- Email: [email protected]
Effects of two low-dose oral contraceptives on serum lipids and lipoproteins: Differential changes in high-density lipoprotein subclasses
Effects of two low-dose oral contraceptives on serum lipids and lipoproteins: Differential changes in high-density lipoprotein subclasses RONALD
M.
SUBIR
KRAUSS
ROY
DANIEL
R.
JOHN
MISHELL,
JR.
CASAGRANDE
MALCOLM
C.
PIKE
und Los Angeles,
Berkeley
Cal{fornia
Oral contraceptives containing otxrorgestrel or norethindrone with ethinyl estradiol were administered by random assignment to 21 menstruating women, matched for anthropometric measurements, age, diet, alcohol consumption, smoking, and exercise habits. Pretreatment and 7-week treatment blood samples were obtained and assayed for serum cholesterol, triglyceride highdensity lipoprotein cholesterol (HDL-C), and total high-density lipoprotein (HDL), HDLPB, HDL,b, and HDL, subclasses by analytic ultracentrifugation. Subjects using the norethindrone oral contraceptive had a significant increase in HDL-C: baseline, 46 mgldl; 7 weeks, 51 mgldl. Values for the subjects using the norgestrel oral contraceptive were not significantly changed: 46 and 44 mg/dl, respectively. Users of the norethindrone oral contraceptive had significant elevations of total HDL and HDL3, while norgestrel oral contraceptive users demonstrated no significant changes. HDL,b increased with the norethindrone oral contraceptive and declined with the norgestrel oral contraceptive. The changes in HDL,b from baseline to treatment were not significant (p > 0.05) but the change with the norethindrone oral contraceptive did differ significantly from that with the norgestrel oral contraceptive (p < 0.02). These changes may indicate oral contraceptive-induced alterations in HDL structure and metabolism that could be related to the risk of development of atherosclerosis. (AM. J. OBSTET. GYNECOL. 145446, 1963.)
From the Donner Labomtory, University of California (Berkeley), and the Departments of Obstetrics and Gynecology and Family und Preventive Medicine, University of Southern Califor& School of Medicine. Finunciul support was provided by the United States Agency for International Development (Grant AIDlpha 1116), the International Development Research Center of Canudu, the Ford Foundation, the Rockefeller Foundation, the GeorgeJ. Hecht Fund, and National institutes of Health Research Grant HL-18574. The content does not necessarily reject the policies of any of the funding sources. This work
us part of the contraceptive sponsored and coordinated by the Committee for Contraception Research of the Council, Inc., New York, New York.
was undertaken
developnent program International Population
Presented at the Twenty-ninth Annual Meeting of the Societyfor Gynecologic Investigation, Dallas, Texas, March
24-27,
1982.
Reprint requests: Dr. Ronald M. Kraws, Dormer Laboratory, University of Cal$ornia, Berkeley, Cal$ornia 94720.
446
IN PREVIOUS STUDIES, contraceptives COntaininglevonorgestrel or norethindrone with ethinyl estradiol have been shown to produce certain estrogen-related metabolic alterations, such as increased serum triglyceride+ 2 and angiotensinogen3* 4 as well as a decrease in antithrombin III.3g 4 In addition, certain progestindominant oral contraceptives and progestin-only therapy have been shown to reduce high-density lipoprotein cholesterol (HDL-C) .*, 5 Recent reports suggest that an inverse relationship exists between HDL-C and the risk of development of coronary heart disease: The higher the HDL-C, the lower the risk for coronary heart disease.6-8 Reduction of the dose of the estrogen as well as the progestin component of oral contraceptives has been attempted in order to reduce these metabolic effects. This study was designed to evaluate the effects of two such low-dose oral contraceptive formulations on serum lipids and lipoproteins, especially 0002-9378/83/040446+07$00.70/O
@ 1983
The C. V. Mosby
Co.
Volume Number
14.5 4
HDL and its subclasses, and to compare, at similar estrogen doses, the effects of a potent progestin (levonorgesrel) with those of a weaker progestin (norethindrone).
Material and methods Subjects. The subjects were chosen from women attending the Family Planning Clinic at Los Angeles County-University of Southern California Medical Center. The five control and 21 oral contraceptive subjects were Caucasians or Mexican-Americans between 20 and 39 years of age, ingesting a standard American or Mexican-American diet, who drank no more than 2 ounces of alcohol per day and were either nonsmokers or smoked less than 15 cigarettes per day. None of the subjects was engaged in strenuous physical exercise or was receiving drugs known to affect lipid or lipoprotein metabolism. None of the experimental subjects had used oral contraceptives or had been pregnant within 3 months of enrollment in the study. The five control subjects had undergone previous tubal sterilization at least 6 months prior to entrance into the study. Materials. The oral contraceptives studied were Ovcon-35 containing 3.5 pg of ethinyl estradiol with 0.4 mg of norethindrone and Lo/Ovral containing 30 pg of ethinyl estradiol with 0.3 mg of dl-norgestrel, of which 0.15 mg was levonorgestrel, the active moiety. Allocation to study medication was randomized; however, one extra woman was inadvertently allocated to the norgestrel oral contraceptive group. All 21 study subjects as well as the five control subjects completed the study. However, analytic ultracentrifuge data were available for only 18 of the study subjects (eight in the norethindrone oral contraceptive group and 10 from the norgestrel oral contraceptive group). Anthropometric measurements. Anthropometric measurements included height in centimeters and weight in kilograms: Ponderal Index, defined as the height divided by the cube root of the weight; Quetelet Index, defined as the weight divided by the height squared times IO”; scapular fold thickness as determined by a micrometer instrument in centimeters; and iliac crest circumference in centimeters. Study design and methods. The oral contraceptives were used in a 3 weeks on/l week off cycle during the treatment period. Baseline blood samples were obtained during the luteal phase of the pretreatment menstrual cycle and in the follicular phase of the first treatment cycle (day 5). The treatment sample was obtained after the third week of the second oral contraceptive cycle. Subjects fasted 12 to 14 hours overnight before blood was sampled. The blood was allowed to clot for 1 hour at room temperature; after
Effects of low-dose oral contraceptives
447
low-speed centrifugation, serum was withdrawn and kept at 4” C for no longer than 2 weeks before measurements were made. Serum cholesterol and triglycerides were measured by enzymatic methods,?, ‘O as was HDL-C after precipitation of serum with sodium phosphotungstate and magnesium chloride.” Low-density lipoprotein cholesterol (LDL-C) was calculated by the equation of Friedewald and associates.‘” Analytical ultracentrifugation was used to measure total mass in very low-density lipoproteins (VLDL, flotation rate [Sy] 20 to 400), intermediate-density lipoproteins (IDL, Sy 12 to 20), low-density lipoproteins (LDL, SP 0 to 12), and high-density lipoproteins (HDL).!’ A computer analysis applied to the HDL determination allowed estimation of three HDL subclasses: HDL2,,, HDLZb, and HDL,.14 The skewness and kurtosis for all of the pretreatment values, both the raw data and logarithmically transformed data, were calculated. For those measurements which appeared to better fit a normal distribution when transformed, logarithms were used in further statistical analyses; otherwise, the raw data were used. For non-normally distributed variables, 95% confidence limits were calculated. The pretreatment values were found not to differ significantly from one another; therefore, the average was taken as the baseline value. The 95q confidence limits of each measured parameter were generated with these baseline values of all study and control subjects and IO additional subjects who were recruited for a parallel study with the same admission criteria (n = 36). The resulting values were used IO define reference ranges that were found to be consistent with previously published results by the same laboratory for a more heterogeneous population of subjects.‘” Statistical significance for differences between baseline and treatment values for each group was determined with paired t tests. The t test for independent samples was used to assess differences in baseline values for oral contraceptive and control groups. To adjust for any individual variation when between-group effects were assessed for any time point, each subject’s baseline value was subtracted from the respective treatment value, and these differences were evaluated for statistical significance, again with the t test. All computations were done with the use of the Statistical Package for the Social Sciences. The levei of significance for statistical tests was set at 0.05.
