Vitamin D deficiency predicts decline in kidney allograft function: a prospective cohort study.

ORIGINAL E n d o c r i n e

ARTICLE R e s e a r c h

Vitamin D Deficiency Predicts Decline in Kidney Allograft Function: A Prospective Cohort Study Yoshitsugu Obi,* Takayuki Hamano,* Naotsugu Ichimaru, Kodo Tomida, Isao Matsui, Naohiko Fujii, Masayoshi Okumi, Jun-ya Kaimori, Koji Yazawa, Yukito Kokado, Norio Nonomura, Hiromi Rakugi, Shiro Takahara, Yoshitaka Isaka, and Yoshiharu Tsubakihara Departments of Geriatric Medicine and Nephrology (Y.O., I.M., H.R., Y.I.), Comprehensive Kidney Disease Research (T.H., Y.T.), Advanced Technology for Transplantation (N.I., J.K., S.T.), and Specific Organ Regulation (Urology) (M.O., K.Y., N.N.), Osaka University Graduate School of Medicine, Suita 565– 0871, Osaka, Japan; Department of Kidney Disease and Hypertension (K.T.), Osaka General Medical Center, Osaka 558 – 0056, Osaka, Japan; Department of Internal Medicine (N.F.), Hyogo Prefectural Nishinomiya Hospital, Nishinomiya 662– 0918, Hyogo, Japan; and Takahashi Clinic (Y.K.), Toyonaka 570 – 0027, Osaka, Japan

Context: Vitamin D, often deficient in kidney transplant (KTx) recipients, has potential immunomodulatory effects. Objective: This study aimed to evaluate whether vitamin D status affects the rate of decline in kidney allograft function. Design, Setting, and Patients: The study included a prospective cohort of 264 ambulatory KTx recipients at a single Japanese center. Main Outcome Measures: We measured the baseline 25-hydroxyvitamin D (25D) concentration and examined its association with annual decline in estimated glomerular filtration rate (eGFR). Secondary outcome was rescue treatment with iv methylprednisolone (IV-MP) as an index of rejection episodes. Results: The mean serum 25D concentration was 17.1 (SD 6.5) ng/mL, and 68.4% patients had vitamin D inadequacy or deficiency. Time after KTx was a significant effect modifier for the association of serum 25D concentration with annual eGFR change and need for IV-MP (P for interaction ⬍ .1). We divided patients according to the median time after KTx (10 y) and found that low vitamin D was significantly associated with a rapid eGFR decline at less than 10 years after KTx but not at 10 or more years after KTx. The same was true for rescue treatment with IV-MP. Overall, propensity score matching showed independent associations of low vitamin D with both outcomes. Stratified matching confirmed pronounced associations at less than 10 years after KTx. Conclusions: Vitamin D deficiency predicts a rapid decline in eGFR and need for IV-MP at less than 10 years after KTx. Future studies are warranted to evaluate the clinical efficacy of vitamin D supplementation. (J Clin Endocrinol Metab 99: 527–535, 2014)

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2014 by the Endocrine Society Received June 4, 2013. Accepted November 18, 2013. First Published Online November 27, 2013

doi: 10.1210/jc.2013-2421

* Y.O. and T.H. contributed equally to this work. Abbreviations: ABMR, antibody-mediated rejection; BMI, body mass index; CKD, chronic kidney disease; CRP, C-reactive protein; 25D, 25-hydroxyvitamin D; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; FGF23, fibroblast growth factor 23; HLA, human leukocyte antigen; IQR, interquartile range; IV-MP, iv methylprednisolone; KTR, kidney transplant recipient; KTx, kidney transplantation; PS, propensity score; TCMR, T cell-mediated rejection.

J Clin Endocrinol Metab, February 2014, 99(2):527–535

jcem.endojournals.org

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 May 2014. at 13:57 For personal use only. No other uses without permission. . All rights reserved.

527

528

Obi et al

Vitamin D and Kidney Allograft Function

llograft failure remains a major concern in kidney transplant recipients (KTRs) despite the recent development of advanced immunosuppressants. Under the current immunosuppressive regimen, antibody-mediated rejection (ABMR) and glomerular diseases are the major causes of allograft failure (1), and for both causes, rescue treatment with immunosuppressive drugs is ineffective (2). Although T cell-mediated rejection (TCMR) has a lower impact on graft failure than ABMR if detected early and treated appropriately, it also poses significant risks (3). Besides promoting bone and muscle health, vitamin D has immunomodulatory effects, which might protect kidney allografts (4). 1,25-Dihydroxyvitamin D inhibits the proliferation and modulates the function of activated T and B cells, and it possibly promotes tolerance by inducing regulatory T cells. Animal transplantation studies have shown the additive effects of active vitamin D compounds on kidney allograft survival when prescribed with immunosuppressants (5, 6). Furthermore, vitamin D deficiency [low serum 25-hydroxyvitamin D (25D) concentrations] has been associated with several autoimmune diseases (7), and cholecalciferol (vitamin D3) supplementation ameliorates disease activities (8, 9). These findings are evidenced by the presence of 1␣-hydroxylase and vitamin D receptors in immune cell types, including antigen-presenting cells and activated T and B cells, wherein 25D is locally converted to 1,25-dihydroxyvitamin D and exerts its immunomodulatory effects. Only a few studies have evaluated the impact of vitamin D deficiency on allograft kidneys in human KTRs (10), although low vitamin D in liver and lung transplant recipients are reported to be independently associated with TCMR (11, 12). Additionally, serum 25D measurement and nutritional vitamin D supplementation are not reimbursed by national health insurance in Japan, and vitamin D deficiency has been overlooked by the general population and medical community. The dietary intake of vitamin D in the Japanese general population is substantially lower than recommended (13), and the prevalence of vitamin supplement use is only 3.4% (14). These findings prompted us to investigate KTRs for the association of vitamin D status with annual changes in estimated glomerular filtration rate (eGFR) and allograft rejection episodes. We also hypothesized that vitamin D deficiency/ inadequacy is more closely associated with these outcomes in short-term survivors of kidney transplantation (KTx), who are at higher risk for rejection episodes than longterm survivors.

