Just as I suspected, it didn't take long. The AHJ article entitled A cardiovascular safety study of LibiGel (testosterone gel) in postmenopausal women with elevated cardiovascular risk and hypoactive sexual desire disorder has been taken down from free full text availability.
Maybe the following will generate a little transparency. As a long time Biosante shareholder who helped fund the research for the article, I feel obligated to share it with other ANI and former Biosante shareholders.
A cardiovascular safety study of LibiGel (testosterone gel) in postmenopausal women with elevated cardiovascular risk and hypoactive sexual desire disorder
•William B. White, MD, Deborah Grady, MD, MPH, Linda C. Giudice, MD, PhD, Scott M. Berry, PhD, Joanne Zborowski, RN, Michael C. Snabes, MD, PhD,
Received 28 June 2011; accepted 26 September 2011. published online 23 November 2011.
Article Outline • Abstract • Background and study rationale • Study design and conduct o General study conduct • Study treatment and procedures o Treatment o Study procedures o Breast and endometrial safety examinations • Statistical considerations o Adaptive design for determination of sample size ? Sample size estimation o Study organization and oversight • Study progress and baseline characteristics • Discussion • Conclusions • Disclosures • References • Copyright
Evaluation of the safety of hormonal preparations for the treatment of female sexual dysfunction is important to assess the benefit-to-risk profile of these drugs and has been strongly encouraged by the Food and Drug Administration. LibiGel (Biosante Pharmaceuticals, Inc., Lincolnshire, IL), a low-dose testosterone gel, is under development for the treatment of hypoactive sexual desire disorder (HSDD) in oophorectomized women. To evaluate the long-term effects of LibiGel on risk for cardiovascular (CV) events, breast cancer, and general safety, a randomized, placebo-controlled clinical study using a novel adaptive design to optimize sample size and power is being conducted. The primary end point of the BioSante LibiGel Safety Study (BLISS) is a composite of CV events including death, nonfatal myocardial infarction, nonfatal stroke, coronary revascularization, hospitalized unstable angina, and venous thromboembolic events. Breast cancer is a coprimary end point. Postmenopausal women (both surgically and naturally) with HSDD and increased risk for CV events will be followed up for up to 5 years postrandomization with an interim data analysis for regulatory approval after the last woman enrolled has been on therapy for at least 12 months. Determination of the number of subjects to enroll is based on an adaptive design that uses interim data to estimate the predictive probability of study success. In agreement with the Food and Drug Administration, LibiGel will be declared safe if the upper limit of the 97.2% CI of the hazard ratio is =2.0 or the upper bound of the 97.2% CI for the absolute difference between CV event rates per 100 person-years is =1% and the observed hazard ratio is =2.0. The BLISS study will define the CV safety profile of low-dose testosterone therapy in the formulation of LibiGel for postmenopausal women with HSDD, and the trial design may provide a paradigm for studies that aim to document long-term safety when the proposed outcome under study is an uncommon adverse event.
Background and study rationale
Female sexual dysfunction is a common clinical problem. The most common form of female sexual dysfunction is hypoactive sexual desire disorder (HSDD), a syndrome of low sexual desire causing personal distress. It is estimated that >40% of women report some type of sexual dysfunction and the prevalence of HSDD is approximately 10% to 12%.1, 2, 3 Currently, there are no pharmaceutical treatments for HSDD that are approved by the US Food and Drug Administration (FDA).
Low-dose testosterone has been under development for the treatment of HSDD for several years and has been shown to be effective.4, 5 Libigel (Biosante Pharmaceuticals, Inc, Lincolnshire, IL) is a 1% testosterone hydroalcoholic gel, a pea size amount of which is applied to a small area of the skin on the upper arm and when applied once daily increases the serum free testosterone in postmenopausal women into the range of younger premenopausal women. Currently, 2 pivotal phase III randomized, placebo-controlled trials are ongoing to evaluate the efficacy and short-term safety of LibiGel in women with HSDD.
