Is Early Discharge Following Isolated Splenic Injury in the Hemodynamically Stable Child Possible?

By James M. Lynch, Henri Ford, Mary J. Gardner, and Eugene S. Weiner Pittsburgh, Pennsylvania

�9 Nonoperative treatment of splenic injury is well ac- cepted. Two questions have not been answered. (1) What is the intensity of monitoring required in the hemodynamically stable patient? (2) How long do patients need to be hospital- ized? Ninety-one patients having computed tomography (CT) or surgically proven splenic injury were treated between September 1986 and September 1991. Excluded from the study were 16 patients requiring operation and 22 patients having multiple system injuries. All operations occurred within 24 hours of admission. No transfusions were required later than 48 hours following admission. The remaining 53 patients (58%} constitute the study group. CT classification of Buntain indicated 6 class I, 21 class II, 24 class III, and 2 class IV injuries. The mean Injury Severity Score (ISS) for the group was 6.98 + 3.43. Serial hematocrits for the patients treated without transfusions were followed until three con- secutive determinations showed no change. The lowest average bematocrit for the nontransfused group was 30.96% -+ 4.47% and occurred on day 2.06 _+ 0.76. Eleven patients (23%} had left.sided pleural effusions that resolved without intervention. One patient had an ileus for 3 days. CT or ultrasound examination was obtained on day 5 to 7 to document healing before the patient was allowed out of bed and discharged. The average hospital stay was 7.06 -+ 2.24 days. Twenty-two patients were initially observed in the intensive care unit (ICU}. Clearly the interval between hema- tocrit stability (average, 2.06 days} and discharge (average, 7.06 days) constitutes a time of minimal nursing care while utilizing bed space and health care dollars. We conclude that in the study group, (1) there were no benefits to ICU monitoring; (2) most patients could have been discharged after day 3; and (3) implementing an early discharge policy is safe and results in health care dollar savings. Copyright �9 1993 by W.B, Saunders Company

INDEX WORDS: Splenic trauma, nonoperative therapy, pedi- atric.

N 'ONOPERATIVE management of splenic inju- ries in children is well accepted with success

rates approaching 90%. TM Accurate fluid resuscita- tion, hemodynamic monitoring with transfusion when necessary, and bed rest constitutes early therapy. We have observed many patients who appear to tolerate these injuries very well and appear ready for dis- charge only a few days after admission. We therefore questioned whether intensive care unit (ICU) admis- sions were appropriate in this group and whether the last few days of forced hospital bed rest for little more than daily hematocrit determinations was of any benefit to the child.

In order to study this problem, we have retrospec- tively reviewed all 91 patients treated for splenic injury at our institution during a 5-year period to

answer the following questions. What is the intensity of monitoring required in the hemodynamically stable child with isolated splenic injury? How long do the children need to be hospitalized? Are Injury Severity Scores (ISS), mechanism of injury data, or computed tomography (CT)-graded severity of injury helpful in deciding who might be candidates for early dis- charge? And is early discharge safe and cost-effective in the situation of an isolated splenic injury?

MATERIALS AND METHODS

There were 91 patients with splenic injuries admitted to the Benedum Pediatric Trauma Service at Children's Hospital of Pittsburgh between September 1986 and September 1991. A retrospective chart review was performed to determine mechanism of injury, injuries received, demographics, hematocrit determina- tions, transfusion, complications, length of stay, and duration of ICU stay for each patient. The Benedum Pediatric Trauma registry was used to establish the ISS.

CT grading of splenic injuries was based on the classification of Buntain et alS: class I injuries have localized capsular disruption or subcapsular hematoma without significant parencbymal injury; class II have single or multiple capsular or parenchymal disrup- tions, transverse or longitudinal, that do not extend into the hilum or involve major vessels. Intraparenchymal hematoma may or may not coexist; class III have deep fractures, single or multiple, transverse or longitudinal, extending into the hilum and involving major blood vessels; class IV have completely shattered or frag- mented spleen or a spleen separated from its normal blood supply at the pedicle.

