Testosterone treatment and risk of venous thromboembolism: population based case-control studyBMJ 2016; 355 doi: https://doi.org/10.1136/bmj.i5968 (Published 30 November 2016) Cite this as: BMJ 2016;355:i5968
All rapid responses
We read with interest the recent case-control study by Martinez et al, comparing rates of testosterone therapy between men with diagnoses of venous thromboembolism (VTE) and matched controls.1 The authors found significantly higher testosterone use among men with venous thromboembolism, largely driven by initial six months of use. These findings will add to the contentious topic of testosterone treatment – a therapy, which has ballooned in use over recent decades for “lifestyle purposes” despite considerable controversy about potential risks. With that said, we do have some comments about this study and its methods.
Most importantly, the authors mention the possibility of “differential identification”—that men using testosterone could have more opportunities to be diagnosed with venous thromboembolism but then state that this is “unlikely to be the case and is unlikely to alter the study findings.” We would argue that men pursuing testosterone therapy for lifestyle purposes could easily have increased opportunities for identification of venous thromboembolism. For example, men who seek out testosterone therapy may visit physicians more often, and may be more likely to seek out diagnostic tests or interventions for symptoms, which could be caused by venous thromboembolism. Perhaps controlling for number of physician visits would allow this to be addressed—a feasible statistical approach. Perhaps further exacerbating this, the study years encompass a period where a link between VTE and testosterone was suspected. Thus, conscientious physicians treating men using testosterone replacement could be more likely to order diagnostic tests or examinations that would identify blood clots.
Second, the authors point out that the rate ratio for men with pathological hypogonadism was not significant for VTE. It is only among those without hypogonadism that an elevated rate ratio was seen. It is not immediately clear why this would be the case, perhaps men with hypogonadism had more physiologic androgen levels while undergoing testosterone treatment? Without information on serum androgen levels it is difficult to say whether this is the case. Presumably the men in this study with and without encompass a wide range of hypogonadal and eugonadal states –information on serum androgen levels both pre and post treatment would help physicians counseling patients on whether to use these medications.
Regardless, professional guidelines agree that it is generally men with hypogonadism who should be receiving testosterone therapy.2 Fewer clinicians would argue that eugonadal men should be getting testosterone therapy –yet it is only among these men that increased rate ratios were seen.
Finally, there are well known limitations in using administrative data for identification of venous thromboembolism.3 4 The authors address this by using a previously validated tool for identifying venous thromboembolism, which appears to have good accuracy within the British database they employed.5 However, the accuracy for diagnoses of hypogonadism and for accurately capturing testosterone prescriptions is less clear. Given the common use of mail order and compounding pharmacies for testosterone replacement could some testosterone prescriptions have been missed from these data? Further, given the complexity involved in appropriately diagnosing hypogonadism, these pathological states could easily be over or under-diagnosed using administrative measures.
Ultimately, this study adds to the vigorous debate on the risks of testosterone therapy. There is no question that testosterone use probably exceeds what is medically necessary and may come with significance harms. However, we would argue that the above issues might limit the relevance of this study’s results for men with true hypogonadism, considering testosterone therapy. Ultimately, prospective, population-based cohorts of men using testosterone therapy will be vital for assessing the absolute and relative risk of VTE and may provide the key information for men deciding whether to start testosterone therapy.
1. Martinez C, Suissa S, Rietbrock S, et al. Testosterone treatment and risk of venous thromboembolism: population based case-control study. Bmj 2016;355:i5968.
2. Seftel AD, Kathrins M, Niederberger C. Critical Update of the 2010 Endocrine Society Clinical Practice Guidelines for Male Hypogonadism: A Systematic Analysis. Mayo Clinic proceedings 2015;90(8):1104-15.
3. Lawson EH, Louie R, Zingmond DS, et al. A comparison of clinical registry versus administrative claims data for reporting of 30-day surgical complications. Annals of surgery 2012;256(6):973-81.
