ARTICLES

A Logistic-Regression Model Provides Novel Guidelines to Maximize the Anti-Acute Rejection Properties of Cyclosporine with a Minimum of Toxicity

Annalisa Perna

Eliana Gotti

Ernesto de Bernardis

Norberto Perico

Giuseppe Remuzzi

A. Perna, E. Gotti, N. Perico, G. Remuzzi, Mario Negri Institute for Pharmacological Research,
Bergamo, Italy

E. Gotti, N. Perico, G. Remuzzi, Division of Nephrology, Ospedali Riuniti di Bergamo,
Bergamo, Italy

E. de Bernardis, Institute of Pharmacology, University of Catania Medical School,
Catania, Italy

Although cyclosporine has become the mainstay of immunosuppression in organ transplantation, there is still no consensus on the criteria to optimize its antirejection activity with minimum toxicity. A clear and objective definition of target cyclosporine trough levels at different times from renal transplantation is still lacking, primarily because of the lack of a model correlating cyclosporine levels with probability of rejection or toxicity. In this study, logistic-regression model was developed that was applied to data collected retrospectively from two postoperative periods, i.e., Days 0 to 9 and 10 to 30, in 135 consecutive cadaveric renal transplant recipients, for a total of 1851 determinations. Only minimum and maximum trough levels were considered for each period. Concentration-response curves were estimated for Days 0 to 9 ( P = 0.0001 for efficacy and P = 0.028 for toxicity) and for Days 10 to 30 ( P = 0.015 for efficacy and P = 0.037 for toxicity). Therapeutic intervals of 330 to 430 ng/mL (parent compound in whole blood) for Days 0 to 9 and 260 to 390 ng/mL for Days 10 to 30 predicted an incidence of acute rejection of 22% and 12%, respectively, with a reasonably low toxicity that primarily consisted of elevation of serum aminotransferases.

Key Words:

Cyclosporine dosage

cyclosporine blood concentration

acute graft rejection

nephrotoxicity

hepatotoxicity

Addition of cyclosporine (CsA) to conventional antirejection protocols has improved the short-term results of organ transplantation ^[1] . However, acute rejection remains a major problem in transplantation and, concerning renal transplants, it is a recognized risk factor for subsequent chronic deterioration of renal function ^{[2] [3] [4]} . On the other hand, CsA is a nephrotoxic agent ^[5] , and doses high enough to avoid acute rejection may per se impair renal function. Thus studies have addressed the issue of the most appropriate strategy for desired immunosuppression with minimum toxicity. Complete pharmacokinetic profiles for area under the time-concentration curve determination are costly and time- and staff-consuming ^[6] , and abbreviated profiles have been proved inaccurate ^[7] . Thus, trough level monitoring was invariably preferred in clinical practice ^[8] . Target values for CsA trough levels at different times from transplant have not been defined in sufficient detail, primarily because of the lack of a model correlating CsA blood levels with probability of rejection or toxicity. Burke and coworkers ^[4] , in a recent retrospective analysis of data on 1663 renal transplant patients, have reported that average CsA trough levels considered to be "high" ( e.g., >250 ng/mL by HPLC on whole blood) at 3 months post-transplant were associated with lower creatinine values in the first and third years, respectively. Lindholm ^[9] recommended a minimum trough level of 150 ng/mL (parent compound in whole blood) during the first postoperative month. However, he studied a very heterogeneous patient population, with both kidney and kidney-pancreas transplantations, and with approximately two-thirds of patients in triple therapy (CsA, azathioprine, steroids) and one-third in double therapy (CsA plus steroids). Other investigators suggested a therapeutic interval of 150 to 250 ng/mL (parent compound in whole blood) for the first few months after transplantation in patients on triple therapy ^{[10] [11] [12]} , but they did not provide the scientific basis for such indication, or try to define recommended levels for the immediate post-transplantation period, when the frequency of rejection is maximal. Here we have developed a mathematical model on the basis of CsA trough levels measured during two different post-transplant periods--Days 0 to 9 and 10 to 30 after surgery--in 135 consecutive renal transplantations, all treated with CsA combined with steroids and azathioprine.

