Practice Management Guidelines for Nonoperative Management of Penetrating Abdominal Trauma

1990 - Present

Updated April 3, 2007

Article #

First Author

Year

Reference Title

Class

Comments

1

Albrecht RM

1999

Stab wounds to the back/flank in hemodynamically stable patients:  evaluation using triple-contrast computed tomography.

III

This is a retrospective study with 79 hemodynamically stable patients who were stabbed in the flank or back.  Triple contrast computed tomography was performed in 58 and was found to be 97.9% accurate in identifying significant injury with only clinical follow up, not operative.  Patients with low-risk scans and no associated injuries were discharged immediately.

Am Surg 1999;65:683-8.

2

Alzamel HA

2005

When is it safe to discharge asymptomatic patients with abdominal stab wounds?

III

This is a chart review of 650 asymptomatic patients with abdominal stab wounds who were admitted for serial examination.  Fifteen of 650 left against medical advice within 6 hours of presentation.  Sixty-eight of 635 underwent exploratory laparotomy.  All patients who needed surgery were identified within 12 hours of presentation.  Twenty-three (33%) underwent surgery within 2 hours; 26 (38%) between 2 to 4 hours; 9 (13%) between 4 and 6 hours; 9 (13%) between 6 &and 10 hours; and 1(1.4%) at 12 hours.  The authors concluded that asymptomatic patients with abdominal stab wounds may be discharged after 12 hours of observation with little likelihood of missed injury.

J Trauma 2005;58:523-5.

3

Arikan S

2005

A prospective comparison of the selective observation and routine exploration methods for penetrating abdominal stab wounds with organ or omentum evisceration.

II

This is a prospective, nonrandomized series of 52 hemodynamically stable patients with abdominal stab wounds and either visceral or omental evisceration, who were treated either with exploratory laparotomy (21) or wound exploration/closure under local anesthesia (31).  Patients with obviously perforated hollow viscera or peritonitis were excluded.  Nineteen of 52 patients had significant injuries.  Seven of 31 patients treated selectively required operation.  Of the 21 patients treated with a routine laparotomy, 33% were nontherapeutic versus 6.45% in the selective group (p < 0.05).  Of the routine laparotomy group, 19% had complications versus 3.2% in the selective group (p > 0.05).  The only complication in the selective group was a case of small bowel obstruction managed nonoperatively.  Mean length of stay was 137 hours in the routine exploration group versus 81 hours in the selective group (p < 0.001), and the mean LOS was 81 hours.  The authors concluded that selective observation is safe and superior to routine laparotomy for the treatment of penetrating abdominal stab wounds with omental evisceration.

J Trauma 2005;58:526-32.

4

Bokhari F

2004

The ultrasound screen for penetrating truncal trauma.

II

This is a prospective blinded pilot study of 49 patients with truncal stab wounds (SW) and gunshot wounds (GSW) evaluated by ultrasonic evaluation of the injured abdominal wall compared to the contralateral uninjured side.  These were compared to diagnostic peritoneal lavages performed in all penetrating injuries to the anterior abdomen or thoracoabdomen; back and flank injuries were worked up with a triple-contrast computed tomography.  A total of 58 injuries were evaluated: 37 SWs and 21 GSWs.  There were 20 true positives, 20 false positive, and 18 true negatives.  There were no false negatives.  The sensitivity and negative predictive value of ultrasound in determining clinically significant truncal visceral injury was 100%; the specificity and positive predictive value were both approximately 50%.  The authors concluded that ultrasonic exam of the injured abdominal walls in truncal penetrating trauma is an excellent screening tool. 

Am Surg 2004;70:316-21.

5

Boyle EM Jr

1997

Diagnosis of injuries after stab wounds to the back and flank.

