Volunteer subjects were recruited by M city public advertising. A participation condition is to be being normal and a postmenopausal woman, and the person that the ovary and the uterus which have an influence on the osteoporosis do not have a history of treatment. Exclusion criteria are the person with a history of the serious disease and the person whom it was judged to be inappropriate by the investigator. They were given detailed information about the study (such as aims and methods) at an orientation session offered prior to enrollment in the study. Written, informed consent was obtained from 97 volunteers. Of these, 83 were selected for statistical analysis. They had undergone all scheduled examinations conducted at the beginning of the study (prior to intervention with investigational products), as well as two and five months after study initiation.

This was a randomized, controlled trial. An identification number was assigned to each subject for anonymization. Subjects were divided into three groups matched for age and body mass index (BMI). Each group underwent a different intervention. In the LECA group, subjects received a Ca-supplemented lemon-juice (lemon) beverage. In the LE group, subjects received a Ca-unsupplemented lemon-juice (lemon) beverage. The third group (control) received no intervention. The placebo was not used in this study. The control group lived a usual life. The investigational products were distributed by Pokka Sapporo Food & Beverage Ltd. using a double-blind protocol. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the Prefectural University of Hiroshima (Approval Number 15MH025-01).

The study was conducted between December 2015 and June 2016. Each subject in the LECA and LE groups consumed one serving (290 mL) of an investigational product every day for five consecutive months. Assessments of physical measurements, BMD, and bone metabolism markers were performed at study initiation (pre-intervention) and after two and five months of intervention.

Table 2 shows the baseline physical characteristics of the study subjects. There were no significant differences in any of these variables among the three subject groups.

There were no significant differences in the BMD of the lumbar spine among the means of the three subject groups (p = 0.541) or between different time points (p = 0.902). However, there was a significant interaction effect (p < 0.001) (Figure 1(a)). When the pre-intervention BMD of the lumbar spine was subtracted from the BMD after five months of intervention, positive values were obtained for the LECA and LE groups, while a negative value was obtained for the control group. The differences in these values between the LECA and control groups and between the LE and control groups were both significant (both p < 0.001) (Figure 1(b)).

Similarly, there were no significant differences in the BMD of the femur among the means of the three subject groups (p = 0.709) or between different time points (p = 0.807). Again, the interaction effect was significant (p < 0.001) (Figure 1(c)). The value obtained by subtracting the pre-intervention BMD from the BMD after five months of intervention was positive for the LECA group. For the LE and control groups, the values obtained by the same method were negative. The difference in these values between the LECA and control groups was significant (p = 0.036). The difference between the LE and control groups was borderline significant (p = 0.064) (Figure 1(d)).

†: vs 153.4 (4.8), n = 2444, by Health and Welfare Statistics Association, p = 0.192271; ‡: vs 53.9 (7.8), n = 2443 by Health and Welfare Statistics Association, p < 0.001.

TRACP-5b concentrations did not show any significant differences among the means of the three subject groups (p = 0.624) or between different time points (p = 0.331). However, the interaction effect was significant (p < 0.001) (Figure 2(a)). When the pre-intervention TRACP-5b level was subtracted from the TRACP-5b level after five months of intervention, a negative value was obtained for the LECA group, whereas positive values were obtained for the LE and control groups. The differences in these values between the LECA and LE groups and between the LECA and control groups were both significant (both p < 0.001) (Figure 2(b)). In the case of u-NTx, its concentrations did not differ significantly among the means of three subject groups or between different time points. There was also no significant interaction effect (Figure 2(c)). Furthermore, there were no significant differences in u-NTx levels before and after the five-month intervention (Figure 2(d)).

As for the concentrations of OC, there were no significant differences among the means of the three subject groups (p = 0.835) or between different time points (p = 0.328). The interaction effect was significant (p = 0.012) (Figure 3(a)). When the pre-intervention OC level was subtracted from the OC level after five months of intervention, a negative value was obtained for all subject groups. However, the value obtained for the LECA group was significantly smaller than that obtained for the LE group (p = 0.008) or control (p = 0.033) group (Figure 3(b)).

