Abstract
Background:
Osteoporotic fractures occur in almost half of patients with a spinal cord injury (SCI) and are associated with significant morbidity and excess mortality. Paralyzed Veterans Administration (PVA) guidelines suggest that adequate calcium and vitamin D intake is important for skeletal health, however, the association of these supplements with osteoporotic fracture risk is unclear.
Objectives:
To determine the association of filled prescriptions for calcium and vitamin D with fracture risk in Veterans with an SCI.
Methods:
The 5897 persons with a traumatic SCI of at least 2 years’ duration (96% male; 4% female) included in the VSSC SCI/D Registry in FY2014 were followed from FY2014 to FY2020 for incident upper and lower extremity fractures. Filled daily prescriptions for calcium or vitamin D supplements for ≥6 months with an adherence ≥80% were examined.
Results:
Filled prescriptions for calcium (hazard ratio [HR] 0.65; 95% CI, 0.54-0.78) and vitamin D (HR 0.33; 95% CI, 0.29-0.38) supplements were associated with a significantly decreased risk for incident fractures.
Conclusion:
Calcium and vitamin D supplements are associated with decreased risk of fracture, supporting PVA guidelines that calcium and vitamin D intake are important for skeletal health in persons with an SCI.
Keywords: calcium and vitamin D supplements, fractures, spinal cord injury
Introduction
It is estimated that 25% to 46% of persons with spinal cord injury (SCI) will experience a fracture during their lifetime.1 These fractures are associated with significant morbidity and a more than three-fold increased risk of death in older men with complete injuries.2
In 2022, the Paralyzed Veterans of America (PVA) convened an expert task force to develop a clinical practice guideline (CPG) for management of osteoporosis in persons with a SCI: the Consortium for Spinal Cord Medicine Clinical Practice Guidelines: Bone Health and Osteoporosis Management in Individuals with Spinal Cord Injury. Consideration of optimal calcium and vitamin D supplementation for skeletal health was one focus area of this CPG. The CPG recommended calcium supplementation as a combination of food and supplements (with preference for food over supplements); men who are 19 to 70 and premenopausal women who are 19 to 50 should receive 1000 mg/day and women over age 50 and men 71 and older should receive 1000 to 1200 mg/day, unless they are hypercalcemic (grade 1B). However, as this CPG acknowledges, these recommendations were largely informed by reports from the able-bodied population due to the lack of data specific to persons with an SCI.3 Despite the paucity of information on the efficacy of calcium and vitamin D supplements for fracture prevention, persons with an SCI are likely to receive prescription calcium supplements after a fracture.4 Therefore, to address the need to understand the relationship of calcium and vitamin D supplements to fracture risk in persons with an SCI, we conducted a retrospective cohort study including all Veterans in the Veterans Health Administration SCI/D Registry (VHA SCIDR) from October 1, 2013 (start of fiscal year [FY] 2014) through September 30, 2020, and determined the association of filled prescriptions for calcium and vitamin D supplements with incident fractures.
Methods
Participants
We included adult (≥18 years) Veterans with a chronic (duration of injury ≥2 years) traumatic SCI in the VHA SCIDR at the start of FY2014; we followed them over the study period of FY2014– FY2020. The VHA SCIDR is a clinical and operational administrative database comprised of Veterans with spinal cord injuries/disorders (SCI/D) receiving care at the VHA SCI/D System of Care.5 The VHA SCIDR cases used in the present study were identified through the VHA SCIDR informatics phenotype platform managed by the VHA SCI/D National Program Office and the VHA Service Support Center (VSSC).5 Only Veterans with complete demographic and SCI-related data were selected. The VHA SCIDR VSSC algorithm platform was accessed on July 13, 2022, and Veterans from the FY2021 (meeting criteria as of October 1, 2012 through September 30, 2021) cumulative cohort were identified.5 The eligibility start date of the VHA SCIDR VSSC algorithm platform was October 1, 2012, and the algorithm has been reported.5 The selection of Veterans for inclusion is illustrated in Figure 1; 5897 Veterans with traumatic SCI were identified through the VHA SCIDR and were used for the present study. This project was reviewed and approved by the Edward Hines Jr. VA Hospital Institutional Review Board.
Figure 1.
Selection of Veterans. AIS = American Spinal Injury Association Impairment Scale; SCI/D = spinal cord injuries/disorders; VHA SCIDR = Veterans Health Administration SCI/D Registry.
Outcomes
An incident fracture was defined as an International Statistical Classification of Diseases and Related Health Problems, 9th revision (ICD-9) or 10th revision (ICD-10), code for an upper or lower extremity fracture. A fracture was considered incident (i.e., a new fracture that occurred during the study period) if there were no encounters with the same three-digit ICD-9 codes within a 120-day time period prior to the identified fracture.6 If there was more than one fracture during the study period, only the first fracture was included in the analyses. If upper and lower extremity fractures occurred on the same day, the Veteran was considered to have an incident lower extremity fracture.
Predictors
Calcium supplements
Filled prescriptions for calcium supplements were obtained from the outpatient pharmacy local and national drug files of Corporate Data Warehouse (CDW). The calcium salts in these supplements were converted to elemental calcium in milligrams by multiplying the dose of the calcium salt by the percentage of elemental calcium within the salt. Only those with at least 6 months (180 days) of continuous filled scripts at an adherence level of at least 80% were included in these analyses as having filled calcium prescriptions, with the average daily dose of calcium calculated over this period. The referent category included those with no filled prescriptions for calcium supplements or filled prescriptions for less than 6 months and/or at an adherence level of less than 80%. An overall analysis of any filled calcium prescription compared with the referent category was done. Within the filled calcium prescription category, dose subgroups of (i) >0 mg to 500 mg, (ii) >500 mg to 1000 mg, or (iii) >1000 mg/ day were compared to the referent category.
