Clinical Chemistry 60:1 68–77 (2014)
Menopausal Hormone Therapy and Cardiovascular Disease Risk: Utility of Biomarkers and Clinical Factors for Risk Stratification Shari S. Bassuk1 and JoAnn E. Manson1*
BACKGROUND: Menopausal hormone therapy (HT) continues to have a clinical role in symptom management, but identifying women for whom benefits will outweigh the risks remains a challenge. Although hormone therapy (HT) is the most effective strategy for ameliorating vasomotor and other symptoms, randomized clinical trials show an unfavorable balance of benefits and risks for many women. However, closer examination of data from these trials suggests that it may be possible to classify women as better or worse candidates for HT by using individual risk stratification. CONTENT: Data from 2 landmark trials—the Women’s Health Initiative (WHI) and the Heart and Estrogen/ progestin Replacement Study (HERS)—suggest an important role for clinical characteristics, serum biomarkers, genomic markers, and gene– environment interactions in developing a personalized approach to the prediction of risk for cardiovascular disease (CVD) events for women while on HT. The available data suggest several characteristics of women who are optimal candidates for HT use: younger age (⬍60 years), recent onset of menopause (⬍10 years), favorable lipid profile (LDL cholesterol ⬍130 mg/dL or LDL/HDL cholesterol ratio ⬍2.5), absence of metabolic syndrome, and absence of factor V Leiden genotype. The identification of other characteristics is an area of active investigation. In addition, women at high risk for venous thromboembolism should avoid systemic HT or choose a transdermal rather than oral delivery route. SUMMARY: Personalized medicine—i.e., the use of the specific biological profile of an individual to guide the choice of treatment—is highly relevant for clinical decision-making regarding HT and offers promise for improved treatment efficacy and safety.
Vasomotor symptoms are a prominent feature of the menopause transition, affecting more than 70% of women and disturbing sleep and diminishing quality of life in 15%–20% (1–3 ). Menopausal hormone therapy (HT)2 is by far the most effective treatment for such symptoms, but its use has sharply declined in recent years (4 ) in response to results from randomized clinical trials showing an unfavorable balance of benefits and risks for many patients (2, 3 ). However, a closer look at data from these trials suggests that it may be possible to identify women for whom HT remains an appropriate option (i.e., those more likely to experience a favorable balance of benefits and risks) by the use of individual risk stratification. It may also be possible to use this information to optimize the choice of type of HT (delivery route, dose, and formulation) for an individual patient. The concept of personalized medicine, that each person has a unique physiology that can guide treatment selection, is highly relevant for clinical decision-making regarding HT and offers promise for improved treatment efficacy and safety. This review, which adapts and expands an earlier review (5 ), summarizes recent findings from 2 landmark randomized clinical trials in the US—the Women’s Health Initiative (WHI) and the Heart and Estrogen/progestin Replacement Study (HERS)— on the role of clinical characteristics, serum/plasma biomarkers, gene markers, and gene– environment interactions in helping to identify good vs poor candidates for HT use with respect to cardiovascular disease (CVD) outcomes. The ultimate aim of such research is to develop a personalized benefit–risk prediction model for systemic HT use to treat moderate-to-severe vasomotor symptoms or prevent osteoporosis in postmenopausal women at high fracture risk who cannot tolerate alternative therapies, the only 2 indications for which systemic HT use is
© 2013 American Association for Clinical Chemistry
Brigham and Women’s Hospital, Harvard Medical School, Boston, MA. * Address correspondence to this author at: Division of Preventive Medicine, Brigham and Women’s Hospital, 900 Commonwealth Ave. East, Boston, MA 02215. Fax 617-731-3843; e-mail [email protected]
Received August 16, 2013; accepted November 1, 2013. Previously published online at DOI: 10.1373/clinchem.2013.202556
Nonstandard abbreviations: HT, hormone therapy; WHI, Women’s Health Initiative; HERS, Heart and Estrogen/progestin Replacement Study; CVD, cardiovascular disease; CEE, conjugated equine estrogens; MPA, medroxyprogesterone acetate; MI, myocardial infarction; VTE, venous thromboembolism; RR, relative risk; WISDOM, Women’s International Study of long Duration Oestrogen after Menopause; OR, odds ratio; Lp(a), lipoprotein(a); IL-6, interleukin-6; MMP-9, matrix metalloproteinase-9; CRP, C-reactive protein; TFPI, tissue factor pathway inhibitor.
