Table 1 Characteristics of studies included in meta-analysis

Adriaanse et al 1995, Holland

At onset of labour, the attending obstetrician applied 10 ml chlorhexidine (CHX) gel around the portio vaginalis and into the fornices. This procedure was repeated after 10 h in case delivery had not yet occurred.

Pregnant women from two hospitals with obstetric services in the city of Nijmegen, the Netherlands.

Known GBS carrier, use of antibiotics during the 4 weeks before admission, planned caesarean section, antepartum foetal death, suspected congenital abnormalities and premature labour.

No significant difference between groups

No significant differences between the three groups, except for the % of neonates admitted to the (special) neonatal care unit (P = 0.012) in the CHX and control group.

No special training given to doctors giving intervention, no protocol given for intervention e.g. timing of washing.

10ml 0.3% CHX gel

Standard care

1020 participating women, 522 were enrolled in one hospital and 498 in the other. Of the 981 analysed mother-infant pairs, 327 were assigned to the chlorhexidine group, 328 to the placebo group and 326 to the control group.

Primary outcomes were vertical GBS transmission to the neonate. Secondary goals were to study the vertical transmission rates of E. coli, S. aureus and C. albicans, and to establish neonatal and maternal morbidity. Neonatal septicaemia, meningitis and pneumonia diagnosed from the positive cultures of blood or CSF or tracheal aspirate.

Burman 1992, Sweden

60mls of solution (CHX or sterile water) was used to flush the anterior fornix, vaginal walls and urethral orifice in a spiral outward motion by a midwife. This was repeated every 6hours until delivery. Flush was counted if birth occurred > 1 hour after flush and no more than 6 hours lapsed between flushes.

Pregnant women who were urogenital carriers of GBS from 10 Swedish hospitals.

Pre term infants (<37 weeks) planned caesarean section, pregnancy complications after the 30th week of gestation requiring hospital admission, twin or multiple pregnancies, suspected congenital abnormality of the infant, known or suspected allergy to CHX, previous invasive GBS infection, antibiotics during the 2 weeks before admission, and antepartum foetal death.

Not analysed

No significant differences seen

No rigorous set procedure for flushing e.g. time taken to flush, no specialist training given to midwives, multiparity pregnancies excluded, group sizes not even, maternal characteristics not determined.

60 ml 2g/l CHX given as 2 30ml ampules via catheter

60ml sterile saline

4483 women 2238 CHX group and2245 saline placebo group

Rate of admission of babies to special-care neonatal units within 48 h of delivery. Admissions for sepsis/meningitis, pneumonia, skin infection, meconium aspiration, surveillance, maladaptation and non-specific problems were included.

Cutland et al 2009 South Africa

Midwives wrapped cotton pads soaked in water or CHX around gloved fingers. Fingers were rotated circumferentially over the cervix and vaginal walls, and the external genitalia wiped

Pregnant women (Aged 15-51) and their neonates born to South African women at Chris Hani- Baragwanathe hospital, Soweto, South Africa

Exclusion criteria were planned caesarean section, antepartum haemorrhage, known severe congenital malformation, intrauterine death, allergy to CHX, face presentation, genital warts or ulcers, full cervical dilatation, and age younger than 15 years.

No significant differences seen

No significant differences seen

No volumes or times of washing stated

0.5% CHX

Autoclaved tap water

8011 mothers 4005 to chx and 4006 to control

Neonatal and maternal Sepsis and vertical transmission GBS within 1st 3 days of life. Neonatal sepsis defined as clinical diagnosis or culture positive. Maternal sepsis defined as admission within 14 days of delivery for endometritis (at least two of uterine tenderness, fever, foul-smelling or purulent lochia, or vaginal discharge), culture confirmed infection of sterile site, or perineal wound infection among vaginal parturients.

Dykes et al 1987, Sweden

Midwives used a compress steeped in the 2g/L CHX solution. Compress turned three times around the cervix then over the vaginal walls using spiral movements outwards. Procedure was repeated twice with new compresses. Fourth compress was pressed against the cervical orifice and then used for washing of labia minora and the introitus. All patients including those in the control group also had a shower using soap containing CHX on admission and had their lower abdomen and external genitalia washed with CHX prior to delivery.

All pregnant women attending the antenatal clinics in the region served by the Department of Obstetrics, University Hospital, Lund, who were GBS positive (urogenital tract) at weeks 32 and 36 and at onset of labour.

Not stated

Not analysed

Not analysed

All participants had CHX wash including controls, no exclusion criteria e.g. for prior antibiotic use.

2g/l CHX

Standard care

78 patients in total 31 in chx and 47 in the control group.

Maternal urogenital colonisation GBS at 4 days post-partum.

Eriksen et al 1997, USA

20 cc of a 0.4% CHX solution was placed around the portio vaginalis and fornices using a syringe. Women in the control group were irrigated with 20 cc of sterile water.

Women admitted to the Lyndon Baines Johnson Hospital, Texas, USA labour and delivery room

Preterm labour, foetal distress, malpresentation, intraamniotic infection, cervical dilatation >6 cm, and known allergy to CHX.

Not reported

Not reported

Patients with prior use of antibiotics not excluded, no protocol for washing procedure

20cc 0.4% CHX

20cc sterile water

947 patients were randomized to CHX (481) or of sterile water (466)

Incidence of neonatal pneumonia, culture proven neonatal sepsis, and use of the antibiotics in the neonate. The diagnosis of neonatal pneumonia was made by the attending physician if the neonate was febrile and had chest radiograph findings consistent with the diagnosis. Neonatal sepsis was diagnosed if the infant had a positive blood or CSF culture, along with a clinical course consistent with sepsis.

