provide a summary of your research and the project overview.
Heart failure (HF) is a major cause of morbidity,
mortality, and resource utilization throughout the world, and its importance is
likely to increase with the aging population. Patients are heterogenous with
respect to etiology, prognosis, and response to therapy, and our ability to
identify high-risk patients versus those likely to improve with medical therapy
remains limited. Approximately 50% of all HF patients have a preserved ejection
fraction (HFpEF). Most HFpEF patients
are women, and there are currently no known effective therapies for HFpEF. In
comparison, HF with reduced ejection fraction (HFrEF), which predominately
affects men, has many established therapies. Characterization of HF subgroups
may provide insight into mechanisms of disease, gender differences, and
facilitate personalized prediction of outcomes and treatment response. I use
emerging bioinformatics technologies including large-scale database systems and
cluster-based analysis applied to clinical and molecular data to identify and
characterize subgroups of HF. The central hypothesis of my work is that
systematic characterization of HF patients will improve HF outcomes by
supporting precision HF management. I propose using
these results to develop personalized treatment plans for HF patients to
maximize likelihood of clinical improvement and minimize unnecessary
the implications for your research? Why is your work important both short term
and long term?
I plan to use the results of my research to develop
personalized treatment plans for HF patients to maximize likelihood of clinical
improvement and minimize unnecessary interventions. In the
short term, the results of my research may provide valuable insight into
mechanisms of HF and identify potential strategies for personalizing HF
management based on patient characteristics. My long-term goal is to develop
technologies that combine electronic health data with molecular profiling data
to identify important HF subgroups and demonstrate efficacy of personalized HF
care. The product of my research will be a patient-centered, reusable
infrastructure to study gender differences in HF, test established predictive
models, and discover novel HF subgroups that can be shared with providers via
an electronic medical record through real-time decision support.
take the above two questions and answer them as if you are explaining them to a
HF is common health problem that affects many people,
particularly the elderly, and frequently results in death within a few years of
diagnosis. Although some HF patients can
be treated with medicines, about half of patients have a type of HF (HFpEF)
that has no proven treatments. Most of
the patients with this kind of HF are women. I believe that we may be able to
use computer-based methods similar to the ones used by Amazon, iTunes, and
Pandora to identify groups of patients that may respond better to certain
medicines. To do this, I combine electronic data stored by patients’ doctors with
genetic testing in order find special patient groups. I then look to see if any of these groups of
HF patients do better than others with specific medicines. If I am successful, I hope that this
information can be built into the computers at doctors’ offices so that doctors
can use it to customize treatment of their HF patients.
is the proposed timeframe for your project? Where would you like to see your
This work has been ongoing for approximately 3 years,
resulting in 3 publications on this specific project (PLOS One, the Journal of
the American College of Cardiology-Heart Failure, American Journal of
Cardiology) in addition to several others in the areas of drug safety. The
manuscript describing the initial HFpEF analysis is currently in revision with
the European Journal of Heart Failure, and we are about to submit a manuscript
to Circulation describing the Effect of β-Blocker On Remodeling and Gene Expression Trial (BORG),
which combines clinical characteristics and molecular profiling to elucidate
mechanisms of response to β-blockers in HF.
Dr. Lindenfeld and I are about to submit a paper of efficacy of
telemonitoring to the Journal of Cardiac Failure, and we have another
manuscript on association of timing of hospital admission with outcomes in HF
that we hope to submit within the next 6 months, possibly to JACC-HF or the
European Journal of HF. Ultimately, I
would also like to write a methods paper detailing the analytic pipeline for
publication in either the Journal of Bioinformatics or the Journal of the
American Medical Informatics Association.
My primary mentor is JoAnn Lindenfeld, and my secondary
UC mentor is Mike Bristow.
you are able to, please share any recent progress or findings to date.
We have identified subtypes of both HFpEF and HFrEF that
have quite different prognosis (survival time) and response to therapies. This includes one group of HFpEF patients
(patients with diabetes mellitus, hypertension, high cholesterol, and kidney
problems) who may respond to the drug irbesartan. We now plan to look at additional variables
that may be useful in identifying patient subgroups and validate the results in
independent clinical trials, most notably the CHARM trial.