Hormone Therapy Risk Stratification
Table 1. Cardiovascular effects of postmenopausal hormone therapy in the overall study population of women aged 50 –79 years in the WHI estrogen–progestin and estrogen-alone trials.a Intervention ⴙ postintervention phasec
Intervention phaseb Estrogen-progestin, RR (95% CI)
Estrogen alone, RR (95% CI)
Estrogen–progestin, RR (95% CI)
Estrogen alone, RR (95% CI)
Deep vein thrombosis
Data from Manson et al. (8 ). Median length of intervention was 5.6 years for estrogen–progestin and 7.1 years for estrogen alone. c Median length of total follow-up was 13 years for both trials. d CHD is defined as nonfatal myocardial infarction or coronary death. b
currently warranted (1 ). HT should be prescribed only for patients with a preference for such therapy. Overview of the WHI and HERS Trials The WHI HT trials included more than 27 000 healthy postmenopausal women aged 50 –79 (mean age, 63). In the estrogen–progestin trial, 16 608 women with an intact uterus were assigned to oral estrogen plus progestin [0.625 mg of conjugated equine estrogens (CEE) plus 2.5 mg of medroxyprogesterone acetate (MPA)] or a placebo daily (6 ). In the estrogen-alone trial, 10 739 women with hysterectomy (40% also had oophorectomy) were assigned to oral estrogen alone (0.625 mg CEE) or a placebo daily (7 ). The sample sizes were chosen to have adequate power to detect an effect of HT on CHD, the primary outcome of interest, and to assess the balance of benefits and risks over the trials’ planned duration of 9 years. Blood samples were collected at baseline and at the 1-year visit. The estrogen– progestin trial was halted after a mean of 5.6 years of treatment, because of a significant increase in breast cancer risk and an unfavorable balance of benefits and risks associated with estrogen–progestin therapy in the study population as a whole (6 ). The estrogen-alone trial was halted after 7.1 years because of an excess risk for stroke that was not offset by a reduced risk for CHD in the treatment group (7 ). Although estrogen with or without progestin lowered the risk for osteoporotic fracture, it offered no other clear benefit in terms of reducing risk for chronic disease. After the trials were stopped, the participants were followed observationally to determine whether, and how rapidly, risks and benefits decline after discontinuation of HT use. As shown in Table 1, women assigned to 5.6 years of estrogen–progestin therapy were 18% more likely to
develop CHD [defined in primary analyses as nonfatal myocardial infarction (MI) or coronary death] than those assigned to placebo, although this risk increase was not statistically significant (8 ). However, during the first year of the trial, there was a significant 80% increase in risk, which diminished in subsequent years (P for trend by time ⫽ 0.03) (8 ). Women assigned to 7.1 years of estrogen alone experienced neither an increase nor decrease in risk for CHD (8 ). This pattern of results was similar for the outcome of total MI (8 ). Women assigned to estrogen–progestin or estrogen alone were about 35% more likely to suffer a stroke than those assigned to placebo (8 ). An increase in risk was found only for ischemic stroke but not hemorrhagic stroke. Assignment to estrogen–progestin was associated with a nearly 2-fold increase in risk for pulmonary embolism and deep vein thrombosis, and assignment to estrogen alone led to a 35%–50% increase in these risks (8 ). Cardiovascular risks associated with active use of HT largely decreased after discontinuation of therapy, although some remained increased when considering the entire length of the study (5.6 years of treatment plus 6.8 years of postintervention follow-up) (8 ). In the 4.1-year HERS, conducted among 2763 women with preexisting CHD (age range 55– 80 years; mean age, 67 years), the incidence of major coronary events was similar in the HT (0.625 mg of oral CEE plus 2.5 mg of MPA daily) and placebo groups (9 ). Participants assigned to estrogen–progestin had a 50% increase in their risk for CHD events during the first year of the trial. This early increase was offset by a decreased risk later in the trial, leading to null results overall (9, 10 ). Assignment to estrogen–progestin was not associated with risk for stroke (9, 10 ) but was associated with significantly increased risk for venous thromboembolism (VTE); the relative risk (RR) was 2.7 Clinical Chemistry 60:1 (2014) 69
Reviews (95% CI, 1.4 –5.0) (11 ). The risk decreased after discontinuation of therapy [2.7 years of postintervention followup; RR, 1.40 (0.64 –3.05)], although it remained increased when the entire length of the study was considered [RR, 2.08 (1.28 –3.40)] (12 ). Blood samples were collected at baseline and at 1 year after randomization.
menopause (within 10 years), favorable lipid profile (LDL cholesterol ⬍130 mg/dL or LDL/HDL cholesterol ratio ⬍2.5), absence of metabolic syndrome, and absence of factor V Leiden genotype. In addition, recent evidence suggests that women at high risk for VTE should either avoid systemic HT or choose a transdermal rather than oral delivery route.
Factors That Influence HT-Associated Cardiovascular Risks
AGE AND TIME SINCE MENOPAUSE ONSET
WHI and HERS investigators conducted a large number of subgroup analyses to identify subgroups of women for which HT may be especially beneficial or harmful. It is important to recognize that, as is the case for many clinical trials, focused subgroup analyses were not emphasized at the outset, and thus the trials were not powered to detect potential interactions. Thus, it is possible that some true interactions, especially those of small or moderate size, were missed (false-negative findings). Conversely, the large number of subgroup comparisons would be expected to yield some statistically significant findings purely by chance, in the absence of true interactions (false-positive findings), suggesting the need for caution in interpretation of the findings. The existence of congruent results across trials and/or independent evidence regarding the biological plausibility of an observed interaction is important in assessing whether an observed interaction is a true or spurious effect. Subgroup analyses have been reported by various WHI investigative teams, including (but not limited to) Manson et al. (8, 13 ), Hsia et al. (14 ), and Rossouw et al. (15, 16 ) for the outcome of CHD; WassertheilSmoller et al. (17 ), Hendrix et al. (18 ), Rossouw et al. (15 ), Kooperberg et al. (19 ), and Manson et al. (8 ) for stroke; and Cushman et al. (20 ), Curb et al. (21 ), and Manson et al. (8 ) for the outcome of VTE. [Subgroup analyses have also been reported by multiple teams in HERS, most extensively by Furberg et al. (22 ) but also by Herrington et al. (23 ), Hulley et al. (12 ), Huang et al. (24 ), and others.] When factors have been examined as potential effect modifiers in multiple reports from the same study (e.g., age and time since menopause in the WHI), we focus on those reports that included the largest number of confirmed endpoints; used a uniform coding scheme for statistical modeling of the factor of interest; and in the case of WHI, combined data from the estrogen–progestin and estrogen-alone trials to increase power. However, the results of various within-study reports regarding specific interactions were generally congruent. Several characteristics that modify risk for CVD events in women while on HT have been identified. As reviewed below, optimal candidates for HT use include women with younger age (⬍60 years), recent onset of
WHI analyses suggest that age or time since menopause onset affects the relation between HT and CHD (Table 2). Although there was no association between estrogenonly therapy and CHD in the overall study population, such therapy was associated with a borderline significant 40% reduction in CHD among women age 50 –59 years, whereas there was no risk reduction or even a slight increase in older age groups (P for trend by age ⫽ 0.08) (8 ). For total MI, estrogen alone was associated with a 45% reduction and a nonsignificant 24% increase in risk among the youngest and oldest women, respectively (P for trend by age ⫽ 0.02), and this pattern of results persisted during the cumulative 13-year follow-up period (including 6.6 years of postintervention observation). Estrogen was also associated with lower levels of coronary artery calcified plaque in the youngest age group (25 ). Although age did not have a similar effect in the estrogen–progestin trial, HTassociated CHD risks increased with years since menopause (P, trend ⫽ 0.08), with a significant 52% increase in risk among women who were ⱖ20 years past menopause. For total MI, estrogen–progestin was associated with a nonsignificant 9% risk reduction among women ⬍10 years past menopause but was associated with a 16% increase in risk among women 10 –19 years past menopause and a 2-fold increase in risk among women ⱖ20 years past menopause (P, trend ⫽ 0.01). Interestingly, the increased coronary risk with HT use at older ages or many years past menopause was particularly prominent among women with vasomotor symptoms (15 ). Vasomotor symptoms did not predict an adverse coronary outcome on HT in younger or recently menopausal women, however. In HERS, with its predominantly older study population, the adverse coronary impact of HT was also pronounced in those with vasomotor symptoms (24 ). Given that vasomotor symptom alleviation is the primary indication for HT, this finding strongly bolsters current clinical recommendations against initiating HT use in older women and suggests caution for longterm continuation of treatment. The apparent 3-way interaction between age, vasomotor symptoms, and HT on risk for CHD illustrates the potential importance of higher-order interactions in more finely pinpointing the benefit-risk balance among women with certain clinical or biomarker profiles, but such analyses remain limited by power considerations.
