|Home | About | Journals | Submit | Contact Us | Français|
This study was designed to (a) estimate the contribution of tacrolimus nephrotoxicity to episodes of renal allograft dysfunction investigated by needle biopsy, (b) describe the temporal evolution of nephrotoxicity and its response to therapy, and (c) ascertain how often renal dysfunction is associated with concurrent extra-renal toxicity.
Patients were selected based on a rising serum creatinine, normal ultrasound, and biopsy findings leading to a reduction in the dose of tacrolimus and a fall in serum creatinine. Twenty two (17%) cases of nephrotoxicity were identified amongst 128 consecutive kidney transplant biopsies with sufficient clinical data for analysis.
There were 13 males and 9 females, 17–75 yr in age. Tacrolimus was administered initially as a 0.075–0.1 mg/kg/d IV continuous infusion followed by an oral dose of 0.15 mg/kg twice daily. The onset of nephrotoxicity in this study occurred 1–156 wk post-operatively. The mean baseline creatinine was 212.2±168.0 μmol/l (range 88.4–875.2) and rose 40.6% ± 14.2% (range 11–66) during episodes of nephrotoxicity (p<0.001). The highest recorded plasma and whole-blood tacrolimus levels during the toxic episodes were respectively 2.7±0.8 ng/ml (range 1.1–3.5) and 31.6±10.6 ng/ml (range 14.5–50.5). The drug levels were considered to be beyond the therapeutic range in 18/22 (82%) patients. The highest tacrolimus level preceeded the rise in serum creatinine in 20 cases by an interval of 1.6 ± 1.8 d. A mean reduction in tacrolimus dosage of 41% ± 21% (range 11–89) led to a 86% ± 18% (range 45–100) fall in the serum creatinine within 1–14 d (p<0.001). Interactions between tacrolimus and clarithromycin, diltiazem, or itraconazole modified the pharmakokinetic parameters in three cases. Serum potassium >5.0 mequiv./l was recorded in 9/22 (41 %) cases. Three or more elevations in blood glucose >7.7 mmol/l (140 mg/dl) were recorded in 4/11 (36%) non-diabetic patients. Hand tremors were seen in two (9%) cases and elevated diastolic blood pressure> 90 mmHg in seven (32%) patients.
In conclusion, tacrolimus nephrotoxicity accounted for 17% of graft dysfunction episodes investigated by biopsy. Concurrent hyperglycemia, hyperkalemia, or tremors were noted in several patients. Nephrotoxicity responded well to reduction in the drug dosage.
Tacrolimus (Fujisawa, USA Inc.) is an effective primary immunosuppressive agent following kidney transplantation (1–7). It has both short-term and long-term advantages over conventional drugs (5). Thus, tacrolimus is associated with less hypertension and hypercholesterolemia than is cyclosporine (1). Nephrotoxicity is well recognized, but with the exception of the Japanese Multicenter trial (6), which describes its occurrence in up to 44% of patients, its incidence has not been specifically determined in most studies on renal transplant recipients. The estimated incidence of tacrolimus nephrotoxicity in liver transplant recipients varies from 18 to 42% (8–11). These latter figures may not be directly applicable to the kidney transplant population, since liver disease impairs tacrolimus metabolism, thereby increasing the potential for toxicity. Furthermore, the diagnosis of tacrolimus nephrotoxicity in liver transplant patients is confounded by renal dysfunction due to sepsis, hepato-renal syndrome, hepatic glomerulopathy and the use of other nephrotoxic drugs. Heart and lung transplant recipients are not subject to some of these variables, but published studies on these patients again do not attempt to define precisely the magnitude of tacrolimus nephrotoxicity in this clinical setting (12, 13). A detailed clinical description of patients undergoing tacrolimus-induced renal dysfunction is also lacking in much of the available literature. Many studies provide comparative data on groups of patients receiving either tacrolimus or cyclosporine without analyzing each episode of toxicity separately (8, 9, 11, 14–16). The purpose of the present study is therefore to estimate the incidence of tacrolimus nephrotoxicity in episodes of renal allograft biopsy investigated by needle biopsy, to describe its temporal evolution and response to therapy in individual patients, and to ascertain how often it is associated with concurrent extra-renal toxicity.
