Vitamin D and the mechanisms, circumstances and consequences of falls in older adults: A case-control study

Abstract

Objectives

To determine i) whether cases of elderly fallers had lower serum 25-hydroxyvitamin D (25OHD) concentration than controls without history of falls; and ii) whether serum 25OHD concentration was associated with specific mechanisms, circumstances and consequences of falls.

Design

Case-control study with a 1:2 ratio.

Setting

Geriatric ward of the University Hospital of Angers, France, between February 2012 and March 2014.

Participants

216 inpatients (72 cases and 144 age- and gender-matched controls).

Measurement

Falls were defined as involuntary events causing the person to the ground or other lower level. The main mechanisms, circumstances and consequences of falls were identified using standardized questionnaires. Vitamin D deficiency was defined as serum 25OHD concentration ≤25nmol/L. Age, gender, body mass index, polypharmacy, use antihypertensive drugs, use psychoactive drugs, disability, cognitive performance, serum concentrations of parathyroid hormone, creatinine and albumin, and season of evaluation were used as potential confounders.

Results

216 participants (72 cases and 144 controls) were included in the study. There was no between-group difference in the prevalence of vitamin D deficiency (P=0.176). After adjusting for confounding factors, vitamin D deficiency was positively associated with falls (OR=4.03, P=0.014). Finally, the fallers with vitamin D deficiency exhibited more often orthostatic hypotension (68.8% against 33.3%, P=0.039) and a history of recurrent falls (85% against 50%, P=0.002) than those without vitamin D deficiency.

Conclusion

This casecontrol study reported that vitamin D deficiency was associated with falls in older inpatients. There was a greater prevalence of orthostatic hypotension and of the reccurrence of falls among fallers with vitamin D deficiency, suggesting that vitamin D may influence the conditions predisposing to falls rather than the fall by itself.

Introduction

More than half of people aged 80 and over experience a fall every year (1). Besides such high incidence, the severity of falls relies on their adverse consequences including the recurrence of falls, bone fractures, fear of falling, loss of independence, institutionnalization, hospitalization and death (2, 3); leading to significant health care costs (3). Another issue is the limited number of long-term effective prevention interventions. Among them is the vitamin D supplementation, which has the merit to be simple, efficient and inexpensive (4).

Besides its classical effects on bone and calcium metabolism, vitamin D exerts a number of nonclassical effects on target organs other than the bones (4, 5, 6, 7, 8). In particular, data accumulated since the original publication by Chapuy et al. (7) showed, despite few inconclusive results (9, 10), a significantly reduced fall rate following vitamin D supplementation among older adults with hypovitaminosis D (4, 5, 6, 7). This effect has been explained mainly by the improvement of the neuromuscular function in the case of adequate vitamin D impregnation (4, 5, 6). Other mechanisms have been also proposed, such as the prevention of orthostatic hypotension (11, 12, 13).

Interestingly, a recent meta-analysis, which objective was to explore the relationship between serum vitamin D concentration and fall onset in the elderly, found that, despite a significant number of randomized controlled trials on this issue, no case-control study has been conducted so far to compare the vitamin D status between cases of elderly fallers and controls without history of falls (8). This step seems yet crucial to characterize the profile of target populations that would most benefit from vitamin D supplements to prevent falls in clinical trials and in clinical routine. Likewise, the mechanisms and circumstances of falls have been scarcely studied according to the vitamin D status; just like the consequences and complications of falls (with the exception of bone fractures) (6, 8). We hypothesized that elderly fallers have a serum concentration of 25-hydroxyvitamin D (25OHD) lower than controls without history of falls. The main objective of this case-control study was to determine whether cases of elderly fallers had lower serum 25OHD concentration, and more often vitamin D deficiency, than age- and gender-matched controls without history of falls. Our secondary objective was to determine, among fallers, whether serum 25OHD concentration was associated with specific mechanisms, circumstances and consequences of falls.

Methods

Participants

We conducted a matched case-control study with a 1:2 ratio. Cases were inpatients aged 75 years and over admitted for a fall to the geriatric acute care unit of Angers University Hospital, France, between February 2012 and March 2014. Controls were age- and gender-matched inpatients admitted to the same unit for another reason and with no history of falls in the preceding year. Triads of 1case:2controls were screened using the computerized registry of Angers University Hospital, and finally included in the analysis after manual review of medical records by one author (GD). Triads were excluded if i) medical records or medical correspondence indicated a history of falls among at least one control in the preceding year (i.e., misclassified triad), and ii) if at least one patient among the triad was receiving vitamin D supplements at the time of the hospitalization. Participants in the study were included after having given their informed consent for research. The study was conducted in accordance with ethical standards set forth in the Helsinki Declaration (1983). The whole project was approved by the local ethics committee.

