Delayed Abdominal Closure in the Management of Ruptured Abdominal Aortic Aneurysms Brant K. Oelschlager,
MD, Edward
M. Boyle, Jr., MD, Kaj Johansen, Seattle,
BACKGROUND: Resuscitative measures associated with ruptured abdominal aortic aneurysm (rAAA) repair may result in massive edema of the bowel, retroperitoneum and abdominal wall. The resulting “abdominal compartment syndrome” may compromise abdominal closure and may be associated with respiratory, renal and cardiovascular deterioration. METHODS: The medical records of 23 patients surviving initial operative repair of a rAAA were retrospectively reviewed. Eight underwent delayed abdominal closure after early approximation with silastic sheets (n = 6) or of the skin only (n = 2). Ultimate outcome, as well as several pulmonary and cardiovascular parameters, were compared with patients undergoing standard primary fascial closure (n = 15). RESULTS: A trend toward improved survival was apparent in the group undergoing delayed abdominal wall closure. Significant improvements in oxygenation and mixed venous oxygen saturation were observed in these patients, and there were fewer late deaths due to multiple organ failure. No patient undergoing delayed abdominal closure developed a graft infection. CONCLUSIONS; As in massively resuscitated trauma victims, delayed laparotomy closure in rAAA patients may confer a physiologic and survival benefit. Am J Sorg. 1997;172:411-415. 0 1997 by Excerpta Medica, Inc.
D
espite a reduction in the operative mortality associated with elective abdominal aortic aneurysm repair to 5% or less,‘,’ ruptured abdominal aortic aneurysms (rAAA) continue to pose a formidable management problem. Mortality rates continue to average 40% to 70% among such patients,3,4 with multiple organ failure (MOF) responsible for many of the deaths in those surviving initial repair. Whether due to inadequate oxygen delivery,’ tissue ischemia and reperfusion injury,6 or altered leukocyte and endothelial interactions,7 inadequate organ perfusion occurring early after rAAA is likely responsible for many of the sequelae of MOF. From
the Department
of Surgery,
Harbor-view
Medical
Center,
Seattle, Washington. Requests for reprints should be addressed to Mark H. Meissner, MD, Department of Surgery, Box 359796, Harbor-view Medical Center, 325 9th Avenue, Seattle, Washington 98195-9796. Presented at the 83rd Annual Meeting of the North Pacific Surgical Association, Seattle, Washington, November 8-9, 1996.
% 1997 by Excerpta All rights reserved.
Medica,
Inc.
MD, PhD, Mark H. Meissner,
MD,
Washington
Although comorbid conditions may be vastly different, many of the challenges encountered in management of rAAA patients are not unlike those in the multiply injured trauma victim. Among trauma patients, visceral edema with subsequent increases in intraabdominal pressure may be associated with respiratory compromise, impaired venous return leading to decreased cardiac output, and decreased end-organ perfusion resulting in renal and hepatic ischemia.‘-” Laparotomy closure may be difficult: forced closure of the fascia against tension increases the likelihood of wound ischemia, necrosis, infection and dehiscence. This “abdominal compartment syndrome” has been extensively described in the trauma literature,“,14 and early data suggest that decompressive laparotomy may benefit some of these patients.‘jmi7 In a similar fashion, massive fluid resuscitation predictably produces substantial visceral and abdominal wall edema in patients surviving repair of a rAAA. Furthermore, the retroperitoneal hematoma surrounding a rAAA may displace the viscera anteriorly, rendering abdominal closure even more difficult. Compromised pulmonary, cardiac and renal function among these elderly patients makes multiple organ failure even more predictable than in younger, healthier trauma victims. Although decompressive laparotomy or preemptive avoidance of fascial closure may benefit polytrauma victims suffering from “abdominal compartment syndrome,“i5-I7 the risk of aortic graft infection would seem to militate against delayed abdominal closure in patients following rAAA. However, we were forced to examine this hypothesis in a series of rAAA patients in whom massive edema prevented primary fascial closure. We thus undertook a retrospective review of rAAAs repaired at Harborview Medical Center (Seattle, WA) over the past 29 months.