Patients were selected over a four year period from July 1991 to July 1995 from a database maintained by the Thomas E. Starzl Transplantation Institute, based on the following criteria: 1. Use of tacrolimus as the primary immunosuppressant after kidney transplantation. 2. Availability of a biopsy performed for rising serum creatinine. Serum creatinine was performed daily in the immediate post-transplant period, biweekly or weekly in the first 8–12 wk after transplantation, and gradually reduced to monthly estimations thereafter. 3. Absence of acute rejection at biopsy leading to reduction in the dose of tacrolimus. 4. Fall of serum creatinine within a 2-wk period following biopsy and reduction in immunosuppression. 5. Clinical exclusion of other causes of graft dysfunction such as vascular and technical problems.
Twenty-two (17%) cases of tacrolimus nephrotoxicity were identified from 128 consecutive biopsies with adequate clinical data available for analysis. The medical records of these patients were analyzed in detail with particular attention to serial changes in serum creatinine, serum potassium, blood glucose, and alterations in the drug dosage. For outpatients, the last available serum creatinine was recorded as the baseline. In hospitalized patients the lowest value within the 14 d preceding the allograft biopsy was regarded as the basal serum creatinine. The highest serum creatinine corresponding to the episode of nephrotoxicity was then recorded, to permit calculation of the percentage rise in creatinine. The post-treatment trough creatinine was defined as the lowest measurement recorded within a 14-d period following reduction of immunosuppression. To quantitate the therapeutic response, the fall in creatinine was expressed as a percentage of the pre-treatment rise above baseline. The statistical significance of the initial rise and subsequent fall in serum creatinine was evaluated by the Mann–Whitney test. Trough levels of tacrolimus were monitored either in plasma (n=10) or in whole-blood (n=12) using published methodology (17, 18). The normal therapeutic range of tacrolimus at our institution is considered to be 0.5–1.5 ng/ml in plasma and 5–20 ng/ml in whole blood.
The clinical profile of the patients studied is presented in Table 1. There were 13 males, 9 females, with an age range of 17–75 yr. The indications for kidney transplantation were diabetes (n=11), polycystic disease (n=3), glomerulonephritis, (n=4), IgA nephropathy (n=2), hypertension (n=1) and nephrolithiasis (n=1). After a single preoperative dose of 0.15 mg/kg, tacrolimus was administered postoperatively as a 0.075–0.1 mg/kg/d intravenous continuous infusion followed by an oral maintenance dose of 0.15 mg/kg twice daily after commencement of oral feeding. The onset of tacrolimus nephrotoxicity in this study occurred 1–156 wk postoperatively. The baseline creatinine was 212.2 ±168.0 μmol/l (range 88.4–875.2) and showed a mean rise of 40.6% ± 14.2% (range 11–66) during episodes of nephrotoxicity (p<0.001). The highest mean plasma and whole-blood tacrolimus levels during the toxic episodes were 2.7 ± 0.8 ng/ml (range 1.1–3.5) and 31.6 ± 10.6 ng/ml (range 14.5–50.5), respectively. The drug levels were considered to be beyond the therapeutic range in 18/22 (82%) patients. The highest tacrolimus level preceded the highest serum creatinine in 16 cases by an interval of 1.6 ± 1.8 d. In the remaining two cases the serum creatinine peaked 1 and 2 d respectively prior to the highest tacrolimus level. The dose of tacrolimus was reduced stepwise until a satisfactory response in serum creatinine was obtained. The mean reduction was 41% ± 21% (range 11–89), and this was statistically significant (p<0.001). No dose reduction was necessary in case 3, since the elevated tacrolimus level fell spontaneously. In two cases a single dose of solumedrol was empirically administered for possible rejection prior to results of the biopsy becoming available. Reduction in the dose of tacrolimus led to a 86% ± 18% (range 45–100) fall in the serum creatinine. The therapeutic response generally commenced within 1–8 d, except in case 8, where a drug interaction with itraconzole delayed the response to 14 d. One or more episodes of hyperkalemia (serum potassium greater than 5.0 mequiv./L) were recorded in 9/22 (41%) cases within 10 d of the peak in serum creatinine. The actual potassium levels fluctuated somewhat, and persistent high values were not always observed. Three or more values of elevated blood glucose exceeding 7.7 mmol/l (140 mg/dl) were recorded in 4/11 (36%) non-diabetic patients. Concurrent use of intravenous steroids adequately explained the hyperglycemia in one patient; in the remainder it was considered to be a transient manifestation of FK 506 toxicity to the pancreas. Hand tremors attributed to tacrolimus neurotoxicity were seen in two (9%) cases (patients 5 and 21). The blood pressure readings in these patients were reviewed to test the notion that tacrolimus exerts its deleterious actions on the kidney by causing vasoconstriction in the renal vascular bed. Elevation in diastolic blood pressure greater than 90 mmHg diastolic was noted in seven (32%) patients. However, day-to-day readings were quite variable, and no progressive rise coincided with the clinical episodes of nephrotoxicity. Infections were seen in 6/22 (27%) patients: #3 (staphylococcal pneumonia), #5 (upper respiratory infection), #8 (Aspergillus in brochoalveolar lavage), #9 (Pseudomonas infection of abdominal wound), #15 (Clostridium difficile colitis) and #21 (acute bronchitis).
