Professional Documents
Culture Documents
BACKGROUND: Optimal management following index laparotomy is poorly defined in secondary peritonitis patients. Although “open abdomen”
(OA), or temporary abdominal closure with planned relaparotomy, is used to reassess bowel viability or severity of contamination,
recent studies demonstrate comparable morbidity and mortality with primary abdominal closure (PC). This study evaluates differences
between OA and PC following emergent laparotomy.
METHODS: Using the Premier database at a quaternary care center (2012–2016), nontrauma patients with secondary peritonitis requiring
emergent laparotomy were identified (N = 534). Propensity matching for PC (n = 331; 62%) or OA (n = 203; 38%) was performed
using variables: Mannheim Peritonitis Index, lactate, and vasopressor requirement. One hundred eleven closely matched pairs (PC:
OA) were compared.
RESULTS: Five hundred thirty-four patients (55.0% female; mean age, 59.6 ± 15.5 years) underwent emergent laparotomy. Of the OA patients,
136 (67.0%) had one relaparotomy, while 67 (33.0%) underwent multiple reoperations. Compared to daytime cases, laparotomies
performed overnight (6 pm–6 am) had more temporary closures with OA (42.8% OA vs. 57.2% PC, p = 0.04). When assessing by
surgeon type, PC was performed in 78.7% of laparotomies by surgical subspecialties compared to 56.7% (p < 0.0001) of
acute care surgeons. After propensity matching, OA patients had increased postoperative complications (71.2% vs. 41.4%,
p < 0.0001), mortality (22.5% vs. 11.7%, p = 0.006), and longer median length of stay (13 vs. 9 days, p = 0.0001).
CONCLUSION: Open abdomen was performed in 38.0% of patients, with one-third of those requiring multiple reoperations. Complications, mortality
rates, and costs associated with OAwere significantly increased when compared to PC. Given these findings, future studies are needed
to determine appropriate indications for OA. (J Trauma Acute Care Surg. 2019;87: 623–629. Copyright © 2019 Wolters Kluwer
Health, Inc. All rights reserved.)
LEVEL OF EVIDENCE: Therapeutic/care management, level IV.
KEY WORDS: Open abdomen; vacuum-assisted closure; on-demand laparotomy; emergency; peritonitis.
(PC), others have demonstrated reduced in-hospital mortality took place at Carolinas Medical Center in Charlotte, NC,
using OA with VAC after adjusting for Acute Physiologic Assess- which serves as a regional quaternary referral center for North
ment and Chronic Health Evaluation (APACHE) scores.12–14 In and South Carolina. The study group included all patients
addition to disease severity, the initial operative approach often with secondary peritonitis who underwent an emergent
depends on the source and degree of contamination, patient's laparotomy between 2012 and 2016. Secondary peritonitis
physiologic stability, and surgeon/institution comfort using was defined as active or impending peritoneal infection
OA.14 Few studies have evaluated patient and surgical conditions secondary to intraabdominal or intrapelvic source of hollow
to determine when and why surgeons may be more inclined to viscus perforation, bowel ischemia/necrosis, or nonbacterial
temporarily close the abdomen. This paucity of data contributes peritonitis.15 Patients demonstrating physical examination
to nonstandard protocols used throughout multiple medical cen- findings consistent with localized or generalized peritonitis
ters and indicates a need for robust study. and diagnoses including closed loop obstruction, volvulus,
The efficacy and optimal scenarios for use of OA are dif- and strangulated or internal hernias were also selected for
ficult to assess owing to significant patient heterogeneity, physi- inclusion given the natural progression to bowel ischemia
cian comfort with OA, and lack of strict indications to guide the and perforation in the absence of surgical treatment. Patients
unfamiliar surgeon. This institutional analysis of patients under- were identified from the database based on International
going emergent laparotomy seeks to evaluate morbidity and Classification of Diseases (ICD), Ninth and Tenth versions,
mortality associated with OA management, using propensity- and Clinical Modification (ICD-9-CM and ICD-10-CM).
matching adjustment for inherent selection differences associ- Patients with specific diagnoses (Fig. 1) who underwent an
ated with each treatment group. Based on literature review, we emergent laparotomy were selected from the database.
hypothesize that the use of OA for surgical management of Emergent laparotomy was defined as an operative procedure
suspected secondary peritonitis is associated with higher in- performed within 48 hours of surgery consultation. Patient’s
hospital complication and mortality rates when compared with records were reviewed by two independent sources (A.K.
