Randomized clinical trial

Use of an intraoperative wound protector to prevent

surgical-site infection after pancreatoduodenectomy:
randomized clinical trial
M. De Pastena , G. Marchegiani, S. Paiella, M. Fontana, A. Esposito, L. Casetti, E. Secchettin,
G. Manzini, C. Bassi and R. Salvia
Department of General and Pancreatic Surgery, Pancreas Institute, University and Hospital Trust of Verona, Policlinico GB Rossi, Piazzale L.A. Scuro,
10, 37134, Verona, Italy
Correspondence to: Dr M. De Pastena (e-mail: m.depastena@gmail.com)

Background: Surgical-site infection (SSI) increases treatment costs, duration of hospital stay and
readmission rate after pancreatic surgery. This study aimed to assess whether a wound protector could
reduce the risk of superficial incisional SSI after pancreatoduodenectomy.
Methods: This RCT included patients undergoing pancreatoduodenectomy at Verona University Hos-
pital, between 2017 and 2018. The experimental group had a dual-ring wound protector, whereas the
control group had standard surgical drapes. The groups were stratified by preoperative biliary stent place-
ment. The primary outcome was the overall rate of superficial SSI.
Results: An interim analysis was conducted after 212 patients had been enrolled; 22 patients (10⋅4
per cent) were excluded owing to inability to complete the pancreatoduodenectomy, or the need for
postoperative reintervention. Some 94 patients (49⋅5 per cent) had a wound protector and 96 (50⋅5
per cent) had standard drapes. There were no differences between groups in demographics, or in
intraoperative findings, pathological data or surgical outcomes. The overall superficial SSI rate was 7⋅4 per
cent, which did not differ between groups (7 per cent in each group; P = 0⋅585). Subanalysis of patients
with a preoperative biliary stent showed a similar outcome (superficial SSI rate 9 versus 8 per cent with
wound protector versus surgical drapes respectively; P = 0⋅536). The trial was stopped prematurely on the
grounds of futility.
Conclusion: Use of a wound protector did not reduce the rate of superficial SSI after pancreatoduo-
denectomy. Registration number: NCT03820648 (http://www.clinicaltrials.gov).

Paper accepted 13 January 2020

Published online 12 March 2020 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.11527

Introduction Various measures of perioperative care have been adopted

to reduce SSIs, including cleaning the skin, hair removal
Surgical-site infection (SSI) is a leading cause of pre- from the intervention area, prevention of intraoperative
ventable morbidity and mortality1 , and accounts for up to hypothermia and perioperative antibiotic therapy8 – 10 . Use
25 per cent of all healthcare-associated infections2 . Pro- of adhesive membrane barriers, protective covers, or a
longed hospitalization, more frequent hospital readmis- wound protector is hypothesized to reduce intraopera-
sions after surgery, and a more than a twofold increase in tive contamination, while preserving the temperature and
costs and mortality are consistently associated with SSI3 . humidity of the surgical wound11 . Earlier studies demon-
SSIs correlate with an increased risk of postoperative pain, strated reduced exposure of the surgical wound to enteric
poor wound healing, aesthetic issues and incisional hernia4 . bacteria in gastrointestinal operations12 . This aim of this
Infective complications are among the most frequent study was to assess whether the use of a wound protector
problems after pancreatic surgery, occurring in at least could reduce the risk of superficial incisional SSI in patients
one-third of patients5 . After pancreatoduodenectomy undergoing PD. Patients requiring biliary drainage are
(PD), the SSI rate may be as high as 40 per cent6 . SSI can known to have a higher rate of SSI13 , so patients in the
delay adjuvant chemotherapy in patients with pancreatic study were stratified according to whether drainage was
cancer, and double the risk of omitting chemotherapy7 . used before surgery.

© 2020 BJS Society Ltd BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
1108 M. De Pastena, G. Marchegiani, S. Paiella, M. Fontana, A. Esposito, L. Casetti et al.

