Article Text

This article has been retracted. Please see:

Amelioration of osteopenia and hypovitaminosis D by 1α-hydroxyvitamin D3 in elderly patients with Parkinson’s disease
  1. Yoshihiro Satoa,
  2. Seiji Manabea,
  3. Haruko Kunoa,
  4. Kotaro Oizumib
  1. aDepartment of Neurology, Futase Social Insurance Hospital, Iizuka, Japan, bFirst Department of Internal Medicine, Kurume University School of Medicine, Kurume, Japan
  1. Dr Yoshihiro Sato, Department of Neurology, Kurume University Medical Center, 155–1 Kokubumachi, Kurume 839–0863, Japan. Telephone 0081 942 22 6111; fax 0081 942 22 6533; email y-sato{at}


OBJECTIVES A high prevalence of hip and other fractures in elderly patients with Parkinson’s disease has been linked to reduced bone mass arising from a defect of renal synthesis of 1, 25-dihydroxyvitamin D (1, 25-[OH]2D). Treatment with 1α-hydroxyvitamin D3 (1α(OH)D3; an active form of vitamin D) was evaluated for maintaining bone mass and reducing the incidence of hip and other non-vertebral fractures in patients with Parkinson’s disease.

METHODS In a double blind, randomised trial, 86 elderly patients with Parkinson’s disease (mean Hoehn and Yahr stage, 3; mean age 70.6 years) were randomised to receive either 1 μg 1α(OH)D3 daily (treatment group, n=43) or a placebo (n=43) for 18 months. Bone mineral densities in the second metacarpals were determined by computed radiographic densitometry. Serum bone turnover indices were measured serially, and incidence of non-vertebral fractures was recorded.

RESULTS Bone mineral densities decreased 1.2% in the treatment group compared with 6.7% in the placebo group during 18 months (p<0.0001). At baseline in both groups, the serum concentration of 1, 25-[OH]2D was reduced. Parathyroid hormone was abnormally increased in 15 patients (17%) and correlated negatively with serum 25-hydroxyvitamin D, indicating compensatory hyperparathyroidism. Eight patients sustained fractures (six at the hip and two at other sites) in the placebo group, and one hip fracture occurred among treated patients (odds ratio 9.8; p=0.0028).

CONCLUSION By increasing serum 1, 25-[OH]2D concentrations, treatment with 1α(OH)D3 can reduce the risk of hip and other non-vertebral fractures in osteoporotic elderly patients with Parkinson’s disease by slowing the loss of bone mineral densities.

  • osteoporosis
  • Parkinson’s disease
  • vitamin D

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Despite a high incidence of hip fractures due to falls in patients with Parkinson’s disease,1-3 especially in elderly female patients (odds ratio 9.4),1 little is known about bone changes in these patients.4 5 Furthermore, hip fractures complicate Parkinson’s disease in many elderly people of both sexes,6-8 and increased frequency of fractures other than the hip has also been reported.3 Our previous study9 showed that a deficiency of 1, 25-dihydroxyvitamin D (1, 25-[OH]2D) (the most active metabolic form of vitamin D) due to a 25-hydroxyvitamin D (25-OHD) deficiency or an age related reduction in the amount of 1 α-hydroxylase with compensatory hyperparathyroidism contributes to reduced bone mineral density in patients with Parkinson’s disease. This reduction in bone mineral density increases the risk of fractures. As there is a defect in the renal synthesis of 1, 25-[OH]2D in Parkinson’s disease,9 fracture risk may be reduced not by the non-active forms of vitamin D supplements such as ergocalciferol or cholecalciferol, but by an active form of vitamin D such as 1α-hydroxyvitamin D3 (1α(OH)D3). Supplementation with calcium plus 1α(OH)D3 (an analogue of calcitriol and a provitamin of 1,25(OH)2D3) or with 1α(OH)D3 alone increases bone mineral density in cortical and trabecular bone.10-13 Supplements decrease the incidence of vertebral fractures in patients with senile and postmenopausal osteoporosis,10-14 and reduce the risk of hip fractures on the hemiplegic side of elderly patients with stroke.15

We conducted an 18 month randomised trial to evaluate the efficacy of 1α(OH)D3 in reducing progression of osteopenia in the second metacarpal and in decreasing non-vertebral fractures in elderly patients with Parkinson’s disease with low serum 1, 25-[OH]2D concentrations. The rate of vertebral fractures was not determined in this study, because many such fractures are asymptomatic among elderly patients with Parkinson’s disease and because the interpretation of spinal radiography may be complicated by osteoarthritis or scoliosis.