After a review of transdermal testosterone, the Division of Reproduction and Urologic Products of the FDA6 requested that long-term effects on risk for cardiovascular (CV) events and breast cancer be evaluated before approval of testosterone for treatment of postmenopausal women with HSDD.6 To address this request and to better understand the long-term risks of low-dose testosterone therapy in women, we developed the BioSante LibiGel Safety Study (BLISS) to evaluate CV and breast safety in postmenopausal women with HSDD.
Study design and conduct
BLISS is a randomized, double-blind, placebo-controlled, parallel group study comparing LibiGel (0.22 g/d of testosterone) and identical placebo gel in postmenopausal women with HSDD and increased CV risk. The primary objective is to demonstrate that the CV risk associated with LibiGel is not higher than placebo. The primary outcome of the study is a composite of adjudicated CV events including CV death, nonfatal myocardial infarction, nonfatal stroke, coronary revascularization, hospitalized unstable angina, and venous thromboembolic events (deep venous thrombosis or pulmonary embolism). Incidence of invasive breast cancer is a coprimary outcome. The key inclusion and exclusion criteria for the trial are outlined in Table I, Table II.
Table I. BLISS inclusion criteria
1. Postmenopausal women at least 50 y of age with a clinical diagnosis of HSDD (DSM-IV) 2. CV disease risk score if =2(adapted from Barett-Connor et al7) using the following point scale:
a. Age, 60 to <70 y (1 point) or =70 y (2 points) b. Diabetes mellitus (2 points) c. Peripheral vascular disease with ankle-brachial index <0.6 (2 points) d. Documented CV disease (myocardial infarction, stroke, hospitalization for unstable angina/acute coronary syndrome, revascularization of the coronary or peripheral circulations) (2 points) e. Present smoker of at least 10 cigarettes per day (or the equivalent) (1 point) f. Hypertension—defined as seated systolic blood pressure =150 mm Hg and/or diastolic blood pressure =95 mm Hg and/or taking antihypertensive medications (for treatment of hypertension)—(1 point) g. Dyslipidemia—low-density lipoprotein >160 mg/dL and/or high-density liproprotein <45 mg/dL with triglycerides >250 mg/dL and/or taking prescribed lipid-lowering medication (1 point) DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition.
Table II. Exclusion criteria of the BLISS trial
1. Treatment with systemic antiandrogens, tamoxifen, or other selective estrogen receptor modulators (treatment and prevention of osteoporosis therapies are acceptable) 2. Use of androgen therapy within 2 m of randomization 3. History of estrogen-dependent neoplasia or any gynecologic cancer 4. History of melanoma of any stage at any time or a history of cancer of any other type within 10 y of randomization (exception is treated basal cell carcinoma of the skin) 5. Screening mammogram with any finding or reason that requires follow-up within 12 m of randomization 6. Acute myocardial infarction, stroke, hospitalization for unstable angina/acute coronary syndrome, or revascularization of the coronary or peripheral circulations within 6 m of randomization 7. Major psychiatric illness 8. Patients who have a life expectancy of <3 y in the judgment of the study physician or who have any condition that, in the opinion of the study physician, would not allow safe participation in the trial (eg, drug addiction and serious hepatic or renal disorders) 9. Patients who have been on an investigational drug in the past 30 d
General study conduct
The study includes a screening period of up to 8 weeks and a 60-month treatment period (Figure 1). The BLISS study complies with the Declaration of Helsinki and subsequent revisions and follows Good Clinical Practice guidelines. Each of the investigative sites is in the United States and Canada, and each obtained approval by an institutional review board, and study participants must review and sign informed consent before any study-related procedure. The study is registered on clinicaltrials.gov (NCT00612742). • o View Large Image o Download to PowerPoint • Figure 1. Study design, visits, and procedures of BLISS.
