Impact of Vitamin D Supplementation on Bone Mineral Density and All-Cause Mortality in Heart Transplant Patients

Vitamin D (VD) deficiency is frequently reported in heart transplant (HT) recipients and routinely supplemented. However, the efficacy of VD supplementation on bone mineral density (BMD) and its association with all-cause mortality is underinvestigated. The VD levels and BMD were studied for two years, and the association of VD and BMD with all-cause mortality risk was investigated. Ninety-six HT patients (38.18 ± 12.10 years old; 74% men) were followed up during VD, Ca, and Mg supplementation. Anthropometric measurements, BMD by Dual-energy X-ray absorptiometry (DEXA) scan, VD concentrations, and related biochemical parameters were analyzed before, 1 year, and 2 years after HT. Despite significant improvement of VD3 and 25-hydroxy VD (25OHVD) levels especially in the men, BMD parameters were insignificantly changed. After 2 years, the all-cause mortality rate was 15.6%. High pretransplant levels of 25OHVD failed to improve the survival probability. Cox’s regression showed a 32.7% increased hazard ratio for each unit increase in body mass index (95% CI: 1.015–1.733, p = 0.038), in the VD-deficient group rather than in the VD-sufficient one. In conclusion, VD supplementation improves the biochemical status, especially in VD-deficient HT. However, its impact on the BMD and mortality was not as usually expected. Further investigation of the disturbed VD metabolism in HT is warranted.


Introduction
Vitamin D deficiency (VD-D) is highly prevalent among patients with end-stage organ failure especially heart failure. Further, VD-D is frequently reported in heart transplant (HT) patients [1,2]. Patients eligible for HT are hardly exposed to sunlight due to frequent hospitalization and defective hepatic metabolism for VD due to heart failure-associated hepatic congestion. The VD status is frequently represented by the level of 25 hydroxyvitamin D (25OHVD). The International Osteoporosis Foundation (IOF) reported the definition of the VD-D by having a serum level of 25OHVD less than 25 nmol/L [3]. However, many cutoff points for defining VD-D were also reported, e.g., serum level of 25OHVD < 30 nmol/L [4], and 25OHVD < 50 nmol/L [5]. The 25 nmo/L was suggested as an arbitrary cutoff value especially with a lack of sufficient evidence for the optimal value for non-skeletal effects of VD [6,7], especially in populations with endemic VD-D such as Saudi Arabia. In Saudi patients with VD insufficiency (25OHVD < 50 nmol/L), while those with proven VD-D (by 25OHVD < 25 nmol/L) were given 50,000 IU as a weekly oral dose for 3 months followed by a 10,000 IU maintenance dose. Calcium (Ca) in an oral dose of 1200 mg/day, and magnesium (Mg) 200 mg/day per oral were also routinely given to all patients. Commitment percentages of study participants on other supplements and medications were shown in Figure 1.

Anthropometric Parameters
Bodyweight (kg) and height (cm) were used for calculating body mass index (BMI) using the formula; BMI = weight (kg)/height (m) 2 . Weight was measured to the nearest 0.1 kg by (Scale-Tronix scale, Chicago, IL, USA) while a stadiometer (Seca Co, Hamburg, Germany) was used for height measurement.

Measurement of BMD
The BMD was measured at scheduled appointments for follow-up at pre-transplant, 1 year, and 2 years post-transplant phases and the reports were used for analysis. BMD was assessed by the DEXA scan using a GE medical system Lunar iDEXA (GE Healthcare, Madison, WI, USA). Two sites were selected: lumbar spine (LS) and femoral neck (FN).

Anthropometric Parameters
Bodyweight (kg) and height (cm) were used for calculating body mass index (BMI) using the formula; BMI = weight (kg)/height (m) 2 . Weight was measured to the nearest 0.1 kg by (Scale-Tronix scale, Chicago, IL, USA) while a stadiometer (Seca Co, Hamburg, Germany) was used for height measurement.

Measurement of BMD
The BMD was measured at scheduled appointments for follow-up at pre-transplant, 1 year, and 2 years post-transplant phases and the reports were used for analysis. BMD was assessed by the DEXA scan using a GE medical system Lunar iDEXA (GE Healthcare, Madison, WI, USA). Two sites were selected: lumbar spine (LS) and femoral neck (FN). Due to the relatively young age of our samples, Z-scores (rather than T-scores) were calculated as standard deviations from the mean of the gender-and age-matched controls. Tertials of the BMD results were created as follows: (a) normal BMD: Z-scores above −1; (b) osteopenia: Z-scores between −1 and −2.5 gm/cm 2 ; and (c) osteoporosis: Z-score below −2.5 gm/cm 2 [23].

