ABSTRACT
Background
Bone loss and fractures are critical post‐kidney transplantation (KT) complications. We evaluated age‐stratified screening practices, treatment patterns, and fracture incidence in a nationwide KT cohort.
Methods
Using the Korean National Health Insurance Service database (2003–2020), we analyzed 34 758 adult KT recipients. Screening practices (bone mineral density [BMD] testing), medication continuity, and incident fractures were assessed across age groups (< 50, 50–64, ≥ 65 years).
Results
During a median follow‐up of 8.4 years, 63.8% of recipients underwent BMD testing, but only 29.9% were screened within the first year post‐transplant. Overall fracture incidence was 1.2/100 person‐years (PY), increasing to 3.9/100 PY in recipients aged ≥ 65 (IRR 5.41 vs. < 50 years). The excess fracture risk in older recipients persisted across all post‐transplant time intervals (adjusted SHR 4.18, 4.23, and 3.61 for < 1, 1–5, and ≥ 5 years, respectively). Despite a 5.2‐ to 7.6‐fold increase in post‐KT fracture risk associated with a pre‐transplant fracture history, screening practices were not consistently intensified among younger patients with prior fractures.
Conclusions
Post‐transplant bone health management is characterized by delayed assessment and insufficient risk‐stratified screening. Risk‐adapted strategies are necessary to improve early screening and sustained therapy, particularly for high‐risk recipients.
Keywords: bone mineral density, epidemiology, kidney transplantation, osteoporotic fractures, practice patterns
1. Introduction
Kidney transplantation (KT) is the preferred treatment for end‐stage renal disease, offering superior survival compared to dialysis [1, 2]. With significant improvements in graft survival, the clinical focus has shifted from preventing rejection to managing long‐term comorbidities [3]. Among these, Chronic Kidney Disease‐Mineral and Bone Disorder (CKD‐MBD) remains a critical yet often under‐managed challenge.
KT recipients experience rapid bone loss due to pre‐existing renal osteodystrophy and the osteotoxic effects of immunosuppressants [4, 5]. Consequently, they face a substantially increased risk of fractures compared with the general population, and fractures in this group are associated with significant morbidity and mortality [6, 7]. Although the evidence varies, observational data suggest that hip and other skeletal fractures occur at rates higher than expected for age‐matched controls [8, 9].
Although the KDIGO guidelines recommend bone mineral density (BMD) screening to assess fracture risk if results will impact treatment decisions [10], clinical practice tends to focus on graft survival as a natural priority, while bone health monitoring is often relatively neglected in the post‐transplant setting. Previous studies have demonstrated considerable variability in post‐transplant BMD testing and its impact on clinical decision‐making [11].
However, comprehensive data on long‐term bone health management in Asian KT recipients remain scarce. Therefore, this study aimed to investigate the patterns of BMD screening, medication continuity, and fracture risks in a nationwide Korean cohort spanning 18 years, providing evidence to bridge the gap in post‐transplant care.
2. Methods
2.1. Data Source and Study Design
This study was designed as a retrospective cohort study utilizing the nationwide claim database provided by the National Health Insurance Service (NHIS) of Korea. The dataset covered the period from January 1, 2002, to December 31, 2020. To ensure a sufficient washout period for ascertaining incident cases and assessing baseline comorbidities, data from 2002 were utilized as a screening and comorbidity assessment window. Consequently, the enrollment of study participants commenced on January 1, 2003, and patients were followed until death, graft failure, or the end of the study period (December 31, 2020).
This study was approved by the Institutional Review Board of Korea University Guro Hospital (IRB No. 2002GR0137). Given the retrospective design and the use of de‐identified nationwide claims data, the requirement for informed consent was waived
2.2. Study Population
We identified adult patients (aged ≥ 18 years) with chronic kidney disease (ICD‐10 codes N18, N19) who underwent KT between January 1, 2003, and December 31, 2020. KT recipients were defined as patients who had a claim for KT surgery (Electronic Data Interchange [EDI] code R3280 series) or were registered with the government's rare/intractable disease code for KT (V005) at least twice. The index date was defined as the date of KT surgery. We excluded patients who were under 18 years of age at the time of transplantation, those with a history of malignancy, and those who underwent parathyroidectomy during the 1‐year lookback period prior to the index date to minimize confounding factors affecting bone metabolism.
2.3. Assessment of Covariates and Pre‐Transplant Bone Health
Baseline demographic and clinical characteristics were assessed using claims data from the 1‐year period preceding the index date. Comorbidities such as hypertension (I10–I15) and diabetes mellitus (E11–E14) were identified based on diagnosis codes and relevant medication prescriptions. Socioeconomic status was evaluated using insurance type (Health Insurance vs. Medical Aid) and income quartiles. To evaluate pre‐transplant bone health status, we assessed the history of osteoporotic fractures within the 1‐year period preceding the index date of KT, and the performance of BMD testing within one year prior to KT. Medication history was defined as the prescription of a drug for ≥ 30 days within the 1‐year pre‐transplant period. We specifically tracked the use of corticosteroids, antidepressants/anxiety medications, and CKD‐MBD related agents, including phosphate binders (calcium acetate, calcium carbonate, and non‐calcium‐based binders), cinacalcet, and vitamin D derivatives.