Study treatment and procedures
Treatment
Participants are randomized in a 1:1 ratio to receive either 1% testosterone gel 0.22 g/d (LibiGel) or an equivalent weight of identical placebo gel daily. Randomization is stratified by use of postmenopausal hormone therapy (none, estrogen, or estrogen + progestogen).
Study procedures
There are 9 clinical visits and 14 telephone contacts during the course of the study (Figure 1). Office visits will occur at screening and randomization and at 3, 6, and 12 months postrandomization and yearly thereafter. Telephone contacts will be completed at week 6, month 9, and at 3-month intervals unless a clinical visit is required. Subjects are instructed to contact sites if they believe that they have experienced a CV event or breast cancer. At each scheduled contact, participants are asked about adverse events and health care and hospitalizations; questioned about specific androgenic side effects, potential CV events, and breast cancer; and undergo other assessments of safety and tolerability as well as laboratory measures.
Efficacy will be evaluated at office visits and telephone contacts using the Subject Global Assessment and Perception of Benefit Questions.5 Participants who discontinue study drug will be encouraged to complete all study visits, examinations, and questionnaires and to report potential CV events and breast cancer.
Breast and endometrial safety examinations
Breast examination and mammography will be performed at baseline and annually. Endometrial biopsy will be performed at baseline and at study years 1, 2, and 5. Participants with vaginal bleeding will undergo uterine evaluation that includes an endometrial biopsy. Study medication will be discontinued in participants with evidence of uterine cancer, complex or adenomatous hyperplasia, or atypia.
Statistical considerations
The primary safety outcome is time to the first occurrence of a CV event. The coprimary safety outcome is time to first incident of invasive breast cancer. In addition, each component of the composite CV outcome, all-cause mortality, as well as breast carcinoma in situ will be evaluated as secondary end points. The primary analysis is Bayesian and will use the posterior distribution for the hazard ratio (HR) and its associated CI and the posterior distribution for the absolute risk difference to compare the CV event rates among women randomized to LibiGel with those randomized to placebo. LibiGel will be declared noninferior to placebo if the upper limit of the 2-sided 97.2% CI of the HR is =2.0 or the upper bound of the 2-sided 97.2% CI for the absolute difference between CV event rates per 100 person-years is =1% and the observed HR is =2.0. The use of a 2-sided 97.2% CI rather than the standard 95% CI compensates for the interim analyses conducted for sample size selection and success being either the HR or the absolute difference.
Adaptive design for determination of sample size
To calculate sample size, we assumed that the CV event rate in the BLISS placebo group would be 2% per year and that the event rates in each group would be equivalent.
The estimate of the CV event rate was based on age-specific rates observed in the Women's Health Initiative Randomized Trial8, 9 and an increased CV risk based on the enrollment criteria requiring established CV risk factors (Table I).
The BLISS trial uses an adaptive “Goldilocks” design10 to determine sample size. After enrollment of 2,500 participants, the posterior distributions of the HR and absolute risk difference are calculated. The final analysis for trial success does not occur until 12 months of follow-up for each subject; therefore, an adaptive decision is addressed for the possibility of stopping enrolling for expected success or for futility using the predictive probability for study success with full follow-up.
The design is completely prospectively created, including the Bayesian modeling, and is conducted by an independent unblinded statistician. If the predictive probability of trial success after 12 additional months of follow-up is a minimum of 0.90, enrollment will cease and participants enrolled at the time recruitment ends will be followed up for a minimum of 1 additional year before the primary analysis of data for regulatory submission. Hence, all randomized patients will be followed up for at least 1 year until the one primary analysis for noninferiority of treatment versus placebo is conducted. After analysis for regulatory submission, patients will be followed up in the trial for a total of 5 years each; at the end of this period, the main analyses will be repeated.
