Preterm birth
Preterm birth defined as birth before 37 complete weeks of gestation causes 60% to 80% of neonatal deaths (Cram 2002; Klein 2004; Leitich 2003; Yost 2000). Other important adverse outcomes of preterm birth include respiratory distress syndrome, intraventricular haemorrhage, leukomalacia, necrotizing enterocolitis, and prolonged hospitalisation (McGregor 1997). Survivors can suffer life‐long complications including cerebral palsy, blindness, and deafness (Kiss 2004; McGregor 1997). The direct and indirect costs of prematurity can be immense (McGregor 1997). The lifetime costs per preterm birth (baby's birthweight less than 2500 grams) have been estimated at £511,614 ($941,640; EUR 766,339) (Kiss 2004; Petrou 2000).

Since 1970, the incidence of preterm deliveries in developed countries has increased from 9% to 11%. Preterm premature rupture of the membranes and spontaneous preterm labour accounts for approximately 80% of preterm deliveries; the remaining 20% are elective deliveries for maternal or fetal reasons (for example, eclampsia) (Yost 2000).

Prematurity and infection
The risk of preterm labour in the presence of maternal infection is thought to be 30% to 50% (Chaim 1997; Cram 2002; Crowther 2005; Kiss 2004; Klein 2004; Reid 2003a; Romero 2002; VIPSG 1995). Urogenital infections, including urinary tract infections, bacterial vaginosis, and yeast vaginitis, affect an estimated one billion women in the world annually (Reid 2001). Urogenital infections have emerged in the last 20 years as an important cause of preterm labour. In fact, it is the only pathological process for which a firm causal link with preterm labour has been established (Chaim 1997; Kiss 2004; Leitich 2003; McGregor 1997; Reid 2001; Reid 2003b; Romero 2002; VIPSG 1995).

Bacterial vaginosis
Bacterial vaginosis is a clinical syndrome of unknown aetiology characterized by an overgrowth of vaginal anaerobes and variable degrees of depletion of the Lactobacilli population (Andreu 2004; Howe 1999; Iannacchione 2004; Reid 2003a; Ugwumadu 1999). In the western world bacterial vaginosis is the leading cause of vaginal discharge (Ugwumadu 1999). Bacterial vaginosis is common in pregnant women with a prevalence of 10% to 20% (Kiss 2004; Reid 2001; Sieber 1998). Adequate treatment in early stages of pregnancy may lead to reduction of prematurity (Howe 1999; Kiss 2004; McGregor 1997; Reid 2001; Romero 2002). When bacterial vaginosis develops, the concentration of many anaerobic or facultative anaerobic species increases up to a thousand fold. These species include Gardnerella vaginalis, Mobiluncus spp, Bacteroides spp, Mycoplasma hominis, and Peptostreptococcus spp (Chaim 1997; Howe 1999; Iannacchione 2004; Leitich 2003; Reid 2003a). The clinical diagnosis of bacterial vaginosis is based on the finding of at least three of the following four signs:
(1) vaginal PH greater than 4.7;
(2) grey homogenous vaginal discharge;
(3) presence of clue cells in a wet mount preparation of vaginal fluid;
(4) positive amine test in which a fishy odour is released after the addition of 10% potassium hydroxide (KOH) to the vaginal fluid (Chaim 1997; Cram 2002; McDonald 2005; Reid 2003a; Reid 2003c; Riggs 2004; Sieber 1998).

Bacterial vaginosis is more likely to be triggered by the decreasing number of Lactobacilli in the vaginal flora leading to relative predominance of pathological organisms (Sieber 1998). These organisms utilize ketoacids to produce amines and virulence factors including cytotoxins, sialidases, mucinases, and collagenases causing epithelial cell damage and creating the typical bacterial vaginosis discharge (Ugwumadu 1999).

