Does Dose Matter When it Comes to Antidepressants and Self-Harm?
Concerns about the possible risk of suicide in young people taking antidepressants are often voiced by experts and parents. But how much of a role does dosage play in the risk of suicide and self-harm among young people on antidepressants?
A group of researchers recently looked to answer this question by examining the relationship between antidepressant dose, age, and risk of deliberate self-harm. The propensity-score matched cohort study looked at healthcare utilization data to create a sample of more than 160,000 people with depression who took the SSRIs fluoxetine (Prozac), sertraline (Zoloft), or citalopram (Celexa) in a 13-year period, from January 1998 to December 2010.
The study group was restricted to new initiators only; that is, those who were taking the antidepressant for the first time. Participants were split into two groups: those ages 10 to 24 and those ages 25 to 64. They were then assigned to one of three dosage categories: Those who took the most common (or “modal”) dose of antidepressant, those who took a lower dose than that, and those who took a higher dose.
Researchers determined their modal dose based on the distribution of doses among all antidepressant initiators in the study. The modal dose of each drug in the age 10 to 24 group according to this study was 20 mg/day for both citalopram and fluoxetine and 50 mg/day for sertraline. Patients taking higher doses than the FDA recommended maximum dose or lower than the FDA recommended lowest dose were excluded from analysis. (It was unclear from the study if or how many children started immediately on the modal/high dose and/or how many were titrated up to it in the study period.)
Next, up to two participants in an age group with the same baseline characteristics (such as comorbidity, treatment history, etc) receiving the most common dose of a drug were matched with one receiving a high dose. This was to help researchers determine what reasonably would have happened to the high dose participants if they had instead been on a regular dose (ie, propensity score matching).
Rates of deliberate self-harm (DSH) were calculated over a one-year period starting with the initiation of medication. The rate of DSH was twice as high in people age 24 and younger who took high-dose therapy then those who took modal doses of SSRIs (31.5 per 1,000 person years vs 14.7 per 1000 person years). There was virtually no difference in self-harm between dosage groups in the older cohort.
Most self-harm events happened within the first three months of starting medication. For every 1,000 10- to 24-year olds, there were approximately seven more self-harm events among the high dose group than the modal dose group in the first 90 days of treatment. The number needed to harm (NNH) was 136 (Miller M et al, JAMA Intern Med 2014;online ahead of print).
CCPR’s TAKE: This study finds that high doses of SSRIs lead to greater risk of acts of deliberate self-harm among young people. However, there are several caveats. First, study design: Propensity matching in kids is very difficult, because it requires the members of the cohort to have done things that can be matched— made previous attempts or been hospitalized, for example, and often kids don’t have that kind of medical history, even if they have severe symptoms. In addition, the paper didn’t really address why the higher dose was chosen by the clinician. It’s possible that the kids with the most subjective distress got the higher initial doses, and therefore it stands to reason that they also had the higher likelihood of self-harm to begin with. Finally, it should be noted that 136 is a high NNH. For example, penicillin has an NNH of 20 for self-reported allergic reaction. However, there’s not a lot of evidence that antidepressants work any better at higher doses anyway, so keeping the dose as low as therapeutically effective is always a good bet.