Increased Bone Resorption During the First Year After Stroke
Background and Purpose—Significant bone mineral density (BMD) reduction occurs in stroke patients on the hemiplegic side compared with the intact side. To elucidate the pathogenesis of hip fractures in this population, we measured serum markers of bone metabolism and BMD in the stroke patients within 1 year (early group) and between 1 and 2 years after onset of hemiplegia (long-term group).
Methods—Sera were collected from 51 patients from the early group and 93 patients from the long-term group. All patients had hemiplegia. Sera were assayed for pyridinoline cross-linked carboxy-terminal telopeptide of type I collagen (ICTP; a bone resorption marker) and bone Gla protein (a bone formation marker). The z score of BMD was determined in both second metacarpals.
Results—Serum ICTP concentrations (ng/mL) were higher in the early group (15.4±4.1) than in the long-term group (6.7±4.4). Bone Gla protein was normal or low in both groups. Multiple regression analysis identified Barthel Index, degree of hemiplegia, and illness duration as independent determinants of ICTP in the early group, whereas Barthel Index, degree of hemiplegia, and serum calcium were determinants of ICTP in the long-term group. There were statistically significant correlations between the z score of the hemiplegic side and age, Barthel Index, degree of hemiplegia, illness duration, 25-hydroxyvitamin D (25-OHD), and ICTP in the early group and between the z score and degree of hemiplegia and 25-OHD level in the long-term group.
Conclusions—The pathogenesis of reduced BMD differed between the early and long-term stroke groups. These results suggest that in the early group, increased bone resorption caused by immobilization was responsible for osteopenia on the hemiplegic side, whereas the degree of hemiplegia and 25-OHD level were the determinants of osteopenia in the long-term group.
Previously we demonstrated that disuse as the result of paralysis and vitamin D deficiency caused by malnutrition, sunlight deprivation, and immobilization-induced hypercalcemia can cause reduced bone mineral density (BMD) on the hemiplegic side compared with the contralateral side in elderly patients over a period of 4 years after a stroke.1 2 3 Because long-standing immobilization can cause hypovitaminosis D and induce normal or low skeletal turnover in poststroke patients, immobilization may be a major cause of this osteopenia.3
The immobilization-induced osteoporosis observed in poliomyelitis or spinal cord injury has an effect on bone modeling and remodeling through an increased activation of remodeling loci leading to increased osteoclastic bone resorption as evidenced by increased urinary calcium excretion4 and a decrease of the osteoblastic stimulus.5 There is a potential for recovery during the active early phase of immobilization osteoporosis that may disappear in the subsequent late inactive phase.4 5 After 12 months, excretion of urinary calcium indicating net bone resorption returns to normal6 or remains elevated.4 On the other hand, from 4% to 15% of hip fractures occur as a late complication of stroke, with ≥79% occurring on the hemiplegic side.7 8 9 10
To prevent hip fracture, the goal of this study was to determine the bone turnover and bone mass in hemiplegic patients at different phases of recovery from stroke. We examined the differences in biochemical indices of bone metabolism and turnover and the changes in bone density between patients in an earlier recovery period (<1 year, early group) and those in a later recovery stage (illness duration between 1 and 2 years, long-term group).
Subjects and Methods
Hemiplegic patients <1 year after a stroke were eligible for this study. Patients included individuals of either sex, 65 years of age or older, who had been examined in one of three hospitals (Kurume University Medical Center, Ohshima Hospital, and Futase Social Insurance Hospital) from June 1997 to November 1997. Patients were excluded if they had received any drug known to alter bone metabolism such as corticosteroids, thyroxine, anticonvulsants, estrogen, or vitamin D before and after the onset of stroke. Other reasons for exclusion included multiple strokes, severe physical immobility (bedridden), or severe renal (>1.5 mg/dL of creatinine) or hepatic insufficiency. Fifty-one physically mobile patients were enrolled in the early group.
Stroke patients with illness duration between 1 and 2 years after onset were defined as the long-term group and were selected by detailed examination of case notes of patients with hemiplegia to match patients of <1-year duration with respect to age, sex, severity of hemiplegia, and Barthel Index (BI).11 The exclusion criteria of the early group also were used in the long-term group. Ninety-three patients with hemiplegia caused by stroke constituted the long-term group. As the day the patient first had hemiplegia was defined as the onset of stroke, the duration of illness was defined as the duration of both hemiplegia and stroke in the two groups. Community-dwelling age-matched volunteers (15 men and 17 women) served as healthy control subjects.
