Establishing Evidence:

A Collection of
Confirmatory Trials
How long-term evidence informs clinical practice
 July 2023
Dear Reader,
It has been 18 months since we launched NEJM Evidence, with a mission to provide clinical investigators and physicians
caring for patients with research and evidence that informs clinical decision-making and to shine a spotlight on key areas
where more study is critical. The original research that NEJM Evidence publishes falls along the full spectrum of human
investigation — from first-in-human studies that offer insight into the safety and potential efficacy of novel therapeutics
to the late-stage confirmatory trials that solidify the evidence base that forms the foundation of clinical practice.

Highlighting the latter of these two research types, this special content collection from NEJM Evidence offers a selection of
confirmatory, multiyear follow-ups of original research breakthroughs, many of which were published in the New England
Journal of Medicine. Each of the NEJM Evidence articles in this collection clarifies and builds on current medical evidence —
to identify something not previously considered, to confirm current clinical practices thereby reinforcing clinical practice,
or to create dialog around where more evidence and study is required. Here are a few highlights:

CONFIRMING DURABILITY OF EFFECT

Over the past decade, immunotherapy has transformed the treatment landscape of advanced melanoma. For example, the
RELATIVITY-047 trial whose results were published in an original research article in the New England Journal of Medicine
(NEJM) in January 2022 showed that the inhibition of two immune checkpoints, LAG-3 and PD-1, provided a greater benefit
with regard to progression-free survival than inhibition of PD-1 alone in patients with previously untreated metastatic or
unresectable melanoma. But were these results durable? NEJM Evidence’s original research article, “Overall Survival and
Response with Nivolumab and Relatlimab in Advanced Melanoma,” confirmed the durable effect on progression-free
survival, and encouraging trends in overall survival and the objective response rate after approximately 6 months of
additional follow-up.

The second article is “Patient-Reported Outcomes 12 Years after Localized Prostate Cancer Treatment,” a 12-year follow-up
of the ProtecT trial report published in NEJM in October 2016 with an update on clinical findings published in NEJM in April
of 2023. What was not included in the NEJM article were the patient-centered outcomes (i.e., the ones that make a huge
difference in a patient’s quality of life). The article published in NEJM Evidence provided the answers to these questions with
data on the medium, and long-term effects of treatments on urinary, sexual, and bowel function, thus enabling patients
newly diagnosed with low or intermediate risk localized prostate cancer to carefully consider and be well-informed about
the trade-offs between treatment harms and prostate cancer progression risks.

Another article in this collection reports the final long-term results of ABCSG-18 trial, that was published in Lancet Oncology
in 2019. That trial was designed to determine whether adjuvant denosumab, administered at a dose of 60 mg every 6 months,
improved bone health and breast cancer outcomes compared with placebo in postmenopausal patients with early hormone
receptor–positive breast cancer. When those data were published, there was concern that the positive impact on survival was
fragile and thus definitive conclusions could not be drawn. The NEJM Evidence article, published in November of 2022, added
2 additional years of median follow-up and approximately 30% more disease free survival events compared with the earlier
report. The new data shows that the subcutaneous administration of adjuvant denosumab 60 mg twice per year led to a
modest benefit in postmenopausal patients receiving this type of therapy; the data also showed there was a continued
benefit on bone fractures.

Two other confirmatory studies — one a 7-year follow-up study on early antiretroviral therapy (ART) for HIV infection and
the other a 5-year follow-up of adjuvant immune checkpoint inhibitor therapy for resected stage III melanoma — confirm

(continued on next page)

781.893.3800 | @nejmevidence
860 winter street, waltham, ma 02451-1413 usa | evidence.nejm.org
NEJM Evidence is a product of NEJM Group, a division of the Massachusetts Medical Society.
(continued from previous page)

the long-term results of two breakthrough original research articles, both published in NEJM in August 2015 and in May 2018
respectively. The HIV infection extension article provided solid evidence that early diagnosis of HIV and prompt initiation
of ART saved lives; data that were needed to confirm the ongoing benefit of early treatment of HIV. The melanoma article
demonstrates that adjuvant therapy with pembrolizumab provides long-term, sustained improvement of recurrence and
distant metastasis-free survival compared to placebo in patients with resected stage III melanoma.

EXTENDING THE TREATMENT WINDOW

Whether phlebotomy alone can adequately maintain target hematocrit in patients with low-risk polycythemia vera (PV)
remains elusive. The trial published in NEJM Evidence, “Ropeginterferon (ropeg) versus Standard Therapy for Low-Risk
Patients with Polycythemia Vera” provided novel evidence that in patients with low-risk PV, treatment with low-dose ropeg
for 2 years in addition to standard treatment with therapeutic phlebotomy and aspirin was more efficacious than standard
treatment alone. Treatment with ropeg consistently maintained hematocrit at the target level with a reduction of phlebotomy
needed without thrombotic complications.

NEJM Evidence is a place for investigators — at all stages of research careers — and clinicians — whether in training now
or in practice for decades — to come together to learn about advances in medicine and clinical trial design. We hope you
enjoy this curated article collection and find NEJM Evidence a valuable resource. Send us your comments anytime at
editorial.evidence@nejm.org … we’d like to hear from you!

Sincerely,
NEJM Evidence Editorial Team

Jeffrey M. Drazen, MD      Chana A. Sacks, MD, MPH

Editor-in-Chief              
Executive Editor
Table of Contents

Advanced Melanoma
NEJM EVIDENCE ORIGINAL ARTICLE
Overall Survival and Response with Nivolumab and Relatlimab in Advanced Melanoma   1

NEJM ORIGINAL ARTICLE

Relatlimab and Nivolumab versus Nivolumab in Untreated Advanced Melanoma   2

Prostate Cancer
NEJM EVIDENCE ORIGINAL ARTICLE
Patient-Reported Outcomes 12 Years after Localized Prostate Cancer Treatment   3

NEJM ORIGINAL ARTICLE

Fifteen-Year Outcomes after Monitoring, Surgery, or Radiotherapy for Prostate Cancer   4

NEJM ORIGINAL ARTICLE

Patient-Reported Outcomes after Monitoring, Surgery, or Radiotherapy for Prostate Cancer   5

Breast Cancer
NEJM EVIDENCE ORIGINAL ARTICLE
Long-Term Outcomes of Adjuvant Denosumab in Breast Cancer   6

ACCOMPANYING NEJM EVIDENCE EDITORIAL

Bone-Modifying Agents in Early Breast Cancer: Making Sense of Conflicting Data   7

HIV Infection
NEJM EVIDENCE ORIGINAL ARTICLE
Long-Term Benefits from Early Antiretroviral Therapy Initiation in HIV Infection   9

ACCOMPANYING NEJM EVIDENCE EDITORIAL

Scientific Inference Does Not Rest on Statistical Testing 10

NEJM ORIGINAL ARTICLE

Initiation of Antiretroviral Therapy in Early Asymptomatic HIV Infection 12

continued on next page

The New England Journal of Medicine and NEJM Evidence are publications of NEJM Group, a division of the Massachusetts Medical Society.
©2023 Massachusetts Medical Society. All rights reserved.
Table of Contents
continued from previous page

Stage III Melanoma

NEJM EVIDENCE ORIGINAL ARTICLE
Five-Year Analysis of Adjuvant Pembrolizumab or Placebo in Stage III Melanoma 13

ACCOMPANYING NEJM EVIDENCE EDITORIAL

Progression/Recurrence-Free Survival 2 in Adjuvant Melanoma 14

NEJM ORIGINAL ARTICLE

Adjuvant Pembrolizumab versus Placebo in Resected Stage III Melanoma 16

Polycythemia Vera
NEJM EVIDENCE ORIGINAL ARTICLE
Ropeginterferon versus Standard Therapy for Low-Risk Patients with Polycythemia Vera 17

The New England Journal of Medicine and NEJM Evidence are publications of NEJM Group, a division of the Massachusetts Medical Society.
©2023 Massachusetts Medical Society. All rights reserved.
evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 1

Published March 22, 2023

NEJM Evid 2023; 2 (4)
DOI: 10.1056/EVIDoa2200239

ORIGINAL ARTICLE

Overall Survival and Response with Nivolumab

and Relatlimab in Advanced Melanoma
Georgina V. Long, M.D., Ph.D.,1 F. Stephen Hodi, M.D.,2 Evan J. Lipson, M.D.,3 Dirk Schadendorf, M.D.,4,5 Paolo A.
Ascierto, M.D.,6 Luis Matamala, M.D.,7 Pamela Salman, M.D.,7 Erika Castillo Gutiérrez, M.D.,8 Piotr Rutkowski, M.D., Ph.D.,9
Helen J. Gogas, M.D.,10 Christopher D. Lao, M.D., M.P.H.,11 Juliana Janoski De Menezes, M.D.,12 Stéphane Dalle, M.D.,
Ph.D.,13 Ana Arance, M.D., Ph.D.,14 Jean-Jacques Grob, M.D.,15 Sarah Keidel, M.B., Ch.B.,16 Anadil Shaikh, M.Sc.,16
Anne Marie Sobiesk, Ph.D.,16 Sonia Dolfi, Ph.D.,16 and Hussein A. Tawbi, M.D., Ph.D.17