Results No significant differences between the control and oral contraceptive groups were noted for age, parity, obesity indices, cigarette smoking, or alcohol con-
448
Krauss et al.
Table
I. Serum
February Am. J. Obstet.
and Iipoprotein
lipids E thinrl estradiol-norethindrae (n = IO}
control (n = 5) Treatment
Baseline
169 6
168 6
160 6
59 39-91
53 38-75
49 43-56
E thinyl estradkl-norgestrel (n = II)
Treatment
II. Serum
181*t 8
177 7
181 6
127-218
60 48-74
67 39-l 14
55 48-63
6%
33-109
52-80
51 44-57
46 42-51
51$6 45-57
46 42-51
40-49
107 5
107 4
101 6
113 7
118 7
123 6
71-152
3.4 3.0-3.9
3.3 2.9-3.8
3.5 3.0-4.0
3.5 3.1-4.0
3.8 3.3-4.3
4.1t.J 3.6-4.6
2.5-5.0
44§
37-63 37-63
significant (p C 0.01).
lipoproteins Ethinyl estradiol-wethindrone
Control (n = 5) Baseline
VLDL (mg/dl): Median 95% CL IDL (mg/dl): Median 95% CL LDL (mgldl): Mean SEM HDL (mgldl): Mean SEM
range
36)
Treatment
95% CL = 95% Confidence limits. *Treatment versus baseline significant (p < 0.01). tA Treatment versus A control significant (p < 0.05). $Treatment versus baseline significant (p < 0.05). $A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel Table
Referace (n =
Baseline
Base&
Cholesterol (mg/dI): Mean SEM Triglycerides (mg/dl): Median 95% CL HDL-C (mg/dl): Median 95% CL LDL-C (mgldl): Mean SEM Cholesterol/HDL-C: Median 95% CL
15, 1983 Gynecol.
Ethinyl estradiul-norgestrel (n = 10)
(n = 8)
Treatment
Baseline
Treatment
Refeence (n =
range
36)
Baseline
Trea&ent
19 9-40
40* 23-68
3-148
17 5-53
3!:8
22 9-55
59 30-l 18
12 4-33
12 4-32
20 10-40
16 lo-28
19 11-32
2-63
7’i3
257 13
253 11
227 10
292~~6 18
293 17
3128 16
162-375
301 10
309 14
298 18
336$/(
292 11
27811 15
224-380
16
*Treatment versus baseline significant (p < 0.05). tTreatment versus baseline significant (p < 0.001). $A Treatment versus A control significant (p < 0.05). $A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel /( A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel sumption, as has been previously reported.16 Baseline serum lipid and lipoprotein levels were not significantly different among the groups, except for lower total LDL in the norethindrone oral contraceptive group (p < 0.05). The results for cholesterol, triglyceride, HDL-C, LDL-C, and the ratio cholesterol/HDL-C are shown in Table I. Baseline measures did not differ significantly
significant (p < 0.05). significant (p < 0.001).
among the groups. Among norethindrone oral contraceptive users cholesterol and HDL-C were significantly increased during treatment. The cholesterol change was significantly greater than values in the control or norgestrel oral contraceptive groups while the HDL-C incremental change with treatment was significantly different from that observed with norgestrel oral contraceptive use. Triglyceride was increased in
Volume Number
145 4
Table
III.
Effects of low-dose oral contraceptives
HDL
subclasses control
Ethinyl -nwethindrone
estmdiol
(n = 5)
HDL, (mgldl): Mean SEM HDL,, (mg/dl): Median 95% CL HDLSb (mg/dl): Median 95% CL
IV. LDL/HDL
Ethinyl &m&l-norgestrel
(n = 8)
(n = IO)
Baseline
Treatmmt
Baseline
135 14
123 11
130 4
165*t 5
127 4
135t 5
9’1-175
122 107-138
126 112-142
119 97-143
123 112-134
106 96-117
98 78-125
79-165
44 24-80
51 18-142
39 22-67
53 38-74
21-60
62$§ 34-l 11
Treatment
3W§
15-139
p = 0.06. significant (p < 0.05).
ratios Ethinyl
Baseline
*Treatment
range
36)
Treatment
Control (n = 5)
LDL/HDL: Median 95% CL IDL/HDL: Median 95% CL LDL + IDL/HDL: Median 95% CL
Refwence (n =
Baseline
*Treatment versus baseline significant (p < 0.05). tA Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel, $Treatment versus baseline, p = 0.08. §A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel
Table
449
Treatment
Ethiql
estradiol-mrethindwme
(n = 8)
estradiol-nargestrel (n = 10)
Baseline
Treatment
Baseline
Treatnwnt
Reference (n =
range
36)
0.84 0.75-0.96
0.83 0.70-0.98
0.76 0.63-0.91
0.79 0.67-0.94
0.99 0.81-1.20
1.13* 0.91-1.39
0.54-l .42
0.03 0.01-0.09
0.02 0.003-0.20
0.04 0.03-0.08
0.05 0.03-O. 11
0.06 0.03-O. 10
0.07 0.04-0.12
0.01-0.2 1
0.89 0.80-0.98
0.87 0.71-1.06
0.81 0.69-0.93
0.86 0.74-1.00
1.05 0.86-1.29
1.21* 0.98-1.50
0.58-l .50
versus baseline significant (p < 0.05).
the norgestrel oral contraceptive group, but the increase did not differ significantly from changes in the control and norethindrone oral contraceptive groups. CholesteroUHDL-C ratios were significantly increased in the norgestrel oral contraceptive group as compared with either pretreatment or control group values. None of these changes, however, resulted in values outside the reference range. VLDL, IDL, LDL, and HDL results are shown in Table II. VLDL was increased only in the norgestrel oral contraceptive group while IDL was unaffected by any formulation. LDL, which was significantly lower at baseline in the norethindrone oral contraceptive group than in the control or norgestrel oral contraceptive groups, increased with treatment in the norethindrone oral contraceptive group. This increase was significant in comparison with baseline values and with changes observed in the control and norgestrel oral contraceptive groups. The HDL levels during treatment were
significantly increased only in the norethindrone oral contraceptive group, and this incremental change with treatment was significantly different from that observed with norgestrel oral contraceptive use. None of the observed changes resulted in values outside the reference range. The results for the HDL subclasses are shown in Table III. HDL3 was significantly increased in the norethindrone oral contraceptive group and this increase differed from the change in the norgestrel oral contraceptive group (p = 0.058). Nonsignificant changes of HDLzb were produced by the norethindrone oral contraceptive (increasing, p = 0.082) and the norgestrel oral contraceptive (decreasing, p = 0.076). The incremental change with treatment was significantly different between these two formulations (p = 0.016). HDL,, levels were not altered by any of the preparations. None of the changes was outside the reference range.
450
Krauss et al.