At each designated analysis of the predictive probability of success, the numbers of CV events and time at risk in the treatment and placebo groups are observed. Assuming exponential times to event, vague gamma priors are applied to the unknown event rates. Based on the posterior distributions, the predicted distributions of the time to CV events can be determined for all patients currently in the study and estimated for possibly newly accrued subjects who will be followed up for an additional year after recruitment ends. The Bayesian predictive probability has the effect of integrating over the uncertainty present in the current distribution of CV events in each treatment group and the uncertainty and randomness of future events. At each interim analysis, the Bayesian predictive probability of study success and the predictive probability of study success at the maximum sample size of 4,000 are calculated.
If the predictive probability for study success with an additional year of follow-up with the current sample size is at least 90%, then recruitment will end for expected success. If the predictive probability of success at the maximum sample size (4000) is small, then the study is stopped for futility. If the chance for study success is neither high nor low, enrollment will continue and another interim analysis will be performed after an additional 2 CV events are observed. Sample size interim analyses will continue after every 2 CV events are observed until the predictive probability indicates success or futility or the maximum number of 4,000 participants is reached. In this latter case, all subjects will be followed up for 1 year, and the primary analysis for success will be conducted.
Sample size estimation
Based on a CV event rate of approximately 0.020 events per subject-year and 300 patients enrolled per month, the expected sample size in this study is 3,011 with an SD of 773. Based on study assumptions and equality of treatment arms, simulations suggest that there is a 75% chance that study enrollment will stop early for expected success, 17% chance that the study will enroll 4,000 patients, and an 8% chance that the study will stop early for futility. The resulting power to determine noninferiority is >90%.
Study organization and oversight
The BLISS study has an executive committee that has the overall responsibility for oversight of study conduct, modifications or revisions to the study protocol, and oversight of public presentations or publications of the study findings. Operations for the study are coordinated by the sponsor's product development department with a contracted clinical research organization. All serious CV events are reported to the operations group of the sponsor and forwarded to a CV end points committee (CEC). Members of the CEC review all serious events and potential CV events and adjudicate CV outcomes. Only adjudicated CV events are included in the analyses for adaptive estimation of sample size and the composite end point.
Breast adverse events, including breast cancer and carcinoma in situ, will be reviewed by a breast end points committee (BEC). The committee will review histology slides, breast biopsy pathology reports, and the clinical case. The diagnosis of the initial local pathologist will not be changed; the committee will provide general oversight and advice regarding breast safety. Patients diagnosed with invasive breast cancer or ductal carcinoma in situ by local pathologists will discontinue study drug and remain in the posttreatment follow-up phase of the study.
Findings of the CEC and BEC will be conveyed by an independent statistician to the data monitoring committee (DMC). The DMC is responsible for safeguarding the interests of study patients, assessing the safety and efficacy of the interventions during the study, and monitoring the overall study conduct. Based on regular reviews of all pertinent study data, including CV and breast cancer events, adverse events, and laboratory data, the DMC will provide recommendations to the executive committee and to the sponsor regarding continuing, stopping, or changing the study. The executive committee will be responsible for promptly reviewing DMC recommendations, advising the sponsor about whether to continue or terminate the study and whether amendments to the protocol or changes in study conduct should be implemented. The various committees for BLISS are listed in Table III.
Table III. BLISS executive committee, DMC, and CEC members
Executive committee: William B White, MD, Farmington, CT (chair); Linda Giudice, MD, PhD, San Francisco, CA; Deborah Grady, MD, San Francisco, CA Michael Snabes, MD, PhD, Lincolnshire, IL (ex officio) DMC: Robert Makuch, PhD, New Haven, CT (chair); Henry R Black, MD, New York, NY; Ronald S Swerdloff, MD, PhD, Los Angeles, CA; Janet Wittes, PhD, Washington, DC (ex officio) CEC: David Waters, MD, San Francisco, CA (chair); Karen Furie, MD, Boston, MA; Priscilla Hsue, MD, San Francisco, CA BEC: Leslie Laufmann, MD, Columbus, OH; Yunn-Yi Chen, MD, San Francisco, CA
Independent statistician: Jason Connor, PhD, Orlando, FL As of May 1, 2011.
