203
Atherosclerosis, 84 (1990) 203-211 Elsevier Scientific Publishers Ireland, Ltd.
ATHERO
04539
Comparison of the lipoprotein and hemostatic changes after a triphasic and a monophasic low dose oral contraceptive in premenopausal middle-aged women Cesare R. Sirtori I, Laura Calabresi ‘, Guido Franceschini ‘, Gemma Gianfranceschi ‘, Francesco Zoppi 2, Sergio Winkler 3, Pasquale Bilotta 4 and Alfonso Zampetti 3 ’ Center E. Grossi Paoletti, Institute of Pharmacological
Sciences, University of Milan, Milan (Italy), ’ Clinical Laboratory 4 Medical Division Akzo, Rome (Italy)
and ’ Division of Obstetrics and Gynaecology, Niguarda Hospital, Milan (Italy), (Received 7 February, 1990) (Revised, received 15 June, 1990) (Accepted 18 June, 1990)
Summary
Metabolic and hemostatic effects of 2 low dose oral contraceptives (OCs), a triphasic (ethinylestradiol + ( - )-norgestrel) and a monophasic (ethinylestradiol + desogestrel) preparation, were compared in a cross-over trial in fertile women over 35 years of age. Both combinations moderately affected plasma lipids, with 17-248 increases of total triglyceridemia. Triglycerides accumulate in low density lipoproteins, thus suggesting the possible formation of an atherogenic lipoprotein particle. Only the monophasic preparation increased high density lipoprotein (HDL)-cholesterol levels significantly, with a rise in HDL, mass and cholesterol. OC treatment led to slight changes in HDL, and HDL, structure, with a rise of the cholesteryl ester and triglyceride contents, indicative of a stimulated cholesterol esterification and reverse transport. Changes in the hemostatic indexes (fibrinogen, antithrombin III and protein C) were negligible. The new low dose OCs, even when prescribed to relatively older women, affect to a relatively small extent lipid/lipoprotein metabolism, with the exception of changes in the low density lipoprotein composition.
Key words: Oral contraceptives; Desogestrel; (-)-Norgestrel; density lipoproteins; Thrombosis
Correspondence to: Prof. Cesare R. Sirtoti, Chair of Clinical Pharmacology, Center E. Grossi Paoletti, Via Balzaretti, 9, 20133 Milano, Italy. Tel.: (business) 02/294X672; (home) 02/76.001.681. 0021-9150/90/%03.50
High density lipoprotein subfractions;
Low
Introduction The use of combination oral contraceptives (OCs) is still rated as a major risk factor for
@ 1990 Elsevier Scientific Publishers Ireland, Ltd.
cardiovascular (CV) disease, including deep vein thrombosis and thromboembolism, as well as ischaemic heart disease and myocardial infarction [l]. Such conclusions are mostly related to the results of US based studies, evaluating very large series of exposed women over many years [2]. These studies have generally considered subjects taking high-dose, progestogen-dominant OCs, inducing a potential atherogenic lipoprotein profile, i.e. with an elevated low density lipoprotein (LDL)/high density lipoprotein (HDL) ratio [3], probably attributable to the androgenic activity of the progestogens. The hypothesis of an excess cardiovascular mortality in current and past users of OCs has been recently questioned [4] and, possibly, only limited to women over 35 who smoke [5,6]. Even such conclusions are, however, open to question, since users of OCs are mostly of a young age and the exposure to OCs may be limited in time. Moreover, the majority of metabolic studies addressing the issue of atherogenic changes in plasma lipoproteins during OC treatment, have been of short duration and, at times, only historical [5]. The recent suggestion that any adverse change in the HDL system is mediated by the progestogen component of hormonal replacement treatments [7], is only based on retrospective data. The present study evaluated two of the newer low dose OC preparations, one monophasic containing desogestrel, and the other triphasic with ethinylestradiol and levonorgestrel. The selected volunteers were all over 35 years of age, but none presented with clear pre-menopausal symptoms (with the possible exception of subjective complaints, headache, occasional menstrual irregularities). Although detailed lipid-lipoprotein studies in this middle-aged group are not currently available, the premenopausal age is known to be characterized by a variety of minor but significant changes in the pituitary-ovary axis [8]. In addition to the standard lipid-lipoprotein parameters, the subfraction pattern of HDL was also examined, because of the known correlation between specific HDL distributions and a change in risk for CV disease [9]. The reported data show interesting differences between the two OC preparations, particularly as relating to the HDL total mass and subfraction distribution, thus suggesting that long