Several ratios were examined where the numerator was LDL, IDL, or LDL + IDL, and the denominator was HDL, HDL,, + 2b, or HDLI (Table IV). HDL,, and HDLzb were combined because the values for HDLzb were small and ratios calculated with this parameter were potentially not reliable. The (LDL + 1DL)IHDL and LDL/HDL ratios were significantly increased from baseline with treatment in the norgestrel oral contraceptive group while IDL/HDL was unaffected. When HDLza + 2b was substituted in the denominator, the results paralleled those for total HDL, while none of the ratios with HDL, as the denominator was changed significantly with treatment (not shown).
Comment In this report we have sought to determine whether changes in serum lipids and lipoproteins are associated with use of two low-dose estrogen-progestin oral contraceptives. The results indicate that changes can occur with use of such agents for a 2-month period, and, as has been shown with higher-dose oral contraceptives,29 5 the results depend on the specific formulation used. With use of the norethindrone oral contraceptive, there were significant increases in total and HDL-C as well as increases in LDL and HDL mass. The increase in LDL, while significantly greater than in control or norgestrel oral contraceptive groups, must be interpreted with caution since the baseline level in the norethindrone oral contraceptive group was significantly lower than in the other groups. The relative increases in HDL-C and total HDL mass were similar (11% and 13%, respectively) and resulted in no changes in ratios of cholesterol/HDL-C or LDL/HDL mass. In contrast to the changes observed with use of the norethindrone oral contraceptive, there were no significant changes in serum total cholesterol, HDL-C, and total LDL and HDL mass with use of the norgestrel oral contraceptive. Despite the absence of significant changes in these individual parameters, there were small but significant increases in ratios of cholesterol/ HDL-C and LDL/HDL mass in the norgestrel oral contraceptive group. Increases in serum triglyceride and VLDL were observed with the norgestrel oral contraceptive but not with the norethindrone oral contraceptive. Interpretation of these data is made difficult by the well-known wide variation in serum VLDL and triglyceride levels in free-living subjects. The median VLDL levels increased more with the norethindrone oral contraceptive than with the norgestrel oral contraceptive, but there was a wide range of changes in the norethindrone oral contraceptive group and the differences were not significant at p < 0.05. In a recent
February Am. J. Obstet.
15, 1983 Gynecoi.
cross-sectional study,” use of combination contraceptive agents containing low doses of d-norgestrel(O.15 to 0.25 mg) was found to be associated with lower levels of total and HDL, cholesterol than were found in control subjects, with smaller and variable differences in HDLS cholesterol, while use of agents containing 1 mg of norethindrone or norethindrone acetate were not associated with changes in either HDL-C fraction. In another report, however, 35 pg of ethinyl estradiol/l mg of norethindrone was found to lower HDL-C and raise serum triglyceride between three and six cycles of use.l* A smaller increase in triglyceride and no change in HDL-C were seen after four cycles of use of an agent containing a smaller dose (0.5 mg) of norethindrone.18 Other investigators have reported a dose-related decrease in HDL-C within 1 month of treatment with ethinyl estradiol-levonorgestrel combinations with significant increases in plasma triglyceride levels at later time points. I9 When these results are compared with those in the present report, it seems likely that the dose of norethindrone used here (0.4 mg) resulted in relative dominance of an estrogen-mediated increase in HDL-C, while the dose of dl-norgestrel was sufficient to reverse this effect but was not sufficient to overcome estrogen-related increases in triglyceride and VLDL.” Differences in metabolic effects of the two oral contraceptive preparations studied here are also revealed by examination of changes in HDL subspecies. The increased HDL mass in norethindrone oral contraceptive users was accounted for primarily by an increase in HDLB mass, with a smaller and statistically insignificant increase in HDLzb and a minimal increase in HDL,,. The changes in both HDL3 and HDLzb differed significantly from those seen with the use of the norgestrel oral contraceptive, and in the case of HDLzb the change was in the opposite direction. These changes in HDL distribution as assessed by analytic ultracentrifugation may reflect alterations in structure and composition of HDL subclasses. Previously it has been reported that HDL triglyceride content is increasedZo and the ratio of apolipoprotein I/ apolipoprotein II is decreasedzl in users of estrogenprogestin combinations. Hormone-induced alterations in structure or composition of the HDL subclasses may also influence the properties of the subclasses as measured by analytic ultracentrifugation. Thus, it cannot be assumed that the procedure used to resolve three components within HDL in untreated subjectsI will yield measurements of these same components in plasma obtained after oral contraceptive treatment. However, irrespective of whether the physical characteristics of the HDL components change after oral contraceptive use, analytic ultracentrifugation provides a
Volume 145 Number
Effects of low-dose oral contraceptives
I
means of assessing changes in the distribution of HDL subspecies with treatment. The norethindrone oral contraceptive-induced increase of HDL in the HDL3 component is consistent with a previous report of higher levels of HDL, in women using oral contraceptives compared with nonuser control subjects. 22 Although a variety of oral contraceptive preparations were used by the women in that study, the majority were norethindrone-estrogen combinations. The norethindrone oral contraceptive-induced release in HDL, is to be contrasted with the findings in postmenopausal estrogen users, in whom increased levels of HDL are confined to the fasterHoating, more buoyant, and lipid-enriched HDLz species, primarily HDL,,.‘“, 24 Thus, norethindrone (but not norgestrel) appears to modify the effect of estrogen on HDL by shifting the increase in HDL mass to more dense particles with slower flotation rates. While the oral contraceptive-related changes in HDLzb did not reach statistical significance, the findings may be indicative of additional metabolic effects of oral contraceptive use. Norgestrel has been reported to lower levels of HDL, cholesterol in conjunction with stimulation of the activity of heparin-releasable hepatic lipase,25 an enzyme which appears to be involved in the catabolism of HDL, lipids. 26. Z7 While other metabolic factors may also be operative,** it is possible that the relative reduction in HDLzh with norgestrel oral contraceptive versus norethindrone oral contraceptive use may involve differences in oral contraceptive effects on hepatic lipase activity. Marked reductions in HDL2b as well as HDL,, have also been shown with use of a norgestrel-containing contraceptive vaginal ring,2s and it is likely that the mechanisms involved are similar to those operative with norgestrel oral contraceptive use. There is as yet insufficient information available to allow assessment of the possible clinical significance of the lipid and lipoprotein changes observed with oral
contraceptive use. For instance, if risk of coronary disease is increased by higher levels of total cholesterol and LDL, as was found here with norethindrone oral contraceptive use, it cannot be concluded from existing data that a proportional increase in HDL is sufficient to offset such a risk. The analysis is further clouded by the absence of information regarding the impact on coronary risk of an increase in the HDL3 component, since existing data suggest that HDL-C, the measure most commonly studied as a “negative risk factor,” is correlated with levels of HDL, but not HDL3?’ Also, it has been reported recently that cholesterol in HDL, was significantly reduced in patients with severe coronarv disease, while levels of HDL3 cholesterol were not rklated to extent of coronary disease.“’ Another consideration in assessing the present results is that effects associated with long-term use of oral contraceptives may differ from those observed with relatively short duration of use, and it is the long-term etl’ects that are most likely to influence the development of atherosclerosis. Finally, the quantitative significance of the increases in total cholesterol/HDL-C: and LDLIHDL ratios observed here with norgestrel oral ccmtraceptive use remains open to question. None of the resultant values were outside the reference ranges. and none were sufficiently high to result in a higher than average risk for coronary disease on the basis of taxisting epidemiologic data.7 Such changes might, howe\.er, compound the effects of other risk factors, such as cigaret.te smoking, and might be indicative of more substantial responses to oral contraceptive use in patients with underlying disorders of lipid metabolism. The contribution of Dr. Frank Lindgren and staff to the analytic ultracentrifuge measurements is greatly appreciated. Lipid analyses were performed with the expert assistance of Dennis Duncan and Janet Selmek. We also thank Linda Stimson and Linda Abe for preparation of the manuscript.
