optimal treatment of adult splenic injuries and whether observation is safe. This report is an attempt to evaluate this trend toward spleen-sparing management. PATIENTS AND METHODS We reviewed the charts of all patients with splenic injuries who were admitted to St. Joseph's Hospital in Phoenix, Arizona, a level I trauma center, from January 1, 1983, through December 31, 1988. Each chart was examined retrospectively, and data were collected regarding date of admission, gender, age, identity of the trauma surgeon, mechanism of injury, hemodynamic stability, trauma indices, associated intra-abdominal injury, grade of splenic injury, performance of diagnostic peritoneal lavage and computed tomography (CT), transfusion requirements, method of treatment, and clinical results. These patients were arbitrarily divided into two groups. The early group included 66 patients admitted from 1983 through 1985~ The late group included 103 patients admitted from 1986 through 1988. We compared these two groups to evaluate the trend toward spleen-sparing management. Significance was tested using chi-square, Fisher's exact test, Wilcoxon's two-sample test, and Kruskal-Wallis analysis of variance where appropriate.
RESULTS The 169 patients ranged in age from 1 month to 74 years old, with a mean age of 28.5 years. There were 26 patients (15%) who were younger than 17 years of age and were considered in the pediatric category. The majority of splenic injuries occurred in young adult males who comprised 72% of the group. Most splenic injuries (91%) were caused by blunt trauma, 54% being the result of automobile or truck accidents, 14% due to motorcycle accidents, 14% the result of pedestrian interactions with From the Trauma Center, St. Joseph'sHospitaland MedicalCenter, moving vehicles, 4% falls, and 5% miscellaneous. PenePhoenix,Arizona. trating trauma comprised 9% of cases, with 5% being Requests for reprints shouldbe addressedto WilliamR. Schiller, gunshot wounds and 4% representing stab wounds. All MD, 2601 East Roosevelt,Phoenix,Arizona85008. Presentedat the 42ndAnnualMeetingof the SouthwesternSurgi- penetrating injuries of the spleen were treated by laparotomy. Associated intra-abdominal injury occurred in 118 cal Congress,La Quinta,California,April 22-25, 1990. 588
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
(70%) of the 169 patients, 78 of whom required operative therapy. Liver damage occurred in 53 patients, making it the organ most frequently damaged. The other organs injured, in decreasing order of frequency, were kidney 22%, diaphragm 16%, colon 14%, mesentery 14%, stomach 11%, small bowel 8%, and pancreas 8%. Splenic injuries were classified according to the criteria described by Buntain and associates [12]. Class I represents capsular or subcapsular disruption without parenchymal injury. In class II injury, parenchymal damage is present without hilar or major vessel involvement. In class III injury, hilar or major vessels are damaged. Class IV refers to fragmentation or devascularization of the spleen. The class of injury was defined at surgery in the operated groups and by CT scan in the nonoperated group. The class of splenic injury in adults correlated well with the salvage rate. Only 13% of class I injuries required splenectomy, 34% were repaired, and more than half did not require operative intervention. The splenectomy rate rose to 38% for class II splenic injuries with 43% requiring repair. Only one class III injury did not require operative intervention, and 21% of injuries in this class were repaired. All adult patients with class IV injured spleens required splenectomy, usually because of instability of vital signs. These differences were statistically significant (p <0.001, chi-square). Analysis of the 143 adult patients revealed that 69 patients (48%) underwent splenectomy, 43 patients (30%) underwent splenorrhaphy, 20 patients (14%) underwent laparotomy without operative treatment of the spleen, and 11 patients (8%) were managed nonoperatively. In the three operated groups, the Injury Severity Score (ISS) was similar. The ISS of the nonoperated group was lower and this approached, but did not achieve, statistical significance (p -- 0.06). The condition of any patient who had a recorded systolic blood pressure of less than 90 mm Hg was defined as unstable, regardless of the response to fluid resuscitation. Only 14% of the patients undergoing splenectomy were hemodynamically stable, compared with 64% of the group treated nonoperatively. The splenorrhaphy and laparotomy without splenic treatment groups fell between these two extremes. The amounts of blood transfused also followed this trend. These differences concerning stability and transfusions were significant (p = 0.0015 and 0.04, respectively). The splenectomy and splenorrhaphy groups had significantly more severe splenic injuries than the other groups (p <0.0001). Compared with the patients who were treated early by laparotomy (splenectomy, splenorrhapy, and laparotomy without operative treatment of the spleen), the nonoperated group was significantly less injured overall as reflected by their lower rate of associated intra-abdominal injuries (p = 0.006), lower ISS (p -- 0.06), lower transfusion requirement (p = 0.001), and higher incidence of hemodynamic stability (p = 0.004). Comparison of the morbidity and mortality of these four groups revealed two deaths in the splenectomy group as a direct result of splenic injury, both in elderly poorrisk patients. There were two wound infections and three intra-abdominal abscesses. Two of the abscesses occurred
in patients with associated colon injury. Two liver injuries and one gastric perforation were missed and required reexploration. One 90-minute delay of operation occurred secondary to difficulty in tomogramic interpretation. There were six respiratory complications. One patient developed pneumonia, which progressed to empyema. Pulmonary embolism occurred in a patient with a pelvic fracture whose postsplenectomy platelet count was greater than 1 million. One of the deaths occurred in an elderly patient who underwent re-exploration for bleeding and died of diffuse intravascular coagulation. Among the 43 patients who underwent splenorrhaphy, no deaths occurred secondary to splenic injury. There were two wound infections, one patient developed atelectasis, one episode of prolonged ileus was observed, and one patient required re-exploration for splenic bleeding, which was treated by splenectomy. There were two operative delays of greater than 90 minutes secondary to false-negative or misinterpreted CT scans. In the 20 patients who underwent laparotomy without operative treatment of the spleen, there were two cases of atelectasis and one delay in operation greater than 90 minutes secondary to a false-negative CT scan. Three of the eleven adults (27%) included in the nonoperative management group eventually underwent surgery because of continued bleeding from their splenic injuries. One patient was scheduled