WorldLink Medical

Neal Rouzier responds to the JAMA article about Men and Testosterone

Neal Rouzier responds to the JAMA article about Men and Testosterone:

Neal Rouzier, MDNEAL ROUZIER, M.D.

JAMA 2013;310(17):1829-1836

On the first day the JAMA article was released I received 500 emails from physicians and patients requesting my opinion of the article that demonstrated an increase in heart attacks and strokes in men treated with testosterone. Today is the second day and I’m afraid to turn on the computer. First of all this was an observational study that was retrospective in nature and this type of study is fraught with compounding biases that are difficult to control as expressed in the discussion section of this study. A randomized controlled trial (RCT) would have much more power than this type of study. Also the problem with an observation study is that it does not prove causation as would an interventional study in a blinded fashion. Therefore observational studies can’t prove causation as well as RCTs and what we should take away from the study is that which the researchers state in the last paragraph, that more studies are necessary before definitive conclusions can be made as to cause and effect. Also, treatment decisions should not be based solely on one study but rather on a trend of studies. Unfortunately the editorial comment section did not express this clearly.

The discussion section of this article mentions that this is the only study that showed this adverse outcome and it was in a select group of individuals. All other studies have shown the opposite outcome, either no effect or protection against heart attacks. Since all other studies show the opposite, and one study does not negate all the other studies, and there were some biases in this study, I would suggest that we do not change anything that we do based on one study with flaws and biases when all other studies demonstrate protection against heart disease and stroke (see attached articles). And this was an observational study which has weaker power than a randomized controlled trial. I’m sure that other experts will voice the same opinion once they review the discussion section of this article as there were many biases and flaws in this study. A review of the index lists the studies that demonstrate protection against heart disease and strokes. In the “Longevity Section” that I present in the Part II course, all of the articles demonstrate improved longevity in those treated with testosterone, but increased morbidity and mortality in those men not treated with testosterone. The WHI study showed that Prempro increased heart attacks and strokes in certain individuals. Subsequent studies have proven that estradiol and progesterone, particularly in younger women, don’t. Perhaps there is a confounding problem in older Veterans with cardiovascular disease that is different from other studies. However, as presented in the Part II course, every study that I review (and there are many) demonstrated a significant improvement in longevity and decreased morbidity and mortality in addition to improvement in all cardiovascular risk factors in men treated with testosterone as opposed to control groups treated with placebo (see attached studies).

Had this study been published years ago, and all subsequent studies since then showed protection against cardiovascular disease, then this study would have probably been ignored and forgotten. However, since it is recent, then we tend to believe it and reject all the past studies that showed the exact opposite outcome. Nevertheless, one study does not negate many other studies that show opposite results and benefits. So I will log this study on the negative side for testosterone results but it is the only such study on this side. This is in contrast to all the other studies that show benefit of testosterone administration. It is interesting that this study appears now, just after I gave 2 lectures to a medical academy this past weekend in Las Vegas. The two one hour lectures were on all the studies of both estrogen and testosterone protecting against heart attacks and strokes. These reviews of the world’s literature demonstrate all the various mechanisms of benefits of hormones in protecting the heart and brain against heart disease, stroke, dementia, and plaque deposition. The data and literature is overwhelming in favor of a protective effect of estrogen in women and testosterone in men. This recent study, although interesting and intriguing, does not change any of the evidence that I presented in these lectures nor does it change my treatment strategies. Until more studies demonstrate the same, I will continue to follow the scientific literature that demonstrates benefit. As per the suggestion from the authors, they state that more study is needed to evaluate these results. I recommend to patients and physicians that they continue the same treatment with both estrogen in women and testosterone in men based on all prior studies that show benefit in spite of this one negative study.

Certain statements in the discussion section of the study deserve comment. The authors do note that other trials and meta-analyses do not demonstrate adverse cardiovascular outcomes. The trend so far in the literature has been a protective effect as trials demonstrated that testosterone therapy improves a number of intermediate outcomes and cardiac risk factors. This new JAMA study is the first and only study to demonstrate harm and should therefore be interpreted carefully in light of all the other studies demonstrating opposite results. In addition, the results of this study differ from a similar retrospective VA study by Shores et al that demonstrated a 39% reduction in mortality risk among patients treated with testosterone which again suggests caution in coming to conclusions only based on the present study. Different confounders and biases might account for the discrepancy. Multiple limitations of this study are noted by the authors that certainly can affect outcomes. All in all, it is an interesting study with unexpected results that are in discordance with all other studies and should not influence current therapy, but one that begs for more study.

