The metabolic impact of oral contraceptives Ronald M. Krauss, MD: and Ronald T. Burkman, Jr., MDb Berkeley, California, and Detroit, Michigan The hormonal components of oral contraceptives exert major effects on plasma lipoprotein metabolism. Estrogens may increase production of plasma triglycerides, leading to increased levels of very low-density lipoproteins, but they may also reduce levels of cholesterol-enriched and potentially atherogenic intermediate- and lOW-density lipoproteins. Furthermore, estrogens increase levels of high-density lipoproteins (HDLs), particularly the HDL2 subspecies, an effect linked to reduced mortality rates from cardiovascular disease in postmenopausal women receiving hormone replacement therapy. All combination oral contraceptives in use in the United States tend to raise levels of plasma triglycerides, lOW-density lipoprotein, and HDL3 to varying degrees. In contrast, changes in HDL and HDL2 reflect the combined effects of estrogen dose and relative androgenicity of the progestin component. Although in general, the lipoprotein changes are greater in magnitude with higher dose oral contraceptive preparations, they can be significant in lower dose preparations as well. Oral contraceptives also affect carbohydrate metabolism, primarily through the activity of progestin. Studies have demonstrated insulin resistance, rises in plasma insulin, and relative glucose intolerance by means of curve analysis of glucose tolerance tests. These effects are far less pronounced with lower dose preparations and with formulations using the newer progestins. (AM J OBSTET GYNECOL 1992;167:1177-84.)
Key words: Metabolic impact, lipid and carbohydrate metabolism, cardiovascular risk A potential relationship may exist between the metabolic effects of oral contraceptives (OCs) and the risk of cardiovascular disease. OCs may influence cardiovascular risk by direct effects on the vessel wall, by altering plasma lipoprotein, carbohydrate, or prostaglandin metabolism, or by altering coagulation parameters. The most significant findings to date concern lipoprotein metabolism. Numerous preclinical and clinical studies have consistently identified changes in lipoprotein metabolism as one of the major risk factors for the development of atherosclerosis.
Lipoprotein metabolism Appreciation of OC-induced changes in lipoprotein metabolism requires assessment of the role various forms of lipoprotein play in the genesis of atherosclerosis (Fig. I). The liver synthesizes a high molecular-weight protein (designated apolipoprotein B-IOO), which is secreted as an integral component of very low-density lipoproteins (VLDL). VLDL contain triglyceride, smaller amounts of cholesterol, and additional smaller proteins identified as apolipoprotein E and apolipoprotein C. VLDL transport endogenously synthesized triglycerides in the blood and are metabolized in tissues by the action of the enzyme lipoprotein lipase, leaving From the Lawrence Berkeley Labomtory, University of California,' and the Department of Gynecology-Obstetrics, Henry Ford Hospital' Reprint requests: Ronald M. Krauss, MD, Molecular Medicine Research, Lawrence Berkelev Laboratory, Donner Lahomtory 465, Berkeley, CA 94720." . 610127040
remnant forms containing reduced triglyceride content that may be picked up and rapidly cleared by the liver. By this means, triglyceride and cholesterol are transported to the periphery and the vehicle returned to the liver for further processing and excretion. However, in Western societies in which a high-fat diet predominates, the transport and clearance system of many persons becomes saturated. As a result, remnant particles remain in plasma, where additional triglyceride is removed, leaving predominantly cholesterol-containing particles known as intermediate-density lipoproteins (IDLs). IDLs are further degraded to low-density lipoproteins (LDL), which contain most of the cholesterol found in human plasma. In this process, the original structural protein, apolipoprotein B-lOO, remains on the LDL particle, whereas the smaller proteins are removed. LDL particles are taken up by specific hepatic receptors. However, when the system is saturated, excess LDL, IDL, and perhaps VLDL remnants are removed more slowly from plasma and may be susceptible to modification by abnormal mechanisms. It is currently believed that oxidation predisposes such particles to accumulate within macrophages in the artery wall. The accumulating lipid deposits interact with platelets, other thrombotic mechanisms, and tissue factors to promote growth of atherosclerotic plaques. High-density lipoproteins (HDLs) play a role in a process termed "reverse cholesterol transport" (Fig. 2). HDLs are formed from proteins, primarily apolipoprotein A-I and apolipoprotein A-II, that are synthesized by the liver and intestine and that combine with
1178
Krauss and Burkman
October 1992 Am J Obstet Gynecol
!
ApoB-100
I
Liver Receptors
Lipoprotein Lipase
VLDL Oxidation Macrophage uptake
Artery
~ Fig. 1. Mechanisms of lipoprotein metabolism and their role in atherogenesis.
Receptors:
Remnant
t
Liver
~jf
HDL
r+l
~(Q)~O IDL
ApOA-I_-=~_ ApoA-1I
LDL
0-
Fig. 2. HDL and reverse cholesterol transport. LCAT, Lecithin-cholesterol acyltransferase. APO, Apolipoprotein.
lipids removed from other lipoproteins and tissues. The cholesterol is esterified by an enzyme known as lecithin-cholesterol acyltransferase. Then additional apolipoprotein A-I and other proteins are incorporated into HDL to form a series of particles of progressively larger size and increased cholesterol content, known as HDL" HDL 2ao and HDL 2b • Greater heterogeneity exists
within these three HDL subclasses, which represent the m~or milestones in cholesterol uptake, with HDL 2b having the greatest transport capacity. At least three pathways exist by which HDLs return cholesterol to the liver. HDL may deliver cholesterol to remnants, which are then either taken up directly by the liver through receptor-mediated pathways or me-
Metabolic impact of oral contraceptives
Volume 167 r\umber 4, Part 2
Liver
............
