Abstract
Background/Objectives:
To assess whether vitamin D supplementation prevents specific fall subtypes and sequelae (e.g., fracture).
Design:
Secondary analyses of STURDY (Study to Understand Fall Reduction and Vitamin D in You)—a response-adaptive, randomized clinical trial.
Setting:
Two community-based research units.
Participants:
688 participants ≥70 years old with elevated fall risk and baseline serum 25-hydroxyvitamin D levels of 10–29 ng/mL.
Intervention:
200IU/d (control), 1000IU/d, 2000IU/d or 4000IU/d of vitamin D3.
Measurements:
Outcomes included repeat falls and falls that were consequential, were injurious, resulted in emergency care, resulted in fracture, and occurred either indoors or outdoors.
Results:
After adjustment for multiple comparisons, the risk of fall-related fracture was greater in the pooled higher doses (≥1000IU/d) group compared to control (hazard ratio [HR]=2.66; 95% confidence interval [CI]:1.18–6.00). While not statistically significant after multiple comparisons adjustment, time to first outdoor fall appeared to differ between the 4 dose groups (unadjusted P for overall difference=0.013; adjusted P=0.259), with risk of a first-time outdoor fall 39% lower in the 1000IU/d group (HR=0.61; 95% CI: 0.38–0.97; unadjusted P=0.036; adjusted P=0.259) and 40% lower in the 2000IU/d group (HR=0.60; 95%CI 0.38–0.97; P=0.037; adjusted P=0.259), each versus control.
Conclusion:
Vitamin D supplementation doses ≥1000IU/d might have differential effects on fall risk based on fall location and fracture risk, with the most robust finding that vitamin D doses between 1000–4000IU/d might increase the risk of first time falls with fractures. Replication is warranted, given the possibility of type 1 error.
Keywords: Indoor, outdoor, fall reoccurrence, vitamin D dose, randomized clinical trial
INTRODUCTION
Falls are an important and costly public health issue for adults aged ≥65 years1. Vitamin D supplementation has been recommended as a means to prevent falls2. However, a 2018 US Preventive Services Task Force (USPSTF) report concluded that current evidence for vitamin D reducing fall risk was inconsistent and suggested possible harm at higher doses3. Most recently, the Study to Understand Fall Reduction and Vitamin D in You (STURDY) trial—designed to examine the effects of different vitamin D doses for fall prevention—demonstrated that vitamin D supplementation ≥1000IU/d did not prevent falls compared to 200IU/d and raised the possibility that doses ≥2000IU/d may increase fall risk4.
While clinical trials often focus on the occurrence of any fall, the circumstances and consequences of falls among older adults vary greatly5. For instance, falling does not always result in injury or require medical care6–8. Yet, individuals who have fallen once are twice as likely to fall again8, increasing the chance for adverse outcomes9. Though older adults spend more time indoors at home10, more than half of falls happen outside11–13. Those who experience indoor falls tend to be frail14, live with disability, are physically inactive15, and self-report worse health. In contrast, those who fall outdoors tend to be more physically active and report good health15. Exploration of the effects of trial interventions on these types of falls is warranted11.
The principal aim of this paper was to evaluate whether the association of vitamin D supplementation with fall risk in the STURDY trial differed by type of fall, namely: 1) first consequential fall, 2) first fall resulting in emergency care, 3) first fall resulting in a fracture, 4) first injurious fall, 5) first indoor fall, and 6) first outdoor fall. The second aim was to assess whether vitamin D supplementation ≥1000IU/d was associated with reduced rates of all falls and subtypes. Participants randomized to daily high-dose vitamin D supplementation (≥1000IU/d) were hypothesized to have decreased incidence and rates of all falls of these types compared to those randomized to 200IU/d.
METHODS
A Johns Hopkins University institutional review board approved the trial protocol. A data and safety monitoring board approved the protocol and monitored the trial. Each participant provided written informed consent.
Setting
The STURDY trial was a seamless, two-stage, Bayesian response-adaptive, dose-finding and confirmatory randomized trial of four vitamin D3 (cholecalciferol) supplement doses—200IU/d (control), 1000IU/d, 2000IU/d, or 4000IU/d—to prevent falls. The primary outcome was time to first fall or death. The trial’s design and main results papers have been published4,16,17. During the dose-finding stage, the 1000IU/d dose was associated with the lowest rate of falls among the three non-control doses and was selected as the best non-control dose to be tested in the confirmatory phase. Participants randomized to 2000IU/d or 4000IU/d were then switched to 1000IU/d, and new enrollees were randomized to 1000IU/d or control. There was no difference in time to first fall for those with experience on best dose (events and observation time limited to time while on 1000IU/d) versus those randomized to control, nor any difference in risk between those in the pooled higher doses group (those randomized to 1000, 2000, or 4000IU/d) versus control, nor between those randomized to 1000IU/d versus control. The trial also reported safety-related fall outcomes, including serious falls (falls resulting in fracture or dislocation or associated with hospitalization or other serious adverse event [SAE]), falls with hospitalization, and serious events potentially related to vitamin D.
Participants
Eligible participants were community-dwelling adults aged ≥70 years with elevated fall risk. Elevated fall risk was self-reported and included ≥1 of the following: an injurious fall in the past year, ≥2 falls in the past year regardless of injury, fear of falling due to balance or walking problems, difficulty maintaining balance, or use of an assistive device when walking. Also, eligibility required a serum 25-hydroxyvitamin D [25(OH)D] concentration of 10–29ng/mL. Exclusion criteria included cognitive impairment (Mini-Mental State Examination score <24); hypercalcemia; kidney, bladder or ureteral stone (1 recent or ≥2 lifetime); and use of personal vitamin D supplement >1000IU/d or calcium supplement >1200mg/d. Participants enrolled throughout the year at two centers, each at approximately 39° latitude.
Treatment and randomization
Rationale and safety for the four doses are explained elsewhere16,17. Duration of pill-taking and follow-up was 2 years or end of the trial, whichever came first. Notably, if the primary outcome occurred, follow-up continued. The probability of assignment to the 200IU/d group was 0.50 throughout the trial. At the start of the trial, each of the higher dose non-control groups (1000, 2000, and 4000IU/d) had equal probability of assignment. During dose-finding, beginning after the 100th participant randomized to a non-control group achieved 6 months of follow-up, assignment probabilities to the non-control doses were adjusted at pre-specified times. Participants and study personnel were masked to randomized dose, occurrence of adaptations, and to the end of dose-finding.
