GPs have been less successful than their secondary care colleagues at discontinuing antipsychotic medication in elderly care home residents, psychiatrists in Leicester have found (Psychiatric Bull 2007; 31:459‐62).

Following warnings from the Committee on Safety of Medicines about an increased risk of adverse events with olanzapine and risperidone, the Royal College of Psychiatrists (RCP) recommended that olanzapine and risperidone should be gradually withdrawn in care home residents. Most patients with dementia in a community mental health team (CMHT) for older people in Leicestershire had their antipsychotic withdrawn (58 per cent) or substituted with a benzodiazepine (16 per cent). This follow‐up study reviewed progress in patients under the care of GPs.

Of 164 residents identified with dementia, 64 had taken an antipsychotic at some time since before the CSM issued its warning. Before the CSM advice, 69 per cent of prescribed antipsychotics were olanzapine or risperidone; afterwards, this dropped to 39 per cent. Of the 19 who continued with these drugs, 15 were under GP review only.

The RCP recommends setting a clear date for review. Fifteen patients had been reviewed less frequently than six‐monthly and, of these, 14 were under GP review. The authors say that patients who experience adverse effects may not be followed up unless they have a concurrent problem requiring GP review.

A record linkage study from The Netherlands has shown that GPs prescribe more benzodiazepines for patients with anxiety if they also have other problems (Family Practice 2007;24:538‐46).

Of 4604 patients with newly diagnosed anxiety, 58 per cent had some form of comorbidity. Compared with patients diagnosed with anxiety alone, those with chronic somatic conditions were more likely to receive a benzodiazepine (effect size 0.44) in the year following diagnosis. For those with psychiatric comorbidity (mostly depression), they were twice as likely to receive not only an antidepressant but also a benzodiazepine. In the case of social problems, patients with a family member who was ill were three times more likely to be prescribed a benzodiazepine. In general, prescribing of psychotropic drugs declined during the nine months after a diagnosis of anxiety but benzodiazepine use persisted.

Real‐life effectiveness is not predicted by randomised trials and pragmatic evaluations of antipsychotics suggest problems with treatment persistence, say pharmacists at the Maudsley Hospital in London.

They documented experience with aripiprazole in 228 patients treated between 2004 and 2005 in an acute mental health unit (Acta Psychiatr Scand 2007;116:461‐6). Of these, 112 (49 per cent) completed six months' treatment, with most discontinuations occurring within the first 12 weeks.

Persistence with treatment correlated with being an outpatient, older age and no prior use of clozapine. The main reasons for discontinuation were lack of effectiveness (39 per cent) and adverse effects (53 per cent), including nausea, extrapyramidal effects, agitation and anxiety. In particular, adverse effects accounted for more withdrawals in the first 60 days than thereafter. The authors say that close attention to early tolerability could improve this.

A meta‐analysis suggests that suicidal ideation is rare during treatment of children with ADHD with atomoxetine though, on the basis of few cases, significantly more frequent than with placebo (J Am Acad Child Adolesc Psychiatry 2008;47:209‐18).The study included 12 placebo‐controlled trials and five comparative trials with methylphenidate in a total of 3230 children with ADHD. Study duration ranged from 6 to 18 weeks.

There were no completed suicides. Suicidal ideation occurred in five children (all boys) taking atomoxetine and there was one nonfatal suicide attempt; there were no cases in children taking placebo. For a suicide‐related event, the number needed to harm with atomoxetine was 227 compared with a number needed to treat of five. The frequency of suicidal events with atomoxetine and methylphenidate were similar.

US psychiatrists say they have reported the first controlled, prospective, longitudinal study of the risk of relapse of bipolar disorder during pregnancy (Am J Psychiatry 2007;164:1817‐24).

Eighty‐nine pregnant women with bipolar disorder taking mood stabiliser therapy (or discontinuing therapy less than six months before conception) were followed up until 12 months after delivery. Among the 62 who discontinued treatment, 85.5 per cent experienced a recurrence and on average spent over 40 per cent of their pregnancy in an illness episode; the mean time to recurrence was nine weeks. Of those who continued treatment, 37 per cent had a recurrence and the mean duration of illness accounted for 9 per cent of their pregnancy; the median time to recurrence was over 41 weeks from the estimated date of conception compared with nine weeks when treatment was discontinued.

Depressive or dysphoric episodes were four times more common than manic or hypomanic symptoms overall. Gradual discontinuation was associated with a lower risk of recurrence than abruptly stopping treatment. Antidepressant use was a strong predictor of recurrence.

The alpha₂‐agonist clonidine, currently licensed for hypertension, has been used in the treatment of ADHD. In two papers (J Am Acad Child Adolesc Psychiatry 2008; 47:180‐8 & 189‐98), US investigators report a trial to determine its efficacy and safety.