METHODS

Study Population and Immunosuppression Protocol

To focus on the immediate post-transplant phase, the first month after surgery has been divided into two periods of increasing length: from Day 0 to 9 (first period) and from Day 10 to 30 (second period). The minimum number of documented daily CsA trough levels for admission to the study was three for the first period and six for the second period. This resulted in the selection of 115 and 90 cases, respectively, from a series of 135 consecutive renal transplantations. Only the minimum and maximum CsA trough levels for each period were considered. Trough levels evaluated during Orthoclone OKT3 (muromonab-CD3; Ortho Pharmaceutical Corp., Raritan, NJ) cycles for steroid-resistant rejections were discarded, because in these cases, CsA doses are generally lowered to reduce the potential for opportunistic infections, whereas the anti-CD3 antibody affords the mainstay of immunosuppression.

Diagnosis of Acute Graft Rejection and Cyclosporine Toxicity

Statistical Analysis

RESULTS

First Period (Days 0 to 9 After Surgery)

Second Period (Days 10 to 30)

Concentration-Response Curves

During the first 9 days, half-maximal efficacy (EC50) is reached at 102 ng/mL (95% CI, 59 to 177), maximum efficacy (EC95) at 330 ng/mL, with a predicted 22% incidence of acute rejection. For the rest of the first month, half-maximal efficacy is attained at 91 ng/mL (95% CI, 57 to 146), maximum efficacy at 260 ng/mL, with a predicted 12% incidence of acute rejection. The threshold values under which rejection is most probable (more than 90%) should be located around 50 ng/mL (EC10, 43 ng/mL for the first and 55 ng/mL for the second period). Half-maximal toxicity (TC50) is extrapolated at 2300 ng/mL for the first and 1500 ng/mL for the second period; 15% toxicity (TC15) is reached at 430 ng/mL until Day 9 and at 390 ng/mL thereafter.

DISCUSSION

Left and right limits of logistic regression curves were constrained to 0 and 1, respectively, assuming that complete CsA withdrawal from triple therapy during the early post-transplant phase would result in rejection and absence of CsA-induced toxicity, whereas CsA overload would very likely lead to toxicity. However, imposing or removing constraints did not substantially change the curves' shape and location.

The therapeutic window for a given drug is generally calculated as the interval between the concentration effective in 99% of patients and that that causes toxicity in 1% of patients (EC99 to TC1); however, because these values overlap in the present situation, we suggest use of the EC95-TC5 interval, i.e., 330 to 430 ng/mL for the first 9 days, and 260 to 390 ng/mL for the rest of the month. No further improvement of the antirejection properties of CsA may be achieved by increasing blood levels beyond these values. In any case, CsA trough levels should never fall under the concentrations active in 90% of patients (EC90), i.e., 240 ng/mL for Days 0 to 9 and 150 ng/mL until the end of the first month. These levels are higher than those suggested elsewhere for the first few months after transplantation in patients on triple therapy (150 to 250 ng/mL) ^{[10] [11] [12]} . Interestingly, a recent report

TABLE 1 -- Distribution of minimum CsA trough levels in cases with or without rejection

Period	Rejection
	Yes				No
	Median	25th Percentile	75th Percentile	N	Median	25th Percentile	75th Percentile	N
Days 0 to 9	134	96	183	54	174	134	219	61
Days 10 to 30	189	140	237	16	223	169	284	74

Figure 1. Estimated concentration-response curves for efficacy versus minimum CsA trough levels and toxicity versus maximum CsA trough levels during Days 0 to 9 ( N = 115) and Days 10 to 30 ( N = 90). E9, efficacy curve for Days 0 to 9; E30, efficacy curve for Days 10 to 30; T9, toxicity curve for Days 0 to 9; T30, toxicity curve for Days 10 to 30.
has documented that at CsA trough levels of 150 to 190 ng/mL, the calcineurin phosphatase activity of peripheral blood leukocytes is still 45 to 55% of control subjects, indicating that higher levels may be required to achieve a more substantial inhibition of IL-2 production ^[24] . The actual incidence of nephrotoxicity in our study was very low (2 to 3%), whereas elevations of aminotransferases were more common, reaching 20% during Days 10 to 30. Indeed, CsA-dependent hepatotoxicity generally appears to be mild and asymptomatic ^[25] , reversible on dosage reduction ^[26] , and not troublesome in patients with normal liver function at the time of transplantation ^[27] . For these reasons, we feel confident that the chosen 15% maximal incidence of toxicity may be well tolerated in a clinical setting. All patients studied had no recorded antecedent liver disease before kidney transplant. One should be aware however of the possible confounding factor of previous liver diseases, such as hepatitis C, in evaluating the actual incidence of CsA-induced hepatotoxicity by this approach.

