Serum Vitamin E Concentrations and Recovery of Physical Function During the Year After Hip Fracture

Abstract

Background.

Poor nutritional status after hip fracture is common and may contribute to physical function decline. Low serum concentrations of vitamin E have been associated with decline in physical function among older adults, but the role of vitamin E in physical recovery from hip fracture has never been explored.

Methods.

Serum concentrations of α- and γ-tocopherol, the two major forms of vitamin E, were measured in female hip fracture patients from the Baltimore Hip Studies cohort 4 at baseline and at 2-, 6-, and 12-month postfracture follow-up visits. Four physical function measures—Six-Minute Walk Distance, Lower Extremity Gain Scale, Short Form-36 Physical Functioning Domain, and Yale Physical Activity Survey—were assessed at 2, 6, and 12 months postfracture. Generalized estimating equations modeled the relationship between baseline and time-varying serum tocopherol concentrations and physical function after hip fracture.

Results.

A total of 148 women aged 65 years and older were studied. After adjusting for covariates, baseline vitamin E concentrations were positively associated with Six-Minute Walk Distance, Lower Extremity Gain Scale, and Yale Physical Activity Survey scores (p < .1) and faster improvement in Lower Extremity Gain Scale and Yale Physical Activity Survey scores (p < .008). Time-varying vitamin E was also positively associated with Six-Minute Walk Distance, Lower Extremity Gain Scale, Yale Physical Activity Survey, and Short Form-36 Physical Functioning Domain (p < .03) and faster improvement in Six-Minute Walk Distance and Short Form-36 Physical Functioning Domain (p < .07).

Conclusions.

Serum concentrations of both α- and γ-tocopherol were associated with better physical function after hip fracture. Vitamin E may represent a potentially modifiable factor related to recovery of postfracture physical function.

HIP fracture is a devastating injury that is increasingly common among the growing population of older adults. One of the most debilitating consequences of hip fracture is a steep and persistent decline in physical function. Recovery of mobility, lower extremity function, and many other critical aspects of physical functioning often exceeds an entire year (1). Due in large part to this decline in physical functioning, only 50% of previously independent elders who suffer a hip fracture will return to independent living after the fracture (2). Consequently, identifying modifiable factors associated with expedited and more complete recovery of physical function is of great importance in preventing this loss of independence. Insufficient nutrition has been identified as one such risk factor for poor recovery after hip fracture (3,4).

Between 50% and 68% of hip fracture patients suffer from some form of malnutrition (5,6). Deficiencies in protein and total caloric intake have been consistently identified after hip fracture and associated with impaired physical recovery (3,4). Recent efforts have shown that postfracture malnutrition extends beyond protein and caloric deficits, as poor dietary summary scores (7) and insufficient micronutrient intake (6) have been identified after fracture. However, postfracture nutritional interventions have focused almost exclusively upon increasing protein intake. Some postfracture protein supplementation interventions have improved activities of daily living function (8) and reduced complications (9) and recovery time in hospital (10), whereas others have failed to improve such outcomes (11). These conflicting results may be due to narrow focus on protein and calories, emphasizing quantity as opposed to quality of nutritional support. Collectively, these findings suggest that a more comprehensive approach to nutritional interventions may be necessary to improve postfracture physical function.

Community-dwelling older adults with higher serum micronutrient concentrations demonstrate better physical function and a delayed onset of disability when compared with similar older adults with lower concentrations (12,13), but such relationships remain largely unexplored after hip fracture. Serum α- and γ-tocopherol, reliable markers of the primary forms of vitamin E obtained through the diet (14,15), have shown particularly strong associations with physical function among older adults (16). Notably, recent work has shown that α-tocopherol was the serum micronutrient most strongly predictive of subsequent decline in physical function as measured by Short Physical Performance Battery scores (14). However, it is not known whether hip fracture patients with higher serum tocopherol concentrations recover physical function faster or more completely than those with lower concentrations. An understanding of this relationship after hip fracture is of particular importance because the decline in physical function following fracture is dramatic and insufficient amounts of vitamin E are likely consumed after fracture. Previous research has shown that serum tocopherols are depleted following traumatic injury (17) and surgery (18), suggesting that vitamin E levels may decline after fracture even among those with adequate dietary intake.

