Levormeloxifene: safety, pharmacodynamics and pharmacokinetics in healthy postmenopausal women following single and multiple doses of a new selective oestrogen receptor modulator

Abstract

Aims

The safety, pharmacodynamics and pharmacokinetics of levormeloxifene, a selective oestrogen receptor modulator (SERM), were investigated in postmenopausal women following single doses and multiple dosing once daily up to 56 days.

Methods

The two randomized, double-blind, placebo controlled studies of six single ascending doses and at four multiple dose levels, respectively, included a total of 104 healthy postmenopausal women. Safety assessments comprised vital signs, ECG, haematology, clinical chemistry and reporting of adverse events. The pharmacodynamic properties were investigated after multiple dosing by assessment of the short-term effects on bone and lipid metabolism and on the hypothalamic-pituitary axis. Blood samples for pharmacokinetic analysis were collected at intervals until 648 h (27 days) after single and multiple dosing.

Results

Levormeloxifene was tolerated well after single doses in the range of 2.5–320 mg and multiple once daily dosing in the range of 20–160 mg. Adverse events reported were generally mild or moderate. The most frequent adverse events after multiple dosing were headache, abdominal pain and leukorrhea with the highest frequency reported after the highest daily dose of 160 mg levormeloxifene. Five weeks of treatment with 20–160 mg levormeloxifene and 8 weeks of treatment with 40 or 80 mg levormeloxifene reduced the biochemical marker of bone turnover, the collagen I C-terminal telopeptide (CrossLaps™) by 44.4% [95% CI: 11.3, 65.1] and 35.5% [95% CI: 14.0, 51.6], respectively, without any dose-dependent decrease in the studied dose range. The total cholesterol and LDL-cholesterol concentrations were significantly reduced by 19–25% and 28–35%, respectively, when compared with placebo. HDL-cholesterol and triglyceride concentrations were not affected. An oestrogen-like effect on the hypothalamic-pituitary axis was observed with approximately 50% reductions of FSH and LH after 8 weeks of treatment. No clinically significant changes of other safety variables were observed. The pharmacokinetic analysis demonstrated a rapid absorption (mean t_max: 2–3 h), a slow elimination (mean t_1/2: 4.8–8.4 days) and dose linearity of C_max and AUC for doses up to 160 mg. As expected for a drug with slow elimination given frequently, the relative fluctuation around the steady state plasma concentration was small and the drug accumulation considerable (R_A: 3–5).

Conclusions

Short-term administration of levormeloxifene in postmenopausal women was well-tolerated at doses that elicited a favourable pharmacodynamic response suggesting oestrogen-like bone preserving and antiatherogenic effects. Little variation of peak-trough plasma concentrations was observed during daily administration due to a plasma half-life of approximately 1 week.

Introduction

Levormeloxifene (NNC 46–0020, (−)-3,4-trans-7-methoxy-2,2-dimethyl-3-phenyl-4-[4-[2-(pyrrolidin-1-yl)ethoxy] phenyl]chromane) is a selective oestrogen receptor modulator (SERM). It is the (−)-enantiomer of ormeloxifene which, under the name Centchroman, has been on the market in India as an oral contraceptive since the late 1980s and is presently under development in India for the treatment of advanced breast cancer [1].

Oestrogen replacement therapy (ERT) is effective in both preventing postmenopausal osteoporosis and reducing the risk of cardiovascular disease [2, 3]. However, without the concomitant administration of progesterone supplements, ERT is associated with an increased stimulation of the endometrium causing hyperplasia and risk of cancer [4–6]. As an alternative to current ERT, levormeloxifene was being developed for the prevention and treatment of postmenopausal osteoporosis.

In an oestrogen-depleted rat model of human osteoporosis an oral dose of 0.5 mg kg⁻¹ of levormeloxifene given three times weekly for 5 weeks was sufficient to maintain bone densities at the level of sham-operated animals [7] and the average steady state plasma concentration of levormeloxifene corresponding to an effective dose was approximately 20–25 ng ml⁻¹. The uteri from ovariectomized (OVX) animals treated with levormeloxifene showed no evidence of epithelial proliferation or glandular stimulation [8–10]. In addition, levormeloxifene was shown to reduce the serum cholesterol in OVX rats [11] and to prevent aortic cholesterol accumulation in the OVX rabbit model [9]. In summary, the preclinical data suggest that levormeloxifene has oestrogenic effects on the skeleton and the cardiovascular system without inducing endometrial hyperplasia.

The objectives of the first two human studies of levormeloxifene were to investigate the tolerability, safety and pharmacokinetics of levormeloxifene and to confirm in man the promising pharmacological profile observed in preclinical studies.

