Keywords
fosfomycin - gastrointestinal surgical procedure - systematic review - postoperative
complications - adverse effects
Introduction
A great challenge of modern medicine is the rise in antimicrobial resistance. The
World Health Organisation ranked antimicrobial resistance as one of the top 10
global public health threats [1]. A possible
solution to this threat could be the re-entry of the use of older antibiotics [2] such as fosfomycin. Fosfomycin, a
phosphoenolpyruvate analogue [3], is a
bactericidal antibiotic agent that interferes with the first step of the bacterial
cell wall synthesis, where it irreversibly inhibits the enzyme enolpyruvyl
transferase [4]. It has a half-life of
1.9–3.9 hours in plasma [5] and is
eliminated by unchanged excretion into the urine [6]. Fosfomycin has shown good tissue penetration, thus, reaching
sufficient concentrations to kill bacteria [5]. Fosfomycin is generally well-tolerated. Common harms reported after
intravenous administration of fosfomycin were gastrointestinal symptoms such as
nausea, vomiting, and diarrhoea, and less than 0.01% of the reported adverse
reactions would be classified as serious adverse events [7]. Fosfomycin has been used for prophylactic
treatment of urinary and abdominal infections and for treatment of multi-drug
resistant Gram-positive and Gram-negative aerobic bacteria and could thus be a
promising antimicrobial agent in gastrointestinal surgery [8]. It is, therefore, relevant to characterize
the clinical experience and evidence for the use of fosfomycin as prophylaxis or
empirical treatment in association with abdominal surgery.
We aimed to investigate if perioperative parenteral administration of fosfomycin
given before or during emergency or elective gastrointestinal surgery could protect
against postoperative infectious complications. Secondly, we aimed to characterize
the use of fosfomycin regarding dose, timing, surgical indication, and harms in
gastrointestinal surgery.
Materials and methods
This systematic review was reported according to Preferred Reporting Items for
Systematic reviews and Meta-Analysis (PRISMA) 2020 guideline [9]. Before data extraction, we registered a
protocol at PROSPERO (registration number: CRD42020201268) [10].
The eligibility criteria were participants undergoing emergency or elective
gastrointestinal surgery having received parenteral administration of fosfomycin
either before or during surgery and prescribed as either empirical antimicrobial
mono- or polytherapy. No comparison group was required, but the study design had to
be an original study covering any type of trial and studies where the number of
participants was ≥5, thus, the exclusion criterion was case reports
reporting on <5 patients.
The information sources included three databases: PubMed (1966–now), Embase
(1974–now), and Cochrane CENTRAL, and there were no limitations on dates or
languages. The search strategy that was used in these three databases was developed
together with a professional research librarian and the last day of the search was
March 24, 2023. The search strategy included search terms for fosfomycin combined
with terms on surgery or surgical procedures. The specific search string for PubMed
was: (((fosfomycin) OR phosphomycin)) AND ((((surgery) OR surgical) OR procedure)
OR
procedures). The adapted search string in Embase was: (fosfomycin.mp. or exp
fosfomycin/ or phosphomycin.mp.) and (exp abdominal surgery/ or exp
biliary tract surgery/ or exp colon surgery/ or exp colorectal
surgery/ or exp elective surgery/ or exp gastrointestinal
surgery/ or exp general surgery/ or exp surgery/ or
surgery.mp. or (operations or operated or operate or operation or operating).mp.).
Lastly, the adapted search string in Cochrane CENTRAL was: (fosfomycin OR
phosphomycin) AND ((operation OR operations OR operating) OR (procedure OR
procedure) OR (surgery OR surgically)). The systematic search was supplemented with
a screening of the reference lists of the included studies for relevant studies
(backward citation search).
The selection process was done using the screening tool Rayyan [11] and was divided into three parts. Firstly,
duplicates were removed. Secondly, two independent reviewers screened the titles and
abstracts of the reports against the eligibility criteria. Thirdly, two independent
reviewers screened full-text articles against the eligibility criteria. Furthermore,
we excluded studies when specific data on patients receiving fosfomycin were not
retrievable, and when fosfomycin was solely administered after gastrointestinal
surgery for sepsis or postoperative infectious complications.