Pathogenesis of ascending infection
The most common pathway for urogenital pathogens to cause preterm labour is the ascending route (McGregor 1997; Romero 2002). The presence of sialidases facilitates bacterial attachment and the breakdown of mucin, while mucinases assist microbial ascent into uterine tissues (Howe 1999; Klein 2004; McGregor 1997). Proteolytic enzymes may act directly on cervical collagen and fetal membrane leading to premature cervical ripening and weakening of the fetal membranes with subsequent preterm premature rupture of the membranes (McGregor 1997). Micro‐organisms may stimulate the host monocytes and macrophages resulting in the production of phospholipase A2. A2 is an enzyme that liberates arachidonic acid from the phospholipids of the membranes leading to the synthesis of prostaglandins E2 and F2a by the placental membranes (Bejar 1981; Chaim 1997; Reid 2004; Riggs 2004; Yost 2000). Similarly, protease toxins activate the deciduas and fetal membranes to produce cytokines such as tumour necrosis factor, interleukin (IL1a, IL1b, IL6, IL8), and granulocyte‐macrophage colony stimulating factor. In response to the activation of local inflammatory reaction, prostaglandins synthesis and release are stimulated (Cram 2002; Klein 2004; Reid 2002a; Riggs 2004; Ugwumadu 1999), which is known to stimulate uterine contractions (Riggs 2004).

Lactobacilli
Lactobacilli are gram positive, catalase negative, non‐sporing rods that dominate vaginal flora (Sieber 1998). Types of Lactobacilli found in the vaginal flora include: Lactobacillus acidophilus, fermentum, crispatus, and jensenii (Reid 2003a). The human vagina is lined by stratified squamous non‐keratinized epithelium. The surface is multilayered and the middle and superficial layers contains glycogen which is set free by the breakdown of superficial cells. Free glycogen is fermented by epithelial cells and by Lactobacilli producing lactic acid, and hydrogen peroxide (Andreu 2004; Reid 2002b; Reid 2003d; Sieber 1998). The presence and dominance of Lactobacillus in the vagina is associated with reduced risk of bacterial vaginosis and urinary tract infection (Reid 2002b). Women who have hydrogen peroxide producing strains of Lactobacilli have a 4% prevalence rate of bacterial vaginosis compared with 32% in women colonized by non‐hydrogen peroxide producing strains and 56% in those without Lactobacilli (Ugwumadu 1999). The mechanism includes blocking pathogen attachment to vaginal epithelium, and producing hydrogen peroxide (H₂O₂) and bacteriocins that inhibits pathogen multiplication (Andreu 2004; Reid 2003a; Riggs 2004; Sieber 1998; Wilks 2004).

Certain Lactobacillus strains are able to colonize the vagina following vaginal suppository use (for example, Vivag®) and may reduce the risk of urogenital infections, and it was suggested that it reduces infant mortality and preterm labour (Reid 2003b; Reid 2003c; Reid 2003d). This raises the question as to whether restoration of Lactobacilli by probiotic therapy can restore the normal flora and improve the chances of having a healthy term pregnancy (Reid 2003a).

Probiotics
Probiotics are defined as live micro‐organisms which when administered in an adequate amount confer a health benefit on the host (Andreu 2004; Elmer 2001; Reid 2003a; Reid 2003d). Probiotics have been shown to displace and kill pathogens and modulate the immune response by interfering with the inflammatory cascade that leads to preterm labour and delivery (Bengmark 2001; Reid 2003a). As yeast can co‐exist with Lactobacilli in the vagina, and few Lactobacillus strains inhibit yeast growth, it is unlikely that probiotics will cure yeast vaginitis. They are more likely to prevent recurrences by restoring the normal flora postantifungal treatment (Reid 2002a; Reid 2002b; Reid 2004). The actual mechanism of action of probiotics in the vagina has not been proven and is probably multifactorial. The production of lactic acid, bacteriocin, and hydrogen peroxide seems to be important and modulation of immunity is another possible mechanism (Reid 2004). The administration of these Lactobacilli by mouth or intravaginally, or both, (Reid 2003a; Reid 2004) has been shown to be safe and effective in reducing or treating, or both, urogenital infections (seeTable 1). The recommended dose, by whatever route of administration, is 109 to 1011 air‐dried or freeze‐dried bacteria (Andreu 2004; Elmer 2001; Reid 2003d).

The current recommendation by the Centre for Disease Control and Prevention (Atlanta, GA, USA) and the UK Drug and Therapeutics Bulletin is to screen and treat bacterial vaginosis in high‐risk pregnancies (Ugwumadu 1999). While antimicrobial agents are quite effective at providing clinical cure for bacterial infections, urogenital pathogen drug resistance is on the increase. Also, drugs have side‐effects including disruption of the protective vaginal flora which create an increased risk of recurrent infections (Reid 2001; Reid 2002b; Reid 2004; Romero 2002; Shennan 2006; Yost 2000). During pregnancy a local treatment restoring the normal acidity and vaginal flora without systemic effects may be preferable to any other treatment.