A total of 122 children aged 7‐12 years with any form of ADHD were randomised to treatment with clonidine or methylphenidate, alone or in combination, or placebo. Doses were titrated over the first eight weeks then maintained for a further eight weeks.

The primary endpoint was the change in score in the Conners Abbreviated Symptom Questionnaire (ASQ) for Teachers. Either drug as monotherapy was superior to placebo but only transiently so for clonidine; the combination was the most effective, though not significantly better than methylphenidate alone. A secondary endpoint, ASQ for Parents, rated clonidine and the combination significantly more effective than methylphenidate.

Clonidine was said to be well tolerated, though it was associated with a higher incidence of bradycardia (18 vs 3 per cent) and more moderate or severe adverse events (79 vs 49 per cent). Children taking clonidine alone were also more likely to gain weight and have higher blood pressure; somnolence (42 vs 7 per cent with methylphenidate or placebo) and apathy (32 vs 14 and 17 per cent respectively) were more frequent. The frequency of these adverse effects was reduced with combination therapy.

Current opinion holds that patients with epilepsy are most likely to become seizure‐free with monotherapy. Neurologists in Germany suspected this did not accurately reflect their clinical experience, so they retrospectively reviewed the records of 500 patients attending their specialist clinic who had achieved seizure control and compared them with 321 patients who had not (Acta Neurologica Scand 2008;117:55‐9). In the 586 patients with symptomatic or cryptogenic epilepsy with focal or secondary generalised seizures, freedom from seizures was more likely among patients with a single seizure type than multiple types. Patients taking monotherapy were more likely to be seizure‐free than those taking two or three drugs (72 vs 57 per cent) but they also had a shorter history of epilepsy. No differences between drug treatments were reported.

Of 217 patients with idiopathic epilepsies, freedom from seizures was again more frequently associated with monotherapy (57 vs 43 per cent) and valproate was superior to lamotrigine (66 per cent seizure‐free vs 44 per cent).

The authors suggest that, in their patients with difficult‐tocontrol epilepsy, idiopathic generalised epilepsy is more frequently refractory than published studies suggest whereas combined therapy seems more effective than expected in symptomatic/cryptogenic epilepsy; no drug was clearly superior.

Epidemiological evidence shows that NSAID use is associated with a reduced risk of Alzheimer's disease (AD) but it is unclear whether APOE genotype influences the effect, or if NSAIDs that selectively lower Aβ₄₂ type of amyloid‐beta peptide confer greater protection.

US investigators have addressed these questions in their analysis of 3229 participants (aged 65 or older) in the Cardiovascular Health Cognition Study (Neurology 2008;70:17‐24). Of these, 37 per cent reported NSAID use, 60 per cent took aspirin at some time and 38 per cent reported taking paracetamol. After four to five years' follow‐up, there were a total of 452 cases of all‐cause dementia of which 231 cases were probable or possible AD and 199 were vascular dementia.

Use of NSAIDs, but not aspirin or paracetamol, was associated with a 24 per cent lower risk of all‐cause dementia and a 37 per cent lower risk of AD. The link between NSAID use and AD risk was not affected by NSAID type, dose or duration of use. Further analysis suggested that this effect was confined to individuals with one or more APOEε4 allele. However, there was no difference in risk associated with the NSAID's Aβ₄₂ lowering activity.

Prompted by the findings of small‐scale studies suggesting that galantamine may improve some aspects of cognitive function in patients with schizophrenia, US psychiatrists have reported a randomised placebo‐controlled trial in 86 patients (Am J Psychiatry 2008; 165:82‐9).

All but six participants were taking an atypical antipsychotic and the remainder were on low‐dose conventional antipsychotics; the mean final dose of galantamine was 23.5mg daily. After 12 weeks' treatment neither galantamine nor placebo significantly changed overall performance in a battery of neuropsychological tests compared with baseline (though there was significant heterogeneity between two tests, with one favouring an effect of galantamine on processing speed). Galantamine had little effect on symptoms with the exception of a reduction in symptoms of alogia.

Based on small studies, orlistat, topiramate, zonisamide and sibutramine are among the few drugs found to have clinically significant effects in the management of binge eating. A trial of sibutramine in 304 participants with binge eating disorders reporting an average of three binge days per week now provides more robust evidence (Am J Psychiatry 2008;165:51‐8).

