Clinical characteristics of new‐onset diabetes after liver transplantation and outcomes

Abstract Background We aimed to identify the characteristics of new‐onset diabetes after liver transplantation (LT) (NODAT) and investigate its impacts on post‐transplant outcomes. Methods Adult LT patients between 2014 and 2020 who used tacrolimus as initial immunosuppression and survived 3 months at least were evaluated. Patients who developed NODAT within 3 months after LT were classified as NODAT group. Also, patients were further classified as history of diabetes before LT (PHDBT) and non‐diabetes (ND) groups. Patient characteristics, post‐LT outcomes, and cardiovascular and/or pulmonary complications were compared. Results A total of 83, 225, and 263 patients were classified into NODAT, PHDBT, and ND groups. The proportion of cholestatic liver disease and rejection within 90 days were higher in NODAT group. Mean serum tacrolimus concentration trough level in the first week after LT was 7.12, 6.12, and 6.12 ng/mL (p < 0.001). Duration of corticosteroids was significantly longer in NODAT compared to PHDBD or ND (416, 289, and 228 days, p < 0.001). Three‐year graft and patient survival were significantly worse in NODAT than ND (80.5% vs. 95.0%, p < 0.001: 82.0% vs. 95.4%, p < 0.001) but similar to PHDBT. Adjusted risks of 3‐year graft loss and patient death using Cox regression analysis were significantly higher in NODAT compared to ND (adjusted hazard ratio [aHR] 3.41, p = 0.004; aHR 3.61, p = 0.004). Incidence rates of cardiovascular or pulmonary complications after LT in NODAT were significantly higher than ND but similar to PHDBT. Conclusion Higher initial tacrolimus concentration and early rejection might be risk factors for NODAT. NODAT was associated with worse post‐transplant outcomes.


| INTRODUC TI ON
In the general population, the prevalence of diabetes is estimated to be up to 4%. 1 Meanwhile, new-onset diabetes after transplantation (NODAT) is a frequent comorbidity for patients who received solid organ transplantation, which is estimated to be up to 7%-28% in patients after liver transplantation (LT). 2 Steroids increase insulin resistance and gluconeogenesis, 3 and calcineurin inhibitors impair insulin secretion from pancreatic βcells. 4Therefore, steroids and calcineurin inhibitors were well-known as risk factors for NODAT. 5,6 addition, various risk factors such as older age, 7 hepatitis C virus (HCV), 8 and acute cellular rejection (ACR) 9 were reported as risk factors for NODAT in other studies.
Diabetes is a risk factor for cardiovascular events 10 and/or infectious diseases. 11In patients who received LT, a history of diabetes increases the risk of mortality following LT, 12 which were often related to atherosclerotic vascular events 13 and/or end-stage kidney disease. 146][17]  Kuo et al. reported   that a history of diabetes was associated with mortality and graft failure but not NODAT. 15According to the large Korean multicenter study, graft survival rates were similar regardless of NODAT. 16Other reports showed that patients with NODAT had reduced survival and an increased incidence of sepsis and chronic renal insufficiency. 17e aim of this study is to identify risk factors for NODAT, and to investigate the impact of the NODAT on graft and patient survival, and cardiovascular events compared to patients who didn't show diabetes or patients who had diabetes before LT.

| Study population
Henry Ford Health (HFH) is an integrated tertiary care center in metropolitan Detroit, Michigan, US.Study protocols were approved by the HFH Institutional Review Board (#15051); requirements for written informed consent were waived due to the de-identified and observational nature of data.Retrospective medical records data were collected for patients who received a liver transplant (LT) between January 2014 and December 2020.Adult patients (≥18 years) who used tacrolimus as initial immunosuppression and survived 3 months at least were eligible for inclusion.Patients who received retransplant or combined transplants with thoracic organs, intestine, and/or pancreas were excluded.Three patient who experienced intraoperative death was excluded.Patients who developed new-onset diabetes after LT (NODAT) within 3 months after LT were classified as NODAT group.Also, patients were further classified as previous history of diabetes before LT (PHDBT) and non-diabetes (ND) groups (Figure 1).

| The definition of NODAT
The definition of NODAT is as follows 18

| Comparison of patient characteristics
Patient characteristics were compared among the three groups (NODAT vs. PHDBT vs. ND).These were also compared between the two groups (NODAT vs. ND).Multivariable analysis was performed to identify the risk factors for NODAT using logistic regression after univariable analysis.

