Introduction

Hip fracture is an important primary cause of disability in the elderly, increasing the burden of disease in the population and leading to substantial loss of healthy life-years in elderly people [1]. Hip fracture patients have a high prevalence of vitamin D deficiency which has been associated with osteoporosis, muscle weakness, higher fall, and fracture risk in the elderly population [2, 3]. In patients with hip fractures, vitamin D deficiency has also been associated with increased short-term and long-term mortality [4,5,6,7,8], increased complications [5], and poorer cognitive status [9].

There are controversies in the literature concerning vitamin D, and the definition of optimal vitamin D status is still unclear in the literature [10]. There is also no defined link between vitamin D levels and propensity for other metabolic bone diseases apart from rickets, such as osteoporosis which is a well-established risk factor for hip fracture.

Currently, studies have demonstrated that low vitamin D levels are associated with poorer functional outcomes, but some of these studies only assessed short-term outcomes [11, 12], had population sizes less than two hundred patients [12,13,14], did not account for pre-fracture baseline function [11], and importantly, did not assess patients utilizing previously defined vitamin D thresholds [12, 14, 15]. Evaluating serum vitamin D thresholds through studying the effect of vitamin D on functional outcomes may help define target levels of serum vitamin D to optimize recovery of hip fracture patients.

Aim

The aim of this study is to assess for any association between vitamin D levels on admission, stratified along previously defined vitamin D thresholds, and 6-month functional outcome scores after hip fracture surgery, after adjusting for baseline function and other confounders.

Hypothesis

Our hypothesis is that severe vitamin D deficiency is an independent risk factor for poorer 6-month functional outcome after surgery for hip fractures.

Methods

Patient selection

The medical records of patients who underwent hip fracture surgery in our institution, Singapore General Hospital, from January 2012 to December 2016, and had vitamin D levels assessed during admission, were reviewed.

We excluded patients who had pathological fractures, prior surgeries in involved hips, and those resulting from high energy trauma.

Retrospective analysis was performed on patients’ demographic data such as age, gender, and clinical parameters such as preoperative vitamin D and haemoglobin levels measured on admission, co-morbidity burden as assessed by Charlson Comorbidity Index (CCI), and type and site of surgery.

Vitamin D definition

Vitamin D in this paper refers to total serum 25-hydroxycholecalciferol, which is reported elsewhere as 25-hydroxyvitamin D or 25(OH) vitamin D or calcifediol. Vitamin D levels were assessed on admission together with other parameters such as serum calcium for bone health assessment and osteoporosis workup.

According to definitions by the Endocrine Society [2], total serum 25-hydroxycholecalciferol levels between of 20–30 ng/ml is termed as vitamin D insufficiency, and that of below 20 ng/ml is termed as deficiency.

Our study utilized ≤10 ng/ml to define severe deficiency; this measure was used in previous studies [4, 5, 16, 17]. Thus, we defined mild vitamin D deficiency using levels of 10–20ng/ml in this study and previous definitions of vitamin D insufficiency (20–30 ng/ml) and normal vitamin D (>30 ng/ml) were retained.

Anaemia definition

Anaemia was defined according to World Health Organization (WHO) gender-based classification. In men, mild anaemia was defined at haemoglobin 11.0–12.9 g/dL; in women, mild anaemia was defined at haemoglobin 11.0–11.9 g/dL. While for both genders, moderate anaemia was defined at haemoglobin 8.0–10.9g/dL and severe anaemia was defined at haemoglobin < 8.0g/dL

Charlson Comorbidity Index

This study used Charlson Comorbidity Index (CCI) to assess co-morbidity burden in hip fracture patients. The CCI score is based on number of conditions including previous myocardial infarction, stroke, and liver disease that are each assigned different weights, with a higher weight representing more severe morbidity. The summation of the weighted comorbidity scores results in a summary score. For statistical analysis, patients in this study were divided into two groups according to their CCI score, CCI 0–1 and CCI ≥2.

Functional outcomes

Functional outcome was measured by the Harris Hip Score and the Parker Mobility Score, while quality of life was assessed using the various domains of the 36-Item Short Form Health Survey (SF36). Harris Hip Score is an assessment of patients’ hip with a maximum score of 100 points. Greater emphasis is placed on function as assessed by activities of daily living, gait (47 points), and pain (44 points), while range of motion (5 points) and absence of deformity (4 points) have less weightage in the Harris Hip Score. Parker Mobility Score measures patients’ ambulatory status in different situations—indoors, outdoors, or whilst doing shopping, and scores patients from 0 to 9. The use of SF36, Parker Mobility Score, and Harris Hip Score to evaluate outcomes after hip fracture surgery has been validated in previous recent studies [13, 18]. Assessment was performed by independent third-party trained physiotherapists from our Orthopaedic Diagnostic Centre. These assessments were done at baseline on admission and at follow up at 6 months post-discharge at the outpatient clinic. Assessments were done independent of the study and were routinely done in all hip fracture patients.