A

Subjects and Methods Study population We examined a prospective cohort of KTRs (UltraSonographical Evaluation of ParaThyroid Hypertrophy study) visit-


ing the outpatient department in Inoue Hospital, Osaka, Japan, using a previously reported study design and sampling procedures (15). Briefly, participation was sought from all ambulatory KTRs; 293 patients (⬃92%) provided written informed consent between August 2007 and June 2008 and were followed up with monthly blood testing until June 2011. All patients had negative results in lymphocyte cytotoxic cross-match tests before KTx. The exclusion criteria for this study were subsequent transplantation (n ⫽ 10), acute rejection episodes within 3 months before enrollment (n ⫽ 5), documented nonadherence to immunosuppressants (n ⫽ 4), refractory urinary tract infection episodes (n ⫽ 2), missing data on 25D concentration (n ⫽ 2), pregnancy or desire to become pregnant (n ⫽ 4), and cancer (n ⫽ 2). The final study population comprised 264 patients (90.1%). Patients were censored if they died, developed end-stage renal disease (ESRD) requiring renal replacement therapy, wished to become pregnant, unexpectedly became pregnant, developed malignancy, and if they were lost to follow up for more than 3 months. This study was approved by our local ethics committee (number 179) and conformed to the Declaration of Helsinki.

Laboratory measurements Blood samples were drawn from each participant at enrollment and stored at ⫺30°C for subsequent analysis. The serum concentrations of creatinine, intact PTH, and the biologically active form of fibroblast growth factor 23 (FGF23) were measured using an enzymatic assay (Roche Creatinine Plus assay; Roche Diagnostics Corp), an immunochemiluminometric assay (ECLusys PTH; Roche Diagnostics), and a sandwich ELISA system (Kainos Laboratories, Inc), respectively. The urinary protein concentration was detected using a dipstick test. Laboratory parameters other than 25D were measured using a Roche Cobas 6000 chemistry analyzer (Roche Diagnostics). Two months after the last follow-up (August 2011), the samples were sent to Kyowa Medex, Inc for serum 25D assay using the DiaSorin LIAISON 25-hydroxy OH Vitamin D TOTAL Assay (DiaSorin, Inc) (16).

Outcome measures The primary outcome was the annual change in the eGFR. The estimated effect of exposure on the rate of kidney function decline is independent of whether the glomerular filtration rate is estimated or measured (17). We used the equation for eGFR provided in the Japanese clinical practice guidelines for patients with chronic kidney disease (CKD) (18). Using the difference in the eGFR values (⌬eGFR) and the number of years between the first and last measurements during the study period, we calculated the annual change in eGFR. For sensitivity analyses, we used the creatinine-based CKD Epidemiology Collaboration (CKD-EPI) equation (19), validated in a heterogeneous population including KTRs (20), along with the Japanese coefficient (21). We also used the last-observation-carried-forward method, wherein the annual eGFR change was calculated by dividing ⌬eGFR by the number of years between enrollment and June 30, 2011. The secondary outcome was rescue treatment with iv methylprednisolone (IV-MP) 250 mg or greater as an index of allograft rejection episodes. After IV-MP, it may be necessary to increase the daily dose of calcineurin inhibitors or antiproliferative agents; change cyclosporine to tacrolimus; change antiproliferative agents to mycophenolate mofetil; or administer deoxyspergualin (22). We did not use monoclonal antibodies or


doi: 10.1210/jc.2013-2421

mammalian target of rapamycin inhibitors. The treatment protocol was consistent with the standard clinical practice in Japan during the study period.

Study variables The primary exposures were serum 25D concentration and vitamin D status categories, as defined by the Institute of Medicine (deficiency, ⬍ 12 ng/mL; inadequacy, ⱖ 12 ng/mL and ⬍ 20 ng/mL; sufficiency, ⱖ20 ng/mL) (23). We accounted for seasonality because serum 25D concentrations show seasonal variation in solar UV-B light, the major source of vitamin D. Using the mean monthly cumulative UV-B dose data (24), we divided the seasons of measurement into 3 categories: low (November through February), intermediate (March, April, September, and October), and high (May through August). During the study period, we did not measure serum 25D concentrations or prescribe nutritional vitamin D. Other possible confounders are listed in Table 1. Active vitamin D compounds prescribed were calcitriol (59%) and alfacalcidol (41%). ABO blood type incompatibility and the number of human leukocyte antigen (HLA) mismatches were divided into three categories (compatible, incompatible, and no information) and four categories (0, 1–3, 4 – 6, and no information), respectively (25). The total calcium concentration was corrected for albumin if the albumin concentration was less than 4.0 g/dL (26).

Statistical analysis Values with normal and nonnormal distributions are expressed as mean ⫾ SD and median [interquartile range (IQR)], respectively. Categorical variables are expressed as proportions. The Kruskal-Wallis and ␹2 tests were used for intergroup comparisons of continuous and categorical variables, respectively. If differences were significant, Cuzick’s test and the ␹2 test were used to analyze trends. Multivariate linear regression with robust variances was used to analyze the association of baseline vitamin D status with annual eGFR decline, considering the potential heteroscedasticity and nonnormality of regression residuals. We hierarchically adjusted for the following confounders: age, gender, body mass index (BMI), UV-B seasonality, time after KTx, and the use of active vitamin D compounds (model 1); these factors in model 1 plus eGFR, positive proteinuria (ⱖ30 mg/dL), hemoglobin, and the number of HLA mismatches (model 2); and these factors in model 2 plus diabetes, donor source (living or cadaveric), laboratory markers [albumin, corrected calcium, phosphate, intact PTH, FGF23, C-reactive protein (CRP)]), and the use of erythropoiesis-stimulating agents (model 3). Linearity was checked using the quadratic term of each continuous variable. Additionally, serum 25D concentration was added to analyses as restricted cubic splines with three interior knots to visualize twodimensional trends. For the secondary outcome, rescue treatment with IV-MP, its cumulative incidence according to vitamin D status was estimated using the Kaplan-Meier method and compared using the log-rank test. We used stratified Cox regressions, assuming that the baseline hazard would differ among primary physicians because IV-MP administration depends on physicians’ clinical judgment. We also censored patients if they changed primary physicians. Proportional hazard assumption was tested by a time-covariate interaction term.


