Vitamin D Status and Outcomes After Renal Transplantation

Abstract

Kidney transplant recipients usually have low vitamin D levels, especially in the early posttransplantation period, but the association between vitamin D status with renal outcomes is not well described in this population. Here, we studied a prospective cohort of 634 kidney recipients who underwent transplantation at a single institution between January 2005 and June 2010. In this cohort, low 25-hydroxyvitamin D concentrations 3 months after transplantation did not predict early death or graft loss but were independently associated with lower measured GFR at 12 months (P=0.001) and higher risk for interstitial fibrosis and tubular atrophy (P=0.01). In contrast, levels of calcium, phosphorus, calcitriol, parathyroid hormone, or fibroblast growth factor-23 were not consistently associated with any of the studied outcomes. In conclusion, low 25-hydroxyvitamin D concentration measured 3 months after transplantation is an independent risk factor for interstitial fibrosis progression and is associated with a lower GFR 1 year after transplantation.

Kidney transplantation is the preferred treatment for the increasing number of patients with ESRD. Transplantation allows patients to avoid the heavy constraints associated with dialysis therapy and is associated with a lower mortality and morbidity than is hemodialysis.¹ Evolution of GFR in transplant recipients appears to be imprinted at an early stage after the surgery. Indeed, the estimated GFR (eGFR) reached 1 year after transplantation is critically associated with allograft outcome.^2,3 This observation prompts research on modifiable factors affecting GFR in the first year after transplantation.

Vitamin D is a critical hormone controlling mineral homeostasis. It promotes phosphate and calcium absorption by the gut and increases calcium reabsorption by the renal distal tubule, thereby providing the positive calcium and phosphorus flux required for bone mineralization. The formation of fully active vitamin D requires two-step hydroxylation of its precursors, either cholecalciferol or ergocalciferol. Hepatocytes mediate the first hydroxylation on carbon 25 to produce 25-hydroxyvitamin D (25[OH]D),⁴ which binds the vitamin D receptor (VDR) with only a modest affinity. The complete activation of vitamin D requires further hydroxylation on carbon 1 by the enzyme CYP27B1, resulting in the formation of calcitriol or 1,25 dihydroxyvitamin D (1,25[OH]D). This last step takes place mainly in the proximal tubular cells of the kidney and is tightly regulated.⁵ Parathyroid hormone (PTH) and hypophosphatemia increase CYP27B1 expression in the proximal tubular cells, whereas the phosphatonin fibroblast growth factor-23 (FGF-23) decreases it.⁶

Recent research has expanded the spectrum of vitamin D actions beyond mineral metabolism. Concordant evidence sustains a wide range of unconventional vitamin D action on the immune, cardiovascular, and renal systems. Of note, vitamin D signaling protects against progression of chronic renal lesions in different models of CKD.^7–15 More recently, calcitriol and paricalcitol, two agonists of VDR signaling, have been shown to lessen albuminuria in patients with IgA nephropathy¹⁶ and diabetic nephropathy,¹⁷ respectively. In addition, two community-based studies and a small observational study in patients with CKD who have not had transplantation found an association between low 25(OH)D levels and the risk of ESRD¹⁸ or eGFR decline.¹⁹ Transplant recipients frequently have low 25(OH)D and 1,25(OH)D levels, especially early after transplantation. Whether these modifiable variables influence kidney function or transplant outcome is unknown.

Since 2005, the standard care of kidney transplant recipients at our institution includes a measurement of GFR by iohexol clearance, an assessment of mineral metabolism hormones, and screening kidney transplant biopsies performed 3 and 12 months after transplantation (Supplemental Figure 1). We took advantage of this close monitoring to assess the link between vitamin D status at 3 months and transplantation outcome. We approached this question with three complementary analyses. First, we explored the association of 3-month vitamin D status with early mortality and allograft loss. Second, we evaluated the link between vitamin D status at 3 months and measured GFR (mGFR) reached 12 months after transplantation. Finally, we analyzed the association between vitamin D status 3 months after transplantation and interstitial fibrosis and tubular atrophy progression between 3 and 12 months.