Eighty six patients with Parkinson’s disease 65 years of age or older (35 men, 51 women; mean age 70.6 years; range 65 to 88) were enrolled in a randomised double blind, placebo controlled study; the subjects had not participated in a previous study.9 The patients lived in the community (Kahanzan district with a population of 210 000), and about half of the patients visited the hospital where the study was performed. Patients were excluded if they showed other known causes of osteoporosis, such as hyperparathyroidism or renal osteodystrophy; impairment of renal, cardiac, or thyroid function; a history of therapy with corticosteroids, oestrogens, calcitonin, etidronate, calcium, or vitamin D for 3 months or longer during the 18 months preceding the study; or even brief treatment of this nature during the 2 months immediately preceding the study. Patients at Hoehn and Yahr stage 5 were excluded because their poor ambulation status largely precluded any chance of fracture. Patients with a history of non-vertebral fracture were also excluded.

The study was approved by the institutional ethics committee, and informed consent was obtained from all study subjects in the presence of a witness.

The patients with Parkinson’s disease were assigned to one of two study groups (treatment (n=43) and placebo (n=43)) by means of a computer generated random number. Neither the patients nor the physicians were aware of the treatment assignment throughout the study period. Patients in the treatment group received a single daily oral dose of 1.0 μg 1α(OH)D3 (active vitamin D3: 1.0 μg tablet) for 18 months. The control group received a placebo once a day. The 1α(OH)D3 and placebo tablets in this trial were identical in size, weight, colour, and taste. Vitamin D or placebo preparations were taken in the morning. No dose adjustments were made at any time in the study. Both groups were studied for 18 months of treatment. The patients’ clinical status was assessed at baseline and followed up every 2 weeks, at which times clinical status was assessed, and non-vertebral fractures were recorded. The frequency of falls was estimated on the basis of interviews, and patients who had fallen at least once in the 3 months before recruitment were defined as “fallers”. The number of falls per subject was also recorded during the 18 months.

The patients completed a questionnaire about diet and exposure to sunlight. The mean weekly dietary vitamin D intake was calculated for each patient, and patients who consumed less vitamin D than the Japanese recommended daily allowance (100 IU) were defined as low dietary consumers of the vitamin. Exposure to sunlight was assessed by the patients or family members in terms of the preceding year, being graded as almost none, less than 15 minutes a week, or longer.

Using a computed x ray densitometer (CXD; Teijin Limited, Tokyo, Japan),16 17 the bone mineral density of the right second metacarpal bone was measured on the day of study entry and 18 months later. The CXD method measures bone density and cortical thickness at the middle of the second metacarpal bone, using a radiograph of the hand and an aluminum step wedge as a standard (20 steps; 1 mm/step). The computer calculated bone mineral density on the basis of the pattern expressed as gradations along the aluminum step wedge; bone mineral density was expressed as the thickness of an aluminum equivalent (mm Al) showing correspondingx ray absorption.