529

In these multivariate analyses, we excluded patients without data on the number of HLA mismatches. BMI and intact PTH concentration were natural log transformed. The time after KTx and dialysis vintage were also log transformed after adding 1 to the raw data. Serum CRP concentration was dichotomized at 0.2 mg/dL because of its extremely skewed distribution. We used a propensity score (PS) of vitamin D sufficiency (25D ⱖ 20 ng/mL) as a data-reduction technique by using a logistic regression model because the study size was too small to adjust for the potential confounders (Table 1) (27, 28). For PS estimation, we used the natural log-transformed serum creatinine concentration instead of eGFR as an index of kidney function, which enabled calculation of the annual eGFR decline using the Japanese equation and the creatinine-based CKD-EPI equation, in the propensity-matched cohort. The missing values for the number of HLA mismatches and ABO incompatibility were categorized as in Table 1 and used as covariates in the logistic regression model. The C index of this model was 0.85. We used optimal matching within the caliper of 25% SD of the logit (PS) (29) Then we conducted robust linear regression analyses for the annual change in eGFR and Cox regression analyses for rescue treatment with IV-MP to evaluate the associations of vitamin D inadequacy/deficiency. In the evaluations of the associations between vitamin D status and the outcomes, values of P ⬍ .05 were considered to indicate statistical significance. Effect modification to these associations by age, sex, BMI, time after KTx, and use of active vitamin D compounds was examined by adding each interaction term with serum 25D concentration to the multivariate analysis (model 1). Values of P ⬍ .10 were considered to indicate statistical significance for these interactions (30, 31). All analyses were conducted using STATA/SE 11.1 for Windows (STATA Corp LP).

Results The mean age of the patients was 49.0 (SD 12.3) years and 61.3% were male. The mean serum 25D concentration was 17.1 (SD 6.5) ng/mL and 68.4% patients had vitamin D inadequacy/deficiency. Patient characteristics by vitamin D status are shown in Table 1. Older age, male gender, low BMI, use of active vitamin D compounds, high UV-B season, high eGFR, high hemoglobin concentration, and low intact PTH concentration were associated with vitamin D sufficiency. Of these, high UV-B season, male gender, and low BMI showed independent relationships with vitamin D sufficiency (P ⬍ .05) in the multivariate logistic regression, whereas high hemoglobin concentration and high intact PTH showed nonsignificant trends (P ⬍ .1). Patients with type 1 diabetes were likely to have vitamin D deficiency. During a median follow-up period of 3.5 (IQR 3.2–3.7) years, eight patients died, 23 developed ESRD requiring maintenance dialysis or subsequent transplantation, and 59 patients received IV-MP. Only one patient was lost to follow-up.


530

Obi et al

Table 1.



Baseline Characteristics of the Patients According to Vitamin D Statusa

Basic information Age, y Male, % BMI Mean arterial pressure, mm Hg Dialysis vintage, y Time since KTx, y ABO incompatibility Compatible, % Incompatible, % No information, % HLA mismatches (A⫹B⫹DR) 0, % 1–3, % 4 – 6, % No information, % Living donor, % Donor age, y Medications Calcineurin inhibitor Cyclosporine, % Tacrolimus, % None, % Antiproliferative agent Azathioprine, % Mycophenolate mofetil, % Mizoribine, % None, % Prednisolone, % ESAs, % Active vitamin D compounds, % RAAS inhibitors, % Medical history Diabetes Type 1, % Type 2, % NODAT, % Previous cancer, % Chronic hepatitis C, % Laboratory data Hemoglobin, g/dL Albumin, mg/dL eGFR, mL/min per 1.73 m2 The Japanese equation The modified CKD-EPI equation Corrected calcium, mg/dL Phosphate, mg/dL Intact PTH, pg/mL 25D, ng/mL FGF23, pg/mL CRP, mg/dL Urinary protein ⱖ 30 mg/dL, % Season of blood draw Low UV-B season, % Intermediate UV-B season, % High UV-B season, %

Deficiency (n ⴝ 64)

Inadequacy (n ⴝ 118)

Sufficiency (n ⴝ 82)

45.0 ⫾ 12.7 32.8 20.4 (18.5–23.1) 93.2 ⫾ 11.9 2.0 (0.8 – 4.3) 10.4 (2.0 –17.8)

50.2 ⫾ 10.8 61.0 22.0 (19.9 –24.7) 90.9 ⫾ 10.1 2.8 (1.3– 6.4) 11.3 (5.3–18.0)

49.8 ⫾ 13.0 84.2 21.1 (19.6 –23.6) 92.1 ⫾ 7.6 1.5 (0.8 – 4.7) 9.0 (4.4 –15.5)

