Low rate of sex-specific analyses in presentations at the Conference of Retroviruses and Opportunistic Infections (CROI) meeting, 2018: Room to improve

INTRODUCTION

Sex differences in HIV acquisition, pathogenesis, reservoir, treatment response, antiretroviral pharmacokinetics and toxicities exist¹ and have been elucidated when studied. However, despite over 50% of the world’s HIV infections occurring among women,² sex differences in HIV treatment and prevention are still unevenly studied. The National Institutes of Health (NIH) “expects that sex as a biological variable will be factored into research designs, analyses, and reporting”, stressing that sex-based data should not only be collected, but analyzed and reported for every study, including those which are pre-clinical.^3,4 However, adequate sex representation is not always achieved^5,6 and, even then, sex delineated outcomes are not always reported.⁷ Moreover, although there has been an increase in the recruitment of women into clinical trials of new drugs over the past two decades,⁸ there is not always adequate representation of women to perform sex-specific reporting for novel therapeutics.⁹ Finally, there is increasing importance placed on distinguishing biological sex from gender, which is self-identified and rooted in social, environmental and cultural factors.¹⁰

Since immunologic differences between men and women are prominent,^11,12 the field of HIV is one in which sex differences may be particularly relevant. The field has made stuttering progress in sex-specific reporting for HIV studies, with important advances in some areas, such as pharmacokinetics,^13,14 while crucial gaps remain in areas such as co-morbidities and cure research.^11,15–17 Regulatory guidelines or mandates can ensure better compliance with policies to collect, analyze and report sex-specific results. Towards that end, the program committee of one of the major scientific conferences on HIV, the Conference on Retroviruses and Opportunistic Infections (CROI), requested investigators submitting abstracts to the 2018 meeting to consider sex as a biological variable. Upon abstract acceptance, primary presenters of all oral abstracts were provided with specific instructions to include sex-stratified analyses when feasible. This paper seeks to examine what proportion of the oral abstracts presented at CROI 2018 showed information on sex distribution when relevant and presented sex-specific analyses.

METHODS:

CROI 2018 was held in Boston, Massachusetts from March 4–7, 2018. Instructions provided to presenters of oral abstracts from the program committee read as follows: “Please consider whether there is substantive evidence of differences in effect by sex or other key demographic groups. If so, a stratified analysis should be made available during the presentation of the abstract at CROI.” Each oral presentation at CROI 2018 was video-taped with the talk and slides made available as a webcast. The Women’s Health Inter-Network Scientific Committee (WHISC) is a working group of two NIH-funded clinical trials networks, the AIDS Clinical Trials Group (ACTG) and International Maternal Pediatric Adolescent AIDS Clinical Trials Network (IMPAACT), focused on HIV among girls and women. For this analysis, members of the WHISC independently reviewed webcasts of CROI 2018 oral scientific sessions with themes relevant to both sexes. In order to optimize data collection, at least two WHISC members reviewed each of the selected oral sessions; any discordant results were resolved by the first, second and last author on this paper via discussion.

Reviewers initially evaluated the following for each oral presentation: 1) whether the abstract’s scientific question/objective was relevant to both sexes (e.g. whether only men-who-have-sex-with-men (MSM) or women were considered), and 2) whether the study included human participants, animals, or was pre-clinical, but still included specimens from humans or animals. If both criteria were met, reviewers then assessed 1) whether the reported study demographics included sex, and 2) if sex-delineated outcomes or sex-stratified analyses were presented. lf results by sex were not presented, reviewers recorded whether an explanation for this omission was provided. Descriptive statistics summarized results across presentations.

RESULTS

Of 83 original oral abstracts presented at CROI 2018, 16 (19%) were deemed relevant to one sex only and were excluded from this analysis (Table 1). Of the remaining 67 oral abstracts relevant to both sexes, 35 (52%) included the distribution of the study sample by sex; 7 (10%) presented sex distributions, albeit mislabeled as “gender”; and 25 (37%) did not present or address sex distributions. Only one abstract among the basic science and animal studies included sex distribution (1/13, 8%). Sex distribution was reported in human observational studies and clinical trials the majority of the time (41/54, 76%).

TABLE 1:

Proportion of abstracts at CROI 2018 presenting sex-specific results

Type of abstract	Number of total (%)	Presented sex distribution	Presented sex-specific results
Relevant to one sex only	16 of 83 (19%)	NA	NA
Relevant to both sexes	67 of 83 (81%)	42 of 67 (63%)	16 of 67 (24%)*
Human studies	54 of 67 (81%)	41 of 54 (76%)	16 of 54(30%)
Animal or cell-based studies	13 of 67 (19%)	1 of 13 (8%)	0 of 13 (0%)

^*Of 51 which did not present sex-specific results, only 8 provided an explanation of why this wasn’t done

Only 16 (24%) of all oral abstracts presented at CROI 2018 relevant to both sexes included sex-stratified analyses or sex-delineated outcomes (Table 1). The remaining 51 (76%) did not and only 8 (12%) provided an explanation for why sex stratification was not presented. Of the 28 presentations summarizing the results of clinical trials, 25 (89%) included sex distribution, but only 6 (21%) presented results by sex.

DISCUSSION:

Despite CROI 2018 providing presentation guidelines consistent with U.S. federal expectations on reporting by sex, more than a third of oral presentations at this conference failed to report sex demographics and only a quarter included sex-stratified analyses. As 52% of adults living with HIV worldwide are women,² determining whether responses and treatment paradigms differ by sex is important. Sex-specific analyses can provide insight into whether HIV prevention, treatment, cure and the management of HIV-associated conditions should be tailored by sex. This analysis of sex-specific reporting among the highest profile presentations at one of the field’s major scientific conference highlights missed opportunities to delineate outcomes by sex and contribute to the knowledge base of sex differences in HIV infection.

Of note, pre-clinical studies at CROI 2018 only rarely reported the sex distribution of the cells or animals used in the study. However, the Institute of Medicine’s landmark report Does Sex Matter? stresses that “Sex.. is an important basic human variable that should be considered when designing and analyzing studies in all areas and at all levels of biomedical and health-related research.”¹⁸ At the cellular level, genetic differences by sex even in somatic cells may have an influence on HIV responses, along with the epigenetic effects of the hormonal milieu. Moreover, sex is not the same as gender.¹⁰ In the CROI 2018 presentations relevant to both sexes, 10% of them mislabeled sex as gender when reporting distributions, whereas, ideally and when possible, both sex and gender should be reported. Starting in 2018, the AIDS Clinical Trials Group (ACTG) updated the uniform demographic information collected to include both gender identity and sex at birth for participants in all network trials. This addition was championed by the National Institute of Allergy and Infectious Diseases (NIAID) Cross Network Transgender Working Group.

In conclusion, our analysis showed that only a quarter of oral scientific presentations at CROI 2018 provided sex-specific analyses, highlighting missed opportunities to contribute to the knowledge base on sex differences in HIV infection. The CROI program committee has aligned its recommendations to abstract presenters with NIH, Institute of Medicine and Food and Drug Administration (FDA) policies, but guidance on considering sex as a biological variable at this and other major HIV-related scientific conferences may need to be accompanied by evaluation and monitoring strategies to ensure compliance. Moreover, investigators should implement planning for considering sex as a biological variable early in study design so that adequate data are available upon study completion for sex-specific reporting. Further education of researchers on the expectations regarding sex-specific reporting is essential to maximize knowledge about sex differences and similarities in HIV and its associated conditions.