Saturday 19 May 1979
COMPARISON OF EFFECTS OF DIFFERENT COMBINED ORAL-CONTRACEPTIVE
FORMULATIONS ON CARBOHYDRATE AND LIPID METABOLISM V. WYNN I. GODSLAND R. NITHTHYANANTHAN
P. W. ADAMS J. MELROSE N. W. OAKLEY
M. SEED Alexander Simpson Laboratory for Metabolic Research, St. Mary’s Hospital Medical School, London W2 1PG
Oral glucose-tolerance tests were performed and fasting serum cholesterol and triglyceride levels measured in 1628 Caucasian women taking combined œstrogen/progestagen oral contraceptives (o.c.) and 577 women not taking o.c. The former were divided into six groups according to the composition of the o.c. Glucose tolerance deteriorated in all o.c. groups containing œstrane progestagens (nortestosterone-derived) or the gonane, norgestrel, but was unaltered by o.c. containing a pregnane progestagen (derived from progesterone). The greatest deterioration was with o.c. containing 75 µg or more œstrogen, and this was associated with impairment of the early insulin response to glucose. In o.c. containing a pregnane progestagen insulin secretion was unaffected. In the remaining o.c. groups insulin secretion was increased; this was most pronounced with the o.c. containing a gonane progestagen. Serum-cholesterol was elevated only with o.c. containing 75 µg or more œstrogen and an œstrane progestagen and tended to be lower in o.c. containing a gonane progestagen. o.c.-induced hypertriglyceridæmia was œstrogen-dose-related, and this effect was potentiated by the pregnane progestagen. The gonane proges-
Summary
tagen
antagonised œstrogen-induced hypertriglyceridæ-
mia.
Introduction
OESTROGEN-PROGESTAGEN oral
contraceptives (o.c.)
produce changes in carbohydrate and lipid metabolism which have been widely studied 1,2 but conflicting results3.4 have hampered the search for pills producing the least metabolic disturbance. Much of the confusion has arisen because insufficient consideration has been given to the selection of subjects, the dose of oestrogen, and the dose and type of progestagen used. Large numbers of subjects are needed for this type of investigation. The present paper reports the results of a metabolic study of 2205 Caucasian women.
Structural and Functional Relationships between Gonadal Steroids Two synthetic oestrogens are used in o.c.-ethinyl restradiol and mestranol (M.E.), the 3-methyl ether of E.o. Claims that M.E. is a weaker oestrogen than E.o. are not borne out by studies of their metabolic effects. Indeed M.E. is inactive until converted to E.o. by demethylation in the liver.5
(E.o.)
Synthetic progestagens are pregnanes (megestrol, chlormadinone, and medroxyprogesterone acetates), oestranes (norethisterone, norethisterone acetate, lynoestranol, ethynodiol diacetate, and norethynodrel), and a gonane (norgestrel). (Norgestrel was originally synthesised as a racemic mixture of which the d-isomer was active and the I-isomer inactive. The W.H.O. Expert Committee on Non-Proprietary Names for Pharmaceutical Substances has recommended that d-norgestrel be referred to as levonorgestrel. For simplicity we have retained the older terminology in this paper.) No contraceptive containing a pregnane progestagen is now recommended because of doubts raised by the occurrence of breast tumours in beagle dogs given this type of progestagen. The cestrane progestagens, more commonly known as 19-nortestosterone (19-N.T.) progestagens, are all metabolised to norethisterone before becoming active ;6 norethisterone is therefore the basic progestagen of this group. The gonane, d-norgestrel, differs from norethisterone in having an angular substitution on the 13-carbon atom instead of a methyl group. It binds avidly to the progesterone receptor and is an active progestagen without metabolic transformation.’7
Subjects (Tables The
577
I and
II)
study groups comprised wishing to start (controls) and 1628 women already taking the pill for a minimum of 3 months (test subjects). The women using o.c. were divided into six groups (table I) to compare the metabolic effects of high (75-150 fLg, group 1), medium (50 jjLg, groups 2, 3, and 4), and low (300 ug, groups 5 and 6) oestrogen concenwomen
o.c.
