Objectives: The typical immunosuppressive regimen of hematopoietic stem cell transplant includes cyclosporine. However, cyclosporine nephrotoxicity is a concern. We studied cyclosporine nephrotoxicity epidemiology in hematopoietic stem cell transplant patients and compared the pattern and urinary levels of the KIM-1 kidney injury molecule versus serum and urine creatinine levels.
Materials and Methods: The study covered 10 months at Namazi Hospital, Shiraz, Iran. All patients met the following criteria: > 15 years old, received allogenic hematopoietic stem cell transplant without history of acute or chronic kidney disease, and scheduled for at least 1 week of cyclosporine treatment. Urinary and serum levels of creatinine, urea, sodium, potassium, magnesium, and the KIM-1 kidney injury molecule were measured on days 0, 3, 5, 7, 10, and 14 of cyclosporine treatment.
Results: Of 42 patients, one-third developed cyclosporine nephrotoxicity (30.95%), and median onset time was 15 days. Hypokalemia and hypomagnesemia were reported in 76.2% and 53.4% of the cohort, respectively. None of the demographic, clinical, and paraclinical parameters was significantly associated with cyclosporine nephrotoxicity. Median duration of hospital stay for patients with cyclosporine nephrotoxicity (41 days) was significantly higher (P < .001) than those without nephrotoxicity (29 days). Area under the curve for receiver operating characteristic showed that accuracy of serum creatinine (0.267; 95% CI, 0.11-0.43) at day 0 of cyclosporine treatment was significantly lower (P = .017) than the accuracy of urine creatinine (0.477; 95% CI, 0.28-0.67) and urine levels of the KIM-1 kidney injury molecule (0.594; 95% CI, 0.41-0.78).
Conclusions: Cyclosporine nephrotoxicity is a common adverse effect in the setting of hematopoietic stem cell transplant and occurs mostly within the first 2 weeks of cyclosporine treatment. Urine KIM-1 kidney injury molecule measurement had no overall superiority and no improved accuracy over serum or urine creatinine measurements for prediction or detection of cyclosporine nephrotoxicity.
Key words : Acute kidney injury, Allogeneic stem cell transplant, Immunosuppressive agents
For many hematologic and nonhematologic diseases, hematopoietic stem cell transplant (HSCT) is the main therapeutic modality and provides either cure or prolonged survival.1 The Eastern Mediterranean Blood and Marrow Transplantation (EMBMT) group from 9 countries of the World Health Organization Regional Office for the Eastern Mediterranean, including Iran, reported 3546 first HSCT for the years 2011 to 2012. Most of the reported HSCT were allogeneic (62%).2 Ramzi and colleagues described their 15-year experience (from May 1993 to October 2008) with HSCT at their center in Shiraz, Iran. A total of 423 patients underwent either allogeneic HSCT (n = 311) or autologous HSCT (n = 112) at this center.3
Cyclosporine, a calcineurin inhibitor, is an important immunosuppressive agent in the therapeutic regimen of HSCT. It is most often used for prevention of graft-versus-host disease in allogeneic HSCT. Although cyclosporine has a major role in HSCT, its use is complicated by a narrow therapeutic index and various adverse reactions, including both acute and chronic nephrotoxicity.4 Cyclosporine acute nephrotoxicity presents as increased serum creatinine, decreased glomerular filtration rate (GFR), and imbalanced electrolytes (eg, hypo- and hyperkalemia and hypomagnesemia).5 Studies have shown that nephrotoxicity in HSCT recipients is common (with the frequency up to 70%), and it has been associated with increased rates of short-term and long-term mortality.6
We considered the nephrotoxic potential of cyclosporine and the lack of data for cyclosporine nephrotoxicity in the HSCT setting, and the aim of the present study was to determine the incidence, onset time, and factors associated with the incidence of nephrotoxicity and electrolyte disturbances of cyclosporine. In addition, we assessed the usefulness of the KIM-1 kidney injury molecule as a novel biomarker of renal function and compared this with the usefulness of serum and urine creatinine measurements for detection of acute kidney injury during treatment with cyclosporine in allogeneic HSCT recipients in Iran.