| Comparison of post-LT outcomes
Three-year graft and patient survival after LT were compared among the three groups.We performed multivariable analyses of risk factors for post-LT 3-year graft loss and patient death.Cardiovascular events, pulmonary complications, and/or the proportion of patients who could stop the medication for diabetes after LT were also compared.

| Statistical analysis
Patient and donor demographic and clinical characteristics were described by the groups, using median and interquartile range (IQR) for continuous variables and numbers and percentages for categorical variables.Continuous variables were compared with the Mann-Whitney U test and categorical variables were compared using the chi-square test.Logistic regression was used for the multivariable analysis to identify the risk factors for NODAT.
Post-transplant graft and patient survival were evaluated using the Kaplan-Meier method and groups were compared using logrank tests.A multivariable Cox regression model assessed hazards of post-transplant graft loss and patient death using factors which had p value <0.157 in univariable analyses. 19p-values <0.05 were considered statistically significant for all analyses.All statistical analyses were completed using SPSS version 27 (IBM, Chicago IL, USA) and R version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria).

TA B L E 1
Comparison of recipient and donor characteristics among patients stratified by the diabetes status.4).NODAT also had a significantly higher proportion of patients who could stop the medication for diabetes after LT compared to PHDBT (51.8% vs. 6.7%, p < 0.001) (Table 5).The cause of 3-year mortality was shown in Table S1.

| DISCUSS ION
In our series, 15% of LT patients developed NODAT, who showed a higher proportion of cholestatic liver disease as their primary liver disease compared to those without diabetes.Multivariable analysis showed tacrolimus trough in the first week, higher HbA1c within normal limit, and early rejection after LT were associated with NODAT.After risk-adjusted analysis, NODAT was a risk factor for graft loss and patient death similar to PHDBT.Of note, NODAT increased risks of post-LT cardiovascular and/or pulmonary complications.
According to a large meta-analysis by Chin et al., the incident rates of NODAT at 3, 6, and 12 months after LT were 15.5%, 16.1%, and 18.3%, respectively. 20It was consistent with our results.It was known that tacrolimus inhibits insulin-mediated inactivation of hepatic glycogenolysis, causes a reduction in human pancreatic ductal cells, and inhibits glucose-stimulated insulin secretion. 21 Meta-analysis using 11 randomized controlled trials showed tacrolimus to be superior to ciclosporin in terms of patient mortality and hypertension, while ciclosporin was superior in terms of NODAT. 24 this study, five patients who received ciclosporin as an initial immunosuppression were excluded from this study, because the limited number of patients with ciclosporin did not allow comparisons with those with tacrolimus.