Abstract
BACKGROUND A phase 2/3 trial — A Study of Relatlimab Plus Nivolumab Versus Nivolumab
Alone in Participants With Advanced Melanoma (RELATIVITY-047) — evaluated nivolumab +
relatlimab as a fixed-dose combination and found a significant progressionfree survival
(PFS) benefit over nivolumab monotherapy in previously untreated unresectable or metastatic
melanoma. We now report updated PFS and safety data and the first results for overall
survival (OS) and objective response rate (ORR).
METHODS Patients were randomly assigned 1:1 to receive nivolumab 480 mg and relatlimab
160 mg fixed-dose combination or nivolumab 480 mg alone, given intravenously every
4 weeks. PFS (primary end point) according to the Response Evaluation Criteria in Solid
Tumors, version 1.1, was assessed by blinded independent central review (BICR). Secondary
end points, tested hierarchically, were OS and then ORR per Response Evaluation Criteria in
Solid Tumors, version 1.1, per BICR.
RESULTS At a median follow-up of 19.3 months, median PFS according to BICR was
10.2 months (95% confidence interval [CI], 6.5 to 14.8) with nivolumab + relatlimab versus
4.6 months (95% CI, 3.5 to 6.4) with nivolumab (hazard ratio, 0.78; 95% CI, 0.64 to 0.94).
Median OS was not reached (NR) (95% CI, 34.2 to NR) with nivolumab + relatlimab versus The author affiliations are listed
at the end of the article.
34.1 months (95% CI, 25.2 to NR) with nivolumab (hazard ratio, 0.80; 95% CI, 0.64 to 1.01;
P=0.059) (prespecified value for statistical significance, P≤0.043). ORRs per BICR were Dr. Long can be contacted at
43.1% (95% CI, 37.9 to 48.4) versus 32.6% (95% CI, 27.8 to 37.7), respectively. Grade 3/4 georgina.long@sydney.edu.au or
treatment-related adverse events were observed in 21.1% of patients treated with nivolumab at Melanoma Institute Australia,
+ relatlimab versus 11.1% treated with nivolumab. 40 Rocklands Rd., North Sydney,
NSW 2060, Australia.
CONCLUSIONS The fixed-dose combination of nivolumab + relatlimab showed consistent
PFS benefit versus nivolumab with approximately 6 months of additional median follow-up. This article was updated on April
The combination treatment did not reach the preplanned statistical threshold for OS, with a 25, 2023, at evidence.nejm.org.
10.3 percentage-point difference in ORR. Grade 3/4 treatment-related adverse events were
more frequent with nivolumab + relatlimab versus nivolumab. (Funded by Bristol Myers Read Full Article
Squibb; ClinicalTrials.gov number, NCT03470922.) at evidence.nejm.org

Back to Table of Contents

2 Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article
ORIGINAL ARTICLE

Relatlimab and Nivolumab versus Nivolumab

in Untreated Advanced Melanoma
Hussein A. Tawbi, M.D., Ph.D., Dirk Schadendorf, M.D., Evan J. Lipson, M.D.,
Paolo A. Ascierto, M.D., Luis Matamala, M.D., Erika Castillo Gutiérrez, M.D.,
Piotr Rutkowski, M.D., Ph.D., Helen J. Gogas, M.D., Christopher D. Lao, M.D., M.P.H.,
Juliana Janoski De Menezes, M.D., Stéphane Dalle, M.D., Ph.D.,
Ana Arance, M.D., Ph.D., Jean-Jacques Grob, M.D., Shivani Srivastava, M.D.,
Mena Abaskharoun, Pharm.D., Melissa Hamilton, M.P.H., Sarah Keidel, M.B., Ch.B.,
Katy L. Simonsen, Ph.D., Anne Marie Sobiesk, Ph.D., Bin Li, Ph.D.,
F. Stephen Hodi, M.D., and Georgina V. Long, M.D., Ph.D.,
for the RELATIVITY-047 Investigators*

A BS T R AC T

BACKGROUND
Lymphocyte-activation gene 3 (LAG-3) and programmed death 1 (PD-1) are distinct The authors’ affiliations are listed in the
Appendix. Dr. Tawbi can be contacted at
inhibitory immune checkpoints that contribute to T-cell exhaustion. The combina-
htawbi@mdanderson.org or at the De-
tion of relatlimab, a LAG-3–blocking antibody, and nivolumab, a PD-1–blocking partment of Melanoma Medical Oncol-
antibody, has been shown to be safe and to have antitumor activity in patients with ogy, Division of Cancer Medicine, Uni-
versity of Texas M.D. Anderson Cancer
previously treated melanoma, but the safety and activity in patients with previ-
Center, 1400 Holcombe Blvd., Houston,
ously untreated melanoma need investigation. TX 77030.

METHODS *The RELATIVITY-047 Investigators are

listed in the Supplementary Appendix,
In this phase 2–3, global, double-blind, randomized trial, we evaluated relatlimab available at NEJM.org.
and nivolumab as a fixed-dose combination as compared with nivolumab alone
Drs. Hodi and Long contributed equally
when administered intravenously every 4 weeks to patients with previously un- to this article.
treated metastatic or unresectable melanoma. The primary end point was progres-
This article was updated on January 6,
sion-free survival as assessed by blinded independent central review. 2022, at NEJM.org.

RESULTS N Engl J Med 2022;386:24-34.

DOI: 10.1056/NEJMoa2109970
The median progression-free survival was 10.1 months (95% confidence interval
Copyright © 2022 Massachusetts Medical Society.
[CI], 6.4 to 15.7) with relatlimab–nivolumab as compared with 4.6 months (95%
CI, 3.4 to 5.6) with nivolumab (hazard ratio for progression or death, 0.75 [95% CI,
Read Full Article at NEJM.org
0.62 to 0.92]; P = 0.006 by the log-rank test). Progression-free survival at 12 months
was 47.7% (95% CI, 41.8 to 53.2) with relatlimab–nivolumab as compared with
36.0% (95% CI, 30.5 to 41.6) with nivolumab. Progression-free survival across key
subgroups favored relatlimab–nivolumab over nivolumab. Grade 3 or 4 treatment-
related adverse events occurred in 18.9% of patients in the relatlimab–nivolumab
group and in 9.7% of patients in the nivolumab group.
CONCLUSIONS
The inhibition of two immune checkpoints, LAG-3 and PD-1, provided a greater
benefit with regard to progression-free survival than inhibition of PD-1 alone in
patients with previously untreated metastatic or unresectable melanoma. Relatlimab
and nivolumab in combination showed no new safety signals. (Funded by Bristol
Myers Squibb; RELATIVITY-047 ClinicalTrials.gov number, NCT03470922.)

24 n engl j med 386;1 nejm.org January 6, 2022 Back to Table of Contents

evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 3

Published March 11, 2023

NEJM Evid 2023; 2 (4)
DOI: 10.1056/EVIDoa2300018

ORIGINAL ARTICLE

Patient-Reported Outcomes 12 Years after Localized

Prostate Cancer Treatment
Jenny L. Donovan, Ph.D., F.Med.Sci.,1 Freddie C. Hamdy, F.R.C.S. (Urol.), F.Med.Sci.,2 J. Athene Lane, Ph.D.,1,3
Grace J. Young, M.Sc.,1,3 Chris Metcalfe, Ph.D.,1,3 Eleanor I. Walsh, M.Sc.,1 Michael Davis, M.Sc.,1
Thomas Steuart-Feilding, B.A.,1 Jane M. Blazeby, F.R.C.S. (Gen. Surg.), F.Med.Sci.,1 Kerry N. L. Avery, Ph.D.,1
Richard M. Martin, B.M.B.S., Ph.D.,1 Prasad Bollina, M.B.B.S., F.R.C.S. (Urol.),4 Andrew Doble, M.S., F.R.C.S. (Urol.),5
Alan Doherty, F.R.C.S. (Urol.),6 David Gillatt, M.S., F.R.C.S. (Urol.),7 Vincent Gnanapragasam, Ph.D., F.R.C.S. (Urol.),8
Owen Hughes,9 Roger Kockelbergh, D.M., F.R.C.S. (Urol.),10 Howard Kynaston, M.D., F.R.C.S. (Urol.),9
Alan Paul, M.D., F.R.C.S. (Urol.),11 Edgar Paez, F.R.C.S. (Urol.),12 Phillip Powell, M.D., F.R.C.S.,12
Derek J. Rosario, M.D., F.R.C.S. (Urol.),13 Edward Rowe, M.D., F.R.C.S. (Urol.),14 Malcolm Mason, M.D., F.R.C.R.,15
James W. F. Catto, Ph.D., F.R.C.S. (Urol.),13,16 Tim J. Peters, Ph.D., F.Med.Sci.,1 Julia Wade, Ph.D.,1 Emma L. Turner,
Ph.D.,1 Naomi J. Williams, Ph.D.,1 Jon Oxley, M.D., F.R.C.Path.,17 John Staffurth, F.R.C.R., F.R.C.P.,18
Richard J. Bryant, Ph.D., F.R.C.S. (Urol.),2 David E. Neal, C.B.E., F.Med.Sci.,2,8 for the ProtecT study group*