Results

Study Population

During a 5-year period, 634 patients had renal transplantation and were assessed 2–4 months after transplantation. Table 1 shows the patients’ characteristics. Overall, 25(OH)D concentrations were low (median concentration, 13 ng/ml; interquartile range, 9–20 mg/nl). The 25(OH)D level was <30 ng/ml in 91.7% of patients (Figure 1A). In contrast, 1,25(OH)D concentrations were within the normal range in most patients (Figure 1B). At 3 months, 25(OH)D correlated significantly with 1,25(OH)D (r=0.29; P<0.001) and PTH (r=−0.28; P=0.001) levels but not with FGF-23, ionized blood calcium, phosphate, or mGFR. The characteristics of the patients with or without vitamin D deficiency (defined as 25[OH]D < 15 ng/ml) 3 months after transplantation are presented in Table 1. It is worth noting that longer time on hemodialysis was associated with 25(OH)D deficiency. In addition, 3 months after transplantation, 1,25(OH)D correlated significantly with PTH (r=0.18; P=0.002), mGFR (r=0.33; P=0.001), and FGF-23 (r=−0.17; P=0.002); correlated weakly with serum phosphate concentration (r=−0.09; P=0.02); and was not correlated with ionized blood calcium concentration.

Table 1.

Patient characteristics 3 months after transplantation

Characteristic	All Patients (n=634)	3-Month 25(OH)D Level < 15 ng/ml (n=348)	3-Month 25(OH)D Level ≥ 15 ng/ml (n=286)	P Valuea
Age at transplantation (yr)	48.3±13.4	48±14	50±13	0.41
Male sex	58.7 (372)	58.0 (202)	59.4 (170)	0.72
Initial nephropathy				0.16
Primary GN	23.2 (147)	24.1 (84)	22.0 (63)
Cystic nephropathy	15.3 (97)	12.9 (45)	18.2 (52)
Uropathy	8.2 (52)	6.9 (24)	9.8 (28)
Diabetes	7.6 (48)	9.2 (32)	5.6 (16)
Vascular nephropathy	6.2 (39)	6.3 (22)	5.9 (17)
Interstitial nephropathy	3.6 (23)	4.6 (16)	2.4 (7)
Other	15.9 (101)	17.0 (59)	14.7 (42)
Unknown	20.0 (127)	19.0 (66)	21.3 (61)
Preemptive transplantation	13.7 (86)	10.7 (37)	17.3 (49)	0.02
Treated by hemodialysis before transplantation	81.2 (510)	84.6 (292)	77.0 (218)	0.02
Treated by peritoneal dialysis before transplantation	3.7 (23)	3.2 (11)	4.2 (12)	0.49
Time on dialysis before transplantation (mo)	35.4 (12.0–73.2)	38.8 (15.2–76.3)	30.0 (7.95–69.4)	0.01
Parathyroidectomy before transplantation	12.6 (80)	12.1 (42)	13.3 (38)	0.65
Repeat transplantation (for first transplantations)	77.7 (491)	77.2 (267)	78.3 (224)	0.73
Living donor	20.8 (131)	17.7 (61)	24.6 (70)	0.04
Donor age (yr)	53±16	53±16	53±16	0.87
Extended donor criteria	51.4 (314)	52.3 (176)	50.2 (138)	0.59
Cold ischemia time (min)	1122 (753–1492)	1151 (851–1476)	1090 (492–1500)	0.15
Delayed graft function	32.3 (205)	35.3 (123)	28.7 (82)	0.07
Induction treatment	95.6 (593)	96.2 (328)	95.0 (265)	0.46
Type of induction treatment				0.78
Basiliximab	69.6 (413)	68.6 (225)	70.9 (188)
Antilymphocytic serum	29.0 (172)	29.9 (98)	27.9 (74)
Others	1.3 (8)	1.52 (5)	1.1 (3)
Calcineurin inhibitor–based regimen	98.5 (605)	98.8 (333)	98.2 (272)	0.65
Cellular acute rejection before 3 mo	8.2 (52)	8.62 (30)	7.69 (22)	0.67
Humoral acute rejection before 3 mo	4.6 (29)	4.02 (14)	5.24 (15)	0.46
Body mass index (kg/m2)	24.1±4.3	24.2±4.53	24.0±4.09	0.55
Systolic BP (mmHg)	136.2±17.9	136.7±17.8	135.7±18.0	0.48
Diastolic BP (mmHg)	77.9±10.7	77.8±11.1	78.1±10.1	0.71
mGFR (ml/min)	57±17	57±17	58±17	0.28
Ionized calcium (mmol/L)	1.27±0.09	1.27±0.09	1.27±0.09	0.21
Phosphate (mmol/L)	0.77±0.21	0.77±0.22	0.77±0.20	0.80
1,25(OH)D (pg/L)	35 (23–48)	31 (19–41)	40 (27–53)	P<0.001
PTH (ng/L)	98 (64–159)	111 (78–198)	81 (55–137)	P<0.001
FGF-23 (RU/ml)	70 (47–116)	67 (44–115)	73 (52–119)	0.23
Cinacalcet treatment	3.6 (23)	3.2 (11)	4.2 (12)	0.49

Values are reported as percentage (number of patients), mean ± SD, or median (interquartile range), as appropriate.

