Scandinavian Journal of Gastroenterology, 2011; 46: 261270

REVIEW ARTICLE

Systematic review and meta-analysis of antibiotic prophylaxis in severe

acute pancreatitis
MATHIAS WITTAU1, BENJAMIN MAYER2, JAN SCHEELE1, DORIS HENNE-BRUNS1,
E. PATCHEN DELLINGER3 & RAINER ISENMANN4
1

Department of General, Visceral and Transplantation Surgery, University Hospital, Ulm, Germany, 2Department of
Biometry, University of Ulm, Ulm, Germany, 3Department of Surgery, Division of General Surgery, University of
Washington, Seattle, WA, USA, and 4Department of Visceral Surgery, St. Anna Virngrundklinik, Ellwangen,
Germany

Abstract
Objective. The incidence of acute pancreatitis varies from 5 to 80 per 100,000 throughout the world. The most common
cause of death in these patients is infection of pancreatic necrosis by enteric bacteria, spurring the discussion of whether or
not prophylactic antibiotic administration could be a benecial approach. In order to provide evidence of the effect of
antibiotic prophylaxis in severe acute pancreatitis (SAP) we performed an updated systematic review and meta-analysis on
this topic. Methods. The review of randomized controlled trials was performed in accordance with the Preferred Reporting
Items for Systematic Reviews and Meta-analysis (PRISMA) statement. We conducted a search of MEDLINE, EMBASE,
and the Cochrane Central Register of Controlled Trials. For assessment of the treatment effects we calculated the risk ratios
(RRs) for dichotomous data of included studies. Results. Fourteen trials were included with a total of 841 patients. The use
of antibiotic prophylaxis was not associated with a statistically signicant reduction in mortality (RR 0.74 [95% CI 0.50
1.07]), in the incidence of infected pancreatic necrosis (RR 0.78 [95% CI 0.601.02]), in the incidence of nonpancreatic infections (RR 0.70 [95% CI 0.461.06]), and in surgical interventions (RR 0.93 [95% CI 0.721.20]).
Conclusion. In summary, to date there is no evidence that supports the routine use of antibiotic prophylaxis in patients
with SAP.

Key Words: Acute pancreatitis, antibiotic prophylaxis, pancreatitis, prophylactic antibiotics

Introduction
The most common cause of death in patients suffering
from severe acute pancreatitis (SAP) is the infection of
pancreatic necrosis by enteric bacteria with mortality
rates of 30% (range 1462%) [1,2], spurring the
discussion of whether or not prophylactic antibiotic
administration could be a benecial approach. Randomized controlled clinical trials (RCTs) on this topic
published in the 1990s have shown lower rates of
infected pancreatic necrosis and reduced mortality
[3,4]. Contrary to this, statistically signicant positive
effects of this prophylaxis could not be conrmed by

three double blind randomized studies, published in

2004, 2007 and 2009 [1,5,6]. Several metaanalyses on this issue were published since the late
1990s and intended to provide reliable evidence.
However, their results ranged from absolutely no
effect of antibiotic prophylaxis to positive effects
regarding mortality, the incidence of infected pancreatic necrosis and the incidence of extrapancreatic
infections.
In order to provide reliable evidence of the effect of
antibiotic prophylaxis in SAP, we performed an
updated systematic review and a meta-analysis on
this topic.

Correspondence: Mathias Wittau, MD, Department of General, Visceral and Transplantation Surgery, University Hospital Ulm, Steinhoevelstr. 9, 89075 Ulm,
Germany. Tel: +49 73150053500. Fax: +49 73150053503. E-mail: mathias.wittau@uniklinik-ulm.de

(Received 13 July 2010; accepted 19 September 2010)

ISSN 0036-5521 print/ISSN 1502-7708 online
2011 Informa Healthcare
DOI: 10.3109/00365521.2010.531486

262

M. Wittau et al.

Methods
Data sources and study selection
A detailed protocol was developed prior to study
initiation. The review of RCTs was performed in
accordance with the Preferred Reporting Items for
Systematic Reviews and Meta-analysis (PRISMA)
statement [7]. We conducted a search of MEDLINE
(1966 to December 2009), MEDLINE in-process
and other non-indexed citations (December 2009),
EMBASE (1980 to December 2009), and the
Cochrane Central Register of Controlled Trials
(4th Quarter 2009) via PubMed and Ovid interfaces.
The following medical subject heading terms
pancreatitis, antibiotics, and randomized controlled trial and text terms were used with Boolean
operators NOT, AND and OR, as demonstrated in Appendix A for MEDLINE. Reference lists
from the trials selected by electronic searching and
conference abstracts from 2005 to 2009 (European
Pancreatic Club, American Pancreatic Association,
International Association of Pancreatology, United
European Gastroenterology Week, and Digestive Disease Week) were hand searched. Google Scholar
was
used to search gray literature sources and Controlledtrials.com/ClinicalTrials.gov for ongoing registered
clinical trials. Search was performed without language
restrictions non-English/non-German manuscripts
were translated.
Inclusion criteria
.
.

.
.