Data extraction was done by one reviewer, and extracted data were checked for
accuracy by another reviewer. Data reported in other languages than English were
translated and discussed thoroughly during extraction. The extracted variables
included: characteristics of the studies, specification of fosfomycin
administration, postoperative complications and mortality, and reports on harms. If
possible, we also extracted data for any of the control groups. An exhausting list
of all variables can be seen in the PROSPERO protocol [10]. Due to sparse reporting of several
variables, these were left out of the final review.
Bias assessment was done independently by two authors or a contributor, so bias
assessment was only done by those not involved in the included study. The tools used
depended on the study design, thus, the risk of bias in randomised controlled trials
(RCTs) was assessed with the Cochrane Handbook risk of bias assessment tool 1 [12], and the risk of bias in observational
cohort studies was assessed with the Newcastle-Ottawa Scale (NOS) [13].
We primarily set out to characterise the use of fosfomycin administered parenterally
before or during gastrointestinal surgery, focusing on postoperative complications
such as surgical site infection (SSI), intraabdominal abscess, sepsis, and mortality
due to infectious complications. Furthermore, we wanted to characterise the use of
fosfomycin in emergency or elective surgery, dose and timing, all-cause mortality,
and harms. Synthesis of results across studies can only be applied if methodological
and clinical heterogeneity is low [14]. We
deemed it impossible in this review, as there were tremendous differences in
diseases, indication for and the type of surgery as well as the dose of fosfomycin
and the comparison groups. Therefore, we used ranges to summarise the data across
included studies and supplemented this with the specific extracted data from each
study presented in tables or figures together with proportions and 95%
confidence intervals (CI), using OpenMeta[Analyst] software [15]. For infectious complications, we presented
data within the two subgroups emergency and elective surgery.
As we did not assess a difference, e. g. neither effect nor risk, we did not
investigate the risk of reporting bias through funnel plots. As no formal
meta-analysis was conducted, the certainty of evidence could not be assessed with
the GRADE approach.
Results
After searching databases and removing duplicates, 3,511 articles were screened in
title and abstract and the selection process is depicted in [Fig. 1]. Finally, 15 unique studies [16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30] were included in this
systematic review as five reports had overlapping populations with the included
studies [31]
[32]
[33]
[34]
[35].
Fig. 1 A flowchart that shows the process of the systematic review
including screening of the articles’ title and abstract, screening
of full-text articles, reasons for exclusion of articles, and the total
number of included studies in this systematic review. *Five reports
were not included as they had overlapping populations with included studies
[31]
[32]
[33]
[34]
[35].
The characteristic of the included studies is depicted in [Table 1]. The majority of the included studies
were RCTs [16]
[20]
[21]
[22]
[23]
[24]
[26]
[27]
[28], three studies were
retrospective cohort studies [19]
[29]
[30],
and three studies were prospective cohort studies [17]
[18]
[25]. Nearly half of the studies were carried
out in the 1980s to early 1990s [21]
[22]
[23]
[24]
[25]
[26]
[27]. Almost half of the studies
were carried out in Scandinavia with four from Sweden [21]
[23]
[24]
[26] and three from Denmark [16]
[18]
[22], respectively. The
remaining studies were from Italy [27]
[28]
[29],
Japan [19]
[20], Germany [17]
[25], and Spain [30]. The included studies reported on a median (range) of 58
(12–517) patients each, totalling 1,029 patients receiving fosfomycin in
connection with gastrointestinal surgery ([Table
1]). The studies included patients of both sexes, however, one study only
included males [18] (Supplementary File
[Table 1]). The age of patients was sparsely
reported. Most of the patients treated with fosfomycin (51%) had undergone
an elective procedure [17]
[19]
[20]
[21]
[22]
[24]
[25]
[26], the most common being colorectal
procedures or cholecystectomy (Supplementary File
[Table 1]). Almost one-fourth of patients
treated with fosfomycin (23%) underwent an emergency procedure [16]
[18]
[23]
[30] including laparoscopic cholecystectomy or
appendectomy. The remaining three studies reported on a mixed population of patients
undergoing either an elective or emergency procedure [27]
[28]
[29].