for laparotomy after initial failure of nonoperative management but died with a severe closed head injury (Glasgow coma score of 4) before the operation could commence. One patient required splenectomy and one was managed by splenorrhaphy. Four of these patients had associated intra-abdominal injuries identified on CT scan but none required laparotomy. Eight of the original 11 adults were successfully managed nonoperatively. One patient required 5 units of blood on the ninth hospital day when he underwent open fixation of a femoral fracture. The follow-up CT scan suggested an increase of intraperitoneal blood. Two other patients required 2 units of blood each. The remaining five patients did not require transfusion; however, one of these was a Jehovah's Witness who bled to a hematocrit of 14%. The three patients in whom observation was not successful received 1, 3, and 3 units of blood, respectively. Only patients whose condition became stable after fluid resuscitation were selected for nonoperative management. The patients in whom nonoperative treatment failed tended to have more unstable conditions and to require more blood, although these differences were not significant. The tendency to conserve injured spleens varied greatly among the trauma surgeons involved. Our data show that the treatment received, the amount of blood transfused, and the ability to save the spleen correlated closely with the severity of the splenic and overall injuries. We also identified 20 patients who were hemodynamically stable, had Buntain class I or II splenic injury, and whose ISS excluding closed head injury was less than 20 who eventually underwent laparotomy. This early laparotomy method of management resulted in a splenic salvage rate of 75% and an average
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER1990 589
WILLIAMS ET AL
TABLE I
Comparison of Two 3-Year Treatment Periods 1983-1985
No. of patients Mean age Male Hemodynamic stability Severity of splenic injury class I II III IV Associated abdominal injury Diagnosis Diagnostic peritoneal lavage Computed tomography Treatment Blood transfusion (Average no. units) Distribution of therapy Splenectomy Splenorrhaphy Laparotomy without splenectomy Nonoperative managemerit Results Splenic salvage Infections Length of stay (average no. days) Overall morbidity Overall mortality
1986-1968 p Value
66 29.9 45 (66%) 15 (23%)
103 26.4 77 (75%) 36 (35%)
18 (27%) 26 (40%) 10 (15%) 12 (18%) 49 (74%)
21 (20%) 55 (53%) 12 (12%) 15 (15%) 69 (67%)
0.32
56 (85%) 5 (8%)
70 (68%) 27 (26%)
0.014 0.003
14
39 (59%) 16 (24%) 10 (15%)
8.4
0.35 0.09
0.36*
0.005
39 (38%) 0,007 34 (32%) 0.28 15 (15%) 0.9
1 (2%)
15 (15%)
0.003
27 (41%) 5 (8 % ) 8.7
63 (61%) 4 (4 %) 8.6
0.01 0.3 0.6
36 (55%) 13 (20%)
44 (43%) 16 (16%)
0.13 0.48
* By chi-squaretest.
blood transfusion volume of 2.0 units. These results are not significantly different from our nonoperative splenic salvage rate of 82% (eight successes and one splenorrhaphy after failed nonoperative management) and average transfusion requirement of 1.4 units. In the 26 cases of pediatric splenic injury, the rate of nonoperative management was much greater than in adults (23% versus 8%), reflecting the more prevalent method of treatment for splenic injuries in children. The children who were not operated on were significantly younger than those children who underwent laparotomy (p = 0.003). A comparison of the two 3-year treatment periods is presented in Table I. There were 66 patients in the early group (1983 to 1985) and 103 patients in the recent group (1986 to 1988). The two groups were similar with respect to age, percentage of hemodynamically stable patients, ISS, morbidity, and mortality. In the recent group, nonoperative management was significantly utilized much more frequently (15% versus 2%; p = 0.003), while the rate of splenectomy significantly decreased (59% versus 38%; p = 0.007). Combining the successful splenorrhaphies, nontherapeutic laparotomies, and successful observation cases produces the rate of splenic salvage. The salvage rate increased from 41% in the early group to 61% in the recent group, a significant difference (p -- 0.01). 590
THE AMERICAN JOURNAL OF SURGERY
Although the recent group's severity of splenic injury and overall severity of injury were not significantly different from those in the early group, they received significantly fewer blood transfusions (p = 0.005). The use of CT significantly increased from 8% in the early group to 26% (p = 0.003), whereas the use of diagnostic peritoneal lavage significantly decreased from 85% to 68% (p = 0.014). Only four abdominal tomegrams were performed in this series prior to 1986. Tomography was performed in a total of 30 cases. Four tomograms were misread, producing an accuracy of 87%. In three patients, scanning delayed necessary operation by greater than 90 minutes. In all, the management of 7 of 30 patients (23%) was complicated or delayed by use of the scanner, but we could not identify CT scan delays in those not requiring early laparotomy. COMMENTS By 1983, Dickerman [13] was able to document 24 cases of overwhelming postsplenectomy infection (OPSI) in adults who underwent splenectomy for trauma. Malangoni [14] recently reviewed the literature regarding OPSI in adults and found 1 death due to OPSI following splenectomy for trauma in 825 cases (0.12%). Luna and Dellinger [15] reviewed 2,531 reported splenectomies for trauma and iatrogenic injury and found 11 deaths (0.43%). In a review of 8 published series involving splenectomy in 5,318 patients, morbidity and mortality were reported as significantly increased, although many of these reports did not have strict controls [3]. One study that compared trauma patients who underwent splenectomy with a control group with an intact spleen that underwent laparotomy for other intra-abdominal injuries and matched for ISS and mechanism of injury found no spleen-related increase in morbidity, mortality, or infections [16]. However, it has recently been shown that transfusion as an independent variable is significantly associated with an increase in acute infectious morbidity, presumably because of an inhibition of immune function
[17-19]. Nonoperative management is the latest development in the effort to avoid splenectomy in adults. Although this strategy is partially the result of successful nonoperative management in children, it is well known that in some instances, the injured adult spleen can heal without serious sequelae and with resultant normal function. This was observed by Billroth [20] and has been documented at autopsy. To a large extent, successful nonoperative management of patients with blunt splenic trauma is a manifestation of sensitive modern diagnostic techniques [51. Attempts to select adult patients who will not require operations for their splenic injuries have met with varying degrees of success. Since 1983, there have been 13 series totaling 2,233 adult splenic injuries and the combined nonoperative rate from the literature was 12% [5-11,2227]. Nonoperative management was selected for 356 of these adults and success rates ranged from 22% to 92%. Adding our 11 patients produces a total of 367 patients, with 88 of these patients eventually requiring laparot-