76.6 18.8 4.7

81.4 13.6 5.1

79.3 14.6 6.1

10.9 68.8 9.4 10.9 82.8 51.5 ⫾ 13.2

9.3 72.9 7.6 10.2 82.4 50.3 ⫾ 13.6

4.9 67.1 14.6 13.4 85.7 48.2 ⫾ 13.7

59.4 32.8 7.8

55.1 34.8 10.2

46.3 43.9 9.8

14.1 65.6 9.4 10.9 98.4 42.2 28.1 62.5

19.5 55.1 16.1 9.3 98.3 34.8 48.3 74.6

23.2 50.0 23.2 3.7 97.6 30.5 46.3 70.7

10.9 1.6 6.3 9.4 6.3

1.7 4.2 13.6 7.6 4.2

1.2 2.4 15.9 7.3 3.7

11.6 ⫾ 1.4 4.3 ⫾ 0.3

12.4 ⫾ 1.8 4.3 ⫾ 0.3

12.8 ⫾ 1.6 4.3 ⫾ 0.3

⬍.001** .931

38.4 ⫾ 14.9 42.2 ⫾ 17.7 9.1 ⫾ 0.5 3.3 ⫾ 0.8 85.6 (58.7–122.8) 9.5 (8.2–10.6) 60.0 (36.2–112.3) 0.10 (0.10 – 0.13) 56.3%

39.9 ⫾ 15.1 43.3 ⫾ 17.2 9.3 ⫾ 0.6 3.1 ⫾ 0.7 65.9 (45.1–101.4) 15.5 (13.9 –18.1) 67.8 (47.2–125.3) 0.10 (0.10 – 0.11) 44.9%

68.8 25.0 6.3

61.0 30.5 8.5

44.3 ⫾ 16.4 47.7 ⫾ 18.6 9.2 ⫾ 0.5 3.2 ⫾ 0.6 59.1 (42.0 – 88.4) 24.4 (21.3–26.5) 63.2 (43.2–113.4) 0.10 (0.10 – 0.10) 52.4% 0 36.6 50.0 13.4

.046* .117 .066 .350 ⬍.001** N/A .591 .448 .299 .001 ** **

P Value .015* ⬍.001** .020† .326 .025† .144 .905

.552

.656 .321 .554

.117

.908 .337 .023* .233 .017 ** † †

.661 .469

†

Abbreviations: ESA, erythropoiesis-stimulating agent; N/A, not applicable; NODAT, new-onset diabetes after transplantation; RAAS, reninangiotensin-aldosterone system. Conversion factors for units: serum intact PTH in picograms per milliliter to picomoles per liter ⫻ 0.1061; serum 25D in nanograms per milliliter to nanomoles per liter ⫻ 2.496; serum calcium in milligrams per deciliter to millimoles per liter ⫻ 0.2495; serum phosphorus in milligrams per deciliter to millimoles per liter ⫻0.3229. a

Data are expressed as mean ⫾ SD, median (IQR), and percentage.

†, P ⬍ .1; *, P ⬍ .05; **, P ⬍ .01 for post hoc trend tents.


doi: 10.1210/jc.2013-2421


531

change (Figure 1B). No significant association was found at 10 or more years after KTx (Figure 1, A and C). Similar interactions were found in the relationships to rescue treatment with IV-MP. Lower vitamin D status was significantly associated with higher risk of IV-MP in patients at less than 10 years after KTx (Figure 2A) but not in the others (Figure 2B). Propensity-matched cohort To adjust rigorously for potential confounders (Table 1), we matched patients according to the PS of vitamin D sufficiency using the whole cohort. In this PS-matched cohort, the median serum 25D concentration was 14.2 (IQR 11.3–16.2) ng/mL and 23.8 (IQR 21.2–26.4) ng/mL in patients with vitamin D inadequacy/ deficiency and vitamin D sufficiency, Figure 1. Association between vitamin D status and annual eGFR change in multivariate regression analyses. A, Adjusted association between vitamin D inadequacy/deficiency and respectively. The former group annual eGFR change at less than 10 years and 10 or more years after transplantation (vitamin D tended to have higher, albeit nonsigsufficiency as the reference). Model 1 was adjusted for age, gender, BMI, UV-B seasonality, time nificant, intact PTH concentrations after KTx, and the use of active vitamin D compounds. Model 2 was adjusted for those variables in model 1 plus eGFR, positive proteinuria (ⱖ30 mg/dL), hemoglobin, and the number of HLA (P ⫽ .082), whereas other baseline mismatches. Model 3 had these variables in model 2 plus diabetes, donor source (living or variables showed nonsignificant difcadaveric), laboratory markers (albumin, corrected calcium, phosphate, intact PTH, FGF23, CRP), ferences (Supplemental Table 1, and the use of erythropoiesis-stimulating agents. Adjusted annual eGFR change and distribution published on The Endocrine Sociof serum 25D concentration at less than 10 years (B) and 10 or more years after transplantation (C) are shown. Robust linear regression with cubic spline functions was applied with adjustments ety’s Journals Online web site at for model 3. Conversion factors for units: serum 25D in nanograms per milliliter to nanomoles http://jcem.endojournals.org). We per liter ⫻ 2.496. D, deficiency; I, inadequacy; S, sufficiency. observed 17 and 8 IV-MPs among patients with vitamin D inadequacy/ Multivariate regression analyses deficiency and vitamin D sufficiency, respectively. ComThe interaction term between serum 25D concentra- pared with vitamin D sufficiency, vitamin D inadequacy/ tion and time after transplantation at enrollment was sig- deficiency was more significantly associated with faster nificant for annual eGFR change in the linear regression annual eGFR decline (Figure 3); addition of the log-transand for rescue treatment with IV-MP in the Cox model (P formed intact PTH concentration into the models did not for interaction ⫽ .099 and .004, respectively). Therefore, significantly affect these relationships. Then we matched we divided the patients into two groups according to the patients based on PS after stratifying them by time after median time after KTx (10 y) and evaluated the associa- KTx (10 years), and found results consistent with the prior tions between vitamin D status and outcomes in each multivariate models; the association of low vitamin D stagroup. tus with a rapid eGFR decline was pronounced at less than With vitamin D sufficiency as the reference, vitamin D 10 years but not significant at 10 or more years after KTx inadequacy and deficiency were independently associated (Figure 3). Again, similar interactions were observed in the with rapid eGFR decline at less than 10 years after KTx in relationship to rescue treatment with IV-MP (Figure 4). all models (P for trend ⬍ .05, Figure 1A). A high number of HLA mismatches also showed a nonsignificant trend Sensitivity analyses For sensitivity analyses regarding allograft function detoward being associated with a rapid decline in eGFR in model 2 (P ⫽ .084). Multivariate regression with cubic cline, we used the creatinine-based CKD-EPI equation spline functions revealed that serum 25D concentration and/or last-observation-carried-forward method. The crehad an almost linear relationship with the annual eGFR atinine-based CKD-EPI equation afforded results similar


