Risk Factors for Development of Post-Transplant Diabetes Mellitus After Kidney Transplantation and Comparison Between Older and Younger Recipients in the Early Post-Transplantation Period: A Single-Center Study

Introduction

Kidney transplantation is considered the most effective treatment for end-stage chronic kidney disease (ESKD) [1], but various complications can occur after this procedure.

Post-transplant diabetes mellitus (PTDM) is one of these complications and is a commonly observed metabolic disorder [2].

PTDM is defined as newly diagnosed diabetes mellitus in the post-transplant setting, irrespective of whether it was present but undetected before transplantation [3]. Diabetes mellitus after transplantation was first described in 1964 among kidney transplant recipients [4], and since then the nomenclature of this disease has changed many times. In 2014, the International Expert Panel, consisting of transplant nephrologists, diabetologists, and clinical scientists, recommended changing the terminology from new-onset diabetes after transplantation (NODAT) to post-transplant diabetes mellitus (PTDM). This change was due to the high prevalence of undiagnosed pre-transplant diabetes mellitus [4].

The incidence of PTDM ranges from 4% to 25% [2], but higher rates have also been reported (up to 40%) [5].

Various modifiable and non-modifiable risk factors for PTDM have been reported. Some of them are the same as risk factors for type 2 diabetes mellitus (DM), and include Black and Hispanic ethnicity, age, elevated BMI, family history of diabetes, and male sex. Other reported risk factors are specific to solid organ transplantation and include hypomagnesemia, a history of biopsy-proven acute rejection (BPAR), use of steroids and calcineurin inhibitors (CNI), cytomegalovirus (CMV) infection, hepatitis C, and certain human leukocyte antigen (HLA) types [6]. Moreover, genetic and epigenetic polymorphisms have also been also associated with PTDM [7].

The occurrence of PTDM has a significant impact on quality of life and mortality. Diabetes after transplantation has been associated with worse patient and graft survival [8–11 and it can also promote other transplant complications such as cardiovascular diseases, infections, and impaired wound healing [12,13].

Although knowledge about risk factors leading to PTDM has improved, there is still a need for further research. Furthermore, patients with PTDM may require tailored management strategies. For example, older individuals with PTDM might be at higher risk of cardiovascular events, which warrants a different approach to follow-up and treatment in this subgroup. This topic has not been comprehensively addressed in the literature, which highlight the need for further investigation.

Thus, the aim of this single-center study, performed at Infant Jesus Clinical Hospital Warsaw, Poland, was to analyze potential risk factors associated with PTDM in kidney transplant recipients. Additionally, we compared older and younger patients with PTDM, seeking to determine differences between these groups and to assess the need for a dedicated approach.

Material and Methods

STUDY POPULATION:

In this retrospective, observational study we analyzed data from patients who underwent kidney transplant (KT) between January 2021 and February 2024. A total of 375 KTs were performed. Patients who received single-organ kidney transplant without prior history of diabetes mellitus were included. The exclusion criteria included multi-organ transplantation, pre-existing diabetes mellitus (DM), conversion from cyclosporine to tacrolimus during the follow-up period (the timing of conversion to cyclosporine varied significantly between individuals, making it difficult to accurately describe and interpret their immunosuppressive exposure), and transferring to a different center during the follow-up period (which resulted from the nationwide mandatory organ allocation framework for kidney transplantation, by which patients were transferred to other centers at various time points and could not be systematically followed by our center, which prevented consistent data collection). We excluded 157 recipients from the study, after applying the exclusion criteria. A homogeneous group of 218 White patients was analyzed.

All data were accessed through medical records and laboratory test results obtained during hospitalizations, outpatient clinic visits, and from transplant registry data. Our data included demographics, comorbidities, transplant characteristics, type of induction therapy (if applicable), treatment of biopsy-proven acute rejection (if applicable), polyomavirus replication, cytomegalovirus replication, laboratory test results (cholesterol, magnesium, triglycerides, tacrolimus, serum creatinine, uric acid), and type of treatment for PTDM. The follow-up period was 6 months. The study was conducted in full accordance with the principles of the Declaration of Helsinki and the Declaration of Istanbul, and informed consent was obtained from all participants.

DEFINITIONS:

Post-transplant diabetes mellitus was diagnosed based on the 2013 International Consensus Meeting on Post-transplant Diabetes Mellitus, and included symptoms of hyperglycemia (polydipsia, polyuria, and unintentional weight loss) with random blood glucose ≥200 mg/dl, fasting plasma glucose ≥126 mg/dl, or two-hour plasma glucose ≥200 mg/dl during oral glucose tolerance test (OGTT). Since hemoglobin A1C (HbA1C) was not recommended as a screening test in the early post-transplant period, it was not used as a diagnostic tool [5]. The diagnosis of PTDM was made only after the patient had been on maintenance immunosuppression treatment for at least 3 months after transplantation). The diagnosis of BPAR was made based on histological findings obtained during the protocol biopsy at 3 months after transplant or when acute rejection was clinically suspected. Biopsy-proven acute rejection was treated with 500 mg of methylprednisolone for 3 consecutive days. Older patients were defined as those aged ≥60 years, based on United Nations definition.

