Hypertension is a major risk factor for cardiovascular morbidity and mortality in post-menopausal women.¹ The menopausal transition accounts for changes that can increase the likelihood of hypertension and other cardiovascular risk factors. Indeed, changes in the endogenous sex hormones and in the physiology of aging itself may affect the cardiac function, cause arterial stiffness and insulin resistance, alter one’s lipid profile, and increase one’s bodyweight and central adiposity.^{1 , 2}

Although menopausal hormone therapy (MHT) is the most effective treatment for vasomotor symptoms and for the genitourinary syndrome of menopause, and is a very effective treatment for the prevention of bone loss and fractures,³ other effects are also involved, and this treatment may be associated with cardiovascular risk markers.^{4 , 5} The risks and benefits of MHT use seem to depend on the type of hormone prescribed, the dose and the duration of use, the route of administration, and the moment at which treatment is begun.³

Findings regarding the effect of MHT on arterial blood pressure in women have been conflicting, with clinical trials reporting either a neutral effect^{6 , 7} or a protective effect with a reduction in blood pressure,^{8 , 9} while others with the same design have suggested a harmful effect with an increase in blood pressure levels.^{10 , 11}

Because most of the studies dealing with this subject were performed with samples of North American and European women, there is a need to evaluate the effect of MHT on blood pressure in Brazilian women. Based on the hypothesis that MHT affects blood pressure levels, this study aimed to evaluate the relationship between MHT use and hypertension in women participating in the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil).

The median age of the 2,138 women participating in the study was 57 years (IQR 53-62). According to self-reports, 1,492 (69.8%) were never users of MHT, while 457 (21.4%) were past users and 189 (8.8%) were current users.

MHT use was more common in women with a BMI <25 kg/m²and triglyceride levels <150 mg/dl, and in less physically inactive women, non-smokers and non-diabetics. Of the past users, 59.7% were ≥60 years of age, while 54.5% of the current users were 50-59 years of age ( Table 1 ).

Table 1

Characteristics	Never Users n (%)	Past Users n (%)	Current Users n (%)	p-value
Age				0.000
40-49 years	140 (9.4)	16 (3.5)	20 (10.5)
50-59 years	900 (60.3)	168 (36.8)	103 (54.5)
≥60 years	452 (30.3)	273 (59.7)	66 (35.0)
Ethnicity/skin color				0.000
Black	272 (19.4)	54 (12.6)	17 (9.1)
Brown	384 (27.4)	114 (26.6)	45 (24.2)
White	747 (53.2)	260 (60.8)	124 (66.7)
Schooling				0.000
High school	805 (54.0)	192 (42.0)	50 (26.5)
University degree	687 (46.0)	265 (58.0)	139 (73.5)
* Excessive alcohol consumption				0.243
No	1.436 (96.5)	444 (97.4)	179 (94.7)
Yes	52 (3.5)	12 (2.6)	10 (5.3)
Smoking				0.000
Never smoked	821 (55.0)	289 (63.2)	106 (56.1)
Former smoker	427 (28.6)	120 (26.3)	68 (36.0)
Smoker	244 (16.4)	48 (10.5)	15 (7.9)
Physical activity				0.001
Inactive	1.204 (81.6)	337 (74.6)	138 (73.4)
Active	271 (18.4)	115 (25.4)	50 (26.6)
Body mass index				0.000
≤ 24.9 kg/m2	502 (33.6)	185 (40.5)	99 (52.4)
25.0 - 29.9 kg/m2	517 (34.7)	183 (40.0)	71 (37.6)
≥ 30.0 kg/m2	473 (31.7)	89 (19.5)	19 (10.0)
Diabetes				0.007
No	1.128 (75.7)	365 (79.9)	160 (84.7)
Yes	363 (24.3)	92 (20.1)	29 (15.3)
Arterial hypertension				0.000
No	880 (59.0)	260 (57.0)	139 (73.5)
Yes	612 (41.0)	197 (43.0)	50 (26.5)
Cardiovascular disease				0.581
No	1.412 (94.8)	431 (94.3)	182 (96.3)
Yes	78 (5.2)	26 (5.7)	7 (3.7)
LDL-cholesterol				0.396
<130 mg/dL	648 (43.4)	215 (47.0)	84 (44.4)
≥130 mg/dL	844 (56.6)	242 (53.0)	105 (55.6)
HDL–cholesterol				0.041
>40 mg/dL	1.452 (97.3)	446 (97.6)	178 (94.2)
≤40 mg/dL	40 (2.7)	11 (2.4)	11 (5.8)
Triglycerides				0.000
<150 mg/dL	1.044 (70.0)	351 (76.8)	158 (84.0)
≥150 mg/dL	448 (30.0)	106 (23.2)	31 (16.0)

* Excessive alcohol consumption: ≥140grams of alcohol/week; HDL: high- density lipoprotein; LDL: low-density lipoprotein.

The prevalence of hypertension was 40.2%. Of the hypertensive women, 71.3% had never used MHT, while 5.8% were current MHT users. Of the normotensive women, 68.8% had never used MHT, while 10.9% were current users.