4. Lawson EH, Zingmond DS, Hall BL, et al. Comparison between clinical registry and medicare claims data on the classification of hospital quality of surgical care. Annals of surgery 2015;261(2):290-6.
5. Martinez C, Cohen AT, Bamber L, et al. Epidemiology of first and recurrent venous thromboembolism: a population-based cohort study in patients without active cancer. Thromb Haemost 2014;112(2):255-63.
Competing interests: No competing interests
It is disappointing to see that once again researchers have failed to realise that hypogonadism confers a well-established additional risk and that will be greater in the first 6 months or until testosterone (T) levels have been normalised for a sustained period.
Co-morbidities that cause DVT also cause low T, and we don’t know if the risk factors or the low T were adequately treated.
From 2001 onwards testosterone treatment would probably have been with Sustanon with all the known problems, or even worse, oral agents. There is no mention of the form of treatment or whether replacement was adequate or possibly over dosed, as can often occur with previously used short acting agents.
The database was set up in 1989, so how accurate was DVT diagnosis in 2001? when the investigators state that 2 DVTs in 18,000 would have made the difference, and path links (re the biochemistry etc) to the GP systems were not available until 2005 in this database.
What is “pathological primary and secondary hypogonadism” - how do you diagnose this from retrospective GP data with without mentioning any blood levels or with no mention of clinical symptoms?
Hypogonadism and ED increase risk and until the patient is stabilised and they will be at greater risk until well established on therapy (as is the case in the early treatment of thyroid deficiency).
The number of events is very small, 11 in 19,000 men, and as stated, this is too small for subgroup analysis, so why do it? The conclusion of this study should surely have been that the rate of VTE in this group of patients was very small, and that Testosterone Replacement Therapy was not associated with increased risk.
It also seems that having taken testosterone reduces risk but how do you become an ex-user to achieve this reduced risk without having taken the testosterone for an initial 3-6 months.
We are also asked to believe that 64 yrs old men with hypogonadism had a sexual dysfunction rate (does this mean erectile dysfunction?) of 14% and the authors supposedly adjusted for this - this is nonsense - the real rate of ED alone in such men of 64yrs would be 80% and low T and ED independently increase risk.
The statement below explains how to make the data fit the conclusion the investigators had decided on in the first place:
Page 7: “Although the rate ratios were not statistically different, the difference in magnitude of effect may be clinically relevant.” We find it strange that, with large numbers of people involved – although the number of events is very low – they do not see statistical significance – evidence based medicine!!
The conclusion was that there was no significance but then comes a statement that it might be clinically significant. This is making false assumptions from the data and will cause unwarranted anxiety and concern. It is disappointing that the BMJ reviewers allowed inappropriate conclusions to be reached. The increasing trend for journals to publish falsely negative findings, particularly related to testosterone, is elegantly discussed in a recent paper by Traish et al.
The following papers provide stronger evidence of long term safety:
Baillargeon J, Urban RJ, Morgentaler A, Glueck CJ, Baillargeon G, Sharma G, Kuo YF (2015) Risk of venous thromboembolism in men receiving testosterone therapy. Mayo Clin Proc 90: 1038–1045
Sharma R, Oni OA, Chen G, Sharma M, Dawn B, Sharma R, Parashara D, Savin VJ, Barua RS, Gupta K (2016) Association between testosterone replacement therapy and the incidence of deep vein thrombosis and pulmonary embolism: A retrospective cohort study of the Veterans Administration Database. Chest 150: 563–571
Wallis CJD, Kirk L. Yuna M. et al. Survival and cardiovascular events in men treated with testosterone replacement therapy: an intention-to-treat observational cohort study. Lancet Diabetes Endocrinol. 2016 Jun;4(6):498-506. doi: 10.1016/S2213-8587(16)00112-1. Epub 2016 May 7
Traish AM, Vance JC, Morgentaler A, Overselling Hysteria Dangerously. EMBO reports, Published online 5th December 2016.
