Increasing the dose intensity of the conditioning regimen prior to allogeneic hematopoietic stem cell transplant: the role of pharmacokinetic monitoring

Allogeneic hematopoietic stem cell transplant (SCT) is a potentially curative treatment for a variety of hematological malignancies. The therapeutic effect of SCT is mediated by both the administration of high-dose chemo-radiotherapy and induction of the graft-versus-malignancy effect (GVM) by immune competent cells in the graft. Historically, more emphasis was given to the dose intensity of the regimen. Studies aiming at increasing myeloablative doses have generally shown a reduced relapse risk after SCT, but were also associated with increased transplant-related mortality (TRM) such that overall survival was not improved [1]. Over the last decade the pendulum turned more toward induction of GVM as the primary goal of SCT. Reduced-intensity conditioning (RIC) regimens were designed to reduce regimen-toxicity and to allow SCT in elderly or medically infirm patients. These regimens are usually based on fludarabine with an alkylator, like busulfan, in about half of the myeloablative dose. RIC achieves constant engraftment in patients previously considered not eligible for SCT; however, relapse rates are often higher than after standard myeloablative conditioning [2]. Reducing toxicity without compromising SCT efficacy could be of significant benefit. Novel reduced toxicity myeloablative regimens, such as the combination of fludarabine with myeloablative doses of busulfan [3] or treosulfan [4], were designed to achieve this goal.

The major regimen-toxicity of high-dose busulfan regimens is veno-occlusive disease of the liver (VOD). Busulfan is metabolized by conjugation with glutathione. In this reaction mediated by glutathione-S-transferase, glutathione is oxidized to glutathione disulfide. Cyclophosphamide metabolites are a major cause of VOD, especially when the liver is depleted of glutathione after prior busulfan therapy. High-dose cyclophosphamide was used historically in myeloablative conditioning. It has a limited antileukemic activity but contributes significantly to transplant toxicity. Its substitution with fludarabine, an equally immunosuppressive agent but with limited toxicity and metabolism that is unrelated to the glutathione pathway, was able to markedly reduce TRM, with no increase in relapse rate. The next step is to build on this less toxic platform and to increase the busulfan dose above the standard myeloablative dose, trying to reduce the relapse rate. In this issue of Leukemia and Lymphoma, O'Donnell et al. report a phase I study of dose-escalated busulfan with fludarabine and determine the dose-limiting toxicity of busulfan in this setting [5].

Monitoring busulfan pharmacokinetics (PK) is an absolute requirement for studies using escalated doses. Historically, busulfan was only available as an oral preparation. Busulfan tablets are highly irritating to the gastric mucosa and emetogenic, and thus it was difficult to control bioavailability. In addition it is associated with erratic intestinal absorption and a large interpatient variability in PK parameters. There was also a significant dose-to-dose variability in the same patient. PK monitoring and dose adjustments are used as means to improve the predictability of busulfan exposure, but have questionable effect due to this high dose-to-dose variability. The intravenous formulation allows 100% bioavailability. PK parameters are linear and stable in the standard therapeutic doses of 0.8–3.2 mg/kg. A standard 0.8 mg/kg dose gives a mean area under the curve (AUC) of 1100–1200 μmol-min [6,7]. The total daily dose can be administered once daily rather than the historical four times daily, with no additional toxicity. There is no drug accumulation between doses, and the AUC in the once-daily dose is approximately four times greater, at a mean of 4800–5000 μmol-min [8]. There is minimal dose-to-dose variability; however, there is still a large interpatient variability in AUC that may be related to age, disease status, concomitant medications, and genetic variability in glutathione-S-transferase, the main enzyme mediating busulfan metabolism. We have recently shown a correlation of blood glutathione levels pre-treatment with busulfan clearance, which may explain part of this variability [7]. Busulfan has a narrow therapeutic window. In the combination with cyclophosphamide it was in the AUC range of 950–1500 μmol-min. Lower AUCs were correlated with engraftment failure and disease relapse, while higher AUCs were associated with toxicity (especially VOD) and graft-versus-host disease (GVHD) [6]. Intravenous busulfan is associated with less VOD than oral busulfan since a high proportion of patients achieve AUC within the therapeutic window, and also due to elimination of the first-pass effect through the liver that may increase the toxicity of the oral preparation. It is conceivable that a therapeutic window will also exist in the combination with fludarabine, but not necessarily with the same AUC levels.

O'Donnell et al. aimed at increasing daily the AUC up to 6800 μmol-min using two approaches [5]. The initial one was calculating PK parameters after a small test-dose given within 8 days of conditioning. This approach failed to predict treatment-dose PK, as clearance was often faster in this smaller dose, and linearity could not be documented. The second approach was to monitor PK of the first dose and adjust the third and fourth doses to the desired average AUC. This approach was successful, as the dose-to-dose variability is usually small and PK is highly linear and predictable in the range of therapeutic doses. The busulfan dose could be safely escalated to 5800 μmol-min; however, dose-limiting severe (often fatal) VOD was seen in five of eight patients targeted to an AUC of 6800 μmol-min.

Interestingly, VOD occurred later than usual on median day + 52, rather than within the first 21 days, but was otherwise not clinically different. The actual maximal AUC was >6000 μmol-min in seven of eight patients with VOD. VOD was best correlated with maximal AUC measured, rather than the total cycle busulfan exposure or average AUC. It is currently unknown what are the PK parameters that best predict toxicity and efficacy. Although it is conceivable that peak levels will determine toxicity, the once-daily dose that is associated with three times higher C_max than the four times-daily regimen is not associated with more VOD [8]. Similarly, RIC regimens that are associated with similar maximal AUC, but half cumulative AUC, are associated with less VOD. In addition we have shown that blood glutathione levels that are correlated with VOD are reduced with therapy, but only in myeloablative conditioning after cumulative exposure [7]. Further studies are required to define the best dose schedule approach. As O'Donnell et al. suggest, other approaches such as continuous administration of busulfan to avoid peaks, and combination of a test dose to predict the first treatment dose with first treatment-dose PK to adjust further dosing, may be promising.

Although increasing the busulfan dose is feasible and possibly desirable to reduce relapse rates, the possible increase in dose is modest, and it remains to be shown whether it is indeed associated with better outcome. The PK targeted dosing is also labor-intensive and requires a dedicated, experienced team, with short turn-around results of busulfan levels and dose calculations, not an easy task for most transplant centers. Other approaches to increase the anti-malignancy intensity of the conditioning regimen include, among others, substitution of fludarabine with a more potent antileukemic nucleoside analog, such as clofarabine. Another approach is adding disease-targeted therapy, such as monoclonal antibodies, demethylating agents, or tyrosine kinase inhibitors, pre- and post-transplant, or various cellular therapy approaches. Such novel regimens are likely to improve SCT outcome in the future and possibly reduce disease recurrence rate, the major cause of treatment failure.