III

This is a retrospective review of 203 patients admitted for stab wounds to the back and flank.  Group I patients were admitted prior to 1989 and had mandatory celiotomy.  Group II patients were managed selectively.  Selectively managed patients with hemodynamic instability, evisceration or acute abdominal symptoms had an IVP followed by immediate laparotomy (Group IIa).  Stable patients without obvious signs suggesting internal injury had either observation alone (IIb), a DPL (IIc), a triple contrast computed tomography scan (CT) after a negative DPL (IId), or CT alone (IIe).  The nontherapeutic laparotomy rate was 85% in Group I.  None of the 34 patients in Group IIb underwent laparotomy, with no missed injuries.  Eighteen of 32 Group IIc patients underwent laparotomy and 49% had a nontherapeutic laparotomy.  Two of 37 patients in Group IId underwent initial exploration with all procedures being therapeutic.  Two of 37 scans were falsely negative.   Two of 28 patients in Group IIe underwent initial exploration and both laparotomies were negative.  One of 28 scans was falsely negative.  The authors conclude that DPL should be the initial diagnostic study in stable patients; if the DPL is negative a triple contrast CT is indicated.  We have a concern for how a DPL will alter CT findings with fluid and air.  This study provides support at the very least for selective management since the therapeutic laparotomy rate increased from 15 to 80%.

J Trauma 1997;42:260-5.

6

Chihombori A

1991

Role of diagnostic techniques in the initial evaluation of stab wounds to the anterior abdomen, back, and flank.

III

This is a retrospective review of 162 patients with stab wounds, 103 with anterior abdominal wounds and 59 with back and flank wounds.  Seventeen of 162 patients with shock, peritonitis, and evisceration were immediately explored.  Most of the remainder underwent diagnostic peritoneal lavage (DPL).  A total of 54 patients were explored with 6 negative laparotomies.  Of 126 DPLs, none were false positive and 1 was false negative.  Of 47 computed tomographic enema scans (CTEs), 3 were interpreted as an indication for angiography.  The overall mortality was 4.3%.  DPL was considered positive if red blood cells were > 2000/mm3, white blood cells were > 500/mm3 or lavage fluid exited the urinary catheter or chest tube.  The authors concluded that their algorithm can be safely applied to patients with penetrating trauma, as 108 of 162 patients were spared laparotomy. 

J Natl Med Assoc 1991;83:137-40.

7

Chiu WC

2001

Determining the need for laparotomy in penetrating torso trauma: a prospective study using triple-contrast enhanced abdominopelvic computed tomography.

II

This is a prospective study of 75 consecutive hemodynamically stable patients with penetrating injury to the torso (lower chest, abdomen or pelvis) without definitive indication for laparotomy who underwent triple contrast computed tomography (CT) interpreted by blinded radiologists.  A positive CT scan was defined as any evidence of peritoneal violation.  In patients with a positive CT, 18 (69%) had laparotomy, two nontherapeutic and one negative.  The remainder with positive CT was successfully managed nonoperatively.  Of the patients with negative CT, 47/49 (96%) were successfully managed nonoperatively and one received a negative laparotomy.  The false-negative CT injury was a left diaphragm injury discovered at laparotomy.  CT accurately predicted whether laparotomy was needed in 71/75 (95%) patients.  The authors note that adjunctive angiography and investigation for diaphragm injury may be prudent. 

J Trauma 2001;51:860-9.

8

Chmielewski GW

1995

Nonoperative management of gunshot wounds to the abdomen.

II

This is a prospective report on 12 patients with a single gunshot wound to the right upper quadrant, stable vital signs, reliable examination, and minimal or no abdominal tenderness.  All were successfully observed.  One nontherapeutic laparotomy was done secondary to abdominal tenderness.

Am Surg 1995;61:665-8.

9

Conrad MF

2003

Selective management of penetrating truncal injuries: is emergency department discharge a reasonable goal?

III

This is a retrospective review of 107 hemodynamically patients with penetrating truncal injuries who had "selective" emergency department (ED) workup consisting of local wound exploration for stab wounds to the anterior abdomen, a triple contrast computed tomography (CT) for penetrating injuries to the back/flank, or a triple contrast CT with sigmoidoscopy and/or cystography (depending on the trajectory) for penetrating wounds to the pelvis.  Gunshot wounds to the anterior abdomen and left-sided thoracoabdominal injuries underwent operative intervention (laparoscopy or laparotomy).  Of the patients who did not receive operative intervention, 62/107 (58%) were discharged home after negative CT in the ED, 18 were managed operatively (for positive CT scan), and 27 were managed nonoperatively.  Two missed injuries were later identified (one hepatic and one small bowel injury) and managed in a delayed fashion without complications.  Follow-up was available in 66% of ED workup patients.  The authors concluded that certain patients having a negative ED workup can be safely discharged home. 