Competing interests: MK has received funding for research, conference attendance, lecturing and advice from the pharmaceutical industry including Astellas, Pfizer, Takeda, Bayer, MSD, BI, Lilly, GSK, AZ, Ferring and Menarini. Editor PCCJ. Also on several NHS advisory boards including the Prostate cancer Risk Management Programme and the Prostate Cancer Advisory Group.
It is clinically demonstrated and widely accepted that low androgen levels/testosterone are correlated to low libido, decreased arousal, orgasmic dysfunction in postmenopausal women and in surgically ovariectomized younger women. 
Furthermore, testosterone replacement therapy proved effective and successful.
The North American Menopause Society and the International Society for the Study of Women's Sexual Health (ISSWSH) officially suggest off-label testosterone therapy in postmenopausal women suffering from these issues. 
Meanwhile, there exist widespread baseline sexual problems in the youth population, in many Countries investigated.
In this research study, published in the Journal of Adolescent Health, 59.2% of young women reported persistent orgasmic dysfunction, 48.3% low satisfaction, sexual problems that did not fluctuate over time (study period of two years), meaning they were established diseases. 
Young Japanese women report sexual aversion in 42.4% of cases. 
This trend leads to complete sexual avoidance. 
All those young and older distressed women are likely to try off-label testosterone treatments at some point in their lives.
As a consequence, I wrote that they too are at risk of associated thromboembolic events.
Retrieving the attached references you requested on PubMed only took 10 minutes.
 Testosterone correlation/efficacy in female orgasmic dysfunction.
 The North American Menopause Society suggests testosterone therapy in postmenopausal women.
 The International Society for the Study of Women's Sexual Health (ISSWSH) also suggests testosterone therapy in postmenopausal women.
Competing interests: No competing interests
After submitting my comment as a rapid response considerable additional hours were spent reviewing the peer-reviewed literature on the topics of exogenous testosterone i.e. testosterone replacement therapy or TRT, as well as on related issues of aromatization of testosterone to estradiol in men receiving TRT.
This issue was initially brought to my attention decades ago when a treatment for prostate cancer involved the use of single-agent anti-androgen (e.g., bicalutamide, eulexin, or nilutamide). In patients treated in this fashion it is common to see reflex increases in testosterone with aromatization to estradiol and the development of significant gynecomastia. Additionally, elevations of estrogen in these patients reflexly increase prolactin. Simply measuring estradiol levels at baseline and during TRT would identify such men at risk & allow for aromatase inhibitor therapy and/or use of a prolactin inhibitor. Rarely, have I seen fellow physicians monitor patients in this manner. There are hundreds of peer-reviewed publications on the effects of both estrogen and prolactin on increasing platelet activation, various clotting factors, as well as decreasing anti-thrombin III (ATIII) in various clinical settings involving both men and women (1-4). Of course, additional key issues relate to the dose-response nature when using TRT, and even something as basic as monitoring serum total and free testosterone using the more accurate laboratory methodologies.
It is therefore crucial in any publication on TRT, such as that of Martinez et al, to make reference to the associations between T, estradiol, and prolactin. Most importantly, the authors should indicate a limitation in their study if the use of aromatase inhibitors, the measurements of estradiol and prolactin, and an assessment of hypercoagulability have not been done. If this is not the case, readers will not fully understand the pros and the cons of the use of TRT. What I have found commonplace in many areas of medical therapeutics is the lack in baseline studies, follow-up monitoring and employing biological "work arounds" to minimize toxicity and enhance the therapeutic index, (i.e., maximize the benefits, minimize the adverse side-effects).
On an entirely different note, the first rapid response to the Martinez et al article written by Dr. Saripanidis commented: "Postmenopausal women who are arbitrarily prescribed and administered testosterone for low libido/orgasmic dysfunction issues, are also at increased risk of venous thromboembolism. Oﬀ-label testosterone prescriptions are a dangerous trend in more older women than older men, unfortunately."