Table 2 shows the crude and the age- and BMI-adjusted associations between MHT use and the presence of hypertension. Current MHT users were significantly less likely to have hypertension (OR=0.59; 95% CI: 0.41-0.85) when compared to never users. This adjusted inverse association persisted even after making further adjustment for the route of administration (data not presented in tables).

Table 2

	Never user n=1492	Past user n=457	Current user n=189
Arterial hypertension n (%)
No	880 (68.8)	260 (20.3)	139 (10.9)
Yes	612 (71.3)	197 (22.9)	50 (5.8)
OR (95%CI)
Crude	1	1.08 (0.88-1.34)	0.51 (0.36-0.72)
Adjusted*	1	0.89 (0.71-1.13)	0.59 (0.41- 0.85)

OR: Odds Ratio; *Adjusted for age and body mass index.

In the comparative analysis of blood pressure levels according to exposure to MHT, considering hypertensive (using or not antihypertensive) and normotensive, results showed that current MHT users had the lowest median systolic blood pressure of 113mmHg, as compared to never users at 118.5mmHg, and to past users at 120mmHg (p=0.001). Furthermore, the upper limit was notably lower ( Figure 1 ). Statistically significant differences were found only between never/current use (p=0.00) and between current use/past use (p=0.00).

Figure 1

Of the current and past users of MHT, the majority had begun treatment at or before 59 years of age, within 10 years of experiencing menopause, and the duration of therapy was up to 5 years, regardless of hypertension. Nevertheless, the proportion of hypertensive women was greatest among those who began MHT after 60 years of age and/or 10 years or more after menopause ( Table 3 ).

Table 3

Characteristics	Normotensive women n (%)	Hypertensive women n (%)	p-value
Age at beginning of MHTa			0.034
<60 years	377(63.3)	219(36.7)
≥60 years	9(40.9)	13(59.1)
Duration of menopause at beginning of MHT			0.000
<10 years	378 (63.7)	215 (36.3)
≥10 years	3 (17.6)	14 (82.4)
Duration of MHT			0.927
<5 years	252(62.4)	152(37.6)
≥5 years	142(62.0)	87(38.0)

a Menopausal hormone therapy. NB: Specifically for this analysis, exclusions of some observations were necessary due to missing data and for this reason, the sum may vary for the different variables.

In the group of current users who had hypertension, the most common MHT type consisted of combined estrogen-progestogen formulations followed by estrogen alone. However, in normotensive users of MHT, tibolone was also widely used, as well as the combined estrogen-progestogen formulations. All the different types of MHT were more common in the normotensive women compared to the hypertensive women. The majority of the women (80.3%) were found to use the oral route of administration; nevertheless, there was no statistically significant association between the route of administration and the presence of hypertension ( Table 4 ).

Table 4

Characteristics	Normotensive women (n=138)	Hypertensive women (n=50)	p-value
	n (%)	n (%)
Type of hormone			0.024*
Estrogen + progestogen	65 (47.1)	18 (36.0)
Estrogens	28 (20.3)	16 (32.0)
Progestogens	3 (2.2)	-
Estrogens + testosterone	5 (3.6)	3 (6.0)
Tibolone	35 (25.4)	8 (16.0)
Others	2 (1.4)	5 (10.0)
Route of administration			0.190
Oral	114 (82.6)	37 (74.0)
Non-oral	24 (17.4)	26.0)

* Fisher’s exact test. NB: One participant was excluded due to missing data.

The results indicate that the use of MHT is not related to arterial hypertension. MHT users were less likely to have hypertension compared to past or never users, regardless of age or BMI. Nevertheless, these findings must be analyzed with caution.

The possibility of women with health problems being less likely to be prescribed hormones, cannot be ruled out. MHT users had a more favorable health profile, being healthier in almost all the parameters evaluated here. A study conducted within the ELSA-Brasil showed that women with at least one clinical contraindication to MHT were less likely to be exposed to this type of medication.²³ Therefore, the prescription of MHT may have been more restricted in the case of women with hypertension, since, although hypertension alone is not a formal contraindication, it is frequently associated with diseases for which hormone use would be contraindicated. Nonetheless, the present results are in agreement with the findings of the Rancho Bernardo Study conducted in California with 1,044 women, in which the blood pressure levels of participants in current use of MHT were lower than those of a control group.²⁴

The Baltimore study, with a 10-year follow-up time, found that although systolic blood pressure levels increased in both the users and non-users of MHT, the increase was less expressive in the users.²⁵ In the present study, differences in median blood pressure levels were also found between users and non-users of MHT, particularly in relation to systolic blood pressure, with a difference of 5.5 mmHg between current and never users. However, in a randomized clinical trial in which variations in blood pressure were determined by ambulatory blood pressure monitoring (ABPM), a decrease was found both in systolic and diastolic blood pressure in MHT users.⁸

A study conducted in Finland evaluated the effect of the different routes of administration of MHT. Although both the oral and transdermal routes of administration resulted in a decrease in daytime systolic blood pressure, this reduction was maintained for longer (6 months) with the oral route. However, that study only analyzed the short-term effect.⁹ In the present sample, although most of the MHT users used the oral route of administration, no significant differences were found between the hypertensive and normotensive women as a function of the route of administration.

