Description of the condition

Secretions from the vagina (known as vaginal discharge) are a normal occurrence and result from cervical secretion and other components derived from micro‐organisms that colonize the vaginal flora, mostly lactobacilli (Mylonas 2011). Lactobacilli play a predominant role in vaginal homeostasis through the production of lactic acid, hydrogen peroxide and bacteriocins, which are compounds that control the overgrowth of pathogenic micro‐organisms (Taylor 2013). An increase in the amount of vaginal discharge is usually associated with the presence of a genital tract infection (Mylonas 2011). Women with an abnormal vaginal discharge most commonly have bacterial vaginosis (BV) (39.6%) or, less frequently, vulvovaginal candidiasis (11%) or trichomoniasis (0.8%); however, the exact etiology of the vaginal discharge can change according to the population (Ángel‐Müller 2012). BV is thought to arise as a result of an imbalance in the normal vaginal flora when lactobacilli decrease in concentration and are replaced by anaerobic and facultative aerobic bacteria (Taylor 2013), such as Gardnerella vaginalis (40%) or, less frequently, by Mycoplasma hominis, Clostridia (Mirmonsef 2012), Peptostreptococcus or Mobiluncus (Mylonas 2011). Although the precise cause of this dysbacteriosis is not completely understood, some factors, such as high sexual activity, a new sexual partner and risky sexual behavior (Bradshaw 2005; Mirmonsef 2012; Mylonas 2011), could contribute to this condition (Mylonas 2011).

BV is found worldwide with a prevalence between 10% and 50%, and is usually higher among sex workers, the HIV‐positive population, black and Hispanic women (Kenyon 2013), and among women with a new or a high number of sexual partners (Marrazzo 2011). Having three or more male sexual partners in the past 12 months has been identified as a risk factor for BV (Marrazzo 2011). Also, having a female sexual partner has been reported to increase the risk of BV by 60% (Smart 2004). The burden of BV is related not only to the prevalence of the condition, but also to a reduction in quality of life, as well as to anxiety and self‐conscious feelings relating to the fear of exclusion due to the foul smell associated with the infection (Bilardi 2013). The bacteria associated with BV may be carried by male genitalia, but men are usually asymptomatic (Nelson 2012). The direct cost of BV care has been evaluated in pregnant women, and is particularly related to the complications of pregnancy that derive from it: in a study evaluating a population from the USA, the Netherlands and the UK it was estimated that the overall cost of BV was around one thousand million USD a year (Oleen‐Burkey 1995). BV has also been associated with a risk of acquiring a sexually transmitted infection (STI), such as pelvic inflammatory disease (Mirmonsef 2012; Taylor 2013), and infections caused by Neisseria gonorrhoeae, Chlamydiia trachomatis and HIV (Mirmonsef 2012), and with a higher risk of developing cervical cancer (Gillet 2012), and even subfertility (Taylor 2013). In addition, BV may increase the risk of adverse perinatal outcomes, such as low birth weight, preterm delivery and very low birth weight (Sangkomkamhang 2008), and may result in approximately 80,000 excess preterm births every year in the USA (Goldenberg 2005).

The Amsel criteria and Gram staining can be used to diagnose BV, and Gram staining is considered the gold standard for this purpose (Workowski 2015). The Nugent scoring system is applied to Gram stains of vaginal smears to visually estimate the numbers of lactobacilli and BV‐associated organisms; a Nugent score of zero to three is considered healthy, four to six is intermediate and a score of seven to ten implies the presence of BV (Brotman 2011). This method has a number of disadvantages: a trained professional is required for the interpretation of results and the Gram stain is unable to detect some micro‐organisms, such as Ureaplasma and Mycoplasma (Taylor 2013). When Gram staining is not possible, the Amsel criteria provide a good clinical tool for the diagnosis of BV, which requires the presence of at least three of the following four criteria (Taylor 2013).

• Homogeneous, thin, white vaginal discharge.

• Presence of clue cells on microscopic examination.

• A pH greater than 4.5.

• A positive ‘whiff test’ (a fishy odour after 10% potassium hydroxide is added).

