Cost-effectiveness of hepatitis C–positive donor kidney transplantation for hepatitis C–negative recipients with concomitant direct-acting antiviral therapy

1 INTRODUCTION

The continuing disparity between demand and supply for kidney transplants has exerted pressure on the transplant community to utilize donors at increased risk of transmission of bloodborne infections such as hepatitis B, human immunodeficiency virus, and hepatitis C virus (HCV). HCV donor antibody (Ab) +/nucleic acid test (NAT) + positive donor (D+) kidneys are now increasingly being offered to HCV NAT+ recipients due to published data that HCV NAT+ kidney transplant recipients (positive recipients [R+]) can significantly cut short their waiting time if they are willing to accept HCV D+ kidneys.1, 2 The shorter waiting time can result in an overall improved mortality without any significant worsening of hepatitis C–related-liver disease.3 Unfortunately, a large proportion of D+ kidneys might never be used due to the unavailability of R+ candidates. Reese et al reported that nearly 65% (out of a total 6546) of all D+ kidneys were discarded between the years 2005 and 2014.4

Traditionally, HCV D+ kidneys have not been offered to HCV− (R−) recipients due to the concerns of viral transmission. This was due to the fact that early small studies reported in the 1990 s suggested a very high transmission rate. A study from the New England Organ Bank on 29 organ transplants (from 13 HCV Ab+ donors; 19 kidneys) showed that chronic hepatitis C developed in 55% of all transplants. After transplant, 67% (of 29) of patients tested positive for HCV by antibody testing and 96% by NAT-based testing.5-7 Nevertheless, our recent analysis of the national Organ Procurement and Transplantation Network (OPTN) data suggested that this practice continued sporadically and ≈700 such HCV Ab+ kidneys were transplanted from 1994 to 2014. Interestingly, even in this era HCV D+/R− kidney transplant patients have improved mortality compared with matched controls on the waitlist who never received a transplant.8 The generalizability of these findings was limited by the fact that HCV NAT testing was not available for a large majority of donors. Thus, it is conceivable that many donor kidneys could have been HCV Ab+/NAT−, resulting in a minimal infectious transmission risk.

The development of direct-acting antiviral (DAA) drugs for the treatment of hepatitis C has changed the landscape of HCV therapy with high rates of sustained virologic response (SVR) even for patients with kidney transplants.9-12 With these advances in therapy, there is a potential of preemptive treatment with DAAs allowing for safe transplantation of D+ kidneys into needy R− recipients to minimize any risk of viral transmission. Early data from some pilot studies suggest that this is feasible.13, 14 While data on safety and efficacy are becoming available, the cost-effectiveness of such a strategy is less clear. Thus, it is necessary and important to assess the cost-effectiveness of this approach, given the fact that DAAs are extremely expensive.

To address this gap in knowledge, we performed a cost-effectiveness analysis comparing 2 strategies: (a) renal transplantation (RT) from a HCV+ donor (D+) into HCV− recipients (R−) followed by immediate DAA therapy vs (b) HCV− recipients (R−) continuing dialysis and waiting for RT from a HCV− donor (D−).

2 METHODS

The clinical and research activities reported here are consistent with the Principles of the Declaration of Istanbul as outlined in the “Declaration of Istanbul on Organ Trafficking and Transplant Tourism.”

This study was conducted from the perspective of a third-party payer. The hypothetical population studied consisted of adult end-stage renal disease (ESRD) patients who were hepatitis C negative (R−) and on the waiting list for a deceased donor kidney. A decision tree model was developed to analyze costs and effectiveness over a 5-year time frame between 2 choices: RT using D+ kidney followed by DAA treatment compared to continuing dialysis and waiting for RT with a D− kidney. The model was developed in Microsoft Excel 2016.

