Atanu Kumar Pati a; Arti Parganiha2007

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [INFLIBNET India Order]On: 7 September 2009Access details: Access Details: [subscription number 909277340]Publisher Informa HealthcareInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Chronobiology InternationalPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713597233

Alterations of the Characteristics of the Circadian Rest-Activity Rhythm ofCancer In-PatientsAtanu Kumar Pati a; Arti Parganiha ab; Anjana Kar a; Rakesh Soni a; Sushmita Roy c; Vivek Choudhary c

a School of Life Sciences, Pt. Ravishankar Shukla University, Raipur, India b INSERM U776, RythmesBiologiques et Cancers, Hôpital Paul Brousse, Villejuif, France c Regional Cancer Center, Pt. JawaharlalNehru Medical College, Raipur, India

Online Publication Date: 01 November 2007

To cite this Article Pati, Atanu Kumar, Parganiha, Arti, Kar, Anjana, Soni, Rakesh, Roy, Sushmita and Choudhary,Vivek(2007)'Alterations of the Characteristics of the Circadian Rest-Activity Rhythm of Cancer In-Patients',ChronobiologyInternational,24:6,1179 — 1197

To link to this Article: DOI: 10.1080/07420520701800868

URL: http://dx.doi.org/10.1080/07420520701800868

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

http://www.informaworld.com/smpp/title~content=t713597233

http://dx.doi.org/10.1080/07420520701800868

http://www.informaworld.com/terms-and-conditions-of-access.pdf

ALTERATIONS OF THE CHARACTERISTICS OF THE CIRCADIAN

REST-ACTIVITY RHYTHM OF CANCER IN-PATIENTS

Atanu Kumar Pati,1 Arti Parganiha,1,2 Anjana Kar,1 Rakesh Soni,1

Sushmita Roy,3 and Vivek Choudhary3

1School of Life Sciences, Pt. Ravishankar Shukla University, Raipur, India2INSERM U776, Rythmes Biologiques et Cancers, Hopital Paul Brousse, Villejuif, France3Regional Cancer Center, Pt. Jawaharlal Nehru Medical College, Raipur, India

The aim of the present study was to evaluate the characteristics of the circadian rest-activity rhythm of cancer patients. Thirty-one in-patients, consisting of 19 males and12 females, were randomly selected from the Regional Cancer Center, PanditJawaharlal Nehru Medical College, Raipur, India. The rest-activity rhythm wasstudied non-invasively by wrist actigraphy, and compared with 35 age-matched appar-ently healthy subjects (22 males and 13 females). All subjects wore an Actiwatch (AW64,Mini Mitter Co. Inc., USA) for at least 4–7 consecutive days. Fifteen-second epochlength was selected for gathering actigraphy data. In addition, several sleep par-ameters, such as time in bed, assumed sleep, actual sleep time, actual wake time,sleep efficiency, sleep latency, sleep bouts, wake bouts, and fragmentation index,were also recorded. Data were analyzed using several statistical techniques, such ascosinor rhythmometry, spectral analysis, ANOVA, Duncan’s multiple-range test, andt-test. Dichotomy index (I , O) and autocorrelation coefficient (r24) were also com-puted. The results validated a statistically significant circadian rhythm in rest-activitywith a prominent period of 24 h for most cancer patients and control subjects.Results of this study further revealed that cancer patients do experience a drasticalteration in the circadian rest-activity rhythm parameters. Both the dichotomyindex and r24 declined in the group of cancer patients. The occurrence of the peak(acrophase, Ø) of the rest-activity rhythm was earlier ( p , 0.001) in cancer patientsthan age- and gender-matched control subjects. Results of sleep parameters revealedthat cancer patients spent longer time in bed, had longer assumed and actual sleepdurations, and a greater number of sleep and wake bouts compared to control subjects.Further, nap frequency, total nap duration, average nap, and total nap duration per1 h awake span were statistically significantly higher in cancer patients than controlsubjects. In conclusion, the results of the present study document the disruption ofthe circadian rhythm in rest-activity of cancer in-patients, with a dampening ofamplitude, lowering of mean level of activity, and phase advancement. These

Submitted December 15, 2006, Returned for revision January 9, 2007, Accepted October 25,2007

Address correspondence to Dr. Atanu Kumar Pati, Ph.D., F.N.A.Sc.School of Life Sciences, Pt.Ravishankar Shukla University, Raipur 492 010, India. E-mail: [email protected] or [email protected]