Materials and Methods
Study setting and sample size
This cross-sectional observational study was performed within a 10-month period from April 2018 to February 2019 at the Bone Marrow Transplant Unit of Namazi Hospital, affiliated with the Shiraz University of Medical Sciences in Shiraz, Iran. The medical ethics committee of the university approved the study (IR.SUMS.REC.1398.088), and written informed consent was obtained from all patients.
For planning, we considered α = .05, power of study = 80% (1 - β = .8), and the average incidence of cyclosporine nephrotoxicity to be 20%; therefore, we calculated that the sample size should be at least 30.
Study population and patient selection
The following criteria were used to evaluate candidates for the study: (1) age > 15 years; (2) no documented history of acute kidney injury (defined as an increase in serum creatinine ≥ 0.3 mg/dL within 48 hours, an increase in serum creatinine by ≥ 1.5 times the baseline within the prior 7 days, or urine volume < 0.5 mL/kg/h for 6 hours)7 or chronic kidney disease (defined as creatinine clearance below 60 mL/min/1.73 m2 or documented history of regular peritoneal or hemodialysis for more than 3 months)8; (3) absence of any prominent clinical condition that may cause or aggravate renal impairment in the setting of HSCT (eg, graft-versus-host disease, hepatic sinusoidal obstruction syndrome, tumor lysis syndrome, sepsis, severe hypovolemia due to diarrhea or vomiting); (4) no documented history of treatment with cyclosporine within the past 14 days; and (5) willingness and ability to receive either oral or intravenous cyclosporine for at least 1 week.
The following data were obtained from the patients’ medical records: (1) demographic data, including age, sex, weight, and height; and (2) clinical data, including underlying hematologic-oncologic disease, conditioning regimen, cyclosporine daily dose, duration of cyclosporine treatment, cyclosporine trough level, types of potentially nephrotoxic medications administered (eg, aminoglycosides, amphotericin B, furosemide), length of hospital stay, and mortality.
Biochemical and serologic measurements
Serum creatinine, sodium, and urea were monitored daily until hospital discharge or death of patient, in accordance with the routine practice of the bone marrow transplant ward at Namazi Hospital. At days 0, 3, 5, 7, 10, and 14 of cyclosporine treatment, spot urine samples were collected for determination of sodium, urea, creatinine, and KIM-1. Urine samples were stored at -80 °C in a freezer (New Brunswick Scientific) until completion of the sampling procedure for the cohort. An autoanalyzer (Shanghai Xunda Medical Instrument Co.) was used for measurement of creatinine, sodium, and urea in both serum and urine. Urine levels of KIM-1 were determined by the double-sandwich enzyme-linked immunosorbent assay technique (Bioassay Technology Laboratory, Shanghai Korain Biotech Co.).
Cyclosporine nephrotoxicity was defined as increase in serum creatinine level to 2-fold or greater from baseline.6 Either fractional excretion of sodium > 2% or fractional excretion of urea > 50% (in cases of loop diuretic treatment) was considered an indicator for acute tubular necrosis (ATN).9 Apart from the above definition for cyclosporine nephrotoxicity, acute kidney injury in the cohort was also determined by the Risk, Injury, Failure, Loss of Kidney Function, and End-Stage Kidney Disease (RIFLE) criteria.10 Baseline GFR was calculated by the abbreviated Modification of Diet in Renal Disease formula.
Electrolyte imbalances were defined as serum potassium < 3 mEq/L (hypokalemia), serum potassium level ≥ 5.5 mEq/L (hyperkalemia), and serum magnesium level < 1.2 mEq/L (hypomagnesemia).11 A urine potassium-to-creatinine ratio above 13 mEq/g in the presence of hypokalemia was considered to be renal potassium wasting.12 Renal magnesium wasting was determined as the fractional excretion of magnesium equal to or above 4% in patients with hypomagnesemia.13
We recorded all interventions for the management of cyclosporine nephrotoxicity, including daily dose reduction, alternate day dosing, discontinuance of the agent, dialysis treatment, and change of medication.