Study design: RCT antibiotic prophylaxis versus

placebo or best supportive care.
Study population: patients with severe acute
pancreatitis. SAP was diagnosed with any of
the following: Ranson score, Acute Physiology
and Chronic Health Evaluation II score, Imrie
score, C-reactive protein levels >120 mg/L,
Balthazar CT grade E [8], Atlanta criteria [9]
or according to the Bangkok Working Party
Report 2002 [10].
Intervention: intravenous prophylactic antibiotic
administration.
Outcome data of mortality and incidence of
infected pancreatic necrosis available.

form: rst author, year of publication, number of

patients in the treatment and control arms, setting
(single-center/multicenter), blinding, name of antibiotic, dose and timing of antibiotic administration,
incidence of mortality, incidence of infected pancreatic necrosis, incidence of non-pancreatic infections,
and need for surgery. Disagreements over values or
analyses were resolved by discussion and arbitration
of a third author (R. I.).

Assessment of internal validity

The internal validity of included RCTs was appraised
by one reviewer and checked by a second using
the criteria of the Cochrane Collaboration (sequence
generation, concealment of allocation, blinding,
incomplete outcome data, and selective outcome
reporting) [11,12].

Data synthesis and analysis

Analyses were performed using the Review Manager
software version 5.0 for Windows (The Cochrane
Collaboration). For assessment of the treatment
effects, we calculated the risk ratios (RRs) for dichotomous data of included studies with their corresponding 95% condence intervals (CIs). To
calculate an overall effect size, we combined the
single RRs with the xed effects MantelHaenszel
method, which assume homogeneity among the
studies [13]. In case of heterogeneity, the random
effects method according to DerSimonian and Laird
was used to calculate the overall effect [14]. The
decision for calculation with the xed or random
effects method was made by the I2 test. An I2 greater
than 50% indicates heterogeneity between the
included studies, and the random effects method
was used for calculation. p 0.05 was considered
statistically signicant. For qualitative evaluation of
publication bias we used funnel plots. The quantitative assessment of the presence of publication bias
was performed with the Egger et al. [15] and
Begg et al. [16] tests. A sensitivity analysis was
conducted for class of antibiotic, risk of bias in the
trials, and blinding.

Data extraction

Results

Data were independently extracted by two reviewers

(M. W. and J. S.). Each article was reviewed to
determine whether it met the predened inclusion
criteria. Reviewers extracted the following data from
each report using a standardized data collection

The process of identifying relevant trials is shown

in Figure 1. A total of 1633 literature records were
retrieved by initial search. After removing duplicates
and articles not related to pancreatitis or antibiotic
prophylaxis or humans, 24 articles were assessed for

Evidence of antibiotic prophylaxis in severe acute pancreatitis

Records identified through database
searching:
EMBASE (n = 1039)
Medline + Medline in process (n = 430)
Cochrane Central Register of Controlled
Trials (n = 159)

263

Additional records identified

through other sources
(n = 5)

Records after duplicates removed

(n = 1426)

Records screened
(n = 1426)

Records excluded
(n = 1402)

Articles assessed for

eligibility
(n = 24)

Articles excluded:
Antibiotics vs. antibiotics (n = 5)
Antibiotic timing (n = 1)
RCT with unclear/different severity
(n = 3)
Subgroup analysis of 2 RCTs (n = 1)

Studies included in
qualitative synthesis
(n = 14)

Studies included in
quantitative synthesis
(meta-analysis)
(n = 14)

Figure 1. Flow diagram: meta-analysis of RCTs in patients with severe acute pancreatitis.

eligibility. Ten articles were judged not eligible: three

early trials including different severities of disease and
ampicillin prophylaxis [1719], one trial on antibiotic
timing [20], and ve trials comparing two or more
different antibiotic drugs or different dosing duration
[2125]. One record was a report of a subgroup
analysis of patients with proven necrosis of two randomized trails [26]. The data of this report were not
included in the analysis. Finally, 14 trials were
included in the qualitative and quantitative analysis
[1,36,27,2835].

quinolone-based prophylaxis imidazole, one trial

used a combination of ceftazidime, amikacin and
imidazole, and one trial used cefuroxime for prophylaxis. The 14 RCTs comprised a total of 841 patients
with SAP. Of these, 420 patients were randomized to
the prophylaxis group and 421 to the control group.
Data regarding mortality, incidence of infection of
pancreatic necrosis and non-pancreatic infections and
surgical intervention are summarized in Table I. No
important baseline differences between the treatment
groups were identied.

