Featured Articles

QuarterWatch™ (2016 Quarter 3 Data) Depression and Suicidal Behaviors: Exploring the Link with Certain Drugs. Serious Injuries from Aspirin

The latest issue of ISMP’s QuarterWatch™ provides a review of drug safety issues reflected in adverse drug events reported to the US Food and Drug Administration (FDA) during the third quarter (Q3) of 2016 and during the previous 12 months. In this report, we examine depression and suicidal thoughts or behaviors as an adverse effect of certain drugs, with a focus on three newer medications not intended to treat psychiatric disorders:

  • Roflumilast (DALIRESP), used to reduce exacerbations in patients with severe chronic obstructive pulmonary disease (COPD)
  • Apremilast (OTEZLA), used to treat psoriatic arthritis and plaque psoriasis
  • Suvorexant (BELSOMRA), used to treat insomnia

In this report, we also examine why so many serious injuries were reported for aspirin.

Adverse Drug Event Report Totals for 2016 Q3

FDA received 259,941 new reports of adverse drug events in 2016 Q3, a 10% decline from the previous quarter and a 26.6% decline from the same quarter in 2015. Of these, 70,942 US cases involved a fatal, disabling, or serious outcome. After decades of steady growth in reported events, report totals have been relatively stable for the past 2 years. However, in the most recent quarter, declines were seen in practically every category, including patient deaths. Despite this, the number of reported cases was still 4.8 times higher than the same quarter 10 years ago. This long-term increase in the number of reports may be attributed to the steady growth of newly approved drugs with new risks, and an increase in manufacturers that are now contacting patients for educational and marketing purposes, leading to greater awareness of adverse events and higher reporting rates.

Drugs, Depression, and Suicidal Behavior

After decades of denial and controversy, today more than 50 drugs have FDA-required warnings about the possible risk of depression and/or suicidal thoughts or behaviors. While the early drug warnings were for antidepressants and mood stabilizers, risks have been identified in drugs used for many other medical conditions, including a malaria drug (mefloquine), an acne medication (ISOtretinoin), and two smoking cessation aids (varenicline, buPROPion). Surprisingly little is known about these events, how often they occur, whether there are any telltale characteristics indicating a drug’s role, and what other factors might contribute to the adverse effect. We explore what can be learned about these adverse events by focusing on three recently approved drugs. 

Roflumilast (Daliresp)

Roflumilast was approved in 2011 and was the first among a new class of drugs that works by inhibiting phosphodiesterase-4 (PDE4), an enzyme prevalent in immune cells and active in the brain. The first drug that inhibited PDE4 was tested as an antidepressant and failed. During roflumilast clinical trials, FDA identified three suicides and two attempted suicides. As a result, the agency required the manufacturer to include a warning in the prescribing information and Medication Guide. Our analysis of the most recent 12 months of postmarketing reports confirms and extends this concern. Among 176 reports of adverse events with roflumilast, there were 11 cases of suicidal or self-injurious behavior, including 7 cases of suicidal ideation, 3 suicide attempts, and 1 completed suicide. To compare roflumilast to other drugs, we focused on reports of suicidal ideation and found that roflumilast accounted for 7.8 times the number of reports that would have been expected based on the combined results for all other drugs during the 12-month period.  

Apremilast (Otezla)

Apremilast, approved in 2014, also works by inhibiting PDE4 and was the second drug to be approved in this class. While FDA was uncertain whether a small number of cases of depression and suicidal thoughts or behaviors seen in clinical trials signaled a risk, the agency still required a cautiously worded warning in the prescribing information. While the preapproval data regarding these adverse effects might have been uncertain, our analysis of the most recent 12 months of reports identified a strong signal for depression—638 cases—which is 6.3 times greater than the number of expected reports. Reports of depression ranked in frequency just below the three most documented adverse effects of PDE4 inhibitors: diarrhea, nausea, and headache. In the 12-month totals, only sodium oxybate (XYREM, a central nervous system depressant used to treat narcolepsy) accounted for more frequent reports of depressed mood disorders than apremilast. We suspect that depression was underestimated in clinical trials due to the small number of patients studied and the failure to specifically ask about depression during patient assessments.

We also identified 68 cases of suicidal and self-injurious behavior, including 61 cases of suicidal ideation, 4 cases of suicidal behavior or attempt, and 3 completed suicides. To compare apremilast with other drugs we focused on the 61 reported cases of suicidal ideation and found this reported 1.6 times more than expected. This large total number of cases in 1 year has scientific weight. In the 12-month totals, only 17 drugs (mostly antidepressants and anticonvulsants/mood stabilizers) accounted for more reports of suicidal ideation. But given the large number of total reports for apremilast, the proportional reporting ratio of 1.6 times the expected number was only moderately elevated.

Suvorexant (Belsomra)

At the time of its approval in 2014, suvorexant was the first drug to block the binding of two newly discovered neuropeptides called orexins, which send wakefulness signals to neural circuits that regulate sleep/wake cycles. Faced with unknown risks of blocking a new kind of neurotransmitters, FDA required the manufacturer to use a specific assessment tool to query study patients about suicidal thoughts and behaviors.1 In a high-dose (30-40 mg) group (which was not approved by FDA), 8 cases of suicidal ideation were identified, compared to none with a placebo or in a lower-dose group (20 mg, which is the maximum approved dose). However, FDA still required a warning that possible side effects included aggressive behavior, hallucinations, worsening depression, and suicidal thoughts or actions. 