Sample Size and Satistical Power
The sample size and statistical power were calculated by G*Power software 3.1.9.4 (University of Kiel, Kiel, Germany), considering medium effect size (f = 0.25), alpha error probability at 0.05, power (1-β error probability) = 0.95, number of repeated measurements = 3, and the number of groups = 2. The estimated total sample size was 44 participants (22/group) and the actual power was 0.9557.

Statistical Analysis
The SPSS tool version 25 (SPSS, IBM, Chicago, IL, USA) was used for processing and analyzing these data. Continuous data were expressed as means ± SD, while dichotomous variables were expressed as percentages and categories. The normal distribution of continuous variables was tested by the Shapiro-Wilk test. For comparison of the three related samples, the Friedman ANOVA test was used with pairwise comparisons. Gender differences and comparisons between groups I and II were done via the Mann-Whitney U test. For survival analysis, Cox's proportional hazard regression analysis was used with 95% CIs. Results were considered statistically significant at p ≤ 0.05. Figure 2 shows participants' flow throughout the study. As shown in Table 1, pretransplant data of our sample showed younger age, significantly lower BMI, and BMD in the women's group. Levels of VD 3 , 25OHVD, PTH, Ca, ALP, Urea, Na, and K were insignificantly different. However, Mg and Cl levels were significantly higher, while creatinine was lower in the women group. Figure 3 shows the cardiovascular events which were diagnosed in our participants. The majority had dilated cardiomyopathy (52.22%), followed by ischemic cardiomyopathy (32.22%), and chemo-induced cardiomyopathy and post-partum cardiomyopathy (1.11%). Due to the relatively young age of our samples, Z-scores (rather than T-scores) were calculated as standard deviations from the mean of the gender-and age-matched controls. Tertials of the BMD results were created as follows: (a) normal BMD: Z-scores above −1; (b) osteopenia: Z-scores between −1 and −2.5 gm/cm 2 ; and (c) osteoporosis: Z-score below −2.5 gm/cm 2 [23].

Sample Size and Satistical Power
The sample size and statistical power were calculated by G*Power software 3.1.9.4 (University of Kiel, Kiel, Germany), considering medium effect size (f = 0.25), alpha error probability at 0.05, power (1-β error probability) = 0.95, number of repeated measurements = 3, and the number of groups = 2. The estimated total sample size was 44 participants (22/group) and the actual power was 0.9557.

Statistical Analysis
The SPSS tool version 25 (SPSS, IBM, Chicago, IN, USA) was used for processing and analyzing these data. Continuous data were expressed as means ± SD, while dichotomous variables were expressed as percentages and categories. The normal distribution of continuous variables was tested by the Shapiro-Wilk test. For comparison of the three related samples, the Friedman ANOVA test was used with pairwise comparisons. Gender differences and comparisons between groups I and II were done via the Mann-Whitney U test. For survival analysis, Cox's proportional hazard regression analysis was used with 95% CIs. Results were considered statistically significant at p ≤ 0.05. Figure 2 shows participants' flow throughout the study. As shown in Table 1, pretransplant data of our sample showed younger age, significantly lower BMI, and BMD in the women's group. Levels of VD3, 25OHVD, PTH, Ca, ALP, Urea, Na, and K were insignificantly different. However, Mg and Cl levels were significantly higher, while creatinine was lower in the women group. Figure 3 shows the cardiovascular events which were diagnosed in our participants. The majority had dilated cardiomyopathy (52.22%), followed by ischemic cardiomyopathy (32.22%), and chemo-induced cardiomyopathy and post-partum cardiomyopathy (1.11%).

