Professor
Brian Andrews NEJM Recommendations for Medical Students and Tutors
Week
of 7th July 2016 (#50)
University
of Notre Dame Australia (Fremantle Campus)
Occasional Editorial Comment
I was unsure as to where to include the five articles
which include the perspective, two on adaptive randomization (AR) in breast
cancer trials, the review on adaptive designs for clinical trials and the relevant
editorial. While these articles focus on the AR design in breast cancer drug
trials, the overall message is that many of the classical phase 2 clinical
trials will be modified in the future, using instead AR trials in phase 2
studies.
This is a timely, about time, and refreshing update in
clinical trials methodology. It is noteworthy that only three medical students
out of 24 found any of these articles or the concept worthy of reading.
I would suggest reading the perspective article thoroughly
as this clearly outlines the principles of this new approach, especially in
testing oncology drugs.
Perspective
STATISTICS IN MEDICINE
I-SPY
2 — A Glimpse of the Future of Phase 2 Drug Development?
The
I-SPY 2 breast cancer trials assess activity of new agents by adaptive
randomization, applying criteria for success on a surrogate end point to
predict success in confirmatory trials. Innovative designs may speed progress
and eliminate waste in drug development.
I have extracted various sections from the perspective
article as these are articulated extremely well:
1.
Traditionally
designed phase 2 trials that test treatments one at a time in heterogeneous
groups of patients have created a traffic jam: there are too many new drugs,
and the signal of a treatment effect can be diluted in these heterogeneous
groups.
2.
Adaptive
multigroup trials such as I-SPY 2 have the potential to answer several
questions simultaneously and more efficiently than traditionally designed
trials.
3.
The
I-SPY 2 platform will be used to compare up to 12 experimental therapies with a
common control in subgroups of breast cancer with 10 distinct biomarker
signatures.
4.
New
drugs can enter the platform (see Figure 1 in the Neratinib study) as they
emerge from phase 1 testing and exit the platform with an estimate of the
chances of future success in a phase 3 trial of prespecified size. This will
reduce cost significantly with a greater chance that successful drugs will
reach the patient faster.
5.
The
platform may be an appealing setting for cooperation among pharmaceutical
companies and academic investigators.
6.
The
entire process, including design and analysis, is carried out dynamically using
Bayesian methods (see Tables 1 and 2).
7.
In
the world of trial design, this new platform is still in its adolescence.
There are many questions that will need to be
answered which include:
1.
How
robust are the adaptive randomization probabilities and the predictive
probabilities of success in a phase 3 trial to misspecifications of the model?
2.
What
are effective ways to communicate to our clinical colleagues the modelling
assumptions used, the potential vulnerability of the model to errors, and the
best ways to explain these designs to trial participants?
3.
What
visual and numerical summaries provide insight into the trial data? Simple
summary statistics such as odds ratios or relative risk can be misleading.
4.
Will
the predicted chance of future success (85%) upset equipoise for trial
investigators or influence the kinds of patients investigators choose to enrol
or not enrol in a future trial?
ORIGINAL ARTICLE
Adaptive
Randomization of Neratinib in Early Breast Cancer
Among
patients with HER2-positive, hormone-receptor–negative locally advanced breast
cancer, the addition of neratinib to standard therapy resulted in higher rates
of pathological complete response, with some higher rates of toxic effects.
This is an important paper (especially Figure 1) which
reviews the concept of an AR design and the concept of the platform to which
test drugs can be added when they become available after successful phase 1
trials. The analysis of results is based on Bayesian probability which considers
multiple variables in each response. Neoadjuvant and adjuvant therapies are
also discussed.
All patients (high risk
clinical stage II or III breast cancer) receive the standard of care, namely
paclitaxel followed by doxorubicin + cyclophosphamide. The patient then undergoes
surgery and the results on the tumour are analysed. Following these phase 2 studies, various
protocols can be established for further phase 3 trials.