In vitro cultures for G. vaginalis are of limited utility because although cultures in almost all women with a symptomatic infection are positive, 50% of the cultures from healthy asymptomatic women are also positive (Mylonas 2011). Finally, in some circumstances a diagnosis can be reached through combining the medical history with the clinical characterization of the discharge during a gynaecological examination (Ministerio de Salud y Protección Social 2013). This approach, the syndromic approach, is based on the assumption that genital tract infections share some common symptoms and signs thus making it possible to assign them a specific etiology without the need for a laboratory test (Ministerio de Salud y Protección Social 2013).

BV is not currently considered a STI but it is well‐recognized that both have many aspects in common (Morris 2001). Several studies have reported an association between BV and the coexistence of STIs, risky sexual behaviours, young age at first sexual intercourse, sharing sexual toys, and a new or multiple male or female partners (Chavez 2009; Fethers 2008; Fethers 2009; Marrazzo 2005; Morris 2001). Furthermore, BV is unusual in women without a history of sexual contact (Fethers 2009), and its prevalence is less frequent in women who report that they regularly use condoms or have a stable sexual partner (Chavez 2009; Fethers 2008); these findings are consistent with a strong association between BV and sexual activity (Fethers 2009).

Additionally, some studies have proposed that the bacteria associated with BV may be carried by the male genitalia (Nelson 2012), based on the microbiological isolation of the micro‐organisms associated with this infection in the flora of the penis and urethra in male partners of infected females (Mandar 2013; Nelson 2012). This is particularly notable in uncircumcized males in whom the presence of a major surface of contact could have a major role as a reservoir, and favour the occurrence of an infection after sexual intercourse (Bukusi 2011; Mehta 2012). Bacteria associated with BV have been isolated from the normal flora of the male coronal sulcus and urethra, and could be related to infection in female partners and to the recurrence of infection after treatment (seen in about 20% of individuals and defined as three or more proven episodes of BV in 12 months) (Colli 1997; Cook 1992). Another mechanism for infection could be the mechanical transfer of perineal enteric bacteria from the flora of the male genitalia through unprotected and protected intercourse (Mandar 2013).

Description of the intervention

As BV is considered to arise from an imbalance in the normal vaginal flora (Taylor 2013), treatment includes the administration of antibiotics, such as nitroimidazoles, lincosamides, macrolides (e.g. erythromycin) and, in some circumstances, penicillins (Oduyebo 2009; Workowski 2015), the objective being to eradicate the abnormal vaginal bacterial flora (Mylonas 2011). 5‐Nitroimidazole derivatives are available for oral and intravaginal administration (Workowski 2015). These medications are absorbed completely after oral intake and the volume of distribution of these agents approximates to that of total body water because they are associated with a low level of protein binding (Lamp 1999). 5‐Nitroimidazole derivatives penetrate well into body tissues and fluids, with the exception of the placenta, and are eliminated in the urine after being metabolized in the liver (Brunton 2011; Trevor 2012). The side effects of 5‐nitroimidazoles are usually mild and the most common are headache, nausea, dry mouth and a metallic taste (Brunton 2011; Trevor 2012). The regimens recommended for the treatment of BV include metronidazole 500 mg orally twice a day for seven days or metronidazole gel 0.75% one full applicator (5 g) intravaginally once a day for five days; tinidazole 2 g orally once daily for one or two days or 1 g orally once daily for five days; or secnidazole 1 g or 2 g orally in single doses (Oduyebo 2009; Workowski 2015).

Clindamycin, the most representative of the lincosamides, is almost completely absorbed following oral administration (Bouazza 2012), with a half‐life of about three hours, and is widely distributed into fluids and tissues (Bouazza 2012; Gatti 1993). More than 80% of this agent is bound to plasma proteins, primarily α1‐acid glycoprotein (Flaherty 1996); it is metabolized in the liver by cytochrome P450 3A4 and is excreted in the urine and bile (Trevor 2012; Wynalda 2003, Yasuda 2008). The most common side effects related to the administration of clindamycin include diarrhoea and, in a small number of individuals, pseudomembranous colitis (Gurwith 1977). Rash can occur in about 10% of individuals; other reactions, such as Stevens‐Johnson syndrome, the reversible elevation of transaminases, granulocytopenia, thrombocytopenia and anaphylactic reactions, are less frequent (Brunton 2011). Regimens suggested for the treatment of BV include clindamycin cream 2% one full applicator (5 mg) intravaginally at bedtime for seven days; clindamycin 300 mg orally twice daily for seven days; or clindamycin ovules 100 mg intravaginally once at bedtime for three days (Workowski 2015).