Changes in VD and BMD throughout the Study Period
After 2 years, 15 participants passed away (33% women), and 6 dropped out (83% women). The data of remaining participants (n = 75; 20% women) are presented in Table 2 and analyzed by the Friedman ANOVA test with pairwise comparisons. Bodyweight and BMI significantly improved, indicating improvement of nutritional status. Femoral BMD in the men's group showed a significant reduction after 1 year. After the second year, it returned to a value like that of the pretransplant status. In women, the three measurements were insignificantly different (p trend > 0.05). In the men's group, levels of 25OHVD and Ca increased progressively throughout the period with a significant reduction of PTH (especially in the first year), and ALP enzyme. However, insignificant changes in their levels were noticed in the women group. The prevalence of VD-D (defined by 25OHVD < 25 nmol/L) is presented in Figure 4. A progressive reduction of the VD_D was noticed, especially in the men's group, indicating sufficient VD supplementation. Despite supplementation, Mg serum level showed progressive reduction (p < 0.05) in men and women (p = 0.74). Percentages of study participants with osteopenia and osteoporosis at the lumbar spine and femoral neck throughout the study period are presented in Figure 5. Longitudinal changes in the 25OHVD, and PTH, as well as BMD at the lumbar spine and femoral neck, are shown in Figure 6. Table 2. Bone mineral density and related biochemical parameters before, 1 year, and 2 years after the heart transplant. Values with different superscripts ( a and b ) mean significant vs. pretransplant phase; * Significant versus the related samples in Group II by Mann-Whitney U test; BMI is body mass index; DEXA is dual-energy X-ray absorptiometry; BMD is bone mineral density; 25OHVD is 25 hydroxyvitamin D; ALP is alkaline phosphatase enzyme; Mg is magnesium.

Survival Analysis Based on VD and BMD
Notably, in Group I with VD-D, Cox's regression analysis showed that for each additional unit of BMI, the hazard increases by about 33% (Table 4). This was not the case in Group II. Besides, in both groups, for each additional unit of 25OHVD or VD3, the hazard ratio (HR) showed insignificant changes (Table 4). Furthermore, the age and BMD parameters had insignificant impacts.

Survival Analysis Based on VD and BMD
Notably, in Group I with VD-D, Cox's regression analysis showed that for each additional unit of BMI, the hazard increases by about 33% (Table 4). This was not the case in Group II. Besides, in both groups, for each additional unit of 25OHVD or VD3, the hazard ratio (HR) showed insignificant changes (Table 4). Furthermore, the age and BMD parameters had insignificant impacts.  Table 3 further reports the changes in study parameters in both study groups. Pretransplant body BMI was insignificantly different between group I and group II. However, at the post-transplant assessment points, the BMI was significantly higher in the VD-sufficient group (group II). Besides, longitudinal changes showed a progressive increase of BMI with time, especially in the VD-deficient group. Group I had significantly lower VD 3 , 25OHVD, and calcium levels. In Group I (VD-D group), significant reductions in BMD parameters were detected after the first year which improved at the second year to be insignificantly different from the pretransplant levels. In Group II, insignificant changes were reported in the three-time points. VD 3 level progressively increased with time due to the supplementation, while the 25OHVD level significantly increased in group I rather than group II. Similarly, levels of iPTH, ALP, calcium, and magnesium showed significant changes in group I rather than group II.

Survival Analysis Based on VD and BMD
Notably, in Group I with VD-D, Cox's regression analysis showed that for each additional unit of BMI, the hazard increases by about 33% (Table 4). This was not the case in Group II. Besides, in both groups, for each additional unit of 25OHVD or VD 3 , the hazard ratio (HR) showed insignificant changes (Table 4). Furthermore, the age and BMD parameters had insignificant impacts.

