Clinical experience with testosterone treatment in sildenafil non-responders

view presentation | view abstract | view curriculum vitae | watch interview video | view interview transcript | print this page

Ridwan Shabsigh, MD, Associate Professor Of Urology, Columbia University, Director, New York Center For Human Sexuality, New York, USA

My involvement with testosterone for the treatment of erectile dysfunction is a long story. However, don’t worry. For the sake of time, I will share with you the short version of a long story. When I left my training at Baylor College of Medicine in Houston, Texas, and joined the Department of Urology at Columbia University, my chairman, Dr. Carl Olsen, had a discussion with me and he said: "what do you think would be new?" and we discussed testosterone effects. He said testosterone works only the brain, on the libido. I said, no, I think it might work on the penis. He said, if you can prove that it works on the penis, you will make a lot of good work that will also affect a lot of patients and will be very interesting from a scientific viewpoint. So that started a long involvement in testosterone. 

The background of that was that animal experiments did confirm in a number of studies, including some studies we did in our laboratory, the dependency of the nitric oxide synthesis of the corpora cavernosa on testosterone. Dr. Abdul Traish did show beautiful data from a number of studies in his laboratory showing the involvement of testosterone at multiple levels of the erectile physiology. Human research, however, as Dr. Buvat has shown, has given unclear conclusions. A significant number of patients seen in the erectile dysfunction practice although they might not have severe hypogonadism, they have what we call low-normal or high-abnormal testosterone, depending on how you like to word this. Furthermore, in recent years, we had anecdotal experiences where we had a patient who had poor response to PDE-5 inhibitors, mainly Sildenafil, and had improvement of the response after treatment with testosterone in those who have low serum testosterone. So, early experiments we did in our laboratory in the mid-90s, we looked at the simple effect of testosterone on the corpora cavernosa on the tissue.

This is looking at apoptosis, or programmed cell death we can see here the effects in the normal control rat, in the three-day post-castration, five-day post-castration, you can see this is an electrophoresis of the DNA from the corpora cavernosa in the rat showing the laddering pattern which is fragmentation of the DNA. This is one piece of evidence of apoptosis in the rat penis following castration. This simply means that a population of cells in the corpora cavernosa depends for its life on testosterone. Now, when I presented this data, my chairman said, well, I have never seen anybody's penis fall off after treatment with Lupron, and that is true. There must be a small number of cells, but they might be a very important number of cells and those cells might be important for function. This is just structural evidence, this is not a functional evidence.

Looking further at some surrogate markers of apoptosis, this is the expression of cell-fated glycoprotein-2. Simply, this is a surrogate marker for apoptosis, programmed cell death, in the rat's penis. You see significant messenger RNA expression after castration as compared to the control rat. 

Well, we wanted to look at the tissue itself, so we did in situ localisation in the corpora cavernosa of the rat. This is a classic presentation in situ of apoptosis. You can see condensed hypochromatic nuclei with vascuolation of the cytoplasm. This is the classic look of apoptosis in the corporal cavernosum. Well, we thought that the animal experiments provided the opportunity of an on-off switch. That means castration, removal of testosterone, replenishment - maybe we can reverse this. So we had an interest to see whether testosterone replenishment of the castrated rat would result in new cell proliferation in new DNA synthesis. 

So we looked at new DNA synthesis using a nucleide surrogate and staining of the nuclei. The dark brown nuclei you see here have nuclei that have synthesised new DNA. New DNA meaning that after replenishment with testosterone, a population of cells did proliferate and this is a combination staining of the nuclei and of the cytoplasm with trichrome showing that in the smooth muscle and in the fibrous tissue, this appears to be a pan-cellular effect, even in the endothelial cells. When we looked at this very carefully, under very high power, we thought that we did have some suspicion that endothelial cells might also be the target. This was something very interesting to us. Here we have a vascular organ dependent on a hormone – a sex hormone.