REFERENCES 1. Beck, P.: Contraceptive steroids: Modifications of carbohydrate and lipid metabolism, Metabolism 42:841, 1973. 2. Krauss, R. M.: Effects of progestational agents on serum lipids and lipoproteins, J. Reprod. Med. f7:503, 1982. 3. Mishell, D. R., Jr., Moore, D. F., Roy, S., Brenner, P. F., and Page, M. A.: Clinical performance and endocrine profiles with contraceptive vaginal rings containing a combination of estradiol and d-norgestrel, AM. J. OBSTET. GYNECOL. 130:55, 1978. 4. Roy, S., Mishel!. D. R., Jr., Gray, G., Donzano-Takano, R., Brenner, P. F., Eide, I., DeQuattro, V., and Shaw, S. T.: Comparison of metabolic and clinical effects of four oral contraceptive formulations and a contraceptive-vaginal ring, AM. J. OBSTET. GYNECOL. 136:920, 1980. 5. Bradley, D. D., Wingerd, J., Petitti. D. B., Krauss, R. M.,
451
6.
7.
8.
9.
and Ramcharan, S.: Serum high density lipoprotein cholesterol in women using oral contraceptives, estrogens and progestins, N. Engl. J. Med. 499: 17. 1978. Miller, G. J.. and Miller, N. E.: Plasma high density lipoprotein concentration and development nt’ ischaemic heart disease, Lancet 1:16, 1975. Castelli, W. B., Dorfe, J, T., Gordon, ‘I-., et al.: HDLcholesterol and other lipids in coronary heart disease: The cooperative lipoprotein phenotyping study, Circulation 55:767, 1977. Gordon, T., Caste&, W. P., Hjortland, M. C.. et al.: High density lipoproteins as a protective factor against coronary heart disease, Am. J. Med. 64:707, 1977. Allain, C. C., Poon, L. S., Chan, C. S. G.. Richmond, W., and Fu, P. C.: Enzymatic determination of rotal serum cholesterol, Clin. Chem. 20:470, 1974.
452
February Am. J. Obstet.
Krauss et al.
10. Bondar, R. J. L., Balwyot, E. S., and Hsu, M.: A totally enzymatic method for the determination of serum triglycerides, Clin. Chem. 23: 1132, 1977. (Abst.) 11. Lopes-Virella, M. F., Stone, P., Ellis, S., and Colwell,.I. A.: Cholesterol determination in high density lipoproteins senarated bv three different methods. Clin. Chem. 23’:882, 1977. 12. Friedwald, W. T., Levy, R. I., and Frederickson, D. S.: Estimation of concentration of low-density lipoprotein cholesterol without use of the preparative ultracentrifuge, Clin. Chem. l&499, 1972. 13. Ewing, A. M., Freeman, N. K., and Lindgren, F. T.: In Paoletti, R., and Kritchevsky, D., editors: Advances in Lipid Research, New York, 1965, Vol. 3, Academic Press, Inc., pp. 25-61. 14. Anderson, D. W., Nichols, A. V., Pan, S. S., and Lindgren, F. T.: High density lipo rotein distribution. Resolution and determination of tK ree major components in a normal population sample, Atherosclerosis 29:161, 1978. 15. Lindgren, F. T., Adamson, G. L., Jensen, L. C., and Wood, P. D.: Lipid and lipoprotein measurements in a normal adult American population, Lipids lOr750, 1975. 16. Roy, S., Krauss, R. M., k&shell, D. R., jr., Casagrande, J., and Pike. M. C.: The Metabolic Effects of a LowEstrogeniiow-Progestogen Oral Contraceptive, New York, 1981, Biomedical Information Corporation. 17. Wynn, V., and Niththyananthan, R.: The effect of progestins in combined oral contraceptives on serum lipids with special reference to high-density lipoproteins, AM. J. OBSTET.GYNECOL. 142:766, 1982. 18. Briggs, M. H., and Briggs, M.: A randomized study of metabolic effects of four low estrogen oral contraceptives. I. Results after six cycles, Contraception 23:463, 1981. 19. Larsson-Cohn, V., fihraeus, L., Wallentin, L., and Zador, G.: Lipoprotein changes may be minimized by proper comnosition of a combined oral contracentive. Fertil. 1 Sterii. 35:172, 1981. 20. Knopp, R. H., Walden, C. E., Wahl, P. W., Hoover, J. J., Warnick, G. R., Albers, J. J., Ogilvie, J. T., and Hazzard, W. R.: Oral contraceptive and postmenopausal estrogen effects on lipoprotein triglycerides and cholesterol in an adult female population: Relationships to estrogen and
21. 22.
23. 24.
25.
26.
27.
28. 29.
30. 31.
15, 1983 Gynecol.
progestin potency, J. Clin. Endocrinol. Metab. 53: 1123, 1981. Cheung, M. C., and Albers, J. J.: The measurement of apolipoprotein A-I and A-II levels in men and women by immunoassay, J. Clin. Invest. 60:43, 1977. Krauss, R. M., Lindgren, F. T., Silvers, A., Jutagir, R., and Bradley, D. D.: Changes in serum high density lipoproteins in women on oral contraceptive drugs, Clin. Chim. Acta 80:465, 1977. Krauss, R. M., Lindgren, F. T., Wingerd, J., Bradley, D. D., and Ramcharan, J.: Effects of estrogens and progestins on high density lipoproteins, Lipids 14: 113,1979. Krauss, R. M., Takaki, N., and Lewis, S. B.: Relation of serum high density lipoproteins (HDL) to gonadotropin and estrogen levels in postmenopausal estrogen users and non-users, in Fifth International Symposium on Arteriosclerosis, Houston, Texas, November, 1979. Tikkinen, M. J., Nikkill, E. A., Kuusi, T., and Sipinen, S.: Different effects of two progestins on plasma high density lipoprotein (HDL,) and postheparin plasma lipase activity, Atherosclerosis 40:365, 1981. Kuusi, T., Kinnunen, P. K. J., and Nikkila, E. A.: Hepatic endothelial lipase antiserum influences rat plasma low and high density lipoproteins in vivo, FEBS Lett. 104: 384, 1979. Kuusi, T., Saarinen, P., and Nikkili, E. A.: Evidence for the role of hepatic endothelial lipase in the metabolism of plasma high density lipoprotein, in man, Atherosclerosis 36:589, 1980. Krauss, R. M.: Regulation of high density lipoprotein levels, Med. Clin. North Am. 66:403. 1982. Roy, S., Krauss, R. M., Mishell, D. R.; Jr., Casagrande, J., and Pike, M. C.: The effects on lipids and lipoproteins of a contraceptive vaginal ring containing levonorgestrel and estradiol, Contraception 24429, 1981. Anderson, D. W.: HDL cholesterol: the variable components, Lancet 1:819, 1978. Miller, N. E., Hammett, F., Saltissi, S., Rao, S., Van Zeller, H., Coltant, J., and Lewis, B.: Relation of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins, Br. Med. J. 282: 1741, 1981.
