Scientific paper

Can failure of percutaneous drainage of postoperative abdominalabscesses be predicted?

Stephane Benoist, M.D., Ph.D.a, Yves Panis, M.D., Ph.D.a*, Virginie Pannegeon, M.D.a,Philippe Soyer, M.D., Ph.D.b, Thierry Watrin, M.D.a, Mourad Boudiaf, M.D.b,

Patrice Valleur, M.D.aaDepartment of Surgery, Service de Chirurgie Generale et Digestive, Hopital Lariboisie`re, 2, Rue Ambroise Pare, 75475 Paris Cedex 10, France

bDepartment of Radiology, Lariboisie`re Hospital, Paris, FranceManuscript received December 5, 2001; revised manuscript April 29, 2002

Abstract

Background: Percutaneous drainage (PD) of complex postoperative abscesses associated with a variety of factors such as multiple locationor enteric fistula remains a matter of debate. Accordingly, this retrospective study was designed to determine the predictive factors for failureof PD of postoperative abscess, in order to better select the patients who may benefit from PD.Methods: From 1992 to 2000, the data of 73 patients who underwent computed tomography (CT)-guided PD for postoperativeintra-abdominal abscess, were reviewed. PD was considered as failure when clinical sepsis persisted or subsequent surgery was needed. Thepossible association between failure of PD and 27 patient-, abscess-, surgical-, and drainage-related variables were assessed using univariateand multivariate analysis.Results: Successful PD was achieved in 59 of 73 (81%) patients. The overall mortality was 3% but no patient died after salvage surgery.Multivariate analysis showed that only an abscess diameter of less than 5 cm (P 0.042) and absence of antibiotic therapy (P 0.01) weresignificant predictive variables for failure of PD.Conclusions: CT-guided PD associated with antibiotic therapy could be attempted as the initial treatment of postoperative abdominalabscesses even in complex cases such as loculated abscess or abscess associated with enteric fistula. 2002 Excerpta Medica, Inc. Allrights reserved.

Keywords: Percutaneous drainage; Radiological guidance; Intra-abdominal abscess; Postoperative abscess

Despite advances in surgical techniques and antibiotic ther-apy, postoperative development of intra-abdominal ab-scesses remains a serious and potentially fatal complicationin abdominal surgery with a mortality rate often ranging upto 30% [1,2]. Early recognition and localization, followedby prompt, and appropriate drainage, are essential factorsfor intra-abdominal abscesses to be treated successfully[3,4]. For many years, surgical drainage has been consid-ered as the best therapeutic option in postoperative abdom-inal abscesses, although it was associated with a high mor-tality rate of 20% to 40% [57]. During the 2 last decades,major changes have occurred in the treatment of abdominalabscesses. With refinements in imaging technology, which

improve early detection and accurate localization of post-operative intra-abdominal fluid collection, ultrasonography-or CT-guided percutaneous drainage (PD) has been success-fully developed [810]. Several retrospective studiesshowed that percutaneous approach is as effective as surgi-cal drainage, and can lead to substantial improvement inprognosis [7,1113]. Therefore, despite the lack of random-ized study comparing percutaneous to surgical drainage ofpostoperative abdominal abscesses drainage, PD has be-come a widely accepted treatment for accessible postoper-ative abscess and especially in case of unilocular abscess[7,11,13,14]. Recently, with improved radiological experi-ence, PD of complex abscesses (ie, loculated, poorly definedcollections, abscesses communicating with enteric fistulas,and interloop abscesses) is commonly performed [15,16].However, for such complex abscess, there is no definiteconsensus regarding the use of PD as the initial treatment.

* Corresponding author. Tel.: 00-33-1-4995-8258; fax: 00-33-1-4995-9102.

E-mail address: yves.panis@lrb.ap-hop-paris.fr

The American Journal of Surgery 184 (2002) 148153

0002-9610/02/$ see front matter 2002 Excerpta Medica, Inc. All rights reserved.PII: S0002-9610(02)00912-1

Accordingly, we performed this retrospective study in orderto determine the predictive factors for failure of PD ofpostoperative abdominal abscess, to better select the pa-tients with postoperative abdominal abscess who may ben-efit from PD as the initial treatment.