Discussion
This study investigated changes in the VD levels and BMD in HT patients for two years and analyzed the association of VD status (by using an arbitrary cutoff value of 25OHVD, i.e., 25 nmol/L), and BMD with all causes-mortality in vitamin D-deficient and -sufficient groups of HT Saudi recipients. Most of our study participants were on VD supplementation at least by a maintenance dose of 10,000 IU/day. This was successful in the reduction of the percentage of the VD deficiency in both men and women's groups (Figure 2). Besides, means of the 25OHVD and VD 3 serum levels were significantly increased progressively in the men's group and Group I, while the rise of their levels in the women's group and group II were insignificant. Indicating that the benefit of VD supplementation is prominent in those with VD-D. Our baseline percentage of the VD-D was much higher than that reported by Stein et al. [24] where severe deficiency (25OHD <25 nmol/L) was found in 16% of heart transplant patients. Moreover, a Slovenian cohort of HT recipients showed 21.3% with severe VD deficiency and 54.7% with mild-to-moderate VD-D. However, these patients were on VD 3 supplementation in a dose of 2000 IU/day and Alfacalcidol of 0.5 µg/day [25]. This frequently reported phenomenon is critical, since the VD-D is linked to post-HT bone loss and fracture possibility, sarcopenia, and may aggravate the immunosuppressive action of corticosteroids or calcineurin inhibitors [26]. Besides, it is associated with periodontal disease and gingival inflammation in HT recipients which impair the nutritional intake [27]. The improvement of VD status after HT especially in the VD-and Ca-supplemented men was in line with a previous report by Gilfraguas et al. [28]. Supplementation in addition to relief of hepatic congestion and improvement of general condition with more mobility and sunlight exposure were the causes of VD status improvement [29]. Unfortunately, this was not the case in the women's group especially in Saudi Arabia where indoor lifestyle and extensive body covering are the traditions.
Bone metabolism and VD status are closely related. Low 25OHVD levels can negatively impact bone turnover biochemical markers. The PTH as an indicator of bone resorption is usually investigated. Baseline measurements of PTH in our sample indicated higher serum levels of iPTH and low normal Ca levels together with low 25OHVD (Tables  2 and 3). This secondary hyperparathyroidism was improved in the men's group and the deficient groups after the first year then relapsed later, while in Group II and women, insignificant changes were detected in the first year and a significant increase in iPTH (with insignificant reduction of Ca level) were detected by the second year. This finding was consistent with previous reports [27,29]. The increase in the 25OHVD level leads to normalization of serum calcium and phosphate levels, nevertheless, serum iPTH level remained high, especially in the women's group and in the second year in the men's group, indicating a status of persistent hyperparathyroidism in the HT recipients. This persistent secondary hyperparathyroidism occurred in both deficient and sufficient groups in the second year. This finding was consistent with previous reports about HT [30,31], renal transplant adults [32], and up to 50% of children's kidney transplants [33]. Persistent secondary hyperparathyroidism may then lead to autonomous hyperplasia of parathyroid glands. Besides, perioperative administration of large amounts of citrate during blood transfusion leading to precipitation of calcium resulting in hypocalcemia-induced hyperparathyroidism. This persistent hyperparathyroidism or tertiary hyperparathyroidism is usually reported after successful renal transplantation. PTH levels usually decline significantly within the first 3-6 months after kidney transplantation due to the reduction of the functional mass of parathyroid glands [34]. Persistent hyperparathyroidism despite normalization of renal functions, and overall survival was reported in 25% of kidney transplant recipients 1-year after the procedure. Medical management and even parathyroidectomy may be required in these cases [35][36][37][38][39][40]. Moreover, it may cause serious consequences such as hypercalcemia, organ calcification, hypophosphatemia, and hypercalciuria [34,41].
Despite VD and Ca supplementation, a significant reduction in femoral neck BMD after 1 year was noticed especially in group I, while all remaining measurements were insignificantly different from the pretransplant status, especially in Group II and women.
These findings were consistent with previous reports [13,42] about both lung and heart recipients. Compared to pretransplant status, Caffarelli et al. [42] found an increase in the incidence rate of vertebral fractures in the first period post-transplantation (9.6% vs. 25.7%). These vertebral fractures were predicted only by the history of any fracture, while in lung transplant recipients, vertebral fractures were predicted by age, BMD at the femur neck, and history of fracture [43]. The transplantation-associated abnormalities in bone metabolism are generally similar regardless of the transplanted organ, pre-existent low BMD, and previous treatment. Typically, bone loss occurs in the first year after the organ transplant, because of immunosuppressive medications, and the long period of immobilization. Supplementation with VD, Ca, and Mg was not sufficient in the improvement of BMD in the HT population. At least in part, tertiary hyperparathyroidism may be the underlying mechanism. In another hand, Calcitriol (1,25(OH)2VD) supplementation in a dose of 0.5-0.75 µg/day for 12 or 24 months in addition to calcium 600 mg/day in comparison with calcium 600 mg/day produced improvement in the femoral neck (but not at lumbar spine) in the calcitriol groups at 12 months [44]. In another trial, Calcidiol (25OHVD) in a dose of 32,000 IU/week showed a mild improvement of about 4.9% only at the lumbar spine in HT patients [45] In our study ALP, and urea significantly decreased in the male group rather than the women's group (Table 2). In the pretransplant phase, congestive hepatopathy and even liver cirrhosis may be evident resulting in impaired hepatic functions such as protein and lipid biosynthesis and decreased ability for detoxification of toxic metabolites. Besides, secretions of hepatic enzymes show an abnormal pattern such as rises in alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) [46]. Post-transplant relief of hepatic congestion greatly correct this abnormal pattern. Moreover, Przybyłowski et al. [10] reported that vitamin D was correlated with kidney functions in heart transplant patients, i.e., improvement of VD status was associated with improvement of renal functions.
Interestingly, Cox's regression analysis showed that in Group I, for each additional unite of 25OHVD, the hazard for all-cause mortality decreases by 7% (HR = 0.930; 95%CI: 0.681-1.270, p = 0.648), while in Group II, for each additional unite of 25OHVD, the hazard decreases by 65.5% (HR = 0.345; 95%CI: 0.650-1.163, p = 0.345). However, these findings were statistically insignificant ( Table 4). The current study finding was in line with Zittermann et al. [18] who reported no effect of 4000 IU/day oral vitamin D supplementation in reduction of mortality in patients with advanced heart failure. After 3 years there was no beneficial latency impact of VD supplementation on all-cause mortality in the same study participants [47]. In a meta-analysis of randomized clinical trials, with >83,000 participants, VD supplementation failed in reducing the risks of major adverse cardiovascular events, stroke, myocardial infarction, cardiovascular disease mortality, or all-cause mortality [48]. In children undergoing hematopoietic stem cell transplant, there was no significant difference in overall survival for those with pretransplant VD deficiency, or sufficiency or optimal level (p = 0.51) [49]. In a renal transplant study, Cox regression analysis showed no significant prediction between 3-month 25OHVD or 3-month 1,25(OH) 2 VD levels and mortality (HR = 0.97; 95% CI: 0.93-1.02, p = 0.27 for 1 25OHVD unit increase, and for 10 units increase it was 0.86; 95% CI: 0.70-1.06, p = 0.16) [50]. On the other hand, Zittermann et al. [17] found that low postoperative levels of 1,25(OH) 2 VD were associated with high 1-year mortality in HT recipients. Furthermore, in a large cohort of kidney transplant recipients, survival was better in recipients with sufficient vitamin D which was measured 10 weeks post-transplant [51]. However, we used a different indicator of the VD status (i.e., 25OHVD), and we used the pretransplant level. Another report from patients with chronic heart failure showed about a 14% reduction of all-cause mortality with a 2.7-fold increment in the 25OHVD Level (95% CI: 1-26%; p = 0.04) [52]. Pediatric reports also stated that VD-D was associated with lower survival on a short-term basis after hematopoietic stem cell transplantation in children [53].
This study's findings indicate a significant increase in the HR of all-cause mortality in the VD-deficient group rather than the VD-sufficient group (HR = 1.327; 95% CI: 1.015-1.733, p = 0.038). Independent of the VD status, a systematic review showed that pretransplant BMI was associated with increased risk of mortality in those with BMI above 30 Kg/m 2 (10% increase in HR) and those above 35 kg/m 2 (by about 24%) [54]. Moreover, Doumouras et al. [55] added the low BMI to the obesity as independent factors for increased mortality in HT recipients. While Nagendran et al. [56] excluded BMI up to 35 kg/m 2 from the factors that worse mortality risk. Our sample's pretransplant BMI was at the range of normal BMI in Group I, and the overweightedness in Group II. In the general population, obesity may affect the association between VD and cardiac disorders. However, VD supplementation failed to reduce the incidence rate of cardiovascular diseases and mortality [57]. In the HT population, the current study tested the BMI and VD levels in the same Cox's regression model and resulted in a significant effect of BMI with an insignificant effect of the VD levels. This indicates that the mortality-increasing effect of the BMI is independent of VD.