2.4. Study Outcomes and Post‐Transplant Management
To provide a comprehensive evaluation of post‐transplant bone health status, we assessed three key domains: Diagnostic surveillance, therapeutic continuity, and clinical fracture outcomes. First, diagnostic patterns were evaluated by analyzing the cumulative incidence and timing of the first BMD screening following transplantation. BMD testing was ascertained using the Korean NHIS procedure codes for bone densitometry (HC34 series). Among recipients who underwent post‐transplant BMD testing, 96.5% received dual‐energy x‐ray absorptiometry (DXA), with the remainder undergoing quantitative computed tomography (QCT, 1.9%) or other modalities (1.7%). Second, we examined the longitudinal trajectories of pharmacological interventions, specifically tracking the initiation, maintenance, and discontinuation of anti‐osteoporotic agents. The medications were categorized into bisphosphonates, selective estrogen receptor modulators (SERMs), denosumab, and teriparatide. Finally, clinical fracture outcomes were identified based on the first occurrence of specific ICD‐10 diagnosis codes for osteoporotic fractures, comprising the four major osteoporotic fracture sites—hip (S72.0, S72.1), vertebra (M48.4, M48.5, S22.0, S22.1, S32.0), wrist (S52.5, S52.6), and humerus (S42.2, S42.3)—together with ankle fractures (S82.3, S82.5), which are considered a minor rather than major osteoporotic fracture site. The ICD‐10 codes used in this study are consistent with those applied in prior Korean nationwide fracture studies using NHIS or HIRA administrative data [12, 13].
2.5. Statistical Analysis
The study population was stratified into three age groups: < 50 years, 50–64 years, and ≥ 65 years. Baseline characteristics were compared across these age groups using analysis of variance (ANOVA) for continuous variables and the Chi‐square test for categorical variables. The incidence rates of fractures and BMD screening were calculated as the number of events per 100 person‐years (PY). To quantify the relative risk between subgroups (e.g., age groups, patients with vs. without fracture history), Incidence Rate Ratios (IRRs) and corresponding 95% confidence intervals (CIs) or p‐values were estimated using Poisson models. Multivariable Cox proportional hazards models for time‐to‐first fracture were constructed, adjusting for age group, pre‐transplant fracture history, sex, inclusion year, socioeconomic status, diabetes mellitus, hypertension, pre‐transplant steroid use, antidepressant/anxiolytic use, calcium/vitamin D supplementation, phosphate binders, osteoporosis medication, cinacalcet use, and pre‐transplant BMD testing. To account for the competing risk of death, Fine‐Gray subdistribution hazard models were additionally performed using the same covariate structure. To assess the robustness of our findings with respect to the pre‐transplant fracture history definition, a sensitivity analysis was performed extending the lookback period from 1 year to 3 years prior to transplantation, restricted to recipients transplanted in 2005–2020 (n = 26 423). Changes in osteoporosis medication regimens from the pre‐ to post‐transplant period were visualized using Sankey diagrams to illustrate treatment continuity and discontinuation patterns. All statistical analyses were performed using SAS Enterprise Guide (version 7.1; SAS Institute Inc., Cary, NC, USA), and a P‐value < 0.05 was considered statistically significant.
3. Results
3.1. Baseline Bone Health Characteristics of KT Recipients
Among 34 758 kidney transplant (KT) recipients, pre‐transplant bone health assessment and pharmacological intervention were infrequently performed (Table 1). A total of 354 (1.0%) had a history of fracture, and 406 (1.2%) had been treated with anti‐osteoporotic medications prior to transplantation. BMD testing was performed in 21.4% of the cohort within one year before transplantation. Age‐stratified analysis revealed that while the ≥ 65 age group had the highest prevalence of pre‐transplant fractures (3.3%) and medication history (3.7%), the BMD testing rate was highest in the 50–64 age group. This indicates a mismatch between fracture‐related risk and screening intensity: older candidates, despite a greater burden of prior fractures, were not preferentially targeted for BMD evaluation during the pre‐transplant period.
TABLE 1.
Baseline demographic and clinical characteristics of kidney transplant recipients stratified by age groups
| Characteristics | Total (N = 34 758) | Age < 50 (n = 19 480) | Age 50–64 (n = 13 014) | Age ≥ 65 (n = 2264) |
|---|---|---|---|---|
| Demographics | ||||
| Age, mean ± SD | 47.1 ± 12.0 | 38.5 ± 7.6 | 56.1 ± 4.1 | 69.5 ± 4.9 |
| Male sex, n (%) | 21 069 (60.6) | 11 694 (60.0) | 7953 (61.1) | 1422 (62.8) |
| Inclusion year | ||||
| 2003–2008 | 14 304 (41.2) | 9728 (49.9) | 4086 (31.4) | 490 (21.6) |
| 2009–2014 | 9125 (26.3) | 4961 (25.5) | 3574 (27.5) | 590 (26.1) |
| 2015–2020 | 11 329 (32.6) | 4791 (24.6) | 5354 (41.1) | 1184 (52.3) |
| Insurance type | ||||
| Health insurance | 29 882 (86.0) | 16 176 (83.0) | 11 606 (89.2) | 2100 (92.8) |
| Medical aid | 4876 (14.0) | 3304 (17.0) | 1408 (10.8) | 164 (7.2) |