HDL
A"""
Intestines
--LCAT--HDL3---~
1179
Shuttles cholesterol phospholipids from chylomicrons, VLDL and peripheral cells
-:--AffecC \ / D,et Progestins t Exercise Hormones Estrogens • Disease states Drugs Genetics Catalyzed by hepatic Gut, endothelial lipase ......... Lipoprotein VLDL ---'--'---~,---... FFA / lipase Liver Affect LDL Liver ~ Insulin, exercise, alcohol, fats, Estrogens t drugs, diet
I
t
Transports To periphery triglycerides
0.05 >0.05
NE (mg)
1.5 1.5
n
Total cholesterol
192 1123
192.5 198.4
57.5 59.7
106 12 72 19
201.8 209.6 208.3 215.4
60.0 53.0 63.3 58.8
I
HDL
1
LDL
121.0 124.2
I
121.1 136.4 123.5 135.4
Triglycerides
84.2 86.7
123.8 120.9 129.0 127.1
From Perlman JA, et al. Smoking. oral contraceptives, and other risk factors for atherosclerotic heart disease. In: Rosenberg M, ed. Smoking and reproductive health. New York: Praeger Medical, 1987:27-35. Reprinted by permission of Greenwood Publishing Group, Inc., Westport, Conn. *Adjusted for age, body mass index, cigarette, and caffeine use.
mains unexplained. We do not yet understand the biologic effects of apolipoprotein A-II; consequently, we cannot interpret the observed OC-induced increases in this protein in terms of cardiovascular disease risk or any other clinical consequences. Nevertheless, it is a phenomenon to be kept in mind as we learn more about this protein. A smaller study: one recently reproduced with a larger number of patients (Roy et aI., unpublished data) reflects results reported by others on lower dose preparations containing either NE or NG (Table V). Triglyceride effects were blunted and relatively small at these lower doses but clearly tended to be dose related. There was no significant change observed in the levels of LDL cholesterol. HDL cholesterol tended to decrease slightly with NG, although not significantly. However, a very significant trend was observed in the HDL2 subfractions. The NG preparation lowered HDL2b and the NE preparation raised it, although neither effect achieved significance at the p = 0.05 level. Another randomized clinical trial 8 with 63 to 70 patients in each of four groups studied preparations with a lower dose of estrogen (30 to 35 /Lg) but similar progestins, ED, NE, and LNG. There was a significant decline in HDL and a nonsignificant increase in apolipoprotein A-I with LNG. Both the ED and NE compounds were associated with increases in apolipoprotein A-I, and the ED compound with a slight increase in HDL," The observed tendencies in these studies confirmed previous reports that androgenic and estrogenic effects on HDL2 , although blunted, are still detectable with low-dose preparations. Relationship to cardiovascular risk
In terms of the effect on cardiovascular disease risk, the most important lipoprotein changes associated with OC use are those that occur in levels of HDL choles-
Table III. Relative progestin potency Estranes NE NEA ED Gonanes NG LNG
5-10 10-20
From Dorflinger LJ. Contraception 1985;31:557-70. terol, particularly the HDL 2b and apolipoprotein A-I components. As noted, these effects vary with the relative proportions of estrogen and progestin in the steroid combination and with the androgenic activity of the progestin component. The effects that occur with use of low-dose preparations are, on average, small. Nevertheless, it behooves the clinician to be aware of these potential changes in the context of the overall health status and specifically the cardiovascular risk status of female patients. Just as there are subsets of women in whom the presence of other risk factors may place them at increased cardiovascular risk with OC use, there will be subsets of women who have amplified metabolic changes with OC use. OC effects in the vessel wall
Recent data from a study of cynomolgus monkeys" suggest that estrogen may protect against the entry of cholesterol into vessel walls. After being fed a high-fat diet for 7 months, 50 female monkeys were randomized to receive no medication (n = 16), an intravaginal ring containing estrogen and LNG (n = 17), or an OC containing EE + LNG (n = 17). They were studied for 24 months while continuing on the same diet. As expected, HDL levels were reduced in the treated monkeys. Surprisingly, however, despite the observed adverse changes in HDL, the prevalence of plaques in the coronary arteries of OC-treated monkeys was lower than
1182
Krauss and Burkman
October 1992 Am J Obstet Gynecol
Table IV. Percentage change in lipids and lipoproteins: Baseline to I year % change Control
Total cholesterol T riglycerides LDL HDL HDL, HDL, Apolipoprotein A-I Apolipoprotein A-II
EE (50 fLg)+:
ED (l mg)
NEA (1 mg)
NG (0.5 mg)
9* 57* 10* 1 4 5 11* 17*
7* 45* 6 3 -3 11* 9* 18*
8* 32* 18* -13* -27* 5 -9 12*
-2 -5 0 -3 -I
-2 -4 -4
From Lipson A, et a\. Contraception 1986;34: 121-34, by permission of Butterworth-Heinemann.
