Despite the benefits of short treatment courses, the use of monotherapy or dual therapy in HIV disease has consistently been plagued by the development of resistance to those drugs. This also appears to be the case with regimens for preventing mother-to-child transmission.

For example, studies have shown that some AZT resistance occured in many women taking AZT monotherapy in pregnancy in American and European studies, with a prevalence of between 9 and 17% (Colgrove 1999; Kully 1999; Palumbo 2001). Some studies have also noted an increased transmission risk in women with AZT resistance. This is of particular concern if the child should become infected, as there is a chance that the child's future treatment options will be limited (Masquelier 2001; Welles 2000).

Restriction of AZT monotherapy to women with low viral loads and better immune function has resulted in a lower incidence of resistance and transmission (Larbalestier 2003). Furthermore, clinically significant AZT resistance, which can take months to develop, has not been observed with shorter courses of AZT (Ekpini 2002).

3TC resistance has been observed in two small studies following its use in short treatment courses. In ANRS 075, for example, 39% of the women developed the M184V resistance mutation. This was more common in women with higher viral loads and lower CD4 cell counts (Mandelbrot 2001). Another study of 19 HIV-infected pregnant women in London that compared AZT monotherapy with the combination of AZT and 3TC found that four of five women on the dual combination rapidly developed resistance to 3TC (Clarke 1999). A study of 32 women given AZT and 3TC reported no resistance mutations in any of the mothers, but NRTI resistance in one of the three infected infants (Chokephaibulkit 2005). As for AZT, the clinical significance of the development of 3TC resistance mutations may be slight as it may dissipate after treatment has stopped.

In contrast, resistance to nevirapine seems to be more serious. Nevirapine is a cornerstone of antiretroviral therapy, particularly in the developing world, and resistance to the drug usually confers cross-resistance to efavirenz (Sustiva), thus eliminating most first-line treatment options.

Initial reports suggested that resistance occurred in 20 to 30% of women exposed to single-dose nevirapine (Eshleman 2001b; Jourdain 2004). However, recent studies using more sensitive techniques to screen for resistance suggest that it is much more common (Flys 2005; Johnson 2005; Kantor 2003; Loubser 2005; Palmer 2005). These methods have also revealed thatv these low-level mutations can persist for up to a year (Eshleman 2005c). The methods used in some of these studies can detect very small minority populations of drug-resistant virus, which appear to occur in as many as 60 to 75% of the women who take it, depending on the study and perhaps the viral sub-type. Some research suggests that nevirapine resistance is commonest in women with HIV-1 subtype C, followed by subtype D and least common in subtype A (Eshleman 2005b).

As long as the resistant virus is in circulation, there is a risk that it can be transmitted to breastfeeding infants. At least one study has found that breastmilk can be a reservoir for nevirapine-resistant virus (Lee 2003).

However, there is evidence that this resistance diminishes over time. In a South African study, which estimated the true rate of resistance to be around 65%, nevirapine resistance was seen one year after delivery in only 25% of the women. Encouragingly, though, HIV DNA in the womens white blood cells did not show any evidence of nevirapine resistance, despite use of the more sensitive laboratory test. This suggests that archiving of the nevirapine resistance mutations may be a rare event and that re-emergence of resistant virus may occur infrequently (Johnson 2005). A substudy of HIVNET 012 showed that fading of nevirapine resistance may occur more rapidly in HIV sub-type D than in sub-type A (Eshleman 2005a).

It is not clear whether the development of low frequency nevirapine resistance mutations will have a clinical impact. Notably, one study has shown that it has little impact on the effectiveness of subsequent use of single dose nevirapine in a second pregnancy (Martinson 2005). However, the effects on the mothers' future treatment options are more worrying. For example, prior exposure to nevirapine did compromise some mother's subsequent response to non-nucleoside reverse transcriptase inhibitor (NNRTI)-based antiretroviral therapy in a sub-study of the PHPT-2 trial, with fewer women who had received single-dose nevirapine achieving undetectable viral loads after starting NNRTI-based therapy around six months after giving birth (Jourdain 2004).

Virologic responses might not be so impaired in women who wait longer before starting antiretroviral therapy, but no clinical trials have adequately addressed this. In one study from Ivory Coast, a median of 17 months passed between when women exposed to single-dose nevirapine and when they began antiretroviral therapy. Although there are no virologic data from this study, the investigators could find no difference in CD4 cell count responses between the those who were exposed to nevirapine and those who were not up to twelve months after starting treatment. However, many of the participants also took AZT and 3TC for a few days after childbirth (Bedikou 2005).