Study progress and baseline characteristics
The BLISS study was initiated in early 2008, and >150 research sites experienced in women's health studies in the United States and Canada have been enrolling patients. All investigative personnel were trained at formal 2-day investigator meetings, site initiation visits, and continuous education on study conduct via clinical monitoring visits, monthly newsletters, and sponsor letters. Patients have been recruited from databases of the investigative sites, referrals, local advertising campaigns, and sponsor-supported, national advertising including Web-based and national television campaigns.
More than 2,800 subjects have been randomized into the study, and recruitment is ongoing as of April 2011. Baseline characteristics of the first 2,889 participants are shown in Table IV. The mean age of patients at enrollment is 58 ± 6 years, and a high proportion have multiple risk factors for CV disease, reflecting an older population with higher CV risk compared with previous studies of low-dose testosterone for HSDD.5, 11 As expected, women in the BLISS study would be at the higher end of the CV risk continuum of those who will be taking the drug for low sexual desire but is an appropriate population to evaluate the CV safety of low-dose testosterone.
Table IV. Baseline characteristics of the study population in BLISS (n = 2889) Age (y) Mean ± SD (range) 58.6 ± 5.6 (45-83) Race/ethnicity, n (%) White 2356 (81.6%) Black 393 (13.6%) Hispanic 89 (3.1%) Asian/Pacific Islander 20 (0.7%) Native American 17 (0.6%) Other/unknown 14 (0.5%) Menopausal status Time since menopause (y) 14.6 ± 9.6 Natural occurrence , n (%) 1103 (54.9) Surgically induced, n (%) 901 (44.8) Unknown, n (%) 5 (0.2) Risk factors for CV points, n (%) Age 60 to 69 y (1 point) 987 (34.2) Age =70 y (2 points) 123 (4.3) Diabetes mellitus (2 points) 608 (21.0) Current smoker (1 point) 569 (19.7) Hypertension (1 point) 1962 (67.9) Dyslipidemia (1 point) 1870 (64.7) Ankle-brachial index <0.6 (2 points) 2 (0.1%) Documented CV disease (2 points) 200 (6.9%) Concomitant medications, n (%) Antihypertensives 1071 (53.3%) Aspirin 522 (26.0%) Cholesterol-lowering agents 1068 (53.2%) Estrogen alone 324 (16.1%) Estrogen/progestins 60 (3.0%) Characteristics of patients randomized between 2008 and second quarter 2011.
Discussion
The BLISS study is the first large, long-term study of the safety of testosterone in postmenopausal women with HSDD. The dose of testosterone chosen for the study, 0.22 g/d, is approximately equal to the ovarian production rate in younger women with normal menstrual cycles.12 Although the tolerability and short-term safety of physiologic doses of testosterone in women have been acceptable for clinical use,5, 11 long-term safety has not been established. The primary safety outcome of BLISS will determine the effect of LibiGel on a broad composite of CV and thromboembolic events. Although the intended treatment group for this product is oophorectomized postmenopausal women aged 30 to 65 years with HSDD, the BLISS study population, of both oophorectomized and naturally postmenopausal women, has been enriched with participants at elevated CV risk (Table IV).
Despite the lack of evidence suggesting that treating women with low-dose testosterone increases risk of CV disease or breast cancer, a study such as BLISS is necessary to address several key issues that are debated in the literature. At this time, there are sparse data in women or men concerning the long-term safety of testosterone treatment. Testosterone is an obligate precursor of estradiol, a hormone that is known to stimulate breast tissue. However, rates of breast cancer are much lower in men than women,13 and testosterone treatment in women blocks estradiol-induced mitosis in breast tissue.14 There is no clear, direct mechanism by which testosterone may increase CV risk in women.15 However, women with polycystic ovarian syndrome have ovarian dysfunction and relatively high serum testosterone levels; these women are at increased risk for obesity, hypertension, and insulin resistance suggesting a link to increased CV risk.16 In contrast, some studies have shown that testosterone may increase insulin sensitivity.17, 18 In addition, men with low testosterone are more likely to have CV disease than men with normal levels of testosterone.19, 20, 21 The results of the BLISS study will address key clinical questions regarding the impact of replacement doses of testosterone on risk for CV disease and breast cancer in postmenopausal women.
