Atherosclerosis,

94 (1992)

147

147- 152

Elsewier Scientific Publishers Ireland, Ltd. All rights reserved. 0021-9150/92/$05.00 Printed and Published in Ireland 0

1992

ATHERO 04821

Hyperhomocyst(e)inemia, anti-estrogen antibodies and other risk factors for thrombosis in women on oral contraceptives V. Beaumonta, M.R. Malinowb2”, G. Sextonc, D. Wilson”, N. Lemorta, B. Upsonb and J.L. Beaumont” “Department

of Cardiology,

Hopital

Henri

Mondor,

(’ The Oregon

Creteil

Health

(France),

Sciences

h Oregon

University,

Regional Portland,

Primate

Research

Center.

Beawrtcm,

OR and

OR (USA)

(Received 5 December, 1991) (Revised, received 27 February, 1992) (Accepted IO March, 1992)

Summary Hyperhomocyst(e)inemia was shown to be associated with vascular occlusion in atherosclertotic patients. We have conducted a study to determine if hyperhomocyst(e)inemia was also related to the vascular events observed in women on oral contraceptives, presumably having little or no atherosclerosis. Two hundred women receiving oral contraceptives were included in the study: 100 were healthy controls and 100 had documented vascular occlusion. Determination of serum homocyst(e)ine and anti-estrogen antibody levels wore performed under blind conditions. They were evaluated in logistic regression models in which age and smoking were also included. Women with vascular occlusion had higher levels of homocyst(e)ine (P < 0.001) and of anti-estrogen antibodies (P < 0.001) when compared to controls. They were also older (P < O.oO]) and more frequently smokers (P < 0.05). The above mentioned variables were, in isolation, independent predictors of vascular occlusion, Moreover, a model assessing those variables and their interactions indicated that the levels of anti-estrogen antibodies and smoking increased the predictability in older women, as well as the levels of age-adjusted homocyst(e)ine. The study suggests that the above factors can identify women at risk and that determination of anti-estrogen antibodies and homocyst(e)ine levels may help to detect women predisposed to vascular occlusions when taking oral contraceptives.

Key words: Hyperhomocyst(e)inemia;

Anti-estrogen antibodies; Smoking; Oral contraception; Thrombosis

Introduction

Correspondence to: Professor Jean-Louis Beaumont, Hopital Henri Mondor, 51 Avenue du Marechal de Lattre de Tassigny, 94010 Creteil, France.

Oral contraception (OC) is associated with an increased risk of vascular complications [ 11. Measures to reduce such a risk have consisted

148

mainly in lowering the estrogen intake, controlling certain risk factors for cardiovascular disease (e.g., high blood pressure, hyperlipidemia, diabetes) and avoiding OC in women with coagulation abnormalities. Although some of the above factors may be moderately increased in users of OC [l], they failed to predict the development of thrombotic complications in healthy women [2]. Subsequently, three factors were reported to be associated with the thrombotic complications of OC, age [l], smoking [3-41 and anti-estrogen antibodies [5]. Since increased levels of plasma homocyst(e)ine were found in patients affected with atherosclerotic occlusions [6- 151, the present study was conducted to test the hypothesis that hyperhomocyst(e)inemia may be an additional risk factor for vascular occlusions in women on OC.

(c) 10 women (mean age 27 i 12 years who had never used OC or other sex steroid hormones were also selected. This group is very small but ex-users had to be excluded, because anti-estrogen antibodies (Abs) once induced were shown to persist after stopping the Pill. Adequate control groups had been studied previously [2-lo] and their results were similar to those of the 10 women of the present study.

Patients and Methods

Homocyst (e) ine measurement Serum levels of homocyst(e)ine, that is, the sum of homocysteine and the homocysteinyl moieties of homocystine and cysteine-homocysteine mixed disultide, whether free or bound, were measured by high pressure liquid chromatography and electrochemical detection as previously described [lO,ll]. Results are expressed in nmol/ml of homocysteine. The within run coefficient of variation of the assay was 1.1%.

Study population Two hundred women who were regular users of OC tablets containing 30 or 50 rg ethinylestradiol, were studied: (a) 100 healthy women (mean age 27.9 f 8.9 years) systematically controlled in the course of oral contraception, and symptom free. (b) 100 women (mean age 32.3 i 9.7 years) with clinically documented vascular occlusion in the course of OC. They were provided by several hospital departments. For each patient we obtained a summary of the hospitalisation record or a documented letter from the physician, and complementary documents: perfusion lung scan for pulmonary embolism, electrocardiogram for myocardial infarction, angiography, CT scan. or magnetic resonance imaging for neurological diseases. Patients were discarded when vascular thrombosis was suspected but could not be ascertained and when other causes of thromboembolism were associated. Finally, among the 100 patients retained 36 had venous and/or pulmonary thrombosis, 6 arterial thrombosis (1 in the leg, 3 in a coronary artery, 2 in a mesenteric artery) and 58 had cerebrovascular occlusive disease, either arterial (52 cases) or venous (6 cases). All serum samples were obtained within the course of the hospitalization.

