Abstract
This observational study examines whether hormone therapy use modifies the association of lipoprotein(a) with coronary heart disease.
Lipoprotein(a) (Lp[a]) is associated with atherosclerotic cardiovascular disease (ASCVD) events with preclinical evidence suggestive of causality.1 Menopausal hormone therapy (HT) has been suggested to lower Lp(a) levels in women by 15% to 20%.2 However, HT also increases C-reactive protein and prothrombin levels, and systemic inflammation may be synergistic with Lp(a) for increased ASCVD risk.3 A previous study of approximately 28 000 women enrolled nearly 30 years ago suggested Lp(a)-associated ASCVD risk was lessened among women using HT.2 Using a large, contemporary cohort, we reexamined whether HT use modifies the association of Lp(a) with coronary heart disease (CHD).
Methods
The Mass General Brigham institutional review board approved these analyses. We included patients in the UK Biobank who were postmenopausal at enrollment from 2006 to 2010; had complete reproductive history and covariate data; lacked preexisting ASCVD, heart failure, aortic stenosis, or venous thromboembolism; and underwent phlebotomy with measurement of Lp(a) at enrollment. Ever use and current use of HT were ascertained from UK Biobank field IDs 2814 and 6153/3546, respectively. Incident CHD during follow-up was ascertained using qualifying International Classification of Diseases codes as used previously.1,4 Cox proportional hazards regression models tested the association of Lp(a) with CHD in the overall sample and stratified by current vs previous vs never use of HT, with adjustment for age, age squared, race, socioeconomic status, current/former tobacco use, type 2 diabetes, natural/surgical menopause at younger than 40 years, body mass index, and systolic blood pressure; aspirin, antihypertensive, and cholesterol-lowering medication use; total and high-density lipoprotein cholesterol; and C-reactive protein. We generated individual-level LPA genetic risk scores using 43 single-nucleotide variation significantly associated with Lp(a) in data sets external to the UK Biobank.1 In sensitivity analyses, we assessed (1) a subcohort propensity score–matched for use of HT (previous and never users each matched 3:1 to current HT users) using all covariates plus LPA genetic risk with the nearest-neighbor algorithm in ‘MatchIt’ (R 3.6.0) and (2) LPA genetic scores in lieu of measured Lp(a). Two-sided P values were statistically significant at .05.
Results
Of 114 028 postmenopausal patients with measured Lp(a), 88 266 met inclusion criteria (mean age, 60.0 years). At enrollment, 4550 (5.2%) were using HT, and 34 396 (39.0%) reported previous HT use. The median Lp(a) was lower in HT users than nonusers, but the median LPA genetic score was higher in HT users (Table 1). After adjustment for age, cholesterol-lowering medication use, and LPA genetic risk, current HT users had lower Lp(a) than previous HT users by 7.5 nmol/L (95% CI, −8.5 to −6.4 nmol/L; P < .001) and never HT users by 7.9 nmol/L (95% CI, −8.9 to −6.8 nmol/L; P < .001).
Table 1. Baseline Characteristics of Study Participants.
