Approaching the Therapeutic Window for Cyclosporine in Kidney Transplantation: A Prospective Study

Cyclosporine (CyA) is the backbone of current immunosuppression protocols in many kidney transplant centers. Therapeutic monitoring of the drug is essential because CyA has a narrow therapeutic window with serious potential side effects. CyA trough blood levels (C₀) have been used widely to monitor CyA dosing. Significant groups of patients experience acute rejection (AR) or acute CyA nephrotoxicity (CyANT) despite maintaining CyA levels within a “therapeutic range” that tends to be variable and center specific (1,2). The area under the blood CyA concentration versus time curve (AUC) is a better measure of systemic drug exposure than C₀. Although earlier reports using Sandimmune (Sandoz Pharmaceuticals, Basel, Switzerland) CyA suggested AUC monitoring was associated with better outcomes (3,4), therapeutic monitoring of CyA based on pharmacokinetic studies and AUC has not gained popularity in large part because of the inconvenience and costs associated with multiple sampling and analysis.

A new galenic formulation of CyA, Neoral (Novartis Pharmaceuticals, Basel, Switzerland), has more predictable and consistent pharmacokinetic characteristics than Sandimmune. It was hoped that with Neoral, C₀ would better reflect true drug exposure, resulting in better immunosuppression with fewer side effects (5,6,7). Subsequent randomized clinical trials in new kidney transplant patients showed a comparable tolerability and side effect profile for the two CyA formulations and a lower but still significant rate of AR (35 to 45%) (8,9,10) with Neoral. Better pharmacokinetic characteristics of Neoral, however, allow AUC determination with fewer blood samplings (11,12,13,14). A 4-h AUC (AUC_0-4), shown to correlate very closely with a full 12-h AUC (r² = 0.94) (11), is a more practical monitoring tool and is very effective in differentiating between patients with normal and impaired absorption of CyA (11). On the basis of this relationship, we analyzed and reported (15) a cohort of 156 de novo kidney transplant recipients who received Neoral-based triple immunosuppression and had an AUC_0-4 measured on days 2 to 4 after starting Neoral. In this cohort, the Neoral dose was monitored by C₀ levels, not AUC_0-4. This early AUC_0-4 displayed a significant correlation with AR and CyANT during the first 3 mo posttransplantation, whereas C₀ did not. An AUC_0-4 range of 4400 to 5500 μg/h per L was associated with the lowest rate of AR and CyANT (15). Given the demonstrated crucial importance of early exposure to therapeutic levels of CyA as measured by AUC_0-4, we conducted this pilot study in which we prospectively monitored and adjusted Neoral doses in de novo renal transplant patients by serial AUC_0-4 determinations rather than by C₀. The objective of the study was to assess efficacy, safety, and feasibility of achieving and maintaining therapeutic ranges of CyA as measured by AUC_0-4 (4400 to 5500 μg/h per L) in the first 3 mo posttransplantation.

Materials and Methods

The study population comprised 59 consecutive first kidney transplant recipients who received a transplant between August 1998 and May 1999 at the Queen Elizabeth II Health Sciences Center in Halifax. All patients were followed for a minimum of 3 mo. Collaboration with in-patient and outpatient nursing staff and with the drug monitoring laboratory facilitated timely blood sampling and same-day AUC reporting, without disruption of other normal routines.

Logistic feasibility was measured by the ratio of AUC performed to AUC planned during the first 2 wk. During this time, each patient had five AUC scheduled, compared with only once a week to once a month afterward. AR was the primary efficacy end point. Primary safety end points were 3-mo serum creatinine and episodes of CyANT. Neurotoxicity and hepatotoxicity were secondary safety end points.

Results

Table 1 shows the characteristics of the patients in the study. The only unusual finding for our center is the unexpectedly high percentage of living related kidney transplants, which was due to an unusual paucity of cadaver kidneys during the study period.