Initial evaluation and management was according to Children's Hospital of Pittsburgh protocol for suspected abdominal injury. During the early years of the study, most patients were admitted to the ICU. In the latter years of the study, some patients were admitted to the regular surgical nursing units. Intravenous fluid therapy, nasogastric (NG) suctioning for 24 hours, frequent vital signs, serial abdominal examinations, serial hematocrit determina- tions, and bed rest were initiated on all patients. Hematocrits were performed every 8 hours until stabilization of three consecutive hematocrits occurred. Hematocrits were then obtained daily. Transfusions were given for hematocrits less than 21% to 25% with the approval of the attending surgeon. All patients were restudied

From the Department of Pediatric Surgery and The Benedum Pediatric Trauma Program, Children's Hospital of Pittsburgh, Pitts- burgh, PA.

Presented at the 1992Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, San Francisco, California, October 9-11, 1992.

Address reprint requests to James M. Lynch, MD, Director, Benedum Pediatric Trauma Program, Department of Pediatric Surgery, Children's Hospital of Pittsburgh, 3705 Fifth Ave, Pittsburgh, PA 15213- 2583.

Copyright �9 1993 by W.B. Saunders Company 0022-3468/93/2810-0035503.00/0

Journal of Pediatric Surgery, Vo128, No 10 (October), 1993: pp 1403-1407 1403

1404

for healing of the splenic lesion at 5 to 7 days postinjury by CT or ultrasound (US) scan. All patients were examined as outpatients with an ultrasound examination at 1 month after discharge. Children were allowed to return to school after 1 week of restricted activity at home. Gym and participation in other sports was prohibited until complete healing of the lesion was demonstrated.

RESULTS

Ninety-one patients had a diagnosis of splenic injury during the 5-year study period. There were 60 males and 31 females. The diagnosis of splenic injury was made by computed tomography in 80 and lapa- rotomy in 11.

Sixteen patients (18%), group I, required early operation for abdominal or brain injury (11 laparoto- mies and 5 craniotomies). Four splenorrhaphies and 3 splenectomies were performed. The four remaining laparotomies were for pancreas, liver (2), or renal artery injuries, and 4 nonbleeding splenic injuries were found. In the five craniotomy patients, the diagnosis of splenic injury was made or confirmed by CT. Fourteen of the patients in this group required transfusion (average, 43.8 mL/kg).

Twenty-two patients (24%), group II, had docu- mented splenic injuries along with at least one other significant organ system injury; predominantly brain or orthopedic. All major injuries were identified during the first 24 hours of hospitalization and all required ICU admission. Nine of these patients required transfusion therapy (average, 12.8 mL/kg).

Group III, the study group, had splenic injuries diagnosed by CT, were hemodynamically stable, re- quired little or no initial fluid resuscitation (<20 mL/kg), and had no other significant associated injuries. Five patients were transfused (average, 9.77 mL/kg). All transfusions were received in the first 48 hours. None were needed afterwards. Two of the transfusions were inadvertently given for hematocrits of 28% and 29% respectively. Forty-eight patients in group III were hemodynamically stable throughout and received no transfusion therapy.

The average age for all patients was 8.41 years (range, 3 months to 17 years). The average age of the study group, group III, was 8.86 years. There were a total of four deaths: two in group I and two in group II. There were no deaths in group III. Mean ISS for all 91 patients was 11.65. ISS for group I averaged 21.50 ___ 10.92. Group II's ISS was 14.14 __. 3.75, and group iii 's i s s was 6.98 _.+ 3.43 (Table 1).

Mechanisms of injury data implied greater injury severity for groups I and II (Tables 1 and 2). Motor vehicle accident (MVA) related injuries accounted for 71% (27/38) of all injuries in the operated and multiply injured groups and only 23% in the study group. Falls and sports-related injuries accounted for

LYNCH ET AL

Table 1. Average Age and Injury Severity Score Related to Group

Group Age (yr) ISS

I 7.50 21.50 II 8.14 14.14 Ill 8.86 6.98

NOTE. Group I, early operation; group II, multiply injured; group III, study group.

49% of the injuries seen in the study group but less than 8% (3 of 38) of those in groups I and II.

The average hematocrit for the 48 patients not transfused in group III was 30.96% - 4.47%. Our policy has been to transfuse if the hematocrit falls to 21% to 25% with the approval of an attending physician.

Complications in group III were infrequent. No episodes of delayed rupture or delayed bleeding were seen. The most common complication was pleural effusion or atelectasis of the left lower lobe of the lung in 11 patients (21%). All of these were asymptom- atic and represented incidental findings on CT exami- nation. No therapeutic interventions were required. One patient experienced prolonged ileus requiring reinsertion of a NG tube. This resolved in 3 days.