^aP value represents tests of significance from t test, Wilcoxon test, chi-squared test, or Fisher exact test, as appropriate.

Figure 1.

Kidney tranplant recipients have low 25(OH)D 3 months after transplantation. Distribution of (A) 25-hydroxy vitamin D (25[OH]D) and (B) 1-25-dihydroxy vitamin D (1,25[OH]D) in kidney recipients 3 months after transplantation. Gray areas indicate target range for 25(OH)D and normal range for 1,25(OH)D in patients with CKD who have not undergone transplantation.

Association of 3-Month 25(OH)D Concentration with Survival and Graft Loss

The median follow-up time was 48.6 months (interquartile range, 3.2–75.6 months). At the end of the study, 19 patients were lost to follow-up, 30 patients had lost their allograft, 28 patients had died with a functioning graft, and 3 patients had died after losing their allograft. The principal causes of death were infections in 12 patients, malignancies in 9 patients, and cardiovascular accidents in 5 patients. Kaplan-Meier survival curves revealed similar risk of death in patients with low or high 3-month 25(OH)D and 3-month 1,25(OH)D (Figure 2). Similarly, Cox regression models showed no significant association between continuous 3-month 25(OH)D or 3-month 1,25(OH)D levels and mortality (hazard ratio [HR], 0.97 [95% confidence interval (CI), 0.93–1.02], P=0.27 for increase of 1 unit of 25[OH]D; HR, 0.86 [95% CI, 0.70–1.06], P=0.16 for increase of 10 units of 1,25[OH]D, respectively). In addition, competitive risk analysis revealed no association between graft loss and 3-month 25(OH)D or 3-month 1,25(OH)D levels (HR, 0.99 [95% CI, 0.95–1.04], P=0.65 for increase of 1 unit of 25[OH]D; HR, 0.98 [95% CI, 0.77–1.25], P=0.86 for increase of 10 units of 1,25[OH]D, respectively).

Figure 2.

Three-month vitamin D levels do not associate with patient survival. Patient survival curves according to (A) 25(OH)D and (B) 1,25(OH)D status, 3 months after transplantation.

Association of 3-Month 25(OH)D with 12-Month mGFR

Among the 634 patients tested 3 months after transplantation, 465 had a second measurement of mGFR at 12 months. The characteristics of the subgroup of patients who were reassessed at 12 months did not differ from those of the whole-study population. Although the overall distribution of mGFR was similar at 3 months and 12 months (Figure 3A), the analysis of individual mGFR variation revealed significant heterogeneity (Figure 3B). Overall, between 3 and 12 months, mGFR worsened (mGFR variation < −0.6 ml/min per month) in 157 (33.7%) patients, improved (mGFR variation > 0.6 ml/min per month) in 132 (28.3%) patients, and remained stable in 176 (37.8%). The results of the univariate analyses of the factors linked to 12-month mGFR are listed in Table 2. Unsurprisingly, 3-month mGFR was a strong predictor of 12-month mGFR (r=0.78; P<0.001). Of note, 12-month mGFR correlated with 3-month 25(OH)D (r=0.13; P=0.006), 3-month 1,25(OH)D (ρ=0.31; P<0.001), and 3-month FGF-23 (ρ=−0.30; P<0.001) (Table 2 and Figure 4). Patients with 3-month 25(OH)D levels <15 ng/ml had a mean 12-month mGFR ± SD of 55±17 ml/min, whereas patients with 3-month 25(OH)D levels >15 ng/ml had a 3-month mGFR of 59±17 ml/min (Figure 4B). Multivariate analyses revealed that patient sex, donor age, 3-month mGFR, 3-month body mass index, and 3-month 25(OH)D (considered as a continuous variable), but not 1,25(OH)D or FGF-23, were independent predictors of 12-month mGFR (Table 3 for 25[OH]D and 1,25[OH]D; data not shown for FGF-23). Consideration of a cutoff value of 15 ng/ml for 3-month 25(OH)D did not modify the significance of this association (Supplemental Table 1). A second measurement of 25(OH)D was also performed at 12 months (Figure 5), along with GFR measurement. Overall, 25(OH)D concentration increased between 3 months and 12 months. However, 25(OH)D concentration remained below the recommended threshold for CKD (30 ng/ml) in 65.7% of patients. In contrast with 3-month 25(OH)D concentration, 12-month 25(OH)D concentration was not correlated with 12-month mGFR (r=0.04; P=0.43).

