Vitamin D Status Appears Unrelated to Fractures in Patients with Chronic Kidney Disease

Background: Although lack of Vitamin D is widespread in chronic kidney disease, data is scarce in the role Vitamin D may play in fractures in such cases. Methods: Retrospective analysis of all patients visiting Nephrology outpatients department in 1 UK District General Hospital over 2 years. Chronic kidney disease was 1.00), p = 0.028. Conclusion: Vitamin D deficiency appears widely prevalent in chronic kidney disease with a third of patients sustaining fractures; total Vitamin D levels however are unrelated to fractures. Prospective interventional studies can help answer if earlier replacement of Vitamin D before chronic kidney disease develops will help improve musculoskeletal health and prevent fractures.


INTRODUCTION
Chronic kidney disease (CKD) affects 5-10% of the world population [1] and is associated with many adverse outcomes including bone disorders and fractures. CKD is characterised into several stages from 1 through to 5 on basis of estimates of glomerular filtration rates (eGFR) [2]. Decreased bone mass and disruption of bone micro-architecture occur early in the course of CKD and worsens with the progressive decline in renal function so that at the time of initiation of dialysis at least 50% of patients have had a fracture [1]. The classic sequence of events begins with a deficit of calcitriol synthesis and retention of phosphorus [3] although more recently an association has been found with the fibroblast growth factor 23 (FGF 23) in response to 'invisible' phosphate retention [4].
In a 2009 survey about 15% of patients in the UK were found to have CKD stages 1-5 [5]. Elderly patient groups are developing end stage renal disease in the UK in record numbers; patients aged over 85 accepted onto the UK dialysis programmes nearly doubled between 2006 and 2011 and the percentage of dialysis patients who are aged greater than 70 years has increased from 19.2% in 2000 to 24.9% in 2012 [6].
The aetiology of fractures in patients with CKD is multifactorial; quantity of bone, osteoporosis with age and alterations in the quality of bone are all suspected factors. While bone quantity and quality can be assessed by laboratory or radiology based techniques, other unmeasured risk factors also likely contributory to fractures in CKD could be poor nutrition, inactivity, medications and increased risk of falling from myopathy or peripheral neuropathy [1]. Elderly and frail patients may have bone disease from age related osteoporosis and therefore are additionally vulnerable to other forms of mineral bone disease associated with the various stages of CKD. The risk of fracture then increases exponentially in this group increasing the burden on stretched NHS resources. Length of stay in an acute care hospital is estimated to be 3 weeks following a hip fracture with a quarter of patients remaining in long-term care institutions for at least a year and a third returning home depend on other people or devices for mobility [7].
As Vitamin D deficiency is known to be widespread in CKD, we hypothesized that increased fracture risk of any kind, fragility related or unrelated, in CKD patients was associated with Vitamin D deficiency.

METHODS
This retrospectively analysed study was undertaken at a UK District General Hospital, Countess of Chester NHS Trust that caters to ~250000 population in West Cheshire, North West UK. We retrospectively analysed all total Vitamin D levels measured in patients visiting the Nephrology OP between July 2011 -May 2013 by accessing records via the Hospital laboratory information technology system, MEDITECH TM . Dates when total Vitamin D was measured were available in all cases from MEDITECH TM . Total Vitamin D was the sum of Vitamin D2 and D3; all results for total Vitamin D were obtained from Birmingham Children's Hospital UK. Total Vitamin D results were then matched to each patient's kidney disease status via their serum creatinine and the eGFR. We also obtained serum albumin, calcium and phosphate and the parathyroid hormone (PTH) level for each patient. Any fracture, fragility related or unrelated that was recorded on MEDITECH TM were then obtained for each patient along with dates of occurrence. A total of 8 fractures occurred between 2001-2009, 1 additional fracture did not have a date of incidence, 11 fractures occurred between 2010-2013. For purposes of this study we analysed all fracture incidences. Sex and age of each patient was recorded. Severe deficiency of total Vitamin D was defined as less than 25nmol/L. Inadequate status of vitamin D was considered to be at a level less than 50nmols/L whereas anything higher than 50nmol/L was assumed to be adequate Vitamin D status. 19 total Vitamin D measurements in this study were during May to September and defined as Summer and the rest during October to April, defined as Winter. As data was irreversibly anonymised we did not seek patient consent or ethics approval for this study.

