A meta-analysis of 20 short term comparative studies of 5 selective serotonin reuptake inhibitors (SSRIs; citalopram, fluoxetine, fluvoxamine, paroxetine and sertraline) has shown no difference in efficacy between individual compounds but a slower onset of action of fluoxetine. There were suggestions that fluoxetine caused more agitation, weight loss and dermatological reactions than the other SSRIs, More patients discontinued fluvoxamine and fewer patients stopped sertraline because of adverse effects than their comparator SSRIs. The mot common adverse reactions to the SSRIs were gastrointestinal (especially nausea) and neuropsychiatric (particularly headache and tremor). Data from the Committee on Safety of Medicines showed more reports of suspected reactions (including discontinuation reactions) to paroxetine, and of gastrointestinal reactions to fluvoxamine and paroxetine. than the other SSRIs during their first 2 years of marketing. Prescription-event monitoring revealed a higher incidence of adverse events related to fluvoxamine than its comparators. There were higher incidences of gastrointestinal symptoms, malaise, sedation and tremor during treatment with fluvoxamine and of sedation, tremor, sweating, sexual dysfunction and discontinuation reactions with paroxetine. Fluoxetine was not associated with a higher incidence of suicidal, aggressive and related events than the other SSRls. Patients have survived large overdoses of each of the compounds, but concern has been expressed over 6 fatalities following overdoses of citalopram. Drug interactions mediated by cytochrome P450 enzymes are theoretically less likely to occur during treatment with citalopram and sertraline, but there is a sparsity of clinical data to support this. Methodological difficulties and price changes do not allow choice for recommendations on the choice of SSRI based on pharmacoeconomic data. Taking into account the strengths and weaknesses of the methods used to compare drugs, guidelines to the selection of individual SSRIs in clinical practice are proposed. Citalopram should be avoided in patients likely to take overdoses. Fluoxetine may not be the drug of first choice for patients in whom a rapid antidepressant effect is important or for those who are agitated. but it may have advantages over other SSRIs in patients who are poorly compliant with treatment and those who have previously had troublesome discontinuation symptoms. Fluvoxamine, and possibly paroxetine, should not be used as first choice in patients especially prone to SSRI-related adverse reactions, while paroxetine should be avoided if previous discontinuation of treatment was troublesome. When in doubt about the risks of drug interactions, citalopram or sertraline should be considered given the lower theoretical risk of interactions.
