Misalignments of rest–activity rhythms in inpatients with schizophrenia

Authors


  • Field: Neurophysiology and psychophysiology

*Manami Kodaka, PhD, Department of Psychogeriatrics, National Institute of Mental Health, National Center of Neurology and Psychiatry, 4-1-1 Ogawahigashi-machi, Kodaira, Tokyo 187-8553, Japan. Email: mkodaka@ncnp.go.jp

Abstract

Aims:  Rest–activity rhythms of human beings generally synchronize to a 24-h time cue. Very few detailed research studies have examined rest–activity rhythms in patients with schizophrenia. The present study aimed to explore (i) rest–activity rhythms in patients with schizophrenia, and (ii) factors relevant to their rhythm characteristics.

Methods:  We selected only inpatients for this research, because the time cue for inpatients was considered more standardized than that of outpatients. Sixteen inpatients with schizophrenia wore an ActiTrac accelerometer-based activity monitor (IM Systems Inc., Baltimore, USA) for eight consecutive days to measure their activity. We used a χ2 periodogram to compute rest–activity rhythms from the activity data, whereby the χ2 value amplitude was regarded as an index of regularity. We conducted non-parametric tests to identify factors relevant to rhythm cycles and patterns.

Results:  Half of the participants exhibited prolonged rest–activity cycles, and 25% also had irregular rest–activity patterns defined by insufficient χ2 value amplitude, even though they were clearly under a 24-h time cue. Participants with misaligned rest–activity rhythms had attended daytime non-medical treatment programs less frequently, and had received more anti-anxiety/hypnotic medications than those with proper rhythms.

Conclusion:  Changes in rest–activity rhythms by optimizing pharmacological and non-pharmacological treatment could improve social adjustment or quality of life in patients with schizophrenia.

FOR INDIVIDUALS WITH schizophrenia, improving social adjustment is a major challenge. In the clinical setting, mental health practitioners have met many patients whose impaired social adjustment seemed to be associated with misaligned rest–activity rhythms. While some studies have observed misaligned rhythms, very few research studies have examined rest–activity rhythms in patients with schizophrenia in great detail.

Several previous studies investigated the relationship between schizophrenia and misaligned circadian rest–activity rhythms. Motor activity levels in non-medicated patients with schizophrenia were significantly lower than those observed in healthy subjects.1 Older patients with schizophrenia stayed in bed longer, reported lower quality night-time sleep, slept more during the day, and had less robust circadian rhythms than age- and gender-matched comparison subjects without psychiatric disorders.2 Wulff et al.3 reported a case study of a patient with schizophrenia for whom the period of rest–activity rhythm was longer than 24 h. Patients with higher cognitive function and milder psychiatric symptoms were reportedly more alert during the day and had fewer awakenings during the night.4 Boivin5 also published a review of studies, which reported misaligned sleep–wake rhythms and endocrine disturbances in schizophrenic patients.

Other studies have examined how medication might influence rest–activity rhythms in patients with schizophrenia. For non-medicated patients with lower activity levels than healthy subjects, administration of antipsychotics further lowered the amount of activity.1 Patients on risperidone experienced longer sleep periods during the day and night compared to patients on haloperidol.1 Finally, patients treated with clozapine had well-ordered rest–activity rhythms compared to those treated with typical neuroleptics.6

A particular aim of the present study was to ascertain whether any patients might exhibit misaligned rest–activity rhythms in a controlled environment. We therefore selected only inpatients for this research, as inpatient time cues are generally more standardized than that of outpatients. We conducted a case series study with our recruited cohort of inpatients with schizophrenia, and determined (i) rest–activity rhythms of patients with schizophrenia, and (ii) factors relevant to their rhythm characteristics.

METHODS

Participants

All participants were patients with schizophrenia hospitalized in a psychiatric unit. We selected inpatients as study participants to improve the validity of the study, as their waking, sleeping, and eating times were regulated in the inpatient ward. We surmised that this regulation would create less variation in time cues influencing rest–activity rhythms than if we had selected people living outside the hospital.

Participant eligibility criteria comprised the following:

  • 1diagnosis of schizophrenia was based on the DSM-IV;
  • 2the patient received inpatient treatment at a psychiatric hospital;
  • 3the patient had a history of schizophrenia longer than 10 years;
  • 4they were over 20 years of age;
  • 5they had been hospitalized for more than one month, and were considered to be stable by an attending psychiatrist;
  • 6they had not changed their main antipsychotic drugs or sleep inducers for at least one month prior to study initiation; and
  • 7they were able to understand the study description and provide informed consent.