Vitamin D supplements
Vitamin D supplements were determined as filled medications for a vitamin D-containing supplement in the outpatient local and national drug files of CDW. The vitamin D content in these supplements was converted to International Units (IU) of vitamin D/day. Only those with at least 6 months (180 days) of continuous filled scripts at an adherence level of at least 80% were included in these analyses, with the average daily dose of vitamin D calculated over this period. The referent category included those with no filled prescriptions for vitamin D supplements or filled prescriptions for less than 6 months and/or at an adherence level of less than 80%. An overall analysis of any filled vitamin D prescription compared with the referent category was done. Within the filled vitamin D prescription category, dose subgroups of (i) >0 to ≤2000 IU, (ii) >2000 IU to ≤4000 IU, and (iii) >4000 IU/day were compared to the referent category.
A separate category of 50,000 IU of ergocalciferol as a single dose was examined without regard to number of scripts or any adherence consideration. The referent group for this high-dose vitamin D group included those with no filled prescriptions for any vitamin D supplements during the study period.
Covariates
Demographics (age, sex, race, ethnicity) and SCI-related data (duration of injury, level of injury, American Spinal Injury Association Impairment Scale [AIS]) were obtained at baseline using SCIDR data. The Charlson Comorbidity Index (CCI) was calculated using the inpatient and outpatient files of CDW and was obtained from the date of the incident fracture, end of the study period, or date of death, whichever was the most recent. Weight was obtained from the vital sign files of CDW and queried at the beginning of each fiscal year. Prevalent fracture was defined as an ICD-9 or ICD-10 code in the inpatient or outpatient files for an upper or lower extremity fracture in the 5 years prior to the start of the study. Filled prescriptions for anticonvulsants, benzodiazepines, opioids, sedatives/hypnotics, corticosteroids, antidepressants, and thiazides were defined as 6 months of continuous filled prescriptions from filled prescription data from outpatient local and national drug and intravenous (IV) drug files in CDW prior to the censor date.
Statistical analysis
In primary analysis, baseline characteristics (including demographics, SCI-related characteristics, comorbidities, and medication data) for all Veterans with SCI/D with and without a fracture were compared. T tests were used to compare continuous variables between the groups with and without fracture. Chi-square tests were used for categorical variable comparisons between the groups with and without fracture. Cox proportional hazards regression analysis was performed to obtain univariate and multivariate hazard ratios. In multivariate analyses, a backwards selection approach using the Akaike Information Criterion was used to determine the best fit model.
Secondary analyses were performed to examine incident fractures stratified by site including upper extremity and lower extremity fractures. Baseline characteristics were summarized using analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. Cox proportional regression analysis was performed to obtain univariate and multivariate hazard ratios, comparing Veterans with upper extremity versus no fractures and lower extremity versus no fractures.
An additional sensitivity analysis was performed using Cox proportional regression analysis to compare Veterans who had any use of 50,000 IU of ergocalciferol to those who did not have any filled prescriptions for vitamin D supplements with respect to any incident upper or lower extremity fracture.
Statistical significance was set at p < .05. All analyses were performed using R version 4.1.2.
Results
There were 5897 adult Veterans with chronic, traumatic SCI in the VHA SCIDR (using the complete longitudinal cohort starting from the start date of the VHA SCIDR VSSC algorithm platform, which was October 1, 2012 through September 30, 2021) with complete demographic and SCI-related data at the start of FY2014. There were 765 patients who had filled prescriptions for calcium supplements (13%) and 2295 patients who had filled prescriptions for vitamin D supplements (38.9%) for at least 6 months at a level of expected medication possession of at least 80%. There were 1593 (27%) patients who had at least one filled prescription for 50,000 IU (i.e., high-dose vitamin D).
Twenty-one percent (n = 1240) of Veterans had at least one incident fracture between FY2014 and FY2020. Baseline characteristics for Veterans with and without a fracture are shown in Table 1. There were statistically significant differences between the groups in regard to age; sex; race; duration of injury; level of injury; AIS; CCI; prevalent fracture; and use of anticonvulsants, benzodiazepines, opioids, sedatives/hypnotics, corticosteroids, anti depressants, calcium supplements, and vitamin D supplements (p < .05 for all). There were no statistically significant differences between the groups regarding change in weight, body mass index (BMI), use of thiazides, and filled prescriptions for 50,000 IU of vitamin D (p ≥ .22 for all).
Table 1.