In addition to BI, clinical severity of the hemiplegia was evaluated with Brunstrom’s staging classification,12 in which a score of 1 is defined as complete paralysis of the finger, arm, or leg, and a score of 6 represents normal strength.
With the use of a computed x-ray densitometer (CXD; Teijin Limited),13 BMD of the second metacarpal was measured in both hands. The CXD method measures bone density at the middle of the second metacarpal by using a radiograph of the hand and an aluminum step wedge as a standard (20 steps, 1 mm/step). The computer compares the gradations of the aluminum step wedge as the thickness of an aluminum equivalent (mm Al) with corresponding x-ray absorption. Thus the estimated BMD is a relative scale. A total of 754 individuals, 245 men and 509 women, 50 to 90 years of age (matched for age and sex) served as control subjects. The standard deviation (SD) for the normal control subjects in each sex and age group was used to calculate the z score. Because the normal values of each individual index vary according to sex and age; we used only the z score, which is unrelated to sex or age, to assess patients’ bone changes.
On the day of bone evaluation, a fasting blood sample was obtained in the 51 patients from the early group, 93 long-term patients, and 32 healthy control subjects. The samples were analyzed for ionized calcium, parathyroid hormone (intact PTH, 1 to 84), intact bone Gla protein (BGP, a bone formation marker),14 pyridinoline cross-linked carboxy-terminal telopeptide of type I collagen (ICTP, a bone resorption marker),15 25-hydroxyvitamin D (25-OHD),1, 25-dihydroxyvitamin D (1, 25-[OH]2D), and creatinine, as previously described.2 3 Because in our previous study of the older, convalescent-stage stroke patients, serum ICTP concentration was low or normal,3 we used serum ICTP as a bone resorption marker. On the basis of previously reported data, 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.
All patients and volunteers were informed of the nature of the study before witnessed consent from each participant. The protocol of the study was approved by the Local Ethics Committee.
Data are presented as mean±SD. Unpaired t tests (continuous variables) were used to assess the significance of differences between the early and long-term groups. Categorical data group differences were tested by x2 analyses. One-way analysis of variance and Fisher’s protected least significant difference were used to assess differences between the two stroke groups and the control subjects. To separate the influence of illness duration on the z scores of the hemiplegic and intact sides in the early group, bilateral measurement of z scores was performed in subgroups with duration of illness of <6 months and illness duration >6 months. The measurements were analyzed by repeated-measures analysis of variance. Spearman’s rank correlation coefficients were calculated to determine the relation between the z score of BMD or ICTP and each variable. Multiple regression analysis was used to estimate the independent effects of predictor variables on the z score of BMD for both sides and ICTP in each stroke group. Values of P<0.05 were considered statistically significant.
Characteristics of Study Subjects
Descriptive characteristics of the patient population are shown in Table 1⇓. No differences were observed between the two stroke groups in terms of age, sex, BI score, degree of hemiplegia, or type of stroke. The duration of illness was significantly shorter in the early group than in the long-term group. The mean illness duration of the patients from the early and long-term groups was 118 and 439 days, respectively. Eighteen (35%) patients in the early group and 31 (33%) patients in the long-term group were able to ambulate without assistance, whereas the remaining patients in both groups could ambulate independently with an assistive device but did not require help from another person. Hemiplegia affected the dominant side in 26 (51%) patients in the early group and 45 (48%) patients in the long-term group. Other than hemiplegia, the following neurologic deficits were observed in the early and long-term groups: hemisensory impairment in 26 and 69, homonymous hemianopsia in 3 and 7, and alternate hemiplegia in 1 and 2 patients, respectively. Aphasia was observed in 6 and 8 patients, respectively.