Participant characteristics at entry
Self-reported sociodemographic and behavioral information included age, sex, race, ethnicity, education, marital status, smoking status, and alcohol consumption, as well as baseline medical conditions. Body mass index (BMI; kg/m2) was calculated using measured weight (digital scale) and height (wall-mounted stadiometer). Physical performance was measured using the Short Physical Performance Battery (SPPB), which consists of three physical tests (balance, walking speed, and chair rises), each scored from 0–418. The SPPB score was calculated as the sum of the test scores and ranges from 0–12, with higher score indicating better physical performance. Frailty was defined as having three or more of the following conditions: 1) underweight or unintentional weight loss; 2) exhaustion; 3) slow gait speed; 4) low physical activity; and 5) weakness19.
Fall ascertainment and classification
The trial used the World Health Organization definition of a fall (any fall, slip, or trip in which the participant lost his or her balance and landed on the floor or ground or at a lower level)23 and implemented three fall surveillance methods: monthly calendars, scheduled contacts (at 1, 3, and 6 months and every 3 months thereafter to 24 months), and ad hoc telephone contacts. Participants were instructed to call the clinic if they fell, and staff asked about fall occurrence when telephoning about missing calendars or after obtaining information about a reported fall. Participants were asked to use the calendars to document each day whether a fall had occurred and to mail calendars monthly. When a participant reported a fall, staff called the participant to obtain additional information including fall date, time of day, location, direction, medical care sought, and type of injury (scrape/cut, bruising, sprain, strain, dislocation, fracture), circumstances of fall (e.g., association with slipping, tripping, etc.), and type of medical care sought if any (doctor visit, urgent care, emergency, hospital).
Information obtained from the participant interview about injury sustained during the fall and medical care received was used to characterize each fall. Fall types of interest were time to first:
consequential fall (fall resulting in an injury or in any medical care sought)
injurious fall (injury of any kind to any body part)
fall with emergency care (fall resulting in care sought at urgent care center, emergency department, or hospital)
fall with fracture (fracture of any body part including broken nose)
indoor fall
outdoor fall
Statistical analysis
Participant characteristics at study entry were examined by randomized treatment assignment using means, standard deviations, medians, and upper and lower quartiles for continuous variables and frequencies and percentages for categorical variables. Fall characteristics by cumulative, first, second, and third or higher falls were reported as frequencies and percentages. The accumulation of falls by time since randomization, stratified by those randomized to either the pooled higher doses (≥1000IU/d) or 200IU/d, was descriptively reported for all falls and for each fall type.
The primary analysis includes all randomized participants. Cox proportional hazards regression models were used to compare the time from randomization to first fall of each type for the pooled higher doses group versus the control group; each model included a single term for treatment group. Participants without a fall of the type being analyzed were censored at their last observation for the event or at the date of the first fall that could not be categorized as to type, whichever occurred first. A competing risk model was not used because mortality was low relative to the large number of falls (1.8/100 person-years; 19 deaths)4.
Negative binomial regression models were used to compare the pooled higher doses group versus the control group on cumulative fall rate for each type of fall; each model included a term for treatment group and an offset term for the participant’s observation time. For these cumulative fall analyses, each participant’s observation time was their duration from randomization to date of last observation for falls, and fall counts were top-coded at the value equal to the 99th percentile of the non-zero fall counts to reduce the influence of outliers.
In a sensitivity analysis, participants randomized to 1000IU/d (the dose identified as best non-control dose for fall prevention) were compared to those randomized to control for the time to first fall outcomes and the cumulative fall rate outcomes using the same methods as above.
The effects of each higher dose versus 200IU/d on falls were evaluated in the 406 participants who were randomized prior to the first adaptation to the randomization probabilities, a cohort for whom treatment assignment was not influenced by previously collected outcome data. The same methods as above were used except that treatment group was represented with 3 terms. For the participants who were switched to the best dose after dose-finding ended, events and observation time are limited to their time prior to switching.
The Benjamini-Hochberg procedure was used to control the false discovery rate to below 1 out of the maximum number of comparisons for each family (table) of analyses presented (1 out of 7 for the tables comparing 2 groups and 1 out of 28 for the tables comparing 4 groups)20. Each table includes the unadjusted or nominal 2-sided P value for each test and the Benjamini-Hochberg adjusted P value and the false discovery rate specified for the family of analyses; an adjusted P value is statistically significant if below the specified false discovery rate. All analyses were conducted in SAS version 9.4 (Cary, NC), STATA version 15 (College Station, TX), or R version 4.0.0 (https://www.r-project.org/).
RESULTS
Study population
The primary analysis population included 688 participants, 349 randomized to the pooled higher doses (≥1000IU/d) group and 339 to the control (200IU/d) group. Baseline characteristics are provided in Table 1. Participants randomized prior to the first adaptation of the randomization probabilities are described in Supplementary Table S1; this subset included 406 participants (67 in each of the three non-control dose groups, and 205 in the control group).
Table 1.
Characteristics of randomized participants at enrollment (abbreviated version)a.
| Pooled Higher Dosesb (N=349) |
200 IU/d (N=339) |
All (N=688) |
|
|---|---|---|---|
| Age (years), mean ± SD | 77.2 ± 5.4 | 77.2 ± 5.4 | 77.2 ± 5.4 |
| Women, no. (%) | 159 (45.6%) | 141 (41.6%) | 300 (43.6%) |
| Black, African-American, no. (%) | 69 (19.9%) | 55 (16.4%) | 124 (18.2%) |
| Hispanic, no. (%) | 5 (1.4%) | 3 (0.9%) | 8 (1.2%) |
| Did not complete high school, no. (%) | 17 (4.9%) | 15 (4.4%) | 32 (4.7%) |
| Married or married-like relationship, no. (%) | 178 (51.0%) | 201 (59.3%) | 379 (55.1%) |
| Body Mass Index (kg/m2), mean ± SD | 30.6 ± 5.6 | 30.4 ± 6.3 | 30.5 ± 6.0 |
| Current cigarette smoker, no. (%) | 12 (3.4%) | 9 (2.7%) | 21 (3.1%) |
| Alcoholic drinks/week, median (Q1, Q3) | 0 (0, 3) | 0 (0, 4) | 0 (0, 3) |
| SPPBc total score, median (Q1, Q3) | 9 (7, 10) | 9 (8, 10) | 9 (8, 10) |
| No. of morbid conditionsd, median (Q1, Q3) | 3 (2, 5) | 3 (2, 4) | 3 (2, 5) |
| Mini-Mental State Examination scoree | |||
| 24–26 | 39 (11.2%) | 37 (10.9%) | 76 (11.0%) |
| 27–30 | 310 (88.8%) | 302 (89.1%) | 612 (89.0%) |
| Frailty statusf, no. (%) | |||
| Robust | 94 (27.0%) | 104 (30.7%) | 198 (28.8%) |
| Pre-frail | 203 (58.3%) | 205 (60.5%) | 408 (59.4%) |
| Frail | 51 (14.7%) | 30 (8.8%) | 81 (11.8%) |
Abbreviations: IU/d = international units per day; SD = standard deviation; Q1 = 25th percentile; Q3 = 75th percentile; SPPB = Short Physical Performance Battery.