PC. Here, we aim to describe surgeon use of OA at a quaternary and L.C.) to confirm the emergent nature of the procedure.
care center including patient demographics, clinical indications, Exclusion criteria included age younger than 18 years, index
and subsequent outcomes. This study adds much-needed data laparotomy at an outside institution, and patients with
regarding use of OA in nontrauma patients. diagnosis of acute or necrotizing pancreatitis given its
propensity for reoperative management.
METHODS
Data Collection
Study Design Data were collected retrospectively from electronic medi-
Following institutional board review approval, an external cal records and operative notes. Complications and cost data
administrative database was retrospectively queried. The study were obtained using the Premier database. Patients were
Figure 1. International Classification of Diseases, Ninth Revision / International Classification of Diseases, Ninth Revision diagnosis codes.
separated into two treatment groups based on approach at index reported as number and frequency, mean with standard deviation,
laparotomy (primary abdominal closure and open abdomen), or median with interquartile range if non-normally distributed.
with the clinical definitions as follows:
Open abdomen (OA)—At the termination of the index Statistical Analysis
laparotomy, the fascia is left open with an Abthera VAC device Descriptive statistics and univariate analysis were performed
placed intra-abdominally and connected to negative pressure to describe patients’ demographics and operative variables. Both
suction with plan for repeat laparotomy before ultimate closure. unplanned relaparotomy and death occurring within 48 hours of
Patients were subsequently returned to the operating room for initial laparotomy occurred in each treatment group; therefore,
reassessment of the peritoneal cavity and all other indicated pro- comparisons were made using an intention-to-treat analysis based
cedures. Timing of relaparotomy was based on clinical stability on the operating surgeon's planned approach at the index laparot-
and the discretion of the managing surgical team, typically oc- omy. Differences in the study population by method of abdominal
curring within 24 to 48 hours. Based on relaparotomy findings, closure during index laparotomy were reported using the χ2 test for
patients either underwent fascial closure or repeat OA with sub- comparing the proportions and the Wilcoxon rank-sum test for
sequent laparotomies as necessary. comparing the medians, with statistical significance set at two-
Primary abdominal closure (PC)—During the index lapa- tailed p < 0.05. Propensity score nearest neighbor matching
rotomy, patient's fascia is reapproximated primarily as definitive was performed for PC or OA based on logistic regression model,
closure and additional reoperation is performed on-demand generating 1:1 closely matched pairs for comparison. Con-
based on clinical suspicion by the surgical team. Indications founding variables used for propensity matching included pre-
for reoperation included clinical deterioration or failure of im- operative MPI score, vasopressor requirements during index
provement due to suspected intraabdominal cause. laparotomy, and lactate level of 3 or greater. Selection of
Perioperative data were recorded and included patients’ propensity-matching variables was based on statistical signifi-
demographics, clinical presentation, surgical indication, postop- cance as well as clinical applicability and relevance to intraoper-
erative complications, and hospital cost. Indication for emergent ative decision to close the abdomen during initial laparotomy.
laparotomy as defined in this study was based on retrospective While other variables also reached statistical significance on
review of intraoperative findings in conjunction with operative univariate analysis, many were institution specific and therefore
notes and preoperative variables, while surgeon preoperative were not included to limit the influence of factors outside the
indication for emergent laparotomy was based on physical surgeon's control. The combination of MPI score, lactate level,
examination findings and clinical signs/symptoms of peritonitis. and vasopressor requirement is not only accessible in the preop-
Surgical indications were categorized into obstruction, ischemia/ erative and intraoperative setting, but also captures the severity
necrosis, contamination/sepsis, and perforation. Patients with of disease and patient’s physiology and thus can be readily eval-
presence of multiple findings, such as closed loop obstruction uated by any surgeon to aid in decision making. Multivariate re-
leading to perforation, were categorized into the latter group ac- gression analysis was also performed on the propensity-matched
cording to their most emergent indication for surgery. Disease se- cohort to evaluate the strength of association between other po-
verity was measured using the Mannheim Peritonitis Index (MPI) tentially confounding variables that also reached statistical sig-
and the APACHE-II score. The MPI includes patient's age, pres- nificance on univariate analysis, including surgical indication
ence of malignancy, duration of preoperative symptoms, source of (perforation, contamination/sepsis, ischemia/necrosis, and
sepsis, quality of intraabdominal exudate, and chronic organ fail- obstruction/other), patient’s point of origin (health care facility
ure.16 The APACHE-II score integrates patient's age, physiologic vs. home), and operative length (minutes). All data were ana-
parameters, and laboratory data within the first 24 hours of inten- lyzed using Statistical Analysis Software, version 9.4 (SAS In-
sive care unit admission.17 Operative details included time of day stitute, Inc., Cary, NC).
