Verstraelen2012 Article AntisepticsAndDisinfectantsFor
Verstraelen2012 Article AntisepticsAndDisinfectantsFor
Verstraelen2012 Article AntisepticsAndDisinfectantsFor
Abstract
Background: The study objective was to assess the available data on efficacy and tolerability of antiseptics and
disinfectants in treating bacterial vaginosis (BV).
Methods: A systematic search was conducted by consulting PubMed (1966-2010), CINAHL (1982-2010), IPA (1970-
2010), and the Cochrane CENTRAL databases. Clinical trials were searched for by the generic names of all antiseptics
and disinfectants listed in the Anatomical Therapeutic Chemical (ATC) Classification System under the code D08A.
Clinical trials were considered eligible if the efficacy of antiseptics and disinfectants in the treatment of BV was
assessed in comparison to placebo or standard antibiotic treatment with metronidazole or clindamycin and if
diagnosis of BV relied on standard criteria such as Amsel’s and Nugent’s criteria.
Results: A total of 262 articles were found, of which 15 reports on clinical trials were assessed. Of these, four
randomised controlled trials (RCTs) were withheld from analysis. Reasons for exclusion were primarily the lack of
standard criteria to diagnose BV or to assess cure, and control treatment not involving placebo or standard
antibiotic treatment. Risk of bias for the included studies was assessed with the Cochrane Collaboration’s tool for
assessing risk of bias. Three studies showed non-inferiority of chlorhexidine and polyhexamethylene biguanide
compared to metronidazole or clindamycin. One RCT found that a single vaginal douche with hydrogen peroxide
was slightly, though significantly less effective than a single oral dose of metronidazole.
Conclusion: The use of antiseptics and disinfectants for the treatment of BV has been poorly studied and most
studies are somehow methodologically flawed. There is insufficient evidence at present to advocate the use of
these agents, although some studies suggest that some antiseptics may have equal efficacy compared to
clindamycin or metronidazole. Further study is warranted with special regard to the long-term efficacy and safety of
antiseptics and disinfectants for vaginal use.
Keywords: Bacterial vaginosis, Antiseptics, Disinfectants, Therapy, Systematic review, Chlorhexidine,
Polyhexamethylene biguanide, Hydrogen peroxide
© 2012 Verstraelen et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the
Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
Verstraelen et al. BMC Infectious Diseases 2012, 12:148 Page 2 of 8
http://www.biomedcentral.com/1471-2334/12/148
antiseptics facilitate the eradication of the anaerobic vagi- mercuric amidochloride, phenylmercuric borate, mercuric
nal microbiota associated with bacterial vaginosis which chloride, mercurochrome, mercury, thiomersal, mercuric
allows for the recolonisation of indigenous lactobacilli. iodide, phenylmercuric borate, silver nitrate, silver, hydro-
Antiseptics generally have a very broad spectrum as they gen peroxide, eosin, propanol, tosylchloramide sodium,
act non-specifically on bacteria through mechanisms such isopropanol, potassium permanganate, sodium hypochlor-
as bacterial cell membrane disruption. In accordance, ite, and ethanol.
there are very few reports on antimicrobial resistance to In addition we accounted for benzydamine which is
these agents. Nonetheless, antiseptics and disinfectants listed under ATC code G02CC03 for “anti-inflammatory
have become unpopular for the purpose of treating vaginal products for vaginal administration”, considering benzy-
infections, presumably because they are often used as damine also has broad antimicrobial activity.
over-the-counter (OTC) products associated with vaginal Only English-language studies were considered for re-
douching, a common practice that has been associated view. Additional studies were searched for by checking
with the occurrence of BV. In many countries, antiseptics cross-references cited in the primary studies. No efforts
are still marketed for the treatment of vaginal infections, were made to identify unpublished studies.
though their efficacy has not been supported by a sound
evidence base. We therefore sought to assess currently
Types of studies
available data on the efficacy and tolerability of antiseptics
Reports were considered eligible if they involved clinical
and disinfectants in the treatment of BV. Through a sys-
trials in which the efficacy and tolerability of antiseptics
tematic literature search, we identified and evaluated clin-
and disinfectants in the treatment of BV was assessed in
ical trials that compared antiseptic and/or disinfectant
comparison to placebo or in comparison to standard
treatment to either placebo or standard antibiotic treat-
antibiotic treatment with metronidazole or clindamycin.
ment in women of any age.