Table 1 Characteristics of the included studies and fosfomycin
administration where n refers to the number of participants undergoing
abdominal surgery. IV: intravenous, IP: intraperitoneal, IM:
intramuscular, n: number of participants, Pro: prospective cohort study,
Q-RCT: Quasi-randomized controlled trial, RCT: randomized controlled
trial, Retro: retrospective cohort study.
|
Authors
|
Year
|
Study design
|
Type of surgery
|
Total n
|
Fosfomycin administration
|
|
n
|
Dose (g)
|
Total dose (g)
|
Route
|
|
Fonnes et al. [16]
|
2020
|
Q-RCT
|
Emergency
|
12
|
6
|
4
|
4
|
IP
|
|
Dorn et al. [17]
|
2019
|
Pro
|
Elective
|
27
|
27
|
8
|
8
|
IV
|
|
Fonnes et al. [18]
|
2019
|
Pro
|
Emergency
|
14
|
14
|
4
|
4
|
IP
|
|
Shinagawa et al. [19]
|
2006
|
Retro
|
Elective
|
162
|
68
|
4
|
6–10
|
IV
|
|
Unemura et al. [20]
|
2000
|
RCT
|
Elective
|
242
|
7
|
2
|
2–16
|
IV
|
|
Andåker et al. [21]
|
1992
|
RCT
|
Elective
|
517
|
259
|
8
|
16
|
IV
|
|
Nøhr et al. [22]
|
1990
|
RCT
|
Elective
|
149
|
72
|
8
|
8
|
IV
|
|
Andåker et al. [23]
|
1987
|
RCT
|
Emergency
|
381
|
190
|
4
|
4/16/64a
|
IV
|
|
Lindhagen et al. [24]
|
1984
|
RCT
|
Elective
|
49
|
26
|
2
|
8
|
IV
|
|
Müller et al. [25]
|
1982
|
Pro
|
Elective
|
40
|
40
|
4
|
4/8b
|
IV
|
|
Lindhagen et al. [26]
|
1981
|
RCT
|
Elective
|
58
|
30
|
2
|
32
|
IV
|
|
Cardia et al. [27]
|
1980
|
RCT
|
Mixed
|
25
|
12
|
1
|
4
|
IM
|
|
Bianca et al.[28]
|
1979
|
RCT
|
Mixed
|
263
|
129
|
1
|
3
|
IM
|
|
Germiniani et al. [29]
|
1979
|
Retro
|
Mixed
|
365
|
120
|
1
|
3
|
IM
|
|
Gallardo et al. [30]
|
1977
|
Retro
|
Emergency
|
29
|
29
|
4–6
|
20–30
|
IM
|
a Patients were divided into three groups (increasing severity of
disease): group A (only one preoperative dose)/ group B (one
preoperative dose and three postoperative doses)/ group C (one
preoperative dose and three postoperative doses for 5 days). b
Patients were divided into two groups: one dose before surgery
(n=23)/two doses 8–10 days after surgery
(n=17).
Fosfomycin and other antimicrobial agents
Regarding the administration of fosfomycin, the median (range) dose of fosfomycin
was 4 g (1–8 g) ([Table 1]).
Fosfomycin was administered before surgery in 73% of studies [17]
[19]
[20]
[21]
[22]
[24]
[25]
[26]
[27]
[28]
[29]
[36], the median was 1 hour
preoperatively but it ranged from 30 minutes to 6 hours preoperatively. Two
studies administered fosfomycin during the procedure [16]
[18]. In two-thirds of the studies, the postoperative course was also
supplemented with fosfomycin. The postoperative regimens listed from fewest to
most administrations were as follows: one administration of fosfomycin was given
in two studies 8 hours postoperatively [21], two administrations were given either at 5- and 12 hours [28]
[37], or 6- and 12 hours [27]
postoperatively, and in two studies the patients received fosfomycin three times
every 8 hours for 24 hours [24] or with
various intervals [36], see [Table 1]. In one study, the patients
received fosfomycin four times daily postoperatively for either 3–7 days
[38]. Lastly, three studies gave
fosfomycin regularly (time interval not reported) for some days [19] up to 2 days [20], or 5 days [30] after surgery.