term use of these drugs in premenopausal women may be associated with different metabolic risk profiles. Materials and methods
Subjects Fourteen healthy premenopausal women, aged between 38 and 43 years (40.2 f 1.7 yrs), were selected among those attending the Ob-Gyn Department of the Niguarda Hospital for OC treatment. None was overweight (* 10% according to the Metropolitan Height and Weight Tables), was engaged in regular exercise programs or had a clinical history of alcohol abuse; 4 smoked regularly (5-20 cigarettes/day). Five women had used OCs before, but discontinued them at least 3 months prior to the start of the present study. All the selected subjects continued their customary way of life, including dietary habits. A Papanicolaou smear was carried out in all before selection, ovulation being confirmed by the determination of progesterone (> 3 ng/ml). Study protocol The study was carried out according to a cross-over protocol. A baseline blood sample was collected between the 10th and 12th day of the menstrual cycle, prior to the start of the study. The selected women were then randomized into 2 groups, 7 in each. One group received a monophasic (M) combination (desogestrel 150 pg and ethinylestradiol 20 pg) for 6 months and then switched to a triphasic (T) treatment (levonorgestrel 50 pg and ethinylestradiol 30 pg for 6 days; levonorgestrel 75 pg and ethinylestradiol 40 pg for 5 days; levonorgestrel 125 pg and ethinylestradiol 30 pg for 10 days) for another 6 months. The second group followed the opposite sequence. The treatments started on the first day of each cycle, defined as the first day of menstrual flow. The subjects were seen every 3 menstrual cycles, between the 10th and 12th day; a gynecological and clinical examination was carried out, including weight and blood pressure determinations, and side effects were recorded. Blood samples were collected at each visit for serum lipid, glucose and uric acid determinations. A complete lipoprotein separation, including HDL subfractions, and the
205
determination of serum apolipoprotein levels were carried out before and after each treatment period. At the same visits, plasma antithrombin III, fibrinogen and C protein levels were measured. The research was carried out according to the Helsinki Declaration, after informed consent of the subjects and approval by the Institutional Review Committee. Laboratory
metho&
Blood samples were drawn after an overnight fast, and serum was prepared by low speed centrifugation. A blood aliquot was taken on 3.8% sodium citrate and plasma isolated for the determination of the hemostatic indexes. Lipoproteins were separated by ultracentrifugation, according to the NIH protocol [lo]; HDL were obtained by selective precipitation with dextran sulfate-MgC12 [ll]. The HDL subfractions were isolated by rate zonal ultracentrifugation using a Beckman SW 41Ti rotor, automatically monitoring the volume of elution (Ve) of each subfraction from the ultracentrifugal gradient [12]. The HDL particles were also separated according to size, by nondenaturing polyacrylamide gradient gel electrophoresis (GGE) using a 4-30% gradient, as previously described [ 131. Total cholesterol (TC) and triglycerides (TG) in sera and isolated lipoproteins were determined by enzymic methods [14,15]. Free cholesterol (FC) [14], phospholipids (PL) [16] and proteins (P) [17] were measured in the isolated HDL subfractions after dialysis against 0.15 M NaCl, 0.02% EDTA, 0.01% NaN,, pH 7.4. Cholesteryl ester (CE) concentration was calculated as (TC-FC) X 1.68, and the total lipoprotein mass as the sum of the different components. Serum apolipoproteins AI, AI1 and B were measured by immunoturbidimetry [18]. Hemostatic indexes were tested in the samples of titrated plasma. Fibrinogen was determined according to [19] and antithrombin III by immunodiffusion [20], using NOR-Partigen plates (Behringwerke, Marburg, F.R.G.). Protein C activity was evaluated by an APTT test following activation by a fraction of the snake venom Agkistrodon contortrix [21], and expressed as percentage of a normal plasma pool. Results are expressed as mean k SD values, if not otherwise stated. Paired comparisons for dif-