Baseline characteristics of study population by any incident fracture status
| Characteristics | Fracture (n = 1240) | No fracture (n = 4657) | p value |
|---|---|---|---|
| Age, years, mean ± SD | 56.5 ± 12.2 | 55.9 ± 12.2 | <.01 |
| Sex, n (%) | <.01 | ||
| Male | 1177 (95) | 4498 (97) | |
| Female | 63 (5) | 159 (3) | |
| Race, n (%) | <.01 | ||
| American Indian or Alaskan Native | 14 (1) | 38 (0.8) | |
| Asian | 3 (0.2) | 30 (0.6) | |
| Black or African American | 211 (17) | 1002 (22) | |
| Native Hawaiian or Other Pacific Islander | 7 (.5) | 52 (1) | |
| White | 996 (80) | 3488 (75) | |
| Multiple | 7 (0.5) | 52 (1) | |
| Change in weight, kg, mean ± SD | -0.83 ± 19.0 | -1.62 ± 23.4 | .22 |
| Body mass index, mean ± SD | 27.2 ± 6.27 | 27.3 ± 6.14 | .66 |
| Duration of injury, years, mean ± SD | 22.4 ± 13.9 | 21.6 ± 13.7 | .03 |
| Level of injury, n (%) | <.01 | ||
| Cervical | 537 (43) | 2290 (49) | |
| Thoracic | 585 (47) | 1813 (39) | |
| Lumbar | 109 (9) | 491 (11) | |
| Sacral | 9 (0.7) | 63 (1) | |
| AIS, n (%) | <.01 | ||
| A | 636 (51) | 1788 (38) | |
| B | 120 (10) | 480 (10) | |
| C | 136 (11) | 625 (13) | |
| D | 340 (27) | 1720 (37) | |
| E | 8 (0.6) | 44 (1) | |
| CCI, mean ± SD | 2.92 ± 2.21 | 2.28 ± 2.01 | <.01 |
| Prevalent fracture, n (%) | 15 (1) | 1225 (26) | <.01 |
| Medications, n (%) | |||
| Anticonvulsants | 770 (62) | 2427 (52) | <.01 |
| Benzodiazepines | 531 (43) | 1644 (35) | <.01 |
| Opioids | 951 (77) | 2972 (64) | <.01 |
| Sedatives/Hypnotics | 343 (28) | 986 (21) | <.01 |
| Corticosteroids | 374 (30) | 1103 (24) | <.01 |
| Antidepressants | 771 (62) | 2369 (51) | <.01 |
| Thiazides | 358 (29) | 1295 (28) | .48 |
| Filled prescriptions for calcium, n (%) | <.01 | ||
| No filled prescriptions or nonadherent for at least 80% for ≥6 months | 1112 (90) | 4020 (86) | |
| At least 80% adherent for ≥6 months | |||
| >0 mg and ≤500 mg | 100 (8) | 543 (12) | |
| >500 mg and ≤1000 mg | 25 (2) | 87 (2) | |
| >1000 mg | 3 (0.2) | 7 (0.2) | |
| Filled prescriptions for vitamin D, n (%) | <.01 | ||
| No filled prescriptions or nonadherent | 979 (79) | 2623 (56) | |
| At least 80% adherent for ≥6 months | |||
| >0 IU and ≤2000 IU | 242 (20) | 1883 (40) | |
| >2000 IU and ≤4000 IU | 19 (1) | 151 (3) | |
| >4000 IU | 0 (0) | 0 (0) | |
| Filled prescriptions for 50,000 IU of Vitamin D, n (%) | 345 (28) | 1248 (27) | .49 |
Note: AIS = American Spinal Injury Association Impairment Scale; CCI = Charlson Comorbidity Index; IU = International Units.
Baseline characteristics for Veterans stratified by upper extremity, lower extremity, and no fractures are shown in Table 2. There were statistically significant differences between the groups with respect to age; sex; race; duration of injury; level of injury; AIS; CCI; prevalent fracture; use of anticonvulsants, benzodiazepines, opioids, sedatives/hypnotics, corticosteroids, and antidepressants; filled prescriptions for calcium; and filled prescriptions for vitamin D groups (p < .05 for all). There were no statistically significant differences between the groups with respect to age, change in weight, use of thiazides, and filled prescriptions of 50,000 IU of vitamin D groups (p > .25 for all).