Finally, BAP concentrations did not show any significant differences among the means of the three subject groups (p = 0.234) or between different time points (p = 0.138), although the interaction effect was significant (p < 0.001) (Figure 3(c)). The value obtained by subtracting the pre-intervention BAP level from the BAP level after five months of intervention was negative for the LECA group. Conversely, the same calculation yielded positive values for the LE and control groups. The differences in these values between the LECA and LE groups

and between the LECA and control groups were both significant (both p < 0.001) (Figure 3(d)).

In this study involving postmenopausal women, there was a significant interaction effect for the concentration of the bone resorption marker TRACP-5b. When the pre-intervention TRACP-5b level was subtracted from the TRACP-5b level after five months of intervention, a negative value was obtained for the LECA group. This value differed significantly from the positive values obtained by the same formula for the LE and control groups. Significant interaction effects were also observed for the concentrations of the bone formation markers OC and BAP. The value obtained by subtracting the pre-intervention OC level from the OC level after five months of intervention was negative for all three subject groups. However, the differences in these values between the LECA and LE groups and between the LECA and control groups were both significant. As for BAP, the value obtained by subtracting the pre-intervention level from the level after five months was negative for the LECA group and positive for the LE and control groups. The differences in these values between the LECA and LE groups and between the LECA and control groups were both significant. The results of previous studies that evaluated the effects of dietary Ca intervention are controversial. Matsumoto et al. [14] found significant decreases in bone resorption markers, whereas Kuwabara et al. [15] identified no significant changes. These discrepancies are partly due to differences in age between study participants. In the report by Kuwabara et al. [15] , 68 institutionalized elderly subjects consumed a jelly containing Ca (200 mg) and vitamin D3 (800 IU) every day for one month. After study completion, the authors found no significant changes in the concentrations of TRACP-5b and BAP, although serum 25(OH)D levels increased significantly. Separately, Toba et al. [16] examined the effect of consumption of the milk basic protein (MBP), a protein assembly with basic isoelectric points, on bone metabolism markers in adult men (mean age 36.2 years). They reported that, after a 16-day intervention, the concentrations of OC and u-NTx showed a significant increase and decrease, respectively. These were both short-term studies, with intervention periods of one month or less. In the present study, each subject regularly received an investigational product for five months, and, upon study completion, subjects in the LECA group had significantly lower levels of both bone resorption and bone formation markers. We hypothesize that when Ca-supplemented lemon beverages are regularly consumed, citric acid promotes the absorption of this mineral through its chelating activity. This would help maintain blood Ca concentrations within the normal range, resulting in lower levels of bone resorption and bone formation markers. In the LECA group, normal levels of blood Ca presumably eliminated the need for supplying this mineral from bone and suppressed osteoclast-mediated bone resorption. This likely inhibited the function of osteoblasts and led to the attenuation of bone formation [11] .

The increase in the BMD of the lumbar spine was calculated for each subject group by subtracting the pre-intervention value from the value obtained after five months of intervention. The results indicated that there was a significant difference in the increase between the LECA and control groups. A similar result was obtained for the BMD of the femur. Bone is constantly renewed through a process called remodeling [17] . This process elegantly regulates the balance between osteoclastic and osteoblastic activity, thereby preserving dynamic bone homeostasis. In other words, normal bone tissue is maintained through the resorption of an appropriate amount of old bone and the subsequent formation of new bone [18] . Changes in bone metabolism are induced during menopause, resulting in high bone turnover in which both bone resorption and bone formation are enhanced. However, bone resorption rates are much higher than bone formation rates, thus causing a reduction in bone mass [4] [19] . In the present study, the maintenance of blood Ca levels likely inhibited bone resorption. This must have suppressed the proliferation and differentiation of osteoblasts and caused the attenuation of high-turnover bone metabolism.

