The Association Between Vitamin D and Acute Rejection in Human Kidney Transplantation: A Systematic Review and Meta-analysis Study

Background: Vitamin D (VitD) deciency is associated with several diseases such as multiple sclerosis, rheumatoid arthritis, respiratory infection, and so forth. In the eld of transplantation (kidney transplantation), some studies reported that patients with VitD deciency are of increased risk of acute rejection, but other studies did not show such a risk. On the other hand, since VitD is a modulatory factor and can reduce the inammatory response, understanding the exact role of it in transplantation may contribute to tolerance condition in these patients. Methods: The electronic databases, including PubMed, Scopus, Embase, ProQuest, Web of Science, and Google Scholar, were searched for eligible studies. In general, 14 studies with a total of 4,770 patients were included in this meta-analysis. Regarding the methodological heterogeneity, we selected a random-effects combination model. Moreover, OR was chosen as an effect size for this study. Results: After the combination of 14 studies, we showed that patients in the VitD-decient group had an 82% increased risk of acute rejection compared with patients in the VitD-sucient group, and this effect was signicant (OR 1.82; 95% condence interval [CI] [1.29, 2.56]; I 2 = 52.3%). This result was signicant, and, regarding the narrow CI, it can be a conclusive result. Study quality and gender variables were the main sources of inconsistent results in the primary studies. Moreover, using meta-regression, we showed that VitD deciency (independent from the estimated glomerular ltration rate (eGFR) of patients) increased the risk of acute rejection. Conclusion: The normal VitD status of patients a few days before and after transplantation can reduce the risk of acute rejection, as it has denite modulatory effects on immune cells. reducing the risk of acute rejection, we suggest that clinicians keep the VitD status of their patients normal, before and after transplantation.


Introduction
Vitamin D (VitD) is a fat-soluble steroid, which participates in bone health through regulating the calcium and phosphate metabolism. VitD is produced after skin exposure to ultraviolet solar irradiation or can be acquired in the diet or vitamin supplements. Cholecalciferol (D3) and ergocalciferol (D2) are two main sources of VitD, which needs two steps of hydroxylation to become biologically active. After the rst hydroxylation, which is performed in the liver, the D2 and D3 are converted to 25-hydroxyvitamin D. It is the main type of VitD in the circulation with a half-life of 2 -3 weeks and is routinely measured to determine the VitD status. The second hydroxylation is performed in the kidney through cytochrome P450 1-α hydroxylase enzyme CYP27B1, leading to the production of the active form of VitD, i.e., 1,25-dihydroxyvitamin D (1)(2)(3)(4).
VitD is a crucial factor for the immune system, as its receptor is expressed on monocytes, macrophages, dendritic cells (DCs), and activated lymphocytes. 25-hydroxyvitamin D can bind to its receptor on mentioned immune cells and be converted to 1,25-dihydroxyvitamin D. Then, 1,25dihydroxyvitamin D increases the production of antimicrobial peptides, such as cathelicidin and defensin (3,(5)(6)(7).
In contrast, the effect of VitD on DCs and T/B lymphocytes is regulatory. DCs and T/B lymphocytes are the major immune cells, initiating the immune responses of the recipient against the allograft (3). The VitD reduces the in ammatory cytokines of T helper 1 (TH1) and TH17 and increases the antiin ammatory cytokines of TH2 and regulatory T cells (Tregs). Moreover, VitD reduces the proliferation, immunoglobulin G (IgG), and immunoglobulin M (IgM) of B lymphocytes (3,6,(8)(9)(10)(11). Another important regulatory effect of VitD is on DCs. DCs are very crucial cells to initiate and stimulate T cells against the allograft. VitD could reduce the antigen presentation capacity of DCs by reducing the major histocompatibility class II (MHCII) and costimulatory molecules. These regulatory effects of VitD on the immune system may contribute to a better function of allograft and reduce the risk of allograft rejection (3,12,13).
VitD de ciency (VitD < 20 ng/ml) (14) is associated with several immune and non-immune diseases and disorders, including multiple sclerosis, rheumatoid arthritis, respiratory infection (3), cystic brosis (15), cardiovascular (16), and hematological diseases (17). Moreover, in the eld of transplantation (kidney transplantation), some studies reported that VitD de ciency remarkably increases the risk of acute rejection (18, 19), while others did not indicate such a risk (20,21). Since VitD de ciency is more common in the kidney transplant recipients compared with the general population (19,22), and there is no consistency regarding the association between VitD concentrations and acute rejection, we aimed to carry out the rst systematic review and meta-analysis to clarify whether VitD-de cient patients are more susceptible to acute rejection.

