Convulsive therapy has been in continuous use for more than 60 years. The clinical literature establishing its efficacy in specific disorders is amongst the most substantial for any medical treatment (Weiner and Coffey 1988; Mukherjee et al. 1994; Krueger and Sackeim 1995; Sackeim et al. 1995; Abrams 1997a). Like other medical treatments, various sources of evidence support the efficacy of ECT in specific conditions. The indications for ECT have been defined by randomized controlled trials comparing ECT to sham interventions or treatment alternatives and similar trials comparing modifications of ECT technique. The indications for ECT have also been supported by reports of uncontrolled clinical series, case studies, and surveys of expert opinion.

The decision to recommend the use of ECT derives from a risk/benefit analysis for the specific patient. This analysis considers the diagnosis of the patient and the severity of the presenting illness, the patient's treatment history, the anticipated speed of action and efficacy of ECT, the medical risks and anticipated adverse side effects, and the likely speed of action, efficacy, and safety of alternative treatments.

2.2. Referral for ECT

2.2.1. Primary use. There is considerable variability among practitioners in the frequency with which ECT is used a first-line or primary treatment or is only considered for secondary use after patients have not responded to other interventions. ECT is a major treatment in psychiatry, with well defined indications. It should not be reserved for use only as a "last resort." Such practice may deprive patients of an effective treatment, delay response and prolong suffering, and may possibly contribute to treatment resistance. In major depression, the chronicity of the index episode is one of the few consistent predictors of clinical outcome with ECT or pharmacotherapy (Hobson 1953; Hamilton and White 1960; Kukopulos et al. 1977; Dunn and Quinlan 1978; Magni et al. 1988; Black et al. 1989b, 1993; Kindler et al. 1991; Prudic et al. 1996). Patients with longer duration of current illness have a reduced probability of responding to antidepressant treatments. The possibility has been raised that exposure to ineffective treatment or to a longer duration of episode actively contributes to treatment resistance (Fava and Davidson 1996; Flint and Rifat 1996).

The likely speed and efficacy of ECT are factors that influence its use as a primary intervention. Particularly in major depression and acute mania, substantial clinical improvement often occurs soon after the start of ECT. It is common for patients to manifest appreciable improvement after one or two treatments (Segman et al. 1995; Nobler et al. 1997). In addition, the time to achieve maximal response is often more rapid than that with psychotropic medications (Sackeim et al. 1995). Besides speed of action, the likelihood of obtaining significant clinical improvement is often more certain with ECT than with other treatment alternatives. Therefore, when a rapid or a higher probability of response is needed, as when patients are severely medically ill, or at risk to harm themselves or others, primary use of ECT should be considered.

Other considerations for the first-line use of ECT involve the patient's medical status, treatment history, and treatment preference. Due to the patient's medical status, in some situations, ECT may be safer than alternative treatments (Sackeim 1993, 1998; Weiner et al. in press). This circumstance most commonly arises among the infirm elderly and during pregnancy (see Sections 6.2 and 6.3). Positive response to ECT in the past, particularly in the context medication resistance or intolerance, leads to early consideration of ECT. At times, patients will prefer to receive ECT over alternative treatments, but commonly the opposite will be the case. Patient preferences should be discussed and given weight prior to making treatment recommendations.

Some practitioners also base a decision for primary use of ECT upon other factors, including the nature and severity of symptomatology. Severe major depression with psychotic features, manic delirium, or catatonia are conditions for which there is a clear consensus favoring early reliance on ECT (Weiner and Coffey 1988).

2.2.2. Secondary use. The most common use of ECT is in patients who have not responded to other treatments. During the course of pharmacotherapy, lack of clinical response, intolerance of side effects, deterioration in the psychiatric condition, the appearance of suicidality or inanition are reasons to consider the use of ECT.

The definition of medication resistance and its implications with respect to a referral for ECT have been the subject of considerable discussion (Quitkin et al. 1984; Kroessler 1985; Keller et al. 1986; Prudic et al. 1990; Sackeim et al. 1990a, 1990b; Rush and Thase 1995; Prudic et al. 1996). At present there are no accepted standards by which to define medication resistance. In practice, when assessing the adequacy of pharmacological treatment, psychiatrists rely upon factors such as the type of medication used, dosage, blood levels, duration of treatment, compliance with the medication regimen, adverse effects, nature and degree of therapeutic response, and type and severity of clinical symptomatology (Prudic et al. 1996). For example, patients with psychotic depression should not be viewed as pharmacological nonresponders unless a trial of an antipsychotic medication has been attempted in combination with an antidepressant medication (Spiker et al. 1985; Nelson et al. 1986; Chan et al. 1987). Regardless of diagnosis, patients who have not responded to psychotherapy alone should not be considered treatment resistant in the context of referral for ECT.

In general, failure of patients with major depression to respond to one or more antidepressant medications trials does not preclude a favorable response to ECT (Avery and Lubrano 1979; Paul et al. 1981; Magni et al. 1988; Prudic et al. 1996). Indeed, compared to other treatment alternatives, the probability of response to ECT among patients with medication-resistant depression may be favorable. This is not to say, however, that medication resistance does not predict clinical outcome of ECT. Patients who have not responded to one or more adequate antidepressant medication trials have a lower probability of responding to ECT compared to patients treated with ECT without having received an adequate medication trial during the index episode (Prudic et al. 1990, 1996; Shapira et al. 1996). In addition, medication-resistant patients may require particularly intensive ECT treatment to achieve symptomatic improvement. Consequently, the bulk of patients who fail to benefit from ECT are likely to also be patients who have received, and not benefited from, adequate pharmacotherapy. The relationship between medication resistance and ECT outcome may be stronger for tricyclic antidepressants (TCAs) than for selective serotonin reuptake inhibitors (SSRIs) (Prudic et al. 1996).