VOLUME 160
DECEMBER 1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
omy, an overall success rate of 76%. Even failures of nonoperative management may be treated successfully by operative splenic salvage without increased morbidity; however, the incidence of splenic repair in this setting has been low. When secondary operative salvage has been attempted, splenic repair was possible in 15 of 50 patients. Length of stay and transfusion requirements are generally higher in patients undergoing delayed laparotomy compared with those undergoing successful nonoperative treatment [5,6,10,22,23]. The controversy therefore persists as to whether patients who meet the criteria for nonoperative management might be better served by early laparotomy with operative splenic salvage [14,22,28]. Although the operative salvage rate after failed nonoperative management is lower than overall reported rates, this may be due to adverse case selection. Fewer blood transfusions might be required if early operative splenic salvage was performed in all patients by theoretically avoiding the increased requirements for unsuccessful nonoperative treatment. Perhaps this strategy would best balance the risks of OPSI against those of transfusion such as hepatitis and acquired immunodeficiency syndrome, and the increased risk of early infection [15,17,19]. Beal and Spisso [28] reviewed splenorrhaphy and found that the risk of blood transfusion necessary to accomplish splenorrhaphy outweighed the risk of OPSI, proposing that expeditious splenectomy would have prevented excessive transfusions. Most series, including our data, reveal that splenectomy patients required more blood than do splenorrhaphy patients and this is probably secondary to the increased magnitude of injuries in the splenectomy group [5,24,26]. Luna and Dellinger [15] reviewed the literature and analyzed the average transfusion requirements for nonoperative management compared with those of early laparotomy. With a post-transfusion hepatitis death rate of 0.14% per unit and using an incidence of death from OPSI following splenectomy ranging from 0.026% to 0.61%, they found that early laparotomy was safer in terms of mortality from OPSI compared with post-transfusion hepatitis. Cogbill et al [6] applied their data (83% nonoperative success rate, a splenic salvage rate of 42% after failed nonoperative management, and transfusion requirements of 1 unit for successful nonoperative management and 3.75 units for failed nonoperative management) and found the late death risk to be the same for both early laparotomy and nonoperative management. Our corroborating data indicate that successful nonoperative therapy resulted in an average transfusion of 1.1 units, failures averaged 2.3 units of blood, and the successful nonoperative management rate was 73%. Whether CT scans can help predict which splenic injuries will require operation is debatable. Buckman et al [21], in their recent retrospective study, found that neither preoperative hypotension nor the anatomic grade of splenic injury significantly correlated with the severity of bleeding found at operation. Nevertheless, if grading of splenic injury is to be useful in the selection of patients for nonoperation, dependability of tomographic information must be assured. In reports in which emergency tomo-
grams were read immediately by experienced radiologists, accuracy rates were equal to or better than those of diagnostic peritoneal lavage [12,29,30]. High-quality CT and skilled interpretation must be readily available without delay. If these considerations are not met, then better results may be obtained by diagnostic peritoneal lavage and early laparotomy with splenic repair.
REFERENCES 1. Singer DB. Postsplenectomy sepsis. Perspect Pediatr Pathol 1973; 1: 285-31l. 2. Upadhyaya P, Simpson JS. Splenic trauma in children. Surg Gynecol Obstet 1968; 126: 781-90. 3. Buntain WL, Lynn HB. Splenorrhaphy: changing concepts for the traumatized spleen. Surgery 1979; 86: 748-60. 4. Millikan JS, Moore EE, Moore GE, Stevens RE. Alternatives to splenectomy in adults after trauma: repair, partial resection, and reimplantation of splenic tissue. Am J Surg 1982; 144:711-6. 5. Mucha P Jr, Daly RC, Farnell MB. Selective management of blunt splenic trauma. J Trauma 1986; 26: 970-9. 6. Cogbill TH, Moore EE, Jurkovich GJ, et al. Nonoperative management of blunt splenic trauma: a multicenter experience. J Trauma 1989; 29: 1312-7. 7. Longo WE, Baker CC, McMillen MA, Modlin IM, DeGutis LC, Zucker KA. Nonoperative management of adult blunt splenic trauma: criteria for successfuloutcome. Ann Surg 1989; 210: 6269. 8. Elmore JR, Clark DE, Isler R J, Horner WR. Selective nonoperative management of blunt splenic trauma in adults. Arch Surg 1989; 124: 581-5. 9. Resciniti A, Fink MP, Raptopoulos V, Davidoff A, Silva WE. Nonoperative treatment of adult splenic trauma: developmentof a computed tomographic scoring system that detects appropriate candidates for expectant management. J Trauma 1988; 128: 82831. 10. Johnson H Jr, Shatney CH. Splenic injuries in adults: selective nonoperative management. South Med J 1986; 79: 5-8. 11. Tom WW, Howells GA, Bree RL, Schwab R, Lucas RJ. A nonoperative approach to the adult ruptured spleen sustained from blunt trauma. Am Surg 1985; 51: 367-71. 12. Buntain WL, Gould HR, Maull KI. Predictability of splenic salvage by computed tomography. J Trauma 1988; 28: 24-34. 13. Dickerman JD. Traumatic asplenia in adults: a defined hazard? Arch Surg 1981; 116: 361-3. 14. Malangoni MA. Early operation for splenic injury. Contemp Surg 1989; 34: 18-21. 15. Luna GK, Dellinger EP. Nonoperative observationtherapy for splenic injuries: a safe therapeutic option? Am J Surg 1987; 153: 462 8. 16. Willis BK, Deitch EA, McDonald JC. The influence of trauma to the spleen on postoperativecomplicationsand mortality. J Trauma 1986; 26: 1073-6. 17. Tartter PI. Blood transfusion and infectious complications following colorectal cancer surgery. Br J Surg 1988; 75: 789-92. 18. Nichols RL, Smith JW, Klein DB, et al. Risk of infection after penetrating abdominal trauma. N Engl J Med 1984; 311: 1065-70. 19. Waymack JP, Robb E, Alexander JW. Effect of transfusion on immune function on a traumatized animal model. II. Effect on mortality rate following septic challenge. Arch Surg 1987; 122: 935-9. 20. Billroth TM. Clinical surgery. Extracts from the reports of surgical practice between the years 1860-1876. Edited and translated by C.T. Dent. London: The New Syndenham Society, 1881: 229. 21. Buckman RF Jr, Dunham CM, Kerr TM, Militello PR. Hypotension and bleedingwith various anatomic patterns of blunt splenic
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER 1990 591
