Two of the most widely prescribed drugs in the world, when taken together, can increase blood glucose levels. That's the finding
of a group of researchers from California, Tennessee, and Massachusetts, the result of a unique effort to spotlight difficult-to-detect
adverse drug reactions.
"The challenge of discovering drug interactions is exacerbated because large clinical trials routinely focus on establishing
the effects of single drugs. Unpredictable combinatorial effects can be identified only through postmarketing surveillance
and signal detection," explained Russ Altman, M.D., Ph.D., a professor of bioengineering, genetics, and medicine; chair of
the Department of Engineering; and director of the Biomedical Informatics Training Program at Stanford University, and his
colleagues in the July Clinical Pharmacology and Therapeutics.
Adverse-event reporting programs may detect such interactions, but if the events are rare—or rarely reported—traditional methods
for detection may fail. "We reasoned that algorithms that identify the presence of an adverse event through an analysis of
aggregate side effects (rather than direct reporting) could be more successful. With this in mind, we developed a novel signal-detection
algorithm for the identification of drug-drug interactions from adverse-event reports in spontaneous reporting systems, using
an analysis of latent signals that indirectly provide evidence for an adverse event," wrote Altman and colleagues.
Nicholas Tatonetti, a doctoral student in biomedical informatics at Stanford University and lead author of the paper, described
for Psychiatric News the group's approach: "The current generation of data-mining algorithms are effective at identifying some single drug effects;
however, there is a greater challenge when trying to find drug-drug interactions. This is because there are often far fewer
reports for a given pair of drugs than for any single drug. We designed an approach that could identify the presence of the
interaction without actually requiring the effect to be reported explicitly. We do so by learning a side-effect profile that
is indicative of a more severe adverse event occurring under the surface, so to speak."
Their approach swiftly found a target, an intriguing suggestion that the lipid-lowering compound pravastatin and the selective
serotonin reuptake inhibitor (SSRI) paroxetine taken together may affect glucose homeostasis, even though neither drug by
itself is typically associated with significant changes in glucose levels, and an increase in glucose levels is not a general
effect of combined therapy with SSRIs and statins.
They focuses on the potential interaction, studying three independent cohorts of patients at academic medical centers with
extensive electronic medical records, in a retrospective observational study.
Then they extracted blood glucose measurements at baseline and random time points during treatment of 449 patients on pravastatin,
982 on paroxetine, and 18 on a combination regimen of paroxetine and pravastatin from the electronic medical record (EMR)
data at Vanderbilt Hospital. They did the same for 632 patients on pravastatin, 780 on paroxetine, and 109 receiving the paroxetine/pravastatin
combination from the EMR data at Partners HealthCare in the Boston area. The combination treatment cohorts showed a significantly
greater change in blood glucose in post-hoc tests compared with either the pravastatin or paroxetine cohorts.
The rapid timeline is perhaps the most exciting part of this story. "The entire process from data mining the FDA's adverse-event
reporting system database to replicating our clinical validation and drafting the paper for submission took just under two
months," said Tatonetti. "I think this highlights the power of these large-scale emerging datasets. We are able to formulate
and validate clinical hypotheses more rapidly than ever before."
How important was their finding? "An analysis of the prescriptions at the three sites from which we drew data revealed that
approximately 6 percent of the patients on pravastatin were also on paroxetine, and approximately 4 percent of patients on
paroxetine were also on pravastatin," said the researchers. "Given that there were approximately 18 million prescriptions
for pravastatin and approximately 15 million for paroxetine in 2009, we estimate that between 550,000 and 715,000 individuals
received prescriptions for combined therapy with these two drugs in the United States that year."
They pointed out the need for a prospective clinical trial in which fasting blood glucose, insulin secretion, and insulin
action are evaluated before and after the start of pravastatin and paroxetine combination therapy. As a first step, they have
already begun in vivo studies in mice, presenting preliminary confirmatory results in their publication.
"We are actively pursuing more in-depth animal trials as well as a prospective clinical trial in collaborations with some
colleagues here at Stanford," said Tatonetti. "We hope to further establish the presence of the effect as well as investigate
alternative combinations of SSRIs and statins to confirm our hypothesis that this interaction is specific to paroxetine and
pravastatin. This is especially interesting to us since it would allow us to make a recommendation to those physicians who
would like to consider alternatives."
The researchers noted that their data do not suggest that the interaction of pravastatin and paroxetine is a general class
effect between all statins and all SSRIs. "The data-mining procedure did not identify other specific statin-SSRI pairs as
being disruptive of glucose homeostatis," they wrote.
Their advice until more information is available about the pravastatin/paroxetine association? "It is important for patients
not to change their own meds," cautioned Altman. "In general, if the paroxetine is working, it probably should not be changed.
Physicians of patients on both medications might first consider whether there is a problem with hyperglycemia; if there is,
they should perhaps consider changing the cholesterol medication. The clinical significance of our findings has yet to be
Tatonetti's work is supported by the Department of Energy's Office of Science Graduate Fellowship Program.
An abstract of "Detecting Drug Interactions From Adverse-Event Reports: Interaction Between Paroxetine and Pravastatin Increases
Blood Glucose Levels" is posted at <www.ncbi.nlm.nih.gov/pubmed/21613990>.