We looked at other studies meanwhile in the literature and Dr. Abdul Traish did present a number of those. Just as an example, this study looked at function. Sodium nitrous oxide is a donor of nitric oxide, the neurotransmitter for erection in the penis. Here, looking at control rats and castrated rats, showing deficiency in nitric oxide and showing increased intra-cavernosum pressure in the control rat, but not in the castrated rat, with a nitric oxide donor. The preliminary chain reaction did demonstrate that messenger RNA levels of nitric oxide synthesis in the penis were greater in the control rats as compared to castrated rats.

Furthermore, the difference in the erectile response between control and castrated was no longer apparent after giving an arginine competitor. We know that arginine is the precursor of nitric oxide. If you give methyl arginine, which is a competitive inhibitor of nitric oxide synthase, then this different is no longer apparent. This gives evidence that nitric oxide pathway both from the inhibition part and also the stimulation part is dependent on testosterone.

This is another elegant study that was done looking at the model of castration, castration plus testosterone, versus sham-operated. They looked here at a number of ways to induce erections in the rat. In the rat, you can induce central erections by apomorphine. Apomorphine has no ill effects on the penis, it has an effect on the dopamine receptors in the brain, generating erections via the central and peripheral nervous systems. They looked at electro-stimulation, which is a peripheral neural vasodilator effect of papaverine. They show that all three types of erections were dependent on testosterone and also staining of nitric oxide containing nerves was decreased in those that were castrated and elevated than those that were replenished.

The other part of the story that attracted our attention was data like this from John Morley, showing the gradual decline in testosterone with age. Not only the gradual decline of total testosterone.

Although data from Tenover and Dr. Wang earlier shared a wonderful review of all this, showing that the bio available, that the free testosterone and the total testosterone were decreased with aging, attracted our attention also in that maybe some of the PDE-5 inhibitors had a relationship to this. So when all this came together from animal data, from the clinical data on testosterone, and also from anecdotal experiences that we had in our clinic.

We approached the pharmaceutical company that makes the testosterone gel with the idea of doing a double-blind, placebo-controlled randomised study on testosterone therapy of patients who failed PDE-5 inhibitor, sildenafil, with testosterone versus placebo. 

We wanted to have the patients, who had not only the failure of PDE-5 inhibitor, but also had the low serum testosterone. One problem we had when we approached the clinical team, and you know the clinical team in planning a study includes a very important person called a statistician. The statistician asked me how do we calculate power? How many patients should we include? I said, well, this is a pilot study, we do not have data on the therapeutic effect, we do not have enough mathematics to design a fully-powered study, we have to make a best-educated guess, and that is what we did.

The study procedures were simple. We took patients, who had erectile dysfunction for more than three months, had low to low-normal total testosterone, that means less than 400 ng/dL, and I submit that this is something that can be debated. It was confirmed in the first four weeks of the study that there were non-responders to Sildenafil, so we did not accept historical lack of response. We wanted to confirm non-response within the study. After confirmation of those, who were not responsive to Sildenafil, they were randomised to open-label Sildenafil plus double-blind gel of testosterone versus placebo.

Evaluations occurred at four, eight, and twelve weeks, looking at a number of parameters, including the primary efficacy endpoint of the erectile function domain score, the secondary efficacy endpoints, including the IIEF domains of sexual desire, orgasmic function, total score satisfaction, including global assessment questions, and also the serum testosterone levels. 

At the entry of the study, 75 patients were randomised although, with time, we lost some of these patients to follow-up, so this was a relatively small study. This is one of the limitations, and was also a study that is still, in my opinion, considered a pilot for further research in the future, and I am happy to learn that other investigators are going in this direction. The mean age was 58, which is similar to many patients with erectile dysfunction, and also I would like to attract your attention to this mean age as similar to a lot of the clinical trials we see on erectile dysfunction. If you look at clinical trials of Sildenafil, Tadalafil and Vardenafil, this is very common. The patients were mostly on the obese side. The aetiology of erectile dysfunction was organic or a combination in the majority of patients, and over 90% of patients had long-term erectile dysfunction, so this is really not a group of trivial or mild erectile dysfunction.