surgery performed, operative duration, source and etiology of
peritonitis, and date of abdominal closure. Intraoperative details RESULTS
collected from anesthesia records also included presence of hypo-
tension (defined as mean arterial pressure <65) and need for Five hundred thirty-four patients [55.0% female with median
blood transfusions or vasopressors. In patients managed with age of 61 years (interquartile range, 50–71 years)] underwent
OA, the number of subsequent operations (after index laparot- emergent laparotomy for secondary peritonitis. Following the
omy) and the procedures performed (i.e., bowel resection, index laparotomy, 203 patients (38.0%) were managed with
ileostomy creation) during relaparotomies were recorded. When OA, while the remaining 331 patients (62.0%) underwent PC.
available from the operative note, the reason for OA technique Of the OA patients, 136 (67.0%) underwent fascial closure dur-
was also noted. Postoperative data were obtained from the Pre- ing the first relaparotomy (second operative procedure), while
mier database and included total hospital length of stay (LOS), 67 (33.0%) had multiple subsequent planned reoperations
hospital billed charges (costs), in-hospital complications, in- (mean, 1.6 ± 1.5).
hospital mortality, discharge destination, and 30-day readmission. Patients’ demographics and surgical characteristics of
Reoperation performed after abdominal fascial closure was de- each group are shown in Table 1. On unmatched univariate anal-
fined as an unplanned laparotomy. ysis, OA patients were older (median age, 63 vs. 59 years,
Primary outcomes included in-hospital complications and p = 0.004) and more often required transfer to our institution
in-hospital mortality. Secondary outcomes included unplanned from an outside facility owing to complexity of care (66.2%
relaparotomy and readmission. Hospital system measures of vs. 28.8%, p < 0.0001). When assessing by surgeon type, OA
interest included LOS and hospital costs. All outcomes were was performed in 21.3% of laparotomies by surgical subspecialties
MPI scores (median MPI, 23 vs. 18; p < 0.0001) and peak lactate
TABLE 1. Baseline Characteristics in Emergent
(median lactate level, 3 vs. 1.6; p < 0.0001). When comparing in-
Laparotomy Patients
dications for laparotomy, 75.2% of patients with ischemia or necro-
Open Abdomen Primary Closure sis had OA performed compared to 24.8% undergoing PC
n = 203 n = 331 p (p < 0.0001). Open abdomen was performed following index lap-
Age, years 63 (54–72) 59 (47–70) 0.004
arotomy in 39.6% of perforations and 18.4% of obstructions
Female s 109 (54.2) 179 (55.4) 0.79
(p < 0.0001). Open abdomen was performed more often in patients
Mannheim Peritonitis Index 23 (16–28) 18 (13–24) <0.0001
undergoing surgery at night (52.7% vs. 43.7%, p = 0.044), with in-
Apache score 18 (13–24) 14 (9–19) <0.0001
traoperative hypotension (58.2% vs. 46.5%, p = 0.018), and those
Lactate (mmol/L) 3 (1.8–5) 1.6 (1.1–2.5) <0.0001
requiring vasopressors (68.5% vs. 23.6%, p < 0.0001) or blood
ED admission 94 (46.8) 216 (66.9) <0.0001
transfusions (30.1% vs. 9.0%, p < 0.0001) during index
laparotomy.
Point of Origin <0.0001
Home 68 (33.8) 230 (71.2)
Unmatched analysis clearly demonstrated worsened
Healthcare facility 133 (66.2) 93 (28.8)
outcomes in the OA group when compared to PC. Mortality
Overnight case (6 pm–6 am) 107 (52.7) 143 (43.7) 0.0441
(24.9% vs. 6.5%, p < 0.0001), any complication (72.6% vs.
Surgeon type <0.0001
39.0%, p < 0.0001), and all subcategories of complication
Acute care surgery 171 (85.5) 224 (67.7)
rates were significantly higher in the OA group (Table 2).
