A number of epidemiological and clinical studies have consistently shown that oral contraceptives are associated with an increased incidence of venous thromboembolism in women of childbearing age.1–3 However, it is becoming more and more obvious that women who develop venous thromboembolism must have some underlying predisposition that leads to hypercoagulability after the introduction of oral contraceptives.4 It is already known that inherited thrombophilia, such as possession of factor V Leiden, G20210A prothrombin, or deficiency of antithrombin, protein C, or protein S, substantially increases the risk of venous thromboembolism in women using oral contraceptives.5–7 Several other novel risk factors, such as inflammatory cytokines and antiphospholipid antibodies, which have already been shown to be associated with venous thromboembolism,8,9 might also be involved in the development of hypercoagulability and venous thromboembolism in users of oral contraceptives. However, data on such associations that might be clinically relevant are lacking.
Thus, the aim of our study was to test the hypothesis that interleukin-6 (IL-6), a procoagulant cytokine,10 and antiphospholipid antibodies, which also have prothrombotic capacity,9 would be elevated and associated with each other in young women who had developed first venous thromboembolism during oral contraceptive use (in the absence of other risk factors of venous thromboembolism such as malignancy, trauma/operation/immobilization, or inherited thrombophilia) in comparison with young women, who had developed first venous thromboembolism when not on oral contraceptives, and with a control group of healthy women.
PATIENTS AND METHODS
The patients and controls were selected from the World Health Organization Collaborative hospital-based, case-control Study of Cardiovascular Diseases and Steroid Hormone Contraception, in which Slovenia was one of the participating countries.1,11 The study was undertaken in 21 centers in 17 countries. The study investigated the risk of venous thromboembolism, myocardial infarction, and stroke associated with combined oral contraceptives. The principal aim of the venous thromboembolism component was to evaluate whether current oral contraceptive use was associated with increased risk of a first deep vein thrombosis, pulmonary embolism, or both. Secondary aims were to evaluate the risk of venous thromboembolism according to type, duration, and past use of oral contraceptives and in different subgroups of women. In each center, a monitoring system was set up to identify all eligible cases. Women were eligible as cases if they were aged 15–49 years, had been admitted to a collaborating hospital between February 1, 1989, and January 31, 1995, and had a discharge diagnosis of deep venous thrombosis (DVT) and/or pulmonary embolism. Those who died within 24 hours of admission, who had a history of stroke, DVT, pulmonary embolism, or acute myocardial infarction, or natural or surgical menopause, or who had a recent history (within 6 weeks) of pregnancy, major illness causing prolonged bed rest, or surgery were excluded. Cases were categorized according to methods used to confirm the diagnosis of DVT or pulmonary embolism as definite (DVT confirmed by venography or duplex scanning or radioisotope studies, and pulmonary embolism confirmed by a ventilation/perfusion scan or angiography), probable, possible, or other. Approximately 50% of DVT and 25% of pulmonary embolism cases were categorized as definite in European centers. For each case, an attempt was made to recruit 3 female controls, matched within a 5-year age band. Controls with 27 predefined diagnoses not associated with oral contraceptives had to be admitted to the same hospital as the given case within the period of 2 weeks before and up to 4 months after the case's admission. They were recruited by the investigators surveying the wards that accepted women with any of the control diagnoses in a predefined random order. Controls were subject to the same exclusion criteria as the cases. The study confirmed an association between oral contraceptive use and venous thromboembolism in European and developing countries (in Africa, Asia, and Latin America): oral contraceptives users had a 3- to 4-fold increased risk of venous thromboembolism. Risk was higher with use of oral contraceptives containing third-generation progesterone, compared with the use of oral contraceptives containing first- or second-generation progesterone. Increased risk was apparent within 4 months of starting oral contraceptives, was unaffected by the duration of the current period of oral contraceptives use, and had disappeared within 3 months of stopping oral contraceptives. In addition, increased body mass index above 25 kg/m2, but not age, smoking, or a history of hypertension (excluding hypertension in pregnancy), was an independent risk factor for venous thromboembolism.
In our Slovenian center, 120 eligible cases with venous thromboembolism were identified and included in the World Health Organization study, of which 48% were on oral contraceptives. We reviewed all of the 120 protocols and selected all those who had objectively confirmed venous thromboembolism (n = 54). We invited them and 108 controls for further analysis. Thirty patients (22 had DVT, 7 DVT and pulmonary embolism, and 1 only pulmonary embolism) and 52 controls responded to our invitation and were willing to participate. The study was approved by the State Ethical Committee on Human Research, and full informed consent was obtained from all patients and controls.