Methods

Prior to the study the clinical trial protocols and written informed consent forms were approved by independent ethics committees of Inveresk Research, Edinburgh, Phase I (Clinical Trials Unit) Limited, Plymouth, Drug Development (Scotland) Limited, Dundee, and Simbec Research Limited, Methyr Tydfil all in the United Kingdom (first human single dose trial) and in the Netherlands of the Stichting Beoordeling Ethiek Bio-Medisch Onderzoek (multiple dose trial). Before entering the trials, the nature of the studies was explained to each participant and written informed consent was obtained. The trials were conducted in accordance with the amended Helsinki Declaration of 1989 [12] and Good Clinical Practice (GCP) described by the Commission of the European Communities in 1990 [13].

All methodological details apply to both trials unless otherwise stated. The two trials were of a randomized, double-blind, placebo controlled, parallel group design. Levormeloxifene was administered orally as levormeloxifene fumarate. Doses were stated in mg of the base. In total, 104 healthy postmenopausal women aged 47–66 years, all having at least a 1 year history of physiological or surgical amenorrhoea and all not receiving hormone replacement treatment (HRT) for the last 3 months before drug administration, participated in the trials. The 17-β-oestradiol (E₂) plasma concentration was below 110 nmol ml⁻¹ and the follicle stimulating hormone (FSH) measured in trial 2 only was above 40 IU l⁻¹. For all subjects body weights were within 20% of ideal weight (Metropolitan Life tables) and ranged from 42 to 90 kg. All subjects were healthy based on a pre-study physical examination, routine haematology and clinical chemistry, urinalysis, screening for drugs of abuse, tests of hepatitis B and C, and HIV and pregnancy test. The dose administration was managed by a nurse, the investigator or coinvestigators.

Trial 1, single dose study

Single ascending doses of 2.5, 10, 30, 80, 160 and 320 mg levormeloxifene were administered to six groups of eight subjects (six active, two placebo). Levormeloxifene was given as a solution of 1 mg ml⁻¹ and the placebo as a solution of 40 mg l⁻¹ quinine sulphate. Hydroxypropyl-β-cyclodextrin was added to both as a solubiliser. The subjects were fasted from 8 h before until 4 h after dose administration. The trial was conducted by Inveresk Clinical Research Limited, Edinburgh in cooperation with three other Clinical Research Organizations in the United Kingdom.

Trial 2, multiple dose study

The trial consisted of two parts: Part 1 with multiple dosing for 35 days, and Part 2 with multiple dosing for 56 days. The objectives of Part 1 were to assess the safety profile, tolerability and pharmacokinetics after ascending oral multiple dosing at three dose levels of levormeloxifene in three groups. The objectives of Part 2 were to gain additional safety information in a larger group of women for a longer period of dosing at two dose levels of levormeloxifene. A secondary objective of both parts of the trial was to investigate any effect of levormeloxifene on a metabolic marker of bone resorption, urinary collagen I C-terminal telopeptide fragments (CrossLaps™) corrected for creatinine excretion. In Part 1, three groups of eight subjects (six active, two placebo) received, respectively, ascending doses of 20, 80 and 160 mg levormeloxifene once daily for 35 days. Based on the results of Part 1, two groups of 16 subjects (12 active, four placebo) were dosed daily for 56 days with 40 and 80 mg levormeloxifene, respectively, in Part 2. The duration of the trial was 69 days for all subjects. The trial products were given as tablets in the morning after an overnight fast and fasting was continued until 0.5 h after dosing.

The trial products were tablets of 10 mg for a dose of 20 mg levormeloxifene, tablets of 40 mg at the other doses and corresponding placebo tablets of similar size and colour.

The trial was performed by Pharma Bio-Research, Zuidlaren, the Netherlands.

Safety assessment and pharmacodynamic evaluation

Vital signs [blood pressure (BP), heart rate (HR), body temperature], ECG (in trial 1 including continuous ECG monitoring 0–4 h after dosing), routine haematology and clinical chemistry, urinalysis, plasma concentrations of E₂, follicle stimulating hormone (FSH), luteinizing hormone (LH, in trial 2 only) and adverse events were evaluated before and at regular intervals during the trials. In trial 2 the concentrations of a marker of metabolic bone turnover, [urinary collagen I C-terminal telopeptide corrected for creatinine (CrossLaps™, Osteometer Biotech, Herlev, Denmark)], was determined by Hôpital Edouard Herriot, Lyons, France before commencement of dosing, after the last dose and at the final visit. For subjects reporting bleeding, an ultrasonic measurement of the endometrium was performed, and for subjects with an endometrium above 8 mm, a biopsy for histopathological examination was removed to establish a precise diagnosis.