The duration of 6 months was chosen in view of previous studies which show that functional recovery after hip fracture may be largely complete in the first 6 months [13] and is sufficient to assess midterm recovery from a hip fracture.

For the SF36, patients are assessed using eight different domains divided into two groups: four physical health domains which include physical functioning (SF36 PF), role limitations as a result of physical health problems (SF36 RP), bodily pain (SF36 BP), and general health (SF36 GH), as well as four mental health domains which include vitality (SF36 VI), social functioning (SF36 SF), role limitations as a result of emotional problems (SF36 RE), and mental health perceptions (SF 36 MH).

Baseline pre-operative Harris Hip Score was unavailable, and Harris Hip Score was only assessed at 6 months, as it is not practical to assess pre-operative deformity and range of motion upon admission, due to acute pain.

Statistical analysis

Analysis was performed to determine any correlation between vitamin D levels, and various demographic and clinical parameters. Spearman correlation was performed between two continuous/ordinal variables to determine any monotonic relationships, while the Mann-Whitney U test was used to determine if median vitamin D levels were similar between groups of different demographic or clinical parameters.

Patients were subsequently divided into 4 groups according to the vitamin D levels previously defined. Univariate analysis was performed using one-way ANOVA to determine if there was any difference in baseline and 6-month functional outcome scores among the different vitamin D groups.

Multivariate analysis with general linear model was performed on the respective 6-month scores that demonstrate statistically significant differences in outcome amongst vitamin D groups, in order to ascertain the association between vitamin D and recovery from hip surgery, while accounting for other confounders.

All analyses were conducted using SPSS Statistics for Windows version 22.0 (IBM Corp., Armonk, NY). A p < 0.05 is considered statistically significant.

Results

Hospital surgical logs identified 664 patients who underwent surgical treatment of osteoporotic hip fractures with available serum total 25-hydroxyvitamin D level recorded between January 2012 and December 2016. The characteristics of our cohort can be seen in Table 1. This study population had a mean age of 77 years; 29% were male and 71% female. Joint-replacing procedures were performed in 55%, and joint-preserving (fixation) procedures in 45%. Of these 664 patients, 9% had severe vitamin D deficiency (≤10 ng/mL), 39% mild vitamin D deficiency (10–20 ng/ml), 42% vitamin D insufficiency (20–30 ng/ml), and 10% normal vitamin D levels (>30ng/ml). Univariate analysis using one-way ANOVA demonstrated that there was a significant difference in 6-month outcome across vitamin D groups as assessed by Parker Mobility Score (p =.003) and SF36 PF (p=.047) (Table 2). There was no significant difference found among vitamin D groups for the other outcome scores.

Table 1 Study cohort characteristics
Full size table
Table 2 Analysis of 6-month and baseline Parker Mobility Score, Harris Hip Score, and 36-Item Short Form Health Survey (SF36) physical functioning (SF36 PF), role limitations as a result of physical health problems (SF36 RP), bodily pain (SF36) general health (SF36 GH) vitality (SF36 VI), social functioning (SF36 SF), role limitations as a result of emotional problems (SF36 RE), and mental health perceptions (SF 36 MH) score according to vitamin D groups using one-way ANOVA test
Full size table

In a separate analysis, increasing age was significantly correlated to poorer Parker Mobility Score and SF36 PF at 6 months. Parker Mobility Score and SF36 PF were both inversely associated with increasing age (p<0.001 for both), and thus, an initial multivariate analysis investigating the association between vitamin D level, age, and functional outcomes was performed which demonstrated that there is a significant difference in both 6-month PMS and SF36 PF between patients in severe vitamin D deficiency group and patients in all the other vitamin D groups when corrected only for patient’s age (Table 3).

Table 3 Age-adjusted multivariate analysis examining association between preoperative serum total 25-hydroxyvitamin D and 6-month Parker Mobility Score and Medical Outcomes Study 36-Item Short Form Survey (SF36) Physical Functioning domain after hip fracture surgery
Full size table

After further multivariate analysis, adjusting also for preoperative anaemia, type of surgery, gender, other co-morbidities, and baseline score, severe vitamin D deficiency was found to be an independent risk factor for lower 6-month Parker Mobility Score and SF36 PF(Table 4). Severe vitamin D deficiency was still found to be an independent risk factor for lower 6-month Parker Mobility Score and SF36 PF in the multivariate analysis when baseline scores were excluded from the analysis. Preoperative anaemia, age, and baseline score were found to be independent risk factors for poorer Parker Mobility Score and SF36 PF after hip fracture surgery. Additionally, CCI was found to be an independent risk factor for Parker Mobility Score after hip fracture surgery.

Table 4 Results of multivariate analysis examining association between preoperative serum total 25-hydroxyvitamin D and 6-month Parker Mobility Score and Medical Outcomes Study 36-Item Short Form Survey (SF36) Physical Functioning domain after hip fracture surgery
Full size table

Discussion

The prevalence of vitamin D deficiency (<20ng/ml) in our cohort of surgically treated hip fracture is 48%, which is significantly lower than reported rates of vitamin D deficiency in other study populations (77–87%) [4, 5, 7]. We postulate that this could be due to our equatorial climate. Vitamin D supplementation, type of dwelling, and seasonality have been shown to affect vitamin D levels of hip fracture patients [19]. There was no association between serum vitamin D levels and other clinical parameters such as age, comorbidities, anaemia, and gender (Table 5), which is consistent with the literature [20].

