Journal of Bone and Mineral Metabolism

, Volume 27, Issue 1, pp 66–75

Randomized controlled study on the prevention of osteoporotic fractures (OF study): a phase IV clinical study of 15-mg menatetrenone capsules

Authors

  • Tetsuo Inoue
    • Aoyama General Hospital
    • Department of Data ScienceInstitute of Statistical Mathematics
  • Hideaki Kishimoto
    • Department of OrthopedicsSan-in Rosai Hospital
  • Toshitaka Makino
    • Japan Cosmetic Industry Association
  • Tetsuro Nakamura
    • Institute for Medical Science of Aging
  • Toshitaka Nakamura
    • Department of Orthopedic SurgeryUniversity of Occupational and Environmental Health
  • Tosiya Sato
    • Department of BiostatisticsKyoto University School of Public Health
  • Kaoru Yamazaki
    • Department of Orthopaedic SurgeryHamamatsu University School of Medicine
Original article

DOI: 10.1007/s00774-008-0008-8

Cite this article as:
Inoue, T., Fujita, T., Kishimoto, H. et al. J Bone Miner Metab (2009) 27: 66. doi:10.1007/s00774-008-0008-8

Abstract

An open-label study with blinded evaluation was performed to compare the preventive effect of a calcium supplement alone (monotherapy) or calcium supplement plus menatetrenone (combined therapy) on fracture in osteoporotic postmenopausal women aged 50 years or older. Patients were randomized to receive monotherapy (n = 2,193) or combined therapy (n = 2,185). Before randomization, the subjects were stratified into a subgroup without vertebral fractures (n = 2,986; no-fracture subgroup) and a subgroup with at least one vertebral fracture (n = 1,392; fracture subgroup). The incidence rate of new vertebral fractures during 36 months of treatment (primary endpoint) did not differ significantly between either subgroup of the two treatment groups. Although the cumulative 48-month incidence rate of new clinical fractures (secondary endpoint) was lower in the combined therapy group, the difference was not significant. There was a lower risk of new vertebral fractures in patients with at least five baseline fractures who received combined therapy. Also, the loss of height was less with combined therapy than with monotherapy among patients 75 years of age or older at enrollment, those whose last menstrual period occurred 30 years or more before enrollment, and those with at least five vertebral fractures at enrollment. Adverse events and adverse reactions were more frequent in the combined therapy group. In conclusion, menatetrenone therapy was not effective for preventing vertebral fractures in the full analysis set of this study, but the results suggested that it may prevent vertebral fractures in patients with more advanced osteoporosis.

Keywords

OsteoporosisFracturesMenatetrenoneControlled studyADL

Introduction

One of the major goals of treatment for osteoporosis is to prevent vertebral and clinical fractures, which have a great impact on the daily life of affected persons.

Several decades ago, Bouckaert et al. [1] found that vitamin K promotes fracture healing. Subsequent studies led to the discovery of γ-carboxyglutamic acid (Gla), an amino acid that binds to ionic calcium with a high affinity assisted by vitamin K [2], and further studies identified and characterized the Gla-containing protein osteocalcin in the bones [3, 4]. Other studies demonstrated that vitamin K has an important role in regulating bone turnover, including bone formation [5, 6]. A reduced circulating vitamin K level has been reported in patients with osteoporotic fractures [79], as well as a raised circulating level of noncalcium-binding uncarboxylated osteocalcin (ucOC) in elderly women with femoral neck fractures [10], indicating that vitamin K has an important role in bone turnover and even contributes to bone strength. In a clinical study of Japanese patients with osteoporosis, vitamin K2 (menatetrenone) prevented further reduction of bone mineral density (BMD) and relieved osteoporotic pain [11].

When the present study was designed in 1995, there was insufficient evidence about the preventive effect of therapy for osteoporosis on vertebral fractures in Japanese patients. Reliable data were also lacking about the incidence rate of clinical fractures in patients who had osteoporosis. These circumstances indicated the need to obtain the aforementioned epidemiological data. Accordingly, the present study was designed to determine whether menatetrenone therapy could prevent vertebral fractures in postmenopausal Japanese women with osteoporosis aged 50 years or older and to obtain an estimate of the incidence rate of clinical fractures in this population.