Pharmacokinetics

In trial 1, blood samples were collected before and at intervals up to 648 h (27 days) after dosing (18 samples in total). In trial 2, blood sampling was performed before the first dose and at intervals up to 24 h after the first dose and weekly thereafter (trough sampling) and again after the final dose at intervals up to 34 or 13 days after last dose. In total 31 and 25 samples were taken in Part 1 and Part 2, respectively. Plasma was separated by centrifugation and assayed for levormeloxifene and 7-desmethyllevormeloxifene, the main metabolite of levormeloxifene, by a validated reversed phase high pressure liquid chromatography assay using solid phase extraction [14]. The lower limits of quantification were 1.5 ng ml⁻¹ and 2.5 ng ml⁻¹ for levormeloxifene and 7-desmethyllevormeloxifene, respectively. Inter- and intra-assay coefficients of variation were < 8% for both compounds at all concentrations except at the limit of quantification where they were 15%. Inter- and intra-assay accuracy calculated as a percentage of the nominal value were between 90 and 114% for both compounds.

Pharmacokinetic data were analysed by noncompartmental techniques using the WinNonlin 1.1 software (Scientific Consulting Inc. Apex, NC, USA). For peak concentration in plasma, C_max, and time to reach peak concentration, t_max, observed values were taken. The apparent elimination rate constant, λ_z, was estimated by log-linear regression analysis on the terminal phase of the plasma concentration vs time curve. The terminal phase was assessed subjectively over at least the final three sampling points with a measured concentration equal to or above the lower limit of quantification. The terminal half-life, t_1/2, was calculated as (ln2)/λ_z. After single doses, the total area under the plasma concentration vs time curve (AUC) was determined by the linear trapezoidal rule from time zero to the last sampling point equal or above the lower limit of quantification, AUC(0, t), plus the residual area as estimated by log-linear extrapolation to infinity. After multiple dosing, the area during one dosing interval (τ: 24 h) was calculated after the first dose, AUC(1, τ), and at steady state AUC_ss(0, τ) using the linear trapezoidal rule. Trough plasma concentrations on the approach to and at steady state were determined at weekly intervals. The observed degree of accumulation as calculated by

on days 35 or 56 was compared with the anticipated accumulation deduced from the equation

The degree of fluctuation of plasma concentrations at steady state was estimated by the fluctuation index

Statistical analysis

The statistical evaluations were performed using the statistical package SAS (v 6.07 or SAS for Windows). The level of significance of all tests was 0.05.

For all endpoints, i.e. vital signs, ECG, FSH, LH, E₂, haematology, clinical chemistry, urinary collagen I C-terminal telopeptide concentration (corrected for creatinine excretion and logarithmic transformed), changes from baseline to each time point separately were analysed using analysis of variance (anova) techniques including terms for dose group and dose level. In addition, a repeated measurement analysis of variance was performed at the given time points. The factors in this model included dose group, dose level, subject within dose level and dose group, time point, dose level × time point interaction and dose group × time point interaction. Furthermore, the overall percent reduction in urinary collagen I C-terminal telopeptide concentration for active treatment compared with placebo was calculated as 100(Y_T– Y_P)/(100 – Y_P), where Y_T is percent reduction for active treatment and Y_P is percent reduction for placebo, with 95% confidence intervals.

The statistical analysis of the pharmacokinetic results in both trials was performed using analysis of variance testing differences between dose levels. For the assessment of dose relationship, linear regression analysis was performed for ln-transformed values of dose adjusted C_max, AUC (trial 1 only), C_ss,min, AUC(1, τ) and AUC_ss(0, τ) (trial 2 Part 1 only) on ln(dose). In trial 2 Part 2 with two dose levels only, dose linearity was not assessed.

Results

Tolerability and adverse events

Trial 1

Forty-eight subjects were randomized and completed the single dose trial. Levormeloxifene was tolerated well after all single doses studied. There was one serious event with four separate symptoms. Five minutes after dosing (10 mg) the subject developed tachycardia (maximum 131 beats min⁻¹), distal parasthesia, bodily discomfort and increased blood pressure (maximum 180/100 mmHg). All events resolved in 30–60 min, but as a similar pattern of events, of milder degree, continued to happen on a daily basis for approximately 1 h each morning, the investigator concluded that they were unlikely to be related to the trial medication. All other adverse events were generally mild to moderate in severity and resolved completely. No maximum tolerated dose was identified, and no differences between the adverse event profiles, in frequency or type of event, after levormeloxifene or placebo were observed. There was no evidence of an increase in the frequency of adverse events with increasing dose (Table 1).

Table 1.

Number of subjects with adverse events (in > 10% of subjects) at ascending single doses of levormeloxifene or placebo.