Figure 3.

Transplanted patients show significant variations in mGFR between 3 and 12 months after transplantation. (A) Distribution of mGFR in kidney recipients 3 months and 12 months after transplantation. (B) Individual variation in mGFR between 3 and 12 months after transplantation.

Table 2.

Association of the baseline variables with mGFR at 12 months in univariate analysis

Characteristic	12-Month mGFR or Correlation Coefficient	P Value
Sex		<0.001
Male	59.6±16.9
Female	53.6±17.1
Age at transplantation	r=−0.26	<0.001
Donor type		<0.001
Deceased donor	55.1±17.3
Living donor	63.6±15.2
Donor age	r=−0.48	<0.001
Preemptive transplantation		0.065
Yes	60.1±16.4
No	56.3±17.3
Cold ischemia time (min)		<0.001
≤250	63.4±15.2
250–1200	57.9±16.9
>1200	52.3±16.8
Delayed graft function		0.07
Yes	54.8±19.2
No	57.9±16.3
Acute cellular rejection before 3 months		0.46
Yes	53.6±13.1
No	60.0±21.7
Acute humoral rejection before 3 months		0.97
Yes	56.8±15.2
No	58.5±16.1
3-month body mass index	r=0.09	0.046
3-month mGFR	r=0.78	<0.001
3-month 25(OH)D	r=0.13	0.006
3-month PTH	ρ=0.07	0.15
3-month FGF-23	ρ=0.30	<0.001
3-month 1,25(OH)D	ρ=0.31	<0.001

Values of 12-month mGFR are expressed in ml/min and reported as mean ± SD.

^aP values represent tests of significance from t test, Wilcoxon test, or Pearson or Spearman correlation tests, as appropriate.

Figure 4.

Low 25(OH)D 3 months after transplantation correlates with lower GFR at 12 months. (A) Correlation between 25(OH)D 3 months after transplantation (3-month 25[OH]D) and measured GFR at 12 months (12-month mGFR) (B) 12-month mGFR according to 3-month 25(OH)D status.

Table 3.

Association of baseline variables with mGFR at 12 months using multiple linear regressions.

Model	Coefficient ± SD	P Value
Full modela plus 3-month 25(OH)D
Intercept	17.1±4.02
Male sex	2.86±1.00	0.004
Donor ageb	−0.22±0.03	<0.001
3-month body mass indexb	0.27±0.12	0.02
3-month mGFRb	0.70±0.03	<0.001
3-month 25(OH)Db	0.17±0.05	0.001
Full modela plus 3-month 1, 25(OH)D
Intercept	19.6±3.95
Male sex	3.34±1.01	0.001
Donor ageb	−0.22±0.04	<0.001
Preemptive transplantation	3.38±1.40	0.02
3-month body mass indexb	0.28±0.12	0.02
3-month mGFRb	0.70±0.03	<0.001

Only the variables independently associated with 12-mo mGFR are presented (final models after backward selection procedure).

^aThe full model includes recipient sex, age at transplantation, donor type, donor age, preemptive transplantation, delayed graft function, 3-month body mass index, and 3-month mGFR.

^bFor an increase of 1 unit of the variable.

Figure 5.

Although 25(OH)D concentrations tend to increase between 3 and 12 months after transplantation, they usually remain below the recommended threshold for CKD patients. Distribution of 25(OH)D 3 and 12 months after transplantation. Gray area indicates target range for 25(OH)D in patients with CKD who have not undergone kidney transplantation.

