- Email: [email protected]
Delayed interventions and mortality in trauma damage control laparotomy
ARTICLE IN PRESS
Delayed interventions and mortality in trauma damage control laparotomy Margaret H. Lauerman, MD,a Joseph Dubose, MD,b Kyle Cunningham, MD,a Brandon Bruns, MD,a Matthew Bradley, MD,a Jose Diaz, MD,a Thomas Scalea, MD,a and Deborah Stein, MD, MPH,a Baltimore, MD, and Sacramento, CA
Background. In damage control laparotomy, operative principles include hemorrhage and contamination control. However, required components of initial damage control laparotomy are unknown, and nonemergency injury repair is sometimes delayed for resuscitation, angiography, or nonabdominal operations. The frequency and effects of delayed interventions are unknown. Methods. A retrospective review of patients undergoing damage control laparotomy at a single, urban trauma center was performed. Interventions initially performed at the second laparotomy were considered delayed interventions. Results. In the study, 330 damage control laparotomy patients survived to reoperation. Of all interventions, 13.9% were first performed at the second laparotomy, including 11.9% of visceral interventions and 27.2% of vascular interventions. Overall, 29.7% of patients underwent an unplanned re-exploration, and 21.8% of patients underwent re-exploration for hemorrhage control. There was no significant increase in mortality (33.3% vs 23.9%, P = .09), intra-abdominal infection (37.9% vs 28.0%; P = .10), anastomotic leak (8.0% vs 5.8%, P = .45), or enterocutaneous fistula formation (9.2% vs 9.1%, P = 1.00) with delayed interventions overall. However, mortality was increased in patients undergoing delayed vascular interventions (59.1% vs 22.8%, P = .003), unplanned re-exploration (45.9% vs 18.1%, P < .001) and re-exploration for hemorrhage control (50.0% vs 19.8%, P < .001). Conclusion. Delayed interventions are common in damage control laparotomy, with abdominal interventions often spread over multiple explorations. Mortality is increased in patients undergoing emergent re-exploration and with delayed repair of major vascular injuries. Ideal treatment of damage control laparotomy patients may include addressing injuries more completely at the first laparotomy instead of deferring care for other priorities. (Surgery 2016;j:j-j.) From the Division of Trauma and Critical Care,a R Adams Cowley Shock Trauma Center, University of Maryland, Baltimore, MD; and Division of Trauma,b Acute Care Surgery and Critical Care, University of California Davis, Sacramento, CA
OPERATIVE MANAGEMENT OF ABDOMINAL INJURIES is only a portion of trauma care. Patient management requires coordination between orthopedic surgeons, neurosurgeons, head and neck surgeons, trauma surgeons, and intensivists, all of whom have different, often competing priorities. Patients may quickly traverse many locations within the hospital, including the trauma bay, interventional radiology, operating room, and intensive care unit (ICU). Indications for damage control
Accepted for publication May 27, 2016. Reprint requests: Margaret H. Lauerman, MD, 22 South Greene St, Baltimore, MD 21201. E-mail: [email protected]. 0039-6060/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.surg.2016.05.044
laparotomy (DCL) reflect this multimodal care, expanding beyond physiologic exhaustion to include second-look laparotomy and expediting care to other procedures or radiographic imaging, at the discretion of expert opinion.1-3 DCL components include contamination and hemorrhage control; however, there are no standardized guidelines as to what specific interventions a trauma DCL should include. Fully addressing nonemergent abdominal injuries may be deferred for higher priority interventions, such as ICU resuscitation or nonabdominal operations. Some abdominal injuries cannot be treated fully in the operating room, such as those requiring angiography. The frequency and consequences of delaying interventions in DCL are unknown. We hypothesized that delayed interventions in DCL would not SURGERY 1
ARTICLE IN PRESS 2 Lauerman et al
increase mortality or rates of postoperative complications given the critically ill and complex nature of these patients. METHODS Approval was first obtained from the Institutional Review Board at the University of Maryland School of Medicine. A retrospective review was performed of trauma patients undergoing a DCL in 2003–2012. Patients undergoing DCL, operations performed at the first and second laparotomy, and indications for DCL were identified through review of the patient medical records after identification of patients from the trauma registry. Inclusion criteria included age 18 years or older and patients undergoing their initial trauma laparotomy at the R Adams Cowley Shock Trauma Center. Only patients undergoing DCL at the index operation were included. All nontrauma laparotomies were excluded. Delayed interventions were defined as interventions for injuries undergoing initial operative therapy at the second laparotomy. Complications from the first operation were excluded from delayed interventions, as were planned, 2-part interventions, such as delayed bowel anastomoses after being left in discontinuity or removal of vascular shunts placed at the previous operation. Patients were considered to have had no delayed interventions if no procedures were performed for a new injury at the second laparotomy; for example, patients undergoing liver packing at the initial laparotomy all would have undergone unpacking at the second operation, which was not considered a delayed intervention. Further hepatic interventions at the second laparotomy only were considered delayed if performed for a new injury not intervened upon at the first laparotomy. We did not investigate timing of bowel anastomosis or ostomy creation specifically, and bowel resections performed at the first laparotomy included both patients left in discontinuity and patients who underwent anastomosis creation. Vascular injuries were defined as injuries to a named vessel in the abdomen. Visceral injuries were defined as injuries to intra-abdominal organs and excluded diaphragm injuries. Interventions performed after the second laparotomy were not investigated. Primary indication for DCL was defined as the main stated reason for DCL. Physiologic exhaustion was defined as hypotension, coagulopathy, acidosis, and/or hypothermia. Need for a second look was defined as deferring fascial closure to allow for a repeat exploration, such as with
Surgery j 2016
concern for bowel ischemia. Expediting care was defined as deferring fascial closure to facilitate nonabdominal operations, angiography, radiographic tests, or a physical exam thought to be a higher priority than fascial closure. For example, a patient could undergo an exploratory laparotomy but have fascial closure deferred for a more rapid transition to angiogram for a pelvic fracture. Statistics were performed using SPSS software (version 21; IBM, Armonk, NY). Bivariate analysis was performed using independent t tests, the Mann-Whitney U test, Fisher exact test, and Pearson v2 testing. Our primary outcome was mortality, and secondary outcomes were postoperative complications. RESULTS During the 10-year study period, 391 patients underwent DCL at their initial trauma laparotomy. Sixty-one (15.6%) patients died prior to reexploration; 330 (84.4%) patients who underwent DCL survived to re-exploration and were included in this review. Overall mean age was 36.7 ± 15.88 and mean injury severity score (ISS) was 34.9 ± 14.89. Patients with delayed interventions had a significantly higher abdominal abbreviated injury scale (AIS) score (3.9 ± 1.13 vs 3.6 ± 1.27, P = .05) and were less likely to have undergone a concurrent thoracotomy (2.3% vs 12.8%, P = .003; Table I). Indications for DCL included physiologic exhaustion in 153 patients (46.4%), expediting care in 90 patients (27.3%), and second-look laparotomy in 31 patients (9.4%). Of patients undergoing DCL for physiologic exhaustion, 42 patients (27.5%) underwent a delayed intervention, and of patients undergoing DCL for a second look, 15 patients (48.4%) ultimately underwent a delayed intervention. In patients who underwent DCL to expedite care to another intervention, specific indications for expediting care included angiography in 52 patients (57.8%), nonabdominal operations in 16 patients (17.8%), urgent computed tomography in 15 patients (16.7%), and neurologic evaluation in 7 patients (7.8%). Of patients undergoing DCL for expediting care, 24 patients (26.6%) underwent an intervention at the second laparotomy. Presenting physiology and injury severity varied by DCL indication. When compared with patients with a DCL indication of physiologic exhaustion, patients with a DCL indication of second-look laparotomy had a significantly lower ISS (25.84 ± 12.95 vs 34.44 ± 14.63, P = .002) but no significant difference in admission systolic blood
ARTICLE IN PRESS Lauerman et al 3
Surgery Volume j, Number j
Table I. Demographic and operative data for patients undergoing damage control laparotomy Overall (n = 330) Demographics Age (mean ± SD) Men, N (%) Penetrating injury, N (%) Presenting SBP, mm Hg (mean ± SD) Markers of injury burden Injury severity score (mean ± SD) Thorax AIS (mean ± SD) Brain AIS (mean ± SD) Spine AIS (mean ± SD) Abdomen AIS (mean ± SD) Presenting GCS (median ± IQR) Resuscitation EBL $2 L during first 24 hr, N (%) Massive Transfusion >10 U, N (%) Massive Transfusion >50 U blood products, N (%) Adjunct procedures Concurrent thoracotomy, N (%) Concurrent angiography, N (%) Pelvis angiography, N (%) Thorax angiography, N (%) Liver angiography, N (%) Primary reason for abbreviated laparotomy Physiologic exhaustion, N (%) Expedited care, N (%) Second look, N (%) Unable to close, N (%) Unclear/other, N (%) Overall intervention Initial vascular intervention, N (%) Initial visceral intervention, N (%)
Delayed intervention (n = 87)
No delayed intervention (n = 243)
P value
36.7 ± 15.88 263 (79.7) 130 (39.4) 116.5 ± 63.23
39.2 ± 18.16 68 (78.2) 31 (35.6) 116.0 ± 28.81
35.8 ± 14.92 195 (80.2) 99 (40.7) 116.6 ± 71.69
.12 .76 .44 .91
34.9 ± 14.89 2.7 ± 1.66 1.0 ± 1.70 0.9 ± 1.33 3.6 ± 1.24 14.0 ± 7.00
36.1 ± 14.89 2.6 ± 1.60 1.1 ± 1.84 0.9 ± 1.25 3.9 ± 1.13 14.0 ± 4.00
34.5 ± 14.91 2.7 ± 1.69 1.0 ± 1.64 0.8 ± 1.36 3.6 ± 1.27 14.0 ± 9.00
.40 .83 .43 .68 .05 .34
133 (40.3) 248 (87.9) 173 (61.3)
38 (43.7) 68 (90.7) 49 (65.3)
95 (39.1) 180 (87.0) 124 (59.9)
.52 .54 .49
33 109 37 7 51
(10.0) (33.0) (11.2) (2.1) (15.5)
2 31 8 0 19
(2.3) (35.6) (9.2) (0) (21.8)
31 78 29 7 32
(12.8) (32.1) (11.9) (2.9) (13.2)
.003 .60 .56 .20 .06
153 90 31 16 40
(46.4) (27.3) (9.4) (4.8) (12.1)
42 24 15 2 4
(48.3) (27.6) (17.2) (2.3) (4.6)
111 66 16 14 36
(45.7) (27.2) (6.6) (5.8) (14.8)
.005
62 (18.7) 284 (86.1)
18 (20.7) 78 (89.7)
44 (18.1) 206 (84.8)
.60 .26
EBL, Estimated blood loss; GCS, Glasgow Coma Scale; IQR, interquartile range; SD, standard deviation; U, units.
