Hypertension After Hormone Therapy Withdrawal
Some evidence suggests that HT improves quality of life, especially for symptomatic women. In some observational and randomized studies, HT has been shown to exert protection against the development of chronic diseases such as osteoporosis, cardiovascular diseases, and diabetes, although randomized trials have been conflicting, particularly with respect to cardiovascular disease.
Our results suggest that discontinuation of HT affects quality of life, particularly occupational satisfaction, for this select group of women. Decisions to discontinue HT may be influenced by multiple confounders, but our sample was fairly uniform with respect to SES, education, access to health care, and other behavioral factors, including exercise, alcohol intake, and smoking. Some studies have shown that hormone users are healthier and wealthier than nonhormone users, although the Nurses' Health Study found minor differences in risk factors and between users and nonusers, with increased cardiovascular risk in nonusers when adjusting for these. Our study consisted mainly of women of high SES status; these findings may apply only to this group, as they represent a typical sample of hormone users. We do not have data on diet, but it is improbable that these women of similar ethnicity, SES status, educational background, and BMI, and from a small geographic area with access to health care differed significantly in their diet. Others have recently published findings that menopausal symptoms are negatively associated with work ability and increase the risk of sickness absence. A negative effect on work outside the home was also documented in the Yale midlife study and commented on by others. It may not be coincidental that the most rapid decline in the rate of women’s work ability occurs at 51 years, the peak time for menopausal symptoms. The highest annual declining rate in work ability index was seen in this study among older women (aged 51 y), in comparison with men, during an 11-year period when a population working in the same occupation (age range, 44-51 y) was followed. The authors stated that "the role of physiological and mental changes associated with menopause among women should be studied." Our study found that women who continued hormones had better quality of life, reported more professional satisfaction, and held more managerial positions. However, these differences were not significant when controlling for age or ability to work. One explanation for greater hormone use among managerial women is that they could be more willing to assume risk for symptom relief.
Discontinuation of HT may also place some women at risk for the development of hypertension, which may be an early indicator of metabolic syndrome, a significant risk factor for cardiovascular disease. When controlling for other predictors of hypertension (age, weight, family history, and heart problems before starting HT), the odds of being on blood pressure medication were 2.289 times greater for those who were not on HT. Estrogen withdrawal may also initiate remodeling leading to atherosclerotic cardiovascular diseases; the first symptom may be heralded by hypertension. The role of hormone replacement in blood pressure and arterial function has been debated for several decades. Women seem to be protected against cardiovascular disease-related death before menopause, with a progressive rise in cardiovascular disease-related deaths after menopause; higher blood pressure levels occur as well, becoming steeper after the age of 62 years. Epidemiologic studies have been conflicting with respect to the effects of menopause or hormone replacement on blood pressure, with some studies suggesting a role for HT in reducing blood pressure whereas others show no effect or even a small increase. However, there have been no publications on blood pressure with HT withdrawal, except for the WHI. Results from the WHI E + P (HT) and estrogen-alone arms showed no difference in extension studies on withdrawal of treatment, but follow-up was conducted for 1.3 years only for the estrogen trial or reported only at baseline for E + P. During active treatment of the WHI trial, blood pressure reports were carried out for 1 year only. However, many of these women did not enroll and receive treatment close to the age of menopause, a time when vascular pathology is minimal and when they are significantly less likely to have complications such as cardiovascular disease or stroke (ie, complications reported in the WHI). In contrast to the WHI, the Nurses' Health Study followed 121,700 nurses prospectively to examine the role of hypertension in the development of cardiovascular events. Careful examination of their data showed that women who used postmenopausal hormones were less likely to have hypertension. A subsequent report showed a small decrease in hypertension and significantly fewer cardiovascular events with hormone use. Women in the Nurses' Health Study started hormones at the normal age of menopause. Timing of intervention with ET to prevent atherosclerosis is crucial in primate studies, with significant protection seen only close to oophorectomy.
