Investigation of Relation between Use of Oral Contraceptives and Thromboembolic Disease. A Further Report

The results of a previous study of the use of oral contraceptives by married women discharged from hospital with a diagnosis of thromboembolic disease in the years 1964–6 were reported by us last year. The present paper adds results relating to patients discharged during 1967 and a few data, that could not be sought previously, for patients discharged with cerebral or coronary thrombosis from three of the hospitals in the earlier period.Of 84 patients with deep-vein thrombosis or pulmonary embolism 42 (50%) had used oral contraceptives during the month preceding the onset of their illness, while only 23 of the 168 controls (14%) had done so. No differences in risk were found either for the types of preparation or for the duration of use. After allowance for age and height, the patients with venous thromboembolism were about 10 lb. (4,535 g.) heavier than the control patients, irrespective of whether they were using oral contraceptives or not. No appreciable difference was found between the smoking habits of patients with and without venous thromboembolism treated during 1967, nor between women who were using oral contraceptives and those who were not. The trend in hospital admissions for venous thromboembolism with time corresponded to the trend in the use of oral contraceptives, and there was no evidence to suggest that the number of admissions was affected by publicity about the risk of using the preparations. Of 19 patients with cerebral thrombosis 11 (58%) had been using oral contraceptives, compared with an expected figure of 3.5 from the experience of the control subjects. All the published data (clinical, angiographic, and post-mortem) show that the thrombosis affects the cerebral arteries rather than the cerebral veins. Of 17 patients with coronary thrombosis 2 (12%) had been using oral contraceptives, compared with an expected figure of 2.1. The patients with coronary thrombosis smoked more than the control patients and were, on average, 8.3 lb. (3,765 g.) heavier than control women of the same age and height.The new evidence strengthens the belief that oral contraceptives are a cause of venous thromboembolism and cerebral thrombosis but does not indicate that they are a cause of coronary thrombosis.     

Oral contraceptive use and venous thromboembolism: absence of an effect of smoking.

We conducted a case-control study to test the hypothesis that women smokers who use oral contraceptives have an increased risk of developing venous thrombosis. Patients and controls were drawn from two sets of hospital patients already included in the Boston Collaborative Drug Surveillance Programme. Sixty patients with uncomplicated thromboembolism were matched with 180 controls with other diagnoses; all were premenopausal women taking oral contraceptives. Patients with conditions that might predispose to thromboembolism or be related to smoking were excluded. We found no association between smoking habits and thromboembolism. Similarly, we found no association between thromboembolism, smoking, and duration of oral contraceptive use. Thus we conclude that differences in fibrinolytic activity between smokers and non-smokers are not major factors in the aetiology of uncomplicated thromboembolism in women using oral contraceptives.     

Case-control study of risk of cerebral sinus thrombosis in oral contraceptive users who are carriers of hereditary prothrombotic conditions

Objective: To investigate whether users of oral contraceptives who are carriers of a hereditary prothrombotic condition (factor V Leiden mutation, protein C, S, or antithrombin deficiency) have an increased risk of cerebral sinus thrombosis. Design: Comparison of a prospective series of cases of cerebral sinus thrombosis with population data. Setting: Neurological teaching hospitals from different regions in the Netherlands (cases) and a representative sample of the non-institutionalised Dutch population (controls). Subjects: 40 women aged 18-54 years with cerebral sinus thrombosis (cases) and 2248 women aged 18-49 years (controls). Main outcome measure: Current use of oral contraceptives at the time of the thrombosis (cases) or at the time of the questionnaire (controls). Prevalences of a hereditary prothrombotic condition in patients and in the population with odds ratios. Results: 34 of 40 (85%) women with cerebral sinus thrombosis used oral contraceptives, versus 1007 of 2248 (45%) of the control women; the age adjusted odds ratio was 13 (95% confidence interval 5 to 37). Seven of 36 patients (19%) had a prothrombotic deficiency, versus 7% expected in the population; this corresponds to a threefold to fourfold increase in risk. In women who used oral contraceptives and also carried a prothrombotic defect, the odds ratio for cerebral sinus thrombosis was about 30 relative to women who had neither risk factor. Conclusion: The use of oral contraceptives and being a carrier of a hereditary prothrombotic condition increase the risk of and interact in a multiplicative way in the development of cerebral sinus thrombosis.