Abstract
BACKGROUND Long-term patient-reported outcomes are needed to inform treatment deci-
sions for localized prostate cancer.
METHODS Patient-reported outcomes of 1643 randomly assigned participants in the ProtecT
(Prostate Testing for Cancer and Treatment) trial were evaluated to assess the functional and Professor Donovan, Professor
quality-of-life impacts of prostatectomy, radiotherapy with neoadjuvant androgen depriva- Hamdy, Professor Lane, Ms.
tion, and active monitoring. This article focuses on the outcomes of the randomly assigned Young, Professor Metcalfe, and
participants from 7 to 12 years using mixed effects linear and logistic models. Professor Neal contributed equally
to this article.
RESULTS Response rates exceeded 80% for most measures. Among the randomized groups
over 7 to 12 years, generic quality-of-life scores were similar. Among those in the prostatec- *A complete list of investigators
tomy group, urinary leakage requiring pads occurred in 18 to 24% of patients over 7 to in the ProtecT trial is provided
12 years, compared with 9 to 11% in the active monitoring group and 3 to 8% in the radio- in the Supplementary Appendix,
therapy group. In the prostatectomy group, 18% reported erections sufficient for intercourse available at evidence.nejm.org.
at 7 years, compared with 30% in the active monitoring and 27% in the radiotherapy groups;
The author affiliations are listed
all converged to low levels of potency by year 12. Nocturia (voiding at least twice per night) at the end of the article. Professor
occurred in 34% in the prostatectomy group compared with 48% in the radiotherapy group Donovan can be contacted at
and 47% in the active monitoring group at 12 years. Fecal leakage affected 12% in the jenny.donovan@ bristol.ac.uk
radiotherapy group compared with 6% in the other groups by year 12. The active monitoring or at Population Health Sciences,
group experienced gradual age-related declines in sexual and urinary function, avoiding Bristol Medical School, Univer-
radical treatment effects unless they changed management. sity of Bristol, Bristol BS8 1QU,
United Kingdom.
CONCLUSIONS ProtecT provides robust evidence about continued impacts of treatments
in the long term. These data allow patients newly diagnosed with localized prostate cancer This article was updated on May
and their clinicians to assess the trade-offs between treatment harms and benefits and en- 10, 2023, at evidence.nejm.org.
able better informed and prudent treatment decisions. (Funded by the UK National Institute
for Health and Care Research Health Technology Assessment Programme projects 96/20/06 Read Full Article
and 96/20/99; ISRCTN number, ISRCTN20141297; ClinicalTrials.gov number, NCT02044172.) at evidence.nejm.org
Back to Table of Contents
4
new england
Establishing Evidence: A Collection of Confirmatory Trials
The
evidence.nejm.org

journal of medicine
established in 1812 April 27, 2023 vol. 388 no. 17

ORIGINAL ARTICLE

Fifteen-Year Outcomes after Monitoring, Surgery,

or Radiotherapy for Prostate Cancer
F.C. Hamdy, J.L. Donovan, J.A. Lane, C. Metcalfe, M. Davis, E.L. Turner, R.M. Martin, G.J. Young, E.I. Walsh,
R.J. Bryant, P. Bollina, A. Doble, A. Doherty, D. Gillatt, V. Gnanapragasam, O. Hughes, R. Kockelbergh,
H. Kynaston, A. Paul, E. Paez, P. Powell, D.J. Rosario, E. Rowe, M. Mason, J.W.F. Catto, T.J. Peters, J. Oxley,
N.J. Williams, J. Staffurth, and D.E. Neal, for the ProtecT Study Group*

a bs t r ac t

BACKGROUND
Between 1999 and 2009 in the United Kingdom, 82,429 men between 50 and 69 The authors’ full names, academic de-
years of age received a prostate-specific antigen (PSA) test. Localized prostate grees, and affiliations are listed in the
Appendix. Dr. Hamdy can be contacted
cancer was diagnosed in 2664 men. Of these men, 1643 were enrolled in a trial to at freddie.hamdy@nds.ox.ac.uk or at the
evaluate the effectiveness of treatments, with 545 randomly assigned to receive Nuffield Department of Surgical Sciences,
active monitoring, 553 to undergo prostatectomy, and 545 to undergo radiotherapy. University of Oxford, Old Road Campus
Research Building, Roosevelt Dr., Heading-
METHODS ton, Oxford OX3 7DQ, United Kingdom.
At a median follow-up of 15 years (range, 11 to 21), we compared the results in *A list of members of the ProtecT Study
this population with respect to death from prostate cancer (the primary outcome) Group is provided in the Supplemen-
tary Appendix, available at NEJM.org.
and death from any cause, metastases, disease progression, and initiation of long-
term androgen-deprivation therapy (secondary outcomes). Drs. Hamdy, Donovan, Lane, Metcalfe,
and Neal contributed equally to this article.
RESULTS
This article was published on March 11,
Follow-up was complete for 1610 patients (98%). A risk-stratification analysis 2023, at NEJM.org.
showed that more than one third of the men had intermediate or high-risk disease
N Engl J Med 2023;388:1547-58.
at diagnosis. Death from prostate cancer occurred in 45 men (2.7%): 17 (3.1%) in DOI: 10.1056/NEJMoa2214122
the active-monitoring group, 12 (2.2%) in the prostatectomy group, and 16 (2.9%) Copyright © 2023 Massachusetts Medical Society.

in the radiotherapy group (P = 0.53 for the overall comparison). Death from any
CME
cause occurred in 356 men (21.7%), with similar numbers in all three groups. Read Full Article at NEJM.org
at NEJM.org
Metastases developed in 51 men (9.4%) in the active-monitoring group, in 26
(4.7%) in the prostatectomy group, and in 27 (5.0%) in the radiotherapy group.
Long-term androgen-deprivation therapy was initiated in 69 men (12.7%), 40 (7.2%),
and 42 (7.7%), respectively; clinical progression occurred in 141 men (25.9%), 58
(10.5%), and 60 (11.0%), respectively. In the active-monitoring group, 133 men
(24.4%) were alive without any prostate cancer treatment at the end of follow-up.
No differential effects on cancer-specific mortality were noted in relation to the
baseline PSA level, tumor stage or grade, or risk-stratification score. No treatment
complications were reported after the 10-year analysis.
CONCLUSIONS
After 15 years of follow-up, prostate cancer–specific mortality was low regardless
of the treatment assigned. Thus, the choice of therapy involves weighing trade-offs
between benefits and harms associated with treatments for localized prostate
cancer. (Funded by the National Institute for Health and Care Research; ProtecT
Current Controlled Trials number, ISRCTN20141297; ClinicalTrials.gov number,
NCT02044172.)
n engl j med 388;17 nejm.org April 27, 2023 1547

Back to Table of Contents

evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 5

The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article
ORIGINAL ARTICLE

Patient-Reported Outcomes after Monitoring,

Surgery, or Radiotherapy for Prostate Cancer
J.L. Donovan, F.C. Hamdy, J.A. Lane, M. Mason, C. Metcalfe, E. Walsh,
J.M. Blazeby, T.J. Peters, P. Holding, S. Bonnington, T. Lennon, L. Bradshaw,
D. Cooper, P. Herbert, J. Howson, A. Jones, N. Lyons, E. Salter, P. Thompson,
S. Tidball, J. Blaikie, C. Gray, P. Bollina, J. Catto, A. Doble, A. Doherty, D. Gillatt,
R. Kockelbergh, H. Kynaston, A. Paul, P. Powell, S. Prescott, D.J. Rosario, E. Rowe,
M. Davis, E.L. Turner, R.M. Martin, and D.E. Neal, for the ProtecT Study Group*

A BS T R AC T

BACKGROUND
Robust data on patient-reported outcome measures comparing treatments for clinically The authors’ full names, academic de-
localized prostate cancer are lacking. We investigated the effects of active monitoring, grees, and affiliations are listed in the
Appendix. Address reprint requests to
radical prostatectomy, and radical radiotherapy with hormones on patient-reported outcomes. Dr. Donovan at the School of Social and
Community Medicine, University of Bris-
METHODS tol, Canynge Hall, 39 Whatley Rd., Bristol
We compared patient-reported outcomes among 1643 men in the Prostate Testing for BS8 2PS, United Kingdom, or at jenny
Cancer and Treatment (ProtecT) trial who completed questionnaires before diagnosis, .donovan@bristol.ac.uk.
at 6 and 12 months after randomization, and annually thereafter. Patients completed * A complete list of investigators in the
validated measures that assessed urinary, bowel, and sexual function and specific ef- Prostate Testing for Cancer and Treat-
ment (ProtecT) Study Group is provided
fects on quality of life, anxiety and depression, and general health. Cancer-related qual- in the Supplementary Appendix, avail-
ity of life was assessed at 5 years. Complete 6-year data were analyzed according to the able at NEJM.org.
intention-to-treat principle. Drs. Donovan, Hamdy, Lane, and Neal
contributed equally to this article.
RESULTS
The rate of questionnaire completion during follow-up was higher than 85% for most This article was published on September
14, 2016, and updated on May 22, 2023,
measures. Of the three treatments, prostatectomy had the greatest negative effect on at NEJM.org.
sexual function and urinary continence, and although there was some recovery, these
N Engl J Med 2016;375:1425-37.
outcomes remained worse in the prostatectomy group than in the other groups DOI: 10.1056/NEJMoa1606221
throughout the trial. The negative effect of radiotherapy on sexual function was greatest Copyright © 2016 Massachusetts Medical Society.
at 6 months, but sexual function then recovered somewhat and was stable thereafter;
radiotherapy had little effect on urinary continence. Sexual and urinary function de- Read Full Article at NEJM.org
clined gradually in the active-monitoring group. Bowel function was worse in the radio-
therapy group at 6 months than in the other groups but then recovered somewhat,
except for the increasing frequency of bloody stools; bowel function was unchanged
in the other groups. Urinary voiding and nocturia were worse in the radiotherapy
group at 6 months but then mostly recovered and were similar to the other groups
after 12 months. Effects on quality of life mirrored the reported changes in function.
No significant differences were observed among the groups in measures of anxiety,
depression, or general health-related or cancer-related quality of life.
CONCLUSIONS
In this analysis of patient-reported outcomes after treatment for localized prostate cancer,
patterns of severity, recovery, and decline in urinary, bowel, and sexual function and as-
sociated quality of life differed among the three groups. (Funded by the U.K. National
Institute for Health Research Health Technology Assessment Program; ProtecT Current
Controlled Trials number, ISRCTN20141297; ClinicalTrials.gov number, NCT02044172.)

n engl j med 375;15 nejm.org October 13, 2016 1425

Back to Table of Contents
6 Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

