Factors affecting HIV therapy choice in women have changed

The first decade of effective anti-HIV therapy has seen the main predictors of regimen choice among HIV-positive women switch from biological to sociodemographic factors, according to a report in the January 15^th edition of Clinical Infectious Diseases. The results come from a wide-ranging study that investigated the patterns of antiretroviral therapy, from initial regimen choice to predictors of switching, among a cohort of HIV-positive women in the US.

While knowledge of the specific issues faced by women with HIV grows, there is a paucity of data concerning antiretroviral therapy in this group. To explore how anti-HIV therapy might be different in women, Golub and colleagues at several American universities tapped into the extensive records of the Women’s Interagency HIV Study (WIHS) to determine patterns of initial regimen choice and switches in regimens, as well as to identify their predictors.

WIHS, established in 1993, is a prospective, multicentre, cohort study of over 2700 HIV-positive women in the US. Researchers study data for almost ten years, which they then divided into three periods: 1996 to 2000, 2000 to 2002 and 2002 to 2005.

The records of 1555 women starting antiretroviral therapy during these periods were studied for initial regimen type (protease inhibitor [PI] based, non-nucleoside reverse transcriptase inhibitor [NNRTI] based or triple nucleoside reverse transcriptase inhibitor [NRTI]), any switches from their initial program and immunological follow-up at years one and two. Switches in regimen were recorded only when an entire drug class was switched, not a drug from one class was switched for a drug of the same class.

Patterns of initial regimen choice

At study start, in July 1996, almost all women began treatment with a PI-based regimen. By March 2000, less than 40% of women started a PI regimen, with NNRTI-based regimens being a more popular choice and triple-NRTI regimens comprising less than 10% of new regimens.

In the second period, from 2000 to 2002, there were no significant changes in the type of regimen started.

In the final period, there was a sharp, but brief increase in the proportion of women starting triple-NRTI regimens. This proportion quickly dwindled so that by the end of the study period, March 2005, women starting antiretroviral therapy were almost equally divided between the PI and NNRTI groups.

Predictors of initial regimen choice

The researchers observed that predictors of regimen type also changed during these periods. From 1996 to 2000, women with a CD4 cell count above 200 cells/mm³ were more than twice as likely to start an NNRTI regimen than a PI regimen compared with women with a lower CD4 cell count. As well, injection drug use was associated with a lower chance of starting an NNRTI regimen.

From 2000 to 2002, biological predictors continued to predominate, with starting an NNRTI regimen associated with lower CD4 cells counts (under 500 cells/mm³) and higher peak viral loads compared with PI regimen starters. Preference for a Spanish-language interview was also associated with NNRTI regimen choice.

In the final period, from 2002 to 2005, sociodemographic factors predominated in choice of initial regimen. Of note, Black and Latina women were more likely to initiate a PI regimen than an NNRTI regimen. They were also more likely to initiate a triple-NRTI regimen than an NNRTI regimen.

Patterns in regimen switching

The researchers included 301 women in their analysis of regimen switching. About half of the group, 157 women, started with an NNRTI regimen, while 93 started a PI regimen. The cumulative hazard of switching by 12 months was 50%, and was 62% by 24 months. Compared with those starting a PI regimen, women who started an NNRTI regimen were slightly, but not significantly less likely to switch, while women starting a triple-NRTI regimen were significantly more likely to switch.

Women who switched were more likely to have started with an NNRTI regimen, be African American and to have already been on antiretroviral therapy. Many expected factors were independently associated with regimen switching, including clinical AIDS after starting treatment, already having taken antiretrovirals, higher viral load, frequent recent visits with a physician and history of seeking mental health treatment.

Impact on immunological response

When researchers evaluated changes in CD4 cell counts, they performed analyses incorporating the identified predictors. In women starting a triple-NRTI regimen had on average a CD4 cell count 90 cells lower after a year of treatment than women starting a PI regimen, after adjusting for potentially confounding factors. There was no different between women starting PI and NNRTI regimens. However at two years, women starting an NNRTI regimen also had a CD4 cell count lower than PI-regimen women, on average 54 cells/mm³ lower.

In discussing their findings, the researchers note that regimen choice and predictors have changed over time. They argue that including these predictors in the analysis of CD4 cell count data allowed them to identify a significant association between initial regimen choice and subsequent immunological response.

“It is clear,” they write, “that a multitude of factors affect a patient’s initial regimen, and we conclude that omitting those from investigations into the effectiveness of therapy has the potential to introduce residual confounding. In this cohort of US women, we were able to avoid such confounding by identifying, and subsequently controlling for, predictors of initial HAART regimen.” In conclusion, they urge future groups to include predictors in the design of future studies assessing the efficacy of HIV therapies in observational settings such as WIHS.