In October 2021, The Lancet Haematology published a small study, with Jeremy Abramson of Boston listed second among its authors, that found the lymphoma drug venetoclax, added to the chemotherapy regimen DA-EPOCH-R, had “an acceptable safety profile” at the dose chosen for further testing. Two months later, at the American Society of Hematology’s annual meeting, Abramson presented the first randomized test of the same pairing in double-hit lymphoma. On the venetoclax arm, six patients had died of adverse events. On chemotherapy alone, one had.

That sequence sits at the center of “Don’t second guess a trial?”, a post Vinay Prasad published on November 1. Prasad, the oncologist who announced in March that he was leaving the top job at the FDA’s biologics center, was answering colleagues who complain that second opinions talk patients out of clinical trials. He calls the notion that a trial is automatically the best treatment a “popular lie among oncologists,” and he argues that a consultant who steers a patient away from a study is sometimes giving the soundest advice that patient will hear.

The complaint he is answering has an obvious logic. Oncology advances through trials, trials need volunteers, and every patient talked out of one is a patient the study does not get. In that accounting the enrollment is the outcome worth protecting. Prasad counts the patient’s odds instead, and venetoclax is the case he uses to show the difference.

Thirty patients and no one to compare them with

The phase 1 study, registered as NCT03036904 and funded by Genentech, which develops venetoclax with AbbVie, enrolled 30 adults with previously untreated aggressive B-cell lymphomas at seven U.S. centers between February 2017 and June 2019. Every one of them got DA-EPOCH-R, an intensive multidrug regimen, plus venetoclax at rising doses until the investigators settled on 600 mg for five days per cycle, a dose picked for tolerability. According to the published abstract, 28 of the 30 had grade 3 or 4 cytopenias, 19 developed febrile neutropenia, and one died of sepsis attributed to treatment. Twenty-eight had a complete response, a rate of 93 percent, and with a single arm there was nothing to set that number against.

Prasad and others objected at the time. Most of these patients had a standard alternative in R-CHOP, a regimen he describes as “well established and curative” and credits with a “65-85% chance of cure,” and he asked why anyone would “skip a less toxic curative option” for a heavier chemotherapy backbone plus an investigational drug that suppresses the bone marrow. Enrolling, in his words, was “a bad decision.” His cure figure applies to the broad group of aggressive lymphomas the phase 1 took in. The randomized test that followed was confined to double-hit lymphoma, a narrower population, and the comparison it ran was with DA-EPOCH-R alone.

Seventy-three patients, two arms

That trial, ALLIANCE A051701, was paid for by the National Cancer Institute. Between October 2019 and September 2020 it randomized 73 patients with newly diagnosed double-hit lymphoma at 41 U.S. hospitals and clinics, 36 to DA-EPOCH-R alone and 37 to the same regimen plus venetoclax at the 600 mg, five-day dose the phase 1 had blessed. Abramson was first author when the full results appeared in The Lancet Haematology this September.

Six patients on the venetoclax arm died during treatment, four from sepsis and two from cardiac arrest, against one on chemotherapy alone. Pharmacy Times reported four later deaths on the venetoclax arm, from a lung infection, COVID-19, septic shock and hypoxia, and one later sepsis death among the controls, which makes the overall count ten to two.

Estimated two-year overall survival was 52 percent with venetoclax and 72 percent without it, a hazard ratio of 2.49 with a 95 percent confidence interval of 1.03 to 6.04. Median progression-free survival was 7.7 months on the combination against 28.4 months on chemotherapy alone, a gap that did not reach statistical significance. The complete-response rate, the phase 1’s headline number, came in at 56 percent with venetoclax and 77 percent without it.

