Impact of High Body Mass Index on The Allograft Outcomes in Kidney Transplant Recipients with Pre-Sensitization to Human Leukocyte Antigen: A Retrospective Cohort Study

Background The aim of this study was to investigate whether high body mass index (BMI) and pre-sensitization to human leukocyte antigen (HLA) in kidney transplant recipients (KTR) had a synergistic impact on the allograft outcomes. Methods From January 2010 to December 2018, 1,290 KTs were performed in Seoul St Mary's Hospital. Of these, 682 cases of ABO compatible KT were enrolled. They were divided into 4 groups (low BMI-non-sensitized, high BMI-non-sensitized, low BMI sensitized, high BMI-sensitized) according to the median BMI value (22.7 kg/m 2 ) and HLA pre-sensitization status (HLA-DSA MFI >3,000). Short-term and long-term allograft outcomes were compared between the groups. Results The rate of late antibody mediated rejection tended to be the highest in the high BMI-sensitized group, and the decline in allograft function in the high BMI-sensitized group was higher than in the other 3 groups. Death-censored graft loss (DCGL) rates and the hazard ratio (HR) for DCGL were the highest in the high BMI-sensitized group (4/21 (19.0%), HR 4.648, P = 0.022) and a signicant interaction was detected between high BMI and HLA pre-sensitization status (P value for interaction = 0.008). Conclusions Our results suggest that pre-sensitization to HLA and high BMI in KTRs might have a synergistic adverse impact on the allograft outcomes. number, type of induction therapy, main immunosuppressant used, and drug used for desensitization. kidney transplantation, ESRD: end stage renal disease, CGN: clinical glomerulonephritis, DSA: donor specic antibody, MFI: mean uorescence intensity stage renal disease; FCXM: ow cytometry-crossmatch; HDL: high-density lipoprotein; HLA: human leukocyte antigen; HLA-DSA: donor specic anti-human leukocyte antigen antibody; HR: hazard ratio; HTN: hypertension; IVIG: intravenous immunoglobulin; KT: kidney transplantation; KTR: kidney transplant recipient; LDL: low-density lipoprotein; MFI: mean uorescence intensity; PJP: pneumocystis jiroveci pneumonia; PRA: panel reactive antibody; TCMR: T-cell mediated rejection; Th: helper T-cell; TPE: total exchange; XM: crossmatch


Background
Obesity is a major public health problem that is increasing worldwide [1], and its prevalence in Korea has increased from 29.2% in 2001 to 34.6% in 2018 [2]. Obesity and metabolic syndrome are recognized as risk factors for the development and progression of chronic kidney disease (CKD) through various mechanisms [3,4]. In a previous nationwide cohort study in Korea, obesity (body mass index (BMI) ≥ 25 kg/m2) was identi ed as an independent risk factor for the progression of CKD [5]. Moreover, obesity also adversely affects the allograft function in kidney transplant (KT) recipients. A previous study reported that BMI ≥ 35 kg/m2 was an independent risk factor for death-censored graft loss (DCGL) and biopsy-proven acute rejection (BPAR) (hazard ratio (HR) 2.43 (95% con dence interval (CI), 1.48-3.99), HR 2.43 (95% CI, 1.07-5.51), respectively) [6]. Another previous study also reported that BMI ≥ 35 kg/m2 was an independent risk factor for overall graft loss (HR 1.22; 95% CI, 1.09-1.38) [7]. A meta-analysis also reported that recipient obesity had a marginally greater risk for DCGL (HR 1.06; 95% CI, 1.01-1.12) [8].
Meanwhile, pre-transplant sensitization to human leukocyte antigen (HLA) is a well-known risk factor associated with adverse allograft outcomes [9]. Pre-sensitization to HLA is not only associated with a high rate of acute antibody mediated rejection (ABMR), but also with the gradual development of chronic allograft tissue injury caused by activation of the humoral immune system. Indeed, the development of chronic antibody mediated rejection (cAMR) was signi cantly higher in patients with pre-sensitization to HLA compared to patients with low immunologic risk [10][11][12]. In addition, both pre-formed persistent donor speci c antibody (DSA) and pre-formed cleared DSA showed an increased risk of graft loss [9].
Another previous study also reported that preformed donor speci c anti-human leukocyte antigen antibody (HLA-DSA) with mean uorescence intensity (MFI) > 3000 had an increased risk for graft loss with a HR of 3.8 (P < 0.001) [13]. Therefore, both obesity and pre-transplant sensitization to HLA in recipients might contribute to the progression of chronic allograft tissue injury and adverse allograft outcomes, although the mechanisms are different. However, it has not been investigated whether both factors have a synergistic effect on allograft outcomes. Hence, we analyzed the short-term and long-term graft outcomes in kidney transplant recipients (KTRs) with high BMI and pre-sensitization to HLA, and investigated the interaction between high BMI and HLA pre-sensitization status in this study