532

Obi et al



spective cohort of KTRs. More than two thirds of the enrolled ambulatory KTRs had vitamin D inadequacy/deficiency at baseline, consistent with previous reports (35–37). Time after KTx was a significant effect modifier for the association of serum 25D concentration with annual eGFR change and IV-MP. In patients at less than 10 years after KTx who have a higher risk of rejection than long-term survivors (38), vitamin D inadequacy/deficiency was significantly associated with a rapid decline in kidney function and a high cumulative probability of rescue treatment with IV-MP. However, no Figure 2. Estimated incidence of rescue treatment with IV-MP according to vitamin D status. significant association was observed Kidney transplant recipients at less than 10 years (A) and 10 or more years after transplantation at 10 or more years after KTx. (B) are shown. D, deficiency; I, inadequacy; S, sufficiency. PS-matched cohort analyses also to those obtained with the Japanese equation (Supplemen- showed that the associations between vitamin D status tal Figures 1 and 2), and the association of vitamin D status and these outcomes were significant overall and confirmed with annual eGFR decline in the last-observation-carried- that these associations were pronounced at less than 10 forward method was weaker but significant (data not years after KTx but nonsignificant at the other periods. shown). These findings suggest that the association between vitaAdditionally, we evaluated the association between semin D status and change in allograft kidney function over rum 25D concentration and annual eGFR change after time might be partly explained by the immunomodulatory excluding patients with diabetes, urinary protein conceneffects of vitamin D. tration of 30 mg/dL or greater, or inflammation (CRP ⱖ Certain findings seem to indicate the clinical relevance 0.2 mg/dL) because serum 25D concentrations are influof the observed association between vitamin D deficiency enced by these factors (32–34). The association remained and kidney function decline. With the average eGFR 1 significant, even in these analyses (P ⬍ .05). year after transplantation and the median years of kidney allograft survival being 60 mL/min per 1.73 m2 and 11 years, respectively (10, 39), some KTRs experience an anDiscussion nual eGFR decline of 5.0 mL/min per 1.73 m2, assuming We assessed the associations of vitamin D status with that renal replacement therapy is started at 10 mL/min per 2 eGFR decline and rescue treatment with IV-MP in a pro- 1.73 m (40). If vitamin D supplementation could ameliorate the annual eGFR decline by 1.0 mL/min per 1.73 m2 in a KTR with vitamin D deficiency [60% of the estimated association (Figure 1A)], allograft failure may be stalled for an additional 2.5 years. Further, PS-matched cohort analyses showed a significant high risk of a 50% increase in serum creatinine concentration, ESRD, or death in patients with low vitamin D (data not shown), although a 50% increase in serum creatinine is not established as an outcome for KTRs unlike its doubling. Adjustments were made for the confounders associated with serum 25D, namely intact PTH, FGF23, and hemoFigure 3. Association between vitamin D inadequacy/deficiency and annual eGFR change in propensity-score matched cohort analyses. globin, in multivariate analyses. Serum intact PTH conResults are obtained from the propensity score matching within the centrations were higher in KTRs with poor vitamin D stawhole cohort, patients at less than 10 years, and those with 10 or tus, as in the general population. The major regulator of more years after transplantation, respectively. D, deficiency; I, inadequacy; S, sufficiency. PTH concentrations in CKD patients is kidney function


