Oral contraceptives: a risk factor for uncontrolled blood pressure among hypertensive women

1. Introduction 

It is estimated that 60 million women around the world use oral contraceptives (OC) [1]. In the United States, approximately 34.5 million women (59%) use contraception. Sterilization and OC are the most common methods (30% and 29% of users, respectively) [2]. The rates in Brazil are similar: 53% of women use contraception, with OC and tubal ligation comprising 85% of the methods of contraception used by Brazilian women [3].

Despite the many attempts to reduce unwanted side effects of OC, some still persist. Increased blood pressure is among the most commonly cited harmful side effects [4], [5] and hypertension is cited as a relative contraindication for OC use [6].

The majority of studies documenting an association between hypertension and OC use involved first generation OC (50 μg or more of estrogen). It was suggested that the incidence of hypertension would be lower with lower estrogen doses [7], [8], but significant rise in blood pressure has also been verified with monophasic OC containing 30 μg of estrogen [9].

There has been disagreement about the component of OC that is responsible for increasing blood pressure. The effect was initially attributed to estrogens alone, but when clinical studies were analyzed, no significant relationship was found between different concentrations of estrogen and change in blood pressure. There is evidence that progestogens may raise blood pressure or reinforce the effect of the estrogens [10], [11].

The wide use of OC and the prevalence of hypertension justify investigations on the subject. In a sample in Porto Alegre, 81.8% (n = 342) of women of reproductive age were using OC, and 5.7% of them were hypertensive (blood pressure [BP] ≥ 160/95 mmHg) [12]. Values should be higher for the 140/90 mmHg criteria. The use of OC may influence blood pressure control in patients attending outpatient clinics. In this study we investigated this possibility, showing that OC use is a risk factor for uncontrolled blood pressure among patients referred to our Hypertension Outpatient Clinic.

2. Materials and methods 

In a cross-sectional manner, we studied a consecutive sample of 171 hypertensive women between 18 and 50 years of age. All patients attending the Hypertensive Outpatient Clinic (Clinical Pharmacology Division of the Hospital de Clínicas de Porto Alegre) between 1989 and 1997 were included. Patients attending this Clinic take part in a large prospectively designed observational study underway since 1989. Some results have been reported elsewhere [13], [14], [15]. Most patients were referred to the clinic for uncomplicated hypertension. All patients submitted to a complete medical examination, with emphasis on aspects related to hypertension, and to a laboratory analysis, including fasting glucose, potassium, creatinine, cholesterol and urinalysis Blood pressure was determined according to standard recommendations [16]: blood pressure was measured with the patient seated, after at least 5 min of rest and with the arm in precordial level. Systolic blood pressure was defined by the first Korotkoff sound and diastolic blood pressure as the fifth sound. Two measures were done at each visit, at least 5 min apart. The average of three blood pressure readings—usual blood pressure—was used for the diagnosis and classification of hypertension.

Hypertension was diagnosed and classified on the first visit only in patients with blood pressure within the normal range (normal blood pressure—systolic blood pressure [SBP] < 140 mmHg and diastolic blood pressure [DBP] < 90 mmHg), or within the severe range (SBP > 210 mmHg or DBP > 115 mmHg), or in any patient with target organ failure.

The other patients were evaluated on two additional visits, with two blood pressure readings taken on each occasion. The average of six blood pressure measurements was used to classify these patients. Those taking medication at the time of the first evaluation were instructed to continue their regimen and were submitted to the same routine. The diagnosis and classification of blood pressure were based on the guidelines issued by the World Health Organization and the International Society of Hypertension [17].

The participants enrolled in this study were asked about current methods of contraception (rhythm method, condom, intrauterine contraceptive device, vaginal diaphragm, or oral hormonal contraceptives). Duration of oral hormonal contraceptive use was categorized as more than or less than 8 years.

Participants were divided into three groups according to the contraceptive method in use at the time of evaluation: current users of OC, users of other contraceptive methods, and nonusers of contraceptive methods.

Study variables were defined as follows: hypertension: SBP ≥ 140 mmHg or DBP ≥ 90 mmHg; moderate–severe hypertension: SBP ≥ 160 mmHg or DBP ≥ 105 mmHg.

Data were processed using the EPI-INFO software (Geneva, Switzerland) with a checking program for consistency and reliability. The analyses were performed with the SPSS software. Categorical variables were analyzed in contingency tables and tested by two-tailed χ2. DBP, SBP and other continuous variables (age, parity, body mass index, waist/hip ratio) from the patients classified by the contraceptive method were compared using analysis of variance and tested for contrasts by the Scheffe’s test.

A logistic regression model (ENTER option) was used to analyze the simultaneous associations between age, body mass index, antihypertensive treatment and use of OC (explanatory variables), including hypertension (SBP ≥ 140 mmHg or DBP ≥ 90 mmHg) as the dependent variable. An ∝ level of 0.05 was regarded as significant and 95% confidence intervals were calculated. The protocol was approved by the Ethics Committee of our Institution. All participants gave written consent to participate in the study.

