The Practical Utility of an Economic Analysis of Calcineurin Withdrawal following Renal Transplantation
Bruce Kaplan* and
Jesse D. Schold
* Department of Surgery, University of Illinois at Chicago, Chicago, Illinois; and Departments of Medicine and Health Services Research, Management and Policy, University of Florida, Gainesville, Florida
Correspondence: Dr. Bruce Kaplan, Department of Surgery, University of Illinois at Chicago, 820 South Wood Street, Chicago, IL 60612. Phone: 312-413-5605; Fax: 312-413-3483; E-mail: kaplanb{at}uic.edu
Economic models can be powerful tools for understanding cost-utility,effectiveness, and informing decision-making processes in healthcare. Limited health care resources necessitate objective evaluationsand the identification of potential efficiencies of medicalinterventions. At base, economic analyses are often about evaluatingchoices, and the identification of cost-savings associated withcertain therapies can potentially be reallocated to other efficacioustreatment strategies. In these respects, economic analyses inthe field of transplantation are important contributions toenhancing the ultimate goal of maximizing patient care. Botha strength and a limitation of economic analyses are that theyoften provide a net result, a bottom line that can be easilyinterpreted in a given health care context; however, as withany model, it is also important to recognize that the conclusionsdrawn from economic evaluations are also often critically dependenton the inputs of the analysis and the assumptions that are made.
The study by Earnshaw et al.1 in this issue of JASN examinesthe cost-utility of calcineurin inhibitor (CNI) withdrawal inde novo renal transplant recipients. The study attempts to estimatethe relative cost-savings associated with CNI withdrawal usinga lifetime Markov model, aggregating data from both clinicaltrial and population-based studies. The model incorporates healthstates and medication costs and concludes that "CNI withdrawalnot only shows potential for long-term clinical benefits butalso is expected to be cost-saving over a patient's life." Clearly,these results may be provocative to the transplantation community;therefore, it is essential to examine critically the analysisto gauge the plausibility of the end results and the likelihoodof a robust finding. Two primary assumptions incorporated inthis study have less than certain information and may significantlyalter the external validity of these findings. The first ofthese relates to the estimated efficacy of the therapeutic strategies.The second relates to the long-term outcomes extrapolated fromthese results over the lifetime of the patient.
One of the core inputs of this analysis is the impact of transplantrecipients’ renal function on costs and health statusassociated with the respective therapeutic strategies. Importantly,the estimated renal function between CNI withdrawal and cyclosporine-basedtherapy is derived from the Rapamune Maintenance Regimen (RMR)study.2 The clear limitation of the selection of this trialas a basis to estimate the effect of the respective treatmentarms is that the control arm is particularly nephrotoxic andlikely represents only the most extreme benefit of CNI withdrawal,and the control arm does not represent a standard of care, isinfrequently used in current practice, and as such provideslittle basis by which to generalize evidence. Given this, thepresumption that similar effects will be attributable to otherforms of maintenance CNI therapies is problematic. In fact,in the era after the RMR trial, additional studies have nowaccumulated data about CNI withdrawal. The Cyclosporine AvoidanceEliminates Serious Adverse Renal-toxicity (CAESAR) study foundno differences in renal function at 1 yr between CNI-based therapiesand a significantly increased acute rejection rate in the withdrawalarm.3 In addition, Larsen et al.4 conducted a trial of CNI withdrawalusing a tacrolimus-based treatment arm and found no differencesin renal function, survival, or acute rejection at 1 yr; therefore,it is important to recognize that the estimated effects describedin this study may exaggerate the immunologic and metabolic risksassociated with the therapeutic regimens, and using data fromone of the more recent (and applicable) studies may dramaticallyalter these findings. In fact, examining some of the resultsof the study critically (see Table 1) indicates that the tacrolimus–mycophenolatemofetil arm was estimated to have the lowest life-years andquality-adjusted life-years among recipients despite other studies’indicating superior efficacy in this arm. This type of inputdata seems to lack some degree of face validity on the basisof empirical evidence, and, as such, it is difficult to discernwhether the selected studies for the analysis alter the estimates,whether the inputs can reliably be aggregated from differentstudies, and whether the cumulative uncertainty associated withthe assumptions for the analysis even lend themselves to conductingany type of rigorous study examining this important topic atthis stage.