Chronobiology International, 24(6): 1179–1197, (2007)Copyright # Informa HealthcareISSN 0742-0528 print/1525-6073 onlineDOI: 10.1080/07420520701800868

1179

Downloaded By: [INFLIBNET India Order] At: 09:51 7 September 2009

alterations of the circadian rhythm characteristics could be attributed to disease,irrespective of variability due to gender, sites of cancer, and timings of therapies.These results might help in designing patient-specific chronotherapeutic protocols.(Author correspondence: [email protected] or [email protected])

Keywords Rest-activity pattern, Circadian rhythm, Head and neck cancer,Reproductive organ cancer, Dichotomy index

INTRODUCTION

The extent of the toxicity and efficacy of a number of anticancer drugs/agents applied in experimental animal models vary with the time of theiradministration (Levi, 2001, 2006). These phenomena have been attributedto the 24 h rhythms of both the host and tumor (Filipski et al., 2002).The clinical relevance of these findings has been amply validated incancer patients (Eriguchi et al., 2003; Kobayashi et al., 2002). However,inter-individual variations in the characteristics of circadian rhythmsamong human cancer patients make it extremely difficult to devise a gen-eralized therapeutic strategy. This is a reason why patient-specific therapyis considered more useful (Mormont & Levi, 2003). Furthermore, asuitable marker rhythm is necessary to ensure effective outcomes of thechronotherapeutic approach (Mormont et al., 2002).

Currently, in clinical chrono-oncology, wrist-actigraphy is used to vali-date the circadian rest-activity rhythm of patients (Chevalier et al., 2003;Levin et al., 2005; Mormont & Waterhouse, 2002). Moreover, thisrhythm has been previously investigated as a potential suitable markerof the internal circadian time organization (Chevalier et al., 2003;Mormont & Waterhouse, 2002). Pati et al. (2006) recently reported thatthe circadian rest-activity rhythm dampens in patients suffering fromhead and neck cancer. Although this information was the first from theIndian subcontinent, it was based on a small sample of patients. The ration-ale of the present study was to gather data on a greater number of cancerpatients from India and to test the hypothesis of the dampening ofcircadian rest-activity rhythm in cancer.

PATIENTS AND METHODS

Biographical and Clinical Characteristics of Subjects

Thirty-one cancer in-patients, 19 males and 12 females, were selectedrandomly from the Regional Cancer Centre, Pt. J.N. Medical College,Raipur, India. Thirteen patients were suffering from cancer in theregion of head and neck, eight in the reproductive organs, and the remain-ing ten in other organs. The performance status (PS) of the patients ranged

A. K. Pati et al.1180


from fully active (0) to completely disabled (4), according to the criteria ofthe Eastern Cooperative Oncology Group (ECOG; Oken et al., 1982). Theinitial evaluation included blood pressure and temperature measure-ments. During the period of this study, the in-patients received eitherintravenous chemotherapy (n ¼ 20), radiotherapy (n ¼ 7), or both(n ¼ 4) in the morning (n ¼ 27) or afternoon (n ¼ 4). Chemotherapy con-sisted of a combination of two or three of the following medications:5-fluorouracil (500–600 mg/m2), cisplatin (40–60 mg/m2), gemcitabine(1000 mg/m2), peclitaxel (180–200 mg/m2), doxorubicin (50–60 mg/m2), and carboplatin (450–600 mg/m2). Cobalt 60 (1.33 Mev) was usedas the energy for radiotherapy. In addition, several other medicationsand blood transfusions were prescribed to minimize treatment sideeffects, such as nausea, vomiting, diarrhea, and anemia. The biographicaland clinical characteristics of the in-patients and apparently healthy sub-jects, those who did not self-report any disease, are summarized inTable 1. This design and conduct of the study complied with the ethicalstandards of Chronobiology International (Touitou et al., 2006).