Posttransplant cumulative survival probabilities of ESRD adult patients completing RT were extracted from the OPTN/Scientific Registry of Transplant Recipients (SRTR) 2016 Annual Report.15 All data were the latest available records for a 2011 ESRD incident patient cohort followed for 5 years. The same data source was used for obtaining cumulative death probabilities of ESRD adult patients undergoing dialysis while waitlisted for RT. These data came from the latest available records of a 2013 ESRD incident patient cohort followed for 3 years, plus an estimation of another 2 years follow-up to fit our 5-year time frame. We used an exponential model to estimate the rates for the last 2 years for 2 reasons: (1) the literature indicates that the annual death rates increase over time for waitlisted dialysis patients16; (2) the exponential model provides an average annual mortality rate of ≈5% over the 5-year frame. This is consistent with expert opinion and literature.15 We then took the cumulative death probabilities and calculated corresponding annual death probabilities (see Supporting Information).

For D+/R− patients, DAA therapy was initiated immediately after the HCV-infected kidney transplant. If initial DAA therapy failed, a second round of treatment was implemented. D−/R− patients remained on dialysis until a HCV-negative (D−) kidney became available. We assumed that the probability of getting a D− kidney was 0 in year (Y) 1, 25% in Y2, 50% in Y3, 75% in Y4%, and 100% by the end of Y5. Base case probabilities used in the model are listed in Table 1.

Effectiveness was measured in expected years of life (YOL) using data from the ESRD incident patients from the OPTN/SRTR 2016 Annual Report. Cost data were collected from public data sources (Table 2). National Average Drug Acquisition Costs were used to estimate DAA costs.17 Costs for other services were based on Medicare reimbursement rates, extracted from either the United States Renal Data System 2016 Annual Report or the Centers for Medicare and Medicaid Services payment database.18-20 Costs included direct expenditures for RT and immediate care post-RT, immunosuppressive therapy, dialysis, and associated medication costs while awaiting RT, and DAA therapy for D+/R− patients. Expenditures for RT and immediate care post-RT were calculated as the sum of Medicare reimbursements for the kidney transplant Diagnosis-Related Group (DRG 652) and corresponding provider service costs. All cost data were adjusted to 2017 US dollars, using the Consumer Price Index (CPI).21, 22 Specifically, the medicinal drugs CPI was used to adjust immunosuppressive medication and dialysis-related medication costs. The inpatient hospital service CPI was used to adjust the RT DRG cost, while the physician’s services CPI was used to adjust costs of corresponding provider services related to RT. The medical care services by other medical professionals CPI were used to adjust dialysis cost. Indirect and intangible costs were not modeled.

3 RESULTS

The decision tree is shown in Figure 1. For those accepting a HCV NAT+ organ, the cost of care (transplant, DAA, immune suppression, and provider care for up to 5 years) and YOL ranged from $123 092 to $218 196 and 3.1 to 5 years, respectively. Costs were lowest in those who did not survive the 5 years and highest in those who survived and required a second course of DAA. Using an expected 94% SVR rate with DAA and published 5-year survival rates (87%), the cost of the D+/R− strategy was ≈$138 000. In comparison, for those who remained on dialysis while waiting for a HCV-negative organ, the costs and YOL ranged from $143 594 and 1.5 years, respectively, in those who died within 2 years while on dialysis to $478 645 and 5 YOL, respectively, for those who were still waiting for an organ on dialysis after 5 years. Because of the cost of dialysis, the shorter the wait time for a HCV-negative organ the lower the cost: $230 927 to $236 667 if transplanted in year 2 compared to $417 689 if transplanted in year 4. The baseline cost-effectiveness analysis found that the D+/R− strategy resulted in an expected 4.8 YOL with an expected total cost of ≈$138 000 compared to 4.7 YOL and ≈$329 000 for the D−/R− strategy.

As shown in Table 3, the D+/R− strategy remained dominant in all 1-way sensitivity analyses including adjustment for RT cumulative death probability (D+/R− was equivalent to or 20% higher than D−/R−), difference in acute rejection probability between strategies (D+/R− was 5%-20% higher than D−/R−), DAA therapy cure rate (75%-94% for the first round; 71%-89% for the second round), death probability after the second round of DAA therapy failure (failure group was equivalent to or 20% higher than the success group), waiting time on dialysis in the D+/R− strategy (0-24 months), cost of kidney transplant ($19 579-$142 913), cost for treating acute rejection ($11 192-$45 686), immunosuppressive medication costs ($4174-$6262), dialysis costs ($76 217-$114 325), DAA treatment costs ($18 692-$93 582), probability of dialysis patients getting a RT in the D−/R− strategy by year (25% at Y1, 50% at Y2, 75% at Y3%, and 100% at Y4 vs base case of 0% at Y1, 25% at Y2, 50% at Y3, 75% at Y4%, and 100% at Y5), including the cost of Y5 RT (vs excluding it in base case), kidney transplant annual death probability (±20% of base value), and dialysis annual death probability while waitlisted for RT in the D−/R− strategy (±20% of base value). All sensitivity analyses resulted in the same or slightly higher YOL with the D+/R− strategy compared to the D−/R− strategy.