| Economic status | ||||
| 1st quartile & medical aid | 10 083 (29.0) | 6321 (32.5) | 3315 (25.5) | 447 (19.7) |
| 2nd quartile | 5575 (16.0) | 3283 (16.9) | 2031 (15.6) | 261 (11.5) |
| 3rd quartile | 7606 (21.9) | 4363 (22.4) | 2770 (21.3) | 473 (20.9) |
| 4th quartile (Highest) | 10 937 (31.5) | 5222 (26.8) | 4678 (36.0) | 1037 (45.8) |
| Missing | 557 (1.6) | 291 (1.5) | 220 (1.7) | 46 (2.0) |
| Comorbidities | ||||
| Hypertension | 25 676 (73.9) | 13 715 (70.4) | 10 121 (77.8) | 1840 (81.3) |
| Diabetes mellitus | 7637 (22.0) | 2529 (13.0) | 4127 (31.7) | 981 (43.3) |
| Dialysis modality | ||||
| Hemodialysis | 16 751 (48.2) | 8486 (43.6) | 6883 (52.9) | 1382 (61.0) |
| Peritoneal dialysis | 3029 (8.7) | 1712 (8.8) | 1156 (8.9) | 161 (7.1) |
| Both | 2593 (7.5) | 1429 (7.3) | 1014 (7.8) | 150 (6.6) |
| No (Pre‐emptive) | 12 385 (35.6) | 7853 (40.3) | 3961 (30.4) | 571 (25.2) |
| Pre‐transplant medication | ||||
| Steroids | 4910 (14.1) | 3262 (16.8) | 1417 (10.9) | 231 (10.2) |
| Antidepressants/Anxiety | 1811 (5.2) | 639 (3.3) | 938 (7.2) | 234 (10.3) |
| Calcium + Vitamin D | 1004 (2.9) | 508 (2.6) | 395 (3.0) | 101 (4.5) |
| CAA | 10 921 (31.4) | 5559 (28.5) | 4582 (35.2) | 780 (34.5) |
| CAC | 10 164 (29.2) | 5396 (27.7) | 4023 (30.9) | 745 (32.9) |
| NCBPB | 5465 (15.7) | 2708 (13.9) | 2400 (18.4) | 357 (15.8) |
| Cinacalcet | 1438 (4.1) | 763 (3.9) | 599 (4.6) | 76 (3.4) |
| Post‐transplant medication (within 1 year) | ||||
| Steroids | 28 501 (82.0%) | 16 300 (83.7%) | 10 680 (82.1%) | 1521 (67.2%) |
| Pre‐transplant Bone Health | ||||
| BMD Test performed | 7424 (21.4) | 3983 (20.5) | 2950 (22.7) | 491 (21.7) |
| Prior Anti‐osteoporotic Rx | 406 (1.2) | 119 (0.6) | 203 (1.6) | 84 (3.7) |
| Bisphosphonates | 332 (1.0) | 106 (0.5) | 155 (1.2) | 71 (3.1) |
| SERM | 55 (0.2) | 10 (0.1) | 33 (0.3) | 12 (0.5) |
| Teriparatide | 1 (0.0) | 0 (0.0) | 0 (0.0) | 1 (0.0) |
| Denosumab | 28 (0.1) | 5 (0.0) | 19 (0.2) | 4 (0.2) |
| Prior Fracture History | 354 (1.0) | 102 (0.5) | 177 (1.4) | 75 (3.3) |
| Hip fracture | 60 (0.2) | 8 (0.0) | 29 (0.2) | 23 (1.0) |
| Vertebral fracture | 71 (0.2) | 13 (0.1) | 31 (0.2) | 27 (1.2) |
| Other fractures | 230 (0.7) | 82 (0.4) | 121 (0.9) | 27 (1.2) |
Note: Data are presented as mean ± standard deviation for continuous variables and number (percentage) for categorical variables. All baseline characteristics differed significantly across age groups (all p <0.001).
Abbreviations: BMD, bone mineral density; CAA, calcium acetate; CAC, calcium carbonate; DM, diabetes mellitus; HTN, hypertension; KT, kidney transplantation; NCBPB, non‐calcium‐based phosphate binders; SERM, selective estrogen receptor modulator
3.2. Post‐Transplant BMD Testing Patterns
During a median follow‐up of 3083 days (IQR 1360–5409), post‐transplant bone health outcomes including BMD screening, medication use, and fracture incidence varied significantly by age (Table 2). Specifically, 63.8% of patients underwent BMD testing (14.6/100 person‐years [PY]). The testing rate was significantly higher in the ≥ 65 age group (23.0/100 PY) than in the 50–64 group (IRR 1.20, 95% CI 1.13–1.27, p < 0.001) and the <50 group (IRR 1.89, 95% CI 1.78–2.00, p < 0.001). However, only 29.9% of the total cohort received their first BMD scan within the first year post‐transplant. The median time to the first BMD test was similar across all age groups (≥ 65 age: 363.0 days; 50–64: 367 days; <50: 382 days), indicating that for more than half of tested recipients the initial assessment was performed at approximately one year post‐transplant or later (Figure 1A).
TABLE 2.
Post‐transplant bone health management patterns and fracture incidence stratified by age groups
| Variables |
Total (N = 34 758) |
Age < 50 (n = 19 480) |
Age 50–64 (n = 13 014) |
Age ≥ 65 (n = 2264) |
|---|---|---|---|---|
| Follow‐up duration | ||||
| Days, median (IQR) | 3083 (1360–5409) | 3989 (1973–5886) | 2313 (1070–4168) | 1238 (568–2372) |
| BMD Screening | ||||
| Performed, n (%) | 22 173 (63.8) | 12 461 (64.0) | 8487 (65.2) | 1225 (54.1) |
| Incidence rate (per 100 PY) | 14.6 | 12.2 | 19.2 | 23.0 |
| Time to 1st test, days (IQR) | 374 (37–1064) | 382 (53–1232) | 367 (25–825) | 363 (45–688) |
| Tested within 1 year * , n (%) | 10 407 (29.9) | 5637 (28.9) | 4152 (31.9) | 618 (27.3) |
| Osteoporosis Medication | ||||
| Initiated post‐KT, n (%) | 5408 (15.6) | 2415 (12.4) | 2580 (19.8) | 413 (18.2) |
| Incidence rate (per 100 PY) | 1.9 | 1.3 | 3.1 | 4.7 |
| Time to initiation, days (IQR) | 1234 (415–2975) | 1820 (509–3664) | 1019 (391–2454) | 579 (215–1496) |
| Fracture Incidence | ||||
| Overall Fracture, n (%) | 3484 (10.0) | 1424 (7.3) | 1702 (13.1) | 358 (15.8) |
| Hip fracture, n (%) | 641 (1.8) | 183 (0.9) | 336 (2.6) | 122 (5.4) |
| Vertebral fracture, n (%) | 1030 (3.0) | 301 (1.6) | 574 (4.4) | 155 (6.9) |
| Other fracture, n (%) | 2226 (6.4) | 1044 (5.4) | 1039 (8.0) | 143 (6.3) |
| Incidence rate (per 100 PY) | 1.2 | 0.7 | 1.9 | 3.9 |
| Time to fracture, days (IQR) | 1990 (834–3512) | 2474 (1160–3988) | 1782 (755–3313) | 1064 (371–2121) |
Note: Data are presented as number (percentage) or median (interquartile range [IQR]). Incidence rates are expressed as events per 100 person‐years (PY).