Better renal function leads to better graft survival among transplantrecipients. This general association has been shown in population-basedanalyses and is of course not a surprising result; however,the appropriate application of these data in this circumstance(and others) is certainly worthy of discussion. At base, theissue is whether this general association can be appropriatelyapplied to novel circumstances, specifically whether elevatedrenal function necessarily translates into improved survival.For instance, given the same logic stream, treating patientswith angiotensin blockers or inhibitors would independentlyimprove graft survival as a function of their impact on renalfunction. Similarly, simple fluid changes can substantiallyaffect estimated renal function, but this artificial event clearlydoes not translate to differences in long-term survival. Theanalysis by Hariharan et al.,5 which was the basis for the estimationin this analysis, demonstrated an association that would bereasonable given that all further injuries to the allograftare evenly and randomly distributed across the population. Inthe case of CNI withdrawal, we certainly cannot assume thatfurther allograft damage (e.g., late antibody-mediated injury)would be randomly distributed when such a large interventionis imposed. Using this type of associative data to make assumptionswhen conditions are nonrandomly changed is problematic and potentiallymisleading. To date, there is no compelling evidence for improvedgraft survival associated with CNI withdrawal, and, as such,extrapolation of selected renal function changes to long-termefficacy and cost-effectiveness may be premature. Furthermore,even accepting the assumptions of the analysis and the long-termextrapolation to results, the estimated cost savings of theCNI withdrawal arm was less than $12,000 over a patient's lifetime,and the estimated benefit in life years was 0.06 yr. If we thenconsider the mathematical variability in the estimates in conjunctionwith the accuracy and precision of the assumptions, then thequestion that arises is what can be reasonably concluded fromthese net results?
At the end of the day, left unanswered for any practical purposes,unfortunately, is the essential question asked by this study.We hope that these data are interpreted only with considerationof the strong caveats and that readers along with media approachthese data with similar restraint. For the time being, one cannotassume a cost saving or a health benefit of CNI withdrawal,and patience for additional data accumulation will likely bethe most important prospective strategy. This important issuerequires—and our patients deserve—studies specificallydesigned to answer this question.
Earnshaw SR, Graham CN, Irish WD, Sato R, Schnitzler MA: Lifetime cost-effectiveness of calcineurin inhibitor withdrawal after de novo renal transplantation.
J Am Soc Nephrol 19
: 1807
–1816, 2008[Abstract/Free Full Text]
Oberbauer R, Kreis H, Johnson RW, Mota A, Claesson K, Ruiz JC, Wilczek H, Jamieson N, Henriques AC, Paczek L, Chapman J, Burke JT, Rapamune Maintenance Regimen Study Group: Long-term improvement in renal function with sirolimus after early cyclosporine withdrawal in renal transplant recipients: 2-Year results of the Rapamune Maintenance Regimen Study.
Transplantation 76
: 364
–370, 2003[CrossRef][Medline]
Ekberg H, Grinyo J, Nashan B, Vanrenterghem Y, Vincenti F, Voulgari A, Truman M, Nasmyth-Miller C, Rashford M: Cyclosporine sparing with mycophenolate mofetil, daclizumab and corticosteroids in renal allograft recipients: The CAESAR Study.
Am J Transplant 7
: 560
–570, 2007[CrossRef][Medline]
Larson TS, Dean PG, Stegall MD, Griffin MD, Textor SC, Schwab TR, Gloor JM, Cosio FG, Lund WJ, Kremers WK, Nyberg SL, Ishitani MB, Prieto M, Velosa JA: Complete avoidance of calcineurin inhibitors in renal transplantation: A randomized trial comparing sirolimus and tacrolimus.
Am J Transplant 6
: 514
–522, 2006[CrossRef][Medline]
Hariharan S, McBride MA, Cherikh WS, Tolleris CB, Bresnahan BA, Johnson CP: Post-transplant renal function in the first year predicts long-term kidney transplant survival.
Kidney Int 62
: 311
–318, 2002[CrossRef][Medline]
Related Article
Lifetime Cost-Effectiveness of Calcineurin Inhibitor Withdrawal After De Novo Renal Transplantation
Stephanie R. Earnshaw, Christopher N. Graham, William D. Irish, Reiko Sato, and Mark A. Schnitzler
J. Am. Soc. Nephrol. 2008 19: 1807-1816.
[Abstract][Full Text][PDF]
Top
Introduction
DISCLOSURES