| Characteristic | No. (%) | |||||
|---|---|---|---|---|---|---|
| Full cohort (N = 88 266) | Propensity score–matched subcohort (n = 31 437) | |||||
| Currently using HT (n = 4550) | Previously used HT (n = 34 396) | Never used HT (n = 49 320) | Currently using HT (n = 4491) | Previously used HT (n = 13 473) | Never used HT (n = 13 473) | |
| Age, mean (SD), y | 57.6 (6.1) | 61.7 (4.6) | 59.0 (5.5) | 57.6 (6.1) | 58.8 (4.7) | 57.6 (5.8) |
| White | 4413 (97.0) | 33 443 (97.2) | 46 535 (94.4) | 4360 (97.1) | 13 104 (97.3) | 13 083 (97.1) |
| Townsend deprivation index, median (IQR) | −2.1 (−3.6 to 0.4) | −2.4 (−3.7 to 0) | −2.3 (−3.7 to 0.2) | −2.1 (−3.6 to 0.4) | −2.2 (−3.7 to 0.3) | −2.2 (−3.6 to 0.4) |
| Smoking status | ||||||
| Current | 491 (10.8) | 1759 (6.0) | 2289 (4.4) | 486 (10.8) | 1509 (11.2) | 1517 (11.3) |
| Former | 1759 (38.7) | 12 904 (43.8) | 18 538 (35.9) | 1735 (38.7) | 5132 (38.2) | 5220 (38.9) |
| Never | 2300 (50.5) | 14 811 (50.3) | 30 804 (59.7) | 2270 (50.5) | 6789 (50.6) | 6697 (49.9) |
| Type 2 diabetes | 42 (0.9) | 519 (1.5) | 688 (1.4) | 42 (0.9) | 147 (1.1) | 115 (0.8) |
| Premature menopause (age <40 y) | 425 (9.3) | 1929 (5.6) | 910 (1.8) | 421 (9.3) | 996 (7.4) | 819 (6.1) |
| Natural | 201 (4.4) | 816 (2.4) | 491 (1.0) | 199 (4.4) | 440 (3.3) | 423 (3.1) |
| Surgical | 224 (4.9) | 1113 (3.2) | 419 (0.8) | 222 (4.9) | 556 (4.1) | 396 (2.9) |
| Aspirin use | 384 (8.4) | 3883 (11.3) | 4095 (8.3) | 381 (8.5) | 1242 (9.2) | 1110 (8.2) |
| Cholesterol-lowering medication use | 387 (8.5) | 5343 (15.5) | 5574 (11.3) | 380 (8.5) | 1257 (9.3) | 1101 (8.2) |
| Antihypertensive medication use | 652 (14.3) | 7592 (22.1) | 8458 (17.1) | 640 (14.5) | 2099 (15.6) | 1882 (14.0) |
| BMI, mean (SD) | 25.9 (4.4) | 27.1 (4.7) | 26.9 (5.1) | 26.0 (4.4) | 26.2 (4.5) | 25.9 (4.6) |
| Systolic blood pressure, mean (SD), mm Hg | 136.6 (19.6) | 141.2 (19.8) | 139.8 (20.5) | 136.6 (19.6) | 137.7 (19.7) | 136.3 (20.2) |
| Cholesterol, mean (SD), mg/dL | ||||||
| Total | 223.4 (40.6) | 235.7 (42.7) | 235.6 (42.3) | 223.4 (40.5) | 227.2 (38.8) | 222.7 (39.2) |
| HDL | 62.1 (15.5) | 62.9 (14.7) | 63.1 (14.7) | 62.1 (15.5) | 62.3 (14.4) | 62.0 (14.2) |
| High-sensitivity C-reactive protein, median (IQR), mg/dL | 0.18 (0.09 to 0.36) | 0.15 (0.07 to 0.30) | 0.13 (0.06 to 0.28) | 0.18 (0.09 to 0.36) | 0.13 (0.07 to 0.29) | 1.12 (0.06 to 0.28) |
| Measured Lp(a), median (IQR), nmol/L | 18.8 (8.8 to 61.2) | 23.3 (10.4 to 60.4) | 23.9 (10.5 to 62.3) | 18.7 (8.8 to 61.1) | 24.8 (10.6 to 67.8) | 25.0 (10.8 to 71.8) |
| Lp(a) genetic risk score, median (IQR), nmol/L | 17.2 (3.7 to 56.8) | 15.1 (1.9 to 41.9) | 15.1 (1.9 to 43.0) | 17.2 (3.7 to 56.8) | 17.0 (3.4 to 53.5) | 18.3 (3.7 to 59.3) |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); HDL, high-density lipoprotein; HT, hormone therapy; Lp(a), lipoprotein(a).
SI conversion factor: To convert cholesterol to millimoles per liter, multiply by 0.0259; C-reactive protein to milligrams per liter, multiply by 10.
Throughout a median (IQR) follow-up of 11.1 (10.4-11.7) years, 3537 incident CHD events occurred. Individuals in the highest vs lowest Lp(a) quartiles had the highest risk of CHD (adjusted hazard ratio, 1.15; 95% CI, 1.05-1.26; P = .003). Hazards associated with greater Lp(a) appeared larger in those who previously used HT (P for interaction = .046) and those currently taking HT (P for interaction = .04) vs those who never used HT (Table 2). Observations were consistent in the propensity-matched subcohort and when examining LPA genetic risk scores in lieu of measured Lp(a), albeit with no significant interactions.