*p < 0.05. Table V. Percentage change in lipids and lipoproteins: Baseline to 7 weeks o/c change Control (n = 5)
Cholesterol Triglyceride LDL cholesterol IDL mass§ LDL mass HDL cholesterol HDL mass HDL, mass HDL" mass HDL2b mass
EE (30-35 fLg) +:
I
NE (0.4 mg) (n = 8)
0 -10 0 0 -2 4 3 -10 3 16
13t 12 12 54 29:j: 11* 13 27* 3 59**
I
NG (0.3 mg) (n = 10)
2 18* 4 6 6 -4 -5 6 -8 -32**
From Krauss RM, et a\. AM] OBSTET GY~ECOL 1983;145:446-52. Change from baseline significant: **p < 0.08; *p < 0.05; tp < 0.01; :j:p < 0.001. §Total lipoprotein mass concentration as measured by analytic ultracentrifugation.
in control monkeys. The investigators suggested that this effect was probably related to a direct protective effect of estrogen on the uptake of LDL cholesterol at the arterial wall. The most recent data provided by the same investigators lO suggest the influence of two opposing OC effects relevant to atherogenesis. In this later study, the expected hierarchy of effects on HDL was observed in monkeys treated with OCs containing either ED or NG; the less androgenic progestin showed less effect than did the stronger progestin. A hierarchy of effects was also evident with regard to prevalence of plaque. Monkeys receiving no medication had the most plaque, those receiving the more potently androgenic progestin had less, and those receiving the least androgenic progestin had still less. The data, based on limited numbers of subjects, suggest that in these monkeys the protective effect of estrogen on the vessel wall is the most important factor influencing the atherogenic process. However, this finding does not negate the adverse lipoprotein effects of progestin that have been demonstrated in other biologic systems.
OC effect on carbohydrate metabolism
Considerably less is known about the effects of OC use on carbohydrate metabolism, and the relationship of these effects to cardiovascular risk, than is known about OC lipoprotein effects. There is a striking paucity of hard data on which to base the determination of appropriate prescribing practices. The difficulty is compounded by the finding of considerable individual variation in effects, as well as by considerable variation in the testing systems used in the studies and in study duration. Furthermore, most observed changes are in the nondiabetic range, compelling investigators to use analyses of area (or incremental area) under the curve in an attempt to evaluate OC impact. Theoretic concerns are based on the well-established observation that diabetic subjects are prone to accelerated atherogenesis and the development of blood vessel disease, on findings indicating that OCs will increase insulin levels and on the demonstration that hyperinsulinism has at least some potential for atherogenesis even in nondiabetic persons. Evaluating the validity of these concerns requires differentiating
Metabolic impact of oral contraceptives
Volume 167 Number 4, Part 2
1183
Table VI. Effects of OCs on carbohydrate metabolism in normal women"·16 Incremental AUG OGs
Glucose
Insulin
G-peptide
Monophasics* Triphasicst
t t
t t
t t
AUG, Area under the curve. *Differential effects: LNG 250/EE 30 > LNG 150/EE 30 > NE 1000/EE 35 > NE 500/EE 35, DG 150/EE 30 tLNG 50-125/EE 30-40 and NE 500-1000/EE 35. No differential effects.
between the effects of OC use in women with and without existing impairment of carbohydrate metabolism. Early studies ll demonstrated that most progestins in high-dose preparations had effects on carbohydrate metabolism in normal women. ED increased peripheral glucose and insulin; NE increased insulin but not glucose; and LNG increased both insulin and glucose most. A large cross-sectional studyI2 of the carbohydrate effects of 30 ILg of EE + 150 ILg of LNG found that impact was related to duration of use; oral glucose tolerance deteriorated, and insulin levels increased over approximately a 3- to 6-month perioa and then stabilized. The Walnut Creek Contraceptive Drug Study,13 which also evaluated the effects of high-dose OCs, showed increases in 1- and 2-hour glucose values, and increases in serum glucose that were dose related-5 to 10 mgldl for each milligram of NE, NEA, or ED, and 18 to 35 mg/dl for each milligram of LNG. Insulin resistance, as manifested by insulin levels after a carbohydrate meal, was increased; again, increases were correlated with the amount and type of progestin, with the more androgenic preparations having a greater effect. Studies using lower dose preparations have given variable results. A recent randomized triaP4 examining the three triphasic preparations currently available in the United States (two containing NE and one LNG) showed slight increases in glucose at 3 to 6 months and slight increases in mean insulin at 3 months. These increases were at a much lower level of magnitude than those observed previously with high-dose preparations. Very little difference was apparent between the effects of the various triphasic compounds. Preliminary data on seven OC formulations from a recent British cross-sectional studyI5 in 1060 OC users and 418 control subjects suggest that even lower dose preparations have detectable effects on insulin resistance as monitored by the incremental areas under the curve for glucose, insulin, and C-peptide (which reflects insulin secretion from the pancreas). 15.16 The data from that study are summarized in Table VI. Although reduced in magnitude compared with the effects of higher dose preparations, the effects of the lower dose
preparations were statistically significant and of potential clinical interest. A hierarchy of effects correlated with the relative biologic activity of the progestin component was demonstrated with the use of monophasic preparations but not with the use of triphasic preparations. Evaluation of the carbohydrate effects of OCs in women with previously impaired carbohydrate metabolism is limited by the small number of women studied and the variety of study designs that have been used. Szabo '7 studied 15 patients with gestational diabetes; 10 were control subjects and five were treated with 50 ILg of mestranol + 1 mg of NE for up to 6 months. Oral glucose tolerance tests deteriorated in all five treated women. Skouby et al. 'S studied 10 patients formerly having gestational diabetes and 8 nondiabetic control subjects who were treated with a preparation of 30 ILg of EE + 150 mg of LNG for 6 months; compared with control subjects, diabetic patients showed increases in plasma insulin but no significant changes in the overall glucose tolerance curve. More recently, Kung'9 studied 14 patients with gestational diabetes and 6 control subjects treated with a LNG triphasic for 6 months; the diabetic patients had higher overall plasma insulin concentrations and greater impairment of glucose tolerance. As we continue to test more women for the potential of gestational diabetes, increasing numbers are identified as being at risk. Considering that changes even within the normal range of an oral glucose tolerance test curve may have some impact on subsequent cardiovascular risk, the need for larger studies of OC impact in women with gestational diabetes is indicated. REFERENCES 1. Wahl P, Walden C, Knopp R, Hoover J. Effect of estrogen/progestin potency on lipid/lipoprotein cholesterol. N EnglJ Med 1983;308:862-7. 2. PerlmanJA, Krauss RM, Ray R, et al. Smoking, oral contraceptives. and other risk factors for atherosclerotic heart disease. In: Rosenberg M, ed. Smoking and reproductive health. New York: Praeger Medical, 1987:27-35. 3. Lipson A, Stoy DB, LaRosa JC, et al. Progestins and oral contraceptive-induced lipoprotein changes: a prospective study. Contraception 1986;34: 121-34. 4. Krauss RM, Roy S, Mishell DR Jr, et al. Effects of low-
Krauss and Burkman
5.