ferent groups were carried out by Student’s t-test [22]; logarithmic transformations of serum TG, VLDL-TC and VLDL-TG values were examined in the statistical analyses in order to normalize the skewed distribution usually observed for these parameters. Since it was not possible to stop contraceptive treatment for a length of time sufficient to induce a return to baseline of the laboratory values, the possibly disturbing effect of a carry over was excluded by adjusting the values according to the method described by Wallenstein and Fisher [23], prior to statistical analyses. Results
Ten of the 14 selected subjects completed the study, 5 in each sequence group. Four dropped out: 2 because of inadequate compliance to the treatment schedule, 1 for recurrent nausea, and 1 because of death by accident. No major differences in the serum lipid changes were observed in the 2 groups following the opposite treatment sequences; data from the 10 subjects, who completed the study, were thus pooled and analysed as a whole group. Minimal changes were recorded for serum cholesterol levels, increasing by 3 and 5% after 3 and 6 months of M treatment and by 5 and 3% after T (Fig. 1, Table 1). Serum TG levels were increased to statistically significant levels with both preparations: by 24 and 20% after 3 and 6 months of M and by 18 and 17% after T. A major difference between the 2 treatments was observed in the HDL-cholesterol levels, increasing by 12 and 6% (P < 0.01) after 3 and 6 months of M and being stable during T treatment (Fig. 1). Minor lipoprotein changes were observed after 6 months of OC administration (Table 1). LDL-cholesterol levels increased by 6 and 2% during M and T treatment, respectively; VLDL-cholesterol decreased by 5% during M and increased by 23% after T. The increase in serum TG levels was essentially due to a rise of LDL-TG (+ 50.6% after M and + 42.5% after T), leading to a selective enrichment of LDL with TG: the LDL-TC/LDL-TG ratio decreased from 6.37 ~fi2.42 to 4.01 f 2.05 after M and to 3.83 k 2.10 after T (both P < 0.05). No major changes were noted in the composition of VLDL and HDL. Serum apolipoprotein B levels
206 Wdl
TABLE 2
230
SERUM APOLIPOPROTEIN LEVELS IN 10 HEALTHY PREMENOPAUSAL WOMEN TAKING A MONOPHASIC OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS
t
Results are expressed as mean f SEM (mg/dl). Baseline
Monophasic
Triphasic
Apo AI Apo AI1
158.8 It 10.0 43.6* 3.4
Apo B
105.1*
177.9 k 6.4 49.3k2.9 * 112.8 f 8.0
172.9 f 10.3 48.2zt 3.2 * 116.4k 9.6
*
L
1
I
0 MONTHS
I
3 OF TREATMENT
6
Fig. 1. Changes in total cholesterol, triglycerides and HDLcholesterol levels in 10 healthy premenopausal women taking a monophasic (0) or a triphasic (0) oral contraceptive combination for 6 months (* P -z0.05; * * P < 0.01).
increased slightly with both OCs (+ 7.3% after M and +10.7% after T; Table 2), the LDLcholesterol/ape B ratio decreasing from 1.24 + 0.21 to 1.22 f 0.16 (M; P > 0.05) and to 1.15 +
7.2
P < 0.05 vs. baseline.
0.24 (T; P < 0.05). The apo AI and AI1 concentrations rose by 12.1 and by 13.3% with M and by 8.9 and 10.8% with T. A comparison between the effects of the 2 contraceptive combinations on serum lipid, lipoprotein and apolipoprotein levels did not show major differences, except for the increase in HDL-C levels, evident only after M treatment (Fig. 1 and Table 1). The HDL subfraction distribution and composition were analyzed after separation by rate zonal uhracentrifugation, identifying 2 subfractions, i.e. HDL, (d= 1.063-1.125 g/ml) and HDL, (d = 1.125-1.210 g/ml) [12], at the start of the study and after 6 months of each treatment. Total HDL, mass increased by 4% after M and decreased by 10% after T (Table 3); the HDL, associated cholesterol was reduced, more significantly after T (-16.4% vs. -5.5%). Both OC combinations increased HDL, mass and associ-
TABLE 1 SERUM LIPID AND LIPOPROTEIN LEVELS IN 10 HEALTHY PREMENOPAUSAL OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS
WOMEN TAKING
A MONOPHASIC
Results are expressed as mean f SEM (mg/dl).
Total cholesterol VLDL-cholesterol LDL-cholesterol HDL-cholesterol Triglycerides VLDL-triglycerides LDL-triglycerides HDL-triglycerides *
Baseline
Monophasic
Triphasic
205.6 f 10.9 13.6* 3.4 128.6k 9.7 63.2& 5.0
216.7 f 10.0 12.9k 2.1 136.4rt 10.4 67.2* 4.2 *
211.4k11.8 16.7& 3.2 131.3 f 12.4 63.2& 6.3 a
82.2 + 16.6 40.3 f 12.3 24.9* 3.2 17.0* 2.2
P < 0.05 vs. baseline; o P < 0.05 vs. monophasic.