M.
SUBIR
KRAUSS
ROY
DANIEL
R.
JOHN
MISHELL,
JR.
CASAGRANDE
MALCOLM
C.
PIKE
und Los Angeles,
Berkeley
Cal{fornia
Oral contraceptives containing otxrorgestrel or norethindrone with ethinyl estradiol were administered by random assignment to 21 menstruating women, matched for anthropometric measurements, age, diet, alcohol consumption, smoking, and exercise habits. Pretreatment and 7-week treatment blood samples were obtained and assayed for serum cholesterol, triglyceride highdensity lipoprotein cholesterol (HDL-C), and total high-density lipoprotein (HDL), HDLPB, HDL,b, and HDL, subclasses by analytic ultracentrifugation. Subjects using the norethindrone oral contraceptive had a significant increase in HDL-C: baseline, 46 mgldl; 7 weeks, 51 mgldl. Values for the subjects using the norgestrel oral contraceptive were not significantly changed: 46 and 44 mg/dl, respectively. Users of the norethindrone oral contraceptive had significant elevations of total HDL and HDL3, while norgestrel oral contraceptive users demonstrated no significant changes. HDL,b increased with the norethindrone oral contraceptive and declined with the norgestrel oral contraceptive. The changes in HDL,b from baseline to treatment were not significant (p > 0.05) but the change with the norethindrone oral contraceptive did differ significantly from that with the norgestrel oral contraceptive (p < 0.02). These changes may indicate oral contraceptive-induced alterations in HDL structure and metabolism that could be related to the risk of development of atherosclerosis. (AM. J. OBSTET. GYNECOL. 145446, 1963.)
From the Donner Labomtory, University of California (Berkeley), and the Departments of Obstetrics and Gynecology and Family und Preventive Medicine, University of Southern Califor& School of Medicine. Finunciul support was provided by the United States Agency for International Development (Grant AIDlpha 1116), the International Development Research Center of Canudu, the Ford Foundation, the Rockefeller Foundation, the GeorgeJ. Hecht Fund, and National institutes of Health Research Grant HL-18574. The content does not necessarily reject the policies of any of the funding sources. This work
us part of the contraceptive sponsored and coordinated by the Committee for Contraception Research of the Council, Inc., New York, New York.
was undertaken
developnent program International Population
Presented at the Twenty-ninth Annual Meeting of the Societyfor Gynecologic Investigation, Dallas, Texas, March
24-27,
1982.
Reprint requests: Dr. Ronald M. Kraws, Dormer Laboratory, University of Cal$ornia, Berkeley, Cal$ornia 94720.
446
IN PREVIOUS STUDIES, contraceptives COntaininglevonorgestrel or norethindrone with ethinyl estradiol have been shown to produce certain estrogen-related metabolic alterations, such as increased serum triglyceride+ 2 and angiotensinogen3* 4 as well as a decrease in antithrombin III.3g 4 In addition, certain progestindominant oral contraceptives and progestin-only therapy have been shown to reduce high-density lipoprotein cholesterol (HDL-C) .*, 5 Recent reports suggest that an inverse relationship exists between HDL-C and the risk of development of coronary heart disease: The higher the HDL-C, the lower the risk for coronary heart disease.6-8 Reduction of the dose of the estrogen as well as the progestin component of oral contraceptives has been attempted in order to reduce these metabolic effects. This study was designed to evaluate the effects of two such low-dose oral contraceptive formulations on serum lipids and lipoproteins, especially 0002-9378/83/040446+07$00.70/O
@ 1983
The C. V. Mosby
Co.
Volume Number
14.5 4
HDL and its subclasses, and to compare, at similar estrogen doses, the effects of a potent progestin (levonorgesrel) with those of a weaker progestin (norethindrone).
Material and methods Subjects. The subjects were chosen from women attending the Family Planning Clinic at Los Angeles County-University of Southern California Medical Center. The five control and 21 oral contraceptive subjects were Caucasians or Mexican-Americans between 20 and 39 years of age, ingesting a standard American or Mexican-American diet, who drank no more than 2 ounces of alcohol per day and were either nonsmokers or smoked less than 15 cigarettes per day. None of the subjects was engaged in strenuous physical exercise or was receiving drugs known to affect lipid or lipoprotein metabolism. None of the experimental subjects had used oral contraceptives or had been pregnant within 3 months of enrollment in the study. The five control subjects had undergone previous tubal sterilization at least 6 months prior to entrance into the study. Materials. The oral contraceptives studied were Ovcon-35 containing 3.5 pg of ethinyl estradiol with 0.4 mg of norethindrone and Lo/Ovral containing 30 pg of ethinyl estradiol with 0.3 mg of dl-norgestrel, of which 0.15 mg was levonorgestrel, the active moiety. Allocation to study medication was randomized; however, one extra woman was inadvertently allocated to the norgestrel oral contraceptive group. All 21 study subjects as well as the five control subjects completed the study. However, analytic ultracentrifuge data were available for only 18 of the study subjects (eight in the norethindrone oral contraceptive group and 10 from the norgestrel oral contraceptive group). Anthropometric measurements. Anthropometric measurements included height in centimeters and weight in kilograms: Ponderal Index, defined as the height divided by the cube root of the weight; Quetelet Index, defined as the weight divided by the height squared times IO”; scapular fold thickness as determined by a micrometer instrument in centimeters; and iliac crest circumference in centimeters. Study design and methods. The oral contraceptives were used in a 3 weeks on/l week off cycle during the treatment period. Baseline blood samples were obtained during the luteal phase of the pretreatment menstrual cycle and in the follicular phase of the first treatment cycle (day 5). The treatment sample was obtained after the third week of the second oral contraceptive cycle. Subjects fasted 12 to 14 hours overnight before blood was sampled. The blood was allowed to clot for 1 hour at room temperature; after
Effects of low-dose oral contraceptives
447
low-speed centrifugation, serum was withdrawn and kept at 4” C for no longer than 2 weeks before measurements were made. Serum cholesterol and triglycerides were measured by enzymatic methods,?, ‘O as was HDL-C after precipitation of serum with sodium phosphotungstate and magnesium chloride.” Low-density lipoprotein cholesterol (LDL-C) was calculated by the equation of Friedewald and associates.‘” Analytical ultracentrifugation was used to measure total mass in very low-density lipoproteins (VLDL, flotation rate [Sy] 20 to 400), intermediate-density lipoproteins (IDL, Sy 12 to 20), low-density lipoproteins (LDL, SP 0 to 12), and high-density lipoproteins (HDL).!’ A computer analysis applied to the HDL determination allowed estimation of three HDL subclasses: HDL2,,, HDLZb, and HDL,.14 The skewness and kurtosis for all of the pretreatment values, both the raw data and logarithmically transformed data, were calculated. For those measurements which appeared to better fit a normal distribution when transformed, logarithms were used in further statistical analyses; otherwise, the raw data were used. For non-normally distributed variables, 95% confidence limits were calculated. The pretreatment values were found not to differ significantly from one another; therefore, the average was taken as the baseline value. The 95q confidence limits of each measured parameter were generated with these baseline values of all study and control subjects and IO additional subjects who were recruited for a parallel study with the same admission criteria (n = 36). The resulting values were used IO define reference ranges that were found to be consistent with previously published results by the same laboratory for a more heterogeneous population of subjects.‘” Statistical significance for differences between baseline and treatment values for each group was determined with paired t tests. The t test for independent samples was used to assess differences in baseline values for oral contraceptive and control groups. To adjust for any individual variation when between-group effects were assessed for any time point, each subject’s baseline value was subtracted from the respective treatment value, and these differences were evaluated for statistical significance, again with the t test. All computations were done with the use of the Statistical Package for the Social Sciences. The levei of significance for statistical tests was set at 0.05.