Methods
Fig. 1 Dual-ring wound protector used during
pancreatoduodenectomy
This study was a prospective, observer-blinded, parallel-
group, intention-to-treat RCT, carried out in accordance
with the CONSORT guidelines14 . The trial was approved
by the Verona Institutional Review Board (CE 1369 cesc)
and registered with ClinicalTrials.gov (NCT03820648).
The study population comprised patients aged at least
18 years, with periampullary pancreatic cancer who had
PD at the Pancreas Institute, Verona University Hospital,
between September 2017 and September 2018. Exclusion
criteria were age less than 18 years and declining informed
consent. Patients were excluded if it was not possible to
complete the planned resection (conversion to gastric
bypass and/or biliary bypass and/or exploratory laparo-
tomy) or a relaparotomy was needed during the hospital
stay. The primary endpoint was the rate of superficial inci-
sional SSI within 30 days of surgery. Secondary endpoints
were analysed for feasibility and device safety, and duration
of hospital stay.

Intervention
All patients underwent routine preoperative diagnos-
tic evaluation during the first clinical visit, to check for XX-Large for incision length 11–17 and 17–25 cm respec-
inclusion and exclusion criteria; after this, they signed tively. At operation, the skin was cleansed with iodinated
informed consent. A rectal swab was taken before surgery antiseptics, or chlorhexidine in patients allergic to iodine.
to look for multidrug-resistant bacteria, with perioperative During the operation, all patients had a bile sample sent
antibiotic prophylaxis based on the result15 . Patients with for culture to check the bacterial flora of the biliary tract.
a negative swab received 2 g ampicillin and 1 g sulbactam A classical Whipple’s procedure was performed, preserv-
intravenously at induction; this was repeated at half the ing the pylorus where possible. The pancreatic stump was
dose every 3 h until the end of surgery. Those with a closed with a single-layer, end-to-side pancreatojejunos-
positive swab had individualized antibiotic prophylaxis. tomy (with or without duct-to-mucosa anastomosis), or by
The sample size and the randomization list were a single-layer pancreatogastrostomy performed through
calculated using a tool available from https://www an anterior surface gastrotomy16,17 . An end-to-side hepati-
.sealedenvelope.com. To avoid selection bias only the cojejunostomy and duodenojejunostomy were performed
head nurse had access to the list. Before the surgery all the on the same jejunal loop in a Child’s fashion. Two laminar
surgeons were blinded to the randomization. On the day of Penrose-type drains were positioned, near the pancreatic
surgery, the research nurse communicated to the surgeon and biliary anastomoses respectively. The surgeon was
the randomization to one of two groups (wound protector responsible for reporting any device damage. The fascia
or standard drapes). The patients were also stratified into was closed by a braided absorbable suture. The skin was
those who had pancreatic duct drainage with a stent, and closed with a skin stapler. The wound was covered by
those without, before randomization, to ensure a balanced a dressing, which was usually removed on postoperative
number of stented patients in each group. The experimen- day 3.
tal group received the Alexis® wound protector (Applied
Medical, Rancho Santa Margarita, California, USA). A
Data collection and definitions
standard 3 M™ Steri-Drape™ polypropylene drape (3 M,
St Paul, Minnesota, USA) was used in the control group. Demographic and clinical details recorded included: age,
The Alexis® is a surgical device for use in retracting and sex, ASA fitness grade, BMI, diabetes mellitus, age-adjusted
protecting an incision during laparoscopic or open surgery Charlson co-morbidity index score18 , smoking habits,
(Fig. 1). The size of the device was based on the length jaundice, preoperative biliary stent and neoadjuvant treat-
of the abdominal incision: Alexis® X-Large or Alexis® ment. Surgical and pathological data were also collected.

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
Prevention of surgical-site infection 1109

Fig. 2 CONSORT diagram for the trial

Assessed for eligibility

at interim analysis
n = 262

Enrolment
Excluded n = 50
Did not meet inclusion criteria n = 50

Randomized
n = 212
Allocation

Allocated to wound protection n = 105 Allocated to standard surgical

Received intervention n = 105 drapes n = 107
Received intervention n = 107
Excluded

Excluded n = 11 Excluded n = 11
Unable to complete PD or need for Unable to complete PD or need for
postoperative reintervention n = 11 postoperative reintervention n = 11
Analysis

Analysed Analysed
n = 94 n = 96

PD, pancreatoduodenectomy.