A full version of this table is included as Supplement Table S1.
Pooled Higher Doses denotes the combined 1000, 2000, and 4000 IU/day groups.
The Short Physical Performance Battery (SPPB) is a 3-part assessment of physical functioning: balance testing, gait-speed testing (timed 4-meter walk at usual pace), and demonstration of ability to complete 5 chair stands; each part is scored 0 to 4 and the total SPPB score (range 0 to 12) is the sum of the 3 subscores. Higher scores indicate better physical function.
14 morbid conditions were queried (have you ever been told by a physician that you had): cancer other than basal cell skin cancer, heart disease, high cholesterol, hypertension, stroke, peripheral vascular disease, chronic lung disease, diabetes, kidney disease, liver disease, connective tissue disease, arthritis, Parkinson’s disease, multiple sclerosis.
The Mini-Mental State Examination is a questionnaire used to assess cognitive impairment; scores range from 0 to 30 and higher scores indicate better cognition. Score < 24 was exclusionary.
Physical frailty phenotype was determined using the criteria of Fried et al.: weight loss, exhaustion, slow gait speed, low grip strength, and low physical activity; status was assessed as frail (3 conditions present), pre-frail (1 or 2 conditions), or robust (no conditions).
Characteristics of falls
Participants reported a total of 1477 falls during follow-up; 423(29%) were first falls, 275(19%) were second falls, and 779(53%) were third or higher falls (Table 2). The detailed fall interview was missing for 7(2%) first falls and 177(12%) of all falls; these 177 falls were reported by 31 participants. Additionally, among the 1300 fall interviews that were obtained, 294(19%) were missing at least 1 item of information. Using all information available, 647 falls were classified as consequential, 614 as injurious, 124 as resulting in emergency care, 31 as resulting in a fracture, 764 as happening indoors, and 525 as happening outdoors. Approximately half of all 1477 falls occurred within 9 months of randomization, and the pattern of fall accumulation over time appeared visually similar between dose groups across fall types (Figure 1).
Table 2.
Characteristics of falls.
| All falls (N=1477) |
First fall (N=423) |
Second fall (N=275) |
Third and higher falls (N=779) |
|
|---|---|---|---|---|
| Fall was consequential a | 647 (50.1%) | 209 (50.2%) | 149 (56.4%) | 289 (47.3%) |
| Unknown | 186 | 7 | 11 | 168 |
| Injurious fall b | 614 (47.6%) | 198 (47.6%) | 143 (54.2%) | 273 (44.8%) |
| Unknown | 187 | 7 | 11 | 169 |
| Fall with emergency care c | 124 (9.6%) | 40 (9.6%) | 30 (11.4%) | 54 (8.8%) |
| Unknown | 183 | 7 | 12 | 164 |
| Fall with fracture d | 31 (2.4%) | 12 (2.9%) | 10 (3.8%) | 9 (1.5%) |
| Unknown | 187 | 7 | 11 | 169 |
| Indoor fall | 764 (59.3%) | 235 (56.6%) | 146 (55.1%) | 383 (62.9%) |
| Unknown | 188 | 8 | 10 | 170 |
| Outdoor fall | 525 (40.7%) | 180 (43.4%) | 119 (44.9%) | 226 (37.1%) |
| Unknown | 188 | 8 | 10 | 170 |
| Location of fall | ||||
| Inside home | 605 (46.9%) | 182 (43.9%) | 109 (41.1%) | 314 (51.6%) |
| Inside but not at home | 127 (9.9%) | 42 (10.1%) | 29 (10.9%) | 56 (9.2%) |
| Street, pavement, sidewalk, parking lot | 256 (19.9%) | 96 (23.1%) | 65 (24.5%) | 95 (15.6%) |
| Grass or unfinished ground | 192 (14.9%) | 56 (13.5%) | 45 (17.0%) | 91 (14.9%) |
| Other | 109 (8.5%) | 39 (9.4%) | 16 (6.4%) | 53 (8.7% |
| Unknown | 188 | 8 | 10 | 170 |
| Summary cause of fall | ||||
| Tripped | 474 (42.0%) | 171 (46.1%) | 91 (39.2%) | 212 (40.3%) |
| Slipped | 163 (14.4%) | 50 (13.5%) | 40 (17.2%) | 73 (13.9%) |
| Lost balance | 180 (15.9%) | 47 (12.7%) | 30 (12.9%) | 103 (19.6%) |
| Felt dizzy or fainted | 90 (8.0%) | 26 (7.0%) | 21 (9.1%) | 43 (8.2%) |
| Misjudged step or misstepped | 79 (7.0%) | 30 (8.1%) | 18 (7.8%) | 31 (5.9%) |
| Legs or knees gave out | 31 (2.7%) | 13 (3.5%) | 3 (1.3%) | 15 (2.9%) |
| Other | 112 (9.9%) | 34 (9.2%) | 29 (12.5%) | 49 (9.3%) |
| Unknown | 348 | 52 | 43 | 253 |
| Medical care sought | ||||
| Doctor visit | 50 (3.9%) | 25 (6.0%) | 10 (3.8%) | 15 (2.4%) |
| Minute clinic | 7 (0.5%) | 3 (0.7%) | 0 (0.0%) | 4 (0.7%) |
| Urgent care | 22 (1.7%) | 8 (1.9%) | 4 (1.5%) | 10 (1.6%) |
| Emergency department | 81 (6.3%) | 28 (6.7%) | 20 (7.6%) | 33 (5.4%) |
| Hospital | 44 (3.4%) | 10 (2.4%) | 13 (4.9%) | 21 (3.4%) |
| Unknown | 182 | 7 | 11 | 164 |
| Type of injury | ||||
| Cut or scrape, no stitches | 277 (21.5%) | 94 (22.6%) | 68 (25.8%) | 115 (19.0%) |
| Cut or scrape needing stitches | 16 (1.2%) | 4 (1.0%) | 4 (1.5%) | 8 (1.3%) |
| Bruised | 306 (23.8%) | 98 (23.6%) | 72 (27.3%) | 136 (22.4%) |
| Strained | 137 (10.7%) | 40 (9.6%) | 33 (12.5%) | 64 (10.6%) |
| Sprained | 34 (2.7%) | 11 (2.6%) | 10 (3.8%) | 13 (2.2%) |
| Dislocated | 4 (0.3%) | 1 (0.2%) | 2 (0.8%) | 1 (0.2%) |
| Fracture | 31 (2.4%) | 12 (2.9%) | 10 (3.8%) | 9 (1.5%) |
| Unknown | 187 | 7 | 11 | 169 |
| Body part injured | ||||
| Head, neck, face | 138 (10.7%) | 45 (10.8%) | 39 (14.8%) | 54 (8.9%) |
| Arm | 136 (10.6%) | 44 (10.6%) | 28 (10.6%) | 64 (10.6%) |
| Wrist | 32 (2.5%) | 15 (3.6%) | 8 (3.0%) | 9 (1.5%) |
| Hand or finger | 103 (8.0%) | 39 (9.4%) | 28 (10.6%) | 36 (6.0%) |
| Hip | 52 (4.0%) | 15 (3.6%) | 16 (6.1%) | 21 (3.5%) |
| Leg | 78 (6.1%) | 25 (6.0%) | 19 (7.2%) | 34 (5.6%) |
| Knee | 187 (14.5%) | 70 (16.8%) | 49 (18.6%) | 68 (11.2%) |
| Ankle | 28 (2.2%) | 12 (2.9%) | 5 (1.9%) | 11 (1.8%) |
| Foot or toe | 26 (2.0%) | 15 (3.6%) | 5 (1.9%) | 6 (1.0%) |
| Other body part | 188 (14.6%) | 49 (11.8%) | 44 (16.7%) | 95 (15.7%) |
| Unknown | 187 | 7 | 11 | 169 |
Note: The detailed fall interview was completely missing for 7 first falls (2% of first falls) and 177 of all falls (12% of all falls); these 177 falls were reported by 31 participants. Among 416 initiated first fall interviews, 78 (19%) were missing at least 1 item of information; among 1300 all fall interviews that were initiated, 294 (23%) were missing at least 1 item of information.