Clinical Parameters

Efficacy. Four recipients of HLA-identical living donor kidney recipients were deliberately excluded from the acute rejection efficacy analysis, leaving 55 patients for analyses regarding AR. In 33 of 55 recipients (60%), AUC_0-4 >4400 μg/h per L was achieved on postoperative day 3 and maintained afterward (group 1). In 22 of 55 recipients (40%), AUC_0-4 on day 3 and/or day 5 (AUC_0-4) was <4400 μg/h per L (group 2). Only 1 patient (3%) in group 1 versus 10 patients (45%) in group 2 had AR (P = 0.0002). Virtually all AR episodes occurred early, within 26 d of transplantation (Figure 1). The only patient in group 1 who had AR was a highly sensitized (panel reactive antibody, 60%) recipient with severe DGF. Protocol biopsy on day 6, while the graft was still not functioning, showed vascular rejection. In group 2, 10 patients had 12 episodes of AR. Eight of them had AR within 20 d posttransplantation. Two patients had AR on days 26 and 87, respectively, when their Neoral dose was decreased because of assumed neurotoxicity (severe tremors) and nephrotoxicity, despite AUC_0-4 in the low target range. Reducing the Neoral dose led to AUC_0-4 values below target levels, which was followed by AR within 14 d, without an interval fall in serum creatinine. Two of these 10 patients had recurrent rejection on days 40 and 54, respectively, both with AUC_0-4 < 4400 μg/h per L. The Neoral dose had been decreased because of severe tremors and assumed nephrotoxicity, but again, the Neoral dose reduction did not lead to a reduction in serum creatinine.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Freedom from rejection during first 3 mo posttransplantation. Except for one patient, achieving target cyclosporine (CyA) exposure as measured by absorption profile area under the curve (AUC_0-4) within 5 d prevented acute rejection. Virtually all rejection episodes occurred within 1 mo.

Stepwise logistic regression analysis of variables that affect AR showed that only AUC_0-4 (P = 0.0001) and DGF (P = 0.04) were significant predictors of AR. Age, gender, C₀, weight-adjusted Neoral dose, HLA mismatch, source of transplantation (cadaver versus living), and type of third immunosuppressive drug were not significant.

Figures 2, 3, and 4 show mean serum creatinine, AUC, and weight-adjusted Neoral dose, respectively, in both group 1 and group 2 patients during the first 3 mo after transplantation. Group 1 patients had lower mean serum creatinine levels throughout (Figure 2), despite higher mean AUC_0-4 for most of the 3 mo (Figure 3). The influence of early AUC_0-4 on AR and 3-mo serum creatinine was seen in both cadaveric and live donor kidney recipients (Table 2), but because of the small number of patients in the cadaveric group, the difference did not reach statistical significance despite a strong trend. The dose of Neoral required to maintain target AUC_0-4 diminished over time to 4.5 ± 1.3 mg/kg per d in group 1 patients and to 5.1 ± 1.4 mg/kg per d in group 2 patients at 3 mo posttransplantation (P = 0.29; Figure 4). The dose of Neoral in 156 historic patients monitored by C₀ in our center was 4.6 ± 1.6 mg/kg per d at 3 mo posttransplantation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mean serum creatinine was lower at all time points during the first 3 mo in patients who reached target CyA exposure within 5 d of transplantation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Mean AUC_0-4 rose quickly with appropriate dose adjustment during the first posttransplantation week, but some patients required 10 d to reach target exposure.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

The Neoral dose was increased in patients with early AUC_0-4 below target to achieve optimal exposure. Mean doses were not significantly different between the two patient groups by 30 d but remained slightly higher at all time points in patients who were unable to reach target exposure by day 5.

View this table:

• View inline
• View popup

Table 2.

Acute rejection and 3-mo serum creatinine in cadaveric and living donor kidney transplant recipients in relation to early CyA exposure^a

Safety. Nephrotoxicity. None of the patients experienced CyANT with AUC_0-4 < 5500 μg/h per L. Four patients had episodes of CyANT while AUC_0-4 was in the range of 6200 to 8500 μg/h per L. In all four patients, CyANT was reversed by appropriate reduction in the Neoral dose.

Multiple regression analysis showed that AUC_0-4 and DGF were the only two independent variables measured that affected serum creatinine at 3 mo. Higher early AUC_0-4 was associated with lower serum creatinine at 3 mo (P = 0.001). Presence of DGF was associated with higher serum creatinine at 3 mo (P = 0.02). Source of transplantation, C₀, weight-adjusted Neoral dose, recipient's diabetes mellitus, HLA mismatch, and type of third immunosuppressive drug were not significant.