Serum magnesium, uric acid, tacrolimus, cholesterol, and triglyceride levels were obtained during the follow-up period and are presented as mean values over the first 6 months. Hypertriglyceridemia was defined as serum triglyceride levels ≥150 mg/dL (reference rage <150 mg/dL), hypomagnesemia as serum magnesium levels <1.6 mg/dL (reference range 1.6–2.4 mg/dL), and hypercholesterolemia as total cholesterol levels ≥190 mg/dL (reference range <190 mg/dL) or current use of lipid-lowering medication. Hyperuricemia was defined as serum uric acid >6.8 mg/dL (reference range 3.4–6.8 mg/dL) or current use of uric acid-lowering agents. Participants were considered to have dyslipidemia and/or hypomagnesemia and/or hyperuricemia if any of the above criteria were met. Both polyomavirus replication and cytomegalovirus replication were monitored in the third and sixth months after transplant or in the presence of symptoms suggestive of viremia.

STATISTICAL ANALYSIS:

Continuous variables were summarized as mean±standard deviation (SD) or median and interquartile range (IQR), while categorical variables were presented as n (%). Normality was evaluated with the Shapiro–Wilk test, supported by assessments of skewness and kurtosis. Levene’s test was applied to assess homogeneity of variances. Comparisons between groups were made with the t test, Mann-Whitney U test, Pearson’s chi-square test, or Fisher’s exact test, as appropriate. Two-step logistic regression analysis was performed to identify risk factors for post-transplant diabetes mellitus. Variables for the multivariate model were selected based on p-value threshold of <0.25 [14], and a stepwise approach was used for final variables selection. Potential confounders were identified a priori based on clinical relevance and literature review, and were included in the multivariable logistic regression model alongside variables meeting the p-value threshold of <0.25 in univariate analysis. This approach allowed adjustment for potential confounding effects when estimating the independent association between each variable and PTDM risk. Logistic regression was chosen due to the binary nature of the outcome (presence/absence of PTDM) and its ability to control for multiple covariates simultaneously. Model fit was assessed using Nagelkerke R2 and Hosmer and Lemeshow goodness of fit (GOF) test. Variance inflation factors (VIF) were calculated to verify multicollinearity. Receiver operating characteristic (ROC) analysis was conducted to evaluate the prognostic performance of the selected variables for PTDM. Optimal cut-offs were indicated with the Youden method. A significance level (alpha) of 0.05 was used for statistical significance. All analyses were performed using R software (R4.1.2).

Results

CHARACTERISTICS OF STUDY GROUPS AND COMPARISON OF STUDY SUBGROUPS:

Of the 375 KT procedures, 218 (58%) patients were included in the study. Among them 131 (60.09%) were male and 87 (39.91%) were female.

In the study group, 189 patients (87%) underwent their first kidney transplantation, while 29 (13%) had received a second transplant. The mean age of all recipients was 45.5 years. Fifty-five patients (25%) developed PTDM (26 women and 29 men; 47.3% vs 52.7%). Of these, 24 (44%) patients were aged ≥60 years (12 women and 12 men).

All patients received a standard immunosuppression protocol consisting of steroids, calcineurin inhibitors (CNI)-tacrolimus, and mycophenolate acid. Seventy-five patients included in the study (34%) received induction therapy of thymoglobulin (ATG), and 24 patients (11%) received basiliximab prior to transplantation.

The group with PTDM was compared with those without PTDM in terms of clinical and laboratory data. Subsequently, patients who developed PTDM were divided into 2 subgroups according to age (< 60 years and ≥60 years), and comparisons between them were conducted.

Compared to patients without PTDM, patients with PTDM were significantly older (MD=11.11, CI95 [7.19; 15.03], p<0.001), and had significantly higher BMI (MD=2.46, CI95 [1.20; 3.71], p<0.001). The proportion of patients with age above or equal 60 years was significantly higher among those with PTDM (43.6% vs 12.9%, p<0.001). A comparison between patients with PTDM and patients without PTDM is presented in Table 1.

