Persistent deficiencies in the measurement and reporting of tobacco use in contemporary genitourinary oncology clinical trials

Smoking is a causal or contributory risk factor for the development of genitourinary cancers [12] . Smoking also exerts continued influence on treatment, recurrence, toxicity, and quality of life outcomes beyond diagnosis 3456 . Although awareness and support to help patients diagnosed with cancer quit smoking has grown recently 789 , there remains a critical deficiency in how smoking-related data are collected and reported in oncology clinical trials [1011] . Since patients often quit or relapse after diagnosis [12] , changes in smoking exposure over time can also result in differential recurrence patterns, altered sensitivity to systemic agents, and changes in overall quality of life among survivors 3456 .

We hypothesized that despite the strong association between smoking and genitourinary cancers, data on participant smoking status are infrequently collected and reported in the publication of contemporary therapeutic genitourinary cancer trials. Therefore, we conducted a systematic review of recent randomized trials to better evaluate this potential gap in knowledge and to call attention to this persistent deficiency in trial design. Improving our understanding of the complex relationship between smoking and therapeutic outcomes will only be possible when tobacco use data are collected and available for study. Identifying deficiencies in clinical trial design will help promote improvements in standardized and systematic data collection and reporting.

To create a manageable and contemporary subset of randomized clinical trials (RCTs) in urologic oncology, we first identified journals with the highest impact in the fields of urology, oncology and general medicine. In May 2020, an experienced Medical Librarian (TR) retrieved the top ten journals ranked by 2018 Impact Factor in each of those three specialties using the most recent Journal Citation Reports (Clarivate Analytics, 2020). This yielded 30 unique indexed journal titles (supplement). Using the Ovid platform, the MEDLINE database was queried for any articles that discussed prostate, urothelial (bladder and upper tract urothelial), or kidney cancer. The search was not limited by language but was limited to articles published from May 2010 to May 2020 and to RCTs. The complete Ovid Medline search is available in the supplement. The titles/abstracts then full text manuscripts were iteratively screened by two independent reviewers (CZ, AM). Studies were included if they studied a therapeutic intervention (i.e. not diagnostic or screening), reported planned analysis of primary data, and were randomized Phase II-IV trials. Phase I or pilot studies, those that reported results of unplanned secondary analyses, and trials of diagnostic/screening interventions were excluded. Data were extracted using a standardized form. Conflicting votes or data were resolved by a third reviewer (RM) with urologic oncology content expertise. Quality of evidence was assessed according to the Oxford Center for Evidence Based Medicine rating scheme to ascertain potential bias and report study quality systematically. This review was conducted and reported according to PRISMA guidelines [13] . Our primary outcome of interest was whether a study collected and reported data on participant smoking status. Secondary outcomes included details about how smoking status was measured, whether the trial arms were balanced with respect to smoking status, and if smoking status was included in data analysis. Additional information collected included clinical trial characteristics such as cancer type, intervention type, outcomes reported, study duration, and number of participants. Data are reported using descriptive statistics.

Our initial search criteria yielded 622 articles. After screening, 354 met criteria and were included in our analysis ( Fig. 1, supplement ). All studies were randomized trials and were of high (Rating 1 or 2) quality according to the Oxford rating scheme. The majority of publications were published in The Journal of Clinical Oncology ( n = 81), European Urology ( n = 73), and The Lancet, Oncology ( n = 60). Prostate cancer trials (59.3%) were the most frequently reported, followed by urothelial (22.3%) and kidney (17.5%). Systemic therapies (46.9%) were the most common interventions and most publications reported outcomes from Phase III studies (69.8%) of patients with either localized (33.9%) or metastatic (33.9%) disease ( Table 1 ). Overall survival (28.8%) and progression-free survival (27.1%) were the most common study primary outcomes. Median study duration was 37 months (IQR 22–61 months) and median number of study participants was 369.5 (IQR 131–917 patients).