Exclusion criteria were as follows:

  • 1the patient had major neuropsychiatric or physical complications; and
  • 2study participation was deemed undesirable by an attending psychiatrist.

A total of 21 patients agreed to participate in the present study. Four retracted their consent during the study, and one did not complete the rhythm measurements due to an anticipated change in hospitals. The final analysis included data collected from 16 patients.

Demographic and clinical characteristics of participants

Table 1 presents demographic and clinical characteristics of participants. Ten participants were on atypical antipsychotics (quetiapine fumarate, risperidone, olanzapine, aripiprazole, or combinations of these), three were on typical antipsychotics (haloperidol), and three were on combinations of atypical (quetiapine or risperidone) and typical (haloperidol) antipsychotics. No participants were being treated for illnesses other than psychiatric illness. One participant stayed out of the hospital overnight during the measurement period.

Table 1.  Demographic and clinical characteristics (n = 16)
 MedianMin.Max.
  1. BPRS, Brief Psychiatric Rating Scale; CP, chlorpromazine; PSQI, Pittsburgh Sleep Quality Index.

Age (years)55.53073
Number of years since the first visit31.0456
BPRS scores (total)63.04873
BPRS scores (positive symptoms)12.0815
BPRS scores (negative symptoms)12.01015
PSQI scores5.0313
CP equivalencies (mg)1050.0100.02967.0
Biperiden equivalencies (mg)2.809.0
Diazepam equivalencies (mg)7.5023.3
Daytime non-medical treatment (times/month)5.0112
Morning exercise (times/month)4.0030
Caffeinated drinks (cups/day)4.3011
 n% 
Sex, male850.0 
Smokers637.5 
Types of schizophrenia (DSM-IV)   
 Unspecified16.2 
 Catatonic16.2 
 Disorganized318.8 
 Paranoid1168.8 

Ethics and informed consent

All patients received a detailed written explanation of the study. After researchers explained the study information to them and guaranteed preservation of their anonymity, they were asked to provide their informed consent and acknowledge their adequate understanding of the study. This study was approved by the ethics committee at Showa University School of Medicine, Japan and conducted in accordance with the provisions of the 1995 version of the Declaration of Helsinki (as revised in 2004).

Measurement of rest–activity rhythms

An ActiTrac accelerometer-based activity monitor (IM Systems Inc., Baltimore, USA)7 was used to measure participant activity. Participants wore the ActiTrac on the non-dominant wrist for eight consecutive days and removed it only to bathe. Participant rest–activity rhythms were then computed from the obtained activity data.

Other measurements

Trained psychiatrists conducted the clinical diagnosis of participants based on the DSM-IV. Each participant's attending psychiatrist evaluated psychiatric symptom severity using the Brief Psychiatric Rating Scale (BPRS)8 (the Japanese version of BPRS9). Participants assessed their subjective sleep quality according to the Pittsburgh Sleep Quality Index (PSQI)10 (the Japanese version of PSQI11). We collected medical information for each participant from their medical charts, and demographic data and information pertaining to the number of daytime non-medical treatment and morning exercise programs attended per month from their nursing records. Chlorpromazine (CP) equivalencies,12,13 biperiden equivalencies,13 and diazepam equivalencies13 were calculated for doses of antipsychotic drugs taken daily by participants. Participants also provided information regarding the number of cigarettes they smoked and caffeinated drinks they consumed per day. Participant smoking or ingestion of caffeinated drinks was not regulated during the study.

Data analysis

The χ2 periodogram14 was applied to activity data to compute rest–activity rhythms. Amplitude of the χ2 value was regarded as an index of regularity, and if it was below 200, we operationally assumed that there was no regular pattern. As rest–activity cycles are generally synchronized to 24 h, those longer than 24 h were considered prolonged.

We used the Mann–Whitney U-test and the Kruskal–Wallis H-test to examine variables associated with rhythm characteristics. The Mann–Whitney U-test compared the groups with proper versus prolonged cycles and the groups with regular versus irregular patterns. The Kruskal–Wallis H-test compared three groups: (i) participants with proper rest–activity circadian rhythms (‘Group 1,’n = 8); (ii) those with regular patterns but with prolonged cycles (‘Group 2,’n = 4); and (iii) those with prolonged cycles as well as irregular patterns (‘Group 3,’n = 4). We compared median age, number of years since the first visit, severity of symptoms, quality of sleep, antipsychotic doses, number of daytime non-medical treatment programs and exercise programs attended per month, and the amount of caffeinated drinks taken per day. We used the Mann–Whitney U-test with a Bonferroni correction as a post-hoc multiple comparison test to analyze the results from the Kruskal–Wallis H-test. A P-value < 0.017 was considered statistically significant. The χ2-test was used to examine the correlations between rhythm characteristics and participant sex or cigarette smoking. We employed P < 0.05 as a significance level for a two-tailed test. spss (ver.16.0; spss Inc., Chicago, IL, USA) statistical package was used for data analyses.