70 Clinical Chemistry 60:1 (2014)
Hormone Therapy Risk Stratification
Table 2. Effects of postmenopausal hormone therapy on cardiovascular outcomes in the WHI estrogen–progestin and estrogen-alone trials, according to age at study entry.a Intervention phaseb Estrogen–progestin
RR (95% CI)
RR (95% CI)
MI 50–59 years
0.55 (0.31–1.00) 0.95 (0.69–1.30)
Stroke 50–59 years
0.99 (0.53–1.85) 1.55 (1.10–2.16)
Pulmonary embolism 50–59 years
1.53 (0.63–3.75) 1.72 (0.94–3.14)
Deep vein thrombosis 50–59 years
1.66 (0.75–3.67) 1.41 (0.87–2.27)
Data from Manson et al. (8 ). Median length of intervention was 5.6 years for estrogen–progestin and 7.1 years for estrogen alone. c CHD is defined as nonfatal myocardial infarction or coronary death. b
Because younger women and those closer to menopause are likely to have healthier arteries than their counterparts who are older or further past menopause, the WHI findings suggest that the net coronary effect of HT may vary according to the initial health of the vasculature. The findings also spurred reanalyses of existing data. For example, investigators in the Nurses’ Health Study, one of the largest and longest-running prospective observational studies in the US, first reported that current use of HT was associated with an approximately 40% reduction in CHD risk in the overall cohort (26 ). On reexamination of the data, they found that the coronary benefit was largely limited to women who started HT within 4 years of menopause onset (27 ). A metaanalysis that pooled data from 21 smaller randomized trials with data from WHI and HERS found that HT was associated with a 30%– 40% reduction in CHD risk in trials that enrolled predominantly younger participants (those aged ⬍60 or ⱕ10
years from menopause onset) but not in trials with predominantly older participants (28 ). Recently reported results from other clinical trials are also relevant. The UK-based Women’s International Study of long Duration Oestrogen after Menopause (WISDOM), which had planned to randomize 22 000 healthy postmenopausal women aged 50 – 69 to 10 years of HT (CEE with or without MPA) or placebo, was halted upon publication of the initial WHI results. Analyses of the available data—5692 participants with a mean age of 63 years were treated for 1 year with HT or placebo—showed an increase in the number of CHD events in the HT (estrogen–progestin, n ⫽ 7; estrogen alone, n ⫽ 4) vs placebo (n ⫽ 0) groups (P value for estrogen–progestin vs placebo comparison ⫽ 0.016) (29 ). It is notable that all but 2 of the 11 CHD events occurred in women who were aged ⬎64 years at trial entry and had at least 1 coronary risk factor. On the other hand, in a 10-year open-label trial in Denmark among 1000 healthy reClinical Chemistry 60:1 (2014) 71
Reviews cently menopausal women aged 45–58 years (mean age, 50 years) at trial entry, HT (estradiol with or without norethisterone acetate) was associated with a significant reduction in the primary endpoint of mortality, MI, or heart failure [16 vs 33 events; RR, 0.48 (0.26 – 0.87)] (30 ). We did not perform an updated metaanalysis to summarize results across the various trials due to the substantial differences in the clinical characteristics of women in these trials and differences in trial methodology. Although the WHI (primary prevention) and HERS (secondary prevention) are by far the 2 largest completed randomized trials of HT in relation to CHD outcomes and combining them could increase statistical power to examine interactions, such a metaanalysis would not achieve the goal of identifying factors that may assist in cardiovascular risk stratification among women now considered appropriate candidates for HT. To combine data from the 2 trials would not be advisable, as it is already clear that women with preexisting heart disease—the population studied in HERS—are not appropriate candidates for HT. In addition, smaller and/or shorter-term trials designed to look at outcomes other than CHD—as well as the prematurely terminated WISDOM trial for primary prevention of CHD (29 )— do not have an appreciable number of clinical cardiovascular events, so their results would be dwarfed by the WHI results in a metaanalysis. Findings of nonhuman primate experiments are also informative. Conjugated estrogens with or without MPA had no effect on coronary artery plaque in cynomolgus monkeys started on this treatment at 2 years (approximately 6 human years) after oophorectomy and well after the establishment of atherosclerosis, whereas such therapy reduced the extent of plaque by 70% when initiated immediately after oophorectomy, during the early stages of atherosclerosis (31 ). Similarly, human imaging trials in women with significant coronary lesions at baseline have not found estrogen to be effective in slowing the rate of arterial narrowing (32–35 ), whereas an imaging trial in which the presence of significant lesions was not a criterion for study entry found that micronized 17␤-estradiol significantly slowed the progression of carotid atherosclerosis (36 ). However, the recently completed Kronos Early Estrogen Prevention Study (KEEPS), a 4-year randomized trial among 729 healthy recently menopausal women with a mean age of 53 years, failed to find that low-dose oral CEE (0.45 mg/day) or transdermal estradiol (50 g/day by weekly patch), administered in combination with oral micronized progesterone (200 mg/day for 12 days/month), prevented progression of carotid intimal medial thickness, although it did find a nonsignificant reduction in accrual of coronary calcium (37, 38 ). Results from a random72 Clinical Chemistry 60:1 (2014)
ized trial testing whether effects of estradiol on the development and progression of atherosclerosis vary according to age at HT initiation are expected in 2014 (39 ). In contrast to the findings for CHD, analyses from the WHI do not suggest that age or time since menopause onset influence the risk for HT-induced stroke or VTE (8, 40 ) (Table 2). However, because younger women have lower absolute risks of CHD, stroke, and VTE than do older women, they also have much lower absolute excess risks associated with HT use. CARDIOMETABOLIC RISK FACTORS AND OTHER CLINICAL CHARACTERISTICS
