Explaining Divergent Levels of Longevity in High-Income Countries (2011)

As discussed in Chapter 1, the slowdown in life expectancy in the United States versus other high-income countries during the period 1980–2005 was particularly pronounced among U.S. women. Thus it makes sense to look for a phenomenon that affects women but not men.

One possibility is the use of hormone therapy (HT) or, as it is often called, hormone replacement therapy. In the early 1980s, HT was prescribed for perimenopausal and postmenopausal women to combat various symptoms of menopause. Through 2002, it was prescribed increasingly often and to increasingly older women to provide protection against bone loss and cardiovascular disease (Goldman, 2010a; Kim et al., 2007). By 1995, about 40 percent of women between the ages of 50 and 74 were on HT, and the number peaked after 2000, when 92 million prescriptions were written in the United States (Hersh et al., 2004).

In 2002, however, researchers directing the Women’s Health Initiative announced that they were halting a randomized controlled trial that was looking at the effects of estrogen plus progestin in healthy postmenopausal women. The researchers had observed that women taking the combination of the two hormones had higher rates of breast cancer, coronary heart disease, stroke, and other diseases than women who had been receiving a placebo (Nelson et al., 2002; Writing Group for the Women’s Health Initiative Investigators, 2002). A second trial that was testing estrogen-only therapy against a placebo was halted in 2004 after data showed that the women receiving the hormone were more likely to suffer a stroke (Women’s Health Initiative Steering Committee, 2004).

Given the widespread use of HT in the United States and the findings

that it increased the risks of developing certain diseases, including coronary heart disease, breast cancer, and stroke, it is natural to ask whether HT might have contributed to the divergence in life expectancy trends between U.S. women and women in many other high-income countries. The question can be broken down into two parts: Does HT increase mortality risk? and Was the use of HT significantly more common in the United States than in other high-income countries? Goldman (2010b) examines both of those questions in a paper prepared for the panel.

The increased risk for coronary heart disease that the Women’s Health Initiative found among older women receiving HT was a surprise because earlier observational studies had found just the opposite (Goldman, 2010b). These earlier studies—which were based on observations of women given HT as part of normal medical practice rather than in randomized controlled trials—showed that women taking HT were generally from 35 to 50 percent less likely to develop heart disease than those not taking HT (Grodstein et al., 2000, 2006; Manson and Bassuk, 2007; Prentice and Anderson, 2008). For example, one meta-analysis of 32 previous observational studies calculated that women who had been treated with estrogen at some point had a 35 percent lower risk of developing coronary heart disease than women who had never been given estrogen (Grady et al., 1992).

By contrast, the Women’s Health Initiative and at least two other randomized controlled trials found an increased risk of coronary problems among women who were given HT. The results of the Women’s Health Initiative were particularly compelling as it involved 27,500 postmenopausal women and thus assembled a very large amount of data from which to draw conclusions about the effects of HT. The controlled trials reinforced some of the findings of the earlier observational studies—the benefits of HT in reducing the risk of colorectal cancers and hip fractures, for example, and an increased risk of breast cancer—but the new finding that HT increased the risk of coronary heart disease and stroke changed medical opinion on HT. Based on the judgment that the risks of HT appeared to outweigh the benefits, guidelines in the United States were modified to recommend against using HT for the prevention of cardiovascular disease in postmenopausal women. As a result, the use of HT has decreased dramatically since 2002 in the United States and other countries around the world (Barbaglia et al., 2009; Guay et al., 2007; Hersh et al., 2004).

Despite the evidence from the controlled trials, however, it appears unlikely that HT resulted in a significant increase in mortality among women in the United States. There are two reasons for this conclusion.

First, although the data indicate that HT—and the estrogen–progestin

regimen in particular—may have caused an increase in heart attack and stroke, there is no evidence that it actually led to an increase in overall death rates. Instead, the overall mortality rate was similar among women who were given the hormones and those who were not. For example, when the data from the Women’s Health Initiative were analyzed, the risk of death among womenS who were given estrogen–progestin HT was actually 2 percent less than for women who were given the placebo, although the difference was not statistically significant (Writing Group for the Women’s Health Initiative Investigators, 2002). A meta-analysis of the data from other randomized controlled trials pointed to a similar conclusion (Salpeter et al., 2004).

The second reason is that it now appears that the findings from the Women’s Health Initiative were likely to have been related to the timing of HT initiation in the trial. In the earlier studies based on observation of women receiving HT prescribed by their physicians, most of the women received treatment starting in early menopause. For example, 80 percent of the participants in the Nurses’ Health Study began HT within 2 to 3 years after beginning menopause (Manson and Bassuk, 2007). Since women in the United States begin menopause, on average, at age 51 (Manson et al., 2007), a large majority of the women in that study would have started HT by the time they were 53 or 54. By contrast, the average age of the women starting HT in the Women’s Health Initiative was 63, and most of the women in the study had begun menopause at least a decade before joining the study (Grodstein et al., 2006). However, these older women may have been more representative of the age group for which HT was being prescribed in later years to prevent fracture and cardiovascular disease.