Table 2.
Baseline characteristics of study population by incident fracture status and location
| Characteristics | Upper extremity fracture (n = 199) | Lower extremity fracture (n = 1041) | No fracture (n = 4657) | p value |
|---|---|---|---|---|
| Age, years, mean ± SD | 55.5 ± 12.2 | 56.7 ± 12.3 | 55.9 ± 12.3 | <.01 |
| Sex, n (%) | ||||
| Male | 189 (95) | 988 (95) | 4498 (97) | .02 |
| Female | 10 (5) | 53 (5) | 159 (3) | |
| Race, n (%) | .004 | |||
| American Indian or Alaskan Native | 2 (1) | 12 (1) | 38 (0.8) | |
| Asian | 2 (1) | 30 (3) | 1 (0.02) | |
| Black or African American | 44 (22) | 167 (16) | 1002 (22) | |
| Native Hawaiian or Other Pacific Islander | 0 (0) | 9 (0.8) | 47 (1) | |
| White | 150 (75) | 846 (81) | 3488 (75) | |
| Multiple | 1 (0.5) | 6 (0.6) | 52 (1) | |
| Change in weight, kg ± SD | 1.01 ± 17.1 | -1.18 ± 17.1 | -1.62 ± 23.4 | .25 |
| Body mass index, mean ± SD | 27.7 ± 6.76 | 27.1 ± 6.17 | 27.3 ± 6.14 | .40 |
| Duration of injury, years, mean ± SD | 18.8 ± 12.9 | 23.1 ± 14.0 | 21.6 ± 13.7 | <.01 |
| Level of injury, n (%) | <.01 | |||
| Cervical | 108 (54) | 429 (41) | 2290 (49) | |
| Thoracic | 62 (31) | 523 (50) | 1813 (39) | |
| Lumbar | 24 (12) | 85 (8) | 491 (11) | |
| Sacral | 5 (3) | 4 (0.4) | 63 (1) | |
| AIS, n (%) | <.01 | |||
| A | 49 (25) | 587 (56) | 1788 (38) | |
| B | 12 (6) | 108 (10) | 480 (10) | |
| C | 23 (12) | 113 (11) | 625 (13) | |
| D | 109 (55) | 231 (22) | 1720 (37) | |
| E | 7 (4) | 2 (0.2) | 56 (1) | |
| CCI, mean ± SD | 2.59 ± 1.86 | 2.98 ± 2.27 | 2.28 ± 2.01 | <.01 |
| Prevalent fracture, n (%) | 2 (1) | 13 (1) | 17 (0.4) | <.01 |
| Medications, n (%) | ||||
| Anticonvulsants | 139 (70) | 631 (61) | 2427 (52) | <.01 |
| Benzodiazepines | 81 (41) | 450 (43) | 1644 (35) | <.01 |
| Opioids | 172 (86) | 779 (75) | 2972 (64) | <.01 |
| Sedatives/Hypnotics | 55 (28) | 986 (95) | 288 (6) | <.01 |
| Corticosteroids | 85 (43) | 289 (28) | 1103 (24) | <.01 |
| Antidepressants | 138 (69) | 633 (61) | 2369 (51) | <.01 |
| Thiazides | 61 (31) | 297 (29) | 1295 (28) | .63 |
| Filled prescriptions for calcium, n (%) | <.01 | |||
| No filled prescriptions or nonadherent | 183 (92) | 929 (89) | 4020 (86) | |
| At least 80% adherent for ≥6 months | ||||
| >0 mg and ≤500 mg | 10 (5) | 90 (9) | 543 (12) | |
| >500 mg and ≤1000 mg | 5 (3) | 20 (2) | 87 (2) | |
| >1000 mg | 1 (0.5) | 2 (0.2) | 5 (0.1) | |
| Filled prescriptions for Vitamin D, n (%) | <.01 | |||
| No filled prescriptions or nonadherent | 154 (77) | 825 (79) | 2623 (56) | |
| At least 80% adherent for ≥6 months | ||||
| >0 IU and ≤2000 IU | 40 (20) | 202 (19) | 1883 (40) | |
| >2000 IU and ≤4000 IU | 5 (3) | 14 (1) | 151 (3) | |
| >4000 IU | — | — | — | |
| Filled prescriptions for 50,000 IU of Vitamin D, n (%) | 54 (27) | 291 (28) | 1248 (27) | .75 |
Note: AIS = American Spinal Injury Association Impairment Scale; CCI = Charlson Comorbidity Index; IU = International Units.