In the WHI, hypertension, diabetes, smoking status, body mass index, aspirin use, and baseline risk for CHD (as assessed by the number of CHD risk factors, for the outcome of CHD) or stroke (Framingham Stroke Score, for the outcome of stroke) did not significantly affect the relation between either estrogen– progestin or estrogen alone and incident CHD (13, 14 ) or stroke (17, 18 ). This was also largely true for statin use (but see discussion in the next section). A history of CVD at study entry also did not appear to be a significant effect modifier (⬍5% of WHI participants had prior CVD). In the WHI estrogen–progestin trial, obesity was a strong risk factor for VTE but did not significantly modify the effect of HT on VTE risk (20 ). Other variables (e.g., smoking, aspirin use, and history of CVD before enrollment) did not modify the effect of HT on risk for VTE (20 ). However, in HERS, aspirin use appeared to attenuate risk for VTE associated with estrogen– progestin; in an analysis of the active treatment plus postintervention follow-up periods, HT-associated RR in aspirin users and nonusers were 1.68 (0.96 –2.92) and 4.23 (1.41–12.7), respectively (P, interaction ⫽ 0.14) (12 ). (It should be noted that differences in findings between the WHI and HERS could be a result of differences in the study populations or could be a result of chance; they are the only 2 trials that are large enough for a meaningful examination of interactions.) In the WHI estrogen-alone trial, body mass index and other variables did not significantly modify the effect of HT on risk for VTE. Although later age at menopause onset was a risk factor for VTE in both WHI (40 ) and HERS (11 ), reproductive and menopause-related variables did not appear to modify the effect of HT on VTE in either trial. SERUM/PLASMA BIOMARKERS
To identify biomarkers that may modify the impact of HT on CHD or ischemic stroke, WHI investigators performed parallel nested case-control studies of these outcomes. The studies included all adjudicated cases of CHD (n ⫽ 359) (16, 41– 44 ) or ischemic stroke (n ⫽
Hormone Therapy Risk Stratification
Table 3. CHD risk in the WHI estrogen–progestin and estrogen-alone trials, according to baseline concentrations of selected markers.a,b OR (95% CI) for HT treatment effect
P for interaction
LDL cholesterol, mg/dL
LDL/HDL cholesterol ratio ⬍2.5
C-reactive protein ⬍2.0
Metabolic syndromec,d No
Waist circumferenced ⱕ88 cm
Data from Bray et al. (41 ) and Wild et al. (42 ). To convert cholesterol concentrations from mg/dL to mmol/L, multiply by 0.2586. c Metabolic syndrome was defined as having 3 or more of the following: waist size ⬎88 cm (⬎80 cm for Asians and Native Americans); systolic blood pressure ⬎130 mm/Hg or diastolic blood pressure ⬎85 mmHg; fasting glucose ⬎100 mg/dL; HDL cholesterol ⬍50 mg/dL; triglycerides ⬎150 mg/dL. d The reported ORs are among participants without CVD, hypertension, or diabetes at baseline. b
205) (19, 45 ) that occurred within the first 4 years of follow-up of the estrogen–progestin and estrogenalone trials. (A similar study of VTE is in progress.) Unless noted, the results summarized here pertain to the combined analysis of the estrogen–progestin and estrogen-only trials. Women without vascular risk factors generally fared better than women with such factors with respect to HT-associated risk for CHD, but a similar pattern was not found for stroke. CHD. As expected, baseline lipids (high LDL cholesterol, low HDL cholesterol, high total cholesterol, high triglycerides) predicted an increased risk for CHD events. In addition, a more favorable lipid profile was associated with better coronary outcomes on HT (Table 3). Among women with a lower LDL/HDL cholesterol ratio (⬍2.5), estrogen with or without progestin led to a 40% lower risk for incident CHD [RR, 0.60 (0.34 –1.06)], whereas among women with a higher baseline ratio (ⱖ2.5), such therapy resulted in a 73% higher risk [RR, 1.73 (1.18 –2.53)] (P, interaction ⫽
0.02) (41 ). Similarly, among women who entered the trial with a lower LDL cholesterol concentration (⬍130 mg/dL or 3.36 mmol/L), estrogen with or without progestin led to a nonsignificantly lower risk for incident CHD [RR, 0.66 (0.34 –1.27)], whereas among participants with a higher baseline cholesterol (ⱖ130 mg/dL), such therapy resulted in a 46% higher risk [RR, 1.46 (1.02–2.10)] (P value for interaction ⫽ 0.03) (41 ). The results were similar when users of statins and other cholesterol-lowering drugs were excluded from the analysis (41 ). Statins may attenuate the coronary risk associated with estrogen–progestin. In the WHI estrogen– progestin trial, the HT-associated RRs of CHD among statin users and nonusers were 0.99 and a statistically significant 1.27, respectively, but the P value for the interaction was nonsignificant (0.44) (13 ). In HERS, the strong adverse effect of HT on CHD seen during the first year of the trial was less pronounced among baseline statin users [RR, 1.34 (0.63–2.86)] than among nonusers [RR, 1.75 (1.02–3.03)] (P for interaction not reported) (23 ). The presence or absence of the metabolic syndrome— defined as having ⱖ3 of the following: waist size ⬎88 cm; systolic blood pressure ⬎130 mm/Hg or diastolic blood pressure ⬎85 mmHg; fasting glucose ⬎100 mg/dL (⬎5.55 mmol/L); HDL cholesterol ⬍50 mg/dL (⬍1.29 mmol/L); triglycerides ⬎150 mg/dL(⬎1.69 mmol/L) —strongly influenced the relation between HT and incident CHD. The effect modification was most pronounced in participants without CVD, hypertension, or diabetes at baseline. In these individuals, the presence of the metabolic syndrome more than doubled the HT-associated CHD risk [odds ratio (OR), 2.26 (1.26 – 4.07)], whereas no association between HT and CHD risk was observed among those without the syndrome [OR, 0.97 (0.58 – 1.36)] (P, interaction ⫽ 0.03) (42 ). Although there was a suggestive interaction for waist size [waist ⬎88 cm, OR ⫽ 1.93 (1.08 –3.44); waist ⱕ88 cm, OR ⫽ 1.03 (0.62–1.70); P, interaction ⫽ 0.12], the individual components of the syndrome did not significantly influence the HT-CHD association. This finding suggests that higher-order interactions between risk markers may be important for individualized risk prediction. Of 86 baseline characteristics examined as potential modifiers of the effect of estrogen–progestin on risk for recurrent CHD events in HERS, only lipoprotein(a) [Lp(a)] (a main-effects predictor of CHD in both HERS and the WHI estrogen–progestin trial) was found to significantly influence HT-associated coronary risk over the entire 4-year randomized treatment period (22 ). HT appeared protective against CHD in women with Lp(a) ⱖ25.3 mg/dL [RR, 0.82 (0.62– 1.08)] but harmful in those with Lp(a) ⬍25.3 [RR, 1.25 Clinical Chemistry 60:1 (2014) 73
