Sample characteristics

The total number of case notes studied was 221, of which 103 were for males (mean age 41 years, range 19-77) and 118 for females (mean age 43.5 years, range 19-72). As regards the severity of intellectual disability, this was mild in 43% of the sample, moderate in 26.2%, severe in 16.3%, profound in 2.3% and borderline intellectual disability in 2.7%; in 9.5% of the sample, the level of intellectual disability was not classified.

The underlying cause of intellectual disability was not known in 77% of the cases. The most common known cause was Down syndrome (10.8%), with the following other main causes: cerebral palsy (3.1%), Angelman syndrome (1.8%), fragile-X syndrome (1.3%), Prader-Willi syndrome (0.9%), birth injury (0.9%) and childhood encephalitis (0.9%).

The majority of service users were either living in supported residential homes (51.8%, n = 114) or with their families (35.7%, n = 79). Of the remaining sample, seven were in-patients (3.2%), five (2.2%) lived independently and others were in adult or other placements.

Case-note analysis revealed a total of 241 treatment episodes. Antidepressants were used for depression in 147 episodes (61.3%), for generalised anxiety disorder in 24 episodes (9.9%) and for obsessive-compulsive disorder in 22 episodes (9.1%). The full range of indicators for the use of antidepressants is presented in Table 1.

Table 1 Diagnoses that prompted the use of antidepressants in the study sample

Diagnosis	Treatment episodes,a n	Total episodes, %
Depression	147	61.3
Generalised anxiety disorder	24	9.9
Obsessive–compulsive disorder	22	9.1
Mixed anxiety and depressive disorder	11	4.6
Bipolar disorder	9	3.8
Behaviour disorder	8	3
Insomnia	4	1.7
Premenstrual tension	3	1.3
Phobias	2	0.8
Personality disorder	2	0.8
Schizoaffective disorder	2	0.8
Panic disorder	1	0.4
Post-traumatic stress disorder	1	0.4
Adjustment disorder	1	0.4
Atypical grief	1	0.4
Hypochondriasis	1	0.4
Mood swings	1	0.4
Schizophrenia	1	0.4

Out of 221 service users, 41 (24%) had a history of epilepsy; 21 had active epilepsy and 20 had not had seizures for 1 year or longer. Of those who had active epilepsy, two-thirds were having fewer than 12 seizures per year and a third were experiencing more than 12 seizures per year. Challenging behaviour was present in 96 (43.4%) of the sample.

Medication used

Selective serotonin reuptake inhibitors (SSRIs) were found to be the most commonly used antidepressants (77%) (Table 2). Mirtazapine and trazodone were used in 8.3% and 5.8% of the total treatment episodes respectively; tricyclics were noted to be used in only 5% of treatment episodes; serotonin noradrenaline reuptake inhibitors (venlafaxine and duloxetine) were used in 4% of episodes. Among the SSRIs, citalopram stood out unchallenged as the most favoured antidepressant (43.2% of the episodes and 56% of the SSRIs use). No monoamine oxidase inhibitors had been used.

Table 2 Types of antidepressants used in the study sample and average dosage

Antidepressant	Frequency	Per cent	Average dose, mg (range)
Citalopram	104	43.2	18.7 (8–40)
Fluoxetine	43	17.8	20.6 (10–60)
Escitalopram	21	8.7	11.10 (5–20)
Mirtazapine	20	8.3	24.20 (15–45)
Trazodone	14	5.8	122.8 (50–400)
Paroxetine	9	3.7	22.5 (10–40)
Venlafaxine	8	3.3	99.2 (75–150)
Sertraline	7	2.9	52.5 (25–100)
Dothiepin	7	2.9	140.5 (25–175)
Clomipramine	2	0.8	69 (10–75)
Duloxetine	2	0.8	30 (30–30)
Fluvoxamine	1	0.4	37.5 (25–50)
Imipramine	1	0.4	25
Amitriptyline	1	0.4	30
Nortriptyline	1	0.4	10

Co-prescription occurred with the majority of the treatment episodes with antidepressants - only 52 treatment episodes (21.5%) were not associated with co-prescription. Antipsychotics (36%), antiepileptic drugs (22.8%), benzodiazepines (14.5%), mood stabilisers (8.3%) and proton pump inhibitors (7.5%) were noted.