RESULTS

Descriptive data of rest–activity rhythms

χ2 periodogram analysis revealed that half of the participants exhibited prolonged (longer than 24 h) rest–activity cycles. Of the eight with prolonged cycles, four (25% of the total) also had irregular rest–activity patterns defined by insufficient χ2 value amplitudes. Figure 1 shows examples of rest–activity rhythm profiles of participants with (a) proper circadian rhythms, (b) regular rest–activity patterns but prolonged cycles, and (c) prolonged cycles and irregular rest–activity patterns.

Figure 1.

Rest–activity rhythm profiles. Examples of rest–activity rhythm profiles computed by χ2 periodogram for participants with (a) proper circadian rhythms, (b) regular rest–activity patterns but prolonged cycles, and (c) prolonged cycles and irregular rest–activity patterns.

Correlations between rhythm characteristics and other variables

Table 2 presents differences in the 12 variables between the groups with proper versus prolonged rest–activity cycles. The median diazepam-equivalent dose was significantly higher among participants with prolonged cycles than those with 24-h cycles. Patients with prolonged cycles had attended significantly fewer non-medical treatments than those with proper cycles. We observed no significant differences in other variables between these two groups. Rest–activity cycles did not differ significantly by sex (χ= 4.0, d.f. = 1, P = 0.05). Of the eight participants with proper cycles, two were male (25%), while six (75%) of the eight with prolonged cycles were male. We observed no significant differences in rest–activity cycles between smokers and non-smokers either (χ= 0, d.f. = 1, P = 1.00). Three (37.5%) of the eight with proper cycles were smokers, while three (37.5%) of the eight with prolonged cycles were smokers.

Table 2.  Factors relevant to rest–activity cycles
 Proper cycle
n = 8
Prolonged cycle
n = 8
UP
Median
  1. Mann–Whitney U-test. BPRS, Brief Psychiatric Rating Scale; CP, chlorpromazine; PSQI, Pittsburgh Sleep Quality Index.

Age (years)555627.50.64
Number of years since the first visit34.52824.00.40
BPRS scores (total)636331.00.92
BPRS scores (positive symptoms)12.51230.50.87
BPRS scores (negative symptoms)121228.00.66
PSQI scores5.5525.50.48
CP equivalencies (mg)1537.595018.50.16
Biperiden equivalencies (mg)2.752.530.50.87
Diazepam equivalencies (mg)517.59.50.02
Daytime non-medical treatment (times/month)6.52.58.50.01
Morning exercise (times/month)14.51.516.50.10
Caffeinated drinks (cups/day)3.54.2528.00.67

Table 3 shows differences in the 12 variables between the groups with regular versus irregular rest–activity patterns. Median diazepam-equivalent dose was significantly higher among the participants with irregular patterns than those with regular patterns. Medians of other variables did not differ significantly between groups. We observed no significant differences in the regularity of rest–activity patterns between sexes (χ= 0, d.f. = 1, P = 1.00). Six (50%) of the twelve with regular patterns were male, and two (50%) of the four with irregular patterns were male. Rest–activity cycles did not differ significantly between smokers and non-smokers either (χ= 3.2, d.f. = 1, P = 0.07), in that six (50%) of the twelve with regular patterns were smokers, and none of the four with irregular patterns were smokers.

Table 3.  Factors relevant to rest–activity patterns
 Regular pattern
n = 12
Irregular pattern
n = 4
UP
Median
  1. Mann–Whitney U-test. BPRS, Brief Psychiatric Rating Scale; CP, chlorpromazine; PSQI, Pittsburgh Sleep Quality Index.