Reviews (0.92–1.70)] (P, interaction ⫽ 0.04). In the initial report on CHD outcomes in the WHI estrogen–progestin trial, a similar pattern was seen, although the findings did not approach statistical significance. For Lp(a) ⬍12, 12– 31, and ⬎31 mg/dL, HT-associated ORs for CHD were 1.87, 1.54, and 0.93, respectively (P, interaction ⫽ 0.44) (13 ). Although the direction of the effect modification contradicts the hypothesis that HT is more benign for women with a more favorable coronary risk profile at baseline, the results fit with the finding that HT significantly reduced Lp(a) concentrations in HERS (46 ). A detailed analysis of the interrelationship between Lp(a), HT, and CHD in the WHI is in process. In the WHI, the cholesterol derivative 27hydroxycholesterol, which acts as an estrogen receptor antagonist, did not independently predict CHD risk, affect the HT–CHD association, or mediate the observed interaction of LDL cholesterol and HT on CHD risk described earlier (44 ). In addition to lipids, the WHI found that several biomarkers related to inflammation, thrombosis, and other processes were associated with increased CHD risk after adjustment for traditional risk factors and randomized assignment to HT. These included higher concentrations of the inflammatory markers interleukin-6 (IL-6), matrix metalloproteinase-9 (MMP-9), and leukocyte count; the thrombotic markers D-dimer, factor VIII, and von Willebrand factor; homocysteine; and fasting insulin (16 ). Although most of these biomarkers did not significantly aid in identifying women more or less likely to experience a CHD event while taking HT, there was a significant interaction for factor VIII (P ⫽ 0.04) and a suggestive interaction for homocysteine (P ⫽ 0.12) (16 ). The interaction for homocysteine became significant during the second 2 years of follow-up (P ⫽ 0.03). The inflammatory markers C-reactive protein (CRP) and E-selectin, the thrombotic markers fibrinogen and prothrombin fragment 1.2, and hematocrit did not significantly modify HT-associated CHD risk. Although women with CRP concentrations ⬍2.0 mg/L were not at increased risk for HT-associated CHD events [RR, 1.01 (0.63–1.62)] and those with CRP concentrations ⱖ2.0 mg/L were at increased risk [RR, 1.58 (1.05–2.39)], the P-value for interaction was not statistically significant (0.16) (41 ). In an attempt to determine the biological processes underlying the early increase in CHD events seen in the WHI and HERS trials, WHI investigators examined the impact of HT on the aforementioned biomarkers after 1 year of treatment (16 ). These analyses included 236 cases of CHD that occurred after the year 1 visit. Estrogen with or without progestin decreased LDL and total cholesterol concentrations; increased HDL cholesterol and triglyceride concentrations; increased CRP and MMP-9 concentrations; decreased 74 Clinical Chemistry 60:1 (2014)
E-selectin concentrations; had no effect on IL-6 concentrations; increased plasmin–antiplasmin complex concentrations; decreased fibrinogen, plasminogen activator inhibitor 1 antigen, homocysteine, glucose, and insulin concentrations; and had no effect on D-dimer, factor VIII, prothrombin fragment 1.2, thrombinactivatable fibrinolysis inhibitor, and von Willebrand factor concentrations. However, none of the changes significantly predicted CHD risk or affected the HT–CHD association after the first year. In HERS, estrogen– progestin had similar effects on 1-year changes in lipids as was observed in the WHI, but these changes were also unrelated to risk for subsequent CHD (46 ). Neither the WHI nor HERS analyses address the relation between very early biomarker changes (i.e., those occurring during the first weeks or months of the trial) and the strongly increased HT-associated CHD risk seen during the first year of the trials. The analyses also do not address the potential synergy of simultaneous HTinduced changes in biomarkers on CVD outcomes. Stroke. In the WHI, several biomarkers predicted ischemic stroke risk, including LDL cholesterol, HDL cholesterol, CRP, IL-6, MMP-9, D-dimer, and thrombinactivatable fibrinolysis inhibitor (19 ). However, only baseline plasmin–antiplasmin complex concentrations interacted significantly with assignment to HT; the interaction was in a paradoxical direction such that higher concentrations of this biomarker were not associated with risk in the estrogen–progestin group [RR, 0.32 (0.13–1.79)] but were associated with increased risk in the placebo group [RR, 7.09 (1.48 –34.03)]. Similar paradoxical trends were observed for baseline IL-6 and D– dimer. Although several biomarkers were significantly affected by HT (see preceding section), these changes did not predict stroke risk among women on HT, with one exception—an increase in D-dimer predicted a higher risk. In a recent nested case control analysis that included 455 cases of ischemic stroke and focused on the effects of 2 additional hemostatic markers, tissue factor pathway inhibitor (TFPI) and activated protein C resistance (normalized activated protein C resistance ratio), WHI investigators found that free TFPI was a significant predictor of higher stroke risk [OR per 1-SD increase, 1.17 (1.01–1.37)] and that HT decreased total TFPI, free TFPI, and TFPI activity and increased normalized activated protein C resistance ratio levels. However, neither baseline concentrations of, nor changes in, these factors influenced the effect of HT on stroke risk (45 ). In-depth plasma proteome profiling studies identified ␤2-microglobulin and insulin-like growth factor binding protein 4 as predictors of incident CHD and stroke in the WHI, respectively. Levels of these proteins
Hormone Therapy Risk Stratification
increased following HT initiation, suggesting that these changes may partly explain effects of HT on CHD and stroke (47 ).