Published November 18, 2022

NEJM Evid 2022; 1 (12)
DOI: 10.1056/EVIDoa2200162

ORIGINAL ARTICLE

Long-Term Outcomes of Adjuvant Denosumab

in Breast Cancer
Michael Gnant, M.D.,1,2 Sophie Frantal, M.Sc.,2 Georg Pfeiler, M.D.,3 Guenther G. Steger, M.D.,4 Daniel Egle, M.D.,5
Richard Greil, M.D.,6 Florian Fitzal, M.D.,7 Viktor Wette, M.D.,8 Marija Balic, M.D.,9 Ferdinand Haslbauer, M.D.,10
Elisabeth Melbinger-Zeinitzer, M.D.,11 Vesna Bjelic-Radisic, M.D.,12,13 Silvia Artner-Matuschek, M.D.,14
Franz Kainberger, M.D.,15 Magdalena Ritter, M.D.,5 Gabriel Rinnerthaler, M.D.,6 Paul Sevelda, M.D.,16 Jonas Bergh,
M.D.,17 Stephanie Kacerovsky-Strobl, M.D.,7,18 Christoph Suppan, M.D.,9 Christine Brunner, M.D.,5 Christine Deutschmann,
M.D.,3 Simon P. Gampenrieder, M.D.,6 Hannes Fohler, M.V.M.,2 Raimund Jakesz, M.D.,7 Christian Fesl, Ph.D.,2
and Christian Singer, M.D.,3 for the Austrian Breast & Colorectal Cancer Study Group*

Abstract
BACKGROUND Adjuvant aromatase inhibitors increase osteoporosis and fractures in pa-
tients with hormone receptor-positive breast cancer. We have previously reported outcomes
of the ABCSG-18 (study 18 from the Austrian Breast & Colorectal Cancer Study Group) trial
showing that adjuvant anti-receptor activator of nuclear factor-κB ligand denosumab treat-
ment counteracts these adverse effects and may improve outcomes. We report here the final
long-term outcomes.
METHODS ABCSG-18 is a prospective, double-blind, placebo-controlled, phase 3 trial in
which 3425 postmenopausal patients with early hormone receptor-positive breast cancer re-
ceiving aromatase inhibitor therapy were randomly assigned in 58 trial centers to receive either
*A complete list of collaborators
denosumab 60 mg or placebo administered subcutaneously every 6 months. The primary end in the Austrian Breast&Colorectal
point was the time to first clinical fracture after randomization. Secondary disease outcome- Cancer Study Group is provided
related end points were disease-free survival (DFS), bone metastasis-free survival (BMFS), and in the Supplementary Appendix,
overall survival (OS). available at evidence.nejm.org.
RESULTS For this final protocol-defined analysis, median follow-up is 8 years (interquartile
The author affiliations are listed at
range, 6 to 9.6 years). There were 309 versus 368 DFS events (hazard ratio, 0.83; 95% confi- the end of the article.
dence interval [CI], 0.71 to 0.97) in the denosumab versus the placebo group, respectively,
resulting in an absolute 9-year DFS benefit of 3.5 percentage points (79.4 vs. 75.9%). Adju- Dr. Gnant can be contacted at
vant denosumab improved BMFS by 2.5 percentage points (88.9 vs. 86.4%; hazard ratio, michael.gnant@meduniwien.ac.at
0.81; 95% CI, 0.65 to 1.00) and OS by 1.0 percentage point (90.9 vs. 89.9%; hazard ratio, or at Comprehensive Cancer Center,
Medical University of Vienna,
0.80; 95% CI, 0.64 to 1.01). No new toxicities for this dose of adjuvant denosumab were
Waehringer Guertel 18-20,
observed.
A-1090 Vienna, Austria.
CONCLUSIONS DFS, BMFS, and OS continued to show benefit in this final long-term
analysis of ABCSG-18. There were no new toxicities. (Funded by Amgen; ClinicalTrials.gov Read Full Article
number, NCT00556374) at evidence.nejm.org

Back to Table of Contents

evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 7

Published November 22, 2022

NEJM Evid 2022; 1 (12)
DOI: 10.1056/EVIDe2200259

EDITORIAL

Bone-Modifying Agents in Early Breast Cancer:

Making Sense of Conflicting Data
Erica T. Warner, Sc.D., M.P.H.,1 and Beverly Moy, M.D., M.P.H.1

T
he use of adjuvant bone-modifying agents to reduce risk of recurrence in patients
with early-stage breast cancer has not been widely embraced because of conflict-
ing data and small absolute benefits. The clinical practice guideline produced
jointly by the American Society of Clinical Oncology and Cancer Care Ontario recom-
mends discussion of risks/benefits with postmenopausal patients with early-stage breast
cancer about adjuvant bisphosphonates and does not recommend use of adjuvant denosu-
mab to prevent breast cancer recurrence.1 In this issue of NEJM Evidence, Gnant et al.2
report the final long-term data from ABCSG-18 (study 18 from the Austrian Breast & Colorec-
tal Cancer Study Group), a prospective, double-blind, placebo-controlled, multicenter, ran-
domized phase 3 trial of postmenopausal women with early-stage hormone receptor–positive
breast cancer. Participants were randomized to receive either denosumab 60 mg subcutane-
ously every 6 months or placebo. The primary end point was first clinical fracture postran-
domization; secondary end points were disease-free survival, bone metastasis–free survival,
and overall survival. An earlier report with a median follow-up of 6.1 years found that the
intervention was associated with improved disease-free survival (hazard ratio, 0.82; 95%
confidence interval, 0.69 to 0.98; absolute benefit, 3.1 percentage points).3

How do the results of this final protocol analysis of ABCSG-18, which showed an absolute
benefit in 9-year disease-free survival of 3.5 percentage points, bone metastasis–free sur-
vival of 1.5 percentage points, and overall survival of 1 percentage point,2 affect our prac-
tice now?

The results of ABCSG-182 need to be taken into context with the contrasting results of the
randomized D-CARE (Study of Denosumab as Adjuvant Treatment for Women With High
Risk Early Breast Cancer Receiving Neoadjuvant or Adjuvant Therapy; ClinicalTrials.gov
identifier, NCT01077154) trial, which showed that adjuvant denosumab did not improve
breast cancer outcomes.4 Differences in trial design and patient populations of the two tri-
als are notable. The larger D-CARE study evaluated denosumab given at a higher dose and
greater frequency in patients with breast cancer at higher risk of recurrence compared with
ABCSG-18. Furthermore, the statistical analysis of ABCSG-18 was descriptive and affected The author affiliations are listed
at the end of the article.
by crossover, and its primary end point was time to first clinical fracture, rather than breast
cancer outcomes.2 Dr. Moy can be contacted at
BMOY@mgh.harvard.edu or
Massachusetts General Hospital,
Results of another important related clinical trial, TEAM-IIB (ISRCTN17633610), have also 55 Fruit Street, Boston, MA,
been recently reported.5 TEAM-IIB showed that adjuvant treatment with the potent oral 02114-2696.

Back to Table of Contents

For personal use only. No other uses without permission. Copyright © 2022 Massachusetts Medical Society.
8 Establishing Evidence: A Collection of Confirmatory Trials
bisphosphonate ibandronate did not improve disease-free
survival. The results of TEAM-IIB now call into question
whether clinicians should rethink altogether the use of
adjuvant bisphosphonates in patients with early-stage
breast cancer.
The ABCSG-18 results2 must also be interpreted with cau-
tion given the stark lack of racial/ethnic diversity in the tri-
al’s study population. Of the 3420 participants enrolled,
just 17 (0.5%) were identified as Asian (n512 [0.4%]),
Hispanic/Latino (n54 [0.2%]), or Black/Afro-Caribbean
(n51 [0.0%]). The authors justify this imbalance by argu-
ing that the lifetime risk of breast cancer is slightly elevated
in White patients (13% lifetime risk) compared with Asian
(11%), Black (12%), or Hispanic/Latino (11%) patients. The
very modest differences in lifetime risk presented are not
consistent with a 99.4% White study population. Some
may argue that because the trial was conducted in Western
Europe, the lack of diversity is understandable. However,
although there are lower proportions of Asian, Hispanic/
Latino, and Black individuals in Western Europe compared
with the United States, the population size is large enough
to support greater enrollment than observed in this trial.
Recruiting and enrolling individuals from underrepre-
sented populations in these settings is not impossible, but it
does take planning and collaboration with clinical sites
and centers where these patients receive care, as well as
engagement of diverse study staff and investigators. It also
requires acknowledgment that recruiting a diverse study
population may take longer. If trials continue to enroll in a
first come, first served manner, we will continue to enroll
people for whom participation is easiest and will see more
trials like this one.
Would a trial conducted in an almost entirely Asian, Black,
Hispanic/Latino, or other population be considered poten-
tially practice-changing, or would there be widespread con-
cerns about generalizability? We argue that the same
standard should apply here — and for all trials that lack
diverse participation. Such trials are prone to develop regi-
mens or tools that are aimed at a single population instead
of those that serve all populations well. Examples of this
abound in the breast cancer literature, including evidence
that the 21-gene recurrence score (Oncotype DX ) is less
V R
accurate at predicting risk of recurrence in Black patients
than it is in White patients.6 U.S. Preventive Services Task
Force breast cancer screening guidelines are best aligned
with the age-specific incidence rates of White women com-
pared with all other racial/ethnic groups.7 As we strive to
improve minority enrollment to cancer clinical trials, at a
NEJM EVIDENCE
For personal use only. No other uses without permission. Copyright © 2022 Massachusetts Medical Society.
evidence.nejm.org
bare minimum, racial/ethnic demographic enrollment is
key information that should be reported in all clinical trials.
Although the final results of ABCSG-18 are compelling for
the population in which it was conducted,2 taken into con-
text with other studies such as D-CARE and TEAM-IIB, it
remains unclear whether adjuvant bone-modifying agents,
especially denosumab, should be widely used in early-stage
breast cancer. As adjuvant systemic anticancer therapies
improve, the modest benefits, if any, of bone-modifying
agents are questionable. An updated Early Breast Cancer
Trialists’ Collaborative Group meta-analysis may help to
guide clinicians with this important clinical question.
Disclosures
Author disclosures are available at evidence.nejm.org.
Author Affiliation
1
Massachusetts General Hospital, Boston
References
1. Eisen A, Somerfield MR, Accordino MK, et al. Use of adjuvant
bisphosphonates and other bone-modifying agents in breast cancer:
ASCO-OH (CCO) Guideline Update. J Clin Oncol 2022;40:787-800.
DOI: 10.1200/JCO.21.02647.
2. Gnant M, Frantal S, Pfeiler G, et al. Long-term outcomes of adjuvant
denosumab in breast cancer. NEJM Evid. 2022;1(12). DOI: 10.1056/
EVIDoa2200162.
3. Gnant M, Pfeiler G, Steger GG, et al. Adjuvant denosumab in post-
menopausal patients with hormone receptor-positive breast cancer
(ABCSG-18): disease-free survival results from a randomised,
double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2019;
20:339-351. DOI: 10.1016/S1470-2045(18)30862-3.
4. Coleman R, Finkelstein DM, Barrios C, et al. Adjuvant denosumab
in early breast cancer (D-CARE): an international, multicentre, rand-
omised, controlled, phase 3 trial. Lancet Oncol 2020;21:60-72. DOI:
10.1016/S1470-2045(19)30687-4.
5. Vliek SB, Noordhoek I, Meershoek-Klein Kranenbarg E, et al. Daily
oral ibandronate with adjuvant endocrine therapy in postmenopausal
women with estrogen receptor-positive breast cancer (BOOG 2006-
04): randomized phase III TEAM-IIB trial [published correction
appears in J Clin Oncol 2022;40:3352]. J Clin Oncol 2022;40:2934-
2945. DOI: 10.1200/JCO.21.00311.
6. Hoskins KF, Danciu OC, Ko NY, Calip GS. Association of
race/ethnicity and the 21-gene recurrence score with breast cancer-
specific mortality among US women. JAMA Oncol 2021;7:370-378.
DOI: 10.1001/jamaoncol.2020.7320.
7. Stapleton SM, Oseni TO, Bababekov YJ, Hung Y-C, Chang DC. Race/
ethnicity and age distribution of breast cancer diagnosis in the United
States. JAMA Surg 2018;153:594-595. DOI: 10.1001/jamasurg.2018.
0035.
2
Back to Table of Contents
evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 9

