Discrepancies on the association between androgen deprivation therapy for prostate cancer and subsequent dementia: meta-analysis and meta-regression

Jae Heon Kim, Bora Lee, Deok Hyun Han, Kyoung Jin Chung, In Gab Jeong and Benjamin I. Chung _

PDF  |  HTML  |  Supplementary Files  |  How to cite

Oncotarget. 2017; 8:73087-73097. https://doi.org/10.18632/oncotarget.20391

Metrics: PDF 1744 views  |   HTML 2220 views  |   ?  


Jae Heon Kim1,2,*, Bora Lee3,4,*, Deok Hyun Han1, Kyoung Jin Chung1, In Gab Jeong1 and Benjamin I. Chung1

1Department of Urology, Stanford University Medical Center, Stanford, CA, USA

2Department of Urology, Soonchunhyang University Hospital, Soonchuhyang University Medical College, Seoul, Korea

3Department of Biostatistics, Clinical Trial Center, Soonchunhyang University Bucheon Hospital, Bucheon, Korea

4Department of Statistics, Graduate School of Chung-Ang University, Seoul, Korea

*These authors contributed equally to this work

Correspondence to:

Jae Heon Kim, email: [email protected]

Keywords: androgen deprivation therapy, prostate cancer, dementia, alzheimer’s disease

Received: June 22, 2017     Accepted: August 04, 2017     Published: August 22, 2017


Limited literature exists on the association between androgen deprivation therapy (ADT) for prostate cancer (PCa) and subsequent dementia and the study conclusions are in conflicts with one another. We searched several cohort databases from 1960 to 2017 for observational or prospective studies that reported on an association between ADT for PCa and subsequent dementia. A meta-analysis was performed to cumulatively determine the association between ADT and dementia including Alzheimer’s disease using an incidence rate ratio (IRR), crude hazard ratio (HR), and adjusted HR. Seven studies were eligible for the meta-analysis, with the inclusion of a total of 90, 543 prostate cancer patients. The pooled overall IRR, crude HR, and adjusted HR were 1.78 [95% confidence interval (CI): 1.51–2.10)], 1.80 (95% CI: 1.05–3.10), and 1.59 (95% CI: 1.16–2.18), respectively. A meta-regression analysis showed that the crude HR was affected by both follow -up duration and lag time in the univariate model (p = < 0.001). However, IRR and adjusted HR were not affected by these moderators. The overall outcomes of IRR, crude HR, and adjusted HR were found to be balanced in the sensitivity analysis. A positive association was demonstrated between ADT and the subsequent incidence of dementia in this meta-analysis. Methodological difference including follow-up duration and the time lag could be related with the discrepancies.

Creative Commons License All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 4.0 License.
PII: 20391