There are many available effective treatment options for patients with locally confined prostate cancer including surgical and radiation modalities. The beneficial effects of androgen deprivation on men with prostate cancer were first realized over 70 years ago, and today, it remains a mainstay of treatment for advanced prostate cancer patients [19].

In 1941, Huggins and Hodges established a strong link between prostate cancer and testosterone. Twenty-five years later, they won a Nobel Prize for their groundbreaking work in prostate cancer treatment. They observed increased levels of acid phosphatase in two patients with metastatic prostate cancer who received exogenous testosterone, although one of the patients was already castrated. They also described disease regression in patients with metastatic prostate cancer after the administration of high doses of estrogens or surgical castration. They correlated this with a decrease in acid phosphatase level. Despite the multiple methodological flaws in the study, most notably the small cohort size, they arrived at the conclusion that high levels of serum testosterone were associated with enhanced prostate cancer growth [20]. With these observations, they provided the first scientific basis for advanced prostate cancer treatment.

Subsequent studies, during the pre-PSA era, supported the correlation between testosterone and prostate cancer progression. Prout et al. reported their results of an observational study of patients with recurrent prostate cancer after castration. The patients were administered exogenous testosterone, and this resulted in disease progression or death [21]. Similarly, Fowler et al. also reported on the adverse effects of testosterone administration to patients with metastatic prostate cancer. However, the majority of patients in this study were already androgen deprived upon receiving testosterone administration [22].

These early observations gave rise to the concept that high testosterone results in growth of prostate cancer and administration of exogenous testosterone to men with prostate cancer is harmful. Further research went beyond the correlation and sought to prove causation. For example, Pollard et al. supported this concept by inducing prostate cancer in rat models by exposing them to exogenous testosterone [23]. While there may be compelling evidence in animal and laboratory models, quality evidence in humans is lacking [24].

Although androgens are important for maintaining the prostate gland, and despite the effectiveness of androgen deprivation therapy against metastatic prostate cancer, the hypothesis of high circulating testosterone levels adversely affecting prostate cancer has been challenged in recent literature. Some studies have found a positive correlation between prostate cancer and testosterone levels; [25, 26] however, the vast majority has found no direct relationship between high levels of testosterone and the pathogenesis of prostate cancer [27, 28].

Roddam et al. has demonstrated the lack of evidence based support for this concept after reviewing 18 prospective studies which evaluated the association between testosterone level and prostate cancer risk. Clinical data from 3,886 men with prostate cancer and 6,438 control subjects were analyzed. They found that serum testosterone was not significantly higher in patients with prostate cancer. Conversely, the incidence of prostate cancer was not significantly increased in patients with high testosterone levels compared to those with low testosterone. Although the validity of some of the individual studies is in question due to methodological issues, this meta-analysis concluded that there was no statistical difference in pre-diagnostic serum androgen concentration between patients who developed prostate cancer and control groups. Therefore, high androgen concentration is not associated with the risk of prostate cancer [28].

Contrary to the above reports, a recently published prospective observational study by Pierorazio et al. evaluated 781 men with hormone measurements over several decades. They observed the likelihood of high risk prostate cancer in men over age 65 was double for each unit rise in the free testosterone index. The authors concluded that higher levels of serum free testosterone are associated with an increased risk of aggressive prostate cancer in older men. This study differed from others in the literature because it considered the effect of long term exposure to circulating androgens and its importance in the development of prostate cancer [29].

The opposite spectrum of the association has also been evaluated. One would expect that if high androgen levels increased the risk of prostate cancer, then it would follow that low androgen levels would decrease the risk. Surprisingly, however, multiple studies have not shown this to be true. There are numerous examples in the literature that have demonstrated low androgen levels not only do not decrease the risk, but may actually be associated with an increased risk of prostate cancer, worse 5-year biochemical relapse-free survival, increased positive surgical margins, worse pathological stage, increased percentage positive biopsy cores, and higher Gleason score [29–32].

In an early study alluding to this concept, Morgentaler et al. observed an increased rate of prostate cancer in hypogonadal men (mean age 58 years) with a normal prostate exam and a PSA  <  4 ng/mL. In follow-up investigations, including a larger sample (345 men) of patients with hypogonadism and PSA  <  4 ng/mL, the same authors found a correlation between the severity of testosterone deficiency and an increased risk for a positive prostate biopsy. There was more than double the risk of positive biopsy in patients with testosterone levels in the lowest tertile compared to those in the highest tertile [33].

