International consensus on post-transplantation diabetes mellitus

ABSTRACT Post-transplantation diabetes mellitus (PTDM) remains a leading complication after solid organ transplantation. Previous international PTDM consensus meetings in 2003 and 2013 provided standardized frameworks to reduce heterogeneity in diagnosis, risk stratification and management. However, the last decade has seen significant advancements in our PTDM knowledge complemented by rapidly changing treatment algorithms for management of diabetes in the general population. In view of these developments, and to ensure reduced variation in clinical practice, a 3rd international PTDM Consensus Meeting was planned and held from 6–8 May 2022 in Vienna, Austria involving global delegates with PTDM expertise to update the previous reports. This update includes opinion statements concerning optimal diagnostic tools, recognition of prediabetes (impaired fasting glucose and/or impaired glucose tolerance), new mechanistic insights, immunosuppression modification, evidence-based strategies to prevent PTDM, treatment hierarchy for incorporating novel glucose-lowering agents and suggestions for the future direction of PTDM research to address unmet needs. Due to the paucity of good quality evidence, consensus meeting participants agreed that making GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) recommendations would be flawed. Although kidney-allograft centric, we suggest that these opinion statements can be appraised by the transplantation community for implementation across different solid organ transplant cohorts. Acknowledging the paucity of published literature, this report reflects consensus expert opinion. Attaining evidence is desirable to ensure establishment of optimized care for any solid organ transplant recipient at risk of, or who develops, PTDM as we strive to improve long-term outcomes.


INTRODUCTION
Post-transplantation diabetes mellitus (PTDM) significantly contributes to morbidity and mortality after solid organ transplantation (SOT).The last International PTDM Consensus Meeting in 2013 consolidated heterogenous clinical practice and suggested standards of care for the screening, diagnosis and management of PTDM [1 ].However, the PTDM field has evolved dramatically since 2013, justifying an update.Research has enhanced our understanding, while expanded therapeutic options in the general population have dramatically shifted treatment algorithms.In this rapidly changing climate, ambitions to improve long-term SOT outcomes require optimized strategies to prevent/manage PTDM that are aligned with the latest scientific updates.
This Meeting Report summarizes proceedings from the 3rd International PTDM Consensus Meeting held in Vienna, Austria, from 6-8 May 2022.The meeting was endorsed by the European Renal Association (Diabesity Working Group) and the European Society for Organ Transplantation (EKITA Working Group).An international expert panel was convened by invitation, comprising 18 transplant clinicians, diabetologists and scientists with an active interest in the field, to deliberate updates to the previous consensus statement relevant for contemporary clinical practice.Invitations were based upon a meeting prerequisite to systematically review existing literature for presentation at open scientific sessions, encouraging debate and discussion.While targeting all SOT recipients, published data are kidney-centric and organspecific considerations are required.After reviewing and reflecting upon the paucity of good quality evidence, consensus opinion agreed that making GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) recommendations would be flawed [2 ].Therefore, our terminological use of 'Opinion Statement' is deliberate to acknowledge this.This position statement reflects the consensus view of expert delegates.Ultimately, attaining this evidence is desirable to ensure establishment of optimized care for any solid organ transplant recipient at risk of, or who develops, PTDM as we strive to improve long-term outcomes.

OPINION STATEMENT 1: PERFORM AN ORAL GLUCOSE TOLERANCE TEST FOR DIAGNOSIS AND SCREENING; START ON THE WAITING LIST
Glucose thresholds for defining diabetes in the general population are based on the probability of developing retinopathy [3 ], but only one study explores this issue post-transplantation [4 ].An oral glucose tolerance test (OGTT) is essential for diagnosis and screening (see Supplementary data, Table S1), as alternatives like haemoglobin A1c (HbA1c) lack diagnostic sensitivity [5 -7 ] and association with adverse outcomes [1 , 8 , 9 ].Patients with impaired glucose tolerance (IGT), exclusively diagnosed by OGTT, or PTDM are at risk for cardiovascular disease [9 ] and premature death [1 , 8 ].Importantly, OGTTs allow earlier identification of at-risk individuals on the waiting list [10 ].When diagnosed early or by 2-h postprandial glucose only, PTDM may have greater chance of reversibility, although this may reflect low reproducibility [11 ].Supplementary data, Table S2 summarizes the published evidence.
Long-term evolution of PTDM is characterized by metabolic variability [7 , 11 , 12 ].Individuals with prediabetes (impaired fasting glucose and/or IGT) or PTDM risk factors will benefit from re-peated (e.g.annual) OGTT testing.If diagnosed early (e.g. 3 months post-operatively), PTDM may need later confirmation.A diagnosis and screening algorithm is proposed (Fig. 1 ) but warrants validation for improvement of outcomes.

