Pharmacokinetics of gestagens: Some problems K. Fotherby, PhD
London, England Various approaches to studying the pharmacokinetics of gestagens and the factors that influence derivation of the parameters are described with levonorgestrel used as an example. Published studies of the pharmacokinetics of levonorgestrel are reviewed, and new information is presented regarding intra- and intersubject variation. Differences between various formulations of levonorgestrel are apparent when the formulations are compared in the sam~ subjects. There is also a marked difference in the parameters when derived under single-dose or steady-state conditions. The role of sex hormone-binding globulin in the metabolism of levonorgestrel is questioned. Large intra- and inter-subject variations in the parameters exist, and a subject may show a large month-to-month variation when one levonorgestrel formulation is used and smaller variations when another formulation is used. This wide variability in the pharmacokinetic parameters, problems that arise in the derivation and interpretation of the parameters, the biologic significance of most of these parameters, and their lack of correlation with pharmacodynamic responses severely limit the usefulness of pharmacokinetic studies of the gestagens. (AM J OBSTET GVNECOL 1990;163:323-8.) Key words: Pharmacokinetics, gestagens, oral contraceptives, variability, sex
hormone-binding globulin
The comments on pharmacokinetics presented in this article relate only to the gestagens used in oral contraceptives. The pharmacokinetics of contraceptive steroids is a specialized topic, as evidenced by the fact that most textbooks and monographs on pharmacology denote little if any attention to this area. The pharmacokinetics of the gestagens show many facets not seen with most other drugs, and these lead to many problems in both the derivation and interpretation of the pharmacokinetic parameters. There are four main approaches to the pharmacokinetic analysis of serum drug concentrations: (1) Simple description describes the changes in serum drug concentrations with time and allows the determination of the rate of absorption and bioavailability, the rate of elimination of the drug, and the steady-state concentrations. (2) Compartment modeling derives equations based on the assumption of one, two, or more body compartments throughout which the drug is distributed and relates changes in serum drug concentrations with time. The assumed compartments may have no physiologic meaning, and as the number of postulated compartments increases, the mathematic complexity increases so that a model based on more than two compartments becomes difficult to deal with. (3) Iterative curve fitting makes no assumptions about compartments, but by computer analysis attempts to derive the From the Royal Postgraauate Medical School. Reprint requests: K. Fotherby, PhD, Royal Postgraauate Medical School, Ducane Road, London W12 ONN, England. 610117476
equation that best describes the changes in serum drug concentrations with time. Such equations will usually be complex and may be difficult to interpret. (4) Systems dynamics considers the body as a system with a large number of subsystems; the flux of the drug through the various subsystems is determined by positive feedback. 1 This approach has not yet been widely applied and not at all so far to steroids. Each approach has it advantages and disadvantages and that most appropriate for the problem under investigation must be chosen. Compartment modeling, which uses a two-compartment open model, has been widely used for the gestagens, but the simple descriptive approach will give the clinician, although maybe not the pharmacologist, most of the information required. This information will usually be as follows: (1) how rapidly the gestagen is absorbed (the time required [T=x] to reach peak concentration in blood); (2) an indication of bioavailability, usually by assessment of the area under the serum concentration-time curve (AUC); (3) how rapidly the gestagen is eliminated from the body, usually by determining the half-life of elimination; this widely used parameter (the time required for the serum gestagen concentration to decrease by 50%) may, however, often be misleading because it is determined by two other parameters, the apparent volume of distribution and clearance, and these may change considerably, denoting marked changes in the metabolism of the gestagen without any changes in the half-life of elimination. (This aspect is discussed in more detail in a forthcoming article. 2 ) However, the
323
324
Fotherby
July 1990 Am J Obstet Gynecol
Table I. Half-life of elimination (hours) of LNG (values taken from published studies, for definition of "recalculated value," see text) ":)
Reference
n
Reported value (mean) SD or range Recalculated value
LNG 150, EE 30
LNG 1000
Formulation
9 5 11.9 1.7 13.1
lO 5 11.4 1.3 12.0
11
5 13.7 8-23 14.1
11
5 12.6 lO-19 13.3
11
5 12.6 9-20 15.4
12 9 8.0 3.2 9.6
3 24 18.0 8-43 18.0
n, Number of estimates.
