Our data from both the HPV screening survey of unselected pregnant women and the prospective cohort study provide for the first time estimates for HPV prevalence, type-specific distribution, mother-to-child transmission rates as well as HPV persistence in pregnant women and in infants in Spain.
Consistent with previous reports (reviewed in [2
]), our data confirm that the risk of transmission of any HPV type from infected mothers to the newborn is shown to be relatively low (9.4%), even lower (2.0%) if type-specific transmission is considered.
Although HPV-DNA detection rates in samples of newborns and infants vary widely in the literature, well conducted prospective studies suggest that the risk of perinatal transmission, although existent, is relatively low. Several studies have tested infants for HPV-DNA or antibodies ([9
] and reviews [2
]). Detection rates in the first 1 or 2 days of life range between 4% and 72% among infants born to women with genital HPV detected during pregnancy, and between 0.6% and 20% among infants delivered by women with no detectable HPV during pregnancy. Rates of detection at 6 weeks vary also widely and they are not always significantly different for infants born to HPV positive or negative mothers.
In a carefully conducted large study [13
] the HPV prevalence among infants born to HPV-positive women was 4% and among infants born to HPV-negative women was 8%. Consistent with our high rates of HPV X detection, this study found that all positive results in the infants were positive for unclassified HPV types and all of them were preceded or followed by HPV negative specimens. This report and our data clearly show that the few HPV infections detected in infants probably represent low-level genital or non-genital HPVs or may represent horizontal transmission. Taken together the evidence from this and other prospective studies [9
] strongly suggests that the risk of perinatal transmission of HPVs although existent is relatively low.
A consistent finding from our cohorts of HPV positive and negative pregnant women and their offspring is the evidence for horizontal transmission. First, we found that up to 16.9% of children born to HPV-negative mothers had HPV infections in the first 24 months of life. This percentage is only slightly lower, and not statistically significantly different, than that observed in infants born to HPV-positive mothers (19.7%). Secondly, we found an association between HPV status in the mother at the 6-week postpartum visit and the HPV status in children at the same visit or thereafter. Thus, at the 6-week post-partum visit, children of mothers' who were HPV-positive at the post-partum visit were about 5 times more likely to test HPV-positive than children of corresponding HPV-negative mothers (27.3% vs. 7.2%; age-adjusted OR = 4.8; 95% CI, 1.36–16.88; p = 0.02). In contrast, no association was found between mothers' HPV status at pregnancy and children's HPV status at any of the visits combined. Thus, all together, the data in Table indicate that the HPV detected at the post-partum visit in the mother is a stronger determinant of HPV infection in the child than the HPV detected during pregnancy, suggesting that horizontal mother-to-child transmission may play a more important role than vertical transmission in determining HPV DNA detection in children. Mothers themselves, relatives, caregivers and fomites harbour HPVs that can be horizontally transmitted to the child, in particular in the first weeks of life when there is a close caring physical contact relationship with the infant. Indeed, other study designs are needed to properly distinguish vertical from horizontal transmission. These studies should include accurate and repeated HPV detection and genotyping of multiple sites from parents, siblings and care givers as well as assays to distinguish between markers of inert HPV DNA detection and markers of active HPV infection.
In assessing HPV positivity in children born to HPV-negative mothers we can not rule out that these mothers were false HPV negatives at pregnancy. We need to take into account that 52% of the HPV-negative mothers came from the high-risk group of women included in the initial cohort study. Thus, increased HPV exposure may increase risk of false negative results which might somehow partially explain HPV transmission among HPV-negative mothers.
Our results do not support a high prevalence of HPV during pregnancy: 6.5% of the unselected group of women was positive for HPV-DNA by consensus PCR. This relatively low HPV prevalence in pregnant women correlates with the low prevalence of HPV infection in the female general population (between 1.3% and 2.4%) [14
] and the low incidence rate of cervical cancer in Spain [15
]. Still, our HPV prevalence estimate among pregnant women is between 3 and 5 times higher than that observed in the female general population, confirming the findings from other studies showing that pregnant women do have a higher HPV-DNA detection rate than un-pregnant women [6
]. It has been argued that immunological or hormonal changes could modulate the rate of HPV positivity and clearance during pregnancy [18
]. While some authors report evidence that pregnancy decreases clearance of high-risk HPV types in the first two trimesters of pregnancy [16
], others question these findings [20
As expected, and consistent with the sexual mode of transmission of HPV infections, we found that an early age at first sexual intercourse, a high number of sexual partners, particularly before the age of 20, a previous history of genital warts or STDs as well as a current squamous intraepithelial lesion in the cervix as diagnosed by cytology, were the strongest determinants of HPV positivity in pregnant women (Table ). These expected associations provide further internal validity to our complex study.
As shown in Table none of the obstetric variables collected in the study, including cesarean section delivery, were associated with HPV positivity in children at any time during follow-up. The effect of cesarean section on HPV transmission among HPV-positive pregnant women could not be assessed due to the low number of HPV-positive children born by cesarean section. Concerning reproductive variables, only ever use of hormonal contraception was associated with a reduced risk of HPV in the child. We do not have any biologically plausible explanation for this inverse association. Since such a relationship has never before been reported one should be cautious in its interpretation. Concerning HPV in the mother the only correlate found for HPV positivity in the child at any point during follow-up was the mother's HPV positivity at the post-partum visit. Again, this finding suggests that HPV at the post-partum visit may be more determinant for child's HPV infections than HPV at pregnancy. It is worth noting that mothers that had HPV status persistence from pregnancy to the post-partum visit had a higher percentage of HPV-positive children (29.2%) as compared to mothers negative at the two visits (15.4%) or HPV positive in only one of the two visits (16.3%). However, the association did not reach statistical significance (Table ).
In interpreting these results it should be considered that our reported HPV type-specific distribution is probably biased towards an overestimated detection of HPV X, as the samples that tested positive with the generic probe were tested with only seven type-specific probes (HPVs 6,11,16,18,31,33, and 39). Thus the high percentage of samples classified as HPV X could be true rare genital HPV types, cutaneous types (unlikely because of the poor efficiency of the primer system in detecting cutaneous types) or still untyped HPVs. Furthermore the MY09/MY11 primer system has been improved by the PGMY primer system in terms of higher reproducibility of primer synthesis as well as detection of mucosal HPVs. Unfortunately, very few samples remained available for re-testing with the newer PGMY system. This limitation may also have resulted in an underestimation of the true underlying type-specific concordance.
In conclusion, our study, conducted in a population at low risk for HPV and cervical cancer, confirms that high-risk HPV genotypes can be vertically transmitted to the child, although the risk of vertical transmission is relatively low. In this study, if we exclude untyped (HPV X) infections, HPV 16 has been found to be the most frequent type detected both in mothers and infants. Infants of women who tested HPV positive at six weeks after delivery are nearly five times more likely to test HPV positive than infants of HPV-negative mothers. HPV persistence in infants is a rare event. Given the substantial HPV positivity observed in children born to HPV-negative mothers, these data suggest that vertical transmission may not be the sole source of HPV infections in children and that horizontal mother-to-child transmission may play also a role. It remains to be seen whether this alternative mode of HPV transmission and acquisition may have an impact in several areas, including vaccination strategies, epidemiological studies, and the clinical management of children with HPV-associated diseases.