EDITORIAL
I-SPY
2 — Toward More Rapid Progress in Breast Cancer Treatment
The editorial reviews the two papers on the AR clinical
drug trials in breast cancer and discusses the interpretation and pros and cons
of the results.
There have been recent papers published in cardiology
in which long-term studies have been carried out on a new device. By the time the results are presented at
meetings and then published, the original device is no longer being used, being
supplanted by a newer device. What a waste of time and money! If a platform
type study could be utilized to add newer devices (presumed to be more effective,
but not proven) where both the new and the newer device could be studied in
parallel, then time, money and earlier use by patients will be accomplished. One wonders how device manufacturers would react
to this type of trial design.
Must Read Articles
None
Articles Recommended for Medical Students
Perspective
Zika
and the Risk of Microcephaly
An
analysis of data from Brazil reveals a strong association between the risk of
microcephaly in a newborn and the risk of Zika virus infection during the
mother's first trimester of pregnancy. The association in the second and third
trimesters was negligible.
In the preceding months the Journal has reviewed data
on the pathology of ZIKV on the nervous system, the epidemiology of ZIKV
infections and, most recently, has discussed a direct causal association
between the virus and clinical disease by fulfilling standard causality
criteria.
Today an article provides data from recent ZIKV outbreaks
in French Polynesia and Bahia, Brazil. The
greatest chance of developing microcephaly or significant neurological
impairment occurs in women who develop a symptomatic infection during the first
trimester and who deliver at term.
The article produces more questions than answers. However, it focuses on our lack of knowledge
and data on the outcomes of asymptomatic ZIKV infections during pregnancy and on
the other long-term, less severe neurological sequelae not encompassed by
microcephaly or marked neurological impairment at birth. The authors indicate
that based on the current data available more cases of significant birth
defects will occur and that ZIKV infection during pregnancy should be avoided.
I believe there is also a significant ethical issue
raised (the individual versus the society as a whole) with playing “Russian
roulette” in endemic areas for sexually active people who avoid responsible
birth control and protection. Education should
be provided by the state for women and men. If not, should society then be held
responsible for the long-term health care costs in caring for neurologically
impaired infants if appropriate care is not taken in prevention?
The knowledge of the ZIKV is still in its infancy, especially
in terms of other modes of viral transmission, the genes associated with the
expression of disease and other organ systems involved long-term in infants
born to infected mothers. Stay tuned and let’s see what the Olympics produces
apart from gold medals.
ORIGINAL ARTICLE
Ticagrelor
versus Aspirin in Acute Stroke or Transient Ischemic Attack
In
this double-blind, randomized trial involving 13,199 patients with ischemic stroke
or transient ischemic attack, ticagrelor was not superior to aspirin in
reducing the rate of stroke, myocardial infarction, or death at 90 days.
As predicted, a majority of MED300 and MED400 students
reviewed this article but frequently did not carefully review the exclusion
criteria. The authors note the lower frequency of primary outcome events
(stroke, myocardial infarction or death within 90 days - usually stroke within
the first two weeks) in this study with aspirin (7.5%) or ticagrelor (6.7%)
compared with historical data using aspirin for secondary stroke prevention. Previous
studies found a recurrent stroke rate significantly higher at between 10-15%.
In previous aspirin studies, patients with surgically amenable carotid artery
disease or major intracerebral vessel disease amenable to radiological
intervention were not excluded in the analyses although they were in this
study.
The data is from an international, randomized, double
blind controlled trial comparing either aspirin or ticagrelor in 13,199
patients with a high-risk TIA or non-severe ischaemic stroke (check exclusion
criteria) who were started on the drug within 24 hours of initial symptoms.
Patients were followed for 90 days with the primary end-point of stroke, MI, or
death. The data showed no significant difference in primary outcome with either
drug i.e. ticagrelor was not superior to aspirin. The hazard ratios for each
treatment should be reviewed to see subtle subgroup differences.
CLINICAL PRACTICE
Polycystic
Ovary Syndrome
The
polycystic ovary syndrome increases the risk of infertility, endometrial
cancer, abnormal glucose metabolism, and dyslipidemia. Strategies such as
lifestyle modification, hair removal, and combined oral contraceptive therapy
and other pharmacotherapies are reviewed.