half-life of elimination has the advantage that it can be determined simply from the plot of serum gestagen concentrations against time; (4) the serum concentration under steady-state conditions, since this will be more relevant to the biologic activity of the gestagen than concentrations after a single dose. However, regarding the gestagens, steady-state conditions have not often been determined until recently. Since these four parameters are the most often derived and useful ones, the factors influencing their determination and interpretation should be considered. Little difficulty usually arises in T max and steady-state conditions so that only half-life of elimination and Aue will be discussed in more detail.
General considerations Of the gestagens widely used in oral contraceptive formulations, levonorgestrel (LNG) will be mainly used as the example because it has the following advantages: (1) More information is available for LNG than for the other gestagens; (2) it does not undergo a first-pass effect so that complications that may arise as a result of metabolism in the gut or liver are avoided; and (3) most published studies have analyzed their data in a similar way in a two-compartment open model. One disadvantage, but one that is common to the other gestagens, is that most of the pharmacokinetic studies reported have involved only small numbers of subjects. Of the more than 17 papers concerned with serum LNG concentrations after administration of various oral contraceptives containing LNG, only two provide data for more than 10 study periods. Goebelsmann et al.3 studied 24 women, and although only six were investigated by Dennerstein et al.,< sampling was performed during each of 3 months of treatment so that 18 sets of data were available per formulation used. All other studies include less than lO and usually five or less subjects. Most studies have analyzed the data in a two-compartment open model. Although LNG has been used for illustrative purposes, the points considered are equally applicable to all other gestagens used in contraceptive formulations. In determining half-life of elimination and AUe, fac-
tors that may need to be taken into account include the following: (1) the dose of the gestagen and the route by which it is administered; (2) the dose, if any, of estrogen with which it is administered; (3) the status of the subjects at the time blood sampling is performed; (4) the number of samples and the time period over which they are taken; (5) whether taken under steadystate or single dose conditions; (6) the procedures used for calculating the pharmacokinetic parameters; and (7) the reliability criteria of the method used to estimate the serum gestagen concentrations; this is particularly important with the low-dose formulations at the longer times after dosing since by 24 hours, serum concentrations in many subjects will be near the limit of sensitivity of the assay. This is important not only for the validity of the 24-hour concentration but also for determining whether this concentration is significantly different from that of samples taken later. In almost all of the studies, sampling has not continued beyond 24 hours, and this may lead to an underestimate of half-life of elimination, particularly for gestagens with a long half-life of elimination (20 hours or more). In such studies half-life of elimination will include the "13-half life" but not the subsequent half-lives of compounds that do not follow the kinetics of a firstorder reaction. Similarly, considerable errors may be encountered in determining the AUe. If sampling is terminated at 24 hours, AUe (0 to 24 hours) can be calculated by the trapezoid rule, but AUe (24 to 00) may be significant for gestagens with a long half-life, and calculation of AUe (0 to 00) by some suitable equation usually assumes, possibly erroneously, that the elimination rate for the period 24 hours to infinity does not differ from that for 0 to 24 hours.
Elimination of LNG Published values for half-life of elimination of LNG are summarized in Table I. Most of the reported mean values are within a narrow range, although in three studies 3 . 5. 6 considerably longer values were obtained. In one of these studies 3 the longer mean value may result from the larger number of subjects studied and consequently the larger range obtained; this range en-
Pharmacokinetics of gestagens
Volume 163 Number 1, Part 2
.--...