Although epidemiologic studies are conflicting, physiological studies of HT for arterial stiffness and structure have been compelling and have been summarized. Surrogate measures of arterial stiffness are reduced with HT. These include reduced pulse wave velocity, improved systemic arterial compliance, and carotid distensiblity. Multiple studies have also found a decrease in vascular intima-media thickness and an attenuation of its development, similar to findings when premenopausal women are compared with postmenopausal women. Other nonendothelial factors may include improved lipid profile on menopausal therapy and enhanced secretion of nitrous oxide, a powerful vasodilator. Smooth muscle proliferation may also be inhibited. Withdrawal of estrogen secretion or therapy, as suggested by these studies, probably causes progressive stiffening of the vasculature, thickening of large arteries, and impaired vascular reactivity and relaxation, which result in increased systolic blood pressure, pulse pressure, and future cardiovascular events, and may also lead to hypertension. Pulse pressure, a surrogate measure of arterial stiffness, is a predictor of future cardiovascular events.
Our participants were also of normal weight. BMI is a strong indicator of endogenous estrogen, and women with higher BMI have higher levels of circulating estrogen. One very large cohort of 290,827 postmenopausal women found coronary protection only in women with lower BMI, where lower endogenous levels of estrogen would be expected. The mean BMI in the E + P arm of the WHI was 28.5 kg/m, and that in the estrogen-alone arm was 30.1 kg/m, whereas our participants averaged 24.0 (4.1) kg/m, similar to the Nurses' Health Study (24.3 kg/m), where coronary protection was observed. However, a recent analysis of a large cross-sectional cohort of women suggests an increased risk of hypertension selfreported by women using HT. Women in this study were not hypertensive before menopause. The baseline observational arm of the WHI also suggested an increased risk of hypertension in current hormone users. Thus, women with higher BMI and more endogenous estrogen and risk factors may be at higher risk for developing hypertension on HT/ET, whereas healthier women with normal BMI may experience a protective effect.
Importantly, medication use was not reported in the WHI studies. Antihypertensive medication use is a meaningful proxy for hypertension diagnosis and is particularly sensitive when accompanied by a diagnosis, which we obtained at interview. Medication history is an appropriate proxy because women who are being treated for hypertension will not have elevated blood pressure upon examination. We did not find differences in blood pressure between the groups of women. Treatment of hypertension is, however, considered an appropriate indicator of hypertension and has been used in recent clinical guidelines. Results from retrospective and other observational studies have been inconsistent with respect to the effects of aging, menopause, and HT on blood pressure. Any changes noted have been observed in the range of 1 to 2 mm Hg. Accuracy of measurements would be a large factor and suggests that these findings are not clinically relevant.
The use of a relatively small observational study to represent a sample selected from a healthy group of women who were highly educated and had normal BMI, as well as the nonapplication of the findings to other groups, was a limitation of this study. With respect to hypertension, a number of potential confounders were examined, but none showed significance in a multivariate-adjusted analysis examining HT status and use of hypertension medicationValthough our numbers were too small to include SES status, as many participants declined to answer questions concerning income. Obese participants composed only 8.3% of the sample, and women with diabetes only represented 2.3% ( Table 6 ). However, these participants are typical of hormone users, and all of these women had been undergoing treatment with HT/ET for a significant length of time. Gynecologic and women’s health practices may also have attracted women without significant health problems. However, the observation that hypertensive medication use in the group deciding to remain off therapy is significantly different may have implications for the pathophysiology of atherosclerotic diseases, suggesting that HT/ET treatment may delay the onset of hypertension in some, particularly in this normal-weight sample. Also, many different preparations of ET and HT were used even in individual women, and no conclusions can be made on the effects of dose, delivery (oral vs transdermal), or role of estrogen or progestin. More women who continued or restarted treatment used transdermal hormones. This may be attributed to a shift in treatment after the WHI and perceptions that transdermal therapies may have vascular advantages. Another limitation of the study was the small sample, which prevented adjusting for multiple variables that influence the outcomes of interest, such as obesity, hyperlipidemia, diabetes, medical morbidity, and SES. However, the sample was relatively homogeneous, and the incidence of diabetes and obesity was very low such that multivariable adjustment was not possible for selected variables. Our findings are preliminary and should be examined in a much larger sample.
A strength of this study is uniformity in the sample and in the decision to discontinue therapy because of an external event at approximately the same time. Sampling from healthcare practices provided a uniform sample of women who had taken HT/ET under the care of a healthcare professional and did not favor employed women, despite our findings in this regard. However, application of these findings requires further study and may apply only to the typical healthy hormone user with access to health care.