Hennequin 1995, Denmark

Vaginal examinations of the treated group were systematically per- formed with gloves lubrified with 5 ml CHX digluconate 1% cream; the control group was examined with uncoated gloves.

Pregnant antenatally screened GBS positive pregnant women attending the labour ward

Not stated

Not reported

Not reported

No exclusion criteria e.g. abx use ruptured membranes etc, no protocol for vaginal examination, no training given

5 ml CHX digluconate 1% cream

Standard care

59 women in total. 28 CHX cream 31 control

Mother Infant GBS transmission.

Pereira et al 2011, Zimbabwe

Vulva cleansing with a 4x4 cotton wool ball soaked in 15-20ml 1% CHX solution followed by vaginal cleansing with another cotton wool ball as described above. The process was repeated from onset every 2 hours.

Pregnant women attending Harare central hospital who had no allergy to CHX, lived in close proximity to the hospital and planned to have a vaginal birth.

None stated

No significant difference between groups

Apgar scores were significantly higher in CHX group. However neonatal outcomes not included as had full body washing.

No exclusion criteria, no training given,

15-20ml 1% CHX

Standard care

502 women in total 2:1 randomisation 334 to chx and 168 to UC. However only 37 women were swabbed for cultures. 5 in UC 32 in chx.

Safety, acceptability and antimicrobial effect of 1% CHX. Maternal vaginal colonisation (any species) was primary antimicrobial effect measured

Rouse et al 1997, USA

Irrigations were performed either during active labour or before planned caesarean delivery by resident physicians and medical students. CHX solution containing bottles were aseptically attached to 12 cm douche nozzles. These were inserted high into the vaginal fornix, and, as completely as possible, discharged the contents of the bottles. Typically, approximately 200 ml of the irrigation was delivered.

Pregnant women at 24 weeks gestation or more at Cooper Green Hospital, hospital in Birmingham, Alabama, serving publicly funded patients

Contraindication to digital cervical examination (e.g., placenta previa), active genital herpes, chorioamnionitis before randomization, or known or suspected allergy to CHX

Significant differences seen in maternal age, nulliparous, meconium and Intrauterine pressure catheter

Not analysed

Prophylactic antibiotics given for early onset neonatal group B streptococcal sepsis for the following risk factors: anticipated delivery before 37 weeks, rupture of membranes >18 hours, history of a prior affected neonate, or intrapartum fever (temperature - >100.0 ° F)

0.2% CHX

200ml sterile water vaginal wash out pre delivery

A total of 1024 patients were enrolled: 508 in the CHX group and 516 in the placebo group.

Primary outcomes Maternal chorioamnionitis and endometritis Other outcomes; UTI and wound infection Neonatal outcomes; Sepsis, hyperbilirubinaemia, Death, necrotizing enterocolitis, supplemental oxygen, APGAR and intraventricular haemorrhage.

Rouse et al 2003, USA

See Rouse 1997 performed every 6 hours (maximum 4 irrigations)

See Rouse 1997. Patients were eligible if they were nulliparous and admitted for delivery at 32 weeks of gestation

See Rouse 1997

No significant difference between groups seen.

Not analysed

Prophylactic antibiotics given See Rouse 1997

See Rouse 1997

See Rouse 1997

1041 participants 525 in chx; 516 in control

Primary outcomes: Maternal infection - chorioamnionitis and endometritis Secondary neonatal outcomes included birth weight, Apgar scores <7, receipt of antibiotics, need for mechanical ventilation, and admission to the neonatal intensive care unit

Stray- Pedersen et al 1999, Norway

Douching started by intravaginal insertion of catheter towards the cervix. The bottle was squeezed while the catheter was retracted slowly. Patient remained supine for 5 min. Process repeated every 6 hours.

Over 9 Months pregnant women were consecutively selected from the Aker University Hospital, Norway. The first 4 months was a reference period and the next five months the intervention period.

None given

No significant difference between groups

No significant difference between groups

Ampicillin was given to women with prolonged delivery > 24 hours

120 ml 0.2% CHX douche

Reference phase standard care. Intervention phase vaginal douche with sterile saline

1989 participants 548 in chx douche 583 control (saline douche) 858 reference group (nothing)

GBS transmission, Maternal outcomes (postpartum UTI and fever) Fever was recorded when temperature exceeded 38.5°C during the first 24 h after delivery, or if the temperature thereafter exceeded 38°C on two occasions at least 4 h apart, provided that other obvious explanations were absent. Neonatal outcomes (Septicaemia, Strep. agalactiae sepsis Respiratory problems and Superficial infections)

Sweeten et al 1997, USA

Women randomized to the study arm received 20 ml of a 0.4% CHX solution. The solution was placed around the portio vaginalis and fornices with a syringe. Women in the control group were irrigated with 20 ml of sterile water.

Women admitted to Lyndon Baines Johnson General Hospital, USA labour and delivery suite at or greater than 36 weeks' gestation

Preterm delivery, foetal distress, malpresentation, intraamniotic infection, cervical dilatation >6 cm and known allergy to CHX.

No significant difference between groups

Not evaluated

No training given, no set protocol e.g. timing

20ml 0.4% CHX

20ml sterile water

CHX group 481 Placebo 466

Maternal outcomes were intraamniotic infection and endometritis. Diagnosis of intraamniotic infection was made if temperature >100°F with two of the following criteria: maternal tachycardia, uterine tenderness, foul-smelling amniotic fluid, maternal leukocytosis, or foetal tachycardia. Diagnosis of endometritis was defined as a postpartum oral temperature >101 ° F, uterine tenderness, and no other source of infection. Patients with a diagnosis of intraamniotic infection could not also be included in the endometritis group.