A consequential fall is a fall that resulted in an injury of any type or a fall after which the participant sought any type of medical care.
An injurious fall is a fall with any type of injury (e.g., bruising, sprain, fracture).
A fall with emergency care is a fall that resulted in a visit to an urgent care center, emergency department, or hospital.
A fall with fracture is a fall with fracture of any body part.
Figure 1. Accumulation of falls over all follow-up of all 688 randomized participants by type of fall.
Each pair of lines shows the accumulation of the specified type of fall as the falls occurred in the months after randomization by treatment group. Pooled Higher Doses includes participants randomized to 1000, 2000, or 4000 IU/day. A fall was consequential if any injury was sustained or any medical care was sought. A fall required emergency care if the participant sought care from an urgent care center, emergency department or hospital. A fall was injurious if the participant sustained any injury (e.g., bruising, sprain, fracture, etc). A fall with fracture was a fall with fracture of any body part.
First fall outcomes for the primary and sensitivity analysis populations
The pooled higher doses group versus the control group had a greater than twofold higher risk of a first fall resulting in a fracture (hazard ratio [HR]=2.66, 95% confidence interval [CI]: 1.18–6.00; unadjusted P=0.019, adjusted P=0.133, below the false discovery rate=0.143), a result that remained statistically significant after adjusting for multiple comparisons (Table 3). In the sensitivity analysis comparing participants randomized to 1000IU/d to the control group (Supplementary Table S2), the two groups did not significantly differ in time-to-fracture fall risk (HR=2.04;95% CI:0.76–5.45; unadjusted P=0.156, adjusted P=0.359). Although not statistically significant, this HR (2.04) is qualitatively similar to the statistically significant HR for the pooled higher doses group versus control (2.66).
Table 3.
Time-to-event analyses of fall outcomes in the primary analysis population.
| Pooled Higher Dosesa (N=349) |
200 IU/d (N=339) |
Hazard Ratio: Pooled Higher Dosesa vs. 200 IU/d (95% CI)b |
Unadjusted Pb | Adjusted P (significant if < 0.143)c | |
|---|---|---|---|---|---|
| First consequential fall d | |||||
| No. of events | 165 | 151 | 1.11 (0.89, 1.38) |
||
| No. of PY | 345.5 | 350.3 | 0.371 | 0.433 | |
| Rate per 100 PY | 47.8 | 43.1 | |||
| First injurious fall e | |||||
| No. of events | 162 | 143 | 1.16 (0.93, 1.45) |
||
| No. of PY | 349.7 | 359.3 | 0.198 | 0.277 | |
| Rate per 100 PY | 46.3 | 39.8 | |||
| First fall with emergency care g | |||||
| No. of events | 49 | 37 | 1.33 (0.87, 2.03) |
||
| No. of PY | 477.3 | 474.3 | 0.195 | 0.277 | |
| Rate per 100 PY | 10.3 | 7.8 | |||
| First fall with fracture g | |||||
| No. of events | 21 | 8 | 2.66 (1.18, 6.00) |
||
| No. of PY | 500.3 | 501.9 | 0.019 | 0.133 | |
| Rate per 100 PY | 4.2 | 1.6 | |||
| First indoor fall | |||||
| No. of events | 171 | 145 | 1.24 (0.99, 1.55) |
||
| No. of PY | 340.7 | 360.3 | 0.057 | 0.194 | |
| Rate per 100 PY | 50.2 | 40.2 | |||
| First outdoor fall | |||||
| No. of events | 118 | 137 | 0.80 (0.63, 1.03) |
||
| No. of PY | 386.4 | 353.1 | 0.083 | 0.194 | |
| Rate per 100 PY | 30.5 | 38.8 |
Abbreviations: IU/d = international units per day; CI = confidence interval; PY = person-years.
Note: Location, injury, and/or medical care information are missing for some falls; the number of first falls that could not be classified as consequential, injurious, with emergency care, with fracture, indoors, or outdoors were 7, 7, 7, 7, 8, and 8, respectively. For each fall type analysis, data from participants who did not have the event were censored at their date of last contact, with the exception that any participant who reported a fall that could not be categorized yes or no as to that fall type was censored at that date if they did not report a fall of that type previous to the fall of unknown type.
Pooled Higher Doses denotes the combined 1000, 2000, and 4000 IU/d groups.
Each hazard ratio and its 95% confidence interval were derived from a Cox regression model with dose group as the single model variable; each unadjusted P value is from the log-rank test.
The Benjamini-Hochberg procedure was used to control the false discovery rate to less than 1/7 (1 out of the maximum number of comparisons in each family of analyses (table) comparing 2 groups); an adjusted P value is statistically significant if < 0.143.
A consequential fall was a fall resulting in injury or in any medical care being sought.
An injurious fall is a fall with any type of injury (e.g., bruising, sprain, fracture).
A fall with emergency care is a fall that resulted in a visit to an urgent care center, emergency department, or hospital.
A fall with fracture is a fall with fracture of any body part.
