Nonnucleoside Reverse Transcriptase Inhibitor Resistance Among Antiretroviral-Naive HIV-Positive Pregnant Women

From 1999 to 2001, the overall prevalence of resistance in the antiretroviral (ART)-naive population in St. Louis, Missouri, was 17%. We sought to determine if resistance testing in ART-naive HIV-positive pregnant women identified resistant mutations, which would modify our initial choice of therapy. A retrospective chart review was performed on all HIV-positive pregnant women seen from January 2000 to December 2001 at a university hospital. There were 72 pregnancies. Twenty-seven of 72 patients were ART naive. Genotype testing was performed in 18 of 27 naive patients. Three of 18 ART-naive patients (17%) had primary resistance (95% CI: 4%-41%) by genotype to NNRTIs. The primary mutation, G190S, conferring resistance to NNRTIs was present in 1 patient. Another had the K103N mutation. One had the K103R mutation, which conferred phenotypic resistance to NNRTIs by 8.3-fold, warranting a change in the initial regimen. In our community, resistance testing in ART-naive pregnant patients is warranted. Switching later to a more complex regimen during pregnancy may adversely affect adherence, resulting in virologic failure. Strategies to avoid prescribing a suboptimal regimen include waiting to initiate ART until the resistance testing results are available and/or beginning ART with a protease inhibitor-based regimen if the patient is already in the third trimester of pregnancy at the time of her initial clinic presentation.

The International AIDS Society-USA Panel (IAS-USA) recommends resistance testing of all HIV-positive pregnant women to help guide the choice of antiretroviral regimens and recommends consideration of resistance testing in antiretroviral-naive patients (¹). The US Public Health Service Task Force recommends resistance testing only in acute HIV infection, virologic failure or suboptimal viral suppression by antiretroviral therapy (ART), or high likelihood of exposure to resistant virus based on community prevalence or source characteristics (²).

In patients who are antiretroviral naive, the prevalence of resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs) is highest among the three classes of antiretrovirals (^3,4). Multiple studies have recently been conducted evaluating the prevalence of resistance in antiretroviral-naive patients in the United States. In a study of antiretroviral-naive patients sampled from six states in the United States (California, Florida, Georgia, Massachusetts, Minnesota, and Pennsylvania), genotypes had mutations associated with decreased susceptibility to nucleoside reverse transcriptase inhibitors (NRTIs) in 14% of patients and to NNRTIs in 16% of patients, and there were primary mutations conferring resistance to protease inhibitors (PIs) in 3% of patients (⁵). The Resistance Study conducted in New York found that 8.8% of treatment-naive patients had high-level resistance mutations (⁶). In Houston, Texas, 11% of naive patients had high-level resistance mutations in the protease or reverse transcriptase genes (⁷).

From 1999 to 2001, the overall prevalence of individuals with resistant virus in the drug-naive population in St. Louis, Missouri, was 17%, of which a third of the patients had resistance to NNRTIs (⁸). Given that nevirapine (NVP)-containing (Viramune; Roxane Laboratories/Boehringer Ingelheim Pharmaceuticals, Ridgefield, CT, U.S.A.) regimens have become standard of care for the pregnant HIV population in St. Louis, we wanted to know if resistance testing in antiretroviral-naive HIV-positive pregnant women identified resistant mutations that would modify our initial choice of therapy.

METHODS

Study Population

A retrospective chart review was performed on all HIV-positive pregnant women seen from January 2000 to December 2001 at a university hospital. Approval was obtained from the Washington University Human Studies Committee to conduct this study. Data collected included demographics, CD4⁺ T-cell count, HIV viral load, HIV genotype and phenotype results, ART history, source contact if known, and pregnancy outcome.

Sensitivity Testing

All sensitivity testing was ordered at the discretion of the health care provider in the infectious disease clinic or high-risk obstetrics clinic. Plasma HIV RNA was determined using the Amplicor HIV-1 Monitor assay (Roche Diagnostics Systems, Branchburg, NJ, U.S.A.). To obtain genotypes, HIV isolates were sequenced by fluorescent dideoxy chain termination (Perkin-Elmer, Wellesley, MA, U.S.A. or Applied Biosystems, ABI, Foster City, CA, U.S.A.). Sequencing results were reported as amino acid changes at positions in HIV protease and reverse transcriptase genes. Phenotypes were obtained using Phenosense (Virologic, San Francisco, CA, U.S.A.). In this analysis, nucleic acid from the patient's reverse transcriptase and PI genes is amplified in a vector viral genome, and the IC₅₀ of the patient's virus is compared with the IC₅₀ of a drug-sensitive similarly processed reference virus control to calculate fold change in susceptibility. The change is calculated as the ratio of the patient IC₅₀ to the drug-sensitive reference IC₅₀.

The patients' genotype results were compared with reported mutations published in the Drug Resistance Mutations in HIV-1, updated in June 2002 by the IAS-USA (⁹). We included all patients' mutations considered to be primary as well as the variant K103R mutation. This amino acid substitution has not been associated with NNRTI resistance; however, because of the strong association of the K103N mutation with NNRTI resistance, the K103R strain was tested for phenotypic susceptibility to NVP. Substantial decrease in susceptibility by phenotype was defined as a ≥10-fold change in the IC₅₀ value compared with the IC₅₀ value for the reference wild-type virus.