Studies were considered eligible if at least 40 study
participants had been enrolled for comparison between
Methods
the two treatment arms.
Objectives
The overall objective is to summarize currently available
data (published up to December 31 2010) on the efficacy Types of participants
and tolerability of antiseptics and disinfectants in the treat- Women of any age diagnosed with bacterial vaginosis
ment of BV, thereby accounting for the quality of the clin- through standardized criteria like Amsel’s or Nugent
ical trials identified through a systematic literature search. criteria [5,6]. No study was excluded because of (pos-
sible) co-infection with sexually transmitted infections.
Search strategy
An electronic search was conducted by consulting the
Types of intervention
following databases: PubMed (1966-2010), CINAHL
We confined our review to antiseptics and disinfectants for
(1982-2010), International Pharmaceutical Abstracts
which the dosage and treatment regimen was specified, and
(IPA) database (1970-2010), and the Cochrane Central
which are listed in the Anatomical Therapeutic Chemical
Register of Controlled Trials.
(ATC) Classification System under the code D08A (“anti-
Clinical trials were searched for by use of the key words
septics and disinfectants”). In addition we accounted for
“bacterial vaginosis” and “non-specific vaginitis” in com-
benzydamine which is listed under ATC code G02CC03 for
bination with “disinfectant”, “antiseptic”, and subse-
“anti-inflammatory products for vaginal administration”,
quently with all generic names of antiseptics and
considering benzydamine also has broad antimicrobial ac-
disinfectants listed in the Anatomical Therapeutic Chem-
tivity. Hence, we excluded any report in which any antisep-
ical (ATC) Classification System under the code D08A
tic product (e.g. herbal medicines) was used for which the
(Antiseptics and disinfectants), i.e. ethacridine lactate,
chemical composition and/or substance dose was not verifi-
aminoacridine, euflavine, aluminium agents, dibrompro-
able. Treatment with an antiseptic or disinfectant was con-
pamidine, chlorhexidine, propamidine, hexamidine, poly-
sidered in any preparation type, any dosage regimen, and
hexanide, boric acid, hexachlorophene, policresulen,
any route of administration.
phenol, triclosan, chloroxylenol, biphenylol, nitrofural,
iodine/octylphenoxypolyglycolether, povidone-iodine, iod-
ine, diiodohydroxypropane, dequalinium, chlorquinaldol, Types of outcome measures
oxyquinoline, clioquinol, benzalkonium, cetrimonium, The primary outcome was cure of bacterial vaginosis as
cetylpyridinium, cetrimide, benzoxonium chloride, dide- assessed through use of the Amsel criteria or Nugent
cyldimethylammonium chloride, benzethonium chloride, criteria at least 7 days following treatment initiation.
octenidine, benzethonium chloride, dodeclonium bromide, Cure was defined as a Nugent score <7 or the presence
Verstraelen et al. BMC Infectious Diseases 2012, 12:148 Page 3 of 8
http://www.biomedcentral.com/1471-2334/12/148
of less than 3 Amsel criteria. The secondary outcome was octenidine (1 RCT), polyhexamethylene biguanide (2
the occurrence of adverse reactions or side effects, with RCTs), and povidone iodine (6 clinical trials including
special regard to vaginal symptoms following local adminis- 5 RCTs). These studies are herewith listed and briefly
tration of antiseptics. discussed according to the antiseptic or disinfectant
under study.
Risk of bias in included studies
As indicated by the PRISMA guidelines no efforts were
made to label the quality of studies in a (semi-)quantitative Benzydamine
matter [7]. Rather, the Cochrane Collaboration tool for We found two randomized clinical trials on treatment of
assessing risk of bias was applied to assess the risk of vari- BV with benzydamine of which none was in accordance
ous sources of bias among the included studies [8]. with the inclusion criteria set forth [10,11].