In this study, we aimed to determine whether serum tocopherol concentrations measured shortly after hip fracture (baseline) and throughout the year postfracture (time-varying) predict the recovery of physical function. Dietary intake of vitamin E is likely to change at various stages of hip fracture recovery due to improving abilities to shop for groceries, prepare healthy meals, and so on. The relationship between vitamin E and physical function may also differ throughout the year after fracture as function generally improves over time. The assessment of associations both at baseline and throughout recovery will elucidate whether such differences in the relationship between vitamin E and physical function over time exist. We hypothesized that higher concentrations of both baseline and time-varying α- and γ-tocopherol would be associated with better physical function during the year after hip fracture.

METHODS

Participants

The study sample composed of participants from the fourth cohort of the Baltimore Hip Studies (BHS4), a randomized clinical trial designed to test the effects of an exercise intervention in reducing bone and muscle loss after hip fracture, described in detail elsewhere (19). Due to the physical demands of participation in an exercise intervention, BHS4 had a stringent set of study inclusion criteria in which only 243 of the 1,276 (19%) screened hip fracture patients were eligible. The study enrolled 180 community-dwelling women aged 65 years and older who were admitted to one of three hospitals in Baltimore with surgical repair of a nonpathological hip fracture. Additional eligibility criteria included the ability to walk independently prior to fracture, no medical conditions contraindicated with exercise, and a score of ≥20 on the Mini-Mental State Examination (20). Assessments consisting of an interview, functional measurements, body composition assessment by dual energy x-ray absorptiometry, and a blood draw were conducted by research nurses after surgical repair at baseline (≤22 days postfracture) and 2, 6 and 12 months postfracture. The final sample for this analysis consisted of the 148 unique BHS4 participants from whom a blood sample was successfully drawn at ≥1 study visit. The study was approved by the Institutional Review Boards of the University of Maryland and individual study hospitals, and participants provided informed consent.

Vitamin E Assessment

Nonfasting blood samples were obtained by venipuncture and processed according to a standardized protocol. Serum samples were stored continuously at −70°C until the time of analyses for α- and γ-tocopherol, which are both stable under these long-term storage conditions (21). Serum α- and γ-tocopherol from these samples were measured by reverse-phase high-pressure liquid chromatography at Johns Hopkins University. The intra-assay and interassay coefficients of variation, respectively, were 7.8% and 11.4% for α-tocopherol and 6.2% and 11.5% for γ-tocopherol.

Physical Function Assessment

The World Health Organization’s International Classification of Functioning, Disability and Health (22) “Part I: Functioning and Disability” informed the selection of physical functioning outcomes. Each physical function measure described below addresses at least one of the following components of the International Classification of Functioning, Disability and Health conceptualization of function and disability: (a) body functions and structures or (b) activities and participation.

Lower Extremity Gain Scale (LEGS): LEGS is a summary measurement of timed performance in nine tasks developed to evaluate lower extremity function in hip fracture patients (23). Performance in each of these tasks was scored on a scale of 0–4, resulting in a total score between 0 and 36. A higher LEGS score indicates better performance.

Six-Minute Walk Distance (6MWD): 6MWD measures ambulatory ability and endurance (24). The 6MWD was conducted according to a standardized protocol in which the maximum distance participants walked in a 6-minute period was assessed on a measured indoor course. Greater distance represents better performance.

Short Form-36 Physical Functioning Domain (SFPF): SFPF is a self-reported summary assessment of physical health (25). A subscale of the comprehensive SF-36 Health Survey, this standardized questionnaire consists of 10 items encompassing physical function. Scores range from 0 to 100, with higher scores representing better function.

Yale Physical Activity Survey (YPAS)—Total Activity Time: YPAS is a self-reported assessment of physical activity (26), a strong indicator of physical function (27), with demonstrated reliability among older adults. The YPAS measures the number of hours spent in all physical activities during a typical week in the last month.

All outcomes were assessed at 2, 6, and 12 months postfracture. Additionally, self-report measures (SFPF and YPAS) were assessed at the baseline visit in reference to the 2 weeks prior to the hip fracture.