Relationship of filled prescriptions for calcium supplements to incident fractures
Table 3 shows the univariate and multivariate Cox hazard regression for incident fractures including calcium supplements. In multivariate analyses, Black race was associated with a significantly decreased risk of incident fracture (hazard ratio [HR] 0.73; 95% CI, 0.63-0.85; p < .001) compared to White race. AIS score was not associated with fracture risk (p ≥ .11 for all). Higher CCI score was associated with an increased risk for fracture (HR 1.10; 95% CI, 1.07-1.13; p < .001). Having a prevalent fracture was associated with nearly a 3-fold risk for incident fracture (HR 3.23; 95% CI, 1.94-5.39; p < .001). Filled prescriptions for opioids (HR 1.57; 95% CI, 1.36-1.81), corticosteroids (HR 1.19; 95% CI, 1.05-1.35), and antidepressants (HR 1.30; 95% CI, 1.15-1.47) were associated with an increased risk for fracture (p ≤ .006 for all). When comparing the referent group to all other calcium dose categories (combined as one group), filled prescriptions for calcium supplements (with at least 80% adherence) were significantly associated with a decreased risk for incident fracture (HR 0.65; 95% CI, 0.54-0.78; p < .001) in multivariate modeling. Filled prescriptions for calcium supplements at doses >0 mg to 500 mg/day were associated with a decreased risk for fracture compared to the referent category (HR 0.59; 95% CI, 0.48-0.73; p < .001). There was no statistically significant relationship between calcium supplements at doses of >500 mg to 1000 mg/day (HR 0.93; 95% CI, 0.62-1.38; p = .7) and >1000 mg/day (HR 1.10; 95% CI, 0.35-3.44; p = .90) with fracture.
Table 3.
Risk factors for any fracture by filled prescriptions for calcium supplements
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| HR | CI | p value | HR | CI | p | |
| value | ||||||
| Age | 1.00 | 0.996-1.008 | .12 | — | — | — |
| Sex | — | — | — | |||
| Male | Referent | — | — | — | ||
| Female | 1.43 | 1.11-1.85 | .005 | 1.40 | 1.08-1.81 | .010 |
| Race | ||||||
| American Indian or Alaskan Native | 1.24 | 0.73-2.10 | .40 | 1.33 | 0.78-2.26 | .3 |
| Asian | 0.38 | 0.12-1.17 | .092 | 0.51 | 0.16-1.57 | .2 |
| Black or African American | 0.76 | 0.65-0.88 | <.001 | 0.73 | 0.63-0.85 | <.001 |
| Native Hawaiian or Other Pacific | 0.69 | 0.36-1.33 | .30 | 0.71 | 0.37-1.37 | .30 |
| Islander | ||||||
| White | Referent | Referent | ||||
| Multiple | 0.51 | 0.24-1.08 | .087 | 0.52 | 0.25-1.11 | .090 |
| Change in weight | 1.00 | 0.999-1.004 | .26 | — | — | — |
| Body mass index | 1.00 | 0.9887-1.007 | .631 | |||
| Duration of injury | 1.00 | 0.9995-1.008 | .09 | — | — | — |
| Level of injury | — | — | — | |||
| Cervical | Referent | — | — | — | ||
| Thoracic | 1.32 | 1.18-1.49 | <.001 | — | — | — |
| Lumbar | 0.95 | 0.77-1.16 | .60 | — | — | — |
| Sacral | 0.64 | 0.33-1.23 | .20 | — | — | — |
| AIS | ||||||
| A | 1.84 | 0.92-3.70 | .80 | 1.77 | 0.88-3.57 | .11 |
| B | 1.34 | 0.64-2.74 | .40 | 1.28 | 0.62-2.63 | .50 |
| C | 1.19 | 0.66-2.74 | >.40 | 1.00 | 0.49-2.31 | .70 |
| D | 1.08 | 0.54-2.19 | .80 | 1.00 | 0.49-2.02 | .11 |
| E | Referent | Referent | ||||
| CCI | 1.12 | 1.09-1.14 | <.001 | 1.10 | 1.07-1.13 | <.001 |
| Prevalent fracture | 3.03 | 1.82-5.04 | <.001 | 3.23 | 1.94-5.39 | <.001 |
| Medications | ||||||
| Anticonvulsants | 1.44 | 1.29-1.62 | <.001 | — | — | — |
| Benzodiazepines | 1.33 | 1.19-1.48 | <.001 | — | — | — |
| Opioids | 1.76 | 1.54-2.01 | <.001 | 1.57 | 1.36-1.81 | <.001 |
| Sedatives/Hypnotics | 1.36 | 1.20-1.54 | <.001 | — | — | — |
| Corticosteroids | 1.34 | 1.19-1.51 | <.001 | 1.19 | 1.05-1.35 | .006 |
| Antidepressants | 1.52 | 1.35-1.70 | <.001 | 1.30 | 1.15-1.47 | <.001 |
| Thiazides | 1.05 | 0.93-1.18 | .50 | — | — | — |
| Filled prescriptions for calcium | ||||||
| No filled prescriptions or nonadherent | Referent | Referent | ||||
| At least 80% adherent for ≥ 6 months | 0.65 | 0.54-0.78 | <.001 | |||
| >0 mg and ≤500 mg | 0.68 | 0.56-0.84 | <.001 | 0.59 | 0.48-0.73 | <.001 |
| >500 mg and ≤1000 mg | 1.03 | 0.69-1.53 | .9 | 0.93 | 0.62-1.38 | .70 |
| >1000 mg | 1.52 | 0.49-4.71 | .5 | 1.10 | 0.35-3.44 | .90 |
| Filled prescriptions for vitamin D | ||||||
| No filled prescriptions or nonadherent | Referent | |||||
| At least 80% adherent for ≥6 months | 0.33 | 0.29-0.38 | <.001 | |||
| >0 IU and ≤2000 IU | 0.37 | 0.32-0.43 | <.001 | |||
| >2000 IU and ≤4000 IU | 0.37 | 0.23-0.58 | <.001 | |||
| >4000 IU | 0 | 0 | ||||
| Filled prescriptions for 50,000 IU of Vitamin D | 1.05 | 0.92-1.19 | .5 | |||
Note: AIS = American Spinal Injury Association Impairment Scale; CCI = Charlson Comorbidity Index; HR = hazard ratio; IU = International Units.
Results were similar for the relationship of calcium supplementation to incident fracture risk in site specific models confined to upper extremity fracture compared with no fracture and to lower extremity fracture compared with no fracture. Calcium supplements at doses >0 mg to ≤500 mg were associated with a decreased risk of fracture (upper extremity HR 0.36; 95% CI, 0.19-0.69; p = .002; lower extremity HR 0.64; 95% CI, 0.52-0.80; p < .001). There was no statistically significant relationship between calcium supplements at doses of >500 mg to 1000 mg and >1000 mg with fracture (p ≥ .6 for both).
Relationship of filled prescriptions for vitamin D supplements to incident fractures
Table 4 shows the univariate and multivariate Cox hazard regression models for incident fractures including vitamin D supplements. In multivariate analyses, Black race was associated with a significantly decreased risk of incident fracture (HR 0.75; 95% CI, 0.65-0.88; p < .001) compared to being of White race. AIS score was not associated with fracture risk (p ≥ .079 for all). Higher CCI score was associated with a significantly increased risk for lower extremity fracture (HR 1.10; 95% CI, 1.07-1.13; p < .001). Having a prevalent fracture was associated with a 3-fold risk for incident fracture (HR 3.17; 95% CI, 2.02-5.62; p < .001). Filled prescriptions for anticonvulsants (HR 1.25; 95% CI, 1.09-1.42), opioids (HR 1.52; 95% CI, 1.31-1.77), corticosteroids (HR 1.23; 95% CI, 1.08-1.39), and antidepressants (HR 1.32; 95% CI, 1.17-1.50) were associated with an increased risk for fracture (p < .001 for all). When comparing the referent group to all other vitamin D dose categories (combined as one group), filled prescriptions for vitamin D (with at least 80% adherence) were significantly associated with a decreased risk for incident fracture (HR 0.33; 95% CI, 0.29-0.38; p < .001) in multivariate modeling. Vitamin D supplements at doses >0 IU to ≤2000 IU was associated with a 3-fold decreased risk for fracture compared to not being on any vitamin D supplementation (HR 0.33; 95% CI, 0.29-0.38; p < .001). Furthermore, filled prescriptions for vitamin D supplements at doses of >2000 IU and ≤4000 IU were also associated with a 3-fold decreased risk for fracture (HR 0.32; 95% CI, 0.20-0.50; p < .001).
Table 4.