	Single dose (Number of subjects per dose group)
	Placebo (n = 12)	2.5 mg (n = 6)	10 mg (n = 6)	30 mg (n = 6)	80 mg (n = 6)	160 mg (n = 6)	320 mg (n = 6)
Headache	3	2	2	4		2	1
Hot flushes	2		2
Abdominal pain			1		1	1	1
Nausea	1		1			1	2
Myalgia				2	1		1
Rhinitis	3	2				2
Subjects with adverse events (%)	83	67	83	100	83	83	67
Number of adverse events/subject	1.8	0.8	6.6	3.7	1.5	1.8	2.0

Trial 2

In Part 1 of this trial, daily administration of the highest dose (160 mg levormeloxifene) resulted in significantly more adverse events per subject than after administration of 20 mg or 80 mg levormeloxifene. Consequently, lower daily doses of 40 mg and 80 mg levormeloxifene were selected for Part 2 of the study. The most frequent adverse events reported with daily dosing of levormeloxifene (20 mg to 160 mg) for up to 56 days were headache, abdominal pain, hot flushes and leukorrhea, the latter in five subjects after the highest dose (Table 2). One severe adverse event, syncope, was reported after treatment with placebo but none after the drug. One subject developed a painful breast tumour, which was histopathologically benign and was withdrawn after 25 days of 80 mg levormeloxifene. The event was considered unrelated to the drug treatment. A second subject was withdrawn after 13 days of placebo treatment as a result of a depressive disorder which was present before the first drug administration.

Table 2.

Number of adverse events per subject during multiple dosing for 35 or 56 days for those events with a frequency ≥0.5 in one dose group.

	Multiple dose (Number of subjects with advise event)
	Placebo	20 mg (n = 6)	40 mg (n = 12)	80 mg (n = 18)	160 mg (n = 6)
Headache	1.1 (7)	1.7 (5)	0.8 (5)	0.4 (4)	2.5 (5)
Abdominal pain	0.1 (2)	0.2 (1)	0.6 (6)	0.7 (6)	1.7 (3)
Pharyngitis	0.2 (3)	0.2 (2)	0.4 (5)	0.2 (4)	0.5 (3)
Fatigue	–	–	0.3 (4)	0.1 (2)	0.5 (2)
Nausea	0.1 (1)	–	0.3 (3)	0.2 (2)	0.5 (2)
Hot flushes	0.3 (2)	–	0.6 (5)	0.1 (2)	0.2 (1)
Leukorrhea	–	0.2 (1)	0.3 (4)	0.2 (3)	1.0 (5)
Myalgia	0.1 (1)	0.5 (3)	0.2 (1)	0.1 (1)	–
Number of adverse events per subject	4.0	5.0	7.4	4.3	11.8

In Part 1, 71 (50%) of a total of 142 events were reported during treatment with 160 mg levormeloxifene compared with 17%, 21%, and 12% during treatment with placebo, 20 mg levormeloxifene and 80 mg levormeloxifene, respectively. Most of the adverse events (83%) were mild. The most frequently reported drug-related adverse events after 160 mg levormeloxifene treatment were headache (four subjects), leukorrhea (five subjects), and abdominal pain (three subjects). In Part 2, most adverse events (92%) were mild, 57% were considered to be possibly or probably related to the trial product, and most of them occurred during active treatment (15%, 46% and 39% after placebo, 40 mg levormeloxifene and 80 mg levormeloxifene, respectively). Abdominal pain was the most frequently observed drug-related adverse event, reported by 46% of subjects taking levormeloxifene compared with 13% of the placebo group.

After withdrawal of levormeloxifene, five subjects reported haemorrhagic vaginal discharge (one subject in each of 20, 40 and 80 mg groups, and two subjects in the 160 mg group), starting 12–33 days after withdrawal of levormeloxifene (80–160 mg) and lasting from 3 h up to 2 weeks. An ultrasonic measurement of the endometrium found a thickened endometrial stripe (> 8 mm) in four of the five subjects (9–15 mm). Histopathological examinations of the endometrium showed a normal endometrium without hyperplasia or atypia.

Pharmacodynamics

Effect on metabolic bone marker, alkaline phosphatase and lipids

The effect on urinary excretion of collagen I C-telopeptide, total alkaline phosphatase, triglycerides, total cholesterol, LDL-cholesterol, HDL-cholesterol, and on the hormones FSH and LH after 35 or 56 days of treatment are given in Tables 3 and 4. Treatment with levormeloxifene resulted in a statistically significant reduction in urinary collagen I C-terminal telopeptide excretion compared with placebo (Table 3). After 35 days of dosing, there was no statistically significant difference in the reduction of collagen I C-terminal telopeptide concentrations between the three active groups and placebo. However, with more subjects per dose group and a treatment period of 56 days, the difference between active groups and placebo became statistically significant. The reduction in the concentration of this biochemical bone marker was 44% in the two active groups compared with 13% in placebo (P = 0.015). The overall percent reduction in collagen I C-terminal telopeptide for active treatment compared to placebo were 44.4% [95% CI: 11.3, 65.1] in Part 1 and 35.5% [95% CI: 14.0, 51.6] in Part 2, calculated from the expression 100(Y_T−Y_P)/(100−Y_P), where Y_T is percent reduction after active treatment and Y_P is percent reduction after placebo. No dose-dependent decrease was observed over the dose range studied. The collagen I C-terminal telopeptide concentrations remained below baseline values after 14 or 35 days of withdrawal (data not shown).