Vitamin D Status and Interstitial Fibrosis and Tubular Atrophy Score Progression

GFR may be influenced by hemodynamic factors that are not related to the progression of renal lesions. Thus, we sought to determine whether the apparent effect of 3-month 25(OH)D on 12-month mGFR was associated with deterioration of the renal allograft parenchyma. On the complex spectrum of allograft renal lesions, progressive interstitial fibrosis and tubular atrophy (IF/TA) is a major factor associated with GFR decline and graft loss.²⁰ Therefore, we tested the association between 3-month 25(OH)D and IF/TA progression. Among the 464 patients, 306 had usable screening biopsy samples obtained at 3 and 12 months. Among them, 37 samples were not analyzed because severe IF/TA was already present on biopsy at 3 months (interstitial fibrosis score + tubular atrophy score > 4), precluding further progression. The factors associated with IF/TA progression were studied in the remaining 265 patients. The characteristics of this subpopulation are presented in Supplemental Table 2. Between 3 months and 12 months, the composite IF/TA score progressed in 114 patients (43%). Factors associated with IF/TA progression are shown in Table 4. The multivariate analysis showed that donor age (P=0.01) and 3-month 25(OH)D concentration (P=0.01) were independently associated with IF/TA progression (Table 5). The increased risk of IF/TA progression in patients with low 25(OH)D was not associated with an increased incidence of biopsy-proven acute rejection in the same period (Supplemental Table 3).

Table 4.

Association of baseline variables with interstitial fibrosis/tubular atrophy score progression in univariate analysis

Characteristic	Nonprogressors (n=151)	Progressors (n=114)	P Valuea
Male sex	54.3 (82)	58.8 (67)	0.47
Age at transplantation (yr)	45.4±12.5	47.7±13.3	0.16
Living donor	27.3 (41)	18.4 (21)	0.09
Donor age (yr)	48.6±16.1	54.4±14.0	0.002
Preemptive transplantation	20.5(31)	11.4(13)	0.048
Cold ischemia time			0.19
<250 min	26.7 (39)	18.0 (20)
250–1200 min	35.6 (52)	35.1 (39)
>1200 min	37.7 (55)	46.8 (52)
Delayed graft function	20.5 (31)	32.5 (37)	0.03
Acute cellular rejection before 3 months	9.30 (14)	11.4 (13)	0.57
Acute humoral rejection before 3 months	7.90 (12)	4.40 (5)	0.24
3-month body mass index (kg/m2)	23.7±3.92	23.6±3.84	0.83
3-month mGFR (ml/min per 1.73 m2)	61.7±13.5	56.3±14.8	0.002
3-month 25(OH)D (ng/ml)	16.8±10.1	13.2±7.00	0.001
3-month 1,25(OH)D (pg/ml)	41.45±20.8	33.82±16.7	0.002
3-month PTH (pg/ml)	84.0 (59.0–118.5)	102.5 (64.0–180.0)	0.01
3-month FGF-23 (RU/ml)	62.0 (44.0–100.0)	64.5 (44.8–102.8)	0.62

Values are reported as percentage (number of patients), mean ± SD, or median (interquartile range), as appropriate.

^aP value represents tests of significance from t test, Wilcoxon test, or Pearson or Spearman correlation tests as appropriate.

Table 5.

Association of baseline variables with interstitial fibrosis/tubular atrophy score progression between 3 months and 12 months in multivariate analysis using multiple logistic regression

Variable	Odds Ratio (95% Confidence Interval)	P Value
Donor agea	1.26 (1.07–1.50)	0.007
3-month 25(OH)Db	0.96 (0.92–0.98)	0.003

Only the variables independently associated with interstitial fibrosis/tubular atrophy score progression are presented (final model after backward selection procedure). The full model includes donor type, donor age, delayed graft function, preemptive transplantation, 3-month GFR, 3-month PTH, and 3-month 25(OH)D.

^aFor an increase of 10 units of the variable.

^bFor an increase of 1 unit of the variable.

Discussion

In this large prospective observational study of patients who had recently received a kidney transplant, low 25(OH)D levels 3 months after transplantation were independently associated with reduced mGFR 9 months later. Although the variation of mGFR was of low magnitude, this effect was robust to different statistical analyses and was consistent with an adverse evolution of chronic tubular and interstitial lesions in patients with low 3-month 25(OH)D concentrations. In contrast, other hormones implicated in mineral metabolism, including 1,25(OH)D, PTH, and FGF-23, were not independent predictors of 12-month mGFR or IF/TA progression. This study is, to our knowledge, the first linking vitamin D status with functional and histologic renal outcome in kidney transplant recipients. These results extend the findings of recent community-based studies linking 25(OH)D with the risk of ESRD and estimated GFR decline.^18,19 Finally, by showing that mGFR decline and IF/TA progression correlated with 25(OH)D rather than 1,25(OH)D levels, this work raises the question of the therapeutic usefulness of vitamin D supplementation in preserving kidney allograft function.