SSI was defined according to the Centers for Disease was used with a two-sided significance level of 5 per cent,
Control and Prevention guidelines19 . Clinicians or nurses and a group sample size of 191 patients; this number of
blinded to the randomization assessed for the presence patients was required in each group to achieve 80 per
of superficial SSI by direct observation at the bedside or cent power to detect a difference between group pro-
wound swab culture. All other postoperative complications portions of 0⋅070. With a 10 per cent drop-out rate, 212
were scored and graded according to the Clavien–Dindo patients in each group were needed. The study design
system; major complications were defined as those with employed a sequential group approach, with adaptive
a Clavien–Dindo grade of III or higher20 . Postoperative re-evaluation of the sample size after interim analysis; this
mortality and morbidity were recorded within 90 days of was done after the enrolment of 50 per cent of the patients.
surgery. Postoperative pancreatic fistula, delayed gastric At the interim analysis, it was determined whether the
emptying and postpancreatectomy haemorrhage were study should continue or be stopped, using two standard
classified in accordance with International Study Group criteria24,25 .
of Pancreatic Surgery definitions21 – 23 . After discharge, Continuous data are reported as median (i.q.r.). Com-
patients were evaluated in a pancreatic outpatient clinic, parative analysis between groups was conducted using the
recording eventual rates of superficial incisional SSI using Mann–Whitney U test for continuous variables, and χ2
the above methods. or Fisher’s exact test for categorical variables. P < 0⋅050
(2-tailed) was considered statistically significant. Data were
analysed using SPSS® version 24.0 for Windows® (IBM,
Statistical analysis
Armonk, New York, USA).
Considering a superficial incisional SSI rate of 10 per cent
at the authors’ centre5 , along with data in the literature Results
suggesting a 50–100 per cent reduction in superficial SSI
with the use of a wound protector, a final rate of 3 per A total of 212 patients were enrolled in the study between
cent was expected. The Z test with unpooled variance September 2017 and September 2018, when an interim

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
1110 M. De Pastena, G. Marchegiani, S. Paiella, M. Fontana, A. Esposito, L. Casetti et al.

Table 1 Clinical details of patients randomized to wound protector or standard surgical draping during pancreatoduodenectomy

Total Wound Standard surgical drapes

(n = 190) protector (n = 94) (n = 96)

Age (years)* 65 (59–71) 63 (57–70) 65 (60–71)

Sex ratio (M : F) 103 : 87 50 : 44 53 : 43
BMI (kg/m2 )* 24⋅2 (22⋅3–26⋅6) 24 (22⋅3–26⋅6) 24⋅4 (22⋅2–26⋅6)
Diabetes 42 (22⋅1) 21 (22) 21 (22)
Smoking 95 (50⋅0) 43 (46) 52 (54)
ASA fitness grade III–IV 39 (20⋅5) 15 (16) 24 (25)
CACI score > 4 98 of 189 (51⋅9) 48 of 93 (52) 50 (52)
Jaundice 116 (61⋅1) 57 (61) 59 (61)
Preoperative biliary stent 107 (56⋅3) 55 (59) 52 (54)
Neoadjuvant therapy 45 (23⋅7) 20 (21) 25 (26)

Values in parentheses are percentages unless indicated otherwise; *values are median (i.q.r.). CACI, Charlson age-related co-morbidity index.

