Introduction

Prostate cancer (PCa) is the second most common cancer in men and shows significant disproportionally in racial prevalence [1]. In the USA, black men have a PCa incidence of 163.8 per 100,000 men, compared to 96.7 and 52.0 for white and American Indians/Alaskan natives, respectively [2]. Age-adjusted death rates are also highest in black men at 36.4 per 100,000 men, more than double that seen in white men (17.8) [2]. Causes for the racial disparity are likely multifactorial, including environmental, socioeconomic, and the underrepresentation of racial minorities in clinical trials.

Current US Food and Drug Administration (FDA) recommendations for clinical trials, introduced in 2016 recommend race reporting with a minimum of five categories; White, American Indian/Alaskan Native, Asian, Black/African American, and White Hawaiian/Pacific Islander [3]. Implementation and subsequent reporting in licensing publications of this recommendation are unknown. Furthermore, for licensing trials for which the USA is the lead participant contributor, it is unknown if these trials are representative of the PCa population. We analyzed racial representation in PCa registration trials for a period of 15 years (10 years before and 5 years after the introduction of FDA guidance) to analyze compliance with current FDA race reporting recommendations.

Methods

A retrospective review of all new molecular entities and subsequent marketing authorizations of PCa drugs from January 2006 to July 2020 was conducted using the FDA website. Clinical trials cited on the drug licensing label for market authorization were identified using the national clinical trial identifier (NCT). The corresponding licensing trial publication was identified through clinicaltrials.gov or PubMed.

We determined whether race was reported in the corresponding licensing publication, including supplementary appendices and compliance with FDA guidance. If race was unreported or only partially reported (defined ≤3 categories), then the study report on clinicaltrials.gov or FDA website was analyzed. Additional information on participant race and recruitment by lead country was obtained to assess proportional representation based on disease population.

Results

We identified 17 new drug registrations (with corresponding licensing publication) for the management of PCa involving ten unique drugs, including degarelix (one new license), cabazitaxel (two new licenses), denosumab (one new license), abiraterone (three new licenses), enzalutamide (four new licenses), radium-223 (one new license), apalutamide (two new licenses), darolutamide (one new license), rucaparib (one new license) and olaparib (one new license). Table 1 shows further information about licensing indication, clinical trial information, race reporting, and overall racial demographics.

Table 1 FDA prostate cancer licensing trials 2006–2020.
Full size table

The race was reported in 9 (52.9%) licensing publications. However, 4 (23.5%) provided limited information (e.g., only reporting frequency of Caucasian participants). Two of these trials had further information in the trial report and two had no further data available. For 8 (47.1%) licensing publications where no race information was reported, seven had further information within the trial report. Precise subgroup analysis by race was performed in only 2 (11.7%) studies, however, a further 9 (52.9%) studies did analyze trial participants by recruitment site continent from which some race data could be extrapolated.

Of the 18,455Footnote 1 participants included in PCa licensing trials, 14,106 (76.3%) were white or Caucasian, 1454 (7.9%) Asian, 528 (2.9%) Black or African American, 88 (0.5%) American Indian or Alaskan Native, 12 (0.1%) Native Hawaiian or other Pacific Islander, 332 (1.8%) other or multiple races and 1949 (10.5%) unknown. Recruitment by country was reported in 11 (64.7%) out of 17 trials. The USA was the leading participant recruiter in 7 (41.1%) trials, which involved 8562 participants. Of which 6998 (81.7%) were white or Caucasian, 641 (7.5%) Asian, 291 (3.4%) Black or African American, 103Footnote 2 (1.2%) other and 529 (6.2%) unknown.

Discussion and conclusion

Guidelines from the International Committee of Medical Journal Editors (ICMJE) recommend, that because the relevance of race is not always known authors should, at a minimum, provide descriptive data. This study found that race reporting in FDA licensing publications does not meet the ICMJE guidance and is reported in only 53% of licensing publicationsFootnote 3. This echoes the overall poor reporting rates observed in other phases 3 PCa clinical trials [4].

Following the introduction of new FDA guidance for race reporting in 2016, the reporting rate is relatively unchanged with 55% of studies meeting requirements. Reporting is also not uniform and harmonization of recording, following FDA guidance would greatly improve population analysis. Furthermore, this is consistent with a recent study by Rencsok et al which found 29 different race or ethnicity categories utilized in 72 PCa prevention, screening, and treatment trials from 1987 to 2016 [5].

A review by the FDA of new drug approvals between 2008 and 2013 found one-fifth of drugs demonstrate differences in exposure and/or response across racial/ethnic groups [3, 6]. Current guidance expects sponsors of clinical trials to enroll participants who reflect the demographics of the clinically relevant populations. This study concurs with the previous findings of significant under-enrollment of non-white participants in PCa trials [4, 5, 7]. Furthermore, in trials in which the USA was the leading participant recruiter, trial populations were not representative of the PCa population. Only one licensing publication acknowledged the underrepresentation of racial groups as a limitation, suggesting that racial disparity in PCa trials needs greater recognition [8]. Overall, there is a preponderance of white participants in PCa trials despite the known racial disproportionality of the disease burden. The recruitment of trial participants reflective of the burden of disease must be a research priority.