6. 7. 8.
9. 10. II. 12.
dose oral contraceptives on serum lipids and lipoproteins: differential changes in high-density lipoprotein subclasses. AM] OBSTET GYNECOL 1983; 145:446-52. Kuusi T, Nikkila EA, Tikkanen M], Sipinen S. Effects of two progestins with different androgenic properties on hepatic endothelial lipase and high density lipoprotein 2. Atherosclerosis 1985; 54: 251. Dorftinger L]. Relative potency of progestins used in oral contraceptives. Contraception 1985;31 :557 -70. Patsch W, Brown SA, Gotto AM, et al. The effect of triphasic oral contraceptives on plasma' lipids and lipoproteins. AM] OBSTET GYNECOL 1989;161:1396-401. Burkman RT, Robinson ]C, Kruszon-Moran D, et al. Lipid and lipoprotein changes associated with oral contraceptive use: a randomized clinical trial. Obstet Gynecol 1988;71:33-8. Adams MR, Clarkson TB, Koritnik DR. Contraceptive steroids and coronary artery atherosclerosis in cynomolgus macaques. Fertil Steril 1987;47:1010-8. Clarkson TB, Shively CA, Morgan TM, et al. Oral contraceptives and coronary artery atherosclerosis of cynomolgus monkeys. Obstet GynecoI1990;75:217-22. Spellacy WN. Carbohydrate metabolism in male infertility and female fertility-control patients. Ferti! Steril 1976;27: 1132-41. Wynn V. Effect of duration oflow-dose oral contraceptive administration on carbohydrate metabolism. AM] OBSTET GYNECOL 1982; 142:739-46.
October 1992 Am J Obstet Gynecol
13. Ramcharan S, ed. The Walnut Creek Contraceptive Drug Study. Bethesda: National Institutes of Health, 19741197611981; DHEW publication no 5 (NIH) 74-522, 76-563,81-564. 14. Bowes WA, Katta LR, Droegemueller W, et al. Triphasic randomized clinical trial: comparison of effects on carbohydrate metabolism. AM ] OBSTET GYNECOL 1989; 161: 1402. 15. Godsland IF, Crook D, Simpson R, et al. The effects of different formulations of oral contraceptive agents on lipid and carbohydrate metabolism. N Engl ] Med 1990;323: 1375. 16. Godsland IF, Crook D, Wynn V. Coronary heart disease risk markers in users of low-dose oral contraceptiv·es. ] Reprod Med 1991 ;36(suppl):226-37. 17. Szabo A], Cole HS, Grimaldi RD. Glucose tolerance in gestational diabetic women during and after treatment with a combination-type oral contraceptive. N Engl] Med 1970;282:646-50. 18. Skouby SO, Molsted-Pedersen L, Kuhl C. Low dosage oral contraception in women with previous gestational diabetes. Obstet Gynecol 1982;59:325-8. 19. Kung AWC, Ma ]TC, Wung BCW, et al. Glucose and lipid metabolism with triphasic oral contraceptives in women with a history of gestational diabetes. Contraception 1987;35:325-8.
Recent advances in understanding clotting and evaluating patients with recurrent thrombosis Barbara M. Alving, MD: and Philip C. Comp, MD, PhD b Washington, D.C., and Oklahoma City, Oklahoma Significant advances have been made in defining the regulatory mechanisms that control blood clotting. These are reviewed, with special attention to the functions of the natural inhibitors antithrombin III, protein C, and protein S. Congenital defiCiencies of these inhibitors as well as acquired abnormalities, such as defective fibrinolysis, and their role in promoting thrombosis are also discussed, as are thrombotic complications of pregnancy. Pregnancy decreases levels of protein S to 40% to 50% of normal levels. The decrease occurs early in pregnancy and persists into the postpartum period; it appears to be a hormonal rather than a dilutional effect. It is not known whether the thrombotic risk associated with pregnancy is increased in women who are congenitally deficient in protein S. Oral contraceptives decrease levels of protein S by about 20%. Women with a personal or family history of thrombOSis should be evaluated for predisposing conditions before they start an oral contraceptive, as should women taking oral contraceptives who develop deep venous thrombosis. (AM J OasTET GYNECOL 1992;167:1184-91.)