Table 3.

Effect of levormeloxifene on collagen I C-telopeptide, total alkaline phosphatase, triglycerides, total cholesterol, LDL-cholesterol, HDL-cholesterol, FSH, and LH after 35 or 56 days of treatment, expressed as % decrease from baseline (mean±s.e. mean).

	Reduction from baseline in percentage
	Day 35 (Part 1)				Day 56 (Part 2)
	Placebo (n = 6)	20 mg (n = 6)	80 mg (n = 6)	160 mg (n = 6)	Placebo (n = 8)	40 mg (n = 12)	80 mg (n = 12)
C-telopeptide*	17 ± 6.2	56 ± 12	48 ± 10	56 ± 5.0	13 ± 5.1	44 ± 7.2	44 ± 5.3
Alkaline phosphatase**	−0.9 ± 4.3	8.3 ± 2.7	11 ± 6.1	25 ± 2.3	4.5 ± 1.7	31 ± 2.6	29 ± 3.5
Triglycerides***	0.3 ± 14	−15 ± 10	−19 ± 7.0	2.1 ± 13	−10 ± 7.9	2.1 ± 7.1	−3.5 ± 5.4
Total cholesterol**	0.6 ± 1.7	20 ± 3.4	24 ± 4.7	25 ± 3.9	−3.7 ± 3.4	23 ± 3.9	23 ± 2.5
LDL-cholesterol**	0.8 ± 5.0	30 ± 6.0	35 ± 5.8	32 ± 4.5	−4.0 ± 4.6	32 ± 4.9	32 ± 3.1
HDL-cholesterol***	−0.0 ± 3.0	−9.8 ± 5.3	−14 ± 3.7	−15 ± 3.3	−0.6 ± 3.9	11 ± 3.7	9.8 ± 5.0
FSH**	−1.5 ± 5.7	45 ± 6.8	34 ± 11	49 ± 4.6	6.0 ± 2.7	55 ± 4.7	54 ± 5.8
LH**	−6.8 ± 15	48 ± 7.4	42 ± 8.6	41 ± 9.7	−5.7 ± 4.7	62 ± 8.6	49 ± 6.7

^*statistical significant difference between active and placebo after 56 days of treatment, P = 0.015, F-test with 2 d.f., overall percentage reduction for active treatment compared with placebo: 44.4% [95% CI: 11.3, 65.1] in Part 1 and 35.5% [95% CI: 14.0, 51.6] in Part 2.

^**statistical significant difference between active and placebo after 35 days and after 56 days of treatment, P < 0.05.

^***not statistically significant after 35 days nor after 56 days of treatment, P > 0.05.

Table 4.

Mean of baseline and mean of absolute change in collagen I C-telopeptide, total alkaline phosphatase, triglycerides, total cholesterol, LDL-cholesterol, HDL-cholesterol, FSH, and LH after 35 or 56 days of treatment with levomeloxifene.

	Baseline values and absolute change from baseline
	Day 35 (Part 1)				Day 56 (Part 2)
	Placebo (n = 6)	20 mg (n = 6)	80 mg (n = 6)	160 mg (n = 6)	Placebo (n = 8)	40 mg (n = 12)	80 mg (n = 12)
C-telopeptide* (µg mmol−1)	365	250	379	313	333	441	321
	−25.1	−135	−191	−166	−50.6	−208	−139
Alkaline phosphatase** (U l−1)	66.3	66.3	61.8	85.5	62.4	78.2	72.3
	0.6	−5.2	−6.6	−21.7	−2.8	−24.3	−21.3
Triglycerides*** (mmol l−1)	1.33	1.17	0.99	1.53	1.09	1.54	1.19
	0.00	0.18	0.20	−0.03	0.11	−0.03	−0.04
Total cholesterol** (mmol l−1)	6.18	5.53	5.70	6.55	5.74	6.13	6.10
	−0.04	−1.12	−1.32	−1.62	0.21	−1.41	−1.37
LDL-cholesterol** (mmol l−1)	4.23	3.47	3.60	4.37	3.44	3.84	3.84
	−0.03	−1.05	−1.19	−1.39	0.14	−1.23	−1.23
HDL-cholesterol*** (mmol l−1)	1.35	1.53	1.65	1.48	1.81	1.59	1.73
	0.00	−0.15	−0.22	−0.22	0.01	−0.17	−0.17
FSH** (IU l−1)	66.4	59.3	67.7	72.0	63.7	77.2	85.0
	1.00	−26.7	−23.4	−35.6	−3.83	−42.3	−45.5
LH** (IU l−1)	25.9	27.3	35.2	23.6	24.2	39.3	30.9
	1.82	−13.2	−15.5	−9.78	0.64	−21.3	−12.5

^*statistical significant difference between active and placebo after 56 days of treatment, P < 0.015, F-test with 2 d.f.