Falls

A fall was defined as an event resulting in a person coming to rest unintentionally on the ground or at other lower level, not as the result of a major intrinsic event or an overwhelming hazard (1).

The following mechanisms and circumstances of falls were identified among cases of elderly fallers from medical records and correspondance: orthostatic hypotension, physical impairment, cognitive impairment, malaise, fall while standing up, fall while walking, place of fall (indoor or outdoor). Orthostatic hypotension was defined as systolic blood pressure drop ≥20mmHg and/or diastolic blood pressure drop≥10mmHg within 3 minutes of erect posture (11, 12, 13). Physical impairment was retained when the Five-Time Sit-to-Stand test was not feasible or pathological (i.e., performed in >15 seconds) during the hospital stay (14). Cognitive impairment was defined as Mini-Mental State Examination (MMSE) score <26, as previously published (15).

The following consequences of falls were noted from standardized questionnaires and medical records: recurrence, extensive cutaneous wounds, vertebral and/or nonvertebral fractures, head trauma, inability to stand up alone, and fear of falling. Recurrent falls were defined as the occurrence of 2 falls or more in the preceding 12 months (1).

Serum vitamin D assessment

Fasting early morning venous blood was collected on the day of admission from resting inpatients. Serum 25OHD concentration was measured by radioimmunoassay (DiaSorin Inc., Stillwater, MN). With this method, there is no lipid interference, which is often observed in other nonchromatographic assays of serum 25OHD concentrations. The intra- and interassay precision were 5.2% and 11.3% respectively (range 30-125 nmol/L in normal adults aged 20-60 years). Vitamin D deficiency was consensually defined as serum 25OHD≤25nmol/L, as previously described (11). All measurements were performed at the University Hospital of Angers, France.

Covariates

The following covariates were included as potential confounders in the statistical models: age, gender, body mass index (BMI), polypharmacy, use of antihypertensive drugs, use of psychoactive drugs, disability, cognitive performance, serum concentrations of parathyroid hormone, creatinine and albumin, and season of evaluation.

All included study participants received a full medical examination by a physician, which consisted of a clinical examination, standardized questionnaires, and anthropometric measurements (i.e., height and weight to calculate the BMI=height/weight², in kg/m²).

Treatments regularly taken were systematically noted from the primary care physicians' prescriptions and sought by questioning the patients and their relatives. Polypharmacy was defined as the use of more than 4 drug classes per day, as previously described (16). Antihypertensive drugs were diuretics, beta-blockers, calcium antagonists, angiotensin conversion enzyme inhibitors, agonists of angiotensin-II receptors or central antihypertensive agents. Psychoactive drugs were benzodiazepines, antidepressants or neuroleptics. Disability was assessed using the AGGIR (Autonomie Gerontologie Groupes Iso-Ressources) score, ranging from 1 to 6 (best) (17). The AGGIR grid is the national standardized instrument aimed at determining the disability of older people in France living in institutions and in the community (17). Cognitive performance was assessed using the MMSE score, ranging from 0 to 30 (best) (15). Serum concentrations of parathyroid hormone, calcium, albumin and creatinine were determined at the time of the 25OHD assay using automated standard laboratory methods at the University Hospital of Angers, France. Finally, the season of evaluation was recorded as follows: spring from March 21 to June 20, summer from June 21 to September 20, autumn from September 21 to December 20, winter from December 21 to March 20.

Statistical analysis

The participants' characteristics were summarized using means and standard deviations (SD) or frequencies and percentages, as appropriate. Firstly, comparisons between cases of elderly fallers and age- and gender-matched controls without history of falls were performed using Student's t-test or the Chi-square test, as appropriate. Secondly, a multivariate logistic regression model was used to examine the association between vitamin D deficiency (independent variable) and the history of fall (dependent variable), while adjusting for potential confounders. Finally, the mechanisms, circumstances and consequences of falls were compared among fallers (n=72), according to vitamin D deficiency using the Chi-square test or Fisher's exact test, as appropriate. P-values <0.05 were considered significant. All statistics were performed using SPSS (v19.0, IBM Corporation, Chicago, IL).

Results

Among 1128 potential matched participants with all data available (mean age 84.9±5.4years), we excluded 156 triads (n=468 participants) because reading of medical records indicated history of falls among at least one control (i.e., misclassified triads), and 148 triads (n=444 participants) because at least one participant among the triad was using vitamin D supplements. In final, 72 triads (n=216 participants composed of 72 cases of elderly fallers and 144 matched controls without history of falls) were included in this analysis (Figure 1).