E
LNG 250, EE 50
LNG 750 13 6 12.0 4.8 13.9
6 6 25.1 21-35 14.5
5 6 26.4 7.0 14.9
4 18
4 18
---c
30.0
8
Q)
0
6
c 0
0
C)
19.3
325
4
Z
-.J
E :J
2
Io-
m
compasses the values found in all other studies except that of Dennerstein et al" The fivefold range for halflife of elimination shown in Table I agrees with the large variation reported for half-life of elimination of norethisterone,7 and such a large variation appears to be a common finding with contraceptive steroids. 8 The longer values reported by Humpel et a1. S • 6 are more difficult to explain but they agree with the values calculated from data reported in Dennerstein et al. 4 In this latter study, six women were given either 250 f.Lg LNG or 250 f.Lg LNG with 50 f.Lg ethinylestradiol (EE) daily for 3 months with blood samples being taken at 2, 8, and 26 hours after dosing on the last day of each treatment period. The subjects were randomized to the initial treatment and crossed-over after 3 months of use. From the serum LNG concentrations in the 8 and 26 hour samples, the approximate half-life of elimination was calculated. As shown in Table I, recalculation of half-life of elimination from published studies on the basis of the 8- and 24-hour serum concentrations gave values that, with the exception of two studies,s. 6 were in good agreement with, although slightly higher than, values obtained from the investigators' pharmaco kinetic analyses. Therefore the values calculated from the concentrations reported by Dennerstein et al. 4 appear to be valid. The values for each treatment month for each subject are shown in Table II. In contrast with the other studies, the values in Table II provide information on half-life of elimination under steady-state conditions and within-subject variability. For some subjects (e.g., subjects 1 and 4, LNG 250 + EE 50; subjects 2 and 5, LNG 250) monthly variations were small, whereas for others (e.g., subjects 2 and 6, LNG 250 + EE 50: subjects 3 and 6, LNG 250) the variation was much wider. The mean intersubject variation was higher (47%) in women receiving LNG 250 + EE 50 than in women receiving LNG 250 (32%), but because of the wide range of these values, the differences would not be statistically significant. It is also of interest and importance that subjects may show a high variability when receiving one formulation and not when receiving the other (e.g., subjects 2 and 5). The intersubject variability for half-life of elimination in this study under steady-state conditions (two-
(j)
0
2h 8h 24h Hours after administration
Fig. 1. Serum LNG concentrations in two groups each of six women receiving LNG formulations daily for 21 days D, Women receiving LNG 250 + EE 50; i:2.). women receiving LNG 250. Values are the mean; vertical lines denote SD. (Data from Dennerstein et al.')
to threefold) is probably slightly lower than that observed in single-dose studies. Bioavailability of LNG
Published values for bioavailability as measured by AUe are summarized in Table III. These values must be related to the dose of LNG, and when this is taken into account, there is reasonably good agreement between the mean values. However, the intersubject variability as shown by the values for SD and the range is high, particularly so in one study.' There is also wide intersubject variation with norethisterone,1 and such large variations seem common to all contraceptive steroids. 8 The recalculated values in Table III refer to determinations of AUe based on serum LNG concentrations at 2, 8, and 24 hours published in the various studies and are calculated for comparison with values similarly obtained from serum concentrations at 2, 8, and 26 hours reported by Dennerstein et al. For most of the studies, this recalculated value is about 75% of the value obtained in the investigators' pharmacokinetic analysis. Therefore the calculation of AUe from the concentrations reported by Dennerstein et al. appears to have some validity. The values for each treatment month for each subject are shown in Table IV. These values provide information on AUe under steady-state conditions and on within-subject variability. As for half-life of elimination, some subjects show only small month-to-month variations, whereas for others, this variability is much larger. The mean intrasubject variation was generally low (\3.8%) in women receiving LNG 250 + EE 50 and significantly lower than that in women using LNG 250 (45.3%). The mean value for LNG 250 + EE 50 in this study (99.7%) is much higher