The pooled higher doses and 200IU/d groups did not differ in time to first outdoor fall. Similarly, the sensitivity analysis showed that the 1000IU/d group did not differ in time to first outdoor fall after multiple comparisons adjustment. However, it is notable that, before multiple comparisons adjustment, the 1000IU/d group had an associated 29% lower risk of a first outdoor fall versus the control group (HR=0.71;95% CI:0.52–0.96; unadjusted P=0.027, adjusted P=0.189, above the false discovery rate=0.143).
There were no significant differences in times to first consequential fall, first injurious fall, first fall requiring emergency care, or first indoor fall between the pooled higher doses and control group, nor between those randomized to 1000IU/d versus control.
Cumulative falls in the primary and sensitivity analysis populations
The rates of all falls, overall and by specific fall type, were not statistically different between the pooled higher doses group and control group (Table 4), nor between only those randomized to 1000IU/d versus the control group (Supplementary Table S3).
Table 4.
Rates of all falls in the primary analysis population.
| Pooled Higher Dosesa (N=349) |
200 IU/d (N=339) |
Rate Ratio: Pooled Higher Dosesa v. 200 IU/d (95% CI)b |
Unadjusted Pb | Adjusted P (significant if < 0.143)c | |
|---|---|---|---|---|---|
| All falls | |||||
| No. of eventsd | 689 | 670 | 1.02 (0.83, 1.25) |
||
| No. of PYe | 535.1 | 531.2 | 0.871 | 0.871 | |
| Rate per 100 PYf | 128.8 | 126.1 | |||
| All consequential falls g | |||||
| No. of eventsd | 343 | 293 | 1.16 (0.93, 1.45) |
||
| No. of PYe | 535.1 | 531.2 | 0.185 | 0.324 | |
| Rate per 100 PYf | 64.1 | 55.2 | |||
| All injurious falls h | |||||
| No. of eventsd | 330 | 275 | 1.19 (0.95, 1.50) |
||
| No. of PYe | 535.1 | 531.2 | 0.133 | 0.310 | |
| Rate per 100 PYf | 61.7 | 51.8 | |||
| All falls with emergency care i | |||||
| No. of eventsd | 66 | 57 | 1.20 (0.77, 1.87) |
||
| No. of PYe | 535.1 | 531.2 | 0.421 | 0.491 | |
| Rate per 100 PYf | 12.3 | 10.7 | |||
| All falls with fracture j | |||||
| No. of eventsd | 21 | 10 | 2.09 (0.96, 4.57) |
||
| No. of PYe | 535.1 | 531.2 | 0.065 | 0.266 | |
| Rate per 100 PYf | 3.9 | 1.9 | |||
| All indoor falls | |||||
| No. of eventsd | 399 | 321 | 1.24 (0.98, 1.57) |
||
| No. of PYe | 535.1 | 531.2 | 0.076 | 0.266 | |
| Rate per 100 PYf | 74.6 | 60.4 | |||
| All outdoor falls | |||||
| No. of eventsd | 241 | 275 | 0.86 (0.66, 1.11) |
||
| No. of PYe | 535.1 | 531.2 | 0.242 | 0.339 | |
| Rate per 100 PYf | 45.0 | 51.8 |
Abbreviations: IU/d = international units per day; CI = confidence interval; PY = person-years.
Note: Location, injury, and/or medical care information are missing for some falls; the number of falls that could not be classified as consequential, injurious, with emergency care, with fracture, indoors, or outdoors were 186, 187, 183, 187, 188, and 188, respectively.
Pooled Higher Doses denotes the combined 1000, 2000, and 4000 IU/d groups.
Each rate ratio, its 95% confidence interval, and its unadjusted P value are derived from a negative binomial model with a term for treatment and an offset term for the participant’s observation time from randomization to date of last contact in the trial.
The Benjamini-Hochberg procedure was used to control the false discovery rate to less than 1/7 (1 out of the maximum number of comparisons in each family of analyses (table) comparing 2 groups); an adjusted P value is statistically significant if < 0.143.
Fall counts have been top-coded at the value equal to the 99th percentile of non-zero fall counts.
Observation time for a participant is the duration from randomization to date of last contact in the trial.
Rate of all events is calculated as the total number of events summed across all participants divided by the total person-time observed summed across all participants.
A consequential fall was a fall resulting in injury or in any medical care being sought.
An injurious fall is a fall with any type of injury (e.g., bruising, sprain, fracture).
A fall with emergency care is a fall that resulted in a visit to an urgent care center, emergency department, or hospital.
A fall with fracture is a fall with fracture of any body part.
Fall outcomes by dose
The cohort of 406 participants randomized prior to any adaptations of the randomization probabilities was used to examine the individual effects of each of the three non-control dose groups versus the control group on the risks of each first fall outcome (Supplementary Table S4) and rates of each cumulative fall outcome (Supplementary Table S5).
After multiple comparisons adjustment, there were no significant differences in time to fall outcomes among the 4 dose groups. Interestingly, analyses unadjusted for multiple comparisons show time to first outdoor fall appeared to differ between the 4 dose groups (unadjusted P=0.013; adjusted P=0.259, above the false discovery rate=0.035). Specifically, the 1000IU/d dose group had an associated 39% lower risk of a first-time outdoor fall (HR=0.61;95% CI:0.38–0.97; unadjusted P=0.036, adjusted P=0.259) and the 2000IU/d dose group had a 40% lower risk (HR=0.60;95%CI 0.38–0.97; unadjusted P=0.037, adjusted P=0.259) compared to the control group, respectively. The 4000IU/d and control groups had similar risk of a first-time outdoor fall.
There were no significant differences in any of the fall rate outcomes among the 4 dose groups after multiple comparisons adjustment. When examining analyses unadjusted for multiple comparisons, the rate of all falls of any type appeared to differ between the groups (unadjusted P=0.028, adjusted P=0.191, above the false discovery rate=0.036); specifically, the 1000IU/d dose group had a lower rate of all falls than the control group (RR=0.67;95%CI: 0.46–0.98; unadjusted P=0.038, adjusted P=0.191), as did the 2000IU/d group (RR=0.60;95%CI 0.41–0.89; unadjusted P=0.010, adjusted P=0.093; Supplementary Table S5). The 4000IU/d group and control group had similar rates. Outdoor fall rates also appeared to differ between groups (unadjusted P=0.006, adjusted P=0.006), again in the 1000IU/d dose group (RR=0.63;95%CI: 0.40–0.98; unadjusted P=0.041, adjusted P-0.191) and 2000IU/d dose group (RR=0.52;95% CI 0.32–0.83; unadjusted P=0.006, adjusted P=0.084) compared to control. Rates of all outdoor falls did not differ between the 4000IU/d and control groups.
