Preclinical and epidemiologic studies reveal sharply disparate carcinogenic profiles of drugs commonly used in psychiatry, offering some degree of reassurance about certain drugs, including many benzodiazepines, but suggesting that caution might be needed in long-term prescribing of atypical antipsychotics.
Antidepressants fall into a middle ground, with available data revealing a potentially high risk of carcinogenicity in 40% of monoamine oxidase inhibitors (MAOIs), 33% of modern antidepressants (including selective serotonin reuptake inhibitors [SSRIs]), and 22% of tricyclic antidepressants.
Dr. Juan F. Gálvez-Flórez and his associates at Tufts Medical Center in Boston reviewed available pharmacologic records and published studies from 1965 to 2009 and found that carcinogenic data exist for two-thirds of psychotropic drugs, but that the research is spotty and often inconclusive.
Despite the shortcomings underlying available data, the body of evidence is deserving of consideration by prescribers and largely unknown to clinicians, said Dr. Gálvez-Flórez at the annual meeting of the American Psychiatric Association in New Orleans.
“Our patients are chronic,” he said. “Patients with bipolar disorder or schizophrenia or depression … take medications for years, for decades. This should be a great concern for everyone in this audience and all practicing psychiatrists.”
Ideally, clinical trials should include carcinogenicity as a primary outcome measure, and longitudinal trials should examine carcinogenicity of classes of drugs and specific agents, in order to give clinicians a better sense of the risk/benefit equation when prescribing psychotropic agents, he recommended.
Dr. Gálvez-Flórez pointed to the huge Women's Health Initiative trial that was necessary to identify substantial, previously unsuspected risks of hormone replacement therapy, and suggested that a similar effort would seemingly be warranted for the most widely prescribed psychotropic drugs.
For now, preclinical data, epidemiologic studies, and a handful of case-control studies provide the only window into many drugs' potential to increase patients' risk of cancer. The few clinical trials exploring carcinogenic risks have produced mixed results and are marred by a failure to control for myriad confounding factors, he said. “This is a summary of our current knowledge,” said coinvestigator Dr. Nassir Ghaemi, director of the mood disorders and psychopharmacology programs at Tufts University, Boston, in a telephone interview. “What this provides is an awareness that the medical risks of these agents cannot be taken for granted.”
Before he became involved in the review, Dr. Ghaemi said he gave no consideration to carcinogenicity of antidepressants, even when prescribing to patients with a history of cancer.
Now, he said he would attempt to choose an agent with the least amount of carcinogenic potential based on preclinical data, assuming other prescribing considerations were equal.
With regard to antidepressants and breast cancer risk, conclusions about class-wide data remain murky.
A large, Canadian case-control study published in 2002 raised concerns of heightened cancer risk among patients prescribed certain tricyclic antidepressants (amoxapine, clomipramine, desipramine, doxepin, imipramine, and trimipramine) but not others (amitriptyline, maprotiline, nortriptyline, protriptyline) (Br. J. Cancer 2002;86:92-7).
In a follow-up study in 2006 by McGill University, Toronto, researchers from the same research group failed to find significant evidence of the increased risk, but nonetheless saw differences between the two groups of tricyclic antidepressants, with the group of drugs originally suspected of carcinogenicity numerically associated with higher breast cancer risk.
Dr. Gálvez-Flórez maintained that such information, while not definitive, might help to guide prescribing decisions.
No Food and Drug Administration–approved tricyclic antidepressant in the Tufts review showed a low potential risk of carcinogenicity (defined as no positive preclinical or clinical studies or only negative studies). And yet, class wide, the percentage of tricyclics with two or more positive studies of carcinogenicity was lower than that of modern antidepressants or MAOIs.
Within the SSRI class, four medications met the Tufts group's criteria of posing a low carcinogenic risk: fluvoxamine, duloxetine, venlafaxine, and nefazodone. “It was quite shocking to us that sertraline, citalopram, and fluoxetine–which are the main drugs we use in psychiatry–have a high risk of carcinogenicity [according to available data],” he said.
Mirtazapine also fit Tuft's criteria for a high risk of carcinogenicity: two or more preclinical studies with evidence of increased cancer risk.
Among older antidepressants, only moclobemide (which does not have FDA approval) had enough data to convincingly suggest a low carcinogenetic potential, although some, including doxepin and nortriptyline, demonstrated negative genotoxicity in available studies.
Among 15 epidemiologic studies of antidepressants, including 2 that assessed amphetamines, 6 had positive findings, 7 had negative findings, and 2 elicited “questionable” conclusions about carcinogenicity.
Drugs used in the treatment of attention-deficit/hyperactivity disorder also possess varying degrees of available data, and variable levels of carcinogenicity potential, based on that limited data.