METHODS Between January 1994 and May 1996, 23 (61%) of 38 patients presenting to Harborview Medical Center with a rAAA survived initial repair of their aneurysm. As the level I trauma center for Seattle, Harborview routinely receives all patients with suspected rAAA after paramedics have established intravenous access, instituted fluid resuscitation and established an airway as appropriate.‘s Among those surviving repair, 8 patients underwent abdominal closure with a prosthetic silastic sheet or by suturing (n = 1) or towel clipping (n = 1) the skin only. Fifteen patients had standard fascial closure. Those undergoing prosthetic abdominal closure had nonreinforced silicone sheets with open Dacron mesh borders (Bentec Medical, Sacramento, CA) interposed between the fascial margins using continuous 0 monofilament suture. The skin was left open in such 0002-9610/97/$17.00 PII SOOO2-9610(97)00081-O
411
TABLE
I Demographic,
Age Lowest SBP in field Duration SBP 5 90 mm Hg Initial hematocrit Fluid volume in field’ Operative time Lowest SBP in OR Duration OR SBP cr 90 mm Hg Lowest pH in OR Lowest temperature in OR Operative fluid volume+
Prehospital
and lntraoperative
Primary Abdominal Closure’ (n = 15)
Delayed Abdominal Closure* (n = 8)
P
82 years (76-84) 70 mm Hg (51-77) 18 min (O-44) 30 (28-33) 1000 cc (850-3500) 218 min (131-330) 81 mm Hg (68-87) 18 min (5-50) 7.27 (7.10-7.38) 33.8% (32.3-34.6) 21,450 cc (20,106-35,250)
74 years (64-80) 69 mm Hg (45-79) 25 min (O-25) 25 (18-34) 700 cc (500-l 400) 230 min (194-304) 77 mm Hg (75-90) 18 min (l-29) 7.29 (7.12-7.38) 32.5% (32.4-34.3) 27,300 cc (15,000-41,740)
0.02 0.94 0.92 0.29 0.25 0.66 0.78 0.51 0.91 0.68 1 .o
’ Median (interquartile range, 25th to 75th percenti/e). + Operative fluid voiume (cm”) = crystalloid (cc) + colloid OR, operating room. SW, systolic blood pressure.
(cc) + autotransfused
patients, with planned reexploration once the patient had stabilized and edema had diminished. The medical records of those surviving initial operation were retrospectively reviewed and data collected regarding their demographic features, prehospital care, intraoperative course, postoperative hemodynamic and respiratory parameters, associated organ failure and ultimate outcome. All patients remained intubated postoperatively and were monitored in the ICU with Swan-Ganz catheters as long as their clinical picture dictated. Detailed hemodynamic and respiratory data were thus available for the first 5 postoperative days. The following data were collected from their records: perioperative fluid requirements, urine output (UO), tidal volume (V,), positive end expiratory pressure (PEEP), peak inspiratory pressure (PIP), inspired oxygen requirements (FiOz), arterial blood gases, days intubated, pulmonary enddiastolic volume index (EDVI), systemic vascular resistance (SVR) and cardiac index (CI).
0
k
rb
2;
ia
&
is2
40
DAYS POST-O!’
Figure. Survival after ruptured abdominal aortic aneurysms (rAAA). Percent survival versus postoperative day is shown for those undergoing primary (closed circles) versus delayed (open squares) abdominal closure. Overall survival was similar; however, there were four late deaths in the primary closure group due to adult respiratory distress syndrome (ARDS). 412
THE
AMERICAN
JOURNAL
OF SURGERYD
Variables
VOLUME
173
blood
(cc) + banked
blood
products
(cc)
Outcome, hemodynamics and pulmonary variables among those undergoing silastic or skin-only closure of the abdominal wall (delayed abdominal closure) were compared with those undergoing standard fascial closure (primary abdominal closure). As most variables were not normally distributed, all data is reported as the median and interquartile range (IQR, 25th to 75th percentile). Comparisons between groups were made using Fischer’s exact test for categorical variables and the Mann-Whitney U test for continuous variables. Statistical significance was defined as a P value cO.05.