Abbreviations: BMD, bone mineral density; KT, kidney transplantation. Other fractures include wrist (S52.5, S52.6), humerus (S42.2, S42.3), and ankle (S82.3, S82.5).
Percentages for BMD screening performed and tested within 1 year are calculated using the full cohort as the denominator. Time to first test is reported among recipients who underwent BMD testing.
FIGURE 1.
Timing and cumulative rates of the first Bone Mineral Density (BMD) assessment after kidney transplantation. (A) The stacked bar chart presents the proportion of patients undergoing their first BMD scan, stratified by age groups (< 50, 50–64, and ≥ 65 years). The total height of each bar represents the overall assessment rate, while the segments indicate the timing of the first scan: within 1 year (dark slate), 1–5 years (steel blue), and more than 5 years post‐transplant (light grey). Numbers within the segments denote the absolute percentage of the total population. (B) Cumulative incidence functions (CIF) for time‐to‐first BMD testing stratified by age group, incorporating death and graft failure as competing events. Differences across age groups were statistically significant (Gray's test p < 0.001).
Cumulative incidence function (CIF) analysis incorporating death and graft failure as competing events demonstrated that the cumulative incidence of BMD testing differed significantly across age groups (Gray's test p <0.001). Notably, despite having the highest fracture risk, the ≥ 65 age group showed the lowest cumulative incidence of BMD testing throughout the follow‐up period, with a long‐term testing rate of approximately 54.1% compared with 64.0% and 65.2% in the < 50 and 50–64 age groups, respectively (Figure 1B).
Among recipients who underwent at least one post‐transplant BMD test (n = 22 173), 61.4% received a second BMD measurement during follow‐up. The median interval between the first and second BMD test was 456 days (approximately 15 months) overall (< 50: 472 days; 50–64: 441 days; ≥ 65: 429 days). Repeat testing was least frequent in the ≥ 65 age group, with 47.4% of elderly recipients not receiving a second BMD measurement, compared with 36.9% and 39.7% in the < 50 and 50–64 age groups, respectively.
3.3. Incidence and Timing of Post‐Transplant Osteoporotic Fractures
The overall incidence of post‐transplant osteoporotic fractures was 1.2/100 PY, with a median time to the first fracture of 1990 days (IQR 834–3512) (Table 2). The incidence rate was 3.9/100 PY in the ≥ 65 age group, 1.9/100 PY in the 50–64 group, and 0.7/100 PY in the < 50 group. The risk in the elderly was 5.4‐fold higher than in those aged < 50 (IRR 5.41 (95% CI 4.82–6.08), p < 0.001).
On multivariable Cox proportional hazards analysis, the age‐associated fracture risk differential remained robust after full adjustment. Compared with recipients aged < 50 years, those aged 50–64 years had a 2.5‐fold higher fracture risk (HR 2.538, 95% CI 2.357–2.733) and those aged ≥ 65 years had a 5.0‐fold higher risk (HR 5.013, 95% CI 4.425–5.679). When death was incorporated as a competing event using the Fine‐Gray model, subdistribution hazard ratios were attenuated but remained significant (SHR for 50–64 vs. < 50: 2.281, 95% CI 2.116–2.458; SHR for ≥ 65 vs. <50: 3.006, 95% CI 2.632–3.433), reflecting the higher mortality rate in elderly recipients that reduces their cumulative probability of experiencing fracture (Table 3).
TABLE 3.
Multivariable Cox proportional hazards and Fine‐Gray subdistribution hazard models for post‐transplant fracture risk
| Cox proportional hazards model | Fine‐Gray subdistribution hazard model | |||||
|---|---|---|---|---|---|---|
| Age group | HR | 95% CI | p‐value | SHR | 95% CI | p‐value |
| < 50 years | ref | ref | ||||
| 50–64 years | 2.538 | 2.357–2.733 | < 0.001 | 2.281 | 2.116–2.458 | < 0.001 |
| ≥ 65 years | 5.013 | 4.425–5.679 | < 0.001 | 3.006 | 2.632–3.433 | < 0.001 |
Note: Models were adjusted for pre‐transplant fracture history, sex, inclusion year, socioeconomic status, diabetes mellitus, hypertension, pre‐transplant steroid use, antidepressant/anxiolytic use, calcium/vitamin D supplementation, phosphate binders, osteoporosis medication, cinacalcet use, and pre‐transplant BMD testing.
The Fine‐Gray subdistribution hazard model incorporated death and graft failure as competing events.
Abbreviations: CI, confidence interval; HR, hazard ratio; SHR, subdistribution hazard ratio.
Furthermore, time‐interval stratified Fine‐Gray analyses demonstrated that the elevated fracture risk associated with older age persisted across all post‐transplant time periods. In recipients aged ≥ 65 years, the SHR relative to those aged < 50 years remained consistently elevated within the first year (SHR 4.181, 95% CI 3.023–5.784), between 1–5 years (SHR 4.233, 95% CI 3.478–5.153), and beyond 5 years post‐transplant (SHR 3.609, 95% CI 2.903–4.487). Recipients aged 50–64 years similarly maintained a stable excess fracture risk across all time intervals (SHR 2.250, 2.538, and 2.324, respectively). These findings indicate that older age confers a sustained, time‐independent fracture vulnerability throughout the post‐transplant course, beyond what is explained by differences in fracture timing alone (Table 4). Fracture site‐specific analyses demonstrated consistent age‐associated risk gradients for both hip and vertebral fractures (Table S1).
TABLE 4.