^**statistical significant difference between active and placebo after 35 days and after 56 days of treatment, P < 0.05.

^***not statistically significant after 35 days nor after 56 days of treatment, P > 0.05.

Following treatment with 160 mg levormeloxifene for 5 weeks, and with 40 mg or 80 mg levormeloxifene for 8 weeks, alkaline phosphatase was significantly reduced by approximately 24–27% as compared with placebo (Table 3). The reductions in total cholesterol and LDL-cholesterol were approximately 19–25% and 28–35%, respectively, after active treatment compared with placebo (Figure 1 and Table 3). There were no statistically significant effects on triglycerides or HDL-cholesterol.

Figure 1.

Serum alkaline phosphatase, total cholesterol and LDL-cholesterol concentrations during 8 weeks of administration of levormeloxifene expressed as a percentage of baseline values (mean±s.e. mean). ▪ placebo, □ 40 mg, ^ 80 mg levormeloxifene.

Vital signs and ECG

No clinically significant alterations of supine and standing blood pressure, pulse or body temperature were found in either of the trials and no clinically relevant changes in the ECG or 4 h ECG monitoring were observed.

Haematology and clinical chemistry

No clinically significant `changes were detected in haematology or clinical chemistry measurements except the expected therapeutic effects on bone and lipid metabolism parameters as described earlier.

Hormones

Following 5 and 8 weeks treatment with levormeloxifene, the concentrations of FSH and LH were significantly reduced by approximately 50% from mean concentrations of FSH of 59–85 IU l⁻¹ and mean concentrations of LH of 24–39 IU l⁻¹ (P < 0.05) (Tables 3 and 4). No effect on plasma oestradiol concentrations was observed (data not shown).

Pharmacokinetics

Trial 1

The pharmacokinetics of levormeloxifene are reported in Table 5 and the corresponding plasma concentrations are shown for the drug and its major metabolite in Figure 2. Levormeloxifene was rapidly absorbed with peak plasma concentrations appearing 1–2 h after dosing in 28 of 35 subjects. The time to reach peak was independent of the dose (P = 0.56). The elimination of levormeloxifene was slow with half-lives ranging from 87 to 250 h (3.6–10.4 days) with an overall harmonic mean of 6.4 days. Differences in mean λ_z values between doses were statistically significant (P = 0.02), the t_1/2 decreasing with increasing doses (8.4 and 6.0 days after lowest and highest dose, respectively). AUC showed dose proportionality, but C_max increased nonproportionally (Figure 4, panel a and c). For the metabolite 7-desmethyllevormeloxifene t_max was 4–6 h. The decline in plasma concentrations of the metabolite was parallel with that of the parent compound (Figure 2) and consequently the range and mean values of the apparent half-life of elimination were very similar to the values for levormeloxifene (range of t_1/2 = 4.3–13.1 days, overall harmonic mean = 6.9 days). The 7-desmethyl-metabolite exhibited dose-proportional pharmacokinetics. The pharmacokinetic parameters for the metabolite are not shown.

Table 5.

Pharmacokinetic parameters for levormeloxifene after administration of single oral doses. Data are presented as mean (± s.d.). NA: insufficient data for estimation of AUC and t_1/2; Numbers in brackets represent the range. The final eight (or more) sampling points were used for the estimation of λ_z.

	Levormeloxifene dose
	2.5 mg	10 mg	30 mg	80 mg	160 mg	320 mg
Cmax (ng ml−1)	5.6 ± 1.3	27 ± 6.2	128 ± 42	443 ± 94	1115 ± 180	2737 ± 607
tmax (h)	2.50 ± 2.07	2.67 ± 1.03	2.50 ± 1.22	1.50 ± 0.55	2.00 ± 1.10	2.0 ± 0.02
t1/2* (days)	NA	8.4 (6.0–10.4)	6.9 (4.7–8.8)	5.8 (4.6–6.6)	5.4 (3.6–8.2)	6.0 (4.5–9.0)
AUC (µg ml−1 h)	NA	3.2 ± 0.6	10 ± 2.8	26 ± 3.1	64 ± 21	129 ± 35

^*harmonic mean+(min-max).

Figure 2.

Mean (s.d.) plasma concentration-time curves of levormeloxifene and its major metabolite after administration of ascending single oral doses (• 2.5 mg, ▪ 10 mg, ▴ 30 mg, ^ 80 mg, □ 160 mg, ▵ 320 mg, n = 6 subjects).

Figure 4.

Dose normalized values of C_max and AUC (trial 1) or AUC(1, τ) alternatively AUC_ss(0, τ) (trial 2) as a function of dose. The lines were generated by linear regression.

