Vitamin D supplementation and mortality risk in chronic kidney disease: a meta-analysis of 20 observational studies

Mineral and bone disorders (MBD) are early and common complications of CKD, and progress as glomerular filtration rate (GFR) declines. Multiple factors contribute to the development and maintenance of CKD-MBD, but principally involve phosphate retention and vitamin D metabolism abnormalities. Kidney Disease: Improving Global Outcomes defines chronic kidney disease-mineral and bone disorder (CKD-MBD) as a systemic syndrome characterized by abnormalities in serum calcium, phosphorus and parathyroid hormone (PTH) concentration, vitamin D metabolism, and bone turnover [1]. This syndrome is common among CKD patients and has been associated with an increased risk of cardiovascular calcification [2, 3] and mortality [4]. The Third National Health and Nutrition Examination Survey (NHANES III) reported 15068 adults patients with vitamin D deficiency and demonstrated a higher prevalence of cardiovascular disease and mortality in untreated patients [5]. An association between vitamin D deficiency and other traditional cardiovascular risk factors, such as hypertension, insulin resistance, diabetes, and dyslipidemia, has also been reported [6, 7]. The recognition of biochemical components of CKD-MBD associated with increased mortality in dialysis patients [8] and in patients with CKD not treated with dialysis [9] has provided an impetus to explore the effect of these factors on survival and associated treatment modalities. Numerous reports have characterized the nonskeletal benefits of vitamin D [10].

Wang et al. and Pittas et al. reported the benefits of vitamin D supplementation on cardiovascular disease (CVD) in the general population [11, 12]. Nutritional vitamin D supplementation has also been reported to be beneficial to CKD patients [13]. Most reviews, however, had few participants, short follow-up, and lacked survival analyses. We conducted a systematic review of the literature to assess whether vitamin D supplements reduced mortality in patients with ESRD on dialysis and patients with CKD not requiring dialysis.

Active vitamin D compounds were associated with a reduced risk of mortality in patients with ESRD on dialysis and patients with CKD not requiring dialysis. Several mechanisms may explain how vitamin D can modify risk for mortality. Vitamin D down regulates the renin-angiotensin system [45], improves insulin secretion and sensitivity [46], inhibits vascular smooth-muscle cell proliferation [47], protects normal endothelial cell function [48], modulates inflammatory processes [49], inhibits anticoagulant activity [50], and inhibits myocardial cell hypertrophy and proliferation [51]. These findings suggest that vitamin D may decrease mortality through multiple pathways. Although the actual mechanism of mortality is unclear, patient death has been associated with vascular calcifications, left-ventricular hypertrophy and left-ventricular dysfunction. The multi-organ protective effects of vitamin D may explain the lower mortality rate found in these patients.

A fixed covariate baseline Cox model was used in the majority of included studies. Only 5 studies used a time-dependent Cox model to analyze the relationship between active vitamin D use and survival. Although a standard baseline Cox proportional regression model is usually used to analyze cohort studies, it may be inadequate to evaluate active vitamin D treatments due to the presence of time-dependent variation in outcome. Higher serum calcium and phosphorus levels were consistently associated with increased risk of death [4, 52]. Elevated serum PTH levels have also been associated with increased mortality [4, 33]. The serum levels of calcium, phosphorus and PTH are affected by vitamin D therapy. Serum levels of calcium, phosphorus and PTH vary during the course of vitamin D therapy and affect patient outcome. These mineral metabolism indexes are recognized as time-dependent confounders. Time-dependent confounders cannot be controlled by conventional survival analysis methods [53]. Marginal structural modeling (MSM) can control for time-dependent confounders affected by prior treatment [54]. Under some conditions, the treatment estimate from a MSM can have the same causal interpretation as an estimate from a randomized clinical trial [55]. Only the Tentori et al. study reported detailed data regarding the survival advantage of patients treated with active vitamin D. The unadjusted baseline Cox model and time-varying MSM models demonstrated a 16% and 22%, respectively, reduction of all-cause mortality associated with active vitamin D treatment. Most studies included in this meta-analysis had some selection bias. For example, the study of Teng et al. [27] had statistical differences in the baseline characteristics of patient age, primary cause of renal failure, body mass index, blood pressure, and intact parathyroid hormone and hemoglobin levels. Several studies included in the meta-analysis used sophisticated statistical techniques, such as adjustment for time-dependent confounders, propensity score-matching or marginal structural models, to mimic the design of randomized controlled trials. Only the characteristics of patients that were treated with vitamin D analogues were known to the researchers. Any confounding factors would be controlled by these statistical methods and the results would be comparable to randomized controlled trials. The problem with the observational studies was that such knowledge was not available. The potential presence of unmeasured confounders prevented any conclusions of causation, even when sophisticated statistical methods were used. The survival advantage associated with active vitamin D treatment occurred in a dose-dependent manner. This phenomenon has been supported by two studies [29, 32]. There has been no well-designed dose gradient study to test this hypothesis. Although we do not have higher quality evidence to prove this association, we believe that vitamin D will improve survival.