Co-administration with another antimicrobial agent was given in some studies and
was most often metronidazole [21]
[22]
[23]
[24]
[26], however, in 46% of studies
fosfomycin was administered as monotherapy [19]
[20]
[25]
[27]
[28]
[29]
[30] (Supplementary File [Table
1]).
Risk of bias across included studies
The bias assessment according to Cochrane Handbook Risk of bias assessment tool 1
[12] can be seen in Supplementary
File Fig. 1. The nine RCTs [16]
[20]
[21]
[22]
[23]
[24]
[26]
[27]
[28] included were assessed, however, one
was a quasi-randomized clinical trial [16], thus resulting in a high risk of bias for the domains'
random sequence generation and allocation concealment. For most domains, there
was unclear risk of bias in 67% to 89% of included RCTs. The
exception was for the domains regarding 1) blinding of participants and 2)
personal and incomplete outcome data where 44% and 0%,
respectively, had unclear risk of bias. It was also in these two domains where
many RCTs had low risk of bias (33–67% of RCTs). Other sources
of bias that were noted concerned conflicts of interest and funding. In total,
88% of studies had no conflicts of interest statement [20]
[21]
[22]
[23]
[24]
[26]
[27]
[28]. In most RCTs, a funding statement was either not reported [20]
[23]
[24] or had insufficient
details on the role of funders and/or drug providers [21]
[22]
[26]
[27]
[28]. In two RCTs [22]
[27], the risk of bias for this domain was
ultimately assessed to be high, see Supplementary File Fig. 1.
For the retrospective [19]
[29]
[30] and prospective [17]
[18]
[25] cohort studies, bias was assessed using NOS [13]. According to this scale, the studies
are graded with a score of zero to nine stars across three categories: 1)
selection, 2) comparability, and 3) outcome. A low number of stars equal a high
risk of bias and vice versa. The assessed studies were given a median of 3 stars
and ranged from 1–5 stars, the bias assessment of the individual studies
can be seen in Supplementary File
[Table 1]. For the category selection,
all studies demonstrated that the outcome SSI was not present at start of study,
thus given a star, but only four studies provided documentation for the
ascertainment of preoperative parenteral administration of fosfomycin and could
be given a star for this [17]
[18]
[25]
[29]. Only one study was
awarded one out of two possible stars for the category comparability [19]. For the last category outcome, none of
the studies could be given a star for the item assessment of outcome as it was
not blinded, not record linked, or not described. Also, follow-up was only long
enough for the outcome SSI to occur in two studies that were each given one star
for this item [18]
[19].
Postoperative infectious complications
The postoperative complications were categorised into four types of
complications: SSI, intraabdominal abscess, sepsis, and death due to infectious
complications. Five studies did not contribute with data as they did not report
on this outcome [17]
[20]
[25]
[30] or they also included
patients with other indications for the antimicrobial treatment than
gastrointestinal surgery, so relevant data could not be extracted [27]
[28]
[29]. There was a very
sparse use of the classification system for postoperative complications
according to the Clavien Dindo classification [39] for studies published after 1992.
The rate of SSI was reported by eight studies [16]
[18]
[19]
[21]
[22]
[23]
[24]
[26] ([Fig. 2a]). SSI was mostly defined as wound
infection with the presence of pus/purulent material [21]
[22]
[23]
[24]
[26]. One study measured the temperature, pulse, and blood pressure
together with clinical findings [19]. The
rates of SSI for patients receiving fosfomycin vs. comparison group ranged from
0–1% vs. 0–5% for emergency procedures and
0–10% vs. 6–30% for elective procedures,
depicted in [Fig. 2a]. The study with the
highest SSI rates (in the comparison group only receiving metronidazole) was
terminated prematurely [24].