Patients and methods

Patients

From January 1992 to June 2000, the data of all patientswho have symptomatic postoperative intra-abdominal col-lection in our unit were reviewed. All collections werediagnosed by CT scan. An abscess was defined as a collec-tion of pus or infected fluid collection identified by CTscan-guided needle aspiration and systematically confirmedby positive microbiological culture. A CT scan-guided PDwas only attempted as initial procedure if abscess wasaccessible to PD without risk of injuries of the pleural spaceor abdominal organ, if diagnosis of diffuse peritonitis wasnot suspected and in absence of septic chock at presentation.In this setting, during the same period, 50 patients withinaccessible abscess or diffuse peritonitis were drained sur-gically [17,18]. Furthermore, 10 patients with sterile fluidcollections were also excluded from the study. Seventy-three patients who fulfilled these selection criteria weretreated by PD and formed the basis for this study. The studygroup comprised 31 women and 42 men, with a mean age of55 17.4 years (range 16 to 84). The different types ofinitial surgical procedure are summarized in Table 1. In 26patients, the initial operation was performed in an emer-gency setting.

An abscess was defined as simple when there was asingle unilocular cavity without evidence of an enteric con-nection on CT scan. Abscesses were defined as complexwhen either a multilocular cavity or multiple abscesses werepresent or when an enteric fistula communicating with theabscess cavity was demonstrated. According to this defini-tion, abscess was single in 52 patients (71%) and complexin 21 (29%). In 24 patients, a fistula tract was present: abiliary fistula was observed in 8 cases, a pancreatic fistula in8 cases, and an enteric fistula in 8 cases. The mean size ofabscess was 7.2 2.6 cm (range 2 to 15 cm). The meaninterval between surgery and the diagnosis of intra-abdom-inal abscess by CT scan was 19 16 days (range 4 to 48).

Percutaneous drainage procedure

All the procedures were performed under local anesthe-sia and CT scan guidance. The catheter was inserted usinga modified Seldinger technique. Commonly a 18-gauge nee-dle catheter was inserted percutaneously into the abscess. Avariable amount of aspirated contents was sent for Gramstain, culture, and biochemical assays. After dilatation, amultiple side hole catheter (Flexima Regular Medi-Tech;

Boston Scientific, Massachusetts), size ranging from 8F to16F, was left within the abscess cavity. Large drain waschosen when viscous material was encountered. Sixty pa-tients (81%) had a single drain placed, 12 patients (7%) hadtwo drains, and 2 patients (2%) had three drains. Aspiratedcontents were pus in 48 cases (66%), infected hematoma in16 cases (22%), infected bile in 6 cases (8%), and clear butinfected fluid in 3 cases (4%). Microbiologic culture showednegative Gram stained organism in 70 cases, anaerobic in 40cases, Enterococcus faecalis in 20, and Staphylococcus au-reus in 7. In 44 cases, two or more organisms were detected.Catheters were irrigated three times daily with 10 mL sterilesaline. Volume of drainage fluid was measured daily. Sixtypatients (82%) were maintained on broad-spectrum antibi-otics. These were initially chosen empirically and weresubsequently changed, if necessary, to provide specific cov-erage based on culture and sensitivity results. The indicationof empiric antibiotic therapy was left to the discretion ofprimary surgeon. Fifty patients were treated by empiricaldouble combination therapy (piperacillin-tazobactam andgentamicin in 45 and a third-generation cephalosporin andgentamicin in 15), and 10 patients by triple combinationtherapy (ampicillin-clavulanat, metronidazole, and gentami-cin).

In all patients, catheters were removed when drainagehad completely ceased and serial CT scan showed evidence

Table 1Type of initial surgical procedure in 73 patients undergoingpercutaneous drainage of postoperative intra-abdominal abscess

No. ofpatients

Colorectal surgery 32 (44)Rectal resection 3Left colonic resection 7Right colonic resection 6Subtotal colectomy with ileo-rectal anastomosis 5Total proctocolectomy with pouch-anal anastomosis 3Hartmans procedure 1Colorectal anastomosis after Hartmans procedure 1Appendectomy 6

Hepatic and biliary surgery 18 (25)Hepatic resection 10Hepatic resection with common bile duct resection 2Choledocoduodenal anastomosis 2Cholecystectomy 4

Pancreatic and splenic surgery 12 (16)Pancreaticoduodenectomy 5Distal pancreatectomy 5Bypass surgery 1Splenectomy 1

Gastric surgery 6 (8)Total gastrectomy 4Partial gastrectomy 2

Small bowel resection 2 (3)Adrenalectomy 2 (3)Mesh repair for incisional hernia 1 (1)

Values in parentheses are percentages.