Key words: Clotting mechanisms, natural anticoagulant systems, recurrent venous thrombosis, deficiencies of natural clotting inhibitors, effects of pregnancy, oral contraceptives
From the Walter Reed Army Institute,' and the University of Oklahoma Depm·tment of Medicine.' Reprint requests: Philip C. Comp, MD, PhD. Thrombosis and Coagulation Laboratory, Oklahoma Memorial Hospital, Oklahoma Citv. OK 73126. 610128056
1184
Most patients with venous thromboses have no welldefined disorder of hemostasis. However, the identification of such abnormalities in some patients will permit rational decisions concerning contraceptive practices and long-term prophylaxis against thrombosis.
Key words: Metabolic impact, lipid and carbohydrate metabolism, cardiovascular risk A potential relationship may exist between the metabolic effects of oral contraceptives (OCs) and the risk of cardiovascular disease. OCs may influence cardiovascular risk by direct effects on the vessel wall, by altering plasma lipoprotein, carbohydrate, or prostaglandin metabolism, or by altering coagulation parameters. The most significant findings to date concern lipoprotein metabolism. Numerous preclinical and clinical studies have consistently identified changes in lipoprotein metabolism as one of the major risk factors for the development of atherosclerosis.
Lipoprotein metabolism Appreciation of OC-induced changes in lipoprotein metabolism requires assessment of the role various forms of lipoprotein play in the genesis of atherosclerosis (Fig. I). The liver synthesizes a high molecular-weight protein (designated apolipoprotein B-IOO), which is secreted as an integral component of very low-density lipoproteins (VLDL). VLDL contain triglyceride, smaller amounts of cholesterol, and additional smaller proteins identified as apolipoprotein E and apolipoprotein C. VLDL transport endogenously synthesized triglycerides in the blood and are metabolized in tissues by the action of the enzyme lipoprotein lipase, leaving From the Lawrence Berkeley Labomtory, University of California,' and the Department of Gynecology-Obstetrics, Henry Ford Hospital' Reprint requests: Ronald M. Krauss, MD, Molecular Medicine Research, Lawrence Berkelev Laboratory, Donner Lahomtory 465, Berkeley, CA 94720." . 610127040
remnant forms containing reduced triglyceride content that may be picked up and rapidly cleared by the liver. By this means, triglyceride and cholesterol are transported to the periphery and the vehicle returned to the liver for further processing and excretion. However, in Western societies in which a high-fat diet predominates, the transport and clearance system of many persons becomes saturated. As a result, remnant particles remain in plasma, where additional triglyceride is removed, leaving predominantly cholesterol-containing particles known as intermediate-density lipoproteins (IDLs). IDLs are further degraded to low-density lipoproteins (LDL), which contain most of the cholesterol found in human plasma. In this process, the original structural protein, apolipoprotein B-lOO, remains on the LDL particle, whereas the smaller proteins are removed. LDL particles are taken up by specific hepatic receptors. However, when the system is saturated, excess LDL, IDL, and perhaps VLDL remnants are removed more slowly from plasma and may be susceptible to modification by abnormal mechanisms. It is currently believed that oxidation predisposes such particles to accumulate within macrophages in the artery wall. The accumulating lipid deposits interact with platelets, other thrombotic mechanisms, and tissue factors to promote growth of atherosclerotic plaques. High-density lipoproteins (HDLs) play a role in a process termed "reverse cholesterol transport" (Fig. 2). HDLs are formed from proteins, primarily apolipoprotein A-I and apolipoprotein A-II, that are synthesized by the liver and intestine and that combine with
1178
Krauss and Burkman
October 1992 Am J Obstet Gynecol
!
ApoB-100
I
Liver Receptors
Lipoprotein Lipase
VLDL Oxidation Macrophage uptake
Artery
~ Fig. 1. Mechanisms of lipoprotein metabolism and their role in atherogenesis.
Receptors:
Remnant
t
Liver
~jf
HDL
r+l
~(Q)~O IDL
ApOA-I_-=~_ ApoA-1I
LDL
0-
Fig. 2. HDL and reverse cholesterol transport. LCAT, Lecithin-cholesterol acyltransferase. APO, Apolipoprotein.
lipids removed from other lipoproteins and tissues. The cholesterol is esterified by an enzyme known as lecithin-cholesterol acyltransferase. Then additional apolipoprotein A-I and other proteins are incorporated into HDL to form a series of particles of progressively larger size and increased cholesterol content, known as HDL" HDL 2ao and HDL 2b • Greater heterogeneity exists
within these three HDL subclasses, which represent the m~or milestones in cholesterol uptake, with HDL 2b having the greatest transport capacity. At least three pathways exist by which HDLs return cholesterol to the liver. HDL may deliver cholesterol to remnants, which are then either taken up directly by the liver through receptor-mediated pathways or me-
Metabolic impact of oral contraceptives
Volume 167 r\umber 4, Part 2
Liver
............
