EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Objectives:This study investigated the effect of prediabetes in long-term deceased-donor renal transplant recipients regarding graft survival, graft function, and evolution of new-onset diabetes after transplant compared with a control group of graft recipients with normal glucose tolerance test results.

Materials and Methods: This was a follow-up trial of 187 deceased-donor renal transplant recipients. Based on oral glucose tolerance test results, the cohort was divided into groups A and B, comprising individuals with normal glucose metabolism (n = 130, 69.9%) and individuals with prediabetes (n = 56, 30.1%). Data are shown as means ± standard errors.

Results: Both groups showed similar total transplant survival (116.8 ± 5.4 vs 114.5 ± 7.4 mo; P = .742) and transplant survival measured since oral glucose tolerance test (58.5 ± 1.4 vs 59.5 ± 1.9 mo; P = .990, Mantel-Cox P = .943). Univariate and multivariate Cox regression analyses showed no association of prediabetes with graft loss. Transplant function changes were similar between cohorts (-3 ± 1 vs -5 ± 2 mL/min/1.73 m² body surface area, using the Chronic Kidney Disease Epidemiology Collaboration formula; P = .538). At 5-year follow-up, recipients with prediabetes had higher hemoglobin A1c than controls (5.99% ± 0.10% vs 5.67% ± 0.04%; P = .002). Prediabetes was associated with a 4.5-fold increased hazard of new-onset diabetes after transplant (P = .021).

Conclusions: Prediabetes was associated with a 4.5-fold higher hazard ratio for new-onset diabetes after transplant but not with reduced graft function or survival.

Key words : Patient outcome, Prediabetic state, Renal transplantation, Transplant outcome

Renal transplant has become the first option for the treatment of end-stage renal disease in many countries, especially in younger patients with this disease. Over the past decades, kidney transplant recipients (KTRs) of organs from deceased donors have profited from improvements in survival rates and quality of life compared with end-stage renal disease patients on dialysis.¹ A leading cause of reduced graft and KTR survival is new-onset diabetes after transplant (NODAT)² as it represents a major risk factor for cardiovascular disease and mortality in this specific patient population.^3,4 In KTRs, oral glucose tolerance testing (oGTT) can unmask high rates of undiagnosed prediabetes.⁵ Prediabetes itself and its subclasses of impaired fasting glucose and impaired glucose tolerance are recognized risk factors for cardiovascular disease and death in the general population.^6-8 Emerging data suggest that prediabetes in the renal transplant setting not only predisposes patients to later advent of NODAT but also represents a major cardiovascular risk factor for graft and KTR outcomes in the early posttransplant period.^9-11 Thus far, numerous investigations have focused on the development of NODAT, its risk factors, and treatment options after renal transplant^12-14; however, less is known about prediabetes and its effects on graft function, graft survival, and the evolution of NODAT in long-term KTRs once the early posttransplant period is over. Therefore, the aim of this study was to investigate and compare the long-term graft function, graft survival, and the evolution of NODAT in deceased-donor renal transplant recipients with prediabetes and those without prediabetes. We have reported on the prevalence of prediabetes in a cohort of 200 long-term KTRs in our outpatient transplant clinic⁵ and followed this cohort for several years. Results from this observational long-term follow-up trial are presented here.

Materials and Methods

Study approval
The study was approved by the local ethics committee, and data were coded in such a manner that participants could not be identified either directly or through linked identifiers. The clinical and research activities reported herein are consistent with the “Principles of the Declaration of Istanbul” as outlined in the “Declaration of Istanbul on Organ Trafficking and Transplant Tourism” and with the Declaration of Helsinki.^15,16