We also published finding expanding the number of
possible risk factors for peripartum cardiomyopathy (PPCM) by looking at a very
large number of hospital records. We
were also able to demonstrate the risk of adverse neonatal events was also
higher when PPCM was present. In the
course of this work, we proposed a risk score for estimating the likelihood
that a patient will develop PPCM.
A similar analysis looking at many hospitalized heart
failure patients over 14 years in the state of New York also revealed strong
associations between survival in HF hospitalization and season (winter is
worst), day of the week (Friday/Saturday is worst), and time of the day
(afternoon/evening is worst). This work
has only been presented in abstract form, but got a substantial amount of press
at the 2013 European Society of Cardiology HF Congress (Lisbon, Portugal).
In the BORG analysis, we have greatly expanded the number
of myocardial genes whose expression appears to be associated with improvement
in heart function. We have also
identified a number of possible predictors of response to β-blockers, including some which
perform markedly better in women than in men.
This study will produce a series of papers, the first of which is nearly
In collaboration with Bosch Healthcare, we analyzed
results from a demonstration study of a telemonitoring device called the Health
Buddy in Medicare patients. The primary
finding was that availability of the Health Buddy was associated with a
significantly improved survival rate, which was primarily seen in those who
actually used the device. This
manuscript is essentially complete and will be submitted after approval by the
Centers for Medicare and Medicaid Services.
With the support of UCH, we have constructed a real-time
decision support system. The pilot
application is calculation of a complex risk score for hospital readmission in
patients hospitalized for HF. We plan to
expand this functionality to permit use of other risk scores and patient subgroup
definitions that are identified in the course of my other work.
are some of your recent publications?
Kao DP, Kreso E, Dale RA,
Gonarow GC, Krantz MJ. Characteristics and outcomes among heart failure
patients with anemia and renal insufficiency with and without blood
transfusions (Public Discharge Data from California 2000-2006). Am J Card. 2011; 107:69-73. PMCID:
Kao DP, Wagner BD, Robertson AD,
Bristow MR, Lowes BD. A personalized BEST: Characterization of latent clinical
classes of nonischemic heart failure that predict outcomes and response to
bucindolol. PLOS One. 2012; 7:e48184.
Kao DP, Davis G, Aleong R,
O’Connor CM, Fiuzat M, Carson PE, Anand IS, Plehn JF, Gottlieb SS, Silver MA,
Lindenfeld JA, Miller AB, White M, Murphy GA, Sauer W, Bristow MR. Effect of
bucindolol on heart failure outcomes and heart rate response in patients with
reduced ejection fraction heart failure and atrial fibrillation. Eur J Heart Fail. 2013; 15324-33. PMCID:
Kao DP, Bucher Bartleson B,
Khatri V, Dart R, Mehler PS, Katz D, Krantz MJ. Trends in reporting of
methadone-associated cardiac arrhythmia, 1997-2011: an analysis of registry
data. Ann Int Med. 2013; 158:735-740.
ao DP, Hsich E, Lindenfeld J.
Characteristics, adverse events, and racial differences among delivering
mothers with peripartum cardiomyopathy. JCHF
2013; 1:409-416. PMCID: PMC3806506.
list some of the recent conferences where you have presented your research or
attended. (limited to those while funded by the CWHR,
Conference Oral Presentations
Epperson LE, Kao DP, Minobe W, Gilbert EM, Lowes BD, Bristow MR. Gene Expression
Changes Associated with β-blocker reverse remodeling in the failing human heart
constitute a β1-adrenergic
receptor gene regulatory network. Scientific Sessions of American Heart
Association, Nov 16-20, 2013.
Krantz MJ, Traut C, Kao DP. Opioid Agonists & Cardiac
Safety: Pharmacovigilance & ECG Implementation. American Association for
the Treatment of Opioid Dependence National Conference, Nov. 11, 2013.
Hinterberg M, Kao DP, Karimpour-Fard A, Sucharov K, Hunter LE, Port JD, Bristow
MR. Myocardial expression of the microRNA dre-mir-133a-5p is associated with
improvement in left ventricular ejection fraction in patients with idiopathic
dilated cardiomyopathy treated with β-blockers. Annual Scientific Session & Expo of Am
Coll Cardiol, March 10, 2013.
DP, Varosy P, Aleong R. Impact of payor status
on presentation, therapies, and outcomes in 31,614 patients hospitalized with
Wolff-Parkinson-White Syndrome. American Heart Association Annual Scientific
Sessions, Nov 5, 2012. Circulation.