RESULTS Among the 23 patients surviving repair of a rAAA during the 29-month interval examined, 5 were women and 18 were men. Median age was 79 (IQR 73 to 84) years. Five patients required initial abdominal wall closure with a silastic sheet because massive bowel and abdominal wall edema and/or retroperitoneal hematoma prevented primary fascial suture. Two additional patients underwent closure of the skin alone, without fascial closure, for the same indication. An eighth patient developed increased intraabdominal pressure associated with elevated peak inspiratory pressures, decreased urine output and cardiovascular compromise on the second postoperative day, and underwent decompressive laparotomy and silastic closure at that time. For the purpose of analysis this patient was considered to be in the primary closure group on the first postoperative day and in the delayed abdominal closure group for the remainder of his hospital course. In comparison with patients undergoing primary abdominal closure, those undergoing delayed fascial closure were significantly (I’ = 0.02) younger. More patients were male among those undergoing primary closure (87%) than delayed abdominal closure (63%), although this difference was not significant (I’ = 0.30). There were no significant differences in other demographic, prehospital or intraoperative variables, including degree and duration of hypotension, temperature, fluid volume administered prior to admission or in the operating room, or operative time (Table I). MAY
1997
1DELAYED TABLE
CLOSURE/OELSCHLAGER
ET AL 1
II Postoperative Postoperative Primary Closure*
Tidal volume (cc/kg) FiO? (%) PEEP PIP pa% (mm Hg) paO,/FiO, ratio PaCO, (mm Hg) * Me&n
ABDOMINAL
(interquartile
9.6 70 20 60 94 151 41 range,
25th
(8.0-12.0) (50-l 00) (5-27) (47-71) (76-123) (119-196) (37-45)
Pulmonary Day
1
Delayed Closure* 9.5 80 20 61 115 195 39
Parameters
(7.0-l 1.6) (50-l 00) (5-28) (45-85) (78-l 68) (108-260) (35-49)
Postoperative P 0.43 0.73 1.0 0.77 0.45 0.39 0.88
Primary Closure* 9.3 50 7.5 49 69 134 46
(7.0-l 0.2) (45-60) (7.5-20) (40-55) (67-71) (124-153) (44-56)
Day 5 Delayed Closure*
8.0 45 15 41 87 160 52
P
(7.2-10.5) (43-70) (9-l 5) (32-52) (79-l 07) (152-204) (45-67)
0.81 0.87 0.80 0.27 0.007 0.03 0.33
to 75th percenble).
Overall mortality was 4 (50%) of 8 patients undergoing delayed fascial closure versus 11 (73%) of 15 patients undergoing primary abdominal closure (P = 0.37). Although this difference in mortality was not statistically significant, further analysis of the two groups discloses some interesting trends (Figure). Mortality on the first postoperative day was greater in the group undergoing primary abdominal closure, but equalized over the ensuing 5 days. No deaths occurred after the fifth postoperative day among patients undergoing delayed abdominal closure. In contrast, there were four late deaths from multiple organ system failure, specifically adult respiratory distress syndrome (ARDS) and respiratory failure, among patients who had undergone primary abdominal closure. On the premise that early relief of abdominal compartment syndrome should improve end-organ perfusion, pulmonary and cardiovascular parameters on days 1 and 5 were examined. Although both groups showed a trend toward decreasing oxygen and PEEP requirements over the first 5 days, there was no significant difference between the two groups with respect to FiOl, PEEP or tidal volume adjusted to body weight (cc/kg) on either day 1 or day 5 (Table II). These ventilator requirements resulted in similar oxygenation ( PaOz), carbon dioxide pressures (PaCOz) and peak inspiratory pressures in the two groups on postoperative day 1. However, despite similar ventilator requirements, oxygenation and PaOZ/FiOZ ratio were significantly better in the delayed closure group on day 5, presumably indicating a decreased incidence of acute lung injury. From a hemodynamic perspective, SVR and CI were not significantly different among those with primarily closed or open abdomens on either days 1 or 5 (Table III). Median mixed venous oxygen saturation (SvOz) tended to be better in the open-abdomen group (69% vs. 61%), although not significantly so (P = 0.19), on postoperative day 1. However, by the fifth postoperative day, median SvO, was significantly (P = 0.007) better among those patients whose abdomens had not been closed primarily (73% vs. 69%). Delayed abdominal closure thus appears to be associated with an improvement in the balance between oxygen delivery and consumption at the end of the early postoperative period. Respiratory failure was defined as ventilator dependence greater than 48 hours and renal failure as an increase in creatinine greater than 1.0 mg/dL above baseline levels, regardless of the eventual need for dialysis. Almost all of those THE AMERICAN