Serum Biochemical Indices and Bone Changes
As previously reported,2 the serum concentrations of 25-OHD and 1, 25-[OH]2D were significantly lower in the two stroke groups than in the control subjects (Table 2⇓). Also, the mean serum calcium levels were significantly higher in the early group than those in the long-term group and in the control subjects, indicating immobilization-induced hypercalcemia. There was no significant change of PTH and creatinine levels between the three groups. The serum ICTP concentrations of the early group patients were significantly higher than those of the long-term group and control subjects, whereas significant differences of these indices were not observed between the long-term group and control subjects. BGP was normal or low in the two stroke groups, although this difference was not statistically significant. As previously reported,1 the z score of the hemiplegic side was significantly lower than on the contralateral side in the two stroke groups and as compared with the control subjects. The z score on the hemiplegic side in the two stroke groups did not differ significantly but was decreased significantly compared with control subjects.
Parathyroid Hormone Concentrations and Serum Levels of 25-OHD
The mean PTH level was significantly higher in patients with deficient levels of 25-OHD (n=70, PTH; 49.1±11.5 pg/mL) than in those with insufficient (n=73, PTH; 29.4±16.1 pg/mL) or sufficient levels of 25-OHD (n=1, PTH; 13 pg/mL) (P<0.0001). This result indicated vitamin D deficiency with compensatory hyperparathyroidism.
Relations Between BMD or ICTP and Each Variable
Spearman’s rank correlation coefficients between the z score of BMD or ICTP and each variable were analyzed together in both the early and long-term groups (Table 3⇓). There were significant correlations between the z score of the hemiplegic side and the patient’s age, BI score, degree of hemiplegia in the finger and leg, illness duration, calcium, 25-OHD, and ICTP, whereas the z score of the intact side correlated with BI score and 25-OHD. For both sides, PTH, BGP, and creatinine did not correlate with the z scores. ICTP correlated with age, BI score, degree of hemiplegia of not only the finger but the leg, illness duration, calcium, and 25-OHD.
When analyzed together in the two stroke groups, serum ionized calcium concentrations correlated positively with ICTP (r=0.322, P<0.0001) and negatively with BI score (r=−0.400, P<0.0001), PTH (r=−0.314, P=0.0002) and 1, 25-[OH]2D (r=−0.243, P=0.0035). These results imply that immobilization-induced hypercalcemia inhibits secondary hyperparathyroidism in hypovitaminosis D and renal synthesis of 1, 25-[OH]2D.
Multiple Regression Analysis
The results of multiple regression analysis, in which age, BI, degree of hemiplegia, duration of illness, calcium, 25-OHD, and ICTP were selected as independent variables and the z score of BMD as the dependent variable, are shown in Table 4⇓. On both the hemiplegic and intact sides in the early group, age, degree of finger paralysis, and 25-OHD correlated significantly with the z score, whereas BI, illness duration, calcium, and ICTP were significantly related to the z score of the hemiplegic side. In the long-term group, 25-OHD correlated with the z score on both sides, whereas the degree of finger paralysis was related to the z score of the hemiplegic side. In the early group, left-right differences of the z score did not show a significant change between subgroups with an illness duration <6 months (n=16) and subgroups with an illness duration >6 months (n=35) (P=0.77, P value of interaction=0.51). This result suggests that the effect of illness duration did not differ between the hemiplegic and intact sides.
In a second multiple regression analysis, age, BI, degree of hemiplegia, calcium, and 25-OHD were selected as independent variables and ICTP as the dependent variable. In the early group, a significant correlation was observed between ICTP and BI, illness duration, and degree of finger and leg paralysis. In the long-term group, BI, degree of finger paralysis, and calcium were related significantly to ICTP (Table 5⇓).
However, little is known about changes of bone metabolism in immobilized stroke patients. Although our previous study of elderly patients with long-standing hemiplegic stroke indicated that long-standing immobilization can cause hypovitaminosis D without compensatory hyperparathyroidism and induce normal or low skeletal turnover,3 bone metabolism in the early stage of stroke recovery had not been examined.