tration ; of the three types of progestagen when combined with 50 ;jLg oestrogen (groups 2, 3, and 4); and of high (500 ug dlnorgestrel, group 5) and low (150 ug d-norgestrel, group 6) concentrations of the gonane progestagen. None of the women were aged over 45, exceeded 150% of their ideal body-weight, had been pregnant within 6 months, or had any evidence of endocrine or metabolic disease. The means (±1 s.D.) for age, 17( ideal body-weight, and duration of o.c. use are shown in table n. The frequency of a family history of diabetes was the same in controls as in test subjects.
Methods Standard oral glucose-tolerance tests (o.G.T.T.) were done with a 1 g/kg glucose load as previously described.’ Fasting serum was used for estimation of triglyceride8 and cholesterol,9 and plasma taken fasting and at half-hourly intervals during the O.G.T.T. was used for glucose1o and insulin estimations." 8125
1046 The area under the 3 h O.G.T.T. curve was used to derive single indices of carbohydrate tolerance.’ Both the "total area" and the "incremental area" (the area above the fasting plasma-glucose value) were calculated. The upper limit of normal for the total and incremental areas respectively have been taken at 800 and 300 units.12 Data were analysed with the SPSS programmes available at the University of London computer centre. In all calculations the skew distributions of serum triglyceride and insulin values were normalised by log transformation. Control data were age-matched for each o.c. group. The percentage of women in the six 5-year age-groups from 15 to 45 in each o.c. group was calculated, and the control data were then weighted so that the contribution to the overall
from women within each 5-year age-group corresponded that within the group of o.c. users with which it was to be
mean
to
compared.
(Table H) The mean ages of the users were higher than those of the controls. Group-1 subjects (high oestrogen) were significantly older than the other o.c. groups. Except for group 2, the mean % ideal body-weight was significantly higher in users of o.c. than in age-matched controls. Group 2 were significantly lighter than group 1, but there were no significant differences between other o.c. Results
TABLE I-COMPOSITION OF ORAL CONTRACEPTIVES
E.D.D.=ethynodiol diacetate. E.o.=ethinyl oestradiol. LYN=Iyncestranol.
MEG=megestrol acetate.
M.E.=mestranol.
N.E.L.=norethynodrel. N.G.=norgestrel.
TABLE II-METABOLIC MEASUREMENTS
I *
I
I
N.E.=norethisterone. N.E.A.=norethisterone acetate.
(MEAN ± IS.D.)
I
Triglyceride and insulin values have been calculated from log-transformed data. t For composition of o.c. taken see table i.
IN PILL USERS AND
CONTROLS*
1047 groups. The o.c. groups differed duration of o.c. use. (table n).
considerably
in the
O.G.T.T. Plasma-glucose Results
(Figs.and 2, Table II) The fasting plasma-glucose level was lower in all six o.c. groups than in their age-matched controls (fig. 1). This effect was smallest with o.c. containing a gonane progestagen, and among these it was significant only in the largest group (group 6). Oral glucose tolerance was impaired in all the o.c. groups except group 2, in which the o.c. contained a pregnane progestagen (fig. 2). The high-oestrogen o.c. (group 1) showed the greatest elevation of the oral glucose-tolerance curve: this was prolonged so that even the 3 h value was significantly higher than that of the agematched controls, and the glucose peak was delayed (fig. 2). This pattern was also found in individual analyses of the four largest individual pill groups within the group-1 population (table i). Groups 3, 4, and 6 also showed elevation of the mean glucose-tolerance curves with delayed plasma-glucose peak, although the mean differences were smaller and were statistically significant only
the earlier part of the o.G.T.T. Group 5 barely differed from its age-matched controls. Fig. 1 summarises these changes in terms of the incremental O.G.T.T. area. By far the greatest deterioration in glucose tolerance was seen in group 1, in which 28% of the women had abnormal incremental O.G.T.T. glucose areas, and these differed significantly from group 3 with the same cestrane-progestagen composition but with a medium oestrogen concentration (P