149S. Benoist et al. / The American Journal of Surgery 184 (2002) 148153

of collapse or significant size reduction of the abscess cav-ity.

PD was considered to be successful if the patients intra-abdominal infection resolved without the need for addi-tional surgery. Drainage was considered as failure in case ofpersisting clinical sepsis and when urgent or emergencysurgery was needed.

Statistical analysis

To determine the predictive factors for failure of PD,patients were divided into two groups, based on whether ornot subsequent surgical drainage was required. Patients whowere successfully treated by PD, constituted the successgroup and those, who required salvage surgical drainage,the failure group.

Univariate analysis was used to examine the relation-ships between failure of PD and the 27 following variables:age, sex, overweight (body mass index 27), diabetes mel-litus, cirrhosis, cardiopulmonary disease (including coro-nary artery disease, history of myocardial infarction andvalvular heart disease), steroid treatment, American Societyof Anesthesiologists (ASA) classification, previous laparot-omy, type of initial surgery, emergency or elective surgery,benign or malignant disease, abscess characteristics in-cluded number, size, complexity, associated fistula, local-ization, type of aspirated content, volume of drainage at day1, and type and number of organisms cultured, intervalbetween initial surgery and drainage, number of drain, sizeof drain, antibiotic therapy and duration of antibiotic ther-apy, leukocytes blood count, and fever on the day of diag-nosis of abscess. Quantitative data were expressed asmean SD (range).

Comparisons between groups were analyzed by the chi-square test with Yates correction and the Mann-Whitney Utest for quantitative and qualitative variables, when appro-priate. All variables associated with drainage failure withP 0.20 in univariate analysis were examined consecu-tively by multivariate analysis using forward stepwise lo-gistic regression. Significance was defined as p 0.05.Statistical analysis was performed using biomedical soft-ware (SPSS, for Windows, 6.0, Chicago, Illinois).

Results

Successful catheter drainage was achieved in 59 of 73(81%) patients. Five of them required repeated drainage. PDfailed in 14 patients (19%) for whom subsequent surgicaldrainage was required. The indications for surgical drainagewere persistent sepsis in 5 patients, persisting fistula despiteeffective drainage in 2, and recurrence of abscess withsepsis after drain removal in 7. In patients with failed PD,the mean interval of time from the initial PD to surgicaldrainage was 11 9 days (range 1 to 30).

The overall mortality rate was 3% (2 patients). One

patient died of multiple system organ failure secondary topulmonary infection 2 months after successful PD of sub-phrenic abscess after partial gastrectomy for bleeding duo-denal ulcer. A CT scan provided evidence of completecollapse of abscess cavity. The other patient, who hadproved successful PD of subphrenic abscess after bypasssurgery for unresectable pancreatic tumor, died from mas-sive tumor bleeding 1 month after surgery. No patient in thefailure group died after reoperation for surgical drainage.One complication related to PD was observed in 1 patient inthe success group. This patient had transient respiratoryfailure after drainage of peripancreatic abscess followingleft colonic resection for diverticulitis. The mean duration ofcatheter drainage in the success group was 13 12 days(range 3 to 80). In the success group, drains were left inplace for a mean of 8.6 3.3 days for patients withoutfistula tract and significantly longer for a mean of 23 17.6days in patients with fistula (P 0.001). The mean hospitalstay after drainage of postoperative abdominal abscess was17 11 days (range 6 to 59) in the success group and 24 13 days (range 10 to 60) in the failure group (P 0.05). Inthe failure group, no patient required additional drainageprocedure for reaccumulated abscess after surgical drain-age.

Univariate analysis showed four variables associatedwith an increased risk of PD failure (Tables 2 and 3): gastricresection as initial surgery (P 0.02), an abscess diameterof less than 5 cm (P 0.01), the absence of antibiotictherapy (P 0.02) and a mean duration time of antibiotictherapy of less than 7 days (P 0.05). Multivariate analysisshowed that an abscess diameter of less than5 cm and theabsence of antibiotic therapy only were significant predic-tive variables for failure of PD (Table 4).