F I G U R E 2
Comparison of post-LT outcome among the three groups stratified by the diabetes status.(A) Three-year graft survival rate in NODAT was significantly lower than in ND (80.5% vs. 95.0%,p < 0.001) but similar to those in PHDBT (80.5% vs. 86.1%,p = 0.30).(B) Threeyear patient survival rate in NODAT was significantly lower than in ND (82.0% vs. 95.4%,p < 0.001) but similar to those in PHDBT (82.0% vs. 87.9%,p = 0.20).
Instead, we focused on the possible effects of initial tacrolimus trough levels on the incidence of NODAT.Our study showed that tacrolimus trough in the first week was an independent risk factor for NODAT, which concurred with the findings from other studies.Song et al. calculated the mean trough concentration of tacrolimus in the year of diabetes diagnosis patients with NODAT or in the last year of the follow-up in patients without NODAT. 25They reported that the mean tacrolimus of patients with NODAT was significantly higher than that of patients without NODAT and maintaining a tacrolimus value below 5.89 ng/mL after LT decreased risks for NODAT. 25 Another important finding of this study was that an early rejection after LT was an independent risk factor for NODAT, which was consistent with the previous study. 9Patients with NODAT had a higher proportion of patients who received corticosteroids for a longer time.This might be related to a higher proportion of patients with cholestatic disease as their primary liver disease in the NODAT group.
Corticosteroids have diabetogenic effects, which are insulin resistance and increased hepatic gluconeogenesis. 27Previous reports showed that withdrawal of glucocorticoids after LT might reduce the risk of NODAT. 16,27It was reported that basiliximab might decrease risks for NODAT due to a decrease of steroids or a dose of tacrolimus. 16,17so, higher HbA1c even within normal limit was an independent risk factor for NODAT, and patients with NODAT showed a higher proportion of severe encephalopathy in univariable analysis although not significant in multivariable analysis compared to patients with NODAT to those without diabetes.Their higher HbA1c might be associated with glucose intolerance before LT.According to the study of 2248 patients who had received LT without pretransplant diabetes based on the National Health Insurance Research Database of Taiwan, encephalopathy was an important preoperative risk factor for NODAT (aHR, 1.54). 28Regarding other risk factors for NODAT, several previous studies have reported male, 29 older age, 16,26,30 family history, 29 HCV infection, 29 NASH, 22,30 high BMI (obesity), 16,29 and graft volume, 16 which were controversial and not consistent in those reports.In our study, BMI, older age, and NASH were significant factors associated with PHDBT but not with NODAT.
We also found that NODAT was a risk factor for post-LT mortality similar to PHDBT and increased risks of post-LT cardiovascular and/or pulmonary complications.LV et al. reported that patients with NODAT showed higher mortality (mortality rate 40% vs. 22%) and an increased incidence of bacterial infection, and chronic renal insufficiency compared to patients without NODAT. 17They also demonstrated that the proportion of lung infection and multiple organ failure were more frequent in the NODAT group as a cause of death after LT, 17 which was consistent with our study.Conversely, other studies reported that NODAT was neither associated with post-LT mortality 15,16 nor post-LT complications. 15According to the study by Moon et

TA B L E 6
Risks for 3-year graft loss and patient death after liver transplantation in patients with NODAT.
achieve remission of type 2 diabetes, but this efficacy was less likely with longer durations of contracting diabetes. 32Our study showed that patients with NODAT were more likely to discontinue diabetic therapies.Future studies would be warranted to elucidate the reversibility of NODAT.
Regarding the poor prognostic factor among the patients who developed NODAT, rejection within 90 days after LT was a significant risk factor for both graft loss and patient death.In all cohorts including PHDBT and ND, interestingly, rejection within 90 days after LT was not a significant prognostic factor in this study.The patients with NODAT who have experienced rejection within 90 days after LT might need close follow-up after LT.
There are several limitations in our study.This is a retrospective, single-center analysis with a small sample size.Consequently,

DR. SHINGO SHIMADA
I think it's very important, but it might be difficult.In this study, rejection was not a significant covariate for graft and patient survival in univariable analysis.Although the rejection at 30 days and at 90 days after liver transplant were only evaluated in this study, we did not include rejection in multivariable analysis.
Thus, we confirmed by bivariable analysis, including NODAT and rejection.
NODAT was associated with worse graft and patient survival compared to non-diabetes, which means the risk was adjusted for acute rejection, and NODAT remained as an independent factor.However, any specific pathological findings were not evaluated in this study.
Instead, we checked that the cause of patient death showed a higher incidence of infection in NODAT patients.

DR. YUKIKO KOSAI-FUJIMOTO
I want to ask you about the screening for the NODAT.In your slides, you showed that there are some diagnosis criteria and there are things like blood sugar and HbA1c.When do you measure those factors in your outpatient clinic?Because especially in the early stage, it seems there are more hurdles to check all the factors in every visit of the patients.