Published February 27, 2023

NEJM Evid 2023; 2 (3)
DOI: 10.1056/EVIDoa2200302

ORIGINAL ARTICLE

Long-Term Benefits from Early Antiretroviral

Therapy Initiation in HIV Infection
INSIGHT Strategic Timing of AntiRetroviral Treatment (START) Study Group*

Abstract
BACKGROUND For people with HIV and CD4+ counts >500 cells/mm3, early initiation of
antiretroviral therapy (ART) reduces serious AIDS and serious non-AIDS (SNA) risk com-
pared with deferral of treatment until CD4+ counts are <350 cells/mm3. Whether excess risk
of AIDS and SNA persists once ART is initiated for those who defer treatment is uncertain.
METHODS The Strategic Timing of AntiRetroviral Treatment (START) trial, as previously
reported, randomly assigned 4684 ART-naive HIV-positive adults with CD4+ counts
>500 cells/mm3 to immediate treatment initiation after random assignment (n=2325) or
deferred treatment (n=2359). In 2015, a 57% lower risk of the primary end point (AIDS, SNA,
or death) for the immediate group was reported, and the deferred group was offered ART. This
article reports the follow-up that continued to December 31, 2021. Cox proportional-hazards
The members of the writing
models were used to compare hazard ratios for the primary end point from randomization
committee and author affiliations
through December 31, 2015, versus January 1, 2016, through December 31, 2021. are listed at the end of the article.
RESULTS Through December 31, 2015, approximately 7 months after the cutoff date from
*A complete list of members of the
the previous report, the median CD4+ count was 648 and 460 cells/mm3 in the immediate
Strategic Timing of AntiRetroviral
and deferred groups, respectively, at treatment initiation. The percentage of follow-up time
Treatment (START) study group
spent taking ART was 95% and 36% for the immediate and deferred groups, respectively,
is provided in the Supplemen-
and the time-averaged CD4+ difference was 199 cells/mm3. After January 1, 2016, the per- tary Appendix, available at
centage of follow-up time on treatment was 97.2% and 94.1% for the immediate and deferred evidence.nejm.org.
groups, respectively, and the CD4+ count difference was 155 cells/mm3. After January 1, 2016,
a total of 89 immediate and 113 deferred group participants experienced a primary end Dr. Lundgren can be contacted
point (hazard ratio of 0.79 [95% confidence interval, 0.60 to 1.04] versus hazard ratio of 0.47 at jens.lundgren@regionh.dk or
[95% confidence interval, 0.34 to 0.65; P<0.001]) before 2016 (P=0.02 for hazard ratio at Rigshospitalet–University of
Copenhagen, CHIP, Section 2100,
difference).
Blegdamsvej 9, DK-2100 Copen-
CONCLUSIONS Among adults with CD4+ counts >500 cells/mm3, excess risk of AIDS and hagen, Denmark
SNA associated with delaying treatment initiation was diminished after ART initiation, but
persistent excess risk remained. (Funded by the National Institute of Allergy and Infectious Read Full Article
Diseases and others.) at evidence.nejm.org

Back to Table of Contents

10 Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

Published February 28, 2023

NEJM Evid 2023; 2 (3)
DOI: 10.1056/EVIDe2300022

EDITORIAL

Scientific Inference Does Not Rest on

Statistical Testing
Thomas A. Trikalinos, M.D., Ph.D.1

T
he Strategic Timing of AntiRetroviral Treatment (START) trial found that in
patients with HIV and CD41 cell counts above 500 cells/mm3 who had not pre-
viously received treatment, immediate initiation of antiretroviral therapy reduced
the risk of serious adverse outcomes compared with delaying treatment initiation only
when the CD41 cell count fell below 350 cells/mm3.1 After the trial’s completion, people
with HIV not receiving therapy were offered the opportunity to start it without regard to
their CD41 cell count. In this issue of NEJM Evidence, the START group reports data from
an additional 5 years of follow-up.2 They found that over this period, the hazard ratio of
severe events in those initially randomly assigned to immediate versus deferred initiation
was 0.79 (95% confidence interval [CI], 0.60 to 1.04). From this estimate, and without ref-
erence to statistical testing, Lundgren et al. conclude that excess risk persists among those
who deferred initial treatment. Any raised eyebrows might arch even higher when I point
out that their protocol prescribed statistical testing: “If the upper limit of the confidence
interval for the hazard ratio is below 1, it provides evidence” that excess risk persists. “If the
CI contains 1, the weight of the evidence” that no excess risk remains “is determined by the
width (precision) of the CI.” (See protocol of Lundgren et al.,2 Section 3.2.1, Primary analy-
ses: Testing the HIV RNA and nadir CD41 hypotheses.2)

Is the scientific conclusion by Lundgren et al.2 of persistent excess risk warranted? I believe
so and suggest not losing sleep over the protocol departure. The protocol need not have
prescribed testing where estimation suffices.3

We often do not distinguish scientific hypotheses from their mathematical operationaliza-

tions — the statistical hypotheses. Perhaps from some sort of reflex or inertia, we play at
applying to all research tasks the stylized description of the scientific process that was pre-
sented to us in physics and physiology laboratories. We teach our mentees to “develop
hypotheses,”4 genuine or perfunctory, for each specific aim in their funding proposals. Like
Sartre’s cafe waiter who plays at being a cafe waiter, we may come across as a tad too pre-
scribed and procedural and a little too “scientist”-like in our research planning.

The author affiliation is listed at

In medicine and public health, scientific inference and statistical inference are not the the end of the article.
same.5 Scientific inference is the broader concept and involves drawing conclusions from
the accumulation of consistent theoretical and empirical knowledge, the cogency of which Dr. Trikalinos can be contacted at
thomas_trikalinos@brown.edu or
rests on the tentative approval of subject matter experts.6 Key for scientific inference is at 121 S. Main St., Providence, RI
repeatability — not in the narrow sense of carbon copy observations across different 02143.

