Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Caesarean sections reduce the maternal-fetal transmission rate of human papillomavirus infection.

Three Part Question

In [a pregnant woman with genital warts] does [elective caesarean section] reduce the [infection rate in her baby]?

Clinical Scenario

You are the neonatal doctor covering the labour ward. A woman who is 38 weeks pregnant and in early labour reveals a recent history of genital warts. She asks the midwife if this could affect her baby and if it should influence the mode of delivery. The midwife is unsure and so asks you

Search Strategy

Medline via PubMed was the primary source of articles.
Medline: Search terms were: [genital warts OR papillom* OR HPV OR condylom*] AND [labour OR labor OR delivery OR caesarean]. A second search using [papillom*] AND [vertical] was also performed.

Limits were: humans and English language. Dates included 1966 to October 2008.

Secondary searches on the Cochrane database, Clinical Evidence and SUMsearch were performed using the same search terms.

Search Outcome

A total of 295 papers were found via PubMed, eight of which were relevant. The references of the above papers were scanned, along with the linked articles, but no further articles were found.

A review (Medeiros) which partly looked at this question, had been carried out in 2005, but due to only partial analysis of data, the original articles were reviewed instead.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Puranen et al,
105 pregnant women, sampled 2 days around delivery (86 asymptomatic, 19 with clinical HPV signs) All babies tested immediately after birth No follow-up Prospective cohort (level 1b)HPV presence in maternal cervical scrapes and neonatal NPA fluidHPV genotype (2, 6, 7, 11, 16, 18, 30, 31, 33, 53, 66)HPV +ve mothers 42/106 (40%)Transmission from HPV +ve mothers 34/42 (81%)VD = 26/31 = 84%CS = 8/12 = 67%(p = 0.24)HPV +ve babies with HPV −ve mothers 6/64HPV transmission in VD: * Symptomatic 16/18 (89%) * Asymptomatic 17/60 (28%) (p<0.01)HPV transmission in CS: * Symptomatic 1/1 (100%) * Asymptomatic 7/26 (27%) (p = 0.3) “Clinical HPV” was defined as either visible condyloma or condylomatous changes in cytological samplesUsed NPA which is poorly sensitive at detecting HPV Samples taken straight after birth so may have been contamination in airways from amniotic fluid There was increased risk of transmission during VD if symptomatic (p<0.01) but not in CS28% of HPV +ve babies were from mothers with no current symptoms, so subclinically infected women can still transmit HPV
Wang et al,
73 pregnant women in 3rd trimester (no information given on presence or absence of symptoms) Not all babies screened (age not defined) (n = 34) No follow-upProspective cohort (level 1b)HPV DNA in fetal membranes, NPA, amniotic fluid and maternal cervical cellsHPV genotype (16, 18, 35)HPV +ve mothers 26/73 (36%)Transmission from HPV +ve mothers 11/26 (42%)VD 7/14 (50%)CS 4/12 (33%)(p = 0.45)HPV +ve babies with HPV −ve mothers 0/47Increased risk of perinatal complications in HPV +ve mothers (p<0.01)Used NPA which is poorly sensitive at detecting HPV Only investigated HPV 16, 18 and 35 Counted positive amniotic fluid as infected baby Degree of HPV positivity varied depending on where sampling was done:fetal membranes = 42.5% +veamniotic fluid = 15.4% +veNPA = 14.7% +ve They don’t comment if all babies from HPV +ve mothers were screened or not, or how they decided which babies to screen
Tseng et al,
301 pregnant women tested at 36–40 weeks (no information given on presence or absence of symptoms)All babies born to HPV +ve mothers at 3–4 days (n = 68) Control group of HPV −ve women and their babies (n = 30) No follow-upProspective cohort (level 2b)HPV DNA in maternal cervical swabs, buccal and genital swabs of neonatesHPV genotype (16, 18)HPV +ve mothers 68/301 (23%)Transmission from HPV +ve mothers 27/68 (40%)VD = 18/35 (51%)CS = 9/33 (27%)Significantly higher rate of HPV 16 and 18 in VD infants (p = 0.042)HPV +ve babies with HPV −ve mothers 3/233Only investigated HPV 16 and 18 Was the only study to exclude anyone from CS group with ruptured membranes, vaginal bleeding or in labour Neonatal swabs taken 3–4 days after delivery (longer time interval than other studies), so may be more reliable indicator of transmission rather than contamination Increased sensitivity in HPV detection in neonatal buccal swabs compared to genital swabs (40% vs 31%)
Shah et al,