COMPARISON OF EFFECTS OF DIFFERENT COMBINED ORAL-CONTRACEPTIVE
FORMULATIONS ON CARBOHYDRATE AND LIPID METABOLISM V. WYNN I. GODSLAND R. NITHTHYANANTHAN
P. W. ADAMS J. MELROSE N. W. OAKLEY
M. SEED Alexander Simpson Laboratory for Metabolic Research, St. Mary’s Hospital Medical School, London W2 1PG
Oral glucose-tolerance tests were performed and fasting serum cholesterol and triglyceride levels measured in 1628 Caucasian women taking combined œstrogen/progestagen oral contraceptives (o.c.) and 577 women not taking o.c. The former were divided into six groups according to the composition of the o.c. Glucose tolerance deteriorated in all o.c. groups containing œstrane progestagens (nortestosterone-derived) or the gonane, norgestrel, but was unaltered by o.c. containing a pregnane progestagen (derived from progesterone). The greatest deterioration was with o.c. containing 75 µg or more œstrogen, and this was associated with impairment of the early insulin response to glucose. In o.c. containing a pregnane progestagen insulin secretion was unaffected. In the remaining o.c. groups insulin secretion was increased; this was most pronounced with the o.c. containing a gonane progestagen. Serum-cholesterol was elevated only with o.c. containing 75 µg or more œstrogen and an œstrane progestagen and tended to be lower in o.c. containing a gonane progestagen. o.c.-induced hypertriglyceridæmia was œstrogen-dose-related, and this effect was potentiated by the pregnane progestagen. The gonane proges-
Summary
tagen
antagonised œstrogen-induced hypertriglyceridæ-
mia.
Introduction
OESTROGEN-PROGESTAGEN oral
contraceptives (o.c.)
produce changes in carbohydrate and lipid metabolism which have been widely studied 1,2 but conflicting results3.4 have hampered the search for pills producing the least metabolic disturbance. Much of the confusion has arisen because insufficient consideration has been given to the selection of subjects, the dose of oestrogen, and the dose and type of progestagen used. Large numbers of subjects are needed for this type of investigation. The present paper reports the results of a metabolic study of 2205 Caucasian women.
Structural and Functional Relationships between Gonadal Steroids Two synthetic oestrogens are used in o.c.-ethinyl restradiol and mestranol (M.E.), the 3-methyl ether of E.o. Claims that M.E. is a weaker oestrogen than E.o. are not borne out by studies of their metabolic effects. Indeed M.E. is inactive until converted to E.o. by demethylation in the liver.5
(E.o.)
Synthetic progestagens are pregnanes (megestrol, chlormadinone, and medroxyprogesterone acetates), oestranes (norethisterone, norethisterone acetate, lynoestranol, ethynodiol diacetate, and norethynodrel), and a gonane (norgestrel). (Norgestrel was originally synthesised as a racemic mixture of which the d-isomer was active and the I-isomer inactive. The W.H.O. Expert Committee on Non-Proprietary Names for Pharmaceutical Substances has recommended that d-norgestrel be referred to as levonorgestrel. For simplicity we have retained the older terminology in this paper.) No contraceptive containing a pregnane progestagen is now recommended because of doubts raised by the occurrence of breast tumours in beagle dogs given this type of progestagen. The cestrane progestagens, more commonly known as 19-nortestosterone (19-N.T.) progestagens, are all metabolised to norethisterone before becoming active ;6 norethisterone is therefore the basic progestagen of this group. The gonane, d-norgestrel, differs from norethisterone in having an angular substitution on the 13-carbon atom instead of a methyl group. It binds avidly to the progesterone receptor and is an active progestagen without metabolic transformation.’7
Subjects (Tables The
577
I and
II)
study groups comprised wishing to start (controls) and 1628 women already taking the pill for a minimum of 3 months (test subjects). The women using o.c. were divided into six groups (table I) to compare the metabolic effects of high (75-150 fLg, group 1), medium (50 jjLg, groups 2, 3, and 4), and low (300 ug, groups 5 and 6) oestrogen concenwomen
o.c.