Age (years)5556.518.00.47
Number of years since the first visit3031.523.50.95
BPRS scores (total)636419.50.58
BPRS scores (positive symptoms)121223.50.95
BPRS scores (negative symptoms)121220.00.62
PSQI scores5416.00.32
CP equivalencies (mg)1087.577515.00.27
Biperiden equivalencies (mg)2.455.58.50.06
Diazepam equivalencies (mg)5205.00.02
Daytime non-medical treatment (times/month)5.52.58.00.05
Morning exercise (times/month)14.51.59.00.06
Caffeinated drinks (cups/day)4521.50.76

Finally, the Kruskal–Wallis H-test revealed that the median diazepam-equivalent dose and attendance at daytime non-medical treatment programs differed significantly among the three groups (χ= 7.172, d.f. = 2, P = 0.03; χ= 6.596, d.f. = 2, P = 0.04). According to the Mann–Whitney U-test results, the median diazepam-equivalent dose was significantly higher (= 1.00, P = 0.008) in Group 3 (median = 20.0 mg) than in Group 1 (median = 5.0 mg). Significant differences were not observed for median values of other variables.

DISCUSSION

Half of the participants had rest–activity cycles longer than 24 h, and a quarter had irregular rhythm patterns (amplitudes of χ2 values <200). Compared to those with 24-h cycles or those with regular rhythm patterns, participants with misaligned rest–activity rhythms had taken more anti-anxiety/hypnotic medications. They also attended daytime non-medical treatment programs less frequently.

In a previous study, Wirz-Justice et al. reported that participants with relatively moderate symptoms of schizophrenia (BPRS score: 27–64) were synchronized to the 24-h time cue.6 In the present study, half of the participants, whose symptoms were relatively severe (BPRS score: 50–70), exhibited misaligned rest–activity rhythms. It follows therefore that an association exists between psychiatric symptoms and misaligned rhythms. In the present study, however, BPRS scores were not significantly associated with rest–activity rhythms. This may be due to the narrow range and higher BPRS scores of participants in this study. A wider range of BPRS scores might reveal a significant association with rhythm patterns and cycles.

Our findings of irregular rest–activity patterns in patients with schizophrenia are consistent with those of previous studies. Wirz-Justice et al.6 reported that two out of seven patients had irregular rest–activity patterns. Other studies have reported on individuals with schizophrenia who had irregular rhythm patterns as measured by the number of daytime naps and the amount of interrupted nighttime sleep.2,4 The present study confirms that patients with schizophrenia possess misaligned rest–activity rhythms.

Participants with prolonged cycles or irregular patterns of rest–activity rhythms had attended daytime non-medical treatment programs less frequently than those with proper cycles or regular rhythm patterns. They had also taken more anti-anxiety/hypnotic medications than those with proper cycles or regular patterns. Although we cannot identify cause-and-effect correlations between rhythm characteristics and anti-anxiety/hypnotic medications or non-medical treatment programs, future studies should address these items for potential targets of intervention.

In addition to the possibility of treatments affecting rest–activity rhythms, we speculate that the disease itself might be associated with misaligned rhythms. This could be addressed in future studies by comparing rhythms exhibited by patients with schizophrenia with those of healthy control subjects. Moreover, due to the lack of a control group, we cannot conclude that our findings are particular to people with schizophrenia and not so for people with other mental disorders.

A serious limitation of this study was that the sample size was very small. Although we found no significant associations between rest–activity rhythms and CP equivalents, biperiden equivalents, morning exercise, sex, and cigarette smoking, the P-values for the significance of the correlations between these variables and rest–activity rhythms were approximately 0.1 or even close to 0.05. This might have been due to a type II error caused by the small sample size. Future studies should use a larger number of subjects in order to further examine possible factors relevant to rest–activity rhythms. A few noteworthy limitations of our study are that it was a cross-sectional study with a small number of participants; some patients had been on relatively high doses of antipsychotic medications; and most participants suffered from relatively severe symptoms.

One advantage of the present study is that we examined differences in measured variables between groups of people with different rest–activity rhythm characteristics. In addition, as all participants were hospitalized, their time cues were well-regulated.

We conclude that optimization of pharmacological and non-pharmacological treatments could lead to changes in rest–activity rhythms. These may, in turn, improve social adjustment or quality of life for patients with schizophrenia.

ACKNOWLEDGMENTS

We thank Dr Haruhisa Ohta and Dr Ryo Akita (Department of Psychiatry, Showa University Karasuyama Hospital), and all the staff of the Department of Nursing, Showa University Karasuyama Hospital, involved in this project, for their assistance.

Funding for this study was provided by the Grant-in-Aid for Scientific Research from the Ministry of Health, Labor and Welfare of Japan [H17-Shougai-Ippan-008].

Ancillary