Protocol and Registration
This systematic review and meta-analysis study was prospectively registered with the National Institute for Health Research (PROSPERO: CRD42020216086). We followed the Reporting Checklist for Meta-analyses of Observational Studies (MOOSE) (23) and Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) (24) statements to report this study. All observational studies (such as cross-sectional, case-control, and cohort), as well as interventional studies (randomized or non-randomized), were included in this study. In contrast, case-report, case-series, narrative reviews, animal studies, and letters were excluded.

Type of Participants
Kidney transplant recipients with and without acute rejection were enrolled in this study, which their pre-or post-transplant 25-hydroxyvitamin D levels were assessed. Recipients features are as follows: (i) biopsy-proven or clinically suspected acute rejection, (ii) received living or deceased donor kidney transplant, (iii) received their transplant for the rst time or more, (iv) have experienced acute rejection up to 12 months after transplantation, and (v) with any age and gender. We excluded studies that did not separate the acute and chronic rejected patients and that measured 1,25-dihydroxyvitamin D instead of 25-hydroxyvitamin D.

Outcome
The outcome is the occurrence of acute rejection in kidney transplant recipients. Acute rejection is any type of cell-mediated or antibody-mediated rejection with any score or grade.

Search Strategy
The electronic databases searched by authors were PubMed, Scopus, Embase, ProQuest, Web of Science, and Google Scholar. Moreover, the gray literature (conferences/congress papers and theses) and reference lists of included studies were searched. The primary studies are from 1 January 1990 up to 31 October 2020. We used the keywords "kidney transplantation," "acute rejection," "25-hydroxyvitamin D," and their synonym for the search syntax. The Medical Subject Headings (MeSH) and free-text method were used to create search syntax. The search syntax of the PubMed database is shown in Supplemental File 1.

Study Selection
The records searched in all databases were exported to the EndNote software, and duplicated records were removed. Authors performed the screening process using titles and abstracts. Then, two reviewers independently read the full text of screened papers and selected those that had the main data for conducting meta-analysis. Any disagreement was resolved by consensus.

Risk of Bias Assessment
The quality assessment of included studies was independently done by two reviewers using the Newcastle-Ottawa Quality Assessment Scale (NOS) for observational studies (25). Moreover, the quality of a trial study was assessed using the Cochrane checklist (26). The modi ed version of NOS checklists for observational studies are shown in Supplemental File 2.

Data Extraction
Two reviewers independently extracted data from included studies using a data extraction form as follows: rst author name, publication year, maintenance therapy, induction therapy (IT), donor type (living vs. deceased), estimated glomerular ltration rate (eGFR), and age and gender of study subjects. Any disagreement was resolved by consensus, and, for incomplete data, we contacted the corresponding authors of the studies.

Data Synthesis and Analyses
The main data for conducting the meta-analysis are as follows (i.e., four numbers): the numbers of non-acute and acute rejection patients in the VitDde cient (VitD < 20 ng/ml) group and the numbers of non-acute and acute rejection patients in the VitD-su cient (VitD ≥ 20 ng/ml) group. We used the term "su cient" for VitD ≥ 20 ng/ml patients in the whole manuscript and gures. According to the type of included studies and main data for performing the meta-analysis, we employed OR as an effect size to be reported in this study. According to the methodological heterogeneity between included studies, we selected a random-effects model as an appropriate combination model (27). The pooled effect size was plotted using a forest plot.
Statistical heterogeneity between included studies was evaluated using the I 2 index (28). The severe statistical heterogeneity was considered I 2 > 50%.

Publication Bias Assessment
We used three methods to assess publication bias as follows: the funnel plot (which is the visual assessment of publication bias), Begg's (29) and Egger's (30) tests (whish statistically assess publication bias), and trim and ll method (which assesses probable missing papers) (31).