Fig. 2 Forest plot of the proportions and 95% confidence
intervals (CI) for a. surgical site infection (SSI) and b.
intraabdominal abscess (IAA) in each study for patients receiving
fosfomycin (left panel) or a comparison regimen (right panel). The
studies have been sub-grouped according to type of surgery e. g.
emergency or elective and with reference in brackets. The study with the
highest SSI rates was terminated prematurely because of more
postoperative infectious complications in patients in the comparison
group that only received metronidazole [24].
The rate of intraabdominal abscess was reported by eight studies [16]
[18]
[19]
[21]
[22]
[23]
[24]
[26] ([Fig. 2b]). An
intraabdominal abscess was mostly diagnosed either by imaging (ultrasonography
or computer tomography) or laparotomy [21]
[22]
[23] or by “clinical and
bacteriological signs of intraabdominal process causing illness” [24]
[26]. The rate of intraabdominal abscesses for patients receiving
fosfomycin vs. comparison group was 0% vs. 0–1% for
emergency procedures and 0–3% vs. 0–10% for
elective procedures ([Fig. 2a]).
The rate of sepsis was reported by four studies [16]
[18]
[21]
[22] (data not shown). Sepsis was defined as e. g.
“clinical, with malaise and fever” [21] or “temperature>38.5
̊C, together with rigors and poor general condition” [22]. No patients were reported to suffer
from sepsis in the two emergency studies with small populations [16]
[18]. For elective procedures, the rate of sepsis for patients
receiving fosfomycin vs. comparison ranged from 1–2% vs.
1–3%.
The mortality rate due to infectious complications was reported by seven studies
[16]
[18]
[19]
[21]
[22]
[23]
[24] (data not shown). The cause of the
reported mortality due to infectious complications was intraabdominal infection
[21] or peritonitis [22]. The rate of mortality due to
infectious complications was 0% regardless of antibiotic regimen for
emergency procedures and ranging from 0–1% for elective
procedures both for patients receiving fosfomycin and patients receiving any
comparison regimen.
Harms
Reports on harms were missing in three of the included studies, reporting on a
total of 256 patients receiving fosfomycin [20]
[28]
[29]. Of the studies that reported on harms,
six of these described that there were no harms due to treatment with
fosfomycin, covering a total of 263 patients [17]
[19]
[22]
[24]
[25]
[26]. Harms of different degrees were
reported in six studies that reported on 510 patients in total [16]
[18]
[21]
[23]
[27]
[30]. An overview of the
reported harms is illustrated in [Fig.
3], however, the harms occurring after discharge was not included [18]. All in all, there were few harms, and
most were related to the gastrointestinal system (n=19). One harm was
probably a serious adverse reaction, although details were sparse in the study,
and it was unclear if the patients had received fosfomycin or the comparison
regimen [27]. Most reported harms were
deemed to be adverse events or reactions.
Fig. 3 Distributions and types of the harms in the studies that
reported harms (total n=510 patients) with reference in
brackets. *not clearly reported whether this was after
fosfomycin or comparison treatment, “Allergic
manifestation” was further elaborated as “modest
hypotension, skin rashes, pruritus and laryngospasm” [27].
Discussion
This systematic review found that perioperative parenteral administration of
fosfomycin was primarily used in the 1980–1990s for a variety of both
elective and emergency gastrointestinal procedures. Often, a dose of 4 g fosfomycin
was administered an hour before surgery together with metronidazole, and this was
followed by one or more postoperative doses. There were few postoperative infectious
complications such as SSIs in patients receiving fosfomycin as well as patients
receiving the comparison antimicrobial agents. Harms were inconsistently reported,
were few, and most were deemed to be adverse events or reactions that were related
to the gastrointestinal system.