3. Results 

The characteristics of participants according to the method of contraception status are shown in Table 1. Among the OC users, 39.3% were unable to give information on what OC they used, 36.6% used low-dose OC (≤30 μg estrogen), and 24.1% used high-dose OC (≥50 μg estrogen). Women using OC were younger than users of other forms of contraception (p = 0.001) and patients not using contraception (p = 0.044). OC users had also experienced fewer gestations than users of other contraceptive methods (p = 0.044).

Table 1. Characteristics of the patients in accordance with the contraceptive method used (mean ± SD)

	Characteristics	OC users (n = 66)	Using other contraceptive methods (n = 26)	Not using contraception (n = 79)	p-value*
	Age (years)	38.3 ± 6.4	44.5 ± 5.5	42.1 ± 7.0	0.0001
	Monthly income (legal minimum salaries)	4.9 ± 3.5	5.7 ± 5.7	6.3 ± 8.7	0.463
	Education (years)	7.2 ± 3.4	7.0 ± 4.4	6.8 ± 4.0	0.884
	Parity	3.2 ± 1.6	4.4 ± 2.3	3.8 ± 2.5	0.034
	Weight (kg)	69.5 ± 14.9	74.1 ± 10.8	75.0 ± 19.7	0.138
	Body mass index (kg/m2)	28.4 ± 5.6	29.9 ± 4.0	30.9 ± 7.3	0.072
	Waist-to-hip ratio	0.86 ± 0.092	0.84 ± 0.084	0.86 ± 0.084	0.509
	Alcohol consumption (g/day)	1.3 ± 3.5	1.4 ± 5.6	2.0 ± 12.9	0.896
	Duration of hypertension (years)	5.0 ± 4.7	7.0 ± 5.8	6.4 ± 6.2	0.191

* Analysis of variance.

DBP was higher in participants using OC and the differences persisted after adjusting for confounding (Table 2). SBP was not different among women using OC, using other methods or not using contraception at all.

Table 2. Blood pressure according to the contraceptive method used by the participants (mean ± S.D.), adjusted for age, body mass index, and antihypertensive drug use

	Blood pressure (mmHg)	Systolic*		Diastolic**
		Observed	Adjusted	Observed	Adjusted
	OC users (n = 66)	155.1 ± 24.2	157.5	100.2 ± 15.9	100.3
	Using other contraceptive methods (n = 26)	154.8 ± 31.9	152.5	93.4 ± 14.7	93.0
	Not using contraception (n = 79)	148.1 ± 24.4	147.9	93.3 ± 14.4	93.5
	p-value	0.208	0.096	0.016	0.023

Analysis of covariance:

* F = 2,379. ** F = 3,844.

OC users had poorer blood pressure control (SBP < 140 mmHg or DBP < 90 mmHg) than users of other contraceptive methods or nonusers (Table 3).

Table 3. Blood pressure control according to the contraceptive method used

	Blood pressure	Normal (SBP < 140 and DBP < 90)		High (SBP ≥ 140 or DBP ≥ 90)
		n	(%)	n	(%)
	OC users (n = 66)	11	(16.7)	55	(83.3)
	Using other contraceptive methods (n = 26)	9	(34.6)	17	(65.4)
	Not using contraception (n = 79)	25	(31.6)	54	(68.4)

p-value for trend = 0.046.

It was also noted that a higher proportion of OC users tended to be classified as stage 2 and 3 hypertension (moderate–severe) according to the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) (Table 4). Table 5 shows that these tendencies are not related to lower frequency of treatment of the OC users.

Table 4. Association between contraceptive method and severe hypertension (stage 2 and 3 according to the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure—JNC VI)

	Severe hypertension	Yes		No
		n	(%)	n	(%)
	OC users (n = 66)	14	(21.2)	52	(78.8)
	Using other contraceptive methods (n = 26)	5	(19.2)	21	(80.8)
	Not using contraception (n = 79)	10	(12.7)	69	(87.3)

p-value for trend = 0.170.

Table 5. Antihypertensive drug use by the participants classified by the use of contraceptive methods (n and %)

	Antihypertensive drug use	Yes		No
		n	(%)	n	(%)
	OC users (n = 64)	43	(67.2)	19	(29.7)
	Using other contraceptive methods (n = 26)	19	(73.1)	7	(26.9)
	Not using contraception (n = 78)	55	(70.5)	21	(26.9)

p-value = 0.91.

Blood pressure was also analyzed in terms of length of OC use, controlled for age (Table 6). A significant association was found between the duration of use of OC and levels of SBP and DBP.

Table 6. Systolic and diastolic blood pressure according to length of OC use, adjusted for age

	Length of use	n	Observed mean	Adjusted mean	p-value
	SBP
	<8 years	31	147.7	147.4	0.019
	>8 years	34	162.4	162.7
	DBP
	<8 years	31	96.5	95.5	0.028
	>8 years	34	104.0	104.9

In the logistic regression model, controlling for age, body mass index and use of antihypertensive drugs, the use of OC was a risk factor for uncontrolled blood pressure (SBP ≥ 140 mmHg and DBP ≥ 90 mmHg) (Table 7).

