http://stroke.ahajournals.org/content/48/9/2341?etoc=
- Editorials
- animal models
- clinical translation
- end point selection
- predictive value
- study design
- translational stroke research
See related article, p 2632
Preclinical
stroke research faces a substantial transition. Past classical rodent
stroke models and study designs revealed numerous potential targets for
novel stroke therapies; yet, subsequent clinical stroke trials failed to
confirm promising preclinical findings. Pharmacological and mechanical
recanalization therapies, representing the only strategies that have
substantially improved acute ischemic stroke outcomes, were largely
developed omitting conventional preclinical methods.1–3 In this issue of Stroke,
a National Institute of Neurological Disorders and Stroke consensus
group comprising leading academic, industry, and Food and Drug
Administration (FDA) experts working at the forefront of stroke research
has recently published guidelines for improved translational studies.4
Distinguishing
between explorative (basic) versus confirmative (translational) studies
has been suggested. Exploratory research uses simpler rodent models but
a broad spectrum of basic science methods to gain comprehensive
information on a putative treatment target. Subsequent confirmative
research adopts study designs similar to those used in clinical stroke
trials and puts more emphasis on predictive stroke models and study end
points using larger and adequate sample sizes for necessary statistical
power.5
The group’s recommendations reflect latest developments and concepts in
the field, aiming to ultimately enhance the predictive value of
preclinical stroke research.
Current Challenges
Key
challenges include the choice of end points, homogeneity, utilization
of imaging, assessment of important comorbid conditions, and
conceptually disentangling the components of neurorecovery. End points
in confirmative stroke research need to reflect central clinical safety
and efficacy readout parameters rather than intermediate outcomes that
are designed to confirm the impact of the therapeutic approach. Because
the accepted outcomes after stroke in the clinic are assessed at 90
days, this means including a surveillance period of at least a month
after intervention.6
This is important because transient functional improvements after
experimental therapy lasting ≤9 weeks have been observed preclinically.7
Shorter observation periods might lead to false-positive or simply
incorrect results. Homogeneity of preclinical stroke models, at best
marginally representing the broad spectrum of cortical, subcortical, and
combined ischemic lesions exhibited by stroke patients, is a key
challenge. In acute stroke, the highly dynamic changes in the infarct
core and particularly in the penumbra should be understood by
preclinical stroke models because the existence and size of the latter
became an important criterion to select individuals who might benefit
from acute intervention both preclinically and clinically.8,9
Influence of age, sex, and common stroke comorbidities, such has
hypertension, diabetes mellitus, or hypocholesteremia, should also be
modeled. Potential interactions between a novel stroke treatment and
comedications required to treat those comorbidities need to be
identified to increase safety and efficacy during the translation
process.10
Finally, reliable discrimination between functional compensation and
recovery is important because rodents have a higher ability to
compensate functional benefits while economic and simple; hence,
frequently performed behavioral tests are often insensitive to such
masking behavior.11
Recent Milestones and Moving Forward
The evolution of endovascular therapy has produced a consistent true ischemia–reperfusion model in human acute ischemic stroke.12
Failure of translation of over a thousand molecules that were proven to
be effective in rodent or small mammal ischemia–reperfusion models may
in large part be because of the fact that, in the recent past, human
ischemic stroke was a permanent focal ischemia model and not a transient
focal ischemic model as believed.2
A fundamental principle of the human endovascular ischemic stroke
trials was the identification of a target vessel occlusion and a tissue
window for patient selection. The same principle may be used in
preclinical stroke research while recognizing possible measurement error
of imaging techniques and their matching with actual pathology to an
intermediate extent. Future therapies must select subjects in both
preclinical and clinical stroke trials on the basis of the tissue window
to target the population of interest.
We do not have
strong clinical examples of medical or device interventions for
neurorecovery yet. However, the components of neurorecovery must be
elucidated, and an implied focus on specific types of recovery is
needed. Focus on the upper limb, lower limb, kinesthesia, language, and
other specific neurological functions will be necessary to understand
how the brain and the patient recover. While function is the principle
pragmatic outcome, some combination of adaptation, rehabilitation, and
true recovery of function will likely occur; each component may respond
to specific intervention(s). This can and should be modeled.
Similarly,
stroke prevention models are lacking for many types of ischemic stroke.
Stroke suffers from causal multiplicity. While much has been learned
about atherosclerosis spurred by research on the coronary circulation,
the causes of cervical artery dissection, some types of cardioembolism,
and lacunar stroke are poorly understood. Modeling stroke prevention by
specific cause of stroke is needed.
Finally, in all
types of models, and particularly in the confirmative concept of
translational research, there is a growing awareness of the need to
emulate strong clinical data methods. Double-blind, randomized trials in
animal models must be used and powered appropriately to detect key
clinical outcomes. Such trials can be multicenter. The impact of big
data and open science can help stroke research if we embrace the
concepts of widely sharing data and techniques, using public data
deposit with standardized data definitions.
Future Solutions and Conclusions
Success
of future translational stroke research will not only critically depend
on focusing on the most appropriate end points but also on addressing
the right patient population. There is a need to think circularly not
only from bench to bedside but also from bedside to bench. Confirmative
preclinical stroke trials should be designed toward the patient
population most likely to be seen in the subsequent clinical trial.
Clinical stroke trials must recruit patients who match the
characteristics of experimental subjects in the preclinical stroke trial
it is based on and only later expand to broader population. When the
preclinical and clinical stroke research is consistent, translational
success will follow.
No comments:
Post a Comment