Synergistic Value from Real-Life Registry Studies and Randomized Controlled Trials in Testosterone Therapy of Testosterone-Deficient Men

The setting of randomized placebo-controlled clinical trials (RCTs) is considered the gold standard for clinical studies and is seen as required to prove the efficacy and (if designed for that purpose) safety of a treatment regimen. In addition to such approaches, data from registries of patients attending single institutions for treatment have been evaluated and published. What value do registry studies add to the knowledge obtained from RCTs?

To this end, one has to take a look at the principles of RCTs that require strict criteria for inclusion and exclusion to generate a sample of patients that is as homogeneous as possible so that the effects of placebo versus true medication can be evaluated. The recently published T-trials may serve as an example, ¹ comparing the effects of testosterone therapy (TTh) with application of placebo in elderly men with functional (age-related) hypogonadism. The T-trials included 790 testosterone-deficient men who fulfilled rather strict selection criteria. ¹ This provided the possibility to delineate the putatively positive effects of TTh versus placebo for a selected number of androgen-dependent target parameters.^1,2

However, although these 790 testosterone-deficient men were included in this RCT, they represent ∼10% of all testosterone-deficient men at that age who were detected during screening for this study. ¹ That means, information was obtained about the efficacy of TTh versus placebo in a very clearly defined homogeneous subgroup of men, whereas this RCT still does not provide information on the effects of possible TTh in a majority of testosterone-deficient men, as 90% of these patients found during the screening process were excluded.

In every day clinical practice, such a selection does not take place. Also men who were testosterone deficient but excluded from the T-trials deserve, require, and, to an unknown extent, also receive TTh. That means, information regarding the effects of TTh in the remaining 90% of testosterone-deficient men who did not meet the criteria for inclusion and were excluded from the T-trials but might have sought help for their clinical condition in a common clinical practice could exist, but information on this is, to a major extent, not accessible. Thus, effects of TTh on these men might actually differ from the effects observed in the T-trials.^3,4

A registry in a single institution or in a set of institutions, collecting data and information on a predefined basis regarding treatment effects of TTh on such patients, is a unique opportunity to make such very valuable data available. An example is a study on the long-term effects of TTh on testosterone-deficient men with type 2 diabetes mellitus. ⁵ Such centers providing a registry usually establish unique relationships between patient and doctor and the patients are involved in decision making regarding their own health: patients are included in the registry as they come to the doctor's office, regardless of fulfilling specific criteria.

Thus, major differences between RCTs and observational studies are “real-world patients” seen in everyday clinical practice versus a carefully selected subset of patients fulfilling lists of inclusion/exclusion criteria. In addition, the duration of an RCT is often limited due to prohibitive costs and time investments on the part of both physicians and patients. For example, the T-trials were carried out for 1 year, with one follow-up year of observation without treatment, whereas in a registry study, duration does not have to be limited and there is no restriction in terms of number of patients. Thus, a registry can obtain data over a long period of time, in some published cases approaching >10 years. Treatment of patients over such a long time period within an RCT setting would be unethical, leaving patients on placebo treatment for a long time period while it is already proven by shorter RCTs that TTh is beneficial for a subset of parameters. These patients on placebo treatment would, therefore, definitely suffer from one set of symptoms to prove that another set of symptoms might also be influenced.

Thus, another form of evidence can be provided from registries that cannot be obtained from an RCT, that means, a set of “real-life” patients observed over a very long period of time will provide other data than a strictly defined subset of patients observed over a much shorter period of time. Other advantages of RCTs remain, however: clearly set time and treatment protocols, stronger adherence to medication (although not fully possible in case of self-administration at home, i.e., using transdermal preparations), source data verification, and external review. The quality of registry data might be improved and increased in credibility, if such concerns (data handling, uniformity of patient care) are made open and accessible as possible.

One question that arises frequently in connection to data derived from registries is concerning the magnitude of the changes in the parameters measured in response to TTh (e.g., weight, waist circumference, glucose, glycated hemoglobin, insulin, and lipid profiles), whereas such effects appear to be much weaker in RCTs. It is most likely attributed to the duration of treatment with TTh and patient adherence to the medication in patients observed in a registry compared with those being treated within an RCT. Such results can be seen for example in the effects of TTh on glucose metabolism.^5,6 Some processes take longer time than a year because they result from target tissue remodeling. In agreement, a recently published RCT that lasted 2 years corroborates the finding of the long-term registry. ⁷ In this trial, marked changes in body composition, that is, loss of fat mass and gain of muscle mass, were observed and actually required the 2-year duration to be achieved. The long-term effect is facilitated by testosterone inhibition of adipogenesis, which requires long duration.^8,9 Similarly, muscle growth and differentiation in response to androgens also are time dependent. Moreover, the physiological changes in bone tissue also improve over time and start to be visible within 1 year but continue to improve.^10–12 Such effects can be suspected also for other androgen-dependent tissues (i.e., liver tissue, endothelium) but are not well documented.

Thus, although RCTs are pivotal to provide evidence on the principal effects of a treatment, in this case TTh, the observed findings are limited to a clearly defined subset of patients over a clearly defined period of time. A frequently used hierarchy of evidence from RCTs down to registry is undoubtedly well suited for questions of pure therapeutic efficacy, but is of limited to no value for other issues, such as the daily-life effectiveness of therapies, the association between exposures that cannot be controlled, and outcomes, or the safety of a therapy or intervention. ¹³ The effects of TTh on all patients, not only those fulfilling strict inclusion criteria over a long period of time (that might expand to a decade and longer), are likely to be different and can be described—and sometimes even detected—by registries. ¹⁴ Such registries should be continuously improved in quality of data acquisition and transparency.

Hence, both RCTs and open “real-life” registries contribute significantly to the clinical knowledge of the effects of TTh. The nature of these approaches is different and, thus, the nature of the obtained information is also different. Both types of information are unique and contribute synergistically to the whole picture of the effects TTh have on men's health. Andrologists require both sources of knowledge to fully understand the effects of TTh for the best care of their patients.