Results No significant differences between the control and oral contraceptive groups were noted for age, parity, obesity indices, cigarette smoking, or alcohol con-
448
Krauss et al.
Table
I. Serum
February Am. J. Obstet.
and Iipoprotein
lipids E thinrl estradiol-norethindrae (n = IO}
control (n = 5) Treatment
Baseline
169 6
168 6
160 6
59 39-91
53 38-75
49 43-56
E thinyl estradkl-norgestrel (n = II)
Treatment
II. Serum
181*t 8
177 7
181 6
127-218
60 48-74
67 39-l 14
55 48-63
6%
33-109
52-80
51 44-57
46 42-51
51$6 45-57
46 42-51
40-49
107 5
107 4
101 6
113 7
118 7
123 6
71-152
3.4 3.0-3.9
3.3 2.9-3.8
3.5 3.0-4.0
3.5 3.1-4.0
3.8 3.3-4.3
4.1t.J 3.6-4.6
2.5-5.0
44§
37-63 37-63
significant (p C 0.01).
lipoproteins Ethinyl estradiol-wethindrone
Control (n = 5) Baseline
VLDL (mg/dl): Median 95% CL IDL (mg/dl): Median 95% CL LDL (mgldl): Mean SEM HDL (mgldl): Mean SEM
range
36)
Treatment
95% CL = 95% Confidence limits. *Treatment versus baseline significant (p < 0.01). tA Treatment versus A control significant (p < 0.05). $Treatment versus baseline significant (p < 0.05). $A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel Table
Referace (n =
Baseline
Base&
Cholesterol (mg/dI): Mean SEM Triglycerides (mg/dl): Median 95% CL HDL-C (mg/dl): Median 95% CL LDL-C (mgldl): Mean SEM Cholesterol/HDL-C: Median 95% CL
15, 1983 Gynecol.
Ethinyl estradiul-norgestrel (n = 10)
(n = 8)
Treatment
Baseline
Treatment
Refeence (n =
range
36)
Baseline
Trea&ent
19 9-40
40* 23-68
3-148
17 5-53
3!:8
22 9-55
59 30-l 18
12 4-33
12 4-32
20 10-40
16 lo-28
19 11-32
2-63
7’i3
257 13
253 11
227 10
292~~6 18
293 17
3128 16
162-375
301 10
309 14
298 18
336$/(
292 11
27811 15
224-380
16
*Treatment versus baseline significant (p < 0.05). tTreatment versus baseline significant (p < 0.001). $A Treatment versus A control significant (p < 0.05). $A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel /( A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel sumption, as has been previously reported.16 Baseline serum lipid and lipoprotein levels were not significantly different among the groups, except for lower total LDL in the norethindrone oral contraceptive group (p < 0.05). The results for cholesterol, triglyceride, HDL-C, LDL-C, and the ratio cholesterol/HDL-C are shown in Table I. Baseline measures did not differ significantly
significant (p < 0.05). significant (p < 0.001).
among the groups. Among norethindrone oral contraceptive users cholesterol and HDL-C were significantly increased during treatment. The cholesterol change was significantly greater than values in the control or norgestrel oral contraceptive groups while the HDL-C incremental change with treatment was significantly different from that observed with norgestrel oral contraceptive use. Triglyceride was increased in
Volume Number
145 4
Table
III.
Effects of low-dose oral contraceptives
HDL
subclasses control
Ethinyl -nwethindrone
estmdiol
(n = 5)
HDL, (mgldl): Mean SEM HDL,, (mg/dl): Median 95% CL HDLSb (mg/dl): Median 95% CL
IV. LDL/HDL
Ethinyl &m&l-norgestrel
(n = 8)
(n = IO)
Baseline
Treatmmt
Baseline
135 14
123 11
130 4
165*t 5
127 4
135t 5
9’1-175
122 107-138
126 112-142
119 97-143
123 112-134
106 96-117
98 78-125
79-165
44 24-80
51 18-142
39 22-67
53 38-74
21-60
62$§ 34-l 11
Treatment
3W§
15-139
p = 0.06. significant (p < 0.05).
ratios Ethinyl
Baseline
*Treatment
range
36)
Treatment
Control (n = 5)
LDL/HDL: Median 95% CL IDL/HDL: Median 95% CL LDL + IDL/HDL: Median 95% CL
Refwence (n =
Baseline
*Treatment versus baseline significant (p < 0.05). tA Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel, $Treatment versus baseline, p = 0.08. §A Treatment ethinyl estradiol-norethindrone versus ethinyl estradiol-norgestrel
Table
449
Treatment
Ethiql
estradiol-mrethindwme
(n = 8)
estradiol-nargestrel (n = 10)
Baseline
Treatment
Baseline
Treatnwnt
Reference (n =
range
36)
0.84 0.75-0.96
0.83 0.70-0.98
0.76 0.63-0.91
0.79 0.67-0.94
0.99 0.81-1.20
1.13* 0.91-1.39
0.54-l .42
0.03 0.01-0.09
0.02 0.003-0.20
0.04 0.03-0.08
0.05 0.03-O. 11
0.06 0.03-O. 10
0.07 0.04-0.12
0.01-0.2 1
0.89 0.80-0.98
0.87 0.71-1.06
0.81 0.69-0.93
0.86 0.74-1.00
1.05 0.86-1.29
1.21* 0.98-1.50
0.58-l .50
versus baseline significant (p < 0.05).
the norgestrel oral contraceptive group, but the increase did not differ significantly from changes in the control and norethindrone oral contraceptive groups. CholesteroUHDL-C ratios were significantly increased in the norgestrel oral contraceptive group as compared with either pretreatment or control group values. None of these changes, however, resulted in values outside the reference range. VLDL, IDL, LDL, and HDL results are shown in Table II. VLDL was increased only in the norgestrel oral contraceptive group while IDL was unaffected by any formulation. LDL, which was significantly lower at baseline in the norethindrone oral contraceptive group than in the control or norgestrel oral contraceptive groups, increased with treatment in the norethindrone oral contraceptive group. This increase was significant in comparison with baseline values and with changes observed in the control and norgestrel oral contraceptive groups. The HDL levels during treatment were
significantly increased only in the norethindrone oral contraceptive group, and this incremental change with treatment was significantly different from that observed with norgestrel oral contraceptive use. None of the observed changes resulted in values outside the reference range. The results for the HDL subclasses are shown in Table III. HDL3 was significantly increased in the norethindrone oral contraceptive group and this increase differed from the change in the norgestrel oral contraceptive group (p = 0.058). Nonsignificant changes of HDLzb were produced by the norethindrone oral contraceptive (increasing, p = 0.082) and the norgestrel oral contraceptive (decreasing, p = 0.076). The incremental change with treatment was significantly different between these two formulations (p = 0.016). HDL,, levels were not altered by any of the preparations. None of the changes was outside the reference range.
450
Krauss et al.