injury in adults. Surg Gynecol Obstet 1989; 169: 206-12. appearance in adults at the original presentation with 22. Malangoni MA, LevineAW, Droege EA, et al. Management subsequent outcome for the individual patient? of injury to the spleen in adults. Results of early operation and observation. Ann Surg 1984; 200: 702-5. Frederick A. Moore (Denver, CO): What is the inci23. Nallathambi MN, Ivatury RR, Wapnir I, et al. Nonoperative dence of hollow visceral injuries for your whole group as management versus early operation for blunt splenic trauma in well as for the group that you observed, and how many adults. Surg Gynecol Obstet 1988; 166: 252-8. 24. Tibi P, Ouriel K, Schwartz SI. Splenic injury in the adult: patients who presented in a delayed manner were treated splenect0my, splenorrhaphy, or nonoperative management. Con- nonoperatively? temp Surg 1985; 26: 73-6. 25. Morgenstern L, Uyeda RY. Nonoperative management of Michael L. Hawkins (Augusta, GA): My impression injuries of the spleen in adults. Surg Gynecol Obstet 1983; 157: is that the salvage rate is much lower if you delay opera513-8. 26. Moss JF, Hopkins WM. Nonoperative management of blunt tion than otherwise. Is this a select group or do you splenic trauma in the adult: a community hospital's experience.J confirm this finding? The second question pertains to Trauma 1987; 27: 315-8. management in terms of activities and bed rest of nono27. Rappaport W, Mclntyre KE, Carmona R. The managementof perated patients with splenic injuries. When can a 17splenic trauma in the adult patient with blunt multiple injuries. year-old boy resume playing football? Finally, in terms of Surg Gynecol Obstet 1990; 170: 204-8. 28. Beal SL, Spisso JM. The risk of splenorrhaphy. Arch Surg transfusion requirements, how do you evaluate patients who have other injuries, specifically orthopedic injuries? 1988; 123: 1158-63. 29. Peitzman AB, Makaroun MS, Slasky BS, Ritter P. Prospective If, for example, a patient has a fractured femur and a study of computed tomography in initial management of blunt fractured tibia and fibula, how do you correlate how abdominal trauma. J Trauma 1986; 26: 585-92. much of the transfusion is related to that versus the 30. Federle MP. Computed tomography of blunt abdominal trau- spleen? ma. Radiol Clin North Am 1983; 21: 461-75. DISCUSSION David V. Felieiano (Rochester, NY): The authors have compared two separate time periods with regards to the percentage of patients who underwent nonoperative management and the percentage who underwent splenic repair in their trauma center in Phoenix. They have concluded that spleen sparing increased by 21% when the second time period of 1986 to 1988 was compared with the 1983 to 1985 time period, and noted that the Injury Severity Score, severity of splenic injury, and severity of associated injuries were equivalent in the two groups. Counting the Western Trauma Association multi-center review by Cogbill et a l and the recent reports from Denver General Hospital, the San Diego Regional Trauma System, Ben Taub General Hospital, and Bellvue Medical Center, this is the fifth report at a major meeting in the last year that documents the two trends in adult splenic management, namely, nonoperative management by computed tomography (CT), and splenorrhaphy instead of splenectomy when operation is performed. I have some questions for the authors. Why did your blood volume replacement change in the second group, when the severity of injuries was the same? Have you consciously avoided transfusion in recent years in light of Luna's data that the risk of the acquired immunodeficiency syndrome from transfusion may exceed the benefits of nonoperative maangement with continued blood transfusion, much as our pediatric surgical colleagues have practiced for more than 15 years? What predictive factors for failed observation in adults were clearly defined by your retrospective study? Can you give us some idea of those adult patients in whom observation is likely to fail, i.e., those hypotensive in the field only, those hypotensive in both the field and the emergency room, or those with large perisplenic hematomas and pelvic collections of blood on first CT scan? And finally, have you been able to correlate the CT 592
Steven H. Miller (Mesa, AZ): I am curious considering the cortical and capsular thinning that occurs in mononucleosis. Were Monospots ever performed on these patients, particularly the patients who were treated conservatively? Among those in whom the Monospot was positive, was conservative treatment successful? Mark D. Williams (closing): Dr. Feliciano, the number of transfusions decreased in the recent group compared with the early group, and we noted that the difference was significant, although the patients had the same severity of both splenic and overall injuries. We believe this is due to the continuing trend and philosophy among our trauma surgeons of trying to limit the risk of transfusion. We weren't able to identify any other factors that would account for it. The one factor that was able to predict failure was hemodynamic instability, and this has been defined variously in different reports. Several reports recently defined it as requiring 1,500 to 2,000 mL of fluid in the emergency department to achieve stability. In our retrospective review, we defined hemodynamic instability as any recorded blood pressure that was less than 90 mm of mercury. We had three failures, and two of those failures were unstable according to that criterion. Our CT experience consisted of 30 scans. Two scans failed to detect class II injuries. One CT scan showed an enlarged spleen in a patient who had a splenic vein thrombosis that required splenectomy. Another scan detected fluid around the spleen. The operative finding was an avulsed splenic pedicle. In our experience, we cannot prospectively use the CT appearance to decide what management these patients need. Dr. Moore, we did not look at the incidence of hollow visceral injury in our patients. You also asked how many patients who underwent nonoperative management had a delayed presentation, and the answer is none. Those patients were excluded from this review.
THE AMERICAN JOURNAL OF SURGERY VOLUME 160 DECEMBER 1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
Dr. Hawkins, the splenic salvage rate after the failure of nonoperative management in a patient is greatly diminished. Regarding management of these patients, it seems that those patients whose condition is stable for 2 or 3 days are going to remain stable. Splenic injury is a marker of the severity of overall injury, and in the great majority of patients, their length of stay is determined by
their other injuries and not by their splenic injury. We suggest that a 17-year-old boy delay returning to football until after 1 year. In answer to your other question, we just looked at their overall transfusion requirement, and more than anything else, their transfusion requirement was a result of their overall injury. Dr. Miller, none of our patients received Monospots.