HDL
A"""
Intestines
--LCAT--HDL3---~
1179
Shuttles cholesterol phospholipids from chylomicrons, VLDL and peripheral cells
-:--AffecC \ / D,et Progestins t Exercise Hormones Estrogens • Disease states Drugs Genetics Catalyzed by hepatic Gut, endothelial lipase ......... Lipoprotein VLDL ---'--'---~,---... FFA / lipase Liver Affect LDL Liver ~ Insulin, exercise, alcohol, fats, Estrogens t drugs, diet
I
t
Transports To periphery triglycerides
0.05 >0.05
NE (mg)
1.5 1.5
n
Total cholesterol
192 1123
192.5 198.4
57.5 59.7
106 12 72 19
201.8 209.6 208.3 215.4
60.0 53.0 63.3 58.8
I
HDL
1
LDL
121.0 124.2
I
121.1 136.4 123.5 135.4
Triglycerides
84.2 86.7
123.8 120.9 129.0 127.1
From Perlman JA, et al. Smoking. oral contraceptives, and other risk factors for atherosclerotic heart disease. In: Rosenberg M, ed. Smoking and reproductive health. New York: Praeger Medical, 1987:27-35. Reprinted by permission of Greenwood Publishing Group, Inc., Westport, Conn. *Adjusted for age, body mass index, cigarette, and caffeine use.
mains unexplained. We do not yet understand the biologic effects of apolipoprotein A-II; consequently, we cannot interpret the observed OC-induced increases in this protein in terms of cardiovascular disease risk or any other clinical consequences. Nevertheless, it is a phenomenon to be kept in mind as we learn more about this protein. A smaller study: one recently reproduced with a larger number of patients (Roy et aI., unpublished data) reflects results reported by others on lower dose preparations containing either NE or NG (Table V). Triglyceride effects were blunted and relatively small at these lower doses but clearly tended to be dose related. There was no significant change observed in the levels of LDL cholesterol. HDL cholesterol tended to decrease slightly with NG, although not significantly. However, a very significant trend was observed in the HDL2 subfractions. The NG preparation lowered HDL2b and the NE preparation raised it, although neither effect achieved significance at the p = 0.05 level. Another randomized clinical trial 8 with 63 to 70 patients in each of four groups studied preparations with a lower dose of estrogen (30 to 35 /Lg) but similar progestins, ED, NE, and LNG. There was a significant decline in HDL and a nonsignificant increase in apolipoprotein A-I with LNG. Both the ED and NE compounds were associated with increases in apolipoprotein A-I, and the ED compound with a slight increase in HDL," The observed tendencies in these studies confirmed previous reports that androgenic and estrogenic effects on HDL2 , although blunted, are still detectable with low-dose preparations. Relationship to cardiovascular risk
In terms of the effect on cardiovascular disease risk, the most important lipoprotein changes associated with OC use are those that occur in levels of HDL choles-
Table III. Relative progestin potency Estranes NE NEA ED Gonanes NG LNG
5-10 10-20
From Dorflinger LJ. Contraception 1985;31:557-70. terol, particularly the HDL 2b and apolipoprotein A-I components. As noted, these effects vary with the relative proportions of estrogen and progestin in the steroid combination and with the androgenic activity of the progestin component. The effects that occur with use of low-dose preparations are, on average, small. Nevertheless, it behooves the clinician to be aware of these potential changes in the context of the overall health status and specifically the cardiovascular risk status of female patients. Just as there are subsets of women in whom the presence of other risk factors may place them at increased cardiovascular risk with OC use, there will be subsets of women who have amplified metabolic changes with OC use. OC effects in the vessel wall
Recent data from a study of cynomolgus monkeys" suggest that estrogen may protect against the entry of cholesterol into vessel walls. After being fed a high-fat diet for 7 months, 50 female monkeys were randomized to receive no medication (n = 16), an intravaginal ring containing estrogen and LNG (n = 17), or an OC containing EE + LNG (n = 17). They were studied for 24 months while continuing on the same diet. As expected, HDL levels were reduced in the treated monkeys. Surprisingly, however, despite the observed adverse changes in HDL, the prevalence of plaques in the coronary arteries of OC-treated monkeys was lower than
1182
Krauss and Burkman
October 1992 Am J Obstet Gynecol
Table IV. Percentage change in lipids and lipoproteins: Baseline to I year % change Control
Total cholesterol T riglycerides LDL HDL HDL, HDL, Apolipoprotein A-I Apolipoprotein A-II
EE (50 fLg)+:
ED (l mg)
NEA (1 mg)
NG (0.5 mg)
9* 57* 10* 1 4 5 11* 17*
7* 45* 6 3 -3 11* 9* 18*
8* 32* 18* -13* -27* 5 -9 12*
-2 -5 0 -3 -I
-2 -4 -4
From Lipson A, et a\. Contraception 1986;34: 121-34, by permission of Butterworth-Heinemann.
*p < 0.05. Table V. Percentage change in lipids and lipoproteins: Baseline to 7 weeks o/c change Control (n = 5)
Cholesterol Triglyceride LDL cholesterol IDL mass§ LDL mass HDL cholesterol HDL mass HDL, mass HDL" mass HDL2b mass
EE (30-35 fLg) +:
I
NE (0.4 mg) (n = 8)
0 -10 0 0 -2 4 3 -10 3 16
13t 12 12 54 29:j: 11* 13 27* 3 59**
I
NG (0.3 mg) (n = 10)
2 18* 4 6 6 -4 -5 6 -8 -32**
From Krauss RM, et a\. AM] OBSTET GY~ECOL 1983;145:446-52. Change from baseline significant: **p < 0.08; *p < 0.05; tp < 0.01; :j:p < 0.001. §Total lipoprotein mass concentration as measured by analytic ultracentrifugation.
in control monkeys. The investigators suggested that this effect was probably related to a direct protective effect of estrogen on the uptake of LDL cholesterol at the arterial wall. The most recent data provided by the same investigators lO suggest the influence of two opposing OC effects relevant to atherogenesis. In this later study, the expected hierarchy of effects on HDL was observed in monkeys treated with OCs containing either ED or NG; the less androgenic progestin showed less effect than did the stronger progestin. A hierarchy of effects was also evident with regard to prevalence of plaque. Monkeys receiving no medication had the most plaque, those receiving the more potently androgenic progestin had less, and those receiving the least androgenic progestin had still less. The data, based on limited numbers of subjects, suggest that in these monkeys the protective effect of estrogen on the vessel wall is the most important factor influencing the atherogenic process. However, this finding does not negate the adverse lipoprotein effects of progestin that have been demonstrated in other biologic systems.