As we see here on the table of the severity of erectile dysfunction, at base line, the majority of patients had either moderate or severe erectile dysfunction.

The first result is the testosterone levels as we see here, presented in nanogram per deciliter, (ng/dL), this is the arm of patients on the testosterone gel. This is the arm of patients on the placebo gel. On the placebo gel, as we see, the mean values remain low or in the high-abnormal range and those were also the ones in the treatment arm, they were elevated, significantly at all points, although there was a slight decline at the end here. This, again, remains to be explained what this slight decline means. Is it because we lost some patients as we followed into the study or whether there were other reasons? I must share with you that in this study, we had single-doses of the gel which was the 50mg. We did not permit flexible dosing or increased dosing.

This is the result at four weeks of a total of 67 patients looking at testosterone plus sildenafil, placebo plus sildenafil. We look here at the erectile function domain. These numbers are the mean change in this data. This is not the absolute number - this is the change. We can see significant changes on the erectile function, we can see significant changes in the orgasmic function, in the overall satisfaction and in the total score in favour or testosterone plus sildenafil as compared with placebo plus sildenafil. Interestingly, sexual desire was not significant in our data. We believe that might be because the sexual desire domain of the IIEF of the international index of erectile function might not be very sensitive and a good instrument for detection and measurement of sexual desire.

We looked further at the data over the four, eight, and twelve weeks, and we can see here that we reach statistical significance only at four weeks. We had statistical trends at eight and twelve weeks in terms of erectile function. We tried to go back and look at the data, looking for explanations. One of the explanations would be the loss of patients as we went into the eighth and twelfth weeks, which might take us out of statistical power for the study, other interpretations are also open.

This is the improvement of response to sildenafil as measured by the global assessment questions. We had three global assessment questions in this study. One of them was asking the patient did the gel improve your response to Viagra? And this was how it was. It was statistically significant at four weeks with the testosterone around 60% and the placebo around 28%. This statistical significance was lost to a trend in the other endpoint.

We can see here the orgasmic function showing again the same pattern. Orgasmic function did improve significantly, as measured by the orgasmic function domain of the IIEF in the four-week measurement in the eighth and twelfth weeks there was only a trend.

So, in conclusion from the study, testosterone-replacement therapy with testosterone gel of 1% did improve the erectile response to sildenafil 100mg. That is another important point - we had all patients here with a maximum dose of sildenafil. We did not permit lower doses of sildenafil. We did the same thing at the four-week run of sildenafil, where we confirmed failure to 100mg of sildenafil, so this is maxing out the dose of the PDE-5 inhibitor. Compared to placebo, testosterone gel 1% in a 5 gram dose also improved orgasmic function. This is another important aspect of the treatment with testosterone versus treatment only with PDE-5 inhibitors because testosterone addresses a number of other sexual domains in addition to erectile function. As a matter of fact, interviewing some of my patients who were in this study, they reported, they said, doctor, my ejaculate volume is higher, my feeling, my thrill, the “wow” felt with orgasm was much improved with the testosterone treatment. Testosterone is efficacious and safe for the treatment of ED in men with low to low-normal testosterone, who have failed prior treatment with Sildenafil 100mg alone.

So, in summary of this short version of a long story, animal and human data proves the role of testosterone in the nitric oxide pathway of the erectile physiology. Testosterone therapy with 1% gel improved response to sildenafil and also intuitively to other PDE-5 inhibitors, tadalafil and vardenafil. There is no reason to expect that that would not be case and further studies will be performed on this. Screening for hypogonadism is indicated in patients with erectile dysfunction in general, but I would say especially those who fail PDE-5 inhibitor therapy. 

Testosterone may be considered for the treatment of erectile dysfunction in men with low to low-normal testosterone, who have failed prior treatment with PDE-5 inhibitor therapy.