Measuring treatment outcome

The outcome of the antidepressant use was noted in terms of clinical improvement, retention or discontinuation and side-effects at follow-up at 6 weeks, 6 months and 12 months (Fig. 1). At 6 weeks, 119 treatment episodes (49.3%) had a positive outcome, with significant improvement in 115 cases and with the patient being described as ‘back to normal’ in four cases. At 6 months, 116 episodes (48.1%) had shown clinical improvement, 109 significant improvement and 7 back to normal. At 12 months, 117 episodes (48.5%) had positive outcomes, with significant improvement in 97 and back to normal in 20 cases. Retention rates were 218 (90.4%), 185 (76.7%) and 159 (65.9%) at 6 weeks, 6 months and 12 months respectively. There was a trend towards a gradual increase in discontinuation of antidepressants over the 1-year follow-up.

Fig 1 Clinical improvement at 6 weeks, 6 months and 12 months' follow-up.

Out of 59 episodes of discontinuation noted, 20 (33.9%) were due to switch to another antidepressant, 14 (23.7%) due to non-response, 12 (20.3%) due to side-effects and 3 (5.1%) due to non-adherence. No reason had been recorded for 10 episodes (16.9%) of discontinuation.

Only 29 episodes of side-effects had been recorded over the 1-year follow-up.

Antidepressant side-effects recorded in the sample

The use of SSRIs was associated with most of the adverse effects noted, reflecting the more prevalent use of these compounds. Four episodes each of sedation, stimulation, weight gain and gastrointestinal symptoms had been recorded, as well as three episodes of increased seizure activity and two episodes of increased anxiety. Citalopram, the most commonly used antidepressant, was associated with 15 episodes of side-effects viz. three episodes of sedation, two episodes of nausea and vomiting, two episodes each of hypomania and increased seizure, and one episode each of nausea, stimulation, hypotension, weight gain, increased anxiety and urinary symptoms. Of other antidepressants used, fluoxetine was responsible for two episodes of stimulation and one episode of increased anxiety; paroxetine was associated with one episode each of stimulation and sweating; dothiepin was associated with an episode of dry mouth and tremors; mirtazapine caused weight gain and sedation in one episode each; and trazodone was associated with weight gain on one occasion.

Study strengths and limitations

The difficulty in carrying out, and an extreme lack of, randomised controlled trials (RCTs) in individuals with intellectual disability is well recognised. Apart from ethical and consent issues, common confounding factors such as polypharmacy and presence of physical conditions such as epilepsy make RCTs expensive, time consuming and often challenging in terms of recruitment. ^{Reference Lennox, Taylor, Rey-Conde, Bain, Purdie and Boyle10}

Data from alternative sources are therefore needed. Our study has the strength of being naturalistic and describing current clinical practice. However, because of the nature of the study - retrospective case-note analysis - it has limitations. As expected, the biggest problem was found to be related to the record keeping and missing data.

Study findings

The results reveal that antidepressants were commonly used in individuals with intellectual disability, both adults and elderly (14 service users (6.3%) were aged above 65 years) for a range of conditions. Comorbidity was common, with the prevalence of epilepsy in our study group similar to that in individuals with intellectual disability in general, ^{Reference Morgan, Baxter and Kerr11} but the prevalence of challenging behaviour was found to be higher in the study sample. ^{Reference Emerson, Kiernan, Alborz, Reeves, Mason and Swarbrick12}

The majority of those receiving antidepressants had mild to moderate intellectual disability, which begs the question whether we are missing depression and anxiety disorders in service users with severe disability.

The recorded clinical outcome was good and even if the missing data were to be considered as treatment failures, nearly half the sample improved while on antidepressant treatment over the 1-year follow-up period. Of course, the study could not say whether this was drug effect or natural history of the disease. Furthermore, we lacked information on important comorbidity such as the presence of autism-spectrum disorder. Information on differential prescribing and/or outcome in this group would have been valuable and merits future research.

Recording of side-effects seemed inconsistent and clinicians should consider more structured recording as we are likely to be missing these important data. Nonetheless, it was found that the majority of the service users received low average dosage of antidepressants. This may suggest that clinicians are cautious about the dosage and may in particular be concerned about potential side-effects.

The study shows that antidepressants are commonly prescribed in adults with intellectual disability. Approximately half of our sample did well in terms of clinical improvement. However, the lack of RCT data in individuals with intellectual disability and mental health problems is a concern.