DISCUSSION
The principal finding of these secondary analyses is that vitamin D3 supplement doses ≥1000IU/d were associated with greater risk of first fall with fracture, but not rate of falls with fracture, in comparison to the 200IU/d control dose. There was no other difference in fall types between the pooled higher doses and control dose. In a subset analysis comparing each higher dose group to control, there was some evidence, albeit limited, that 1000–2000IU/d doses were associated with a reduced risk and rates of outdoor falls versus the control dose. Yet, 1000–4000IU/d doses were not associated with risk nor rate of indoor falls versus the control dose. Collectively, our results raise the possibility that high-dose vitamin D supplementation and fall risk might vary by fall subtypes.
Consistent with previous literature, none of the doses were associated with fall risk or rate for fall types unrelated to fracture. In a 2018 systematic review of 81 trials, Bolland and colleagues showed that vitamin D supplementation, largely given at levels of ≥800IU/d, did not reduce the risk of falls or fractures21. At the same time, the USPSTF published an updated report stating that evidence so far does not support an association between vitamin D supplementation and reduced fall risk in older adults across seven trials employing varying dosing levels and schedules (e.g., 700IU/d to 500,000 IU/year). Our secondary analysis extends STURDY’s findings by showing that ≥1000IU/d doses of vitamin D did not lead to a decrease or increase of repeated falls, consequential falls that result in any injury or emergency care seeking, and falls stratified by occurring either indoors or outdoors.
A growing body of evidence indicates that doses ≥1000IU/d of vitamin D3 might increase the risk of fall-related fractures22–25. The recent USPSTF report documented inconsistent findings but emphasized that one trial of high dose, annual vitamin D supplementation might raise the rate of fall-related adverse outcomes26. In this trial, Sanders and colleagues showed that a single, annual, oral dose of 500,000IU vitamin D was associated with increased risk of self-reported falls and radiologically confirmed fractures in 2256 women aged 70 years and older22. In 2017, Smith and colleagues showed that doses ranging from 4000–4800IU/d increased incidence of falls when compared to placebo in 140 vitamin D insufficient, postmenopausal women aged 57–90 years23. In 2019, Burt and colleagues showed that vitamin D supplementation at 4000IU/d or 10,000IU/d led to lower radial bone mineral density compared to 400IU/d in women aged 55–70 years24. In the paper reporting the primary results of STURDY, the pooled higher doses (≥1000IU/d) group, compared to 200IU/d, was associated with higher risk of a serious fall and higher risk of fall resulting in a hospitalization—associations primarily driven by the 2000 and 4000IU/d dose groups4.
One possible explanation for the observed deleterious effects is that high-dose vitamin D3 supplementation increases calcitrol (the active metabolite form of vitamin D) and bone resorption markers24 that overstimulate bone resorption27,28. Another possibility is that high-dose vitamin D supplementation suppresses parathyroid hormone (PTH) and inhibits normal regulation of circulating serum calcium and impairs PTH-mediated bone remodeling29. Mechanistic investigations are warranted to understand the potential causes by which daily doses ≥1000IU/d increase fall-related fractures.
Analyses that compared each individual higher dose group to control showed no associations with any fall type after adjusting for multiple comparisons. However, the results unadjusted for multiple comparisons suggest a possibility that both 1000IU/d and 2000IU/d doses of vitamin D3 might reduce outdoor fall risk, respectively. This novel yet preliminary finding might help to explain the mixed evidence of vitamin D’s beneficial role towards fall reduction. In a 2009 systematic review of eight clinical trials, Bischoff and colleagues showed that 700–1000IU/d of vitamin D supplementation reduced the risk of any type of falls by 19%30. Since then, many trials showed no benefit of vitamin D on falls21, including STURDY, which showed that doses ≥1000IU/d led to no reduction in first-time falls and increased risk of serious falls and increased risk of falls with hospitalization4. Smith and colleagues showed in a secondary analysis that doses between 1600–3200IU/d resulted in a significant decrease in fall rates when compared to placebo among 140 vitamin D insufficient, postmenopausal women aged 57–90 years23.
If the effect of 1000–2000IU/d of vitamin D3 on reducing outdoor falls is a true effect, the biological mechanism remains to be elucidated. One possible explanation is that vitamin D at 1000–2000IU/d reduces PTH levels either by direct action on the parathyroid gland to suppress PTH transcription rate or indirectly by increasing absorption of calcium to induce intermittent exposure to PTH that promotes genesis and maintenance of osteoblasts29,31. That, in combination with age-related restriction of outdoor activity32 and individual-level behavioral factors associated with falling outdoors15 may also partly explain vitamin D’s observed association with reduced outdoor falls. Yet, this finding is inconsistent with the principle finding that doses ≥1000IU/d were associated with increased risk of first time fall-related fractures and were not associated with indoor falls. Therefore, the outdoor fall findings are inconclusive and might be due to chance.
Our study has limitations. Participants received 200IU/d of vitamin D, designed to achieve the recommended dietary allowance of 800IU/d, rather than a placebo. Second, the response-adaptive design of STURDY limited the size of the subset available for comparison of each dose group to each other; hence, analyses of falls in the 4 dosage groups are underpowered, especially for fall subtypes. Third, falls rely on self-report. Fourth, our findings might have resulted from false discovery, because of the large number of comparisons across various fall types and doses. Strengths include the diverse study population with both elevated fall risk and low 25(OH)D levels; high adherence; rigorous fall assessment through monthly calendars, quarterly contacts, and ad hoc telephone reports; and detailed information related to context and consequences of each fall.
In conclusion, vitamin D supplementation doses ≥1000IU/d might have differential effects on fall risk based on fall location and fracture risk, with the most robust finding that vitamin D doses between 1000–4000IU/d might increase the risk of first time falls with fractures. Replication is warranted, given the possibility of type 1 error.
Supplementary Material
Table S1. Characteristics of randomized participants at enrollment (full version).
Table S2. Time-to-event analyses of first fall outcomes in only those randomized to 1000 or 200 IU/d.
Table S3. Rates of all falls in only those randomized to 1000 or 200 IU/d by type of fall.
Table S4. Time-to-event analyses of first fall outcomes in the 4-group analysis population.
Table S5. Rates of all falls in the 4-group analysis population by type of fall.
KEY POINTS
Most vitamin D supplement trials have documented that high-dose vitamin D supplementation does not prevent falls.
Our findings suggest that high doses of vitamin D might affect the risk of certain fall types differently.
Specifically, vitamin D doses 1000–4000IU/d were observed to increase the risk, but not the rate, of falls resulting in fracture compared with the 200IU/d control dose.
Why does this paper matter?
Our paper raises the concern that vitamin D doses at ≥1000IU/d might increase the risk of first time falls resulting in fractures.
ACKNOWLEDGMENTS
We thank all the participants of STURDY for their contributions to the trial. Members of the STURDY Collaborative Research Group are as follows:
Johns Hopkins University Centers
Welch Center for Prevention, Epidemiology and Clinical Research: Lawrence J. Appel, MD, MPH (chair); Nicole Cronin, MA; Stephen P. Juraschek, MD, PhD; Scott McClure, MS; Christine M. Mitchell, ScM; Timothy B. Plante, MD, MHS.