Key messages

• The use of oral contraceptives is associated with an increased risk of cerebral venous sinus thrombosis
• This risk of cerebral venous sinus thrombosis in women who use oral contraceptives is larger if there is an additional hereditary prothombotic factor (protein C, S, or antithrombin deficiency, factor V Leiden mutation)
• The association between oral contraceptives, thrombophilia, and deep vein thrombosis is also valid for cerebral sinus thrombosis
• Women do not need to stop using oral contraceptives as the absolute risk of cerebral sinus thrombosis is very small

     

Postoperative Thromboembolism and the Use of Oral Contraceptives

We have previously reported the results of a retrospective study of the relationship between oral contraceptives and deep vein thrombosis and pulmonary embolism occurring in the absence of an evident predisposing cause. The present paper describes a similar investigation of the relationship between oral contraceptives and thromboembolism following certain surgical procedures.Of 30 patients with thromboembolism, 12 (40%) had been using oral contraceptives during the month preceding surgery, while only 9 of 60 matched controls (15%) had been doing so. From these data it is estimated that the risk of postoperative thromboembolism is increased threefold to fourfold by the use of contraceptive steroids.It is recognized that few of the episodes of deep vein thrombosis and pulmonary embolism that occur postoperatively are detectable clinically, and that not all of these are recorded in discharge summaries or diagnostic indexes. Despite these considerations, we have been unable to detect any important bias in the data which could account for the observed association between oral contraceptives and the development of postoperative thromboembolism.     

Altered haemorheology in oral-contraceptive users.

The haemorheological profile of the menstrual cycle was determined in 12 women who did not take oral contraceptives and compared with that in two groups of women (n = 8 and n = 30) who had been taking oral contraceptives for at last six months. Packed cell volume, platelet count, erythrocyte deformability, plasma fibrinogen concentration, and plasma and whole-blood viscosity varied cyclically throughout the menstrual cycle in the 12 non-users. This variation was abolished by the use of oral contraceptives, and the values of these indices were raised by an amount likely to predispose to thrombosis.     

Effect of oral contraceptives and pregnancy on erythrocyte deformability and surface charge.

Erythrocyte deformability and surface charge were studied in normal premenopausal women, oral contraceptive users, and pregnant women. The increased incidence of thrombosis in women taking oral contraceptives could not be explained by decreased erythrocyte deformability or surface charge. However, the decreased erythrocyte deformability of late pregnancy may relate to thrombosis during this period and to increased hemolysis in patients with certain hemoglobinopathies.     

Effect of ethyloestrenol on fibrinolysis in the vessel wall.

Forty-nine patients with decreased fibrinolytic activity in the vessel walls or a decreased release mechanism, or both, were treated with ethyloestrenol for three to 17 months. Forty-five of the patients had had recurrent, phlebographically verified, deep venous thrombosis (DVT) and four had arterial thrombosis. Ethyloestrenol 8 mg/day was given to 31 patients and 4 mg/day was given to 12. The remaining six patients had been treated with a combination of phenformin and ethloestrenol. The phenformin was withdrawn but they were kept on ethyloestrenol 8 mg/day. Another 15 patients with a normal fibrinolytic system--four with recurrent DVT and 11 with severe arteriosclerosis--were given ethyloestrenol 8 mg/day. The spontaneous fibrinolytic activity, local fibrinolytic activity during standardised venous occlusion of the arms, and fibrinolytic activity of the vessel walls increased significantly after treatment with ethyloestrenol 8 mg/day for three months. No further increase occurred after three months, and ethyloestrenol 4 mg/day had no effect. No values rose significantly in the patients with a normal fibrinolytic system. One patient suffered a recurrence within three months of treatment, before the fibrinolytic system became normal. In one patient the fibrinolytic defect reappeared after 10 months in spite of continued treatment. Two of the three women of fertile age developed irregular cycles and intermenstrual bleeding, which disappeared when the treatment was withdrawn. No other side effects were observed.     