patients surviving longer than 48 hours postoperatively developed respiratory failure (delayed closure 1OO%, primary closure 89%). Although patients undergoing delayed abdominal closure tended to be intubated longer (median 9.5 days vs. 1.5 days), this difference was largely influenced by the high early mortality among patients undergoing primary abdominal closure. When expressed as the percentage of hospital stay during which the patient was intubated, duration of intubation was similar (P = 0.50) among those undergoing delayed (median 43%, IQR 29% to 66%) or primary abdominal closure (median 61%, IQR 26% to 99%). Although urine output tended (P = 0.07) to be better on post-operative day 1 in those with open abdomens ( 124 c&r, IQR 19 to 148) than in those whose abdomens had been closed primarily (58 cc/hr, IQR 5 to 84), it was equivalent by postoperative day 5. Three patients in each group developed renal failure (open abdomen 38%, closed abdomen ZOO/), with 2 patients in the delayed closure group and 1 in the primary closure group requiring dialysis. Median length of hospitalization was 49 (IQR 37 to 52) days among those undergoing delayed abdominal closure, versus 3 1 (IQR 19 to 53) days in those with primary closure (P = 0.25). None of the 4 delayed-closure patients dying within the first 5 postoperative day5 underwent definitive fascial closure. Among the survivors, 2 patients had their abdominal wounds closed with skin grafts 14 and 28 days postoperatively. The remaining 2 patients underwent fascial closure 6 and 10 days, after initial rAAA repair. None of the patients closed with silastic sheets or with skin only developed intraabdominal or graft infections, although 1 patient developed an enterocutaneous fistula following hospital discharge.
COMMENTS Abdominal wall closure following laparotomy serves to prevent evisceration, protect the abdominal viscera, diminish peritoneal fluid losses and prevent bacterial contamination of the peritoneal cavity. Despite these obvious advantages, primary fascial closure in patients requiring massive fluid resuscitation can cause significant problems. Closing a wound under tension may lead to tissue necrosis, wound sepsis, dehiscence and evisceration. Furthermore, the attendant increase in intraabdominal pressure may have deleterious effects upon pulmonary and cardiovascular function as well as visceral perfusion.“.‘“.‘s-“’ JOURNAL
OF
SURGERY@)
VOLUME
173
MAY
1997
413
TABLE
III Postoperative Postoperative Primary Closure*
EDVI SVR Cardiac Index svop * Median
25th
2.6 (2.1-3.3) 69 (59-73)
THE AMERICAN
Day
5
P
Primary Closure*
0.30 0.82
132 (121-138) 839 (606-933)
98 (86-l 15) 849 (382-l 158)
0.014 0.85
0.84 0.19
3.5 (2.5-4.2) 69 (66-70)
3.8 (3.1-5.1) 73 (72-78)
0.51 0.007
Delayed Closure*
P
to 75th pefcentde)
The “abdominal compartment syndrome” has been well documented, both clinically and experimentally. Richardson and Trinkle’ showed that increasing abdominal pressure causes a rise in peak inspiratory pressure and impairment of ventilation. Animal studies suggest that this is due to decreased pulmonary compliance resulting from decreased diaphragmatic excursion in the face of high abdominal pressures.” Increased intraabdominal pressure may also compromise cardiovascular function. At abdominal pressures as low as 10 to 15 mm Hg, cardiac output is compromised by a decrease in venous return resulting from elevated intrathoracic pressure and increased SVR.*,” Fietsam24 demonstrated a 25% to 80% increase in cardiac output among rAAA patients undergoing decompressive laparotomy for increased intraabdominal pressure. Increased abdominal pressure also hinders visceral blood flow, reducing perfusion of the liver, kidneys and gastrointestinal mucosa. Diebel et a125 demonstrated that intraabdominal pressures above 20 mm Hg significantly reduced both mesenteric and mucosal blood flow. The resulting ischemia-reperfusion syndrome not only causes intravascular fluid losses, but also releases the systemic inflammatory mediators that may cause or exacerbate multiple organ failure.z6 Despite the hazards of high intraabdominal pressures, maintaining the domain of the abdominal contents is of obvious importance. Initial use of