Table 2. Association of Lp(a) With Incident Coronary Heart Diseasea.
| Factor | Quartile of measured Lp(a) or Lp(a) genetic risk | Per 50 nmol/L of Lp(a) | |||
|---|---|---|---|---|---|
| First | Second | Third | Fourth | ||
| Measured Lp(a), full cohort b | |||||
| Overall | |||||
| Events/person-years of follow-up | 832/239 748 | 824/239 940 | 903/238 945 | 978/239 351 | 3537/957 984 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.99 (0.90-1.09) | 1.06 (0.96-1.16) | 1.15 (1.05-1.26)c | 1.07 (1.04-1.11)d |
| Current HT use | |||||
| Events/person-years of follow-up | 38/15 211 | 41/12 339 | 45/9575 | 51/12 295 | 175/49 421 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.28 (0.82-2.00) | 1.66 (1.08-2.58)c | 1.56 (1.02-2.38)c | 1.11 (0.93-1.24) |
| Previous HT use | |||||
| Events/person-years of follow-up | 389/93 121 | 403/93 658 | 409/94 240 | 486/91 581 | 1687/372 609 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.05 (0.91-1.20) | 1.04 (0.91-1.20) | 1.27 (1.11-1.45)d | 1.11 (1.06-1.16)d |
| Never used HT | |||||
| Events/person-years of follow-up | 405/131 406 | 380/133 944 | 449/135 130 | 441/135 475 | 1675/535 954 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.91 (0.79-1.05) | 1.01 (0.89-1.16) | 1.00 (0.87-1.14) | 1.03 (0.98-1.08) |
| Measured Lp(a), propensity-matched cohort e | |||||
| Overall | |||||
| Events/person-years of follow-up | 247/85 811 | 246/85 912 | 307/85 562 | 317/85 732 | 1117/343 016 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.99 (0.83-1.18) | 1.19 (1.01-1.41)c | 1.24 (1.05-1.47)c | 1.08 (1.03-1.14)c |
| Current HT use | |||||
| Events/person-years of follow-up | 38/15 087 | 41/12 350 | 46/10 086 | 49/11 268 | 174/48 791 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.26 (0.81-1.96) | 1.59 (1.03-2.45)c | 1.60 (1.05-2.45)c | 1.07 (0.92-1.23) |
| Previous HT use | |||||
| Events/person-years of follow-up | 113/35 779 | 117/36 795 | 140/38 238 | 159/36 556 | 529/147 368 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.00 (0.77-1.30) | 1.10 (0.86-1.42) | 1.33 (1.04-1.69)c | 1.12 (1.04-1.21)c |
| Never used HT | |||||
| Events/person-years of follow-up | 96/34 946 | 88/36 767 | 121/37 237 | 109/37 908 | 414/146 858 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.86 (0.64-1.15) | 1.13 (0.86-1.48) | 1.00 (0.76-1.32) | 1.03 (0.94-1.13) |
| Lp(a) genetic risk score f | |||||
| Overall | |||||
| Events/person-years of follow-up | 851/236 682 | 816/228 818 | 880/243 630 | 951/237 022 | 3498/946 151 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.99 (0.90-1.09) | 1.00 (0.91-1.10) | 1.14 (1.04-1.25)c | 1.07 (1.04-1.11)d |
| Current HT use | |||||
| Events/person-years of follow-up | 26/10 554 | 43/11 667 | 50/12 409 | 55/1303 | 174/48 791 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.72 (1.05-2.81)c | 1.80 (1.12-2.90)c | 1.72 (1.08-2.75)c | 1.09 (0.94-1.25) |
| Previous HT use | |||||
| Events/person-years of follow-up | 424/93 284 | 364/90 448 | 409/94 153 | 474/90 541 | 1671/368 437 |
| Hazard ratio (95% CI) | 1 [Reference] | 0.89 (0.77-1.02) | 0.96 (0.84-1.10) | 1.17 (1.03-1.34)c | 1.11 (1.06-1.17)d |
| Never used HT | |||||
| Events/person-years of follow-up | 401/132 844 | 409/126 703 | 421/137 067 | 422/132 320 | 1653/528 934 |
| Hazard ratio (95% CI) | 1 [Reference] | 1.06 (0.93-1.22) | 0.99 (0.86-1.14) | 1.07 (0.93-1.23) | 1.04 (0.98-1.09) |
Abbreviation: Lp(a), lipoprotein(a).