326
Fotherby
July 1990 Am J Obstet Gynecol
Table II. Half-life of elimination (hour) of LNG in a group of six women during each of 3 months of treatment with LNG, 250 fJ-g + EE, 50 fJ-g, or LNG, 250 fJ-g, without estrogen (variation percent denotes percent variation of monthly values in relation to subject's mean value, variation ratio denotes highest monthly value divided by lowest monthly value) LNG 250 + EE 50
Formulation Subject Mol Mo2 Mo3 Mean Variation % Variation ratio Overall mean and range
1 22.5 22.8 28.5 24.6 23.2 1.3
2 54.0 36.0 31.8 40.6 54.7 1.7
4 3 27.0 20.8 35.4 19.8 18.0 23.5 19.5 28.6 14.3 41.6 1.5 1.2 30.0 (18.0-54.0)
LNG 250 5 37.8 23.6 20.1 27.2 65.1 1.9
6 45.0 20.1 52.7 39.3 82.9 2.6
1 16.2 19.6 15.1 17.0 26.5 1.3
2 16.4 16.0 19.4 17.3 19.6 1.2
3 4 18.8 28.7 16.8 17.3 17.3 23.7 20.9 19.9 56.9 32.2 1.7 1.4 19.3 (15.1-28.7)
5 18.6 17.4 15.9 17.3 15.6 1.2
6 17.2 27.3 26.3 23.6 42.8 1.6
Table III. Bioavailability of LNG (AUe in ng/ml/hr) (values taken from published studies, for definition of "recalculated values," see text) LNG 250, EE 50
Formulation Reference n Reported value (mean) SD or range Recalculated value
9 5 36.2 12.8 37.4
10 5 20.4 3.0 24.8
12 9 20.2 7.8 15.2
6 6 32.3 13.7-48.8 23.5
3 24 35.0 9-120 27.0
13 6 116 41 86.2
4 18
4 18
24.3
99.7
n, Number of estimates.
Table IV. Bioavailability (AUe in ng/mllhr) of LNG (for details see legend to Table II)
Subject Mo 1 Mo2 Mo3 Mean Variation % Variation ratio Overall mean and range
I
LNG 250 + EE 50
Formulation 1 89.2 78.4 88.6 85.4 12.6 1.1
2 106.3 82.0 112.5 100.3 30.4 1.4
3 4 68.1 134.0 149.5 73.4 134.1 67.9 139.2 69.8 11.1 7.9 1.1 1.1 99.7 (67.9-149.5)
5 107.6 102.8 96.4 102.3 10.9 1.1
than that (36.2%) found by others 9 ; this is probably because compared with the latter value the former is obtained under steady-state conditions. Also, again probably because the values in Table IV are obtained under steady-state conditions, the intersubject variation was lower than that seen in single-dose studies. Role of sex hormone-binding globulin in LNG pharmacokinetics
LNG is known to bind to sex hormone-binding globulin (SHBG) with a greater affinity than other gestagens, and about 50% of the LNG in serum is bound to this protein! The SHBG concentration in serum can be greatly increased by estrogen administration, and in several studies the higher levels of LNG in serum when the gestagen is given with estrogen compared with
6 96.0 101.8 106.2 101.3 10.1 1.1
LNG 250 1 37.7 21.7 18.8 26.1 72.4 2.0
2 20.6 17.3 12.9 16.9 45.5 1.6
3 4 15.4 27.5 36.9 22.1 34.3 17.0 28.9 22.2 74.4 46.4 1.6 2.4 24.3 (15.4-37.7)
5 28.8 22.4 24.4 25.2 25.4 1.3
6 25.7 25.9 27.8 26.5 7.9 1.1
those when the same dose of LNG is given without estrogen have been attributed to binding of LNG to the increased levels of SHBG:· 11. 14 Examples of this phenomenon are shown in Figs. 1 and 2. ~n Fig. 2 it will also be noted that about 8 days are required to achieve steady-state levels of LNG. The mean value for half-life of elimination from the studies shown in Table I is 15 hours so that steady-state levels of LNG should have been attained in about 3 days. The rise in SHBG concentrations induced by estrogen occurs gradually from the first day of dosing until about day 10 when a plateau is reached,15 and this has been used as further evidence that the slow attainment of steadystate levels of LNG is associated with the rise in serum SHBG concentrations. However, these suppositions appear untenable. Four studies of changes in serum
Pharmacokinetics of gestagens 327
Volume 163 Number I, Part 2
VIIIIIIIIIII1711777777777777777J E (Jl Q
3000
c
o
oJ
!1J I...
oJ
C
ill
U
C
2UOO
o
u
\.J Z
-1
E :J
I...