pressure (129.77 ± 27.56 vs 116.77 ± 82.79, P = .12) or abdominal AIS (3.39 ± 1.05 vs 3.76 ± 1.20, P = .08). When comparing patients undergoing DCL for physiologic exhaustion to patients undergoing DCL for expediting care, expediting care patients had a higher ISS (38.80 ± 13.88 vs 34.44 ± 14.63, P = .02) and no significant difference in systolic blood pressure (111.14 ± 38.05 vs 116.77 ± 82.79, P = .47) or abdominal AIS (3.64 ± 1.29 vs 3.76 ± 1.20, P = .50). At the initial laparotomy, 62 patients (18.7%) underwent a vascular procedure and 284 patients (86.1%) underwent a visceral procedure. Most commonly, these comprised splenectomy in 96 patients (29.1%), small bowel resection in 85 patients (25.6%), hepatic packing in 69 patients (20.9%), and colectomy in 69 patients (20.9%; Table II). The rates of vascular and visceral procedures at the first laparotomy for those with and
without delayed interventions were not significantly different (20.7% vs 18.1%, P = .60 and 89.7% vs 84.8%, P = .26, respectively). At the second laparotomy, 87 patients (26.4%) underwent a delayed intervention, with 76 patients (23.0%) and 22 patients (6.7%) undergoing delayed visceral and vascular interventions, respectively. Incidence of delayed interventions did not change during the study period, with 42 patients (25.3%) undergoing a delayed intervention in 2008–2012 and 45 patients (27.4%) undergoing delayed intervention prior to 2008. The most common visceral procedures at the second laparotomy were colectomy in 19 patients (5.6%), small bowel resection in 10 patients (3.0%), and splenectomy in 9 patients (2.7%), and the most common vascular procedures were other retroperitoneal procedures in 9 patients (2.7%), inferior vena cava (IVC) ligation/repair in 4 patients (1.2%), superior
ARTICLE IN PRESS 4 Lauerman et al
Surgery j 2016
Table II. Abdominal visceral interventions required at the first and second laparotomy, associated mortality, and the percentage of delayed procedures Rates of specific procedures (n = 330)
Colon Colon repair, N (%) Colectomy, N (%) Colostomy, N (%) Small bowel Small bowel repair, N (%) Small bowel resection, N (%) Duodenum Duodenal repair, N (%) Stomach Gastric repair, N (%) Gastrectomy, N (%) Hepatic intervention Hepatic debridement, N (%) Hepatectomy, N (%) Hepatic packing, N (%) Hepatic cautery, N (%) Suture hepatorrhaphy, N (%) Cholecystectomy, N (%) Spleen Splenectomy, N (%) Splenorrhaphy, N (%) Pancreatectomy Pancreatectomy, N (%) Kidney Nephrectomy, N (%) Renal repair, N (%) Bladder Bladder repair, N (%)
Mortality associated with specific procedures (n = 330)
Initial laparotomy
Second laparotomy
Initial laparotomy
Second laparotomy
Percentage of procedures delayed
29/330 (8.8) 69/330 (20.9) 1/330 (0.3)
8/330 (2.4) 19/330 (5.6) 1/330 (0.3)
7/29 (24.1) 13/69 (18.8) 1/1 (100)
3/8 (37.5) 10/19 (52.6) 0/1 (0)
8/37 (21.6) 19/88 (21.6) 1/2 (50)
25/330 (7.6)
6/330 (1.8)
3/25 (12.0)
0/6 (0)
6/31 (19.4)
85/330 (25.6)
10/330 (3.0)
21/85 (24.7)
21/330 (6.4)
1/330 (0.3)
6/21 (28.6)
0/1 (0)
1/22 (4.5)
21/330 (6.4) 3/330 (0.9)
2/330 (0.6) 1/330 (0.3)
4/21 (19.0) 0/3 (0)
1/2 (50.0) 0/1 (0)
2/23 (8.7) 1/4 (25.0)
0/330 (0)
4/330 (1.2)
—
27/330 (8.2) 69/330 (20.9)
2/330 (0.6) 1/330 (0.3)
11/27 (40.7) 15/69 (21.7)
42/330 (12.7)
0/330 (0)
14/42 (33.3)
—
0/42 (0)
29/330 (8.8)
0/330 (0)
7/29 (24.1)
—
0/29 (0)
0/6 (0)
6/12 (50)
2/10 (20.0)
0/4 (0) 1/2 (50) 1/1 (100)
10/95 (10.5)
4/4 (100) 2/29 (6.9) 1/70 (1.4)
6/330 (1.8)
6/330 (1.8)
1/6 (16.7)
96/330 (29.1) 5/330 (1.5)
9/330 (2.7) 0/330 (0)
31/96 (32.3) 2/5 (40.0)
4/9 (44.4) —
17/330 (5.2)
4/330 (1.2)
3/17 (17.6)
2/4 (50.0)
4/21 (19.0)
27/330 (8.2) 6/330 (1.8)
3/330 (0.9) 0/330 (0)
9/27 (33.3) 0/6 (0)
0/3 (0) —
3/30 (10.0) 0/6 (0)
14/330 (4.2)
3/330 (0.9)
4/14 (28.6)
1/3 (33.3)
3/17 (17.6)
mesenteric artery (SMA) repair/thrombectomy in 3 patients (0.9%), and superior mesenteric vein (SMV) repair in 3 patients (0.9%). The second laparotomy often was performed in an emergent fashion, with 98 patients (29.7%) undergoing an unplanned return to the operating room. Seventy-two patients (21.8%) underwent reexploration for hemorrhage control. Patients who underwent delayed interventions were more likely to undergo an unplanned re-exploration, with 33 patients (37.9%) with a delayed intervention
9/105 (8.5) 0/5 (0)
undergoing an unplanned re-exploration compared with 65 patients (26.7%) without a delayed intervention (P = .06). Similarly, 26 patients (29.9%) with a delayed intervention underwent re-exploration for hemorrhage control compared with 46 patients (18.9%) without a delayed intervention (P = .05). Rates of emergent re-exploration and reexploration for hemorrhage control varied between patients with delayed visceral and delayed vascular interventions. Twenty-five patients
ARTICLE IN PRESS Lauerman et al 5
Surgery Volume j, Number j
Table III. Abdominal vascular interventions required at the first and second laparotomy, associated mortality, and the percentage of delayed interventions Rates of specific procedures (n = 330)
Retroperitoneal Aortic repair, N (%) IVC ligation/repair, N (%) Other retroperitoneal N (%) Mesenteric SMA repair/thrombectomy, N (%) SMV repair, N (%) IMA ligation, N (%) IMV ligation/repair, N (%) Visceral Hepatic vein ligation, N (%) Portal vein repair, N (%) Hepatic artery repair, N (%) Splenic vein ligation, N (%) Splenic artery repair/ligation, N (%) Renal vein repair/ligation, N (%) Adrenal vein ligation, N (%) Pelvic Iliac vein repair/ligation, N (%) Iliac artery ligation/repair/shunt, N (%)
Mortality associated with specific procedures (n = 330)
Initial laparotomy
Second laparotomy
Initial laparotomy
Second laparotomy
Percentage of procedures delayed
3/330 (0.9) 18/330 (5.5) 4/330 (1.2)
— 4/330 (1.2) 9/330 (2.7)
2/3 (66.7) 3/18 (16.7) 1/4 (25.0)
— 2/4 (50.0) 5/9 (55.6)
0/3 (0) 4/22 (18.2) 6/9 (66.7)
1/330 9/330 1/330 2/330
(0.3) (2.7) (0.3) (0.6)
3/330 (0.9) 3/330 (0.9) — —
0/1 4/9 0/1 2/2
(0) (44.4) (0) (100)
2/3 (66.7) 3/3 (100) — —
3/4 3/12 0/1 0/2
(75.0) (25.0) (0) (0)
1/330 (0.3) 3/330 (0.9) — — 1/330 (0.3) 5/330 (1.5) 1/330 (0.3)
2/330 (0.6) — 1/330 (0.3) 2/330 (0.6) 1/330 (0.3) — 1/330 (0.3)
0/1 (0) 0/3 (0) — — 1/1 (100) 1/5 (20.0) 0/1 (0)
1/2 (50.0) — 0/1 (0) 1/2 (50.0) 1/1 (100) — 1/1 (100)
2/3 0/3 1/1 2/2 1/2 0/5 1/2
(66.7) (0) (100) (100) (50.0) (0) (50.0)
20/330 (6.1) 9/330 (2.7)
2/330 (0.6) —
3/20 (15.0) 2/9 (22.2)
2/2 (100) —
2/22 (9.1) 0/9 (0)
IMA, Inferior mesenteric artery; IMV, inferior mesenteric vein.