On the other hand, the concern that high early doses of CsA may be a risk factor for long-term renal dysfunction ^{[5] [28]} has not been substantiated by a study specifically designed to address this point ^[4] . In the latter study, patients with the highest whole-blood CsA trough levels ( > = 250 ng/mL by HPLC analysis) at 3 months post-transplant had better renal function at 1 and 3 yr as compared with those with lower CsA levels. Consistent with the above findings, in heart-transplanted patients given CsA as a chronic antirejection therapy after an initial decline at 2 to 3 yr from surgery ^[29] , renal function did not worsen further at 5 to 6 yr ^[30] , indicating that CsA renal toxicity is self-limited. This observation is supported by previous studies in 200 cardiac transplant recipients by Myers and Newton ^[31] . They found that after 12 months, both low- and high-dose CsA treatment was associated with depression of GFR below the values recorded in another 100 recipients who had never been exposed to CsA. In contrast, there are reasons to believe that acute rejection reduces nephron units, which in the long-term may increase the risk of hyperfiltration damage ^{[32] [33]} . Thus the tendency to give low doses of CsA, under the assumption of avoiding early and late renal complications, is not substantiated by current findings. Instead, higher CsA levels during the first 30 post-transplant days, although remarkably reducing the incidence of acute rejection, may better preserve renal function over the long term ^{[4] [34]} . Our findings are in line with a recent observation by Nankivell et al. ^[35] in 92 consecutive renal allograft recipients receiving triple therapy and studied for episodes of renal dysfunction with respect to whole-blood CsA concentration within the first 100 days post-transplantation. They found a good specificity in the prediction of CsA nephrotoxicity by high (>400 ng/mL) levels and of acute rejection by low (<150 ng/mL) CsA levels, although the sensitivity was relatively low. However, in this study, 59% of rejection episodes occurred within the "therapeutic" (150 to 400 ng/mL) range for CsA, a value markedly higher with respect to what we have predicted by our present model. The apparent inconsistency between the two proposed regimens (that admittedly derives from different methodological approaches to the problem) is, however, easy to reconcile if one considers that the "therapeutic" range proposed by the above study ( i.e., 150 to 400 ng/mL) is so wide that it includes a considerable number of patients who have trough levels well below what was shown by our mathematical 

CsA is metabolized in the liver by enzymes of the cytochrome P450 3A subfamily to more than 30 metabolites ^[36] . They include first-generation metabolites, those generated by metabolism of the native compound CsA in one position (AM1, AM9, AM4N), and second-generation metabolites, those generated by further metabolism of other metabolites. Few and conflicting data are available on the potential immunosuppressive activity of these metabolites, and whether they contribute at least in part to the well-known toxic effect of the native compound is still a matter of debate ^{[37] [38] [39]} . The present logistic-regression model is based on measurement of whole-blood CsA concentration by RIA using a specific monoclonal antibody that detects the circulating level only of the parent drug but not of its metabolites. Therefore, the results we obtained reflect the toxic/therapeutic profile of CsA " per se." On the basis of the results of the study presented here, we suggest that the maximal antirejection effect of CsA could be attained at concentrations of 330 to 430 ng/mL for the first 9 days after surgery, and 260 to 390 ng/mL for Days 10 to 30. The model predicts no further improvement of antirejection activity for trough levels of CsA higher than as indicated above. Within these therapeutic intervals, the model predicts a rather low frequency of toxic reactions. The incidence of acute rejection episodes predicted by the present model is considerably lower than that reported in most series, which, by utilizing more conservative protocols, have an incidence of acute rejection episodes ranging between 47% ^[4] and 60% ^[2] in the first 30 postoperative days. It goes without saying that the theoretical findings of this model should be validated by means of a prospective trial to verify whether the therapeutic intervals proposed here, if tested against more conventional protocols, did indeed offer the predicted degree of protection. The results of the study presented here, if confirmed prospectively, may have a major impact in the management of organ transplants other than those of the kidney.