London, England Various approaches to studying the pharmacokinetics of gestagens and the factors that influence derivation of the parameters are described with levonorgestrel used as an example. Published studies of the pharmacokinetics of levonorgestrel are reviewed, and new information is presented regarding intra- and intersubject variation. Differences between various formulations of levonorgestrel are apparent when the formulations are compared in the sam~ subjects. There is also a marked difference in the parameters when derived under single-dose or steady-state conditions. The role of sex hormone-binding globulin in the metabolism of levonorgestrel is questioned. Large intra- and inter-subject variations in the parameters exist, and a subject may show a large month-to-month variation when one levonorgestrel formulation is used and smaller variations when another formulation is used. This wide variability in the pharmacokinetic parameters, problems that arise in the derivation and interpretation of the parameters, the biologic significance of most of these parameters, and their lack of correlation with pharmacodynamic responses severely limit the usefulness of pharmacokinetic studies of the gestagens. (AM J OBSTET GVNECOL 1990;163:323-8.) Key words: Pharmacokinetics, gestagens, oral contraceptives, variability, sex
hormone-binding globulin
The comments on pharmacokinetics presented in this article relate only to the gestagens used in oral contraceptives. The pharmacokinetics of contraceptive steroids is a specialized topic, as evidenced by the fact that most textbooks and monographs on pharmacology denote little if any attention to this area. The pharmacokinetics of the gestagens show many facets not seen with most other drugs, and these lead to many problems in both the derivation and interpretation of the pharmacokinetic parameters. There are four main approaches to the pharmacokinetic analysis of serum drug concentrations: (1) Simple description describes the changes in serum drug concentrations with time and allows the determination of the rate of absorption and bioavailability, the rate of elimination of the drug, and the steady-state concentrations. (2) Compartment modeling derives equations based on the assumption of one, two, or more body compartments throughout which the drug is distributed and relates changes in serum drug concentrations with time. The assumed compartments may have no physiologic meaning, and as the number of postulated compartments increases, the mathematic complexity increases so that a model based on more than two compartments becomes difficult to deal with. (3) Iterative curve fitting makes no assumptions about compartments, but by computer analysis attempts to derive the From the Royal Postgraauate Medical School. Reprint requests: K. Fotherby, PhD, Royal Postgraauate Medical School, Ducane Road, London W12 ONN, England. 610117476
equation that best describes the changes in serum drug concentrations with time. Such equations will usually be complex and may be difficult to interpret. (4) Systems dynamics considers the body as a system with a large number of subsystems; the flux of the drug through the various subsystems is determined by positive feedback. 1 This approach has not yet been widely applied and not at all so far to steroids. Each approach has it advantages and disadvantages and that most appropriate for the problem under investigation must be chosen. Compartment modeling, which uses a two-compartment open model, has been widely used for the gestagens, but the simple descriptive approach will give the clinician, although maybe not the pharmacologist, most of the information required. This information will usually be as follows: (1) how rapidly the gestagen is absorbed (the time required [T=x] to reach peak concentration in blood); (2) an indication of bioavailability, usually by assessment of the area under the serum concentration-time curve (AUC); (3) how rapidly the gestagen is eliminated from the body, usually by determining the half-life of elimination; this widely used parameter (the time required for the serum gestagen concentration to decrease by 50%) may, however, often be misleading because it is determined by two other parameters, the apparent volume of distribution and clearance, and these may change considerably, denoting marked changes in the metabolism of the gestagen without any changes in the half-life of elimination. (This aspect is discussed in more detail in a forthcoming article. 2 ) However, the
323
324
Fotherby
July 1990 Am J Obstet Gynecol
Table I. Half-life of elimination (hours) of LNG (values taken from published studies, for definition of "recalculated value," see text) ":)
Reference
n
Reported value (mean) SD or range Recalculated value
LNG 150, EE 30
LNG 1000
Formulation
9 5 11.9 1.7 13.1
lO 5 11.4 1.3 12.0
11
5 13.7 8-23 14.1
11
5 12.6 lO-19 13.3
11
5 12.6 9-20 15.4
12 9 8.0 3.2 9.6
3 24 18.0 8-43 18.0
n, Number of estimates.