Trial 2

Steady state concentrations of levormeloxifene and 7-desmethyllevormeloxifene were achieved approximately 28 days after commencement of daily dosing (Figure 3). As in the single dose trial, the absorption of levormeloxifene was rapid (mean t_max: 1.5–5.6 h) and the elimination slow (Table 6). The individual half-life of elimination after multiple dosing ranged from 77 to 201 h (3.2–8.4 days) with overall harmonic mean of 5.2 days. The elimination half-life was not dose-dependent and mean values per dose group ranged from 4.8 to 5.9 days. Plasma concentrations of the metabolite peaked 12–24 h after dosing and the elimination rate from plasma was almost identical with that of the parent compound. For both compounds the relative fluctuations around the steady-state concentrations were small (Figure 3), with a fluctuation index of no more than 0.9 for any subject. The AUC(1, τ), AUC_ss(0, τ), C_ss, min and C_max values on days 1, 35 and 56, respectively, were proportional to the dose (Figure 4, panel b and d-C_ss, min data not shown). Other pharmacokinetic parameters calculated in this trial were independent of the dose (t_max: Part 1: P = 0.69; Part 2: P = 0.94-λ_z; Part 1: P = 0.70; Part 2: P = 0.53). The actual accumulation in plasma concentrations over time was smaller than that predicted at all doses (Table 6). The significance of this finding is somewhat unclear as the half-life of elimination at the 80 mg and 160 mg doses remained constant and independent of the dosing period.

Figure 3.

Mean (s.d.) plasma concentration-time curves of levormeloxifene and its major metabolite after administration of ascending multiple oral doses for 35 days (Part 1, n = 6 subjects, •, ^ 20 mg, ▪, □ 80 mg, ▴, ▵ 160 mg) and 56 days (Part 2, n = 12 subjects, •, ^ 40 mg, ▪, □ 80 mg). Filled symbols: levormeloxifene; unfilled symbols: 7-desmethyllevormeloxifene.

Table 6.

Pharmacokinetic parameters for levormeloxifene after administration of multiple oral doses. Data are presented as mean (± s.d.). NA: insufficient data for estimation of t_1/2; Numbers in brackets represent the range. After 35 and 56 days of dosing at least the final nine and three (or more) sampling points were used for the estimation of λ_z, respectively.

		Levormeloxifene dose
	30 days of dosing (part 1)			56 days of dosing (part 2)
		20 mg	80 mg	160 mg	40 mg	80 mg
Cmax (ng ml−1)	1st dose	71 ± 10	387 ± 98	769 ± 256	182 ± 48	377 ± 58
	last dose	254 ± 74	1081 ± 328	2031 ± 618	477 ± 146	879 ± 229
tmax (h)	1st dose	3.00 ± 1.55	1.85 ± 1.16	1.83 ± 1.17	2.42 ± 2.23	2.00 ± 1.53
	last dose	3.17 ± 3.06	5.60** ± 10.29	2.51 ± 2.07	1.50 ± 1.24	2.01 ± 1.54
t1/2* (days)	1st dose	NA	NA	NA	NA	NA
	last dose	5.9 (5.0–7.0)	5.4 (4.6–7.0)	5.5 (4.3–6.3)	5.1 (3.6–8.4)	4.8 (3.2–7.9)
AUC(0, τ) (µg ml−1 h)	1st dose	1.0 ± 0.2	4.8 ± 0.9	9.0 ± 2.4	2.4 ± 0.4	4.9 ± 0.8
	last dose	4.7 ± 1.4	19.8 ± 8.5	33.4 ± 10.5	8.5 ± 2.6	15.2 ± 4.2
RA	actual	4.7 ± 0.9	3.9 ± 1.2	3.7 ± 0.4	3.5 ± 0.9	3.1 ± 0.7
	theoretical	9.1 ± 1.0	8.6 ± 1.8	8.6 ± 1.0	8.3 ± 2.0	7.9 ± 2.0

^*harmonic mean+(min-max).

^**t_max value is 5.60 h due to a single outlier of 24 h.

Discussion

Levormeloxifene was administered to postmenopausal women as single doses of up to 320 mg and as multiple doses of up to 160 mg once daily for 35 or 56 days. The multiple dose study included investigation of the pharmacodynamic properties of levormeloxifene.

Levormeloxifene was tolerated well after all single and multiple doses and no maximum tolerated dose was achieved. In the single dose study there were no differences between the adverse event profiles after levormeloxifene and placebo, in frequency or type, even after the highest dose of 320 mg. However, after multiple dosing the highest dose group receiving 160 mg experienced almost double the number of adverse events compared with those receiving lower doses. The most frequent adverse event was headache. Subjects in the highest dose group reported this event approximately 2–3 times during the treatment period compared with a frequency of 0.4–0.7 in the other dose groups. Besides leukorrhea, adverse events were not dose related. Vaginal bleeding after withdrawal was observed in one or two subjects in all dose groups. The observed increased endometrial stripe determined by ultrasound is compatible with previous reports of increased volume of stroma in the endometrium of levormeloxifene treated rats accompanied by inactive endometrial epithelium and atrophic endometrial glands [7–10]. In subsequent clinical phase II studies in the daily dose range of 1.25–40 mg levormeloxifene, increase of the endometrial stripe was found at most dose levels. However no proliferative changes or hyperplasia were observed. After initiation of the clinical phase III studies, the clinical development of levormeloxifene was interrupted due to other adverse events, such as incontinence and uterine prolapse. Such events were not reported in the present studies of short duration.