Covariates

A number of potential confounders of the relationship between vitamin E and physical function were assessed. Body fat percentage (dual energy x-ray absorptiometry) and depressive symptoms [Geriatric Depression Scale (28)] were measured at each study visit. The following covariates were assessed at baseline only: age at time of fracture, years of education, comorbidity [Charlson Comorbidity Index (29)], Mini-Mental State Examination score, history of osteoarthritis, prefracture disability [instrumental activities of daily living (30) and lower extremity physical activities of daily living (1)], prefracture physical function (SFPF), and prefracture physical activity (YPAS).

Statistical Analysis

Descriptive analyses were performed to characterize the study population and describe the distributions of the vitamin E concentrations and physical function outcomes throughout the year after hip fracture.

Two sets of analyses were performed to examine the relationship between tocopherols and physical function. One analysis related baseline tocopherol concentrations to physical function at 2, 6, and 12 months postfracture. The second analysis related time-varying tocopherol concentrations to physical function at these same time points. The first analysis revealed whether tocopherol concentrations measured shortly after hip fracture (mean = 11 days postfracture) predict physical function during the year after fracture. The second analysis assessed whether postfracture tocopherol concentrations relate to physical function measures throughout the recovery period. The two sets of analyses were performed because tocopherol concentrations may vary over time in the hip fracture population, due to both the relatively short serum half-lives of α-tocopherol [57–72 hours (31,32)] and γ-tocopherol [13 hours (31)] and potential changes in dietary patterns during hip fracture recovery.

Baseline tocopherols and function: Generalized estimating equations (33) were used to determine the associations between tocopherol concentrations and physical function throughout the year after hip fracture while adjusting for baseline covariates and accounting for within-patient correlation. A square-root transformation was applied to YPAS because it was right-skewed.

Time-varying tocopherols and function: Marginal structural models (34) were utilized to account for the potential for time-dependent confounding. The marginal structural models were fit using weighted generalized estimating equations, where the weight was the estimated inverse probability density of tocopherols. Weights were estimated by linearly regressing tocopherol concentrations on the covariates described previously as well as prefracture YPAS and SFPF. The probability density of the observed tocopherols was then calculated assuming tocopherols follow a normal probability density with the fitted tocopherol values and mean squared error from the linear regression as the mean and variance, respectively. The weights were applied to the generalized estimating equations models, which contained the tocopherol concentrations, time since fracture, and time-by-tocopherol interaction terms.

We also performed a sensitivity analysis lagging tocopherol concentrations by one visit in the regression models (function at visit v was regressed on tocopherols at visit v − 1). Additional sensitivity analyses were conducted to account for the possibility that vitamin E concentrations were simply a marker for a healthy diet in general. These measures included adjusting for markers of fruit and vegetable intake (total serum carotenoids: sum of α-carotene, β-carotene, β-cryptoxanthin, lutein, zeaxanthin, and lycopene), other potential measures of malnutrition (serum albumin and 25-hydroxyvitamin D), and baseline meal preparation abilities in the regression models. Nonlinear tocopherol terms (log-transformations, quadratics, and cubics) were also included in regression models to assess potential changes in the relationship with physical function over time, which did not improve model fit (results not shown). Finally, BHS4 participants excluded from statistical analysis due to missing vitamin E data were compared to those included in the study on all exposures of interest. All analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, NC).

RESULTS

Table 1 provides descriptive characteristics of the study population at each visit. The study sample consisted of 148 female hip fracture patients, with a mean age of 82.3 years (standard deviation [SD] = 7.0). At baseline, participants had mean Charlson Comorbidity Index of 1.1 points (SD = 1.3), body mass index of 24.1 kg/m² (SD = 4.2), and Mini-Mental State Examination score of 24.1 (SD = 4.2). In the 2 weeks prior to fracture, participants had mean instrumental activities of daily living and lower extremity physical activities of daily living of 1.2 (SD = 1.4) and 2.0 (SD = 2.1), respectively.

Table 1.

