In this study, we used a national health insurance claims database established by the KNHIS, which includes all claims data provided by the KNHIS and Medical Aid programs. Data extracted from the KNHIS database were considered representative of the entire South Korean population, and the details of this database have been previously described (18). Depending on their occupations, all insured Koreans undergo an annual or biennial health examination that is supported by the KNHIS. Anonymized data are publicly available from the National Health Insurance Sharing Service and can be accessed at https://nhiss.nhis.or.kr/bd/ab/bdaba000eng.do. The study protocol was approved by the Institutional Review Board of Chonnam National University Hospital (CNUH-EXP-2022-274) and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The requirement for written informed consent was waived by the review board because anonymous and de-identified information was used for analysis.

Initially, 38, 227 patients who underwent kidney transplantation from 2002 to 2016 were identified. Of these, we included patients who had undergone health checkups from 2009 to 2017 because the questionnaire form changed in 2009. The index date was the date of the first health check-up after 2009. We excluded those aged <20 years, and those with graft failure or death before the index date and within 1-year of follow-up. We also excluded subjects with missing health examination data. Finally, 14, 249 KTR were included in the study and were followed-up from the index date to the date of graft failure during the follow-up period, death, loss of health insurance qualification, or the end of the study period (December 31, 2019). A detailed enrollment flowchart is shown in Figure 1.

BP was measured by trained clinicians at least twice, using a mercury or automatic sphygmomanometer with the participants in a sitting position following a minimum of 5 min of rest in the index date. Hypertension was defined as SBP ≥ 130 mmHg or diastolic BP (DBP) ≥ 80 mmHg in the health examination database or a history of using antihypertensive medications according to the 2021 KDIGO BP guidelines and 2017 American College of Cardiology/American Heart Association guidelines (15, 19). Moreover, participants were classified into five hypertension groups as follows: (1) normal BP (<120/80 mmHg, patients with no prior diagnosis of hypertension); (2) elevated BP (120–129/ < 80 mmHg, but those with no prior diagnosis of hypertension); (3) incident hypertension without medication (≥130/80 mmHg, but not taking antihypertensive medications; (4) controlled hypertension (<130/80 mmHg, patients diagnosed with and taking medication for hypertension); and (5) uncontrolled hypertension (≥130/80 mmHg, patients diagnosed with and taking medication for hypertension). Participants were also classified into five groups based on their measured SBP: (1) <100 mmHg; (2) 100–119 mmHg; (3) 120–129 mmHg; (4) 130–139 mmHg; (5) ≥140 mmHg for SBP; DBP (1) <70 mmHg; (2) 70–79 mmHg; (3) 80–89 mmHg; (4) 90–99 mmHg; (5) ≥100 mmHg, as well as pulse pressure (PP) defined by SBP minus DBP (1) <40 mmHg; (2) 40–49 mmHg; (3) 50–59 mmHg; (4) 60–69 mmHg; (5) ≥70 mmHg. For each participant, the body mass index was calculated by dividing the body weight (kg) by the height squared (m²). We defined obesity as a body mass index ≥ 25 kg/m² according to the WHO recommendations for Asian populations (20). The estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease formula. Data on age, sex, health behaviour-related factors, and other definitions of smoking status, alcohol consumption, regular exercise, and diagnosis for diabetes, cardiovascular disease and dyslipidemia are described in Supplementary Table S1.

The study outcome was incident death-censored graft failure, defined as the presence of hemodialysis, peritoneal dialysis, or kidney re-transplantation. Patients with death-censored graft failure were identified using a combination of International Classification of Diseases Tenth Revision, Clinical Modification (ICD-10-CM) codes (N18–19, Z49, Z94.0, and Z99.2) and a special code (V001, procedure-related outpatient care or inpatient treatment on the day of hemodialysis; V003, peritoneal dialysis) at least three times during 3 months, and kidney transplantation code (V005). We excluded patients with a dialysis code on the same date as an acute kidney failure code (N17.9). In the event of death with a functioning graft, the follow-up period was censored at the date of death.