Methods

Subjects

The subjects of this study were postmenopausal women aged 50 years or older who had primary osteoporosis according to the Diagnostic Criteria for Primary Osteoporosis (published in 1995 by the Osteoporosis Diagnostic Criteria Committee of the Japanese Society for Bone and Mineral Research). Both inpatients and outpatients were eligible. According to the aforementioned criteria, a diagnosis of osteoporosis was made on the basis of radiographic evidence of a nontraumatic vertebral fracture together with either grade I or higher bone atrophy or a lumbar BMD T-score equal to or lower than −1.5 SD of the young adult mean (YAM). In the absence of radiographic evidence of vertebral fractures, either grade II or higher bone atrophy or a lumbar BMD T-score equal to or less than −2.5 SD of the YAM was considered to be diagnostic of osteoporosis. Bone atrophy was graded on radiographs of the spine as follows: grade 0 meant no changes of the vertebral trabecular pattern; grade I meant a decrease of transverse trabeculae (corresponding to early osteoporosis); grade II meant a decrease of longitudinal as well as transverse trabeculae; and grade III meant further trabecular loss or obscure longitudinal trabeculae [12].

In this study, osteoporosis was assessed from plain X-ray films of the spine and lumbar BMD data submitted at the time of provisional enrollment. Patients with confirmed osteoporosis were divided into two subgroups, depending on whether they had at least one vertebral fracture (the fracture subgroup) or no vertebral fractures (the no-fracture subgroup) at enrollment.

Major exclusion criteria included current treatment with warfarin potassium (menatetrenone is contraindicated in patients receiving warfarin therapy), hypercalcemia, any condition that could interfere with the identification of vertebral fractures, previous treatment with a bisphosphonate, and any condition that might cause poor absorption of hydrophobic drugs (e.g., biliary tract obstruction).

Study design and treatments

This was a multicenter, randomized, nonblinded, parallel group controlled study that compared patients receiving a calcium supplement alone (monotherapy group) with patients receiving a calcium supplement plus menatetrenone (combined therapy group). Before randomization, the subjects were also stratified into subgroups with and without vertebral fractures.

The calcium supplement was oral calcium l-aspartate (1.2 g/day) or dibasic calcium phosphate (3 g/day). Dose modification was allowed according to the patient’s age and response. Menatetrenone was administered orally at 45 mg/day (divided into three postprandial doses of 15 mg each). Patients who were using antiosteoporotic drugs other than calcium at the time of screening could be enrolled after a washout period of at least 8 weeks. All the subjects were scheduled to be treated for 3 years and then followed up for a further 1 year. During the additional follow-up period, no restrictions were placed on their medical or nonmedical treatment for osteoporosis (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-008-0008-8/MediaObjects/774_2008_8_Fig1_HTML.gif
Fig. 1

Study design. ADL activities of daily living

Information on the incidence of clinical fracture was collected at 1 year after the completion of protocol treatment to increase the precision of statistical analysis, because it was anticipated beforehand that the incidence of clinical fractures would not be high enough during 3 years of follow-up. It was planned to collect data on the incidence of clinical fracture over a 4-year period, rather than to compare the efficacy of the two treatments.

The effect of protocol therapy did not subside immediately after the completion of treatment, so comparison of the two protocols was still considered to be valid at 1 year after the end of therapy because the effect was thought to fade gradually.

Assessments

At the time of enrollment, the clinical profile of each subject was investigated with regard to the following: their initials, medical record number, inpatient/outpatient status, date of birth, occupation, diagnosis, time after menopause, number of deliveries, history of ovariectomy, concomitant illnesses, previous illnesses, and previous treatment for osteoporosis.

The primary efficacy variable was the incidence rate of new vertebral fractures after entry into the study, and the secondary efficacy variable was the incidence rate of new clinical fractures. Other efficacy variables were fracture-related changes in the activities of daily living (ADL) and changes of height and BMD.

New vertebral fractures that occurred between the fourth thoracic vertebra (T4) and fourth lumbar vertebra (L4) were counted unless there was a history of severe trauma that could have caused nonosteoporotic fracture in a young adult. To identify new vertebral fractures, radiography was performed at 3, 12, 24, and 36 months after the start of treatment, as well as whenever a new fracture was suspected. According to the criteria proposed by the Longevity/Osteoporosis Study Group sponsored by the Ministry of Health and Welfare, a vertebral fracture was defined as being present when the central height of a vertebral body was less than 80% of its anterior or posterior height or when the anterior height was less than 75% of its posterior height. If a vertebral body showed a decrease of its anterior, central, and posterior height by 20% or more (≥20%) compared with the vertebral body above or below, this was also interpreted as a vertebral fracture. Even in the absence of such vertebral deformities, a diagnosis of vertebral fracture could also be made on the basis of radiographic evidence of bone cortex discontinuity in a patient with a clinically new fracture. A new vertebral fracture was identified as one that was detected after at least 3 months of follow-up in a vertebral body having no radiographic evidence of fracture at baseline or as a 20% or more reduction of the height of a vertebral body with radiographic evidence of a fracture at baseline. Subjects were examined for fractures by the same member of the bone evaluation committee throughout the study period.