97.9 f 40.2k 37.6f 18.1zt
13.9 * 7.8 5.0 * 2.1
%.4* 14.6 * 42.7* 9.3 35.6* 4.9 * 18.1k 1.8
207
I
HDL Fig. 2. Mean rate zonal ultracentrifugation profiles of HDL subfractions from 10 healthy premenopausal women before (. .) and after 6 months of treatment with a monophasic and a triphasic (- - -) oral contraceptive combi( -) nation.
ated cholesterol, the effect being more evident after M (+18.9 and + 12.3% respectively, vs. + 14.0 and +8.8%). Due to these modifications the HDL,/HDL, mass and cholesterol ratios decreased from 0.42 + 0.17 and 0.60 + 0.23 to 0.37 of:0.16 and 0.52 f 0.24 (M; both P > 0.05) and to 0.32 + 0.15 and 0.46 kO.22 (T; both P < 0.05), respectively. The elution profiles from the rate zonal ultracentrifugal gradient clearly showed an increase of material eluting in the HDL, density range after both treatments (Fig. 2); the flotation rate of HDL, was significantly increased after M (Table 3) and did not change after T. The HDL particle size distribution was analyzed by non-denaturing GGE. By this technique, 5 HDL sub-
TABLE
2b
2a 3a 3b3c
Fig. 3. Mean polyacrylamide gradient gel electrophoretic profiles of HDL subfractions from 10 healthy premenopausal women before (. . .) and after 6 months of treatment with a monophasic () and a triphasic (- - -) oral contraceptive combination.
classes have been identified [24], 3 in the HDL, density range, namely HDL,,, HDL,, and HDL,, (with decreasing size), and 2 in the HDL, interval, HDL,, and HDL,,. In all the examined subjects, HDL,, predominated, both before and after treatment (Fig. 3). The HDL,, content was much more variable at baseline, and was not modified by either treatment (Fig. 3). The mean size of HDL,, particles increased slightly after both M and T, with no changes in HDL, diameter (Table 3). Both treatments markedly affected the composition of HDL, and HDL, (Table 4). The FC content was lowered and CE increased; the FC/CE ratio decreased from 0.32 + 0.13 to 0.20 f 0.03 (M) and to 0.18 + 0.04 (T) (both P < 0.05)
3
SERUM LEVELS, WOMEN TAKING Results are expressed
ELUTION VOLUME AND SIZE OF HDL SUBFRACTIONS A MONOPHASIC OR A TRIPHASIC ORAL COMBINATION
IN 10 HEALTHY FOR 6 MONTHS
PREMENOPAUSAL
as mean f SEM.
HDL, Total mass (mg/dl) Associated cholesterol Elution volume (ml) Size (run)
(mg/dl)
HDL, Total mass (mg/dl) Associated cholesterol Elution volume (ml) size (rim)
(mg/dl)
Baseline
Monophasic
Triphasic
124.2 f 17.1 23.7 f 3.1 7.98&- 0.08 10.19 f 0.02
129.0 + 22.4 f 7.98* 10.42+
112.0 f 20.5 19.8 f 3.5 * 7.97* 0.05 10.34* 0.04 *
303.8 k 22.2 39.9 f 2.7 4.73* 0.10 8.61 f 0.06
357.9 k27.1 * + 44.8 f 2.8 *** 4.91 f 0.05 * * 8.64f 0.03
* P < 0.05; * * P < 0.005; * * * P -z0.001 vs.baseline.
15.4 2.7 0.08 0.05 *
349.7 f 32.7 * 43.4 f 3.7 * 4.80+ 0.09 8.61 f 0.03
208 TABLE 4 PERCENTAGE COMPOSITION OF HDL SUBFRACTIONS IN 10 HEALTHY PREMENOPAUSAL WOMEN TAKING A MONOPHASIC OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS Results are expressed as mean f SD.
HDL, FC% CE% TGI PL% P%
Baseline
Monophasic
Triphasic
6.3f1.4 20.9 + 3.0 4.0 f 1.8 30.5 k 1.6 38.3k2.0
4.2kO.3 ** 21.9k2.6 5.0 f 1.6 28.3 k 1.5 * 40.6 i 2.9
4.1*0.8 ** 22.1 f 2.3 5.0*2.5 26.2* 2.5 * * * 0 42.1 f 2.6 * * *
3.8f0.5 15.8k1.8 1.5f0.8 25.3 k 1.9 53.5 f 2.4
2.3f0.2 *** 17.4* 1.1 * 2.8kO.5 * 24.4& 1.5 53.6k 1.9
2.3kO.3 *** 17.0+1.3 3.1f1.4 *** 23.Okl.4 **O 54.4k2.2
Discussion
HDL, FCI CE% TG% PLZ P%
and to 0.058 * 0.012 (T; P < 0.005 vs baseline and P < 0.05 vs M) in HDL,. No significant changes in the hemostatic indexes (Table 5) occurred with either treatment, only fibrinogen levels being slightly reduced after M. Glucose, uric acid and sitting blood pressure were also not modified. A slight increase of body weight was found with both preparations, from 56.9 f 6.34 to 58.8 &-6.8 kg after M and to 58.2 + 7.2 after T (both P -c0.05). Occasional nausea and slight edema at the upper extremities was reported by 4 participants.
* P < 0.05; ** P < 0.005; ***P
Atherosclerosis, 84 (1990) 203-211 Elsevier Scientific Publishers Ireland, Ltd.
