A rose by any other name would smell as sweet.

This erudite paper raises major issues of definition of disorders, description of syndromes, and indirectly of clinical management [1]. The nub of the matter, pharmacologically and clinically, relates to the detailed phenomenological dissection of the syndromes that accompany discontinuation of a drug given in a therapeutic context, and the pharmacological mechanisms involved. Briefly, the clinical syndromes include:

  • 1Rebound, where the initial target symptoms reappear in exaggerated form. This has been studied most intensively with hypnotic drugs, because the polysomnogram provides a precise measure of sleep stages [2].
  • 2Relapse, in which the initial disorder recurs because the medication has suppressed but not ‘cured’ the disorder, and natural remission has not supervened.
  • 3Recurrence: this is the onset of a new episode of illness after a period of normality, usually arbitrarily defined.
  • 4Withdrawal syndromes, which have a characteristic if not pathognomonic collection of symptoms and signs in a well-defined temporal relation with stopping the drug. By arbitrary definition, clinicians will expect to see at least three new symptoms, i.e. not reported previously by the patients.
  • 5Pseudo-withdrawal: one syndrome which is often overlooked is pseudo-withdrawal in which the patient, warned that discontinuation is imminent, reports vague symptoms.
  • 6Dependence is generally regarded as a physiological state implied from the emergence of withdrawal symptoms, divided traditionally into physical and psychological dependence.
  • 7Discontinuation syndromes are descriptions of all of the above.
  • 8Addiction, abuse, etc. can be associated with withdrawal, but contain a kaleidoscope of features such as self-destructive drug-related life-style, social deterioration, cravings and high but not inevitable rate of relapse on discontinuation.

Margrethe Nielsen and her associates rightly emphasize the changes over time in various definitions and how these have altered the perception and management of drug users; but these changes reflect the attitudes of those involved in this field, sometimes based on political and economical biases, and not always on data. Indeed, the evidence base in the area of ‘addiction’ is always limited, at least clinically but not in animal experiments, because of the inability to assign subjects randomly to drug, placebo and perhaps no-treatment groups. For example, the rate of successful withdrawal in drug users is influenced strongly by whether or not the individual is in therapy ([3], p. 65–67).

The basic thesis of this paper is that benzodiazepines (BZDs) and selective serotonin re-uptake inhibitors (SSRIs) are associated with similar reactions which meet the usual criteria for withdrawal and are part of a dependence syndrome. Juggling the changes in definition is an academic exercise and many clinicians would go straight to the practical implications; namely, how frequent, severe, prolonged and disruptive of everyday life are these syndromes? Importantly, how easy are they to avoid or minimize by tapering and how can they be managed successfully? The authors concentrate on the symptom patterns, reporting that the symptoms were very similar for 37 of 42 identified withdrawal reactions. However, this obscures differences in symptom spectra, relating mainly to the distress caused by each individual symptom. For example, perceptual hypersensitivities bedevil the patient attempting to withdraw from BZDs, and these may be protracted [4].

BZD withdrawal has been recognized since the 1980s [5] and SSRI withdrawal since the 1990s [6]. There are many general similarities [7]. Drugs in both classes differ widely in the frequency and severity of withdrawal. Not everyone suffers withdrawal. The time–courses of withdrawal are similar. Relapse and resumption of medication are common. Management relies on tapering, psychological treatment and social support. Tapering is not fully effective [8,9]. Thus, BZD and SSRI withdrawal reactions are very similar in their clinical impact despite the differences in the underlying pharmacology, but there is one major difference. Withdrawal reactions, and by implication the state of dependence, are common in a proportion of both BZD and SSRI users, despite their being maintained on normal therapeutic dosages. SSRI users rarely escalate their doses, nor do they seek illicit supplies. Similarly, the bulk of BZD users are maintained on therapeutic doses by their prescribers. However, some do escalate their doses, becoming high-dose users with severe dependence. Also, the BZDs are well recognized as drugs of abuse, either on their own or as adjuncts to polydrug abuse with diamorphine and cocaine [10]. This is an important difference between the SSRIs and the BZDs.

Finally, we should not lose sight of another fundamental difference. Despite withdrawal reactions, most SSRIs have a favourable risk/benefit ratio. By and large, BZDs do not meet this criterion and should be avoided wherever possible [11]. We must be careful not to blur the distinctions between the two classes of drugs and discourage the careful use of SSRIs as antidepressants and anxiolytics.

Declarations of interest