Study characteristics

Study validity

Of the 14 included RCTs (Table I), nine were singlecenter and ve mulitcenter trials. Eight trials use a
carbapenem-based prophylaxis, four studies a

The internal validity of the trials was heterogeneous

(Table II). Three trials were judged having low risk of
bias and 11 trials having high risk of bias. Three

1993

1995

1996

1997

2001

2002

2003

2004

2007

2007

2009

2009

2009

2009

Pederzoli et al. [3]

Sainio et al. [4]

Delcenserie et al. [29]

Schwarz et al. [30]

Nordback et al. [31]

Spicak et al. [32]

Spicak et al. [33]

Isenmann et al. [1]

Rkke et al. [27]

Dellinger et al. [5]

Barreda et al. [34]

Garcia-Barrasa et al. [6]

Xue et al. [35]

Yang et al. [28]

Number

Number

Double- blinded

Number

Double- blinded

Number

Double-blinded

Number

Number

Number

Number

Number

Number

Number

Blinded

Antibiotic type
Imipenem
None
Cefuroxime
None
Ceftazidime + Amikacin + Metronidazole
None
Ooxacin + Metronidazole
None
Imipenem (primary)
Imipenem (secondary)
Ciprooxacin + Metronidazole
Ciprooxacin + Metronidazole (on demand)
Meropenem
Meropenem (on demand)
Ciprooxacin + Metronidazole
None
Imipenem
None
Meropenem
None
Imipenem
None
Ciprooxacin
None
Imipenem
None
Imipenem
None

Data provided by author after personal communication (O. Rkke).

Year

Study

Table I. Characteristics of included studies.

41
33
30
30
11
12
13
13
25
33
33
30
20
21
58
56
36
37
50
50
24
34
22
19
29
27
28
26

n
3
4
1
7
1
3
0
2
2
5
5
3
4
5
3
4
3
4
10
9
0
0
4
2
3
4
2
3

Mortality,
n
5
10
9
12
0
3
8
7
1
6
1
0
3
6
7
5
3#
6#
9
6
3
2
8
8
8
10
6
8

Infected necrosis,
n
12
11
7
14
0
3
NR
NR
2
5
6
7
4
5
10
6
3
3
13
10
4
2
11
8
9
9
4
5

Surgery,
n
6
16
NR
NR
0
4
4
6
NR
NR
NR
NR
NR
NR
13
13
3
12
16
24
7
5
6
8
NR
NR
13
9

Non-pancreatic
infections, n

single-center

single-center

single-center

single-center

multicenter

multicenter

multicenter

multicenter

multicenter

single-center

single-center

single-center

single-center

single-center

Setting

264
M. Wittau et al.

Evidence of antibiotic prophylaxis in severe acute pancreatitis

265

Table II. Assessment of internal validity.

Study/Author

Year

Adequate sequence
generation

Concealed
allocation

Blinding

Addressing incomplete
outcome data

Free from selective

outcome reporting

Pederzoli et al. [3]

Sainio et al. [4]
Delcenserie et al. [29]
Schwarz et al. [30]
Nordback et al. [31]
Spicak et al. [32]
Spicak et al. [33]
Isenmann et al. [1]
Rkke et al. [27]
Dellinger et al. [5]
Barreda et al. [34]
Garcia-Barrasa et al. [6]
Xue et al. [35]
Yang et al. [28]

1993
1995
1996
1997
2001
2002
2003
2004
2007
2007
2009
2009
2009
2009

Yes
Unclear
Yes
Unclear
Unclear
No
Unclear
Yes
Yes
Yes
Unclear
Yes#
Yes
Yes

Unclear
Yes
Unclear
Unclear
Unclear
Unclear
Unclear
Yes
Unclear
Yes
Unclear
Yes#
Unclear
Unclear

No
No
No
No
No
No
No
Yes
No
Yes
No
Yes
No
No

Yes
Yes
Unclear
Yes
Yes
Unclear
Unclear
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Details provided by author after personal communication (J. Busquets).

studies were double-blinded whereas 11 trials were

unblinded.

bias regarding the outcome parameter mortality

(p = 0.11).

Evaluation of publication bias

Mortality

We found no clear evidence of publication bias

(Figure 2). Only the funnel plot of the outcome
parameter mortality did not show absolute symmetry, suggesting a possibility of publication bias.
The Begg et al. [16] test indicated the presence of publication bias (p = 0.023) while the Egger
et al. [15] test did not show a publication

There were 841 patients involved in 14 trials

(Figure 3). Ninety-six patients died, 41/420 (9.7%)
in the prophylactic antibiotic group and 55/421 (13%)
in the control group. In the trial of Barreda et al. [34]
no patient died. Therefore, in this trial the RR was not
estimable. The use of antibiotic prophylaxis was not
associated with a statistically signicant reduction in

SE (log[RR])

0.5

1.5

RR

2
0.2
Figure 2. Funnel plot: infected necrosis.

0.5

266

M. Wittau et al.

Study or Subgroup

Risk Ratio
Prophylaxis
Control
Events Total Events Total Weight M-H, Fixed, 95% Cl

Barreda et al 09
Delcenserie et al 96
Dellinger et al 07
Garcia-Barrasa 09
Isenmann et al 04
Nordback et al 01
Pederzoli et al 93
Rokke et al 07
Sainio et al 95
Schwarz et al 97
Spicak et al 02
Spicak et al 03
Xue et al 09
Yang et al 09

24
11
50
22
58
25
41
36
30
13
33
20
29
28

0
1
10
4
3
2
3
3
1
0
5
4
3
2

0
3
9
2
4
5
4
4
7
2
3
5
4
3

34
12
50
19
56
33
33
37
30
13
30
21
27
26

420
421
Total (95% Cl)
Total events
41
55
Heterogeneity: Chiz = 7.25, df = 12 (p = 0.84); Iz = 0%
Test for overall effect: Z = 1.59 (p = 0.11)