For those patients and physicians that are unfamiliar with the current literature on testosterone therapy, I have included 3 attachments that review various categories of hormone replacement. First are studies that review mortality in men treated with testosterone compared to control groups. Studies show improved survival in treated men versus untreated men. There are fewer heart attacks, cancer, and reduced mortality in men treated with testosterone (in contrast to the current study). Other studies go on to prove that low levels of testosterone increase morbidity and mortality in contrast to men with testosterone levels in the higher quartiles. Low levels of testosterone are predictive of an increase in all-cause mortality (CAD, CVD, cancer). So where would you like your levels to be? Other studies show that there was no increased risk of cardiac events in men treated with testosterone (in contrast to the current study).

The second attachment lists all the articles that demonstrate all the physiologic benefits of testosterone administration on cholesterol, lipoproteins, insulin sensitivity, diabetes, inflammatory cytokines, endothelial dysfunction, atherosclerosis, blood pressure, memory loss, Alzheimer’s disease, mood, strength, energy, muscle mass, fat mass, osteoporosis, ED, sexual function, and all-cause mortality. Do you really want to stop the testosterone based on only one negative study? I’m not! What are the consequences of stopping or not taking it? Read the foregoing.

The third attachment reviews beneficial effects on quality of life as well as disease protection. It is amazing the data on reduction of body fat, insulin levels, diabetes, inflammation, and vascular disease. “Testosterone serves to maintain health in every system of the body.” Levels of testosterone in the low to mid-normal range are associated with an increase in illness as listed above.” And don’t forget (pun intended) the protection against Alzheimer’s disease.

Respectfully submitted, Neal Rouzier

Attachment 1

Attachment 2

Attachment 3

Read Dr. Josh Trutt’s response to the JAMA article here.

Read Dr. Mike Clark’s response to the JAMA article here.

Find out more about Dr. Rouzier here.

Share this post

Comments (13)

  • Linden Dillin, M.D. Reply

    As a graduate of Part I and Part II, and a big fan of Dr. Rouzier, I am nevertheless sometimes surprised when Dr. Rouzier states “all the articles show …”. In this blog, Dr. Rouzier refers to this recent negative report (I will not dignify it by calling it a study) on testosterone as “the first and only study to demonstrate harm”.

    There are other negative reports.

    The following is a url for an article at the NEJM that was given to a diabetic, approximately 60 year old friend of mine by his internist, who scared my friend out of supplementing with testosterone.

    Still a fan.

    Linden Dillin, M.D.

    November 19, 2013 at 2:22 am
    • neal rouzier Reply

      Dr. Dillin is correct in that there was a negative study published in the NEJM. Oops. I didn’t know that. In this study 200 older men with severe disease, HTN, DM, hyperlipidemia were given testosterone and some had an increase in MI. This was an observational study and not RCT or protected from biases. Perhaps I can publish my observational study with over 1,000 men treated for up to 15 years and not one suffered an MI. I wonder if they would accept it? Probably not as it is a positive report.
      What can be gleamed from this small observational study is that negates hundreds of other studies over the last 50 years that showed protection and benefit. Thank goodness his patient’s PMD caught this and stopped the testosterone thereby preventing him from having a heart attack! 🙂

      November 24, 2013 at 2:43 pm
  • Susan Steadman, RN, BSN, MSN -FNP Student Reply

    I never doubted it for a minute. The JAMA study was very weak, disappointing coming from professionals who should know more about the strength oo weakness of some kinds of data. I’m a nurse, and even I know better than that! Thanks, Neil, for supporting us!

    November 19, 2013 at 12:28 pm
  • Mike Clark, Ph.D. Reply

    I am not even sure why we are comparing this VA study with prior studies. At Natural Bio Health in Austin, Texas, we studied the report and found that the actual results, using the numbers provided in the VA study, showed that the veterans on testosterone had LESS RISK of heart disease, of death and of stroke. See our blog at

    The headlines in the LA Times commented on the VA study as follows: “The new research found that among 8,709 older men who were assessed for the possibility of blocked arteries, those taking testosterone were 30% more likely to suffer an adverse event — a stroke, a heart attack or death. Melissa Healy, LA Times, November 5, 2013, 1:24 p.m.”

