Data from the Korean National Health Insurance Service (NHIS), which provides mandatory universal coverage to approximately 97% of the Korean population (the Medical Aid Program covers the remaining 3% of the population), were used in this nationwide retrospective cohort study. The NHIS database comprises two databases: an eligibility database (containing information such as age, sex, type and severity of disability, socioeconomic variables, income level, and type of eligibility) and a medical treatment database (containing medical bills submitted by medical service providers). Several epidemiological studies have used data from the Korean NHIS database [15–17].

The NHIS conducts an annual or biennial national health screening program that comprises a standardized self-report questionnaire regarding health behaviors, anthropometric measurements, and laboratory tests. Furthermore, BMI, waist circumference, blood pressure, fasting blood glucose, serum creatinine, triglyceride, and urine protein levels are measured after overnight fasting in accordance with the South Korean Association of Laboratory Quality Control guidelines. To receive reimbursement, medical institutions must adhere to NHIS protocols and undergo certification for quality control procedures [18].

This study was approved by the Institutional Review Board of the Samsung Medical Center (IRB No. 2023-01-006) and adhered to the Declaration of Helsinki. The requirement for obtaining informed consent from the patients was waived owing to the use of anonymized and de-identified data.

A total of 19,598 patients who had undergone KT between 2004 and 2017 were identified. A combination of the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) code Z94.0 (Kidney transplant status) and a special code for KT (V005) was used to define KT status [19]. The NHIS has registered patients through the Rare and Incurable Disease Registration Program since 2006 to identify patients requiring additional assistance for the management of rare or intractable diseases, including KT status. Patients under this registration are beneficiaries of special medical aid. KTRs who participated in the national health screening program between 2009 and 2017 (N = 9,134) were eligible for inclusion in the present study. Participants with preexisting graft failure (n = 220) and missing values (n = 371) were excluded. Furthermore, a lag period of 1 year was applied to exclude newly diagnosed cases of graft failure or death within 1 year of the national health screening date (n = 57). Thus, the final analysis included 8,486 participants. The participant selection process is shown in Figure 1.

KTRs were divided into three groups: non-DM, early PTDM, and preexisting DM. Early PTDM was defined as DM newly diagnosed between 3 months and 1 year after KT. Patients who were diagnosed with DM within the first 3 months post-transplant but did not have DM thereafter were classified as non-DM [20]. Preexisting DM was defined as a diagnosis of DM prior to KT. The diagnosis of DM was defined based on the recorded use of at least one oral antidiabetic agent or insulin, with ICD-10 codes E11–E14, or a fasting glucose level of ≥126 mg/dL.

The glomerular filtration rate (eGFR) was determined by assessing the serum creatinine-based eGFR calculated using the Modification of Diet in Renal Disease equation [21]. The dipstick test was used to assess proteinuria, with the detection of traces classified as negative. The presence of at least one claim per year for antihypertensive agents with ICD-10 codes I10–I13 or I15 or a systolic/diastolic blood pressure of ≥140/90 mmHg was defined as hypertension (HTN). The presence of at least one claim per year for antihyperlipidemic agents with ICD-10 code E78 or total cholesterol levels of ≥240 mg/dL was defined as dyslipidemia. A BMI of ≥25 kg/m² was defined as obesity. The ICD-10 codes F32 or F33 indicated a diagnosis of depression. Smoking status was categorized as never, former, or current smoker. Alcohol consumption was categorized as none, moderate (<30 g/day), and heavy (≥30 g/day). Regular exercise was defined as engaging in moderate physical activity for >30 min ≥5 times per week or engaging in strenuous physical activity for >20 min ≥3 times per week. The low-income group comprised those in the bottom 25% of earners. The requirement for rituximab, thymoglobulin, and intravenous glucocorticoids for >3 days; intravenous immune globulin; or plasmapheresis following readmission after KT was defined as acute rejection therapy.

The primary outcome measures were graft failure and all-cause mortality. Graft failure, defined as receiving dialysis at least 3 months post-KT (hemodialysis > at least 25 times or peritoneal dialysis at least 3 months), was assessed using death-censored graft failure. The follow-up period commenced on the date of the national health screening examination and ended at the occurrence of the outcome on 31 December 2021.

Subgroup analyses were conducted based on age (<40, 40–49, 50–59, 60–69, and ≥70 years old), sex, low income, smoking status (never or former vs. current), alcohol consumption, regular exercise, hypertension, dyslipidemia, depression, proteinuria, hospitalization within 1 year post-KT, and acute rejection therapy within 1 year post-KT.