Hypofibrinolysis and thrombophilia

The fibrinolytic capacity was assessed in 18 healthy subjects and in 8 patients each with non-idiopathic venous thrombosis, idiopathic venous thrombosis and myocardial infarction after intravenous administration of 1-deamino-8-D-arginine vasopressin (DDAVP) (0.4 microgram/kg) in comparison to venous occlusion. In healthy subjects the results obtained by either stimulus were approximately in agreement. Compared to the control group, in patients with non-idiopathic venous thrombosis the fibrinolytic capacity was not changed either after venous occlusion or after administration of DDAVP. In 5 out of 8 patients with idiopathic venous thrombosis the capacity was significantly reduced both after venous occlusion and after administration of DDAVP. In 4 out of 8 patients with myocardial infarction the capacity was significantly below the limit after administration of DDAVP while it was not after venous occlusion. In determining the fibrinolytic capacity DDAVP proved to be superior to venous occlusion.     

Two different mechanisms in patients with venous thrombosis and defective fibrinolysis: low concentration of plasminogen activator or increased concentration of plasminogen activator inhibitor.

Fibrinolytic components after venous occlusion and concentrations of tissue plasminogen activator inhibitor were studied in 100 consecutive patients with confirmed recurrent deep vein thrombosis or pulmonary embolism. After 20 minutes of venous occlusion the fibrinolytic response was decreased in 33 patients, as measured both amidolytically with S-2251 and on fibrin plates. Two different mechanisms responsible for the poor fibrinolytic response could be distinguished. Twenty two of the patients in whom the response was poor released normal amounts of tissue plasminogen activator antigen, as assayed by immunoradiometric assay, but had appreciably increased concentrations of tissue plasminogen activator inhibitor. The 11 other patients in whom the response was poor had both low tissue plasminogen activator activities and low tissue plasminogen activator antigen concentrations but normal concentrations of tissue plasminogen activator inhibitor. The results show not only that defective synthesis or release of tissue plasminogen activator may be important in the pathogenesis of venous thrombosis but also that a large group of patients with thrombosis have an increased concentration of the inhibitor to tissue plasminogen activator.     

Antithrombotic therapy in patients with known risk factors for thromboembolism

In about 50% of the cases of spontaneous deep vein thrombosis a congenital deficiency of an inhibitor of coagulation or an insufficient fibrinolytic mechanism can be detected. In arterial thromboembolism a connection with hyperactive platelets or with a diminished availability of tissue plasminogen activator can be found in about 70%. However, in these cases the defect which provokes thrombosis is mostly acquired and is connected with hyperlipidemia and/or with atherosclerotic alterations of the vessel wall. A study on patients with thromboembolic tendency and detectable risk factors was carried out. A total of 470 patients could be observed for 2 years under an adequate antithrombotic prophylaxis. The occurrence of thromboembolic episodes 2 years prior to prophylaxis and 2 years under prophylaxis was compared. In venous cases thrombosis could be controlled almost completely by coumarins when the underlying cause was a deficient plasmatic inhibitor. In patients with diminished fibrinolysis there was only a partial effect of oral anticoagulants. A better result could be obtained when pentosan polysulfate was administered. In arterial thromboembolism the results of prophylaxis were less convincing. The efficacy of ASA in patients with an increased platelet function was only moderate. In addition, ASA hat to be discontinued in about 20% of the patients because of gastrointestinal problems. Pentosan polysulfate in patients with a diminished fibrinolytic capacity had a fairly good effect and resulted in a 60% reduction of thromboembolic manifestations. It is shown that an exact diagnosis of the underlying deficiency which is likely to cause thrombosis can also improve the efficacy and the specificity of prophylaxis.     

Post-"pill" amenorrhoea--cause or coincidence?

The relationship of contraceptive history to diagnostic category of amenorrhoea was analysed in 131 consecutively investigated cases of secondary amenorrhoea. Amenorrhoea occurred in 52 patients immediately after discontinuing the oral contraceptive. Twenty-two had had amenorrhoea before oral contraceptive treatment and 23 patients before the episode of non-contraceptive-related amenorrhoea investigated here. When these cases were excluded from analysis there was no significant difference in the distribution of any of the diagnostic categories between those who had used the oral contraceptive and those who had not. The results suggest that using oral contraceptives does not cause subsequent amenorrhoea.     