Am Surg 2003;69:266-73.

10

Demetriades D

1991

Gunshot wound of the abdomen: role of selective conservative management.

II

This is a prospective series of 41 patients with minimal or equivocal abdominal signs after a gunshot wound (GSW) to the abdomen who were observed nonoperatively.  Seven of the 41 required delayed laparotomy within 4 hours to 4 days (3 colon injuries, 3 small bowel injuries, 1 liver injury); of these, two developed wound infection, one with abdominal dehiscence.  The authors conclude that carefully selected patients with abdominal GSWs can be safely managed nonoperatively.

Br J Surg 1991;78:220-2.

11

Demetriades D

1993

Non-therapeutic operations for penetrating trauma: early morbidity and mortality.

II

This is a prospective series of 372 operations performed on 368 patients with penetrating injuries to the abdomen, chest, neck and extremities.  There were 46 negative or non-therapeutic operations.  Eleven percent of patients with nontherapeutic operations developed major complications due to anesthesia or operation (pancreatitis, aspiration pneumonia, wound infection, DVT, pneumonia).  Hospital length of stay was 4.1 days for those with uncomplicated nontherapeutic operations and 21.2 days for those with complications.  The authors concluded that nontherapeutic operations for penetrating trauma carry a significant morbidity rate and they advocate a policy of selective conservatism. 

Br J Surg 1993;80:860-1.

12

Demetriades D

1999

Gunshot injuries to the liver: the role of selective nonoperative management.

III

This is a retrospective review of gunshot wounds to the liver.  Sixteen stable patients were selected for nonoperative management.  Five patients in the observed group underwent delayed laparotomy for peritonitis (four patients with liver injuries) and abdominal compartment syndrome (one patient who had received six units of blood in violation of the recommended policy).  Except for a missed right diaphragm injury, there were no missed injuries in the 16 patients.  One patient in the group with delayed laparotomy had multiple complications from abdominal compartment syndrome, and one patient in the nonoperative group developed a biloma, which was successfully drained percutaneously.

J Am Coll Surg 1999;188:343-8.

13

Demetriades D

1997

Selective nonoperative management of gunshot wounds of the anterior abdomen.

II

This is a prospective study on gunshot wounds to the anterior abdomen using observation if the patient was stable, without peritonitis, and without severe head or spinal cord injury.  One hundred six patients were in this group, with 14 undergoing delayed operation (13 for increasing tenderness and one for continued bleeding) of which five were therapeutic. Four of these patients had colon injuries managed by primary repair.  Only one of these had a subsequent complication: a psoas abscess that required percutaneous drainage.  One patient was observed for 48 hours in violation of the protocol and developed abdominal compartment syndrome and acute respiratory distress syndrome.  The sensitivity of the initial negative physical examination was 97.1%.   The mean hospital stay in the group with nontherapeutic operations was 6.4 days, and the complication rate was 27.6%.  Of the total of 309 patients in the series, 92 (29.8%) were successfully managed nonoperatively.

Arch Surg 1997;132:178-83.

14

Easter DW

1991

A prospective, randomized comparison of computed tomography with conventional diagnostic methods in the evaluation of penetrating injuries to the back and flank.

II

Sixty-one patients with penetrating back and flank injuries were randomized into conventional testing or computed tomography (CT) testing, if they did not need immediate operation. Specificities were 96% for CT and 93% for conventional testing.  The false positive rate for laparotomy based on physical exam was 43%.  The numbers in this study are small and conventional diagnostic modalities were compared to outdated ones including cystogram and barium enema.  The conclusion is that CT may be a good way to work up patient, but there are not a lot of patients to support this.  Like the Demetriades studies, those who avoided surgery accrued lower costs.

Arch Surg 1991;126:1115-9.

15

Ertekin C

2005

Unnecessary laparotomy by using physical examination and different diagnostic modalities for penetrating abdominal stab wounds.