RISK FACTORS OF PTDM – LOGISTIC REGRESSION ANALYSIS:

In the univariate analysis, advanced age significantly increased the odds of PTDM (OR=1.07, CI95 [1.04; 1.10], p<0.001). Additionally, patients aged 60 years or above had 5-fold higher odds of PTDM than patients below 60 years (OR=5.24, CI95 [2.60; 10.68], p<0.001). Other risk factors associated with PTDM occurrence were higher BMI (OR=1.15, CI95 [1.07; 1.25], p<0.001), hypertriglyceridemia (OR=1.01, CI95 [1.00; 1.01], p<0.001), hypercholesterolemia (OR=1.01, CI95 [1.00; 1.02], p<0.001), and hypomagnesemia (OR=2.34, CI95 [1.19; 4.57], p=0.013). No correlation between sex, type of donor, type of dialysis, hyperuricemia, induction therapy, presence and treatment of acute rejection, and mean tacrolimus level was observed. Figure 1 presents boxplot charts illustrating the distribution of variables significantly different between patients with PTDM and those without PTDM.

Multivariate logistic regression model confirmed that age had a significant impact on the odds of PTDM. Each additional year increased the odds of PTDM by 8% (OR=1.08, CI95 [1.04; 1.11], p<0.001). The odds of PTDM were 12% higher for each 1 kg/m2 increase in BMI (OR=1.12, CI95 [1.02; 1.23], p=0.026). The concentration of triglycerides slightly influenced the odds of PTDM (OR=1.00, CI95 [1.00; 1.01], p=0.046). Hypertriglyceridemia increased the odds of PTDM by 4-fold (OR=3.54, CI95 [1.12; 13.93], p=0.045). Higher magnesium concentrations reduced the odds of PTDM by 88% (OR=0.12, CI95 [0.03; 0.46], p=0.003). Table 2 presents the outcomes of logistic regression analysis for PTDM.

RECEIVER OPERATING CHARACTERISTICS (ROC) ANALYSIS:

Receiver operating characteristics (ROC) analysis was conducted to evaluate the predictive ability of selected parameters for PTDM. The highest AUC (area under the curve), which referred to best prognostic properties, was found for age (AUC=0.733, CI95 [0.658; 0.809]) with a cut-off of 48.5 years. Patients above the cut-off were predicted to develop PTDM with a sensitivity of 71% and a specificity of 69%. AUC values for other variables ranged from 0.638 (presence of hypertriglyceridemia) to 0.693 (concentration of triglycerides), indicating moderate predictive ability. The results are summarized in Table 3.

COMPARISON OF PATIENTS WITH PTDM AGED ≥60 YEARS AND PATIENTS WITH PTDM AGED <60 YEARS:

The younger group was compared with older group in terms of induction therapy, type of PTDM treatment (insulin vs oral medications), infection occurrence, cytomegalovirus (CMV) replication, and polyomavirus (BKV) replication, creatinine level after 6 months, presence of biopsy-proven acute rejection, and levels of cholesterol and triglycerides. No significant difference was found between patients with PTDM aged ≥60 years and patients with PTDM aged <60 years (p>0.05). Results are presented in Table 4.

Discussion

Post-transplant diabetes mellitus is a common complication after kidney transplantation. In our study, 25% developed this metabolic disorder, which is quite high, but it agrees with other studies [12,15]. The incidence of PTDM ranges from 10% to 25% [2,16]. This wide variation may result from lack of a standard definition of PTDM, duration of follow-up, and the presence of modifiable and non-modifiable risk factors in kidney transplant recipients in homogenous cohorts. The present study found that the risk factors advanced age, higher BMI, hypertriglyceridemia, hypercholesterolemia, and hypomagnesemia were associated with PTDM, consistent with previous studies [17].

Since age is a well-known non-modifiable risk factor of type 2 diabetes mellitus [18], it is not surprising that we found a higher risk of PTDM for KT patients. Patient age was the strongest risk factor, with a cut-off of 48.5 years. Additionally, the subgroup of older recipients (aged 60 years or above) had 5 times higher odds of PTDM than patients below 60 years (p<0.001). Comparable conclusions were reported in other studies [19–21].

Consistent with trends in the general population, elevated BMI was a significant risk factor for PTDM in our study. The mechanisms responsible for insulin resistance in obese (BMI >30 kg/m2) and overweight (BMI >25 kg/m2) patients are not fully understood, but may be the consequence of a chronic inflammatory state caused by excessive fat tissue, which stimulates macrophage recruitment to adipocytes and the release of proinflammatory adipokines, leading to downregulation of insulin signaling [22]. Furthermore, adipose tissue produces tumor necrosis factor-alpha (TNF-α) and its activation is associated with insulin resistance due to reduced expression of insulin-sensitive transporters [23].

Some studies suggest that post-transplant weight gain is also a risk factor of PTDM [21,24]. Another important aspect might be the body fat distribution. Cron et al demonstrated that PTDM was strongly associated with central obesity [25]. In a study performed by von Düring et al, visceral fat tissue was correlated with PTDM and hyperglycemia early after transplantation [26]. Thus, it might be essential to monitor not only BMI, but also waist circumference in KT recipients.