Table 2 Intraoperative and pathological data in 190 patients undergoing pancreatoduodenectomy

Total Wound protector Standard surgical drapes

(n = 190) (n = 94) (n = 96) P†

Duration of surgery (min)* 415 (352–470) 405 (360–471) 414 (346–465) 0⋅714‡
Major vessel resection 26 of 189 (13⋅8) 16 of 93 (17) 10 (10) 0⋅126
Blood loss (ml)* 450 (290–707) 400 (290–700) 460 (293–700) 0⋅631‡
Blood transfusion 25 of 189 (13⋅2) 11 of 93 (12) 14 (15) 0⋅366
Pathology 0⋅132
PDAC 97 (51⋅1) 41 (44) 56 (58)
pNET 18 (9⋅5) 11 (12) 7 (7)
IPMN 23 (12⋅1) 12 (13) 11 (11)
Ampullary tumour 24 (12⋅6) 16 (17) 8 (8)
Cholangiocarcinoma 17 (8⋅9) 11 (12) 6 (6)
Other malignancy 6 (3⋅2) 2 (2) 4 (4)
Other benign 5 (2⋅6) 1 (1) 4 (4)

Values in parentheses are percentages unless indicated otherwise; *values are median (i.q.r.). PDAC, pancreatic ductal adenocarcinoma; pNET, pancreatic
neuroendocrine tumour; IPMN, intraductal papillary mucinous neoplasm. †χ2 or Fisher’s exact test, except ‡Mann–Whitney U test.

analysis was performed in accordance with the study pro- wound protector and standard drapes groups respectively;
tocol. Twenty-two patients (10⋅4 per cent) were excluded P = 0⋅324) in the study cohort. In terms of intraoperative
as it was not possible to complete the PD, or reinterven- and pathological variables, the two groups had similar
tion was required during the postoperative course (Fig. 2). operating times, rates of vascular resection, and final
Fourteen patients needed reoperation during the hospital pathology (Table 2). Complication rates and other surgical
stay, of whom four had superficial SSI. The reinterventions outcomes were similar in the two groups (Table 3). No
were done for abdominal abscess (6 patients), bleeding (4), patient sustained any damage or complication as a result
obstruction (3) and perforation (1). of the wound protection device.
The final study population comprised 190 patients: 94
patients (49⋅5 per cent) in the wound protector group and Primary endpoint
96 (50⋅5 per cent) in the standard drapes group. There The overall superficial incisional SSI rate was 7⋅4 per
were no differences between groups in demographic data cent, and was similar in the two groups (both 7 per cent;
(Table 1). Preoperative rectal swabs were positive in 36 P = 0⋅585). Jaundiced patients who had preoperative stent-
patients (18⋅9 per cent), who received tailored antibiotic ing had similar rates of superficial SSI (8⋅4 per cent over-
prophylaxis. Stratification balanced the rate of preop- all) with no advantage from wound protection (9 versus
erative jaundice (61 per cent in both groups; P = 0⋅513) 8 per cent with wound protector versus surgical drapes
and preoperative biliary stenting (59 and 54 per cent in respectively; P = 0⋅536).

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
Prevention of surgical-site infection 1111

Table 3 Complications and other postoperative outcomes in 190 patients undergoing pancreatoduodenectomy

Total Wound protector Standard surgical drapes

(n = 190) (n = 94) (n = 96) P†

Any complication 106 (55⋅8) 54 (57) 52 (54) 0⋅379

Postoperative pancreatic fistula 21 (11⋅1) 13 (14) 8 (8) 0⋅164
Biliary fistula 11 (5⋅8) 7 (7) 4 (4) 0⋅262
Abdominal collection 52 (27⋅4) 26 (28) 26 (27) 0⋅529
Delayed gastric emptying 22 (11⋅6) 10 (11) 12 (13) 0⋅431
Postpancreatectomy haemorrhage 29 (15⋅3) 15 (16) 14 (15) 0⋅475
Clavien–Dindo grade ≥ III 25 (13⋅2) 11 (12) 14 (15) 0⋅355
Infectious complications 57 (30⋅0) 29 (31) 28 (29) 0⋅462
Sepsis 21 (11⋅1) 9 (10) 12 (13) 0⋅351
Recovery in ICU 14 (7⋅4) 6 (6) 8 (8) 0⋅407
Superficial incisional SSI 14 (7⋅4) 7 (7) 7 (7) 0⋅585
Organ/space SSI 29 (15⋅3) 14 (15) 15 (16) 0⋅376
Duration of hospital stay (days)* 8 (7–16) 8 (7–14) 8 (6–13) 0⋅137‡
Mortality 2 (1⋅1) 2 (2) 0 (0) 0⋅246

Values in parentheses are percentages unless indicated otherwise; *values are median (i.q.r.). SSI, surgical-site infection. †χ2 or Fisher’s exact test, except
‡Mann–Whitney U test.