Models are adjusted for age, age squared, race, socioeconomic status, smoking status, type 2 diabetes, natural and surgical premature menopause status, body mass index, systolic blood pressure, aspirin use, antihypertensive medication use, cholesterol-lowering medication use, total cholesterol, high-density lipoprotein cholesterol, C-reactive protein, and (in overall models only) current vs previous vs never use of hormone therapy.
First quartile: 3.80-10.31 nmol/L; second quartile: 10.32-23.35 nmol/L; third quartile: 23.36-61.49 nmol/L; fourth quartile: 61.50-189.00 nmol/L.
P < .05.
P < .001.
First quartile: 3.80-10.39 nmol/L; second quartile: 10.40-23.90 nmol/L; third quartile: 23.91-68.69 nmol/L; fourth quartile: 68.70-189.00 nmol/L.
First quartile: <2.15 nmol/L; second quartile: 2.15-15.04 nmol/L; third quartile: 15.05-42.99 nmol/L; fourth quartile: ≥43.00 nmol/L.
Discussion
In a primary prevention cohort of 90 000 postmenopausal patients in midlife, higher Lp(a) values were associated with increased risk of incident CHD. Consistent with prior observations,2 HT use was associated with modest Lp(a) reduction. However, we found no evidence of lower Lp(a)-associated risks in HT users vs nonusers, in contrast with prior work.2 Differences between study findings likely stem from stark differences in HT use patterns before and after publication of the Women’s Health Initiative and other randomized trials.5 HT is not recommended for the indication of ASCVD risk reduction, but current consensus holds that HT is reasonable for treatment of vasomotor symptoms in female individuals who are younger than 60 years, less than 10 years from menopause, and without high estimated ASCVD risk.6
This study has limitations. As the study population was more than 95% White, whether findings extend to individuals of other races and ethnicities requires further study. Overall, our findings suggest that Lp(a) levels remain prognostic in individuals taking menopausal HT. Clinicians should not be dissuaded from measuring Lp(a) in postmenopausal individuals if clinically indicated irrespective of HT use.
References
- 1.Trinder M, Uddin MM, Finneran P, Aragam KG, Natarajan P. Clinical utility of lipoprotein(a) and LPA genetic risk score in risk prediction of incident atherosclerotic cardiovascular disease. JAMA Cardiol. 2020;6(3):1-9. doi: 10.1001/jamacardio.2020.5398 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Suk Danik J, Rifai N, Buring JE, Ridker PM. Lipoprotein(a), hormone replacement therapy, and risk of future cardiovascular events. J Am Coll Cardiol. 2008;52(2):124-131. doi: 10.1016/j.jacc.2008.04.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Puri R, Nissen SE, Arsenault BJ, et al. Effect of C-reactive protein on lipoprotein(a)-associated cardiovascular risk in optimally treated patients with high-risk vascular disease: a prespecified secondary analysis of the ACCELERATE trial. JAMA Cardiol. 2020;5(10):1136-1143. doi: 10.1001/jamacardio.2020.2413 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Honigberg MC, Zekavat SM, Aragam K, et al. Association of premature natural and surgical menopause with incident cardiovascular disease. JAMA. 2019;322(24):2411-2421. doi: 10.1001/jama.2019.19191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Rossouw JE, Anderson GL, Prentice RL, et al. ; Writing Group for the Women’s Health Initiative Investigators . 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(3):321-333. doi: 10.1001/jama.288.3.321 [DOI] [PubMed] [Google Scholar]
- 6.O’Kelly AC, Michos ED, Shufelt CL, et al. Pregnancy and reproductive risk factors for cardiovascular disease in women. Circ Res. 2022;130(4):652-672. doi: 10.1161/CIRCRESAHA.121.319895 [DOI] [PMC free article] [PubMed] [Google Scholar]