The severity of the injury to the spleen was docu- mented by CT scan and graded according to Bun- tain's classification. 5 There were 6 class I, 21 class II, 24 class III, and 2 class IV injuries. One class IV injury was inflicted by a horse and the other by an all-terrain vehicle. Twenty-six of the 53 injuries (49%) were considered severe by this grading system.

Follow-up CT or US examinations were obtained on day 5 to 7 following injury on all patients prior to discharge from the hospital. All patients had at least one CT or US examination as part of their follow-up as an outpatient. All CT and US examinations showed healing of the injury as evidenced by: (1) decreasing size of the intraparenchymal hematoma; (2) decreas- ing size of the perisplenic hematoma; and/or (3) stability or repair of the extent or depth of the fracture(s).

The average hospital stay for group III was 7.06 --- 2.24 days. Twenty-two patients were initially admitted to the ICU in the earlier years of the study. A total of 54 ICU days were used by group III patients. There were no therapeutic interventions performed in the

Table 2. Mechanism of Splenic Injury Related to Group

Group MVA Sports Fall Bike Other

I 13 (81%1 0 O 2 (13%) 1 (6%) II 14 (64%) 2 (9%) 1 (4%) 2 (9%) 3 (14%) Ill 12 (23%) 14 (26%) 12 (23%) 8 (15%) 7 (13%)

NOTE. Group I, early operation; group II, multiply injured; group Ill, study group.

EARLY DISCHARGE AFTER SPLENIC TRAUMA 1405

ICU and no apparent benefit for any of the patients. The cost of an ICU bed for 1 day at our institution is $2,000.

DISCUSSION

We have defined a group of patients with isolated splenic injuries who appear not to benefit from ICU care or prolonged hospitalization. These patients required minimal fluid resuscitation (less than 20 mL/kg crystalloid), showed early stabilization of he- matocrits, and have mechanisms of injury indicating less severe injury, resulting in lower ISS scores. The group was relatively easy to identify early in their hospital course. Those children with hemodynami- cally unstable splenic injuries requiring operation and those requiring operation for other life threatening reasons, became evident within the first few hours of hospitalization (group I). Patients with multiple or- gan system injuries also were identified early, within the first 24 hours, by use of CT scans and other diagnostic modalities (group II). This leaves those patients with isolated splenic injuries (group III). Five of these received transfusions of which three were necessary. All indicated transfusions were given within 48 hours of admission.

All patients with splenic injuries are restricted to bed rest to protect the spleen from being reinjured and to prevent early disruption of the hematoma in the area of injury. This empirically makes sense during the initial days following injury. However, scientific evidence that prolonged bed rest is benefi- cial appears to be lacking. Ein's landmark paper stated that patients were kept for 2 to 3 weeks in the hospital, on bed rest. 6 We suspect that the length of observation was partially in response to the great criticism that nonoperative therapy generated in that decade. With the firm success of nonoperative therapy established at a number of centers treating children, the hospital stay has dropped to 7 to 10 days. 1,7,8 We have been unable to find any literature discussing why 7 to 10 days is mandatory if the patient is stable earlier.

The young spleen benefits from certain structural characteristics, which if left undisturbed for 24 to 48 hours, may favor the success of nonoperative manage- ment. A thicker splenic capsule is present in children in which an abundance of myoepithelial cells favor hemostatic control of hemorrhage. 9 In addition, the lack of systemic atherosclerosis and degenerative changes in the arterioles of the spleen, allow for greater contraction and retraction of the damaged arterioles leading to more effective local hemostasis. 1~ As long as the major arteries or veins to the spleen are not transected, one would expect significant and

efficient clot formation in blood vessels. Once these vessels are effectively sealed by the normal hemo- static mechanisms of the body, it makes little sense that bed rest will effect further healing.

In populations of mild to moderate injury, fibrino- lytic activity increases early in the trauma course but returns to normal within 24 hours. 11 Our study group of patients had a mean ISS of 6.98 _+ 3.43, indicating mild to moderate injury. Therefore, there was no reason to believe that, in the study group, rebleeding secondary to known changes in the coagulation sys- tem following trauma would be a factor. Once the clot formed, retracted, and became organized, further bleeding should not occur. The time from injury to stable hematocrit of 2.06 days implies that this was the case in these patients.