Participant Characteristics

Variable (units)	Mean (SD) [n]
	Baseline	2 Months	6 Months	12 Months
Age at fracture (years)	82.0 (6.9) [148]	–––	–––	–––
Education (years)	12.2 (3.7) [146]	–––	–––	–––
Charlson Comorbidity Index (number of comorbidities)	1.1 (1.3) [148]	–––	–––	–––
Mini-Mental State Examination (score)	26.8 (2.6) [148]	–––	–––	–––
Body mass index (kg/m2)	23.9 (4.2) [143]	–––	–––	–––
Body fat (%)	33.5 (9.2) [97]	34.4 (9.4) [103]	34.4 (9.7) [122]	36.2 (9.7) [114]
Instrumental activities of daily living (no. of limitations)*	1.1 (1.3) [143]	2.8 (1.4) [116]	2.0 (1.5) [126]	1.9 (1.6) [124]
Lower extremity physical activities of daily living (no. of limitations)*	1.9 (2.0) [148]	6.5 (2.8) [132]	4.5 (2.8) [134]	4.1 (2.8) [132]
Geriatric Depression Scale (score)	2.4 (2.6) [147]	3.4 (2.8) [130]	2.9 (2.6) [134]	3.3 (2.9) [132]
Nutritional markers
α-tocopherol (μmol/L)	44.0 (11.9) [96]	51.8 (16.6) [95]	50.5 (15.9) [108]	50.6 (15.1) [92]
γ-tocopherol (μmol/L)	8.3 (4.2) [96]	7.7 (3.6) [95]	8.1 (4.1) [108]	7.6 (3.9) [92]
Total carotenoids (μmol/L)	1.4 (0.6) [96]	1.8 (1.0) [95]	2.0 (1.3) [108]	2.0 (1.4) [92]
Vitamin D (ng/mL)	22.7 (13.5) [100]	24.7 (9.2) [110]	25.8 (9.3) [123]	26.1 (10.0) [111]
Albumin (g/dL)	3.6 (0.7) [136]	–––	–––	–––
Physical function outcomes
Lower Extremity Gain Scale (points)	–––	22.4 (7.4) [117]	26.0 (7.0) [133]	30.4 (6.0) [124]
Six-Minute Walk Test (m)	–––	148.9 (78.7) [101]	188.1 (84.0) [117]	194.2 (89.8) [115]
Total Activity Time of Yale Physical Activity Survey (YPAS) (h)*	25.0 (18.1) [147]	12.8 (12.9) [130]	18.5 (13.6) [134]	17.0 (13.6) [133]
Short Form-36 Physical Function Domain (points)*	65.7 (26.0) [148]	29.2 (20.7) [131]	43.0 (25.6) [134]	47.7 (28.2) [133]

Note: *Baseline values refer to 2-week period prior to hip fracture.

Mean serum α-tocopherol was lowest at baseline, increased by 2 months, and remained relatively constant throughout the remainder of the study. Conversely, serum γ-tocopherol was highest at baseline, decreased slightly by 2 months, and was fairly constant throughout the other study visits. All participants with vitamin E data had both α- and γ-tocopherol concentrations. Participants who were missing vitamin E data at baseline were older (83.6 vs 81.2 years, p = .02) and demonstrated worse prefracture physical function (SFPF: 58.7 vs 68.7 points, p = .01; YPAS: 1.1 vs 1.6 h/wk, p = .002). There were no other meaningful differences between participants with and without vitamin E data.

With the exception of YPAS, which decreased slightly from 6 to 12 months, there was a monotonic improvement in all the mean physical function outcomes throughout the year after hip fracture. Within individuals, 106 of the 112 participants (94.6%) who had LEGS scores at both 2 and 12 months postfracture showed an improvement during this period. Most participants also improved their 6MWD (62 of 75 [82.7%]) and SFPF (99 of 134 [73.8%]) scores from 2 to 12 months postfracture. However, only 78 of 133 participants (58.6%) had higher YPAS scores after the same period of time.

Figures 1 and 2 illustrate the adjusted associations between baseline concentrations of α- and γ-tocopherol, respectively, and the physical function outcomes at 2, 6, and 12 months after hip fracture. The associations presented in these graphs represent the difference in physical function per SD of tocopherol and the 95% confidence interval. The baseline SDs of α- and γ-tocopherol were 15.3 and 3.8 μmol/L, respectively.