DP, Epperson LE, Karimpour-Fard, Gilbert EM,
Lowes BD, Hunter LE, Bristow MR. Transcription
of Select Homeobox Genes Increases in Patients with Heart Failure-Reduced
Ejection Fraction Who Have Improvements in Left Ventricular Ejection Fraction
on β-blocker Therapy. American Heart Association Annual Scientific
Sessions, Nov 5, 2012. Circulation.
Kao DP, Varosy P, Nguyen DT,
Tzou W, Katz D, Shculler J, Sung R, Steckman D, Sauer WH, Aleong R. Predictors
of adverse events in 24,890 patients undergoing lead extraction. Scientific
Sessions, American Heart Association, Nov 16-20,2013.
Kao DP, Varosy P, Nguyen DT,
Tzou W, Katz D, Shculler J, Sung R, Steckman D, Sauer WH, Aleong R. Temporal
trends in mortality and adverse events in 24,890 patients undergoing lead
extraction over a decade. Scientific Sessions, American Heart Association, Nov.
Kao DP, Epperson LE, Meyer
L, Ferguson D, Zolty R, Weller JD, Gilbert EM, Lowes BD, Bristow MR. Serial
gene expression changes associated with reverse remodeling in dilated
cardiomyopathies: results of the effects of beta-blockers on remodeling and
gene expression in the failing human heart (BORG) Trial. Scientific Sessions,
American Heart Association, Nov 16-20, 2013.
Kao DP, Epperson LE,
Karimpour-Fard A, Gilbert EM, Lowes BD, Hunter LE,
Bristow MR. Expression levels of select mRNA triads are predictive of response
to beta-blocker therapy in patients with reduced ejection fraction heart
failure. European Society of Cardiology Heart Failure Congress 2013.
Kao DP, McIlvennan CK, Page
RL, Lindenfeld J. Impact of day, month, and hour of admission on inpatient outcomes
in 949,907 hospitalizations for congestive heart failure. European Society of
Cardiology Heart Failure Congress 2013.
and how did you become interested in research?
I have been
involved in biomedical research almost continuously since 1996. I worked in 2 primary labs during my
undergraduate education and received a Bachelor of Science in Biomedical
Engineering. Since then, I have
participated in basic science/molecular genetics, behavioral science, clinical
intervention, clinical imaging, phenotyping and now applied informatics
research. I have always enjoyed the
process of discovery, but it was not until I had the opportunity to complete a
post-doctoral fellowship in Biomedical Informatics Research at Stanford
University that I found a research strategy that really suited by experience
and ability. I find the diversity of
studies that the informatics skill set allows me to conduct to be extremely
powerful, and through my work, I hope to contribute to unlocking the potential
of electronic clinical data to improve health.
has/have been the most influential mentor(s) in your career? Why
Roy Ziegelstein, MD (Johns
Hopkins) – He is thoughtful and honest, and was the first mentor I had where I
was sure that his primary interest was my success. I aspire to be like Roy when I am in the
Norman Rizk, MD (Stanford
Hospital) – He is transparent, honest, and supportive. He gave excellent advice, even when it was
hard to hear. Again, I was sure that his
priority was my well-being and su
JoAnn Lindenfeld, MD
(University of Colorado) – She is fearless, and selfless in her advocacy. She is also excited by new and challenging
ideas, and she is willing pursue them, even in the face of skepticism. She is
also one of the best clinical cardiologists I’ve worked with.
Michael Bristow, MD PhD
(University of Colorado) – He is tenacious and determined in the pursuit of new
knowledge. From him, I have learned to
do the next best experiment, even if it won’t change the world and look out for
the surprising findings because they are often the true discoveries.
other grants/funding opportunities are you currently pursuing?
NHLBI R21 – Secondary Dataset Analyses
in Heart, Lung, and Blood Disease and Sleep Disorders, PI: Kao (submitted, Nov.
NHLBI K08 – Mentored Clinical
Scientist Development Award, PI: Kao (due February, 2014)
Butcher Seed Grant – Co-investigator:
Carsten Görg, (due January, 2014)
Fogarty R25 – Global Health Research
and Research Training eCapacity Initiative; PI: James Hakim, University of
Zimbabwe (due May, 2014)
Colleen McIlvennan and David Kao, 2014 American College of Cardiology national conference,