Women were recruited to the study in the chronic stable phase of disease 6 months to 6 years (mean 3.5 years) after the acute event. Clinical examination was performed on the same day as blood sampling in 1995. All patients and controls were examined within a period of 2 months.
Blood samples were drawn from the antecubital vein between 7 and 9 am, with minimal venous stasis. Women were fasting overnight and rested before blood sampling for 20 minutes in a sedentary position. Plasma was prepared by centrifugation at 2,000g and 4°C for 30 minutes, snap frozen in liquid nitrogen, and stored at –70°C until analyzed. Routine laboratory measurements and measurements of fibrinolytic and coagulation tests were made on sera and/or plasma of all patients and controls (we performed the tests in 1995). However, measurements of IL-6 and antiphospholipid antibodies were made in 2001. Sera of all patients and of 37 controls were available for these measurements.
The metabolic variables (fasting glucose, triglycerides, total cholesterol, high-density lipoprotein [HDL] cholesterol, and low-density lipoprotein [LDL] cholesterol) were determined by routine biochemical methods. Patients and controls were screened for the presence of antinuclear antibodies (ANA), extractable nuclear antigens (ENA), and anti-DNA antibodies by routine screening tests from our Laboratory of Immunology. Tests for antithrombin, protein C deficiency were done by a kinetic spectrophotometric assay (Berichrom-Protein C and Berichrom Antithrombin II, Behring, Marburg/Lahn, Germany), and the presence of resistance to activated protein C was determined according to Dahlback et al.12 Plasminogen activator inhibitor-1 (PAI-1) antigen was determined using commercial enzyme-linked immunosorbent assay (ELISA) kits (Imulyse PAI-1, Biopool, Sweden) according to the instruction manual. Interleukin-6 was measured with commercial ELISA kits (Endogen, Pierce, Rockford, IL) according to the instruction manual. The detection limit of the assay was 1 pg/mL. Anticardiolipin and antiphosphatidylserine antibodies of immunoglobulin G (IgG) and M (IgM) isotypes were measured by ELISA kits used for routine measurements in our immunological laboratory as described previously.13 Anti-β2-glycoprotein I antibodies of IgG and IgM isotypes were measured by modified ELISA as described previously.14 Results for all measurements of antiphospholipid antibodies were expressed in optical density units OD405 × 1,000 (OD405). The individuals doing these analyses were masked with respect to the patient's status.
Normally distributed continuous clinical, biochemical, and hemostatic variables of patients and controls were compared by the t test for independent samples. Normality was evaluated by visual inspection of histograms and by means of the Kolmogorov-Smirnov test. Continuous variables that were not distributed normally were expressed as medians with ranges between the first and third quartile, and differences between groups were assessed by the Mann Whitney U test. The χ2 or Fisher exact test, depending on the number of cases included in the analysis, was used for comparison of differences between patients and controls for discrete variables. Depending on the distribution, Pearson or Spearman correlation coefficients were calculated to test associations between different variables. Women from the oral contraceptive group were younger than women from the non–oral contraceptive group at the time of blood sampling, and thus all comparisons were adjusted for age. Analysis of covariance was used to test the influence of age and other covariates on the differences in measured parameters between groups. Statistical analyses were performed by the Statistica for Windows computer program (StatSoft, Inc, Tulsa, OK). A P value < .05 was considered statistically significant.
In patients (n = 30), no significant differences were found compared with controls (n = 52) in the frequency of hypertension (7% versus 20%), hyperlipidemia (3% versus 0%), or smoking (33% versus 36%) at the time of the acute event. At the time of clinical examination none of the women was acutely ill or had clinical evidence of autoimmune disease. Neither patients nor controls were using aspirin or oral anticoagulants at least 1 month before blood sampling. Neither patients nor controls were taking oral contraceptives at the time of blood sampling, and there were no differences between patients and controls in either smoking or menstrual status at the time of blood sampling. However, before the acute event, significantly more patients had been taking oral contraceptives compared with controls (16/30, 56% versus 2/37, 5.4%, P < .000). All women had been taking second-generation oral contraceptives (containing low doses of ethynylestradiol and levonorgestrel); neither patients nor controls had used third-generation oral contraceptives containing desogestrel. No women had taken first-generation oral contraceptives in their lifetimes. For further analysis, patients were divided into 2 subgroups: the oral contraceptives group, women who were users of oral contraceptives before the acute event (n = 16), and non–oral contraceptives group, women with venous thromboembolism who had not used oral contraceptives before the acute event (n = 14).
