NIAGARA FALLS, Ont.—Prostate cancer (PCa) patients who maintain extremely low testosterone levels while on continuous androgen deprivation therapy (ADT) are at lower risk for the development of castrate-resistant PCa (CRPC).
A secondary analysis of the PR-7 trial presented at the Canadian Association of Urology’s 2013 annual meeting showed patients on continuous ADT whose serum testosterone levels stayed at or below 20 ng/dL had only half the risk of developing castrate-resistant PCa compared with those whose levels did not fall below 50 ng/dL. Maintaining very low testosterone levels also was associated with a significantly longer span between the development of CRPC and death.
“This is one of the most important studies I’ve ever done,” lead investigator Laurence Klotz, MD, of the Sunnybrook Health Sciences Centre in Toronto, told Renal & Urology News. “It’s as good data as we’re ever going to get on this question because you’re not going to have a pure population like this probably ever again.”
Dr. Klotz was the lead investigator of the PR-7 trial, which involved 1,386 men with a PSA level greater than 3 ng/mL persisting for more than a year after primary or salvage radiotherapy for localized PCa (N Engl J Med 2012;367:895-903). The men were randomized to either intermittent or continuous ADT. The men used a range of luteinizing hormone-releasing hormone agonists during the trial. The median follow-up was 6.9 years. The median overall survival rates were similar in both groups, as were the seven-year cumulative PCa-mortality rates.
Two previous small, retrospective studies had suggested that very low levels of testosterone – below 20 ng/dL (0.7 mM/dL), which is true castration – are associated with prolonged time to PCa progression (J Urol 2007;178:1290-1295 and BJU Int 2010;105:648-651).
“But both of those studies were small and retrospective – really relatively poor quality of evidence,” Dr. Klotz told meeting attendees.
Dr. Klotz and several co-investigators focused on data from the three measurements of serum-testosterone levels obtained during the first year of the PR-7 trial from the 626 men in the continuous ADT arm.
Median serum testosterone levels were 20 ng/dL or less in 53% of subjects, 20-50 ng/dL in 42%, and 50 ng/dL or greater in 5%. The median testosterone levels were significantly associated with patients’ age: older patients had lower testosterone levels. The median levels did not correlate with PSA levels, or with grade or extent of PCa.
“To my surprise, the result was in fact strongly positive for an effect of testosterone level on time to castration resistance,” Dr. Klotz said.
His team defined castration resistance as a progressive rise in PSA above 4 ng/dL with testosterone levels of 85 ng/dL or lower.
Patients whose median testosterone level failed to drop below 50 ng/dL had a nearly twofold increased risk of developing CRPC compared with those whose testosterone level remained at 20 ng/dL or less, and a 41% increased risk compared with subjects whose testosterone levels were 20-50 ng/dL.
Furthermore, the time to development of CRPC was almost three times more likely in the group with the lowest testosterone level compared with patients who had the highest testosterone levels, and 30% more likely compared with the 20-50-ng/dL group.
The time to CRPC by maximum testosterone level showed a similar trend but was not statistically significant.
“The implication is that if the testosterone on average is low, that’s important, but the mini-flare might not be that important,” Dr. Klotz said.
Cause-specific survival did not reach statistical significance, “probably due to lack of events,” Dr. Klotz said, although there was the same trend to worse results in men who failed to drop their testosterone to very low levels.
“The conclusion is you have to measure testosterone and you have to, we think, manage patients according to their testosterone [levels],” Dr. Klotz told attendees. “If you have failure of testosterone suppression you should consider changing the drug, perhaps adding an anti-androgen. Testosterone as well as PSA should be monitored on a regular basis in these patients. And there’s obviously a goldmine of studies to do with the PR-7 study data related to the predictive value of T, for example in patients on the intermittent arm, and those are ongoing.”