(32.9%) with a delayed visceral intervention underwent an emergent re-exploration compared with 73 patients (28.7%) without a delayed visceral intervention, and 18 patients (23.7%) with a delayed visceral intervention underwent reexploration for hemorrhage control compared with 54 patients (21.3%) without a delayed visceral intervention (P = .48 and P = .64, respectively). Conversely, 17 patients (77.3%) with a delayed vascular intervention underwent emergent reexploration compared with 81 patients (26.3%) without a delayed vascular intervention, and 16 patients (72.7%) with a delayed vascular intervention underwent a re-exploration for hemorrhage control compared with 56 patients (18.2%) without a delayed vascular intervention (P < .001 for both). A considerable portion of the total interventions performed in DCL patients were undertaken at the second laparotomy. Of the 775 total interventions, 108 (13.9%) were delayed, with 28 (27.2%) of vascular interventions and 80 (11.9%) of visceral interventions performed at the second laparotomy. For specific procedures, 10.5% of small bowel resections, 21.6% of colectomies, 19.0% of pancreatectomies, 8.5% of
splenectomies, 17.6% of bladder repairs, and 10.0% of nephrectomies were undertaken at the second laparotomy. Delayed management of vascular injuries occurred as well, with 18.2% of IVC, 75.0% of SMA, and 25.0% of SMV interventions performed at the second operation (Table III). When patients with delayed interventions and those without were compared overall, there was no significant difference in mortality (33.3% vs 23.9%, P = .09), intra-abdominal infection (37.9% vs 28.0%; P = .10), anastomotic leak (8.0% vs 5.8%, P = .45), septic shock (11.5% vs 9.9%, P = .68), or enterocutaneous fistula formation (9.2% vs 9.1%, P = 1.00). Achievement of delayed primary fascial closure was similar (62.9% vs 61.3%, P = .88; Table IV). Overall median duration of stay and ICU median duration of stay was not increased significantly in patients with delayed interventions (22.0 ± 27.40 days vs 24.1 ± 23.40 days, P = .79 and 13.8 ± 20.00 days vs 12.3 ± 19.00 days, P = .38, respectively). Even when excluding the 78 patients (23.6%) with a brain AIS $3, there was no significant difference in overall mortality, acute renal failure, anastomotic leak,
ARTICLE IN PRESS 6 Lauerman et al
Surgery j 2016
Table IV. Outcomes in patients undergoing damage control laparotomy Delayed intervention (n = 87) Acute renal failure, N (%) Anastomotic leak, N (%) ARDS, N (%) Fistula, N (%) Intra-abdominal infection, N (%) Pneumonia, N (%) Pulmonary embolism, N (%) Septic shock, N (%) Fascial closure, N (%) Mortality, N (%)
25 7 11 8 33 12 5 10 39 29
(28.7) (8.0) (12.6) (9.2) (37.9) (13.8) (5.7) (11.5) (62.9) (33.3)
No delayed intervention (n = 243) 49 14 18 22 68 33 15 24 125 58
(20.2) (5.8) (7.4) (9.1) (28.0) (13.6) (6.2) (9.9) (61.3) (23.9)
P value .10 .45 .18 1.00 .10 1.00 1.00 .68 .88 .09
ARDS, Acute respiratory distress syndrome.
acute respiratory distress syndrome (ARDS), enterocutaneous fistula, intra-abdominal infection, pneumonia, septic shock, or delayed primary fascial closure (data not shown). However, when examining the subset of 79 patients with vascular injuries, 13 patients undergoing a delayed vascular intervention died (59.1%) compared with 13 patients (22.8%) not undergoing a delayed vascular intervention (P = .003). When examining only the 291 patients with visceral injuries, 24 patients (31.6%) undergoing a delayed visceral intervention died compared with 49 patients (22.8%) without a delayed visceral intervention (P = .165). Mortality also was increased with unplanned reexplorations. Forty-five patients (45.9%) undergoing an unplanned re-exploration died compared with 42 patients (18.1%) not undergoing an unplanned re-exploration (P < .001). Mortality was similarly increased in patients undergoing reexploration for hemorrhage control, as 36 patients (50.0%) undergoing re-exploration for hemorrhage control died compared with 51 patients (19.8%) who did not undergo re-exploration for hemorrhage control (P < .001). When examining the subset of patients with visceral injuries only, 39 patients (44.8%) undergoing an unplanned re-exploration died, compared with 34 patients (16.7%) not undergoing an unplanned reexploration (P < .001). For patients with vascular injuries only, 17 patients (42.5%) undergoing an emergent re-exploration died compared with 9 patients (23.1%) not undergoing an emergent reexploration (P = .094). DISCUSSION Despite multiple decades of experience with DCL, exact components of the initial laparotomy remain unknown. The decision to terminate the
initial laparotomy has depended on the burden of intra-abdominal and extra-abdominal injuries, as well as presence of hypotension, coagulopathy, acidosis, and hypothermia. The decision may be different for each patient. Missed injuries in abdominal trauma are an uncommon occurrence, with a rate of 1.3–7%,4-6 although these rates do not reflect use of DCL. Intentionally delayed interventions, such as vascular shunts and delayed bowel anastomosis, occur frequently in trauma; the rate of vascular shunt use in this review (3.6%) is comparable with the 2.6% reported in the literature.7,8 Prior to this review, the frequency of delayed interventions overall in DCL, and how abdominal interventions were distributed, was unknown. The rate of delayed intervention (13.9%) in DCL is a significant number. Rates of many specific, delayed interventions seem high, such as an 8.5% rate of delayed splenectomy, 21.6% rate of delayed colon repair and colectomy, and 19.4% rate of delayed small bowel repair. This is the first report of interventions at the second laparotomy of which we are aware, so previous rates are not available for comparison. It cannot fully be determined retrospectively whether injuries treated in a delayed fashion were delayed intentionally from the index operation, were missed at the index operation due to the overall injury burden, were angiographic complications, or represent progression of disease; interventions at the second laparotomy likely represent a combination of these causes. Additionally, mortality for some injuries such as splenectomy at 32.3%, small bowel resection at 24.7%, and colectomy at 18.8% are substantial and are seemingly out of proportion to their severity. It is likely for many injuries that another, more severe injury or physiologic derangement was the true
ARTICLE IN PRESS Surgery Volume j, Number j
cause of the mortality. Calculating mortality associated with specific interventions is confounded given concurrent injuries. It also is difficult to determine the relevance of specific injuries retrospectively amid the overall injury burden. However, delayed interventions seem to have a substantial clinical effect, as patients were frequently re-explored in an emergent fashion. This was especially true for patients with delayed vascular interventions given the high rate of re-exploration for hemorrhage. Whether the injury was missed, the injury progressed, or the intervention was delayed intentionally, the decision that led to a delayed intervention may be incorrect, and a more comprehensive first laparotomy may be a more optimal technique. Mortality associated with delayed vascular injuries was increased significantly, as was mortality with re-exploration in an emergent fashion. Similarly, while mortality was not increased for patients with visceral injuries undergoing delayed interventions, patients with visceral injuries re-explored emergently had substantially increased mortality. The mortality with some specific, delayed interventions was also substantial, especially for major vascular injuries, such as the IVC, SMA, and SMV, and for the colon, stomach, spleen, and pancreas. Our institution does not have standardized guidelines for use of DCL, and the decision to address an injury at the index laparotomy is up to the discretion of the operating surgeon. Similarly, the decision to utilize DCL, proceed to angiography, pursue an extra-abdominal operation, or prioritize ICU resuscitation is done in conjunction with our trauma anesthesiologists and consulting surgical services. Physiology, pressor requirements, transfusion requirements, oxygenation, and laboratory values (if available) are components of this decision. Damage control resuscitation has increased in use over time, with an emphasis on blood product transfusion over massive crystalloid resuscitation. Crystalloid administration is associated with inability to achieve delayed primary fascial closure, likely due to visceral edema.9 Despite the advent of damage control resuscitation, there was no substantial change in delayed interventions over time. Rates of intra-abdominal infection or enterocutaneous fistula formation, which are of high concern in DCL patients, were not significantly increased by delayed intervention.10,11 Delay in management of intra-abdominal pathology overall did not decrease the rate of achieving delayed primary fascial closure. Fascial closure is dependent
