Estrogen Replacement on High-Density Lipoprotein Cholesterol in Women with Coronary Disease

Abstract & Commentary

Synopsis: Of the 10 estrogen-receptor polymorphisms, one was associated with increased estrogen action as evidenced by a doubling of the expected HDL increase during estrogen replacement with conjugated equine estrogens.

Source: Herrington DM, et al. N Engl J Med. 2002; 346:967-974.

In this study, herrington and colleagues asked whether sequence variants in the gene encoding estrogen receptor alpha modified the effect of hormone replacement therapy (HRT) on levels of high-density lipoprotein cholesterol in postmenopausal women. The women were participants in the Estrogen Replacement and Atherosclerosis (ERA) trial. A total of 309 unrelated postmenopausal women with established coronary artery disease (CAD) were randomly assigned to receive 0.625 mg of oral conjugated equine estrogen (CEE) per day, CEE plus 2.5 mg medroxyprogesterone acetate per day, or placebo. Participants were followed an average of 3.2 years for progression of angiographically defined CAD. Plasma samples were collected annually after an overnight fast. There were 8 known polymorphisms of estrogen receptor alpha and 2 more were discovered. Increases in HDL cholesterol following either HRT arm were greatest in women who were homozygous for the less common alleles in intron 1, particularly the IVS (intervening sequence)1-401 polymorphism. There was insufficient power to determine if this increase in HDL afforded a better clinical result, ie, fewer coronary events or regression of CAD on angiography. However, the increase in HDL in that group was more than twice the increase observed in other women. Based on previous data, the observed increase would be expected to lower the risk of coronary events by 26-39%. Further, in 2 small clinical trials, the effects of estrogen on bone mineral density were greatest in women with the IVS1-401 polymorphism, indicating greater estrogen action. One mechanism may be that the polymorphism allows for a transcription factor known as myb. If so, this may increase estrogen action by altering transcription rates or the stability of mRNA transcripts.

Comment by Sarah L. Berga, MD

I realize that this is a complicated paper. It introduces a concept that most people would rather not know about, which is that not all women are the same at the molecular level. This field is generically referred to as "pharmacogenetics," meaning that if 2 people have the same circulating concentration of a drug, they will not necessarily have the same response to the same drug level because of differences in their molecular machinery. If not all estrogens are the same and not all women respond equivalently to a given estrogen, doesn’t that make it exceedingly difficult to know what to do when prescribing HRT? And isn’t the prescribing physician on a collision course with managed care companies who limit the time the physician has to spend with patients, the testing that can be done, and the types of estrogens that the patient is entitled to use? Further, if the circulating concentration doesn’t provide sufficient guidance, how are we to tell if the estrogen so given is achieving the desired effect? In essence, we must resort to the bioassay, ie, we will have to assess surrogate outcome measures for medical events and make sure that there is change in the appropriate direction. But to do that well, we must have relatively reliable surrogate markers for important clinical conditions likely to beset the patient. Staying well is getting ever so much more complicated!

While pharmacogenetics is complicated, it does introduce another layer of potential rationality into the world of prescribing HRT. However, the potential is presently more latent rather than actual. What should we do in the meantime? First, we need to try to define as best we can the conditions that the patient might be at risk for developing as she ages. Some of these are generic; some are specific to individual patients. Second, we need to fall back on physiological benchmarks. As a general rule in endocrine replacement, one wants to replicate physiology. What usually hinders the attainment of this goal is the lack of a suitable product or the cost or cumbersomeness of using that product. Fortunately, for ERT, we do have relatively easy-to-use products that closely replicate physiology, ie, transdermal estradiol patches. We know that the ovary, unlike the thyroid or the pineal, produces almost identical amounts of estradiol during an ovulatory menstrual cycle in all women. However, since HRT is not strictly giving hormones for a disease state, it is not clear if we should try to mimic the ovarian levels of a 25 year old or if we should aim for a somewhat lower level, but one that accomplishes the individually derived clinical objectives. At present, I am voting for the latter course.

Dr. Berga is Professor and Director, Division of Reproductive Endocrinology and Infertility, University of Pittsburgh.