ATHERO
04539
Comparison of the lipoprotein and hemostatic changes after a triphasic and a monophasic low dose oral contraceptive in premenopausal middle-aged women Cesare R. Sirtori I, Laura Calabresi ‘, Guido Franceschini ‘, Gemma Gianfranceschi ‘, Francesco Zoppi 2, Sergio Winkler 3, Pasquale Bilotta 4 and Alfonso Zampetti 3 ’ Center E. Grossi Paoletti, Institute of Pharmacological
Sciences, University of Milan, Milan (Italy), ’ Clinical Laboratory 4 Medical Division Akzo, Rome (Italy)
and ’ Division of Obstetrics and Gynaecology, Niguarda Hospital, Milan (Italy), (Received 7 February, 1990) (Revised, received 15 June, 1990) (Accepted 18 June, 1990)
Summary
Metabolic and hemostatic effects of 2 low dose oral contraceptives (OCs), a triphasic (ethinylestradiol + ( - )-norgestrel) and a monophasic (ethinylestradiol + desogestrel) preparation, were compared in a cross-over trial in fertile women over 35 years of age. Both combinations moderately affected plasma lipids, with 17-248 increases of total triglyceridemia. Triglycerides accumulate in low density lipoproteins, thus suggesting the possible formation of an atherogenic lipoprotein particle. Only the monophasic preparation increased high density lipoprotein (HDL)-cholesterol levels significantly, with a rise in HDL, mass and cholesterol. OC treatment led to slight changes in HDL, and HDL, structure, with a rise of the cholesteryl ester and triglyceride contents, indicative of a stimulated cholesterol esterification and reverse transport. Changes in the hemostatic indexes (fibrinogen, antithrombin III and protein C) were negligible. The new low dose OCs, even when prescribed to relatively older women, affect to a relatively small extent lipid/lipoprotein metabolism, with the exception of changes in the low density lipoprotein composition.
Key words: Oral contraceptives; Desogestrel; (-)-Norgestrel; density lipoproteins; Thrombosis
Correspondence to: Prof. Cesare R. Sirtoti, Chair of Clinical Pharmacology, Center E. Grossi Paoletti, Via Balzaretti, 9, 20133 Milano, Italy. Tel.: (business) 02/294X672; (home) 02/76.001.681. 0021-9150/90/%03.50
High density lipoprotein subfractions;
Low
Introduction The use of combination oral contraceptives (OCs) is still rated as a major risk factor for
@ 1990 Elsevier Scientific Publishers Ireland, Ltd.
cardiovascular (CV) disease, including deep vein thrombosis and thromboembolism, as well as ischaemic heart disease and myocardial infarction [l]. Such conclusions are mostly related to the results of US based studies, evaluating very large series of exposed women over many years [2]. These studies have generally considered subjects taking high-dose, progestogen-dominant OCs, inducing a potential atherogenic lipoprotein profile, i.e. with an elevated low density lipoprotein (LDL)/high density lipoprotein (HDL) ratio [3], probably attributable to the androgenic activity of the progestogens. The hypothesis of an excess cardiovascular mortality in current and past users of OCs has been recently questioned [4] and, possibly, only limited to women over 35 who smoke [5,6]. Even such conclusions are, however, open to question, since users of OCs are mostly of a young age and the exposure to OCs may be limited in time. Moreover, the majority of metabolic studies addressing the issue of atherogenic changes in plasma lipoproteins during OC treatment, have been of short duration and, at times, only historical [5]. The recent suggestion that any adverse change in the HDL system is mediated by the progestogen component of hormonal replacement treatments [7], is only based on retrospective data. The present study evaluated two of the newer low dose OC preparations, one monophasic containing desogestrel, and the other triphasic with ethinylestradiol and levonorgestrel. The selected volunteers were all over 35 years of age, but none presented with clear pre-menopausal symptoms (with the possible exception of subjective complaints, headache, occasional menstrual irregularities). Although detailed lipid-lipoprotein studies in this middle-aged group are not currently available, the premenopausal age is known to be characterized by a variety of minor but significant changes in the pituitary-ovary axis [8]. In addition to the standard lipid-lipoprotein parameters, the subfraction pattern of HDL was also examined, because of the known correlation between specific HDL distributions and a change in risk for CV disease [9]. The reported data show interesting differences between the two OC preparations, particularly as relating to the HDL total mass and subfraction distribution, thus suggesting that long
