Before the Conclusion | After Approval: Drugs Reexamined · Case Five
Can a withdrawn drug be approved again? Mylotarg offers a real and complicated answer.
It received U.S. accelerated approval in 2000. In 2010, it left U.S. commercial distribution after a subsequent trial failed to confirm clinical benefit and raised safety concerns; formal withdrawal of approval took effect in 2011. In 2017, the FDA approved it again. In the seven years between commercial withdrawal and reapproval, the molecule the drug targeted did not change. How it was used clinically, which patients might benefit and the evidence supporting those judgments did change. [1,2,7]
To understand this reapproval, we need a more specific question: are we evaluating a molecule, or the use of that molecule in a particular patient population, at a particular dose and in a particular combination?
An attempt to deliver a cytotoxic drug to target cells
Mylotarg's generic name is gemtuzumab ozogamicin. It is an antibody–drug conjugate, usually abbreviated ADC. Its antibody component recognizes CD33, and the attached cytotoxic component damages the target cell. [1]
CD33 is present on the surface of many acute myeloid leukemia cells, giving researchers an opportunity to use an antibody to help deliver a drug. Once inside the cell, the released cytotoxic component can cause DNA damage. Normal cells and tissues may also be affected, however. Targeted delivery does not mean an absence of systemic toxicity. [3]
The initial approval in 2000 was for patients aged 60 or older with CD33-positive acute myeloid leukemia in first relapse who were not considered candidates for other cytotoxic chemotherapy. These patients had very limited treatment options. The initial evidence came primarily from remissions observed in single-arm studies, rather than long-term benefit thoroughly established by large randomized comparisons. [1,2]
Accelerated approval allowed earlier access to treatment while leaving a responsibility for further verification. After seeing leukemia go into remission, researchers still needed to determine how long remission lasted, whether patients benefited and how much serious harm the treatment itself caused.
Why a subsequent trial changed its fate
SWOG's S0106 study compared two induction regimens. One added Mylotarg to daunorubicin and cytarabine; the other used chemotherapy. The study did not show the expected benefit from the Mylotarg-containing regimen, and deaths during induction were higher. [4]
One feature needs careful explanation: the chemotherapy backbones were not identical. The Mylotarg group received a lower dose of daunorubicin. Participants were also younger, previously untreated patients, unlike the older patients with relapsed disease covered by the original approval. [4]
These differences affect how the failure can be interpreted. It cannot be reduced to a claim that every Mylotarg-containing regimen is necessarily without value in every patient population and with every dosing schedule. But for a drug whose clinical benefit still required confirmation, this unfavorable result, together with existing safety concerns such as serious liver injury, did not provide sufficient support for maintaining the original approval. [1,2]
In 2010, the company announced the end of U.S. commercial distribution and asked the FDA to withdraw approval. Formal FDA withdrawal took effect on November 28, 2011. Commercial discontinuation and the legal termination of approval occurred on different dates. Leaving the market reflected the benefit–risk judgment supported by the evidence then available. It also clarified what subsequent research needed to answer. [7]
Not all patients suddenly stopped treatment on the same day. Pfizer's announcement allowed those already taking the drug or holding a prescription to discuss completing treatment with their doctors. For new patients, physicians would need to submit an investigational new drug application, or IND. This was an arrangement for clinical access, not a resolution of the questions about effectiveness. [8]
Why might the dosing schedule change treatment outcomes?
The early approved monotherapy regimen used a relatively high single dose: 9 milligrams per square meter of body surface area, repeated two weeks later. The S0106 combination-chemotherapy trial used a different schedule, including a single dose of 6 milligrams per square meter. These were not the same regimen to begin with. [2,4]
One important later change was lower, fractionated dosing. In a pivotal study of newly diagnosed leukemia, 3 milligrams per square meter was given on each of days 1, 4 and 7 during induction, in combination with chemotherapy. [5]
Why might a smaller individual dose retain efficacy? Drug binding to its target, cellular uptake and changes in blood concentrations do not simply increase in direct proportion to the dose. Giving more at once may increase peak exposure and toxicity without producing a corresponding increase in cell killing.
Pharmacokinetic research therefore suggested that lower, fractionated doses might maintain activity at the target while reducing some harms. This explanation required clinical support; a plausible model alone could not establish success. Later studies tested whether the new approach could provide a better balance of benefits and risks. [2]
How the new evidence accumulated
The French ALFA-0701 study compared lower, fractionated Mylotarg plus chemotherapy with chemotherapy alone. Its main analysis included 271 previously untreated patients aged 50 to 70. The number originally randomized differed from the number in the final analysis, a distinction to remember when reading the results. [5]
The study showed improved event-free survival. This endpoint included events such as induction failure, leukemia relapse and death; it did not simply ask whether a particular bone marrow examination showed remission. By the blinded independent review reported in the final publication, median event-free survival was 13.6 months versus 8.5 months. [5]
The final overall survival comparison in this trial, however, did not reach statistical significance. Its results therefore cannot be rewritten as independently and definitively proving longer life. Its main evidence was improvement in a composite outcome of treatment failure, relapse or death.
