When a gold standard is made of tin

In its briefing document prepared for the September 1, 2009 meeting of the Oncology Drug Advisory Committee, the US Food and Drug Administration stated that many elderly patients enrolled in a registration trial of the novel agent Clofarabine for the treatment of acute myelogenous leukemia (AML) “were suitable for standard induction chemotherapy or other intensive chemotherapy and would have benefited from such therapy.”^2pp16 Furthermore, the Agency concluded that “approval of new drugs for initial treatment of AML (should be) based on results of randomized controlled trials,”^2pp16 presumably against standard, dose-intensive induction chemotherapy. The question, however, for physicians who treat older patients with AML, is whether standard-induction chemotherapy, the canonical combination of continuous-infusion cytarabine administered over 7 days and 3 doses of anthracycline (3 + 7), constitutes an appropriate “gold standard” against which novel agents should be compared.^3,4  The article from Kantarjian and colleagues in this issue of Blood goes a long way in answering this fundamental question, with both clinical and regulatory implications.

In their study, the authors from the University of Texas M. D. Anderson Cancer Center analyzed the outcome of myelosuppressive, cytarabine-dense, induction chemotherapy administered over an 18-year period to 446 patients over age 70. Precisely because these patients were treated at a single center, and presumably would have been in sufficiently good health to make it to a leukemia unit with expertise in delivering intensive supportive care, the study is able to assess chemotherapy regimens deemed standard for younger patients and in clinical use for over 30 years. Among these patients, only 16 had favorable-risk, core-binding factor–related AML. For the remaining 430 patients, risk factors typical in older patients with AML characterized the group. Unlike younger patients for whom standard infusional cytarabine and anthracycline induction achieves complete remission rates in the range of 70% to 80%, 45% of the elderly patients in the Kantarjian study achieved complete remission, and a staggering 36% of patients died within the first 2 months of the remission-induction period, an interesting and new metric that clarifies the continued risk of dying for this vulnerable population well beyond the traditional 30-day assessment period. The 8-week mortality statistic, whether or not complete remission is achieved, dispensed with terms such as “treatment-related mortality” or death due to induction and allowed the authors to develop a risk model of 8-week mortality in this vulnerable elderly population. Median survival of the whole group was only 4.6 months, and only 28% of patients were alive 1 year after initiating treatment. The results of this study were not invalidated by the epoch of accrual; in fact, the rate of complete remission after 2000 was inferior to the rate before 2000, and no meaningful changes in supportive care over the study period resulted in any differences in survival.

Given the sobering results achieved with standard dose-intensive induction therapy, several investigational agents have sought regulatory approval on the basis of an unmet medical need for the management of AML in the elderly.^5,–7  The strategy has not proven successful. And yet, conducting a phase 3 trial of a novel agent against something like the “gold standard” of 3 + 7 has proven very difficult; physicians and their patients do seem to know what the standard is, and thus resist randomization. The AML14 trial conducted by the Medical Research Council in the United Kingdom was aimed at patients over age 60. The trial had 2 general approaches to treatment: either the standard intensive chemotherapy approach or a nonintensive approach. In case of uncertainty as to which line of therapy to take, patients were to be randomized to one or the other approach. Each approach had a randomization option.^8,9  A total of 1485 patients were accrued to the trial, but of these, only 8 entered the randomized standard versus nonintensive approach,⁹  suggesting that patients or their physicians rejected the concept of being randomized between a myelosuppressive versus low-dose induction regimen. Many in the community of clinical investigators in AML have therefore been criticized for having an “unproven assumption…that elderly patients with AML, especially those with additional poor-risk features, are assumed not to benefit from standard, available therapy.”^10pp9 The Kantarjian article goes far to confirm that assumption as correct and that many vulnerable elderly patients with AML indeed do not benefit from standard-induction chemotherapy. It is fair to conclude that the “gold standard” of induction therapy for many elderly patients with AML is no standard at all.