Study Design
This study was a retrospective, observational and single-center study. Between January 2010 and December 2018, 1290 KTs were performed in Seoul St. Mary's Hospital. Of these, 412 cases received the kidney from a deceased donor, 195 cases were ABO incompatible KT, and 1 KTR had both legs amputated; these cases were excluded from the study. Finally, 682 KTRs were included in this analysis. The distribution of BMI of the KTRs is presented in Supplemental Figure S1 and the median BMI value was 22.7 kg/m 2 . Patients with BMI >22.7 kg/m 2 were categorized in the high BMI group, while others were classi ed in the low BMI group. The cases were de ned as pre-sensitized to HLA when the MFI value of HLA-DSA at baseline was higher than 3,000 [13] while if the value was below 3,000, they were de ned as non-sensitized. Based on the above classi cations, KTRs were divided into 4 groups (low BMI-nonsensitized, high BMI-non-sensitized, low BMI-sensitized, high BMI-sensitized) as presented in Figure 1.
This study followed the Declaration of Helsinki and was approved by the Institutional Review Board of Seoul St. Mary's Hospital (XC15RIMI0072K).
HLA Typing, HLA Antibodies, and DSA HLA typing and HLA antibodies were measured as described previously [14,15]. Brie y, HLA-A, B, DR, DQB1 typing was performed using the deoxyribonucleic acid molecular typing method using sequencespeci c oligonucleotide probes with LIFECODES HLA SSO typing kits (Immucor, Stamford, CT). Lifecodes LSA Class I and Class II kits (Gen-Probe Transplant Diagnostic Inc., Stamford, CT) or LABScreen Single Antigen (One Lambda Inc., A Thermo Fisher Scienti c Brand, Canoga Park, CA) were used for detecting HLA antibodies in the recipient sera. According to the manufacturer's instructions, 10 μL of each serum sample was used, and the uorescence intensities of the samples were measured using a Luminex 200 system (Luminex Corp., Austin, TX).
Desensitization Protocols for Pre-sensitized Patients The desensitization protocol in our center has been described previously [16][17][18]. Brie y, the desensitization protocol for HLA pre-sensitized patients consisted of rituximab, total plasma exchange (TPE) and intravenous immunoglobulin (IVIG). Two weeks to one month prior to the transplantation, rituximab was administered, and total plasma exchange using 5% albumin and fresh frozen plasma was performed seven times. The TPE frequency was controlled based on the MFI titer of HLA-DSA. IVIG was administered at a dose of 100 mg/kg over a period of 1 hour after every TPE. In all patients who underwent desensitization, prophylactic agents were used to prevent pneumocystis jirovecii pneumonia (PJP) and cytomegalovirus (CMV) infection. If the crossmatch (XM) test of T-cell complement dependent cytotoxicity (CDC) was positive or HLA-DSA was positive, and the MFI of HLA-DSA did not decrease adequately after 3 cycles of TPE, bortezomib based protocol was used, in which bortezomib was administered for four times in addition to the desensitization protocol.