In order to directly study the severity of injury to the spleen itself as a predictive factor in the success of early discharge, we graded the CT findings according to Buntain's criteria: CT grading of splenic injuries has been used to predict successful nonoperative management in adults. 12 We expected that the major- ity of injuries in this group would fall in the mild (class I or II) category. However, the findings were almost equally divided between the mild and severe; 6 class I and 21 class II versus 24 class III and 2 class IV. Several studies in the adult literature have shown that CT grading tends to underestimate the severity of the spleen injuries. 12,13 It appears that while CT is valu- able in the diagnosis of splenic injuries and identifica- tion of other solid organ injuries, therapy should still be based on the clinical response to the injury. We have not operated on a splenic injury because of CT appearance alone) 4 Therefore, CT findings were not very helpful in determining the success of early management and discharge. These findings lead us to believe that one may see marked disruption of the gross appearance of the spleen even with trivial trauma, yet the tissue distruction along the lines of injury is probably minimal. Whereas, when the mecha- nism of injury is severe (as in an MVA-related injury where massive kinetic energy can be delivered to the body), the CT scan is unable to quantify the marked tissue distruction along the obvious (by CT) planes of injury.

The lower ISS scores and less severe mechanisms of injury (Tables 1 and 2) argue that less tissue distruction occurred in the study group. Both ISS and the mechanism of injury have been used in the pediatric population to predict success of nonopera- tive management: Falls and sports related injuries accounted for 49% of the injuries in the study group (group III) but only 3 of 38 injuries (8%) in the seriously injured groups (groups I and II). MVA-

1406 LYNCH ET AL

related injuries accounted for 71%, (27 of 38) of the seriously injured patients and only 23% of the study group.

We had no unexpected complications in the study group except for one case of prolonged ileus. There were no patients who had "late bleeds." The average posttrauma day when the hematocrit reached its nadir was day 2.06. We confirmed that no further drops in hematocrit occurred by studying daily hema- tocrits until discharge. The average day of discharge was 7.06. Therefore, on the average, there were 5 days used to confirm that the hematocrit was stable. If we were to allow two days to confirm that the hematocrit was stable, than we could potentially discharge some children on day 4, thus saving 3 days of room charges.

In the present atmosphere of cost containment in health care, substantial savings could result if both ICU and total hospital days were reduced. There is little argument that group I and group II patients require ICU care in the first few days following injury. However in the group III patients, 54 ICU hospital days were used with little benefit. The cost of an ICU

bed in our institution is 2�89 times our regular room rate ($2,000 v $800). Therefore, in this group, a potential savings of close to $250,000 in health care costs attributable to room rates alone could occur.

Based on these findings, our own approach to the care of the hemodynamically stable child has changed. Children with hemodynamically stable CT class I, II, or III isolated splenic injuries are observed in the regular nursing units. We are still concerned about observing class IV injuries, even if stable, outside of the ICU setting! Once the child's hematocrit stabi- lizes, a US examination is obtained to confirm that no further bleeding or disruption has occurred. Two more hematocrits are obtained during the next 24 to 36 hours, and if there is no further drop, the child is discharged home on restricted activity and fol- lowed-up as previously described. Since instituting this policy in a limited number of patients, we have observed no untoward events. We are in the process of a prospective study to confirm that this program of earlier discharge following mild to moderate isolated splenic injury continues to be safe and cost effective.

REFERENCES

1. Pearl RH, Wesson DE, Spence IA, et al: Splenic injury: A 5-year update with improved results and changing criteria for conservative management. J Pediatr Surg 24:428-431, 1989

2. Cohen RC: Blunt splenic trauma in children. A retrospective study of non-operative management. Aust Paediatr J 18:211-215, 1982

3. Kakkasseril JS, Steward D, Cox JA, et al: Changing treatment of pediatric splenic trauma. Arch Surg 117:758-759, 1982

4. Wesson DE: Abdominal injuries in children. Can J Surg 27:472-474, 1984

5. Buntain WL, Gould HR, Maull KI: Predictability of splenic salvage by computed tomography. J Trauma 28:24-31, 1988

6. Ein SH, Shandling B, Simpson JS, et al: Nonoperative management of traumatized spleen in children: How and why. J Pediatr Surg 13:117-119, 1978

