PerspectiveIn the Journals Plus

Androgen deprivation therapy does not increase risk for Alzheimer’s disease, dementia

Show Citation

November 14, 2017

Older men treated with androgen deprivation therapy for prostate cancer appeared to be at no greater risk for Alzheimer’s disease or dementia, according to a study published in Journal of Clinical Oncology.

“These data suggest that ADT presents no hazard of Alzheimer’s disease to men aged 67 years or older who are enrolled in Medicare,” Clement Joseph McDonald, MD, senior investigator and scientific director at Lister Hill National Center for Biomedical Communications at the National Library of Medicine, and colleagues wrote. “Indeed, ADT appears slightly protective [against Alzheimer’s disease].”

Prostate cancer accounted for 21% of all male cancers in 2016 and for 8% of all male cancer deaths.

Kevin T. Nead

In previous studies published in Journal of Clinical Oncology and JAMA Oncology, Kevin T. Nead, MD, MPhil, radiation oncology resident at Perelman School of Medicine at University of Pennsylvania, and colleagues reported a strong association between ADT and the risk for both Alzheimer’s disease (HR = 1.88) and dementia (HR = 2.17). Nead and colleagues noted that the decreased levels of testosterone associated with ADT served as risk factors for Alzheimer’s disease.

In this study, McDonald and colleagues analyzed data from 1.2 million fee-for-service Medicare beneficiaries diagnosed with prostate cancer from 2001 to 2014 to determine the unadjusted and adjusted risks for Alzheimer’s disease. Average observation time was 5.5 years.

Of the 1.2 million men, 35% received ADT. Of those treated, 8.9% (n = 109,815) developed Alzheimer’s disease and 18.8% (n = 223,765) developed dementia. Another 26% of men died without developing Alzheimer’s disease and 33% died without developing dementia.

ADT recipients demonstrated higher unadjusted rates of Alzheimer’s disease (17 vs. 15.5 per 1,000 person-years) and all-cause mortality (73 vs. 51.6 per 1,000 person-years) than those who did not receive ADT. Unadjusted rates of dementia (38.5 vs. 32.9 per 1,000 person-years) and all-cause mortality (60.2 vs. 40.4 per 1,000 person-years) also were higher among recipients in a parallel analysis that analyzed dementia as the event of interest.

However, after researchers adjusted their analysis for other cancer therapies and other covariates, ADT recipients had no increased risk for Alzheimer’s disease (subdistribution HR [SHR] = 0.98; 95% CI, 0.97-0.99) and only a 1% increased risk for dementia (SHR = 1.01; 95% CI, 1.01-1.02).

Patients treated with ADT appeared more likely to die before progression to Alzheimer’s disease (SHR= 1.24; 95% CI, 1.23-1.24) or dementia (SHR = 1.26; 95% CI, 1.25-1.26).


Researchers observed no dose effect on the risks for either disease.

These results contrast with the findings reported by Nead and colleagues, which showed a statistically significant association between ADT and risk for Alzheimer’s disease and dementia.

“The differences in conclusions between this study and the study by Nead [and colleagues] could be a consequence of different choices of methods, definitions and data source,” McDonald and colleagues wrote. “However, this study had a larger patient sample (1.2 million vs. 17,000), longer follow-up (median, 4.9 years vs. 2.7 years) and more complete ascertainment than the study by Nead [and colleagues].”

In an accompanying editorial, Nead wrote that one risk of pooling heterogeneous data is that the true effect can be averaged out, and it may not be appropriate to derive a summary estimate if there is high heterogeneity.

Another challenge is the reliance on claims or electronic medical record data, which are captured primarily to aid reimbursements and clinical care rather than research, Nead wrote.

“This variability in reported studies is balanced by the biologic plausibility for an association between ADT and dementia,” Nead wrote. “Androgens modulate neuron growth, function and maintenance. Low testosterone predicts future dementia. Low testosterone and ADT are associated with elevated -amyloid protein levels.

“In men with dementia, testosterone supplementation improves spatial and verbal memory,” Nead added. “ADT has been linked to impaired cognitive performance within 6 months of ADT initiation in prospective studies, though this also is controversial. Finally, decreased testosterone levels and ADT have been associated with cardiometabolic diseases, which also have been linked to dementia risk. Ultimately, given the biologic plausibility of this association and the disagreement among existing analyses, a higher level of evidence, such as prospective evaluation, is needed to answer whether ADT is causally associated with an increased risk [for] dementia.” – by Chuck Gormley



Nead KT, et al. J Clin Oncol. 2016;doi:10.1200/JCO.2015.63.6266.

Nead KT, et al. JAMA Oncol. 2017;doi:10.1001/jamaoncol.2016.3662.


Disclosures: The Intramural Research Program of the National Library of Medicine of the NIH funded this study. The authors and Nead report no relevant financial disclosures.

itj+ Perspective


ADT is administered to patients with locally advanced or metastatic disease. Locally advanced disease is generally treated with radiotherapy. Patients treated with radiotherapy receive ADT prior to radiation treatment and then are followed for a short time. However, short time is an arbitrary period — it can be between 6 months and 18 months. Long-term ADT — which consists of subcutaneous injections every 3 months — is prescribed to patients with metastatic disease. Because this is a systemic treatment, there are side effects. Dementia is not known to be a significant cause of ADT. This study by Baik and colleagues, and a previous study by Nead and colleagues, both investigate the association between ADT and dementia; however, a couple of issues remain.

Prostate cancer is a disease of the elderly; so is dementia. As patients get older, their risks for having dementia and prostate cancer increase. So, it is not easy to demonstrate whether there is a direct causal association between the two, because there are too many possible reasons for both diseases. To account for the known possible factors, statistical adjustments are made; however, we do not know all the factors, and these databases do not have granular data.

Secondly, metastatic patients are not candidates for surgery or radiation, meaning that systemic treatment is the only option. Any patient treated with ADT will eventually succumb to resistance and, eventually, patients will die. Many patients who receive ADT are not followed long enough to be diagnosed with dementia. However, this cannot be addressed by retrospective studies in literature, and prospective studies are needed to validate these results.

Further, large databases often have missing or inaccurate data that are filtered out before analyses are made. In these studies, researchers used disease codes to identify patients. However, the fact that a patient does not have a disease code does not mean he or she doesn’t have the disease. It may be a recording mistake, or these patients simply may have not been seen by a neurologist yet. Therefore, accuracy of data is always an issue.

The bottom line is Nead and colleagues took few cofactors into consideration due to the limited data they had from their institutional database. Also, they did not analyze radiotherapy patients separately, who receive treatment for a limited time. Furthermore, ADT is a general term and there are different types of mechanisms. On the other hand, Baik and colleagues used a much larger Medicare database and performed a much more comprehensive list of factors associated with dementia. They had a better study design. Still, their study design also had limitations.

Finally, ADT is a systemic treatment that reduces the testosterone level of the blood to undetectable levels. There are side effects associated with this treatment. Evidence suggests that there is not an association between ADT and dementia. It should also be noted that Nead and colleague reported an association but not a direct causal relationship. It is not possible to come to a definitive conclusion with either study.

Alp Tuna Beksac, MD

Icahn School of Medicine at Mount Sinai

Disclosure: Beksac reports no relevant financial disclosures.