The difference in cost between the 2 strategies, however, was substantial in all sensitivity analyses, with R+/D− strategy being cost-saving. The cost-saving for each sensitivity analysis was compared to the base case cost-savings. The variables with the largest impact on cost-savings between the 2 strategies in the sensitivity analyses were expected waiting time for a D+ kidney, reduced wait time for a D− kidney, and total cost of dialysis. In the base case, the expected cost of the D+/R− strategy was $191 282 less than that for the D−/R− strategy. Compared to the base case, the D+/R− group would have an additional saving of around $55 000 compared to the D−/R− group if total dialysis costs were 20% higher than the base case. As the expected waiting time for a D+ kidney increased, the cost-savings for the D+/R− strategy decreased: it was about $65 000 lower if waiting time was 1 year and almost $150 000 lower if waiting time was 2 years. On the other hand, a 1-year reduction in wait time for D−/R− patients to get a D− kidney decreased the cost advantage of the D+/R− strategy by about $80 000. For all sensitivity analyses, however, the cost of the D+/R− strategy was less than that of the D−/R− strategy.

The difference between total expected costs of the D+/R− strategy and that of the D−/R− strategy decreased as the cost of RT or DAA therapy increased. Therefore, additional analysis was conducted to determine the breakeven cost of RT and DAA therapy. The results indicated that, all other factors being equal, the total cost of RT would have to be $1 085 258 or the cost of DAA therapy $235 188 to make the expected costs of the D+/R− and D−/R− strategies equal. The breakeven costs are far higher than our baseline estimates of the costs of these treatments.

4 DISCUSSION

The advent of DAA therapy has created a new paradigm for the therapy of hepatitis C. In addition to providing a cure for a large majority of patients infected with HCV, these drugs have opened a pathway for the use of HCV viremic donor organs into HCV-negative recipients followed by DAA therapy to eliminate the risk of donor-derived HCV disease. Early pilot studies suggest that this is clinically feasible and excellent sustained virologic responses can be achieved both for kidney as well as other solid-organ transplants. In addition, it was reported that between 29% and 82% of the potential HCV− kidney transplant recipients on the waiting list might be willing to accept HCV+ kidneys under certain scenarios.27 In an analysis using contemporaneous cost data from Medicare (in US dollars), we now report that in addition to providing extra years of life for patients, this technique results in lower costs.

When compared to staying on dialysis and waiting for an HCV-negative donor kidney, the baseline cost-effectiveness analysis found that the D+/R− strategy resulted in an average of about one tenth of an additional year of life with a cost-saving of approximately $190 000 per patient. The D+/R− strategy remained dominant even after sensitivity analyses (Table 3) for a variety of variables. It should be noted that the use of yearly waitlist mortality in our study (done using only a 5-year time frame) does not capture the long-term benefits of transplantation extending beyond 5 years. In addition, it also does not capture the mortality of patients who were never waitlisted (or removed from the waitlist after waiting a few years) because they were considered too sick for transplantation. Thus, we believe that in the long-term the HCV D+/R− strategy might result in additional YOL than suggested by our results.

As expected, the difference between total calculated costs of the D+/R− strategy and that of the D−/R− strategy decreased as the cost of RT or DAA therapy increased. Since an important focus of the D+/R− strategy is the cost of DAA therapy, we took a very conservative approach in our sensitivity analysis. Even an increase in the cost of DAA therapy by 50% did not change the dominance of the D+/R− strategy. Our results indicate that, all other factors being equal, the cost of RT would have to be about a million USD or the cost of DAA therapy about $235 000 USD to make the expected costs of the D+/R− and D−/R− strategies equal. In fact in the actual market, it is currently apparent that the new DAA therapy combinations are less costly than the older ones and even the older ones have higher rebates than available previously.