II

This study evaluated primarily anterior stab wounds in a prospective fashion.  There was a significantly decreased negative laparotomy rate with selective management with multiple diagnostics including local wound exploration, echocardiography, colonoscopy, computed tomography, diagnostic peritoneal lavage, diagnostic laparoscopy, and intravenous pyelogram.  Clinical examination, however, was the primary tool used to differentiate those patients requiring operation.  Seventy-nine percent were successfully managed nonoperatively, but too many adjunctive studies are included.  There was no change in mortality but a higher morbidity in delayed laparotomies (more than 8 hours later) versus early laparotomies (36% vs 27%).

Emerg Med J 2005;22:790-4.

16

Ginzburg E

1998

The role of computed tomography in selective management of gunshot wounds to the abdomen and flank.

III

This is a retrospective study of 83 patients using triple contract computed tomography (CT) to rule out injury after a gunshot wound to abdomen or flank.  CT scans were classified as positive, equivocal or negative.  Positive studies (15) received laparotomy, except for four right upper quadrant wounds treated nonoperatively.  Negative studies (53) were observed for 23 hours, with a 100% true negative rate.  Equivocal studies (15) received cavitary endoscopy (11), laparotomy (2), local wound exploration (1), or observation (1).  Only one patient (the observed patient in the equivocal group) had a missed colon injury when using this protocol, which was repaired primarily.

J Trauma 1998;45:1005-9.

17

Gonzalez RP

2001

Abdominal stab wounds: diagnostic peritoneal lavage criteria for emergency room discharge.

II

This is a prospective study of 90 patients investigating stab wounds only.  If diagnostic peritoneal lavage (DPL) showed less than 1000 red blood cells/mm3 (44 patients), patients were sent home or admitted for reasons unrelated to the abdominal wound. Thirty-four were discharged immediately and had no complications with 85% follow up.  One of four patients with evisceration required operation. This study provides good evidence for a threshold of 1000 RBC/mm3 as lower limit for DPL.  There were no complications in patients immediately discharged, and 21% (8/38) of those observed needed surgery, of which 63% (5/8) were therapeutic.

J Trauma 2001;51:939-43.

18

Grossman MD

1998

Determining anatomic injury with computed tomography in selected torso gunshot wounds

III

This is a retrospective chart review.  Computed tomography (CT) was used to look at trajectories through chest or abdomen or pelvis.  Twenty of 37 abdominopelvic CTs were negative without subsequent complication.  Eight of the remaining 17 needed therapeutic laparotomies, and the rest were observed. There were no complications or missed injuries.

J Trauma 1998;45:446-56.

19

Hasaniya N

1994

Early morbidity and mortality of non-therapeutic operations for penetrating trauma.

III

This was a retrospective study to look at complications of non-therapeutic laparotomies (n=230; 21.7%) after penetrating trauma.  The rate was significant at 8.2% and those with complications had longer stays than those without.  One patient with a major thoracic injury died secondary to complications related to a nontherapeutic laparotomy.  The conclusion is to attempt to avoid these surgeries by selective management.

Am Surg 1994;60:744-7.

20

Heyns CF

1992

Selective surgical management of renal stab wounds.

III

This is a retrospective review of 95 patients with renal stab wounds.  Patients with stab wounds and hematuria were selected for surgical exploration if they had signs of severe blood loss, an associated intra-abdominal laceration, or a major abnormality on an intravenous urogram.  Sixty patients were in the nonoperative management group, and 35 were in the operative group. Only 4 patients underwent nontherapeutic laparotomy. Complications, however, developed in 12 of the 60 patients (20%) in the nonoperative group and consisted mainly of secondary hemorrhage caused by an arteriovenous fistula or pseudoaneurysm.  Management consisted of embolization in 6, nephrectomy in 2, heminephrectomy in 1, open ligation of a fistula in 1, and spontaneous resolution in 2.  The authors concluded by stating that certain groups should be more aggressively selected for surgery, and that angioembolization may be a useful adjunct to nonoperative management.

Br J Urol 1992;69:351-7.

21

Himmelman RG

1991

Triple-contrast CT scans in penetrating back and flank trauma.

II

A negative triple contrast computed tomography scan has 100% sensitivity for retroperitoneal injury.  Eighty-eight patients were enrolled.  Five of nine high-risk scans went to surgery; two had injuries.  None of the 77 non-high risk scans had complications.

J Trauma 1991;31:852-5.