An important point of this study is the use of a direct measurement of GFR rather than a creatinine-based GFR estimation. Creatinine-based equations to estimate GFR are an invaluable tool for clinician. However, creatinine concentration may be affected by factors other than GFR, such as variation in creatinine tubular secretion or creatinine production (i.e., increase in muscle mass). This latter point may be of crucial importance in assessing the effect of vitamin D on GFR. Concordant evidence suggests that vitamin D affects striated muscle cell biology.²¹ Indeed, invalidation of VDR in mice reduced muscle diameters,²² and consistent studies found an association between low 25(OH)D and reduced muscle strength in elderly people.²¹ In addition, two studies reported an association between low 25(OH)D levels and fatty degeneration of thigh and rotator cuff muscles.^23,24 This potential effect of vitamin D on muscle biology is likely to affect creatinine production. Indeed, an increase in creatininemia due to enhanced creatinine production without variation in GFR has been reported under calcitriol therapy.²⁵ Thus, the effect of vitamin D on creatinine production represents a potential important bias in studies using creatinine-based equations to estimate GFR or the albumin-to-creatinine ratio to evaluate variation in albuminuria. In this study, the direct measurement of GFR avoids this bias and strengthens the association between poor vitamin D status and adverse evolution of GFR.

This study also has limitations. First, because of the low incidence of death and allograft loss, this analysis is underpowered regarding renal survival. Also of note, the relatively short follow-up achieved so far precludes conclusions regarding the effects of 25(OH)D level on GFR variation after the first year of transplantation. Extended patient follow-up is mandatory to answer this important issue. Finally, plasma 25(OH)D concentrations vary depending on sun exposure, dietary load, and supplementation. Therefore, 3-month 25(OH)D may not adequately reflect the overall exposure to 25(OH)D during this 9-month period. However, our results indicating that 3-month 25(OH)D but not 12-month 25(OH)D correlated with kidney function at 12 months point at a determinant role of 25(OH)D exposure in the early posttransplantation period. To determine whether the repletion of 25(OH)D stocks at 12 months affects the later evolution of kidney function requires a longer follow-up. Keeping in mind those limitations, these results strongly suggest that vitamin D deficiency has a deleterious effect on the renal allograft in the early posttransplantation period.

We can speculate on the mechanisms underlining the effect of vitamin D on kidney function and IF/TA progression. In the context of renal transplantation, it is well established that the recipient immune response against the kidney allograft is an important contributor to renal deterioration. Vitamin D modulates both innate and adaptive immune response,²⁶ and few observations have suggested that vitamin D could modulate allogeneic response. Calcitriol administration was shown to reduce class II major histocompatibility complex and co-stimulatory molecule expression in peripheral blood mononuclear cells from kidney transplant recipients.²⁷ Additionally, alloreactive T cells are more frequently observed in vitamin D–deficient patients awaiting kidney transplantation.²⁸ In contrast to those studies, our results indicate that low 25(OH)D concentration is not linked to an increased occurrence of acute rejection. Thus, the apparent effect of vitamin D deficiency on allograft function is unlikely to be the consequence of a stronger alloimmune response.

Aside from modulating immune response, VDR signaling reduces the progression of renal lesions in different nonimmune models of CKD. Active vitamin D analogues ameliorate renal function in a rat model of chronic allograft nephropathy,²⁹ in subtotally nephrectomized rats,^7,30–32 and in other models of CKD.^8,10,33 Although VDR-independent actions of vitamin D analogues have been demonstrated, recent experiments applying unilateral ureteral obstruction on VDR^−/− mice have revealed that VDR signaling played an instrumental role in preventing CKD progression.^13,15 The molecular mechanisms responsible for the positive effect of vitamin D in CKD remain unclear because various consequences of VDR signaling have been reported, including preservation of podocyte function,³⁴ reduction of glomerular inflammation, and reduction of tubular cell proliferation.⁷ VDR signaling also interferes with important pathways of CKD progression, including the renin-angiotensin system (RAS),³⁵ epidermal growth factor receptor,³⁶ and TGF-β signaling.^29,32 Increased expression of renin and RAS activation appear to be major contributors to the increased sensitivity of VDR^−/− mice to unilateral ureteral obstruction.¹³ However, the additive effects of VDR agonists on RAS blockage in both rodents³⁷ and human CKD^16,38 also argue for RAS-independent mechanisms. The exact contribution of the different cell populations targeted by vitamin D remains unclear so far, and further study using lineage-specific VDR invalidation is mandatory to clarify VDR function in CKD progression.