OPINION STATEMENT 2: BE AWARE OF LONG-TERM CONSEQUENCES OF PREDIABETES AND PTDM
PTDM is associated with overall graft loss [13 ], cardiovascular events [8 , 14 ] and all-cause mortality [8 ], while microvascular complications are less studied [4 ] and patient-reported outcomes are scarce.Some studies observe no association with patient/graft survival [15 , 16 ], but this discrepancy might be influenced by heterogenous cohorts, diagnostic criteria or methodological differences.Importantly, the association of prediabetes with mortality and cardiovascular events should be appreciated [9 ].Other longterm consequences of PTDM require evaluation.For example, diabetes is associated with several cancers in the general population but data with PTDM are limited.A recent cohort analysis has observed an association between PTDM and future renal cell cancer [17 ], consistent with observations from a Danish cohort describing increased risk for cardiovascular and cancer-related mortality in SOT recipients with pre-transplant diabetes or PTDM [18 ].

OPINION STATEMENT 3: PRIORITIZE CLINICAL ATTENTION TO 'AT RISK' GROUPS
SOT recipients are at risk for the development of prediabetes/PTDM, but certain patients have a disproportionately higher risk.Early identification of this high-risk group is crucial to ensure that resources are directed to the most vulnerable, who may be amenable to intervention.This 'at-risk' group can be classified by clinical phenotypes or novel risk prediction methods like polygenic risk scores (PRS).The latter estimates an individual's genetic liability for a specific disease according to their genotypic profile and has been studied after liver and kidney transplantation [19 ].PRS are associated with pre-transplant type 2 diabetes and post-surgery PTDM.PRS in liver donors, but not kidney donors, was an independent risk factor for PTDM development and a combined liver donor/recipient PRS improved PTDM prediction over-and-above a clinical variable model alone.Further research is recommended to identify the optimal way to identify at-risk groups.

OPINION STATEMENT 4: CONSIDER UNDERLYING PATHOMECHANISM OF PTDM DEVELOPMENT AND THE INTER-RELATIONSHIP BETWEEN β-CELL DYSFUNCTION AND METABOLIC STRESS
PTDM arises from an interaction between pre-transplant and post-transplant risk factors ( Supplementary data, Fig. S1).Many pre-transplant risk factors are common to type 2 diabetes (i.e.obesity, metabolic syndrome), but immunosuppression is the most important post-transplant risk factor.Pre-transplant risk factors may identify individuals at risk from immunosuppressioninduced β-cell toxicity amenable to intervention, supporting the use of waiting-list screening.
Mechanistically a combination of pancreatic β-cell dysfunction and insulin resistance are predisposing factors for PTDM, with superimposed immunosuppression accelerating preexisting damage [20 ].A mechanistic approach is depicted in Supplementary data, Fig. S2 according to an animal model of calcineurin inhibitor (CNI)-induced toxicity, potentiating similar cellular damage induced by obesity and insulin resistance, which indicates common pathways in β-cell dysfunction [20 ].Importantly, this principle has been corroborated with slightly different pathways in human islets and pancreas transplant biopsies [21 ].Tacrolimus induces β-cell damage provoked by the glucolipotoxicity state secondary to multi-factorial insults, pathogenic pathways [e.g.mammalian target of rapamycin (mTOR) pathway] [22 ] responsible for β-cell maintenance and function [20 ].Furthermore, low-grade inflammatory stress is associated with early occurrence of PTDM [23 ] and early post-transplant mortality in general [24 ].Thus, a 'two-hit' hypothesis combining transplantation-induced β-cell insult on a background of metabolic stress converging in a dysfunctional synergy is an attractive hypothesis for the development of prediabetes/PTDM.However, other confounders must not be overlooked.For example, Halden et al .demonstrated infusion of the incretin hormone glucagon-like peptide 1 (GLP-1) during fasting and hyperglycaemic conditions in patients with PTDM compared with normal glucose tolerance, rectified pathophysiological defects like hyperglucagonemia, and diminished first-and second-phase insulin secretion [25 ].