term use of these drugs in premenopausal women may be associated with different metabolic risk profiles. Materials and methods
Subjects Fourteen healthy premenopausal women, aged between 38 and 43 years (40.2 f 1.7 yrs), were selected among those attending the Ob-Gyn Department of the Niguarda Hospital for OC treatment. None was overweight (* 10% according to the Metropolitan Height and Weight Tables), was engaged in regular exercise programs or had a clinical history of alcohol abuse; 4 smoked regularly (5-20 cigarettes/day). Five women had used OCs before, but discontinued them at least 3 months prior to the start of the present study. All the selected subjects continued their customary way of life, including dietary habits. A Papanicolaou smear was carried out in all before selection, ovulation being confirmed by the determination of progesterone (> 3 ng/ml). Study protocol The study was carried out according to a cross-over protocol. A baseline blood sample was collected between the 10th and 12th day of the menstrual cycle, prior to the start of the study. The selected women were then randomized into 2 groups, 7 in each. One group received a monophasic (M) combination (desogestrel 150 pg and ethinylestradiol 20 pg) for 6 months and then switched to a triphasic (T) treatment (levonorgestrel 50 pg and ethinylestradiol 30 pg for 6 days; levonorgestrel 75 pg and ethinylestradiol 40 pg for 5 days; levonorgestrel 125 pg and ethinylestradiol 30 pg for 10 days) for another 6 months. The second group followed the opposite sequence. The treatments started on the first day of each cycle, defined as the first day of menstrual flow. The subjects were seen every 3 menstrual cycles, between the 10th and 12th day; a gynecological and clinical examination was carried out, including weight and blood pressure determinations, and side effects were recorded. Blood samples were collected at each visit for serum lipid, glucose and uric acid determinations. A complete lipoprotein separation, including HDL subfractions, and the
205
determination of serum apolipoprotein levels were carried out before and after each treatment period. At the same visits, plasma antithrombin III, fibrinogen and C protein levels were measured. The research was carried out according to the Helsinki Declaration, after informed consent of the subjects and approval by the Institutional Review Committee. Laboratory
metho&
Blood samples were drawn after an overnight fast, and serum was prepared by low speed centrifugation. A blood aliquot was taken on 3.8% sodium citrate and plasma isolated for the determination of the hemostatic indexes. Lipoproteins were separated by ultracentrifugation, according to the NIH protocol [lo]; HDL were obtained by selective precipitation with dextran sulfate-MgC12 [ll]. The HDL subfractions were isolated by rate zonal ultracentrifugation using a Beckman SW 41Ti rotor, automatically monitoring the volume of elution (Ve) of each subfraction from the ultracentrifugal gradient [12]. The HDL particles were also separated according to size, by nondenaturing polyacrylamide gradient gel electrophoresis (GGE) using a 4-30% gradient, as previously described [ 131. Total cholesterol (TC) and triglycerides (TG) in sera and isolated lipoproteins were determined by enzymic methods [14,15]. Free cholesterol (FC) [14], phospholipids (PL) [16] and proteins (P) [17] were measured in the isolated HDL subfractions after dialysis against 0.15 M NaCl, 0.02% EDTA, 0.01% NaN,, pH 7.4. Cholesteryl ester (CE) concentration was calculated as (TC-FC) X 1.68, and the total lipoprotein mass as the sum of the different components. Serum apolipoproteins AI, AI1 and B were measured by immunoturbidimetry [18]. Hemostatic indexes were tested in the samples of titrated plasma. Fibrinogen was determined according to [19] and antithrombin III by immunodiffusion [20], using NOR-Partigen plates (Behringwerke, Marburg, F.R.G.). Protein C activity was evaluated by an APTT test following activation by a fraction of the snake venom Agkistrodon contortrix [21], and expressed as percentage of a normal plasma pool. Results are expressed as mean k SD values, if not otherwise stated. Paired comparisons for dif-