Continuous variables are described as the mean ± standard deviation and categorical variables are presented as numbers with proportions. Intergroup differences were tested using the chi-squared test or Student's t-test, as appropriate. The incidence rates of graft failure are presented as the number of cases calculated per 1,000 person-years. The cumulative incidence probability of graft failure was estimated using the Kaplan–Meier method, and between-group comparisons of the resulting curves were subjected to univariate analysis via the log-rank test. Multivariable analyses were performed using Cox proportional hazard regression models, and calculated hazard ratios (HRs) with 95% confidence intervals (CIs). The proportional hazards assumption was tested visually with the Schoenfeld residual plots. Model 1 was adjusted for age and sex. Model 2 was adjusted for age, sex, income level, smoking status, alcohol consumption status, physical activity, eGFR, obesity, diabetes mellitus, cardiovascular disease and dyslipidemia. Model 3 included all covariates in Model 2, along with the use of antihypertensive medications (diuretics, calcium channel blockers, β-blockers, α-blocker, angiotensin converting enzyme inhibitors, and/or angiotensin receptor blockers). Smooth cubic spline HR curves for the graft failure were plotted after adjusting for all covariates (Model 3). Subgroup analyses were conducted according to age, sex, smoking status, and diabetes as well as the duration from kidney transplantation to BP measurement. Interaction terms were added to test for effect modification across the subgroups. All statistical analyses were performed using the Statistical Analysis System (SAS) software (version 9.4; SAS Institute, Cary, NC, USA). All significance tests were 2-tailed and P-values < 0.05 were considered statistically significant.

The mean baseline age of the participants was 50.9 years, and 58.2% were men. The baseline characteristics of the study population according to hypertension are presented in Table 1. Of the total population, 11,209 (78.7%) KTR were diagnosed with hypertension. Among those with hypertension, 2949 (26.3%), 3050 (27.2%), and 5210 (46.5%) participants had incident hypertension, controlled, and uncontrolled hypertension, respectively. Participants with hypertension were more likely than those without hypertension to be male, older, smokers, take regular exercise, obese, and with a higher prevalence of diabetes, cardiovacular disease and dyslipidemia than those without hypertension.

During a mean follow-up period of 6.0 ± 2.9 years, 1934 (13.6%) participants developed graft failure. The incidence rates of graft failure were 12.2 and 25.3 (per 1,000 person-years) in patients without and with hypertension, respectively. The incidence rates of graft failure according to hypertension groups were 11.7, 13.8, 18.9, 22.8, and 31.3 for normal BP, elevated BP, incident hypertension, and controlled and uncontrolled hypertension with antihypertensive medications, respectively (Table 2). After adjusting for confounding factors (Cox Model 2), hypertensive patients had a significantly higher risk of graft failure than those without hypertension (adjusted HR, 1.703; 95% CI, 1.482–1.957). In the five hypertension groups, adjusted HRs for each group were 1 (reference), 1.198, 1.461, 1.590, and 2.127, respectively. Uncontrolled hypertension in the antihypertensive group had the highest risk for graft failure (adjusted HR, 2.127; 95% CI, 1.799–2.515). Kaplan-Meier curves for the incidence probability of graft failure according to hypertension and the five groups are shown in Figure 2, and similar results were obtained.

Participants were also classified based on SBP, DBP, and PP levels. The incidence rates and adjusted HRs (Cox Model 3) of graft failure were remarkably increased with an increase in the SBP, DBP, and PP in each group compared with the reference group (Table 3). These associations were confirmed by smooth HR curve analyses even after multivariable adjustments (Figure 3).

For subgroup analyses according to age, the participants were classified into <40, 40–65, and ≥65 years. In all subgroup analyses according to age, sex, smoking, and diabetes mellitus, hypertension was consistently associated with the risk of graft failure, and there was no significant difference between the subgroups (Figure 4).

We also analyzed the association between the five hypertension groups, the SBP and DBP groups, and PP groups with graft failure among subgroups (Supplementary Tables S2−S5). The adjusted HRs indicated no significant differences between participants regardless of age group or sex. However, the increased risk of graft failure according to advanced hypertension groups and high SBP and DBP was significantly higher in participants without diabetes than in those with diabetes.

To determine the association between the duration from kidney transplantation to BP measurement and graft failure, participants were classified into two groups; <5 years and ≥5 years based on this duration. The results were found to be consistent in both subgroups of participants (Supplementary Table S6).