I have spent the last few hours on PubMed http://www.ncbi.nlm.nih.gov/pubmed and via Google Scholar looking for references to confirm this important statement, but cannot find such a citation(s). I believe that any input to a medical journal should have associated literature citations & welcome comments with citations by other readers relating to all relevant issues. "We get too soon old, and too late smart."
1. Wallaschofski H, Lohmann T, Hild E, et al: Enhanced platelet activation by prolactin in patients with ischemic stroke. Thromb Haemost 96:38-44, 2006.
2. Raaz D, Wallaschofski H, Stumpf C, et al: Increased prolactin in acute coronary syndromes as putative Co-activator of ADP-stimulated P-selectin expression. Horm Metab Res 38:767-72, 2006.
3. Haring R, Friedrich N, Volzke H, et al: Positive association of serum prolactin concentrations with all-cause and cardiovascular mortality. Eur Heart J 35:1215-21, 2014.
4. Norris LA, Brosnan J, Bonnar J, et al: Inhibitors and activation markers of the haemostatic system during hormone therapy: a comparative study of oral estradiol (2 mg)/ dydrogesterone and estradiol (2 mg)/ trimegestone. Thromb Haemost 100:253-60, 2008.
Competing interests: No competing interests
The authors have made a valuable contribution in alerting us to the increased risk of VTE (venous thromboembolism) associated with the use of TRT (testosterone replacement therapy), and especially in the first 6 months of such treatment.1 As a medical oncologist specializing in prostate cancer since 1983 & involved with the hormonal aspects of this disease, I have found that the vast majority of physicians who employ TRT do not understand the endocrinology involved in TRT.
Baseline assessments of gonadal function using both free (fT) and total testosterone (T) are often not done. Interval monitoring of not only these levels but of associated metabolites & related endocrine values are rarely done. If TRT results in significant elevations of T then it is common for T to be metabolized to estradiol via aromatization (T → E2).2 It would not be unexpected to see an effect of elevated levels of E2 to occur especially in the first months of TRT. Increased estrogen levels induce hypercoagulability by causing anti-thrombin III (ATIII) deficiency. Many physicians believe that aspirin or non-steroidal drugs will lessen the risk of VTE in such an environment but these are anti-platelet agents and do not impact ATIII deficient states. Estrogen also elevate Factors VII & X and other etiologic factors related to the increased risk of VTE.
There is a dose relationship to the adverse side-effects of estrogens relating to VTE and it would also be expected that the dose of TRT & the associated levels of T & fT would be relevant to such side effects. I would anticipate the same to be true as to the effect of TRT on the stimulation of erythropoiesis with resultant increase in hematocrit. Studies such as the one by Martinez et al do not evaluate the specific medical records of the patients under study, or at the very least I cannot find evidence that indicated that levels of T, fT, E2, CBC or other relevant lab test results.
Lastly, elevations in estrogen routinely elevate prolactin levels and the latter also is associated with hypercoagulability and risk of ischemic stroke, coronary syndromes via enhanced platelet aggregation. 3
1. Martinez C, Suissa S, Rietbrock S, et al: Testosterone treatment and risk of venous thromboembolism: population based case-control study. BMJ 355:i5968, 2016.
2. Oettel M: Testosterone metabolism, dose-response relationships and receptor polymorphisms: selected pharmacological/toxicological considerations on benefits versus risks of testosterone therapy in men. Aging Male 6:230-56, 2003.
3. Wallaschofski H, Lohmann T, Hild E, et al: Enhanced platelet activation by prolactin in patients with ischemic stroke. Thromb Haemost 96:38-44, 2006.
Competing interests: No competing interests
Postmenopausal women who are arbitrarily prescribed and administered testosterone for low libido/orgasmic dysfunction issues, are also at increased risk of venous thromboembolism.
Off label testosterone prescriptions are a dangerous trend in more older women than older men, unfortunately.
Competing interests: No competing interests