Clinical Parameters and Outcomes
KTRs' age, height, weight, history of diabetes mellitus (DM) and hypertension (HTN), cause of end stage renal disease (ESRD), previous dialysis modality, previous dialysis period, and previous KT history were collected as baseline demographic characteristics. In addition, fasting glucose, total cholesterol, triglyceride, high density lipoprotein (HDL)-cholesterol, low density lipoprotein (LDL)-cholesterol, and hemoglobin A1c levels were obtained from the pre-transplant investigations. As a pre-transplant immunoassay, the results of XM test using CDC and ow cytometry-crossmatch (FCXM), HLA-DSA and MFI results by Luminex Single Antigen Assay, and panel reactive antibody (PRA) titers were obtained.
Transplantation information was collected for mismatch number, type of induction therapy, main immunosuppressant used, and drug used for desensitization.
We analyzed the incidence of BPAR within 1-year of transplantation (early acute rejection), CMV infection, BK viremia, and PJP rates as short-term clinical outcomes in the 4 groups. The variables used for analyzing long-term clinical outcomes comprised BPAR incidence after 1-year of transplantation (late acute rejection), cAMR, and biopsy-proven calcineurin inhibitor (CNI) toxicity rates. DCGL and patient death rates were also analyzed.
Allograft kidney biopsy ndings were interpreted according to Banff classi cation 2009. BPAR was diagnosed with allograft biopsy as suitable for acute T-cell medicated rejection (TCMR) and acute ABMR criteria according to Banff classi cation. Similarly, cAMR and biopsy proven CNI toxicity were diagnosed with allograft biopsies according to the Banff classi cation [19]. Death-censored allograft survival duration was de ned as the period from KT to dialysis or preemptive KT, except for patient death in a functioning allograft. Patient survival duration was de ned as the period from KT to death due to any cause. Data of changes in allograft function based on serum creatinine levels was collected until 4 years after KT.
The primary outcome of this study was to compare the impact of BMI on DCGL in HLA non-sensitized and sensitized patients. We compared the DCGL rate and death censored allograft survival duration in the 4 groups and analyzed the interaction between high BMI and HLA pre-sensitization status. Secondary outcomes of this study were early acute rejection, CMV infection, BK viremia, PJP, late acute rejection, cAMR, biopsy proven CNI toxicity, patient death rates, and estimated glomerular ltration rate (eGFR) based on Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI) equation [20].

Statistical Analysis
All continuous variables were expressed as mean ± standard deviation. If the variables followed the normal distribution, an analysis of variance (ANOVA) test was performed. If the variables showed nonnormal distribution, a Kruskal-Wallis test was performed. All categorical variables were compared using the chi-square test or Fisher's exact test and expressed as proportions. Multivariate Cox hazard regression model analysis was performed to determine the risk factors affecting DCGL and to investigate the interaction between high BMI and HLA pre-sensitivity. Cumulative survival rates were analyzed during the follow-up period in the 4 groups. Causes of DCGL and patient death were compared using the chi-square test or Fisher's exact test. The mean eGFR (CKD-EPI) and standard deviation in the 4 groups were evaluated. All statistical analyses were performed using the SPSS ® software, version 24 (IBM Corporation, Armonk, NY, USA) and Microsoft Excel 2016.