ferent groups were carried out by Student’s t-test [22]; logarithmic transformations of serum TG, VLDL-TC and VLDL-TG values were examined in the statistical analyses in order to normalize the skewed distribution usually observed for these parameters. Since it was not possible to stop contraceptive treatment for a length of time sufficient to induce a return to baseline of the laboratory values, the possibly disturbing effect of a carry over was excluded by adjusting the values according to the method described by Wallenstein and Fisher [23], prior to statistical analyses. Results
Ten of the 14 selected subjects completed the study, 5 in each sequence group. Four dropped out: 2 because of inadequate compliance to the treatment schedule, 1 for recurrent nausea, and 1 because of death by accident. No major differences in the serum lipid changes were observed in the 2 groups following the opposite treatment sequences; data from the 10 subjects, who completed the study, were thus pooled and analysed as a whole group. Minimal changes were recorded for serum cholesterol levels, increasing by 3 and 5% after 3 and 6 months of M treatment and by 5 and 3% after T (Fig. 1, Table 1). Serum TG levels were increased to statistically significant levels with both preparations: by 24 and 20% after 3 and 6 months of M and by 18 and 17% after T. A major difference between the 2 treatments was observed in the HDL-cholesterol levels, increasing by 12 and 6% (P < 0.01) after 3 and 6 months of M and being stable during T treatment (Fig. 1). Minor lipoprotein changes were observed after 6 months of OC administration (Table 1). LDL-cholesterol levels increased by 6 and 2% during M and T treatment, respectively; VLDL-cholesterol decreased by 5% during M and increased by 23% after T. The increase in serum TG levels was essentially due to a rise of LDL-TG (+ 50.6% after M and + 42.5% after T), leading to a selective enrichment of LDL with TG: the LDL-TC/LDL-TG ratio decreased from 6.37 ~fi2.42 to 4.01 f 2.05 after M and to 3.83 k 2.10 after T (both P < 0.05). No major changes were noted in the composition of VLDL and HDL. Serum apolipoprotein B levels
206 Wdl
TABLE 2
230
SERUM APOLIPOPROTEIN LEVELS IN 10 HEALTHY PREMENOPAUSAL WOMEN TAKING A MONOPHASIC OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS
t
Results are expressed as mean f SEM (mg/dl). Baseline
Monophasic
Triphasic
Apo AI Apo AI1
158.8 It 10.0 43.6* 3.4
Apo B
105.1*
177.9 k 6.4 49.3k2.9 * 112.8 f 8.0
172.9 f 10.3 48.2zt 3.2 * 116.4k 9.6
*
L
1
I
0 MONTHS
I
3 OF TREATMENT
6
Fig. 1. Changes in total cholesterol, triglycerides and HDLcholesterol levels in 10 healthy premenopausal women taking a monophasic (0) or a triphasic (0) oral contraceptive combination for 6 months (* P -z0.05; * * P < 0.01).
increased slightly with both OCs (+ 7.3% after M and +10.7% after T; Table 2), the LDLcholesterol/ape B ratio decreasing from 1.24 + 0.21 to 1.22 f 0.16 (M; P > 0.05) and to 1.15 +
7.2
P < 0.05 vs. baseline.
0.24 (T; P < 0.05). The apo AI and AI1 concentrations rose by 12.1 and by 13.3% with M and by 8.9 and 10.8% with T. A comparison between the effects of the 2 contraceptive combinations on serum lipid, lipoprotein and apolipoprotein levels did not show major differences, except for the increase in HDL-C levels, evident only after M treatment (Fig. 1 and Table 1). The HDL subfraction distribution and composition were analyzed after separation by rate zonal uhracentrifugation, identifying 2 subfractions, i.e. HDL, (d= 1.063-1.125 g/ml) and HDL, (d = 1.125-1.210 g/ml) [12], at the start of the study and after 6 months of each treatment. Total HDL, mass increased by 4% after M and decreased by 10% after T (Table 3); the HDL, associated cholesterol was reduced, more significantly after T (-16.4% vs. -5.5%). Both OC combinations increased HDL, mass and associ-
TABLE 1 SERUM LIPID AND LIPOPROTEIN LEVELS IN 10 HEALTHY PREMENOPAUSAL OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS
WOMEN TAKING
A MONOPHASIC
Results are expressed as mean f SEM (mg/dl).
Total cholesterol VLDL-cholesterol LDL-cholesterol HDL-cholesterol Triglycerides VLDL-triglycerides LDL-triglycerides HDL-triglycerides *
Baseline
Monophasic
Triphasic
205.6 f 10.9 13.6* 3.4 128.6k 9.7 63.2& 5.0
216.7 f 10.0 12.9k 2.1 136.4rt 10.4 67.2* 4.2 *
211.4k11.8 16.7& 3.2 131.3 f 12.4 63.2& 6.3 a
82.2 + 16.6 40.3 f 12.3 24.9* 3.2 17.0* 2.2
P < 0.05 vs. baseline; o P < 0.05 vs. monophasic.
97.9 f 40.2k 37.6f 18.1zt
13.9 * 7.8 5.0 * 2.1
%.4* 14.6 * 42.7* 9.3 35.6* 4.9 * 18.1k 1.8
207
I
HDL Fig. 2. Mean rate zonal ultracentrifugation profiles of HDL subfractions from 10 healthy premenopausal women before (. .) and after 6 months of treatment with a monophasic and a triphasic (- - -) oral contraceptive combi( -) nation.