OPINION STATEMENT 5: CHOOSE AN IMMUNOSUPPRESSION REGIMEN FOR OPTIMIZATION OF PATIENT AND GRAFT SURVIVAL
Despite the association between immunosuppression and PTDM, de novo regimens should not be routinely modified to reduce PTDM risk or adjusted after PTDM development.However, for selected patients, tailored immunosuppression may be justified if development of diabetes outweighs other risks.Patient-specific factors, immunological considerations and competing risks must all be factored when choosing immunosuppression on a personalized basis.
No robust data link induction therapy directly to PTDM risk.However, lymphocyte-depletion therapies (e.g.thymoglobulin, alemtuzumab) can facilitate lower exposure to maintenance CNIs and steroids which can reduce PTDM risk.
Regarding CNIs, Torres et al .randomized 128 de novo kidney transplant recipients (KTRs) at high-risk for PTDM but low immunological risk to: (i) tacrolimus and rapid steroid withdrawal, (ii) cyclosporine and steroid maintenance, or (iii) tacrolimus with steroid maintenance [26 ].All arms received basiliximab and steroids.Patient/graft survival and graft function were similar between study arms, with tacrolimus and steroid maintenance providing the best balance between risk for PTDM versus acute rejection.There is limited evidence supporting conversion of CNI in established PTDM.In a randomized controlled trial (RCT) involving 87 KTRs, conversion from tacrolimus to cyclosporine significantly improved glycaemic control with no increased risk for acute rejection [27 ].Late changes to immunosuppressive regimens may alleviate PTDM but this requires further evaluation to ensure glycaemic benefits outweigh long-term allograft risks.There is not enough evidence to support using different tacrolimus formulations, such as immediate versus prolonged release, but results from ongoing studies are awaited (see Supplementary data, Table S3).
Belatacept has a favourable metabolic risk profile, including less PTDM [28 ], in comparison with CNIs and different regimens have been explored in RCTs including KTRs [29 ].Belatacept is an acceptable alternative to CNIs to reduce PTDM in low immunological-risk patients if logistical and cost implications are surmountable.Any studies to explore efficacy in non-renal SOT recipients should ensure data capture of PTDM as a secondary outcome.
Although mTOR inhibitors are diabetogenic, incidence of PTDM is not significantly increased by their use which may reflect reduced CNI exposure.A recent meta-analysis evaluating the combination of CNI plus mTOR inhibitors in de novo KTRs observed no increase of 1-year PTDM versus CNI plus antiproliferative agents in 13 studies [ n = 4561 participants; relative risk 1.16, 95% confidence interval (CI) 0.97-1.38,P = .10][30 ].These results were confirmed in the TRANSFORM (TRANSplant eFficacy and safety Outcomes with an eveRolimus-based regiMen) study, a 24-month, prospective, open-label trial in 2037 de novo KTRs randomized to receive everolimus with reduced-exposure CNI versus mycophenolate with standard-exposure CNI [31 ].No difference in PTDM in-cidence was observed (risk ratio 1.09, 95% CI 0.87-1.37)with comparable efficacy and graft function.
There is no evidence to suggest any glycaemic risk from antiproliferative agents such as mycophenolate mofetil or azathioprine.
Regarding steroids, a previous Cochrane analysis published in 2016 observed similar rates of mortality, graft loss and PTDM comparing regimens of steroid avoidance/withdrawal (stratified before or after 14 days, respectively) versus steroid maintenance, but higher rates of rejection [32 ].In an updated analysis incorporating post-2016 RCTs of steroid avoidance [33 , 34 ], lower rates of PTDM are now observed in steroid avoidance versus maintenance (risk ratio 0.70, 95% CI 0.56-0.88,P = .002)but with similar mortality, graft loss and rejection observations to before (see Supplementary data, Fig. S3).However, the HARMONY study contributes a large effect size but is flawed by overreliance on HbA1c for PTDM diagnosis in the context of anemia rates between 27% and 39% across study arms [33 ].Early steroid withdrawal may have differential impact stratified by age, with older SOT recipients in a population-cohort study demonstrating more favourable responses to steroid withdrawal (e.g.lower PTDM and mortality) but increased risk for rejection [35 ].Balancing PTDM versus graftrelated concerns with steroid avoidance/withdrawal is essential, although patient/graft survival should take priority.In a causal estimation effects registry analysis including 6070 KTRs, steroid withdrawal within 18 months post-transplantation was associated with increased risk of graft loss compared with steroid maintenance [36 ].If a steroid avoidance regimen is desired then induction therapy with lymphocyte depletion should be considered.