Urinary collagen I-C-telopeptide is liberated from bone during bone resorption and correlates to bone resorptive activity [15]. In the present studies after 8 weeks of levormeloxifene treatment the urinary collagen I-C-telopeptide was reduced by 25–28% compared with placebo. A similar trend was also observed after 5 weeks of treatment at the lowest dose of 20 mg levormeloxifene, but the difference from placebo did not reach statistical significance. The effect of levormeloxifene on bone metabolism was also supported by a significant reduction after treatment for 5 and 8 weeks in the nonspecific marker of bone formation, total serum alkaline phosphatase, which is partly produced by osteoblasts but also is synthesised in liver, kidneys and gut [16]. However, the more specific assay [17] for the skeletal variant was not used. These observed effects of levormeloxifene are both considered indicative of a bone preserving action [16]. The effect of levormeloxifene on collagen I-C telopeptide is similar to those of other antiresorptive drugs like HRT [18], raloxifene [19], and alendronate [20]. For a direct comparison of efficacy between levormeloxifene and other antiresorptive treatment modalities, a longer term treatment with lower doses of levormeloxifene is required. Also the lowering effect on serum total cholesterol and LDL-cholesterol without affecting the concentrations of triglycerides and HDL-cholesterol exceeds what has been reported for HRT [21] or other SERMs [19], and is comparable with data on lipid lowering statins [22]. These changes in the lipid profile resulting from treatment with levormeloxifene may have a beneficial effect on the incidence of cardiovascular diseases [21].

The plasma concentrations of LH and FSH were reduced following multiple dosing with levormeloxifene. The effect was similar to that of oestrogens [23], and by thus this may be indicative of an oestrogen agonist-like action on the hypothalmo-pituitary axis.

Similar pharmacokinetic data were obtained in both trials. Plasma concentration curves indicated a rapid absorption of levormeloxifene with peak concentrations detectable within the first 4 h after dosing in almost all subjects. Levormeloxifene and its metabolite 7-desmethyllevormeloxifene, the latter with no in-vitro binding properties to the oestrogen receptor [S Bain, private communication] were eliminated slowly from the plasma. The half-life of levormeloxifene in the single dose trial decreased with increasing doses, but remained constant and independent of dose and treatment duration in the multiple dose study. The mean t_1/2 was highest at 8.4 and 6.9 days after the lowest single doses (10 and 30 mg, respectively), whereas for the remaining eight dose groups t_1/2 was in the narrow range of 4.8–6.0 days. With an increase in the elimination constant at higher doses in the single dose study, we would also expect to find a correspondingly lower exposure. However the statistical analysis did not indicate any deviation from dose proportionality of drug exposure in the single or multiple study up to 160 mg, as measured by AUC, AUC_ss(0, τ), C_ss,min and AUC(1, τ). On the other hand C_max data showed a greater increase than expected at higher doses, however, suggesting a change in distribution with dose.

Steady state plasma concentrations of levormeloxifene were achieved approximately 4 weeks after commencement of daily dosing. The relative fluctuations of plasma concentrations around the plateau were small and C_ss, max values in all subjects were no more than double the value of the corresponding trough concentration. Therefore, in future clinical trials, once daily dosing in contrast to once weekly dosing, should be considered more favourable from a safety point of view. The latter would give rise to pronounced fluctuations in plasma concentration at steady state with the risk causing adverse events immediately after dosing. Following 35 days of dosing with 20 mg of drug, the average steady state concentration of levormeloxifene in subjects dosed was 230 ng ml⁻¹. Using oestrogen-depleted OVX rats, a dose of 0.5 mg kg⁻¹ of levormeloxifene given three times weekly p.o. for 5 weeks (11) corresponding to an average steady state plasma concentration (C_ss, av) of levormeloxifene of approximately 20–25 ng ml⁻¹, was sufficient to maintain bone densities at the sham level. Assuming that the antiresorptive effect is correlated to steady-state concentrations of levormeloxifene and is species independent, an effective human daily dose should be about 2 mg, which is 10 times less than the lowest dose used in this multiple dose study.

In conclusion, the studies provide the first data on the short-term administration of levormeloxifene to postmenopausal women and suggest that the drug has oestrogen-like bone preserving and antiatherogenic effects.