We performed all statistical analyses with the Statistical Package for the Social Sciences (SPSS, version 20). Continuous data were expressed as either means ± SD or as medians with interquartile range. Categorical variables were reported as percentages. Chi-square or the Fisher exact test was used to evaluate possible associations among categorical variables. Continuous variables were analyzed by either an independent t test or the Mann-Whitney U test. We used the univariate and multivariate logistic regression analyses with odds ratio (OR) and 95% CI to determine associated factors of cyclosporine nephrotoxicity. In univariate analysis, we assessed the possible association with cyclosporine nephrotoxicity (as dependent variable) of each independent variable, including age, sex, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), baseline GFR value, daily dose of oral and intravenous cyclosporine, conditioning regimen, coadministration of aminoglycosides, amphotericin B, vancomycin, and furosemide. Variables with P < .05 were thereafter considered together for multivariate logistic regression analysis. We used 1-way analysis of variance with repeated measures for comparison of the mean values of serum creatinine, urine creatinine, and urine KIM-1 at days 0, 3, 5, 7, 10, and 14 of treatment within and between patients with and without cyclosporine nephrotoxicity. We used the receiver operating characteristic curves of sensitivity and specificity to compare the accuracy for detection of cyclosporine nephrotoxicity by the above markers on treatment days 0, 3, 5, 7, and 10, and relevant data were expressed in terms of area under the curve (AUC) and 95% CI. Statistical significance in all analyses was defined as P < .05.
During the study period, there were 48 patients in the primary screen. Among them, 6 individuals were younger than 15 years old, and these patients were excluded. Therefore, 42 patients were finally recruited into the study.
More than one-half of the cohort were men (54.8%). The mean age of the study population was 32.29 ± 9.37 years. The 3 most common underlying hematologic-oncologic diseases were acute myeloid leukemia (38.1%), acute lymphoid leukemia (31%), and aplastic anemia (7.1%). The mean baseline GFR value was 122.05 ± 33.08 mL/min. Conditioning regimens of the study population were busulfan + cyclophosphamide (n = 37), antithymocyte globulin + cyclophosphamide (n = 4), and busulfan + fludarabine (n = 1). Amphotericin B, vancomycin, and furosemide were administered to 39, 20, and 4 patients, respectively. No patients received aminoglycoside or acyclovir. The mean daily dose of oral cyclosporine was 241.96 ± 57.72 mg, and the mean daily dose of intravenous cyclosporine was 168.81 ± 30.06 mg.
Among the study population, 13 patients (30.95%) developed cyclosporine nephrotoxicity. The median onset time of cyclosporine nephrotoxicity was 15 days (range, 5-35 days). In contrast, no patient experienced ATN. We applied the RIFLE criteria to classify patients into categories of risk (n = 15; 35.71%), injury (n = 11; 26.19%), and failure (n = 1; 2.38%). We used univariate analysis to select for sex (OR, 4.73; 95% CI, 1.05-17.42; P = .043), BMI (OR, 0.82; 95% CI, 0.69-0.99; P = .034), average oral daily dose of cyclosporine (OR, 0.97; 95% CI, 0.95-0.99; P = .004), and average intravenous daily dose of cyclosporine (OR, 0.93; 95% CI, 0.88-0.98; P = .004). We concurrently considered these variables in the multivariate logistic regression model, and none of these was significantly associated with cyclosporine nephrotoxicity (Table 1).
The median length of hospital stay for patients with cyclosporine nephrotoxicity was 41 days (interquartile range, 8.75 days) and for patients without nephrotoxicity was 29 days (interquartile range, 4 days). This difference was both clinically and statistically significant (P < .001). The mortality rate was higher in patients with cyclosporine nephrotoxicity (7.69%) than in patients without nephrotoxicity (3.44%). Although this difference appears large (more than double), it was not statistically significant (P = .561). Dose adjustment in the case of cyclosporine nephrotoxicity was performed in 4 patients. No patients received emergent hemodialysis or continuous renal replacement therapy, and there were no cases for which nephrotoxicity required discontinuance of cyclosporine treatment.
More than three-fourths of the study population (76.2%) experienced at least 1 episode of hypokalemia, and one-half of the study population (52.4%) experienced at least 1 episode of hypomagnesemia. In contrast, no patients developed hyperkalemia. Potassium wasting was observed in 31% of the cohort. However, magnesium wasting was not identified in any patient. For patients with cyclosporine nephrotoxicity the median daily potassium supplementation was 54.67 mEq (interquartile range, 25.74 mEq), and for patients without nephrotoxicity this was 54.11 mEq (interquartile range, 28.13 mEq). The mean daily magnesium supplementation in patients with and without cyclosporine nephrotoxicity was 16.67 ± 3.5 mEq and 14.68 ± 5.12 mEq, respectively. These values were comparable between the 2 groups.