Assessment of the Characteristics of Circadian Rest-Activity

Rhythm

The circadian rest-activity rhythm was monitored non-invasively bywrist actigraphy (Actiwatch, Model AW64, Mini Mitter Co., Inc.) worn

TABLE 1 Biographical and Clinical Characteristics of Cancer In-patients and Control Subjects

Attribute Description

Number of patients (M/F) 31 (19/12)Age (yrs), median/range 43/17–75Site of cancer Head and neck (n ¼ 13), Reproductive organs

(n ¼ 8), or Other organs (n ¼ 10)Specific site of cancer Alveolus (3), Buccal mucosa (4), Supra glottis (1),

Esophagus (1), Larynx (2), Tongue (1), Lung (1),Cervix (3), Ovary (2), Penis (1), Vagina (1), Testis(1), Intestine (1), Liver (1), Rectum (2), Soft tis-sue (1), Thigh (3), Blood (2)

Metastasis cases 7Cancer stage 2 or 3 or 4ECOG performance status 4- PS (3), 3- PS (4), 2- PS (5), 1- PS (8), 0- PS (11)Treatment

Chemotherapy/radiotherapy/both 20/7/4Drugs used for chemotherapy 5-Fluorouracil, Cisplatin, Gemcitabine, Peclitaxel,

Doxorubicin, and Carboplatin.Source of radiation Cobalt 60Symptomatic drugs Granisetron, Paracitamol, FilgrastimCycles of chemotherapy/radiotherapy 1 (15), 2 (6), 3 (4), 6 (3)/20 (3)

Number of control subjects (M/F) 35 (22/13)Age (yrs), median/range 35/19–61

Circadian Rest-Activity Rhythm of Cancer Patients 1181


on the non-dominant arm over a continuous four day span, when thein-patients were admitted to receive either chemotherapy, radiotherapy,or both. Pre- and post-treatment wrist activity data could not be collectedbecause of the following reasons: only seriously ill patients are admitted tothe in-patient ward without prior notice; most of the patients come fromdistant and remote areas, are illiterate, and fail to furnish valid postaladdresses; and most of them do not have phone/email facilities at theirresidences. Thirty-five apparently healthy subjects, 22 males and 13females, also wore the Actiwatch for at least 4–7 consecutive days ontheir non-dominant arm and formed the control group. Both patientsand healthy subjects gave a written informed consent.

The actigraph consists of an accelerometer (a piezoelectric sensordevice capable of converting movement into measurable electricalsignals), a filter that excludes external vibrations by selectively registeringvalues only in the 1.0 to 2.5 Hz range, a programmable memory into whichit stores the resulting values, and a user interface program that allows forits configuration. Data were collected at 15 sec epochs. The duration ofmonitoring was in compliance with the recommendations of the AmericanAcademy of Sleep Medicine (1997) and an American Academy of SleepMedicine Report (Littner et al., 2003).

Study of Sleep Parameters

The following sleep parameters were also studied in both groups asdetermined by the application of the Actiware-Sleep Software (MiniMitter Co., Inc.) to the data:

. Time in Bed (TIB). The amount of time spent in bed, derived as the differ-ence between the get up time and bedtime.

. Assumed Sleep (AS). The difference in time between the Sleep End andSleep Start times. This parameter was calculated automatically usingvalues derived from the sleep scoring algorithm of the Actiware-Sleepsoftware.

. Actual Sleep Time (AST). The amount of time between Sleep Start andSleep End, scored as sleep according to the Actiware-Sleep algorithm.It is determined by the summation of the number of epochs that donot exceed the sensitivity threshold and multiplying that value by theepoch length in min. AST is expressed in h and min.

. Actual Wake Time (AWT). The amount of time between Sleep Start andSleep End, which is scored as wake according to the Actiware-Sleep soft-ware. It is determined by the summation of the number of epochs thatexceed the sensitivity threshold and multiplying that value by theepoch length in min. AWT is expressed in h and min.



. Sleep Efficiency (SE). An index of the amount of time in bed actually spentsleeping. It is determined by the division of the actual sleep time by timein bed and multiplying the result by 100.

. Sleep Latency (SL). The period of time required for sleep onset after retir-ing to bed. Sleep latency is the period between Bed Time and SleepStart.

. Sleep Bout (SB). The number of continuous blocks of sleep calculatedbetween Sleep Start and Sleep End.

. Wake Bout (WB). A measure of the intervening waking periods betweenSleep Start and Sleep End. It is indicative of fitful sleep if short sleepbouts alternate with short waking bouts.

. Fragmentation Index (FI). An index of restlessness. It indicates the extentto which sleep is disturbed in an individual.