Several ratios were examined where the numerator was LDL, IDL, or LDL + IDL, and the denominator was HDL, HDL,, + 2b, or HDLI (Table IV). HDL,, and HDLzb were combined because the values for HDLzb were small and ratios calculated with this parameter were potentially not reliable. The (LDL + 1DL)IHDL and LDL/HDL ratios were significantly increased from baseline with treatment in the norgestrel oral contraceptive group while IDL/HDL was unaffected. When HDLza + 2b was substituted in the denominator, the results paralleled those for total HDL, while none of the ratios with HDL, as the denominator was changed significantly with treatment (not shown).
Comment In this report we have sought to determine whether changes in serum lipids and lipoproteins are associated with use of two low-dose estrogen-progestin oral contraceptives. The results indicate that changes can occur with use of such agents for a 2-month period, and, as has been shown with higher-dose oral contraceptives,29 5 the results depend on the specific formulation used. With use of the norethindrone oral contraceptive, there were significant increases in total and HDL-C as well as increases in LDL and HDL mass. The increase in LDL, while significantly greater than in control or norgestrel oral contraceptive groups, must be interpreted with caution since the baseline level in the norethindrone oral contraceptive group was significantly lower than in the other groups. The relative increases in HDL-C and total HDL mass were similar (11% and 13%, respectively) and resulted in no changes in ratios of cholesterol/HDL-C or LDL/HDL mass. In contrast to the changes observed with use of the norethindrone oral contraceptive, there were no significant changes in serum total cholesterol, HDL-C, and total LDL and HDL mass with use of the norgestrel oral contraceptive. Despite the absence of significant changes in these individual parameters, there were small but significant increases in ratios of cholesterol/ HDL-C and LDL/HDL mass in the norgestrel oral contraceptive group. Increases in serum triglyceride and VLDL were observed with the norgestrel oral contraceptive but not with the norethindrone oral contraceptive. Interpretation of these data is made difficult by the well-known wide variation in serum VLDL and triglyceride levels in free-living subjects. The median VLDL levels increased more with the norethindrone oral contraceptive than with the norgestrel oral contraceptive, but there was a wide range of changes in the norethindrone oral contraceptive group and the differences were not significant at p < 0.05. In a recent
February Am. J. Obstet.
15, 1983 Gynecoi.
cross-sectional study,” use of combination contraceptive agents containing low doses of d-norgestrel(O.15 to 0.25 mg) was found to be associated with lower levels of total and HDL, cholesterol than were found in control subjects, with smaller and variable differences in HDLS cholesterol, while use of agents containing 1 mg of norethindrone or norethindrone acetate were not associated with changes in either HDL-C fraction. In another report, however, 35 pg of ethinyl estradiol/l mg of norethindrone was found to lower HDL-C and raise serum triglyceride between three and six cycles of use.l* A smaller increase in triglyceride and no change in HDL-C were seen after four cycles of use of an agent containing a smaller dose (0.5 mg) of norethindrone.18 Other investigators have reported a dose-related decrease in HDL-C within 1 month of treatment with ethinyl estradiol-levonorgestrel combinations with significant increases in plasma triglyceride levels at later time points. I9 When these results are compared with those in the present report, it seems likely that the dose of norethindrone used here (0.4 mg) resulted in relative dominance of an estrogen-mediated increase in HDL-C, while the dose of dl-norgestrel was sufficient to reverse this effect but was not sufficient to overcome estrogen-related increases in triglyceride and VLDL.” Differences in metabolic effects of the two oral contraceptive preparations studied here are also revealed by examination of changes in HDL subspecies. The increased HDL mass in norethindrone oral contraceptive users was accounted for primarily by an increase in HDLB mass, with a smaller and statistically insignificant increase in HDLzb and a minimal increase in HDL,,. The changes in both HDL3 and HDLzb differed significantly from those seen with the use of the norgestrel oral contraceptive, and in the case of HDLzb the change was in the opposite direction. These changes in HDL distribution as assessed by analytic ultracentrifugation may reflect alterations in structure and composition of HDL subclasses. Previously it has been reported that HDL triglyceride content is increasedZo and the ratio of apolipoprotein I/ apolipoprotein II is decreasedzl in users of estrogenprogestin combinations. Hormone-induced alterations in structure or composition of the HDL subclasses may also influence the properties of the subclasses as measured by analytic ultracentrifugation. Thus, it cannot be assumed that the procedure used to resolve three components within HDL in untreated subjectsI will yield measurements of these same components in plasma obtained after oral contraceptive treatment. However, irrespective of whether the physical characteristics of the HDL components change after oral contraceptive use, analytic ultracentrifugation provides a
Volume 145 Number
Effects of low-dose oral contraceptives
I
means of assessing changes in the distribution of HDL subspecies with treatment. The norethindrone oral contraceptive-induced increase of HDL in the HDL3 component is consistent with a previous report of higher levels of HDL, in women using oral contraceptives compared with nonuser control subjects. 22 Although a variety of oral contraceptive preparations were used by the women in that study, the majority were norethindrone-estrogen combinations. The norethindrone oral contraceptive-induced release in HDL, is to be contrasted with the findings in postmenopausal estrogen users, in whom increased levels of HDL are confined to the fasterHoating, more buoyant, and lipid-enriched HDLz species, primarily HDL,,.‘“, 24 Thus, norethindrone (but not norgestrel) appears to modify the effect of estrogen on HDL by shifting the increase in HDL mass to more dense particles with slower flotation rates. While the oral contraceptive-related changes in HDLzb did not reach statistical significance, the findings may be indicative of additional metabolic effects of oral contraceptive use. Norgestrel has been reported to lower levels of HDL, cholesterol in conjunction with stimulation of the activity of heparin-releasable hepatic lipase,25 an enzyme which appears to be involved in the catabolism of HDL, lipids. 26. Z7 While other metabolic factors may also be operative,** it is possible that the relative reduction in HDLzh with norgestrel oral contraceptive versus norethindrone oral contraceptive use may involve differences in oral contraceptive effects on hepatic lipase activity. Marked reductions in HDL2b as well as HDL,, have also been shown with use of a norgestrel-containing contraceptive vaginal ring,2s and it is likely that the mechanisms involved are similar to those operative with norgestrel oral contraceptive use. There is as yet insufficient information available to allow assessment of the possible clinical significance of the lipid and lipoprotein changes observed with oral
contraceptive use. For instance, if risk of coronary disease is increased by higher levels of total cholesterol and LDL, as was found here with norethindrone oral contraceptive use, it cannot be concluded from existing data that a proportional increase in HDL is sufficient to offset such a risk. The analysis is further clouded by the absence of information regarding the impact on coronary risk of an increase in the HDL3 component, since existing data suggest that HDL-C, the measure most commonly studied as a “negative risk factor,” is correlated with levels of HDL, but not HDL3?’ Also, it has been reported recently that cholesterol in HDL, was significantly reduced in patients with severe coronarv disease, while levels of HDL3 cholesterol were not rklated to extent of coronary disease.“’ Another consideration in assessing the present results is that effects associated with long-term use of oral contraceptives may differ from those observed with relatively short duration of use, and it is the long-term etl’ects that are most likely to influence the development of atherosclerosis. Finally, the quantitative significance of the increases in total cholesterol/HDL-C: and LDLIHDL ratios observed here with norgestrel oral ccmtraceptive use remains open to question. None of the resultant values were outside the reference ranges. and none were sufficiently high to result in a higher than average risk for coronary disease on the basis of taxisting epidemiologic data.7 Such changes might, howe\.er, compound the effects of other risk factors, such as cigaret.te smoking, and might be indicative of more substantial responses to oral contraceptive use in patients with underlying disorders of lipid metabolism. The contribution of Dr. Frank Lindgren and staff to the analytic ultracentrifuge measurements is greatly appreciated. Lipid analyses were performed with the expert assistance of Dennis Duncan and Janet Selmek. We also thank Linda Stimson and Linda Abe for preparation of the manuscript.