Our study also showed that elevated triglyceride levels were associated with PTDM, perhaps due to the association between hypertriglyceridemia and insulin resistance, which can then lead to diabetes [27].

Hypomagnesemia has been found to be related to increased risk of PTDM, although the underlying mechanism remains unclear [12,28]. Lower magnesium level impacts insulin signaling [29], but it also might be the effect of calcineurin inhibitor treatment, which is considered to be a risk factor for PTDM [30]. Moreover, Augusto et al reported that pre-transplant, rather than post-transplant, hypomagnesemia was an independent risk factor of PTDM [31]. The same results were shown by Xu et al [32]. In our study, post-transplant hypomagnesemia was an independent risk factor of development of PTDM, although pre-transplant serum magnesium levels should also be taken into consideration in further research.

Our study revealed that hypercholesterolemia was associated with PTDM. Sinangil et al, also revealed positive correlation between elevated total cholesterol level and LDL-C (low-density lipoprotein cholesterol) in patients with PTDM [33]. On the contrary some studies found out that the rise of TG/HDL-C (triglyceride/high-density lipoprotein cholesterol) ratio and lower HDL-C were increasing risk of diabetes mellitus in KT recipients [22,28].

The impact of cholesterol and its fractions on PTDM might be because excess cholesterol accumulation leads to β-cell dysfunction, thus impairing glucose tolerance, and affecting insulin secretion. Moreover, islet cholesterol deposition can cause increased islet amyloid polypeptide aggregation, and increased islet amyloid formation, thus further deteriorating β-cell function and affecting glucose homeostasis [26,34,35].

Numerous studies suggest that immunosuppression therapy, particularly calcineurin inhibitors, and steroids, may contribute to PTDM development in a dose-dependent manner [30,32,36]. However, in our study, no correlation was found between mean post-transplant tacrolimus levels, additional steroid doses (used to treat BPAR), and PTDM. This may be due to the short follow-up period, and the low incidence of BPAR, which has limited statistical significance.

In the final stage of the study, we divided the group who developed PTDM into 2 subgroups based on age (≥60 years of age, and <60 years of age), and then compared them.

There was no significant difference in regard to BPAR, CMV infection, BKV infection, or type of PTDM treatment (insulin vs oral medications). Mean creatinine level at the end of follow-up period was 1.6 mg/dl for older patients, and 1.4 mg/dl for younger patients, which in our opinion is similar, and acceptable outcome. Revanur et al, in a retrospective study, revealed that survival of patients over the age of 55 years with PTDM was similar to the control group. On the contrary, KT recipients under 55 years of age with PTDM were associated with a much higher risk of death. No differences in graft survival or acute rejection were found [8]. Comparison between older and younger KT recipients with PTDM has not been extensively studied. More research on differences between PTDM patients is needed, ideally with a longer follow-up, thus an adequate patient-specific approach can be performed. We plan to address this in our future work.

This study has a number of limitations. Firstly, it is a single-center study with only 218 White participants, from which 55 developed PTDM, which limits extrapolation of the results to other populations. Moreover, many patients (42%) were excluded due to variable timing of tacrolimus-to-cyclosporine conversion or transfer to other facilities, which precluded consistent follow-up and could limit the generalizability of our findings.

This was a retrospective study with database analysis; therefore, the reliability or lack of available data limits the scope of the results.

The follow-up period was relatively short compared to other studies; therefore, some factors (eg, immunosuppression) might have long-term pro-diabetogenic effects which were not observed in this study. Additional studies with prolonged observation are needed to validate these findings, and we are already planning a follow-up study with extended observation. Lastly, we enrolled a relatively small group of older and younger patients with PTDM (24 and 31 recipients, respectively).

Conclusions

Advanced age had the strongest association with PTDM. Elevated BMI, hypomagnesemia, and hypercholesterolemia also increased the risk of PTDM. No significant differences in terms of serum creatinine level, CMV infection, BKV infection, BPAR, or type of PTDM treatment (insulin vs oral medications) were detected in younger or older recipients with PTDM. PTDM influences patient and graft survival, and increases risk of cardiovascular diseases. More research is necessary to establish modifiable risk factors to help prevent PTDM [8,9,37]. Additionally, KT recipients with non-modifiable risk factors should be regularly screened for PTDM, and aggressive treatment is needed if they develop diabetes, to minimise the risk of complications. Furthermore, close monitoring and management of modifiable risk factors, such as BMI and cholesterol and triglyceride levels, in high-risk patients is crucial, highlighting the importance of personalized preventive strategies in clinical practice. Practical recommendations, including screening protocols and lifestyle interventions, may help clinicians mitigate PTDM risk. More research comparing older and younger patients with PTDM is needed, focusing on long-term risks such as BPAR, CMV/BKV replication, and cardiac events, thereby providing a better and more individualized approach.