doi: 10.1210/jc.2013-2421


533

vitamin D on allograft function; the inhibitory effect of vitamin D on the nuclear factor-␬B pathways and angiotensin II production may have contributed to changes in kidney allograft function over time. This discrepancy between their study and ours may be because their study was conducted in an early period after KTx when patients received highdose immunosuppressants, whereas most of our patients were long-term recipients receiving low-dose immuFigure 4. Estimated incidence of rescue treatment with IV-MP in propensity score-matched nosuppressants. Thus, the immunocohort analyses. Results are obtained from the propensity score matching within the whole cohort (A), patients at less than 10 years (B), and those at 10 or more years after transplantation modulatory effects of vitamin D (C), respectively. D, deficiency; I, inadequacy; S, sufficiency. could be observed 1 year after KTx when chronic ABMR plays a major rather than serum 25D concentrations (15, 41, 42). Inrole in deteriorating allograft kidney function (1). Indeed, deed, the strong relationship between the concentrations a recent randomized placebo-controlled trial revealed that of intact PTH and 25D was substantially attenuated after vitamin D supplementation improves disease activity in adjustment for eGFR in multivariate analyses. With repatients with systemic lupus erythematosus, which is a gard to FGF23, its serum concentration showed no difchronic autoimmune disease involving autoantibodies (8). ferences according to vitamin D status. Low serum calNotably, we could not rule out the possibility of residcium concentration might have counteracted the increase ual confounding and the effect of unmeasured confoundin the serum FGF23 concentration caused by the low ers because this was an observational study. First, the cateGFR in patients with vitamin D deficiency (43). High egorization of continuous variables such as urinary hemoglobin concentrations in patients with vitamin D sufprotein and serum CRP concentration may have led to ficiency can be explained by the high prevalence of men overestimation of the impact of vitamin D status. Howand high eGFR rather than the serum CRP concentration ever, the association of vitamin D status with annual eGFR (44), which was similar across the groups. However, CRP concentrations might be suppressed during immunosup- change remained almost unchanged, even after excluding pression. IL-6, a proinflammatory cytokine that indirectly patients with high urinary protein concentrations (ⱖ30 inhibits iron use by inducing liver production of hepcidin, mg/dL) or inflammation (serum CRP concentration ⱖ 0.2 mg/dL). Second, incomplete adjustment for continuous may correlate with vitamin D status and hemoglobin. Despite the adjustment for the baseline use of active variables such as hemoglobin and intact PTH concentravitamin D compounds, their effect remains unclear. When tions might have affected the estimated impact of vitamin estimating a drug effect, it is necessary to follow up pa- D status. However, we confirmed the presence of linear tients since treatment initiation and to adjust for pretreat- assumptions for these variables using quadratic terms and ment variables in multivariate analyses. However, because the propensity-matched cohort yielded results consistent most patients receiving active vitamin D compounds had with the multivariate models. Lastly, serum 25D concenbeen taking them for several years at baseline, the baseline trations may reflect the patients’ general health status parameters used in the multivariate analyses had been in- since vitamin D is produced by sunlight exposure. Howfluenced by active vitamin D compounds. Therefore, we ever, this is unlikely because vitamin D status showed no association with rapid eGFR decline and rescue treatment could not evaluate their effects in this study. Bienaime et al (10) reported that low serum 25D con- with IV-MP in recipients at 10 or more years after KTx centration 3 months after KTx is associated with a low despite their serum 25D concentrations being similar to GFR 1 year after transplantation. Our findings indicated those at less than 10 years after KTx. This study has several other limitations. First, serum that this association could be applied to long-term recipients and also showing that the impact of vitamin D status 25D concentrations were measured only at baseline, is more significant in patients with recent KTx. In contrast, thereby precluding the evaluation of whether a change in Bienaime et al showed a nonsignificant association be- vitamin D status was associated with allograft outcomes. tween vitamin D status and rejection episodes and impli- However, the impact of baseline serum 25D concentracated a nonimmune response in the protective effect of tions was unlikely to be artificially modified because we


534

Obi et al


did not measure serum 25D concentrations or prescribe nutritional vitamin D during the study period. Additionally, the prevalence of vitamin supplement use is only 3.4% in Japan, irrespective of whether the nutritional supplements administered contain vitamin D (14). Although seasonal variation in serum 25D concentrations may have biased the results, we adjusted for the season of blood draw, which is a common practice in epidemiological studies. Second, we measured serum 25D concentrations using the Diasorin Liaison assay. Although a nondifferential measurement error in exposure leads to a bias toward the null (45), we found significant associations with allograft outcomes. Possible explanations are that serum 25D concentration obtained from liquid chromatography with tandem mass spectrometry and RIA correlated well (26, 46, 47) and that the associations of serum 25D concentration with allograft outcomes were sufficiently strong for detection even with this assay. Further studies using the liquid chromatography with tandem mass spectrometry method are required to estimate the accurate strength of the association with allograft outcomes. Finally, because this cohort was originally designed to evaluate the progression of persistent hyperparathyroidism in KTRs and not allograft rejection (15), data on renal histology as a proof of rejection, which is ideally necessary before IV-MP, were unavailable. Therefore, we could not distinguish between TCMR and ABMR in each rejection episode, and the need for IV-MP might not have necessarily reflected an actual rejection episode. However, both ABMR and TCMR contribute to subsequent allograft loss, and the Organ Procurement and Transplantation Network registry indicated that the need for rescue treatment is a good marker of rejection because it significantly impacts graft loss (38). Despite the limitations mentioned above, six of nine criteria of Hill (48) were fulfilled in this study, suggesting a causal association of vitamin D status with allograft function decline and rejection episodes. First, the association of vitamin D status was evaluated longitudinally (temporality). It showed a relevant effect size (strength), and a relationship between baseline vitamin D concentration and clinical outcomes (biological gradient). The renoprotective effects of vitamin D have been studied in many laboratory investigations, including rodent kidney transplant models (49, 50). Moreover, the association between vitamin D status and allograft kidney function noted in this study concurs with the findings of a previous observational study (consistency) (10). In conclusion, vitamin D deficiency/inadequacy in KTRs independently predicted eGFR decline and rescue treatment for IV-MP as an index of rejection episodes. The effect of treating vitamin D deficiency cannot be estimated


from this study because no patients received vitamin D supplements in this study. Therefore, future studies are warranted to test whether vitamin D supplementation early after KTx could improve allograft outcomes. Accordingly, we have launched a multicenter, open-label randomized controlled trial (number NCT01817699) with a 2 ⫻ 2 factorial design to evaluate the effects of anemia correction and/or vitamin D supplementation in KTRs.

Acknowledgments We thank all the staff at the outpatient section of the kidney transplantation department in Inoue Hospital for cooperating in the implementation of this study. Special thanks are also due to Yuko Watanabe, the kidney transplant coordinator, and Yuji Yamauchi (Department of Clinical Laboratory, Inoue Hospital). Address all correspondence and requests for reprints to: Yoshitaka Isaka, MD, Department of Geriatric Medicine and Nephrology, Osaka University Graduate School of Medicine, B6 2–2 Yamada-oka, Suita 565– 0871, Osaka, Japan. E-mail: [email protected]. This work was supported by Grant 17– 6 from the Osaka Medical Research Foundation for Incurable Diseases (to Y.O.). Disclosure Summary: T.H. has served as a consultant to Kyowa Medex Co Ltd (Tokyo, Japan), which measured serum 25D concentrations in this study. Kyowa Medex Co Ltd was not involved in the data analysis or the manuscript writing. None of the other authors had a potential conflict of interest.

