Objectives: Kidney transplant recipients are at increased risk for avascular necrosis due to steroid use and accompanying comorbidities. Concerning risk factors, uncertainty still exists. We evaluated the clinical characteristics and risk factors of avascular necrosis in kidney transplant recipients.
Materials and Methods: Symptomatic avascular necrosis was found by magnetic resonance imaging in 33 of 360 kidney transplant patients between 2005 and 2021. The patients’ clinical characteristics, biochemical testing, and medications were evaluated.
Results: We found the frequency of avascular necrosis to be 9.7% during the follow-up period. If the total steroid dosage used was more than 4 g in the first 3 months, the risk of developing avascular necrosis increased 4.08 times, and the presence of cytomegalovirus disease increased the risk by 4.03 times. Avascular necrosis was observed bilaterally in 60.6% of cases and at the femoral head in 66.7%. The frequency of avascular necrosis was highest in the first and second years posttransplant.
Conclusions: We found that avascular necrosis appears most frequently in the first 2 years after kidney transplant and the most important risk factors are cumulative steroid dose and cytomegalovirus disease. In the follow-up of kidney transplant patients, it is important to use low-dose steroid doses if possible. Of note, preventing the development of cytomegalovirus disease by screening and prophylaxis for cytomegalovirus is also important in reducing the development of avascular necrosis.
Key words : Bulent Kaya, Saime Paydas, Mustafa Balal, Burak Mete, Tolga Kuzu
In addition to renal osteodystrophy, osteoporosis, bone fracture, and avascular necrosis (AVN) are seen in the posttransplant period in kidney transplant recipients.1 Avascular necrosis is an important bone complication after kidney transplant that can have a considerable negative effect on a patient’s function and eventually result in femoral collapse.2 The incidence of AVN has gradually decreased with the development of modern immunosup-pressants and the reduction of steroid doses. Compared with earlier reports, the incidence of AVN has decreased markedly to about 5%.3-7 Avascular necrosis is associated with a variety of conditions and treatments. Decreased perfusion leading to ischemia and eventually infarction of the bone and marrow are common pathophysiological mechanisms among most etiologies.2 Corticosteroids are the most well-known important risk factor,8 although risk factors remain generally unclarified.9,10 Spontaneous regression is limited in AVN patients, and most patients require arthroplasty.11 We investigated the risk factors and frequency of AVN in our kidney transplant recipients.
Materials and Methods
All patients who underwent kidney transplant in our university hospital and who had follow-up in the nephrology outpatient clinic between 2005 and 2021 were included in our study (n = 360). Patients who had missed laboratory data and had sickle cell anemia (n = 194) were not compared with the AVN-positive group for laboratory parameters. Medical records such as risk factors for AVN, detection time, affected areas, and treatment were obtained from patients’ files and the hospital information systems. Biopsy-proven acute rejection and immunosuppressive treatments over patient follow-up were recorded. Demographic factors (age, body mass index, duration of dialysis, dialysis type, alcohol and smoking), viral infections (BK virus, cytomegalovirus [CMV]), comorbidities (diabetes mellitus, hypertension), antihypertensive drugs, donor and recipient characteristics, and immunosuppressive treatments were evaluated.
Evaluation of bone mineral density and avascular necrosis
Patients with discomfort, mobility restrictions, and walking problems after orthopedic examination received magnetic resonance imaging to confirm AVN. Bone mineral density (BMD) was measured with a dual-energy radiography absorptiometry placed on the lumbar spine and femoral neck at the time of diagnosis.
To determine risk factors among AVN patients, we collected information on complete blood count and biochemical tests after transplant, including creatinine, uric acid, calcium, phosphorus, parathormone (PTH), 25(OH) vitamin D, C-reactive protein, and lipids. Serum creatinine-based estimated glomerular filtration rate (eGFR) was used to calculate glomerular filtration rate.x12
According to KDIGO clinical practice guidelines for kidney transplantation, antithymocyte globulin and basiliximab should be used as induction therapies in patients with high and low immunological risk, respectively. The combination of tacrolimus (or cyclosporine A), mycophenolate mofetil (MMF) (azathioprine), and steroid constituted the main-tenance therapy.13,14 In the induction treatment, 500 mg methylpred-nisolone was intravenously administered on the first day and 100 mg/day for the next 5 days. In the following days, levels were gradually reduced, and the maintenance dose (4 mg/day) was reached at month 3. Target trough levels of tacrolimus were 10 to 12 ng/mL for the first 3 months, 8 to 10 ng/mL for 3 to 6 months, and 6 to 8 ng/mL thereafter. Targeted cyclosporine levels (hour 2) were 800 to 1200 ng/mL in the first 3 months, 800 to 1000 ng/mL from 3 to 6 months, and 400 to 600 ng/mL thereafter. The target dose of MMF was 2 × 360 mg in patients under 65 kg and 2 × 720 mg in patients above 65 kg. Acute cellular rejection was treated with intravenously methylprednisolone at 250 to 500 mg for 3 to 5 days (±antithymocyte globulin, 3-5 mg/kg).