There was more than one study. Researchers subsequently combined individual patient data from five randomized trials involving 3,325 patients, including the unfavorable S0106. The pooled analysis found that adding Mylotarg reduced relapse and showed an overall survival benefit. Benefits were more pronounced in some cytogenetic risk groups; this cannot be generalized into identical benefit for every patient. [6]
The key is that the new regimens and multiple trials contributed together to the judgment. Earlier unfavorable results remained part of the evidence. Researchers needed to explain why they differed from other studies and what those differences meant in practice.
Reapproval also came with clear conditions
In 2017, the FDA reapproved Mylotarg for adults with newly diagnosed CD33-positive acute myeloid leukemia, and for adults and children aged two or older with relapsed or refractory disease. This describes the scope of approval at that time; later indication expansions require separate verification. [1]
The supporting evidence differed between uses. ALFA-0701 mainly supported the combination regimen in previously untreated patients. Evidence for relapsed or refractory disease included MyloFrance-1, which used lower, fractionated monotherapy. The results of one trial cannot serve as the same proof of efficacy for every approved use. [1,2]
Reapproval did not make the risks disappear. Mylotarg still carries a boxed warning for severe hepatotoxicity, including hepatic veno-occlusive disease, also called sinusoidal obstruction syndrome. These injuries can be life-threatening and still require careful consideration in treatment planning. [3]
A return to the market therefore did not make it safe for all patients. The new approval rested on specific patient selection, regimens, supporting studies and risk management.
Does reapproval mean that withdrawal was a mistake?
This is the obvious challenge: if it was approved again in 2017, was withdrawal in 2010 wrong? Reapproval alone cannot settle that judgment.
In 2010, the question was whether the existing evidence sufficiently supported the original use. In 2017, it was whether newly submitted evidence supported new conditions of approval. These questions are related, but arose in different evidential circumstances. New dosing schedules and clinical studies were themselves important reasons why the later judgment changed.
We can, of course, still ask whether early development placed too much weight on remission rates, whether dose exploration was sufficient, why the confirmatory trial used a different population and chemotherapy schedule, and whether convincing new evidence could have arrived sooner. These are all questions that can be explored through specific records.
But later success cannot make earlier failure irrelevant. Nor should an earlier failure prevent evaluation of a new regimen. A judgment that remains open to change must do both.
Returning to the question that began this series
Across these five stories, the reasons for a change in conclusions differed. Avacopan asks us to examine how evidence was generated. Relyvrio asks us to face early hopes that did not survive further testing. Vioxx brought serious risks into a more complete assessment. Bevacizumab shows that different uses of the same drug need separate proof.
Mylotarg takes the question a step further: testing evidence need not leave only two choices, continued use or permanent abandonment. Research can also help us change how a drug is used, narrow the eligible population and find treatment conditions under which the risk is worth taking.
My judgment is that what deserves protection is a decision-making process that can be traced, tested and changed in response to new evidence. Previous approval does not remove the need for further testing; previous withdrawal does not replace evaluation of new evidence.
Patients need treatments they have reason to trust. Those reasons should specify the people, regimen, outcomes and costs involved.
The five drug stories leave one final question: what happened to the losses borne by people who had already used these drugs? Reapproval may help later patients, but does not automatically resolve earlier harms. The next article, “After a Drug Is Withdrawn, Who Bears Patients' Losses?”, will compare treatment transitions, financial arrangements, patient compensation and accountability.
Main sources
[1] FDA, September 1, 2017: Mylotarg reapproval announcement. Covers the original approval, the background to withdrawal and the scope of reapproval at the time.
[2] FDA researchers, 2018: Summary of the approval evidence for relapsed or refractory disease. Includes early dosing, pharmacokinetic rationale, MyloFrance-1 and safety concerns.
[3] DailyMed: U.S. prescribing information for Mylotarg. Used to check the mechanism and boxed warning. Doses in this article explain historical studies and are not individual dosing instructions.
[4] Petersdorf et al., Blood, 2013: SWOG S0106 randomized trial. Includes the different daunorubicin doses between groups and the unfavorable results.
[5] Lambert et al., Haematologica, 2019: Final efficacy and safety results of ALFA-0701. The 13.6 and 8.5 months in the text use this paper's blinded independent review. The prescribing information and other FDA analyses report EFS figures based on different definitions or assessment methods; this article does not mix those figures.
[6] Hills et al., Lancet Oncology, 2014: Individual patient data meta-analysis of five randomized trials.
[7] FDA / U.S. Federal Register, November 28, 2011: Formal withdrawal of Mylotarg approval. This contemporaneous legal document records the application in 2010 and formal withdrawal in 2011. It is the basis for distinguishing commercial withdrawal from formal withdrawal of approval.
[8] Pfizer, June 21, 2010: Discontinuation of commercial supply and arrangements for existing patients.
Sources checked through October 28, 2026. This article is for the exchange of medical and scientific information and does not provide individual diagnostic or treatment advice. Discuss specific treatment changes with your doctor.
Beyond the conclusion, there are still questions worth pursuing.
Medicine · Science · Ideas and Reading
Follow Before the Conclusion to stay curious and think independently together.
Text and image usage statement
The text of this article is original work by Before the Conclusion. You are welcome to share a link to this webpage. Without separate permission, please do not reproduce or excerpt the text by copying, reposting or republishing screenshots. Unless otherwise credited, accompanying illustrations may not be separately reposted, copied, modified or used commercially. Rights in cited material belong to their respective owners.