When Discontinuing SSRI Antidepressants Is a Challenge: A Response to Letters to the Editor

To the Editor: We appreciate the opportunity to respond to the letters to the editor from Fava and Cosci and Hengartner et al. in response to our article (1–3). These letters are welcome additions to the discussion regarding management of antidepressant medication discontinuation.

We fully agree with Fava and Cosci (2) that discontinuing antidepressant medication(s) may be far more challenging than prescribing such medication, which is what prompted our article. Our intent was to make note of the multitude of factors to be considered when discontinuing antidepressant medications. We also agree that it is likely that individual patients may have more persistent or troubling withdrawal symptoms, the exact rate of which is not easily discernable from the extant literature based on clinical trials. In fact, Fava et al. (4) state in the discussion section of their article that the “withdrawal syndrome typically occurs within a few days from drug discontinuation and lasts a few weeks. However, many variations are possible, including late onset and/or longer persistence of disturbances.” The issue of typical duration of discontinuation symptoms was brought up by Hengartner et al. in their letter (3). We agree with them that extant literature may underestimate the duration of discontinuation symptoms owing to short periods of assessment after discontinuation of antidepressants. In fact, the Davies and Read systematic review cited by Hengartner et al. also found that most studies assessed discontinuation symptoms for only a few weeks (5).

In the first paragraph of their letter, Fava and Cosci suggest that we argue discontinuation symptoms occur only in the context of abrupt discontinuation. In fact, elsewhere in the article we do state:

Factors not shown to predict development of discontinuation symptoms include … length of medication taper;

With paroxetine, discontinuation symptoms may also occur if doses are skipped for a few days;

Any patient who either discontinues or markedly reduces the antidepressant medication dosage is at risk of discontinuation symptoms; and

Gradual taper may reduce the severity of discontinuation symptoms as compared with abrupt discontinuation.

We also included information about persistence of discontinuation symptoms beyond a few weeks: “However, some reports have documented persistence of symptoms for up to 1 year.” Other clinical situations listed by Fava et al. that might warrant interruption in antidepressant treatment were also listed in our article:

Beyond the acute phase, changes to the patient’s antidepressant medication regimen may be warranted to address residual symptoms, persistent functional impairments, loss of initial therapeutic effect (antidepressant tachyphylaxis), occurrence or prevention of relapse or recurrence, side effects, affordability, or anticipated drug-drug interactions with medications for other medical conditions.

In addition, we addressed the very concerns raised by Fava and Cosci regarding the conceptualization of discontinuation symptoms as “isolated, self-limiting manifestations” (2) in the section of our article titled “What are the Best Strategies to Manage Discontinuation Symptoms?” In this section, we recommended strategies for prevention as well as treatment. Moreover, we recommend that clinicians need to enhance shared decision making by improving patient education about major depressive disorder, its treatments, and the risk of discontinuation symptoms.

Although we recognize that discontinuation symptoms can be challenging, the alternative of not prescribing antidepressant medications is also not a viable option for those who are likely to benefit. The need for carefully well-designed clinical trials specifically addressing discontinuation cannot be overemphasized. For example, in a recent review article, Horowitz and Taylor argue that taper should be even slower and that the minimum medication dosage preceding discontinuation should be much lower than the therapeutic minimum in order to prevent the occurrence of discontinuation symptoms (6). Such strategies need to be tested prospectively. To get an accurate assessment of prevalence and duration of discontinuation symptoms in clinical practice, large naturalistic studies are needed where patients with depression are enrolled and followed prospectively as they get care during routine clinical care (e.g., ClinicalTrials.gov identifier: NCT02919280). Furthermore, measurement-based practice of systematically assessing discontinuation symptoms and documenting these in clinical practice may facilitate survey of these symptoms through electronic health records (e.g., ClinicalTrials.gov identifier: NCT02697487), which will likely prove to be more accurate than web-based surveys as cited by Hengartner et al. from the Davies and Read article.