The meta-analysis detected slight differences in survival associated with different analogues of active vitamin D. The baseline case mixed and MICS Cox models detected a 5% lower mortality with paricalcitol treatment than with calcitriol treatment. This slight survival difference may be explained by differential effects of calcitriol and its analogue, paricalcitol on vascular calcification. In vitro studies have demonstrated that calcitriol is a growth factor for vascular smooth muscle cells, while the analogue, paricalcitol, is not [56]. In vivo studies have shown that vitamin D sterols have a differential effect on vascular calcification. 1-α-hydroxy vitamin D (calcitriol) was associated with greater vascular calcification than paricalcitol, even though there was equivalent suppression of PTH in these animal models [57]. Only two well-designed cohort studies or randomized controlled trials, Teng et al. [25] and Tentori et al. [30], have evaluated the mortality risk associated with different active vitamin D analogues. Further studies are needed to clarify the survival difference before one vitamin D analogue is recommended over another in clinical practice.

Three studies included in the meta-analysis reported mortality risk associated with different mean daily or weekly doses of vitamin D. In the Naves-Diaz et al. study, the maximum reduction of mortality occurred when the mean daily dose of oral calcitriol was less than 0.25 ug. This survival benefit was lost as the mean daily calcitriol dose was increased to more than 1.0 ug. This dose-dependent benefit effect was also reported with paricalcitol. Kalantar-Zadeh et al. reported patients treated with mean weekly doses of 1.0 ug to 5.0 ug of paricalcitol. Mean weekly doses of paricalcitol above 15.0 ug were associated with an 18% reduction of mortality risk. A possible explanation is that low-dose vitamin D exerts weaker anti-vascular calcification effects than higher doses in CKD patients. High doses of vitamin D could be associated with adverse effects, such as hypercalcemia, that would overwhelm its protective effects.

There were several limitations in our meta-analysis. First, only a few of the included studies used a time-dependent or marginal structural model to analyze the follow-up data. The majority of studies had limited power to draw a definitive conclusion on the effects of vitamin D supplements on all-cause or cardiovascular mortality. Second, there was high heterogeneity in the meta-analysis. Sample size and publication year were the sources of heterogeneity. Third, the possible sources of heterogeneity could not be carefully examined. This included observational studies of the use of recombinant erythropoietin to correct anemia and studies of phosphorus binders to ameliorate hyperphosphatemia in patients with CKD that showed beneficial effects on mortality, CVD outcome, and progression of renal disease. Fourth, we did not seek to identify unpublished studies and several studies were excluded because the published data were not suitable for meta-analysis.

Active vitamin D compounds used to treat abnormal calcium, phosphorus and PTH levels in patients with either ESRD on dialysis or CKD not requiring dialysis. Active vitamin D compound treatment was associated with decreased all cause and cardiovascular mortality. Low dose active vitamin D compounds were associated with improved survival. Large, well designed randomized trials of active vitamin D supplements with different doses are needed to elucidate the role of vitamin D supplementation in reducing mortality.