Lauerman et al 7
on many factors including timing of reoperation, number of re-explorations, resuscitation amounts and type, and closure strategy.9,11-18 A delayed intervention is only one factor in achieving fascial closure and may not be as important as the previously mentioned factors. Need for a delayed intervention also may be a marker for a more severely injured patient, beyond what is measured by ISS and AIS, and mortality associated with delayed interventions may thus be biased. Surgeons likely decide to proceed with earlier resuscitation in patients with qualitatively worse injuries or physiology, thus placing patients with a higher risk of mortality at a greater likelihood of undergoing a delayed intervention. This is difficult to account for retrospectively. Of note, of patients who underwent DCL for a second-look laparotomy, 48.4% ultimately had an intervention at the second laparotomy. While it is difficult to validate specific indications for DCL, this high rate of delayed interventions seems to support utilization of second-look laparotomy at the discretion of the surgeon. Limitations of this review include inability to categorize injury severity retrospectively and to determine cause of mortality. We did not examine the timing of bowel anastomosis or anastomotic technique. Limitations of this study also include that resuscitation patterns were not included, which can affect outcomes in DCL, as resuscitation patterns have been previously studied by our group.9 Laboratory and physiologic parameters were not included as well. Previous studies of DCL have reported a mean ISS in the range of 18 35.10,11,19-22 Our ISS mean is on the higher end of this range, which may limit generalizability of our results. Additionally, surgeon experience (both seniority of the attending and seniority of the participating fellows or residents) and variability in surgeon practice patterns were not included, both of which may influence intraoperative decision-making to proceed with a DCL and to delay procedures until re-exploration. Care of critically injured patients requires prioritizing many different injuries and resuscitation needs. In patients undergoing DCL, abdominal interventions frequently are spread over multiple laparotomies, with a sizable number of abdominal operations performed at the second laparotomy. While delayed injury care does not increase morbidity or mortality overall, mortality is increased in patients undergoing delayed vascular interventions and in patients re-explored emergently. Optimal care may include more completely addressing injuries during the index laparotomy instead of
ARTICLE IN PRESS 8 Lauerman et al
deferring care for other priorities. Future prospective work will aid in delineating the exact components of damage control operation.
Surgery j 2016
12.
REFERENCES 1. Stone HH, Strom PR, Mullins RJ. Management of the major coagulopathy with onset during laparotomy. Ann Surg 1983;197:532-5. 2. Rotondo MF, Schwab CW, McGonigal MD, Phillips GR 3rd, Fruchterman TM, Kauder DR, et al. ‘Damage control’: an approach for improved survival in exsanguinating penetrating abdominal injury. J Trauma 1993;35:375-82; discussion 382-3. 3. Diaz JJ Jr, Cullinane DC, Dutton WD, Jerome R, Bagdonas R, Bilaniuk JW, et al. The management of the open abdomen in trauma and emergency general surgery: part 1-damage control. J Trauma 2010;68:1425-38. 4. Schnuriger B, Lam L, Inaba K, Kobayashi L, Barbarino R, Demetriades D. Negative laparotomy in trauma: are we getting better? Am Surg 2012;78:1219-23. 5. Velmahos GC, Demetriades D, Toutouzas KG, Sarkisyan G, Chan LS, Ishak R, et al. Selective nonoperative management in 1,856 patients with abdominal gunshot wounds: should routine laparotomy still be the standard of care? Ann Surg 2001;234:395-402; discussion 402-3. 6. Sorkey AJ, Farnell MB, Williams HJ Jr, Mucha P Jr, Ilstrup DM. The complementary roles of diagnostic peritoneal lavage and computed tomography in the evaluation of blunt abdominal trauma. Surgery 1989;106:794-800; discussion 800-1. 7. Anjaria DJ, Ullmann TM, Lavery R, Livingston DH. Management of colonic injuries in the setting of damage-control laparotomy: one shot to get it right. J Trauma Acute Care Surg 2014;76:594-8; discussion 598-600. 8. DuBose JJ, Savage SA, Fabian TC, Menaker J, Scalea T, Holcomb JB, et al. The American Association for the Surgery of Trauma PROspective Observational Vascular Injury Treatment (PROOVIT) registry: multicenter data on modern vascular injury diagnosis, management, and outcomes. J Trauma Acute Care Surg 2015;78:215-22; discussion 222–3. 9. Bradley M, Galvagno S, Dhanda A, Rodriguez C, Lauerman M, DuBose J, et al. Damage control resuscitation protocol and the management of open abdomens in trauma patients. Am Surg 2014;80:768-75. 10. Miller RS, Morris JA Jr, Diaz JJ Jr, Herring MB, May AK. Complications after 344 damage-control open celiotomies. J Trauma 2005;59:1365-71; discussion 1371-4. 11. Dubose JJ, Scalea TM, Holcomb JB, Shrestha B, Okoye O, Inaba K, et al. Open abdominal management after damage-control laparotomy for trauma: a prospective observational American Association for the Surgery of Trauma
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
multicenter study. J Trauma Acute Care Surg 2013;74:11320; discussion 1120-2. Tieu BH, Cho SD, Luem N, Riha G, Mayberry J, Schreiber MA. The use of the Wittmann Patch facilitates a high rate of fascial closure in severely injured trauma patients and critically ill emergency surgery patients. J Trauma 2008;65:865-70. Harvin JA, Mims MM, Duchesne JC, Cox CS Jr, Wade CE, Holcomb JB, et al. Chasing 100%: the use of hypertonic saline to improve early, primary fascial closure after damage control laparotomy. J Trauma Acute Care Surg 2013;74: 426-30; discussion 431-2. Pommerening MJ, DuBose JJ, Zielinski MD, Phelan HA, Scalea TM, Inaba K, et al. Time to first take-back operation predicts successful primary fascial closure in patients undergoing damage control laparotomy. Surgery 2014; 156:431-8. Burlew CC, Moore EE, Biffl WL, Bensard DD, Johnson JL, Barnett CC. One hundred percent fascial approximation can be achieved in the postinjury open abdomen with a sequential closure protocol. J Trauma Acute Care Surg 2012;72:235-41. Cothren CC, Moore EE, Johnson JL, Moore JB, Burch JM. One hundred percent fascial approximation with sequential abdominal closure of the open abdomen. Am J Surg 2006;192:238-42. Garner GB, Ware DN, Cocanour CS, Duke JH, McKinley BA, Kozar RA, et al. Vacuum-assisted wound closure provides early fascial reapproximation in trauma patients with open abdomens. Am J Surg 2001;182:630-8. Miller PR, Thompson JT, Faler BJ, Meredith JW, Chang MC. Late fascial closure in lieu of ventral hernia: the next step in open abdomen management. J Trauma 2002;53:843-9. Bradley MJ, Dubose JJ, Scalea TM, Holcomb JB, Shrestha B, Okoye O, et al. Independent predictors of enteric fistula and abdominal sepsis after damage control laparotomy: results from the prospective AAST Open Abdomen registry. JAMA Surg 2013;148:947-54. Miller PR, Meredith JW, Johnson JC, Chang MC. Prospective evaluation of vacuum-assisted fascial closure after open abdomen: planned ventral hernia rate is substantially reduced. Ann Surg 2004;239:608-14; discussion 614-6. Ott MM, Norris PR, Diaz JJ, Collier BR, Jenkins JM, Gunter OL, et al. Colon anastomosis after damage control laparotomy: recommendations from 174 trauma colectomies. J Trauma 2011;70:595-602. Burlew CC, Moore EE, Cuschieri J, Jurkovich GJ, Codner P, Nirula R, et al. Who should we feed? Western Trauma Association multi-institutional study of enteral nutrition in the open abdomen after injury. J Trauma Acute Care Surg 2012;73:1380-7; discussion 1387-8.