OPINION STATEMENT 6: EMPHASIZE LIFESTYLE MODIFICATION TO ALL PATIENTS; CONSIDER MEDICAL OR SURGICAL INTERVENTION FOR TREATMENT OF OBESITY; USE INTERMITTENT EXOGENOUS INSULIN INTERVENTION EARLY POST-TRANSPLANTATION FOR POST-OPERATIVE HYPERGLYCAEMIA
Since the last meeting report [1 ], various groups have summarized suggestions on PTDM prevention [37 -41 ].These include: (i) dietary modification; (ii) physical exercise/training; (iii) pharmacological intervention; (iv) immunosuppression modification; (v) bariatric surgery; (vi) performing OGTTs pre-transplant for targeted intervention; and (vii) other measures including manipulation of microbiota.Meeting participants agreed any opinion regarding prevention would intuitively become stronger with increasing PTDM risk.
Regarding (i), uncertainty exists about the best dietary intervention [42 ], as observational evidence only supports Mediterranean diets [43 ] or increased vegetable intake [44 ].With (ii), the CAVIAR (Comparing glycaemic benefits of Active Versus passive lifestyle Intervention in kidney Allograft Recipients) RCT implemented a graded exercise program with active dietician intervention (versus leaflet advice), which did not improve pathophysiological markers of glucose metabolism but reduced PTDM incidence [45 ].An observational study demonstrated higher physical activity levels lowered risk of PTDM, and cardiovascular and allcause mortality [46 ].Although better evidence is desirable, meeting participants agreed that lifestyle modification, combining measures (i) and (ii), should be emphasized post-transplantation based upon evidence from the general population [47 ].As for (iii), meeting participants agreed early exogenous insulin administration could be considered for PTDM prevention despite a recent RCT not reaching its primary endpoint (1-year PTDM incidence) [48 ].This agreement acknowledged that the odds for overt PTDM at 1-year were significantly reduce in the adjusted per-protocol analysis only [48 ], and was also based on an earlier RCT (cited in previous meeting report) [1 ].However, higher hypoglycaemia rates with this approach must be acknowledged [48 ] and enthusiasm may be influenced by inpatient length of stay post-operatively.An ongoing multicentre RCT testing early administration of vildagliptin for PTDM prevention is underway ( Supplementary data, Table S3) [49 ], but another RCT was recently published demonstrating that post-operative sitagliptin was safe but did not lead to significant improvement in OGTT-derived 2-h glucose at 3 and 6 months post-transplantation [50 ].
The most controversial issue with PTDM prevention is immunosuppression tailoring for SOT patients at higher PTDM risk as per (iv), which is addressed under Opinion Statement 5. Meeting participants agreed further research is warranted to investigate immunosuppression modification strategies to prevent or treat PTDM.
Concerning (v), there is convincing evidence that bariatric surgery is beneficial for individuals with morbid obesity and chronic kidney disease (CKD), including those already waitlisted or seeking eligibility [51 , 52 ].In kidney transplant candidates with obesity (e.g.body mass index ≥35 kg/m 2 ) refractory to lifestyle intervention, consider surgical or medical intervention which will enable successful transplantation and may aid PTDM prevention.A non-randomized study reported zero cases of PTDM in 12 non-diabetic KTRs transplanted after post-laparoscopic sleeve gastrectomy, in comparison with 3 of 18 patients from a matched non-laparoscopic sleeve gastrectomy control group (statistically not significant) [53 ].As an alternative, GLP-1 receptor agonists might be a promising pharmacological option for individuals with advanced CKD and obesity who are transplant candidates.Studies are pending to determine feasibility ( Supplementary data, Table S3).
Regarding measures (vi) and (vii), Hap et al .performed OGTTs among 80 waitlisted kidney transplant candidates and recom-mended a low carbohydrate diet, lifestyle modification and increased physical activity to 31 patients with dysglycaemia (with 28/31 showing attenuated glucose metabolism throughout the 12-month observational period post-transplant) [54 ].These results align with several measures highlighted above showing that behavioural factors such as motivation are important to enable PTDM prevention.