5.2%
16.2%
3.9%
7.3%
7.8%
8.0%
7.1%
12.6%
4.5%
5.7%
8.8%
7.5%
5.6%

Not estimable
0.36 [0.04, 3.00]
1.11 [0.49, 2.50]
1.73 [0.35, 8.41]
0.72 [0.17, 3.09]
0.53 [0.11, 2.50]
0.60 [0.15, 2.51]
077 [0.19, 3.20]
0.14 [0.02, 1.09]
0.20 [0.01,3.80]
1.52 [0.40, 5.81]
0.84 [0.26, 2.69]
0.70 [0.17, 2.84]
0.62 [ 0.11, 3.41]

100.0%

0.74 [0.50, 1.07]

Risk Ratio
M-H, Fixed, 95% Cl

0.2
0.5
1
2
5
Favours Experimental Favours Control

Figure 3. Forest plot of relative risk; [95% condence intervals]: mortality.

mortality (xed effect model, pooled RR 0.74 [95%

CI 0.501.07]). There was no heterogeneity between
the trials (p = 0.84; I2 = 0%).

was not associated with a statistically signicant

reduction in the incidence of infected pancreatic
necrosis (xed effect model, pooled RR 0.78 [95%
CI 0.601.02]). There was no heterogeneity between
the trials (p = 0.58; I2 = 0%).

Infected necrosis
Data of the incidence of infected necrosis were available in all trials (Figure 4). Of 841 patients, 160 developed infected pancreatic necrosis, 71/420 (17%) in
the prophylactic antibiotic group and 89/421 (21%) in
the control group. The use of antibiotic prophylaxis

Non-pancreatic infection
Nine trials reported data of the incidence of nonpancreatic infections (Figure 5) [1,3,5,6,2730,34].
A total of 563 patients were included in the analysis.

Risk Ratio
Prophylaxis
Control
Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% Cl
Barreda et al 09
Delcenserie et al 96
Dellinger et al 07
Garcia-Barrasa 09
Isenmann et al 04
Nordback et al 01
Pederzoli et al 93
Rokke et al 07
Sainio et al 95
Schwarz et al 97
Spicak et al 02
Spicak et al 03
Xue et al 09
Yang et al 09

3
0
9
8
7
1
5
3
9
8
1
3
8
6

24
11
50
22
58
25
41
36
30
13
33
20
29
28

2
3
6
8
5
6
10
6
12
7
0
6
10
8

Risk Ratio
M-H, Fixed, 95% Cl

34 1.8% 2.13 [0.38, 11.76]

12 3.7% 0.15 [0.01, 2.70]
50 6.6% 1.50 [0.58, 3.90]
19 9.4% 0.86 [0.40, 1.85]
56 5.6% 1.35 [0.46, 4.01]
33 5.7% 0.22 [0.03, 1.71]
33 12.2% 0.40 [0.15, 1.06]
37 6.5% 0.51 [0.14, 1.90]
30 13.2% 0.75 [0.37, 1.51]
13 7.7% 1.14 [0.59, 2.22]
30 0.6% 2.74 [0.12, 64.69]
21 6.4% 0.53 [0.15, 1.82]
27 11.4% 0.74 [0.35, 1.61]
26 9.1% 0.70 [0.28, 1.74]

420
421 100.0%
Total (95% Cl)
Total events
71
89
Heterogeneity: Chiz = 11.39, df = 13 (p = 0.58); Iz = 0%
Test for overall effect: Z = 1.81 (p = 0.07)

0.78 [0.60, 1.02]

0.2
0.5
1
2
5
Favours Experimental Favours Control

Figure 4. Forest plot of relative risk; [95% condence intervals]: infected pancreatic necrosis.

Evidence of antibiotic prophylaxis in severe acute pancreatitis

Prophylaxis
Study or Subgroup
Barreda et al 09
Delcenserie et al 96
Dellinger et al 07
Garcia-Barrasa 09
Isenmann et al 04
Pederzoli et al 93
Rokke et al 07
Schwarz et al 97
Yang et al 09

Risk Ratio

Control

Risk Ratio

Events Total Events Total Weight M-H, Random, 95% Cl

7
0
16
6
13
6
3
4
13

24
11
50
22
58
41
36
13
28

5
4
24
8
13
16
12
6
9

34
12
50
19
56
33
37
13
26

9.7%
2.0%
17.3%
11.6%
14.4%
12.2%
8.2%
9.9%
14.6%

283
280 100.0%
Total (95% Cl)
Total events
68
97
Heterogeneity: Tauz = 0.20; Chiz = 17.12, df = 8 (p = 0.03); Iz = 53%
Test for overall effect: Z = 1.67 (p = 0.10)

267

M-H, Random, 95% Cl

1.98 [0.71, 5.51]

0.12 [0.01, 2.01]
0.67 [0.41, 1.10]
0.65 [0.27, 1.53]
0.97 [0.49, 1.90]
0.30 [0.13, 0.68]
0.26 [0.08, 0.84]
0.67 [0.24, 1.82]
1.34 [0.69, 2.60]
0.70 [0.46, 1.06]

0.2
0.5
1
2
5
Favours Experimental
Favours Control

Figure 5. Forest plot of relative risk; [95% condence intervals]: non-pancreatic infections.