Serum 25-hydroxyvitamin D (25-OHD) was determined using a competitive protein binding assay, and 1, 25-dihydroxyvitamin D (1, 25-[OH]2D) was determined by a radioreceptor assay using calf thymus receptor (Nichols Institute Diagnostics, San Juan Capistrano, CA, USA; normal range at 65 to 75 years in our laboratory, 18.5 to 24.7 ng/ml for 25-OHD and 40.4 to 58.8 pg/ml for 1, 25-(OH)2D). Ionised calcium in serum prepared freshly under anaerobic conditions was measured using an ion-selective electrode and an ionised calcium analyser (NOVA Biochemical, Newton, MA, USA; normal range in elderly people 1.242 to 1.307 mmol/l). Serum intact parathyroid hormone (PTH) was measured by a radioimmunoassay (RIA; Nichols Institute Diagnostics; normal range for elderly people, 32 to 41 pg/ml). Intact bone Gla protein (BGP), an osteoblastic bone formation marker,18 was measured with an established enzyme immunoassay using antibodies to the N-terminal and C-terminal regions of human BGP (Teijin Diagnostics, Tokyo, Japan; normal range in elderly people 2.5 to 11.5 ng/ml). The pyridinoline cross linked carboxyterminal telopeptide of type I collagen (ICTP), an osteoclastic bone resorption marker,19 was measured by RIA (Orion Diagnostica, Oulunsalo, Finland; normal range for elderly subjects 6.0 to 8.2 ng/ml).

All statistical procedures were performed using the Statview 4.11 software package (Abacus Concepts, Berkeley, CA). Data are presented as the mean (SD). Each variable between placebo and treatment groups was compared by an unpaired t test. Baseline differences of categorical data were tested by χ2analysis. To examine the correlation between serum 25-OHD concentration and PTH concentration, Spearman’s rank correlation coefficients (SRCCs) were calculated. One way analysis of variance (ANOVA) and Fisher’s protected least significant difference were used to assess the differences of PTH concentrations between the three Parkinson’s disease groups categorised by serum 25-OHD concentrations. For bone mineral density measurements, individual values and laboratory values were computed and expressed as a percentage change from baseline. The two groups were then compared with respect to bone mineral density and the laboratory values by using the Wilcoxon rank sum test. Difference of fracture rate between the two groups during the 18 months was tested by Fisher’s exact test. p Values<5% were considered statistically significant.



Demographic and baseline clinical features of study patients are presented in tables 1 and 2. At entry, groups were comparable in age, sex, duration of illness, Hoehn and Yahr stage, body mass index, percentage of fallers, sunlight exposure, dietary intake of vitamin D, bone mineral density, and serum biochemical markers of bone metabolism. The 37% to 40% incidence of fallers in the two groups was similar to those previously reported for patients with Parkinson’s disease with a mean Hoehn and Yahr stage of 3.0.20 In addition, many patients in both groups had been exposed to sunlight less than 15 minutes a week because of reduced mobility. The mean vitamin D intake in both groups was also less than the Japanese recommended daily allowance. Bone mineral density and serum biochemical markers of bone metabolism at the baseline were not significantly different between the two groups. Of the 86 patients initially enrolled, three in the 1α(OH)D3 group and three in the placebo group left the study for non-compliance with the study regimen (one in each group), loss to follow up (one in each group), intercurrent illness (one placebo patient), or death (one treated patient).

Table 1

Demographic and baseline clinical characteristics of patients with Parkinson’s disease at study entry

Table 2

Bone mineral density and various serum biochemical tests in the vitamin D and placebo groups at baseline and after 18 months


As shown in table 2, bone mineral density in both groups declined significantly below baseline values over 18 months. The percentage changes in bone mineral density were −1.2 (SD 3.7) in the treatment group and −6.7 (SD 2.6) in the placebo group. The difference between the groups was significant (p<0.0001, Wilcoxon rank sum test).

At baseline, patients in both groups had low serum 25-OHD and 1, 25-[OH]2D concentrations, normal serum concentrations of ionised calcium, normal serum PTH concentrations, normal serum BGP concentrations, and low normal serum ICTP concentrations (table 2).

During the 18 month study period, the percentage change differences between the two groups were significant in increasing 1, 25-[OH]2D and in decreasing ionised calcium, PTH, BGP, and ICTP in the treatment group. The difference in the change in 25-OHD concentration between the two groups was not significant (table 2). In the treatment group, the serum calcium concentration increased in 17 patients.