In the 1987 double-blinded randomised controlled trial
Results by Ventolini et al. [11] a regime of intravaginal 0.1% benzy-
Through the Boolean search of the electronic damine hydrochloride applied twice daily for 10 days was
databases PubMed, CINAHL, IPA, and Cochrane compared nor to placebo, neither to standard antibiotic
CENTRAL, a total of 262 articles on BV and anti- treatment. In addition no standard criteria were applied for
septics and disinfectants were identified. Two add- the diagnosis of BV neither at recruitment nor at follow-up.
itional articles were retrieved during the background Hence we considered this study not assessable for the pur-
literature study [9,10]. Eventually, 15 full-text papers pose of the present study.
were assessed for eligibility. Of these, four rando- Hay et al. conducted a small double-blinded randomised
mised controlled trials (RCTs) met the inclusion cri- controlled trial in which 15 patients were enrolled of which
teria set forth. The search and selection strategy is 8 used 140 ml of 0.1% benzydamine vaginal douche twice a
presented as a PRISMA flow diagram in Figure 1. day for 7 days and 7 patients applied a similar regimen with
The 15 clinical trials identified involved treatment placebo douches [10]. Diagnosis of BV was made through
with one or more of the following antiseptics and dis- Amsel criteria and Nugent score at enrolment and cure
infectants: benzydamine (2 RCTs), chlorhexidine was defined as a normal Gram stain and absence of clue
(2 RCTs), dequalinium chloride (1 RCT), hydrogen cells on wet mount. Since the sample size was below the
peroxide (3 clinical trials including 1 RCT), lower limit set forth, we did not include the study.
Studies included in
qualitative synthesis
(n = 4)
vaginal hydrogen peroxide instillation in curing BV was as polyhexanide. One study was considered not assess-
assessed [17]. able [19], whereas another large study was withheld [20].
In a very well-designed triple-blinded randomised con- In the 2003 single-blinded randomised controlled trial
trolled trial, Chaithongwongwatthana et al. investigated by Gerli et al., 133 patients were enrolled of which 59
the effectiveness and safety of a single hydrogen perox- received a 2% polyhexamethylene biguanide gel solution
ide vaginal douche in treating bacterial vaginosis [17]. In and 51 administered 2% clindamycin cream for 7 days
this study, 142 patients with BV were enrolled in a Thai [19]. However, neither at enrolment nor at follow-up
outpatient gynaecologic clinic. Diagnosis was made were all patients systematically assessed through Amsel’s
according to Amsel’s criteria. Randomisation relied on or Nugent’s criteria and therefore we consider the study
computer-generated random number scheme and both not assessable to the purpose of the present study.
subjects and investigators were blinded to treatment. In The 2008 study by Minozzi et al. was a multi-centred
addition double dummies were used to ensure blinding. single (investigator)-blinded randomised controlled trial
In the experimental treatment arm patients (n = 72) were in which 740 BV patients were enrolled [20]. Patients
treated with a single douche with 20 ml of a 3% hydro- were considered eligible if they fulfilled all Amsel’s cri-
gen peroxide solution in lithotomy position for 3 min- teria and were then randomly assigned to one of two
utes, whereas in the control arm patients (n = 70) treatment arms, i.e. a single dose of 100 ml of a 10%
received 4 tablets of 500 mg metronidazole. The dummy polyhexamethylene biguanide vaginal gel solution
treatments consisted of a vaginal douche with clean (n = 371) or 2% clindamycin cream for 7 days (n = 369).