Risk factors for any fracture by filled prescriptions for vitamin D supplements
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| HR | CI | p value | HR | CI | p value | |
| Age | 1.00 | 0.996-1.008 | .12 | — | — | — |
| Sex | — | — | — | |||
| Male | Referent | |||||
| Female | 1.43 | 1.11-1.85 | .005 | |||
| Race | ||||||
| American Indian or Alaskan | 1.24 | 0.73-2.10 | .40 | 1.38 | 0.81-2.34 | .20 |
| Native | ||||||
| Asian | 0.38 | 0.12-1.17 | .092 | 0.50 | 0.16-1.54 | .20 |
| Black or African American | 0.76 | 0.65-0.88 | <.001 | 0.75 | 0.65-0.88 | <.001 |
| Native Hawaiian or Other Pacific | 0.69 | 0.36-1.33 | .3 | 0.80 | 0.41-1.54 | .5 |
| Islander | ||||||
| White | Referent | Referent | ||||
| Multiple | 0.51 | 0.24-1.08 | .087 | 0.50 | 0.24-1.06 | .070 |
| Change in weight | 1.00 | 0.999-1.004 | .26 | — | — | — |
| Body mass index | 1.00 | 0.9887-1.007 | .631 | |||
| Duration of injury | 1.00 | 0.9995-1.008 | .09 | — | — | — |
| Level of injury | ||||||
| Cervical | Referent | Referent | — | — | ||
| Thoracic | 1.32 | 1.18-1.49 | <.001 | — | — | — |
| Lumbar | 0.95 | 0.77-1.16 | .6 | — | — | — |
| Sacral | 0.64 | 0.33-1.23 | .20 | — | — | — |
| AIS | ||||||
| A | 1.84 | 0.92-3.70 | .80 | 1.88 | 0.93-3.80 | .079 |
| B | 1.34 | 0.64-2.74 | .40 | 1.36 | 0.66-2.80 | .40 |
| C | 1.19 | 0.66-2.74 | >.40 | 1.19 | 0.58-2.45 | .60 |
| D | 1.08 | 0.54-2.19 | .80 | 1.05 | 0.52-2.14 | .90 |
| E | Referent | Referent | ||||
| CCI | 1.12 | 1.09-1.14 | <.001 | 1.10 | 1.07-1.13 | <.001 |
| Prevalent fracture | 3.03 | 1.82-5.04 | <.001 | 3.17 | 2.02-5.62 | <.001 |
| Medications | ||||||
| Anticonvulsants | 1.44 | 1.29-1.62 | <.001 | 1.25 | 1.09-1.42 | <.001 |
| Benzodiazepines | 1.33 | 1.19-1.48 | <.001 | — | — | — |
| Opioids | 1.76 | 1.54-2.01 | <.001 | 1.52 | 1.31-1.77 | <.001 |
| Sedatives/Hypnotics | 1.36 | 1.20-1.54 | <.001 | — | — | — |
| Corticosteroids | 1.34 | 1.19-1.51 | <.001 | 1.23 | 1.08-1.39 | <.001 |
| Antidepressants | 1.52 | 1.35-1.70 | <.001 | 1.32 | 1.17-1.50 | <.001 |
| Thiazides | 1.05 | 0.93-1.18 | .50 | — | — | — |
| Filled prescriptions for calcium supplements | ||||||
| No filled prescriptions or nonadherent | Referent | |||||
| At least 80% adherent for ≥6 months | 0.65 | 0.54-0.78 | <.001 | |||
| >0 mg and ≤500 mg | 0.68 | 0.56-0.84 | <.001 | |||
| >500 mg and ≤1000 mg | 1.03 | 0.69-1.53 | .9 | |||
| >1000 mg | 1.52 | 0.49-4.71 | .5 | |||
| Filled prescriptions for vitamin D supplements | ||||||
| No filled prescriptions or nonadherent | Referent | Referent | ||||
| At least 80% adherent for ≥ 6 months | 0.33 | 0.29-0.38 | <.001 | |||
| >0 IU and ≤2000 IU | 0.37 | 0.32-0.43 | <.001 | 0.33 | 0.29-0.38 | <.001 |
| >2000 IU and ≤4000 IU | 0.37 | 0.23-0.58 | <.001 | 0.32 | 0.20-0.50 | <.001 |
| >4000 IU | 0 | 0 | 0 | 0 | ||
| Filled prescriptions for 50,000 IU of Vitamin D | 1.05 | 0.92-1.19 | .5 | |||
Note: AIS = American Spinal Injury Association Impairment Scale; CCI = Charlson Comorbidity Index; HR = hazard ratio; IU = International Units.
Findings were similar for lower extremity fractures with vitamin D supplements. Doses >0 IU to ≤2000 IU were associated with a 3-fold decreased risk for fracture compared to the referent group (HR 0.33; 95% CI, 0.28-0.38; p < .001) in multivariate analyses. Additionally, filled prescriptions for vitamin D supplements of >2000 IU and ≤4000 IU were associated with decreased risk for lower extremity fractures (HR 0.27; 95% CI, 0.16-0.46; p < .001). For upper extremity fractures only low doses of vitamin D supplements >0 IU to ≤2000 IU were associated with a 3-fold decreased risk for fracture compared to the referent group (HR 0.31; 95% CI, 0.22-0.44; p < .001.
In multivariable models examining any fractures, use of vitamin D at doses of 50,000 IU was significantly associated with a decreased risk for incident fractures compared to no filled prescriptions for any vitamin D supplement (HR 0.66; 95%, 0.57-0.77; p < .001).
Discussion
In Veterans with a chronic traumatic SCI, filled prescriptions for calcium and vitamin D supplements for at least 6 months at a high level of adherence were associated with a decreased risk for fractures, including both upper and lower extremity fractures. There was a dose relationship between calcium supplementation and fracture risk, with doses up to 500 mg at a time, and not higher doses, associated with reduced fracture risk. This dose relationship may be because the percentage of calcium absorbed depends not only on the source of calcium but also on the total amount of elemental calcium consumed at one time; as the amount increases, the percentage absorbed decreases. Absorption from supplements is highest with doses of 500 mg or less.7 Calcium and vitamin D supplementation at this level of adherence was infrequent, as filled prescription data suggested that less than 15% were adherent to calcium supplementation and less than half to vitamin D supplementation for at least 6 months.
