Efficacy and safety of body weight-adapted oral cholecalciferol substitution in dialysis patients with vitamin D deficiency

Whether individuals with and without kidney disease suffering from vitamin D deficiency need vitamin D substitution and how this substitution should be given are matters of debate. When determining the optimal vitamin D dose it is crucial to understand the kinetics of conversion of vitamin D to 25OHD₃. At present, the exact liver enzyme or enzymes for 25-hydroxylation are still unknown, but the most likely candidates are CYP2R1, CYP3A4 and CYP27A [12‐14].

Heaney described a biphasic pattern of 25-hydroxylation. At low doses of cholecalciferol the compound is completely converted to 25OHD₃, meaning that cholecalciferol concentrations remain low and 25OHD₃ serum concentrations rise rapidly [15]. When daily cholecalciferol doses above 2000 IU are given, the enzymes become saturated, serum cholecalciferol rises and some of it is stored in fat tissue, and 25OHD₃ serum levels continue to rise slowly. The 25OHD₃ breakpoint of this biphasic response is at about 35 ng/mL, which is practically identical to the level, below which PTH levels begin to rise in response to low 25OHD₃ levels [16]. This study clearly explains why 25OHD₃ increases rapidly under low doses of cholecalciferol and that supraphysiologic doses above 2000 UI per day cause an only marginal further increase. Although these results were obtained in healthy individuals, there is no reason to believe that CKD patients would behave differently. In fact, published studies of cholecalciferol substitution in CKD and hemodialysis patients are in line with this observation [9, 10, 17‐22].

We are not aware of any study in dialysis patients investigating a body weight-based oral cholecalciferol dosing regimen. Body weight is the most important determinant of the 25OHD₃ response to a given cholecalciferol dose [11, 23]. This phenomenon may be partly explained by the fact that in patients with higher body weight and more body fat an increasing proportion of administered vitamin D is stored in fat tissue and not available for 25-hydroxylation [24]. Earlier studies with oral vitamin D2 (ergocalciferol) supplementation have shown an inverse correlation between peak serum vitamin D2 levels and body-mass-index or obesity, respectively [24] . In older (≥64 years) healthy persons, but not in younger adults (20–40 years) a significant negative association between baseline 25OHD₃ and BMI or adiposity was found. When these persons were supplemented in a randomized placebo-controlled trial (cholecalciferol 600 IU/day vs placebo), the 25OHD₃ response to supplemental cholecalciferol was negatively associated with baseline BMI (the mean adjusted change in 25OHD₃ decreased by approximately 2.6 ng/mL with every 5 kg/m² increase in BMI) [25]. Therefore, we believe that the body-weight adapted supplementation should not be adjusted to lean body weight and might possibly be even more precise when corrected for BMI or fat mass. This very interesting point should be further investigated in future studies, ideally with quantifying the vitamin D content in adipose tissue.

Van Groningen calculated the loading dose of cholecalciferol needed to reach a 25OHD₃ level of 30 ng/mL using the formula IU = 100 x (30-25OHD₃) x kg BW. Using this formula, the mean dose for our patients would have been 152.000 IU or a weekly dose of 5850 IU for 26 weeks [23]. In a recent meta-analysis Zittermann et al. showed an exponential dose requirement when trying to achieve higher 25OHD₃ levels [11]. They found a logarithmic association between cholecalciferol dose per kg body weight and the increment in 25OHD_3. For a 70-year-old person with a body weight of 75 kg the calculated daily dose to achieve a target of 20 ng/mL was 308 IU and for a target of 30 ng/mL the daily dose was 1460 IU (corresponding to a weekly dose of 10.220 IU). The weekly dose we used based on the recommendation of 100 IU per kg body weight lies between these calculated doses. There are other studies of comparable vitamin D doses in dialysis patients. Armas et al. used a weekly cholecalciferol dose of 10.333 IU in twenty patients for 15 weeks. Serum 25OHD₃ increased by 23.6 ng/mL [9]. Jean et al. tested a mean daily dose of 640 IU (400–1200 IU) in 149 hemodialysis patients over six months [10]. They found an impressive increase in 25OHD₃ of 38.4 ng/mL. This favorable result compared to ours and that of Armas et al. may be explained by the use of daily low doses as compared to weekly higher doses. These frequent low doses possibly allow for more quantitative 25-hydroxylation in the liver [15]. These and our study clearly show that these rather low doses are effective for reaching 25OHD₃ levels at or close to the target range of 30 ng/mL. From these data we do not believe that daily doses in excess of 3.500 IU are necessary to correct vitamin D deficiency in dialysis patients, as recently suggested [26]. Of note, these dose recommendations apply only to cholecalciferol and not to ergocalciferol treatment. As ergocalciferol seems to be less effective in increasing 25OHD₃ levels, higher doses may be required when using vitamin D₂ [27]. Although the optimal 25OHD₃ level for CKD and dialysis patients is unknown, a target of 30 ng/mL may be reasonable for maximal PTH suppression. Using our dosing regimen the mean 25OHD₃ level was 26 ng/ml, but only a minority of patients achieved the 30 ng/mL goal. Therefore, it may be necessary to increase the weekly dose especially in those patients with very low basal levels [23]. Keeping in mind the ongoing debate about erythropoietin and iron therapy in dialysis patients, we are cautious about recommendations of high and unphysiological cholecalciferol or ergocalciferol doses. It may be wise to use the lowest possible vitamin D dose to achieve a (yet to be clearly defined) 25OHD₃ target level.