OC effect on carbohydrate metabolism
Considerably less is known about the effects of OC use on carbohydrate metabolism, and the relationship of these effects to cardiovascular risk, than is known about OC lipoprotein effects. There is a striking paucity of hard data on which to base the determination of appropriate prescribing practices. The difficulty is compounded by the finding of considerable individual variation in effects, as well as by considerable variation in the testing systems used in the studies and in study duration. Furthermore, most observed changes are in the nondiabetic range, compelling investigators to use analyses of area (or incremental area) under the curve in an attempt to evaluate OC impact. Theoretic concerns are based on the well-established observation that diabetic subjects are prone to accelerated atherogenesis and the development of blood vessel disease, on findings indicating that OCs will increase insulin levels and on the demonstration that hyperinsulinism has at least some potential for atherogenesis even in nondiabetic persons. Evaluating the validity of these concerns requires differentiating
Metabolic impact of oral contraceptives
Volume 167 Number 4, Part 2
1183
Table VI. Effects of OCs on carbohydrate metabolism in normal women"·16 Incremental AUG OGs
Glucose
Insulin
G-peptide
Monophasics* Triphasicst
t t
t t
t t
AUG, Area under the curve. *Differential effects: LNG 250/EE 30 > LNG 150/EE 30 > NE 1000/EE 35 > NE 500/EE 35, DG 150/EE 30 tLNG 50-125/EE 30-40 and NE 500-1000/EE 35. No differential effects.
between the effects of OC use in women with and without existing impairment of carbohydrate metabolism. Early studies ll demonstrated that most progestins in high-dose preparations had effects on carbohydrate metabolism in normal women. ED increased peripheral glucose and insulin; NE increased insulin but not glucose; and LNG increased both insulin and glucose most. A large cross-sectional studyI2 of the carbohydrate effects of 30 ILg of EE + 150 ILg of LNG found that impact was related to duration of use; oral glucose tolerance deteriorated, and insulin levels increased over approximately a 3- to 6-month perioa and then stabilized. The Walnut Creek Contraceptive Drug Study,13 which also evaluated the effects of high-dose OCs, showed increases in 1- and 2-hour glucose values, and increases in serum glucose that were dose related-5 to 10 mgldl for each milligram of NE, NEA, or ED, and 18 to 35 mg/dl for each milligram of LNG. Insulin resistance, as manifested by insulin levels after a carbohydrate meal, was increased; again, increases were correlated with the amount and type of progestin, with the more androgenic preparations having a greater effect. Studies using lower dose preparations have given variable results. A recent randomized triaP4 examining the three triphasic preparations currently available in the United States (two containing NE and one LNG) showed slight increases in glucose at 3 to 6 months and slight increases in mean insulin at 3 months. These increases were at a much lower level of magnitude than those observed previously with high-dose preparations. Very little difference was apparent between the effects of the various triphasic compounds. Preliminary data on seven OC formulations from a recent British cross-sectional studyI5 in 1060 OC users and 418 control subjects suggest that even lower dose preparations have detectable effects on insulin resistance as monitored by the incremental areas under the curve for glucose, insulin, and C-peptide (which reflects insulin secretion from the pancreas). 15.16 The data from that study are summarized in Table VI. Although reduced in magnitude compared with the effects of higher dose preparations, the effects of the lower dose
preparations were statistically significant and of potential clinical interest. A hierarchy of effects correlated with the relative biologic activity of the progestin component was demonstrated with the use of monophasic preparations but not with the use of triphasic preparations. Evaluation of the carbohydrate effects of OCs in women with previously impaired carbohydrate metabolism is limited by the small number of women studied and the variety of study designs that have been used. Szabo '7 studied 15 patients with gestational diabetes; 10 were control subjects and five were treated with 50 ILg of mestranol + 1 mg of NE for up to 6 months. Oral glucose tolerance tests deteriorated in all five treated women. Skouby et al. 'S studied 10 patients formerly having gestational diabetes and 8 nondiabetic control subjects who were treated with a preparation of 30 ILg of EE + 150 mg of LNG for 6 months; compared with control subjects, diabetic patients showed increases in plasma insulin but no significant changes in the overall glucose tolerance curve. More recently, Kung'9 studied 14 patients with gestational diabetes and 6 control subjects treated with a LNG triphasic for 6 months; the diabetic patients had higher overall plasma insulin concentrations and greater impairment of glucose tolerance. As we continue to test more women for the potential of gestational diabetes, increasing numbers are identified as being at risk. Considering that changes even within the normal range of an oral glucose tolerance test curve may have some impact on subsequent cardiovascular risk, the need for larger studies of OC impact in women with gestational diabetes is indicated. REFERENCES 1. Wahl P, Walden C, Knopp R, Hoover J. Effect of estrogen/progestin potency on lipid/lipoprotein cholesterol. N EnglJ Med 1983;308:862-7. 2. PerlmanJA, Krauss RM, Ray R, et al. Smoking, oral contraceptives. and other risk factors for atherosclerotic heart disease. In: Rosenberg M, ed. Smoking and reproductive health. New York: Praeger Medical, 1987:27-35. 3. Lipson A, Stoy DB, LaRosa JC, et al. Progestins and oral contraceptive-induced lipoprotein changes: a prospective study. Contraception 1986;34: 121-34. 4. Krauss RM, Roy S, Mishell DR Jr, et al. Effects of low-
Krauss and Burkman
5.