In addition to the described correlations between low testosterone level and cancer risk, there is also evidence for the associations with worse pathological determinants of prostate cancer [34–36]. Zhang et al. measured serum total and free testosterone levels in 164 patients with high or moderate grade prostate cancer. The levels in patients with high grade prostate cancer were significantly lower than in patients with moderate grade cancer [37]. Another study, by Massengill et al., retrospectively analyzed the records from a large group of patients who had undergone radical prostatectomy for prostate cancer. They found that preoperative testosterone was significantly lower in patients with non-organ confined disease compared to those with organ confined disease. They concluded that pretreatment total testosterone level was a predictor of extraprostatic disease in patients with localized prostate cancer [38].

It has been postulated that aggressive prostate cancer can exist and may grow with a low intraprostatic DHT concentration such as that induced by 5-alpha-reductase medication [39]. As previously discussed, treatment with 5-alpha-reductase inhibitors reduces the intraprostatic concentration of DHT. The results of the Prostate Cancer Prevention Trial established a link between androgen metabolism and prostate cancer. Inhibition of 5-alpha-reductase and, by direct correlation, reduction in intraprostatic DHT levels, are associated with decreased incidence of low grade prostate cancer and slightly higher incidence of high grade cancer [17]. Bologna et al. assessed the effects of finasteride, cyproterone, and hydroxyflutamide on prostate cancer tissue culture and reported that cancer cells stimulated by testosterone and DHT show growth only at low concentrations of androgens. If higher concentrations were used, prostate cancer cell growth was inhibited [40].

Nishiyama et al. reported that patients with a Gleason score of 7–10 prostate cancer had lower concentrations of intraprostatic DHT compared to Gleason 6 cancers. However, there was no difference in serum androgens between the patient groups. They concluded that low DHT in cases of aggressive prostate cancer is probably sufficient to activate AR expression and propagate tumor growth [41].

Prostate cancer regression at castration levels of testosterone is undeniable. Cancer growth and progression after administration of testosterone to patients with prostate cancer and castration levels of testosterone is also an observable fact. However, it has never been reliably demonstrated that an increase in testosterone levels, above a near castrate level, causes any significant incremental growth in prostate cancer. To explain these observations, investigators have introduced the “saturation model,” which proposes that there is a certain serum androgen concentration where all of the available AR’s are bound to androgen [41]. The receptors become saturated with ligand and any additional androgen above this saturation point is unable to further stimulate prostate cancer growth. This saturation point occurs at low concentrations of androgen, slightly above the castrate range. This interaction between androgen and the AR is similar to other biological systems in which receptors become saturated by their ligands. There are two phases in the saturation model. At or below near castrate levels of testosterone, prostate cancer cells are the most sensitive to androgens such that androgen is the rate limiting factor in the AR activation pathway. Above this level, prostate cells exhibit little if any growth changes and become indifferent to higher levels of androgens [41].

A study by Marks et al. gives strong support for the saturation effect on the AR. Forty-four patients with hypogonadism were randomized to testosterone replacement or placebo for 6 months. While there was an appropriate increase in serum testosterone levels in the study group patients, no change was observed in prostatic tissue testosterone or DHT levels. In addition, there was no change in tissue biomarkers, gene expression, prostate histology, or incidence of cancer between the groups. PSA levels increased slightly in both groups, although it was higher at baseline in the group receiving testosterone [42].

As described previously, a number of studies have established an association between high-risk prostate cancer and low serum testosterone levels. While there is a correlation, there is not enough evidence to conclude if one is cause or consequence of the other. In an attempt to explain this association, a suppression theory has been proposed. Based on measurements of preoperative and postoperative testosterone levels in men undergoing radical prostatectomy, Miller et al. hypothesized that prostate cancer cells may secrete a substance that interferes with the normal secretion of testosterone by the testis. Investigators also measured estradiol, free testosterone, DHT, FSH, and LH. There was an increase in the postoperative levels of all these hormones except for DHT, which was decreased relative to the preoperative levels. Given these findings, the authors suggested that there may be a factor, such as inhibin or DHT, secreted by prostate cancer cells causing suppression at the hypothalamic–pituitary axis, ultimately resulting in a decrease in testosterone production [43].