Comments

Whatever the technique of drainage used, postoperativeintra-abdominal abscess remains a dreaded complication[19]. PD could settle this problem, thus obviating additionalsurgery. However, PD has to be considered only for locatedabscess but is not indicated in case of diffuse peritonitis,which requires a surgical approach [17,18].

Ultrasonography and CT scan allow an early diagnosis ofpostoperative abscess and imaging-guided PD has thus beenused extensively in the treatment of postoperative abdomi-nal abscess [2022]. However, controversy remains regard-ing percutaneous treatment of complex abscesses associatedwith a variety of factors such as multiple location, entericfistula. Several authors consider that abscess can benefitfrom PD only in selected cases [19,2325]. Furthermore,some authors reported that the mortality rate of salvagelaparotomy for failed PD could be extremely high, rangingup to 50%, which is higher than those usually reported aftersurgical drainage used as the initial procedure [19,26,27].Our study, which includes patients with complex postoper-

150 S. Benoist et al. / The American Journal of Surgery 184 (2002) 148153

ative abdominal abscess, showed that CT-guided PD wassuccessful in the majority of cases (81%). Multivariate anal-ysis revealed that the absence of antibiotic therapy and anabscess diameter of less than 5 cm were the only twoindependent factors for failure of PD. In addition, the mor-tality rate of salvage surgical drainage for failure of PD wasnil. Therefore, we consider that PD associated with antibi-otic therapy could be attempted as the initial treatment ofpostoperative abscesses even in complex cases such as loc-ulated abscess or associated with enteric fistula.

Our success rate of PD is lower than those usually re-ported by many authors [8,1214,21,28,29]. However, mostof previously published series included patients who havenot only postoperative abscess, but also abscess secondaryto diverticulitis or Crohns disease, which both constitute adifferent situation. If we only consider the results of PD ofpostoperative abscess, our success rate compares favorablywith those reported by others, ranging from 43% to 80%[9,10,19,25,3033].

The most appropriate management of postoperative ab-scess is not clearly defined in the literature. In particular,indications and contraindications for PD are not well estab-lished so far. Ideally, this should be addressed by a pro-spective randomized trial, which is, however, not yet avail-

able. Knowledge of the predictive factors for failure of PDof postoperative intra-abdominal abscess may be helpful inselecting the patients who may benefit from this procedure.Several authors have identified by univariate analysis pre-dictive factors for failure of PD, including enteric fistula,multiple or loculated abscess, large size of abscess, presenceof necrotic tissue and pancreatic localization [4,7,8,10,15,20,21,27]. To our knowledge, our study is the first one, inwhich predictive factors were determined by multivariate

Table 2Univariate analysis of patient characteristics related to failure ofpercutaneous drainage of postoperative intra-abdominal abscess in 73patients

Successgroup(n 59)

Failuregroup(n 14)

Pvalue

Age (years) 55 17 53 18 0.75*Sex (female/male) 26/33 5/9 0.73Overweight (BMI 27) 10 (17) 2 (14) 0.85Cardiopulmonary disease 6 (10) 1 (7) 0.84Diabetes mellitus 5 (8) 0.66Cirrhosis 2 (3) 0.81Previous laparotomy 6 (10) 1 (7) 0.80Steroid treatment 3 (5) 2 (14) 0.48ASA status 2 13 (22) 4 (28) 0.79Malignant disease 23 (39) 7 (50) 0.71Initial surgery as emergency 19 (32) 7 (50) 0.38Type of initial surgery

Colorectal surgery 25 (42) 7 (50) 0.71Hepatic and biliary surgery 17 (29) 1 (7) 0.18Pancreatic and splenic surgery 11 (19) 1 (7) 0.58Gastric surgery 2 (3) 4 (29) 0.013Others 4 (7) 1 (7) 0.64

Interval between surgery andpercutaneous drainage (days)

20 15 26 23 0.22*

Leukocytes blood count15,000

40 (68) 10 (71) 0.88

Fever 38.5C 52 (88) 12 (86) 0.81Values in parentheses are percentages.* Mann-Whitney U test. On the day of percutaneous drainage.BMI body mass index; ASA American Society of Anesthesiolo-

gists.