Figure 1 depicts the changing pattern of serum creatinine, urine creatinine, and urine KIM-1 values at the studied time points during cyclosporine treatment. The overall change in the mean value of serum creatinine during the course of cyclosporine treatment was statistically significant among patients with and without cyclosporine nephrotoxicity (P < .001). However, the overall changes in the mean values of serum creatinine compared between 2 groups (0.087; 95% CI, -0.47 to 0.22) were not significant (P = .195). The overall changes in the mean value of urine creatinine levels during the course of cyclosporine treatment were not statistically significant within or between 2 groups (P = .132 and P = .799, respectively). Interestingly, the overall change in the mean value of urine KIM-1 levels within patients with and without cyclosporine nephrotoxicity was statistically significant (P = .037). However, the mean difference in mean urine KIM-1 between 2 groups (0.087; 95% CI, -0.74 to 0.28) was comparable (P = .372).
Analysis of receiver operating characteristic curves of sensitivity and specificity indicated that AUC of serum creatinine (0.267; 95% CI, 0.11-0.43) at day 0 of cyclosporine treatment was significantly lower (P = .017) than that of urine creatinine (0.477; 95% CI, 0.28-0.67) and urine KIM-1 (0.594; 95% CI, 0.41-0.78). However, the accuracy of studied markers for detection of cyclosporine nephrotoxicity was comparable at other time points, including at days 3, 5, 7, 10, and 14 (Figure 2 and Table 2).
In the present study, the frequency of cyclosporine nephrotoxicity was 30.95%. Three other studies have also considered the nephrotoxicity of cyclosporine in hematologic-oncologic settings in Iran. Tavakoli Ardakani and colleagues evaluated the utilization pattern of cyclosporine in 35 patients who received allogenic HSCT at the Taleghani Hospital in Tehran. Nephrotoxicity was observed in 20% of the cohort.14 Another study in the same hospital reported that 4 of 22 cyclosporine recipients (18.18%) developed nephrotoxicity.15 Finally, a prospective study with a follow-up of 10 years at the Hashemi Nejad Hospital in Tehran showed that, among 14 HSCT patients who developed glomerulopathy after transplant, cyclosporine nephrotoxicity was detected in 1 patient (7%).16 Da Silva and colleagues, in a systematic review of 19 studies, investigated the role of cyclosporine on the occurrence of nephrotoxicity after allogeneic HSCT. The incidence of cyclosporine nephrotoxicity in the included studies was more than 30%.6 The difference in the rate of cyclosporine nephrotoxicity in different studies may be justified by the variation in the conditioning regimen, the duration of patient follow-ups, the nephrotoxic agents that were coadministered (eg, aminoglycosides and amphotericin B), and the definition of nephrotoxicity. With regard to the definition of cyclosporine nephrotoxicity, the definition we used in the present survey (ie, increase in serum creatinine level equal to or greater than 2-fold from the patient’s baseline value) is among the most common in recent studies.6
Hypokalemia was detected in more than three-fourths (76.2%) of the study population and hypomagnesemia in more than one-half (52.4%). In addition, no patient experienced a hyperkalemic episode. Electrolyte disorders are common adverse effects of cyclosporine, mostly reported from solid-organ transplant settings.5 Data on this topic in HSCT patients are rare. Caliskan and colleagues reported 4 cases of hyperkalemia due to cyclosporine in allogenic HSCT recipients.17 Hyperkalemia has been also reported in 2 patients with advanced renal cell carcinoma who underwent HSCT during the course of cyclosporine treatment.18 In an observational study of 36 allogenic HSCT patients who received either cyclosporine or tacrolimus, Aisa and colleagues reported that all patients experienced at least 1 episode of hypomagnesemia within the first 4 weeks after transplant.19 In the present study, we found no cases of hyperkalemia, and this is a novel finding that has not previously been reported for HSCT. We coadministered amphotericin B (92.86% of patients), vancomycin (47.6%), and furosemide (9.52%), and this may partially explain this novel result. In this regard, Karimzadeh and colleagues reported that the frequency of hypokalemia in their hematology-oncology and bone marrow transplant centers was 45% in patients treated with amphotericin B and 40.38% in patients treated with vancomycin.20,21