Figure 1.

Flow diagram of the selection of participants

As indicated in Table 1, the 25OHD concentration in the whole sample was 37.7±22.7 nmol/L (mean±standard deviation). Respectively the mean 25OHD concentration was 34.3±23.0 nmol/L among cases of elderly fallers, and 36.4±22.5 nmol/L among controls without history of falls (P=0.510). The prevalence of vitamin D deficiency was 55.6% among cases, and 45.8% among controls (P=0.176). There was no significant difference between cases and controls regarding the other characteristics studied (Table 1).

Table 1.

Characteristics and comparison^* of cases of elderly fallers (n=72) and age- and gender-matched controls without history of falls (n=144)

	Total cohort (n=216)	Fall
		Yes (n=72 cases)	No (n=144 controls)	P-value*
Demographical and clinical measures
Age, years	84.9±5.4	84.9±5.4	84.9±5.4	1.000
Female gender, n (%)	96 (44.4)	32 (44.4)	64 (44.4)	1.000
Body mass index, kg/m2	26.8±5.8	26.5±6.0	26.9±5.7	0.624
Polypharmacy†, n (%)	165 (80.5)	54 (78.3)	111 (81.6)	0.567
Use antihypertensive drugs‡, n (%)	172 (79.6)	57 (79.2)	115 (79.9)	0.905
Use psychoactive drugs||, n (%)	110 (50.9)	40 (55.6)	70 (48.6)	0.336
MMSE score, /30	19.7±6.0	19.2±6.1	19.9±6.0	0.425
Iso Ressource Group, /6	3.5±1.4	3.3±1.5	3.6±1.3	0.170
Serum measures
Vitamin D
25OHD concentration, nmol/L	37.7±22.7	34.3±23.0	36.4±22.5	0.510
Vitamin D deficiency§, n (%)	105 (49.1)	40 (55.6)	65 (45.8)	0.176
Parathyroid hormone concentration, pg/mL	38.9±27.5	34.4±19.7	41.4±30.4	0.098
Creatinine concentration, µmol/L	95.9±50.2	92.7±46.0	94.5±52.3	0.510
Albumin concentration, g/L	31.7±5.2	31.9±4.4	31.6±5.5	0.748
Season of blood collection¶				0.290
Spring, n (%)	49 (22.7)	16 (22.2)	33 (22.9)
Summer, n (%)	61 (28.2)	16 (22.2)	45 (31.3)
Autumn, n (%)	48 (22.2)	21 (29.2)	27 (18.8)
Winter, n (%)	58 (26.9)	19 (26.4)	39 (27.1)

Data presented as mean±standard deviation where applicable. 25OHD: 25-hydroxyvitamin D; MMSE: Mini-Mental State Examination

^*comparisons between participants with and without history of falls based on Student’s t-test or Chi-square test, as appropriate

^†more than 4 drug classes used per day

^‡diuretics, beta-blockers, calcium antagonists, angiotensin conversion enzyme inhibitors, agonists of angiotensin-II receptors and central antihypertensive agents

^||benzodiazepines, antidepressants or neuroleptics

^§serum 25-hydroxyvitamin D ≤ 25 nmol/L

^¶spring from 21 March to 20 June, summer from 21 June to 20 September, autumn from 21 September to 20 December, winter from 21 December to 20 March; P-value significant (i.e., P<0.05) indicated in bold.

Table 2 reports the results of the multivariate logistic regression examining the association between vitamin D deficiency and falls after adjustment for potential confounders. Vitamin D deficiency was positively associated with falls (odds ratio (OR)=4.03, P=0.014). Additionally, falls were positively associated with the serum concentrations of parathyroid hormone (OR=0.97, P=0.033) and creatinine (OR=1.01, P=0.028).

Table 2.

Multivariate logistic regression showing the cross-sectional association between vitamin D deficiency* (independent variable) and falls (dependent variable) adjusted for potential confounders† (n=216)

Lastly, Table 3 reports the prevalence of the mechanisms, circumstances and consequences of falls according to the vitamin D status (i.e., vitamin D deficiency versus no vitamin D deficiency) among cases of elderly fallers. Comparisons showed a greater prevalence of orthostatic hypotension among fallers with vitamin D deficiency (68.8% versus 33.3%, P=0.039), and also a greater prevalence of recurrent falls in this group (85.0% versus 50.0%, P=0.002). The other mechanisms, circumstances and consequences of falls did not differ between the groups.

Table 3.