OPINION STATEMENT 7: USE THE NOVEL AGENTS; PERSONALIZE GLUCOSE-LOWERING THERAPY BASED UPON A PATIENT-DEPENDENT HIERARCHY
Cardiovascular outcome trials using glucose-lowering treatment in KTRs are lacking.Novel agents, sodium-glucose co-transporter 2 (SGLT2) inhibitors and GLP-1 receptor agonists, now dominate diabetes treatment guidelines [55 ].Meeting participants agreed that novel agents are under-utilized for PTDM management due to limitations of transplant-specific evidence (see Tables 1 A/ 1 B).However, prescribing is sub-optimal even in diabetic kidney disease patients in whom there are clear treatment benefits as per national/international recommendations [56 ].This reflects a disconnect between clinical guidelines and real-world prescribing.Available transplant studies do not currently indicate a clear safety risk, which is why our personal view is more enthusiastic in comparison with recent KDIGO guidance on diabetes and CKD recommending more cautious adoption [57 ].Meeting participants agreed targeted PTDM studies are desirable but adoption should not be delayed based on current evidence.Meeting participants also agreed that initiation of glucose-lowering agents will be reliant upon accessibility.However, if accessibility is not an issue, then a patient-dependent hierarchy (Fig. 2 ) is advisable.

Table 1B:
Retrospective studies on glucose-lowering agents after kidney transplantation or SOT including kidney.mortality.Metformin may be an appropriate choice for solid organ transplant recipients at low risk for adverse cardio-renal outcomes or if access to novel anti-diabetics is an issue.However, for solid organ transplant recipients at moderate to high risk for adverse cardio-renal outcomes with no accessibility issues, the consensus opinion was novel anti-diabetic therapies should be strongly considered before metformin.SGLT2 inhibitors can be used for the treatment of PTDM once stable graft function is achieved [62 ].Initiation should be influenced by comorbidities like heart failure (supporting use) and significant urosepsis or severe mycotic genital infection risk (discouraging use), although current studies have not shown increased urinary tract infection risk with SGLT2 inhibitors (see Tables 1 A and 1 B).Enthusiasm for early post-operative commencement will be influenced by local urological practices (e.g.length of post-operative urinary catheter placement, ureteric stent removal).Improvement of glycaemic control may vary based on kidney function (less effective at lower eGFR) [62 ].Awareness of the risk for euglycaemic diabetic ketoacidosis is critical, especially in patients with insulin deficiency.SGLT2 inhibitors should be suspended if fasting is required or during an acute illness.