Delayed interventions and mortality in trauma damage control laparotomy Margaret H. Lauerman, MD,a Joseph Dubose, MD,b Kyle Cunningham, MD,a Brandon Bruns, MD,a Matthew Bradley, MD,a Jose Diaz, MD,a Thomas Scalea, MD,a and Deborah Stein, MD, MPH,a Baltimore, MD, and Sacramento, CA
Background. In damage control laparotomy, operative principles include hemorrhage and contamination control. However, required components of initial damage control laparotomy are unknown, and nonemergency injury repair is sometimes delayed for resuscitation, angiography, or nonabdominal operations. The frequency and effects of delayed interventions are unknown. Methods. A retrospective review of patients undergoing damage control laparotomy at a single, urban trauma center was performed. Interventions initially performed at the second laparotomy were considered delayed interventions. Results. In the study, 330 damage control laparotomy patients survived to reoperation. Of all interventions, 13.9% were first performed at the second laparotomy, including 11.9% of visceral interventions and 27.2% of vascular interventions. Overall, 29.7% of patients underwent an unplanned re-exploration, and 21.8% of patients underwent re-exploration for hemorrhage control. There was no significant increase in mortality (33.3% vs 23.9%, P = .09), intra-abdominal infection (37.9% vs 28.0%; P = .10), anastomotic leak (8.0% vs 5.8%, P = .45), or enterocutaneous fistula formation (9.2% vs 9.1%, P = 1.00) with delayed interventions overall. However, mortality was increased in patients undergoing delayed vascular interventions (59.1% vs 22.8%, P = .003), unplanned re-exploration (45.9% vs 18.1%, P < .001) and re-exploration for hemorrhage control (50.0% vs 19.8%, P < .001). Conclusion. Delayed interventions are common in damage control laparotomy, with abdominal interventions often spread over multiple explorations. Mortality is increased in patients undergoing emergent re-exploration and with delayed repair of major vascular injuries. Ideal treatment of damage control laparotomy patients may include addressing injuries more completely at the first laparotomy instead of deferring care for other priorities. (Surgery 2016;j:j-j.) From the Division of Trauma and Critical Care,a R Adams Cowley Shock Trauma Center, University of Maryland, Baltimore, MD; and Division of Trauma,b Acute Care Surgery and Critical Care, University of California Davis, Sacramento, CA
OPERATIVE MANAGEMENT OF ABDOMINAL INJURIES is only a portion of trauma care. Patient management requires coordination between orthopedic surgeons, neurosurgeons, head and neck surgeons, trauma surgeons, and intensivists, all of whom have different, often competing priorities. Patients may quickly traverse many locations within the hospital, including the trauma bay, interventional radiology, operating room, and intensive care unit (ICU). Indications for damage control
Accepted for publication May 27, 2016. Reprint requests: Margaret H. Lauerman, MD, 22 South Greene St, Baltimore, MD 21201. E-mail: [email protected]. 0039-6060/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.surg.2016.05.044
laparotomy (DCL) reflect this multimodal care, expanding beyond physiologic exhaustion to include second-look laparotomy and expediting care to other procedures or radiographic imaging, at the discretion of expert opinion.1-3 DCL components include contamination and hemorrhage control; however, there are no standardized guidelines as to what specific interventions a trauma DCL should include. Fully addressing nonemergent abdominal injuries may be deferred for higher priority interventions, such as ICU resuscitation or nonabdominal operations. Some abdominal injuries cannot be treated fully in the operating room, such as those requiring angiography. The frequency and consequences of delaying interventions in DCL are unknown. We hypothesized that delayed interventions in DCL would not SURGERY 1
ARTICLE IN PRESS 2 Lauerman et al
increase mortality or rates of postoperative complications given the critically ill and complex nature of these patients. METHODS Approval was first obtained from the Institutional Review Board at the University of Maryland School of Medicine. A retrospective review was performed of trauma patients undergoing a DCL in 2003–2012. Patients undergoing DCL, operations performed at the first and second laparotomy, and indications for DCL were identified through review of the patient medical records after identification of patients from the trauma registry. Inclusion criteria included age 18 years or older and patients undergoing their initial trauma laparotomy at the R Adams Cowley Shock Trauma Center. Only patients undergoing DCL at the index operation were included. All nontrauma laparotomies were excluded. Delayed interventions were defined as interventions for injuries undergoing initial operative therapy at the second laparotomy. Complications from the first operation were excluded from delayed interventions, as were planned, 2-part interventions, such as delayed bowel anastomoses after being left in discontinuity or removal of vascular shunts placed at the previous operation. Patients were considered to have had no delayed interventions if no procedures were performed for a new injury at the second laparotomy; for example, patients undergoing liver packing at the initial laparotomy all would have undergone unpacking at the second operation, which was not considered a delayed intervention. Further hepatic interventions at the second laparotomy only were considered delayed if performed for a new injury not intervened upon at the first laparotomy. We did not investigate timing of bowel anastomosis or ostomy creation specifically, and bowel resections performed at the first laparotomy included both patients left in discontinuity and patients who underwent anastomosis creation. Vascular injuries were defined as injuries to a named vessel in the abdomen. Visceral injuries were defined as injuries to intra-abdominal organs and excluded diaphragm injuries. Interventions performed after the second laparotomy were not investigated. Primary indication for DCL was defined as the main stated reason for DCL. Physiologic exhaustion was defined as hypotension, coagulopathy, acidosis, and/or hypothermia. Need for a second look was defined as deferring fascial closure to allow for a repeat exploration, such as with
Surgery j 2016
concern for bowel ischemia. Expediting care was defined as deferring fascial closure to facilitate nonabdominal operations, angiography, radiographic tests, or a physical exam thought to be a higher priority than fascial closure. For example, a patient could undergo an exploratory laparotomy but have fascial closure deferred for a more rapid transition to angiogram for a pelvic fracture. Statistics were performed using SPSS software (version 21; IBM, Armonk, NY). Bivariate analysis was performed using independent t tests, the Mann-Whitney U test, Fisher exact test, and Pearson v2 testing. Our primary outcome was mortality, and secondary outcomes were postoperative complications. RESULTS During the 10-year study period, 391 patients underwent DCL at their initial trauma laparotomy. Sixty-one (15.6%) patients died prior to reexploration; 330 (84.4%) patients who underwent DCL survived to re-exploration and were included in this review. Overall mean age was 36.7 ± 15.88 and mean injury severity score (ISS) was 34.9 ± 14.89. Patients with delayed interventions had a significantly higher abdominal abbreviated injury scale (AIS) score (3.9 ± 1.13 vs 3.6 ± 1.27, P = .05) and were less likely to have undergone a concurrent thoracotomy (2.3% vs 12.8%, P = .003; Table I). Indications for DCL included physiologic exhaustion in 153 patients (46.4%), expediting care in 90 patients (27.3%), and second-look laparotomy in 31 patients (9.4%). Of patients undergoing DCL for physiologic exhaustion, 42 patients (27.5%) underwent a delayed intervention, and of patients undergoing DCL for a second look, 15 patients (48.4%) ultimately underwent a delayed intervention. In patients who underwent DCL to expedite care to another intervention, specific indications for expediting care included angiography in 52 patients (57.8%), nonabdominal operations in 16 patients (17.8%), urgent computed tomography in 15 patients (16.7%), and neurologic evaluation in 7 patients (7.8%). Of patients undergoing DCL for expediting care, 24 patients (26.6%) underwent an intervention at the second laparotomy. Presenting physiology and injury severity varied by DCL indication. When compared with patients with a DCL indication of physiologic exhaustion, patients with a DCL indication of second-look laparotomy had a significantly lower ISS (25.84 ± 12.95 vs 34.44 ± 14.63, P = .002) but no significant difference in admission systolic blood
ARTICLE IN PRESS Lauerman et al 3
Surgery Volume j, Number j
Table I. Demographic and operative data for patients undergoing damage control laparotomy Overall (n = 330) Demographics Age (mean ± SD) Men, N (%) Penetrating injury, N (%) Presenting SBP, mm Hg (mean ± SD) Markers of injury burden Injury severity score (mean ± SD) Thorax AIS (mean ± SD) Brain AIS (mean ± SD) Spine AIS (mean ± SD) Abdomen AIS (mean ± SD) Presenting GCS (median ± IQR) Resuscitation EBL $2 L during first 24 hr, N (%) Massive Transfusion >10 U, N (%) Massive Transfusion >50 U blood products, N (%) Adjunct procedures Concurrent thoracotomy, N (%) Concurrent angiography, N (%) Pelvis angiography, N (%) Thorax angiography, N (%) Liver angiography, N (%) Primary reason for abbreviated laparotomy Physiologic exhaustion, N (%) Expedited care, N (%) Second look, N (%) Unable to close, N (%) Unclear/other, N (%) Overall intervention Initial vascular intervention, N (%) Initial visceral intervention, N (%)
Delayed intervention (n = 87)
No delayed intervention (n = 243)
P value
36.7 ± 15.88 263 (79.7) 130 (39.4) 116.5 ± 63.23
39.2 ± 18.16 68 (78.2) 31 (35.6) 116.0 ± 28.81
35.8 ± 14.92 195 (80.2) 99 (40.7) 116.6 ± 71.69
.12 .76 .44 .91
34.9 ± 14.89 2.7 ± 1.66 1.0 ± 1.70 0.9 ± 1.33 3.6 ± 1.24 14.0 ± 7.00
36.1 ± 14.89 2.6 ± 1.60 1.1 ± 1.84 0.9 ± 1.25 3.9 ± 1.13 14.0 ± 4.00
34.5 ± 14.91 2.7 ± 1.69 1.0 ± 1.64 0.8 ± 1.36 3.6 ± 1.27 14.0 ± 9.00
.40 .83 .43 .68 .05 .34
133 (40.3) 248 (87.9) 173 (61.3)
38 (43.7) 68 (90.7) 49 (65.3)
95 (39.1) 180 (87.0) 124 (59.9)
.52 .54 .49
33 109 37 7 51
(10.0) (33.0) (11.2) (2.1) (15.5)
2 31 8 0 19
(2.3) (35.6) (9.2) (0) (21.8)
31 78 29 7 32
(12.8) (32.1) (11.9) (2.9) (13.2)
.003 .60 .56 .20 .06
153 90 31 16 40
(46.4) (27.3) (9.4) (4.8) (12.1)
42 24 15 2 4
(48.3) (27.6) (17.2) (2.3) (4.6)
111 66 16 14 36
(45.7) (27.2) (6.6) (5.8) (14.8)
.005
62 (18.7) 284 (86.1)
18 (20.7) 78 (89.7)
44 (18.1) 206 (84.8)
.60 .26
EBL, Estimated blood loss; GCS, Glasgow Coma Scale; IQR, interquartile range; SD, standard deviation; U, units.