Figure 1.

Difference in physical function per SD (15.3 μmol/L) baseline α-tocopherol.

Figure 2.

Difference in physical function per SD (3.8 μmol/L) baseline γ-tocopherol.

Participants with higher baseline α-tocopherol had better average physical function (Figure 1). All cross-sectional 2-, 6-, and 12-month point estimates of this relationship were positive for each outcome at all time points. The global associations, assessing the longitudinal relationships with physical function including timextocopherol interactions, were statistically significant for 6MWD (α-tocopherol: p = .007) and YPAS (p = .002). Baseline α-tocopherol was also positively associated with faster average improvement in 6MWD and YPAS (p for Time × α-Tocopherol interaction < .008).

Figure 2 shows that there were no consistently meaningful relationships between baseline γ-tocopherol and physical function.

Figures 3 and 4 depict the adjusted associations between time-varying α- and γ-tocopherol, respectively, and the physical function outcomes at 2, 6, and 12 months after hip fracture. The time-varying SDs of α- and γ-tocopherol were 14.6 and 4.0 μmol/L, respectively.

Figure 3.

Difference in physical function per SD (14.6 μmol/L) time-varying α-tocopherol.

Figure 4.

Difference in physical function per SD (4.0 μmol/L) time-varying γ-tocopherol.

Higher concentrations of α-tocopherol throughout the year after fracture were positively associated with 6MWD, LEGS, YPAS, and SFPF (p < .0001) (Figure 3). The associations between α-tocopherol and physical function were strongest at 12 months after fracture for all outcomes. Time-varying α-tocopherol was also positively associated with faster average improvement in 6MWD, LEGS, and SFPF (p for Time × α-Tocopherol interaction < .003).

Figure 4 reveals that higher concentrations of γ-tocopherol throughout the year after fracture were also positively associated with LEGS, YPAS, and SFPF (p < .03). In contrast to both baseline and time-varying α-tocopherol, there was a negative change in 6MWD over time by time-varying γ-tocopherol (p for Time × γ-Tocopherol interaction < .0001); that is, participants with higher γ-tocopherol had greater decreases over time.

Figure 5 reveals that lagging time-varying α-tocopherol concentrations in the physical function models resulted in similar, and in some cases stronger, positive associations with all outcomes (p < .0007). Lagging γ-tocopherol did not meaningfully alter the associations with physical function.

Figure 5.

Difference in physical function per SD (14.6 μmol/L) time-varying lagged α-tocopherol.

Though the associations were slightly attenuated, inclusion of the serum markers of a healthy diet (carotenoids, albumin, and 25-hydroxyvitamin D) and the instrumental activity of daily living item capturing meal preparation limitations in the regression models did not appreciably change any of the relationships between baseline or time-varying tocopherols and physical function (results not shown).

DISCUSSION

Higher concentrations of vitamin E were associated with better physical function in this sample of community-dwelling, female hip fracture patients. Both α- and γ-tocopherol, the two primary forms of vitamin E, were positively associated with measured and self-reported physical function outcomes at points throughout the year after hip fracture. However, there were meaningful differences in the associations with physical function by the form and time of measurement of vitamin E. Stronger associations with physical function were noted for the α-tocopherol form than the γ-tocopherol form and for time-varying as opposed to baseline vitamin E.

The mechanism through which vitamin E may exert beneficial effects on physical function in hip fracture patients has not been established, though there are a number of potential explanations. As a potent class of lipid-soluble antioxidants, vitamin E may help quench the excessive oxidative stress resultant from the trauma of the hip fracture injury and subsequent surgical repair. Oxidative stress has been shown to damage muscle tissue, and antioxidants may protect against impairments in physical function induced by damaged muscle (35). Vitamin E may be of particular importance to muscle as α-tocopherol supplementation has been found to reduce muscle damage during surgery (36) and low serum concentrations of α-tocopherol have been associated with markers of sarcopenia among older adults (21). The relationships between excessive oxidative stress, damage to muscle tissue, and physical function are suggestive of a potential explanation for our findings.