Endogenous estrogen is believed to act through a physiological mechanism that can promote a reduction in arterial blood pressure via a vasodilatory effect, such as an increase in nitric oxide, inhibition of the renin-angiotensin system, a reduction in the transcription of angiotensin-converting enzyme, and the regulation of vasoconstrictors, such as endothelin.^{2 , 26} Nevertheless, despite this apparent benefit of endogenous sex hormones on women’s cardiac health, studies on the effect of the exogenous use of these substances on blood pressure levels have generated conflicting results.

A cross-sectional Australian study that included women of 45-75 years of age found that hormone use was associated with a significantly greater likelihood of having hypertension.²⁷ Furthermore, the Women’s Health Initiative (WHI) clinical trial, which evaluated women of 50-79 years of age, found that MHT led to a small increase in systolic blood pressure over a follow-up period of approximately 5.2 years.¹⁰ Conversely, neither the Postmenopausal Estrogen/Progestin Intervention (PEPI), which followed women of 45-64 years of age over a three-year period,²⁸ nor a study conducted in Denmark⁶ with women of the same age group, found any effect of MHT on arterial blood pressure.

The differences found in the previous studies could be explained first by the variations in the populations, whose ages ranged from 45 to 79 years. Secondly, the regimens, dosages, and hormonal formulations differed, and follow-up times ranged from 6 months to 10 years.^{6 , 8 - 10} Finally, the definitions of hypertension and methods of blood pressure measurement varied, with home monitoring,^{10 , 25 , 28} ABPM,^{8 , 9} and self-reporting²⁷ being used.

In those clinical trials that reported an association between MHT and a reduction in blood pressure or a neutral effect, sample sizes were small, the women enrolled were younger, and participants were followed up for a maximum period of one year.^{7 , 8} Conversely, in those in which an increase in blood pressure was reported, studies tended to have larger sample sizes, involve longer periods of follow-up (up to five years), and be conducted in older women¹⁰ or in women with prior coronary heart disease.¹¹

In addition to the association found in the present study between MHT and a lesser likelihood of hypertension, the low prevalence of MHT is noteworthy. Only 8.8% of the women were current users, a finding that was expected when bearing in mind that the data from this study were generated some years after the publication of the Heart and Estrogen/Progestin Replacement Study (HERS) and the WHI study. Those publications emphasized the risks of MHT and contributed to a considerable reduction in its use, with restrictions for the prescription of MHT and the establishment of criteria for treatment.^{10 , 11}

The pattern of MHT use seen here is in agreement with current recommendations, since most users were under 60 years of age, when the risk-benefit ratio of MHT appears to be more favorable, had initiated therapy within 10 years of the menopause, and had used the treatment for periods of up to 5 years.^{3 , 4 , 29} In the women who initiated hormone use later, the frequency of hypertension was found to be greater; however, those women constituted a minority in this sample. The fact that the recommended time limits are being respected probably offers some protection to MHT users.

The MHT users in the present study had better health conditions, a healthier lifestyle and a better education level. A similar profile was found in a study conducted in Pennsylvania.³⁰ Bearing in mind the pattern of health indicators among the users of MHT, the possibility has to be taken into account that the association between hormone use and a lesser likelihood of hypertension could have been affected by the health profile of these women and not only by the effect of MHT.

Despite the apparent benefit of MHT found here, it is important to emphasize that in accordance with current recommendations, MHT is only indicated for the treatment of the vasomotor symptoms of menopause and not as a strategy to prevent cardiovascular diseases and their risk factors.^{3 , 4}

One of the strengths of the present study is its substantial sample size and the fact that the sample consisted of women from three large geographic regions of the country. Nevertheless, caution is required when making generalizations, since the ELSA-Brasil, despite its robust sample and the known similarities between the results of this study and those of population-based surveys conducted in Brazil, consists of civil servants who are not representative of the general public insofar as their sociodemographic characteristics are concerned.

Other limitations include the methodological impossibility of evaluating reverse causality in the associations observed here, as well as a possible memory bias with respect to the data concerning menopause and the beginning of MHT use, which were obtained using questionnaires. Nevertheless, any bias that may have occurred would be minimal, since the menopause is an important event in women’s lives. In addition, some factors not evaluated, such as sodium intake, kidney function, and the dose used in the hormone regimens, could have led to residual confounding.

These results suggest that current MHT use is not related to hypertension, particularly in women with a healthy lifestyle and those under 60 years of age; however, future studies may clarify the effect of MHT on arterial blood pressure. Despite the ethical issues surrounding studies on MHT due to the delicate risk/benefit balance, longitudinal studies may be more appropriate to evaluate this association and may also include the possibility of identifying long-term effects following the end of MHT use.