Conclusions
This study tracked the changes of the VD3, 25OHVD, and BMD in VD-supplemented heart transplant recipients for 2 years. Further, it investigated the association of 25OHVD, and BMD with all-cause mortality, based on an arbitrary cutoff value of 25OHVD equal 25 nmo/L. Supplementation with VD 3 10,000 IU daily dose, Ca 1200 mg/day, and Mg 200 mg/day were being effective in elevating the serum level of 25OHVD especially in vitamin D deficient HT recipients; however, no significant impact was detected on the preservation of the BMD at measured sites, or on correction of tertiary hyperparathyroidism. Interestingly, the 25OHVD failed to ameliorate the all-cause mortality hazard ratio.

Limitations
The main limitation of this study was the lack of a placebo-controlled group. Instead, we used the comparison of VD deficient (Group I) vs. VD sufficient (Group II). The limited number in the women's group is also a considerable limitation that may affect the obtained results. However, the number of female candidates is usually less than males in many centers all over the world. Another important limitation is the missing of about 21.8% of our participants at the end of the study; either by death (15.6%) or by no-show (6.25%). Missing data are usually common in longitudinal studies. Besides, we did not measure the 1,25(OH)2VD levels and considered the commonly used indicator for VD status which is 25OHVD.

Data Availability Statement:
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.