ACKNOWLEDGMENTS

REFERENCES

2. Lindholm A, Ohlman S, Albrechtsen D, Tufveson G, Persson H, Persson NH: The impact of acute rejection episodes on long-term graft function and outcome in 1347 primary renal transplants treated by 3 cyclosporine regimens. Transplantation 1993;56:307-315.

3. Basadonna GP, Matas AJ, Gillingham KJ, et al.: Early versus late acute renal allograft rejection: Impact on chronic rejection. Transplantation 1993;55:993-995.

4. Burke JF, Pirsch JD, Ramos EL, et al.: Long-term efficacy and safety of cyclosporine in renal-transplant recipients. N Engl J Med 1994;331:358-363.

5. Myers BD: Cyclosporine nephrotoxicity. Kidney Int 1986;30:964-974.

6. Land W: Optimal use of cyclosporine in clinical organ transplantation. Transplant Proc 1987;19:130-135.

7. Gaspari F, Ruggenenti P, Torre L, Bertocchi C, Remuzzi G, Perico N: Failure to predict cyclosporine area under the curve using a limited sampling strategy. Kidney Int 1993;44:436-439.

8. Grevel J, Welsh MS, Kahan BD: Cyclosporine monitoring in renal transplantation: Area under the curve monitoring is superior to trough-level monitoring. Ther Drug Monit 1989;11:246-289.

9. Lindholm A: A prospective study of cyclosporine monitoring in renal transplantation. Transplant Proc 1990;22:1260-1263.

10. Kivisto KT: A review of assay methods for cyclosporin. Clinical implications. Clin Pharmacokinet 1992;23:173-190.

11. Kahan BD, Shaw LM, Holt D, Grevel J, Johnston A: Consensus document: Hawk's Cay meeting on therapeutic drug monitoring of cyclosporine. Clin Chem 1990;36:1510-1516.

12. Shaw LM: Advances in cyclosporine pharmacology, measurement, and therapeutic monitoring. Clin Chem 1989;35:1299-1308.

13. Ponticelli C, Tarantino G, Montagnino G, et al.: A randomized trial comparing triple-drug and double-drug therapy in renal transplantation. Transplantation 1988;45:913-918.

14. Solez K, Marcussen N, Flynn GF, Beschorner WE, Racusen LC, Burdick JF: Pathology of acute tubular necrosis and acute rejection: Observations on early systematic renal transplant biopsies. In: Burdick JF, Racusen LC, Solez K, Williams GM, Eds. Kidney Transplant Rejection: Diagnosis and Treatment. 2nd Ed. New York: Marcel Dekker; 1992:373-383.

15. Dallal GE: LOGISTIC: A logistic regression program for the IBM PC. Amer Statistician 1988;42:272.

16. Sherrod PH. Nonlinear statistical regression program, version 2.4. 1993.

17. Liu J: FK506 and cyclosporin: Molecular probes for studying intracellular signal transduction. Trends Pharmacol Sci 1993;14:182-188.

18. Awni WM: Pharmacodynamic monitoring of cyclosporin. Clin Pharmacokinet 1992;23:428-448.

19. Henry ML, Bowers VD, Fanning WJ, Sommer BG, Ferguson RM: Cyclosporine levels are not helpful. In: Kahan BD, Ed. Cyclosporine: Nature of the Agent and Its Immunological Actions. Philadelphia: Grune and Stratton; 1988:419-421.

20. Kasiske BL, Heim-Duthoy K, Venkateswara RK, Awni WM: The relationship between cyclosporine pharmacokinetic parameters and subsequent acute rejection in renal transplant recipients. Transplantation 1988;46:716-722.

