Contemporary Nationwide Patterns of Active Surveillance Use for Prostate Cancer

In 2011, the US Preventative Services Task Force discouraged routine prostate cancer screening. Despite an initial decline in prostate-specific antigen (PSA) screening, clinical practice has not changed appreciably.¹ Supporters of PSA screening argue that the potential risks of overdiagnosis are mitigated by active surveillance.² Active surveillance—careful monitoring of prostate cancer with selective intervention based on apparent progression—is an option for localized low-risk prostate cancer.³ While no consensus on active surveillance eligibility exists, it is usually reserved for patients with low-risk disease.⁴ We sought to examine active surveillance use and its predictors in American men with low-risk prostate cancer using the National Cancer Data Base, a comprehensive facility-based cancer registry capturing 70% of incident cancer diagnoses in the United States.

We used 2010-2011 data from the US National Cancer Data Base, the most recent data available that was also unique for its inclusion of an active-surveillance-specific identifier. Institutional review board exemption was granted by University Hospitals Case Medical Center, where the data analysis was conducted. We identified men with biopsy-proven clinical N0/M0 prostate cancer during this period. Low-risk disease was defined by the following known criteria: (1) modified Epstein (clinical stage, ≤T1c; Gleason score, ≤6; PSA level, <10; and ≤2 [or <33%] positive biopsy cores); (2) D’Amico (clinical stage, ≤T2a; Gleason score, ≤6; PSA level, <10 ng/mL); and (3) Klotz (clinical stage, ≤T2a; Gleason score, ≤6 [or ≤3 + 4 = 7 for men ≥70 y]; PSA level, <10 ng/mL [or <15 ng/mL for men ≥70 y]). In men who met the modified Epstein criteria, we used multivariate logistic regression to determine the likelihood of undergoing active surveillance, accounting for diagnosis year, race, residential area, income, education, insurance, age, Charlson Comorbidity Index score, hospital type, prostate cancer volume, and geography. Statistical tests were conducted using SAS, version 9.1 (SAS Institute Inc). P <.01 was considered statistically significant.

Of 189 768 patients with prostate cancer, 75 546 (39.8%), 54 070 (28.5%), and 20 377 (10.7%) were determined to be eligible for active surveillance by the Klotz, D’Amico, and modified Epstein criteria, respectively. In practice, 6.5%, 7.4%, and 12.1% of these men received active surveillance (Figure 1). Increasing age was most strongly associated with active surveillance use (60-64 y vs <50 y: OR, 1.77; 95% CI, 1.35-2.32; 65-69 y vs <50 y: OR, 2.38; 95% CI, 1.80-3.16; >70 y vs <50 y: OR, 3.83; 95% CI, 2.88-5.11). Being uninsured or treated in the Northeast also predicted its use (OR, 3.26; 95% CI, 2.33-4.54 and OR, 2.16; 95% CI, 1.90-2.46, respectively) (Figure 2). Additional positive predictors of active surveillance use included diagnosis in 2011 (OR, 1.45; 95% CI, 1.32-1.59) and treatment at an academic hospital (OR, 1.84; 95% CI, 1.44-2.35) or a hospital with a high volume of patients with prostate cancer (OR, 1.33; 95% CI, 1.16-1.54). Being healthy, African-American, or more educated were significantly but less strongly associated with its use (OR, 1.56; 95% CI, 1.33-1.82; OR, 1.23; 95% CI, 1.05-1.43 and OR, 1.32; 95% CI, 1.07-1.62, respectively).

Only 12.1% of men (2466 of 20 377) with very-low-risk prostate cancer received active surveillance. Expanding on prior evidence that observation is used infrequently in men with low-risk disease,⁵ this study establishes that active surveillance use is low. Usage increases as the inclusion criteria for surveillance become more stringent, ie, less likely to miss significant disease. While active surveillance is aptly applied to elderly men, its use is sporadic, confined to academic and regional hospitals, and strongly influenced by nonclinical factors, including the patient’s insurance provider. Patient preference may influence use, especially in certain demographic groups. Despite ongoing adoption, use of active surveillance must increase substantially to effectively reduce the overtreatment of screening-detected prostate cancer.⁶

This study has several limitations. Selection bias related to the National Cancer Data Base’s hospital-based data set may cause potential underrepresentation of active surveillance use in the outpatient setting. Because the data set is somewhat dated, it may not accurately reflect recent urological patterns. Nonetheless, this study represents, to our knowledge, the most up-to-date analysis of active surveillance trends, and its predictors, in a large nationally diverse cohort. Uniquely, our study is generalizable to men of all ages, including younger men who may benefit more in the long term with active surveillance. Last, the treatment-specific identifier that we used minimizes misclassification bias.

Corresponding Author: Hui Zhu, MD, ScD, Louis Stokes Cleveland Veterans Affairs Medical Center, 10701 E Blvd, Cleveland, OH 44106 (hui.zhu@va.gov).

Published Online: June 29, 2015. doi:10.1001/jamainternmed.2015.2835.

Author Contributions: Dr Zhu had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Maurice and Abouassaly contributed equally to this work.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Maurice, Abouassaly, Zhu.

Drafting of the manuscript: Maurice, Abouassaly, Zhu.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Maurice, Abouassaly.

Administrative, technical, or material support: Abouassaly, Zhu.

Study supervision: Abouassaly, Kim, Zhu.

Conflict of Interest Disclosures: None reported.

Disclaimer: The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methods used or the conclusions drawn from these data by the investigators.