Patients were randomised to sibutramine 15mg daily or placebo. Discontinuation rates were high (33 and 43 per cent respectively) with no clear differences in reasons why. Over 24 weeks, binge frequency decreased substantially in both groups (to less than one binge per week) though more with sibutramine (a reduction of 2.7 vs 2.0 binges per week). Sibutramine was also superior on secondary endpoints such as abstinence, weight loss (mean loss 4.3 vs 0.8kg) and global improvement. Adverse effects were more frequent with sibutramine and included headache, dry mouth, constipation, insomnia and dizziness.

The cytotoxic antibiotic mitoxantrone is effective as an induction agent in the treatment of aggressive relapsing‐remitting multiple sclerosis, a French study has shown (J Neurol Neurosurg Psychiatry 2008;79:52‐6).

One hundred consecutive patients were treated with mitoxantrone 20mg plus methylprednisolone 1g monthly for six months, after which 73 received maintenance therapy with mitoxantrone, interferon beta, azathioprine, methotrexate or glatiramer acetate.

In the 12 months following initiation of induction, the annual relapse rate was cut by 91 per cent, 78 per cent of patients were free of relapses and 64 per cent experienced reduced disability. During five years' follow‐up, the relapse rate remained low (around 30‐40 per cent) and the improvement in disability persisted. The risk of adverse effects was acceptable, with three cases of left ventricular dysfunction and one of acute myeloid leukaemia.

In the 1990s there was strong debate about the possible role of the mercury‐based preservative thiomersal in the pathogenesis of autism. It had been present in some vaccines for children and estimates of potential total exposure as part of the recommended immunisation schedule in the USA were shown by the US Food and Drug Administration (FDA) to exceeded safety standards. The amount of thiomersal in US vaccines was subsequently reduced to trace levels by 2001.

In 2004, an influential review rejected a causal link between thiomersal and autism. However, it acknowledged that epidemiological evidence of a temporal link between vaccine use and an increase in autism could only be refuted by monitoring the incidence of autism in the wake of discontinued exposure to thiomersal. That study has now been reported (Arch Gen Psychiatry 2008;65:19‐24).

Epidemiologists in California reviewed State records of autism diagnoses between 1995 and 2007. They found a sustained increase in prevalence across all age groups, with no evidence of variation at the times when the principal vaccines were reformulated. The authors suggest that earlier data purporting to show the opposite were based on incomplete case ascertainment.

A qualitative study in a Southampton general practice has identified some common themes to explain why patients continue or stop long‐term treatment with an SSRI (Family Practice 2007;24:570‐5).

Interviews with 17 patients (representing only one‐fifth of patients invited) revealed reluctance to begin treatment among many. Once started, some were uncertain about the benefits of continuing whereas others were fearful of the consequences of stopping or of withdrawal symptoms. Some spoke of how they valued the GP's input but others said their treatment was not frequently reviewed and they felt alienated by receiving medication on repeat prescription. The authors say that patients want regular review and they want to discuss the merits of continuing treatment with their doctor. GPs should be proactive, not wait to be contacted by the patient.

The first data from the quetiapine prolonged release (XL) major depressive disorder (MDD) and generalised anxiety disorder (GAD) clinical development programmes were presented at the 7th International Forum on Mood and Anxiety Disorders (IFMAD) in Budapest in December 2007.

All studies were randomised and double‐blind and the first study involved 612 patients with MDD. After six weeks' treatment, scores on the Montgomery‐Åsberg Depression Rating Scale (MADRS)were reduced for both the 150mg daily dose (−14.81; p<0.001) and the 300mg daily dose (−15.29; p<0.001) versus placebo (−11.18; p<0.001). Significant improvement was seen within eight days of starting treatment (p<0.01) compared with placebo. In an active‐control arm, patients who received duloxetine (60mg daily) had significant improvement at week six, but no significant response by day eight (p=0.3).

The second study (Phase III), also in patients with MDD, examined quetiapine XL as add‐on to antidepressants. After six weeks of treatment, MADRS scores were significantly reduced for both the 150mg daily dose (−15.26; p<0.01) and the 300mg daily dose (−14.94; p<0.01) versus placebo (−12.21). Onset of action was rapid, with significant differences from placebo apparent from week one (p<0.001).

The third study (Phase III), in patients with GAD, quetiapine XL monotherapy significantly reduced Hamilton Anxiety scale (HAM‐A) scores after eight weeks' treatment at doses of 50mg daily (−13.95, p<0.05) and 150mg daily (−15.96, p<0.001) compared with placebo (12.30). Statistically significant separation from placebo in the HAM‐A total score was evident by day four with quetiapine XL 50mg daily (p<0.001) and 150mg daily (p<0.05). Patients who received paroxetine (20mg daily) in an active‐control arm of the trial had significant improvement at week eight (p<0.01), but not at day four. Copyright © 2008 Wiley Interface Ltd