Table 4. Selected biomarkers under study as potentially useful for risk stratification for hormone therapy decision making.a
VTE. In comparison to CHD and stroke, fewer biomarkers have been examined as potential modifiers of the association between HT and VTE. In the WHI estrogen–progestin trial, baseline lipids and statin use did not modify the association. In HERS, although statin use reduced VTE risk by 60%, HTassociated risks for VTE remained increased in both users and nonusers of statins (23 ). Factor V Leiden is discussed below.
• Biochemical markers: • Lipids (serum LDL cholesterol, LDL/HDL cholesterol ratios, triglyceride levels, lipoprotein(a), 27hydroxycholesterol, apolipoprotein levels) • Inflammatory markers (high-sensitivity C-reactive protein, interleukin-6, tumor necrosis factor alpha, leukocyte count) • Adipokines (adiponectin, leptin, retinol binding protein-4) • Endothelial markers (E-selectin, P-selectin, intercellular adhesion molecule 1, vascular cell adhesion molecule 1)
Several genetic polymorphisms were examined as predictors of CHD and stroke in the aforementioned nested case control studies within the WHI (16, 19 ). In these analyses, glycoprotein IIIa leu33pro predicted risk for CHD (but not stroke) but did not clearly modify the effect of HT on CHD risk, while variants in the genes for estrogen receptor ␤-A1730G, factor V Leiden, factor XIII val34leu, glycoprotein 1b␣-M145T, methylenetetrahydrofolate reductase, plasminogen activator inhibitor 1, integrin alpha2– 807, and prothrombin 20210 did not predict risk for CHD and/or stroke nor identify women who were at greater risk for these events while on HT. In more recent analyses of WHI data, estrogen receptor polymorphisms attenuated the effect of HT on plasmin–antiplasmin, a marker of coagulation and fibrinolysis, but not cardiovascular events (43 ). Genetic variation in the coagulation factor XIII subunit A region and in the proprotein convertase subtilisin kexin 9 region interacted with HT to affect stroke risk (48 ). In the initial report on VTE outcomes in the WHI estrogen–progestin trial, factor V Leiden was strongly predictive of VTE and also interacted with HT to augment VTE risk (20 ). DELIVERY ROUTE, DOSE, AND FORMULATION OF HT
The WHI and HERS trials were limited in that they studied only 1 method of delivery, dose, and formulation of estrogen and progestin; thus, their results are not necessarily generalizable to other hormone preparations. With transdermal therapy, it may be possible to avoid the increased synthesis of clotting factors and the increase in triglycerides and CRP that may contribute in part to adverse cardiovascular outcomes. Indeed, observational studies have revealed that transdermal estrogen may be less likely than oral estrogen to precipitate VTE (49, 50 ), suggesting that women with risk factors for VTE should choose the former route of delivery if opting for HT. Observational findings also indicate that lower doses of estrogen (one-third to onehalf of conventional doses) may be less likely to
• Glucose tolerance markers (fasting glucose, insulin, homeostatic model assessment-insulin resistance [HOMA-IR], insulin-like growth factor-1, and biomarkers of metabolic syndrome) • Matrix metalloproteinases • Hemostatic markers (D-dimer, factor VIII, von Willebrand factor, homocysteine, fibrinogen, tissue factor pathway inhibitor or acquired activated protein C resistance) • Sex steroid hormone levels, sex hormone binding globulin level • Genetic markers: • Factor V Leiden • Glycoprotein IIIa leu33pro • Gene variants in ABO blood group • Estrogen and progesterone receptor polymorphisms • Gene variants related to sex hormone biosynthesis, metabolism, and/or signaling • Genome-wide association studies (GWAS) and exome sequencing for gene discovery a
Reproduced from Manson (5 ), with permission from Elsevier.
increase stroke risk (51 ). However, evidence from randomized trials is lacking. Similarly, there is a dearth of trials on the impact of alternative hormone formulations, including “bioidentical” hormones, on clinical cardiovascular events. Although more research is needed to address these issues, it may be possible to develop individualized risk prediction models to distinguish women most likely to benefit from transdermal or low-dose HT, compared with oral HT in conventional doses, from women who should avoid any type of HT. Future Directions and Conclusion Exploratory analyses of an array of biomarkers and genetic variants that hold promise for risk stratification to enhance clinical decision-making regarding the use of HT are ongoing in the WHI, other randomized triClinical Chemistry 60:1 (2014) 75
Reviews als, and observational studies (Table 4). Genome-wide association studies are also underway and will aid discovery of gene variants that may modify HT-associated risks. Additional clinical factors are also being examined. This review has focused on studies of potential interactive effects in relation to cardiovascular outcomes, but parallel efforts to understand factors that modify HT-associated risks for breast cancer (52, 53 ) and other diseases are also areas of active investigation. The overall benefit–risk balance of HT is complex and appears to depend on the unique clinical and biological profile of the individual user. New research suggests that it may be possible to identify women who are better or worse candidates for HT, as well as to select the optimal delivery route, dose, and formulation of such therapy, by the use of individual benefit–risk stratification. Data from the WHI and HERS trials suggest an important role for clinical characteristics, serum biomarkers, genomic markers, and gene– environment interactions in developing a personalized approach to the prediction of HT outcomes. Achieving this goal would
improve the quality of healthcare for the millions of women whose lives are adversely impacted by moderate to severe menopausal symptoms.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest: Employment or Leadership: J.E. Manson, guest editor, Clinical Chemistry, AACC. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: None declared. Research Funding: S.S. Bassuk, NIH; J.E. Manson, NIH. Expert Testimony: None declared. Patents: None declared.