Statistical Analysis

The collected data were analyzed using the SPSS software package (SPSS, Chicago, IL, U.S.A), and statistical analysis was performed using SAS (SAS Institute, Cary, NC, U.S.A.) to determine the prevalence of clinically significant resistance among antiretroviral-naive pregnant HIV-positive women.

RESULTS

There were a total of 72 pregnancies during the study period. Of the 72 patients, 27 were ART naive. Characteristics of the patients are shown in Table 1. Twenty-four of the ART-naive patients carried to term. Genotypes were performed on 38 of 72 total patients and 18 of 27 antiretroviral-naive patients at the discretion of the health care provider. Thirty-four of the 72 did not have a genotype performed, and 6 of 34 had insufficient plasma viral loads for resistance testing.

Table 1

Of the 18 ART-naive pregnant women who had genotypes performed, none had primary mutations associated with resistance to PIs or NRTIs. Three (17%) had primary resistance (95% CI: 4%-41%) by genotype to NNRTIs. Of these patients, 1 had the primary mutation G190S, conferring resistance to NNRTIs. Another had the K103N mutation. The third had the K103R mutation with phenotypic resistance to NVP and delavirdine (DLV; Agouron Pharmaceuticals, San Diego, CA, U.S.A.) by 8.3-fold, warranting a change in the initial prescribed antiretroviral regimen. Viruses of the 18 ART-naive patients had secondary resistance mutations. A brief clinical summary of the 3 ART-naive pregnant women infected with HIV-resistant strains is presented in Table 2.

Table 2

DISCUSSION

The aim of this study was to evaluate the prevalence of clinically significant resistance among ART-naive pregnant HIV-positive women in St. Louis during the years 2000 and 2001. Resistance testing revealed primary resistance to NNRTIs in 17% of the ART-naive pregnant HIV-positive women. One had the primary mutation G190S, conferring resistance to NNRTIs. Another had the K103N mutation. One had the K103R mutation, which has not been thought to result in resistance to NNRTIs; however, phenotypic testing revealed 8.3-fold resistance to DLV and NVP. This is below the current consensus that ≥10-fold change in the IC₅₀ value results in a substantial decrease in susceptibility to a drug. The need to evaluate the clinical significance of phenotypic decreases in sensitivity despite the lack of recognized resistance mutations has been suggested in the past (^5,10,11). Some experts suggest that intermediate level (4- to 10-fold) phenotypic resistance to NNRTIs in the presence of a genotype suggesting sensitivity does not influence treatment outcome negatively (^10,12). Nevertheless, considering the potential consequences to the newborn of giving a pregnant HIV-positive woman presenting at 30 weeks of gestation a regimen that may be ineffective, clinicians decided to switch to a PI-based regimen. All 3 of the women had a decrease in HIV viral load on a suboptimal regimen, consisting primarily of dual therapy with NRTIs. Although this treatment can result in a drop in viral load, lack of prolonged benefit and rapid development of NRTI mutations are well documented, making this a poor treatment choice.

In a previously described study, the overall prevalence in St. Louis of resistance in the drug-naive male and female population was 17% in 1999 through 2001; a third of these cases involved resistance to NNRTIs (⁸). The results of this study in naive pregnant women support the impression that the prevalence of primary resistance in patients with HIV infection is increasing in this region. Causes for this increase, if real, are most likely multifactorial. NNRTI-based regimens are typically easy to take and are fairly well tolerated, making them attractive choices to health care providers and patients, resulting in wide use. Over 70% of persons infected with HIV in the United States are sexually active after they learn they are infected (¹³). Although many engage in safer sex practices, a considerable number of infected persons continue to engage in unprotected sexual behaviors that place others at risk for acquiring resistant strains of HIV and place themselves at risk for contracting secondary HIV infections and/or other sexually transmitted diseases. Other factors such as intravenous drug use may also contribute to the transmission of primary mutations to others.

In our community, resistance testing in antiretroviral-naive HIV-positive pregnant women is warranted. The importance of maximally suppressing viral replication in all pregnant women to reduce the risk of perinatal transmission is well documented (²). The presence of HIV primary resistance mutations in the patients reported here was associated with a high probability of not achieving an undetectable viral load at the time of delivery. Other factors could have been the presentation of the 3 patients at ≥30 weeks of gestation and the time spent waiting on genotype results while the patient continued to take a suboptimal regimen. Switching to a more complex PI-based regimen may have adversely affected adherence, resulting in virologic failure at the time of delivery. Although mother-to-child transmission rates are low in cohorts that have not used resistance testing, the risk of continued mutation accrual and the impact on future ARV regimen choices in this population must be taken into consideration, particularly in geographic areas that have a high prevalence of resistance among antiretroviral-naive persons. Strategies to avoid prescribing a suboptimal regimen include waiting to initiate ART until the resistance testing results are available and/or beginning ART with a PI-based regimen if the patient is already in the third trimester of pregnancy at time of her initial clinic presentation.

Acknowledgments:

The authors acknowledge Lisa Garrett and Debbie Gase, perinatal coordinators, for their assistance in data collection.