tration ; of the three types of progestagen when combined with 50 ;jLg oestrogen (groups 2, 3, and 4); and of high (500 ug dlnorgestrel, group 5) and low (150 ug d-norgestrel, group 6) concentrations of the gonane progestagen. None of the women were aged over 45, exceeded 150% of their ideal body-weight, had been pregnant within 6 months, or had any evidence of endocrine or metabolic disease. The means (±1 s.D.) for age, 17( ideal body-weight, and duration of o.c. use are shown in table n. The frequency of a family history of diabetes was the same in controls as in test subjects.
Methods Standard oral glucose-tolerance tests (o.G.T.T.) were done with a 1 g/kg glucose load as previously described.’ Fasting serum was used for estimation of triglyceride8 and cholesterol,9 and plasma taken fasting and at half-hourly intervals during the O.G.T.T. was used for glucose1o and insulin estimations." 8125
1046 The area under the 3 h O.G.T.T. curve was used to derive single indices of carbohydrate tolerance.’ Both the "total area" and the "incremental area" (the area above the fasting plasma-glucose value) were calculated. The upper limit of normal for the total and incremental areas respectively have been taken at 800 and 300 units.12 Data were analysed with the SPSS programmes available at the University of London computer centre. In all calculations the skew distributions of serum triglyceride and insulin values were normalised by log transformation. Control data were age-matched for each o.c. group. The percentage of women in the six 5-year age-groups from 15 to 45 in each o.c. group was calculated, and the control data were then weighted so that the contribution to the overall
from women within each 5-year age-group corresponded that within the group of o.c. users with which it was to be
mean
to
compared.
(Table H) The mean ages of the users were higher than those of the controls. Group-1 subjects (high oestrogen) were significantly older than the other o.c. groups. Except for group 2, the mean % ideal body-weight was significantly higher in users of o.c. than in age-matched controls. Group 2 were significantly lighter than group 1, but there were no significant differences between other o.c. Results
TABLE I-COMPOSITION OF ORAL CONTRACEPTIVES
E.D.D.=ethynodiol diacetate. E.o.=ethinyl oestradiol. LYN=Iyncestranol.
MEG=megestrol acetate.
M.E.=mestranol.
N.E.L.=norethynodrel. N.G.=norgestrel.
TABLE II-METABOLIC MEASUREMENTS
I *
I
I
N.E.=norethisterone. N.E.A.=norethisterone acetate.
(MEAN ± IS.D.)
I
Triglyceride and insulin values have been calculated from log-transformed data. t For composition of o.c. taken see table i.
IN PILL USERS AND
CONTROLS*
1047 groups. The o.c. groups differed duration of o.c. use. (table n).
considerably
in the
O.G.T.T. Plasma-glucose Results
(Figs.and 2, Table II) The fasting plasma-glucose level was lower in all six o.c. groups than in their age-matched controls (fig. 1). This effect was smallest with o.c. containing a gonane progestagen, and among these it was significant only in the largest group (group 6). Oral glucose tolerance was impaired in all the o.c. groups except group 2, in which the o.c. contained a pregnane progestagen (fig. 2). The high-oestrogen o.c. (group 1) showed the greatest elevation of the oral glucose-tolerance curve: this was prolonged so that even the 3 h value was significantly higher than that of the agematched controls, and the glucose peak was delayed (fig. 2). This pattern was also found in individual analyses of the four largest individual pill groups within the group-1 population (table i). Groups 3, 4, and 6 also showed elevation of the mean glucose-tolerance curves with delayed plasma-glucose peak, although the mean differences were smaller and were statistically significant only
the earlier part of the o.G.T.T. Group 5 barely differed from its age-matched controls. Fig. 1 summarises these changes in terms of the incremental O.G.T.T. area. By far the greatest deterioration in glucose tolerance was seen in group 1, in which 28% of the women had abnormal incremental O.G.T.T. glucose areas, and these differed significantly from group 3 with the same cestrane-progestagen composition but with a medium oestrogen concentration (P