Based on safety issues that have emerged with other new drugs for a variety of therapeutic indications, scrutiny surrounding the extent of exposure before regulatory approval has increased.22, 23, 24 However, it is difficult to demonstrate safety when adverse events of interest (such as CV events and cancer) are uncommon in the population to be treated with the new drug.6, 22
The standard approach to establishing safety in a placebo-controlled study is to determine an HR with an adequate number of adverse events to yield CIs narrow enough to provide assurance to regulators, practitioners, and patients that a certain level of harm can be “ruled out.” However, unlike major studies of proposed treatments for CV disease that typically enroll very high-risk patients with event rates of 3% to 8% per year,25, 26 the CV event rate in populations targeted for treatment of symptoms may never be high enough to achieve high levels of confidence using standard study designs. Creative proposals22 include ruling out a certain magnitude of risk in participants with relatively higher CV risk with extrapolation to lower-risk patients. The BLISS study is enrolling participants with moderately elevated CV risk because it would be difficult to study only women with high CV risk by virtue of age, risk factors, medical history (such as a previous CV event), who would not be the intended treatment population for low-dose testosterone to treat HSDD. In the case of BLISS, the use of a risk difference (attributable risk) statistic to rule out an upper bound of the CI of =1% is planned in the event that noninferiority is impossible to ascertain with standard HRs and hypothesis testing.
There are additional aspects of the BLISS study that make it unique among recent women's health studies. First, the study not only is attempting to rule out a preapproval level of risk after an average of 2 years (range, 1-4 years) of treatment but also includes a postapproval commitment to follow participants for a total of 5 years each including a period after the primary regulatory analysis and potential approval. In addition, use of an adaptive design to determine sample size and predictive probability to determine correct enrollment have rarely been used in studies of this nature. Furthermore, the indication for LibiGel will initially be in oophorectomized women, but the study is also enrolling naturally menopausal women to explore the safety of LibiGel in a broader population and is based on agreement with FDA. Finally, the CV composite outcome in the BLISS study is expanded compared with other recent studies of CV safety of noncardiac agents, including venous thromboembolic events, because testosterone is an obligate precursor of estradiol, a molecule that can be thrombogenic.
Conclusions
The BLISS study is an important and novel study for establishing the safety of LibiGel testosterone therapy in postmenopausal women with HSDD. The operational principals of this study may be applicable to conducting safety studies for other drugs in other populations with low event rates. The study has been ongoing since 2008 and has been actively enrolling and randomizing participants. The BLISS study is designed to document the efficacy and safety of LibiGel for treatment of HSDD, a common condition for which there is no FDA-approved pharmaceutical treatment in women.
Disclosures Funding source—The sponsor of this study is BioSante Pharmaceuticals, Inc, Lincolnshire, IL.
