Perinatal Acquisition of Drug-resistant HIV-1 Infection: Mechanisms and Long-term Outcome

Constance Delaugerre; Marie-Laure Chaix; Stephane Blanche; Josiane Warszawski; Dorine Cornet; Catherine Dollfus; Veronique Schneider; Marianne Burgard; Albert Faye; Laurent Mandelbrot; Roland Tubiana; Christine Rouzioux


Retrovirology. 2009;6(2):85 

In This Article


In France, early strategies intended to prevent vertical HIV transmission are now considered suboptimal until the recommendations of HAART in 2004.[5] Indeed, newborns are at a high risk of acquiring drug resistant variants emerging from their primary HIV-1 infection under antiretroviral selective pressure.[19]

In this study, we retrospectively detected resistance mutations in 20% of children born between 1997 and 2004 who were enrolled in the ANRS-EPF cohort. Interestingly, the same frequency (7 of 34, 20%) was noted in the same cohort during the period 1994–1996,[7] even though the rate of vertical transmission was lower in the more recent period. However, whereas only zidovudine resistance was detected in 1994–1996, more varied resistance profiles were found in 1997–2004, owing to the increased diversity of antiretroviral combinations used to treat pregnant HIV-1-infected women. Resistance to NRTI remained predominant throughout the study period. The most frequent mutations were those associated with resistance to zidovudine and lamivudine, which are the only antiretroviral drugs licensed for use in neonates. Only 3% of the children (n = 2) harbored variants resistant to NNRTI, compared to 12% in American studies,[13,14] probably owing to more widespread use of NNRTI-containing regimens to treat pregnant women in the USA.[20] In our study, only one child had PI resistance mutations, reflecting the recent recommendation of PI-containing regimens for PMTCT and a higher genetic barrier to resistance with ritonavir-boosted PI-containing regimens.

In most of the children studied here, the resistance profiles were related to antenatal and post partum antiretroviral drug exposure. This contrasts with the lack of relationship between antiretroviral drug resistance in newborns and perinatal antiretroviral exposure observed in New York State.[13,14] However, no information on maternal antiretroviral treatment and no maternal resistance genotyping were available in the latter studies.

The comparison of the maternal and neonatal drug resistance profiles pointed to two different mechanisms of acquisition of resistant variants by infants in the perinatal period (Figure 4). First, the infant could acquire drug-resistant variants directly from the mother (A), in one of two situations: i) the dominant variant in the mother also became dominant in the child, ii) a minor resistant variant transmitted by the mother was selected in the child during perinatal antiretroviral prophylaxis, particularly in the case of drugs such as nevirapine and lamivudine that have a low genetic barrier to resistance. Indeed, a single mutation is enough to confer high-level resistance to lamivudine or nevirapine. Moreover, selective pressure in the fetus is facilitated by the high transplacental diffusion of both these drugs.[21,22] Resistant mutations were detected early in infant lymphocytes. Clonal and longitudinal analyses showed that primary acquisition of resistant viruses was associated with long-term persistence in the infant's cellular reservoir; no matter what the subsequent treatment was.

Figure 4.

Mechanisms of antiretroviral resistance acquisition in HIV-1-infected newborns. Wild-type viruses are shown in green and resistant viruses in red. The length of the yellow-to-red arrow indicates the duration of perinatal prophylaxis and thus the risk of resistance selection.

In the second mechanism, the newborn initially acquires wild-type virus from the mother (B) (figure 4). Drug resistance can then arise during suboptimal zidovudine prophylaxis. Cloned viruses from the infants' cellular compartment were indeed wild-type, and wild-type viruses re-emerged when prophylaxis ended. Alternatively, minor resistant variants circulating in the mother may be undetectable at the clonal level in maternal samples, and/or resistant strains present in the female genital tract could be different from those circulating in the plasma.[23]

Persaud et al. reported that drug-resistant HIV-1 in perinatally infected infants can fully populate the resting CD4+ T cell reservoir early in the course of infection and persist for years in replication-competent form.[24] Moreover, resistance acquisition and long-term persistence have been described after PMTCT with a single dose of nevirapine or lamivudine in resource-poor settings.[25–27] This long-term persistence in the cellular reservoir is reminiscent of the situation described in adults initially infected by resistant viruses.[28–32] As in adults, new resistance mutations can be acquired during suboptimal treatment with residual viral replication.[31] Our results underline the advantages of using HAART for PTMTC instead of suboptimal regimens that include drugs with a low genetic barrier to resistance and a long pharmacological half-life, as currently used in developing countries.

In the second mechanism, withdrawal of zidovudine prophylaxis led to the re-emergence of wild-type virus that had been archived during the primary infection. Once again, this resembles the situation in adults who acquire drug-resistant viruses during antiretroviral failure and in whom a dominant wild-type viral population re-emerges when antiretroviral therapy is stopped.[33]

Our clonal analysis suggests that all archived viruses arising from the first mechanism are resistant (100% resistant cellular clones in children #9 and #11), compared to about 10% resistance in those arising from the second mechanism (10% resistant cellular clones in child #12).

Importantly, the main difference between primary-infection in infant and adults was the use of sub-optimal antiretroviral prophylaxies in infants that could select for resistant viruses if the infection occurs.

We observed mutations associated with resistance to at least one antiretroviral drug in six children (10%), with NRTI resistance in four, NNRTI resistance in two, and PI resistance in one. Recently, Lockman et al. showed that virologic failure of Triomune® was more frequent in infants who were previously exposed to a single dose of nevirapine rather than a placebo.[34] In contrast, Persaud et al. reported that RT resistance-associated mutations did not preclude the suppression of HIV-1 replication after 24 weeks of lopinavir/ritonavir-based HAART.[24] This result together with our findings supports the use of boosted-PI regimens in children with resistance mutations or unknown resistance status.

In conclusion, our findings support resistance genotyping for children at diagnosis of HIV-1 infection, before treatment initiation, including children born to untreated mothers.[35] This approach could avoid jeopardizing drug treatment efficacy as demonstrated in adults.[36] Importantly, resistance testing in both the infant's plasma and lymphocytes would help to show whether resistance is likely to persist, with major implications for long-term treatment.

Our results also support current French recommendations to perform resistance genotyping in HIV-1-infected pregnant women in order to formulate both maternal and neonatal antiretroviral prophylaxis.[5] Finally, it is essential to use HAART and to avoid suboptimal regimens because early resistance acquisition can have drastic long-term consequences.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.