Clinical fractures included those of the humerus, femur, and radius, as well as vertebral fractures that were associated with trauma severe enough to have caused a nonosteoporotic fracture in a young adult. To identify clinical fractures, radiography was performed whenever a new fracture was suspected after the start of treatment.

Subjects who developed new vertebral or clinical fractures discontinued the study treatment, but remained in the study for further follow-up. No restrictions were placed on their therapy for osteoporosis after the study treatment was discontinued.

BMD was determined by microdensitometry or dual-energy X-ray absorptiometry at the start of treatment, as well as 12, 24, and 36 months afterward. Microdensitometry measured the BMD and cortical width at the middle of the second metacarpal with reference to the density of an aluminum step wedge on the same X-ray film [13].

Assessment of fractures and BMD was done by the bone evaluation committee using blinded radiographs and BMD charts to avoid evaluator bias and to standardize the assessment process. For monitoring of safety, clinical adverse events and adverse drug reactions were investigated, and routine laboratory tests were performed.

Subject registration and randomization

Subjects were registered and randomized by a central organization. After obtaining informed consent to participation in the study from a candidate subject, the investigator sent vertebral radiographs and lumbar BMD data to the data center. At the data center, the eligibility of each candidate subject was assessed by evaluating the submitted information, and subjects whose eligibility was confirmed were randomized to the monotherapy or combined therapy groups after stratification by the presence or absence of baseline vertebral fractures. To ensure the objectivity of data obtained during this nonblinded study, the assigned treatment was only disclosed to the ethical/data monitoring committee, which was responsible for ensuring the safety of the study. Monitoring of the study and reporting on its progress was delegated to a contract research organization, and the sponsor and all members of the committees involved in the study (other than the ethical/data monitoring committee) were blinded to the treatments assigned to the individual subjects.

Subjects were recruited during the period from April 1996 to March 2001.

Statistical analysis

The target sample size was a total of 4,000 patients (2,000 per treatment group), including 1,400 in the no-fracture subgroup and 600 in the fracture subgroup receiving each treatment. This sample size was calculated from the results of a phase III controlled study of menatetrenone [11] and was expected to provide adequate statistical power (more than 80%) to demonstrate, at a two-tailed alpha level of 10%, a predicted relative risk reduction of 25% (fracture subgroup) or 20% (no-fracture subgroup) by addition of menatetrenone therapy relative to the estimated incidence of new vertebral fractures in the respective calcium monotherapy subgroups (20 and 10 per 100 patient-years, respectively). This sample size also allowed for an anticipated withdrawal rate of 20% during the 3-year treatment period.

For analysis of the primary endpoint, the incidence rate of new vertebral fractures in each fracture subgroup and treatment group was calculated by the patient-year method [14], and incidence rates were compared by the Chi-square test. Secondary analyses included standard survival analysis of the time until the occurrence of any new clinical fracture. The stratified log–rank test was used for comparisons between treatment groups with the fracture subgroup as a stratification factor. In these analyses, the alpha level for significance was set at 10%.

This study aimed to assess the efficacy of add-on menatetrenone therapy by comparison with calcium alone. Because add-on therapy was unlikely to be inferior to monotherapy in terms of the prevention of fractures, a one-tailed hypothesis was considered to be appropriate for the present study, and the one-tailed level of significance was set at 5% (i.e., two-tailed significance of 10%).

The incidence rate of adverse events and adverse drug reactions was estimated by the patient-year method and was compared between the two treatment groups using the Chi-square test (the alpha level was unspecified in the study protocol).

Three interim analyses were performed, of the primary endpoint (new vertebral fractures) and of adverse events, after enrolling 2,000 patients, at the completion of enrollment, and at 18 months after the completion of enrollment.