half-life of elimination has the advantage that it can be determined simply from the plot of serum gestagen concentrations against time; (4) the serum concentration under steady-state conditions, since this will be more relevant to the biologic activity of the gestagen than concentrations after a single dose. However, regarding the gestagens, steady-state conditions have not often been determined until recently. Since these four parameters are the most often derived and useful ones, the factors influencing their determination and interpretation should be considered. Little difficulty usually arises in T max and steady-state conditions so that only half-life of elimination and Aue will be discussed in more detail.
General considerations Of the gestagens widely used in oral contraceptive formulations, levonorgestrel (LNG) will be mainly used as the example because it has the following advantages: (1) More information is available for LNG than for the other gestagens; (2) it does not undergo a first-pass effect so that complications that may arise as a result of metabolism in the gut or liver are avoided; and (3) most published studies have analyzed their data in a similar way in a two-compartment open model. One disadvantage, but one that is common to the other gestagens, is that most of the pharmacokinetic studies reported have involved only small numbers of subjects. Of the more than 17 papers concerned with serum LNG concentrations after administration of various oral contraceptives containing LNG, only two provide data for more than 10 study periods. Goebelsmann et al.3 studied 24 women, and although only six were investigated by Dennerstein et al.,< sampling was performed during each of 3 months of treatment so that 18 sets of data were available per formulation used. All other studies include less than lO and usually five or less subjects. Most studies have analyzed the data in a two-compartment open model. Although LNG has been used for illustrative purposes, the points considered are equally applicable to all other gestagens used in contraceptive formulations. In determining half-life of elimination and AUe, fac-
tors that may need to be taken into account include the following: (1) the dose of the gestagen and the route by which it is administered; (2) the dose, if any, of estrogen with which it is administered; (3) the status of the subjects at the time blood sampling is performed; (4) the number of samples and the time period over which they are taken; (5) whether taken under steadystate or single dose conditions; (6) the procedures used for calculating the pharmacokinetic parameters; and (7) the reliability criteria of the method used to estimate the serum gestagen concentrations; this is particularly important with the low-dose formulations at the longer times after dosing since by 24 hours, serum concentrations in many subjects will be near the limit of sensitivity of the assay. This is important not only for the validity of the 24-hour concentration but also for determining whether this concentration is significantly different from that of samples taken later. In almost all of the studies, sampling has not continued beyond 24 hours, and this may lead to an underestimate of half-life of elimination, particularly for gestagens with a long half-life of elimination (20 hours or more). In such studies half-life of elimination will include the "13-half life" but not the subsequent half-lives of compounds that do not follow the kinetics of a firstorder reaction. Similarly, considerable errors may be encountered in determining the AUe. If sampling is terminated at 24 hours, AUe (0 to 24 hours) can be calculated by the trapezoid rule, but AUe (24 to 00) may be significant for gestagens with a long half-life, and calculation of AUe (0 to 00) by some suitable equation usually assumes, possibly erroneously, that the elimination rate for the period 24 hours to infinity does not differ from that for 0 to 24 hours.
Elimination of LNG Published values for half-life of elimination of LNG are summarized in Table I. Most of the reported mean values are within a narrow range, although in three studies 3 . 5. 6 considerably longer values were obtained. In one of these studies 3 the longer mean value may result from the larger number of subjects studied and consequently the larger range obtained; this range en-
Pharmacokinetics of gestagens
Volume 163 Number 1, Part 2
.--...