In the present survey, there was no significant association between different demographic and clinical characteristics of patients and cyclosporine nephrotoxicity. This also was true for cyclosporine trough level despite its higher value in patients with cyclosporine nephrotoxicity than those without nephrotoxicity. This recent finding was in contrast to at least 1 study in HSCT that clearly demonstrated a correlation between serum levels of cyclosporine and creatinine (P < .001).22 Although the relationship between some aspects of cyclosporine nephrotoxicity (eg, afferent arteriole vasoconstriction) and its dose, especially more than 5 mg/kg/day, has been reported,23,24 the significant association of its either daily dose or trough level with cyclosporine nephrotoxicity is not conclusive. Nephrotoxicity of calcineurin inhibitors including cyclosporine can occur even in patients with cyclosporine trough levels within the therapeutic range or even on the low side.25 For example, in a retrospective cohort study of 212 kidney transplant patients who received a modified formulation of cyclosporine (Neoral), the authors reported that neither dose nor trough concentration of cyclosporine was a predictor of cyclosporine nephrotoxicity.26
In their systematic review, Da Silva and colleagues reported that the coadministration of nephrotoxic agents, especially aminoglycosides (P = .01) and amphotericin B (P < .01), is an independent risk factor for cyclosporine nephrotoxicity in HSCT patients.6 Similarly, 2 studies from Iran indicated that most of the patients who developed nephrotoxicity had received either amphotericin B or aminoglycosides concurrently with cyclosporine.14,15 However, we found no significant association between coadministration of nephrotoxic agents and cyclosporine nephrotoxicity. This may be partially caused by the presence of type II error secondary to limited sample size. Also, it is noteworthy that none of our patients was given aminoglycosides during the study period.
Results of receiver operating characteristic curve analyses demonstrated that urine KIM-1 had no superiority over serum or urine creatinine in detection of cyclosporine nephrotoxicity. In addition, repeated measures analyses revealed that the changing pattern of urine KIM-1 during 2 weeks of cyclosporine treatment did not differ significantly between patients with and without nephrotoxicity. A number of experimental studies have assessed urine KIM-1 as a marker of kidney function in the setting of cyclosporine nephrotoxicity.27-32 For example, in a rat model of cyclosporine nephrotoxicity, urine KIM-1 increased both in the early (7 days) and late phases (21 days) of nephrotoxicity.29 Sanchez-Pozos and colleagues also demonstrated that cyclosporine (15 mg/kg/day subcutaneously for 15 days) was associated with tubular apoptosis and tubulointerstitial fibrosis along with an increase in the expression of KIM-1 in the kidney tissue.30 Despite the promising results of animal investigations, there are few clinical studies on KIM-1 as a marker of kidney function in the setting of drug nephrotoxicity, and the results from those studies are inconclusive. Two studies in pediatric patients showed that urine KIM-1 was superior to urine N-acetyl-D-glucosaminidase, neutrophil gelatinase-associated lipocalin, and interleukin-18 for detection of cisplatin and gentamicin nephrotoxicity.33,34 In the study by Vaidya and colleagues, 2 patients with malignant mesothelioma developed postoperative acute kidney injury as a result of intraoperative local cisplatin treatment.35 Both preoperative and 6-hour postoperative urine samples were negative for KIM-1. However, 24 hours after surgery, peak urinary KIM-1 levels (800 pg/mL) were identified.35 Positive findings have been also reported by Shinke and colleagues in lung cancer patients who received cisplatin.36 In contrast to these studies, 2 other investigations by Karimzadeh and colleagues failed to show any superiority of urine KIM-1 over either serum or urine creatinine measurements for detection of amphotericin B and vancomycin nephrotoxicity.37,38 The negative results in our present study can be attributed to small sample size, limited measurements of urine KIM-1, and the use of certain serum creatinine cut points rather than the use of a more standardized method for determination of nephrotoxicity. Also, it is noteworthy that none of our patients developed ATN during the course of cyclosporine treatment. That is, the fact that there was no significant difference in the level of urine KIM-1 between patients with and without cyclosporine nephrotoxicity may be because of the absence of ATN, as KIM-1 generally is secreted and released from the proximal tubule cells in the case of ATN.39 Interestingly, Saburi and colleagues reported that, among 50 allogenic HSCT patients treated with cyclosporine, not one experienced a significant increase in the fractional excretion of sodium within 4 weeks after transplant. The authors concluded that the effect of cyclosporine on fractional excretion of sodium is minimal.40