Results
Comparison of baseline clinical and laboratory parameters in the 4 groups according to BMI and HLA presensitization status The mean follow-up duration of the patients included in this study was 61.8 months. The baseline characteristics of the 4 groups are shown in Table 1. In terms of clinical demographics, the age at KT was signi cantly lower in the low BMI-non-sensitized group (44.8 years, P < 0.001, for all), and the proportion of male patients was higher in the non-sensitized groups, especially in the high BMI-non-sensitized group (235/319, 73.7%, P < 0.001). Average BMI and the percentage of patients with DM and HTN were higher in the high BMI groups than in the low BMI groups (P < 0.001). Mean fasting glucose, triglyceride level, and hemoglobin A1c level was signi cantly higher (P < 0.001, P = 0.001, and P < 0.001, for all), and HDLcholesterol level was lower in both high BMI groups (P < 0.001, for each) compared to that in the low BMI groups. Percentage of patients with previous KT history was highest in the low BMI-sensitized group (13/40, 32.5%, P < 0.001), and the percentage of patients in whom DM was the cause of ESRD was signi cantly higher in both high BMI groups (P < 0.001). Of the types of dialysis modality, the percentage of peritoneal dialysis was highest in the high BMI-non-sensitized group (47/319, 14.7%, P = 0.016). In the baseline immunologic test, the number of patients with positive XM test was signi cantly higher in the HLA sensitized groups than in the non-sensitized groups (P < 0.001); however, no difference in the rate of positivity was seen between patients with low BMI and high BMI within the HLA sensitized groups (P = 0.396). Use of anti-thymocyte globulin as induction therapy and percentage of desensitization therapy were signi cantly higher in the HLA sensitized groups compared to that in the non-sensitized groups (P < 0.001, for each).
Comparison of the short-term clinical outcomes, long-term clinical outcomes, and allograft functions among the 4 groups according to BMI and pre-sensitization status Table 2 shows the short-term and long-term clinical outcomes in the 4 groups. In the short-term clinical outcomes, the rate of early ABMR was signi cantly higher in the HLA sensitized groups (P < 0.001) but there was no signi cant difference between those with low and high BMI within the sensitized groups. CMV infection and BK viremia rates tended to be higher in the HLA sensitized groups; however, the difference was not signi cant (P = 0.110, P = 0.208, respectively). Early TCMR and PJP rates did not show statistically signi cant differences between the 4 groups (P = 0.458). In long-term clinical outcomes, the rate of late ABMR tended to be highest in the high BMI-sensitized group although the difference was not signi cant (2/21, 9.5%, P = 0.154). There were no signi cant differences in the rates of late TCMR, cAMR, and biopsy-proven CNI toxicity between the 4 groups. Categorical variables are shown as proportions. * P < 0.05 versus low BMI-non-sensitized group, † P < 0.05 versus high BMI-non-sensitized group, ‡ P < 0.05 versus low BMI-sensitized group, § P < 0.05 versus high BMI-sensitized group.
BMI: body mass index, ABMR: acute antibody mediated rejection, TCMR: T-cell mediated rejection, CMV: cytomegalovirus, PJP: pneumocystis jiroveci pneumonia, CNI: calcineurin inhibitor Figure 2 shows the change in mean eGFR during the follow-up period after transplantation in the 4 groups. Although eGFR did not show statistically signi cant differences between the 4 groups, the decline in eGFR was higher in the high BMI-sensitized group compared to that in the other 3 groups after 1 year of KT. Moreover, in this group the eGFR was less than 60 ml/min/1.73 m 2 after 3 years of KT, while it was more than 60 ml/min/1.73 m 2 in the other 3 groups after 4 years of KT.
Comparison of DCGL rate and its causes in the 4 groups according to BMI and pre-sensitization status In all, 44 DCGL events occurred; of these, 14 were in the low BMI-non-sensitized group (14/302, 4.6%), 22 in the high BMI-non-sensitized group (22/319, 6.9%), and 4 each in the low BMI-sensitized group (4/40, 10.0%), and high BMI-sensitized group (4/21, 19.0%) ( Table 3). There were no statistically signi cant differences in the causes of DCGL between the 4 groups. In the cumulative death-censored allograft survival rate curve, the graft survival rate decreased signi cantly in the sensitized groups and it was poorest in the high BMI-sensitized group (Fig. 3)  Categorical variables are shown as proportions. * P < 0.05 versus low BMI-non-sensitized group, † P < 0.05 versus high BMI-non-sensitized group, ‡ P < 0.05 versus low BMI-sensitized group, § P < 0.05 versus high BMI-sensitized group.  In all 21 patients died; however, there was no signi cant difference in the rate of death between the 4 groups; 8 deaths occurred in the low BMI-non-sensitized group (8/302, 2.6%), 12 in the high BMI-nonsensitized group (12/319, 3.8%), 1 in the low BMI-sensitized group (1/40, 2.5%), and no patient death occurred in the high BMI-sensitized group (Table 3). When comparing the causes of death between the groups, there were no signi cant differences (Table 4).