All individuals in RRPF database (via otolaryngologists)138 JORRP casesRetrospective cohort (level 2b)Presence of JORRP (at <14 years)Type of deliveryIn JORRP cases 4.6 times fewer than expected CS for that year (6 vs 27) (p<0.001)3/6 CS who developed JORRP were contacted and all ruptured membranes pre-delivery 6 CS when 27 expected Other 3 JORRP cases after CS may have also ruptured membranes pre-delivery but no data available Although study open to bias as retrospective data collection, other types of study are difficult to do as rare condition
Tenti et al,
711 pregnant asymptomatic women, sampled just before delivery NPA from all babies in delivery room 18 months’ follow-upProspective cohort (level 1b)HPV DNA in maternal cervicovaginal lavage and neonotal NPAHPV genotype (6, 11, 16, 18, 33)HPV +ve mothers 37/711 (5%)Transmission from HPV +ve mothers 11/37 (30%)VD = 11/29 (38%)CS = 0/8 (0%)(p = 0.04)HPV +ve babies with HPV −ve mothers 0/674Used NPA which is poorly sensitive at detecting HPV Samples taken straight after birth so may have been contamination in airways from amniotic fluid All HPV +ve babies were negative by 5 weeks ROM of >2 h pre-delivery significantly increased HPV transmission (p = 0.009). (ROM >4 h, p = 0.001)
Bandyopadhyay et al,
135 pregnant women with no history of genital warts, sampled 10–15 days pre-delivery All babies at 1–4 days of age 12 months’ follow-up for all positive subjectsProspective cohort (level 2b)HPV DNA in cervical cells of mothers, buccal cells of infantsHPV genotype (6, 11, 16, 18, 31, 33)HPV +ve mothers 38/135 (28.1%)Transmission from HPV +ve mothers 7/38 (18%)VD = 2/11 (18%)CS = 5/27 (19%)(p = 0.68)HPV +ve babies with HPV −ve mothers 7/97Used consensus primer for HPV detection which can be less sensitive than type-specific primers Babies born via VD had greater chance of acquiring infection from HPV +ve than HPV −ve mothers (p = 0.028), but no difference in CS (p = 0.46)
Silverberg et al,
3033 births with maternal history of genital warts in pregnancy (from Danish national registry)57 cases of JORRP from ENT medical records Total of 1 206 213 births in period Retrospective cohort (level 2b)Absolute risk of JORRPMode of deliveryIncreased delivery timeRisk of JORRP with maternal HPV disease = 6.9/1000 births – 230 times increased relative risk than if no warts history (0.03/1000)Expected CS rate 15.7%Actual CS rate 12.5%Relative hazard on multivariate analysis was not significantly different:VD = 1CS = 0.95Delivery time >10 h = twice transmission risk (multivariate analysis)Relied on correct coding in previous discharge summaries to make sure no cases missed Although there was no significant difference in CS rate, there were only 7 CS in the study, of which only 3 were electively performed The low CS rate in Denmark reduces the effectiveness of study power
Smith et al,
574 pregnant women, sampled just before delivery: 432 had no HPV history, 64 had genital warts history, 109 had previous dysplasia or cervical cancer (no information given on presence or absence of symptoms at time of study)68 fathers sampled Babies sampled a mean of 65 h after birth (time range not specified in paper)6-month follow-upProspective cohort (level 2b)HPV DNA from genital and oral swabs from women and neonates, oral swabs from men HPV genotype (24 in total: 6, 11, 13, 16, 18, 31, 33, 35, 38, 39, 44, 51, 53, 54, 56, 58, 59, 61, 66, 68, 69, 70, 83, 84 and 6 un-named types)HPV +ve mothers 172/574 (30%)Transmission from HPV +ve mothers 6/172 (3%)VD = 8/148 (5%)CS = 1/24 (4%)(p = 0.64)HPV +ve babies with HPV −ve mothers 3/4023/9 HPV +ve babies at birth attended follow-up. All were HPV −ve on oral swabNo babies from HPV +ve fathers were positive but small sample size Follow-up incomplete, but no negative babies became positive Less than 1% type-specific concordance in vertical transmission and overall low rate of HPV transmission (3%) They report a non-significant increased rate of neonatal HPV infections in mothers with a history of cervical cancer (14.3% vs 1.6%, p = 0.1), but no raw data given and not separated by delivery method


There are more than 100 genotypes of human papillomavirus, of which at least 40 are known to infect the genital area. Although seropositivity is documented to be highest in the second decade of life (Chen), most studies investigating the age-specific prevalence of human papillomavirus infection (HPV) only include women aged over 18. During pregnancy there is increased virus activation with larger lesion size, higher infection rates and the potential for maternal–fetal transmission. Maternal HPV positivity significantly increases the risk of transmission to the baby, with overall transmission reported to be between 4% and 87% (Syrjänen). This can result in neonatal infection with progression to genital or oral lesions with potential malignant transformation, and laryngeal lesions causing life-threatening respiratory obstruction.