In addition, either the patients or their attendant recorded the bedtime,wake-up time, and meal timings in a diary on daily basis. The time tobed was verified from the data obtained from the Actiware-Sleep software.The apparently healthy subjects recorded the above parameters them-selves. The data on sleep parameters were transferred to a MicrosoftExcel worksheet for further analyses.

In addition, the frequency of naps (NF), the total time spent napping(TN), the average nap length (AN), and the total time spent napping per1 h awake (TN/A) time were determined by Actiware-Sleep software. Anap was defined as suspected sleep periods of less than 1 h during theawake-span.

Statistical Analyses

The circadian rhythm characteristics, such as average of the rhythmicfunction (mesor, M: rhythm-adjusted 24-h mean), amplitude (A, one-halfof the difference between the highest and the lowest value of the 24 hcosine approximation fit to the data), and peak or acrophase (Ø, timingof the highest value of the rhythmic function) were estimated from log-transformed data (Nelson et al., 1979) at two different fixed windows(i.e., t ¼ 24 h and t ¼12 h). The additional 12 h period was especiallyselected because bimodality in the activity pattern was discerned in the per-iodograms/actograms of most of the subjects. Harmonic means were cal-culated for M, A, and Ø obtained at both windows. The prominentperiod (t) of the studied variable was determined by Power SpectrumAnalysis technique (De Prins et al., 1986), which is applicable to timeseries with or without missing data. Autocorrelation analyses, which donot presume the shape of a cyclic function, were also applied to derivethe autocorrelation coefficient at t ¼ 24 h (i.e., r24) to obtain anothermeasure of the regularity of the activity pattern over the 24 h. The



dichotomy index, I , O, differences in activity distribution between thedaily activity and rest spans were also computed. The dichotomy indexI , O is the percentage of the 1 min activity values measured while thepatient is in bed that is inferior to the median value when the patient isout of bed. The value of this index can vary between 0% and 100%. Inthe case of a marked circadian rhythm with complete rest at night andhigh activity during daytime, I , O reaches 100%. Other statistical tech-niques, such as descriptive statistics, ANOVA followed by Duncan’s mul-tiple-range test, and t-test were also used when pertinent (COSTAT,CoHort Software, Version 4.02, #1990).

RESULTS

Rest-Activity Rhythm

Figure 1 shows the actograms of the rest-activity patterns in cancer in-patients and control subjects. The apparently healthy male and female sub-jects displayed a distinct and regular day-night pattern in wrist activity;more activity was displayed during the day than night (see Figure 1Band 1D). However, in the male and female cancer in-patients this distinc-tion and regularity was less marked (see Figure 1A and 1C). The cancer in-patients also showed daytime napping. The level and intensity of activitywere higher in the male and female control subjects than male andfemale cancer patients.

Rhythm Detection

A statistically significant circadian rhythm (t ¼ 24 h) in rest-activity wasdocumented in all control subjects and cancer in-patients, irrespective ofthe sites of the cancer. Most of the subjects also exhibited statistically signifi-cant 12 h rhythmicity in rest-activity. The harmonic means of the mesors,amplitudes, and acrophases obtained separately for the 24 and 12 hperiods were computed and are reported in Table 2. Figure 2 illustratesthe best-fitting cosine curves in selected cancer patients (n ¼ 18) andcontrol subjects (n ¼ 10). Spectral analyses revealed the period of therest-activity rhythm to be exactly 24 h in most control subjects andcancer in-patients. However, in four cancer patients and one controlsubject, frequency multiplication of the rest-activity rhythm was witnessed,with a period length of 12 or 6 h. Figure 3 shows illustrative examples ofcircadian and non-circadian prominent periods in the rest-activityrhythm of four cancer patients (see Figure 3A–3D) and two control sub-jects (Figure 3E and 3F). In these examples, several smaller peaks wereobserved in cancer patients, whereas in control subjects, only a secondarypeak was observed at 12 h. Although control subjects had a prominent



FIGURE 1 Illustrative examples of double-plotted rest-activity profile (actogram) in cancer in-patients(A and C) and apparently healthy subjects (B and D). Abscissa depicts clock hour. Each row representstwo 24 h spans (i.e., for a given day and the following day). The height of the black marks in each rowindicates the level of activity (counts in relative unit) for the corresponding time in the abscissa. Acto-grams of cancer patients reveal marked activity in the nighttime.