Prior reports of the association of calcium and vitamin D with fracture risk in those with a traumatic SCI are confined to bone mineral density (BMD) and do not include fractures as an outcome.8 Others only examine levels of 25-hydroxyvitamin D and not pharmacological vitamin D supplementation.9 In contrast with our report, a cross-sectional study in Scandinavian men with traumatic SCI found that neither calcium nor vitamin D supplements were independently associated with BMD.8 In a prospective 1-year study including persons with traumatic SCI, suboptimal 25-hydroxyvitamin D levels were not significantly associated with fall-induced fractures.9 In support of our findings, a randomized clinical trial reported that administration of a vitamin D analog (1-alpha-hydroxyvitamin D2) to 40 subjects with chronic complete motor SCI for 24 months significantly increased leg BMD on dual-energy x-ray absorptiometry (DXA).10 It is difficult to discern whether or not the addition of calcium and vitamin D supplementation to pharmacological therapies for osteoporosis in randomized controlled trials in sublesional osteoporosis might have affected the study findings, as vitamin D was inconsistently measured and replaced among these trials.11
Mechanistically, it may be that the favorable association of calcium and vitamin D supplementation with fracture risk in our cohort reflects lower bone turnover in these individuals. In support of this, Hummel et al.12 reported that in persons with a traumatic SCI, the threshold of 25-hydroxyvitamin D needed for suppression of intact PTH is approximately 94 nmol/L and that secondary hyperparathyroidism is associated with elevated bone resorption. Total calcium intake is low in persons with a traumatic SCI, with most not meeting dietary reference intake recommendations; suboptimal levels of 25-hydroxyvitamin D are present in the majority of patients with acute and chronic SCI.13-15 It is possible that calcium and vitamin D supplements filled by this cohort suppressed PTH, which in turn slowed the rate of bone turnover and protected against fracture.
Similar to our findings that low doses of vitamin D supplements may be beneficial for fracture prevention in persons with a traumatic SCI, we found that intermittent use of high-dose vitamin D of 50,000 IU was also associated with decreased fracture risk. These results are in contrast to a meta-analysis in the able-bodied population, suggesting that high-dose vitamin D supplementation in older adults is ineffective in preventing fractures.16 This is an area that requires further study as one study suggested that some providers prescribe intermittent high (“mega”) doses of vitamin D.17 It may be high-dose vitamin D supplements are being prescribed to correct hypovitaminosis D, as our previous work has indicated that hypovitaminosis D is quite common in this population.18
There is substantial interest in understanding the value of calcium and vitamin D supplementation for bone health on the part of both patients and clinicians. In semistructured interviews of Veterans with traumatic and nontraumatic SCI/D who had experienced at least one fracture of the femur, tibia/fibula, or hip in the prior 18 months, a major theme that emerged was that patients were very interested in further research on the role of calcium and vitamin D supplements, specifically whether these were effective at reducing fracture risk.19 VA providers reported that they frequently prescribe calcium and vitamin D supplements for sublesional osteoporosis.17
This study has several important strengths. To date, this study includes the largest cohort of Veterans with traumatic SCI to examine the relationship of prescriptions for calcium and vitamin D supplements with incident fractures. Important covariates were considered in analyses including overall health as measured by the CCI. We were able to include both upper and lower extremity fractures and different dosing regimens for calcium and vitamin D supplements.
We also recognize a few limitations to our study. To start, we did not have information on dietary sources of calcium or vitamin D. This is particularly important, as it would be expected that calcium and vitamin D supplementation would be most effective in those with lower dietary intakes, as it is likely that total supplements (dietary plus supplements) are what is responsible for efficacy. We also did not have information on 25-hydroxyvitamin D levels. Serum levels of 25-hydroxyvitamin D largely reflect production from the skin upon exposure to ultraviolet B (UVB) or direct supplementation, as there are few dietary sources of vitamin D.20 Secondly, using administrative databases, we can only know if a prescription was filled, not if it was taken as prescribed. We did not include use of over-the-counter calcium or vitamin D supplements, calcium and or vitamin D supplements that might have been prescribed outside the VHA, or information on fractures outside the VHA as we did not have access to this information. We also did not include information on physical activity levels, diet, or lifestyle practices such as tobacco and alcohol use as these were not uniformly collected within the SCIDR. Approximately 30% of Veterans with traumatic and nontraumatic SCI/D are dual users of VHA and Medicare services.21 Finally, the VHA began use of ICD-10 codes in fiscal year 2015; however, these ICD-10 fracture codes have not been validated for persons with an SCI receiving care at the VHA. We did not examine any safety issues related to calcium and vitamin D supplementation in particular, constipation, and renal/bladder stones, which are particularly problematic for this population.22-24
In conclusion, our findings support the Consortium for Spinal Cord Medicine Clinical Practice Guidelines: Bone Health and Osteoporosis Management in Individuals with Spinal Cord Injury in that they suggest that calcium and vitamin D are important for skeletal health in persons with a traumatic SCI.3 Further studies examining the relationship of dietary intakes of calcium and vitamin D with fracture risk, the optimal total doses of calcium and vitamin D for fracture prevention, the safety of calcium and vitamin D supplements in persons with a traumatic SCI, and these considerations in the nontraumatic SCI/D population are needed.
Acknowledgments
The authors are grateful to Dr. Jennifer Sippel for her assistance with the VSSC SCI/D Registry (SCIDR).
Funding Statement
Financial Support This project was funded by the American Legion, 2020. Brian Le received the VHA VISN7 Research Development Award through Veterans Affairs through the Charlie Norwood VA Medical Center, Augusta Biomedical Research Corporation. Hongyan Xu received funding from VHA VISN7, Veterans Affairs, through the Charlie Norwood VA Medical Center.
Footnotes
Conflicts of Interest
The authors declare no conflicts of interest.
