Cover Image
close this bookHIV in Pregnancy: A Review (UNAIDS, 1999, 67 p.)
close this folderSECTION A : HIV IN PREGNANCY
close this folderMother-to-child transmission
View the document(introduction...)
View the documentFactors affecting mother-to-child transmission of HIV-1

Factors affecting mother-to-child transmission of HIV-1

Transmission from mother to child of HIV is affected by a number of factors, not all of which have been fully elucidated. These can be divided into viral, maternal, obstetrical, fetal and infant factors as demonstrated in Table 1.

Table 1:Factors affecting mother-to-child transmission of HIV-1137, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141


Viral load

Viral genotype and phenotype

Viral resistance


Maternal immunological status

Maternal nutritional status

Maternal clinical status

Behavioural factors

Antiretroviral treatment


Prolonged rupture of membranes (> 4 hours)

Mode of delivery

Intrapartum haemorrhage

Obstetrical procedures

Invasive fetal monitoring




Multiple pregnancy



Gastrointestinal tract factors

Immature immune system

Viral factors

Viral load

Transmission is increased in the presence of high levels of maternal viraemia. Clinical observations of increased transmission in these situations, such as in advanced disease and at the time of seroconversion, are supported by the presence of high levels of p24 antigenaemia142,143. With the development of new techniques for the measurement of the virus, such as quantitative Polymerase Chain Reaction (PCR) DNA and RNA, an association has been shown between the maternal viral load and the risk of transmission from mother to child144,145,146,147,148,149,150,151,152,153. More than half of the women with viral loads of>50 000 RNA copies per ml at the time of delivery have been shown to transmit the virus145,154. A New York study showed a mean viral load of 16 000 RNA copies/ml in transmitters and 6600 RNA copies/ml in non-transmitters145. Women in this study with measurable viral loads were almost six times more likely to transmit than those in whom the virus was undetectable, after controlling for the CD4+ count. In a French study, transmission rates increased with increasing viral load: 12% in those with less than 1000 copies/ml compared with 29% in those with more than 10 000 copies/ml144. Few studies have shown a threshold viral load for transmission and it appears that it can occur at low viral levels, for reasons which are not well understood, but which probably reflect the multiple influences acting on mother-to-child transmission145,154,155.

The local viral load in cervico-vaginal secretions and in breast milk may also be an important determinant of transmission risk intrapartum and through breastfeeding156,157,158. HIV-1 levels in these fluids have been shown in most studies to be correlated with CD4+ count and plasma viral load156,159,160,161. The presence of sexually transmitted infections or other causes of inflammation, vitamin A deficiency and local immune response may affect viral shedding162. In Rwanda, postnatal transmission was associated with the presence of HIV-1 infected cells in breast milk161.

Maternal antiretroviral therapy during pregnancy is thought to reduce transmission partly through the reduction of viral load, although the mechanism may also include post-exposure prophylaxis in the child after birth, as the use of zidovudine has been shown to reduce transmission at all levels of maternal viral load111,146,163. Combination antiretroviral therapy may be more effective in preventing transmission due to the greater reductions in viral load, but no large-scale study results are available yet on this164.

Viral genotype and phenotype

A number of HIV-1 sub-types or clade groups have been identified, with differing geographical distributions165. There is little evidence on the effect of sub-type on infection or transmission, although some studies have shown an increased in-vitro ability of sub-type E to infect epithelial cells from the vagina and cervix156,166. The subtype may affect the cell tropism of the virus, and in turn the infectivity, in-utero, through genital infection or in breast milk.

Most studies on viral variants in mothers and children have demonstrated that the strains in the infant are a distinct subset of maternal virus, although the major maternal variant has also been shown to be transmitted149, 156, 131, 167, 168, 169, 170, 171. Different viral phenotypes show differing tissue tropism. Macrophage-tropic non-syncytium-inducing (NSI) viral isolates appear to be preferentially transmitted to children even when the dominant maternal strains are syncytium inducing (si)149,156,131,172,173,174. There may be a difference in disease progression for the child related to the viral strain. Rapid/high virus isolates have been associated with transmitting mothers whereas slow/low virus isolates were associated with non-transmitting mothers175,176.

Increased strain diversity in the mother may theoretically influence the rate of transmission. Repeated exposure to different viral strains through pregnancy, occurring through unprotected intercourse may be the mechanism responsible for the observed increase in transmission in these cases177. The development of resistance to zidovudine during pregnancy has been shown to be infrequent, but concern has been expressed that the possible development of resistant strains of HIV-1 in women receiving zidovudine monotherapy during pregnancy may result in higher transmission rates in subsequent pregnancies178,179. Since the risk of resistance emergence increases with the duration of treatment and since resistance to AZT usually emerges after 3-4 months of treatment, the emergence of resistance with short-regimen of zidovudine (1 month) is very low and less likely compared to long regimen of zidovudine.