Moschonis et al. [20] observed a significant increase in the BMD of the lumbar spine in postmenopausal women who had received a 12-month intervention consisting of lifestyle counseling and the consumption of dairy products. Aoe et al. [21] also reported a similar observation in healthy postmenopausal women who had consumed MBP for six months. In the present study, the novel finding was that the changes in bone metabolism markers caused by the regular consumption of Ca-supplemented lemon beverages were reflected in BMD levels.

Since estrogen deficiency causes postmenopausal osteoporosis, inhibitors of bone resorption (such as estrogen and bisphosphonate preparations) have been used for the prevention and treatment of osteoporosis. In fact, Wallach et al. [22] found that risedronate treatment significantly increased or maintained BMD of the lumbar spine and femur in patients receiving corticosteroid therapy. Furthermore, according to a report by Yoshida et al. [23] , serum NTx levels were greatly reduced in early postmenopausal women within one month following six-month raloxifene treatment. The authors also mentioned the inhibitory effect on bone resorption. However, while estrogen-based hormone replacement therapy can reduce the risk of fracture and colon cancer, it reportedly increases the risk of breast cancer, thrombosis, coronary artery disease, and stroke [24] . In addition, although bisphosphonate preparations can increase bone mass and prevent bone resorption and fractures, their dosing instructions are not easy to follow. An example is that, upon arising, patients need to swallow bisphosphonate tablets whole with a full glass (approximately 180 mL) of water on an empty stomach, without lying down and consuming anything other than water for at least 30 min afterwards. Thus, one of the disadvantages of bisphosphonate therapy is poor patient compliance. Indeed, several groups have previously shown suboptimal adherence to daily oral bisphosphonate therapy [25] [26] [27] . Another problem associated with the pharmacological intervention of osteoporosis is that an individual must be diagnosed with this disease (i.e., BMD < 70% of the young adult mean) to receive a prescription for bisphosphonate. Considering the importance of preventive medicine in the management of osteoporosis, it would be essential to develop an alternative to conventional osteoporosis drugs that would be effective even for individuals whose BMD is 70% or higher. One can drink the Ca-supplemented lemon beverage investigated in this study regardless of the presence or absence of a disease or diagnosis. There are no instructions or restrictions as to when or how the beverage should be taken. Thus, it can be easily consumed regularly for an extended period of time.

In this study, a survey on individual diet remains to be done. Nevertheless, a new experimental result was achieved by ingesting the test beverage while having a normal life.

The physical measurements of each subject selected for the present analysis were similar to the mean values obtained from an age-matched Japanese population [28] . There were no significant difference between the average height of the subjects and Japanese women in their age group, but the subjects were below average in weight. This difference, however, is so tiny that we can generalize it. However, muscle strength varies among individuals based on a variety of elements, including their physical conditions, motivation, physical ability, nutritional status, and genetic factors [29] . Thus, it is difficult to measure strength without taking these factors into account. This illustrates a limitation of human research. Another limitation of the present study was that the participants were not hospitalized patients or institutionalized individuals. They were volunteers interested in a health study related to BMD measurement. Thus, the present subjects could have been biased towards healthy people or health-conscious individuals. Nevertheless, a novel finding suggesting that the continuous consumption of Ca-supplemented lemon beverages is effective in preventing the loss of BMD after menopause was obtained. This conclusion was based on the changes that occurred after five months of intervention in both bone metabolism markers and the BMD of the lumbar spine and femur. We expect our new discovery to become instrumental in developing preventive strategies for osteoporosis in the future. Research conducted by Nii et al. [30] has previously demonstrated that, in rats, the consumption of shirasuboshi together with sudachi (Citrus sudachi) juice leads to significant decreases in the concentrations of urinary pyridinoline and deoxypyridinoline and inhibits bone resorption. They have also reported that, in healthy male students, a diet containing shirasuboshi and sudachi juice promotes the absorption of minerals such as Ca and magnesium [31] . These observations suggest that citric acid in lemons also increased the absorption of dietary Ca in the present study. Further dietary surveys are necessary to substantiate this hypothesis.