Table 3Univariate analysis of abscess and drainage characteristics related tofailure of percutaneous drainage of postoperative intra-abdominalabscess in 73 patients

Successgroup(n 59)

Failuregroup(n 14)

Pvalue

Multiple abscesses 14 (24) 3 (21) 0.82Size of abscess 5 cm 12 (20) 10 (71) 0.0008Complex abscess* 16 (27) 5 (36) 0.69Associated fistula 17 (29) 7 (50) 0.23Localization of abscess

Subphrenic 24 (41) 4 (28) 0.61Subhepatic 18 (31) 3 (21) 0.78Paracolic gutter 25 (42) 6 (42) 0.83Pancreatic lesser sac 5 (8) 2 (14) 0.85

Type of aspirated contentPus 38 (64) 10 (71) 0.79Hematoma 13 (22) 3 (21) 0.82Bile 5 (8) 1 (8) 0.86Clear fluid 3 (5%) 0.9

Initial volume of drainage(mL)

169 146 225 288 0.28

Multiple cultured organism 35 (59) 9 (64) 0.5Multiple drains 12 (20) 2 (14) 0.87Size of drain 12F 15 (25) 3 (21) 0.92Antibiotic therapy 52 (88) 8 (57) 0.019Duration of antibiotic

therapy 7 days11/52 (21) 5/8 (62) 0.047

Values in parentheses are percentages.* Abscesses were defined as complex when either a multilocular cavity

or multiple abscesses were present or when an enteric fistula communicat-ing with the abscess cavity was demonstrated.

Several abscesses had multiple localization. Mann-Whitney U test. On the day of percutaneous drainage.

Table 4Multivariate analysis of predictive factors for failure of percutaneousdrainage of postoperative intra-abdominal abscess in 73 patients

Oddsratio

95%Confidenceintervals

Pvalue

Gastric surgery 2.09 0.7810.5 0.18Hepatic and biliary surgery 1.87 0.538.6 0.27Size of abscess 5 cm 7.45 3.0113.88 0.01No antibiotic therapy 3.82 1.788.21 0.037Duration of antibiotic therapy7 days

1.36 0.4327.4 0.55

151S. Benoist et al. / The American Journal of Surgery 184 (2002) 148153

analysis. Although the present work is one of the largestseries of patients with postoperative abdominal abscesstreated by PD, the small number of patients make analysisof variables responsible for failure tenuous. It would beprobably interesting to check if similar result would beobserved in a larger cohort of patients.

If the efficiency of empiric antibiotic therapy has beendemonstrated in cases of postoperative diffuse peritonitis[34], to our knowledge, there is no data about the role ofantibiotic therapy in specific case of postoperative intra-abdominal abscesses treated by PD. Several authors [4,8,12,15,21,25] use it routinely as a part of treatment whileothers do not use it or neglect to notice it [3,11,12,19]. Thecurrent data support their systematic use for a prolongedperiod. Initially it should cover usual digestive organisms,ie, anaerobic and negative Gram stained, and subsequentlyadapted to culture results.

In our study, we found, surprisingly, that a small abscessdiameter appeared as an independent predictive factor forfailure of PD. Conversely, previous studies reported largesize of abscess as a cause of failed PD [14,15]. Severalhypotheses may be raised to explain our result. First, smallabscess may be more difficult to localize and less accessiblefor insertion of drain. Second, small abscess may be moreprone to drain displacement resulting in incomplete drain-age [33]. This hypothesis is further supported by the factthat the 10 patients with small abscess in the failure groupwere reoperated for persistent sepsis or sepsis recurrenceafter drain removal, probably because of incomplete drain-age. Some authors have reported that small abscess afterappendicectomy could be successfully managed with anti-biotics alone [35,36]. However, in our study, if we onlyfocus on patients with an abscess diameter of less than 5 cm,the use of antibiotic therapy was similar in the success (8 of12, 75%) and failure group (7 of 10, 70%). We consider thatantibiotics alone are not sufficient for the treatment of smallbut symptomatic postoperative abscess. In our study, alower but still clinically significant rate of success were seenin patients with small abscess (12 of 22 patients; 55 %).Furthermore, mortality rate was nil in patients who requiredsubsequent surgical drainage after failure of PD. Therefore,for patients with small abscess, we consider that when anabscess seems easily accessible on CT scan, PD could bereasonably attempted and surgical drainage should be indi-cated only in patients without clinical improvement within24 to 48 hours after PD.