The majority of previous studies defined PTDM to include DM diagnosed within the first 3 months after KT. For comparison, we performed an additional sensitivity analysis using this conventional definition, classifying all patients diagnosed with DM within 1 year post-transplant as having PTDM.

Categorical and continuous variables are presented as numbers (percentages) and mean ± standard deviation or median (interquartile range), respectively. Intergroup comparisons were conducted using a t-test (for continuous variables) and a Chi-square test (for categorical variables). The incidence rate (IR) was presented per 1,000 person-years. Cumulative graft failure and all-cause mortality were expressed using Kaplan–Meier curves and compared using the log-rank test. Hazard ratios (HR) with 95% confidence intervals (CI) were estimated using Cox proportional hazards regression analysis to estimate the risk of graft failure and all-cause mortality. Multivariable analysis was conducted for the following variables: sex, age, low income, smoking status, alcohol consumption, regular exercise, HTN, dyslipidemia, depression, BMI, eGFR, the time from KT to screening, proteinuria, induction therapy, and acute rejection within 1 year post-KT. All statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, North Carolina, USA). A two-sided P-value of <0.05 was considered statistically significant.

Table 1 presents the characteristics of KTRs at the time of participating in the national health screening program. Of the 8,486 KTRs, 1,032 (12.2%) and 2,416 (28.5%) KTRs had early PTDM and preexisting DM, respectively. The mean interval between KT and health screening was 3.04 ± 2.41 years. The mean age was 50.1 ± 10.4 years, and 5,062 (59.63%) of the KTRs were men. Patients in the preexisting DM group were the oldest and most likely to be men, followed by those in the early PTDM group. The BMI and waist circumference were highest in the preexisting DM group, followed by the early PTDM group. Systolic blood pressure was lowest in the non-DM group and similar between the early PTDM and preexisting DM groups; however, diastolic blood pressure and the proportion of HTN were highest in the early PTDM group. The eGFR was similar across the three groups, whereas the early PTDM group had the highest proportion of proteinuria. Total HDL-, and LDL-cholesterol levels were highest in the non-DM group; however, triglyceride levels and the proportion of dyslipidemia were the highest in the early PTDM group. The proportion of KTRs with low income was the lowest in the non-DM group. The median (interquartile range) follow-up durations for graft failure and all-cause mortality were 6.12 (4.22–8.40) years and 6.00 (4.06–8.14) years, respectively.

Table 2 presents the treatment history during the first year post-KT. The proportion of KTRs who did not receive induction therapy was the highest in the non-DM group (14.4%), followed by the early PTDM group (10.3%) and the preexisting DM group (6.3%). The proportion of KTRs who underwent thymoglobulin induction in the preexisting DM group (11.0%) was significantly higher than that in the non-DM (6.8%) and early PTDM (6.6%) groups. Acute rejection therapy commenced within 1 year post-KT in 7.34% of KTRs. The incidence of acute rejection was the highest in the early PTDM group (10.9%), followed by the preexisting DM (8.3%) and non-DM (6.2%) groups.

The mean number of outpatient department visits within 1 year post-KT was the highest in the preexisting DM group (29.8 ± 16.5), followed by the early PTDM (25.6 ± 12.6) and non-DM (23.0 ± 12.4) groups. Multiple hospitalizations (≥2 times) within 1 year post-KT were similar in the early PTDM (51.8%) and preexisting DM (52.0%) groups, followed by the non-DM (38.2%) group.

Compared with the non-DM group, the early PTDM group exhibited no association with the risk of graft failure but a significant association with the risk of all-cause mortality (adjusted hazard ratio [aHR] 1.044, 95% CI 0.833–1.309 for graft failure and 1.309, 95% CI 1.015–1.689 for all-cause mortality) (Table 3). Figures 2A,B present the cumulative incidences of all-cause mortality and graft failure according to the diabetic status 1 year post-KT. Compared with that in the non-DM group, an increased risk of graft failure and all-cause mortality was observed in the preexisting DM group (Table 3).

The risk of graft failure and all-cause mortality was evaluated according to changes in diabetic status from 1 year post-KT to the time of health screening (Table 4). An analysis was conducted after excluding KTRs who underwent health screening within 1 year post-KT to distinguish between early diabetic status and the changes in diabetic status. While 8.1% of non-DM patients progressed to PTDM, 33.5% of those with early PTDM and 38.6% of those with preexisting DM regressed to a non-DM status. KTRs who remained non-diabetic from KT until health screening served as the reference group.