References 1. North American Menopause Society. The 2012 Hormone Therapy Position Statement of the North American Menopause Society. Menopause 2012;19:257–71. 2. Santen RJ, Allred DC, Ardoin SP, Archer DF, Boyd N, Braunstein GD, et al. Postmenopausal hormone therapy: an Endocrine Society scientific statement. J Clin Endocrinol Metab 2010;95:S1– S66. 3. Manson JE, Bassuk SS. The menopause transition and postmenopausal hormone therapy. In: Longo DL, Fauci AS, Kasper DL, Hauser SL, Jamison JL, Loscalzo J, eds. Harrison’s Principles of Internal Medicine. New York: McGraw-Hill; 2011. 4. Sprague BL, Trentham-Dietz A, Cronin KA. A sustained decline in postmenopausal hormone use: results from the National Health and Nutrition Examination Survey, 1999 –2010. Obstet Gynecol 2012;120:595– 603. 5. Manson JE. The role of personalized medicine in identifying appropriate candidates for menopausal estrogen therapy. Metabolism 2013; 62 Suppl 1:S15–19. 6. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women’s Health Initiative randomized controlled trial. JAMA 2002;288:321–33. 7. Anderson GL, Limacher M, Assaf AR, Bassford T, Beresford SA, Black H, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women’s Health Initiative randomized controlled trial. JAMA 2004;291:1701–12. 8. Manson JE, Chlebowski RT, Stefanick ML, Aragaki AK, Rossouw JE, Prentice RL, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women’s Health Initiative randomized trials. JAMA 2013;310:1353– 68.
76 Clinical Chemistry 60:1 (2014)
9. Hulley S, Grady D, Bush T, Furberg C, Herrington D, Riggs B, Vittinghoff E. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/progestin Replacement Study (HERS) Research Group. JAMA 1998; 280:605–13. 10. Grady D, Herrington D, Bittner V, Blumenthal R, Davidson M, Hlatky M, et al. Cardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/progestin Replacement Study follow-up (HERS II). JAMA 2002;288: 49 –57. 11. Grady D, Wenger NK, Herrington D, Khan S, Furberg C, Hunninghake D, et al. Postmenopausal hormone therapy increases risk for venous thromboembolic disease. The Heart and Estrogen/progestin Replacement Study. Ann Intern Med 2000; 132:689 –96. 12. Hulley S, Furberg C, Barrett-Connor E, Cauley J, Grady D, Haskell W, et al. Noncardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/progestin Replacement Study follow-up (HERS II). JAMA 2002;288: 58 – 66. 13. Manson JE, Hsia J, Johnson KC, Rossouw JE, Assaf AR, Lasser NL, et al. Estrogen plus progestin and the risk of coronary heart disease. N Engl J Med 2003;349:523–34. 14. Hsia J, Langer RD, Manson JE, Kuller L, Johnson KC, Hendrix SL, et al. Conjugated equine estrogens and the risk of coronary heart disease: the Women’s Health Initiative. Arch Intern Med 2006; 166:357– 65. 15. Rossouw JE, Prentice RL, Manson JE, Wu L, Barad D, Barnabei VM, et al. Effects of postmenopausal hormone therapy on cardiovascular disease by age and years since menopause. JAMA 2007;297: 1465–77. 16. Rossouw JE, Cushman M, Greenland P, Lloyd-
Jones DM, Bray P, Kooperberg C, et al. Inflammatory, lipid, thrombotic, and genetic markers of coronary heart disease risk in the Women’s Health Initiative trials of hormone therapy. Arch Intern Med 2008;168:2245–53. Wassertheil-Smoller S, Hendrix SL, Limacher M, Heiss G, Kooperberg C, Baird A, et al. Effect of estrogen plus progestin on stroke in postmenopausal women: the Women’s Health Initiative: a randomized trial. JAMA 2003;289:2673– 84. Hendrix SL, Wassertheil-Smoller S, Johnson KC, Howard BV, Kooperberg C, Rossouw JE, et al. Effects of conjugated equine estrogen on stroke in the Women’s Health Initiative. Circulation 2006;113:2425–34. Kooperberg C, Cushman M, Hsia J, Robinson JG, Aragaki AK, Lynch JK, et al. Can biomarkers identify women at increased stroke risk? The Women’s Health Initiative Hormone Trials. PLoS Clin Trials 2007;2:e28. Cushman M, Kuller LH, Prentice R, Rodabough RJ, Psaty BM, Stafford RS, et al. Estrogen plus progestin and risk of venous thrombosis. JAMA 2004;292:1573– 80. Curb JD, Prentice RL, Bray PF, Langer RD, Van Horn L, Barnabei VM, et al. Venous thrombosis and conjugated equine estrogen in women without a uterus. Arch Intern Med 2006;166:772– 80. Furberg CD, Vittinghoff E, Davidson M, Herrington DM, Simon JA, Wenger NK, Hulley S. Subgroup interactions in the Heart and Estrogen/ Progestin Replacement Study: lessons learned. Circulation 2002;105:917–22. Herrington DM, Vittinghoff E, Lin F, Fong J, Harris F, Hunninghake D, et al. Statin therapy, cardiovascular events, and total mortality in the Heart and Estrogen/Progestin Replacement Study (HERS). Circulation 2002;105:2962–7. Huang AJ, Sawaya GF, Vittinghoff E, Lin F, Grady D. Hot flushes, coronary heart disease, and hor-
Hormone Therapy Risk Stratification
mone therapy in postmenopausal women. Menopause 2009;16:639 – 43. Manson JE, Allison MA, Rossouw JE, Carr J, Langer RD, Hsia J, et al. Estrogen therapy and coronary artery calcification. N Engl J Med 2007; 356:2591– 602. Grodstein F, Manson JE, Colditz GA, Willett WC, Speizer FE, Stampfer MJ. A prospective, observational study of postmenopausal hormone therapy and primary prevention of cardiovascular disease. Ann Intern Med 2000;133:933– 41. Grodstein F, Manson JE, Stampfer MJ. Hormone therapy and coronary heart disease: the role of time since menopause and age at hormone initiation. J Womens Health 2006;15:35– 44. Salpeter SR, Walsh JM, Greyber E, Salpeter EE. Brief report: coronary heart disease events associated with hormone therapy in younger and older women. A meta-analysis. J Gen Intern Med 2006;21:363– 6. Vickers MR, MacLennan AH, Lawton B, Ford D, Martin J, Meredith