Back to Table of Contents

For personal use only. No other uses without permission. Copyright © 2023 Massachusetts Medical Society.
evidence.nejm.org
settings but as consistent and repeatable regularities that
point to the same “best explanation” and support the
same theory.7 This “inference to the best explanation” is
not an ironclad implication. The hazard ratio of 0.79 does
not logically imply that excess risk persists; it can be a
chance finding or the result of residual confounding. How-
ever, as long as the totality of the evidence does not
request its revision, the induction of “persistent excess
risk” is warranted as the simplest and most plausible
explanation.8
Statistical testing is a narrower concept. It requires a math-
ematical operationalization of the scientific question of
whether excess risk persists and further distills analyses as
a binary answer about “statistical significance,” which is use-
ful only when some explicit decision is contemplated.3,9
The correspondence of scientific questions to mathematical
statements may be less of an issue in disciplines such as
high-energy physics, for example, where scientific theories
are explicit mathematical statements and can be more
straightforwardly translated to statistical tests in experimen-
tal data. However, biomedical theories are not spelled out as
mathematical statements, and a multitude of nonequivalent
mathematical operationalizations is possible. For example,
the protocol of Lundgren et al.2 operationalized “persistent
excess risk” as a hazard ratio less than 1 after completion
of the phase of START reported in the New England Journal
of Medicine in 2015.1 Different operationalizations might
involve a hazard ratio smaller than some biologically or
clinically relevant threshold, a different period (e.g., from
randomization), or another analysis (e.g., using some mea-
sure of the total exposure to low CD41 cell count levels).
Statistical testing, where we judge whether a CI includes a
threshold value or a P value meets a cutoff, is useful when
the evidence at hand has a direct bearing on decision-
making.9 When actions are decided, thresholds should be
explicitly defined with reference to study goals and con-
sidering the consequences of incorrect decisions.3 No
imminent decision about scientific theories or patient
management looms in this case. It seems that the mem-
bers of the field, who understand HIV infection patho-
physiology and the relevant literature, had already
inferred that excess risk remains when deferring therapy
NEJM EVIDENCE
For personal use only. No other uses without permission. Copyright © 2023 Massachusetts Medical Society.
Establishing Evidence: A Collection of Confirmatory Trials 11
based on prior and less robust studies and before the
results of Lundgren et al.2 The current Joint United
Nations Programme on HIV/AIDS targets are that by
2025, 95% of people with HIV will know their status, 95%
of those diagnosed will be receiving treatment irrespective
of their CD41 cell counts, and 95% of those treated will
have viral suppression.10 This makes simple good sense.
Disclosures
Author disclosures are available at evidence.nejm.org.
Author Affiliation
1
Departments of Health Services, Policy & Practice and of Biostatistics
and Center for Evidence Synthesis in Health, Brown University School
of Public Health, Providence, RI
References
1. Lundgren JD, Babiker AG, Gordin F, et al. Initiation of antiretroviral
therapy in early asymptomatic HIV infection. N Engl J Med 2015;373:
795-807. DOI: 10.1056/NEJMoa1506816.
2. Lundgren JD, Babiker AG, Sharma S, et al. Long-term benefits from
early antiretroviral therapy initiation in HIV infection. NEJM Evid.
2023;2(3). DOI: 10.1056/EVIDoa2200302.
3. Benjamini Y, De Veaux RD, Efron B, et al. ASA President’s Task Force
statement on statistical significance and replicability. Chance 2021;34:
10-11. DOI: 10.1080/09332480.2021.2003631.
4. Monte AA, Libby AM. Introduction to the specific aims page of a grant
proposal. Acad Emerg Med 2018;25:1042-1047. DOI: 10.1111/acem.
13419.
5. Deming WE. On probability as a basis for action. Am Stat 1975;29:
146-152.
6. Hubbard R, Haig BD, Parsa RA. The limited role of formal statistical
inference in scientific inference. Am Stat 2019;73(Suppl 1):91-98. DOI:
10.1080/00031305.2018.1464947.
7. Douven I. Abduction. The Stanford Encyclopedia of Philosophy. The
Metaphysics Research Lab, Stanford University. Revised May 18, 2021
(https://plato.stanford.edu/archives/sum2021/entries/abduction/).
8. Harman GH. The inference to the best explanation. Philos Rev 1965;
74:88-95. DOI: 10.2307/2183532.
9. Jaynes ET. Probability theory: the logic of science. New York: Cam-
bridge University Press, 2003. DOI: 10.1017/CBO9780511790423.
10. Frescura L, Godfrey-Faussett P, Ali Feizzadeh A, El-Sadr W, Syarif
O, Ghys PD. Achieving the 95 95 95 targets for all: a pathway to
ending AIDS. PLoS One 2022;17:e0272405. DOI: 10.1371/journal.
pone.0272405.
2
Back to Table of Contents
12
new england
The
Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

journal of medicine
established in 1812 August 27, 2015 vol. 373 no. 9

ORIGINAL ARTICLE

Initiation of Antiretroviral Therapy in Early Asymptomatic

HIV Infection
The INSIGHT START Study Group*

a bs t r ac t

BACKGROUND
Data from randomized trials are lacking on the benefits and risks of initiating The members of the writing group (Jens
antiretroviral therapy in patients with asymptomatic human immunodeficiency D. Lundgren, M.D. [cochair], Abdel G.
Babiker, Ph.D. [cochair], Fred Gordin,
virus (HIV) infection who have a CD4+ count of more than 350 cells per cubic M.D. [cochair], Sean Emery, Ph.D., Birgit
millimeter. Grund, Ph.D., Shweta Sharma, M.S., An-
chalee Avihingsanon, M.D., David A.
METHODS Cooper, M.D., Gerd Fätkenheuer, M.D.,
We randomly assigned HIV-positive adults who had a CD4+ count of more than Josep M. Llibre, M.D., Jean-Michel Moli-
na, M.D., Paula Munderi, M.D., Mauro
500 cells per cubic millimeter to start antiretroviral therapy immediately (imme- Schechter, M.D., Robin Wood, M.D.,
diate-initiation group) or to defer it until the CD4+ count decreased to 350 cells Karin L. Klingman, M.D., Simon Collins,
per cubic millimeter or until the development of the acquired immunodeficiency H. Clifford Lane, M.D., Andrew N. Phil-
lips, Ph.D., and James D. Neaton, Ph.D.
syndrome (AIDS) or another condition that dictated the use of antiretroviral [INSIGHT PI]) of the INSIGHT START
therapy (deferred-initiation group). The primary composite end point was any seri- Study Group assume responsibility for
ous AIDS-related event, serious non–AIDS-related event, or death from any cause. the overall content and integrity of this
article. The affiliations of the members
RESULTS of the writing group are listed in the Ap-
pendix. Address reprint requests to Dr.
A total of 4685 patients were followed for a mean of 3.0 years. At study entry, the Lundgren at the Department of Infec-
median HIV viral load was 12,759 copies per milliliter, and the median CD4+ tious Diseases, Rigshospitalet, Univer-
count was 651 cells per cubic millimeter. On May 15, 2015, on the basis of an sity of Copenhagen, Blegdamsvej 9,
2100 Copenhagen Ø, Denmark, or at
interim analysis, the data and safety monitoring board determined that the study jens.lundgren@regionh.dk.
question had been answered and recommended that patients in the deferred-initi-
* A complete list of members in the Stra-
ation group be offered antiretroviral therapy. The primary end point occurred in tegic Timing of Antiretroviral Treatment
42 patients in the immediate-initiation group (1.8%; 0.60 events per 100 person- (START) Study Group is provided in the
years), as compared with 96 patients in the deferred-initiation group (4.1%; 1.38 Supplementary Appendix, available at
NEJM.org.
events per 100 person-years), for a hazard ratio of 0.43 (95% confidence interval
[CI], 0.30 to 0.62; P<0.001). Hazard ratios for serious AIDS-related and serious This article was published on July 20, 2015,
at NEJM.org.
non–AIDS-related events were 0.28 (95% CI, 0.15 to 0.50; P<0.001) and 0.61 (95%
CI, 0.38 to 0.97; P = 0.04), respectively. More than two thirds of the primary end N Engl J Med 2015;373:795-807.
DOI: 10.1056/NEJMoa1506816
points (68%) occurred in patients with a CD4+ count of more than 500 cells per Copyright © 2015 Massachusetts Medical Society.
cubic millimeter. The risks of a grade 4 event were similar in the two groups, as
were the risks of unscheduled hospital admissions. Read Full Article at NEJM.org
CONCLUSIONS
The initiation of antiretroviral therapy in HIV-positive adults with a CD4+ count
of more than 500 cells per cubic millimeter provided net benefits over starting
such therapy in patients after the CD4+ count had declined to 350 cells per cubic
millimeter. (Funded by the National Institute of Allergy and Infectious Diseases
and others; START ClinicalTrials.gov number, NCT00867048.)
n engl j med 373;9 nejm.org August 27, 2015 795

Back to Table of Contents

evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 13

Published September 10, 2022

NEJM Evid 2022; 1 (11)
DOI: 10.1056/EVIDoa2200214

ORIGINAL ARTICLE

Five-Year Analysis of Adjuvant Pembrolizumab

or Placebo in Stage III Melanoma
Alexander M.M. Eggermont, M.D., Ph.D.,1,2 Michal Kicinski, Ph.D.,3 Christian U. Blank, M.D., Ph.D.,4 Mario Mandala, M.D.,5
Georgina V. Long, M.D., Ph.D.,6 Victoria Atkinson, M.D.,7 Stéphane Dalle, M.D.,8 Andrew Haydon, M.D.,9
Andrey Meshcheryakov, M.D.,10 Adnan Khattak, M.D.,11 Matteo S. Carlino, M.D., Ph.D.,12 Shahneen Sandhu, M.D.,13
James Larkin, M.D., Ph.D.,14 Susana Puig, M.D., Ph.D.,15 Paolo A. Ascierto, M.D.,16 Piotr Rutkowski, M.D., Ph.D.,17
Dirk Schadendorf, M.D., Ph.D.,18,19 Marye Boers-Sonderen, M.D., Ph.D.,20 Anna Maria Di Giacomo, M.D., Ph.D.,21
Alfonsus J.M. van den Eertwegh, M.D., Ph.D.,22 Jean-Jacques Grob, M.D.,23 Ralf Gutzmer, M.D.,24 Rahima Jamal, M.D.,25
Alexander C.J. van Akkooi, M.D., Ph.D.,26 Paul Lorigan, M.D.,27 Dmitri Grebennik, M.D.,28 Clemens Krepler, M.D.,28
Sandrine Marreaud, M.D.,3 Stefan Suciu, Ph.D.,3 and Caroline Robert, M.D., Ph.D.29