In our study cinacalcet therapy was associated with a higher increase in 25OHD₃. This association has not been described until now. Cinacalcet has been shown to reduce FGF-23 levels in dialysis patients independently from its effect on iPTH [28‐30]. In animal models, FGF-23 induces expression of CYP24A1 or 24-hydroxylase in the liver [31, 32]. It could be speculated that induction of CYP24A1 by FGF-23 would lead to enhanced 24-hydroxylation of 25OHD₃ to 24,25OHD₃, the first step in 25OHD₃ catabolism and therefore contribute to vitamin D deficiency in CKD. However, 24,25OHD₃ levels have been found to be rather low in this group of patients despite elevated FGF-23 [32‐34]. Nevertheless, we speculate that reduction of FGF-23 and consecutive decrease of CYP24A1 activity and 24-hydroxylation can explain our observation. Clearly, further studies are needed to confirm this finding and mechanism in dialysis patients.

The lower response in peritoneal dialysis patients is most likely explained by lower compliance with home-based intake than in hemodialysis patients, who were given their medication at the dialysis ward resulting in a 100 % compliance rate.

A still unresolved question is the optimal 25OHD₃ target level for CKD and dialysis patients. The recent report from the Institute of Medicine states that a 25OHD₃ serum level of 16 ng/mL covers the requirements of half the general population and a level of 20 ng/mL of 97.5 % of the population [35]. According to the umbrella review by Theodoratou et al. a lowering of PTH is the only indication for vitamin D substitution in CKD and hemodialysis patients. These data, however, are based on studies using mainly active vitamin D compounds [4, 36]. There were no other health-related effects of vitamin D therapy in this population. In agreement with these findings, a recent small study in hemodialysis patients also showed no positive effects of cholecalciferol substitution on muscle strength, functional capacity, quality of life and pulse wave velocity [37]. Active vitamin D therapy lowers PTH by 19.7 pg/mL (95 % CI 34.3-5.2) in CKD patients not on dialysis and by 78.4 pg/mL (119.4-37.1) in dialysis patients [3]. Since PTH levels increase, at least in the general population, as serum 25OHD₃ drops below 30 ng/mL, we suggest that this would be the appropriate target level for CKD and dialysis patients. A very recent publication, however, suggests that this level might be too low for CKD stage 3 to 5 patients, and that a maximum effect on PTH requires levels between 42 to 48 ng/mL [38]. The central question is whether, as shown for active vitamin D compound, native vitamin D is also able to lower PTH levels. We observed a significant decline in PTH levels in our patients. A significant decline in PTH has not been uniformly found in studies of vitamin D substitution in CKD patients. In general, patients in studies without a significant PTH response had iPTH levels below 200 pg/mL [9, 39, 40], whereas in studies of a significant PTH decline reported basal levels were above 200 pg/mL [10, 19, 41‐43]. This hypothesis is strengthened by our observation that baseline PTH is the best predictor of a PTH decline under cholecalciferol substitution. Based on our results and the published literature it can be concluded that native vitamin D substitution in patients with low 25OHD₃ levels is able to cause some PTH reduction, albeit not to a degree as seen with active vitamin D compounds. However, the risk of developing hypercalcemia or hyperphosphatemia is lower when using native vitamin D as shown in our study. Recently Armas et al. found an unchanged calcium absorption despite a high cholecalciferol supplementation of 20,000 IU weekly for 12- to 13 weeks [44]. Considering the fact that at present PTH lowering is the only goal for cholecalciferol substitution in dialysis patients, it remains doubtful whether patients with low PTH levels are candidates for native vitamin D substitution at all.

One argument against native vitamin D substitution in dialysis patients has been that most of them already receive activated vitamin D compounds and therefore would not profit from additional vitamin D substitution. Our results, however, suggest that cholecalciferol substitution further decreases PTH in patients already given low doses of calcitriol. Whether this effect is also present in patients receiving calcitriol doses above 0.25 μg per day needs further investigation.

Our study has some limitations and strengths. First, the number of participants is rather small, but still included more than twice the number of dialysis patients with cholecalciferol supplementation compared to recently published randomized controlled trials [9, 37, 40]. Second, the study lacks a control group without cholecalciferol substitution, although the retrospective analysis of the 25OHD₃ course in the year prior to the cholecalciferol substitution found no change in 25OHD₃ without substitution. We also did not measure 1,25OHD₃ levels, which would have allowed us to determine whether these levels were associated with outcomes, especially iPTH levels. In addition, we also did not measure FGF-23. One major strength of this study is that treatment with phosphate binders, active vitamin D compounds and cinacalcet as well as dialysate calcium was kept strictly constant during the whole treatment period. For this reason, we are sure that the observed iPTH changes were only a consequence of cholecalciferol substitution.