Statistical Analysis
Continuous patient demographic characteristics -including eGFR and Vitamin D -were recorded as mean (standard deviation) or median (interquartile range) depending on their distribution. Categorical variables were recorded as frequency (percentage). Spearman's correlation coefficient (r s ) was used to assess the relationship between continuous variables. Univariate logistic regression was used to assess the relationship between each variable and incidence of fracture (yes/no). A stepwise multivariable model selection procedure, based on Akaike's information criterion (AIC), was employed in order to find a final multivariable model. The sensitivity of the final model to the inclusion and exclusion and covariates and relevant two-way interactions was also assessed using AIC. The Hosmer-Lemeshow goodness-offit test was used to assess the adequacy of model fit. All point estimates are reported with accompanying 95% confidence intervals (CI). All statistical analyses were conducted using Stata 13 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP).  Fig. 1). This result persisted irrespective of eGFR (interaction test, p=0.971), i.e. there is no evidence that relationship between total Vitamin D level and risk of fracture is affected by eGFR. eGFR was however found to be negatively associated with risk of fracture, OR 0.96 (0.93, 1.00), p = 0.028 (Fig. 1). More intuitively this gives a risk of fracture for patients with an eGFR of 20, 40, 60 and 80 as 42%, 26%, 14% and 7% respectively.

RESULTS
There was no evidence of a relationship between eGFR and total Vitamin D levels (r s = -0.03, p = 0.8066, Fig. 2). No other variables were found to be univariately associated with fracture (Table 2) although sex was close to being statistically significant, OR 0.27 (95% CI: 0.07, 1.07) p = 0.063, with 3 out of 21 males (14%) and 17 out of 45 females (38%) experiencing a fracture. Although sex and eGFR were selected in the final multivariable model, sex offered a negligible improvement in model fit according to Akaike's information criterion (AIC) and therefore only eGFR was retained.

DISCUSSION
This small retrospective study demonstrates that in the North West of UK in patients with CKD, Vitamin D levels are unassociated with fractures; other than a low eGFR there are no other factors in the chemical biochemistry profile that appear to influence this outcome.
The most recent UK national diet and nutrition survey showed 20% suffer from severe Vitamin D deficiency with 60% insufficient of Vitamin Din both adults and children [8]. Vitamin D supplementation can potentially reduce falls and non-vertebral fractures, including those at the hip. This benefit is dose-dependent with doses of 700 to 1000 IU Vitamin D per day reducing falls by 19% and doses of 482 to 770 IU Vitamin D per day reducing non-vertebral fractures by 20% and hip fractures by 18% [9].   The largest observational study of an unselected cohort of patients with varying levels of kidney dysfunction, prevalence of abnormalities of vitamin D metabolites, iPTH, and Ca and P was published in 2007 [12]. Other studies in the field were from USA [13,14,15]. Levin's study [12] demonstrated low 1, 25 OH2 D3 in 13% in those with eGFR >80 ml/min, >60% in those with eGFR <30 ml/min. In this study a Vitamin D deficiency or replete state appeared unassociated with eGFR levels.
Whilst the association between fractures and patients with end stage renal disease has been explored to some extent [16]  Whether general Vitamin D replacement would transpire into fruitful musculoskeletal health continues to remain unclear; in one metaanalysis in this field the authors conclude skeletal effects of vitamin D supplementation have not been studied [19]. Urgent interventional/longitudinal studies are required to see if early Vitamin D replacement before onset of disease states alters musculoskeletal health both in the wider population but also in those that are destined to develop CKD over time.
As all retrospective studies do this too has several drawbacks. This was a single centre study with few patients. As this study was undertaken retrospectively there was no sample size estimate -we also believe post-hoc sample size calculations are inappropriate. We did however limit the number of variables permitted into the multivariable analysis in order to satisfy the 10 events per variable rule of thumb. Total Vitamin D levels were only available in a few patients and it is possible the patient sample in the final analysis may not be fully representative of other Renal Out Patient Departments. Laboratory results were obtained at single time points and CKD status or Vitamin D status could have been unrelated to their Orthopaedic incidents. All fractures were included for analysis; although 9/20 fractures occurred earlier than 2010, excluding these 9 fractures that occurred before 2010 did not affect the interpretation of the results. As undiagnosed Vitamin D deficiency or CKD are chronic long term conditions and may have been present before presentation to Nephrology Out Patients we justified including all fractures into the study. We did not specifically study the types of fracture, fragility related or unrelated as numbers of patients were too small to achieve statistical significance.

CONCLUSION
Although widespread, Vitamin D deficiency does not appear to be associated with fractures in patients with CKD; Vitamin D deficiency is also unrelated to stages of underlying CKD. Almost a third of CKD patients sustain fractures and only the stage of renal disease appears to be predictive of fractures in patients with CKD. If fractures are unrelated to Vitamin D levels in patients with CKD can Vitamin D replacement in the wider population help improve musculoskeletal health to a degree that will then prevent fractures in elderly or in those that are predestined to develop CKD over time? Only long term prospective population based interventional Vitamin D studies could help answer such a question.