KTRs who transitioned from non-DM to PTDM or from early PTDM to non-DM showed a significantly higher risk of graft failure in the univariable analysis. However, in the multivariable model, these transition groups did not exhibit a statistically significant increase in risk of graft failure, although transition from non-DM to PTDM tended to be associated with an increased risk (non-DM → PTDM: aHR 1.314, 95% CI 0.934–1.848; early PTDM → non-DM: aHR 1.221, 95% CI 0.818–1.824). Patients who remained in the PTDM status (early PTDM → PTDM) did not show a significantly elevated risk of graft failure. Among KTRs with preexisting DM, those who remained diabetic had a significantly higher risk of graft failure compared with those who remained non-diabetic throughout (aHR 1.545, 95% CI 1.231–1.940), whereas those who transitioned to non-DM did not show a significantly increased risk (aHR 1.298, 95% CI 0.948–1.779). Figure 2C presents the cumulative incidence of graft failure according to the changes in diabetic status.

KTRs who transitioned from non-DM to PTDM, along with those with persistent early PTDM, exhibited a significantly higher risk of all-cause mortality (non-DM → PTDM: aHR 1.646, 95% CI 1.080–2.510; early PTDM → PTDM: aHR 1.755, 95% CI 1.325–2.377) (Table 4). In contrast, KTRs who transitioned from early PTDM to non-DM showed no significant association with all-cause mortality. Among those with preexisting DM, the risk of all-cause mortality was increased regardless of subsequent changes in diabetic status, although those who transitioned from preexisting DM to non-DM exhibited a relatively lower HR for all-cause mortality compared with those who remained diabetic. Figure 2D presents the cumulative incidence of all-cause mortality according to the changes in diabetic status.

A sensitivity analysis conducted after excluding KTRs who underwent health screening within 2 years post-KT yielded comparable results, except that transitioning from non-DM to PTDM was significantly associated with graft failure (aHR 1.473, 95% CI 1.025–2.118) (Supplementary Table S1).

Figure 3 and Supplementary Tables S2, S3 present the results of the subgroup analyses. The relationship between early PTDM and graft failure or all-cause mortality exhibited no significant interaction with respect to age, sex, income, smoking status, alcohol consumption, regular exercise, HTN, dyslipidemia, proteinuria, hospitalization, or acute rejection therapy.

When the first 3 months post-KT were included in the definition of early PTDM, the number of early PTDM cases increased by 921, from 1,032 to 1,953. Under this definition, the HRs of early PTDM were attenuated and were not significantly associated with the risk of either graft failure or all-cause mortality (Supplementary Table S4). In the analysis of changes in diabetic status, HRs were amplified in both directions; that is, higher HRs became higher, and lower HRs became lower. Consequently, KTRs who transitioned from non-DM to PTDM showed a significantly increased risk of graft failure and all-cause mortality (non-DM → PTDM: aHR 1.578, 95% Cl 1.074–2.320 for graft failure and aHR 1.784, 95% Cl 1.100–2.893 for all-cause mortality). Notably, patients who transitioned from early PTDM to non-DM demonstrated an even lower risk of all-cause mortality compared with those who consistently remained non-DM (early PTDM → non-DM: aHR 0.678, 95% Cl 0.464–0.992) (Supplementary Table S5).

We further analyzed all 24,524 KTRs who underwent KT between 2004 and 2020, regardless of their participation in the national health screening program, after excluding those with missing values (n = 536) and those who died <3 months post-KT (n = 309). Supplementary Table S6 summarizes the characteristics of all KTRs who underwent KT between 2004 and 2020; the characteristics were similar to those of KTRs who participated in the health screening. Treatment history patterns of all KTRs according to the diabetic status were also similar to those of KTRs who participated in the national health screening program. However, the proportions of KTRs who underwent thymoglobulin induction (13.51% in all KTRs vs. 7.95% in KTRs who participated in the national health screening program) and acute rejection therapy within 1 year (10.69% in all KTRs vs. 7.34% in KTRs who participated in the health screening) tended to be higher. Furthermore, the mean number of hospitalizations within 1 year was higher among all KTRs compared with that among those who participated in the national health screening program (2.05 in all KTRs vs. 1.68 in KTRs who participated in the national health screening program).

Compared with non-DM, early PTDM was associated with an increased risk of graft failure and all-cause mortality in all KTRs who underwent KT between 2004 and 2020 (aHR 1.524, 95% CI 1.344–1.729 for graft failure and 1.339, 95% CI 1.153–1.535 for all-cause mortality) (Table 5).