Sex, age, race and intervention type in clinical studies of HIV cure: a systematic review

Abstract

This systematic review was undertaken to determine the extent to which adult subjects representing sex (female), race (nonwhite), and age (>50 years) categories are included in clinical studies of HIV curative interventions and thus, by extension, the potential for data to be analyzed that may shed light on the influence of such demographic variables on safety and/or efficacy. English-language publications retrieved from PubMed and from references of retrieved papers describing clinical studies of curative interventions were read and demographic, recruitment year, and intervention-type details were noted. Variables of interest included participation by sex, age, and race; changes in participation rates by recruitment year; and differences in participation by intervention type. Of 151 publications, 23% reported full demographic data of study enrollees, and only 6% reported conducting efficacy analyses by demographic variables. Included studies recruited participants from 1991 to 2011. No study conducted safety analyses by demographic variables. The representation of women, older people, and nonwhites did not reflect national or international burdens of HIV infection. Participation of demographic subgroups differed by intervention type and study location. Rates of participation of demographic groups of interest did not vary with time. Limited data suggest efficacy, particularly of early therapy initiation followed by treatment interruption, may vary by demographic variables, in this case sex. More data are needed to determine associations between demographic characteristics and safety/efficacy of curative interventions. Studies should be powered to conduct such analyses and cure-relevant measures should be standardized.

FIG. 1.

Mean number of participants in each type of intervention. Error bars indicate±1 standard error.

FIG. 2.

Number of studies of each intervention type by epoch. The recruitment year variable was divided into four epochs: pre-combination antiretroviral therapy (cART) (up to and including 1995), triple cART (1996–2002), multi-cART (2003–2008), and post-Berlin patient (2009–present). The most recent epoch was not included in this analysis as there were too few studies included in this review. *p<0.05, **p<0.007.

FIG. 3.

Participation of women and nonwhites in studies conducted entirely within or outside the United States. ⁺p<0.08, **p<0.001.

FIG. 4.

Sex differences in participation by intervention. One-sample binomial z-tests were conducted using the proportion of total subjects per intervention as the expected value (black bars) and proportion of women as the observed value (gray bars). Proportions are represented as percentages. **p<0.007.

FIG. 5.

(A) Mean percentage of women recruited into each type of intervention. Error bars indicate±1 standard error. Lines indicate significant (at p<0.05) differences between interventions from pairwise t-tests. (B) p-values of pairwise t-tests showing differences between interventions in the mean percentage of women participants. *p<0.05, **p<0.007.

FIG. 6.

Race differences in participation by intervention. One-sample binomial z-tests were conducted using the proportion of total subjects per intervention as the expected value (black bars) and proportion of nonwhites as the observed value (gray bars). Proportions are represented as percentages. **p<0.007.

FIG. 7.

Mean percentage of nonwhites recruited into each type of intervention. Error bars indicate±1 standard error. No nonwhites were reported as participants in reactivation interventions. Pairwise t-tests revealed no significant differences between interventions.