Anthropometric and metabolic parameters of patients and controls at the time of clinical examination and blood sampling are presented in Table 1. Women from the oral contraceptives group were younger, leaner, and had lower levels of insulin compared with controls, whereas women from the non–oral contraceptives group were more obese and had higher insulin and PAI-1 concentrations compared with both controls and women from the oral contraceptives group (Table 1). Women from the oral contraceptives group did not have significantly different anthropometric and metabolic parameters compared with their age-matched controls (data not shown).
Autoimmune screening tests (ANA, ENA, and anti-DNA antibodies) were negative in all patients and controls. Tests for thrombophilia showed that neither patients nor controls had antithrombin or protein C deficiency; 3 patients and 1 control were resistant to activated protein C (difference not significant).
Concentrations of IL-6 are shown in Table 2. They were above the detection limit in 68% of controls and 75% of patients. In the oral contraceptives group, values of IL-6 were higher than in controls (P = .05), the difference remaining significant after adjustment for parameters measured at the time of blood sampling: body mass index (BMI), waist/hip ratio, age, smoking status, arterial blood pressure, cholesterol, HDL and LDL cholesterol, triglycerides, and blood glucose (Table 2). In the oral contraceptives group, IL-6 did not correlate with either metabolic parameters or PAI-1 antigen, whereas in the non–oral contraceptives group, IL-6 correlated with BMI (r = 0.53, P = .06).
Median concentrations of antiphospholipid antibodies with ranges between the first and third quartiles are shown in Table 3. Immunoglobulin M anti-β2-glycoprotein I levels were significantly higher in the oral contraceptives group and anticardiolipin IgG in the non–oral contraceptives group compared with controls. Immunoglobulin M anti-β2-glycoprotein I correlated significantly negatively with BMI (r = –0.59, P < .001), waist/hip ratio (r = –0.51, P < .005), insulin (r = –0.57, P < .001), and PAI-1 antigen (r = –0.62, P < .001). Anticardiolipin IgG correlated significantly negatively with HDL cholesterol in the non–oral contraceptives group (r = –0.68, P < .01).
When we tested associations between antiphospholipid antibodies and IL-6, we found a positive correlation between IL-6 and IgG anti-β2-glycoprotein I in lean patients (BMI ≤ 25 kg/m2) who used oral contraceptives at the time of the acute event (n = 10) (r = 0.56, P < .05). In the oral contraceptives group, lean women who had high levels of IL-6 (over the third quartile) (n = 5) also had higher IgG anti-β2-glycoprotein I compared with women with low levels of IL-6 (n = 5) (median interquartile range 20 [0–26] versus 6 [–13 to –6.5] OD450, P < .05) and with lean controls (n = 18) (median interquartile range 20 [0–26] versus 3 [–14 to –9] OD450, P < .05).
In our study we explored the presence of increased levels of IL-6 and antiphospholipid antibodies, both of which have prothrombotic capacity,9,10 in young women who had developed first venous thromboembolism during use or non-use of oral contraceptives. In addition, we explored whether any associations between IL-6 and antiphospholipid antibodies exist.
Our results lead us to construct a new hypothesis about thrombosis potential in young women with oral contraceptive–related venous thromboembolism. However, we are aware of the possible difficulties of interpreting these results because of 2 important points that can occur in epidemiological studies. The first is that significant associations might be found only by chance because of the measurement of a large number of variables in several groups with small numbers of patients. The second is that, in patients with a history of thrombosis, it is impossible to know whether differences between patients and controls are related to or the consequence of the clinical event.
In young women who had suffered venous thromboembolism during the use of oral contraceptives, we found increased levels of IL-6 and anti-β2-glycoprotein I antibodies, whereas, in women who had suffered idiopathic venous thromboembolism (and did not use oral contraceptives) and in controls, IL-6 and anti-β2-glycoprotein I antibodies were not increased. Because venous thromboembolism is a multicausal disease as proposed by Rosendaal15 in his dynamic, age-dependent model of thrombosis potential, we hypothesized that increased production of IL-6 and antiphospholipid antibodies in women with oral contraceptive–related venous thromboembolism could represent additional risk factors for development of venous thromboembolism (Fig. 1).
Besides its relation to bacterial infections, IL-6 was found to be related to age, parameters of obesity, smoking status, arterial blood pressure, cholesterol, HDL and LDL-cholesterol, triglycerides, and blood glucose.16–23 Although, in our cohort of women with idiopathic venous thromboembolism, IL-6 was related to obesity, increased levels of IL-6 in women with oral contraceptive–related venous thromboembolism were not associated with obesity and other factors known to increase IL-6. Thus, it appears that other factors must be involved in high IL-6 production in these women.