Our analysis of 12 months of reports ending with 2016 Q3 further supports the association of suvorexant with suicidal thoughts and behaviors. Among 2,290 adverse event reports for suvorexant, we identified 40 cases of suicidal or self-injurious behavior, including 19 cases of suicidal ideation, 11 attempted suicides, and 8 completed suicides. Reports of suicidal ideation occurred 3.1 times more frequently than expected. In addition, 38 cases described depressed mood disorders.

While QuarterWatch™ normally examines computer excerpts of adverse drug events, which exclude narrative detail, a tragic case reported directly to ISMP—combined with two cases of suicidal ideation from clinical trials—provide new insights. The case reported to ISMP involved a 56-year-old man who became aggressive and hostile with his wife, which was uncharacteristic for him, soon after he started taking 20 mg of suvorexant provided in sample packs from his physician. Because of his work schedule, he sometimes did not take a bedtime dose if he could not remain in bed for at least 7 hours, as instructed. On the days after he omitted a dose, the emotional effects were so substantial that he raised the issue with his doctor, but no changes were made in his treatment. His relationship with his wife was severely strained after he struck her without provocation one night, something he had never done before. A few days later, he committed suicide. 

This case shared four key features with 26 other cases involving varenicline (CHANTIX) and aggression, violence, and suicidality, that were investigated in 2010: 1) an inexplicable and unprovoked act of aggression or violence; 2) a victim of the violent act that was anyone nearby; 3) an onset soon after starting treatment; and 4) a suicidal act was a common resolution.2 However, one possible difference emerged in the suvorexant case. The man who committed suicide had frequently missed doses and had assaulted his wife after he had skipped a dose of medication. His personal records showed that he had taken suvorexant for a total of 13 nights over 26 days. In addition, FDA reviews describe two premarket cases of suicidal ideation: one occurred after missing a dose, another a few days after drug discontinuation. 


Depression and suicidal thoughts or behaviors are more challenging to detect and evaluate than other types of harm with clear physical symptoms or laboratory results. Identifying depression and suicidal ideation depends on how patients describe their thoughts and mood, and how systematically they are asked about these problems. Given these challenges, FDA evaluations prior to approval were impressive in the detailed accounting of the cases identified, and in the critical evaluation of these data as a signal for suicidal thoughts and behaviors. FDA reviewers were alert to the possibility that new PDE4 drugs for respiratory and skin disorders could have psychiatric side effects given their mechanism of action. And given inconclusive results in preapproval studies, the agency correctly opted for warnings. Increased use of FDA’s specific assessment tool for learning about suicidal thoughts and behaviors can improve our knowledge with future drugs, provided that the number of exposed patients is sufficiently large.

For roflumilast and apremilast, our assessment of postmarketing reports confirms and extends the evidence that the PDE4 inhibitor mechanism of action is linked to an increase in the risk of depression and suicidal thoughts or behaviors. In addition, a strong signal for depression linked to apremilast seen in reports suggests this effect may have been underestimated in clinical trials. Our findings for suvorexant confirm the risk of suicidal thoughts or behaviors already identified, and show the need to further investigate the potential of increased risk of these events when treatment is interrupted. Such events are likely to have been underestimated in clinical studies because patients were encouraged to take the drug without interruption during the treatment period. The prescribing information for all three drugs should be updated to include the additional information from postmarketing surveillance, and enhanced patient warnings are needed for some of the drugs; otherwise, patients might never suspect a drug was responsible for an unexpected change in mood or behavior.

Serious Injuries from Aspirin

Given that many would place aspirin near the bottom of a list of potentially dangerous drugs, we were surprised to discover it at or near the top of several rankings of risk in the case reports for 2016 Q3. It ranked first among all the drugs monitored by QuarterWatch™ for reported serious injuries to patients 75 years and older, and fourth among all drugs in reported serious injuries to US patients of all ages. The rankings remained high even when vague reports, consumer reports, and reports without appropriate investigation of patient deaths were excluded.

On further investigation, we found that most reports of serious injuries involved patients who were taking a combination of aspirin (mostly low-dose) and an anticoagulant, most notably rivaroxaban (XARELTO), warfarin (COUMADIN), apixaban (ELIQUIS), and enoxaparin (LOVENOX). Details to determine the appropriateness of the dual therapy were not provided in the reports. Overall, aspirin was the primary suspect drug in 2,134 reported cases during 2016 Q3, including 169 deaths and 1,137 gastrointestinal hemorrhages.


The risks of combining agents that inhibit blood clotting are high. FDA should sponsor a reanalysis of the extensive clinical trials data on anticoagulants with a goal of identifying and guiding the appropriate use of dual therapy with aspirin or other antiplatelet agents.   

The full QuarterWatch™ report with references can be found here.

What is QuarterWatch™?

QuarterWatch™ is the publication of an independent ISMP surveillance program that monitors adverse drug events reported to FDA by manufacturers, health professionals, and the public. The agency releases, for research and data analysis, excerpts of all domestic and foreign reports it receives into the FDA Adverse Event Reporting System (FAERS). The goal is to identify signals that may represent important drug safety issues which often require further investigation to determine their frequency and establish a causal relationship to the suspect drug.


  1. US Food and Drug Administration. Guidance for industry: suicidal ideation and behavior: prospective assessment of occurrence in clinical trials. August 2012.
  2. Moore TJ, Glenmullen J, Furberg CD. Thoughts and acts of aggression/violence toward others reported in association with varenicline. Ann Pharmacother. 2010;44(9):1389–94.