Subdistribution hazard ratios for fracture risk by age group and post‐transplant time interval (adjusted Fine‐Gray model)
| Post‐transplant follow‐up < 1 year | Post‐transplant follow‐up 1–5 years | Post‐transplant follow‐up ≥ 5 years | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Age group | SHR | 95% CI | p‐value | SHR | 95% CI | p‐value | SHR | 95% CI | p‐value |
| < 50 years | ref. | — | ref. | — | ref. | — | |||
| 50–64 years | 2.250 | 1.763–2.871 | < 0.001 | 2.538 | 2.230–2.890 | < 0.001 | 2.324 | 2.104–2.568 | < 0.001 |
| ≥ 65 years | 4.181 | 3.023–5.784 | < 0.001 | 4.233 | 3.478–5.153 | < 0.001 | 3.609 | 2.903–4.487 | < 0.001 |
Note: Models were adjusted for pre‐transplant fracture history, sex, inclusion year, socioeconomic status, diabetes mellitus, hypertension, pre‐transplant steroid use, antidepressant/anxiolytic use, calcium/vitamin D supplementation, phosphate binders, osteoporosis medication, cinacalcet use, and pre‐transplant BMD testing.
Death and graft failure were incorporated as competing events.
Abbreviations: CI, confidence interval; SHR, subdistribution hazard ratio.
3.4. Patterns of Osteoporosis Medication Use and Discontinuity
The post‐transplant osteoporosis medication initiation rate was 1.9/100 PY (15.6% of the total cohort). Although the ≥ 65 age group had the highest treatment rate (4.7/100 PY), among the 406 patients receiving pre‐transplant osteoporosis therapy (1.2% of the total cohort), 44.8% did not receive osteoporosis medication within the first post‐transplant year. Discontinuation was most frequent in the elderly (50.6%) (Figure 2). Maintenance rates did not significantly differ by drug class (Bisphosphonates 55.1%, Denosumab 25.0%, SERM 36.4%; p = 0.304).
FIGURE 2.
Transitions in osteoporosis medication from pre‐ to post‐kidney transplantation. The Sankey diagrams illustrate the longitudinal flow of pharmacological treatment across three age cohorts: < 50, 50–64, and ≥ 65 years. Note: For clarity of visualization and to focus on active clinical transitions, patients who did not receive any anti‐osteoporotic treatment during both the pre‐ and post‐transplant periods (none‐to‐none) were excluded from this diagram. The width of the flows is normalized to 100% of the active treatment/transition group within each cohort. Color Key: Bisphosphonate (Blue), Denosumab (Red), SERM (Green), Teriparatide (Orange), and None (Grey).
Among patients without prior osteoporosis medication, those who initiated post‐transplant therapy had a higher fracture risk (adjusted SHR 1.811, 95% CI 1.643–1.996, p < 0.001). Among prior users (n = 406), no significant difference in fracture risk was observed by post‐transplant treatment continuation status (adjusted SHR 1.192, 95% CI 0.760–1.871, p = 0.445), though the limited sample size precluded a definitive conclusion.
3.5. Temporal Trends in Bone Health Management
Calendar‐period stratified analyses demonstrated a gradual but incomplete improvement in bone health management over time. The rate of BMD testing within 1 year post‐transplant increased progressively from 25.7% in 2003–2008 to 36.2% in 2018–2020, and osteoporosis medication initiation rates increased from 1.7 to 4.7/100 PY overall, with the most pronounced increase in elderly recipients (≥ 65 years: 3.9 to 8.3/100 PY). However, even in the most recent era (2018–2020), fewer than 40% of recipients received BMD testing within the first post‐transplant year. Overall fracture incidence did not decline across the study period (1.1 to 1.6/100 PY), likely reflecting the progressive demographic shift toward older, higher‐risk recipients over time (2003–2008: 3.4% aged ≥ 65; 2018–2020: 11.6%) (Table S2).
3.6. Impact of Pre‐Transplant Fracture on Post‐Transplant Outcomes
Pre‐transplant fracture history was strongly associated with an increased risk of post‐transplant fracture across all age groups (≥ 65 years: IRR 5.17, 95% CI 3.46–7.49; 50–64 years: IRR 5.67, 95% CI 4.47–7.11; < 50 years: IRR 7.58, 95% CI 5.39–10.37; all p < 0.001) (Figure 3A). Despite this markedly elevated risk, screening intensity did not consistently differ according to fracture history (Figure 3B). In recipients aged ≥ 65 years, those with prior fractures underwent more frequent BMD testing than those without (38.4 vs. 22.6 per 100 PY; IRR 1.70, 95% CI 1.24–2.28, p < 0.001). However, among recipients aged < 65 years, BMD testing rates were not significantly higher in those with prior fractures (50–64 years: IRR 1.16, 95% CI 0.96–1.40, p = 0.114; < 50 years: IRR 1.26, 95% CI 0.99–1.59, p = 0.053).
FIGURE 3.
Disparity between post‐transplant fracture risk and diagnostic BMD screening rates. (A) Post‐transplant Fracture Risk: Demonstrates a stark contrast in actual fracture incidence. Patients with a prior fracture history exhibit a significantly higher risk (up to 7.6‐fold higher incidence) compared to those without. (B) Post‐transplant BMD Testing: Shows the incidence rates of BMD screening according to pre‐transplant fracture history. While screening rates slightly increase with age, the disparity between patients with (dark blue) and without (light grey) prior fractures is not statistically significant in younger cohorts (ns). Note: *** indicates p < 0.001.
In a sensitivity analysis extending the fracture history lookback period to 3 years prior to transplantation, results were consistent with the primary analysis (HR for ≥ 65 vs. < 50: 4.876, 95% CI 4.198–5.663; SHR: 3.256, 95% CI 2.780–3.813). The association between pre‐transplant fracture history and subsequent fracture risk remained strong under the 3‐year definition (IRR 9.82, 6.42, and 5.37 in the < 50, 50–64, and ≥ 65 groups, respectively; all p <0.001), while BMD testing rates were again not significantly higher in recipients with a prior fracture among those aged < 65 years (IRR 1.27, p = 0.115 in the < 50 group; IRR 1.06, p = 0.577 in the 50–64 group), in contrast to the ≥ 65 group (IRR 1.95, p <0.001) (Table S3).