References 1. Sellares J, de Freitas DG, Mengel M, et al. Understanding the causes of kidney transplant failure: the dominant role of antibody-mediated rejection and nonadherence. Am J Transplant. 2012;12:388 –399. 2. Bartel G, Schwaiger E, Bohmig GA. Prevention and treatment of alloantibody-mediated kidney transplant rejection. Transpl Int. 2011;24:1142–1155. 3. Wu O, Levy A, Briggs A, Lewis G, Jardine A. Acute rejection and chronic nephropathy: a systematic review of the literature. Transplantation. 2009;87:1330. 4. Courbebaisse M, Souberbielle JC, Thervet E. Potential nonclassical effects of vitamin D in transplant recipients. Transplantation. 2010; 89:131–137. 5. Redaelli CA, Wagner M, Gunter-Duwe D, et al. 1␣,25-Dihydroxyvitamin D3 shows strong and additive immunomodulatory effects with cyclosporine A in rat renal allotransplants. Kidney Int. 2002; 61:288 –296. 6. Gregori S, Casorati M, Amuchastegui S, Smiroldo S, Davalli AM, Adorini L. Regulatory T cells induced by 1␣,25-dihydroxyvitamin D3 and mycophenolate mofetil treatment mediate transplantation tolerance. J Immunol. 2001;167:1945–1953. 7. Plum LA, DeLuca HF. Vitamin D, disease and therapeutic opportunities. Nat Rev Drug Discov. 2010;9:941–955. 8. Abou-Raya A, Abou-Raya S, Helmii M. The effect of vitamin D supplementation on inflammatory and hemostatic markers and disease activity in patients with systemic lupus erythematosus: a randomized placebo-controlled trial. J Rheumatol. 2013;40:265–272.


doi: 10.1210/jc.2013-2421

9. Soilu-Hanninen M, Aivo J, Lindstrom BM, et al. A randomised, double blind, placebo controlled trial with vitamin D3 as an add on treatment to interferon ␤-1b in patients with multiple sclerosis. J Neurol Neurosurg Psychiatry. 2012;83:565–571. 10. Bienaime F, Girard D, Anglicheau D, et al. Vitamin D status and outcomes after renal transplantation. J Am Soc Nephrol. 2013;24: 831– 841. 11. Bitetto D, Fabris C, Falleti E, et al. Vitamin D and the risk of acute allograft rejection following human liver transplantation. Liver Int. 2010;30:417– 444. 12. Lowery EM, Bemiss B, Cascino T, et al. Low vitamin D levels are associated with increased rejection and infections after lung transplantation. J Heart Lung Transplant. 2012;31:700 –707. 13. Sasaki S. Dietary reference intakes (DRIs) in Japan. Asia Pac J Clin Nutr. 2008;17:420 – 444. 14. Nanri A, Foo LH, Nakamura K, et al. Serum 25-hydroxyvitamin d concentrations and season-specific correlates in Japanese adults. J Epidemiol. 2011;21:346 –353. 15. Tomida K, Hamano T, Ichimaru N, et al. Dialysis vintage and parathyroid hormone level, not fibroblast growth factor-23, determines chronic-phase phosphate wasting after renal transplantation. Bone. 2012;51:729 –736. 16. Wagner D, Hanwell HE, Vieth R. An evaluation of automated methods for measurement of serum 25-hydroxyvitamin D. Clin Biochem. 2009;42:1549 –1556. 17. Lewis J, Greene T, Appel L, et al. A comparison of iothalamate-GFR and serum creatinine-based outcomes: acceleration in the rate of GFR decline in the African American Study of Kidney Disease and Hypertension. J Am Soc Nephrol. 2004;15:3175–3183. 18. Imai E, Horio M, Nitta K, et al. Estimation of glomerular filtration rate by the MDRD study equation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol. 2007;11:41–50. 19. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604 – 612. 20. Stevens LA, Schmid CH, Zhang YL, et al. Development and validation of GFR-estimating equations using diabetes, transplant and weight. Nephrol Dial Transplant. 2010;25:449 – 457. 21. Horio M, Imai E, Yasuda Y, Watanabe T, Matsuo S. Modification of the CKD epidemiology collaboration (CKD-EPI) equation for Japanese: accuracy and use for population estimates. Am J Kidney Dis. 2010;56:32–38. 22. Amemiya H. 15-Deoxyspergualin: a newly developed immunosuppressive agent and its mechanism of action and clinical effect: a review. Japan Collaborative Transplant Study Group for NKT-01. Artif Organs. 1996;20:832– 835. 23. Ross AC, Taylor CL, Yaktine AL, Del Valle HB. Dietary Reference Intakes for Calcium and Vitamin D. Washington, DC: National Academy Press; 2011. 24. Japan Meteorological Agency. Mean daily cumulative dose of ultraviolet-B radiation by month. http://www.data.kishou.go.jp/obsenv/uvhp/uvb_monthave_tsu.html. Accessed January 12, 2013. 25. The Japanese Society for Clinical Renal Transplantation TJSft. Annual Progress Report from the Japanese Renal Transplant Registry, 2010. Part III: results from 2009 Recipient Follow-up Survey. Isyoku. 2010;45:608 – 620. 26. KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD). Kidney Int Suppl. 2009;113:S1–S130. 27. Peduzzi P, Concato J, Feinstein AR, Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol. 1995;48:1503–1510. 28. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol. 2007;165: 710 –718.