Mortality and morbidity rates are two important issues tobe addressed before determining the best therapeutic optionfor treatment of postoperative abdominal abscess. In ourstudy, overall mortality rate was 3%, which is similar to therate usually reported, ranging from 0% to 16% [3,4,1012,21,27]. In addition in our study, 2 patients died from causes, which were not related to PD with no missed abscesses. Asin other reports [4,11,13,21,31], the rate of complicationsdirectly attributable to PD was low in our study (ie, less than10%). In particular no enteric fistula resulting from drain

misplacement occurred. Moreover, mortality and morbidityrates favorably compared with those usually reported aftersurgical drainage [5,6,37].

In conclusion, the results of our study suggest that, inabsence of diffuse peritonitis, CT scan-guided PD of post-operative intra-abdominal abscesses could be indicated infirst instance whenever feasible as a safe and relativelyatraumatic procedure with few complications. It should beassociated with a prolonged intravenous antibiotic therapyas a part of treatment. Neither the presence of multiloculatedabscesses, multiple abscesses, nor intestinal fistulas pre-cludes the use of PD.

References

[1] Fry DE, Garrison N, Heitsch RC, et al. Determinants of death inpatients with intraabdominal abscesses. Surgery 1980;88:51723.

[2] Skaut T, Nystrom PO, Carlsson C. Severity of illness in intra-abdom-inal infection: a comparison of two indexes. Arch Surg 1985;120:1528.

[3] Levison MA, Zeigler D. Correlation of Apache II score, drainagetechnique and outcome in postoperative intra-abdominal abscess.J Am Coll Surg 1991;172:8994.

[4] Halasz NA, van Sonnenberg E. Drainage of intraabdominal ab-scesses. Tactics and choices. Am J Surg 1983;146:11215.

[5] Butler JA, Huang J, Wilson SE. Repeated laparotomy for postoper-ative intra-abdominal sepsis. An analysis of outcome predictors. ArchSurg 1987;122:7026.

[6] Hinsdale JG, Jaffe BM. Re-operation for intra-abdominal sepsis.Indications and results in modern critical care setting. Ann Surg1984;199:316.

[7] Malangoni MA, Shumate CR, Hollis AT, Richardson JD. Factorsinfluencing the treatment of intra-abdominal abscesses. Am J Surg1990;159:16771.

[8] Gerzof SG, Robbins AH, Johnson WC, et al. Percutaneous catheterdrainage of abdominal abscesses: a five-year experience. N EnglJ Med 1981;305:6537.

[9] Martin EC, Karlson KB, Fankhuchen EI, et al. Percutaneous drainageof postoperative intraabdominal abscesses. AJR Am J Roentgenol1982;138:1315.

[10] Walters R, Herman CM, Neff R, et al. Percutaneous drainage ofabscesses in the postoperative abdomen that is difficult to explore.Am J Surg 1985;149:6236.

[11] Johnson WC, Gerzof SG, Robbins AH, Nabseth DC. Treatment ofabdominal abscesses. Comparative evaluation of operative drainageversus percutaneous catheter drainage guided by tomography or ul-trasound. Ann Surg 1981;194:51018.

[12] Hemming A, Davis NL, Robins E. Surgical versus percutaneousdrainage of intra-abdominal abscesses. Am J Surg 1991;161:5935.

[13] Brolin RE, Nosher JL, Leiman S, et al. Percutaneous catheter versusopen surgical drainage in the treatment of abdominal abscesses. AmSurg 1984;50:1028.

[14] Haaga JR. Imaging intraabdominal abscesses and nonoperative drain-age procedures. World J Surg 1990;14:2049.

[15] Bernini A, Spencer MP, Wong D, et al. Computed tomography-guided percutaneous abscess drainage in intestinal disease. Factorsassociated with outcome. Dis Colon Rectum 1997;40:100913.