This study demonstrated that accelerated bone metabolism with increased bone resorption occurs in patients with poststroke hemiplegia within 1 year, as evidenced by the high serum concentrations of ICTP and normal or low BGP. Multiple factors may be accountable for the increased bone resorption and uncoupled high bone turnover. Immobilization as the result of hemiplegia in the early group may be the major cause of increased bone resorption, because physical activity as assessed by BI and the degree of hemiplegia in the finger and leg were independent determinants of ICTP in the early group. The illness duration was also a determinant of ICTP in the early group, indicating that a high rate of bone resorption may occur after early recovery from a stroke. On the other hand, BI, degree of finger paralysis, and calcium were the independent determinants of ICTP in the long-term group. This result implies that a weak uncoupled state persists for >1 year after a stroke. Compensatory hyperparathyroidism in vitamin D deficiency may not account for the high ICTP level in the early group, because immobilization-induced hypercalcemia inhibits PTH secretion. Consequently, PTH did not correlate with the z score of either the hemiplegic or intact side. The hypercalcemia in the early group compared with the long-term group resulting in severe inhibition of 1, 25-[OH]2D production may explain the discrepancy between the two stroke groups. Normal or low BGP may be caused by a severe 1, 25-[OH]2D deficiency resulting from hypercalcemia-induced inhibition of its production; 1, 25-[OH]2D is needed to enhance the synthesis of osteoblast-derived BGP and of matrix BGP.
Independent determinants of the z score of the hemiplegic side were age, BI, degree of finger paralysis, illness duration, calcium, 25-OHD, and ICTP in the early group, whereas degree of finger paralysis and 25-OHD were determinants of the hemiplegic z score in the long-term group. The fact that the z score of the hemiplegic side in the two stroke groups did not differ significantly indicates that a great deal of bone loss occurs within 1 year. The reason for the heavy loss of bone may be explained by the fact that seven independent determinants affected the z score of the hemiplegic side in the early group, whereas only two of them influenced the hemiplegic z score in the long-term group. These differences between the two stroke groups are important in skeletal management of stroke patients with different illness durations. In addition to age, physical inactivity as the result of immobilization and paralysis may be the primary cause of the decreased hemiplegic z score. Illness duration, calcium, 25-OHD, and ICTP in the early group may be influenced by physical activity. In the nondisabled population of age ranging from 70 to 95 years, bone remodeling may almost reach an equilibrium, resulting in a steady rate of bone loss.19 It is noteworthy that the older stroke patients in the early group showed higher skeletal turnover and that age was a determinant of the z score of both sides.
A comparison was not made between metacarpal BMD determined by CXD and femoral neck BMD measurements performed by dual-energy x-ray absorptiometry in disabled stroke patients. In the present study, however, we found a positive correlation between Brunstrom’s stage in the leg and the observed BMD in the finger. Also, the degree of leg paralysis correlated with ICTP in the early group. Our previous study demonstrated that the degree of BMD 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.20 Reduction in the second metacarpal BMD in stroke patients, therefore, may reflect a decrease in hip BMD.
Although a longitudinal study would have been desirable to assess continuous changes of bone and biochemical parameters that occur during the early stages of a hemiplegic stroke, the present cross-sectional study demonstrated significant differences in BMD and in biochemical indices between the two stroke groups. During the early recovery period after a stroke, especially within the first year, a clinical trial of agents inhibiting bone resorption such as bisphosphonate21 or calcitonin22 is needed to assess whether they can decrease bone resorption and prevent further bone loss.
This study was supported by grants from the Sumitomo and Eisai Corporations.
- Received December 30, 1997.
- Revision received April 24, 1998.
- Accepted April 24, 1998.
- Copyright © 1998 by American Heart Association
Sato Y, Maruoka H, Oizumi K, Kikuyama M. Vitamin D deficiency and osteopenia in the hemiplegic limbs of stroke patients. Stroke. 1996;27:2183–2187.
Minaire P. Immobilization osteoporosis: a review. Clin Rheumatol. 1989;8:95–103.
Mulley G, Espley AJ. Hip fracture after hemiplegia. Postgrad Med J. 1979;55:264–265.
Brunstrom S. Motor testing procedures in hemiplegia based on sequential recovery stages. Am J Phys Ther. 1966;46:357–375.
Slovic DM, Gundberg CM, Neer RM, Lian JB. Clinical evaluation of bone turnover by serum osteocalcin measurement in a hospital setting. J Clin Endocrinol Metab. 1984;58:228–230.
Sato Y, Maruoka H, Oizumi K. Amelioration of hemiplegia-associated osteopenia over 4 years following stroke by 1 α-hydroxyvitamin D3 and calcium supplementation. Stroke. 1997;28:736–739.