The strength of our survey was shown by our consideration of the different aspects of cyclosporine nephrotoxicity while simultaneously including frequency, pattern (both glomerular and tubular), onset time, and risk factors; the clinical outcome in the HSCT setting; and, for the first time in humans, the comparison of urine KIM-1 versus classic markers of renal function with regard to detection of cyclosporine nephrotoxicity. In contrast, the major drawbacks of this study were the relatively small sample size (mostly due to exclusionary criteria), the limited number of urine KIM-1 measurements for each patient (mostly due to limits on financial resources), the use of certain serum creatinine cut points rather than a standardized method (such as inulin as an exogenous agent, or kidney biopsy) to prove nephrotoxicity (for which we were limited by both practical and financial restrictions), and our inability to exclude all factors that may be involved in kidney injury in the HSCT setting. However, with regard to the first limitation (small sample size), the calculated statistical power of our study is 84.33%, which indicates that our study was not underpowered. With regard to the last limitation (factors for HSCT kidney injury), although our patients received a number of nephrotoxic agents (amphotericin B, vancomycin, furosemide), these agents were comparable between patients with and without nephrotoxicity.
Our data demonstrated that about one-third of patients with allogenic HSCT (30.95%) developed cyclosporine nephrotoxicity, with a median onset time of 15 days. None of the studied demographic, clinical, and paraclinical characteristics of the study population was significantly associated with cyclosporine nephrotoxicity. During the course of cyclosporine treatment in hospital, hypokalemia was detected in more than three-fourths of the study population (76.2%) and hypomagnesemia in more than one-half (52.4%). However, no case of hyperkalemia was observed in cyclosporine recipients. No patient received emergent hemodialysis or continuous renal replacement therapy, and no patient required discontinuance of cyclosporine treatment to mitigate nephrotoxicity. Length of hospital stay for patients with cyclosporine nephrotoxicity was both clinically and statistically higher than the length of hospital stay for patients without cyclosporine nephrotoxicity. Finally, urine KIM-1 was not superior nor more accurate than serum or urine creatinine levels for prediction or detection of cyclosporine nephrotoxicity. Nevertheless, further clinical studies with more precise measurement time points of renal function biomarkers and exclusion of all possible confounders of nephrotoxicity are needed to determine the definite role of urine KIM-1 in both prediction and assessment of cyclosporine nephrotoxicity in the setting of HSCT.
DOI : 10.6002/ect.2020.0123
From the 1Department of Clinical Pharmacy, Faculty of Pharmacy, Shiraz
University of Medical Sciences, Shiraz, Iran; and the 2Hematology Research
Center and Department of Internal Medicine, Shiraz University of Medical
Sciences, Shiraz, Iran
Acknowledgements: The authors thank all the staff of the Bone Marrow Transplant Unit at Namazi Hospital for their kind cooperation. This article was prepared from the data of the PharmD thesis of Mahtab Jafari. This study was financially supported by the Vice Chancellery of Research and Technology Affairs of Shiraz University of Medical Sciences with the grant number 97-01-05-16717. Other than described above, the authors have not received any additional funding or grants in support of the presented research or for the preparation of this work and have no potential declarations of interest. Author contributions are as follows: Karimzadeh and Ramzi designed the study; Karimzadeh and Davani-Davari analyzed the data and drafted the manuscript; Jafari selected patients and gathered data; and Ramzi interpreted the clinical results and reviewed the manuscript; all authors read and approved the manuscript.
Corresponding author: Mani Ramzi, Hematology Research Center and Department of Internal Medicine, Shiraz University of Medical Sciences, Shiraz, Iran
Phone: +98 713 647 32 39
Table 1. Comparison of Different Demographic, Clinical, and Paraclinical Characteristics of the Study Population (n = 42) Between Patients With and Without Cyclosporine Nephrotoxicity
Table 2. Area Under the Curve for Serum Creatinine, Urine Creatinine, and Urine Kidney Injury Molecule at Studied Time Points of Cyclosporine Treatment in Prediction and Detection of Cyclosporine Nephrotoxicity