535

29. Rosenbaum PR. Optimal matching for observational studies. J Am Statist Assoc. 1989;84:1024. 30. Duncan JL, Harrild KA, Iversen L, Lee AJ, Godden DJ. Long term outcomes in men screened for abdominal aortic aneurysm: prospective cohort study. BMJ. 2012;344:e2958. 31. Selvin S. Statistical Analysis of Epidemiologic Data. Oxford, United Kingdom: Oxford University Press; 2004. 32. Tanaka H, Hamano T, Fujii N, et al. The impact of diabetes mellitus on vitamin D metabolism in predialysis patients. Bone. 2009;45: 949 –955. 33. Dobnig H, Pilz S, Scharnagl H, et al. Independent association of low serum 25-hydroxyvitamin d and 1,25-dihydroxyvitamin d levels with all-cause and cardiovascular mortality. Arch Intern Med. 2008; 168:1340 –1349. 34. Reid D, Toole BJ, Knox S, et al. The relation between acute changes in the systemic inflammatory response and plasma 25-hydroxyvitamin D concentrations after elective knee arthroplasty. Am J Clin Nutr. 2011;93:1006 –1011. 35. Ewers B, Gasbjerg A, Moelgaard C, Frederiksen AM, Marckmann P. Vitamin D status in kidney transplant patients: need for intensified routine supplementation. Am J Clin Nutr. 2008;87:431– 437. 36. Stavroulopoulos A, Cassidy MJ, Porter CJ, Hosking DJ, Roe SD. Vitamin D status in renal transplant recipients. Am J Transplant. 2007;7:2546 –2552. 37. Querings K, Girndt M, Geisel J, Georg T, Tilgen W, Reichrath J. 25-hydroxyvitamin D deficiency in renal transplant recipients. J Clin Endocrinol Metab. 2006;91:526 –529. 38. Lentine KL, Gheorghian A, Axelrod D, Kalsekar A, L’Italien G, Schnitzler MA. The implications of acute rejection for allograft survival in contemporary US kidney transplantation. Transplantation. 2012;94:369 –376. 39. Lamb KE, Lodhi S, Meier-Kriesche HU. Long-term renal allograft survival in the United States: a critical reappraisal. Am J Transplant. 2011;11:450 – 462. 40. National Institute of Diabetes and Digestive and Kidney Diseases. US Renal Data System, USRDS 2012 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2012. 41. Isakova T, Wahl P, Vargas GS, et al. Fibroblast growth factor 23 is elevated before parathyroid hormone and phosphate in chronic kidney disease. Kidney Int. 2011;79:1370 –1378. 42. Nakano C, Hamano T, Fujii N, et al. Combined use of vitamin D status and FGF23 for risk stratification of renal outcome. Clin J Am Soc Nephrol. 2012;7:810 – 819. 43. Rodriguez-Ortiz ME, Lopez I, Munoz-Castaneda JR, et al. Calcium deficiency reduces circulating levels of FGF23. J Am Soc Nephrol. 2012;23:1190 –1197. 44. Kendrick J, Targher G, Smits G, Chonchol M. 25-Hydroxyvitamin D deficiency and inflammation and their association with hemoglobin levels in chronic kidney disease. Am J Nephrol. 2009;30:64 –72. 45. Jurek AM, Greenland S, Maldonado G, Church TR. Proper interpretation of non-differential misclassification effects: expectations vs observations. Int J Epidemiol. 2005;34:680 – 687. 46. Saenger AK, Laha TJ, Bremner DE, Sadrzadeh SM. Quantification of serum 25-hydroxyvitamin D(2) and D(3) using HPLC-tandem mass spectrometry and examination of reference intervals for diagnosis of vitamin D deficiency. Am J Clin Pathol. 2006;125:914 –920. 47. Tsugawa N, Suhara Y, Kamao M, Okano T. Determination of 25hydroxyvitamin D in human plasma using high-performance liquid chromatography-tandem mass spectrometry. Anal Chem. 2005;77: 3001–3007. 48. Hill AB. The environment and disease: association or causation? Proc R Soc Med. 1965;58:295–300. 49. Li YC. Renoprotective effects of vitamin D analogs. Kidney Int. 2010;78:134 –139. 50. Stein EM, Shane E. Vitamin D in organ transplantation. Osteoporos Int. 2011;22:2107–2118.


Vitamin D Status and Kidney Function Decline in HIV-Infected Men: A Longitudinal Study in the Multicenter AIDS Cohort Study.

Neck circumference predicts renal function decline in overweight women: A community-based prospective cohort study.

High Serum Level of β2-Microglobulin in Late Posttransplant Period Predicts Subsequent Decline in Kidney Allograft Function: A Preliminary Study.

Vitamin D deficiency predicts prostate biopsy outcomes.

Vitamin D Status, Muscle Strength and Physical Performance Decline in Very Old Adults: A Prospective Study.

Vitamin D Deficiency Strongly Predicts Adverse Medical Outcome Across Different Medical Inpatient Populations: Results From a Prospective Study.

Cardiac autonomic neuropathy predicts renal function decline in patients with type 2 diabetes: a cohort study.

Hyperhomocysteinemia predicts renal function decline: a prospective study in hypertensive adults.

Prevalence and risk factors for efavirenz-based antiretroviral treatment-associated severe vitamin D deficiency: A prospective cohort study.

Vitamin D deficiency aggravates ischemic acute kidney injury in rats.

Maternal vitamin D status and infant outcomes in rural Vietnam: a prospective cohort study.

Vitamin D status and severity of Clostridium difficile infections: a prospective cohort study in hospitalized adults.

Vitamin D deficiency and length of pediatric intensive care unit stay: a prospective observational study.

Vitamin D and activated vitamin D in tuberculosis in equatorial Malaysia: a prospective clinical study.

Circulating Concentrations of Vitamin B6 and Kidney Cancer Prognosis: A Prospective Case-Cohort Study.

Risk factors for renal function decline in adults with normal kidney function: a 7-year cohort study.

Vitamin D in preterm infants: A prospective observational study.

Vitamin D Status and Rates of Cognitive Decline in a Multiethnic Cohort of Older Adults.

Vitamin-D deficiency predicts infections in young north Indian children: A secondary data analysis.

Understanding vitamin D deficiency.

Association between maternal vitamin D deficiency and small for gestational age: evidence from a meta-analysis of prospective cohort studies.

Vitamin D deficiency.

The clinical frailty scale predicts functional decline and mortality when used by junior medical staff: a prospective cohort study.

Vitamin D deficiency in adolescents.