Discussion
Some evidence suggests that HT improves quality of life, especially for symptomatic women. In some observational and randomized studies, HT has been shown to exert protection against the development of chronic diseases such as osteoporosis, cardiovascular diseases, and diabetes, although randomized trials have been conflicting, particularly with respect to cardiovascular disease.
Our results suggest that discontinuation of HT affects quality of life, particularly occupational satisfaction, for this select group of women. Decisions to discontinue HT may be influenced by multiple confounders, but our sample was fairly uniform with respect to SES, education, access to health care, and other behavioral factors, including exercise, alcohol intake, and smoking. Some studies have shown that hormone users are healthier and wealthier than nonhormone users, although the Nurses' Health Study found minor differences in risk factors and between users and nonusers, with increased cardiovascular risk in nonusers when adjusting for these. Our study consisted mainly of women of high SES status; these findings may apply only to this group, as they represent a typical sample of hormone users. We do not have data on diet, but it is improbable that these women of similar ethnicity, SES status, educational background, and BMI, and from a small geographic area with access to health care differed significantly in their diet. Others have recently published findings that menopausal symptoms are negatively associated with work ability and increase the risk of sickness absence. A negative effect on work outside the home was also documented in the Yale midlife study and commented on by others. It may not be coincidental that the most rapid decline in the rate of women’s work ability occurs at 51 years, the peak time for menopausal symptoms. The highest annual declining rate in work ability index was seen in this study among older women (aged 51 y), in comparison with men, during an 11-year period when a population working in the same occupation (age range, 44-51 y) was followed. The authors stated that "the role of physiological and mental changes associated with menopause among women should be studied." Our study found that women who continued hormones had better quality of life, reported more professional satisfaction, and held more managerial positions. However, these differences were not significant when controlling for age or ability to work. One explanation for greater hormone use among managerial women is that they could be more willing to assume risk for symptom relief.
Discontinuation of HT may also place some women at risk for the development of hypertension, which may be an early indicator of metabolic syndrome, a significant risk factor for cardiovascular disease. When controlling for other predictors of hypertension (age, weight, family history, and heart problems before starting HT), the odds of being on blood pressure medication were 2.289 times greater for those who were not on HT. Estrogen withdrawal may also initiate remodeling leading to atherosclerotic cardiovascular diseases; the first symptom may be heralded by hypertension. The role of hormone replacement in blood pressure and arterial function has been debated for several decades. Women seem to be protected against cardiovascular disease-related death before menopause, with a progressive rise in cardiovascular disease-related deaths after menopause; higher blood pressure levels occur as well, becoming steeper after the age of 62 years. Epidemiologic studies have been conflicting with respect to the effects of menopause or hormone replacement on blood pressure, with some studies suggesting a role for HT in reducing blood pressure whereas others show no effect or even a small increase. However, there have been no publications on blood pressure with HT withdrawal, except for the WHI. Results from the WHI E + P (HT) and estrogen-alone arms showed no difference in extension studies on withdrawal of treatment, but follow-up was conducted for 1.3 years only for the estrogen trial or reported only at baseline for E + P. During active treatment of the WHI trial, blood pressure reports were carried out for 1 year only. However, many of these women did not enroll and receive treatment close to the age of menopause, a time when vascular pathology is minimal and when they are significantly less likely to have complications such as cardiovascular disease or stroke (ie, complications reported in the WHI). In contrast to the WHI, the Nurses' Health Study followed 121,700 nurses prospectively to examine the role of hypertension in the development of cardiovascular events. Careful examination of their data showed that women who used postmenopausal hormones were less likely to have hypertension. A subsequent report showed a small decrease in hypertension and significantly fewer cardiovascular events with hormone use. Women in the Nurses' Health Study started hormones at the normal age of menopause. Timing of intervention with ET to prevent atherosclerosis is crucial in primate studies, with significant protection seen only close to oophorectomy.