Gynycologyic oncology 2 (1.0) 23 (7.0)
As expected, hospital LOS (13 vs. 8 days, p < 0.0001) and
Hepatopancreatobiliary 10 (5.0) 11 (3.3)
hospital cost ($41,759 vs. $18,591, p < 0.0001) were signifi-
Minimally invasive surgery 4 (2.0) 40 (12.1)
cantly higher in the OA group. Unplanned laparotomy oc-
Surgical oncology 11 (5.5) 18 (5.4)
curred at a significantly higher rate in PC compared to OA
Transplant surgery 1 (0.5) 12 (3.6)
patients (3.0% vs. 7.0%, p = 0.048). There was no difference
Indication for laparotomy <0.0001
in readmission.
Perforation 61 (30.5) 93 (29.5)
Following propensity matching, 111 closely matched
Contamination/sepsis 20 (10.0) 18 (5.7)
pairs (OA:PC) were compared. Importantly, differences in age,
Ischemia 79 (39.5) 26 (8.3)
MPI score, APACHE score, lactate level, and overnight surgery
Obstruction 40 (20.0) 178 (56.5)
became nonsignificant. Conversely, point of origin, surgeon spe-
Cloudy/Purulent exudate 74 (36.8) 99 (30.0) 0.104
cialty, and procedure indication remained significantly different
Fecal exudate 46 (22.9) 39 (11.8) 0.0007
(Table 3). Open abdomen patients demonstrated increased rates
Intraoperative hypotension 89 (58.2) 151 (46.5) 0.018
of postoperative complications (71.2% vs. 41.4%, p < 0.0001)
Vasopressor(s) during 122 (68.5) 77 (23.6) <0.0001
(Table 4). Specifically, there were higher rates of gastrointestinal
index laparotomy (15.3% vs. 5.4%, p = 0.015) and infectious complications
Blood transfusion 46 (30.1) 29 (9.0) <0.0001 (23.4% vs. 14.4%, p = 0.015). Open abdomen patients also
Operative duration, mins 85 (55–126) 107 (76–157) <0.0001 had longer hospital LOS (median, 13 days vs. 9 days;
p = 0.0001) and were more likely to be discharged to a facility
n (%) or median (interquartile range).
instead of home (58.6 vs. 45.9%, p = 0.006). There was no lon-
ger a significant difference in unplanned relaparotomy (2.7% vs.
compared to 43.3% (p < 0.0001) of acute care surgeons. Preopera- 8.1%, p = 0.07), and readmissions among patients surviving to
tively, OA patients had increased severity of illness as indicated by discharge (20.7% vs. 17.7%, p = 0.608) were similar between
referral center, many patients requiring readmission are rou- 5. Rezende-Neto J, Rice T, Abreu ES, Rotstein O, Rizoli S. Anatomical, phys-
tinely transferred back to our facility from regional hospitals par- iological, and logistical indications for the open abdomen: a proposal for a
new classification system. World J Emerg Surg. 2016;11:28.
ticularly within 30 days of a major surgery. Future studies such 6. Diaz JJ, Cullinane DC, Dutton WD, et al. The management of the open ab-
as the multi-institutional COOL Trial, an active randomized con- domen in trauma and emergency general surgery: part 1-damage control.
trolled trial led by the World Society of Emergency Surgery, J Trauma. 2010;68(6):1425–1438.
could provide further insight.21,22 Comparisons of long-term 7. Miller RS, Morris JA, Diaz JJ, Herring MB, May AK. Complications after
outcomes in subgroups with specific indications, such as relook 344 damage-control open celiotomies. J Trauma. 2005;59(6):1365–1374.
for bowel ischemia, may also be of additional value. 8. Yuan Y, Ren J, He Y. Current status of the open abdomen treatment for intra-
abdominal infection. Gastroenterol Res Pract. 2013;2013:532013.
9. van Ruler O, Mahler CW, Boer KR, et al. Comparison of on-demand vs
planned relaparotomy strategy in patients with severe peritonitis: a random-
CONCLUSIONS ized trial. JAMA. 2007;298(8):865–872.