Strength
GLP-1 receptor agonists are preferable in patients with obesity.Several non-randomized published reports indicate an acceptable safety profile with no increased rejection or graft failure risk, although gastrointestinal side effects are common.Appropriate education is required for patients who are initiated on incretin mimetics with emphasis on slow dose up-titration to improve tolerance, and suspension of treatment with acute illness [25 ].
Insulin should be used for treatment of post-operative hyperglycaemia.For stable patients, oral or non-insulin injectable agents (and their combination) are preferable unless diabetes control cannot be achieved.Of note, data on the glucose-lowering effect of basal insulin in KTRs exist for basal neutral protamine Hagedorn (NPH)-insulin alone [48 ], the peak effect of which can be matched to the glucose peak exhibited by KTRs in the afternoon.
Dipeptidylpeptidase 4 (DPP4) inhibitors are safe but demonstrate no cardio-renal benefit.Thiazolidinediones are better options than sulfonylureas and meglitinides (both have risk of hypoglycaemia), and no evidence exists for alpha-glucosidase inhibitors.Meeting participants agree these drug classes have the lowest priority for clinical use.
In summary, and in view of the pros and cons for each pharmacological therapy, meeting participants agreed that any decision to initiate one glucose-lowering agent versus another is best guided by a patient-dependent hierarchy (shown in Fig. 2 ) if accessibility is not an issue.Personalization of glucose-lowering therapy is essential, with treatment goals depending on comorbidities, awareness of hypoglycaemia risk and allograft function.

OPINION STATEMENT 8: INCREASE COLLABORATIVE RESEARCH BETWEEN ACADEMIC MEDICINE, MULTI-DISCIPLINARY CLINICAL TEAMS, INDUSTRY PARTNERS AND PATIENTS
Exclusion of SOT recipients from pioneering cardiovascular and renal outcome trials of new glucose-lowering agents has resulted in sub-optimal uptake post-transplantation. Observational studies and RCTs relating to PTDM are in progress (see Supplementary data, Table S3), but more are required and should target at-risk groups for maximum benefit.Patient-reported outcomes, health economic analyses and cost effectiveness models are lacking and require dedicated studies and incorporation as secondary outcomes into RCTs where feasible (suggested PTDM clinical trial endpoints in Supplementary data, Table S4).Lack of robust PTDM data capture by national transplant registries limits the ability to ascertain PTDM-associated outcomes [63 ].Acquiring these data should be encouraged to improve our understanding of long-term outcomes with record linkage.Collaboration between healthcare professionals, academic groups, industry and patient groups is essential.
Finally, most published research is after kidney transplantation, but PTDM is a complication affecting all SOT recipients with prevalence rates between 20% and 40% in heart, lung and liver transplant recipients [64 ].In a Danish SOT cohort ( n = 959), the highest incidence of PTDM is seen 46-365 days post-transplantation. SOT recipients with PTDM had higher risk for all-cause mortality (1.89, 95% CI 1.17-3.06),with cardiovascular and cancer-related causality more common than in nondiabetic SOT recipients [18 ].More studies are warranted in nonrenal transplant cohorts.While most of this report is valid across SOT cohorts, bespoke differences may be apparent between different solid organ settings to justify organ-specific versus organgeneric recommendations.

CONCLUSION
PTDM is a complex and multi-factorial post-transplant complication, spanning a continuum of disease that may begin prior to transplantation in many cases.This Meeting Report summarizes proceedings from the 3rd International PTDM Consensus meeting, reflecting expert opinion.Optimizing long-term outcomes after SOT, with attenuation of both premature mortality and/or graft loss, is a clinical priority.Therefore, improving our diagnosis, prevention and management of PTDM should be considered an integral component of long-term post-transplant care.

Figure 1 :
Figure 1: Five aspects of risk assessment for and diagnosis of PTDM and IGT.

Table 1A :
Prospective studies on glucose-lowering agents after kidney transplantation.

Table 1B :
ContinuedBoth tables contain studies from patients with disorders of the glucose metabolism that became known after transplantation (hyperglycaemia/PTDM/IGT). If studies were entirely conducted with patients who had type 2 diabetes before transplantation, they were not listed.