pressure (129.77 ± 27.56 vs 116.77 ± 82.79, P = .12) or abdominal AIS (3.39 ± 1.05 vs 3.76 ± 1.20, P = .08). When comparing patients undergoing DCL for physiologic exhaustion to patients undergoing DCL for expediting care, expediting care patients had a higher ISS (38.80 ± 13.88 vs 34.44 ± 14.63, P = .02) and no significant difference in systolic blood pressure (111.14 ± 38.05 vs 116.77 ± 82.79, P = .47) or abdominal AIS (3.64 ± 1.29 vs 3.76 ± 1.20, P = .50). At the initial laparotomy, 62 patients (18.7%) underwent a vascular procedure and 284 patients (86.1%) underwent a visceral procedure. Most commonly, these comprised splenectomy in 96 patients (29.1%), small bowel resection in 85 patients (25.6%), hepatic packing in 69 patients (20.9%), and colectomy in 69 patients (20.9%; Table II). The rates of vascular and visceral procedures at the first laparotomy for those with and
without delayed interventions were not significantly different (20.7% vs 18.1%, P = .60 and 89.7% vs 84.8%, P = .26, respectively). At the second laparotomy, 87 patients (26.4%) underwent a delayed intervention, with 76 patients (23.0%) and 22 patients (6.7%) undergoing delayed visceral and vascular interventions, respectively. Incidence of delayed interventions did not change during the study period, with 42 patients (25.3%) undergoing a delayed intervention in 2008–2012 and 45 patients (27.4%) undergoing delayed intervention prior to 2008. The most common visceral procedures at the second laparotomy were colectomy in 19 patients (5.6%), small bowel resection in 10 patients (3.0%), and splenectomy in 9 patients (2.7%), and the most common vascular procedures were other retroperitoneal procedures in 9 patients (2.7%), inferior vena cava (IVC) ligation/repair in 4 patients (1.2%), superior
ARTICLE IN PRESS 4 Lauerman et al
Surgery j 2016
Table II. Abdominal visceral interventions required at the first and second laparotomy, associated mortality, and the percentage of delayed procedures Rates of specific procedures (n = 330)
Colon Colon repair, N (%) Colectomy, N (%) Colostomy, N (%) Small bowel Small bowel repair, N (%) Small bowel resection, N (%) Duodenum Duodenal repair, N (%) Stomach Gastric repair, N (%) Gastrectomy, N (%) Hepatic intervention Hepatic debridement, N (%) Hepatectomy, N (%) Hepatic packing, N (%) Hepatic cautery, N (%) Suture hepatorrhaphy, N (%) Cholecystectomy, N (%) Spleen Splenectomy, N (%) Splenorrhaphy, N (%) Pancreatectomy Pancreatectomy, N (%) Kidney Nephrectomy, N (%) Renal repair, N (%) Bladder Bladder repair, N (%)
Mortality associated with specific procedures (n = 330)
Initial laparotomy
Second laparotomy
Initial laparotomy
Second laparotomy
Percentage of procedures delayed
29/330 (8.8) 69/330 (20.9) 1/330 (0.3)
8/330 (2.4) 19/330 (5.6) 1/330 (0.3)
7/29 (24.1) 13/69 (18.8) 1/1 (100)
3/8 (37.5) 10/19 (52.6) 0/1 (0)
8/37 (21.6) 19/88 (21.6) 1/2 (50)
25/330 (7.6)
6/330 (1.8)
3/25 (12.0)
0/6 (0)
6/31 (19.4)
85/330 (25.6)
10/330 (3.0)
21/85 (24.7)
21/330 (6.4)
1/330 (0.3)
6/21 (28.6)
0/1 (0)
1/22 (4.5)
21/330 (6.4) 3/330 (0.9)
2/330 (0.6) 1/330 (0.3)
4/21 (19.0) 0/3 (0)
1/2 (50.0) 0/1 (0)
2/23 (8.7) 1/4 (25.0)
0/330 (0)
4/330 (1.2)
—
27/330 (8.2) 69/330 (20.9)
2/330 (0.6) 1/330 (0.3)
11/27 (40.7) 15/69 (21.7)
42/330 (12.7)
0/330 (0)
14/42 (33.3)
—
0/42 (0)
29/330 (8.8)
0/330 (0)
7/29 (24.1)
—
0/29 (0)
0/6 (0)
6/12 (50)
2/10 (20.0)
0/4 (0) 1/2 (50) 1/1 (100)
10/95 (10.5)
4/4 (100) 2/29 (6.9) 1/70 (1.4)
6/330 (1.8)
6/330 (1.8)
1/6 (16.7)
96/330 (29.1) 5/330 (1.5)
9/330 (2.7) 0/330 (0)
31/96 (32.3) 2/5 (40.0)
4/9 (44.4) —
17/330 (5.2)
4/330 (1.2)
3/17 (17.6)
2/4 (50.0)
4/21 (19.0)
27/330 (8.2) 6/330 (1.8)
3/330 (0.9) 0/330 (0)
9/27 (33.3) 0/6 (0)
0/3 (0) —
3/30 (10.0) 0/6 (0)
14/330 (4.2)
3/330 (0.9)
4/14 (28.6)
1/3 (33.3)
3/17 (17.6)
mesenteric artery (SMA) repair/thrombectomy in 3 patients (0.9%), and superior mesenteric vein (SMV) repair in 3 patients (0.9%). The second laparotomy often was performed in an emergent fashion, with 98 patients (29.7%) undergoing an unplanned return to the operating room. Seventy-two patients (21.8%) underwent reexploration for hemorrhage control. Patients who underwent delayed interventions were more likely to undergo an unplanned re-exploration, with 33 patients (37.9%) with a delayed intervention
9/105 (8.5) 0/5 (0)
undergoing an unplanned re-exploration compared with 65 patients (26.7%) without a delayed intervention (P = .06). Similarly, 26 patients (29.9%) with a delayed intervention underwent re-exploration for hemorrhage control compared with 46 patients (18.9%) without a delayed intervention (P = .05). Rates of emergent re-exploration and reexploration for hemorrhage control varied between patients with delayed visceral and delayed vascular interventions. Twenty-five patients
ARTICLE IN PRESS Lauerman et al 5
Surgery Volume j, Number j
Table III. Abdominal vascular interventions required at the first and second laparotomy, associated mortality, and the percentage of delayed interventions Rates of specific procedures (n = 330)
Retroperitoneal Aortic repair, N (%) IVC ligation/repair, N (%) Other retroperitoneal N (%) Mesenteric SMA repair/thrombectomy, N (%) SMV repair, N (%) IMA ligation, N (%) IMV ligation/repair, N (%) Visceral Hepatic vein ligation, N (%) Portal vein repair, N (%) Hepatic artery repair, N (%) Splenic vein ligation, N (%) Splenic artery repair/ligation, N (%) Renal vein repair/ligation, N (%) Adrenal vein ligation, N (%) Pelvic Iliac vein repair/ligation, N (%) Iliac artery ligation/repair/shunt, N (%)
Mortality associated with specific procedures (n = 330)
Initial laparotomy
Second laparotomy
Initial laparotomy
Second laparotomy
Percentage of procedures delayed
3/330 (0.9) 18/330 (5.5) 4/330 (1.2)
— 4/330 (1.2) 9/330 (2.7)
2/3 (66.7) 3/18 (16.7) 1/4 (25.0)
— 2/4 (50.0) 5/9 (55.6)
0/3 (0) 4/22 (18.2) 6/9 (66.7)
1/330 9/330 1/330 2/330
(0.3) (2.7) (0.3) (0.6)
3/330 (0.9) 3/330 (0.9) — —
0/1 4/9 0/1 2/2
(0) (44.4) (0) (100)
2/3 (66.7) 3/3 (100) — —
3/4 3/12 0/1 0/2
(75.0) (25.0) (0) (0)
1/330 (0.3) 3/330 (0.9) — — 1/330 (0.3) 5/330 (1.5) 1/330 (0.3)
2/330 (0.6) — 1/330 (0.3) 2/330 (0.6) 1/330 (0.3) — 1/330 (0.3)
0/1 (0) 0/3 (0) — — 1/1 (100) 1/5 (20.0) 0/1 (0)
1/2 (50.0) — 0/1 (0) 1/2 (50.0) 1/1 (100) — 1/1 (100)
2/3 0/3 1/1 2/2 1/2 0/5 1/2
(66.7) (0) (100) (100) (50.0) (0) (50.0)
20/330 (6.1) 9/330 (2.7)
2/330 (0.6) —
3/20 (15.0) 2/9 (22.2)
2/2 (100) —
2/22 (9.1) 0/9 (0)
IMA, Inferior mesenteric artery; IMV, inferior mesenteric vein.