Elective caesarean sections reduce vertical transmission of herpes simplex, HIV and hepatitis B, but the role of elective caesarean section in preventing HPV transmission remains controversial. A 1996 survey of obstetricians and otolaryngologists on the benefits of a caesarean section in preventing a longer term complication of early HPV infection (juvenile-onset recurrent respiratory papillomatosis, JORRP) found no firm consensus: 23% of respondents would recommend an elective caesarean section, 19% would not recommend and 58% had no preference (Kosko).

Vertical transmission can occur at any point from the sperm at fertilisation (Lai)to delivery, but the clinical significance remains unknown as long-term studies differentiating between simple colonisation and significant infection have not been carried out and some studies show HPV persistence for more than 3 years after birth (Rintala ), while others show clearance at less than 2 months (Tenti).

The six studies which compared delivery method to rate of vertical transmission encompassed different populations (symptomatic and asymptomatic women), used differing methodologies (eg, neonatal buccal swabs are more sensitive than nasopharyngeal aspirates and genital swabs (Tseng, Smith, Rintala) and Southern blot confirmation and reamplification is more sensitive than just PCR in virus detection and typing) and different timings of specimen collection. This makes combining these studies for interpretation difficult. Three studies (Wang, Tseng, Tenti) showed a significant reduction in HPV transmission from positive mothers after caesarean section rather than vaginal delivery, but three (Puranen,Bandyopadhyay,Smith) showed no difference.

When all six studies are combined there is no statistically significant difference in vertical HPV transmission from HPV positive mothers between delivery methods (23% transmission after caesarean section vs 27% after vaginal delivery, p = 0.29) and the number needed to treat (NNT) is 30, which is a large number of “unprotective” caesarean sections to prevent the transmission of HPV to one baby. However, only one study (Tseng) excluded women with ruptured membranes from the caesarean group and their results showed a significant increase in HPV transmission after vaginal delivery compared to elective caesarean section prior to membrane rupture (p = 0.042). The NNT from this study is 5 with an absolute risk reduction of 24% after caesarean section. As a membrane rupture of only 2 h pre-delivery significantly increases transmission rate (Tenti),membrane rupture pre-caesarean could have confounded results in all other studies. Other potentially biasing factors not considered in the studies included maternal viral load at delivery (shown to increase transmission risk when high, Kaye) and length of second stage (when infants would be in most contact with infected maternal tissues).

With regard to the initial question, only one of the six prospective studies actively distinguished between women with and without clinical signs of HPV infection(Puranen), while three gave no clinical information at all. Therefore, potentially only 19 out of 1899 women in the combined studies were symptomatic at delivery. Within this small subgroup, babies of women who were symptomatic and delivered vaginally were significantly more likely to test positive for HPV (p<0.01). This increased transmission rate was not seen in symptomatic women delivering via caesarean section, although small total numbers (n = 1 in symptomatic caesarean section group) make the data un-interpretable. So although evidence may exist that caesarean section can reduce the maternal–fetal transmission rate of HPV, this is mainly from information gathered in asymptomatic women, or from studies which did not actively recruit symptomatic women. However, these data are still very useful as an estimated 80% of women acquire HPV before aged 50 (CDC), with only 1% having clinically visible genital warts at any one time (with only 1% having clinically visible genital warts at any one time (Koutsky),and the babies of such asymptomatic carriers are still at risk of virus transmission and associated future complications.

Two retrospective studies looked at the incidence of JORRP in relation to delivery method. Shah et al showed that JORRP cases had a higher vaginal delivery rate than expected based on the normal pattern of deliveries for that year and the three positive cases after caesarean section whom they were able to contact had the confounding factor present of ruptured membranes prior to delivery. Silverberg et al linked increased delivery time with increased risk of developing JORRP but found no difference between mode of delivery, although the small power limited the study.

HPV strains in mothers and neonates were often disconcordant (Puranen, Smith) and some HPV positive neonates were born to HPV negative mothers, representing the natural fluctuation of HPV infection in the mothers with previous fetal transmission, paternal infection from sperm cells, post-natal horizontal spread from other sources in the limited period before testing or maternal infection in tissues not studied (eg, placenta). Only four of the six prospective studies investigated HPV genotypes 6 and 11 which are associated with 90% of genital warts, and although all six did look for HPV strain 16, which is seen in 50% of cervical cancers, none looked for all three of the other commonest cancer-associated genotypes (18, 31, 45). Not all HPV strains are present equally in the population – the commonest in these studies being 16, 33 and 51 (Puranen, Smith) – but unfortunately no study separated out transmission rates of each individual genotype by delivery method.