water in the metronidazole group and of identical pla- Patients were then re-assessed through Amsel’s and
cebo tablets similar to the metronidazole tablets in the Nugent’s criteria at 21 to 30 days following treatment
hydrogen peroxide group. Follow-up was performed at initiation. Eventually data on 347 patients were evalu-
2 weeks following treatment. Cure was defined as the able; this is 46.9% of patients constituting the inception
absence of at least 3 Amsel criteria. Assessors were cohort. Hence, although attrition was high, the authors
blinded for the treatment allocation. The cure rates were gave a detailed description for non-inclusion in the final
62.5% in the hydrogen peroxide group and 78.6% in the per protocol analysis for 147 of the 196 non-included
metronidazole group (see Table 1). Patients in the patients in the polyhexamethylene biguanide group and
metronidazole group had significantly more gastrointes- for of all 197 non-included patients in the clindamycin
tinal side effects (48.6% versus 13.9%, p < 0.001), whereas group. Cure was defined among others as the absence of
patients in the hydrogen peroxide group had signifi- all four, of three and of two of the Amsel’s criteria and
cantly more mild vaginal irritation (33.3% versus 14.3%, was obtained in 64.3, 87.5, and 90.6% versus 63.2, 83.2,
p = 0.008). and 91.2% patients in the polyhexamethylene biguanide
and clindamycin groups respectively (see Table 1). Re-
Octenidine conversion to a normal microbiota (Nugent score <4)
In a large Serbian RCT, the efficacy of an octenidine was achieved in 56.5 and 57.7% (see Table 1) of patients
hydrochloride/phenoxyethanol spray as compared to in the polyhexamethylene biguanide and clindamycin
standard treatment with 500 mg metronidazole vaginal groups, while absence of clue cells was recorded in 87.5
tablets for 7 days was assessed among 450 patients en- and 84.0%, respectively. Thirty-seven patients in the
rolled in an outpatient gynecology clinic and diagnosed polyhexamethylene biguanide group reported 29 study
with BV according to Amsel’s criteria [18]. Unfortu- medication-related adverse events (AEs) compared to 29
nately, the authors did not specify how cure was defined. patients in the clindamycin group who reported 22 study
In particular, the authors state that study participants medication-related AEs (P = 0.3).
were “examined gynecologically” and that “control
smears were taken for further bacteriologic analysis” Povidone iodine
[18]. We therefore repeatedly tried to contact the We found six clinical trials in which povidone iodine
authors with a request for more detailed information; was evaluated for its efficacy in the treatment of BV.
however we did not get any response. The otherwise One trial was a non-controlled study [21]. One study
very well designed study could therefore not be entailed a comparison with benzydamine [11], another
included. study involved a comparison with dequalinium chloride
[14], and a third one compared povidone iodine to a
Polyhexamethylene biguanide probiotic [22]. The remainder of studies were two
We identified two controlled clinical trials on treatment placebo-controlled studies [9,23].
of BV with polyhexamethylene biguanide [19,20], a bi- In the 1982 study by Dattani et al. patients were ran-
guanide antiseptic that is actually listed in the Anatom- domised to use either a 200 mg povidone-iodine con-
ical Therapeutic Chemical (ATC) Classification System taining pessary either an identical looking placebo
Verstraelen et al. BMC Infectious Diseases 2012, 12:148 Page 6 of 8
http://www.biomedcentral.com/1471-2334/12/148
pessary during night and morning for two weeks [23]. assessed through use of the Cochrane Collaboration tool
As no standard criteria (Amsel’s criteria or Nugent cri- for assessing risk of bias [8] and is presented in Table 2.
teria) in this era were used either at inclusion or at fol- With regard to side-effects, no mention is made of any
low-up, this study was discarded. adverse effects in the study on the chlorhexidine pessary
Van der Meijden et al. [9], also compared a 200 mg [12]. In the study on chlorhexidine gel [13] it was merely
povidone-iodine containing pessary inserted into the va- mentioned that, overall, six women (20%) complained of
gina in the morning and evening for 5 consecutive days mild transient burning after chlorhexidine vaginal gel
with placebo pessaries. As no standard criteria (Amsel’s application. In the hydrogen peroxide trial [17], patients
criteria or Nugent criteria) in this era were used at inclu- in the metronidazole group had significantly more
sion or at follow-up, this study was therefore also con- gastrointestinal side effects (48.6% versus 13.9%,
sidered not assessable. p < 0.001), whereas patients in the hydrogen peroxide
In summary, of the 15 clinical trials identified through group had significantly more mild vaginal irritation
a systematic literature search, merely four RCTs with at (33.3% versus 14.3%, p = 0.008). Finally, in the polyhexa-
least 40 study participants involved the comparison of a methylene biguanide study [20] it was reported that
vaginal antiseptic to placebo or to standard antibiotic thirty-seven patients in the polyhexamethylene biguan-
treatment and thereby relied on standard criteria in ide group (10.3%) reported 29 study medication-related
assessing BV and cure rates. Main characteristics of the adverse events (AEs) compared to 29 patients in the
latter studies are displayed in Table 1. clindamycin group (7.9%) who reported 22 study
The four selected RCTs compared antiseptics to medication-related AEs and that the number of patients
metronidazole or clindamycin. In two studies control who reported AEs and study medication-related AEs
treatment consisted of a single oral dose of 2 g metro- were not statistically different between the two groups
nidazole, in one study of a 7-day regime of 500 mg (P = 0.386 and 0.336 respectively). A detailed overview of
metronidazole intravaginally, and in one study of 2% AEs was given in this report and no significant differ-
clindamycin cream for 7 days. Only the latter two treat- ences were observed for any given AE, including vaginal
ment modes are CDC recommended regimens [3]. symptoms. So, overall in two of the three studies that
Three studies showed non-inferiority of the antiseptics reported AEs elevated rates of vaginal irritation or burn-
applied in comparison to antibiotic treatment. In one ing were recorded.