    Since this study was contrary to hundreds of previous studies, we checked the math in this study and came to the correct conclusion based on the very numbers provided by the study. The study itself stated “In the three-year period they were followed, 25.7% of the men taking testosterone supplementation had suffered a stroke, a heart attack or death. Among the men who did not take testosterone, 19.9% suffered one of those outcomes. Of 7486 patients not receiving testosterone therapy, 681 died, 420 had MIs, and 486 had strokes. Among 1223 patients receiving testosterone therapy, 67 died, 23 had MIs, and 33 had strokes.”

    Do your own calculations and then ask how and why the VA study came up with their numbers. We come up with vastly different numbers.

    Using the data provided in the VA study, Dr. Joseph Feste of Natural Bio Health calculated that 21.19 % of the men who did not take testosterone had events while only 10% of the men who took testosterone had events. This is directly contrary to the VA study as published by JAMA. Dr. Feste has written to the editorial board of JAMA and asked them to correct their numbers. They responded as follows:
    “Thank you for contacting JAMA about the testosterone paper. We are working on correcting it right away.” Stacy L. Christiansen, MA, Managing Editor, JAMA

    My calculations (feel free to test them): Using the VA study numbers as shown in the chart below, I derived a different set of numbers. The numbers based on the data provided by the VA study are as follows:
    •NOT using testosterone, 9.1% died. Using testosterone, 5.4% of the men died.
    •NOT using testosterone, 5.6% had MIs. Using testosterone, 1.9% of the men had MIs.
    •NOT using testosterone, 6.5% has strokes. Using testosterone, 2.7% of the men had strokes.

    Total % of total Event/Classification
    7486 86% No testosterone therapy
    681 9.1% Men died
    420 5.6% Had MIs
    486 6.5% Had strokes

    1223 14% Testosterone therapy
    67 5.5% Men died
    23 1.9% Had MIs
    33 2.7% Had strokes
    Mike Clark, Ph.D., Natural Bio Health

    November 19, 2013 at 3:16 pm
  • Joseph R. Feste, MD Reply

    I have been exchanging emails with the managing editor of the J.A.M.A. finally being told that I would have to contact the author in order to have the data changed and submission of an explanation as to the discrepancies. I am surprised that the journal reviewers of the manuscript didn’t notice the errors in the data. I have reviewed dozens of manuscripts in the past and was obligated to make sure the data was correct and the proper conclusions were submitted. In this article there is obviously erroneous data resulting in conclusions that have been verified to be the opposite of the authors’. I have attempted to contact the author, Rebecca Vigen, but I was told that she is no longer at UT Southwestern in Dallas, Texas. I will continue to make an attempt to speak with her so I can get an explanation as to how they arrived at their conclusions.

    November 19, 2013 at 4:42 pm
  • EUGENE O’Neill, M.D. Reply

    In all my statistics courses, the opening remarks were” if you wish to see a poor use of statistical method…… a Medical Article!!!!!”
    Mark Twain said ” there are liars, dam liars, and then there are statiisticians in that order!!”
    This article is flawed by ignoring the very simple first order of business….the incidence of all morbities and mortality are far less in the supplemented group than they are untreated group……not even close and statistically close!!
    When one takes this into account, the rest of the ariticle is garbage any way you pitch it.
    And one wonders why the AMA and its journal is not revered in the medical community.
    Articles , so poorly done and then released to the iindisciplined mainstream media as endorsed by JAMA does our profession a disservice, not to mention setting back bonified research in the field of medical care that we aspire to legitimize.

    November 19, 2013 at 10:06 pm
  • Zoraida Navarro, MD Reply


    Just like white slave owners forbade that Blacks be educated and learn to read, it now appears that JAMA is trying to keep doctors stupid and ill informed all while protecting the interest of BIG PHARMA. No need to subscribe to conspiracy theories, the reasons are more than obvious, testosterone is cheaper than diabetic prescriptions, stroke meds, MRI’s etc. And healthy strong citizens are more likely to revolt and demand democracy and fight for it than sick, weak, scared people so BIG GOV doesn’t want that either.

    Thanks Neil for having the guts to say what the rest of us are too busy, scared and tired to point out.

    November 20, 2013 at 12:08 am
  • Lorena Hillman M.D. Reply

    My thanks to Drs. Clark and Feste for taking the time to post the calculations and correct percentage values. On my initial glance at this study, my very first impression was that the conclusions about adverse events did not seem consistent with the numbers put forth in the article. I made a mental note to myself to come back for more detailed reading and to do my own calculations. I am also shocked that such obvious errors were unnoticed by the authors and editors and actually went to print in JAMA! In light of the high media attention given to this article, JAMA readers and the general public deserve a timely response to Dr. Feste’s inquiry.

