The possible association of gastrointestinal bleeding and treatment with selective serotonin reuptake inhibitors (SSRIs) has long been known. SSRIs decrease the concentration of serotonin in platelets dramatically (> 80%). This leads to reduced platelet aggregation and results in increased tendency to bleed. The combination with non-steroidal anti-inflammatory drugs (NSAIDs), platelet inhibitors, or warfarin increases the risk further. It was not clear so far, however, how rapid after the initiation of treatment with SSRIs the risk of bleeding increases and to which this increase in risk amounts. These issues were addressed in a remarkable study by Yen-Po Wang and collaborators from Taiwan published in the January issue of the American Journal of Psychiatry (Wang et al., Am J Psychiatry 2014; 171: 54-61).
The scientists analyzed from the Taiwan National Health Insurance Research Database (NHIRD) the data of all psychiatric patients who had suffered an upper gastrointestinal bleeding between 1998 and 2009. Data on their treatment with antidepressants were also available. Here, the authors in their analysis differentiated SSRIs, tricyclic antidepressants, SNRIs, MAOIs and “other” antidepressants. The authors used the so-called case- crossover design, in which each patient serves as his own control. This design eliminates the role of confounding variables such as smoking or alcohol consumption, the presence of which is often not documented in such registers. The odds ratio for the risk of upper gastrointestinal bleeding associated with a particular drug was calculated as the ratio of the number of patients exposed to that compound in the 14-day period prior to the bleeding (index date) to the number of patients who were exposed to the drug only in the control period (during days 15-28 prior to the index date). Observation periods of 7 days prior to the index date (with the previous 7 days serving as control period) and 28 days (with the previous 28 days serving as control period) were chosen for sensitivity analyzes.
5500 of the 187 117 patients in the database had a upper gastrointestinal bleeding in the 12-year observation period. SSRIs were the only compounds, which increased the risk for upper gastrointestinal bleeding, and this increase in risk was obvious in all three observation periods. Even treatment with an SSRI for just one week increased the risk significantly. When the analysis was based on the 7-day observation window, an odds ratio of 1.67 was found (95 % confidence interval 1.23 to 2.26) . Based on the 28-day period, the odds ratio was 1.67 (95 % confidence interval 1.34 to 2.08). The largest increase in risk was observed for the 14-day observation window (odds ratio 1.84, 95 % confidence interval 1.42 to 2.40). For SNRIs, which were administered much less, there was an increase in risk at the border of statistical significance. I assume that a significantly increased risk of bleeding would have beed detected for SNRIs, if the number of cases treated with this group of compounds had been higher.
Across all periods of observation, the authors calculated an odds ratio for a upper gastrointestinal bleeding associated with SSRI treatment of 1.77. This value was not substantially lower than for NSAIDs (OR = 1.97) and aspirin (OR = 1.92). The administration of the platelet aggregation inhibitor aspirin in addition to an SSRI increased the risk of bleeding only slightly (OR = 2.07). The most significant increase in risk of upper gastrointestinal bleeding was found for the combination of an SSRI with an NSAID (OR = 3.44). Benzodiazepines or non-benzodiazepine hypnotics did not increase the risk of bleeding. Subgroup analyzes further showed that SSRIs increased the risk of bleeding in men but not in women .
This large study indicates that treatment with SSRIs (and possibly also of SNRIs) poses a risk, which is often neglected in clinical practice. Clinicians should act with special caution when administering those compounds in combination with NSAIDs.
This post is also available in: German