Article Text

Sleep benefit in Parkinson's disease
  1. Universitätsklinik für Neurologie, Anichstrasse 35, A-6020 Innsbruck, Austria
  2. Centro Neurológico, Hospital Francés, La Rioja 951, 1221 Buenos Aires, Argentina
  1. Dr Birgit Högl birgit.ho{at}
  1. Universitätsklinik für Neurologie, Anichstrasse 35, A-6020 Innsbruck, Austria
  2. Centro Neurológico, Hospital Francés, La Rioja 951, 1221 Buenos Aires, Argentina
  1. Dr Birgit Högl birgit.ho{at}
  1. Royal United Hospital, Combe Park, Bath BA1 3NG, UK

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    We read with great interest the report by Batemanet al on sleep benefit in Parkinson's disease.1 The authors state that no objective study has been undertaken about sleep benefit. We must object to that, as our group has recently published an extensive objective study about sleep benefit, which includes objective motor examinations, levodopa plasma concentration determinations, and polysomnographies.2Moreover, we would like to briefly discuss some of the authors′ findings in the light of our own results.

    The authors studied 20 patients with the motor part of the UPDRS, at waking (apparently twice), and after medication during “on” and “off”. At each rating, the patients completed an “activities of daily living” (ADL) questionnaire. The authors also administered ADL questionnaires to heterogenous groups of outpatients.

    In their results, the authors comment on six out of 16 patients of the first series: “When they awoke these patients performed as well as when they were “on” due to medication”. Unfortunately, they give no data at all, such as mean UPDRS scores and ranges, for baseline state and “on” time schedules and hours of the ratings. Furthermore, no information is provided to account for the remaining four patients who were included in the first series but do not appear in the results.

    The findings by Bateman et al contrast with our own results, in that patients with sleep benefit performed only slightly better in the morning compared with those without. A clear “on” compared with baseline was found in our study both in patients with and without sleep benefit after intake of their regular medication. We concluded from our data that sleep benefit was much smaller than expected. A morning baseline function as good as a drug induced “on”, as described by Bateman et al, would be indeed a sleep benefit of considerable magnitude. On the other hand, a drug induced “on” similar to the morning baseline state could theoretically also point to an insufficiently drug treated patient group. In any case, as no data are given to allow comparisons, one is confined to speculate why sleep benefit could be so much greater in British patients than the Argentinian population, where sleep benefit, although objectively existing, was quite a subtle phenomenon.

    The authors further state that a “strong correlation” was found between ADL and UPDRS, and conclude that ADL may serve as a “more objective” instrument to measure sleep benefit. Unfortunately they do not indicate if the correlation was found at any point in time or if all evaluations were lumped together, as no correlation index or graph is given.

    In the second sample of the study, 113 patients completed an ADL questionnaire at three points in time (at waking, best, and worst) before any drug intake. This was done at home. The authors determined that sleep benefit was present when the mean ADL score difference between best and worst was more than 12—that is, when strong variations occurred in baseline score before medication. The validity of this arbitrary definition deserves some discussion. Firstly, to take this variation as a criterion for sleep benefit may lead to a confusion with motor fluctuations. As the ADL score has a maximum of 52 points, an absolute score difference of 12 as a prerequisite for sleep benefit will lead to the exclusion of patients with smaller fluctuations irrespective of sleep benefit. So their own definition could have biased the authors' finding that patients with sleep benefit had a younger disease onset, longer disease duration, and more frequent use of bromocriptine. All this might also occur in a fluctuating subgroup of patients and an association of sleep benefit with fluctuations has been previously described.3

    Secondly, it is necessary to be cautious in considering ADL questionnaires as an objective measure to determine the presence of sleep benefit. The patient rating is based on how well he thinks he could perform at a given moment, and as we pointed out in our study, a large difference between self perceived motor function and objective motor function may occur in sleep benefit.

    Finally, we would like to add a word of caution and remind the authors that it would be wise to avoid deriving the measure of “objective duration” of sleep benefit from three scales filled in at home by patients without any further instructions than to fill them at waking, and during best and worst before drug intake.

    As the only dopamine agonist mentioned in this study is bromocriptine, we would be grateful to know what year the study was conducted in.

    In any case, although “little is known about sleep benefit”1 any study concerning this phenomenon should certainly attempt to increase knowledge and avoid a further increment of confusion.


    Bateman replies:

    The difference in results is due to a different definition of sleep benefit. Sleep benefit as defined in our paper refers to mobility as good as “on” on waking, which wears off over a variable period. Högl et al define sleep benefit as “self perceived mobility in the morning before drug intake as better than during the rest of the day.”1-1

    The purpose of the first part of our study was to verify the existence of sleep benefit as we had defined it, particularly in view of the findings of Högl et al that “patients with sleep benefit had a small improvement between night and morning” and “sleep benefit patients were clearly in the “off” state during baseline motor examination”. Our inpatient study showed that six out of the 16 “who were studied from the moment of waking” performed as well as when they were “on” due to medication. Subsequently they spontaneously turned “off” to an identical state to “off” after medication. Four patients could not be studied from the moment of waking as they awoke before the investigator! We clearly found that sleep benefit is as good as “on” after medication and wears off, not an intermediate stable state between “on” and “off” as Högl et al have defined it, by subsequently giving these patients their normal medication on the same day and monitoring their response by half hourly UPDRS and ADL scores.

    Aware of the patient's misperceptions about sleep benefit, we wished to confirm as objectively as possible the findings from our outpatient questionnaire by using the ADL rating scales. The inpatient study showed a correlation between motor UPDRS “on” score and ADL “on” score of r=0.72,t=4.42, p<0.001 and sleep benefit ratingsr=0.62, t=3.35, p<0.01. The ADL maximum score is 125. It consists of 25 items that can be rated on a five point scale. Dr P Brown, originator of the scale, suggested that a change of 12 would be sufficient to confirm sleep benefit. As we were aware that sleep benefit, confirmed by personal observation, can represent a substantial change in motor performance, this criterion seemed reasonable. The correlation between motor UPDRS and ADL scores in our study was good, showing that the ADL scores are generally a reliable measure, although there will inevitably be exceptions.

    Our study showed that sleep benefit as we defined it was generally a feature of patients with young onset Parkinson's disease. A 73 year old patient described in their paper, with disease onset at 62, would be unlikely to have sleep benefit as we defined it. Their paper, as their figure 2 shows, refers to a different phenomenon.


    1. 1-1.