THE AMERICAN JOURNAL OF SURGERY VOLUME 160 DECEMBER1990 593
Treatment of splenic injuries has evolved over the past decade to reflect more effort to conserve function of the spleen. Records of 169 patients admitted over a 6-year period were identified as documenting the treatment of splenic injuries. We collected data regarding patient age, gender, degree of hemodynamic stability, number of units of blood required, severity of splenic injury, Injury Severity Score, and results of treatment. There were 143 adults (age greater than 16 years) and 26 pediatric patients (age less than 17 years), with mean age in the 2 groups of 31.6 and 11.4 years, respectively. Males comprised 72% of the group, and blunt injury occurred in 154 of the 169 patients. In the adults, spleneetomy, splenorrhaphy, laparotomy without operative treatment of the spleen, and nonoperative management were observed 48%, 30%, 14%, and 8% of the time and in the pediatric group 31%, 27%, 19%, and 23% of the time, respectively. By using operative splenic repair techniques and increased use of nonoperative management, the splenic salvage rate has increased in the last 6 years from 41% to 61% without an increase in morbidity and mortality. Incidence of spleen salvage correlated with severity of spleen and overall injury and cardiovascular stability.
he evaluation and treatment of splenic injuries have T changed over the years, partially due to the documented incidence of serious infection after removal [1]. Since postsplenectomy sepsis is a significant risk in children, nonoperative management is now the usual practice for pediatric injuries involving the spleen [2]. At the same time, splenorrhaphy has become a standard alternative method of managing adult splenic trauma [3,4]. Recently, successful nonoperative management of blunt adult splenic injury has been reported and selectively advocated [5-11]. However, in most institutions, controversy exists over the optimal treatment of adult splenic injuries and whether observation is safe. This report is an attempt to evaluate this trend toward spleen-sparing management. PATIENTS AND METHODS We reviewed the charts of all patients with splenic injuries who were admitted to St. Joseph's Hospital in Phoenix, Arizona, a level I trauma center, from January 1, 1983, through December 31, 1988. Each chart was examined retrospectively, and data were collected regarding date of admission, gender, age, identity of the trauma surgeon, mechanism of injury, hemodynamic stability, trauma indices, associated intra-abdominal injury, grade of splenic injury, performance of diagnostic peritoneal lavage and computed tomography (CT), transfusion requirements, method of treatment, and clinical results. These patients were arbitrarily divided into two groups. The early group included 66 patients admitted from 1983 through 1985~ The late group included 103 patients admitted from 1986 through 1988. We compared these two groups to evaluate the trend toward spleen-sparing management. Significance was tested using chi-square, Fisher's exact test, Wilcoxon's two-sample test, and Kruskal-Wallis analysis of variance where appropriate.
RESULTS The 169 patients ranged in age from 1 month to 74 years old, with a mean age of 28.5 years. There were 26 patients (15%) who were younger than 17 years of age and were considered in the pediatric category. The majority of splenic injuries occurred in young adult males who comprised 72% of the group. Most splenic injuries (91%) were caused by blunt trauma, 54% being the result of automobile or truck accidents, 14% due to motorcycle accidents, 14% the result of pedestrian interactions with From the Trauma Center, St. Joseph'sHospitaland MedicalCenter, moving vehicles, 4% falls, and 5% miscellaneous. PenePhoenix,Arizona. trating trauma comprised 9% of cases, with 5% being Requests for reprints shouldbe addressedto WilliamR. Schiller, gunshot wounds and 4% representing stab wounds. All MD, 2601 East Roosevelt,Phoenix,Arizona85008. Presentedat the 42ndAnnualMeetingof the SouthwesternSurgi- penetrating injuries of the spleen were treated by laparotomy. Associated intra-abdominal injury occurred in 118 cal Congress,La Quinta,California,April 22-25, 1990. 588
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
(70%) of the 169 patients, 78 of whom required operative therapy. Liver damage occurred in 53 patients, making it the organ most frequently damaged. The other organs injured, in decreasing order of frequency, were kidney 22%, diaphragm 16%, colon 14%, mesentery 14%, stomach 11%, small bowel 8%, and pancreas 8%. Splenic injuries were classified according to the criteria described by Buntain and associates [12]. Class I represents capsular or subcapsular disruption without parenchymal injury. In class II injury, parenchymal damage is present without hilar or major vessel involvement. In class III injury, hilar or major vessels are damaged. Class IV refers to fragmentation or devascularization of the spleen. The class of injury was defined at surgery in the operated groups and by CT scan in the nonoperated group. The class of splenic injury in adults correlated well with the salvage rate. Only 13% of class I injuries required splenectomy, 34% were repaired, and more than half did not require operative intervention. The splenectomy rate rose to 38% for class II splenic injuries with 43% requiring repair. Only one class III injury did not require operative intervention, and 21% of injuries in this class were repaired. All adult patients with class IV injured spleens required splenectomy, usually because of instability of vital signs. These differences were statistically significant (p <0.001, chi-square). Analysis of the 143 adult patients revealed that 69 patients (48%) underwent splenectomy, 43 patients (30%) underwent splenorrhaphy, 20 patients (14%) underwent laparotomy without operative treatment of the spleen, and 11 patients (8%) were managed nonoperatively. In the three operated groups, the Injury Severity Score (ISS) was similar. The ISS of the nonoperated group was lower and this approached, but did not achieve, statistical significance (p -- 0.06). The condition of any patient who had a recorded systolic blood pressure of less than 90 mm Hg was defined as unstable, regardless of the response to fluid resuscitation. Only 14% of the patients undergoing splenectomy were hemodynamically stable, compared with 64% of the group treated nonoperatively. The splenorrhaphy and laparotomy without splenic treatment groups fell between these two extremes. The amounts of blood transfused also followed this trend. These differences concerning stability and transfusions were significant (p = 0.0015 and 0.04, respectively). The splenectomy and splenorrhaphy groups had significantly more severe splenic injuries than the other groups (p <0.0001). Compared with the patients who were treated early by laparotomy (splenectomy, splenorrhapy, and laparotomy without operative treatment of the spleen), the nonoperated group was significantly less injured overall as reflected by their lower rate of associated intra-abdominal injuries (p = 0.006), lower ISS (p -- 0.06), lower transfusion requirement (p = 0.001), and higher incidence of hemodynamic stability (p = 0.004). Comparison of the morbidity and mortality of these four groups revealed two deaths in the splenectomy group as a direct result of splenic injury, both in elderly poorrisk patients. There were two wound infections and three intra-abdominal abscesses. Two of the abscesses occurred