Recently, it has been shown that increased IL-6 production could be induced by oral contraceptive use and activation of endothelial cells with anti-β2-glycoprotein I antibodies.24–26 Furthermore, it has been shown that IL-6 induces production of antiphospholipid antibodies by B lymphocytes.27–28 Taking this into account, we hypothesized that in women with oral contraceptive–related venous thromboembolism who had increased IL-6 and anti-β2-glycoprotein I antibodies, production of IL-6 and anti-β2-glycoprotein I antibodies could be triggered by oral contraceptives (Fig. 2). This mechanism might lead to additive or supra-additive effects of IL-6 and anti-β2-glycoprotein I antibodies with oral contraceptive use (Fig. 1).
According to this hypothesis, it is of interest whether a correlation between IL-6 and anti-β2-glycoprotein I antibodies exists in our patients. To avoid the influence of increased BMI, which is associated with increased IL-6, we selected only “lean” patients and controls (with BMI ≤ 25 kg/m2) to test this association. We found a significant association between IL-6 and anti-β2-glycoprotein I antibodies in patients who used oral contraceptives at the time of the acute event, but not in non–oral contraceptive users and controls. Further significant differences in anti-β2-glycoprotein I antibodies were found between lean oral contraceptive patients who had IL-6 above and below the third quartile of the controls. These results are consistent with our hypothesis on the supra-additive thrombosis potential of a combination of increased IL-6 and anti-β2-glycoprotein I in women with oral contraceptive–related venous thromboembolism (Fig. 1), as well as with a possible explanation of increased IL-6 and anti-β2-glycoprotein I production during oral contraceptive use in high IL-6 producers (Fig. 2). However, we are aware that this hypothesis is not convincingly supported by our data because of the small sample size. New studies on a larger number of patients should be conducted in the future.
We found significantly increased IL-6 and anti-β2-glycoprotein I levels in young women with oral contraceptive–related venous thromboembolism. Further, in lean patients, increased levels of IL-6 correlated with anti-β2-glycoprotein I. These findings support the hypothesis that, in susceptible women, use of oral contraceptives induces production of IL-6, which stimulates production of anti-β2-glycoprotein I, which in turn further stimulates IL-6 production. Thus, the prothrombotic profile is aggravated and could facilitate the occurrence of venous thromboembolism. This new hypothesis should be further explored to try to define the group of young women susceptible to development of venous thromboembolism during the use of oral contraceptives. It might be that high levels of IL-6 before or after the introduction of oral contraceptives could identify these patients. This hypothesis remains to be elucidated in further prospective studies.
1. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Venous thromboembolic disease and combined oral contraceptives: results of an international multicentre case-control study. Lancet 1995;346:1575–82.
2. Nightingale AL, Lawrenson RA, Simpson EL, Williams TJ, MacRae KD, Farmer RD. The effects of age, body mass index, smoking and general health on the risk of venous thromboembolism in users of combined oral contraceptives. Eur J Contracept Reprod Health Care 2000;5:265–74.
3. Farmer RD, Lawrenson RA, Todd JC, Williams TJ, MacRae KD, Tyrer F, et al. A comparison of the risks of venous thromboembolism disease in association with different combined oral contraceptives. Br J Clin Pharmacol 2000;49:580–90.
4. Hannaford P. Cardiovascular events associated with different combined oral contraceptives: a review of current data. Drug Saf 2000;22:361–71.
5. Vandenbroucke JP, Koster T, Briet E, Reitsma PH, Bertina RM, Rosendaal FR. Increased risk of venous thrombosis in oral contraceptive users who are carriers of factor V Leiden mutation. Lancet 1994;344:1453–7.
6. Martinelli I, Taioli E, Bucciarelli P, Akhavan S, Mannucci PM. Interaction between the G20210A mutation of the prothrombin gene and oral contraceptive use in deep vein thrombosis. Arterioscler Thromb Vasc Biol 1999;19:700–3.
7. Bloemenkamp KW, Rosendaal FR, Helmerhorst FM, Vanderbroucke JP. Higher risk of venous thrombosis during early use of oral contraceptives in women with inherited clotting defects. Arch Intern Med 2000;160:49–52.
8. Van Aken BE, den Heijer M, Bos GMJ, van Deveter SJH, Reitsma PH. Recurrent venous thrombosis and markers of inflammation. Thromb Haemost 2000;83:536–9.
9. Greaves M. Antiphospholipid antibodies and thrombosis. Lancet 1999;353:1348–53.
10. Stouthard JM, Levi M, Hack CE, Veenhof CHN, Romijn JA, Sauerwein HP, et al. Interleukin-6 stimulates coagulation, not fibrinolysis, in humans. Thromb Haemost 1996;76:738–42.
11. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. A multinational case-control study of cardiovascular disease and steroid hormone contraceptives: description and validation of methods. J Clin Epidemiol 1995;48:1513–47.
12. Dahlback B, Carlsson M, Svensson PJ. Familial thrombophilia due to a previously unrecognized mechanism characterized by poor anticoagulant response to activated protein C: prediction of a cofactor to activated protein C. Proc Natl Acad Sci U S A 1993;90:1004–8.
13. Stegnar M, Božič B, Peternel P, Kveder T, Vene N, Rozman B. Prevalence of antiphospholipid antibodies in deep vein thrombosis and their relationship to blood coagulation and fibrinolysis. Thromb Res 1991;63:433–43.
14. Cˇučnik S, Ambrožič A, Božič B, Skitek M, Kveder T. Anti-β2-glycoprotein I ELISA: methodology, determination of cut-off values in 434 healthy Caucasians and evaluation of monoclonal antibodies as possible international standards. Clin Chem Lab Med 2000;38:777–83.
15. Rosendaal FR. Venous thrombosis: a multicausal disease. Lancet 1999;353:1167–73.
16. Abram M, Vučkovič D, Wraber B, Dorič. Plasma cytokine response in mice with bacterial infection. Mediators Inflamm 2000;9:229–34.
17. Tomšič-Matić M, Derganc M, Wraber B, Primožič J. Interleukin-6 (IL-6) and soluble receptors for interleukin-2 (sIL-2R) in the diagnosis of early severe infection in the critically ill newborns. Pflugers Arch 2000;440(suppl):R75–7.
18. Wei J, Xu H, Davies JL, Hemmings GP. Increase of plasma IL-6 concentration with age in healthy subjects. Life Sci 1992;51:1953–6.
19. Castell JV, Gomez-Lechon MJ, David M, Andus T, Geiger T, Trullenque R, et al. Interleukin-6 is the major regulator of acute phase protein synthesis in adult human hepatocytes. FEBS Lett 1989;242:237–9.
20. Bastard JP, Maachi M, Van Nhieu JT, Jardel C, Bruckert E, Grimaldi A, et al. Adipose tissue IL-6 content correlates with resistance to insulin activation of glucose uptake both in vivo and in vitro. J Clin Endocrinol Metab 2002;87:2084–9.
21. Jonkers IJ, Mohrschladt MF, Westendorp RG, van der Laarse A, Smelt AH. Severe hypertriglyceridemia with insulin resistance is associated with systemic inflammation: reversal with bezafibrate therapy in a randomized controlled trial. Am J Med 2002;112:275–80.
22. Kern PA, Ranganathan S, Li C, Wood L, Ranganathan G. Adipose tissue tumor necrosis factor and interleukin-6 expression in human obesity and insulin resistance. Am J Physiol Endocrinol Metab 2001;280:E745–51.
23. Laimer M, Ebenbichler CF, Kaser S, Sandhofer A, Weiss H, Nehod H, et al. Markers of chronic inflammation and obesity: a prospective study on the reversibility of this association in middle-aged women undergoing weight loss by surgical intervention. Int J Obes Relat Metab Disord 2002;26:659–62.
24. Bonafe M, Olivieri F, Cavallone L, Giovagnetti S, Mayegiani F, Cardelli M, et al. A gender-dependent genetic predisposition to produce high levels of IL-6 is detrimental for longevity. Eur J Immunol 2001;31:2357–61.
25. Salkeld BD, MacAulay JC, Ball RW, Cannon JG. Modulation of body temperature, interleukin-6 and leptin by oral contraceptive use. Neuroimmunomodulation 2001;9:319–25.
26. Arai T, Yoshida K, Kaburaki J, Inoko H, Ikeda Y, Kawakami Y, et al. Autoreactive CD4+ T-cell clones to β2-glycoprotein I in patients with antiphospholipid syndrome: preferential recognition of the major phospholipid-binding site. Blood 2001;98:1889–96.
27. Del Papa N, Guidali L, Spatola L, Bonara P, Borghi MO, Tincani A, et al. Relationship between anti-phospholipid and anti-EC antibodies: β2
glycoprotein I mediates the antibody binding to endothelial membranes and induces the expression of adhesion molecules. Clin Exp Rheumatol 1995;13:179–85.
28. Del Papa N, Guidali L, Sala A, Buccellati C, Khamshata MA, Ichikawa K, et al. Endothelial cells as targets for antiphospholipid antibodies: human polyclonal and monoclonal anti-β2
-glycoprotein I antibodies react in vitro with endothelial cells through adherent β2
glycoprotein I and induce endothelial activation. Arthritis Rheum 1997;40:551–61.