The interim analysis was performed as planned and, con- classification of SSIs is needed to assess the real impact of
sidering these results, the criteria for futility were met. A complications after pancreatic surgery. Deep organ/space
revised sample size analysis showed that more than 10 000 SSI is usually related to postoperative pancreatic fistula,
patients would be needed in each group. A time trend which is less easily modifiable in an infection control
analysis was also performed to examine possible fluctua- programme29 .
tion in the superficial SSI rate. Three-month intervals were The superficial SSI rate reported here (7⋅4 per cent) is
assessed, but no significant difference in the superficial SSI extremely low compared with data from the literature6,30 .
rate was detected. In accordance with the Verona Institu- This could be due to the rigorous definition and assessment
tional Review Board, the trial was stopped. of the primary outcome (superficial SSI). Another explana-
Analysing bacterial cultures for SSI, half the infections tion is the role played by implementation of a programme
were caused by Gram-negative Enterobacteriaceae (7 of infection reduction: preoperative antibiotic prophylaxis
patients, mostly with Escherichia coli and Klebsiella pneu- guided by the results of rectal swab15 . This was intended
moniae). Also detected were: Gram-positive cocci, divided to tackle the emerging complication of resistant infection
into skin contaminants (4 patients) and gut contaminants after pancreatic surgery. The improvement in antibiotic
(2); and Aeromonas hydrophila caviae (1). A match between prophylaxis could have led to the low SSI rate.
the intraoperative bile culture and wound swab was present The trial interim analysis suggested that it was futile
in only four of 14 patients. to continue the study, and that proof of effectiveness of
the wound protection device was unlikely to have been
Discussion achieved. Wound protection in pancreatic surgery has been
well investigated, and similar studies31 – 33 have reported
The trial demonstrated that the use of a dual-ring wound contrasting results. The discrepancy could be explained by
protector was both feasible and safe during PD. According the multifactorial process of the development of SSIs, so
to the interim analysis, however, the wound protector did changing a single factor (wound protection) may not show
not reduce the superficial SSI rate compared with standard significant results.
surgical drapes in this setting. In this study, which was terminated prematurely, the use
Despite improvements in perioperative care, SSIs are of a dual-ring wound protector during PD did not reduce
still responsible for the majority of postoperative com- superficial SSI compared with standard surgical drapes
plications after PD, reaching a rate of 40 per cent even in a high-volume centre. Wound protection might still
at high-volume pancreatic centres26 . Several enhanced be useful as part of an infection reduction programme,
recovery programmes focus on reduction in SSIs27,28 , which should be directed at local issues in the reduction
but SSI development is a multifactorial process. A clear of infectious complications.

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
1112 M. De Pastena, G. Marchegiani, S. Paiella, M. Fontana, A. Esposito, L. Casetti et al.