Results

Disposition of the subjects

A total of 4,378 patients were enrolled, including 2,193 assigned to the monotherapy group and 2,185 assigned to the combined therapy group. Among them, 4,015 patients (2,016 in the monotherapy group and 1999 in the combined therapy group) were included in the efficacy evaluation set, and 4,031 patients (2,022 and 2,009, respectively) were included in the safety evaluation set. The primary reasons for excluding 363 patients from efficacy evaluation were (1) failure to start the study treatment and cancellation of enrollment and (2) detection of ineligibility during the early phase of treatment and discontinuation of further follow-up. The primary reasons for excluding 347 patients from safety evaluation were (1) failure to start the study treatment and cancellation of enrollment and (2) no evidence that the study treatment was actually started. Of the 4,378 patients randomized, 2,986 had no vertebral fractures at baseline (no-fracture subgroup) and 1,392 had at least one vertebral fracture at baseline (fracture subgroup) at enrollment (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-008-0008-8/MediaObjects/774_2008_8_Fig2_HTML.gif
Fig. 2

Disposition of subjects for efficacy and safety evaluation

Compliance

Of the 4,378 patients randomized, more than 90% in each group took the study treatment almost exactly as prescribed, only missing occasional doses throughout the study.

Patient characteristics

The baseline clinical profile of the patients evaluated for efficacy showed a difference between the two treatment groups with respect to the serum calcium level and pain during exercise in the no-fracture subgroup, whereas the severity and duration of spontaneous pain at rest differed in the fracture subgroup. The mean age was about 6 years older and the mean time after menopause was about 6 years longer in the fracture subgroup than in the no-fracture subgroup at baseline. The baseline clinical profile of the efficacy evaluation population is summarized in Table 1.
Table 1

Clinical profile of the patients evaluated for efficacy

Variable

No-fracture subgroup

Fracture subgroup

Monotherapy group

Combined therapy group

P value

Monotherapy group

Combined therapy group

P value

Age (years)

67.5 ± 8.61 (1,381)

67.6 ± 8.59 (1,372)

0.570

74.3 ± 7.89 (635)

74.0 ± 8.06 (627)

0.354

Time after menopause (years)

18.5 ± 9.74 (1,374)

18.5 ± 9.45 (1,368)

0.686

25.3 ± 8.90 (633)

25.0 ± 9.01 (620)

0.570

Height (cm)

150.79 ± 5.628 (1,358)

150.64 ± 5.589 (1,366)

0.330

146.79 ± 6.598 (617)

146.76 ± 6.911 (610)

0.876

BMI (kg/m2)

22.88 ± 3.309 (1,358)

22.76 ± 3.312 (1,336)

0.314

22.44 ± 3.407 (616)

22.57 ± 3.319 (610)

0.673

Serum calcium level (mg/dl)

9.18 ± 0.546 (1,270)

9.14 ± 0.489 (1,267)

0.045

9.12 ± 0.493 (590)

9.08 ± 0.503 (582)

 0.199

Datae are mean ± SD (no. of patients in parentheses) after excluding “missing or unknown” data

P values were obtained by analysis using the two-sample Wilcoxon test

Vertebral fractures

Table 2 shows the incidence rate of new vertebral fractures, which was the primary endpoint. In the no-fracture subgroup, the incidence rate of new vertebral fractures (per 100 patient-years) was similar for monotherapy and combined therapy (2.56 vs. 2.74; P = 0.666, χ2 test). In the fracture subgroup, the incidence rate of new vertebral fractures (per 100 patient-years) was also similar between treatments (13.76 vs. 13.80; P = 0.929, χ2 test. The overall incidence rate of new vertebral fractures (per 100 patient-years) was 5.74 in the monotherapy group compared with 5.87 in the combined therapy group.
Table 2

Incidence rate of new vertebral fractures (patients without a history of severe trauma)

 

No-fracture subgroup

Fracture subgroup

Monotherapy group

Combined therapy group

Monotherapy group

Combined therapy group

Patients evaluated for efficacy

1,381

1,372

635

627

Patients evaluated for vertebral fractures

1,122

1,117

516

502

Patient-years

2,779

2,770

1,105

1,094

Patients with new vertebral fractures

71

76

152

151

Incidence rate of new vertebral fractures

2.56

2.74

13.76

13.80

P value

0.666

0.929

The incidence rate of new vertebral fractures is expressed as the number of fractures per 100 patient-years