Serum 25-OHD concentration was defined as deficient when <10 ng/ml, insufficient when 10 to 20 ng/ml, and sufficient when >20 ng/ml.9 At baseline, the mean PTH concentration was significantly higher in patients with deficient concentrations of 25-OHD (n=47; PTH, 40.7 (SD 22.8) pg/ml) and in patients with insufficient concentrations of 25-OHD (n=29; PTH, 34.9 (SD 19.0) pg/ml) than in those with sufficient concentrations of 25-OHD (n=10; PTH, 20.4 (SD 5.1) pg/ml) (p=0.008 and 0.04 respectively). In 15 patients (17%), PTH concentrations were abnormally raised (above an upper limit of 41 pg/ml, which represents the mean+1SD of control subjects). All 15 patients had deficient or insufficient 25-OHD concentrations.

A negative correlation existed between the serum 25-OHD concentration and PTH concentration (SRCC, r=−0.204, p=0.030), indicating compensatory hyperparathyroidism in patients with hypovitaminosis D.


During the 18 months of the study, fractures caused by falls occurred in eight of 40 patients in the placebo group (six with hip fractures, including three women at Hoehn and Yahr stage 4, one woman at stage 3, and two men at stages 3 and 4; two other fractures at the radius and ankle occurred including one woman and one man at stage 4). One hip fracture occurred in the treatment group in a woman at stage 4 (p=0.003). The odds ratio for non-vertebral fractures among patients in the placebo group compared with those in the vitamin D3 group was 9.8 (95% confidence interval, 4.7 to 20.2). The number of non-vertebral fractures per 1000 patient-years was 17 and 167 for the 1α(OH)D3 and placebo groups respectively. There was no significant difference between the two groups in the number of falls per subject during the 18 months (1.3 (SD 1.9) in the placebo group and 1.4 (SD 1.8) in the vitamin D group).


The provitamin 1α(OH)D3 is converted in the liver to 1, 25(OH)2 D3; the vitamin D derivative (calcitriol) reduces the rate of new vertebral fractures in postmenopausal osteoporosis.21Treatment with 1α(OH)D3 has been shown to stimulate intestinal calcium absorption and bone mineralisation, as evidenced by reports of bone mineral density increases in lumbar vertebrae, the radius, and the greater trochanter after treatment.10-13 Treatment can decrease the rate of vertebral fractures in elderly and postmenopausal women,10-14 as well as hip fractures on the hemiplegic side of elderly patients after stroke.15 In addition, administration of 1 μg 1α(OH)D3/day depresses bone turnover and increases the bone mineral density in patients with senile osteoporosis.14 The results of the present study suggest that use of 1α(OH)D310 11 14 in patients with Parkinson’s disease may be beneficial.

Firstly, we showed that among the elderly patients with Parkinson’s disease who completed the 18 month study, the number of hip fractures and other non-vertebral fractures was 17.5% lower among the patients treated with 1α(OH)D3 than among those who received placebo. The incidence of non-vertebral fractures during the 18 months among the placebo group was as high as 20.0%, indicating a fracture rate of 167/1000 patient-years. By contrast, in a previous study the non-vertebral fracture rate in elderly men and women who were living in the community was reported to be 6.5%/18 months,22 and the fracture rate/1000 patient-years has been reported to be 116 in elderly women.23 The high incidence of non-vertebral fractures in elderly patients with Parkinson’s disease may be attributable to osteopenia due to hypovitaminosis D and a high prevalence of falls. The beneficial effect of 1α(OH)D3 in reducing the high incidence of non-vertebral fractures in this population is noteworthy. As the number of falls was similar in the two groups during the observation period, 1α(OH)D3 may prevent non-vertebral fractures even in patients with Parkinson’s disease with frequent falls.