in patients with associated colon injury. Two liver injuries and one gastric perforation were missed and required reexploration. One 90-minute delay of operation occurred secondary to difficulty in tomogramic interpretation. There were six respiratory complications. One patient developed pneumonia, which progressed to empyema. Pulmonary embolism occurred in a patient with a pelvic fracture whose postsplenectomy platelet count was greater than 1 million. One of the deaths occurred in an elderly patient who underwent re-exploration for bleeding and died of diffuse intravascular coagulation. Among the 43 patients who underwent splenorrhaphy, no deaths occurred secondary to splenic injury. There were two wound infections, one patient developed atelectasis, one episode of prolonged ileus was observed, and one patient required re-exploration for splenic bleeding, which was treated by splenectomy. There were two operative delays of greater than 90 minutes secondary to false-negative or misinterpreted CT scans. In the 20 patients who underwent laparotomy without operative treatment of the spleen, there were two cases of atelectasis and one delay in operation greater than 90 minutes secondary to a false-negative CT scan. Three of the eleven adults (27%) included in the nonoperative management group eventually underwent surgery because of continued bleeding from their splenic injuries. One patient was scheduled for laparotomy after initial failure of nonoperative management but died with a severe closed head injury (Glasgow coma score of 4) before the operation could commence. One patient required splenectomy and one was managed by splenorrhaphy. Four of these patients had associated intra-abdominal injuries identified on CT scan but none required laparotomy. Eight of the original 11 adults were successfully managed nonoperatively. One patient required 5 units of blood on the ninth hospital day when he underwent open fixation of a femoral fracture. The follow-up CT scan suggested an increase of intraperitoneal blood. Two other patients required 2 units of blood each. The remaining five patients did not require transfusion; however, one of these was a Jehovah's Witness who bled to a hematocrit of 14%. The three patients in whom observation was not successful received 1, 3, and 3 units of blood, respectively. Only patients whose condition became stable after fluid resuscitation were selected for nonoperative management. The patients in whom nonoperative treatment failed tended to have more unstable conditions and to require more blood, although these differences were not significant. The tendency to conserve injured spleens varied greatly among the trauma surgeons involved. Our data show that the treatment received, the amount of blood transfused, and the ability to save the spleen correlated closely with the severity of the splenic and overall injuries. We also identified 20 patients who were hemodynamically stable, had Buntain class I or II splenic injury, and whose ISS excluding closed head injury was less than 20 who eventually underwent laparotomy. This early laparotomy method of management resulted in a splenic salvage rate of 75% and an average
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER1990 589
WILLIAMS ET AL
TABLE I
Comparison of Two 3-Year Treatment Periods 1983-1985
No. of patients Mean age Male Hemodynamic stability Severity of splenic injury class I II III IV Associated abdominal injury Diagnosis Diagnostic peritoneal lavage Computed tomography Treatment Blood transfusion (Average no. units) Distribution of therapy Splenectomy Splenorrhaphy Laparotomy without splenectomy Nonoperative managemerit Results Splenic salvage Infections Length of stay (average no. days) Overall morbidity Overall mortality
1986-1968 p Value
66 29.9 45 (66%) 15 (23%)
103 26.4 77 (75%) 36 (35%)
18 (27%) 26 (40%) 10 (15%) 12 (18%) 49 (74%)
21 (20%) 55 (53%) 12 (12%) 15 (15%) 69 (67%)
0.32
56 (85%) 5 (8%)
70 (68%) 27 (26%)
0.014 0.003
14
39 (59%) 16 (24%) 10 (15%)
8.4
0.35 0.09
0.36*
0.005
39 (38%) 0,007 34 (32%) 0.28 15 (15%) 0.9
1 (2%)
15 (15%)
0.003
27 (41%) 5 (8 % ) 8.7
63 (61%) 4 (4 %) 8.6
0.01 0.3 0.6
36 (55%) 13 (20%)
44 (43%) 16 (16%)
0.13 0.48
* By chi-squaretest.
blood transfusion volume of 2.0 units. These results are not significantly different from our nonoperative splenic salvage rate of 82% (eight successes and one splenorrhaphy after failed nonoperative management) and average transfusion requirement of 1.4 units. In the 26 cases of pediatric splenic injury, the rate of nonoperative management was much greater than in adults (23% versus 8%), reflecting the more prevalent method of treatment for splenic injuries in children. The children who were not operated on were significantly younger than those children who underwent laparotomy (p = 0.003). A comparison of the two 3-year treatment periods is presented in Table I. There were 66 patients in the early group (1983 to 1985) and 103 patients in the recent group (1986 to 1988). The two groups were similar with respect to age, percentage of hemodynamically stable patients, ISS, morbidity, and mortality. In the recent group, nonoperative management was significantly utilized much more frequently (15% versus 2%; p = 0.003), while the rate of splenectomy significantly decreased (59% versus 38%; p = 0.007). Combining the successful splenorrhaphies, nontherapeutic laparotomies, and successful observation cases produces the rate of splenic salvage. The salvage rate increased from 41% in the early group to 61% in the recent group, a significant difference (p -- 0.01). 590
THE AMERICAN JOURNAL OF SURGERY
Although the recent group's severity of splenic injury and overall severity of injury were not significantly different from those in the early group, they received significantly fewer blood transfusions (p = 0.005). The use of CT significantly increased from 8% in the early group to 26% (p = 0.003), whereas the use of diagnostic peritoneal lavage significantly decreased from 85% to 68% (p = 0.014). Only four abdominal tomegrams were performed in this series prior to 1986. Tomography was performed in a total of 30 cases. Four tomograms were misread, producing an accuracy of 87%. In three patients, scanning delayed necessary operation by greater than 90 minutes. In all, the management of 7 of 30 patients (23%) was complicated or delayed by use of the scanner, but we could not identify CT scan delays in those not requiring early laparotomy. COMMENTS By 1983, Dickerman [13] was able to document 24 cases of overwhelming postsplenectomy infection (OPSI) in adults who underwent splenectomy for trauma. Malangoni [14] recently reviewed the literature regarding OPSI in adults and found 1 death due to OPSI following splenectomy for trauma in 825 cases (0.12%). Luna and Dellinger [15] reviewed 2,531 reported splenectomies for trauma and iatrogenic injury and found 11 deaths (0.43%). In a review of 8 published series involving splenectomy in 5,318 patients, morbidity and mortality were reported as significantly increased, although many of these reports did not have strict controls [3]. One study that compared trauma patients who underwent splenectomy with a control group with an intact spleen that underwent laparotomy for other intra-abdominal injuries and matched for ISS and mechanism of injury found no spleen-related increase in morbidity, mortality, or infections [16]. However, it has recently been shown that transfusion as an independent variable is significantly associated with an increase in acute infectious morbidity, presumably because of an inhibition of immune function