Methods Serum levels of antiethinylestradiol antibodies (anti-estrogen Abs) were determined on fresh serum samples. Additional aliquots were coded, kept frozen at -20°C and sent under dry ice to the Laboratory of Dr. Malinow for homocyst(e)ine determinations. All the tests were performed under blind conditions.

Anti-estrogen Abs measurement Determination of anti-estrogen Abs was performed by immuno-assay with tritiated ethinylestradiol t3HEE). The specific antiEE activity was estimated by the binding of ‘HEE to the immunoglobulin of the serum to be tested and previously prepared by precipitation in ammonium sulfate at 25% saturation. Because of the usually high level of non-specific binding, radioimmuno assay (RIA) was measured as the difference between two assays. In the first assay, 100 ~1 of a stock solution prepared with ‘HEE (54.7 Wmmol) was added to 500 pg of the protein to be tested. In the second assay, an excess of unlabeled EE was added previously to the protein, in order to saturate the non-binding sites. Further details concerning the technique have been described earlier 151. The radioactivity was expressed in counts per minute (counts/min). The difference

149

between the two assays expressed the quantity of ‘HEE bound to the specific and saturable sites of the immunoglobulins. The within run coefficient of variation for the assay was 8%. Statistical

analysis

Statistical evaluation was performed using the SAS statistical package (Statistical Analysis System, Cary, NC), on 3 continuous variables, age and plasma levels of anti-estrogen Abs and homocyst(e)ine, as well as on one categorical variable (presence or absence of cigarette smoking). History of vascular occlusion (see above) defined the 2 groups of subjects. Means, standard deviations and standard errors were calculated for the continuous variables. Homocyst(e)ine levels were transformed using log (In) transformation to improve normality of the distribution. Student’s two sample c-statistic, or the Mann Whitney U-test (anti-estrogen Abs), were used to assess differences between the two groups. Chi-square statistics were used to assess differences in the categorical variable. Correlations between variables were established with Pearson’s correlation coefficient. Two-tailed statistics were used throughout. In the subsequent analysis, age-adjusted log (In) transformed homocyst(e)ine was used because of the strong dependency of homocyst(e)ine on age, as established by a regression analysis of the two variables. Of the 200 women surveyed, 167 had

information on all variables and their data were entered into a stepwise logistic regression analysis [ 16,171 The variables were added sequentially according to the magnitude of their statistical significance. The stepwise logistic regression analysis was repeated entering the variables, as well as two-way interactions between the variables. Tables for both regression analysis were constructed including only those variables, or their interactions, that met the 0.05 significance level for entry into the model. Finally, we calculated the prediction characteristics of the two logistic regression models, i.e., their sensitivity, specificity, accuracy and false positive and false negative rates. Results Table 1 shows results in patients and controls on OC. There was a significant difference in age (32.3 f 9.7 years in patients vs. 27.9 f 8.9 years in controls; P < 0.001). Mean levels of anti-estrogen Abs were higher in patients than in controls (189 f 176 and 94 * 176 counts/min, respectively; P < 0.001). When healthy users and non-users of OC were compared, levels of anti-estrogen Abs differed significantly (94 f 176 and -40 f 101 countsimin, respectively ; P = 0.004). Homocyst(e)ine levels were higher in patients than in controls (10.70 f 6.60 nmol/ml vs. 7.78 i 2.70 nmol/ml, respectively; P < 0.001). OC use did

TABLE I CHARACTERISTICS

OF WOMEN ON ORAL CONTRACEPTIVES P

Thrombosis Absent n Age (years) Estrogen dose (&day) 30 (n) 50 (n) Anti-estrogen Abs (counts/min) Homocyst(e)ine (nmol/ml) Smokers (%)b aMean f SD. bData not available in 21 of the

100 21.9

Present

f

8.9”

86 14 94 zt 176 7.78 f 2.70 38

100 32.3 zt

9.7

19 21 189 zt 176 10.70 f 6.60 56

group without and 11 of the group with occlusion.

Hyperhomocyst(e)inemia, anti-estrogen antibodies and other risk factors for thrombosis in women on oral contraceptives.

Hyperhomocyst(e)inemia was shown to be associated with vascular occlusion in atherosclerotic patients. We have conducted a study to determine if hyper...
576KB Sizes 0 Downloads 0 Views