Hepatotoxicity. Twenty-two patients (37%) had transient elevations of aminotransferase, alanine aminotransferase, and/or bilirubin. All patients were asymptomatic. None of the patients had the Neoral dose reduced specifically because of elevated liver enzymes. By adjusting the Neoral dose according to AUC_0-4, however, 20 of the 22 patients had normal liver enzymes at the end of the first month and another by the end of the second month. One patient, although completely asymptomatic, had persistently high bilirubin levels.

Neurotoxicity. None of the patients had seizures or reported severe headaches. Five patients (8%) developed moderate to severe tremors after transplantation. In one patient, tremors resolved spontaneously despite increasing Neoral doses. Three patients had tremors associated with AUC_0-4 > 5500 μg/h per L, and all improved on achieving target AUC_0-4 range with decreasing Neoral doses. Only one patient (1.7%) had persistent tremors, even with AUC_0-4 in the low target range.

Discussion

C₀ is used widely for monitoring Neoral dosing, although it correlates poorly with AR and CyANT (1,15). Despite improved pharmacokinetic characteristics of Neoral, with current C₀-based dosing and monitoring regimens, AR occurs in almost one third of kidney transplant recipients (8,9,10). With effective treatment of AR, early graft loss from AR is uncommon. However, AR is still an important predictor of chronic rejection and long-term graft survival (18,19,20).

Using the original Sandimmune CyA preparation, CyA exposure measured by AUC had been reported to affect AR rate and/or graft survival in kidney transplant recipients (3,4). Lindholm and Kahan (3) dosed CyA-Sandimmune by serial AUC monitoring in 160 kidney transplant recipients and showed that average CyA blood concentration of <400 ng/ml on the first posttransplant pharmacokinetic study, usually on postoperative day 7, was associated with higher AR rate and poorer graft survival. Only approximately one third of patients achieved the targeted AUC levels, and even in this group the 1-yr AR rate was 34%. No correlation was found between pharmacokinetic parameters and CyANT. In an extension of that study, Senel et al. (4) reported that an average CyA blood concentration >550 ng/ml on day 7 after renal transplantation was associated with higher graft survival for up to 6 yr posttransplantation, although it did not influence the incidence of acute rejection. C₀ did not correlate well with graft or rejection-free survival.

Despite an abundance of information in the literature regarding correlation between limited sampling AUC and AUC_0-12 for Neoral, there are very limited data on the influence of AUC_0-12 or of any limited sampling strategy on the incidence of AR in kidney transplant recipients. Also, we were unable to find any reports that compared the differential influence of AUC_0-4 and AUC_0-12 on clinical events posttransplantation.

On the basis of excellent correlation established between AUC_0-4 and AUC_0-12 in stable kidney transplant recipients (11) and also because it provides an accurate estimate of CyA bioavailability from the Neoral formulation and was demonstrated to be a useful tool for the determination of CyA absorption characteristics in a broad range of renal transplant recipients (11), our group recently reported a retrospective study of 156 first kidney transplant recipients who received Neoral-based triple immunosuppression and who were monitored by C₀ but had AUC_0-4 measured during the first posttransplantation week. We demonstrated that achieving an AUC_0-4 range of 4400 to 5500 μg/h per L as early as days 2 to 4, even with C₀ monitoring, was associated with the lowest rate of AR (7%) and CyANT (6%) at 3 mo posttransplantation. Although patients received different immunosuppressive regimens, the association among AUC_0-4, AR, and CyANT was independent of which third immunosuppressive agent was used. Again, C₀ correlated very poorly with AR (15).