Patients evaluated for vertebral fractures were a subset of the efficacy evaluation population in whom radiographic studies were done

P values were derived from analysis of the patient-year data using the χ2 test

Clinical fractures

The Kaplan–Meier estimate of the cumulative incidence rate of clinical fractures over the entire 48-month follow-up period was 2.5% in the monotherapy group and 2.1% in the combined therapy group (P = 0.436, stratified log–rank test). The respective estimates for the no-fracture subgroups were 1.7 and 1.5% (P = 0.757, log–rank test), and those for the fracture subgroups were 4.4 and 3.4% (P = 0.431, log–rank test). For all subjects and for both subgroups, the cumulative incidence rate of clinical fractures was lower in the combined therapy group than in the monotherapy group, but the differences were not statistically significant (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-008-0008-8/MediaObjects/774_2008_8_Fig3_HTML.gif
Fig. 3

Cumulative incidence rate of clinical fractures. mos. Months

Other efficacy variables

Stratified analysis of vertebral fractures

The incidence rate of new vertebral fractures was subjected to analysis after stratification of the subjects by age, time after menopause, body mass index (BMI), and the number of vertebral fractures at enrollment. The subsets constructed were <65 years, ≥65 and <75 years, and ≥75 years for age; <15 years, ≥15 and <30 years, and ≥30 years for time after menopause; <18.5, ≥18.5 and <25, and ≥25 for BMI; and <5 and ≥5 for the number of vertebral fractures at enrollment.

Among patients with at least five vertebral fractures at enrollment, the incidence rate of new vertebral fractures (per number of subjects evaluated) was significantly lower in the combined therapy group than in the monotherapy group (20.31 vs. 33.16; P = 0.029, χ2 test on the basis of the patient-year method). None of the subsets examined showed a significantly lower incidence rate of vertebral fractures with monotherapy than with combined therapy.

Height

The decrease of height at each time was compared between the treatment groups using an alpha level of 10%. In the fracture subgroup, the decrease of height at 12 months was significantly smaller with combined therapy than with monotherapy (P = 0.059). A significantly smaller decrease of height with combined therapy than monotherapy was also noted in the following patient subsets: patients aged 75 years or more at enrollment from the entire study population (P = 0.009 at 3 months, P = 0.020 at 6 months, P = 0.001 at 12 months, and P = 0.086 at 36 months) and from the fracture subgroup (P = 0.011 at 3 months, P = 0.043 at 6 months, and P = 0.003 at 12 months); patients whose last menstrual period had occurred ≥15 and <30 years before enrollment from the no-fracture subgroup (P = 0.095 at 24 months); patients whose last menstrual period had occurred ≥ 30 years before enrollment from the entire study population (P = 0.036 at 3 months, P = 0.049 at 6 months, P = 0.001 at 12 months, and P = 0.009 at 36 months), the no-fracture subgroup (P = 0.100 at 6 months and P = 0.074 at 36 months), and the fracture subgroup (P = 0.081 at 3 months, P < 0.001 at 12 months, and P = 0.083 at 36 months); patients with a baseline BMI ≥ 25 from the entire population (P = 0.081 at 12 months) and from the fracture subgroup (P = 0.024 at 12 months); and patients with at least five vertebral fractures at enrollment (P = 0.071 at 3 months, P = 0.014 at 6 months, P = 0.034 at 12 months, and P = 0.022 at 24 months). None of the subsets examined showed a significantly smaller loss of height with monotherapy than with combined therapy.

Figure 4 shows the changes of height over time in patients aged 75 years or more at baseline, a subset of the subjects who were more likely to benefit from combined therapy.
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-008-0008-8/MediaObjects/774_2008_8_Fig4_HTML.gif
Fig. 4

Stratified analysis of height decrease in patients aged 75 years or older at baseline

ADL

A significantly (P < 0.10) higher rate of improvement of the following parameters was noted with combined therapy than with monotherapy: walking performance of the entire study population (P = 0.090 at 3 months and P = 0.073 at 12 months) and the fracture subgroup (P = 0.062 at 12 months); severity of spontaneous low back pain at rest in the entire population (P = 0.008 at 6 months) and the no-fracture subgroup (P = 0.016 at 6 months); and duration of spontaneous low back pain at rest in the entire population (P = 0.073 at 3 months and P = 0.004 at 6 months) and the no-fracture subgroup (P = 0.014 at 6 months).