This is an excellent review on the polycystic ovarian
syndrome (PCOS) with key points well summarized in the enclosed box. It should
be read by all MED300 before their O&G rotation and the hyperlink stored
for the future.
Figure 1 is excellent for demonstrating the basic
pathophysiology of the hyperandrogenemia seen in PCOS.
Specifically review the
presented case and the section at the end on conclusions and specific
recommendations for this patient. In the section of Diagnosis, there is a
detailed discussion of the differential diagnosis and other conditions which
need to be excluded before a definitive diagnosis of PCOS can be made.
For those who had forgotten most of what they once
knew, this is a valuable review of the basic science, ovarian cholesterol
metabolism, mechanisms of hyperandrogenism, and, in particular, a balanced
approach to the management of PCOS and the expected outcomes with the therapies
available.
Recommended
learning:
1. Review
pituitary, ovarian, and adrenal axis in terms of uterine function and the
menstrual cycle.
2. Review
the PCOS and other causes of increased androgen levels.
IMAGES IN CLINICAL MEDICINE
Eyelid
Melanoma
An
87-year-old man presented with a stye on his left upper eyelid. Despite conservative
interventions, the stye had persisted and the skin had become noticeably darker
over several weeks. His risk factors included a history of prolonged sun
exposure while working as a lifeguard.
This is an interesting clinical picture of an eyelid
melanoma in an 87 year old male. Considering the possible complexities of the therapy
the patient may have been offered, there is no wonder he decided not to return
for follow up.
Recommended
learning: Review the clinical presentations of melanoma,
prognostic factors, pathology, principles of surgical management and advances
in adjuvant and newer therapies.
IMAGES IN CLINICAL MEDICINE
Dubin–Johnson
Syndrome
A
48-year-old woman scheduled to receive a laparoscopic cholecystectomy underwent
a preoperative evaluation. A laparoscopic exploration, shown in a video,
revealed a smooth liver with normal consistency and morphology but with a
grossly black appearance.
This is a very interesting and well described case of
Dubin-Johnson (autosomal recessive) with excellent clinical pictures, including
video, and typical histopathology. The authors demonstrate loss of the MRP2
receptor on the hepatocyte which leads to blockage of the transport of
conjugated bilirubin from the canalicular pole of the hepatocyte into the bile
canaliculus.
The result of the laparoscopy and biopsy were
predictable based on the clinical history and normal LFTs except for the
elevated conjugated bilirubin. This was a well, 48 year old woman who had had
intermittent jaundice for at least 20 years associated with conjugated hyperbilirubinemia
(not unconjugated as seen in the much more common entity Gilberts’ disease)
with no other impairment in liver function.
Recommended
learning:
1. Review
the pathophysiology and causes of hyperbilirubinemia.
2. Review
Gilbert’s disease.
CLINICAL PROBLEM-SOLVING
A
Bruising Loss
A
32-year-old woman presented to her physician with a 3-week history of
spontaneous bruising on her arms, legs, and back. The bruising began shortly
after she had had sore throat, coryza, and malaise for several days, symptoms
that had resolved without intervention.
This is a very interesting discussion of the
differential diagnosis of spontaneous bruising in a 32 year old female which
should be read by most medical students during their clinical years. While the
cause of this acquired coagulopathy is uncommon (an autoantibody against factor
VIII), the principles of management are interesting. Further the authors
omitted a rare cause of a factor VIII autoantibody which is primary amyloidosis
involving the spleen.
It is interesting that only one of 24 students took the
time to review this article. I believe this was due to the inappropriate and
“trying too hard” title which turned the students off.
The first words in a paper that the reader sees is the title and thus it is imperative that it
capture the interest of the reader to proceed further into the article. The
process of title selection takes time and consideration with multiple options
considered and shared with co-authors.
It is generally considered that important aspects of a
title should include:
1. Sufficient
information about the subsequent work.