[17-19]. Nonoperative management is the latest development in the effort to avoid splenectomy in adults. Although this strategy is partially the result of successful nonoperative management in children, it is well known that in some instances, the injured adult spleen can heal without serious sequelae and with resultant normal function. This was observed by Billroth [20] and has been documented at autopsy. To a large extent, successful nonoperative management of patients with blunt splenic trauma is a manifestation of sensitive modern diagnostic techniques [51. Attempts to select adult patients who will not require operations for their splenic injuries have met with varying degrees of success. Since 1983, there have been 13 series totaling 2,233 adult splenic injuries and the combined nonoperative rate from the literature was 12% [5-11,2227]. Nonoperative management was selected for 356 of these adults and success rates ranged from 22% to 92%. Adding our 11 patients produces a total of 367 patients, with 88 of these patients eventually requiring laparot-
VOLUME 160
DECEMBER 1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
omy, an overall success rate of 76%. Even failures of nonoperative management may be treated successfully by operative splenic salvage without increased morbidity; however, the incidence of splenic repair in this setting has been low. When secondary operative salvage has been attempted, splenic repair was possible in 15 of 50 patients. Length of stay and transfusion requirements are generally higher in patients undergoing delayed laparotomy compared with those undergoing successful nonoperative treatment [5,6,10,22,23]. The controversy therefore persists as to whether patients who meet the criteria for nonoperative management might be better served by early laparotomy with operative splenic salvage [14,22,28]. Although the operative salvage rate after failed nonoperative management is lower than overall reported rates, this may be due to adverse case selection. Fewer blood transfusions might be required if early operative splenic salvage was performed in all patients by theoretically avoiding the increased requirements for unsuccessful nonoperative treatment. Perhaps this strategy would best balance the risks of OPSI against those of transfusion such as hepatitis and acquired immunodeficiency syndrome, and the increased risk of early infection [15,17,19]. Beal and Spisso [28] reviewed splenorrhaphy and found that the risk of blood transfusion necessary to accomplish splenorrhaphy outweighed the risk of OPSI, proposing that expeditious splenectomy would have prevented excessive transfusions. Most series, including our data, reveal that splenectomy patients required more blood than do splenorrhaphy patients and this is probably secondary to the increased magnitude of injuries in the splenectomy group [5,24,26]. Luna and Dellinger [15] reviewed the literature and analyzed the average transfusion requirements for nonoperative management compared with those of early laparotomy. With a post-transfusion hepatitis death rate of 0.14% per unit and using an incidence of death from OPSI following splenectomy ranging from 0.026% to 0.61%, they found that early laparotomy was safer in terms of mortality from OPSI compared with post-transfusion hepatitis. Cogbill et al [6] applied their data (83% nonoperative success rate, a splenic salvage rate of 42% after failed nonoperative management, and transfusion requirements of 1 unit for successful nonoperative management and 3.75 units for failed nonoperative management) and found the late death risk to be the same for both early laparotomy and nonoperative management. Our corroborating data indicate that successful nonoperative therapy resulted in an average transfusion of 1.1 units, failures averaged 2.3 units of blood, and the successful nonoperative management rate was 73%. Whether CT scans can help predict which splenic injuries will require operation is debatable. Buckman et al [21], in their recent retrospective study, found that neither preoperative hypotension nor the anatomic grade of splenic injury significantly correlated with the severity of bleeding found at operation. Nevertheless, if grading of splenic injury is to be useful in the selection of patients for nonoperation, dependability of tomographic information must be assured. In reports in which emergency tomo-
grams were read immediately by experienced radiologists, accuracy rates were equal to or better than those of diagnostic peritoneal lavage [12,29,30]. High-quality CT and skilled interpretation must be readily available without delay. If these considerations are not met, then better results may be obtained by diagnostic peritoneal lavage and early laparotomy with splenic repair.
REFERENCES 1. Singer DB. Postsplenectomy sepsis. Perspect Pediatr Pathol 1973; 1: 285-31l. 2. Upadhyaya P, Simpson JS. Splenic trauma in children. Surg Gynecol Obstet 1968; 126: 781-90. 3. Buntain WL, Lynn HB. Splenorrhaphy: changing concepts for the traumatized spleen. Surgery 1979; 86: 748-60. 4. Millikan JS, Moore EE, Moore GE, Stevens RE. Alternatives to splenectomy in adults after trauma: repair, partial resection, and reimplantation of splenic tissue. Am J Surg 1982; 144:711-6. 5. Mucha P Jr, Daly RC, Farnell MB. Selective management of blunt splenic trauma. J Trauma 1986; 26: 970-9. 6. Cogbill TH, Moore EE, Jurkovich GJ, et al. Nonoperative management of blunt splenic trauma: a multicenter experience. J Trauma 1989; 29: 1312-7. 7. Longo WE, Baker CC, McMillen MA, Modlin IM, DeGutis LC, Zucker KA. Nonoperative management of adult blunt splenic trauma: criteria for successfuloutcome. Ann Surg 1989; 210: 6269. 8. Elmore JR, Clark DE, Isler R J, Horner WR. Selective nonoperative management of blunt splenic trauma in adults. Arch Surg 1989; 124: 581-5. 9. Resciniti A, Fink MP, Raptopoulos V, Davidoff A, Silva WE. Nonoperative treatment of adult splenic trauma: developmentof a computed tomographic scoring system that detects appropriate candidates for expectant management. J Trauma 1988; 128: 82831. 10. Johnson H Jr, Shatney CH. Splenic injuries in adults: selective nonoperative management. South Med J 1986; 79: 5-8. 11. Tom WW, Howells GA, Bree RL, Schwab R, Lucas RJ. A nonoperative approach to the adult ruptured spleen sustained from blunt trauma. Am Surg 1985; 51: 367-71. 12. Buntain WL, Gould HR, Maull KI. Predictability of splenic salvage by computed tomography. J Trauma 1988; 28: 24-34. 13. Dickerman JD. Traumatic asplenia in adults: a defined hazard? Arch Surg 1981; 116: 361-3. 14. Malangoni MA. Early operation for splenic injury. Contemp Surg 1989; 34: 18-21. 15. Luna GK, Dellinger EP. Nonoperative observationtherapy for splenic injuries: a safe therapeutic option? Am J Surg 1987; 153: 462 8. 16. Willis BK, Deitch EA, McDonald JC. The influence of trauma to the spleen on postoperativecomplicationsand mortality. J Trauma 1986; 26: 1073-6. 17. Tartter PI. Blood transfusion and infectious complications following colorectal cancer surgery. Br J Surg 1988; 75: 789-92. 18. Nichols RL, Smith JW, Klein DB, et al. Risk of infection after penetrating abdominal trauma. N Engl J Med 1984; 311: 1065-70. 19. Waymack JP, Robb E, Alexander JW. Effect of transfusion on immune function on a traumatized animal model. II. Effect on mortality rate following septic challenge. Arch Surg 1987; 122: 935-9. 20. Billroth TM. Clinical surgery. Extracts from the reports of surgical practice between the years 1860-1876. Edited and translated by C.T. Dent. London: The New Syndenham Society, 1881: 229. 21. Buckman RF Jr, Dunham CM, Kerr TM, Militello PR. Hypotension and bleedingwith various anatomic patterns of blunt splenic