study, a single douche with 20 ml of a 3% hydrogen per-
oxide was significantly less effective than a single 2 g
oral dose of metronidazole. Discussion
Only one RCT was double-blinded (and actually We sought to assess and summarize currently available
triple-blinded) through the use of double dummies [17]. data on the efficacy and tolerability of antiseptics and
Randomisation procedure was reported to involve a disinfectants in treating BV. Overall, we identified 15
computer-generated random number series in two stud- clinical trials that dealt with treatment of BV with anti-
ies [13,17]. Allocation concealment was not mentioned septics and disinfectants. Taking into account the limita-
explicitly in any study, except one [17]. As to selection tions of the studies, it was found that some of the
bias, in none of the studies it was actually explicitly antiseptics tested, i.e. chlorhexidine and polyhexamethy-
reported that consecutive patients were enrolled. In one lene biguanide were equally effective to antibiotic treat-
large trial, attrition was very high albeit well- ment, whereas in one study, a single douche of hydrogen
documented [20]. Follow-up ranged from 7 up to peroxide was slightly, though significantly less effective
28 days. In one out of the four RCTs no mention is than a single 2 g oral dose of metronidazole.
made of recording of side effects [12]. Risk of selection, Eleven studies were discarded, because the studies had
performance, detection, attrition, and reporting bias was no RCT design, because diagnosis of BV at inclusion
and/or on follow-up did not rely on standard criteria like Authors' Contributions
Amsel’s or Nugent’s criteria, or because the control HV conceived of the study, collected the literature, assessed the data and
drafted the manuscript. RV helped to draft the manuscript. KR assessed the
treatment arm did not involve placebo or standard anti- data and helped to draft the manuscript. MT helped to draft the manuscript.
biotic treatment. All authors read and approved the final manuscript.
Of the four studies included, only one RCT was
double-blinded (and actually triple-blinded) through the Acknowledgements
We thank Prof. Dr. Michael Chikindas for his assistance in collecting the
use of double dummies. Risk of selection bias and of literature. We thank Thomas O’Leary for critical review of the manuscript.
performance bias was generally high. Risk of attrition
bias was high in one study. Reporting bias was presum-
Author details
ably low. It should be added that the risk of publication 1
Department of Obstetrics & Gynaecology, Faculty of Medicine and Health
bias is also presumably high, especially since published Sciences, Ghent University, De Pintelaan 185, B-9000, Ghent, Belgium.
2
reports often involved commercial antiseptic products International Centre or Reproductive Health, Faculty of Medicine and Health
Sciences, Ghent University, Ghent, Belgium.
marketed for vaginal use.
Follow-up in all studies was limited, as has previously Received: 15 December 2011 Accepted: 25 May 2012
been noted with antibiotic treatment studies on BV [4]. Published: 28 June 2012
doi:10.1186/1471-2334-12-148
Cite this article as: Verstraelen et al.: Antiseptics and disinfectants for the
treatment of bacterial vaginosis: A systematic review. BMC Infectious
Diseases 2012 12:148.