4. Discussion
This nationwide cohort study, based on an 18‐year longitudinal dataset, demonstrates that bone health management in kidney transplant (KT) recipients is characterized by delayed early evaluation and insufficient risk‐stratified screening. The overall osteoporotic fracture incidence was 1.2 per 100 person‐years (PY). However, fewer than one‐third of recipients (29.9%) underwent their first BMD assessment within the first year after transplantation, with a median time‐to‐first test of 363–382 days. Importantly, although a history of pre‐transplant fracture was associated with a markedly elevated risk of post‐transplant fracture (5.2–7.6‐fold increase), BMD testing rates were not consistently higher in these high‐risk individuals, particularly among recipients aged < 65 years. Together, these findings suggest that risk‐stratified screening may not be fully integrated into routine post‐transplant care, despite the clearly identifiable fracture burden.
The overall fracture incidence observed in our cohort (1.2 per 100 person‐years [PY]) falls within the range reported in previous nationwide Korean studies. In a KOTRY‐based nationwide cohort of 4134 kidney transplant recipients, Eum et al. reported 63 fractures over 12 441 PY (approximately 0.51 per 100 PY), with a 5‐year cumulative incidence of 2.10% [14]. In another population‐based study using the HIRA database, Kim et al. reported a fracture incidence of 25.2 per 1,000 PY (2.52 per 100 PY) among kidney transplant recipients within the broader kidney replacement therapy population [15]. While our incidence lies between these two national estimates, our study extends the observation period from 2003 to 2020, thereby capturing temporal shifts in transplant practice, evolving immunosuppressive strategies, and the demographic transition toward older transplant recipients. The wide range across these estimates likely reflects differences in cohort composition and follow‐up duration rather than true variation in underlying fracture risk.
The distribution of fracture events differed by age group, with a greater proportion of fractures occurring early after transplantation in older recipients. However, this pattern is largely explained by the substantially shorter follow‐up duration in older recipients (median 1,238 days in the ≥ 65 group vs. 3989 days in the < 50 group; Table 2), which mechanically limits the accumulation of late fracture events in this group. Multivariable Fine‐Gray analyses accounting for competing risks and differential follow‐up demonstrated that the relative fracture risk associated with older age remained consistently elevated across all post‐transplant time intervals, indicating that elderly recipients face a sustained, time‐independent excess fracture risk throughout their post‐transplant course. This underscores the need for continuous—rather than early‐only—bone health monitoring in elderly transplant recipients. Despite these clear risks, the median time to the first BMD test exceeded one year (363–382 days). Furthermore, among recipients who underwent initial BMD testing, only 61.4% received a repeat measurement during follow‐up, with the lowest repeat testing rate observed in the ≥ 65 age group (47.4% with no repeat testing)—the group at highest fracture risk. Although several specialty guidelines address post‐transplant bone health, specific and operationalized screening timelines remain insufficient or inconsistently defined [16, 17]. While the 2017 KDIGO CKD‐MBD guidelines recommend BMD testing in transplant recipients with risk factors for osteoporosis, and suggest consideration of antiresorptive therapy within the first 12 months in patients with low BMD and eGFR > 30 mL/min/1.73m2 [10], these recommendations are not uniformly implemented in real‐world practice. Although eGFR values were not available in the NHIS database, the majority of recipients with a functioning graft would be expected to meet the eGFR > 30 mL/min/1.73m2 threshold, suggesting that the observed undertreatment is not primarily attributable to renal function contraindications. Notably, while KDIGO advocates for “early assessment” of fracture risk in all transplant recipients [17, 18], no specific timeline is defined—an ambiguity that may itself contribute to the delayed screening patterns observed in our cohort. Furthermore, although the 2017 KDIGO CKD‐MBD guidelines and the 2021 European Consensus Statement recommend BMD assessment in patients with CKD stage 5D prior to transplantation [18, 19]—providing a critical baseline for post‐transplant management—only 21.4% of recipients in our cohort underwent BMD testing within 1 year prior to transplantation, suggesting that this ideal clinical workflow is not consistently implemented in practice. Together, these findings underscore a clinically meaningful gap: In many patients, BMD evaluation occurs after the phase of rapid early bone loss, potentially reducing the opportunity for timely preventive intervention.
Temporal analyses revealed a gradual improvement in bone health management over the study period, with BMD screening within 1 year increasing from 25.7% to 36.2% and osteoporosis medication initiation rates rising from 1.7 to 4.7/100 PY. However, even in the most recent era, fewer than 40% of recipients received BMD testing within the first post‐transplant year, suggesting persistent underimplementation of guideline‐recommended bone health surveillance. Moreover, overall fracture incidence did not decline across eras (1.1 to 1.6/100 PY), likely reflecting the concurrent demographic shift toward older, higher‐risk recipients—with the proportion aged ≥ 65 years increasing from 3.4% in 2003–2008 to 11.6% in 2018–2020, consistent with global and domestic trends toward older transplant recipients [20, 21]. These findings suggest that the absolute burden of post‐transplant fractures may continue to grow unless bone health management is substantially intensified in parallel with the aging of the transplant population.