TWO-YEAR OVERALL SURVIVAL
52percent
DA-EPOCH-R plus venetoclax
72percent
DA-EPOCH-R alone
Estimated two-year overall survival in the randomized double-hit lymphoma cohort of ALLIANCE A051701. Source: The ASCO Post: Excess mortality halts trial of venetoclax added to DA-EPOCH-R in double-hit lymphoma
OVERALL SURVIVAL HAZARD RATIO (HR)
Figure data
ResultValue (HR)Lower boundUpper bound
Venetoclax arm vs chemotherapy alone2.491.036.04
Values above the no-effect line mean a higher risk of death on the venetoclax arm. Source: Pharmacy Times: Trial ended early when venetoclax plus DA-EPOCH-R did not improve outcomes in double-hit lymphoma

The dosing records suggest why. DA-EPOCH-R is meant to be intensified cycle by cycle in patients who can take it, and the investigators reported that cytopenias and infections on the venetoclax arm got in the way. Doses could be escalated in 70 percent of the control arm but only 22 percent of the venetoclax arm, and 70 percent of controls completed all six cycles against 51 percent on the combination. Grade 3 or worse sepsis ran 23 percent with venetoclax and 13 percent without, while grade 3 or 4 febrile neutropenia was close, 43 percent on the venetoclax arm and 37 percent on the control arm. The trial cannot prove that the lost chemotherapy is what killed people, but the likeliest reading is that the added pill crowded out the part of the treatment that does the curing.

WHAT THE ADDED DRUG COST THE CHEMOTHERAPY
WHAT THE ADDED DRUG COST THE CHEMOTHERAPY
MeasureDA-EPOCH-R alonePlus venetoclax
Dose escalation possible70%22%
Completed all six cycles70%51%
Severe sepsis13%23%
Treatment delivery and serious infection by arm in the randomized double-hit cohort. Source: The ASCO Post: Excess mortality halts trial of venetoclax added to DA-EPOCH-R in double-hit lymphoma

The investigators did not soften it. Venetoclax “clearly enhances the toxicity of DA-EPOCH-R, and so this combination is not recommended,” they wrote, and the double-hit cohort was closed early because of the excess deaths.


Prasad’s explanation for how studies like the phase 1 get approved and filled is about incentives, and he spreads them around. Companies want market share. Investigators have careers to build, and “no one gets promoted by opposing bad trials.” Universities make money from trials. Institutional review boards, he writes, include members who know little about oncology, a gap he concedes is hard to fault them for. Nothing in that list requires anyone to be corrupt; it only requires that everyone at the table has a reason to say yes except the patient, who is the one being dosed.

The funding in this case follows that shape without proving anyone’s motive. A manufacturer paid for the study that had one arm and reported a 93 percent complete-response rate. Taxpayers paid for the study that had two arms and reported the deaths. Prasad extends the argument to incurable cancers, where trials that require progression on earlier drugs can push doctors to “burn through” approved therapies to qualify patients, and where he wants a second opinion in the room before that sequencing is decided.

The oncologists lamenting second opinions are asking patients to take an investigational drug on top of a standard regimen, and venetoclax shows how a promising single arm can collapse once a control arm is counted. A consultant who asks to see the comparison before the patient signs is doing what the randomized trial eventually did, only sooner.

Jeremy Abramson is on both papers: second author on the one that found an acceptable safety profile, first author on the one that told doctors not to use the combination. Only the second one had a control arm.

Sources

  1. Vinay Prasad, “Don’t second guess a trial?”
  2. The ASCO Post: Excess mortality halts trial of venetoclax added to DA-EPOCH-R in double-hit lymphoma
  3. Pharmacy Times: Trial ended early when venetoclax plus DA-EPOCH-R did not improve outcomes in double-hit lymphoma
  4. Rutherford, Abramson, Bartlett et al., venetoclax with DA-EPOCH-R, single-arm phase 1 study, The Lancet Haematology (2021)
  5. Oncologie.nu: ASH 2021 report on added toxicity and mortality with venetoclax plus DA-EPOCH-R in double-hit lymphoma
  6. ClinicalTrials.gov: ALLIANCE A051701, NCT03984448
  7. ClinicalTrials.gov: Venetoclax plus DA-EPOCH-R phase 1, NCT03036904