Comparisons of the mechanisms, circumstances and consequences of falls according to vitamin D deficiency^* among fallers (n=72)

	All fallers (n=72)	Vitamin D deficiency*		P-Value†
		Yes (n=40)	No (n=32)
Mechanisms/Circumstances
Orthostatic hypotension‡ (n=34)	17 (50)	11 (68.8)	6 (33.3)	0.039
Physical impairment||	19 (70.4)	32 (80.0)	22 (68.8)	0.601
Cognitive impairment§	55 (76.4)	31 (77.5)	24 (75.0)	0.875
Malaise	13 (18.1)	6 (15.0)	7 (21.9)	0.451
Fall while standing up	15 (21.1)	8 (20.5)	7 (21.9)	0.889
Fall while walking	41 (57.7)	21 (53.8)	20 (62.5)	0.463
Outdoor fall	6 (8.7)	3 (6.0)	3 (15.8)	0.336
Consequences
Recurrence¶	47 (65.3)	34 (85.0)	16 (50.0)	0.002
Extensive cutaneous wounds	9 (12.5)	4 (10.0)	5 (15.6)	0.498
Fractures	17 (23.6)	11 (27.5)	6 (18.8)	0.385
Vertebral fractures	4 (5.6)	2 (5.0)	2 (6.3)	1.000
Nonvertebral fractures	13 (18.1)	9 (22.5)	4 (12.5)	0.273
Head trauma	43 (59.7)	24 (60.0)	19 (59.4)	0.957
Inability to stand up alone	16 (22.2)	9 (22.5)	7 (21.9)	0.949
Fear of falling	5 (6.9)	2 (5.0)	3 (9.4)	0.650

Data presented as n (%)

^*serum 25-hydroxyvitamin D ≤ 25 nmol/L

^†comparisons based on the Chi-square test or Fisher’s exact test, as appropriate;

^‡: decrease in systolic blood pressure ≥20mmHg and/or diastolic blood pressure ≥10mmHg within 3 minutes of erect posture

^||Five-Time Sit-to-Stand test not feasible or pathological

^§Mini-Mental State Examination score < 26

^¶≥2 falls in preceding year; P-value significant (i.e., P<0.05) indicated in bold.

Discussion

The main finding of this case-control study in elderly inpatients was that, although there was no significant difference in 25OHD concentration between fallers and non-fallers, vitamin D deficiency was positively associated with falls while considering the potential confounders. This result suggests a relationship that could be more complex than previously expected. Specifically we found a greater prevalence of orthostatic hypotension and of the recurrence of falls among fallers with vitamin D deficiency, suggesting that the effects of vitamin D on falls could involve the conditions predisposing to falls rather than the occurrence of falls by itself.

Despite a large number of studies on the association between vitamin D and falls in older adults, we provide here, to the best of our knowledge, the first case-control study examining the 25OHD concentration between cases of elderly fallers and controls without history of falls. In the present study, the multivariate model, but not the univariate model, was consistent with previous observational cross-sectional (18,19) and longitudinal studies (20,21), and showed a positive association between vitamin D deficiency and falls. In general, multivariate models adjusted for potential confounders should be preferred to univariate models because they account for multiple potential confounders, making these models more complex and closer to reality (22). However, the lack of significant assocation in univariate model here suggests that the relationship between vitamin D deficiency and falls may be not as direct as expected. Several explanations may be proposed. First, it is possible that there is actually no link between vitamin D and fall. However, this assumption is not supported by the fact that the multivariate model highlighted a significant association, and it is also refuted by previous literature, especially previous randomized clinical trials that reported a prevention of falls after vitamin D supplementation with a high level of evidence (5,7). Second, it is also possible that our study was underpowered, with a high beta risk that a false null hypothesis was accepted. Third, as all participants in this study were frail geriatric inpatients, our result may have been distorted by the relatively low 25OHD concentrations, which might result in a floor effect. Finally, it may be argued that the majority of fallers in our study had fallen for a reason other than vitamin D deficiency, but for example due to their high degree of frailty. This would mean that, in these geriatric inpatients, serum vitamin D explains only a limited fraction of falls, and probably some specific falls with a specific mechanism.