This systematic review has several strengths. We performed a systematic search for
articles after a medical research librarian had been consulted to help ensure a
broad and specific literature search. We had no language bias, as all relevant
articles no matter the language were included. A protocol was registered at PROSPERO
[10] to keep stringency, thoroughness, and
transparency through the conduct of the systematic review. Furthermore, registering
a protocol at PROSPERO reduced the risk of selective reporting. The screening of
articles was conducted independently by blinded reviewers, hence, not influencing
each other in the screening process. Finally, we reported according to PRISMA 2020
guideline [9]. However, this review also has
some limitations. Despite our best efforts, one report [40] found by searching the reference list of
included report with overlapping data [33]
could not be retrieved despite expert assistance from the Royal Danish Library. It
was an abstract from 1988 on 371 participants in a controlled clinical trial that
possibly could have contributed with data [40]. We had no language bias, but some information or nuances could have been
lost during translation due to the inclusion of all languages. Some of this
systematic review’s limitations were due to a lack of transparency in the
reporting of the included studies. Most of the studies were conducted in the
1980–1990s, thus before the implementation of reporting guidelines such as
STROBE [41] for cohort studies, CONSORT [42] for RCTs, and ClinPK statement [43] for pharmacokinetic studies. This was
especially evident for the risk of bias assessment for RCTs where most domains had
unclear risk of bias due to insufficient information in 67% to 89%
of the included RCTs. For bias assessment of the cohort studies, the total number
of
awarded stars was low due to the lack of a comparison group (comparability can be
awarded up to two stars). Furthermore, there was often no description of how the
outcomes were assessed, and follow-up was not long enough for SSIs to occur [13]. Also, harms and postoperative
complications were sparsely reported and not always well-defined by the authors for
instance by using definitions by ICH-GCP [44]
or the Clavien-Dindo classification of complications [39].
This systematic review provides an important overview of the use of fosfomycin in
gastrointestinal surgery that could become relevant, e. g. due to the
emerging resistance to currently used antimicrobial agents. In urology, fosfomycin
has been used as antimicrobial prophylaxis during prostate biopsies [45] as fluoroquinolones were associated with
harms and emerging resistance [46]. A
meta-analysis of 1,239 patients undergoing prostate biopsy found that fosfomycin
compared with fluoroquinolones halved the risk of infectious complications [45]. The combination of fosfomycin and
metronidazole could be a potential option to consider in gastrointestinal surgery
as
prophylaxis or empiric treatment in conjunction with surgical source control. For
now, however, the use of systemic fosfomycin is restricted by the European Medical
Agency as a reserve agent for the treatment of serious infections [47]. However, systemic fosfomycin has been used
for several indications, resulting in few harms, and these were mainly
gastrointestinal such as diarrhoea and nausea (5%) [7] as also seen in this systematic review. One
single oral dose of fosfomycin to treat uncomplicated urinary tract infections in
adults has been widely used for several years, and this indication was left
untouched by the European Medicines Agency [47]. Oral fosfomycin is generally well tolerated and side effects are
mainly gastrointestinal [48]. Recently,
fosfomycin was confirmed to be safe also in pregnant women regarding the risk of
congenital anomalies in a larger register-based French study where more than 2,700
women received fosfomycin during their first trimester [49]. All in all, the European Medicines Agency
currently allows oral fosfomycin in uncomplicated urinary tract infections and
prostate biopsies [47].
Conclusion
There were few postoperative infectious complications after perioperative parenteral
administration of fosfomycin in various gastrointestinal procedures, though the
studies were primarily published in the 1980–1990s. One dose of 4 g
fosfomycin sometimes supplemented with a few postoperative doses was often used
together with metronidazole. Harms were few and mild but inconsistently
reported.
Other information
Protocol and registration
A protocol was registered at PROSPERO, registration number: CRD42020201268,
before data extraction [10].
Authorship contribution statement
Authorship contribution statement
Siv Fonnes: Conceptualization, Formal analysis, Investigation, Writing
– Original Draft, Writing – Review & Editing,
Visualization. Masja Klindt Fonnes: Validation, Formal analysis,
Investigation, Writing – Original Draft, Writing – Review
& Editing. Barbara Juliane Holzknecht: Conceptualization, Writing
– Review & Editing. Jacob Rosenberg: Conceptualization,
Writing – Review & Editing, Supervision.