References 1. Laumann E, Paik A, Rosen R. Sexual dysfunction in the United States: prevalence and predictors. JAMA. 1999;281:537–544
2. Shifren JL, Monz BU, Russo PA, et al. Sexual problems and distress in United States women: prevalence and correlates. Obstet Gynecol. 2008;112:970–978
3. West SL, D'Aloisio AA, Agans RP, et al. Prevalence of low sexual desire and hypoactive sexual desire disorder in a nationally representative sample of US women. Arch Intern Med. 2008;168:1441–1449
4. Buster JE, Kingsberg SA, Aguirre O, et al. Testosterone patch for low sexual desire in surgically menopausal women: a randomized trial. Obstet Gynecol. 2005;105:944–952
5. Simon J, Braunstein G, Nachtigall L, et al. Testosterone patch increases sexual activity and desire in surgically menopausal women with hypoactive sexual desire disorder. J Clin Endocrinol Metab. 2005;90:5226–5233
6. Shames D, Monroe SE, Davis D, et al. Regulatory perspective on clinical trials and endpoints for female sexual dysfunction, in particular, hypoactive sexual desire disorder: formulating recommendations in an environment of evolving clinical science. Int J Impot Res. 2007;19:30–36
7. Barrett-Connor E, Mosca L, Collins P, et al., Raloxifene Use for The Heart (RUTH) Trial Investigators Effects if raloxifene on cardiovascular events and breast cancer in postmenopausal women. N Engl J Med. 2006;355:125–137
8. Hsia J, Langer RD, Manson JE, et al. Conjugated equine estrogens and coronary heart disease: the Women's Health Initiative. Arch Intern Med. 2006;166:357–365
9. Writing Group for the Women's Health Initiative Investigators. Risks and benefits of estrogen plus progesterone in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial. JAMA. 2002;288:321–333
10. Chappell R, Branch LG. The Goldilocks dilemma in survey design and its solution. J Clin Epidemiol. 1993;46:309–312
11. Davis SR, Moreau M, Kroll R, et al. Testosterone for low libido in postmenopausal women not taking estrogen. N Engl J Med. 2008;359:2005–2017
12. Longscope C. Adrenal and androgen secretion in normal females. Clin Endocrinol Metab. 1986;15:213–228
13. Giordano SH, Buzdar AU, Hortobagyi GN. Breast cancer in men. Ann Intern Med. 2002;137:678–687
14. Hofling MA, Hirschberg AL, Skoog L, et al. Testosterone inhibits estrogen/progestogen-induced breast cell proliferation in postmenopausal women. Menopause. 2007;14:183–190
15. Braunstein GD. Safety of testosterone treatment in postmenopausal women. Fertil Steril. 2007;88:1–17
16. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril. 2004;81:19–25
17. Miller KK, Biller BM, Schaub A, et al. Effects of testosterone therapy on cardiovascular risk markers in androgen-deficient women with hypopituitarism. J Clin Endocrinol Metab. 2007;92:2474–2479
18. Heufelder AE, F Saad MC, Bunck L, et al. Fifty-two–week treatment with diet and exercise plus transdermal testosterone reverses the metabolic syndrome and improves glycemic control in men with newly diagnosed type 2 diabetes and subnormal plasma testosterone. Andrology. 2009;30:726–733
19. Laughlin GA, Barrett-Connor E, Bergstrom J. Low serum testosterone and mortality in older men. J Clin Endocrinol Metab. 2008;93:68–75
20. Meldrum DR, Gambone JC, Morris MA, et al. The link between erectile and cardiovascular health: The canary in the coal mine. Am J Cardiol. 2011;108:599–606
21. Traish AM, Guay A, Feeley R, et al. The dark side of testosterone deficiency: I. Metabolic syndrome and erectile dysfunction. J Androl. 2009;30:10–22
22. Borer JS, Pouleur H, Abadie E, et al. Cardiovascular safety of drugs not intended for cardiovascular use: need for a new conceptual basis for assessment and approval. Eur Heart J. 2007;28:1904–1909
23. White WB, Faich G, Whelton A, et al. Comparison of thromboembolic events in patients treated with celecoxib, a cyclooxygenase-2 specific inhibitor, versus ibuprofen or diclofenac. Am J Cardiol. 2002;89:425–430
24. White WB, Pepine CJ, Weber MA. The potential role of nitric oxide in cardiovascular safety when treating patients with osteoarthritis and hypertension. Am J Med. 2009;122:S23–S25
25. ONTARGET Investigators . Telmisartan, ramipril , or both in patients at high risk for vascular events. N Engl J Med. 2008;358:1547–1559
26. Wiviott SD, Braunwald E, McCabe CH, et al. Prasugrel versus clopidigrel in patients with acute coronary syndromes. N Engl J Med. 2008;357:2001–2015
