Dont let sleep be a trial

Sleep medicine is an area of growing interest, but before you can treat a disorder you need to diagnose it. Here we learn why forty winks in a £1000 a night sleep clinic may not be the best way to discover symptoms

ALTHOUGH sleep medicine has developed over the last twenty or so years it is still a relatively new speciality. As an area of medicine, it is one of growing interest, both for companies wanting to reduce the side-effects of medication and for those wanting to address the large unmet need for safe and effective treatments – hypnotics, alerting compounds, etc. 

Sleep trials can be problematic for a number of different reasons. Firstly, there are over 89 sleep disorders listed by the International Classification of Sleep Disorders and these may have very different symptoms, to accurately characterise the type of sleep disorder needs specialist knowledge. Secondly, recording the symptoms can be very subjective as there are wide variations in perception and unfamiliar surroundings can affect sleep patterns.

Additionally, studies with patients can be time consuming, especially if the subjects are brought in one at a time to a specialist sleep clinic and demand for these resources is great. An alternative is to monitor the patients remotely and then collate and analyse the data in the sleep lab. This approach allows a single unit to run a trial across multiple recruitment sites.

The Restless Legs Syndrome (RLS) is common and characterised by an irresistible

About 80% of RLS sufferers have involuntary leg movements, which continue into, sleep. These leg movements are stereotypical and often last about 1-5 seconds and occur periodically every 10 – 90 seconds for much of the night. This is known as a periodic leg movements disorder (PLMD). 

RLS and PLMD are associated with low ferritin or can be caused by some drugs, and there is a high incidence in patients with end stage renal failure, Parkinson’s disease (PD) and other neurological diseases. The aetiology of RLS/PLMD is unclear although there is evidence suggesting a genetic component in primary disease and that low ferritin and neuropathy cause secondary RLS.

The role of iron is probably related to the dopamine reuptake system and it is of note that RLS symptoms worsen at the nadir of the dopamine circadian curve.

Dopaminergic agents are the most successful treatment for both primary and secondary RLS/PLMD, suggesting a problem with the dopamine system. However the primary disease form of RLS/PLMD is not associated with cell loss or neurodegenerative disease

The clinician’s observation of disease severity and the sufferer’s impressions are frequently used to describe pre-sleep motor restlessness, but these are very subjective.

To address this, in recent years the international RLS study group has developed diagnostic criteria1 and severity rating scales for the wake features of the RLS/PLMD continuum so as to standardise research and record efficacy outcomes in clinical trials.

To date, outcome measurements usually include a rating of symptoms or a surrogate measurement of movement conducted during a complex, slightly invasive, sleep study (polysomnography).

Electromyography (EMG) has been described as the gold standard measurement of periodic leg movements made during polysomnography. This involves an overnight study in a controlled sleep lab and detection of increased electrical activation of the anterior tibialis muscles. However depolarisation from this muscle is not always responsible for movement and leg movements can occur with no EMG activity from this particular muscle. EMG is, in this context, used as a surrogate measurement of movement.

Polysomnography is a multi-parametric study that is used to record in detail multiple biophysiological changes that occur in the human body when the person is asleep.  It is a very precise technique, but it is expensive (about £800-£1000 a night plus the cost of data analysis) and is only available in dedicated sleep clinics. These facilities are in great demand and access to polysomnography can be limited.

Scheduling volunteers recruited to clinical trials for polysomnography in hospitals is thus difficult and usually results in the need for many sites, or extended study periods. This has implicit cost and quality consequences for study sponsors and delays the drug development programme.

An alternative is the use of actimetry technology or movement detectors, which are emerging as a method of accurately measuring bi-lateral leg movements2 in a clinical trial. Unlike EMG, these measurements can be made in the home within the subjects usual routine and behaviours. These more naturalistic studies can be done when convenient to the subject, over three consecutive nights. This reduces the cost and burden of scheduling research and trials into the usual busy schedule of clinical departments and removes the “lab effect” which is a frequent complaint of subjects when sleeping away from home with many sensors and electrodes attached to them in the sleep laboratory.