REFERENCES 1. Beck, P.: Contraceptive steroids: Modifications of carbohydrate and lipid metabolism, Metabolism 42:841, 1973. 2. Krauss, R. M.: Effects of progestational agents on serum lipids and lipoproteins, J. Reprod. Med. f7:503, 1982. 3. Mishell, D. R., Jr., Moore, D. F., Roy, S., Brenner, P. F., and Page, M. A.: Clinical performance and endocrine profiles with contraceptive vaginal rings containing a combination of estradiol and d-norgestrel, AM. J. OBSTET. GYNECOL. 130:55, 1978. 4. Roy, S., Mishel!. D. R., Jr., Gray, G., Donzano-Takano, R., Brenner, P. F., Eide, I., DeQuattro, V., and Shaw, S. T.: Comparison of metabolic and clinical effects of four oral contraceptive formulations and a contraceptive-vaginal ring, AM. J. OBSTET. GYNECOL. 136:920, 1980. 5. Bradley, D. D., Wingerd, J., Petitti. D. B., Krauss, R. M.,
451
6.
7.
8.
9.
and Ramcharan, S.: Serum high density lipoprotein cholesterol in women using oral contraceptives, estrogens and progestins, N. Engl. J. Med. 499: 17. 1978. Miller, G. J.. and Miller, N. E.: Plasma high density lipoprotein concentration and development nt’ ischaemic heart disease, Lancet 1:16, 1975. Castelli, W. B., Dorfe, J, T., Gordon, ‘I-., et al.: HDLcholesterol and other lipids in coronary heart disease: The cooperative lipoprotein phenotyping study, Circulation 55:767, 1977. Gordon, T., Caste&, W. P., Hjortland, M. C.. et al.: High density lipoproteins as a protective factor against coronary heart disease, Am. J. Med. 64:707, 1977. Allain, C. C., Poon, L. S., Chan, C. S. G.. Richmond, W., and Fu, P. C.: Enzymatic determination of rotal serum cholesterol, Clin. Chem. 20:470, 1974.
452
February Am. J. Obstet.
Krauss et al.
10. Bondar, R. J. L., Balwyot, E. S., and Hsu, M.: A totally enzymatic method for the determination of serum triglycerides, Clin. Chem. 23: 1132, 1977. (Abst.) 11. Lopes-Virella, M. F., Stone, P., Ellis, S., and Colwell,.I. A.: Cholesterol determination in high density lipoproteins senarated bv three different methods. Clin. Chem. 23’:882, 1977. 12. Friedwald, W. T., Levy, R. I., and Frederickson, D. S.: Estimation of concentration of low-density lipoprotein cholesterol without use of the preparative ultracentrifuge, Clin. Chem. l&499, 1972. 13. Ewing, A. M., Freeman, N. K., and Lindgren, F. T.: In Paoletti, R., and Kritchevsky, D., editors: Advances in Lipid Research, New York, 1965, Vol. 3, Academic Press, Inc., pp. 25-61. 14. Anderson, D. W., Nichols, A. V., Pan, S. S., and Lindgren, F. T.: High density lipo rotein distribution. Resolution and determination of tK ree major components in a normal population sample, Atherosclerosis 29:161, 1978. 15. Lindgren, F. T., Adamson, G. L., Jensen, L. C., and Wood, P. D.: Lipid and lipoprotein measurements in a normal adult American population, Lipids lOr750, 1975. 16. Roy, S., Krauss, R. M., k&shell, D. R., jr., Casagrande, J., and Pike. M. C.: The Metabolic Effects of a LowEstrogeniiow-Progestogen Oral Contraceptive, New York, 1981, Biomedical Information Corporation. 17. Wynn, V., and Niththyananthan, R.: The effect of progestins in combined oral contraceptives on serum lipids with special reference to high-density lipoproteins, AM. J. OBSTET.GYNECOL. 142:766, 1982. 18. Briggs, M. H., and Briggs, M.: A randomized study of metabolic effects of four low estrogen oral contraceptives. I. Results after six cycles, Contraception 23:463, 1981. 19. Larsson-Cohn, V., fihraeus, L., Wallentin, L., and Zador, G.: Lipoprotein changes may be minimized by proper comnosition of a combined oral contracentive. Fertil. 1 Sterii. 35:172, 1981. 20. Knopp, R. H., Walden, C. E., Wahl, P. W., Hoover, J. J., Warnick, G. R., Albers, J. J., Ogilvie, J. T., and Hazzard, W. R.: Oral contraceptive and postmenopausal estrogen effects on lipoprotein triglycerides and cholesterol in an adult female population: Relationships to estrogen and
21. 22.
23. 24.
25.
26.
27.
28. 29.
30. 31.
15, 1983 Gynecol.
progestin potency, J. Clin. Endocrinol. Metab. 53: 1123, 1981. Cheung, M. C., and Albers, J. J.: The measurement of apolipoprotein A-I and A-II levels in men and women by immunoassay, J. Clin. Invest. 60:43, 1977. Krauss, R. M., Lindgren, F. T., Silvers, A., Jutagir, R., and Bradley, D. D.: Changes in serum high density lipoproteins in women on oral contraceptive drugs, Clin. Chim. Acta 80:465, 1977. Krauss, R. M., Lindgren, F. T., Wingerd, J., Bradley, D. D., and Ramcharan, J.: Effects of estrogens and progestins on high density lipoproteins, Lipids 14: 113,1979. Krauss, R. M., Takaki, N., and Lewis, S. B.: Relation of serum high density lipoproteins (HDL) to gonadotropin and estrogen levels in postmenopausal estrogen users and non-users, in Fifth International Symposium on Arteriosclerosis, Houston, Texas, November, 1979. Tikkinen, M. J., Nikkill, E. A., Kuusi, T., and Sipinen, S.: Different effects of two progestins on plasma high density lipoprotein (HDL,) and postheparin plasma lipase activity, Atherosclerosis 40:365, 1981. Kuusi, T., Kinnunen, P. K. J., and Nikkila, E. A.: Hepatic endothelial lipase antiserum influences rat plasma low and high density lipoproteins in vivo, FEBS Lett. 104: 384, 1979. Kuusi, T., Saarinen, P., and Nikkili, E. A.: Evidence for the role of hepatic endothelial lipase in the metabolism of plasma high density lipoprotein, in man, Atherosclerosis 36:589, 1980. Krauss, R. M.: Regulation of high density lipoprotein levels, Med. Clin. North Am. 66:403. 1982. Roy, S., Krauss, R. M., Mishell, D. R.; Jr., Casagrande, J., and Pike, M. C.: The effects on lipids and lipoproteins of a contraceptive vaginal ring containing levonorgestrel and estradiol, Contraception 24429, 1981. Anderson, D. W.: HDL cholesterol: the variable components, Lancet 1:819, 1978. Miller, N. E., Hammett, F., Saltissi, S., Rao, S., Van Zeller, H., Coltant, J., and Lewis, B.: Relation of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins, Br. Med. J. 282: 1741, 1981.