Maternal factors

Maternal immunological status

Transmission from mother to child is more likely with decreased maternal immune status, reflected by low CD4+ counts, low CD4+ percentages or high CD4+/CD8 ratios5,180. These in turn may be markers for higher viral loads, as opposed to risk factors in themselves, although an interaction between viral load and immune response may be present. In the European Collaborative Study (ECS), there was an increased risk of mother-to-child transmission where maternal CD4+ counts were below 700/cubic mm142. Transmission increased almost linearly in this study with decreasing CD4+ counts181. Several other studies have noted similar associations181,182,183,184. In the WITS study, the association between low CD4+ percentages and transmission was only seen in women without persistently positive viral cultures. Where there was at least one negative culture and high CD4+ cell percentages, transmission rates were in the range of 1-4%182.

There have been conflicting results about the role of neutralizing antibodies in preventing transmission. Some studies have shown that high levels of maternal neutralizing antibody are associated with lower rates of transmission, while in others no association was observed185,186,187,188. Women who transmit in-utero may have lower levels of autologous neutralizing antibody than those who do not transmit, or those women where transmission occurs intrapartum131. Antibody to the V3 loop of HIV-1 envelope gpl20 has not been shown to be protective, neither do antibody-dependent cellular cytotoxicity (ADCC) antibodies appear to be protective189,190,191. One report has correlated maternal antibodies to the carboxy region of the gp41 envelope glycoprotein with lack of vertical transmission192. The involvement of specific T-cell immunity in the pathogenesis of mother-to-child transmission has yet to be determined.

Little is known about the role of mucosal HIV-1 antibodies and viral shedding in the genital tract which may affect intrapartum transmission rates156,160. Infection through breastfeeding has been associated with a lack of IgM and IgA in breast milk193,194.

Maternal nutritional factors

Serum vitamin A levels in HIV-1 positive mothers have been correlated with the risk of transmission in a Malawi study. The mean vitamin A level in those mothers who transmitted the virus to their children was significantly lower than in those who did not transmit. Women with vitamin A levels below 1.4 umol/1 had a 4.4-fold increased risk of transmission, which dropped with increasing vitamin A levels195. One US study showed no relationship between low vitamin A levels and transmission196, while another cohort study did show a correlation197. The mechanism of vitamin A effect is uncertain, but the influence of vitamin A on the integrity of the vaginal mucosa or placenta and the immune stimulatory properties of the vitamin have been suggested 162, 198. Alternatively, low vitamin A levels may be a marker for other deficiencies or behavioural factors, which influence transmission. Other micronutrients, including zinc and selenium, have been suggested as having a possible role.

Behavioural factors

Several behavioural factors have been associated with an increased rate of transmission from mother to child. These include cigarette smoking199,200 and maternal hard drug use119,201,202.

Unprotected sexual intercourse during pregnancy has been linked to an increased risk of mother-to-child transmission. A transmission rate of 30% was shown in women who had more than 80 episodes of unprotected sex during pregnancy compared with 9.1% in those with no unprotected intercourse203. A similar association is suggested in two African studies204,205. This may be due to an increased concentration or strain diversity of HIV-1, or the effect of cervical or vaginal inflammation or abrasions. An increase in chorioamnionitis has previously been reported linked to sexual activity in pregnancy206, and this may be an alternative mechanism. The presence of sexually transmitted infection during pregnancy has been correlated with increased risk of transmission207, and STIs have been shown to increase viral shedding in cervico-vaginal secretions32.

Placental factors

Placental factors have been implicated in transmission of the virus from mother to child116, 208,209,210,211.Placental infection with HIV-1 has been reported and Hofbauer cells and possibly trophoblasts express CD4+ and are thus susceptible to infection212. An association between increased transmission and the presence of chorioamnionitis was described early in the epidemic. Other placental infections and non-infectious conditions such as abruptio placentae have also been implicated162,213,214. Breaks in the placental surface can occur at any stage of pregnancy and may be related to transmission, although the significance of these may, in turn, depend upon the maternal viral load215. Smoking and drug use, both associated with increased transmission, may exert this effect through placental disruption116. In areas of high malaria prevalence, infection of the placenta is common in pregnancy. Placental P. falciparum infestation has been associated with poorer survival in infants born to HIV-1 positive mothers in Malawi, which may represent increased transmission rates2'6 and with higher rates of transmission from mother to child in Kenya217.

