A large U.K. cohort study has shown that tricyclics may be safer for treating depression in people older than 65 years than are the more commonly prescribed selective serotonin reuptake inhibitors.
The findings, published online Aug. 2 in BMJ (doi:10.1136/bmj.d4551), are striking, considering how frequently SSRIs are prescribed to people aged 65 and older. Current guidelines published by the U.K.’s National Institute for Health and Clinical Excellence promote SSRIs as a first choice among drug treatments for adults with depression, including older adults.
The results are surprising "mainly because it isn’t what we expected from the guidance. There’s a general feeling that the SSRIs are safer," the study’s lead investigator, Carol Coupland, Ph.D., of the University of Nottingham (England), said in an interview.
Compared directly with tricyclic antidepressants, SSRIs were associated with significantly higher rates of all-cause death (adjusted hazard ratio, 1.32), stroke or transient ischemic attack (HR, 1.15), fractures (HR, 1.26), falls (HR, 1.27), hyponatremia (HR, 1.44), and epilepsy or seizures (HR, 1.80). There was no significant difference seen for attempted suicide or self-harm (HR, 1.27).
Absolute risks for all-cause mortality over 1 year were 7.04% for patients not taking antidepressants, 8.12% for those taking TCAs, 10.61% for SSRIs, and 11.43% for other antidepressants.
In this study the TCAs tended to be prescribed at smaller doses than the SSRIs. Of the group taking TCAs, 70% were taking less than half the defined daily dose, compared with 13.8% of those on SSRIs. The researchers noted significant trends (P less than .01) corresponding with dosages of TCAs and SSRIs for all-cause mortality, falls, and epilepsy or seizures.
All 60,746 subjects in Dr. Coupland and colleagues’ linked cohort study were aged 65-100 years (mean age, 75 years; 67% were women) and had been diagnosed with depression by their primary care practitioners in 1996-2008. Of these, 89% were prescribed at least one antidepressant drug, which 55% of the time was an SSRI. About a third of the subjects who had been prescribed a drug received a tricyclic antidepressant. The rest (13.5%) received drugs in other classes; only 0.2% of patients in the study were prescribed MAO inhibitors – too small a group to enter into the analysis.
In all, 11 antidepressant drugs (5 SSRIs, 4 TCAs, and 2 in other categories) accounted for 96% of prescriptions in the cohort, and median duration of treatment was 1 year. The researchers evaluated a host of adverse outcomes both for the drug class and for individual drugs.
The cohort presented with a variety of comorbidities typical for the age bracket, and investigators adjusted results for the potential confounding effects of these diseases and the drugs used to treat them, along with smoking status, socioeconomic status, and other potential confounding factors.
The results, Dr. Coupland said, "suggest possibly an area for concern. We don’t want to be alarmist. It’s very new research and needs to be confirmed by other studies," particularly randomized controlled trials.
The investigators did note significant differences in adverse outcome risks even among drugs in the same class. One drug in the tricyclic group, trazodone, was associated with the highest absolute risk for all-cause death in the cohort (12.44% in 1 year) and one of the highest for attempted suicide or self-harm (1.2% in 1 year).
Three SSRIs – citalopram, escitalopram, and fluoxetine – were shown associated with increased risk of hyponatremia, while two other SSRIs, paroxetine and sertraline, were not.
The two non-SSRI, non-TCA drugs – mirtazapine and venlafaxine – were shown associated with increased risk of suicide or self-harm along with all-cause death and stroke or transient ischemic attack, with venlafaxine additionally associated with higher risk of fracture and epilepsy or seizures.
Dr. Coupland and colleagues noted several limitations for their study, mainly its observational design and the accompanying susceptibility to confounding by indication, channeling bias, and residual confounding.
However, in an editorial (10.1136/bmj.d4660) accompanying their study, Dr. Ian B. Hickie of the Brain and Mind Research Institute at the University of Sydney, said the findings have "clear implications for more informed prescribing and enhanced clinical monitoring" regardless of the study design.
Given the potential harms outlined in the study, Dr. Hickie wrote, "the decision to prescribe for an older person with depression should not be taken lightly."
The study was funded by the U.K.’s National Institute for Health Research. Dr. Coupland and colleagues declared no conflicts of interest. Coauthors Richard Morriss of the University of Nottingham disclosed receiving past financial support for speaking at meetings sponsored by pharmaceutical companies. Editorialist Dr. Hickie disclosed no conflicts of interest related to his editorial, but acknowledged participation at industry-sponsored depression awareness seminars.