Our university hospital’s ethics committee provided the approval (no. June 3, 2022-123/19). Our research was carried out in accordance with the Declaration of Helsinki.
We used SPSS version 22 software to evaluate the data. For normally distributed variables, we used the Shapiro-Wilks test. Values are shown as an arithmetic mean and standard deviation. Nonparametric tests were utilized for both normally distributed data and nonparametric data. We used the chi-square test, t test, Mann-Whitney U test, binary logistic regression analysis, multinomial logistic regression analysis, and Kaplan-Meier survival analysis. Results were considered statistically significant if P was .05 or less.
Of 360 patients, we excluded patients with missing laboratory data and sickle cell anemia (n = 194). Of 166 remaining kidney transplant patients, 33 were diagnosed with AVN, and these patients were compared with 133 patients without AVN (Table 1). None of the 194 excluded patients had AVN. Among 166 patients, 118 (71.1%) were male, with a mean age of 35.7 ± 11.8 years. In the median follow-up of 2286 days (range, 70-5600 d), 9 patients (5.4%) died. In our study, the incidence of AVN was 9.2% (33/360). Avascular necrosis was bilateral in 20 of 33 patients (60.6%) and unilateral in 13 of 33 patients (39.4%). Disease was localized to the femoral head in 22 patients (66.7%), the knee in 7 patients (21.2%), and both the femoral head and knee joint in 4 patients (12.1%). Median time to detection for AVN was 292 days (range, 70-292 d). Among 33 patients, AVN was detected within the first year posttransplant in 22 patients (66.7%) and within the first 2 years in 28 patients (84.8%). Most of the 33 patients with AVN were treated conservatively, with the exception of 6 patients who underwent total hip replacement. The follow-up period ranged from 1 to 16 years. Our patients had no kidney dysfunction or surgerical complications after kidney transplant. In the AVN group, the prevalence of CMV disease was significantly higher than in the group without AVN (P = .009). There was no significant difference in biochemical parameters between groups with and without AVN (Table 2). In univariate logistic regression analyses to predict AVN risk factors, cumulative steroid doses over 4000 mg increased the risk 5.85 times, CMV disease increased the risk 5.68 times, and each 0.266 unit increase in hemoglobin decreased the risk 1.30 times (odds ratio = 7.67). In a multivariate analysis performed using the factors identified as significant in the univariate analyses, cumulative steroid dose and the presence of CMV disease were shown to be significant predictors of AVN. Steroid dosage over 4000 mg increased the risk of AVN by 4.08 times, and the presence of CMV disease increased the risk by 4.03 times (Table 3). Kaplan-Meier analysis evaluated the cumulative steroid dose and the presence of CMV disease with the risk of developing AVN. The mean time to develop-ment of AVN was observed to be shorter in patients treated with a steroid dose of 4000 mg or higher and in patients with CMV disease (P < .001) (Table 4). In a multinomial logistic regression analysis to predict whether AVN is unilateral or bilateral, the model fit was adequate (Pearson P = .848, deviance P = .859), and the factors included in the model (steroid dosage and CMV disease) affected bilateral AVN (Table 5). The risk of bilateral AVN was 8.19 times lower in individuals who received a total steroid dosage of 4000 mg and 7.93 times lower in people who did not have CMV disease ((Figure 1) and (Figure 2)).