One hundred sixty-ve patients developed nonpancreatic infections, 68/283 (24%) in the prophylactic antibiotic group and 97/280 (35%) in the
control group. The use of antibiotic prophylaxis
was not associated with a statistically signicant
reduction in the incidence of non-pancreatic infections (random effect model, pooled RR 0.70 [95% CI
0.461.06]). There was a statistically signicant heterogeneity between the trials (p = 0.03; I2 = 53%).

A total of 815 patients were included in the analysis.

One hundred seventy-three patients underwent surgery, 85/407 (21%) in the prophylactic antibiotic
group and 88/408 (21.5%) in the control group.
The use of antibiotic prophylaxis was not associated
with a statistically signicant reduction in surgical
interventions (xed effect model, pooled RR 0.93
[95% CI 0.721.20]). There was no heterogeneity
between the trials (p = 0.67; I2 = 0%).

Surgical intervention

Sensitivity analysis

Data of the number of surgical interventions were

available in 13 trials (Figure 6) [1,36,2729,3135].

In order to determine whether class of antibiotic, risk

of bias in the trials, or blinding had a signicant effect

Study or Subgroup

Risk Ratio
Prophylaxis
Control
Events Total Events Total Weight M-H, Fixed, 95% Cl
1.8%
3.7%
11.1%
9.6%
6.8%
4.8%
13.6%
3.3%
15.6%
8.2%
5.4%
10.4%
5.8%

2.83 [0.56, 14.24]

0.15 [0.01, 2.70]
1.30 [0.63, 2.68]
1.19 [0.61, 2.33]
1.61 [0.63, 4.13]
0.53 [0.11, 2.50]
0.88 [0.45, 1.73]
1.03 [0.22, 4.76]
0.50 [0.24, 1.06]
0.78 [0.29, 2.06]
0.84 [0.26, 2.69]
0.93 [0.44, 1.99]
0.74 [0.22, 2.47]

Total (95% Cl)

407
408 100.0%
Total events
85
88
Heterogeneity: Chiz = 9.42, df = 12 (p = 0.67); Iz = 0%
Test for overall effect: Z = 0.55 (p = 0.58)

0.93 [0.72, 1.20]

Barreda et al 09
Delcenserie et al 96
Dellinger et al 07
Garcia-Barrasa 09
Isenmann et al 04
Nordback et al 01
Pederzoli et al 93
Rokke et al 07
Sainio et al 95
Spicak et al 02
Spicak et al 03
Xue et al 09
Yang et al 09

4
0
13
11
10
2
12
3
7
6
4
9
4

24
11
50
22
58
25
41
36
30
33
20
29
28

2
3
10
8
6
5
11
3
14
7
5
9
5

34
12
50
19
56
33
33
37
30
30
21
27
26

Figure 6. Forest plot of relative risk; [95% condence intervals]: surgical intervention.

Risk Ratio
M-H, Fixed, 95% Cl

0.2
0.5
1
2
5
Favours Experimental Favours Control

268

M. Wittau et al.

Study or Subgroup
Barreda et al 09
Delcenserie et al 96
Dellinger et al 07
Garcia-Barrasa 09
Isenmann et al 04
Nordback et al 01
Pederzoli et al 93
Rokke et al 07
Sainio et al 95
Schwarz et al 97
Spicak et al 02
Spicak et al 03
Xue et al 09
Yang et al 09

Risk Ratio
Prophylaxis
Control
Events Total Events Total Weight M-H, Fixed, 95% Cl
3
0
9
8
7
1
5
3
9
8
1
3
8
6

24
11
50
22
58
25
41
36
30
13
33
20
29
28

2
3
6
8
5
6
10
6
12
7
0
6
10
8

34
12
50
19
56
33
33
37
30
13
30
21
27
26

0.0%
0.0%
30.5%
43.6%
25.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%

2.13 [0.38, 11.76]

0.15 [0.01, 2.70]
1.50 [0.58, 3.90]
0.86 [0.40, 1.85]
1.35 [0.46, 4.01]
0.22 [0.03, 1.71]
0.40 [0.15, 1.06]
0.51 [0.14, 1.90]
0.75 [0.37, 1.51]
1.14 [0.59, 2.22]
2.74 [0.12, 64.69]
0.53 [0.15, 1.82]
0.74 [0.35, 1.61]
0.70 [0.28, 1.74]

130
125
Total (95% Cl)
Total events
24
19
Heterogeneity: Chiz = 0.95, df = 2 (p = 0.62); Iz = 0%
Test for overall effect: Z = 0.63 (p = 0.53)

100.0%

1.18 [0.70, 2.01]

Risk Ratio
M-H, Fixed, 95% Cl

0.2
0.5
1
2
5
Favours Experimental Favours Control

Figure 7. Forest plot of relative risk; [95% condence intervals]: infected pancreatic necrosis trials with low risk of bias.

on strength and direction of results, we performed a

sensitivity analysis.