Discussion
In this study, we analyzed the short-term and long-term allograft outcomes according to high BMI and HLA pre-sensitization status of the recipients. We found that the high BMI-sensitized group showed the most signi cant decline in allograft function among 4 groups, and had the highest DCGL rates. In addition, high BMI-sensitized status was found to be an independent risk factor for DCGL. There was also a signi cant interaction between high BMI and pre-transplant HLA sensitivity. Our results suggest that increased immunologic risk and high BMI might induce synergistic effects leading to adverse allograft outcomes.
We divided the high BMI and low BMI groups based on the median BMI value of 22.7 kg/m 2 of the patients included in this study. According to the World Health Organization's International Obesity Task Force recommendation, when analyzing co-morbidities risk, the Asian population sets the BMI value of obesity to 25 kg/m 2 and the BMI value of overweight to 23 kg/m 2 [21]. The median BMI value in this study, 22.7 kg/m 2 , was approximately close to the Asian population overweight cutoff value, and most patients in the high BMI group met the Asian overweight criteria.
In the comparison of baseline characteristics, the rates of DM and HTN as a comorbidity were higher in the high BMI groups, and laboratory parameters such as fasting glucose, triglyceride, HDL-cholesterol, and hemoglobin A1c also showed suitable differences as metabolic syndrome as expected [22]. As previously reported, the rate of previous KT history was higher in the sensitized groups [23]. Lastly, the rate of peritoneal dialysis prior to transplantation was signi cantly higher in the high BMI group than in the low BMI group. It is thought that weight gain occurs due to glucose absorption from the peritoneal dialysate in patients undergoing peritoneal dialysis [24].
In the comparison of short-term outcomes, the rates of CMV infection and BK viremia tended to be higher in the HLA pre-sensitized groups. These could be due to immunosuppression caused by desensitization therapy despite adequate prophylaxis in the sensitized groups. In a nationwide cohort study in Korea, which we published previously, desensitization therapy has also been found to be a signi cant risk factor for infection related mortality [25]. Early ABMR rates were higher in the HLA pre-sensitized groups, and the rates of late ABMR also tended to be higher in the sensitized groups. These results were consistent with previous studies, which showed that acute rejection rate is higher despite appropriate desensitization therapy in those with HLA pre-sensitization [17,25].
In the comparison of long-term outcomes, a signi cant decline in the allograft function was seen in the high-BMI-pre-sensitized group. In addition, adjusted allograft DCGL showed the poorest outcome in this group, and high BMI and HLA pre-sensitivity showed signi cant interaction. These could be due to several reasons. First, the nephron mass of the donated kidney might be relatively inadequate for the high BMI recipients. Simply put, it is possible that a physiologic mismatch between the recipient's metabolic demand and the donated kidney's nephron mass has an adverse allograft outcome [26]. Brenner et al. have previously proposed a nephron underdosing theory for chronic allograft failure. They suggested that the transplantation of kidneys with a relatively large nephron mass compared to the recipients' metabolic demand might lead to tolerance for future immunologic challenges, ischemic events, and CNI toxicities [27]. Hence, in high BMI recipients with high metabolic demand, kidneys with relatively small nephron mass might be more susceptible to immunologic damage when accompanied by HLA pre-sensitivity with immunologic risk, resulting in poor allograft outcomes.
Second, obesity is known to induce alloimmune dysregulation by decreasing the adiponectin level and increasing the leptin level. Adiponectin is known to inhibit B-cell lymphopoiesis, macrophage activation, Tcell proliferative response, and responses of helper T-cell (Th)-1 and Th-2 [28], and it has been reported that serum adiponectin levels decrease with visceral obesity [29]. In contrast, leptin is known to increase T-cell response, thymocyte survival signal, and activities of neutrophils, macrophages, and natural killer cells [28], and it has been reported to be correlated with adipose tissue mass [30]. In addition, obesity is known to induce metabolic in ammation [31,32], leading to atherosclerosis, vascular injury, and related comorbid conditions [33][34][35]. Especially in patients with high BMI-sensitized status, the presence of HLA-DSA might be an important factor in triggering the immune reaction. These might account for the high rate of late ABMR in the high BMI-sensitized group in our study.
Although BMI did not show a statistically signi cant impact on acute rejection, graft loss due to acute and chronic rejection was 75.0% in the high BMI-sensitized group, which was higher than that in the low BMI-sensitized group (50.0%) in the analysis of DCGL causes. In addition, in the high BMI-non-sensitized group, graft loss due to acute and chronic rejection was 50.0%, which was higher than that in the low BMI-sensitized group (28.6%). These results possibly indicate that when a rejection event occurs in a kidney with a relatively small nephron mass, the number of remaining functional nephrons might be smaller than in the kidney with a large nephron mass, leading to allograft failure.  Table S1). In addition, this study is important because this is the rst paper to analyze the relationship between the recipient BMI and HLA pre-sensitization status. The second limitation of the study is that it focuses on the recipients' condition before KT and did not analyze whether the patients developed obesity after KT. After KT, prednisolone and other immunosuppressive drugs are administered which can lead to weight gain. It is not possible to rule out that obesity developed after KT affects allograft outcomes.

Conclusions
High BMI and HLA sensitization before KT synergistically affect the long-term allograft outcomes in terms of the decline in allograft function and allograft survival in KT recipients. Our results suggest that active reduction and careful monitoring of BMI might be necessary to improve allograft outcomes especially in patients with high immunologic risk.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by a grant from the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (No. HI20C0317). This manuscript is a researchers-led paper with no speci c participation by funders.

Author Contributions
YP and BHC developed the study ideas; HL, EJK, and SL contributed to data collection; THB, JWM, HEY and EJO contributed to literature review and data interpretation; CWY contributed to data validation and analysis; YP contributed to generation of tables and gures; YP and BHC analyzed the data and wrote the manuscript, and all authors have read and approved the nal manuscript.  Cumulative graft survival rate curve among the 4 groups according to BMI and HLA pre-sensitization status for death censored graft loss Cumulative graft survival rates were signi cantly lower in high BMIsensitized and low BMI-sensitized groups compared to that in the low BMI-non-sensitized group (vs. low BMI-non-sensitized; high BMI-non-sensitized, HR