7. Delius RE, Frankel W, Coran AG: A comparison between operative and nonoperative management of blunt injuries to the liver and spleen in adult and pediatric patients. Surgery 106:788- 793, 1989

8. Muehrcke DD, Kim SH, McCabe CJ: Pediatric splenic trauma: Predicting the success of nonoperative therapy. Am J Emerg Med 5:109-112, 1987

9. Shackford SR, Molin M: Management of splenic injuries. Surg Clin North Am 70:595-620, 1990

10. Morgenstern L, Uyeda RY: Nonoperative management of injuries of the spleen in adults. Surg Gynecol Obstet 157:513-518, 1983

11. Kapsch D, Metzler M, Harrington M, et al: Fibrinolytic response to trauma. Surgery 95:473-478, 1984

12. Malangoni MA, Cue Jl, Fallat ME, et al: Evaluation of splenic injury by computed tomography and its impact on treat- ment. Ann Surg 211:592-599, 1990

13. Umlas SL, Cronan JJ: Splenic trauma: Can CT grading systems enable prediction of successfull nonsurgical treatment? Radiology 17:481-487, 1991

14. Brick SH, Taylor GA, Potter BM, et al: Hepatic and splenic injury in children: Role of CT in the decision for laparotomy. Radiology 165:643-646, 1987

D i s c u s s i o n

C. Musemeche (Albuquerque, NM): This is a long overdue study which demonstrates the rationale and safety of the early discharge of stable patients follow- ing isolated splenic injury. Though widely practiced, this information has not been previously published. This paper, therefore, is another milestone in the evolution of nonoperative therapy of splenic trauma.

My first question concerns the practice of perform- ing multiple CT scans or ultrasounds on these pa- tients. Since you state in the paper that the CT appearance of the spleen did not influence therapy, I am curious about why these studies were performed in clinically stable and at times asymptomatic pa- tients.

EARLY DISCHARGE AFTER SPLENIC TRAUMA 1407

My second question concerns one of the objectives of your paper, which was to determine the intensity of monitoring necessary with an isolated splenic injury. Because the criteria for admission to the ICU was not delineated, I suspect it was not uniform through the study period. I am concerned, therefore, that this information could be extrapolated to conclude that ICU monitoring is not justified in any patients with isolated splenic injury.

J. Langer (St Louis, MO): This is a very important study, but I am not sure that your conclusion is necessarily established by your data, since you kept children in the hospital in bed for 7 days. There's no guarantee that when you send a child home that they're going to stay in bed. Perhaps a better way to do the study would be to let children out of bed in the hospital and see whether there's any difference.

D. Collins (San Diego, CA): I've had four patients who have bled secondarily after they apparently stopped bleeding from a ruptured spleen that was not operated upon. Two occurred in the first week while they were still in the hospital and two later on after they went home. The most recent one just a few months ago was 3 weeks after his initial injury and bled severely. It was on a day that I had allowed him to go back to school at the mother's request. He was just sitting in art class with his hands folded in his lap when he got the pain.

I think it may be all right to let these children go home soon, but I only let them do it if the parents are

very reliable and live nearby. I tell them to stay in bed, whether they do or not, for a month after the injury.

J. Lynch (response): In regard to the multiple CT or ultrasound scanning, in the early part of the study there was no real protocol to study patients prior to discharge and follow-up. Now we do ultrasounds on all patients when we consider them to be stable and when we see them in follow-up. We do not follow them with CT scanning anymore. We continue CT scanning as the initial study.

As for the criteria for admission to the ICU, we now use the presence of multiple organ system injury. This may be either long bone fracture with splenic injury, multiple intraabdominal organ injury, associ- ated thoracic or head injuries, or transfusion require- ments in the emergency room or at an outlying hospital. If they have been stabilized with lactated Ringer's solution, we feel that it's probably safe to put them on the floor. We haven't yet had any untoward events.

As for the question of whether we should get these children up and out of bed and mobile in the hospital prior to discharge, it's a good question. That's what we're doing now prior to discharge without any problems so far.

As regards the patients going home early and the four noted patients who bled secondarily, we like to have our patients on bed rest at home at least for a week and then we liberalize their activities after seeing them in follow-up.