Carbapenem-based prophylaxis
Eight trials used carbapenem-based prophylaxis
[3,5,27,28,31,3335]. This prophylaxis was not associated with a statistically signicant reduction in mortality (RR 0.80 [95% CI 0.501.27]), the incidence of
infected pancreatic necrosis (RR 0.69 [95% CI 0.48
1.01]), the incidence of non-pancreatic infections
(random RR 0.69 [95% CI 0.351.36]), and in surgical interventions (RR 1.00 [95% CI 0.711.40]).

Risk of bias in the trials and blinding

Three trials were of low risk of bias and doubleblinded [1,5,6]. There was not statistically signicant
reduction in mortality (RR 1.09 [95% CI 0.582.08]),
the incidence of infected pancreatic necrosis (RR 1.18
[95% CI 0.702.01]) (Figure 7), the incidence of
non-pancreatic infections (RR 0.75 [95% CI 0.52
1.08]), and in surgical interventions (RR 1.34 [95%
CI 0.862.09]).
Discussion
Antibiotic prophylaxis in SAP is a controversial issue
since nearly two decades. Several randomized trials
focused on this topic with contrary results ranging

from a positive effect on mortality, incidence of

infected necrosis and non-pancreatic infection to
absolute no effect on these outcome parameters. In
order to clarify this important issue we performed an
updated meta-analysis including all 2009 published
RCTs. In patients with SAP antibiotic prophylaxis is
not associated with a statistically signicant decrease
in mortality, the incidence of infected pancreatic
necrosis, the incidence of non-pancreatic infection,
or the need for surgery. Further, the sensitivity analyses could not nd any statistically signicant effect
on the direction of results regarding carbapenembased antibiotic prophylaxis (six trials with imipenem
and two trials with meropenem), risk of bias in the
trials, or blinding. These ndings are consistent with
those of Jafri et al. [36] for type of antibiotics and
higher-quality studies but, however, in contrast to
those of Villatoro et al. [37] for the parameter infected
pancreatic necrosis with imipenem prophylaxis.
Villatoro et al. found a statistically signicant decrease
of infected pancreatic necrosis with imipenem in an
analysis of three trails whereas in our analysis with
carbapenem prophylaxis no statistically signicant
decrease could be seen, nevertheless with borderline
signicance. Although this meta-analysis indicate a
borderline signicance for the outcome parameter
infected pancreatic necrosis with a pooled RR of 0.78
[95% CI 0.601.02], the likelihood that there was
statistically signicance for the important outcomes
of mortality or surgery was not even borderline
close to signicance, especially in higher-quality
double-blinded trials. Thus, the real effect for the

Evidence of antibiotic prophylaxis in severe acute pancreatitis

patient, dying or needing an operation, never comes
close to signicance.
Further, the relationship between methodological
quality of the trials and the outcome, as described by
de Vries et al. [38] could be conrmed by our sensitivity analysis, i.e. the higher quality trials were those
with the least potential effect of antibiotic prophylaxis,
indicating less likelihood of missing a real effect.
One possible limitation of this meta-analysis is
pooling data of patients without proven pancreatic
necrosis and patients with proven pancreatic necrosis.
But, even meta-analyses with the inclusion criterion
ceCT-proven necrosis included, at least in parts,
data of patients in which necrosis was not conrmed
by contrast-enhanced computer tomography (ceCT)
[39,36,37,4042]. Just one example, Sainio et al. [4]
stated in the result section of their publication that
. . .Five patients in the antibiotic group and seven in
the non-antibiotic group had plain CT only. . ., i.e.
only 25 patients in the antibiotic group and 23 patients
in the non-antibiotic group had ceCT-proven necrosis
and not 30 patients per group, included in these metaanalyses.
For a physician it is important to know whether an
antibiotic prophylaxis is benecial for a patient with
SAP, not only for a patient with ceCT-proven pancreatic necrosis. In clinical daily routine, it is often
impossible to perform a ceCT scan, e.g. due to
impaired kidney function.
A second possible limitation could be the timing of
antibiotic prophylaxis. L. Besselink et al. focused in a
cohort study on this topic and found that half of the
relevant infections occur in the rst few days of the
disease and that these infections, especially bacteraemia, are associated with an increased risk of developing infected necrosis and death [43]. On the basis
of these ndings, Besselink et al. [43] concluded that
prophylactic strategies should focus on early intervention. However, Manes and colleagues also addressed
this issue in a recent published randomized trial and
found only a statistically signicant decrease of extrapancreatic infection with antibiotic prophylaxis at
hospital admission [20]. The analysis of our included
trials showed in most cases imprecise data of timing of
prophylaxis (within 3 days, within 2 days, etc.)
and start of antibiotic prophylaxis ranged in the mean
from 1.5 to 5 days. Thus, a conclusive answer could
not be drawn from these data.
In summary, to date there is no statistically significant evidence that supports the routine use of antibiotic prophylaxis in SAP. However, in case of newly
developed sepsis or SIRS, newly developed failure of
two or more organ systems, proven pancreatic or
extrapancreatic infection, and an increase in serum
C-reactive protein in combination with evidence of

269

pancreatic or extrapancreatic infection an antibiotic

treatment should be considered, as recommended by
Isenmann et al. [1]. Nordback et al. demonstrated
that this on demand treatment might avoid or
postpone surgery [31].
Further, it is possible that a subgroup of patients
prot from an antibiotic prophylaxis, e.g. with large
pancreatic necrosis, but this is not evidence-based till
now. Thus, this issue and the impact of timing of
antibiotic prophylaxis should be the subject of further
well designed RCTs.
Declaration of interest: The authors report no
conicts of interest. The authors alone are responsible
for the content and writing of the paper.