Secondly, 1α(OH)D3 was shown to attenuate further bone loss, as evidenced by slowing of metacarpal bone mineral density reduction in patients receiving 1α(OH)D3. Other reports on senile or postmenopausal osteoporosis, however, have shown that treatment with 1α(OH)D3 increased bone mineral density during an observation period between 7 months and 5 years.10-13 24 Loss of bone mineral density of the femoral neck, spine, and total body in elderly patients of both sexes without treatment has been reported to be less than 1% even after 3 years.22 In the present study, we found that bone loss is greater in patients with Parkinson’s disease. Bone mineral density decreased 6.7% and 1.2% in the placebo and 1α(OH)D3 groups, respectively. This result implies that, in addition to senile osteoporotic changes, vitamin D deficiency with compensatory hyperparathyroidism can cause osteopenia in elderly patients with Parkinson’s disease. The present study showed a strong negative correlation between 25-OHD and PTH, with an abnormally raised PTH value in 15 patients when the 25-OHD concentrations were deficient or insufficient. Thus although the PTH value at baseline was normal, compensatory hyperparathyroidism existed in these patients, especially in the setting of low 25-OHD concentrations with normocalcaemia.

These bone mineral density measurements were determined by CXD of the second metacarpal. CXD approaches dual energyx ray absorptiometry (DXA), the accepted standard method of bone mass measurement,25 in precision. The precision errors (coefficient of variation (CV)) of the CXD technology were 0.2 to 1.2% for bone mineral density. Matsumotoet al 26 have compared the accuracy of CXD and DXA in bone mineral density determination. In 251 healthy women, they measured the bone mineral density in the metacarpal by CXD and used DXA in the vertebrae (L2-L4), femoral neck, and the skeleton; nearly identical bone mineral density was found by the two methods. Other investigators have found metacarpal bone density to correlate with lumbar bone mineral density determined by DXA.27Our previous study in stroke patients also showed that the degree of bone mineral density reduction in the second metacarpal of the hemiplegic side, as determined by the CXD method used in the present study, correlated with the risk of hip fractures on that side.15 Therefore, reduction in bone mineral density in the second metacarpals in patients with Parkinson’s disease, as found in this study, most likely reflects a generalised decrease in bone mineral density in the whole skeleton.

Thirdly, we demonstrated that in patients with Parkinson’s disease 1α(OH)D3 increases serum concentrations of 1, 25-[OH]2D. Conversely, such treatment decreases serum concentrations of calcium, PTH, BGP, and ICTP. This effect implies that treatment with 1α(OH)D3 inhibits development of compensatory hyperparathyroidism and diminishes bone turnover. Decreased serum calcium in the treatment group was unexpected, as 1α(OH)D3 usually increases serum calcium concentrations.15 Inhibition of hyperparathyroidism by the 1α(OH)D3 may overshadow the increase of calcium. Indeed, the serum calcium concentration increased in only 17 patients in the treatment group.

In addition, at baseline, the intact BGP and ICTP concentrations were not increased despite the presence of compensatory hyperparathyroidism. The fact that normocalcaemia was accompanied by a reduction in both 1, 25-[OH]2D and bone mineral density in patients with Parkinson’s disease indicates that vitamin D deficiency related osteopenia is caused not by hypocalcemia but by a direct inhibitory effect on osteoblasts, which regulate mineralisation by producing BGP. This hypothesis may explain the normal BGP concentrations at baseline. Also, normal ICTP concentrations may be caused by the very low concentration of 1, 25-[OH]2D, which at higher concentrations would activate osteoclastic bone resorption. None the less, bone turnover was depressed by inhibition of PTH secretion. The mechanism by which 1α(OH)D3 inhibits further bone mineral density loss and prevents fractures seems to involve stimulation of bone mineralisation and inhibition of compensatory hyperparathyroidism rather than stimulation of intestinal calcium absorption.

We found that 1α(OH)D3 slows the decrease in bone mineral density and loss of bone mass and prevents non-vertebral fractures in Parkinson’s disease. Combining 1α(OH)D3 with an agent which increases bone mass, such as calcitonin28 or bisphosphonate,29 30may increase bone mass while preventing non-vertebral fractures. Such study could include subgroup analysis by age to decide at what age therapy is required and would require a large community based trial over a longer period using DXA.


Supplementary materials

Linked Articles

  • Retraction
    BMJ Publishing Group Ltd