Subgroup Analyses
We performed some subgroup analyses to nd out potential sources of statistical heterogeneity observed between included studies. The other purpose of subgroup analyses was to know whether the pooled effect size was in uenced by different variables in each subgroup analysis.
Subgroup analyses were performed as follows:

Quality of Studies
Studies with high quality (i.e., ≥ 6) were compared with those with low quality (i.e., < 6).

Induction Therapy
Studies with induction therapy (IT) were compared with those without IT.

Age
Studies with patients ≤ 47 years were compared with those with patients > 47 years in each group (VitD < 20 ng/ml and VitD ≥ 20 ng/ml) separately.

Gender
Studies with more male patients were compared with those with fewer male patients.

Donor Type
Studies with living donor more than 50% were compared with those with living donor less than 50% in each group (VitD < 20 ng/ml and VitD ≥ 20 ng/ml) separately.

Meta-Regression Analysis
Regarding the low number of studies that reported the eGFR of their patients, we decided to do meta-regression to evaluate whether eGFR affect the pooled effect size.

Sensitivity Analysis
We utilized the one-out-remove method to do sensitivity analysis. In this method, the pooled effect size was recalculated after removing the effect of each study. This method was used to know whether there was any difference between the results after removing a study. If yes, how much does the recalculated pooled effect size change?
This study was analyzed using STATA Version 13 (StataCorp, College Station, TX, USA).

Study Selection
After searching the electronic databases, we identi ed a total of 9,398 records. Literature search procedure is shown in Figure 1. After screening the records according to the title and abstract, 924 records remained, and, after removing duplicated records, 766 records remained. After full-text assessment of 766 records, 14 records included in the meta-analysis. Finally, we included 14 studies (18-22, 32-40) with a total of 4,770 kidney transplant recipients in this systematic review and meta-analysis.

Study Characteristics
Studies included in this meta-analysis are from 2005 to 2020. Among 14 included studies, 10 studies were cohort studies, three were cross-sectional, and one was a randomized clinical trial (RCT). Eleven studies reported the patients' characteristics in two groups (i.e., VitD-de cient and -su cient groups), while three studies did not. These studies reported several variables including immunosuppressive agents (14 studies), age (11 studies), gender (11 studies), donor type (10 studies), and eGFR (nine studies) (see Table 1). Moreover, the quality of all 14 included studies is reported in Supplemental File 3.

The Association Between VitD De ciency and Acute Rejection
Patients in the VitD-de cient group had an 1.82-fold increased risk of acute rejection (82%) compared with patients in the VitD-su cient group, and this effect was signi cant (OR 1.82; 95% con dence interval [CI] [1.29, 2.56]; I 2 = 52.3%). CI was narrow, which indicated that this result was conclusive and can be generalized to a larger population (Figure 2).

Publication Bias Assessment
For the publication bias assessment, three methods were employed. As a rst method, the funnel plot showed a homogeneous pattern and therefore non-considerable publication bias (Supplemental Figure 1). As a second method, in line with the previous method, Begg's and egger's tests showed non-considerable publication bias as well (p = 0.913 and p = 0.744, respectively). Egger's graph is shown in Supplemental Figure 2. Finally, as a third method, the trim and ll method added one paper to our studies; however, the added study did not affect the pooled effect size (Figure 3 and Supplemental File 4).

Subgroup Analyses
As it was shown in Figure 1, primary studies had different results. We performed some subgroup analyses to nd potential sources of this statistical heterogeneity. Moreover, subgroup analyses let us analyze studies that are similar together on the aspect of at least one variable. The results of subgroup analyses are summarized in Table 2.

Quality of Studies
High-quality studies were compared with low-quality ones. There were six studies in the high-quality subgroup and eight studies in the low-quality one.
High-quality studies showed 2.16-fold increased risk of acute rejection.  Figure 7).

Meta-Regression Analysis
Nine studies reported the rate of eGFR. We performed a meta-regression analysis on eGFR in each group separately. The eGFR of patients did not affect the pooled effect size. In other words, VitD de ciency independent from the eGFR of patients increased the risk of acute rejection (Figure 7).