Study population
This study represents a long-term follow-up investigation of a cohort of 187 renal allograft recipients with prediabetes and its respective control group.⁵ Individuals in whom NODAT was diagnosed by oGTT had been excluded a priori from our current analysis. Furthermore, the maintenance immunosuppressive drug regimen in those diagnosed with prediabetes on oGTT had not been changed. Instead, dietary and behavioral counseling had been implemented. Of the original 187 recipients, 1 patient was lost to follow-up, 3 patients had no recorded hemoglobin A1c (HbA1c) levels, and 42 patients died during follow-up from oGTT to study termination at the end of follow-up. Therefore, for this report, 186 transplant recipients could be reevaluated for graft loss and 141 for evolution of glucose metabolism measured by HbA1c levels at study termination (Figure 1). The patients were classified, based on oGTT results, as having normal glucose metabolism (130 patients; 69.9%) or prediabetes (56 patients; 30.1%). Oral glucose tolerance testing was performed on average 59.4 months after successful renal transplant. The cohort consisted of 115 male (61.8%) and 71 female patients (38.2%); the mean age was 45.1 ± 13.1 years, and the mean time on end-stage renal disease replacement therapy before transplant was 75.2 ± 58.2 months. In our study group, 157 patients (84.4%) received their first renal allograft, 24 (12.9%) their second, and 5 (2.7%) their third.

The immunosuppressive regimen at enrollment consisted of cyclosporine + mycophenolate mofetil (16.1%), cyclosporine + mycophenolate mofetil + steroids (17.2%), cyclosporine + steroids (37.1%), tacrolimus + mycophenolate mofetil (5.4%), tacrolimus + mycophenolate mofetil + steroids (7.5%), and tacrolimus + steroids (16.7%). Causes of renal failure were chronic glomerulonephritis (37.1%), autosomal dominant polycystic kidney disease (9.1%), interstitial nephritis and chronic reflux (12.9%), hypertension (5.9%), other (14.5%), and unknown (20.4%). Mean creatinine and HbA1c levels at study inclusion were 1.88 ± 0.69 mg/dL (estimated glomerular filtration rate according to the Chronic Kidney Disease Epidemiology Collaboration formula of 42.6 ± 16.6 mL/min/1.73 m² body surface area) and 5.6 ± 0.5%. Sixty-one patients (32.8%) started renal replacement therapy for end-stage renal disease before 1992, 56 (30.1%) from 1992 to 1995, and 69 (37.1%) from 1996 on.

Measurement of hemoglobin A1c and creatinine levels
Hemoglobin A1c and creatinine levels were measured using a Roche Modular P analyzer (Roche Diagnostic, Mannheim, Germany), according to the manufacturer’s instructions. Levels of HbA1c on follow-up were determined using an immunoturbidimetric assay (Tina-quant HbA1c 2.Gen, Roche Diagnostics) that had been standardized according to the International Federation of Clinical Chemistry (mmol/mol). The HbA1c results were automatically converted to values (%) based on the Diabetes Control and Complications Trial/NGSP program via a correction formula. Creatinine levels were measured using an enzymatic method (Roche Diagnostics). The assay was standardized against isotope dilution mass spectrometry reference measurements.

Classification of hemoglobin A1c levels
For classification of HbA1c levels indicative of normal glucose metabolism, prediabetes, or diabetes, current American Diabetes Association criteria (www.diabetes.org, accessed on May 25, 2016) were applied as follows: normal was ≤ 5.6% (38 mmol/mol), prediabetes was ≥ 5.7% to ≤ 6.4% (39 mmol/mol to 47 mmol/mol), and diabetes (NODAT) was ≥ 6.5% (48 mmol/mol).

Data acquisition
All data were drawn from clinical records or electronic databases (eg, the hospital laboratory data system) or the Eurotransplant electronic resource (www.eurotransplant.org).

Statistical analyses
The primary outcome variable was renal allograft survival, determined as a combined endpoint of graft failure or death with a functioning graft, as related to patients with normal glucose metabolism (group A) versus patients with prediabetes (group B). The secondary outcome parameters were renal graft function and evolution of glucose metabolism measured by HbA1c levels. Descriptive statistics such as means, standard errors, frequencies, and proportions were generated to describe the total study population. Kaplan-Meier analyses were performed to evaluate transplant survival. Cox regression analysis was performed to investigate the effect of prediabetes on graft survival. The Mann-Whitney U test and the chi-square test were applied to assess continuous and categorical cohort characteristics between patients with normal glucose tolerance and prediabetes at the end of the study period. Multivariate binary logistic regression was performed to assess the effects of prediabetes at study entry on the evolution of NODAT on follow-up. Diagnostics to assess the model fit and to detect outliers were conducted. Statistical analyses were performed using SPSS for Windows (SPSS: An IBM Company, version 22.0, IBM Corporation, Armonk, NY, USA).