(32.9%) with a delayed visceral intervention underwent an emergent re-exploration compared with 73 patients (28.7%) without a delayed visceral intervention, and 18 patients (23.7%) with a delayed visceral intervention underwent reexploration for hemorrhage control compared with 54 patients (21.3%) without a delayed visceral intervention (P = .48 and P = .64, respectively). Conversely, 17 patients (77.3%) with a delayed vascular intervention underwent emergent reexploration compared with 81 patients (26.3%) without a delayed vascular intervention, and 16 patients (72.7%) with a delayed vascular intervention underwent a re-exploration for hemorrhage control compared with 56 patients (18.2%) without a delayed vascular intervention (P < .001 for both). A considerable portion of the total interventions performed in DCL patients were undertaken at the second laparotomy. Of the 775 total interventions, 108 (13.9%) were delayed, with 28 (27.2%) of vascular interventions and 80 (11.9%) of visceral interventions performed at the second laparotomy. For specific procedures, 10.5% of small bowel resections, 21.6% of colectomies, 19.0% of pancreatectomies, 8.5% of
splenectomies, 17.6% of bladder repairs, and 10.0% of nephrectomies were undertaken at the second laparotomy. Delayed management of vascular injuries occurred as well, with 18.2% of IVC, 75.0% of SMA, and 25.0% of SMV interventions performed at the second operation (Table III). When patients with delayed interventions and those without were compared overall, there was no significant difference in mortality (33.3% vs 23.9%, P = .09), intra-abdominal infection (37.9% vs 28.0%; P = .10), anastomotic leak (8.0% vs 5.8%, P = .45), septic shock (11.5% vs 9.9%, P = .68), or enterocutaneous fistula formation (9.2% vs 9.1%, P = 1.00). Achievement of delayed primary fascial closure was similar (62.9% vs 61.3%, P = .88; Table IV). Overall median duration of stay and ICU median duration of stay was not increased significantly in patients with delayed interventions (22.0 ± 27.40 days vs 24.1 ± 23.40 days, P = .79 and 13.8 ± 20.00 days vs 12.3 ± 19.00 days, P = .38, respectively). Even when excluding the 78 patients (23.6%) with a brain AIS $3, there was no significant difference in overall mortality, acute renal failure, anastomotic leak,
ARTICLE IN PRESS 6 Lauerman et al
Surgery j 2016
Table IV. Outcomes in patients undergoing damage control laparotomy Delayed intervention (n = 87) Acute renal failure, N (%) Anastomotic leak, N (%) ARDS, N (%) Fistula, N (%) Intra-abdominal infection, N (%) Pneumonia, N (%) Pulmonary embolism, N (%) Septic shock, N (%) Fascial closure, N (%) Mortality, N (%)
25 7 11 8 33 12 5 10 39 29
(28.7) (8.0) (12.6) (9.2) (37.9) (13.8) (5.7) (11.5) (62.9) (33.3)
No delayed intervention (n = 243) 49 14 18 22 68 33 15 24 125 58
(20.2) (5.8) (7.4) (9.1) (28.0) (13.6) (6.2) (9.9) (61.3) (23.9)
P value .10 .45 .18 1.00 .10 1.00 1.00 .68 .88 .09
ARDS, Acute respiratory distress syndrome.
acute respiratory distress syndrome (ARDS), enterocutaneous fistula, intra-abdominal infection, pneumonia, septic shock, or delayed primary fascial closure (data not shown). However, when examining the subset of 79 patients with vascular injuries, 13 patients undergoing a delayed vascular intervention died (59.1%) compared with 13 patients (22.8%) not undergoing a delayed vascular intervention (P = .003). When examining only the 291 patients with visceral injuries, 24 patients (31.6%) undergoing a delayed visceral intervention died compared with 49 patients (22.8%) without a delayed visceral intervention (P = .165). Mortality also was increased with unplanned reexplorations. Forty-five patients (45.9%) undergoing an unplanned re-exploration died compared with 42 patients (18.1%) not undergoing an unplanned re-exploration (P < .001). Mortality was similarly increased in patients undergoing reexploration for hemorrhage control, as 36 patients (50.0%) undergoing re-exploration for hemorrhage control died compared with 51 patients (19.8%) who did not undergo re-exploration for hemorrhage control (P < .001). When examining the subset of patients with visceral injuries only, 39 patients (44.8%) undergoing an unplanned re-exploration died, compared with 34 patients (16.7%) not undergoing an unplanned reexploration (P < .001). For patients with vascular injuries only, 17 patients (42.5%) undergoing an emergent re-exploration died compared with 9 patients (23.1%) not undergoing an emergent reexploration (P = .094). DISCUSSION Despite multiple decades of experience with DCL, exact components of the initial laparotomy remain unknown. The decision to terminate the
initial laparotomy has depended on the burden of intra-abdominal and extra-abdominal injuries, as well as presence of hypotension, coagulopathy, acidosis, and hypothermia. The decision may be different for each patient. Missed injuries in abdominal trauma are an uncommon occurrence, with a rate of 1.3–7%,4-6 although these rates do not reflect use of DCL. Intentionally delayed interventions, such as vascular shunts and delayed bowel anastomosis, occur frequently in trauma; the rate of vascular shunt use in this review (3.6%) is comparable with the 2.6% reported in the literature.7,8 Prior to this review, the frequency of delayed interventions overall in DCL, and how abdominal interventions were distributed, was unknown. The rate of delayed intervention (13.9%) in DCL is a significant number. Rates of many specific, delayed interventions seem high, such as an 8.5% rate of delayed splenectomy, 21.6% rate of delayed colon repair and colectomy, and 19.4% rate of delayed small bowel repair. This is the first report of interventions at the second laparotomy of which we are aware, so previous rates are not available for comparison. It cannot fully be determined retrospectively whether injuries treated in a delayed fashion were delayed intentionally from the index operation, were missed at the index operation due to the overall injury burden, were angiographic complications, or represent progression of disease; interventions at the second laparotomy likely represent a combination of these causes. Additionally, mortality for some injuries such as splenectomy at 32.3%, small bowel resection at 24.7%, and colectomy at 18.8% are substantial and are seemingly out of proportion to their severity. It is likely for many injuries that another, more severe injury or physiologic derangement was the true
ARTICLE IN PRESS Surgery Volume j, Number j
cause of the mortality. Calculating mortality associated with specific interventions is confounded given concurrent injuries. It also is difficult to determine the relevance of specific injuries retrospectively amid the overall injury burden. However, delayed interventions seem to have a substantial clinical effect, as patients were frequently re-explored in an emergent fashion. This was especially true for patients with delayed vascular interventions given the high rate of re-exploration for hemorrhage. Whether the injury was missed, the injury progressed, or the intervention was delayed intentionally, the decision that led to a delayed intervention may be incorrect, and a more comprehensive first laparotomy may be a more optimal technique. Mortality associated with delayed vascular injuries was increased significantly, as was mortality with re-exploration in an emergent fashion. Similarly, while mortality was not increased for patients with visceral injuries undergoing delayed interventions, patients with visceral injuries re-explored emergently had substantially increased mortality. The mortality with some specific, delayed interventions was also substantial, especially for major vascular injuries, such as the IVC, SMA, and SMV, and for the colon, stomach, spleen, and pancreas. Our institution does not have standardized guidelines for use of DCL, and the decision to address an injury at the index laparotomy is up to the discretion of the operating surgeon. Similarly, the decision to utilize DCL, proceed to angiography, pursue an extra-abdominal operation, or prioritize ICU resuscitation is done in conjunction with our trauma anesthesiologists and consulting surgical services. Physiology, pressor requirements, transfusion requirements, oxygenation, and laboratory values (if available) are components of this decision. Damage control resuscitation has increased in use over time, with an emphasis on blood product transfusion over massive crystalloid resuscitation. Crystalloid administration is associated with inability to achieve delayed primary fascial closure, likely due to visceral edema.9 Despite the advent of damage control resuscitation, there was no substantial change in delayed interventions over time. Rates of intra-abdominal infection or enterocutaneous fistula formation, which are of high concern in DCL patients, were not significantly increased by delayed intervention.10,11 Delay in management of intra-abdominal pathology overall did not decrease the rate of achieving delayed primary fascial closure. Fascial closure is dependent