We found that orthostatic hypotension was the sole mechanism of fall that was significantly more frequent among fallers with vitamin D deficiency than among those without vitamin D deficiency. This result is consistent with the few previous studies on this topic (11, 12) and one meta-analysis (13). It was proposed that vitamin D deficiency may cause dysfunction of arterial wall cells with subsequent impairments in arterial compliance and blood pressure control (23). Both endothelial cells and vascular smooth muscle cells (VSMCs) are equipped with 1-alpha-hydroxylase, enabling the synthesis of the active form of vitamin D (1,25-dihydroxyvitamin D), which can then bind and act locally on the vitamin D receptors (VDRs) present in the VSMCs (24, 25). Furthermore, it was also reported that vitamin D is involved in neuronal health and function, and regulates the synthesis of various neuromediators such as the noradrenaline, which is responsible for vasoconstriction in response to hypotension (26). Thus vitamin D deficiency may alter the baroreflex neural arc with consequent inefficient short-term adaptive response to standing up. This may be particularly true in diabetes mellitus (27) -a condition commonly accompanying vitamin D deficiency (28)- as vitamin D deficiency may be involved in the development of diabetic peripheral neuropathy (28). In contrast, we found no difference of the circumstances of falls according to vitamin D status. In particular, there were no more falls during transfers or during walking or outdoors in the case of vitamin D deficiency. This was consistent with the finding that, in the studied sample, there was also no between-group difference in the prevalence of physical impairments. The latter result is yet more controversial as chronic gait disorders have previously been reported with vitamin D deficiency (29). It is commonly accepted that vitamin D is involved in health and function of skeletal muscle cells and of neurons of the central nervous system (6, 26), notably in motor coordination and attention (30). In this perspective, the lack of relationship between vitamin D deficiency and cognitive impairments is also unexpected (30), and could be explained by the high prevalence of cognitive impairment in the studied sample of geriatric frail inpatients, with the risk of a ceiling effect. It is also possible that the MMSE, which is a composite cognitive score, has missed an association (if any) between vitamin D deficiency and domain-specific impairments in attention and information processing speed, as previously reported (31).

Recurrence of falls was the sole consequence of falls that was significantly more frequent among fallers with vitamin D deficiency than among those without vitamin D deficiency. This result confirms a previous study by Snijder and al. (20), who reported that recurrent falls were associated with lower serum 25OHD concentrations. The first fall may be a coincidental fall with an identifiable extrinsic cause, whereas further recurrent falls are more associated with intrinsic mecanisms of falls (1, 32, 33). Here, our results confirm the finding of an association between vitamin D and recurrent falls in older adults, and also provide a rationale for the involvement of a specific intrinsic mechanism of falls, namely the orthostatic hypotension.

The strengths of the present study include i) the originality of the design in a well-documented research area, ii) the detailed description of the participants' characteristics allowing the use of regression models to measure adjusted associations, and iii) the standardized collection of data from a single research center. In addition, enrolling patients in a single center allowed to centralize the 25OHD assays in a single laboratory and to avoid inter-laboratory variability. Regardless, our study also has a few limitations. Firstly, the case-control design of our study is less robust than a prospective longitudinal cohort study. Secondly, participants were included from a single center in France and the study cohort was restricted to 216 geriatric inpatients, who might be unrepresentative of the general population of older adults. Thirdly, cases and controls were not matched on body mass index and comorbidities, which can be confounding factors; hence the advantage of using a multivariate logistic regression model. Fourthly, important covariables such as muscular strength were not assessed in this study, which may limit the interpretation of our results.

Conclusion

We found that vitamin D deficiency was associated with falls in geriatric inpatients, but only after adjustment for confounders. This suggests that the relationship between vitamin D and falls could be more complex than previously expected. Specifically we found a greater prevalence of orthostatic hypotension and of the reccurrence of falls among fallers with vitamin D deficiency, suggesting that vitamin D may influence more the conditions predisposing to falls than the fall by itself, and that falls could actually be a proxy measure for this change. Further case-control studies with various senior cohorts and larger sample size are warranted to assess a direct effect of vitamin D deficiency on fall mechanisms and risk, and to further characterize the profile of target populations that would most benefit from vitamin D supplements to prevent falls.

Acknowledgments: We are grateful to the participants for their cooperation. We also thank Melinda Beaudenon, MSc, and Romain Simon, MSc, from the Research Center on Autonomy and Longevity (CeRAL), France, for daily assistance.

Conflict of interest and disclosures: The authors report no conflict of interests.

Funding: None.

Ethical standard: Participants in the study were included after having given their informed consent for research. The study was conducted in accordance with the ethical standards set forth in the declaration of Helsinski (1983). The study protocol was approved by the local Ethical committee.

Authors contribution: − Duval has full access to the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analyses. − Study concept and design: Annweiler and Duval. − Acquisition of data: Paré, Duval, Gautier and Annweiler. − Analysis and interpretation of data: Duval and Annweiler. − Drafting of the manuscript: Duval and Annweiler. − Critical revision of the manuscript for important intellectual content: Paré, Gautier, Walrand and Dinomais. − Obtained funding: Not applicable. − Statistical expertise: Duval. − Administrative, technical, or material support: Annweiler. − Study supervision: Annweiler.