Although the association of early AUC_0-4 with AR and CyANT was observed in patients who were monitored by CyA C₀, the feasibility and safety of dosing Neoral to maintain an AUC_0-4 range of 4400 to 5500 μg/h per L in the first 3 mo after kidney transplantation was unknown. The purpose of the current study was to assess the feasibility, safety, and efficacy of AUC_0-4 monitoring in different Neoral-based triple immunosuppressive protocols. For this reason, no attempt was made to change the ongoing immunosuppressive protocols during this study. AUC_0-4 rather than C₀ was used to monitor Neoral dosing. This was achieved without significant inconvenience to patients, to laboratory staff, or to nursing staff on the wards or in the outpatient clinic. Requirements for abbreviated AUC monitoring were reviewed with laboratory and nursing staff before initiating the study, and their collaboration allowed potential pitfalls to be identified and eliminated beforehand. As a consequence, 94% of scheduled AUC were actually completed, 88% on the scheduled days. Missed or incomplete AUC were usually due to patients' being off the ward for other investigations. Virtually none of the infrequent later AUC was missed. Although for the purpose of the study five blood samples were drawn, only three blood levels—C₁, C₂, and C₃—proved to be necessary to calculate accurately AUC_0-4 for clinical decision making. The C₀ and C₄ samples were used to confirm the validity of 3-point sampling versus 5-point sampling in calculating AUC_0-4 (r² = 0.99) and to archive data for later determination of whether even fewer sampling points later after transplantation would reflect adequately true CyA exposure. Changes in the Neoral dose were rational and made in direct proportion to AUC_0-4 values.

None of the patients, including eight patients with DGF, experienced CyANT with an AUC_0-4 < 5500 μg/h per L. Achieving and maintaining AUC_0-4 levels of 4400 to 5500 μg/h per L were not associated with deterioration of renal function at 3 mo. On the contrary, early exposure to CyA as measured by AUC_0-4 had a very significant inverse relationship with 3-mo serum creatinine. This effect might be related to prevention of subclinical immunologic injury to the graft with proper CyA exposure (21,22). The absence of nephrotoxicity may also be related in part to rapid and appropriately rational reduction of the Neoral dosage to prevent prolonged overexposure to CyA. This is possible with AUC monitoring but not with C₀ monitoring, because the latter does not reflect accurately true drug exposure, and low or “normal” C₀ levels may actually be associated with high exposure to CyA. This may help explain the anachronistic findings of nephrotoxicity with “low” C₀ and excellent renal function with “high” C₀. It is also interesting that at the end of 3 mo, the mean Neoral dose in mg/kg in patients who were monitored by AUC_0-4 was similar to the mean Neoral dose in our historic patients who were monitored by C₀. Although the mean dose was similar, the distribution of dosage was more appropriately allocated in AUC-monitored patients because AUC monitoring allows for more rational adjustment of the Neoral dose to maintain optimal CyA exposure. In addition, with AUC monitoring, clinical hepatotoxicity was not seen in our 59 patients. Elevated liver enzymes were seen in 37% of patients, but it was asymptomatic in all and normalized with Neoral dose adjustments designed to maintain CyA exposure within the target range. Similarly, neurotoxicity was not a clinical problem when CyA exposure was maintained within the target AUC range.

Our experience also suggests that rapidly achieving and maintaining AUC_0-4 within the 4400 to 5500 μg/h per L range may effectively prevent AR. In this group, only one highly sensitized patient with severe DGF had early vascular AR. Prevention of this type of AR arguably might not be possible with any currently available oral immunosuppressive agents, all of which act by inhibiting T-cell activation or function in some way. The poor correlation between C₀ and AUC_0-4 may help to explain the occurrence of AR with “high” or target C₀ levels and its absence in many patients with “low” C₀. Interestingly, four patients developed AR when their Neoral dose was decreased because of assumed toxicity despite an AUC_0-4 at the low end of the target range. This decrease in dose, attributable to early inexperience with AUC_0-4 monitoring, resulted in AUC_0-4 <4400 μg/h per L. The subsequent AR occurred without any previous reduction in serum creatinine, suggesting that CyANT had not been present before dose reduction. All four of these patients were in group 2, with early subthreshold exposure to CyA during the first 5 postoperative days. This experience emphasizes the importance of not only rapidly reaching adequate levels of CyA exposure as measured by AUC_0-4 but also maintaining it, especially in patients who may have experienced early donor-specific T-cell activation as a result of inadequate early exposure to CyA.

This study shows for the first time that with Neoral, therapeutic drug monitoring by AUC_0-4 is feasible; safe in terms of renal, hepatic, and neurologic function; and efficacious in prevention of AR. By AUC_0-4 monitoring, rapidly achieving and maintaining proper CyA drug exposure may essentially eliminate AR and CyANT. If confirmed by subsequent larger experience, these findings may have a significant impact on our understanding and management of CyA immunosuppression in clinical renal transplantation.