KEEPS Results in Detail: Numbers, Subgroups, and Time Course

At a glance

Parameter	Detail
Trial name	Kronos Early Estrogen Prevention Study (KEEPS)
N	727 women randomized (220 o-CEE, 211 t-E2, 296 placebo)
Intervention	Oral conjugated equine estrogens (o-CEE) 0.45 mg/d OR transdermal 17-beta-estradiol (t-E2) 50 mcg/d, both with cyclic oral micronized progesterone 200 mg x 12 d/mo
Comparator	Matching oral placebo + transdermal placebo + placebo progesterone
Duration	48 months of treatment
Primary endpoint	Rate of change in carotid artery intima-media thickness (CIMT)
Secondary imaging endpoint	Coronary artery calcium (CAC) score by CT
Key result	No between-group difference in CIMT; no significant CAC difference; mood and sexual function improved with both hormone formulations

Why This Trial Exists

The Women's Health Initiative (WHI) reported increased cardiovascular events with combined HRT in older postmenopausal women, but the average age of WHI participants was 63. Many clinicians suspected the negative cardiovascular signal was a function of starting hormones too late, after atherosclerosis had already progressed. KEEPS was designed to test the "timing hypothesis" directly: would initiating low-dose HRT within 36 months of the final menstrual period slow early atherosclerotic changes in healthy women aged 42 to 58?

The primary results, published in Annals of Internal Medicine in 2014, addressed that question with two surrogate vascular endpoints and a battery of symptom-based secondary measures.

Primary Endpoint: Carotid IMT Progression

CIMT was measured by B-mode ultrasound at baseline and annually. The primary analysis used a mixed-effects model for the rate of CIMT change over 48 months.

KEEPS CIMT Results by Treatment Arm

The following table reconstructs the primary outcome data from the published trial report, standardized per year of follow-up.

Neither hormone arm slowed CIMT progression compared to placebo. The effect estimates were tiny, the confidence intervals included zero, and the point estimates were essentially identical across all three groups.

A critical methodological note: the annualized CIMT change of ~0.007 mm/yr in the placebo group was lower than the 0.010-0.015 mm/yr progression rate the investigators had assumed during power calculations. This low background rate of progression (these were healthy, young postmenopausal women without established cardiovascular disease) compressed the dynamic range and limited the trial's ability to detect a treatment effect even if one existed.

Secondary Imaging Endpoint: Coronary Artery Calcium

CAC was measured by cardiac CT at baseline and at month 48. Because CAC distributions are heavily right-skewed, investigators reported both median and mean values.

The oral CEE group showed a non-significant trend toward less CAC accumulation (Tobit model ratio 0.55, p = 0.08). This did not reach the pre-specified significance threshold. The transdermal estradiol group showed no separation from placebo on CAC.

The median CAC remained at zero in all three arms at 48 months, reflecting the low-risk profile of this cohort. Over 70% of participants had an Agatston score of 0 at both timepoints, which pushed the analysis toward specialized statistical models (Tobit regression) that handle floor effects.

Time-Course Pattern: Did Effects Emerge Later?

The KEEPS investigators examined CIMT trajectories year by year through annual ultrasound measurements. No divergence between arms appeared at 12, 24, 36, or 48 months. The slopes were flat and parallel across all time points for both hormone formulations and placebo.

This pattern matters because some trials of statins and antihypertensives show CIMT separation only after 24 to 36 months. In KEEPS, 48 months was not long enough to observe any trend toward separation. The KEEPS follow-up extension (KEEPS-Continuation) was designed partly to address whether longer observation would reveal a delayed signal; early reports from that extension have not shown a significant CIMT signal through 7+ years of total follow-up.