Table 7. Logistic regression model using uncontrolled blood pressure (SBP ≥ 140 mmHg or DBP ≥ 90 mmHg) as the dependent variable

	Variables	OR	CI (95%)	p-value
	Age	1.03	0.97–1.08	0.302
	BMI	0.99	0.94–1.05	0.841
	Antihypertensive drug use	0.76	0.34–1.69	0.507
	OC users	2.67	1.17–6.05	0.018

OR = odds ratio; CI = confidence interval; BMI = body mass index.

4. Discussion 

The majority of studies reporting an association between OC use and blood pressure rising were planned for other purposes and detected elevated blood pressure only in secondary analyses. Most of them examined the development of hypertension in previously normotensive women, and a few examined women previously diagnosed as hypertensive. We are not aware of any study about the association between OC use and blood pressure involving patients attending an outpatient clinic for management of hypertension, at least in Brazil.

Despite the lack of well-conducted studies, the recommendation to suspend OC use in women who have developed hypertension is frequently cited in the literature [8]. In women with essential hypertension, there seems to be less concern with OC use, because the change in blood pressure is usually described as mild. In the present study, the majority of the patients had previous diagnosis of hypertension, but only 22.7% of OC users had been advised to change their form of contraception.

It is important to note that the small rises in blood pressure caused by combined OC may occur on top of already high levels in hypertensive women, resulting in misclassification of these patients. This may also determine the need for pharmacological treatment in patients whose blood pressure could otherwise have been controlled by nonpharmacological means.

The first observations on the subject, in 1967, described significant improvement or complete remission of hypertension in 75% of hypertensive women who suspended high-dose estrogen OC use [18]. These results point to a probable association between OC use and prevalence of hypertension.

In the present study, DBP was approximately 7 mmHg higher in OC users when compared to users of other contraceptive methods and nonusers of contraception, independently of confounding. SBP was also higher in OC users, although the difference was not significant.

There was poor blood pressure control, as well as a higher proportion of moderate–severe hypertension (JNC VI stage 2 and 3), among OC users. Our OC users were younger than patients from the other groups, and could have hypertension at more severe stages. Despite the fact that the incidence of hypertension increases with age, sometimes younger patients have more severe disease, frequently due to secondary hypertension. Because we excluded patients with secondary hypertension, and controlled for age and other confounders in the multivariate analysis, it is unlikely that the association between OC and hypertension could have been confounded by age. Also, these tendencies were not due to lower frequency of antihypertensive drug use by OC users. Fisch et al. [19] found similar results in a transversal study, showing a significant rise in DBP (average of 1.3 mmHg) among OC users, and higher frequency of hypertension, even after control for age. SBP was also significantly higher among OC users (average of 5 mmHg).

It should be emphasized that the nondirectional nature of our cross-sectional study does not make it possible to affirm causality. Only a cohort study could investigate the association over time. Some well-designed cohort studies [20], [21] found significant rises in blood pressure among OC users.

The observation that SBP and DBP are significantly higher in those who have used OC for more than 8 years, even after controlling for age, suggests indirectly that exposure to OC is a risk factor for hypertension. Similar findings were seen in the cohort of Chasan-Taber et al. [22], where those using OC for more than 6 years presented higher risk of developing hypertension (relative risk: 2.1; 95% confidence interval, 1.6–2.7). In the same way, the study of Cook et al. [21] found an association between DBP and duration of OC use (average of 0.5 mmHg per year), although the change was not significant when compared to the initial blood pressure values.

The association between DBP and OC use was shown to be independent of potential biases in the multivariate analysis. Similarly, there was a significant and independent association between OC use and the diagnosis of uncontrolled hypertension in logistic regression models.

A controlled, multicentric study performed by WHO [23] found significant rise in SBP and DBP after 1 year of OC use, and a cumulative incidence of hypertension of 0.78 per 100,000 women-year among OC users, compared to zero in the control group (p = 0.034). Chasan-Taber et al. [22] observed a 2-fold higher risk of development of hypertension among current OC users compared to nonusers.

Some of the findings of our investigation are original or sparsely reported in the literature. With the exception of the study of Narkiewicz [24], this study appears to be one of the first to focus on the behavior of blood pressure in hypertensive OC users. It is also one among few studies that found significant changes in DBP, which found a significant correlation between length of OC use and blood pressure levels, and also poor blood pressure control among OC users. Furthermore, it is the first study in Brazilian women.

Our study has some weaknesses. The nondirectionality of the data collection, in terms of both exposure and outcomes, makes it impossible to draw any conclusions about causality. The available evidence and the biological plausibility reinforce the interpretation that OC use is a risk factor for raised blood pressure, but only prospective studies will be able to define the precise nature of the association.