In 10–20 years’ time the new HPV vaccination should have a knock-on effect by eliminating HPV genotypes 6, 11, 16 and 18 infection in mothers, and thereby reducing vertical transmission.

Editor Comment

CS, caesarean section; HPV, human papillomavirus; HPV +ve (−ve), HPV detected (not detected) by screening methodology used in that study; JORRP, juvenile-onset recurrent respiratory papillomatosis; NPA, nasopharyngeal aspirate; ROM, rupture of membranes; RRPF, recurrent respiratory papillomatosis foundation; VD, vaginal delivery.

Clinical Bottom Line

There is a significant risk of vertical human papillomavirus (HPV) transmission, even from asymptomatic women, which increases with length of time between membrane rupture and delivery. (Grade A)

Large obstructive condylomas necessitate a caesarean section if they cause a physical obstruction to vaginal delivery. (Grade A)

Although it is premature to routinely offer a caesarean section to all HPV positive women, once other factors such as maternal morbidity after caesarean section have been considered, there is a role for maternal choice if the mother’s membranes are still intact and they have additional risk factors such as high HPV viral load. (Grade B)


  1. Medeiros LR, Moraes Ethur AB, Hilgert JB. Vertical transmission of the human papillomavirus: a systematic quantitative review. Cad Saude Publica 2005;21:1006–15.
  2. Puranen MH, Yliskoski MH, Saarikoshi VS, et al. Exposure of an infant to cervical human papillomavirus infection of the mother is common. Am J Obstet Gynecol 1997;176:1039–45.
  3. Wang X, Zhu Q, Rao H. Maternal-fetal transmission of human papillomavirus. Chin Med J (Engl) 1998;111:726–7.
  4. Tseng CJ, Liang CC, Soong YK, et al. Perinatal transmission of human papillomavirus in infants: relationship between infection rate and mode of delivery. Obstet Gynecol 1998;91:92–6.
  5. Shah KV, Stern WF, Shah FK, et al. Risk factors for juvenile onset recurrent respiratory papillomatosis. . Pediatr Infect Dis J 1998;17:372–6
  6. Tenti P, Zappatore R, Migliora P, et al. Perinatal transmission of human papillomavirus from gravidas with latent infection. Obstet Gynecol 1999;93:475–9.
  7. Bandyopadhyay S, Sen S, Majumdar L, et al. Human papillomavirus infection among Indian mothers and their infants. Asian Pac J Cancer Prev 2003;4:179–84.
  8. Silverberg MJ, Thorsen P, Lindeberg H, et al. Condyloma in pregnancy is strongly predictive of juvenile-onset recurrent respiratory papillomatosis. Obstet Gynecol 2003;101:645–52.
  9. Koutsky LA . Epidemiology of genital human papillomavirus infection. Am J Med 1997; 102(5 Suppl 1):3–8.
  10. Centers for Disease Control and Prevention. Genital HPV infection – CDC fact sheet. Atlanta, GA: Centers for Disease Control and Prevention, 2004
  11. Kaye JN, 2. Cason J, 3. Pakarian FB, 4. et al Viral load as a determinant for transmission of human papillomavirus type 16 from mother to child. J Med Virol 1994;44:415–21.
  12. Rintala MA, Grénman SE, Järvenkylä ME, et al . High-risk types of human papillomavirus (HPV) DNA in oral and genital mucosa of infants during their first 3 years of life: experience from the Finnish HPV Family Study. Clin Infect Dis 2005; 41:1728–33.
  13. Lai YM, Yang FP, Pao CC Human papillomavirus deoxyribonucleic acid and ribonucleic acid in seminal plasma and sperm cells. Fertil Steril 1996;65:1026–30.
  14. Kosko JR, Derkay CS Role of caesarean section in prevention of recurrent respiratory papillomatosis – is there one? Int J Pediatr Otorhinolaryngol 1996;35:31–8.
  15. Syrjänen S, Puranen M . Human papillomavirus infections in children: the potential role of maternal transmission. Crit Rev Oral Biol Med 2000;11:259–74.
  16. Chen CJ, Viscidi RP, Chuang CH, et al . Seroprevalence of human papillomavirus types 16 and 18 in the general population in Taiwan: implication for optimal age of human papillomavirus vaccinations. J Clin Virol 2007;38:126–30.
  17. Smith EM, Ritchie JM, Yankowitz J, et al . Human papillomavirus prevalence and types in newborns and parents. Sex Transm Dis 2004;31:57–62.