During the 16-year follow-up period, the incidence of AVN was 9.2% in our kidney transplant recipients. We also found that cumulative steroid dosage and CMV disease were the most important factors in the development of AVN. We also observed that the annual incidence of AVN after transplant was 0.84 percent (9.2/11 years). This incidence was similar to that shown in recent publications (?1%).10,15 Before the 2000s, more kidney transplant patients had AVN. In a retrospective analysis that included 750 patients observed for 27 years (1968-1995), the incidence of AVN was 11.2%.16 A frequency of 24% was reported in 262 renal transplant recipients between 1971 and 1982.4 However, the incidence of AVN has decreased below 5% in recent years.5,6,10 For example, 1.5%, 2.2%, and 3.1% were detected in 300, 805, and 681 renal transplant recipients, respectively.6,7,10 The frequency of AVN was thought to decrease with the introduction of new immuno-suppressive drugs and the reduction of the steroid dose.x5,10 In our investigation, a slightly higher rate (9.2%) than that reported in recent publications was found. The frequency or center differences in AVN prevalence may be affected by donor-recipient risk status, immunosuppressive treatment protocols (including steroids), and posttransplant duration. Musculoskeletal complications such as AVN are well known in kidney transplant recipients.6 Avascular necrosis most frequently affects the hip joint, but it can affect practically any bone. Trauma, alcohol use, and steroids are the 3 most significant causes. Sickle cell anemia, radiation exposure, and Gaucher disease are among other factors.2,7,17,18 Possible contributors to the pathogenesis of AVN include fat embolism, vascular thrombosis, and osteocyte death.2,19 We detected no significant difference between people with and without AVN with regard to primary renal disease and/or concomitant disease. However, we did not determine the cause of end-stage kidney disease in 78 (47%) of our patients. On the other hand, only 11 people (6.6%) in the whole group had diabetes mellitus. As a result, we may have overlooked key risk factors like diabetes.20 More detailed research is necessary to solve this issue. The median duration to diagnose AVN after transplant in our study was 292 days. However, we also had patients with AVN diagnosed in as little as 70 days and as much as 2298 days. In 2 other studies, the median time of AVN detection for 765 and 750 kidney transplant patients was 25 months and 20.5 months, as in our research.15,16 However, AVN may appear years after transplant; for example, in a study of 326 kidney transplant patients, AVN was diagnosed at a median of 3.5 years after transplant (range, 0.5-13 years).21 Studies with a short follow-up duration may overlook the diagnosis of AVN that may arise in the late period, thus decreasing the likelihood of recognizing the true incidence of AVN.6 In our study, for example, AVN was found in 22 patients (66.7%) during the first year of post-transplant and in 6 (18.1%) patients between the first and second years. However, in 1 patient who had intravenous pulse steroid therapy after rejection, AVN was found 2298 days later. In summary, about 85% of our AVN patients were recognized to have AVN during the first 2 years, as also reported by Felten and colleagues.10 However, diagnosis can happen years after transplant if high-dose pulse steroid therapy is used. Avascular necrosis was seen in the femoral head and knee of our kidney transplant patients. It was more common in the femoral head, as in previous research.10,21,22 We did not identify it in humeral heads or wrists of any of the study patients. Total hip replacement was performed in 6 patients (18.2%) with AVN. Before the surgery, 4 of the 6 patients could walk with walkers; after surgery, all no longer required assistive devices. Only 1 of 6 patients reported hip pain after lengthy walks (about 1 km). Total hip replacement seemed safe and improved functional status of patients without causing any renal dysfunction or any other complications. The remaining 27 patients (81.8%) were followed conservatively, including taking bisphosphonates and vitamin D. This suggests that early diagnosis and palliative care may delay the need for arthroplasty. The incidence of osteonecrosis in individuals who are using glucocorticoids has been reported to be between 5% and 40%, with larger doses of glucocorticoids and long-term treatment carrying a higher risk. However, short-term exposure to high doses can lead to osteonecrosis, including intra-articular injection.17,23 The cumulative steroid dose in the first 2 months after kidney transplant has been associated with AVN, but the risk has been shown to rise even earlier, within the first 2 weeks.24,25 Similar to our research, Felten and colleagues reported that the cumulative total steroid dose in the first 3 months after transplant increased the risk of AVN (by 1.55 times).10 We found that AVN risk increased 4.08 times in patients who received 4 g or above of cumulative steroid therapy in the third month after transplant, with association for bilateral AVN. The total amount of steroids necessary to cause AVN is unclear. For example, in a meta-analysis (23?561 people, 57 studies), the risk of AVN increased 2-fold for a cumulative dose greater than 10 g compared with less than 10 g.26 In another study, the risk of AVN was significantly increased in patients who received a mean posttransplant first-year