A striking finding in our study is the unique association between 25(OH)D as opposed to 1,25(OH)D and allograft function at 12 months and IF/TA progression from 3 months to 1 year after transplantation. 1,25(OH)D is less commonly measured than is 25(OH)D. Indeed, 1,25(OH)D status has not been assessed in most studies demonstrating a link between vitamin D deficiency and important clinical endpoints, including cardiovascular disease,³⁹ mortality,⁴⁰ certain types of cancer,⁴¹ and, of note, ESRD¹⁸ and GFR decline.¹⁹ Because 25(OH)D and 1,25(OH)D were systematically measured in this study, we could discriminate the relative association of these two measures with renal outcome. Our finding that low concentration of 25(OH)D but not of 1,25(OH)D predicted GFR decline and IF/TA progression in the first year after kidney transplantation raises important questions. Because of its 500-fold lower affinity for VDR, 25(OH)D is usually considered an inactive compound that reflects vitamin D stocks, whereas circulating 1,25(OH)D is thought to represent biologically active vitamin D.⁴² Thus, the predominant association of 25(OH)D with the evolution of kidney function retrieved in our study may appear paradoxical. However, a significant role for 25(OH)D concentration in regulating important cellular processes in the context of normal 1,25(OH)D concentrations is well established for PTH secretion.⁴³ In parathyroid cells, VDR stimulation downregulates PTH gene transcription,⁴⁴ and specific VDR ablation in parathyroid cells in mice results in an increase in PTH level.⁴⁵ Strikingly, in most conditions in humans, and in this study, PTH levels correlate negatively with 25(OH)D but not with 1,25(OH)D concentration,^46–49 suggesting that circulating 25(OH)D is the main determinant of VDR signaling in parathyroid cells. The molecular mechanisms potentially accounting for this effect of 25(OH)D concentration on VDR signaling include direct binding of 25(OH)D to VDR and its local conversion to 1,25(OH)D. Although of low affinity, the direct binding of 25(OH)D to VDR has been shown to result in VDR signaling. Indeed, the administration of high-dose cholecalciferol partially rescued the phenotype of 1α-hydroxylase–deficient mice and patients.⁵⁰ However, supraphysiologic concentrations of 25(OH)D are needed to achieve efficient VDR signaling in this context. Thus, as the majority of the patients in our study had normal or even high calcitriol concentration, it is unlikely that direct binding of VDR by 25(OH)D significantly contributes to VDR signaling in this population. Local conversion of 25(OH)D to calcitriol is allowed by the local expression of the 1α-hydroxylase enzyme in various tissues including parathyroid;^51–53 immune cells;⁵⁴ and, in the kidney, tubular cells and possibly podocytes.⁵⁵ This locally produced calcitriol plays an important role as an intracrine or paracrine activator of the VDR in monocytes,⁵⁶ cancer cell lines,⁵⁷ and parathyroid cells.⁴³ Whether the same mechanism is involved in the effect of vitamin D on CKD progression has not been assessed but represents an interesting explanation of the predominant association of 25(OH)D concentration with 12-month mGFR.

The potential therapeutic effects of nutritional vitamin D supplementation represent an important corollary of our study. The potential interest of cholecalciferol administration was recently documented in patients with ESRD in whom a significant effect on hyperparathyroidism and left ventricular hypertrophy was recorded.⁵⁸ In contrast, a recent randomized controlled trial showed no effect of the active vitamin D compound paricalcitol on cardiac structure or function in patients with CKD.⁵⁹ Whether these discrepancies reflect a stronger effect of cholecalciferol supplementation on cardiac dysfunction in CKD remains to be tested. Of note, cardiomyocytes in rats express CYP27B1 and respond to cholecalciferol in vitro.⁶⁰

Studies from our group and other have already shown that cholecalciferol supplementation is a safe way to increase 25(OH)D concentrations in kidney transplant recipients, but the small number of treated patients included in these studies and their short follow-up do not allow any conclusion on the effect of these treatments on kidney function.^61–63 Our results should encourage randomized controlled trials evaluating the effect of early cholecalciferol supplementation on renal allograft function and provide valuable information on the magnitude of the expected effects of such treatments on renal function and histologic features.