The superior antioxidant properties of α-tocopherol may account for the fact that this form of vitamin E was generally more strongly associated with physical function than γ-tocopherol. Another possible explanation is that measurement of serum γ-tocopherol is less accurate than α-tocopherol, and less precise assessments could potentially result in tempered associations with physical function. Additionally, although there are many healthy dietary sources of γ-tocopherol such as walnuts and pecans, the primary source of γ-tocopherol in the U.S. diet is hydrogenated soybean oil (37). Hydrogenated soybean oil is highly inflammatory (38) and is contained in many processed and generally unhealthy foods. However, despite the potential for measurement issues and unhealthy sources of dietary intake, γ-tocopherol was positively associated with physical function at several specific points throughout the year after fracture, suggesting that regular intake of γ-tocopherol may be associated with better physical function during hip fracture recovery.

This was the first exploration of the relationship between vitamin E and physical function at multiple time points within a cohort of older adults. The short half-lives of α- and γ-tocopherol and the potential for changes in dietary intake over time suggest that more frequent measurements are necessary to assess the relationship between regular vitamin E consumption and the continual physical functioning of older adults. This may be a particularly important consideration among hip fracture patients, whose dietary patterns are likely to change at different stages during the postfracture year based on varying places of residence, the ability to prepare meals, and so on. We have addressed this limitation of previous studies of vitamin E and physical function by considering the unique relationships at baseline and 2-, 6-, and 12-month follow-up visits. Our study of serum tocopherols, considered the “gold standard” of dietary vitamin E intake (39), also minimizes the potential for recall bias associated with most questionnaire-based measures, thereby strengthening the dietary inferences of this analysis.

There were also a number of limitations to this work that merit consideration. First, our findings may not be generalizable to all hip fracture patients. The BHS4 cohort was exclusively female and relatively healthy and highly functioning. Thus, the associations noted between vitamin E and physical function are likely to be conservative estimates, as the typically less robust hip fracture patient would have more room to demonstrate functional improvement. A related concern was that participants with low baseline or prefracture physical function might have less room for functional decline, which could affect the associations with vitamin E. In order to address this possibility, we included baseline YPAS and SFPF by vitamin E interaction terms, which did not improve model fit. However, because both YPAS and SFPF refer to the 2-week period prior to hip fracture, the possibility for physical function floor effects after fracture cannot be ruled out. Another limitation is that serum lipid data were not collected in BHS4. Not adjusting for lipids may introduce a source of confounding, as serum lipids are strong predictors of vitamin E concentrations and low levels of lipids are associated with poor physical function (40). Major differences in lipid concentrations are unlikely within BHS4, however, due to the relative homogeneity of this cohort. As an observational study, another limitation is that the direction of the associations between vitamin E and physical function are not entirely clear. However, low serum vitamin E has previously been associated with incident decline in physical function (14). The lagged associations between vitamin E and physical function were also often stronger than those at the same visit, providing further support that higher vitamin E concentrations may improve recovery of physical function. A final limitation is that the positive associations noted between serum tocopherol concentrations and physical function may represent the effects of general nutritional status as opposed to vitamin E specifically. A number of measures were taken to address this concern, including adjusting for several additional dietary biomarkers (serum carotenoids, albumin, and vitamin D) in the regression models exploring the relationship between vitamin E and physical function. The inclusion of these other markers of a healthy diet did not appreciably change the results, suggesting that vitamin E may play a role in physical function. However, we were unable to account for total energy intake and other factors that may be related to both serum vitamin E concentrations and physical function, and these meaningful dietary characteristics should be addressed in future studies among hip fracture patients.

The associations uncovered in this study between both baseline and time-varying vitamin E concentrations and physical function suggest that older adults at high risk for hip fracture as well as those currently recovering from this injury could potentially benefit from vitamin E consumption. However, these findings will need to be confirmed in clinical trials before increased vitamin E intake can be recommended for hip fracture patients.

FUNDING

This work was supported by the National Institute on Aging at the National Institutes of Health (R01 AG018668, R37 AG09901, T32 AG00262, K23 AG027746, P60 AG12583, and P30 AG028747).