[16] Gerzof SG, Johnson WC, Robbins AH, Nabseth DC. Expanded cri-teria for percutaneous abscess drainage. Arch Surg 1985;120:22732.

[17] Alves A, Panis Y, Pocard M, et al. Management of anastomoticleakage after non diverted large bowel resection. J Am Coll Surg1999;189:5549.

152 S. Benoist et al. / The American Journal of Surgery 184 (2002) 148153

[18] Benoist S, Panis Y, Alves A, Valleur P. Impact of obesity on surgicaloutcomes after colorectal resection. Am J Surg 2000;179:27581.

[19] McLean TR, Simmons K, Svensson LG. Management of postopera-tive intra-abdominal abscesses by routine percutaneous drainage.Surg Gynecol Obstet 1993;176:16777.

[20] Lambiase RE, Deyoe L, Cronan JJ, Dorfman GS. Percutaneous drain-age of 335 consecutive abscesses: results of primary drainage with1-year follow-up. Radiology 1992;184:16779.

[21] Goletti O, Lippolis PV, Chiarugi M, et al. Percutaneous ultrasound-guided drainage of intraabdominal abscesses. Br J Surg 1993;80:3369.

[22] Levison MA. Percutaneous versus open operative drainage of intra-abdominal abscesses. Infect Dis Clin North Am 1992;6:52544.

[23] Civardi G, Fornari F, Cavanna L, et al. Ultrasonically guided percu-taneous drainage of abdominal fluid collections: a long term study ofits therapeutic efficacy. Gastrointest Radiol 1990;15:24550.

[24] Aeder MI, Wellman JL, Haaga JR, Hau T. Role of surgical andpercutaneous drainage in the treatment of abdominal abscesses. ArchSurg 1983;118:27380.

[25] Sunshine J, McConnell DB, Weinstein CJ, et al. Percutaneous ab-dominal abscess drainage. Portland area experience. Am J Surg 1983;145:61518.

[26] Reynaert MS, Dardenne AN, Jacquemart JL, Kestens PJ. Percutane-ous drainage of infected abdominal collections. Guiding principlesand results. Acta Chir Belg 1983;83:3537.

[27] Brolin RE, Flancbaum L, Ercoli FR, et al. Limitations of percutane-ous catheter drainage of abdominal abscesses. Surg Gynecol Obstet1991;173:20310.

[28] Mueller PR, Simeone JF, Butch RJ, et al. Percutaneous drainage ofsubphrenic abscess: a review of 62 patients. AJR Am J Roentgenol1986;147:123740.

[29] Van Gansbeke D, Matos C, Gelin M, et al. Percutaneous drainage ofsubphrenic abscesses. Br J Radiol 1989;62:12733.

[30] Bouali K, Magotteaux P, Jadot A, et al. Percutaneous catheter drain-age of abdominal abscess after abdominal surgery. Results in 121cases. J Belg Radiol 1993;76:1114.

[31] Glass CA, Cohn I. Drainage of intraabdominal abscesses. A compar-ison of surgical and computerized tomography guided catheter drain-age. Am J Surg 1984;147:31517.

[32] Deveney CW, Lurie K, Deveney KE. Improved treatment of intra-abdominal abscess. A result of improved localization, drainage, andpatient care, not technique. Arch Surg 1988;123:112630.

[33] Lang EK, Springer RM, Glorioso LW, Cammarata CA. Abdominalabscess drainage under radiologic guidance: causes of failure. Radi-ology 1986;159:32936.

[34] Montravers P, Gauzit R, Muller C, et al. Emergence of antibiotic-resistant bacteria in cases of peritonitis after intraabdominal surgeryaffects the efficacy of empirical antimicrobial therapy. Clin Infect Dis1996;23:48694.

[35] Gorenstein A, Gewurtz G, Serour F, Somekh E. Postappendectomyintra-abdominal abscess: a therapeutic approach. Arch Dis Child1994;70:4002.

[36] Heloury Y, Baron M, Bourgoin S, et al. Medical treatment of post-appendectomy intraperitoneal abscesses in children. Eur J PediatrSurg 1995;5:14951.

[37] Fry DE, Clevenger FW. Reoperation for intra-abdominal abscess.Surg Clin North Am 1991;71:15974.

153S. Benoist et al. / The American Journal of Surgery 184 (2002) 148153