a silastic sheet or the skin alone to approximate the abdominal wound solves the problem of both high intraabdominal pressures and maintenance of domain.15-17*27~2H Such silicone sheets are less adherent than most mesh products, and are therefore less likely to cause serosal reactions and damage to adjacent organs.” They also are associated with less peritoneal fluid losses than mesh products. The goal of such an approach is to return to the operating room for definitive abdominal closure once the patient has stabilized physiologically and visceral edema has decreased substantially. Is a policy of delayed abdominal closure helpful in the management of patients with rAAA? Although our observations are suggestive, the question is not definitively answered. This is a small retrospective study, with relatively few patients surviving beyond 5 days. Most patients in this study required an alternative to fascial closure during their initial aneurysm operation: only one patient had his abdominal wound reopened after developing an abdominal compartment syndrome. Although most demographic and prehospital factors were not significantly different between those undergoing delayed and primary abdominal closure, 414
Parameters Postoperative
80 (41-l 02) 939 (706-l 845)
2.6 (1.7-3.4) 61 (56-65) range,
Hemodynamic 1
Delayed Closure*
92 (79-101) 1206 (820-l 687)
(interquartile
Day
JOURNAL
OF
SURGERY@
VOLUME
173
it is likely that most patients were selected for delayed abdominal closure because they were more critically ill at completion of rAAA repair. Such a selection bias would seem to favor a poorer, not better, outcome among those undergoing delayed abdominal closure. On the other hand, the significantly younger age of patients undergoing delayed abdominal closure must be considered in interpreting our results. Despite these shortcomings, our results do suggest a decrease in very early (postoperative day 1) mortality among patients undergoing delayed abdominal closure, although such gains tend to equalize over the ensuing few days. Perhaps more important is the suggestion that late mortality due to MOF may be reduced with delayed abdominal closure. The likelihood of improved late outcome seems plausible given the findings of decreased pulmonary damage (improved P/F ratio) and improved tissue oxygenation (Sv02) that is present after early postoperative resuscitation. Notably, these are precisely the parameters that are maximized routinely in critically ill patients in the surgical intensive care unit. These trends are comparable to, and consistent with, those demonstrated in more controlled studies.20x2’ Among 6 patients undergoing prosthetic abdominal closure after rAAA repair, Akers et al’” reported immediate improvements in intraabdominal pressure, central venous pressure, peak inspiratory pressure and urine output following interval decompressive laparotomy in 2 patients. Similarly, Fietsam et a124demonstrated a dramatic difference between preand postdecompressive cardiopulmonary parameters among 4 patients developing an abdominal compartment syndrome a mean of 17 hours after rAAA repair. Our study does demonstrate that patients in whom fascial closure is deferred have a mortality rate no worse, if not lower, than those undergoing primary closure. Although a theoretical risk, intraabdominal or graft infection did not occur. In patients with substantial respiratory and hemodynamic compromise associated with abdominal closure, delayed abdominal closure seems an acceptable alternative to the morbid (and potentially lethal) consequences of abdominal compartment syndrome. Furthermore, although not directly addressed by this study, these alternative closure techniques may have some benefit in abbreviating the patient’s operative time and facilitating early transport to the intensive care unit for rewarming and correction of the acidosis and coagulopathy that inevitably develop.16 Lastly, this approach also avoids the complications of wound neMAY
1997
DELAYED
crosis and dehiscence that may follow forced closure of the fascia against tension. Derived from a retrospective observational study of an intervention rendered obligatory by clinical circumstances, our data are not definitive. However, they provide a tantalizing suggestion of a survival benefit associated with delayed abdominal closure in rAAA patients. Prospective identification of those factors that might predict the utility of delayed fascial closure after rAAA would seem to be the next investigative step.