E
LNG 250, EE 50
LNG 750 13 6 12.0 4.8 13.9
6 6 25.1 21-35 14.5
5 6 26.4 7.0 14.9
4 18
4 18
---c
30.0
8
Q)
0
6
c 0
0
C)
19.3
325
4
Z
-.J
E :J
2
Io-
m
compasses the values found in all other studies except that of Dennerstein et al" The fivefold range for halflife of elimination shown in Table I agrees with the large variation reported for half-life of elimination of norethisterone,7 and such a large variation appears to be a common finding with contraceptive steroids. 8 The longer values reported by Humpel et a1. S • 6 are more difficult to explain but they agree with the values calculated from data reported in Dennerstein et al. 4 In this latter study, six women were given either 250 f.Lg LNG or 250 f.Lg LNG with 50 f.Lg ethinylestradiol (EE) daily for 3 months with blood samples being taken at 2, 8, and 26 hours after dosing on the last day of each treatment period. The subjects were randomized to the initial treatment and crossed-over after 3 months of use. From the serum LNG concentrations in the 8 and 26 hour samples, the approximate half-life of elimination was calculated. As shown in Table I, recalculation of half-life of elimination from published studies on the basis of the 8- and 24-hour serum concentrations gave values that, with the exception of two studies,s. 6 were in good agreement with, although slightly higher than, values obtained from the investigators' pharmaco kinetic analyses. Therefore the values calculated from the concentrations reported by Dennerstein et al. 4 appear to be valid. The values for each treatment month for each subject are shown in Table II. In contrast with the other studies, the values in Table II provide information on half-life of elimination under steady-state conditions and within-subject variability. For some subjects (e.g., subjects 1 and 4, LNG 250 + EE 50; subjects 2 and 5, LNG 250) monthly variations were small, whereas for others (e.g., subjects 2 and 6, LNG 250 + EE 50: subjects 3 and 6, LNG 250) the variation was much wider. The mean intersubject variation was higher (47%) in women receiving LNG 250 + EE 50 than in women receiving LNG 250 (32%), but because of the wide range of these values, the differences would not be statistically significant. It is also of interest and importance that subjects may show a high variability when receiving one formulation and not when receiving the other (e.g., subjects 2 and 5). The intersubject variability for half-life of elimination in this study under steady-state conditions (two-
(j)
0
2h 8h 24h Hours after administration
Fig. 1. Serum LNG concentrations in two groups each of six women receiving LNG formulations daily for 21 days D, Women receiving LNG 250 + EE 50; i:2.). women receiving LNG 250. Values are the mean; vertical lines denote SD. (Data from Dennerstein et al.')
to threefold) is probably slightly lower than that observed in single-dose studies. Bioavailability of LNG
Published values for bioavailability as measured by AUe are summarized in Table III. These values must be related to the dose of LNG, and when this is taken into account, there is reasonably good agreement between the mean values. However, the intersubject variability as shown by the values for SD and the range is high, particularly so in one study.' There is also wide intersubject variation with norethisterone,1 and such large variations seem common to all contraceptive steroids. 8 The recalculated values in Table III refer to determinations of AUe based on serum LNG concentrations at 2, 8, and 24 hours published in the various studies and are calculated for comparison with values similarly obtained from serum concentrations at 2, 8, and 26 hours reported by Dennerstein et al. For most of the studies, this recalculated value is about 75% of the value obtained in the investigators' pharmacokinetic analysis. Therefore the calculation of AUe from the concentrations reported by Dennerstein et al. appears to have some validity. The values for each treatment month for each subject are shown in Table IV. These values provide information on AUe under steady-state conditions and on within-subject variability. As for half-life of elimination, some subjects show only small month-to-month variations, whereas for others, this variability is much larger. The mean intrasubject variation was generally low (\3.8%) in women receiving LNG 250 + EE 50 and significantly lower than that in women using LNG 250 (45.3%). The mean value for LNG 250 + EE 50 in this study (99.7%) is much higher