Abstract
BACKGROUND In the previously reported primary analyses of this phase 3 trial,
12 months of adjuvant pembrolizumab resulted in significantly longer recurrence- and
distant metastasis-free survival than placebo in patients with resected high-risk stage III
melanoma. To confirm the stability of these benefits, longer-term data were needed.
METHODS We randomly assigned 1019 patients to receive 200 mg of pembrolizumab or pla-
cebo intravenously every 3 weeks for a total of 18 doses (approximately 1 year) and had previ-
ously reported data with a 15-, 36-, and 42-month median follow-up. We now report data at a
median follow-up of 4.9 years. We report a number of outcomes, including recurrence-free
survival in the overall population and in the subgroup of patients with cancer who were
positive for the programmed death-ligand 1 (PD-L1). Distant metastasis-free survival was Drs. Suciu and Robert contributed
a secondary end point. equally to this article.
RESULTS In the overall intention-to-treat population, pembrolizumab was still associated
The author affiliations are listed
with longer recurrence-free survival than placebo (5-year rate of recurrence-free survival,
at the end of the article.
55.4% [95% confidence interval (CI), 50.8 to 59.8] vs. 38.3% [95% CI, 33.9 to 42.7]; hazard
ratio for recurrence or death, 0.61 [95% CI, 0.51 to 0.72]) and a longer distant metastasis- Dr. Eggermont can be con-
free survival (5-year rate of distant metastasis-free survival, 60.6% [95% CI, 56.0 to 64.9] tacted at alexander.eggermont@
vs. 44.5% [95% CI, 39.9 to 48.9]; hazard ratio for distant metastasis or death, 0.62 [95% CI, prinsesmaximacentrum.nl or at
0.52 to 0.75]). Similar findings were obtained in the subgroup of 853 patients with Princess Máxima Center and Uni-
PD-L1–positive tumors. versity Medical Center Utrecht,
Heidelberglaan 25, Utrecht 3584
CONCLUSIONS The 5-year analysis of adjuvant therapy with pembrolizumab resulted in a CS, the Netherlands.
sustained improvement in the long-term recurrence- and distant metastasis-free survival
compared with placebo in patients with resected stage III melanoma. (Funded by Merck & Read Full Article
Co., Inc.; ClinicalTrials.gov number, NCT02362594, and EudraCT number, 2014-004944-37.) at evidence.nejm.org

Back to Table of Contents

14 Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

Published October 25, 2022

NEJM Evid 2022; 1 (11)
DOI: 10.1056/EVIDe2200240

EDITORIAL

Progression/Recurrence-Free Survival 2 in
Adjuvant Melanoma
Ryan C. Augustin, M.D.,1 and Jason J. Luke, M.D.1

T
he long-term outcome of patients with stage III melanoma — that is, melanoma that
has spread to nearby lymph nodes, lymphatics, or skin — who have received treat-
ment with immune checkpoint inhibitors is of substantial interest. The article by
Eggermont et al.1 published in this issue of NEJM Evidence reports 5-year outcomes from the
stage III melanoma trial, KEYNOTE-054, which compared pembrolizumab (anti–programmed
cell death protein 1 [PD-1]) with placebo. The data show durable recurrence-free survival
(RFS) and distant metastasis–free survival (DMFS). They do not report 5-year overall survival
(OS), citing insufficient statistical power, but present a potential surrogate for OS, progression/
recurrence-free survival 2 (PRFS2). PRFS2 is defined as the time from initial random assign-
ment to the second objective disease progression or recurrence. Whether PRFS2 is an appro-
priate end point will be a point of debate, although a growing literature seems to support it.

Immuno- and BRAF-targeted therapies have revolutionized the management of metastatic

melanoma, leading to their investigation in the setting of high-risk patients after surgical
resection.2 Both nivolumab (anti–PD-1) and pembrolizumab have demonstrated long-term
improvements in RFS versus ipilimumab (anti–cytotoxic T-lymphocyte–associated protein 4)
in patients with stage III disease.3,4 In addition, an improved RFS was observed in patients
with resectable, BRAF-mutant melanoma who were treated with combined dabrafenib plus
trametinib (BRAF plus MEK small-molecule inhibitors).5 In patients with IIB/C melanoma,
pembrolizumab has also demonstrated an improved 1.5-year RFS compared with placebo.6

Among the trials noted above, adjuvant treatment provides a 10 to 15% absolute improvement
in RFS and DMFS, although reports with respect to any OS benefit are inconsistent. A trial that
compared ipilimumab with placebo in high-risk patients with stage III melanoma did demon-
strate an OS benefit,7 as did the adjuvant study of dabrafenib plus trametinib versus placebo,
at least with early follow-up.5 However, the enrollment periods for both of these trials were
before the widespread availability of anti–PD-1 therapy for patients who experienced progres-
sion to metastatic disease. This calls into question whether the introduction of anti–PD-1
therapy minimized the OS benefit observed in those studies. Both of the aforementioned trials
comparing nivolumab or pembrolizumab with ipilimumab in the adjuvant stage III setting The author affiliation is listed at
the end of the article.
have demonstrated minimal difference in OS. As noted above, 5-year OS was not calculated in
KEYNOTE-054 given the lack of deaths required to facilitate the final analysis.1 Dr. Luke can be contacted at
lukejj@upmc.edu or at UPMC
Hillman Cancer Center, 5150
Whereas OS generally serves as an important and often primary end point in oncology, robust Centre Ave., Room 1.27C,
evaluation of OS has become seemingly out of reach in melanoma because of increasing lines Pittsburgh, PA 15232.

Back to Table of Contents

For personal use only. No other uses without permission. Copyright © 2022 Massachusetts Medical Society.
evidence.nejm.org
and types of therapies. Improvement in RFS has been the
long-standing regulatory end point for adjuvant clinical
trials in melanoma, and oncology more broadly, despite
meta-analyses showing that RFS correlates poorly with OS.
An end point that could bridge the gap between RFS and OS
may be PRFS2. In contrast to OS, PRFS2 has the potential to
reduce the confounding effect of successive lines of therapy.
In addition, PRFS2 has been shown to have a stronger corre-
lation with OS (r50.67) compared with RFS (r50.21) across
multiple tumor types.8 Using this metric, the KEYNOTE-054
investigators described the 5-year PRFS2 as significantly
improved with pembrolizumab versus placebo (68.2% vs.
55.5%, hazard ratio50.65, 95% confidence interval50.53
to 0.80). Given the reduced time between the second
recurrence and death compared with the first recurrence,
along with inherent challenges in calculating OS itself, the
KEYNOTE-054 authors promote PRFS2 as a key end point
that augments RFS. An interesting analysis would be to
analyze PRFS2 from previous adjuvant melanoma trials
where an OS benefit was observed as a validation technique.
Whereas patients obviously prioritize maximization of OS,
survey data consistently demonstrate that patients value a
reduction of relapse nearly as much.9 This is true even in
situations where RFS does not extend OS10 and suggests
that the risk of recurrence is an ongoing physiological
harm that deserves mitigation. With increasing postprog-
ression survival, long-term disease stability as measured
through PRFS2 may better align with patient preferences,
more strongly correlate with OS, and have the potential
for consistent measurement in future trials.
Overall, KEYNOTE-054 provides long-term data to support
the foundational role of anti–PD-1 therapy in high-risk,
resected melanoma. Given the impending role for both
neoadjuvant and combination immunotherapy strategies
in resectable populations, the most applicable result of the
current report may be its demonstration of PRFS2 as an inno-
vative and clinically meaningful end point for future trials.
Disclosures
Author disclosures are available at evidence.nejm.org.
NEJM EVIDENCE
For personal use only. No other uses without permission. Copyright © 2022 Massachusetts Medical Society.
Establishing Evidence: A Collection of Confirmatory Trials 15
Author Affiliation
1
UPMC Hillman Cancer Center and University of Pittsburgh,
Pittsburgh
References
1. Eggermont AMM, Kicinski M, Blank C, et al. Five-year analysis of
adjuvant pembrolizumab or placebo in stage III melanoma. NEJM
Evid 2022;1(11). DOI: 10.1056/EVIDoa2200214.
2. Eggermont AMM, Hamid O, Long GV, Luke JJ. Optimal systemic
therapy for high-risk resectable melanoma. Nat Rev Clin Oncol
2022;19:431-439. DOI: 10.1038/s41571-022-00630-4.
3. Ascierto PA, Del Vecchio M, Mandal
a M, et al. Adjuvant nivolumab
versus ipilimumab in resected stage IIIB–C and stage IV melanoma
(CheckMate 238): 4-year results from a multicentre, double-blind,
randomised, controlled, phase 3 trial. Lancet Oncol 2020;21:1465-1477.
DOI: 10.1016/S1470-2045(20)30494-0.
4. Grossmann KF, Othus M, Patel SP, et al. Final analysis of overall sur-
vival (OS) and relapse-free-survival (RFS) in the intergroup S1404
phase III randomized trial comparing either high-dose interferon (HDI)
or ipilimumab to pembrolizumab in patients with high-risk resected
melanoma. J Clinical Oncol 2021:39;9501. (https://ascopubs.org/doi/
abs/10.1200/JCO.2021.39.15_suppl.9501).
5. Dummer R, Brase JC, Garrett J, et al. Adjuvant dabrafenib plus
trametinib versus placebo in patients with resected, BRAFV600-
mutant, stage III melanoma (COMBI-AD): exploratory biomarker
analyses from a randomised, phase 3 trial. Lancet Oncol 2020;21:
358-372. DOI: 10.1016/S1470-2045(20)30062-0.
6. Luke JJ, Rutkowski P, Queirolo P, et al. Pembrolizumab versus pla-
cebo as adjuvant therapy in completely resected stage IIB or IIC mel-
anoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial.
Lancet 2022;399:1718-1729. DOI: 10.1016/S0140-6736(22)00562-1.
7. Eggermont AMM, Chiarion-Sileni V, Grob JJ, et al. Adjuvant ipilimu-
mab versus placebo after complete resection of stage III melanoma:
long-term follow-up results of the European Organisation for Research
and Treatment of Cancer 18071 double-blind phase 3 randomised
trial. Eur J Cancer 2019;119:1-10. DOI: 10.1016/j.ejca.2019.07.001.
8. Woodford RG, Zhou DDX, Kok PS, et al. The validity of progression-
free survival 2 as a surrogate trial end point for overall survival. Cancer
2022;128:1449-1457. DOI: 10.1002/cncr.34085.
9. Livingstone A, Agarwal A, Stockler MR, Menzies AM, Howard K, Morton
RL. Preferences for immunotherapy in melanoma: a systematic review.
Ann Surg Oncol 2020;27:571-584. DOI: 10.1245/s10434-019-07963-y.
10. Seghers PALN, Wiersma A, Festen S, et al. Patient preferences for treat-
ment outcomes in oncology with a focus on the older patient-a systematic
review. Cancers (Basel) 2022;14:1147. DOI: 10.3390/cancers14051147.
2
Back to Table of Contents
16 Establishing Evidence: A Collection of Confirmatory Trials evidence.nejm.org