Acknowledgements colorectal surgery: a randomized clinical trial. Dis Colon

Rectum 2010; 53: 1374–1380.
The authors thank Applied Medical (Rancho Santa Mar- 12 Edwards JP, Ho AL, Tee MC, Dixon E, Ball CG. Wound
garita, California, USA) who supplied the dual-ring wound protectors reduce surgical site infection: a meta-analysis of
protectors for the study. The company was not involved in randomized controlled trials. Ann Surg 2012; 256: 53–59.
the analysis of the trial. 13 De Pastena M, Marchegiani G, Paiella S, Malleo G,
Disclosure: The authors declare no other conflict of interest. Ciprani D, Gasparini C et al. Impact of preoperative biliary
drainage on postoperative outcome after
pancreaticoduodenectomy: an analysis of 1500 consecutive
References cases. Dig Endosc 2018; 30: 777–784.
14 Schulz KF, Altman DG, Moher D; CONSORT Group.
1 Wenzel RP. Health care-associated infections: major issues
CONSORT 2010 statement: updated guidelines for
in the early years of the 21st century. Clin Infect Dis 2007;
reporting parallel group randomised trials. BMJ 2010; 340:
45(Suppl 1): S85–S88.
c332.
2 Ban KA, Minei JP, Laronga C, Harbrecht BG, Jensen EH,
15 De Pastena M, Paiella S, Azzini AM, Marchegiani G,
Fry DE et al. American College of Surgeons and Surgical
Malleo G, Ciprani D et al. Preoperative surveillance rectal
Infection Society: surgical site infection guidelines, 2016.
swab is associated with an increased risk of infectious
J Am Coll Surg 2017; 224: 59–74.
complications in pancreaticoduodenectomy and directs
3 Zimlichman E, Henderson D, Tamir O, Franz C, Song P,
antimicrobial prophylaxis: an antibiotic stewardship
Yamin CK et al. Health care-associated infections: a
strategy? HPB (Oxford) 2018; 20: 555–562.
meta-analysis of costs and financial impact on the US health
16 Bassi C, Butturini G, Salvia R, Crippa S, Falconi M,
care system. JAMA Intern Med 2013; 173: 2039–2046.
Pederzoli P. Open pancreaticogastrostomy after
4 Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton
DJ. The impact of surgical-site infections in the 1990s: pancreaticoduodenectomy: a pilot study. J Gastrointest Surg
attributable mortality, excess length of hospitalization, and 2006; 10: 1072–1080.
extra costs. Infect Control Hosp Epidemiol 1999; 20: 725–730. 17 Salvia R, Malleo G, Marchegiani G, Butturini G, Esposito A,
5 De Pastena M, Paiella S, Marchegiani G, Malleo G, Bassi C. Pancreaticoduodenectomy with harmonic focust
Ciprani D, Gasparini C et al. Postoperative infections curved shears for cancer. Dig Surg 2014; 31: 249–254.
represent a major determinant of outcome after 18 Chang CM, Yin WY, Wei CK, Wu CC, Su YC, Yu CH et al.
pancreaticoduodenectomy: results from a high-volume Adjusted age-adjusted Charlson comorbidity index score as a
center. Surgery 2017; 162: 792–801. risk measure of perioperative mortality before cancer
6 Ceppa EP, Pitt HA, House MG, Kilbane EM, Nakeeb A, surgery. PLoS One 2016; 11: e0148076.
Schmidt CM et al. Reducing surgical site infections in 19 Berríos-Torres SI, Umscheid CA, Bratzler DW, Leas B,
hepatopancreatobiliary surgery. HPB (Oxford) 2013; 15: Stone EC, Kelz RR et al.; Healthcare Infection Control
384–391. Practices Advisory Committee. Centers for Disease Control
7 Valle JW, Palmer D, Jackson R, Cox T, Neoptolemos JP, and Prevention guideline for the prevention of surgical site
Ghaneh P et al. Optimal duration and timing of adjuvant infection, 2017. JAMA Surg 2017; 152: 784–791.
chemotherapy after definitive surgery for ductal 20 Dindo D, Demartines N, Clavien PA. Classification of
adenocarcinoma of the pancreas: ongoing lessons from the surgical complications: a new proposal with evaluation in a
ESPAC-3 study. J Clin Oncol 2014; 32: 504–512. cohort of 6336 patients and results of a survey. Ann Surg
8 Swenson BR, Hedrick TL, Metzger R, Bonatti H, Pruett 2004; 240: 205–213.
TL, Sawyer RG. Effects of preoperative skin preparation on 21 Bassi C, Marchegiani G, Dervenis C, Sarr M, Abu Hilal M,
postoperative wound infection rates: a prospective study of 3 Adham M et al.; International Study Group on Pancreatic
skin preparation protocols. Infect Control Hosp Epidemiol Surgery (ISGPS). The 2016 update of the International
2009; 30: 964–971. Study Group (ISGPS) definition and grading of
9 Sugiura T, Uesaka K, Ohmagari N, Kanemoto H, postoperative pancreatic fistula: 11 years after. Surgery 2017;
Mizuno T. Risk factor of surgical site infection after 161: 584–591.
pancreaticoduodenectomy. World J Surg 2012; 36: 22 Wente MN, Bassi C, Dervenis C, Fingerhut A, Gouma DJ,
2888–2894. Izbicki JR et al. Delayed gastric emptying (DGE) after
10 Berenguer CM, Ochsner MG Jr, Lord SA, Senkowski CK. pancreatic surgery: a suggested definition by the
Improving surgical site infections: using National Surgical International Study Group of Pancreatic Surgery (ISGPS).
Quality Improvement Program data to institute Surgical Surgery 2007; 142: 761–768.
Care Improvement Project protocols in improving surgical 23 Wente MN, Veit JA, Bassi C, Dervenis C, Fingerhut A,
outcomes. J Am Coll Surg 2010; 210: 737–741, 741–743. Gouma DJ et al. Postpancreatectomy hemorrhage (PPH): an
11 Reid K, Pockney P, Draganic B, Smith SR. Barrier wound International Study Group of Pancreatic Surgery (ISGPS)
protection decreases surgical site infection in open elective definition. Surgery 2007; 142: 20–25.