Increased tissue-plasminogen activator (t-PA) levels in patients under oral anticoagulant therapy

The fibrinolytic system was investigated in 30 patients under oral anticoagulant therapy, and in 23 control patients not receiving oral anticoagulants. Patients under oral anticoagulant therapy had significantly higher tissue-plasminogen activator (t-PA) antigen levels than patients in the control group. Mean t-PA levels before venous occlusion were 18.4 ng/ml in the anticoagulated patients vs. 7.9 ng/ml in the control patients (p less than 0.001). After venous occlusion for 10 minutes, t-PA levels were 45.0 ng/ml in the anticoagulated patients and 24.2 ng/ml in the control patients (p less than 0.01). Plasminogen activator inhibitor (PAI) capacity was not significantly different in the two groups before venous occlusion (VO) but differed slightly (p less than 0.05) after VO. The net decrease in euglobulin lysis time (ELT) after venous occlusion (= ELT before VO - ELT after VO), indicating the relative potency of the fibrinolytic activity in blood, was also significantly higher in the anticoagulated patients (median 240 min vs. 125 min, p less than 0.001). These data indicate that oral anticoagulant therapy increases the fibrinolytic activity in blood, and thus may have an additional therapeutic effect in addition to anticoagulation.     

Fibrinolytic capacity of arm and leg veins after femoral shaft fracture and acute myocardial infarction.

The local fibrinolytic activity generated in the leg and arm veins during venous occlusion (fibrinolytic capacity) and the systemic fibrinolytic activity were measured at intervals in 11 patients after fracture of the femoral shaft and in 11 patients after acute myocardial infarction. In both groups the fibrinolytic capacity of the leg veins and the systemic fibrinolytic activity were significantly reduced two days after the onset of tissue injury. The fibrinolytic capacity of the arm veins was not altered. These results provide a possible explanation for the predilection of venous thrombosis for the leg veins after accidental trauma and acute myocardial infarction.     

Involuntary movements in patients taking oral contraceptives

Involuntary movements developed in five women taking oral contraceptives. In one, the sudden onset of a unilateral disturbance suggested a cerebral thrombosis; this case is considered to be a further example of the increased risk of cerebrovascular disease associated with oral contraception. The four other patients suffered a relapse of Sydenham''s chorea between one and four months after starting an oral contraceptive regimen. Possibly an underlying vascular mechanism was responsible for these relapses.     

Characteristics affecting fibrinolytic activity and plasma fibrinogen concentrations.

As part of a study to determine the extent to which the haemostatic system is implicated in the onset of clinically manifest ischaemic heart disease, characteristics influencing fibrinolytic activity (FA) and plasma fibrinogen concentrations were examined in 1601 men aged 18-64 and 707 women aged 18-59 in several occupational groups in North-west London. In men FA noticeably decreased till the age of about 58, when there was a small rise. In women a small increase in FA between 18 and about 40 was followed by a slightly larger fall between 40 and 59. There was a pronounced negative association of FA with obesity. FA was significantly less in smokers than non-smokers, though the effect was not large. FA increased with alcohol consumption. FA in men appeared to be greatest in the lower social classes, and men on night shift had poorer FA than those on day work. FA was greater in women using oral contraceptives than in those not using these preparations. In both sexes FA increased with exercise, but there were no associations between any of the characteristics studied and the increase. Plasma fibrinogen concentrations increase with age and obesity, are higher in smokers than non-smokers, and fall with alcohol consumption. In women the concentrations are higher in those using oral contraceptives. The general epidemiology of FA and plasma fibrinogen concentrations suggests that they may well be implicated in the pathogenesis of ischaemic heart disease.     