The serum half‐life of erythromycin is approximately two hours (Josefsson 1982; Zuckerman 2011). The drug is metabolized in the liver by cytochrome P450 3A4 and 2% to 5% is eliminated in an active form in the urine (Trevor 2012; Zuckerman 2011). Cholestatic hepatitis is the most notable side effect of macrolides and hepatotoxicity has also been observed (Derby 1993). Large doses have been associated with abdominal cramps, nausea, vomiting and diarrhoea (Zuckerman 2011). Among the recognized allergic reactions are fever, eosinophilia and skin eruptions, which may occur alone or in combination; all disappear shortly after therapy is suspended (Brunton 2011). Erythromycin for the treatment of BV is used at a dose of 333 mg oral three times daily for 14 days in combination with metronidazole (Brocklehurst 2013).

Some studies have used the aminopenicillin amoxicillin for the treatment of BV (Amit 2013; Brocklehurst 2013). This agent is a semi‐synthetic penicillin with a half‐life of almost 80 minutes, which binds to proteins in the plasma and is excreted renally (Brocklehurst 2013). The most important side effects associated with amoxicillin are nausea, vomiting, rash and antibiotic‐associated colitis. The use of ampicillin is now generally avoided due to the emergence of ampicillin‐resistant bacteria in individuals with BV (Amit 2013).

How the intervention might work

To date, the evidence regarding whether to treat or not the sexual partner of a woman with BV is inconclusive (Mehta 2013; Potter 1999). As frequently‐used antibiotics for treating BV are widely distributed into the fluids and tissues of the human body (Bouazza 2012; Brunton 2011; Gatti 1993; Trevor 2012), the treatment of sexual partners could eliminate the abnormal flora present in the male genital area that may promote relapse in, or reinfection of, women with BV (Bradshaw 2006).

The 5‐nitroimidazole derivatives are agents with antiprotozoal and antibacterial properties (Amit 2013). When a 5‐nitroimidazole enters the cellular environment, it is reduced by the pyruvate ferredoxin oxidoreductase system present in bacterial mitochondria, which modifies its chemical structure (Amit 2013). Pyruvate ferredoxin oxidoreductase generates adenosine triphosphate (ATP) via the oxidative decarboxylation of pyruvate; in the presence of a structurally modified 5‐nitroimidazole, the electrons that would be transferred to hydrogen ions during the generation of ATP are captured by the nitro group (Amit 2013). This disruption of ATP production promotes the formation of intermediate compounds and free radicals that are highly toxic to the bacteria (Brunton 2011). The lincosamides bind to the 50S subunit of bacterial ribosomes and consequentially suppress bacterial protein synthesis (Brunton 2011; Trevor 2012). The macrolides (e.g. erythromycin) are bacteriostatic agents that inhibit protein synthesis by binding reversibly to the 50S ribosomal subunits of sensitive micro‐organisms (Brunton 2011; Trevor 2012).

The administration of any effective and safe intervention to the sexual partner of a woman with BV could offer the advantages of decreasing the recurrence of infection, thus breaking the chain of infection and possibly decreasing the prevalence of STIs, reducing the burden associated with the condition and impacting positively on the sexual and reproductive health of the woman (Mehta 2013).

Why it is important to do this review

This systematic review is of paramount importance as some studies suggest that BV exhibits a behaviour similar to that of an STI (Potter 1999). However, there is still some uncertainty regarding whether recurrence of BV in women could be due to a failure to eradicate the causative organism(s) or a consequence of reinfection by the sexual partner (Bradshaw 2006). Currently, the treatment of sexual partners of women with BV is not recommended (Workowski 2015), but this is based on the critical appraisal of individual clinical trials (none of which provide conclusive evidence regarding the efficacy and safety of male sex partner treatment for women with BV (Mehta 2013)) rather than a synthesis of the available clinical evidence (Mehta 2012; Potter 1999). If the treatment of sexual partners were found to reduce the rate of recurrence of BV then this could reduce costs. However, it would need to be weighed against the possible risk of antibiotic resistance that could arise from the additional multiple treatments such a policy would involve. There is therefore a need for high‐quality systematic reviews to improve the clinical management of BV (Madhivanan 2013). This systematic review will facilitate the synthesis of the current evidence, and recognize its strengths and weakness, address the uncertainty of the current knowledge and make it possible to assess the effectiveness and safety of this intervention.