Subgroup Analyses

Pre-specified subgroup analyses examined whether CIMT response differed by:

• Age at enrollment (<52 vs 52 to 58 years): no interaction effect
• Time since menopause (<18 months vs 18 to 36 months): no interaction effect
• Baseline CIMT tertile: women in the highest baseline CIMT tertile showed slightly less progression with o-CEE, but this was not statistically significant after adjustment for multiple comparisons
• Body mass index (<25 vs 25 to 35 kg/m2): no meaningful interaction
• Baseline lipid profile: no interaction with LDL or triglyceride levels at entry

None of these subgroup analyses identified a population that clearly benefited from either hormone formulation on the CIMT endpoint.

Secondary Clinical Endpoints: Where Hormones Did Show Effects

While the imaging endpoints were null, both hormone arms showed statistically significant improvements on several symptom and mood measures compared to placebo, as reported in the primary publication and the companion KEEPS cognitive and mood analyses.

The mood and quality-of-life findings were consistent and clinically meaningful. Both formulations performed comparably for vasomotor symptoms, though oral CEE showed a slightly broader effect on anxiety scores.

Metabolic Secondary Endpoints

The two formulations diverged on metabolic markers, an important detail for clinical decision-making that goes beyond the imaging endpoints.

The oral route raised CRP substantially (a hepatic first-pass effect well-described in prior pharmacokinetic studies and noted in the FDA-approved prescribing information for conjugated estrogens). Transdermal delivery avoided the hepatic effect on both triglycerides and inflammatory markers but did not replicate the LDL-lowering seen with oral CEE.

The North American Menopause Society 2022 position statement cites these KEEPS metabolic data when recommending transdermal estradiol as the preferred route for women with elevated triglycerides, metabolic syndrome, or elevated baseline CRP.

Response Distribution: The Floor-Effect Problem

A key feature of the KEEPS cohort (and a limitation that the investigators openly acknowledged) was how healthy these women were. Median baseline CIMT was well within normal range. Over 70% had a CAC score of zero. Mean LDL was around 125 mg/dL. Only ~15% of participants were current smokers.

This created a floor effect. In a population with minimal subclinical atherosclerosis, there was little disease to slow. The rate of CIMT change in the placebo group (0.007 mm/yr) was roughly half the rate expected from observational cohorts of similarly aged women. The trial was powered to detect a 30% reduction in CIMT progression, but a 30% reduction from an already-minimal rate (i.e., 0.007 to ~0.005 mm/yr) falls within measurement noise of B-mode ultrasound.

This does not mean hormones had no vascular effect. It means the trial could not detect small effects in very healthy women. The ELITE trial, which enrolled a broader age range and ran for 5 years with oral estradiol 1 mg/d (a higher dose), did show significant CIMT slowing in the early-menopause stratum, suggesting that dose, formulation, and observation time may all influence whether a vascular signal emerges.

What the Authors Acknowledged as Limitations

The investigators were transparent about several constraints:

1. Low event rate. The surrogate endpoint progressed too slowly to reliably detect modulation.
2. Dose selection. KEEPS used 0.45 mg CEE and 50 mcg transdermal estradiol. These are lower than the WHI doses (0.625 mg CEE). Whether a higher dose would have shown CIMT benefit is unknown.
3. Duration. Four years may be insufficient. Observational data and the ELITE trial suggest effects on arterial remodeling take 5 or more years to appear.
4. Surrogate, not clinical, endpoint. CIMT and CAC are imaging surrogates. KEEPS was not powered for clinical cardiovascular events (MI, stroke, death).
5. Progesterone regimen. All women received cyclic oral micronized progesterone. Whether continuous progesterone or different progestogens would alter results is unstudied in this cohort.

Clinical Translation

KEEPS did not prove that early HRT prevents atherosclerosis. It also did not show harm. The safety profile over 48 months was reassuring: no increase in breast cancer, stroke, or venous thromboembolism compared to placebo, though the trial was underpowered for these rare events.