References
[1] Isenmann R, Rnzi M, Kron M, Kahl S, Kraus D,
Jung N, et al. Prophylactic antibiotic treatment in patients
with predicted severe acute pancreatitis a placebocontrolled, double-blind trial. Gastroenterology 2004;126:
9971004.
[2] Banks PA, Freeman ML. The Practice Parameters Committee of the American College of Gastroenterology: practice
guidelines in acute pancreatitis. Am J Gastroenterol 2006;
101:2379400.
[3] Pederzoli P, Bassi C, Vesentini S, Campedelli A.
A randomized multicenter clinical trial of antibiotic prophylaxis of septic complications in acute necrotizing pancreatitis
with imipenem. Surg Gynecol Obstet 1993;176:4803.
[4] Sainio V, Kemppainen E, Puolakkainen P, Taavitsainen M,
Kivisaari L, Valtonen V, et al. Early antibiotic treatment in
acute necrotizing pancreatitis. Lancet 1995;346:6637.
[5] Dellinger EP, Tellado JM, Soto NE, Ashles SW, Barie PS,
Dugernier T, et al. Early antibiotic treatment for severe acute
necrotizing
pancreatitis:
randomized,
double-blind,
placebo-controlled study. Ann Surg 2007;245:67483.
[6] Garcia-Barrasa A, Borobia F, Pallares R, Jorba R, Poves I,
Busquets J, et al. A double-blind, placebo-controlled trial of
ciprooxacin prophylaxis in patients with acute necrotizing
pancreatitis. J Gastrointest Surg 2009;13:76874.
[7] Moher D, Liberati A, Tetzlaff J, Altman DG, The
PRISMA Group (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA Statement.
Open Med 2009;3(3);123130.
[8] Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH. Acute
pancreatitis: value of CT in establishing prognosis. Radiology
1990;174:3316.
[9] Bradley EL III. A clinically based classication system for
acute pancreatitis. Summary of the International Symposium
on Acute Pancreatitis, Atlanta, Ga, September 11 through
13, 1992. Arch Surg 1993;128:58690.
[10] Toouli J, Brooke-Smith M, Bassi C, Carr-Locke D,
Telford J, Freeny P, et al. Working party of the program
committee of the Bangkok World Congress of Gastroenterology 2002. Guideline for the management of acute pancreatitis. J Gastroenterol Hepatol 2002;17(Suppl):s1539.
[11] Higgins JPT, Green S. Cochrane handbook for systematic
reviews of interventions 5.0.1. Chichester: JohnWiley &
Sons; 2008.

270

M. Wittau et al.

[12] Higgins JPT, Green S. Cochrane handbook for systematic

reviews of interventions 5.0.2 [updated September 2009].
The Cochrane Collaboration; 2008.
[13] Mantel N, Haenszel W. Statistical aspects of the analysis of
data from retrospective studies of disease. J Natl Cancer Inst
1959;22:71948.
[14] DerSimonian R, Laird N. Meta-analysis in clinical trials.
Control Clin Trials 1986;7:17788.
[15] Egger M, Davey SG, Schneider M, Minder C. Bias in
meta-analysis detected by a simple, graphical test. BMJ
1997;315:62934.
[16] Begg CB, Mazumdar M. Operating characteristics of a rank
correlation test for publication bias. Biometrics 1994;50:
1088101.
[17] Craig RM, Dordal E, Myles L. The use of ampicillin in acute
pancreatitis. Ann Intern Med 1975;83:8312.
[18] Howes R, Zuidema GD, Cameron JL. Evaluation of prophylactic antibiotics in acute pancreatitis. J Surg Res 1975;
18:197200.
[19] Finch WT, Sawyers JL, Schenker S. A prospective study to
determine the efcacy of antibiotics in acute pancreatitis.
Ann Surg 1976;183:66771.
[20] Manes G, Uomo I, Menchise A, Rabitti PG, Ferrara EC,
Uomo Gl. Timing of antibiotic prophylaxis in acute pancreatitis. A controlled randomized study with meropenem. Am J
Gastroenterol 2006;101:134853.
[21] Manes G, Rabitti PG, Menchise A, Ricco E,
Balzano A Uomo G. Prophylaxis with meropenem of septic
complications in acute pancreatitis: a randomized, controlled
trial versus imipenem. Pancreas 2003;27;7983.
[22] Maravi-Poma E, Gener J, Alvarez-Loma F, Olaechea P,
Blanco A, Dominguez-Munoz JE. Early antibiotic treatment
(prophylaxis) of septic complications in severe acute necrotizing pancreatitis: a prospective, randomized, multicenter
study comparing two regimes with imipenem-cilastatin.
Intensive Care Med 2003;29:197480.
[23] Bassi C, Falconi M, Talamini G, Uomo G, Papaccio G,
Dervenis C, et al. Controlled clinical trial of peoxacin versus
imipenem in severe acute pancreatitis. Gastroenterology
1998;115:151317.
[24] Swidnicka-Siergiejko A, Wereszczyska-Siemiatkowska U,
Daabrowski A. The efcacy of different modes of antibiotic prophylaxis in the prevention of complications in
severe acute pancreatitis. Gastroenterologia Polska 2007;
14:17983.
[25] Grezebieniak Z, Rudnicki J, Skalski A. Evaluation of selected
antibiotics in the treatment of acute pancreatitis. Adv Clin
Exp Med 2003;12:42333.
[26] Spicak J, Hejtmankova S, Cech P, Hoskovec D, Kostka R,
Lefer J, et al. Antibiotic prophylaxis in large pancreatic
necrosis: multicenter randomized trial with ciprooxacin
and metronidazole or meropenem. Gastroenterology 2004;
126:A-229.
[27] Rkke O, Harbitz TB, Liljedal J, Pettersen T, Fetvedt T,
Heen LO, et al. Early treatment of severe pancreatitis with
imipenem: a prospective randomized clinical trial. Scand J
Gastroenterol 2007;42:7716.
[28] Yang X, Deng LH, Xue P, Zhao L, Jin T, Wan M, et al.
Non-preventive use of antibiotics in patients with severe