SK, et al. Main morbidities recorded in the Women’s International Study of long Duration Oestrogen after Menopause (WISDOM): a randomised controlled trial of hormone replacement therapy in postmenopausal women. BMJ 2007;335:239. Schierbeck LL, Rejnmark L, Tofteng CL, Stilgren L, Eiken P, Mosekilde L, et al. Effect of hormone replacement therapy on cardiovascular events in recently postmenopausal women: randomised trial. BMJ 2012;345:e6409. Mikkola TS, Clarkson TB. Estrogen replacement therapy, atherosclerosis, and vascular function. Cardiovasc Res 2002;53:605–19. Herrington DM, Reboussin DM, Brosnihan KB, Sharp PC, Shumaker SA, Snyder TE, et al. Effects of estrogen replacement on the progression of coronary-artery atherosclerosis. N Engl J Med 2000;343:522–9. Hodis HN, Mack WJ, Azen SP, Lobo RA, Shoupe D, Mahrer PR, et al. Hormone therapy and the progression of coronary-artery atherosclerosis in postmenopausal women. N Engl J Med 2003;349: 535– 45. Waters DD, Alderman EL, Hsia J, Howard BV, Cobb FR, Rogers WJ, et al. Effects of hormone replacement therapy and antioxidant vitamin supplements on coronary atherosclerosis in postmenopausal women: a randomized controlled trial. JAMA 2002;288:2432– 40. Angerer P, Stork S, Kothny W, Schmitt P, von
Schacky C. Effect of oral postmenopausal hormone replacement on progression of atherosclerosis : a randomized, controlled trial. Arterioscler Thromb Vasc Biol 2001;21:262– 8. Hodis HN, Mack WJ, Lobo RA, Shoupe D, Sevanian A, Mahrer PR, et al. Estrogen in the prevention of atherosclerosis. A randomized, doubleblind, placebo-controlled trial. Annual Intern Med 2001;135:939 –53. Harman SM. Effects of oral conjugated estrogen or transdermal estradiol plus oral progesterone treatment on common carotid artery intima media thickness (CIMT) & coronary artery calcium (CAC) in menopausal women: initial results from the Kronos Early Estrogen Prevention Study (KEEPS) [Abstract]. Menopause 2012;19:1365. Kronos Early Estrogen Prevention Study. News: frequently asked questions. http://www. keepstudy.org/news/keeps_faq.cfm (Accessed December 2013). ELITE: Early Versus Late Intervention Trial With Estrogen. Hodis HN, principal investigator. http:// clinicaltrials.gov/ct/show/NCT00114517?order⫽1 (Accessed December 2013). Canonico M, Plu-Bureau G, O’Sullivan MJ, Stefanick ML, Cochrane B, Scarabin PY, Manson JE. Age at menopause, reproductive history, and venous thromboembolism risk among postmenopausal women: the Women’s Health Initiative hormone therapy clinical trials. Menopause [Epub 2013 Jun 10]. Bray PF, Larson JC, LaCroix AZ, Manson J, Limacher MC, Rossouw JE, et al. Usefulness of baseline lipids and C-reactive protein in women receiving menopausal hormone therapy as predictors of treatment-related coronary events. Am J Cardiol 2008;101:1599 – 605. Wild RA, Wu C, Curb JD, Martin LW, Phillips L, Stefanick M, et al. Coronary heart disease events in the Women’s Health Initiative hormone trials: effect modification by metabolic syndrome: a nested case-control study within the Women’s Health Initiative randomized clinical trials. Menopause 2013;20:254 – 60. Rossouw J, Bray P, Liu J, Kooperberg C, Hsia J, Lewis C, et al. Estrogen receptor polymorphisms and the vascular effects of hormone therapy. Arterioscler Thromb Vasc Biol 2011;31:464 –9. Rossouw JE, Prentice RL, Manson JE, Aragaki AK, Hsia J, Martin LW, et al. Relationships of coronary heart disease with 27-hydroxycholesterol, lowdensity lipoprotein cholesterol, and menopausal
hormone therapy. Circulation 2012;126:1577– 86. 45. Rossouw JE, Johnson KC, Pettinger M, Cushman M, Sandset PM, Kuller L, et al. Tissue factor pathway inhibitor, activated protein C resistance, and risk of ischemic stroke due to postmenopausal hormone therapy. Stroke 2012;43:952–7. 46. Shlipak MG, Simon JA, Vittinghoff E, Lin F, Barrett-Connor E, Knopp RH, et al. Estrogen and progestin, lipoprotein(a), and the risk of recurrent coronary heart disease events after menopause. JAMA 2000;283:1845–52. 47. Prentice RL, Paczesny S, Aragaki A, Amon LM, Chen L, Pitteri SJ, et al. Novel proteins associated with risk for coronary heart disease or stroke among postmenopausal women identified by indepth plasma proteome profiling. Genome Med 2010;2:48. 48. Huang Y, Ballinger DG, Stokowski R, Beilharz E, Robinson JG, Liu S, et al. Exploring the interaction between SNP genotype and postmenopausal hormone therapy effects on stroke risk. Genome Med 2012;4:57. 49. Canonico M, Oger E, Plu-Bureau G, Conard J, Meyer G, Levesque H, et al. Hormone therapy and venous thromboembolism among postmenopausal women: impact of the route of estrogen administration and progestogens: the ESTHER study. Circulation 2007;115:840 –5. 50. Canonico M, Fournier A, Carcaillon L, Olie V, Plu-Bureau G, Oger E, et al. Postmenopausal hormone therapy and risk of idiopathic venous thromboembolism: results from the E3N cohort study. Arterioscler Thromb Vasc Biol 2010;30: 340 –5. 51. Grodstein F, Manson JE, Stampfer MJ, Rexrode K. Postmenopausal hormone therapy and stroke: role of time since menopause and age at initiation of hormone therapy. Arch Intern Med 2008; 168:861– 6. 52. Prentice RL, Huang Y, Hinds DA, Peters U, Pettinger M, Cox DR, et al. Variation in the FGFR2 gene and the effects of postmenopausal hormone therapy on invasive breast cancer. Cancer Epidemiol Biomarkers Prev 2009;18:3079 – 85. 53. Mackey RH, Fanelli TJ, Modugno F, Cauley JA, McTigue KM, Brooks MM, et al. Hormone therapy, estrogen metabolism, and risk of breast cancer in the Women’s Health Initiative Hormone Therapy Trial. Cancer Epidemiol Biomarkers Prev 2012;21:2022–32.
Clinical Chemistry 60:1 (2014) 77