Furthermore, our study highlights a significant clinical blind spot regarding younger recipients with a history of pre‐transplant fracture. Despite a 7.6‐fold increased risk of subsequent fracture, this group did not receive significantly more BMD screenings compared to those without such history. A key challenge in this context is the limited availability of validated risk stratification tools for younger transplant recipients: FRAX is validated only for individuals aged 40–90 years and cannot be applied to younger patients, leaving clinicians with fewer evidence‐based methods to identify high‐risk individuals in this age group. Although FRAX has been validated to predict fracture risk in KT recipients, its discriminative accuracy for major osteoporotic fractures is modest (AUC 0.62) [22], and its practical application is further restricted by a national reimbursement system that prioritizes absolute BMD T‐scores over clinical risk factors. A risk‐stratified approach that incorporates pre‐transplant fracture history—regardless of patient age—as a trigger for intensified post‐transplant bone health surveillance may help address this gap.
The clinical consequences of post‐transplant fractures extend beyond bone health. A recent Korean national cohort study demonstrated that incident osteoporotic fractures after KT were associated with 1.4‐fold increased death‐censored graft failure and 2.2‐fold higher mortality risk compared to non‐fracture KT recipients, underscoring fractures as a critical threat to both graft and patient survival [23]. Unlike many unpredictable post‐transplant complications, osteoporotic fractures are a manageable risk through proactive CKD‐MBD screening and the optimization of anti‐osteoporotic therapy [24]. While the primary goal of transplantation is graft survival, the prevention of fractures is essential to preserve the patient's functional status and quality of life [23, 25].
A key constraint of this study is the absence of individual‐level BMD T‐score values, which limits our ability to determine whether the observed management patterns reflect informed clinical decision‐making or true gaps in care. Nevertheless, the primary objective was to characterize systemic patterns of bone health care delivery at the national level—an objective achievable without individual BMD data. Future studies linking claims data with clinical registry data containing BMD results and FRAX scores would be needed to more precisely characterize the relationship between individual fracture risk and management decisions.
This study has several limitations. First, the NHIS claims database does not capture actual BMD T‐score values or diagnostic classifications; therefore, we were unable to determine the proportion of tested patients classified as osteoporotic, nor whether treatment discontinuation reflected intentional de‐escalation or a true management failure. Second, a valid comparison of fracture risk by treatment status was limited by indication bias, as post‐transplant osteoporosis therapy was preferentially prescribed to higher‐risk patients. Furthermore, the subgroup of patients with prior osteoporosis medication use was too small (n = 406) to draw definitive conclusions regarding the benefit of treatment continuation after transplantation; larger studies with adequate power would be required to address this question. Third, fracture ascertainment relied solely on ICD‐10 diagnosis codes without radiographic confirmation, and some misclassification cannot be excluded—particularly for vertebral fractures identified incidentally. Fourth, post‐transplant immunosuppression data including cumulative corticosteroid doses and calcineurin inhibitor type were not available in the claims database, which may have introduced residual confounding. Fifth, biochemical CKD‐MBD markers such as serum PTH and 25(OH)D, as well as pre‐transplant dialysis duration, were not captured and represent potential unmeasured confounders. Sixth, ESKD etiology could not be reliably ascertained from the NHIS claims database due to potential misclassification inherent in administrative coding, and was therefore not included as a covariate in our models.
In this long‐term nationwide cohort, post‐transplant bone health management in Korea was characterized by delayed early risk assessment, and insufficient risk‐stratified screening. Despite a measurable fracture burden—particularly the early clustering of fractures in elderly recipients and the markedly elevated risk among those with prior fractures—early BMD evaluation and maintenance of osteoporosis therapy were not consistently implemented. These findings highlight the need for more explicit, transplant‐specific bone health recommendations that clearly define the timing of early BMD assessment and strategies for maintaining pharmacologic therapy in high‐risk patients. Strengthening guideline clarity and promoting their implementation in routine clinical practice may help bridge the gap between fracture risk and preventive care in kidney transplant recipients.
Author Contributions
Ji Eun Kim: Conceptualization, Methodology, Software, Data curation, Formal analysis, and Writing – original draft. Chan Mi Park: Data curation, Formal analysis, and Visualization. Seung Min Song: Supervision and Validation. Min Woo Kang: Supervision and Validation. Hyo Jin Kim: Supervision and Validation. Eun Jung Cho: Supervision and Validation. Gang Jee Ko: Supervision and Validation. Young Joo Kwon: Conceptualization, Supervision, Project administration, and Writing – review & editing. All authors have read and agreed to the published version of the manuscript.
Funding
Ji Eun Kim was supported by a grant from the Korean Nephrology Research Foundation (Grant No. Q2427931, 2024) and by a grant from the Patient‐Centered Clinical Research Coordinating Center (PACEN), funded by the Ministry of Health & Welfare, Republic of Korea (Grant No. RS‐2025‐02304904).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting information: ctr70704‐sup‐0001‐Tables S1‐S3.xlsx
Data Availability Statement
The data used in this study are available from the National Health Insurance Service (NHIS) through a formal application process, but restrictions apply to their public availability.