Although epidemiologic studies are conflicting, physiological studies of HT for arterial stiffness and structure have been compelling and have been summarized. Surrogate measures of arterial stiffness are reduced with HT. These include reduced pulse wave velocity, improved systemic arterial compliance, and carotid distensiblity. Multiple studies have also found a decrease in vascular intima-media thickness and an attenuation of its development, similar to findings when premenopausal women are compared with postmenopausal women. Other nonendothelial factors may include improved lipid profile on menopausal therapy and enhanced secretion of nitrous oxide, a powerful vasodilator. Smooth muscle proliferation may also be inhibited. Withdrawal of estrogen secretion or therapy, as suggested by these studies, probably causes progressive stiffening of the vasculature, thickening of large arteries, and impaired vascular reactivity and relaxation, which result in increased systolic blood pressure, pulse pressure, and future cardiovascular events, and may also lead to hypertension. Pulse pressure, a surrogate measure of arterial stiffness, is a predictor of future cardiovascular events.
Our participants were also of normal weight. BMI is a strong indicator of endogenous estrogen, and women with higher BMI have higher levels of circulating estrogen. One very large cohort of 290,827 postmenopausal women found coronary protection only in women with lower BMI, where lower endogenous levels of estrogen would be expected. The mean BMI in the E + P arm of the WHI was 28.5 kg/m, and that in the estrogen-alone arm was 30.1 kg/m, whereas our participants averaged 24.0 (4.1) kg/m, similar to the Nurses' Health Study (24.3 kg/m), where coronary protection was observed. However, a recent analysis of a large cross-sectional cohort of women suggests an increased risk of hypertension selfreported by women using HT. Women in this study were not hypertensive before menopause. The baseline observational arm of the WHI also suggested an increased risk of hypertension in current hormone users. Thus, women with higher BMI and more endogenous estrogen and risk factors may be at higher risk for developing hypertension on HT/ET, whereas healthier women with normal BMI may experience a protective effect.
Importantly, medication use was not reported in the WHI studies. Antihypertensive medication use is a meaningful proxy for hypertension diagnosis and is particularly sensitive when accompanied by a diagnosis, which we obtained at interview. Medication history is an appropriate proxy because women who are being treated for hypertension will not have elevated blood pressure upon examination. We did not find differences in blood pressure between the groups of women. Treatment of hypertension is, however, considered an appropriate indicator of hypertension and has been used in recent clinical guidelines. Results from retrospective and other observational studies have been inconsistent with respect to the effects of aging, menopause, and HT on blood pressure. Any changes noted have been observed in the range of 1 to 2 mm Hg. Accuracy of measurements would be a large factor and suggests that these findings are not clinically relevant.
The use of a relatively small observational study to represent a sample selected from a healthy group of women who were highly educated and had normal BMI, as well as the nonapplication of the findings to other groups, was a limitation of this study. With respect to hypertension, a number of potential confounders were examined, but none showed significance in a multivariate-adjusted analysis examining HT status and use of hypertension medicationValthough our numbers were too small to include SES status, as many participants declined to answer questions concerning income. Obese participants composed only 8.3% of the sample, and women with diabetes only represented 2.3% ( Table 6 ). However, these participants are typical of hormone users, and all of these women had been undergoing treatment with HT/ET for a significant length of time. Gynecologic and women’s health practices may also have attracted women without significant health problems. However, the observation that hypertensive medication use in the group deciding to remain off therapy is significantly different may have implications for the pathophysiology of atherosclerotic diseases, suggesting that HT/ET treatment may delay the onset of hypertension in some, particularly in this normal-weight sample. Also, many different preparations of ET and HT were used even in individual women, and no conclusions can be made on the effects of dose, delivery (oral vs transdermal), or role of estrogen or progestin. More women who continued or restarted treatment used transdermal hormones. This may be attributed to a shift in treatment after the WHI and perceptions that transdermal therapies may have vascular advantages. Another limitation of the study was the small sample, which prevented adjusting for multiple variables that influence the outcomes of interest, such as obesity, hyperlipidemia, diabetes, medical morbidity, and SES. However, the sample was relatively homogeneous, and the incidence of diabetes and obesity was very low such that multivariable adjustment was not possible for selected variables. Our findings are preliminary and should be examined in a much larger sample.
A strength of this study is uniformity in the sample and in the decision to discontinue therapy because of an external event at approximately the same time. Sampling from healthcare practices provided a uniform sample of women who had taken HT/ET under the care of a healthcare professional and did not favor employed women, despite our findings in this regard. However, application of these findings requires further study and may apply only to the typical healthy hormone user with access to health care.
Source...