Table 5 Investigation of associations between patient characteristics and vitamin D levels
Full size table

Our study found an association between lower vitamin D levels and poorer baseline ambulatory function based on Parker Mobility Score and quality of life in physical functioning (Tables 2 and 3). The latter measure encompasses carrying out activities of daily living such as groceries, bathing, and dressing. In addition, severe vitamin D deficiency (<10ng/ml) was an independent risk factor for lower Parker Mobility Score and SF36 PF at 6 months after hip fracture surgery in this group, even after accounting for the baseline differences above, meaning that the detrimental effect of severe vitamin D is sustained till at least midterm recovery after a hip fracture.

Our findings build upon the findings of previous studies. Pioli found an association between severe vitamin D deficiency and recovery of pre-fall ambulatory status at 6 months [15]. Di Monaco found an association between severe vitamin D deficiency and a poorer Barthel Index score, which assesses independence in activities of daily living (ADL), after discharge from rehabilitation facilities [11]. Other studies by Leboff and Fischer found associations between severe vitamin D deficiency and poorer lower limb strength over 1 year, and this may indirectly affect ambulatory function and quality of life [13, 14].

While the above studies were able to demonstrate an association between vitamin D deficiency and outcomes after hip fracture surgery, most of these studies were only able to demonstrate this association at a single cut-off value around the range of 6–10ng/ml or severe vitamin D deficiency, similar to that found in our study [12,13,14,15].

However, the study by Di Monaco was able to demonstrate a difference in outcome between mild vitamin D deficiency and vitamin D insufficiency, as they used validated vitamin D classes to determine their prognostic capability of the functional recovery after hip fracture [11]. They found that there was a stepwise decrease in ADL when comparing patients with vitamin D insufficiency to those with mild deficiency, and then again between patients who had mild deficiency to those with severe deficiency. There was however no difference in ADL between patients with insufficiency and normal vitamin D levels. Hence, they recommended that patients with vitamin D deficiency should receive supplementation but that there was minimal evidence from their study to increase the threshold for vitamin D supplementation beyond 20ng/ml [11].

We were unable to demonstrate such a difference in outcome between patients with mild vitamin D deficiency and vitamin D insufficiency. This may be due to the fact that their outcomes were assessed at discharge from rehabilitation facility, whilst ours were assessed at 6 months post-operatively. Their study also utilized a different set of validated outcome scores from this study.

In patients who sustain hip fractures, the majority of which require surgical intervention, Vitamin D deficiency is related to increased complications including pneumonia [5], thromboembolic events [5], and poorer cognitive status [9] which affects rehabilitation after hip fracture surgery. Baseline vitamin D function may also affect recovery after hip fracture through its skeletal and extra-skeletal effects. Vitamin D deficiency is linked to reduced muscle strength and function [3], increased fracture comminution and hence slower recovery [20], impaired bone formation in patients with hip fracture [21], and altered bone turnover markers which are surrogate measures of fracture healing [22]. Our study consolidates the point that pre- and post-hip fracture quality of life is affected by severe Vitamin D deficiency, largely in the domains of physical functioning, consistent with the known effects of Vitamin D on the musculoskeletal system. This is especially important as we strive to develop a healthy and active aging population who may remain socioeconomically viable well into their senior years.

Hence, understanding the association between low vitamin D levels and recovery from hip fracture can help clinicians to identify patients who are at risk for poorer recovery as a target for focused rehabilitation and potentially nutritional supplementation. Further studies are still required to confirm the benefit of vitamin D supplementation in the hip fracture population, many of whom have severe vitamin D deficiency.

Strengths and limitations

This is the study of a large cohort of elderly adults with hip fracture that demonstrates the negative association between severe preoperative serum vitamin D deficiency and midterm functional recovery after hip fracture surgery even after accounting for other confounders. Functional recovery and quality of life were assessed using validated tools. We have a follow-up rate of about 92% of patients who were able to return for assessment at 6 months.

This study however was unable to account for patients who received vitamin D supplementation before their fracture and was also unable to study the effect of vitamin D supplementation after surgical treatment for hip fractures. Our study design was only able to elucidate the association between vitamin D at admission and functional recovery, but unable to determine causality.

Conclusion

Prevalence of vitamin D deficiency is high in surgically treated hip fracture patients. Preoperative severe vitamin D deficiency has been shown to be an independent risk factor for slower recovery of function and quality of life after hip fracture surgery. Prospective studies are required to confirm the association between vitamin D and outcomes after hip fracture surgery, to re-define optimal vitamin D levels, and to investigate if vitamin D supplementation can improve recovery in patients with hip fractures.