REFERENCES 1. Kazmers A, Jacobs L, Perkins A, et al. Abdominal aortlc aneurysm repair in Veterans Affairs medical centers. ] Vnsc Surg, lY96;23:191~200. 2. Clark ET, Gewertz BL, Basiouny HS, Zarins CK. Current results of elective aortic reconstruction for aneurysmal and occlusive disease. J Curdiouasc Surg (Torino). 1990;3 1:438-441. 3. Gloviczki I’, P am) Iera P, Mucha I’ Jr, et al. Ruptured abdominal aortic aneurysms: repair should not be denied. J Va.x Swrg. 1992;15:851-859. 4. McReady RA, Slderya H, Pittman JN, et al. Ruptured aortic aneurysms ma private teachmg hoqxtal, R decade’s experience (19801989). Ann Vast Surg. 1993;7:225-228. 5. Shoemaker WC, Appel W. Kram HB, et al. Prospective trial of suprxxxmal values of surviwrs as therapeutic goals in high risk surgical patients. Chest. 1988;44:1176. 6. Zimmerman BJ, Granger DN. Keperfusion injury. 9ur~Clm i\iurth Am. 1992;72:65. 7. Koike K, Moore FA, Moore EE, et al. Gut ischemia mediates lung inlury hy xanthine oxidase dependent neutrophil mechamsms. J Scq Res. 1992:32:723. 8. Richardson JD, Trinkle JK. H emodynamic and respiratory alteratwn with increased intra-ahdommal pressure. ] Stug Res. 1976;20:401-404. 9. Kron IL, Harman PK, Nolan SP. The measurement ~)f intraabdominal pressure as a criterion for &k,minal re-explL,ration. Ann Surg. 1984;199:28-30. IO. Harsch S, Kelly KM, Benjamm J, et al. The hemodynamic effects of increased intra-abdominal m the canine model. Crit Care Med. 1986;15:423. 11. Cullen DJ, Coyle ]I’, Teplick R, Long MC. Cardiovascular, pulmonary. and renal effects of masclvely mueased intra-ahdomnxal pressure in critically ill patients. Cric Care Med. 1989; 17:l 18-12 I. 12. Obeid F, Saha A, Fath J, et al. Increases in inn-a-abdominal pressure affect pulmonary compliance. Arch Surf. 1995;1995:544-548. 13. Saxe JM. Ledgerwood AM, Lucas CE. Management elf the Jlfficult abdominal closure. Surg Clm North Am. 1993;73:243-251. 14. Bendahan J, Coetree CJ, Papagiaopoulos C, Muller R. Abdominal compartment syndrome. J Trauma. 1995;38:152-153.
THE AMERICAN
ABDOMINAL
CLOSUREIOELSCHLAGER
ET AL j
Ii. I-llrschherg A, Mattox RL. Plxnd reopratlon for severe trauma. Ann Sur,~. 1995;222:1, 2. 16. Rurch JM, Ortiz VR, Richardson KJ, et al. Abbreviated laprotomy and planned rcoperation fcor critically injured patients. Ann Surg. 1992;215:476-482. 17. Morris JA. Eddy VA, Bhnman JA, vt ,rl. The staged celiotomy for trauma: issues in unpackmg and I-cc~ltlstruction. Ann surg. 1993;217:576-586. 18. J<)hansen K. Kohler TR, Nicholls SC:, et al. Ruptured ahdominal aOrt1c aneurysm: the Harhorview cxperiencr. J b’ox Sur,y. 1991; 13:240-247. 19. Smith PC, Twcddell JS, Bessey PQ. Alternative approaches to abdominal wound closure in severely inJurcd patients with massive visceral edema. J Trauma. 1992;32:16-20. 2C. Kashtan J, Green JF, Parsons EQ, Holcrc~ft JW. Hemodynamic effects of increased nhdommal pressure. J Sur%y Rec. 1981;30:249255. 21. Barnes GE, Laine GA, Glnrn PY. Cardiova.~cular responses to elevauon of mtra-abdominal hydrostatic pressure. 4m .I Physiol. 1985;248:RZC9%R213. 22. Mutch T, L,mun WJ, Emtwe LJ, ct .)I. Ahdom~nal distension altera regional pleural pressures and cheit \vall mechanics in pigs in VIVO. J &pl Phyioi. 1991;7@:261 l-2618. 23. Ho HS, Gunther RA, and Wolfe BM. Intraperlroneal carbon dioxide insufflation and car~~iopulmon;lr\ functIoni. Arch Surg. 1992;127:928-Y ii. 24. Fletsam R, Villalha M, c‘l1 clver JL. Clark K. Intra-abdominal compartment syndrome as a complic;ltli)n of ruptured abdominal Scqy~n. 1989;55:396aortic aneurysm repair. The Amermm 402. 25. Die&l LN. l)ulchavsky SA, WI~$
JOURNAL
OF
SURGERY”
VOLUME
173
MAY
1997
415