1185


TABLE 2 Summary of the Characteristics of Circadian Rest-Activity Rhythm of Cancer In-patients and Control Subjects. Rhythm Parameters were Computed at FixedWindows with t ¼ 24 h or t ¼ 12 h. The Harmonic Means of Each Parameter Obtained at t ¼ 24 h and t ¼ 12 h were also Calculated

Cancer patients Control subjects Cancer vs.control: tvalue ( pvalue)Variable Mean + SE (n) Median (range) Mean + SE (n) Median (range)

Rhythm parameters at t ¼ 24 h24 h average, M 00.98 + 00.05 (31) 00.99 (00.17, 01.67) 01.64 + 00.03 (35) 01.66 (01.34, 01.94) 11.48 (0.001)Amplitude, A 00.48 + 00.03 (31) 00.47 (00.09, 00.91) 00.92 + 00.02 (35) 00.95 (00.60, 01.19) 11.33 (0.001)Acrophase, Ø in h 13.81 + 00.26 (31) 13.90 (10.10, 17.50) 15.29 + 00.19 (35) 15.00 (13.70, 18.50) 04.65 (0.001)

Rhythm parameters at t ¼ 12 h24 h average, M 00.98 + 00.05 (31) 01.00 (00.17, 01.67) 01.66 + 00.03 (35) 01.69 (01.37, 02.00) 11.77 (0.001)Amplitude, A 00.29 + 00.02 (31) 00.30 (00.04, 00.53) 00.53 + 00.02 (35) 00.51 (00.31, 00.84) 08.32 (0.001)Acrophase, Ø in h 08.40 + 00.12 (31) 08.50 (07.00, 09.70) 09.04 + 00.14 (35) 08.90 (07.60, 11.20) 03.45 (0.001)

20.40 + 00.12 (31) 20.50 (19.00, 21.70) 21.04 + 00.14 (35) 20.90 (19.60, 23.20)Rhythm parameters (harmonic mean)

24-h average, M 00.98 + 00.05 (31) 00.99 (00.17, 01.67) 01.65 + 00.03 (35) 01.68 (01.36, 01.97) 11.69 (0.001)Amplitude, A 00.35 + 00.02 (31) 00.35 (00.06, 00.59) 00.66 + 00.02 (35) 00.65 (00.42, 00.89) 11.36 (0.001)Acrophase, Ø in h 10.41 + 00.14 (31) 10.30 (09.00, 11.90) 11.35 + 00.16 (35) 11.10 (09.90, 13.40) 04.56 (0.001)

11

86


period of 24 h, most of them exhibited a secondary peak at 12 h. Thisbimodality in the rest-activity rhythm was drastically suppressed amongthe cancer patients.

Circadian 24 h Average (Mesor)

The averages of the mesors for each group obtained at the fixedwindows of 24 h, 12 h, and their harmonic means are shown in Table 2.The results of cosinor rhythmometry indicated inter-individual differencesin the level of 24-h average activity in cancer in-patients and control sub-jects. Results showed a statistically significant ( p , 0.001) difference inthe harmonic means of the mesors of the activity rhythm between cancerin-patients and control subjects, independent of gender (see Figure 4A).This difference was also statistically significantly validated when activity

FIGURE 2 Representative examples of cosine-fitted curve in cancer in-patients and control subjectsbased on harmonic means of mesors, amplitudes, and acrophases obtained independently at fixed win-dows with periods equal to 24 h and 12 h (A: male control, B: male patient, C: female control, D: femalepatient). Same scale has been used in each plot. The following cosine function was used in the singlecosinor method: Yti ¼ MþA cos (vti þ Ø) þ ei, where Yti ¼ the value of the fitted cosine function attime ti. M ¼ mesor (rhythm-adjusted mean), which is equal to the arithmetic mean when measure-ments are taken at equidistant time intervals. A ¼ amplitude (one-half the difference between the high-est and the lowest value of the rhythmic function). v ¼ angular frequency (degrees per unit time, with3608 representing a complete cycle). t ¼ time when measurements are taken. Ø ¼ peak (acrophase:timing of the highest value with reference to local midnight). ei ¼ uncontrollable random errorassumed to be independent normal and deviates with means zero and common variance. Figures indi-cate all cancer in-patients and control subjects exhibited a significant circadian rhythm in rest-activity;however, the amplitude considerably dampened in most cancer in-patients, irrespective of gender, ascompared to their respective apparently healthy counterparts. The bimodality in rest-activity rhythmalso appeared to be dampened in cancer patients.



count, obtained at the fixed window with t ¼ 24 h, was expressed asaverage activity per 15 min (see Figure 4B).