One possible explanation for the different effects of HT at different ages is that estrogen supplements may have varying effects on the heart depending on the stage of atherosclerosis. Some researchers have hypothesized that in the early stages of atherosclerosis, estrogen may have a beneficial effect because it improves lipid and endothelial function, but in the later stages, when the arteries have developed more serious lesions, the estrogen may cause clotting or a rupturing of the plaque in the arteries (Manson and Bassuk, 2007). If so, the timing of the HT becomes critical.

Since the release of the Women’s Health Initiative findings, a number of studies have examined this possibility. In general, these studies have reanalyzed the data from previous reports by looking at the differences in how HT affected women at different ages, specifically those who started the treatment around the time of menopause versus those who started much later. Most of these reanalyses suggest that while women who begin HT well after menopause may have an increased risk of heart disease and stroke, those who start around the time of menopause do not, and for them the HT does appear to protect somewhat against heart disease (Goldman, 2010b).

However, at least one recent analysis did not find a significant difference in rates of coronary disease or in overall mortality between women who started HT near menopause and those who started later (Prentice et al., 2009), so questions remain about the exact role of timing of HT.

The news is not all good. The data do imply that HT increases certain risks for women, such as the risk of breast cancer, and this is true among women of every age. And indeed, when HT use was cut dramatically after the results of the Women’s Health Initiative were announced, the incidence of breast cancer dropped in the United States, as well as a number of other countries. However, the decreases were relatively small—a 7 percent drop in the United States, for example (Ravdin et al., 2007)—and even if HT was responsible for an increase in breast cancer deaths, those deaths were likely offset, at least in part, by a decline in deaths from other causes, such as colon cancer.

Thus, Goldman concludes, there is to date little evidence that women who start HT around the time that menopause begins are at greater risk of developing heart disease or that they are more likely to die when all causes of death are considered. Thus, it appears unlikely that HT played a role in the diverging life expectancies examined in this report. Still, for the sake of completeness, it is worthwhile to examine the second question: Was HT use significantly higher in the United States than in other countries?

In the early 1980s the World Health Organization began its MONICA, or Multinational Monitoring of Trends and Determinants in Cardiovascular Disease, project. As part of that project, women aged 45–64 from 32 separate populations in 20 countries were asked whether they had used HT in the previous month. The lowest percentage—0 percent—was found in Moscow, and the highest—42 percent—in Newcastle, Australia, and Halifax, Canada. The average among women from four communities in the United States was 38 percent—above the average in this data set, but below the numbers for Australia and Canada and approximately equal to those for France, Germany, and Iceland (Lundberg et al., 2004).

Various other studies have also indicated that, while a large percentage of U.S. women have used HT, the percentage of women in several other countries has been comparable. One study of French women aged 50–69 found that more than half of the women used HT (Gayet-Ageron et al., 2005), while a second study reported that, for the period 1998–2001, the percentage of French women using HT was twice that of U.S. women (Schneider, 2002). Another study showed that women in the United Kingdom received HT at about the same rate as women in the United States

(Townsend and Nanchahal, 2005). And while the women in some countries, such as Japan, did receive HT at a much lower rate than those in the United States, there clearly were a number of countries with both extensive HT usage and a rapid growth in female life expectancy during the crucial 1980–2005 period (Goldman, 2010b).

The cross-country comparisons are not definitive because of various issues with the data, as well as the fact that HT is administered differently in different countries—orally versus via a transdermal patch or gel. Some Europeans are much more likely to use the latter alternatives. For example, estimates are that 70–80 percent of women in France use these methods (Canonico et al., 2007; Ringa et al., 2005; Varas-Lorenzo et al., 1998). The recent literature suggests that orally administered estrogens are more likely to result in cardiovascular risk (e.g., elevated C-reactive protein [CRP] levels or an increase in triglycerides) than nonorally administered estrogens (L’Hermite et al., 2008; Vrablik et al., 2008). Evidence also suggests that orally but not transdermally administered HT is related to a higher risk of thrombolytic events in postmenopausal women (Scarabin et al., 2003). These findings open up the possibility that HT could have had more negative health consequences in the United States than in other countries, but the evidence is not strong enough to draw any firm conclusions (Goldman, 2010b).

Still, Goldman concludes that there is little evidence supporting the hypothesis that HT played a part in the divergence of life expectancy trends among high-income countries. She bases her conclusions on three factors. First, the data indicate that HT does not appear to increase all-cause mortality risk. Second, when HT is timed as it typically is—that is, when it is begun near the onset of menopause—it does not appear to increase the risk of heart disease and may actually decrease the risk for some women. And third, HT does not appear to have been any more common among U.S. women than among women in several other countries where life expectancy continued to increase steadily throughout the second half of the 20th century (Goldman, 2010b).