The study was terminated 66 months after oGTT, with mean follow-up time in the surviving KTRs of 63.5 ± 14.8 months.

Renal transplant survival
At the end of the study period, 42 patients (22.6%) reached the combined endpoint of transplant failure or death with a functioning graft. Table 1 shows the graft survival from transplant to end of follow-up in months, and Figure 2 shows the respective box plot graph. Figure 3 shows the Kaplan-Meier analysis. In essence, we could not detect any difference in renal graft survival between renal transplant recipients with prediabetes and those with normal glucose metabolism.

Risk factor analysis of graft loss
On univariate Cox regression, prediabetes was not shown to be a risk factor for early graft loss (P = .729). Multivariate Cox regression was performed to search for further risk factors of graft loss with inclusion of the following variables into the model fit: prediabetes versus normal glucose metabolism on oGTT (P = .970), sex (P = .085), dialysis duration before transplant (P = .313), and age at transplant (P = .333).

Effect of prediabetes on renal graft function and hemoglobin A1c levels
We also analyzed the effects of prediabetes on renal graft function and HbA1c levels (Table 2). The groups did not differ with respect to sex (with Fisher exact test, P = .438) and immunosuppressive regimen (cyclosporine versus tacrolimus; with Fisher exact test, P = .256) (data not shown in Table 2). The change in transplant function over the study period was equal in both cohorts (-3 ± 1 vs -5 ± 2 mL/min/1.73 m2 body surface area, using the Chronic Kidney Disease Epidemiology Collaboration formula; P = .538), although there seemed to be a tendency for a faster decline in the KTRs with prediabetes.

Hemoglobin A1c levels were significantly
higher in the group with prediabetes at the end of follow-up, and, although no effect on graft function could be determined, prediabetes was associated with a higher incidence of NODAT (16.3% vs. 4.1%; with chi-square test, P = .003) on long-term follow-up (Table 3). Logistic regression showed an independent effect of prediabetes at the time of oGTT on the evolution of NODAT approximately 5 years later (P = .021, hazard ratio of 4.5).

In contrast to the well-known negative effects of NODAT, the impact of prediabetes in KTRs on long-term kidney function, evolution of NODAT, and allograft outcome have not yet been definitely described.

The main findings of our study are that (1) prediabetes as defined by oGTT was associated with a 4.5-times higher likelihood of later evolution of NODAT in long-term renal transplant recipients; (2) over a follow-up of 5 years, prediabetes was associated neither with reduced graft function nor with increased rates of graft failure or patient mortality; and (3) KTRs with normal oGTT results carried a relatively low risk of de novo diabetes, once the early post-transplant period was over.

Numerous studies have demonstrated the association of NODAT and early posttransplant hyperglycemia with significantly higher hazard ratios for major cardiovascular events, graft dysfunction, graft loss, and patient death.^17-21 As a consequence, a paradigm shift of treatment recommendations regarding changes in glucose metabolism in KTRs has been suggested.²²Current investigations indeed have pointed toward a reduced risk of NODAT on major adverse events, presumably due to more thorough blood glucose monitoring posttransplant.^23,24 In line with these more recent reports, our data could not determine any association of prediabetes with reduced transplant function or graft and patient outcomes. Hyperglycemia as caused by overt diabetes can lead to insulin resistance, dyslipidemia, platelet activation, or hypertension.²⁵ Because prediabetes does not induce the same amount of unfavorable glucometabolic changes as those in overt diabetes, our results seem to be pathophysiologically plausible. Taking this and the tendency toward a faster decline of graft function shown in prediabetic KTRs into account, one might argue that a follow-up of 5 to 10 years to investigate patient and graft outcome data of KTRs with prediabetes is too short given that even overt diabetes in the general population needs decades of active disease before relevant adverse events manifest.