The findings in our study corroborate that of the previous expert opinions as well as the data published from the United Kingdom recently.28 Trotter et al reported that while the use of DAA (sofusbuvir/ledipasvir) therapy increased the cumulative cost of transplantation at 5 years by £38 979 per patient, transplantation remained cost-effective compared to hemodialysis at 5 years by approximately £20 000 per patient.29 The primary mechanism for the realization of these cost-savings is by reduction in dialysis time. For example, Sawinski et al assessed HCV-positive recipients who received either a HCV D+ or a D− kidney.30 They reported that on average, wait times went down nearly by half (485 days vs 969 days) in case HCV kidney recipients received D+/R+ transplants as opposed to waiting for a D−/R+ transplant. This resulted in a cost-saving of $38 000 to $386 000 per patient, depending upon local wait times.

There are limitations to this analysis, as we used several assumptions for the derivation of our cost-effectiveness models. We assumed that survival probabilities after RT were the same for D+/R− and D−/R− patients. While there is scant clinical evidence to support or refute this assumption, we did do an additional sensitivity analysis assuming a 20% higher probability of cumulative death in D+/R− patients compared to D−/R− patients. This did not change our results significantly (Table 3). In addition, the 2 recently reported HCV D+/R− pilot trials did not report an increase in acute rejection, increased infections, early graft loss, or patient mortality in recipients of HCV D+/R− kidneys, thus supporting our assumption.13, 14 In fact, since HCV+ donors are much more likely to be younger, it is also possible that long-term graft survival might be better for this population than estimated.1, 4 It should also be noted that our data were limited to patients insured by Medicare, and thus findings may not be generalizable to patients with private health insurance. However, there is no reason to believe that these data would not be applicable to those with non-Medicare coverage.

Our analyses only included costs directly related to the transplant or to dialysis. That is, we did not include other health care costs—such as the cost of managing hypertension, diabetes, bone disease, anemia, access complications, etc—incurred by patients on dialysis or after transplant. If such costs were substantially higher for transplant patients than for dialysis patients, then our cost-savings might be overestimates. However, our sensitivity analyses indicate that these differences would have to be very large—around $190 000 per patient in most cases—to eliminate the cost-savings of the D+/R− strategy.

For the purpose of this analysis, we conservatively assumed that all patients would get HCV infection. Infection transmission has ranged anywhere from 0% to 100% based upon the previous data accumulated over the last 3 decades.5, 7, 31 Thus, it is possible that some patients can avoid HCV transmission and might not require DAA therapy. This could potentially result in additional cost-savings. We also assumed that DAA treatment would result in cures for 94% of patients. This assumption is supported by previous studies using DAA to treat HCV after RT.8, 9, 16 Further sensitivity analyses using a lower cure probability and a higher death probability for patients who failed the second round of DAA therapy did not alter the results significantly (Table 3). Finally, we assumed that 100% of HCV D+/R− patients will receive a kidney transplant between 3 and 24 months from listing. Although these data are corroborated by previous evidence as well as our own ongoing study on HCV D+/R− kidney transplants (# NCT03249194), it might not be true for some parts of the country.1 On the other hand, we did use a conservative assumption that all HCV D−/R− patients will receive a kidney transplant by the end of the fifth year on the waiting list. Since wait times can be much longer in many parts of the country, this could have resulted in an underestimation of the cost-effectiveness of the HCV D+/R− approach. Finally, this is a population-based study in which we looked at averages across the dialysis population rather than at specific patient subtypes. It was not our intention to provide estimates for specific patient types, nor do we have the data to support this.

In conclusion, our analysis performed using a third-party payer’s perspective suggests that utilization of HCV donor–positive kidney transplants for HCV-negative recipients is cost-effective. This strategy provides for equal or slightly reduced mortality and substantially reduced costs by limiting dialysis time.

DISCLOSURE

The authors of this manuscript have no conflicts of interest to declare as described by the American Journal of Transplantation.