Concise Methods

Study Population

All patients underwent transplantation and were followed in the Department of Renal Transplantation at Necker Hospital, Paris, France. Since September 2005, the standard longitudinal care of kidney recipients includes the measurement of GFR and mineral metabolism variables 3 and 12 months after transplantation. Between June 6, 2005, and June 7, 2010, all patients who were assessed 3 months (from 2 to 4 months) after transplantation were included in this study. Patients in whom the first GFR measurement was performed >4 months after transplantation (generally because of ongoing complications necessitating patient hospitalization) were excluded. The characteristics of the recipient, the donor, the transplantation procedure, and the outcome were prospectively collected and computerized in the Données Informatisées et Validées en Transplantation cross-validated French database (http://www.divat.fr/).⁶⁴ The collected data included information on the recipient (age, sex, primary cause of CKD, previous time spent on dialysis), the donor (age, sex, deceased or living donor, cold ischemia time), the immunosuppressive regimen, and the occurrence of acute rejection or delayed graft function, defined as the need for hemodialysis during the first week after transplantation. In addition, height, body weight, and BP were systematically recorded at the time of GFR measurements.

Biochemistry

GFR was measured by iohexol clearance or inulin clearance when allergy to iohexol was suspected on the basis of patient history. Patients were infused with iohexol (Omnipaque 360) or inulin (Inutest 25%) as previously described.⁶⁵ Blood and urine samples were collected every hour for 5 hours, and iohexol (or inulin), as sodium, potassium, calcium, phosphorus, creatinine, urea, glucose, and uric acid concentrations were measured. Iohexol and inulin concentrations were determined by HPLC and an enzymatic method, respectively, as previously described.⁶¹

25(OH)D and 1,25(OH)D concentrations were measured by radioimmunoassay (DiaSorin, Stillwater, MN) and radioimmunoassay after immunoextraction (IDS Ltd, Boldon, United Kingdom), respectively. Serum PTH concentration was measured with an immunochemiluminescent assay performed on the Elecsys analyzer (Roche Diagnostics, Meylan, France). Total calcium, ionized calcium, phosphate, and creatinine concentrations were measured using standard methods. FGF-23 plasma concentration was measured with an ELISA (Immutopics, Inc., San Clemente, CA) that recognizes the C-terminal part of the peptide.

Analysis of Renal Graft Biopsy Specimens

In our department, the standard care of kidney recipients includes renal graft screening biopsies that are routinely performed at months 3 and 12 after transplantation, with results scored according to the Banff classification.⁶⁶ Briefly, cortical interstitial fibrosis and tubular atrophy areas were evaluated semiquantitatively, and the corresponding IF and TA scores were determined according to Banff classification (0, <5% of the cortical area; 1, 6%–25%; 2, 26%–50%; and 3, >50%). Progression of IF/TA between 3 and 12 months was defined as an increase in score of at least 1 point. Inadequate biopsy samples (with fewer than seven glomeruli and one artery) were excluded from analysis.

Statistical Analyses

Results are expressed as numbers and percentages for categorical variables and as mean ± SD or median (range) for continuous variables. Correlations between biologic measures were assessed using Pearson (r) or Spearman (ρ) correlation coefficients.

Survival curves were computed by the Kaplan-Meier method and compared by the log-rank test to select prognostic factors of overall survival. Overall survival was defined as time from transplantation to death or date of last follow-up. Cox regression models were used for multivariate analyses. Allograft loss was analyzed with a Gray model that considered death as a competing event.

Effect of biologic and clinical variables on 12-month mGFR was assessed using Pearson or Spearman correlation coefficients for quantitative variables and t test (or Wilcoxon test) for categorical variables. Independent predictive factors of 12-month mGFR were assessed using multivariate linear regression model.

Finally, effect of biologic and clinical variables on interstitial fibrosis and tubular atrophy progression was assessed using a chi-squared test for qualitative variables and a t test for quantitative variables. Independent predictive factors of progression were assessed using a multivariate logistic regression model.

For all multivariate regression models, P<0.10 was used for the variable entry criteria. Then, a backward selection procedure was applied. Only the factors associated with the considered outcome with P<0.05 were kept in the final model. A forward selection procedure was used to control the robustness of the final model.

Statistical analysis was performed using the R package, version 2.10.

Disclosures

None.