With respect to the evolution of NODAT, however, our observational data indicated a 4.5-times increased risk for the evolution of posttransplant diabetes mellitus in KTRs with prediabetes. On the other hand, patients with normal glucose tolerance carried a relatively low risk of de novo diabetes over the 5-year follow-up. In a recent analysis of serial oGTTs in 672 KTRs over 5 years posttransplant, Porrini and colleagues reported similar outcomes.²⁶ In their prospective trial, 28% of patients showed oGTT results indicative of prediabetes and 21% of these patients developed late posttransplant diabetes mellitus during follow-up, similar to our result of approximately 16%. They reported a much higher prevalence of NODAT at 60 months after transplant, at 33.6%, versus 7.8% at about 120 months posttransplant in our cohort. The relatively low incidence of NODAT in our cohort is likely the result of 2 factors: first, patient selection due to our study protocol, with exclusion from analyses of KTRs with incident NODAT during our initial investigation, and, second, the long follow-up time where occult NODAT in the deceased KTRs could not be ruled out.

Still a matter of debate is how to regularly monitor glucose metabolism in KTRs most optimally. Here, the better accuracy of oGTT5,26 might be weighed against cost effectiveness and convenience of regular (eg, quarterly) measurements of HbA1c levels or afternoon blood glucose level estimations.²⁷In the general nondiabetic population, a positive association of HbA1c levels with aging has been reported.²⁸ Interestingly, although not associated with aging, an even greater effect of time on HbA1c levels in nondiabetic KTRs has been demonstrated.²⁹ Therefore, and for reasons of diagnostic accuracy, we currently still advocate for yearly oGTT for the exact diagnoses of glucometabolic impairments in KTRs.

Nevertheless, regular measurements (eg, every 12 wk) of HbA1c and/or random blood sugar measurements (fasting or in the afternoon) are helpful in the early detection of an impaired glucose metabolism after kidney transplant. In prediabetic long-term KTRs, we do not consider changing the immunosuppressive regimen; instead, we recommend counseling on dietary habits, weight reduction, and physical activity.

Limitations and strengths
Major limitations of our study include its retrospective design, relatively small study cohort, and the possible bias of patient selection at study entry. Although HbA1c measurements alone have been reported to yield similar sensitivity to that of oGTT from 3 months posttransplant,²⁷ regular oGTT measurements, despite increased costs and reduced patient convenience, might have substantially improved our data on evolution of NODAT.

Because 42 patients died before study termination without definite evaluation of glucose metabolism by oGTT before death, the 4.5-fold risk calculation for the evolution of NODAT in the remaining patients with prediabetes may be distorted.

Finally, our data represent a single-center experience in transplant recipients of mainly white origin. Therefore, these results may not be applicable to other transplant centers or patients of a different race.

Strengths of this study include the long follow-up of a formerly reported well-defined patient cohort and the analyses of patient and transplant survival as important patient outcome data. Therefore, we believe that this observational study makes a valuable contribution by suggesting that, although prediabetes carries a relatively low risk with respect to graft function and outcomes, it represents a major and amenable risk factor for the evolution of NODAT in KTRs on long-term follow-up.

During a follow-up of 5.3 years after oGTT, prediabetes was not associated with reduced graft function or increased rates of graft loss or patient mortality in KTRs on maintenance immunosuppressive medication. Nevertheless, prediabetes was associated with a 4.5-fold higher likelihood of later evolution of diabetes. Serial oGTT or HbA1c measurements in every KTR could be valuable to exactly define the glucometabolic status of long-term transplant recipients for further risk stratification and to improve counseling of KTRs by transplant physicians on modifiable risk factors to prevent NODAT. Prospective trials with follow-up times longer than 5 to 10 years may be necessary to unmask deleterious effects of prediabetic states on graft and patient outcome parameters in this special patient population.

Volume: 15 Issue: 6 December 2017

FULL TEXT