REFERENCES
- 1.Champs APS, Maia GAG, Oliveira FG, de Melo GCN, Soares MMS. Osteoporosis-related fractures after spinal cord injury: A retrospective study from Brazil. Spinal Cord. 2020;58(4):484–489. doi: 10.1038/s41393-019-0387-9. [DOI] [PubMed] [Google Scholar]
- 2.Carbone LD, Chin AS, Burns SP, et al. Mortality after lower extremity fractures in men with spinal cord injury. J Bone Miner Res. 2014;29(2):432–9. doi: 10.1002/jbmr.2050. [DOI] [PubMed] [Google Scholar]
- 3.Paralyzed Veterans of America Bone Health and Osteoporosis Management in Individuals with Spinal Cord Injury; Clinical Practice Guidelines for Health Care Providers. 2022 [Google Scholar]
- 4.Akhigbe T, Chin AS, Svircev JN, et al. A retrospective review of lower extremity fracture care in patients with spinal cord injury. J Spinal Cord Med. 2015;38(1):2–9. doi: 10.1179/2045772313Y.0000000156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sippel JL, Daly JE, Poggensee L, et al. Modernization of a large spinal cord injuries and disorders registry: The Veterans Administration experience. Arch Rehabil Res Clin Transl. 2022;4(4):100237. doi: 10.1016/j.arrct.2022.100237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Logan WC, Jr., Sloane R, Lyles KW, Goldstein B, Hoenig HM. Incidence of fractures in a cohort of veterans with chronic multiple sclerosis or traumatic spinal cord injury. Arch Phys Med Rehabil. 2008;89(2):237–43. doi: 10.1016/j.apmr.2007.08.144. [DOI] [PubMed] [Google Scholar]
- 7.US Standing Committee on the Scientific Evaluation of Dietary Reference Intakes Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. Washington, DC: National Academies Press; 1997. [PubMed] [Google Scholar]
- 8.Kokorelis C, Gonzalez-Fernandez M, Morgan M, Sadowsky C. Effects of drugs on bone metabolism in a cohort of individuals with traumatic spinal cord injury. Spinal Cord Ser Cases. 2019;5:3. doi: 10.1038/s41394-018-0146-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Jorgensen V, Slettahjell HB, Skavberg Roaldsen K, Kostovski E. Carboxy terminal collagen crosslinks as a prognostic risk factor for fall-related fractures in individuals with established spinal cord injury. Spinal Cord. 2019;57(11):985–991. doi: 10.1038/s41393-019-0322-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bauman WA, Spungen AM, Morrison N, Zhang RL, Schwartz E. Effect of a vitamin D analog on leg bone mineral density in patients with chronic spinal cord injury. J Rehabil Res Dev. 2005;42(5):625–34. doi: 10.1682/jrrd.2004.11.0145. [DOI] [PubMed] [Google Scholar]
- 11.Soleyman-Jahi S, Yousefian A, Maheronnaghsh R, et al. Evidence-based prevention and treatment of osteoporosis after spinal cord injury: A systematic review. Eur Spine J. 2018;27(8):1798–1814. doi: 10.1007/s00586-017-5114-7. [DOI] [PubMed] [Google Scholar]
- 12.Hummel K, Craven BC, Giangregorio L. Serum 25(OH) D, PTH and correlates of suboptimal 25(OH)D levels in persons with chronic spinal cord injury. Spinal Cord. 2012;50(11):812–6. doi: 10.1038/sc.2012.67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Miyatani M, CBC. Loewenberger E., MCFa, J.D A The dietary intakes of calcium and bone health related nutrients among individuals with and without spinal cord injury. J Nutr Ther. 2014;2014(3):103–113. doi: 10.6000/1929-5634.2014.03.02.9. [DOI] [Google Scholar]
- 14.Koutrakis NE, Goldstein RL, Walia P, et al. Vitamin D, diet, and lifestyle in a chronic SCI population. Spinal Cord. 2019;57(2):117–127. doi: 10.1038/s41393-018-0148-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Javidan AN, Sabour H, Latifi S, et al. Calcium and vitamin D plasma concentration and nutritional intake status in patients with chronic spinal cord injury: A referral center report. J Res Med Sci. 2014;19(9):881–884. [PMC free article] [PubMed] [Google Scholar]
- 16.Zheng YT, Cui QQ, Hong YM, Yao WG. A meta-analysis of high dose, intermittent vitamin D supplementation among older adults. PLoS One. 2015;10(1):e0115850. doi: 10.1371/journal.pone.0115850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Weaver FM, Etingen B, Guihan M, et al. Spinal cord injury providers' perspectives on managing sublesional osteoporosis. J Spinal Cord Med. 2019:1–7. doi: 10.1080/10790268.2019.1704552. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Le B, Ray C, Gonzalez B, et al. Laboratory evaluation of secondary causes of bone loss in Veterans with spinal cord injury and disorders. Osteoporos Int. 2019;30(11):2241–2248. doi: 10.1007/s00198-019-05089-1. [DOI] [PubMed] [Google Scholar]
- 19.Etingen B, Carbone LD, Guihan M, et al. Lower extremity fracture prevention and management in persons with spinal cord injuries and disorders: The patient perspective. J Spinal Cord Med. 2021:1–11. doi: 10.1080/10790268.2021.1907675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wacker M, Holick MF. Sunlight and Vitamin D: A global perspective for health. Dermatoendocrinol. 2013;5(1):51–108. doi: 10.4161/derm.24494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hatch MN, Etingen B, Raad J, Siddiqui S, Stroupe KT, Smith BM. Dual utilization of Medicare and VA outpatient care among Veterans with spinal cord injuries and disorders. J Spinal Cord Med. 2022:1–9. doi: 10.1080/10790268.2022.2027321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Craven BC, Cirnigliaro CM, Carbone LD, Tsang P, Morse LR. The pathophysiology, identification and management of fracture risk, sublesional osteoporosis and fracture among adults with spinal cord injury. J Pers Med. 2023;13(6) doi: 10.3390/jpm13060966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Welk B, Fuller A, Razvi H, Denstedt J. Renal stone disease in spinal-cord-injured patients. J Endourol. 2012;26(8):954–9. doi: 10.1089/end.2012.0063. [DOI] [PubMed] [Google Scholar]
- 24.Glickman S, Kamm MA. Bowel dysfunction in spinal-cord-injury patients. Lancet. 1996;347(9016):1651–1653. doi: 10.1016/s0140-6736(96)91487-7. [DOI] [PubMed] [Google Scholar]