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd
Prevention of surgical-site infection 1113

24 DeMets DL, Lan KK. Interim analysis: the alpha spending 30 Poruk KE, Lin JA, Cooper MA, He J, Makary MA,
function approach. Stat Med 1994; 13: 1341–1352. Hirose K et al. A novel, validated risk score to predict
25 Lachin JM. A review of methods for futility stopping based surgical site infection after pancreaticoduodenectomy. HPB
on conditional power. Stat Med 2005; 24: 2747–2764. (Oxford) 2016; 18: 893–899.
26 Morikane K. Epidemiology and risk factors associated with 31 Liu JB, Baker MS, Thompson VM, Kilbane EM, Pitt HA.
surgical site infection after different types of hepatobiliary Wound protectors mitigate superficial surgical site
and pancreatic surgery. Surg Today 2017; 47: 1208–1214. infections after pancreatoduodenectomy. HPB (Oxford)
27 Joliat GR, Sauvain MO, Petermann D, Halkic N, 2019; 21: 121–131.
Demartines N, Schäfer M. Surgical site infections after 32 Bressan AK, Aubin JM, Martel G, Dixon E, Bathe OF,
pancreatic surgery in the era of enhanced recovery Sutherland FR et al. Efficacy of a dual-ring wound
protocols. Medicine (Baltimore) 2018; 97: e11728. protector for prevention of surgical site infections after
28 Lavu H, Klinge MJ, Nowcid LJ, Cohn HE, Grenda DR, pancreaticoduodenectomy in patients with intrabiliary
Sauter PK et al. Perioperative surgical care bundle reduces stents: a randomized clinical trial. Ann Surg 2018; 268:
pancreaticoduodenectomy wound infections. J Surg Res 35–40.
2012; 174: 215–221. 33 Lawrence SA, McIntyre CA, Pulvirenti A, Seier K,
29 Elliott IA, Chan C, Russell TA, Dann AM, Williams JL, Chou Y, Gonen M et al. Perioperative bundle to reduce
Damato L et al. Distinction of risk factors for superficial surgical site infection after pancreaticoduodenectomy: a
versus organ-space surgical site infections after pancreatic prospective cohort study. J Am Coll Surg 2019; 228:
surgery. JAMA Surg 2017; 152: 1023–1029. 595–601.

© 2020 BJS Society Ltd www.bjs.co.uk BJS 2020; 107: 1107–1113

Published by John Wiley & Sons Ltd