DR. SHINGO SHIMADA
Most patients come to our hospital after discharge, at 1 month, 3 months, and 6 months after liver transplant.During visits to our hospital, the family medicine or hepatologist does a close follow-up.
So, we can check those factors at the outpatient clinic.
: 1) two posttransplant fasting plasma glucose levels ≥126 mg/dL ≥30 days apart; 2) oral hypoglycemic agent use for ≥30 consecutive days after transplantation; 3) insulin therapy for ≥30 consecutive days after transplantation; 4) hemoglobin A1c (HbA1c) ≥6.5% on at least one occasion after F I G U R E 1 Flow chart of study population selection.transplantation.Patients who have at least one of four criteria are diagnosed as NODAT.2.3 | Covariates Categorical variables included: recipient/donor gender; recipient/ donor race (White, Black, Hispanic, other); etiology of end-stage liver disease (hepatitis B virus [HBV], hepatitis C virus [HCV], nonalcoholic steatohepatitis [NASH], cholestatic disease, alcoholrelated liver disease); Karnofsky score at LT (10-30%, 40%-60%, or 70%-100%); presence of severe/moderate grade ascites at LT (y/n); grade III/IV encephalopathy at LT (y/n); dialysis at LT (y/n); mechanical ventilation at LT (y/n); presence of hepatocellular carcinoma (HCC) at LT (y/n); recipient family history of diabetes within second degree relative (y/n); type of liver graft (whole or partial/split); use of donation after circulatory death (DCD) donor liver graft (y/n); donor cause of death (trauma, anoxia, cerebrovascular accident [CVA], or other); rejection within 90 days after LT (biopsy proven) (y/n); steroid pulse treatment within 90 days after LT (y/n); reoperation within 30 days after LT (y/n); and readmission within 30 days after LT (y/n); patients who could stop corticosteroids (y/n).Recipient/donor age at LT, recipient/donor body mass index (BMI) at LT, recipient white blood cell (WBC) at LT, hemoglobin (Hb) at LT, platelet (Plt) at LT, total cholesterol (T-cho) at LT, HbA1c at LT, albumin (Alb) at LT, recipient model for end-stage liver disease (MELD) score at LT, recipient warm ischemia time (WIT), recipient cold ischemia time (CIT), amount of blood loss at LT, operative time at LT, amount of intraoperative transfusion (red blood cell [RBC], fresh frozen plasma [FFP], platelet concentrates [PC]), hospital stay days after LT, ICU stay days after LT, mean tacrolimus trough in the first week after LT, duration of corticosteroids after LT were used as continuous variables.
we could not adjust patient backgrounds among groups and could not evaluate the dose effects of corticosteroids after LT because the dose of corticosteroids might fluctuate during post-LT course due to increase or decrease, and steroid pulse treatment.Also, the duration of NODAT was not evaluated.Despite these limitations, this study provides important insights into the risk stratification for NODAT and its impact on post-LT outcomes.In conclusion, NODAT was associated with worse posttransplant outcomes.Since high initial tacrolimus concentration and episodes of early rejection were considered as risk factors for NODAT, careful immunosuppression management would be important to decrease its risk.Recently, LT patient populations have become older and the number of patients with NASH has been increasing.Because risks of cardiopulmonary complications were higher in patients with NODAT, pretransplant assessments and risk stratifications for possible underlying modalities in those patients would be crucial.Further investigations regarding the long-term disease course of NODAT would be warranted to better understand its prognosis in LT patients.biopsy-proven rejection during the course.We are also aware that steroid dosage should affect NODAT.We can evaluate the base daily dosage of steroids.However, in some cases, additional dosage is given due to rejection or other reasons.It was difficult to evaluate the total dosage of steroids objectively.PROFESSOR TAKEAKI ISHIZAWAI would like to congratulate Dr. Shimada for tackling this important issue in the current situation of liver transplantation, where lifetime of post-transplant patients is getting longer and longer.Regarding your presentation, I am a little concerned about the overlaps between the development of early acute rejection and postoperative diabetes because the early acute rejection is known as an independent risk factor.Rejection followed by the strong antiimmunotherapy is a possible risk factor for development of diabetes and also poor prognosis.Is it possible to show us any findings which could suggest a direct association of NODAT with a poor postoperative graft or patient survival, for example, the pathological proof of the fat infiltration on the transplanted liver or something?Or was the other factor [something] like severe infection rather than liver function associated with the poor postoperative survival?
Comparison of complications after liver transplantation.
26gi et al. reportedthat a tacrolimus trough level ≥8 ng/mL 3 months after LDLT was an independent risk factor for NODAT in the multivariable analysis.26Abbreviations:aHR,adjustedhazard ratio; CI, confidence interval; CVA, cerebrovascular accident; HCC, hepatocellular carcinoma; ICU, intensive care unit; LT, liver transplant; MELD, model for endstage liver disease; ND, non-diabetes; NODAT, new-onset diabetes after transplantation; PHDBT, prior history of diabetes before transplantation.TA B L E 3 Risks for 3-year graft loss and patient death after liver transplantation.TA B L E 4