2. A
title no more than 12 words in length.
3. Considering
how much information should be in the title? Should it summarize the results of
the study? For example in the study on aspirin and ticagrelor for secondary
prevention in ischaemic stroke, should the title have been “Ticagrelor and
Aspirin are equally effective in preventing a second stoke in patients
presenting with an ischaemic stroke?” Apart from exceeding the 12 word limit,
does the title provide too much information with the reader then not inclined
to read further?
4. Try
breaking the title into two components with the second following the first, but
separated by a colon. For example, I wrote a review article many years ago:
“Scleroderma: Advances from Osler to the Present.”
5. Refraining
from hyperbole and “attempts to be too cute.” Don’t try too hard like “A
Bruising Loss.”
Important Articles Related to Mechanisms of Disease and
Translational Research
CLINICAL IMPLICATIONS OF BASIC RESEARCH
Corralling
Colonic Flagellated Microbiota
A
study of a mouse deficient in the gene Lypd8 showed that the Lypd8
protein prevents flagellated bacteria from invading the inner mucosal lining of
colonic epithelium.
This article discusses a role of the mechanical mucous
layer in the colon in protecting the host from potentially invasive gut
microbiota. There is an excellent Figure 1 which demonstrates the protective
role of a protein Lypd8, which is secreted by colonic epithelial cells. This
protein binds to the flagellae of motile bacteria at the junction of the loose
outer mucous layer with the firm inner mucous layer stopping bacteria entering
the inner mucous layer. A Lypd8
knock-out mouse model resulted in a mouse phenotype which was readily subjected
to colitis and patients with inflammatory bowel disease also exhibited reduction
in Lypd8 production.
Recommended
learning: Review the role of host protective mechanisms within
the GI tract, including mechanical barriers, and the role of the innate and adaptive
immune systems.
Other Articles which should interest medical students
SPECIAL ARTICLE
State
Legal Restrictions and Prescription-Opioid Use among Disabled Adults
In this analysis of Medicare data and a data set of
state laws, adoption of legislation to restrict the prescribing and dispensing
of opioid medications was not associated with reductions in potentially hazardous
use of opioids among disabled Medicare beneficiaries.
This is a timely article on the effect(s) of laws
introduced in the US to limit the epidemic of prescription narcotic abuse and to
reduce opioid overdose.
In this study, the laws introduced were state specific
and varied significantly between states participating from 2006 through 2012.
The population studied were disabled Medicare beneficiaries aged from 21 to 64 years.
Eight laws that were reviewed in this study and are illustrated best in Figure
1, together with the percentage of beneficiaries subjected to the laws.
The results to
date are disappointing from the legislative point of view. The researchers showed
no significant changes in opioid outcomes, such as a decline in the number of
prescriptions prescribed for morphine equivalent daily dose of 120 mg, or drug
overdosage.
The authors discuss the limitations of the study, and
from their findings it becomes apparent that any future similar study should
take into consideration the following:
1. A
large number of participants must be included and followed for prolonged
periods, possibly up to 10 years (this study was for seven years).
2. A
robust statistical analysis must be carried out.
3. A nationwide
study is more meaningful than individual state studies.
4. The
law(s) must remain stable over the duration of the study without any law changes.
5. Limiting
the amount of opioid in each prescription only appears to lead to more frequent
visits to the GP or pain specialist, which significantly increases the total cost
to the health care system, and adds more work to overworked GPs in particular
without changing the net prescribing habits.
6. Any
laws introduced should be based on evidence and should not be “feel good” or
“something needs to be done” laws.
This is an extremely challenging area where changes are
probably not likely to be seen in the short-term future.
The only hope I see for future change is early education of medical students and
doctors in training. Training should include the ability to make the appropriate
diagnosis. This skill should include an understanding of the natural course of
the pain syndrome, and learning appropriate scientific, pharmacological, and
clinical pain management for both acute and particularly early chronic pain
syndromes before inappropriate prescribing habits become established. The
future lies in medical student education in alleviating this problem.