Conventional core biopsies of the allograft were performed in 21 cases, and a fine needle aspiration biopsy in one case. Histopathological review of the core biopsies showed tubular vacuolization in 20/21 (95 %), myocyte vacuolization in 15/21 (71%), arteriolar hyalinization in 7/21 (33%), striped fibrosis in 2/21 (10%) and glomerular capillary fibrin thrombi in 1/21 (15%) specimens, respectively. A minimal interstitial infiltrate involving less than 25% of the surface area was present in 11/21 (52%) biopsies. Rare foci of tubulitis (1–2 lymphocytes per tubular cross-section were demonstrable in 4/21 (19%) cases. The needle aspiration specimen showed tubular vacuolization with no evidence of immune activation.
Tacrolimus allograft nephrotoxicity is widely recognized, but to our knowledge, only one study has attempted to estimate its incidence in kidney transplant recipients (6). That investigation based the diagnosis of nephrotoxicity on the clinical course of the patient and graft biopsy, but the actual criteria used were not explicitly stated. It was found that renal toxicity was dependent on the dosing regimen and prior clinical experience with the drug. Thus, this study recorded a 44% incidence of nephrotoxicity when tacrolimus was administered at a dose of 0.3 mg/kg/d. Subsequently, with additional experience in the use of the tacrolimus a reduction in incidence to 20.5% was observed (6).
In the present study, we have observed a 17% incidence of nephrotoxicity in renal transplant biosies. The actual incidence of nephrotoxicity is probably higher, since tacrolimus dosage reduction is at times performed without a biopsy. The initial maintainence dose of tacrolimus in our patients was 0.15 mg/kg twice daily. The ultimate maintainence dose, which was fine tuned by serial determinations of plasma or whole-blood levels of tacrolimus, varied in different patients. It is important to stress that the incidence of nephrotoxicity reported here refers to the use of tacrolimus as a primary immunosuppressant after renal transplantation. In patients receiving intravenous tacrolimus as “rescue” therapy for refractory renal or hepatic allograft rejection, initial nephrotoxicity is seen in nearly all patients (19–21). The definition of nephrotoxicity used in this study required a fall in creatinine in response to reduction in the maintainence dose of tacrolimus. Hence, only reversible episodes of tacrolimus nephrotoxicity were identified. Other investigators have described patients with apparently persistent nephrotoxicity (6, 10). However, it is not always clear whether the clinically non-responsive cases sustained additional adverse effects such as dehydration, acute tubular necrosis, drug reactions or transplant glomerulopathy to explain the persistent graft dysfunction. Progressive drug induced renal dysfunction may occur in patients receiving continuous tacrolimus immunosuppression, since the nephrotoxic and antirejection actions of this drug are mechanistically related (22). However, recognition and distinction of such insidiously developing toxicity from chronic rejection is difficult on clinical grounds. Chronic tacrolimus nephrotoxicity could be better studied in liver transplant recipients without the confounding influence of chronic rejection. However, these patients are predisposed to renal dysfunction caused by sepsis, hypotension, hepato-renal syndrome, and glomerulopathy associated with liver disease, all of which would need to be distinguished from tacrolimus nephrotoxicity. Studies of chronic tacrolimus toxicity in heart and lung transplant recipients should likewise control for elevations in serum creatinine due to nephrotoxic antibiotics and congestive heart failure existing at the time of transplantation or developing subsequently.