cumulative steroid dose of 12.5 g compared with 6.5 g.16 Steroids are important in AVN because they have a negative impact on osteoblasts and stimulate osteoclast activity, resulting in rapid and early bone loss.27 These findings suggest that steroid doses should be kept as low as possible in the posttransplant period, and patients who are taking long-term steroid therapy in the pretransplant period should be monitored for AVN (eg, patients with antiglomerular basement membrane disease-associated vasculitis, systemic autoimmune diseases, and second or third kidney transplant). We found that CMV disease raised the incidence of AVN by 4.03 times. Furthermore, CMV disease was one of the most significant factors in the development of bilateral AVN. One of the etiologies for AVN is viral infections.28-30 Multiple organ systems, including the musculoskeletal system, are affected directly or indirectly by CMV infection. The risk of venous thrombosis and the acceleration of atherosclerosis have both been linked to CMV infection.31-33 The underlying process might be arterial-venous ischemia, followed by necrosis in the affected region. In addition to discontinuing or decreasing antimetabolites and/or calcineurin inhibitors, an increased dose of steroids may be used to avoid rejection during acute CMV infection. However, the relationship between CMV infection and AVN has not been well-explored in the literature. Because of pretransplant renal osteodystrophy and posttransplant medications, the bone structure after a transplant is heterogeneous. Decreasing the pretransplant high levels of PTH and phosphorus and increasing vitamin D and calcium levels may provide some improvement in the posttransplant period.1 In the early posttransplant period, we found no link between PTH, 25(OH) vitamin D, calcium, and phosphorus levels and AVN. However, we did not follow PTH levels prospectively in the posttransplant period as performed by Felten and colleagues.10 Prospective studies are required on this issue. Cyclosporine can cause high-turnover bone loss and tacrolimus increases bone resorption and leads to bone loss.34 Cyclosporine may cause osteonecrosis due to its vasoconstrictive effects and sirolimus by potentiating the calcineurin inhibitor effect or by affecting the lipid profile.35 Reports have both stated that AVN is related to cyclosporine treatment15,36 or not, such as in our study.6,10 In contrast to immunosuppressive drug types like calcineurin inhibitors and MMF, the cumulative steroid dose in our study had the greatest effect on AVN. In renal transplant patients with glucocorticoid-induced osteoporosis, BMD loss occurs more rapidly (6%-12%) in the first year and more slowly (3% loss) thereafter.37 However, within the first 3 months after steroid therapy, the risk of fracture increases up to 75% without significant reduction in BMD, suggesting that glucocorticoids have adverse effects on bone that cannot be detected by bone densitometry.38 The relationship between low BMD and AVN is unclear.39,40 In our study, we did not find a relationship between BMD levels after kidney transplant and AVN. We also found that low hemoglobin levels increased the risk of AVN; however, this was not seen in multivariate analysis. Lower hemoglobin was found related with AVN in renal transplant recipients.41 Avascular necrosis is a known risk factor in sickle cell anemia patients; however, people with sickle cell anemia were not included in our study’s analysis.42 There are a few potential explanations for the relationship between anemia and AVN: inade-quate blood supply to the tissue, which might result in ischemia, inflammation, or an accidental finding. Single-center, retrospective, and inadequate patient populations are the key limitations of our study. Inadequate patient numbers also restricted analysis of potential risk factors, including alcohol use and diabetes. Magnetic resonance imaging was performed only in symptomatic patients, and routine screening was not performed to detect these patients. As a result, people with asymptomatic AVN may have been ignored. In conclusion, we found that AVN after kidney transplant is most common in the first 2 years. Cumulative steroid dose (>4 g) at first 3 months after kidney transplant and CMV disease are significant predictors of posttransplant AVN. Thus, clinicians should focus CMV and AVN screening in this high-risk population and, if possible, prescribe lower steroid dosages to avoid AVN.
DOI : 10.6002/ect.2022.0345
From the 1Cukurova University Faculty of Medicine, Department of Nephrology; and the 2Cukurova University Faculty of Medicine, Department of Public Health, Adana, Turkey
Acknowledgements: The authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential conflicts of interest.
Corresponding author: Bulent Kaya, Cukurova University Faculty of Medicine, Department of Nephrology, 013330, Sar?cam, Adana, Turkey
Table 1. Demographic, Clinical, and Laboratory Features of Patients
Table 2. Laboratory Results of Study Patients
Table 3. Univariate and Multivariate Logistic Regression Analyses of Predictors of Avascular Necrosis
Table 4. Kaplan-Meier Analysis for Month 3 Cumulative Steroid Dose and Cytomegalovirus Disease
Figure 1. Relationship Between Month 3 Cumulative Steroid Dose and Avascular Necrosis
Figure 2. Relationship Between Month 3 Cumulative Steroid Dose and Cytomegalovirus Disease
Table 5. Multinomial Logistic Regression Analysis for Unilateral Versus Bilateral Avascular Necrosis