By Amanda Gardner
MONDAY, July 7 (HealthDay News) -- The widely prescribed class of antidepressants known as SSRIs may be associated with bleeding in the upper gastrointestinal tract, especially when taken with painkillers called nonsteroidal anti-inflammatory drugs, a new study says.
Although reports of this effect from the selective serotonin reuptake inhibitors family of antidepressants -- which includes Prozac, Paxil and Zoloft -- have been noted before, some outside experts said the evidence is still far from conclusive.
"Certainly, doctors should be advising patients of this finding, but it's far from definitive," said Dr. Ewald Horwath, professor of psychiatry, epidemiology and public health at the University of Miami Miller School of Medicine. "The risk of discontinuing antidepressants and getting depression and its consequences are much greater than this drug."
But, other experts said that the public health impact of the new finding, if confirmed, could be significant.
"SSRIs are the most commonly prescribed psychotropic drug in the U.S.," said Dr. Norman Sussman, a psychopharmacologist at New York University Langone Medical Center and associate dean of the post-graduate program at New York University School of Medicine. "This could be a major pharmaceutical issue."
The findings were published in the July issue of Archives of General Psychiatry. The study was led by researchers at the Spanish Agency for Medicines and Healthcare Products in Madrid. The research was supported by grants from the drug company AstraZeneca.
Some previous research has linked SSRIs with upper gastrointestinal tract bleeding. Other research, however, has found no such association.
For the new study, the researchers looked at 1,321 patients with upper GI bleeding who had been referred to a hospital or a specialist, and compared them with 10,000 "control" patients with no GI bleeding.
More people who had GI bleeding (5.3 percent) were taking SSRIs, compared to 3 percent of the controls -- an increased risk of 60 percent. Similarly, 1.1 percent of patients with GI bleeding were taking Effexor (venlafaxine, an antidepressant related to SSRIs), compared with 0.3 percent in the control group, an almost threefold increased risk.
And there was an almost fivefold increased risk of GI bleeding among users of SSRIs and NSAIDs or corticosteroids, and a ninefold increased risk among those taking NSAIDs and SSRIs without acid-suppressing agents, according to the study.
Both NSAIDs, a class of pain relievers that includes ibuprofen and naproxen, and corticosteroids have known negative effects on the GI tract.
There was also a "suggestion" of an interaction of SSRIs with antiplatelet drugs, which prevent blood clots from forming.
The study authors did not find a significant association for other types of antidepressants.
Based on these numbers, for every 2,000 people treated with one of these antidepressants, there would be one case of upper GI tract bleeding, higher than in the general population. But when NSAID drugs are added, there would be one case of GI bleeding among every 250 people taking the antidepressants, the study authors said.
"There has been a bit of controversy about the significance of this. This is the best study to date to look at this," said Dr. Joseph White, associate professor of internal medicine at Texas A&M Health Science Center College of Medicine and chief of laboratory medicine at Scott & White Clinic. "The risk probably does appear real, but it's slight, unless you're taking concomitant medications. If you're taking these antidepressants along with other medications, keep it in mind and discuss with your physician."
The study's lead author, Dr. Francisco J. de Abajo, said: "The increased risk of bleeding associated with SSRIs and venlafaxine is already included in the data sheets of the products. There have been several editorials and revisions on this matter that are referred to in my article. What is still a matter of controversy is the interaction with NSAIDs and the use of acid-suppressing agents as an effective minimization measure. The contribution of our article points in that direction."