Safety evaluation

The incidence of adverse events during the entire 48-month follow-up period was higher in the combined therapy group than in the monotherapy group (10.06 vs. 9.18; P = 0.178, χ2 test). The incidence of adverse drug reactions was also higher in the combined therapy group than in the monotherapy group (3.61 vs. 2.86; P = 0.040, χ2 test) (Table 3).
Table 3

Incidence of adverse drug reactions (only the first adverse reaction was counted in each patient)

 

No-fracture subgroup

Fracture subgroup

Entire population

Monotherapy group

Combined therapy group

Monotherapy group

Combined therapy group

Monotherapy group

Combined therapy group

Number of patients evaluated for safety

1,385

1,381

637

628

2022

2,009

Number of patients with adverse drug reactions

89

117

50

54

139

171

Total patient-year exposure

3364.12

3268.19

1501.30

1470.26

4865.42

4738.44

Incidence of adverse drug reactions

2.65

3.58

3.33

3.67

2.86

3.61

P value

0.031

0.618

0.040

The incidence of adverse drug reactions is expressed as the number of patients with any adverse drug reaction per 100 patient-years

All adverse drug reactions were included in the analysis unless the date of onset was unknown

P values were derived from analysis using the χ2 test

Serious adverse events occurred in 271 patients (138 from the monotherapy group and 133 from the combined therapy group), and there were 109 deaths (57 and 52, respectively). Serious adverse events and deaths showed a similar incidence in the two groups.

Discussion

This clinical study involved more than 4,000 Japanese women with osteoporosis. The two treatment groups were fairly well matched with regard to baseline characteristics, and the between-group differences of some variables probably reflected the enhancement of statistical power resulting from a large sample size.

In the monotherapy group, the incidence rate of new vertebral fractures (per 100 patient-years) was 2.6 for patients without existing vertebral fractures and 13.8 for patients with fractures at enrollment. The incidence rate of new vertebral fractures in each of the two subgroups was about one-fourth and two-thirds of the respective estimates (10 and 20 per 100 patient-years) used to design the study. In clinical studies of calcium supplements conducted between 1989 and 1994, the incidence rate of new vertebral fractures (per 100 patient-years) for patients with at least one existing vertebral fracture was reported to be 27.7 by Orimo et al. [15], 31.5 by Tilyard et al. [16], 33.3 by Aloia et al. [17], and 82.3 by Gallagher et al.[18]. The incidence rate of new vertebral fractures in the fracture subgroup from the present study (13.8) was lower than reported previously for osteoporotic patients and higher than that demonstrated by population-based studies (1.1 in the European Prospective Osteoporosis Study, or EPOS) [19] and 1.2 in a study conducted by Fujiwara et al. [20]. This result suggests that many of the patients enrolled in the present study might have had mild osteoporosis and a lower risk of developing fractures, probably because of the use of less specific diagnostic criteria for osteoporosis in this study, which were only based on radiographic grading of bone atrophy and disregarded various other risk factors for fractures. In contrast, supplementation of radiographic findings by assessment of risk factors for the occurrence of fracture (such as existing fractures) was done in the previous studies that were used as the basis for setting our target sample size.

Some clinical studies of activated vitamin D3 have suggested that a low baseline incidence rate of new vertebral fractures may lead to failure to demonstrate the efficacy of treatment for inhibition of fractures [21]. Thus, failure to show a between-group difference in the incidence rate of new vertebral fractures in the present study might also be the result of the generally low risk of new vertebral fractures in the patient population. The subgroup of patients with advanced osteoporosis, i.e., those who had at least five existing vertebral fractures, had a similar incidence rate of new vertebral fractures during calcium monotherapy to that reported previously. In these subjects, the combination of calcium and menatetrenone was shown to significantly reduce the risk of new vertebral fractures compared with calcium therapy alone.

Analysis by the Kaplan–Meier method showed that the 36-month cumulative incidence rate of clinical fractures during treatment with a calcium supplement was 2.7% in patients with one or more existing fractures at baseline; this was a lower rate than reported previously for patients on placebo therapy, being about one-third of that reported by Harris et al. (8.4%) [22] and about one-sixth of that reported by Reginster et al. (16.0%) [23].