Nephrotoxicity episodes were associated with elevated plasma or whole-blood in tacrolimus levels in 18/22 patients. In other investigations, good correlation between blood tacrolimus levels and graft dysfunction have been observed by some authors (23, 24), but not by others (11, 25). The percent dose reduction necessary to restore allograft function varied, reflecting the known variability of tacrolimus pharmacokinetics in individual patients (26). Drug interactions appeared to modify further the disposition of tacrolimus in three cases. Thus, case 8, which required the greatest dose reduction (89%), was receiving itraconazole, a drug known to compete with tacrolimus for metabolism by the hepatic microsmal cytochrome P450 system (27). Inhibition of tacrolimus detoxification by itraconazole would explain why this patient required such a drastic reduction in dose, and showed a lag period of 14 d before reversal of nephrotoxicity could be observed clinically. Diltiazem, another drug biotransformed by the liver microsomes (28), was used in patient # 10, who required a 66% dose reduction in tacrolimus. An extremely high level of whole-blood tacrolimus (50.5 ng/ml) was recorded in case 5, who received clarithromycin, a macrolide antibiotic structurally related to both erythromycin and tacrolimus. Interactions between clarithromycin and tacrolimus have not been previously observed to the best of our knowledge, but erythromycin is known to inhibit competitively the metabolism of tacrolimus by the cytochrome P-450 system. We have previously reported a kidney transplant recipient in whom the plasma Tacrolimus level increased from 1.3 to 8.5 ng/ml within 4 d of starting erythromycin (29).
The reported incidence of hyperkalemia in Tacrolimus treated kidney transplant patients varies from 27 to 67% (6, 10, 19, 30–32). Its specific incidence in patients with nephrotoxicity is not mentioned in these studies. In the current study, one or more values of elevated serum potassium occurred in 9/22 (41 %) cases with renal dysfunction. Hyperkalemia as an isolated finding without other evidence of impaired kidney function has been described in 9% of patients (31). The mechanisms responsible for hyperkalemia are not well understood, but an effect of tacrolimus on mineralocorticoid secretion and altered mineralocorticoid activity at the renal tubules have been proposed. Clinically, hyperkalemia induced by Tacrolimus usually responds readily to dietary restriction, potassium binding resins and fludrocortisone.
Altered glucose metabolism is another recognized toxicity of Tacrolimus. Changes in the peripheral sensitivity to insulin and/or response of islet cells to blood glucose lead to hyperglycemia in 25–35% of transplanted subjects (24, 30). Post-transplant diabetes mellitus defined as persistently high blood sugar with an abnormal glucose tolerance test is seen in 4–22% of patients (2, 30, 33, 34). Evaluation of this side effect was difficult in 11 of our cases with known insulin-dependent diabetes mellitus, since by definition these subjects were hyperglycemic even before they received Tacrolimus. Considering only patients transplanted for diseases other than diabetes, 4/11 patients with Tacrolimus nephrotoxicity had hyperglycemia defined as blood glucose exceeding 7.7 mmol/l (140 mg/dl) on at least three occasions (35). Intravenous methylprednisolone had been administered empirically to one patient while the result of an allograft needle biopsy was pending. The remaining 3 patients were on stable maintenance doses of steroids, and the hyperglycemia was likely a manifestation of Tacrolimus toxicity.
In summary, this study has shown that reversible Tacrolimus nephrotoxicity accounts for 17% of renal allograft dysfunction episodes investigated by needle biopsy. The diagnosis of Tacrolimus nephrotoxicity was based on rigorous criteria, namely a rise in serum creatinine requiring biopsy, absence of histopathologic changes of acute rejection, and clinical response to reduction in the dose of Tacrolimus. It was noted that Tacrolimus toxicity could occur both early and late post-transplant. Plasma or whole-blood Tacrolimus levels were typically high at the time of clinical diagnosis, and the peak Tacrolimus level preceded the peak in serum creatinine. A reduction in dosage led to an improved serum creatinine within 1–14 d. Hyperkalemia and hyperglycemia were noted in several cases during the episodes of nephrotoxicity.
This work was supported by the Pathology Education & Research Foundation, Pittsburgh, PA, and has been presented in abstract form at the International Transplantation Society Meeting at Barcelona, Spain, August, 1996.