ated cholesterol, the effect being more evident after M (+18.9 and + 12.3% respectively, vs. + 14.0 and +8.8%). Due to these modifications the HDL,/HDL, mass and cholesterol ratios decreased from 0.42 + 0.17 and 0.60 + 0.23 to 0.37 of:0.16 and 0.52 f 0.24 (M; both P > 0.05) and to 0.32 + 0.15 and 0.46 kO.22 (T; both P < 0.05), respectively. The elution profiles from the rate zonal ultracentrifugal gradient clearly showed an increase of material eluting in the HDL, density range after both treatments (Fig. 2); the flotation rate of HDL, was significantly increased after M (Table 3) and did not change after T. The HDL particle size distribution was analyzed by non-denaturing GGE. By this technique, 5 HDL sub-
TABLE
2b
2a 3a 3b3c
Fig. 3. Mean polyacrylamide gradient gel electrophoretic profiles of HDL subfractions from 10 healthy premenopausal women before (. . .) and after 6 months of treatment with a monophasic () and a triphasic (- - -) oral contraceptive combination.
classes have been identified [24], 3 in the HDL, density range, namely HDL,,, HDL,, and HDL,, (with decreasing size), and 2 in the HDL, interval, HDL,, and HDL,,. In all the examined subjects, HDL,, predominated, both before and after treatment (Fig. 3). The HDL,, content was much more variable at baseline, and was not modified by either treatment (Fig. 3). The mean size of HDL,, particles increased slightly after both M and T, with no changes in HDL, diameter (Table 3). Both treatments markedly affected the composition of HDL, and HDL, (Table 4). The FC content was lowered and CE increased; the FC/CE ratio decreased from 0.32 + 0.13 to 0.20 f 0.03 (M) and to 0.18 + 0.04 (T) (both P < 0.05)
3
SERUM LEVELS, WOMEN TAKING Results are expressed
ELUTION VOLUME AND SIZE OF HDL SUBFRACTIONS A MONOPHASIC OR A TRIPHASIC ORAL COMBINATION
IN 10 HEALTHY FOR 6 MONTHS
PREMENOPAUSAL
as mean f SEM.
HDL, Total mass (mg/dl) Associated cholesterol Elution volume (ml) Size (run)
(mg/dl)
HDL, Total mass (mg/dl) Associated cholesterol Elution volume (ml) size (rim)
(mg/dl)
Baseline
Monophasic
Triphasic
124.2 f 17.1 23.7 f 3.1 7.98&- 0.08 10.19 f 0.02
129.0 + 22.4 f 7.98* 10.42+
112.0 f 20.5 19.8 f 3.5 * 7.97* 0.05 10.34* 0.04 *
303.8 k 22.2 39.9 f 2.7 4.73* 0.10 8.61 f 0.06
357.9 k27.1 * + 44.8 f 2.8 *** 4.91 f 0.05 * * 8.64f 0.03
* P < 0.05; * * P < 0.005; * * * P -z0.001 vs.baseline.
15.4 2.7 0.08 0.05 *
349.7 f 32.7 * 43.4 f 3.7 * 4.80+ 0.09 8.61 f 0.03
208 TABLE 4 PERCENTAGE COMPOSITION OF HDL SUBFRACTIONS IN 10 HEALTHY PREMENOPAUSAL WOMEN TAKING A MONOPHASIC OR A TRIPHASIC ORAL CONTRACEPTIVE COMBINATION FOR 6 MONTHS Results are expressed as mean f SD.
HDL, FC% CE% TGI PL% P%
Baseline
Monophasic
Triphasic
6.3f1.4 20.9 + 3.0 4.0 f 1.8 30.5 k 1.6 38.3k2.0
4.2kO.3 ** 21.9k2.6 5.0 f 1.6 28.3 k 1.5 * 40.6 i 2.9
4.1*0.8 ** 22.1 f 2.3 5.0*2.5 26.2* 2.5 * * * 0 42.1 f 2.6 * * *
3.8f0.5 15.8k1.8 1.5f0.8 25.3 k 1.9 53.5 f 2.4
2.3f0.2 *** 17.4* 1.1 * 2.8kO.5 * 24.4& 1.5 53.6k 1.9
2.3kO.3 *** 17.0+1.3 3.1f1.4 *** 23.Okl.4 **O 54.4k2.2
Discussion
HDL, FCI CE% TG% PLZ P%
and to 0.058 * 0.012 (T; P < 0.005 vs baseline and P < 0.05 vs M) in HDL,. No significant changes in the hemostatic indexes (Table 5) occurred with either treatment, only fibrinogen levels being slightly reduced after M. Glucose, uric acid and sitting blood pressure were also not modified. A slight increase of body weight was found with both preparations, from 56.9 f 6.34 to 58.8 &-6.8 kg after M and to 58.2 + 7.2 after T (both P -c0.05). Occasional nausea and slight edema at the upper extremities was reported by 4 participants.
* P < 0.05; ** P < 0.005; ***P
