Pharmacoeconomics of Empirical Antifungal Use in Febrile Neutropenic Hematological Malignancy and Hematopoietic Stem Cell Transplant Patients

Stuart J Turner; Sharon CA Chen; Monica A Slavin; David CM Kong


Expert Rev Pharmacoeconomics Outcomes Res. 2013;13(2):227-235. 

In This Article

Pharmacoeconomics of Empiric Antifungal Treatment

Given the high cost of treating an IFI, there have been several pharmacoeconomic studies investigating the cost-effective treatment regimens for empirical antifungal therapy.[40–50] These studies are summarized in Table 1. Search terms used for this review included 'caspofungin', 'LAmB', 'voriconazole', 'cost–effectiveness', 'cost', 'pharmacoeconomics', 'quality-adjusted life year (QALY)', 'empiric therapy' and 'invasive fungal infection' and the search was limited to English. The references associated with each article were also reviewed for further potential articles.

Rubio-Terrés and Grau reported that the main cost drivers for the treatment of IFI include inpatient/outpatient costs (e.g., length of stay in hospital), nursing care, home care, medication costs, treatment of adverse effects and toxicity costs.[51] In particular, increased hospital stay was the primary driver (47% of cost), particularly if it involved admission to an intensive care unit; an intensive care unit bed is estimated to cost up to 3.6 times as much as a general ward bed per day.[51] The authors also estimated that if a patient is admitted to hospital for a reason other than treatment of an IFI, and then develops an IFI during their episode of care, the patient can incur additional costs up to €15,000 per episode (based on Candida infection).[51] Indeed, the local 'fungal' epidemiology may also influence the cost of care, given that the fungal epidemiology at a local level will influence, for example, the choice of antifungal agent prescribed, length of treatment and duration of hospitalization and types of monitoring or diagnostic tests.

Almost all pharmacoeconomic evaluations have compared either caspofungin or voriconazole to LAmB, with only one study simultaneously analyzing all three treatment options (Table 1).[49] A three-way comparison by Romá-Sánchez et al.,[49] while providing a useful contribution to the body of evidence supporting voriconazole use, employs an 'in-house' defined outcome measure instead of the preferred composite outcome as used in pivotal efficacy studies,[18] making it difficult to generalize and compare the findings. Furthermore, the study was a relatively small, retrospective, observational study with only six patients receiving voriconazole, and little consideration was given to the inclusion of the cost associated with major adverse effects such as nephrotoxicity with LAmB.[49]

Collins et al. retrospectively reviewed the hospital records for patients receiving voriconazole and LAmB.[44] The investigators employed decision-tree analysis techniques. This study, however, is limited by its inclusion of only 'nephrotoxicity' as the clinical outcome beside successful treatment. Thus, significant bias results against LAmB, given that voriconazole has much less risk of renal toxicity. The authors concluded that voriconazole was associated with a 27% reduction in treatment cost compared with LAmB, with this difference becoming almost negligible when the rates of nephrotoxicity were imported from large multicenter trials.

Another study, which investigated the rates of nephrotoxicity as a sole clinical outcome, by Wingard et al. compared caspofungin and LAmB.[47] This study utilized published data on drug acquisition costs and rates of impaired renal function. There was no consideration given to the variable effectiveness of each agent or the costs associated with inpatient treatment and monitoring. Thus, it was not surprising for caspofungin, which has a lower drug acquisition cost and reduced incidence of nephrotoxicity (hence less attributable cost), to be the dominant alternative. A more robust comparison would have been to include other direct medical costs, such as some form of a daily hospitalization cost.

Similarly, Kaskel et al. compared the difference between caspofungin and LAmB by evaluating the difference in cost due to nephrotoxicity and the resultant extended hospitalization.[45] Omission of other relevant outcomes has weakened the study's findings and, thus, generalizability of the data. This study also did not explore beyond first-line therapy in scenarios where either caspofungin or LAmB were ineffective first choices, an important consideration in costing studies. Incorporating costs associated with alternative antifungal therapies would assist in determining an estimate for the overall treatment cost for each agent, given that neither agent has a remarkably high success rate in pivotal trials.[18] It is noteworthy that in practice, the first-line antifungal agent is often not changed as a consequence of persistent fever – a composite outcome measure used routinely in major clinical trials.[22] Some experts have argued that the short time horizon in the composite outcome (7 days) does not allow for all fevers to resolve.[52] Hence, the inclusion of costs associated with alternative agents after the failure of first-line therapy may be appropriate in the setting of drug toxicity, but not in persistent fever.

Other economic studies have utilized what has become the standard 5-point composite outcome measure.[40,41,43,46,48] These composite outcome measures are absence of breakthrough infection, survival for 7 days beyond the end of therapy, no premature discontinuation of therapy due to drug toxicity or lack of efficacy, resolution of fever and successful treatment of baseline fungal infections.[22] A cost–effectiveness study that excluded baseline fungal infection resolution as part of a composite outcome was performed by Shehab et al..[50] This study evaluated voriconazole and LAmB from a USA hospital's perspective and has the benefit of being based on day-to-day clinical practice and data, but with small sample size. The baseline fungal infections were also not described. It failed to evaluate the two alternative antifungal agents within the same time period, but by comparing the use of LAmB prior to and voriconazole after implementation of a new neutropenia treatment algorithm may have had an impact on the success rate and resource usage in each arm.

An Australian study performed three similar evaluations; voriconazole versus LAmB, caspofungin versus LAmB and caspofungin versus voriconazole.[40–42] They are novel in that an expert panel was used to describe treatment algorithms beyond the first-line empiric therapy; that is, which alternative antifungal was used in various circumstances of treatment failure. The cost of these alternative therapies was included to further consolidate the entire treatment cost. Another unique feature for the study comparing voriconazole and LAmB was that it included consideration in the base-case and sensitivity analyses of the use of orally administered voriconazole, which can reduce overall treatment cost for this agent.[40] This is the only study that reported LAmB was economically superior to voriconazole for empiric therapy of IFI.[40] Al-Badriyeh et al.'s most recent study comparing caspofungin and voriconazole is the first study to directly perform an economic evaluation comparing these two agents in empiric therapy of IFI, with caspofungin having a higher probability (65.5%) of providing cost-savings.[42] Of interest is the apparent relationship between these three agents that can be derived from these studies, which utilized similar methodology. While caspofungin was preferred over LAmB and slightly in favor of voriconazole, LAmB was preferential compared with voriconazole. This may have been influenced by local treatment algorithms and the choice/associated cost of alternative therapies used in the instance of initial treatment failure.

Stam et al. (Italy),[46] Bruynesteyn et al. (UK)[43] and Naik et al. (Sweden)[48] compared caspofungin and LAmB economically utilizing the 5-point composite outcome measure, and QALYs. These studies included the costs for minor adverse effects such as chills and nausea, an aspect not considered in other studies. While this provides more complete resemblance to a real-life scenario, it is probably unnecessary and time-prohibitive owing to the relatively small change in total treatment cost that would be attributed to these adverse effects. These studies all concluded that caspofungin was superior, consistent with the findings of Al-Badriyeh et al..[41]

Importantly, existing data suggest a need for further pharmacoeconomic studies, given the limitations with existing economic studies as discussed previously. The lack of robust clinical data required to support economic analysis is a hurdle that needs to be resolved. Drug cost alone is not sufficient to describe the most favorable agent, nor is it always feasible or realistic to include every cost of treatment into a comparative economic analysis. Differences in healthcare systems and treatment algorithms require work in this area to be carefully applied to an individual clinician's setting and practice.