The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article
ORIGINAL ARTICLE

Adjuvant Pembrolizumab versus Placebo

in Resected Stage III Melanoma
Alexander M.M. Eggermont, M.D., Ph.D., Christian U. Blank, M.D., Ph.D.,
Mario Mandala, M.D., Georgina V. Long, M.D., Ph.D., Victoria Atkinson, M.D.,
Stéphane Dalle, M.D., Andrew Haydon, M.D., Mikhail Lichinitser, M.D.,
Adnan Khattak, M.D., Matteo S. Carlino, M.D., Ph.D., Shahneen Sandhu, M.D.,
James Larkin, M.D., Susana Puig, M.D., Ph.D., Paolo A. Ascierto, M.D.,
Piotr Rutkowski, M.D., Dirk Schadendorf, M.D., Ph.D., Rutger Koornstra, M.D.,
Leonel Hernandez-Aya, M.D., Michele Maio, M.D., Ph.D.,
Alfonsus J.M. van den Eertwegh, M.D., Ph.D., Jean-Jacques Grob, M.D., Ph.D.,
Ralf Gutzmer, M.D., Rahima Jamal, M.D., Paul Lorigan, M.D., Nageatte Ibrahim, M.D.,
Sandrine Marreaud, M.D., Alexander C.J. van Akkooi, M.D., Ph.D., Stefan Suciu, Ph.D.,
and Caroline Robert, M.D., Ph.D.

A BS T R AC T

BACKGROUND
The programmed death 1 (PD-1) inhibitor pembrolizumab has been found to prolong The authors’ affiliations are listed in the
progression-free and overall survival among patients with advanced melanoma. We con- Appendix. Address reprint requests to Dr.
Eggermont at the Gustave Roussy Can-
ducted a phase 3 double-blind trial to evaluate pembrolizumab as adjuvant therapy in cer Campus Grand Paris and University
patients with resected, high-risk stage III melanoma. Paris-Saclay, 94805 Villejuif, France, or at
alexander.eggermont@gustaveroussy.fr.
METHODS
Drs. Suciu and Robert contributed equal-
Patients with completely resected stage III melanoma were randomly assigned (with ly to this article.
stratification according to cancer stage and geographic region) to receive 200 mg of
This article was published on April 15, 2018,
pembrolizumab (514 patients) or placebo (505 patients) intravenously every 3 weeks for and updated on April 24, 2018, at NEJM.org.
a total of 18 doses (approximately 1 year) or until disease recurrence or unacceptable
N Engl J Med 2018;378:1789-801.
toxic effects occurred. Recurrence-free survival in the overall intention-to-treat popula- DOI: 10.1056/NEJMoa1802357
tion and in the subgroup of patients with cancer that was positive for the PD-1 ligand Copyright © 2018 Massachusetts Medical Society.

(PD-L1) were the primary end points. Safety was also evaluated.
Read Full Article at NEJM.org
RESULTS
At a median follow-up of 15 months, pembrolizumab was associated with significantly
longer recurrence-free survival than placebo in the overall intention-to-treat population
(1-year rate of recurrence-free survival, 75.4% [95% confidence interval {CI}, 71.3 to 78.9] vs.
61.0% [95% CI, 56.5 to 65.1]; hazard ratio for recurrence or death, 0.57; 98.4% CI, 0.43 to
0.74; P<0.001) and in the subgroup of 853 patients with PD-L1–positive tumors (1-year rate
of recurrence-free survival, 77.1% [95% CI, 72.7 to 80.9] in the pembrolizumab group and
62.6% [95% CI, 57.7 to 67.0] in the placebo group; hazard ratio, 0.54; 95% CI, 0.42 to 0.69;
P<0.001). Adverse events of grades 3 to 5 that were related to the trial regimen were reported
in 14.7% of the patients in the pembrolizumab group and in 3.4% of patients in the placebo
group. There was one treatment-related death due to myositis in the pembrolizumab group.
CONCLUSIONS
As adjuvant therapy for high-risk stage III melanoma, 200 mg of pembrolizumab ad-
ministered every 3 weeks for up to 1 year resulted in significantly longer recurrence-
free survival than placebo, with no new toxic effects identified. (Funded by Merck;
ClinicalTrials.gov number, NCT02362594; EudraCT number, 2014-004944-37.)

n engl j med 378;19 nejm.org May 10, 2018 1789

Back to Table of Contents
evidence.nejm.org Establishing Evidence: A Collection of Confirmatory Trials 17

Published May 15, 2023

NEJM Evid 2023; 2 (6)
DOI: 10.1056/EVIDoa2200335

ORIGINAL ARTICLE

Ropeginterferon versus Standard Therapy

for Low-Risk Patients with Polycythemia Vera
Tiziano Barbui, M.D.,1 Alessandro Maria Vannucchi, M.D.,2 Valerio De Stefano, M.D.,3 Alessandra Carobbio, M.Sc.,1
Arianna Ghirardi, Ph.D.,1 Greta Carioli, Ph.D.,1 Arianna Masciulli, Ph.D.,1 Elena Rossi, M.D.,3 Fabio Ciceri, M.D.,4
Massimiliano Bonifacio, M.D.,5 Alessandra Iurlo, M.D.,6 Francesca Palandri, M.D.,7 Giulia Benevolo, M.D.,8
Fabrizio Pane, M.D.,9,10 Alessandra Ricco, M.D.,11 Giuseppe Carli, M.D.,12 Marianna Caramella, M.D.,13 Davide Rapezzi, M.D.,14
Caterina Musolino, M.D.,15 Sergio Siragusa, M.D.,16 Elisa Rumi, M.D.,17,18 Andrea Patriarca, M.D.,19 Nicola Cascavilla,
M.D.,20 Barbara Mora, M.D.,21 Emma Cacciola, M.D.,22 Carmela Mannarelli, M.Sc.,2 Giuseppe Gaetano Loscocco, M.D.,2
Paola Guglielmelli, M.D.,2 Francesca Gesullo, M.Sc.,2 Silvia Betti, M.D.,3 Francesca Lunghi, M.D.,4 Luigi Scaffidi, M.D.,5
Cristina Bucelli, M.D.,6 Nicola Vianelli, M.D.,7 Marta Bellini, M.D.,23 Maria Chiara Finazzi, M.D.,23,24 Gianni Tognoni, M.D.,25
and Alessandro Rambaldi, M.D.23,24

Abstract
BACKGROUND Whether phlebotomy alone can adequately maintain target hematocrit in
patients with low-risk polycythemia vera (PV) remains elusive.
METHODS In a phase 2 open-label randomized trial, we compared ropeginterferon alfa-2b
(ropeg; 100 μg every 2 weeks) with phlebotomy only regarding maintenance of a median
hematocrit level (≤45%) over 12 months in the absence of progressive disease (primary end
point). In follow-up, crossover to the alternative treatment group was allowed if the primary
end point was not met.
RESULTS In total, 127 patients were enrolled (ropeg: n=64; standard group: n=63). The
primary end point was met in 81% and 51% in the ropeg and standard groups, respectively.
The author affiliations are listed at
Responders continued the assigned treatment until month 24 and maintained response
the end of the article.
in 83% and 59%, respectively (P=0.02). Ropeg responders less frequently experienced
moderate/severe symptoms (33% vs. 67% in the standard group) and palpable splenomegaly Dr. Tiziano Barbui can be contacted
(14% vs. 37%) and showed normalization of ferritin levels and blood counts. Nonresponders at tbarbui@fondazionefrom.it
at 12 months crossed over to the standard (n=9) or ropeg (n=23) group; in patients switched or at Fondazione per la Ricerca
to ropeg only, 7 of 23 met the response criteria in 12 months, and phlebotomy need was Ospedale di Bergamo (FROM),
high (4.7 per patient per year). Discontinuation because of adverse events occurred in seven Papa Giovanni XXIII Hospital,
patients treated with ropeg. Piazza Organizzazione Mondiale
della Sanità 1, 24127 Bergamo,
CONCLUSIONS In this 24-month trial, ropeg was superior to phlebotomy alone in main- Italy.
taining hematocrit on target. No dose-limiting side effects or toxicities were noted; 9.2% of
patients on ropeg and no patients on standard treatment developed neutropenia. (Funded by Read Full Article
AOP Health and others; ClinicalTrials.gov number, NCT03003325.) at evidence.nejm.org

Back to Table of Contents

 evidence.nejm.org

evaluate the reasons.

understand the results.

Discover NEJM Evidence, a new approach to clinical trial

design and execution.
NEJM Evidence gives you an insider’s view on how clinical
trials are designed, how evidence is generated, and the
subsequent implications for diagnostic and treatment
decisions.

SUBSCRIBE NOW and receive unlimited subscriber access

to NEJM Evidence for one full year, 12 monthly digital issues
along with continual publication of insightful, innovative
articles, cases, podcasts, and videos, all accessible on
evidence.nejm.org.

“ ,
Clinical research is a really complex game. We re
going to help readers understand how clinical trials
are put together, run, and analyzed — and how they
impact practice.
” j e f f r e y m . d r a z e n , md
editor-in-chief

Subscribe today at STORE.NEJM.ORG/EVIDENCE