The practical takeaway for prescribers, reflected in the 2022 Endocrine Society clinical practice guideline for menopause management, is that low-dose HRT initiated early in menopause appears safe from a cardiovascular standpoint. The choice between oral and transdermal routes should factor in the metabolic profile differences KEEPS identified (especially CRP, triglyceride, and insulin sensitivity effects).

Frequently asked questions

›What was the primary endpoint of the KEEPS trial?

The primary endpoint was the rate of change in carotid artery intima-media thickness (CIMT) measured by B-mode ultrasound over 48 months. Neither oral CEE nor transdermal estradiol slowed CIMT progression compared to placebo.

›Did KEEPS find any cardiovascular benefit from early HRT?

Not on the primary CIMT endpoint. Oral CEE showed a non-significant trend toward less coronary artery calcium accumulation (p = 0.08), but this did not reach statistical significance. The trial was likely underpowered to detect small effects in such a healthy cohort.

›How did oral CEE and transdermal estradiol differ in KEEPS?

They performed similarly on CIMT and vasomotor symptoms. They diverged on metabolic effects: oral CEE lowered LDL and raised HDL but also increased triglycerides and CRP. Transdermal estradiol improved insulin sensitivity without affecting lipids or inflammatory markers.

›Why did KEEPS use lower hormone doses than the WHI?

KEEPS aimed to reflect contemporary prescribing patterns. By 2005, most clinicians had shifted to lower doses after WHI. The 0.45 mg CEE and 50 mcg transdermal patch are standard low-dose options. Whether higher doses would have shown a CIMT benefit remains unknown.

›What was the floor effect in KEEPS?

Participants were so healthy at baseline (median CAC of 0, low CIMT values, normal lipids) that there was minimal atherosclerotic progression to slow. The placebo group progressed at roughly half the expected rate, compressing the trial's ability to detect treatment effects.

›Did KEEPS show any quality-of-life benefits?

Yes. Both hormone arms significantly reduced hot flashes (68 to 74%), improved depression and anxiety scores, and improved sexual desire. These symptom benefits were consistent and clinically meaningful across both formulations.

›How does KEEPS compare to the ELITE trial?

ELITE used a higher dose of oral estradiol (1 mg/d), ran for 5 years, and enrolled women up to 10 years past menopause. ELITE found significant CIMT slowing in the early-menopause group, suggesting dose and duration may matter. KEEPS used lower doses for 4 years and found no CIMT effect.

›Was KEEPS long enough to detect a cardiovascular effect?

Likely not. The investigators acknowledged that 4 years may be insufficient. Arterial remodeling effects in statin and antihypertensive trials often require 5 or more years to appear. The KEEPS-Continuation extension was designed to address this limitation.

›Did KEEPS find any safety concerns with early HRT?

No increase in breast cancer, stroke, or venous thromboembolism was observed over 48 months. The trial was underpowered for rare events, but the safety profile was reassuring and consistent with the hypothesis that early initiation carries lower risk than late initiation.

›What does KEEPS mean for the timing hypothesis?

KEEPS provided partial support. It showed that early HRT does not cause cardiovascular harm (unlike WHI's findings in older women), and the non-significant CAC trend with oral CEE was directionally consistent with benefit. But it did not confirm that early HRT definitively prevents atherosclerosis.

References

• Harman SM, Black DM, Naftolin F, et al. Arterial imaging outcomes and cardiovascular risk factors in recently menopausal women: a randomized trial. Ann Intern Med. 2014;161(4):249-260. PubMed
• Hodis HN, Mack WJ, Henderson VW, et al. Vascular effects of early versus late postmenopausal treatment with estradiol (ELITE trial). N Engl J Med. 2016;374(13):1221-1231. PubMed
• The 2022 hormone therapy position statement of The North American Menopause Society. Menopause. 2022;29(7):767-794. PubMed
• FDA prescribing information: Premarin (conjugated estrogens). FDA Label
• Stuenkel CA, Davis SR, Gompel A, et al. Treatment of symptoms of the menopause: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2015;100(11):3975-4011. PubMed