Bi-lateral leg actigraphy is an acceptable alternative to polysomnography for the assessment of RLS and PLMD. Other actigraphic methods can also be used to demonstrate proof of concept or drug efficacy, or can be used to monitor side effects or adverse events such as motor restlessness or wake after sleep onset.

This large multi-centre study has demonstrated that a single laboratory can manage multiple home leg actigraphy assessments and that assessments can be scheduled within just a few days of patient enrolment. This presents significant practical and economic advantages over complex polysomnography and offers great potential as a clinical research tool.

Assessing the Feasibility of investigating involuntary leg movements in a home environment

Method
26 GP sites recruited adults with a diagnosis of primary RLS (normal ferritin and no neurological disease or use of drugs likely to induce RLS/PLMD) to a placebo controlled 12-week drug study.

 A sub group of patients in this double blind study were identified for home actigraphy assessment at baseline then at 6 weeks and 12 weeks. GPs agreed study nights and a planned bedtime with each patient then advised the sleep lab of these details.

Assessment packs were posted to the patient at baseline then 6 and 12 weeks into the study period.

Each home assessment pack included two actimeters (Actiwatch, Cambridge Neurotechnology Ltd / Respironics Inc) marked left leg and right leg, adhesive tape, user instructions and a simple card to log bedtimes and nocturnal behaviour.

Actiwatches were programmed to commence recording one hour before the planned bedtime specified by each patient. The devices were set to sample movements at 32 Hz and store a value of acceleration (integrated sum of 64 samples) every two seconds. The devices recorded for 12 hours from the user specified start time  (one hour before their planned bedtime) for three consecutive nights yielding 64,800 values of movement.

All assessment packs were posted to patients with a stamped self addressed envelope supplied for the patients to return completed assessments. Data was transferred from the Actiwatch via wireless communication to a standard desktop PC and dedicated software was used to apply an algorithm, which identified candidate PLMs. One person reviewed all patient records and applied PLM scoring criteria and data validation techniques as used in polysomnographic evaluation of PLM. A PLM index (PLMI), the number of PLMs /hour, was determined for unilateral and bi-lateral PLMs were recorded for three consecutive nights at baseline, week 6 and week 12.  

Results
94 patients were recruited by GPs within a 90-day enrolment period. All had baseline three-night home actigraphy study within five days of enrolment and 68 patients had repeat home actigraphy after a 12-week treatment phase. Two assessments were either lost in the post or not returned by the patients, one night of data was lost because of a battery failure and one Actiwatch failed due to a technical failure (<1%).

This low level of failures was exceeded by patients deviating from study protocol e.g. wearing the Actiwatches on the wrong night. Only 10 assessments were posted by special delivery.

Actiwatch data files were copied to the study sponsor and hard copy results and an excel spreadsheet were added to each patients case report file and the study file.

Discussion
There are published standards for the use of actimetry (actigraphy) in sleep, circadian and movement disorders3, and it is important to recognise the value and limitations of these measurements in the evaluation of new medicines.

The results from the study illustrated above show that Actiwatch is a robust device and that studies can be scheduled to start several days after the Actiwatch were programmed to record. Bilateral leg actigraphy over three consecutive nights eliminates error of unilateral measurement and offers valuable information about the night-to-night variability of involuntary leg movements as well as practical advantages (Table 2). Furthermore the objective measurement of PLMI (referenced to time in bed) may be a more sensitive measurement of both disease severity and change in severity compared to a clinician’s or the patients impression of sensory discomfort and motor activity in the wake and sleep state.  

Figure 2: Example of leg movements recorded overnight. The top panel shows all movements of the right leg within the 12-hour study period. The periods at the start and end of this period indicate the sepf reported time out of bed and are excluded from analysis. The lower panel is a 15-minute section taken at about midnight and shows periodic movements.