326
Fotherby
July 1990 Am J Obstet Gynecol
Table II. Half-life of elimination (hour) of LNG in a group of six women during each of 3 months of treatment with LNG, 250 fJ-g + EE, 50 fJ-g, or LNG, 250 fJ-g, without estrogen (variation percent denotes percent variation of monthly values in relation to subject's mean value, variation ratio denotes highest monthly value divided by lowest monthly value) LNG 250 + EE 50
Formulation Subject Mol Mo2 Mo3 Mean Variation % Variation ratio Overall mean and range
1 22.5 22.8 28.5 24.6 23.2 1.3
2 54.0 36.0 31.8 40.6 54.7 1.7
4 3 27.0 20.8 35.4 19.8 18.0 23.5 19.5 28.6 14.3 41.6 1.5 1.2 30.0 (18.0-54.0)
LNG 250 5 37.8 23.6 20.1 27.2 65.1 1.9
6 45.0 20.1 52.7 39.3 82.9 2.6
1 16.2 19.6 15.1 17.0 26.5 1.3
2 16.4 16.0 19.4 17.3 19.6 1.2
3 4 18.8 28.7 16.8 17.3 17.3 23.7 20.9 19.9 56.9 32.2 1.7 1.4 19.3 (15.1-28.7)
5 18.6 17.4 15.9 17.3 15.6 1.2
6 17.2 27.3 26.3 23.6 42.8 1.6
Table III. Bioavailability of LNG (AUe in ng/ml/hr) (values taken from published studies, for definition of "recalculated values," see text) LNG 250, EE 50
Formulation Reference n Reported value (mean) SD or range Recalculated value
9 5 36.2 12.8 37.4
10 5 20.4 3.0 24.8
12 9 20.2 7.8 15.2
6 6 32.3 13.7-48.8 23.5
3 24 35.0 9-120 27.0
13 6 116 41 86.2
4 18
4 18
24.3
99.7
n, Number of estimates.
Table IV. Bioavailability (AUe in ng/mllhr) of LNG (for details see legend to Table II)
Subject Mo 1 Mo2 Mo3 Mean Variation % Variation ratio Overall mean and range
I
LNG 250 + EE 50
Formulation 1 89.2 78.4 88.6 85.4 12.6 1.1
2 106.3 82.0 112.5 100.3 30.4 1.4
3 4 68.1 134.0 149.5 73.4 134.1 67.9 139.2 69.8 11.1 7.9 1.1 1.1 99.7 (67.9-149.5)
5 107.6 102.8 96.4 102.3 10.9 1.1
than that (36.2%) found by others 9 ; this is probably because compared with the latter value the former is obtained under steady-state conditions. Also, again probably because the values in Table IV are obtained under steady-state conditions, the intersubject variation was lower than that seen in single-dose studies. Role of sex hormone-binding globulin in LNG pharmacokinetics
LNG is known to bind to sex hormone-binding globulin (SHBG) with a greater affinity than other gestagens, and about 50% of the LNG in serum is bound to this protein! The SHBG concentration in serum can be greatly increased by estrogen administration, and in several studies the higher levels of LNG in serum when the gestagen is given with estrogen compared with
6 96.0 101.8 106.2 101.3 10.1 1.1
LNG 250 1 37.7 21.7 18.8 26.1 72.4 2.0
2 20.6 17.3 12.9 16.9 45.5 1.6
3 4 15.4 27.5 36.9 22.1 34.3 17.0 28.9 22.2 74.4 46.4 1.6 2.4 24.3 (15.4-37.7)
5 28.8 22.4 24.4 25.2 25.4 1.3
6 25.7 25.9 27.8 26.5 7.9 1.1
those when the same dose of LNG is given without estrogen have been attributed to binding of LNG to the increased levels of SHBG:· 11. 14 Examples of this phenomenon are shown in Figs. 1 and 2. ~n Fig. 2 it will also be noted that about 8 days are required to achieve steady-state levels of LNG. The mean value for half-life of elimination from the studies shown in Table I is 15 hours so that steady-state levels of LNG should have been attained in about 3 days. The rise in SHBG concentrations induced by estrogen occurs gradually from the first day of dosing until about day 10 when a plateau is reached,15 and this has been used as further evidence that the slow attainment of steadystate levels of LNG is associated with the rise in serum SHBG concentrations. However, these suppositions appear untenable. Four studies of changes in serum
Pharmacokinetics of gestagens 327
Volume 163 Number I, Part 2
VIIIIIIIIIII1711777777777777777J E (Jl Q
3000
c
o
oJ
!1J I...
oJ
C
ill
U
C
2UOO
o
u
\.J Z
-1
E :J
I...