Sensitivity Analysis
Sensitivity analysis showed that if the studies are separately omitted, the pooled effect size does not change remarkably. Therefore, the pooled effect size had a robust result (Figure 8 and Supplemental File 5).

Discussion
A total of 14 studies with 4,770 kidney transplant recipients were eligible to be included in this study. This is the rst systematic review and metaanalysis study, which assesses the association between VitD concentration and the risk of acute rejection in kidney transplant recipients. The results of primary studies regarding the association between VitD and the risk of acute rejection were not conclusive. Some of them indicated that VitD de ciency is associated with acute rejection occurrence, but others not. These different statistical results may result from the distinct characteristics of studies' subjects. Thus, we performed different subgroup analyses to nd the potential sources of statistical heterogeneity observed between primary studies.
In general, after the combination of all studies (14 studies and 4,770 patients), we showed that patients with VitD de ciency had an 82% increased risk of acute rejection. This result was not in uenced by random errors, and, regarding the narrow CI, we can conclude that this result is conclusive and can be generalized to a larger population. Further, we indicated that several variables separately reduced the statistical heterogeneity (Table 2). For instance, study quality and gender of patients were the most important factors that reduced the heterogeneity. As an example, in the low-quality category, the heterogeneity was reduced from 52.3% to 0.0%; further, in studies with more male patients, the heterogeneity was reduced from 52.3% to 0.0% and 5.0% in the de cient and su cient groups, respectively.
The study quality and male gender also affected the pooled effect size. High-quality studies increased the risk of acute rejection from OR 1.82 to OR 2.91, while low-quality studies reduced the risk of acute rejection from OR 1.82 to OR 1.41. In other words, the association between VitD and the risk of acute rejection is higher in high-quality studies compared with low-quality ones. Thus, according to high quality studies, VitD-de cient patients had a 2.9-fold increased risk of acute rejection, which was a signi cant and conclusive result.
As mentioned above, another factor that affected the pooled effect size was the male gender. Studies with more male patients increased the pooled effect size from OR 1.82 to OR 2.85, which, regarding the CI, this result is inconclusive. On the other hand, studies with fewer male patients reduced it from OR 1.82 to OR 1.65.
Similar results were observed when VitD-de cient or -su cient groups had more male patients ( Table 2). The result of more-male-patients subgroup in the de cient group was signi cant and partially conclusive. However, the result of more-male-patients subgroup in the su cient group was inconclusive. Consequently, it could be implied that male patients with VitD de ciency are more susceptible to acute rejection compared with female patients with VitD de ciency.
The age of patients was another factor that affected the pooled effect size. Patients with VitD de ciency ≤ 47 years showed a 2.23-fold increased risk of acute rejection, which was signi cant and partially conclusive. Consequently, besides the male gender variable, patients with VitD de ciency ≤ 47 years are more susceptible to acute rejection compared with patients with VitD de ciency > 47 years.
Using meta-regression, we assessed the effect of eGFR on the pooled effect size. This analysis showed that patients with VitD de ciency independent from their eGFR had an increased risk of acute rejection.
After transplantation, the in ammatory reactions are initiated against the allograft. Because of danger signals and innate allorecognition, innate immunity is immediately activated against the allograft and thereby result in the activation of adaptive immunity. These in ammatory responses have a negative effect on the transplantation outcome and could be a reason for acute rejection (41)(42)(43)(44)(45)(46). Given the anti-in ammatory role of VitD (which was stated in the Introduction section), VitD non-de ciency (≥ 20 ng/ml) may modulate such in ammation and reduce the risk of acute rejection.
In summary, the result of this systematic review and meta-analysis study indicated that patients with VitD de ciency had an increased risk of acute rejection. In addition, study quality and male gender were the main sources of statistical heterogeneity observed in the primary studies. Regarding the modulatory effect of VitD and also its role in reducing the risk of acute rejection, we suggest that clinicians keep the VitD status of their patients normal, before and after transplantation.

Con ict interest statement
The authors declare no con ict of interest.   The forest plot of the association between VitD de ciency and acute rejection The subgroup analysis according to the quality of studies The subgroup analysis according to the gender of patients The meta-regression analysis according to the eGFR of VitD-de cient or -su cient patients