6. 7. 8.
9. 10. II. 12.
dose oral contraceptives on serum lipids and lipoproteins: differential changes in high-density lipoprotein subclasses. AM] OBSTET GYNECOL 1983; 145:446-52. Kuusi T, Nikkila EA, Tikkanen M], Sipinen S. Effects of two progestins with different androgenic properties on hepatic endothelial lipase and high density lipoprotein 2. Atherosclerosis 1985; 54: 251. Dorftinger L]. Relative potency of progestins used in oral contraceptives. Contraception 1985;31 :557 -70. Patsch W, Brown SA, Gotto AM, et al. The effect of triphasic oral contraceptives on plasma' lipids and lipoproteins. AM] OBSTET GYNECOL 1989;161:1396-401. Burkman RT, Robinson ]C, Kruszon-Moran D, et al. Lipid and lipoprotein changes associated with oral contraceptive use: a randomized clinical trial. Obstet Gynecol 1988;71:33-8. Adams MR, Clarkson TB, Koritnik DR. Contraceptive steroids and coronary artery atherosclerosis in cynomolgus macaques. Fertil Steril 1987;47:1010-8. Clarkson TB, Shively CA, Morgan TM, et al. Oral contraceptives and coronary artery atherosclerosis of cynomolgus monkeys. Obstet GynecoI1990;75:217-22. Spellacy WN. Carbohydrate metabolism in male infertility and female fertility-control patients. Ferti! Steril 1976;27: 1132-41. Wynn V. Effect of duration oflow-dose oral contraceptive administration on carbohydrate metabolism. AM] OBSTET GYNECOL 1982; 142:739-46.
October 1992 Am J Obstet Gynecol
13. Ramcharan S, ed. The Walnut Creek Contraceptive Drug Study. Bethesda: National Institutes of Health, 19741197611981; DHEW publication no 5 (NIH) 74-522, 76-563,81-564. 14. Bowes WA, Katta LR, Droegemueller W, et al. Triphasic randomized clinical trial: comparison of effects on carbohydrate metabolism. AM ] OBSTET GYNECOL 1989; 161: 1402. 15. Godsland IF, Crook D, Simpson R, et al. The effects of different formulations of oral contraceptive agents on lipid and carbohydrate metabolism. N Engl ] Med 1990;323: 1375. 16. Godsland IF, Crook D, Wynn V. Coronary heart disease risk markers in users of low-dose oral contraceptiv·es. ] Reprod Med 1991 ;36(suppl):226-37. 17. Szabo A], Cole HS, Grimaldi RD. Glucose tolerance in gestational diabetic women during and after treatment with a combination-type oral contraceptive. N Engl] Med 1970;282:646-50. 18. Skouby SO, Molsted-Pedersen L, Kuhl C. Low dosage oral contraception in women with previous gestational diabetes. Obstet Gynecol 1982;59:325-8. 19. Kung AWC, Ma ]TC, Wung BCW, et al. Glucose and lipid metabolism with triphasic oral contraceptives in women with a history of gestational diabetes. Contraception 1987;35:325-8.
Recent advances in understanding clotting and evaluating patients with recurrent thrombosis Barbara M. Alving, MD: and Philip C. Comp, MD, PhD b Washington, D.C., and Oklahoma City, Oklahoma Significant advances have been made in defining the regulatory mechanisms that control blood clotting. These are reviewed, with special attention to the functions of the natural inhibitors antithrombin III, protein C, and protein S. Congenital defiCiencies of these inhibitors as well as acquired abnormalities, such as defective fibrinolysis, and their role in promoting thrombosis are also discussed, as are thrombotic complications of pregnancy. Pregnancy decreases levels of protein S to 40% to 50% of normal levels. The decrease occurs early in pregnancy and persists into the postpartum period; it appears to be a hormonal rather than a dilutional effect. It is not known whether the thrombotic risk associated with pregnancy is increased in women who are congenitally deficient in protein S. Oral contraceptives decrease levels of protein S by about 20%. Women with a personal or family history of thrombOSis should be evaluated for predisposing conditions before they start an oral contraceptive, as should women taking oral contraceptives who develop deep venous thrombosis. (AM J OasTET GYNECOL 1992;167:1184-91.)
Key words: Clotting mechanisms, natural anticoagulant systems, recurrent venous thrombosis, deficiencies of natural clotting inhibitors, effects of pregnancy, oral contraceptives
From the Walter Reed Army Institute,' and the University of Oklahoma Depm·tment of Medicine.' Reprint requests: Philip C. Comp, MD, PhD. Thrombosis and Coagulation Laboratory, Oklahoma Memorial Hospital, Oklahoma Citv. OK 73126. 610128056
1184
Most patients with venous thromboses have no welldefined disorder of hemostasis. However, the identification of such abnormalities in some patients will permit rational decisions concerning contraceptive practices and long-term prophylaxis against thrombosis.