Supplementary material available online

Appendix A.

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

acute pancreatitis treated with integrated traditional Chinese

and Western medicine therapy: a randomized controlled
trial. J Chin Interg Med 2009;7:3303.
Delcenserie R, Yzet T, Duccroix JP. Prophylactic antibiotics
in treatment of severe acute alcoholic pancreatitis. Pancreas
1996;13:198201.
Schwarz M, Isenmann R, Meyer H, Beger HG. Antibiotika
bei nekrotisierender Pankreatitis. Ergebnisse einer kontrollierten Studie. Dtsch Med Wochenschr 1997;122:35661.
Nordback I, Sand J, Saaristo R, Paajanen H. Early treatment
with antibiotics reduces the need for surgery in acute necrotizing pancreatitisa single-center randomized study.
J Gastrointest Surg 2001;5:11318.
Spicak J, Hubaczova M, Antos F, Bartova J, Cech P,
Kasalicky M, et al. Antibiotics in the treatment of acute
pancreatitis-ndings from a randomized multi-center prospective study. Ces Slov Gastroent Hepatol 2002;56:1839.
Spicak J, Hejtmankova S, Hubaczova M, Antos F, Bartova J,
Cech P, et al. Antibiotic prophylaxis of infectious complications of acute pancreatitis-the results of a randomised study
by meropenem. Ces Slov Gastroent Hepatol 2003;57:2227.
Barreda L, Targarono J, Milian W, Portugal J, Sequeiros J,
Pando E, et al. Is the prophylactic therapy with imipenem
effective for patients with pancreatic necrosis? Acta Gastroenterol Latinoam 2009;39:249.
Xue P, Deng LH, Zhang ZD, Yang XN, Wan MH,
Song B, et al. Effect of antibiotic prophylaxis on acute
necrotizing pancreatitis: results of a randomized controlled
trial. J Gastroenterol Hepatol 2009;24:73642.
Jafri NS, Mahid SS, Idstein SR, Hornung CA, Galandiuk S.
Antibiotic prophylaxis is not preventive in severe acute pancreatitis: a systematic review and meta-analysis. Am J Surg
2009;197:80613.
Villatoro E, Mulla M, Larvin M. Antibiotic therapy for
prophylaxis against infection of pancreatic necrosis in acute
pancreatitis. Cochrane Database Syst Rev 2010:CD002941.
de Vries A, Besselink MG, Buskens E, Ridwan B,
Schipper M, van Erpecum KJ, et al. Randomised controlled
trials of antibiotic prophylaxis in severe acute pancreatitis:
relationship between methodological quality and outcome.
Pancreatology 2007;7:5318.
Mazaki T, Ishii Y, Takayama T. Meta-analysis of prophylactic antibiotic use in acute necrotizing pancreatitis. Br J
Surg 2006;93:67484.
Villatoro E, Bassi C, Larvin M. Antibiotic therapy for prophylaxis against infection of pancreatic necrosis in acute
pancreatitis. Cochrane Database Syst Rev 2006:CD002941.
Bai Y, Gao J, Zou DW, Li Z. Prophylactic antibiotics cannot
reduce infected pancreatic necrosis and mortality in acute
necrotizing pancreatitis: evidence from a meta-analysis of
randomized controlled trials. Am J Gastroenterol 2008;
103:104110.
Heinrich S, Schafer M, Rousson V, Clavien PA. Evidencebased treatment of acute pancreatitis: a look at established
paradigms. Ann Surg 2006;243:15468.
Besselink MG, van Santvoort HC, Boermeester MA,
Nieuwenhuijs VB, van Goor H, Dejong CHC, et al. Timing
and impact of infections in acute pancreatitis. Br J Surg 2009;
96:26773.

Copyright of Scandinavian Journal of Gastroenterology is the property of Taylor & Francis Ltd and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.