21. Perico N, Ruggenenti P, Gaspari F, et al.: Daily renal hypoperfusion induced by cyclosporine in patients with renal transplantation. Transplantation 1992;54:56-60.

22. Perico N, Dadan J, Remuzzi G: Endothelin mediates the renal vasoconstriction induced by cyclosporine in the rat. J Am Soc Nephrol 1990;1:76-83.

23. Takeguchi N, Ichimura K, Koike M, Matsui W, Kashiwagura T, Kawahara K: Inhibition of the multidrug efflux pump in isolated hepatocyte couplets by immunosuppressants FK506 and cyclosporine. Transplantation 1993;55:646-650.

24. Batiuk TD, Pazderka F, Halloran PF: Cyclosporine-treated renal transplant patients have only partial inhibition of calcineurin phosphatase activity. Transplant 

25. Kassianides C, Nussenblatt R, Palestine AG, Mellow SD, Hoofnagle JH: Liver injury from cyclosporine A. Dig Dis Sci 1990;35:693-697.

26. Ramos EL, Tilney LN, Ravenscraft MD: Clinical aspects of renal transplantation. In: Brenner BM, Rector FC, Eds. The Kidney. Philadelphia: W.B. Saunders Company; 1994:2361-2407.

27. Morris PJ. Cyclosporine. In: Morris PJ, Ed. Kidney Transplantation: Principles and Practice. Philadelphia: W.B. Saunders Company; 1988:285-317.

28. Feutren G, Mihatsch MJ, et al.: Risk factors for cyclosporine-induced nephropathy in patients with autoimmune diseases. N Engl J Med 1992;326:1654-1660.

29. Bertani T, Ferrazzi P, Schieppati A, et al.: Nature and extent of glomerular injury induced by cyclosporine in heart transplant patients. Kidney Int 1991;40:243-250.

30. Ruggenenti P, Perico N, Amuchastegui CS, Ferrazzi P, Mamprin F, Remuzzi G: Following an initial decline, glomerular filtration rate stabilizes in heart transplant patients on chronic cyclosporine. Am J Kidney Dis 1994;24:549-553.

31. Myers BD, Newton L: Cyclosporine-induced chronic nephropathy: An obliterative microvascular renal injury. J Am Soc Nephrol 1991;2[Suppl 1]:S45-S52.

32. Brenner BM, Cohen BM, Mildford EL: In renal transplantation, one size may not fit all. J Am Soc Nephrol 1992;3:162-169.

33. Brenner BM, Meyer TW, Hostetter TH: Dietary protein intake and the progressive nature of renal disease. The role of hemodynamically mediated glomerular injury in the pathogenesis of progressive glomerular sclerosis in aging, renal ablation, and intrinsic renal disease. N Engl J Med 1982;307:652-659.

34. Salomon D, Brunson M, Vansickler J, et al.: A retrospective analysis of late renal graft function: Correlation with mean cyclosporine levels and lack of evidence for chronic cyclosporine toxicity. Transplant Proc 1991;23:1018-1019.

35. Nankivell BJ, Hibbins M, Chapman JR: Diagnostic utility of whole blood cyclosporine measurements in renal transplantation using triple therapy. Transplantation 1994;58:989-996.

36. Christians U, Sewing KF: Cyclosporin metabolism in transplant patients. Pharmacol Ther 1993;57:291-345.

37. Lucey MR, Kolars JC, Merion RM, Campbell DA, Aldrich M, Watkins PB: Cyclosporin toxicity at therapeutic blood levels and cytochrome P-450 IIIA. Lancet 1990;335:11-15.

38. Christians U, Kohlhaw K, Budniak J, et al.: Ciclosporin metabolite pattern in blood and urine of liver graft recipients. I. Association of ciclosporin metabolites with nephrotoxicity. Eur J Clin Pharmacol 1991;41:285-291.

39. Rosano TG, Pell MA, Freed BM, Dybas MT, Lempert N: Cyclosporine and metabolites in blood from renal allograft recipients with nephrotoxicity, rejection or good renal function: Comparative high-performance liquid chromatography and monoclonal radio-immunoassay studies. Transplant Proc 1988;20:330-334.