The findings of this study demonstrate significant poten- 10. Sartelli M, Abu-Zidan FM, Ansaloni L, et al. The role of the open abdomen
procedure in managing severe abdominal sepsis: WSES position paper.
tial for adverse outcomes and overuse of health care resources World J Emerg Surg. 2015;10:35.
with OA, even when using a selective approach. The implica- 11. Schein M. Surgical management of intra-abdominal infection: is there any
tions of our results and previous studies further highlight the evidence? Langenbecks Arch Surg. 2002;387(1):1–7.
need for determining appropriate use of OA and planned 12. Bleszynski MS, Chan T, Buczkowski AK. Open abdomen with negative
relaparotomy approach. With data to suggest OA may not pro- pressure device vs primary abdominal closure for the management of surgi-
cal abdominal sepsis: a retrospective review. Am J Surg. 2016;211(5):
vide the benefit that operating surgeons perceive, a closer look 926–932.
at the added value of relaparotomy is warranted. Future studies 13. Rausei S, Dionigi G, Boni L, et al. Open abdomen management of intra-
directed at identifying predictors of ongoing intra-abdominal abdominal infections: analysis of a twenty-year experience. Surg Infect
sepsis during relaparotomy and evaluating the incidence of addi- (Larchmt). 2014;15(3):200–206.
tional procedures performed during second look laparotomy 14. Robledo FA, Luque-de-León E, Suárez R, et al. Open versus closed manage-
may help pinpoint a subset of patients that benefit from ment of the abdomen in the surgical treatment of severe secondary peritoni-
tis: a randomized clinical trial. Surg Infect (Larchmt). 2007;8(1):63–72.
relaparotomy. 15. Holzheimer R. Management of secondary peritonitis. In: Mannick JA, ed.
AUTHORSHIP Surgical Treatment: Evidence-Based and Problem-Oriented. Munich,
Germany: Zuckschwerdt; 2001.
A.K., M.B.G., and L.C. reviewed the literature. A.K., B.T.H., B.D., and K.K. 16. Muralidhar VA, Madhu CP, Sudhir S, Srinivasarangan M. Efficacy of
designed and conceptualized the study. A.K., L.C., M.B.G., and T.P. col- Mannheim Peritonitis Index (MPI) score in patients with secondary peritoni-
lected the data collection. A.K., T.P., L.C., and K.K. analyzed and interpreted tis. J Clin Diagnostic Res. 2014;8(12):NC01–NC03.
the data. A.K., M.B.G., and K.K. drafted the manuscript, A.K., B.T.H., B.D., 17. Viehl CT, Kraus R, Zürcher M, Ernst T, Oertli D, Kettelhack C. The acute
and K.K. made critical revision to the manuscript. physiology and chronic health evaluation II score is helpful in predicting
the need of relaparotomies in patients with secondary peritonitis of colorectal
DISCLOSURE origin. Swiss Med Wkly. 2012;142:w13640.
Dr Heniford has prior grants and/or honoraria from W.L. Gore, Allergan, 18. Kreis BE, de Mol van Otterloo AJ, Kreis RW. Open abdomen management: a
and Stryker. All the remaining authors declare no conflicts of interest. review of its history and a proposed management algorithm. Med Sci Monit.
This research did not receive any specific grant from funding agencies in 2013;19:524–533.
the public, commercial, or not-for-profit sectors. 19. Higa G, Friese R, O'Keeffe T, Wynne J, Bowlby P, Ziemba M, Latifi R,
Kulvatunyou N, Rhee P. Damage control laparotomy: a vital tool once
overused. J Trauma Inj Infect Crit Care. 2010;69(1):53–59.
REFERENCES 20. Opmeer BC, Boer KR, van Ruler O, et al. Costs of relaparotomy on-demand
1. Ogilvie W. The late complications of abdominal war wounds. Lancet. 1940; versus planned relaparotomy in patients with severe peritonitis: an economic
2:253–256. evaluation within a randomized controlled trial. Crit Care. 2010;14(3):R97.
2. Rotondo MF, Schwab CW, McGonigal MD, et al. 'Damage control': an ap- 21. Kirkpatrick AW, Coccolini F, Ansaloni L, et al. Closed Or Open after Source
proach for improved survival in exsanguinating penetrating abdominal in- Control Laparotomy for Severe Complicated Intra-Abdominal Sepsis (the
jury. J Trauma. 1993;35(3):375–383. COOL trial): study protocol for a randomized controlled trial. World J Emerg
3. Coccolini F, Montori G, Ceresoli M, et al. The role of open abdomen in non- Surg. 2018;13:26.
trauma patient: WSES Consensus Paper. World J Emerg Surg. 2017;12(1):39. 22. Coccolini F, Catena F, Montori G, et al. IROA: the international register of
4. Wittmann DH, Schein M, Condon RE. Management of secondary peritoni- open abdomen: an international effort to better understand the open abdomen:
tis. Ann Surg. 1996;224(1):10–18. Call for participants. World J Emerg Surg. 2015;10:37.