As this study was planned in 1995, when data were still lacking on the incidence rate of clinical fractures in the Japanese population, it might not have been adequately powered to detect a statistically significant treatment effect on the incidence rate of clinical fractures (a secondary efficacy endpoint). The results of a meta-analysis of clinical studies of vitamin K2 were recently published [24], indicating a reduction of the risk of vertebral as well as nonvertebral fractures by treatment with this vitamin. In addition, Knapen et al. [25] reported that vitamin K2 supplementation increased the circumference and bone strength indices of the femoral neck. The lower incidence rate of clinical fractures in our combined therapy group compared with our monotherapy group, even though the difference was not significant, also suggests that vitamin K2 may provide an additional therapeutic benefit by reducing the risk of nonvertebral fractures.

Osteoporotic patients often suffer from loss of height as a consequence of kyphosis or lordosis, which is accelerated by an increase in the severity or number of vertebral body deformities. It has also been reported that a decrease of height is a predictor of fractures. A recent double-blind controlled study compared etidronate (a first-generation bisphosphonate) and risedronate (a third-generation bisphosphonate) in Japanese patients with osteoporosis and showed that risedronate significantly prevented a decrease of height compared to etidronate and also had a beneficial effect on ADL [26].

In the present study, combined therapy prevented a decrease of height compared with monotherapy in patients aged 75 years or more at enrollment, those whose last menstrual period occurred 30 years or more before enrollment, and those having at least five existing vertebral fractures at enrollment. This effect was associated with improvement of ADL, although there was no significant difference in the incidence rate of new vertebral fractures between the two groups.

Previous studies have revealed that the blood level of undercarboxylated osteocalcin (ucOC) shows a significant positive correlation with age [27] and that a higher circulating level of vitamin K is required to reduce blood ucOC in women aged ≥ 70 years than in younger women [28]. These findings suggest that older individuals are more likely to develop vitamin K deficiency or impaired utilization of this vitamin. Taken together with the known positive correlation between age and the prevalence of multiple vertebral fractures [29], this may account for the efficacy of adding menatetrenone to calcium supplementation that was demonstrated in several subsets of the present study population, i.e., patients aged 75 years or more at enrollment, those whose last menstrual period occurred 30 years or more before enrollment, and those with at least five vertebral fractures at enrollment.

Osteoporotic vertebral fractures can cause a decrease of height and chronic back pain, thereby adversely affecting both the quality of life (QOL) and ADL [30]. Recently, an increasing number of studies have used QOL and ADL in addition to BMD and the incidence rate of fractures as important measures when assessing the efficacy of medications for the treatment and prevention of osteoporosis. In the present study, combined therapy with calcium and menatetrenone was shown to improve the ADL of patients with osteoporosis, as a higher percentage of patients receiving combined therapy experienced improvement of several ADL parameters during the first 12 months of the study compared with those receiving calcium supplementation alone.

Although combined therapy caused more frequent adverse reactions, there were no major concerns in relation to serious adverse events. Some patients died during the 4-year follow-up period, but none of the deaths was related to the study treatment.

Limitations

This study aimed to assess the efficacy of add-on menatetrenone therapy by comparison with calcium alone. Because add-on therapy was unlikely to be inferior to monotherapy in terms of the prevention of fractures, a one-tailed hypothesis was considered to be appropriate for the present study, and the one-tailed level of significance was set at 5% (i.e., two-tailed significance of 10%).

Although a 5% level of significance has been used in most recent large-scale clinical studies, the LRC-CPPT study used a one-tailed level of 5% for statistical analysis [31]. The Mega Study, which was conducted before the present study, also used a two-tailed 10% level of significance [32]. Thus, at the time when this study was designed, some large-scale overseas studies were being conducted with a two-tailed significance level of 10% for statistical analysis. Accordingly, selection of a two-tailed level of significance of 10% for statistical analysis was not unusual at that time, but would probably not be done if the study were being planned today.

In conclusion, this was the first large-scale clinical study conducted in Japanese patients with postmenopausal osteoporosis to obtain evidence about the preventive effect of antiosteoporotic therapy on vertebral fractures and definitive data on the incidence rate of clinical fractures in this patient population. We hope that the results of this study will be helpful for the design of future clinical studies in Japanese patients with this condition.

Acknowledgments

Research funds were provided by the Japanese Ministry of Health, Labor and Welfare for the first 2 years of the study, and thereafter the study was funded by Eisai Co. Ltd. (Tokyo, Japan). We thank all the study participants, physicians, and co-medical staff.

Copyright information

© The Japanese Society for Bone and Mineral Research and Springer 2008