References
- 1. Wolfe R. A., Ashby V. B., Milford E. L., et al., “Comparison of Mortality in all Patients on Dialysis, Patients on Dialysis Awaiting Transplantation, and Recipients of a First Cadaveric Transplant,” New England Journal of Medicine 341, no. 23 (1999): 1725–1730. [DOI] [PubMed] [Google Scholar]
- 2. Laupacis A., Keown P., Pus N., et al., “A Study of the Quality of Life and Cost‐Utility of Renal Transplantation,” Kidney International 50, no. 1 (1996): 235–242. [DOI] [PubMed] [Google Scholar]
- 3. Hariharan S., Johnson C. P., Bresnahan B. A., Taranto S. E., McIntosh M. J., and Stablein D., “Improved Graft Survival After Renal Transplantation in the United States, 1988 to 1996,” New England Journal of Medicine 342, no. 9 (2000): 605–612. [DOI] [PubMed] [Google Scholar]
- 4. Evenepoel P., Claes K., Meijers B., et al., “Bone Mineral Density, Bone Turnover Markers, and Incident Fractures in De Novo Kidney Transplant Recipients,” Kidney International 95, no. 6 (2019): 1461–1470. [DOI] [PubMed] [Google Scholar]
- 5. Canalis E., Mazziotti G., Giustina A., and Bilezikian J. P., “Glucocorticoid‐Induced Osteoporosis: Pathophysiology and Therapy,” Osteoporosis International 18, no. 10 (2007): 1319–1328. [DOI] [PubMed] [Google Scholar]
- 6. Naylor K. L., McArthur E., Leslie W. D., et al., “The Three‐Year Incidence of Fracture in Chronic Kidney Disease,” Kidney International 86, no. 4 (2014): 810–818. [DOI] [PubMed] [Google Scholar]
- 7. Chen H., Lai Y. R., Yang Y., et al., “High Risk of Osteoporosis and Fracture Following Solid Organ Transplantation: A Population‐Based Study,” Front Endocrinol (Lausanne) 14 (2023): 1167574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Sidibe A., Auguste D., and Desbiens L. C., “Fracture Risk in Dialysis and Kidney Transplanted Patients: A Systematic Review,” JBMR Plus 3, no. 1 (2019): 45–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Jaikaransingh V., “Bone Health and Fracture Prevention After Kidney Transplantation,” Journal of Clinical and Translational Endocrinology 36 (2024): 100345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ketteler M., Block G. A., Evenepoel P., et al., “Executive Summary of the 2017 KDIGO Chronic Kidney Disease‐Mineral and Bone Disorder (CKD‐MBD) Guideline Update: What's Changed and Why It Matters,” Kidney International 92, no. 1 (2017): 26–36. [DOI] [PubMed] [Google Scholar]
- 11. Naylor K. L., Zou G., Leslie W. D., et al., “Frequency of Bone Mineral Density Testing in Adult Kidney Transplant Recipients From Ontario, Canada: A Population‐Based Cohort Study,” Canadian Journal of Kidney Health and Disease 3 (2016): 2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Ahn S. H., Park S. M., Park S. Y., et al., “Osteoporosis and Osteoporotic Fracture Fact Sheet in Korea,” Journal of Bone Metabolism 27, no. 4 (2020): 281–290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Yu T. Y., Cho H., Kim T. Y., Ha Y. C., Jang S., and Kim H. Y., “Utilization of Osteoporosis‐Related Health Services: Use of Data From the Korean National Health Insurance Database 2008‐2012,” Journal of Korean Medical Science 33, no. 3 (2018): e20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Eum S. H., Kim D. W., Lee J. H., et al., “Time‐Varying Risk Factors for Incident Fractures in Kidney Transplant Recipients: A Nationwide Cohort Study in South Korea,” Journal of Clinical Medicine 12, no. 6 (2023): 2337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kim Y., Lee E., Lee M. J., Park B., and Park I., “Characteristics of Fracture in Patients Who Firstly Starts Kidney Replacement Therapy in Korea: A Retrospective Population‐Based Study,” Scientific Reports 12, no. 1 (2022): 3107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Camacho P. M., Petak S. M., Binkley N., et al., “American Association of Clinical Endocrinologists/American College of Endocrinology Clinical Practice Guidelines for the Diagnosis and Treatment of Postmenopausal Osteoporosis‐2020 Update,” Endocrine Practice 26, no. 1 (2020): 1–46. [DOI] [PubMed] [Google Scholar]
- 17. Kidney Disease: Improving Global Outcomes Transplant Work G , “KDIGO Clinical Practice Guideline for the Care of Kidney Transplant Recipients,” American Journal of Transplantation 9, no. 3 (2009): S1–155. [DOI] [PubMed] [Google Scholar]
- 18. Kidney Disease: Improving Global Outcomes CKDMBDUWG “KDIGO 2017 Clinical Practice Guideline Update for the Diagnosis, Evaluation, Prevention, and Treatment of Chronic Kidney Disease‐Mineral and Bone Disorder (CKD‐MBD),” Kidney International Supplements 7, no. 1 (2017): 1–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Evenepoel P., Cunningham J., Ferrari S., et al., “European Consensus Statement on the Diagnosis and Management of Osteoporosis in Chronic Kidney Disease Stages G4‐G5D,” Nephrology, Dialysis, Transplantation 36, no. 1 (2021): 42–59. [DOI] [PubMed] [Google Scholar]
- 20. Park S., Kim M., Kim J. E., et al., “Characteristics of Kidney Transplantation Recipients Over Time in South Korea,” Korean Journal of Internal Medicine 35, no. 6 (2020): 1457–1467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Li Y., Menon G., Wu W., et al., “Evolving Trends in Kidney Transplant Outcomes among Older Adults: A Comparative Analysis Before and During the COVID‐19 Pandemic,” Transplant Direct 9, no. 12 (2023): e1520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Naylor K. L., Leslie W. D., Hodsman A. B., Rush D. N., and Garg A. X., “FRAX Predicts Fracture Risk in Kidney Transplant Recipients,” Transplantation 97, no. 9 (2014): 940–945. [DOI] [PubMed] [Google Scholar]
- 23. Jang Y., Kim J. E., Park J., et al., “Clinical Significance of Incident Osteoporotic Fractures After Kidney Transplantation: A National Korean Cohort Study,” American Journal of Kidney Diseases 85, no. 4 (2025): 409–420. [DOI] [PubMed] [Google Scholar]
- 24. Henry N., Hildebrand S., and Cunningham J., “Management of Osteoporosis in Chronic Kidney Disease,” Clinical Kidney Journal 18, no. 10 (2025): sfaf228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Salter M. L., Liu X., Bae S., et al., “Fractures and Subsequent Graft Loss and Mortality Among Older Kidney Transplant Recipients,” Journal of the American Geriatrics Society 67, no. 8 (2019): 1680–1688. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting information: ctr70704‐sup‐0001‐Tables S1‐S3.xlsx
Data Availability Statement
The data used in this study are available from the National Health Insurance Service (NHIS) through a formal application process, but restrictions apply to their public availability.