Center on Aging and Health: Rita R. Kalyani, MD, MHS; David L. Roth, PhD; Jennifer A. Schrack, PhD; Sarah L. Szanton, PhD; Jacek K. Urbanek, PhD; Jeremy Walston, MD; Amal A. Wanigatunga, PhD, MPH.
STURDY Data Coordinating Center: Sheriza N. Baksh, PhD; Amanda L. Blackford, ScM; Shumon Chattopadhyay, MSE; Lea T. Drye, PhD; John Dodge; Cathleen Ewing; Sana Haider, BS; Stephanie C. Holland, BS; Rosetta Jackson; Andrea Lears, BS; Curtis Meinert, PhD; David M. Shade, JD; Michael Smith, BS; Alice L. Sternberg, ScM; James Tonascia, PhD; Mark L. Van Natta, MHS; Annette Wagoner.
Clinical Field Centers, Johns Hopkins University
George W. Comstock Center for Public Health Research and Prevention: Erin D. Michos, MD, MHS (Site PI); J. Denise Bennett; Pamela Bowers; Josef Coresh, MD, PhD; Patricia Crowley, MS; Tammy Crunkleton; Briana Dick, BA; Rebecca Evans, RN; Mary Godwin; Lynne Hammann; Deborah Hawks; Karen Horning; Erika Hull; Brandi Mills; Melissa Minotti, MPH; Leann Raley; Amanda Reed, MS; Rhonda Reeder, RN; Cassie Reid; Melissa Shuda; Adria Spikes; Rhonda Stouffer; Kelly Weicht.
ProHealth Clinical Research Unit: Edgar R. Miller III, MD, PhD (Site PI); Caroline Abbas; Bernellyn Carey, BS; Jeanne Charleston, RN; Syree Davis; Naomi DeRoche-Brown; Debra Gayles; Sherlina Holland, MPH; Ina Glenn-Smith; Duane Johnson; Mia Johnson; Eva Keyes; Kristen McArthur; Danielle Santiago; Chanchai Sapun; Valerie Sneed; Lee Swartz, MBA; Letitia Thomas.
Clinical Core Research Laboratory, University of Maryland School of Medicine: Robert H. Christenson, PhD; Show-Hong Duh, PhD; Heather Rebuck.
Data and Safety Monitoring Board (DSMB): Clifford Rosen, MD (chair, Maine Medical Center Research Institute); Thomas Cook, PhD (University of Wisconsin); Pamela Duncan, PhD (Wake Forest Baptist Health); Karen E. Hansen, MD, MS (2016–2019, University of Wisconsin); Anne Kenny, MD (2014–2016, University of Connecticut); Sue Shapses, PhD, RD (Rutgers University).
National Institute on Aging (NIA): Judy Hannah, PhD; Sergei Romashkan, MD.
Office of Dietary Supplements (ODS): Cindy D. Davis, PhD; Christopher T. Sempos, PhD
Consultants: Jack M. Guralnik, MD, PhD (University of Maryland School of Medicine); J.C. Gallagher, MD (Creighton University School of Medicine).
Conflict of Interest
Authors report NIH/NIA grant funding supporting their work and additional relevant NIH support as described in the Funding section.
SPONSOR’S ROLE
The funding sources for this study played no role in the design or conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
FUNDING
Findings were presented at the annual 2020 Gerontological Society of America. STURDY was funded by the National Institute on Aging (U01AG047837) with support from the Office of Dietary Supplements, the Mid-Atlantic Nutrition Obesity Research Center (P30DK072488), and the Johns Hopkins Institute for Clinical and Translation Research (UL1TR003098). Dr. Juraschek was supported by a National Institute of Diabetes and Digestive and Kidney Diseases training grant (T32DK007732) and a National Heart, Lung, and Blood Institute career development award (K23HL135273). Dr. Wanigatunga was supported by the Johns Hopkins Older Americans Independence Center (P30AG021334) and the Johns Hopkins Alzheimer’s Disease Resource Center for Minority Aging Research (P30AG059298). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
This paper is subject to the NIH Public Access Policy.
Trial Registration: ClinicalTrials.gov: NCT02166333
REFERENCES
- 1.Bergen G, Stevens MR, Burns ER. Falls and Fall Injuries Among Adults Aged ≥65 Years — United States, 2014. Morbidity and Mortality Weekly Report. 2016;65(37):993–998. doi: 10.2307/24858985 [DOI] [PubMed] [Google Scholar]
- 2.Bischoff-Ferrari HA, Dawson-Hughes B, Willett WC, et al. Effect of Vitamin D on Falls: A Meta-analysis. JAMA. 2004;291(16):1999–2006. doi: 10.1001/jama.291.16.1999 [DOI] [PubMed] [Google Scholar]
- 3.Guirguis-Blake JM, Michael YL, Perdue LA, Coppola EL, Beil TL. Interventions to Prevent Falls in Older Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2018;319(16):1705–1716. doi: 10.1001/jama.2017.21962 [DOI] [PubMed] [Google Scholar]
- 4.Appel LJ, Michos ED, Mitchell CM, et al. The Effects of Four Doses of Vitamin D Supplements on Falls in Older Adults. Ann Intern Med. 2020;174(2):145–156. doi: 10.7326/M20-3812 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Stevens JA, Mahoney JE, Ehrenreich H. Circumstances and outcomes of falls among high risk community-dwelling older adults. Inj Epidemiol. 2014;1(1):5. doi: 10.1186/2197-1714-1-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Davalos-Bichara M, Lin FR, Carey JP, et al. Development and Validation of a Falls Grading Scale. J Geriatr Phys Ther. 2013;36(2). doi: 10.1519/JPT.0b013e31825f6777 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Schwenk M, Lauenroth A, Stock C, et al. Definitions and methods of measuring and reporting on injurious falls in randomised controlled fall prevention trials: a systematic review. BMC Medical Research Methodology. 2012;12(1):50. doi: 10.1186/1471-2288-12-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.O’Loughlin JL, Robitaille Y, Boivin J-F, Suissa S. Incidence of and Risk Factors for Falls and Injurious Falls among the Community-dwelling Elderly. Am J Epidemiol. 1993;137(3):342–354. doi: 10.1093/oxfordjournals.aje.a116681 [DOI] [PubMed] [Google Scholar]
- 9.Summary of the Updated American Geriatrics Society/British Geriatrics Society Clinical Practice Guideline for Prevention of Falls in Older Persons. Journal of the American Geriatrics Society. 2011;59(1):148–157. doi: 10.1111/j.1532-5415.2010.03234.x [DOI] [PubMed] [Google Scholar]