Lauerman et al 7
on many factors including timing of reoperation, number of re-explorations, resuscitation amounts and type, and closure strategy.9,11-18 A delayed intervention is only one factor in achieving fascial closure and may not be as important as the previously mentioned factors. Need for a delayed intervention also may be a marker for a more severely injured patient, beyond what is measured by ISS and AIS, and mortality associated with delayed interventions may thus be biased. Surgeons likely decide to proceed with earlier resuscitation in patients with qualitatively worse injuries or physiology, thus placing patients with a higher risk of mortality at a greater likelihood of undergoing a delayed intervention. This is difficult to account for retrospectively. Of note, of patients who underwent DCL for a second-look laparotomy, 48.4% ultimately had an intervention at the second laparotomy. While it is difficult to validate specific indications for DCL, this high rate of delayed interventions seems to support utilization of second-look laparotomy at the discretion of the surgeon. Limitations of this review include inability to categorize injury severity retrospectively and to determine cause of mortality. We did not examine the timing of bowel anastomosis or anastomotic technique. Limitations of this study also include that resuscitation patterns were not included, which can affect outcomes in DCL, as resuscitation patterns have been previously studied by our group.9 Laboratory and physiologic parameters were not included as well. Previous studies of DCL have reported a mean ISS in the range of 18 35.10,11,19-22 Our ISS mean is on the higher end of this range, which may limit generalizability of our results. Additionally, surgeon experience (both seniority of the attending and seniority of the participating fellows or residents) and variability in surgeon practice patterns were not included, both of which may influence intraoperative decision-making to proceed with a DCL and to delay procedures until re-exploration. Care of critically injured patients requires prioritizing many different injuries and resuscitation needs. In patients undergoing DCL, abdominal interventions frequently are spread over multiple laparotomies, with a sizable number of abdominal operations performed at the second laparotomy. While delayed injury care does not increase morbidity or mortality overall, mortality is increased in patients undergoing delayed vascular interventions and in patients re-explored emergently. Optimal care may include more completely addressing injuries during the index laparotomy instead of
ARTICLE IN PRESS 8 Lauerman et al
deferring care for other priorities. Future prospective work will aid in delineating the exact components of damage control operation.
Surgery j 2016
12.
REFERENCES 1. Stone HH, Strom PR, Mullins RJ. Management of the major coagulopathy with onset during laparotomy. Ann Surg 1983;197:532-5. 2. Rotondo MF, Schwab CW, McGonigal MD, Phillips GR 3rd, Fruchterman TM, Kauder DR, et al. ‘Damage control’: an approach for improved survival in exsanguinating penetrating abdominal injury. J Trauma 1993;35:375-82; discussion 382-3. 3. Diaz JJ Jr, Cullinane DC, Dutton WD, Jerome R, Bagdonas R, Bilaniuk JW, et al. The management of the open abdomen in trauma and emergency general surgery: part 1-damage control. J Trauma 2010;68:1425-38. 4. Schnuriger B, Lam L, Inaba K, Kobayashi L, Barbarino R, Demetriades D. Negative laparotomy in trauma: are we getting better? Am Surg 2012;78:1219-23. 5. Velmahos GC, Demetriades D, Toutouzas KG, Sarkisyan G, Chan LS, Ishak R, et al. Selective nonoperative management in 1,856 patients with abdominal gunshot wounds: should routine laparotomy still be the standard of care? Ann Surg 2001;234:395-402; discussion 402-3. 6. Sorkey AJ, Farnell MB, Williams HJ Jr, Mucha P Jr, Ilstrup DM. The complementary roles of diagnostic peritoneal lavage and computed tomography in the evaluation of blunt abdominal trauma. Surgery 1989;106:794-800; discussion 800-1. 7. Anjaria DJ, Ullmann TM, Lavery R, Livingston DH. Management of colonic injuries in the setting of damage-control laparotomy: one shot to get it right. J Trauma Acute Care Surg 2014;76:594-8; discussion 598-600. 8. DuBose JJ, Savage SA, Fabian TC, Menaker J, Scalea T, Holcomb JB, et al. The American Association for the Surgery of Trauma PROspective Observational Vascular Injury Treatment (PROOVIT) registry: multicenter data on modern vascular injury diagnosis, management, and outcomes. J Trauma Acute Care Surg 2015;78:215-22; discussion 222–3. 9. Bradley M, Galvagno S, Dhanda A, Rodriguez C, Lauerman M, DuBose J, et al. Damage control resuscitation protocol and the management of open abdomens in trauma patients. Am Surg 2014;80:768-75. 10. Miller RS, Morris JA Jr, Diaz JJ Jr, Herring MB, May AK. Complications after 344 damage-control open celiotomies. J Trauma 2005;59:1365-71; discussion 1371-4. 11. Dubose JJ, Scalea TM, Holcomb JB, Shrestha B, Okoye O, Inaba K, et al. Open abdominal management after damage-control laparotomy for trauma: a prospective observational American Association for the Surgery of Trauma
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
multicenter study. J Trauma Acute Care Surg 2013;74:11320; discussion 1120-2. Tieu BH, Cho SD, Luem N, Riha G, Mayberry J, Schreiber MA. The use of the Wittmann Patch facilitates a high rate of fascial closure in severely injured trauma patients and critically ill emergency surgery patients. J Trauma 2008;65:865-70. Harvin JA, Mims MM, Duchesne JC, Cox CS Jr, Wade CE, Holcomb JB, et al. Chasing 100%: the use of hypertonic saline to improve early, primary fascial closure after damage control laparotomy. J Trauma Acute Care Surg 2013;74: 426-30; discussion 431-2. Pommerening MJ, DuBose JJ, Zielinski MD, Phelan HA, Scalea TM, Inaba K, et al. Time to first take-back operation predicts successful primary fascial closure in patients undergoing damage control laparotomy. Surgery 2014; 156:431-8. Burlew CC, Moore EE, Biffl WL, Bensard DD, Johnson JL, Barnett CC. One hundred percent fascial approximation can be achieved in the postinjury open abdomen with a sequential closure protocol. J Trauma Acute Care Surg 2012;72:235-41. Cothren CC, Moore EE, Johnson JL, Moore JB, Burch JM. One hundred percent fascial approximation with sequential abdominal closure of the open abdomen. Am J Surg 2006;192:238-42. Garner GB, Ware DN, Cocanour CS, Duke JH, McKinley BA, Kozar RA, et al. Vacuum-assisted wound closure provides early fascial reapproximation in trauma patients with open abdomens. Am J Surg 2001;182:630-8. Miller PR, Thompson JT, Faler BJ, Meredith JW, Chang MC. Late fascial closure in lieu of ventral hernia: the next step in open abdomen management. J Trauma 2002;53:843-9. Bradley MJ, Dubose JJ, Scalea TM, Holcomb JB, Shrestha B, Okoye O, et al. Independent predictors of enteric fistula and abdominal sepsis after damage control laparotomy: results from the prospective AAST Open Abdomen registry. JAMA Surg 2013;148:947-54. Miller PR, Meredith JW, Johnson JC, Chang MC. Prospective evaluation of vacuum-assisted fascial closure after open abdomen: planned ventral hernia rate is substantially reduced. Ann Surg 2004;239:608-14; discussion 614-6. Ott MM, Norris PR, Diaz JJ, Collier BR, Jenkins JM, Gunter OL, et al. Colon anastomosis after damage control laparotomy: recommendations from 174 trauma colectomies. J Trauma 2011;70:595-602. Burlew CC, Moore EE, Cuschieri J, Jurkovich GJ, Codner P, Nirula R, et al. Who should we feed? Western Trauma Association multi-institutional study of enteral nutrition in the open abdomen after injury. J Trauma Acute Care Surg 2012;73:1380-7; discussion 1387-8.