THE AMERICAN JOURNAL OF SURGERY VOLUME160 DECEMBER 1990 591
injury in adults. Surg Gynecol Obstet 1989; 169: 206-12. appearance in adults at the original presentation with 22. Malangoni MA, LevineAW, Droege EA, et al. Management subsequent outcome for the individual patient? of injury to the spleen in adults. Results of early operation and observation. Ann Surg 1984; 200: 702-5. Frederick A. Moore (Denver, CO): What is the inci23. Nallathambi MN, Ivatury RR, Wapnir I, et al. Nonoperative dence of hollow visceral injuries for your whole group as management versus early operation for blunt splenic trauma in well as for the group that you observed, and how many adults. Surg Gynecol Obstet 1988; 166: 252-8. 24. Tibi P, Ouriel K, Schwartz SI. Splenic injury in the adult: patients who presented in a delayed manner were treated splenect0my, splenorrhaphy, or nonoperative management. Con- nonoperatively? temp Surg 1985; 26: 73-6. 25. Morgenstern L, Uyeda RY. Nonoperative management of Michael L. Hawkins (Augusta, GA): My impression injuries of the spleen in adults. Surg Gynecol Obstet 1983; 157: is that the salvage rate is much lower if you delay opera513-8. 26. Moss JF, Hopkins WM. Nonoperative management of blunt tion than otherwise. Is this a select group or do you splenic trauma in the adult: a community hospital's experience.J confirm this finding? The second question pertains to Trauma 1987; 27: 315-8. management in terms of activities and bed rest of nono27. Rappaport W, Mclntyre KE, Carmona R. The managementof perated patients with splenic injuries. When can a 17splenic trauma in the adult patient with blunt multiple injuries. year-old boy resume playing football? Finally, in terms of Surg Gynecol Obstet 1990; 170: 204-8. 28. Beal SL, Spisso JM. The risk of splenorrhaphy. Arch Surg transfusion requirements, how do you evaluate patients who have other injuries, specifically orthopedic injuries? 1988; 123: 1158-63. 29. Peitzman AB, Makaroun MS, Slasky BS, Ritter P. Prospective If, for example, a patient has a fractured femur and a study of computed tomography in initial management of blunt fractured tibia and fibula, how do you correlate how abdominal trauma. J Trauma 1986; 26: 585-92. much of the transfusion is related to that versus the 30. Federle MP. Computed tomography of blunt abdominal trau- spleen? ma. Radiol Clin North Am 1983; 21: 461-75. DISCUSSION David V. Felieiano (Rochester, NY): The authors have compared two separate time periods with regards to the percentage of patients who underwent nonoperative management and the percentage who underwent splenic repair in their trauma center in Phoenix. They have concluded that spleen sparing increased by 21% when the second time period of 1986 to 1988 was compared with the 1983 to 1985 time period, and noted that the Injury Severity Score, severity of splenic injury, and severity of associated injuries were equivalent in the two groups. Counting the Western Trauma Association multi-center review by Cogbill et a l and the recent reports from Denver General Hospital, the San Diego Regional Trauma System, Ben Taub General Hospital, and Bellvue Medical Center, this is the fifth report at a major meeting in the last year that documents the two trends in adult splenic management, namely, nonoperative management by computed tomography (CT), and splenorrhaphy instead of splenectomy when operation is performed. I have some questions for the authors. Why did your blood volume replacement change in the second group, when the severity of injuries was the same? Have you consciously avoided transfusion in recent years in light of Luna's data that the risk of the acquired immunodeficiency syndrome from transfusion may exceed the benefits of nonoperative maangement with continued blood transfusion, much as our pediatric surgical colleagues have practiced for more than 15 years? What predictive factors for failed observation in adults were clearly defined by your retrospective study? Can you give us some idea of those adult patients in whom observation is likely to fail, i.e., those hypotensive in the field only, those hypotensive in both the field and the emergency room, or those with large perisplenic hematomas and pelvic collections of blood on first CT scan? And finally, have you been able to correlate the CT 592
Steven H. Miller (Mesa, AZ): I am curious considering the cortical and capsular thinning that occurs in mononucleosis. Were Monospots ever performed on these patients, particularly the patients who were treated conservatively? Among those in whom the Monospot was positive, was conservative treatment successful? Mark D. Williams (closing): Dr. Feliciano, the number of transfusions decreased in the recent group compared with the early group, and we noted that the difference was significant, although the patients had the same severity of both splenic and overall injuries. We believe this is due to the continuing trend and philosophy among our trauma surgeons of trying to limit the risk of transfusion. We weren't able to identify any other factors that would account for it. The one factor that was able to predict failure was hemodynamic instability, and this has been defined variously in different reports. Several reports recently defined it as requiring 1,500 to 2,000 mL of fluid in the emergency department to achieve stability. In our retrospective review, we defined hemodynamic instability as any recorded blood pressure that was less than 90 mm of mercury. We had three failures, and two of those failures were unstable according to that criterion. Our CT experience consisted of 30 scans. Two scans failed to detect class II injuries. One CT scan showed an enlarged spleen in a patient who had a splenic vein thrombosis that required splenectomy. Another scan detected fluid around the spleen. The operative finding was an avulsed splenic pedicle. In our experience, we cannot prospectively use the CT appearance to decide what management these patients need. Dr. Moore, we did not look at the incidence of hollow visceral injury in our patients. You also asked how many patients who underwent nonoperative management had a delayed presentation, and the answer is none. Those patients were excluded from this review.
THE AMERICAN JOURNAL OF SURGERY VOLUME 160 DECEMBER 1990
NONOPERATIVEMANAGEMENTOF SPLENICINJURIES
Dr. Hawkins, the splenic salvage rate after the failure of nonoperative management in a patient is greatly diminished. Regarding management of these patients, it seems that those patients whose condition is stable for 2 or 3 days are going to remain stable. Splenic injury is a marker of the severity of overall injury, and in the great majority of patients, their length of stay is determined by
their other injuries and not by their splenic injury. We suggest that a 17-year-old boy delay returning to football until after 1 year. In answer to your other question, we just looked at their overall transfusion requirement, and more than anything else, their transfusion requirement was a result of their overall injury. Dr. Miller, none of our patients received Monospots.
THE AMERICAN JOURNAL OF SURGERY VOLUME 160 DECEMBER1990 593