- 10.Matz CJ, Stieb DM, Davis K, et al. Effects of Age, Season, Gender and Urban-Rural Status on Time-Activity: Canadian Human Activity Pattern Survey 2 (CHAPS 2). Int J Environ Res Public Health. 2014;11(2):2108–2124. doi: 10.3390/ijerph110202108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kelsey JL, Procter-Gray E, Hannan MT, Li W. Heterogeneity of Falls Among Older Adults: Implications for Public Health Prevention. Am J Public Health. 2012;102(11):2149–2156. doi: 10.2105/AJPH.2012.300677 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li W, Keegan THM, Sternfeld B, Sidney S, Quesenberry CP, Kelsey JL. Outdoor Falls Among Middle-Aged and Older Adults: A Neglected Public Health Problem. Am J Public Health. 2006;96(7):1192–1200. doi: 10.2105/AJPH.2005.083055 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mänty M, Heinonen A, Viljanen A, et al. Outdoor and indoor falls as predictors of mobility limitation in older women. Age Ageing. 2009;38(6):757–761. doi: 10.1093/ageing/afp178 [DOI] [PubMed] [Google Scholar]
- 14.Bergland A, Jarnlo G-B, Laake K. Predictors of falls in the elderly by location. Aging Clin Exp Res. 2003;15(1):43–50. doi: 10.1007/BF03324479 [DOI] [PubMed] [Google Scholar]
- 15.Kelsey JL, Berry SD, Procter-Gray E, et al. Indoor and Outdoor Falls in Older Adults Are Different: The Maintenance of Balance, Independent Living, Intellect, and Zest in the Elderly of Boston Study. Journal of the American Geriatrics Society. 2010;58(11):2135–2141. doi: 10.1111/j.1532-5415.2010.03062.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Michos ED, Mitchell CM, Miller ER, et al. Rationale and Design of the Study to Understand Fall Reduction and vitamin D in You (STURDY): A Randomized Clinical Trial of Vitamin D Supplement Doses for the Prevention of Falls in Older Adults. Contemp Clin Trials. 2018;73:111–122. doi: 10.1016/j.cct.2018.08.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Michos ED, Mitchell CM, Miller ER, et al. Corrigendum to Rationale and design of the Study to Understand Fall Reduction and vitamin D in You (STURDY): A randomized clinical trial of Vitamin D supplement doses for the prevention of falls in older adults. Contemp Clin Trials. 2020;90:105936. doi: 10.1016/j.cct.2020.105936 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. JGerontol. 1994;49(2):M85–M94. [DOI] [PubMed] [Google Scholar]
- 19.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. JGerontolA BiolSciMedSci. 2001;56(3):M146–56. [DOI] [PubMed] [Google Scholar]
- 20.Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological). 1995;57(1):289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
- 21.Bolland MJ, Grey A, Avenell A. Effects of vitamin D supplementation on musculoskeletal health: a systematic review, meta-analysis, and trial sequential analysis. The Lancet Diabetes & Endocrinology. 2018;6(11):847–858. doi: 10.1016/S2213-8587(18)30265-1 [DOI] [PubMed] [Google Scholar]
- 22.Sanders KM, Stuart AL, Williamson EJ, et al. Annual High-Dose Oral Vitamin D and Falls and Fractures in Older Women: A Randomized Controlled Trial. JAMA. 2010;303(18):1815–1822. doi: 10.1001/jama.2010.594 [DOI] [PubMed] [Google Scholar]
- 23.Smith LM, Gallagher JC, Suiter C. Medium doses of daily vitamin D decrease falls and higher doses of daily vitamin D3 increase falls: A randomized clinical trial. The Journal of Steroid Biochemistry and Molecular Biology. 2017;173:317–322. doi: 10.1016/j.jsbmb.2017.03.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Burt LA, Billington EO, Rose MS, Raymond DA, Hanley DA, Boyd SK. Effect of High-Dose Vitamin D Supplementation on Volumetric Bone Density and Bone Strength: A Randomized Clinical Trial. JAMA. 2019;322(8):736–745. doi: 10.1001/jama.2019.11889 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ginde AA, Blatchford P, Breese K, et al. High-Dose Monthly Vitamin D for Prevention of Acute Respiratory Infection in Older Long-Term Care Residents: A Randomized Clinical Trial. Journal of the American Geriatrics Society. 2017;65(3):496–503. doi: 10.1111/jgs.14679 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Guirguis-Blake JM, Michael YL, Perdue LA, Coppola EL, Beil TL. Interventions to Prevent Falls in Older Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2018;319(16):1705–1716. doi: 10.1001/jama.2017.21962 [DOI] [PubMed] [Google Scholar]
- 27.Nakamichi Y, Udagawa N, Suda T, Takahashi N. Mechanisms involved in bone resorption regulated by vitamin D. The Journal of Steroid Biochemistry and Molecular Biology. 2018;177:70–76. doi: 10.1016/j.jsbmb.2017.11.005 [DOI] [PubMed] [Google Scholar]
- 28.Takeda S, Yoshizawa T, Nagai Y, et al. Stimulation of Osteoclast Formation by 1,25-Dihydroxyvitamin D Requires Its Binding to Vitamin D Receptor (VDR) in Osteoblastic Cells: Studies Using VDR Knockout Mice. Endocrinology. 1999;140(2):1005–1008. doi: 10.1210/endo.140.2.6673 [DOI] [PubMed] [Google Scholar]
- 29.Wein MN, Kronenberg HM. Regulation of Bone Remodeling by Parathyroid Hormone. Cold Spring Harb Perspect Med. 2018;8(8):a031237. doi: 10.1101/cshperspect.a031237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, et al. Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials. BMJ. 2009;339:b3692. doi: 10.1136/bmj.b3692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Esbrit P, Alcaraz MJ. Current perspectives on parathyroid hormone (PTH) and PTH-related protein (PTHrP) as bone anabolic therapies. Biochemical Pharmacology. 2013;85(10):1417–1423. doi: 10.1016/j.bcp.2013.03.002 [DOI] [PubMed] [Google Scholar]
- 32.Eronen J, von Bonsdorff MB, Törmäkangas T, et al. Barriers to outdoor physical activity and unmet physical activity need in older adults. Preventive Medicine. 2014;67:106–111. doi: 10.1016/j.ypmed.2014.07.020 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Characteristics of randomized participants at enrollment (full version).
Table S2. Time-to-event analyses of first fall outcomes in only those randomized to 1000 or 200 IU/d.
Table S3. Rates of all falls in only those randomized to 1000 or 200 IU/d by type of fall.
Table S4. Time-to-event analyses of first fall outcomes in the 4-group analysis population.
Table S5. Rates of all falls in the 4-group analysis population by type of fall.