    November 20, 2013 at 4:27 am
  • Screven Edgerton M.D. Reply

    Thank you Neal and Worldlink. It is extremely helpful to have a statement from you when this so called “literature” makes a big media splash. There is no doubt my patients will be coming in with this article in hand.

    November 20, 2013 at 2:53 pm
  • Josh Trutt, MD Reply

    It seemed a bit too simple that they could have misread their own numbers that badly, so I wrote to the authors. They stand by their numbers, as they say it has to do with how Kaplan-Meier curves work:

    “If you’re familiar with how a Kaplan-Meier curve is estimated, each time that an event occurs, the probability of remaining event-free is calculated based on {100% minus [the number of events at that time divided by the number at risk of having an event at that time] x100%}. The number of men at risk in the Testosterone Therapy (TT) group is always much smaller at any given time than the number of men at risk in the group that never receives TT. So one event in the TT group leads to a smaller % remaining event-free compared to the % that would be calculated when one event occurs in the never-treated group. Within each of the TT and no TT group, the % are then multiplied over all of the event times that occur separately in each group to arrive at the overall Kaplan-Meier curve in each group. The result is a lower cumulative % remaining event-free in the TT group, or, conversely, a higher cumulative percentage of events in that group.
    Anna E. Barón, PhD | Professor
    University of Colorado Anschutz Medical Campus
    Biostatistics and Informatics | Colorado School of Public Health

    I think what they are saying is that the raw numbers don’t take into account the length of time each person was on testosterone, which differed:
    i.e. it is NOT true that everyone in the TT group used testosterone for, say one year. If that WERE the case, then the raw event numbers would be all you needed, and this study would show a better outcome for the TT group. But because some people had events despite not being on testosterone for very long, it seems that the Kaplan Meier curve bears out a small negative effect for TT.

    The larger point remains, which is that the patients in this study had quite a high incidence of heart disease– much higher than those in the similar JCEM study of 2011. There is also little follow up in these patients: did their hematocrits rise? Did their blood viscosity increase? In a patient population with over 50% heart disease, I would be watching that pretty carefully. I don’t think it’s necessary to disparage JAMA, as most likely their statisticians knew what they are doing (with what little data they were given). The TOM study also showed bad outcomes in elderly, debilitated patients. So, there may be a trend there: testosterone may be best used as a preventative, before men develop all the sequelae of androgen deficiency. Certainly, there are many positive studies using testosterone specifically in patients with CAD, but the longest follow up I have seen in those studies was one year. I review some of those in my comment on this article on my blog:

    Overall this study reinforces the fact that testosterone should be prescribed not as “injectable Viagra” with no follow-up, but rather in conjunction with appropriate diet and exercise advice, by doctors committed to providing proper dosing guidelines and follow-up. Properly prescribed, there is little doubt as to the benefits in preventing metabolic syndrome and maintaining a healthy vigorous lifestyle with advancing age.

    November 22, 2013 at 1:00 am
    • neal rouzier Reply

      Nice review Josh but unfortunately it is water under the bridge as the harm is already done and the public will view BHRT with a black eye. With time it will be forgotten as soon as the next positive study is published. Oddly many patients contacted me and I sent them my response. So far no one has stopped their testosterone or wanted to stop. They just wanted my blessing that it was OK to continue as they all voiced that they would not stop no matter what the study showed. I wonder what that says?

      November 24, 2013 at 2:51 pm
  • MachineGhost Reply

    The Kaplan-Meier curve estimates, by design, don’t take into account the levels of the testosterone replaced. Everyone seems to be making an assumption that the levels were replaced to physiologically optimal levels as would be expected by any competent practitioner or informed patient. They were not and were still sub-optimal! So either the sub-optimal levels showed there WAS a protective effect in the raw data or it was NOT strong enough to overcome the Kaplan-Meier curve estimator. You decide.

    November 23, 2013 at 12:58 pm
  • Lorena Hillman M.D. Reply

    Yes, after I posted above I also concluded that such obvious errors couldn’t possibly have been published with so many collaborators. Josh, I greatly appreciate your efforts in obtaining clarification about the Kaplan-Meier methodology as well as Neal’s feedback on his patients’ responses. It would indeed be interesting if a duplicate study were performed with all TT treated patients reaching optimal levels and having appropriate management and follow up. However, I am still mulling over how to have a lucid discussion about the study with a couple of new patients(age 55/59) with hx of prior MI…

    December 5, 2013 at 7:43 am

Leave a comment

Your email address will not be published.