Possible vascular-bed-specific role of interleukin-6 in young women with a history of myocardial infarction, lacunar cerebral infarction and deep vein thrombosis

The level of interleukin-6 (IL-6), a cytokine with prothrombotic properties, and its associations with metabolic, coagulation and fibrinolytic parameters were investigated in 68 young women (23-49 years, mean 40 years old) six months to six years after myocardial infarction (MI, N=22), lacunar cerebral infarction (LACI, N=16) and deep vein thrombosis (VT, N=30); all women were in the reproductive period, aged <45 years at the time of acute thrombotic event. Forty-seven age-matched women comprised a control group. Basic and multivariate analysis disclosed different patterns of IL-6 increase in all three groups of patients. In the MI group IL-6 was significantly elevated independently of factors known to increase IL-6 levels; the increase was most pronounced in patients with high lipoprotein(a). This result suggests a prothrombotic association of lipoprotein(a) with IL-6. In the whole LACI group IL-6 was not significantly increased. However, patients with elevated levels of IL-6 had abdominal obesity and elevated fibrinogen, suggesting the possibility that this combination might represent a specific risk profile. In VT group elevated IL-6 level was found in the group of previous users of oral contraceptives (OC). This might be relevant, since it is known that OC could importantly increase (previously elevated) IL-6 level in selected women. Our results suggest the hypothesis that the role of IL-6 might be vascular-bed-specific and further propose that increased IL-6 level might represent a novel, non-classical risk factor for development of MI, LACI and VT in specific subgroups of young women, which have to be clarified in further studies.     

Intracranial venous thrombosis complicating oral contraception

Four days after the onset of a severe headache a 22-year-old woman who had been taking oral contraceptives for less than three weeks had a convulsion, followed by right hemiparesis. Other focal neurologic signs and evidence of raised intracranial pressure appeared, and she became comatose on the seventh day. A left craniotomy revealed extensive cerebral venous thrombosis. She died the next day. On postmortem examination extensive thrombosis of the superior sagittal sinus and draining cerebral veins, and multiple areas of cerebral hemorrhage and hemorrhagic infarction were seen. Some of the superficial cerebral veins showed focal necrosis of their walls, and the lateral lacunae of the superior sagittal sinus contained proliferating endothelial cells. The adrenal veins were also thrombosed. The significance of these findings is discussed. The literature on cerebrovascular complications of oral contraception, particularly cerebral venous thrombosis, is reviewed.     

Blood pressure and contraceptive use

In a survey of 461 women routinely attending family planning clinics those taking oral contraceptives had significantly higher mean systolic and diastolic blood pressures than those using non-hormonal contraception. There appeared to be a dose-response relation of blood pressure to the progestogen component of two oral contraceptives with an identical 30 μg ethinyloestradiol component. This supports the idea that the progestogen as well as the oestrogen component has an aetiological role in the rise in blood pressure. There was a significant correlation of blood pressure with duration of current use of oral contraceptive but not with total duration of use. There was also a significant negative correlation of blood pressure with time since oral contraceptives were last taken, and women who had stopped using oral contraceptives over a month previously had similar blood pressures to those who had never taken them. In women taking oral contraceptives those who had either a history of hypertension in pregnancy or a family history of hypertension had significantly higher mean blood pressures than those who did not. Both systolic and diastolic blood pressures correlated independently with weight and body mass index, but controlling for the effect of this and age did not affect the above relations. No significant differences in mean blood pressures were found between different ethnic groups, and there was no relation of blood pressure to reported marital state, social class, parity, smoking, or alcohol use.Any oral contraceptive that has a less adverse effect on blood pressure has implications for general prescribing policy; thus even small differences in the progestogen contents of low-dose oestrogen pills may be important.     

Ovarian neoplasms,functional ovarian cysts,and oral contraceptives.

The incidence of ovarian neoplasms and functional ovarian cysts diagnosed at laparotomy or laparoscopy among the 17,000 women taking part in the Oxford Family Planning Association contraceptive study was investigated. Epithelial cancer of the ovary was only 25% as common among those who had ever taken oral contraceptives as those who had never done so (95% confidence interval 8% to 67%). There was little evidence of any important association between use of oral contraceptives and benign teratoma or cystadenoma. Functional cysts of the ovary occurred much less commonly in women who had recently (in the six months preceding diagnosis) taken combined oral contraceptives (but not in those who had taken progestogen only oral contraceptives) than in those who had never taken oral contraceptives or had taken them in the past. This protective effect was more pronounced for corpus luteum cysts (78% reduction; 95% confidence interval 47% to 93%) than for follicular cysts (49% reduction; 95% confidence interval 20% to 70%). It is estimated that about 28 (95% confidence interval 16 to 35) operations for functional ovarian cysts are avoided among every 100,000 women who take oral contraceptives each year.