Obstetric factors

With the majority of mother-to-child transmission occurring at the time of labour and delivery, obstetric factors are important determinants of transmission. Suggested mechanisms for intrapartum transmission of HIV-1 include direct skin and mucous membrane contact between the infant and maternal cervico-vaginal secretions during labour, ingestion of virus from these secretions, and ascending infection to the amniotic fluid113,207. HIV-1 in cervico-vaginal secretions may be raised four-fold during pregnancy218. The higher rate of infection in first-born twins may be due to longer exposure of the infants to infected secretions219.

Several obstetric factors have been implicated, although results are not consistent across studies with regard to the relative importance of different obstetric factors. In the French perinatal cohort study, preterm delivery, intrapartum haemorrhage and obstetric procedures were related to transmission risk207. Other factors such as the use of fetal scalp electrodes, episiotomy, vaginal tears and operative delivery have been implicated in some studies but not in others113,120,181,214,220.

The duration of labour does not appear to be as important as the duration of rupture of membranes182,221. Prolonged rupture of membranes has been associated with increased risk of transmission in a number of studies and is an important risk factor119,207,222. In an American study, duration of ruptured membranes of over four hours nearly doubled the risk of infection, regardless of the eventual mode of delivery119.

Delivery by Caesarean section has been shown to be protective in some prospective follow-up studies, but not in all181,119,223,224,225. This has now been confirmed in a randomized controlled trial in Europe226. A Swiss study showed an additive protective effect of elective Caesarean section for women receiving antiretroviral treatment227. In France, women who received long-course antiretroviral treatment in pregnancy and had an elective Caesarean section had a transmission rate of less than 1%228.

Fetal factors

Fetal genetic factors may play a part in transmission. Little is known yet about the role of genetic factors such as the CCR-5 delta32 deletion and human leukocyte antigen (HLA) compatibility of mother and infant in the determination of transmission risk229,230,231. Concordance between infant and maternal HLA has been associated with increased risk of transmission232.

Preterm infants have higher reported rates of transmission of HIV-1 in several studies111,181,220,233. Women with low CD4+ counts are more likely to have preterm deliveries, which may influence this finding. The higher rates of infection seen in firstborn twins have been widely reported and have formed part of the evidence for the role of intrapartum transmission234,235. This effect is more pronounced in vaginally delivered twins, where a two fold increase in infection is seen in first born twins over second born, but is also present in twins delivered by Caesarean section219.

Other fetal factors may include co-infection with other pathogens, fetal nutrition and fetal immune status132.

Infant factors

Breastfeeding is responsible for a high proportion of mother-to-child transmission in developing countries, where 30% or more of perinatal HIV infections will occur through breast milk. This is less common in the developed world, where most HIV-positive women will not breastfeed. Breast milk contains both cell associated and free virus, the amount of which may be related to the immune suppression of the mother and vitamin A levels159,236. Other protective factors are also present in breast milk, including mucins, HIV antibodies, lactoferrin, and secretory leukocyte protease inhibitor (SLPI)111,161,234,132.

A meta-analysis of studies of transmission through breastfeeding showed the additional risk of transmission through breastfeeding to be between 7 and 22%, equivalent to a doubling of transmission rates 237. That is to say that, in breastfeeding populations, between one third and one half of mother-to-child transmission occurs during breastfeeding. A Soweto study has shown transmission rates of 18% in formula fed infants compared with 42% in breastfed238. Rates are higher when the mother sero-converts during breastfeeding, where the estimated additional risk is around 30%237,239. The risk of breast milk transmission may also depend upon other factors, such as maternal disease stage, breast abscesses, mastitis, nipple cracks, patterns of breastfeeding (i.e. exclusive or mixed), maternal vitamin A and oral thrush in the child111,193. A Zimbabwe study showed that 31% of breastfeeding mothers of HIV-1 infected children had active nipple disease240.

Late postnatal transmission, after the age of six months, has been described in a number of studies111,220,130,241. In Abidjan, 12% of infants born to HIV-1 positive mothers were diagnosed as HIV-infected after the age of six months but may have been infected earlier242.

The risks of postnatal transmission may also be related to other factors in the newbom. HIV entry may occur through the gastro-intestinal tract following ingestion of virus in utero or at birth83,132. There is decreased acidity, decreased mucus, lower IgA activity and thinned mucosa in the newborn gastro-intestinal tract, which may facilitate transmission"''234'132. The newbom immune system may also be deficient in macrophage and T cell immune response132, increasing the susceptibility to infection. At least part of the effect of antiretroviral drugs in pregnancy appears to be due to a post exposure prophylaxis effect after birth113.

Further information on HIV transmission and breastfeeding can be found in the joint UNAIDS, UNICEF and WHO publication A review of HIV transmission through breastfeeding (WHO/FRH/NUT/CHD/98.3/UNAIDS 98.5)/UNICEF/PD/NUT/(J) 98.1.