Circadian Amplitude

Variation in the inter-individual as well as group circadian amplitudesof the rest-activity rhythm was observed in both groups. The harmonicmeans of the amplitude of the rest-activity rhythm in the cancer in-patientswere significantly lower ( p , 0.001) than those of the control subjects, irre-spective of gender (see Figure 4C).

Circadian Peak (Acrophase)

The harmonic means of the circadian acrophase of the rest-activityrhythm of the subjects were localized mostly in the morning hours. A stat-istically significant ( p , 0.001) difference was observed for the averagetiming of the peaks between the two groups. The acrophase of the rest-activity rhythm occurred earlier, by about 1 h, in the cancer in-patientsthan the controls (see Figure 4D).

FIGURE 3 Illustrative examples of prominent periods of the 24 h rest-activity rhythm in two femalecancer in-patients (A and B), two male cancer in-patients (C and D), and two-control subjects (E: female,F: male). One control subject and four cancer in-patients exhibited a period less than 24 h.



Dichotomy Index (I < O)

The mean dichotomy index was drastically lower in the cancer in-patients (female cancer patients vs. female controls, p , 0.001; malecancer patients vs. male controls, p , 0.001; all cancer patients vs. all con-trols, p , 0.001) compared with the control subjects (see Figure 4E).

Autocorrelation Coefficient (r24)

The autocorrelation coefficients at t ¼ 24 h, the r24, were computedfor each cancer patient and control subject. In each subject, the r24value was statistically significant. However, statistically significant differ-ences were not obtained when the averages of r24 values of the cancerpatients and control subjects were compared as a group (see Figure 4F).Nonetheless, the group mean r24 was lower in the cancer in-patientsthan control subjects (Figure 4F).

FIGURE 4 24 h average based on harmonic means (A), average level of activity per 15 min intervals (B),amplitude based on harmonic means (C), and acrophse based on harmonic means (D) of the circadianrest-activity rhythm of control subjects and cancer in-patients, mean dichotomy index, I , O (E), andautocorrelation coefficient, r24 (F). For all the mentioned rhythm parameters, statistically significant(��p , 0.001) differences were noticed between the cancer in-patients and control subjects, except r24.



Peak (Acrophase) Map

The harmonic means of the peaks (acrophase) of the rest-activity circa-dian rhythm in 35 control subjects (22 males and 13 females) and 31 cancerin-patients (19 males and 12 females) are plotted in Figure 5A and 5B. Themaps reveal inter-individual variability in the acrophase of the 24 h rest-activity rhythm among both the cancer in-patients and apparentlyhealthy control subjects.

Sleep and Nap Parameters

Table 3 depicts the means and standard errors of TIB, AS, AST, AWT,SE, SL, SB, WB, FI, NF, TN, AN, and TN/A of the cancer in-patients andcontrol subjects. Cancer patients spent a longer time in bed and had longerassumed and actual sleep durations than control subjects. The average

FIGURE 5 Acrophase map showing the timings (harmonic mean) of occurrences in the circadian rest-activity rhythm of individual control subjects (A) and cancer in-patients (B).



duration of the sleep and wake bouts was significantly longer in thecancer in-patients than controls. In addition, nap frequency, total napduration, average nap, and total nap duration per 1 h awake-time werestatistically significantly higher in cancer patients than control subjects(see Table 3).

DISCUSSION

A noticeable difference between the rest-activity circadian pattern ofhealthy subjects and cancer in-patients was observed. The normal day-night variability in cancer patients was impaired. Comparable impairmentof the normal day-night pattern in the activity has been documented inpatients suffering from breast cancer (Roscoe et al., 2002), metastatic color-ectal cancer (Chevalier et al., 2003; Mormont & Waterhouse, 2002), lungcancer (Levin et al., 2005), and head and neck cancer (Pati et al., 2006).Healthy subjects exhibit a high amplitude 24 h rest-activity rhythm charac-terized by high level activity during the daytime (waking) hours and verylow to zero-level activity during the night (rest). The onset and offset ofactivity timings of the healthy subjects can be well differentiated and areconsistent from one day to the next. However, in the cancer in-patients,this distinction was less marked. These observations corroborate with thefindings of several earlier studies (Chevalier et al., 2003; Levin et al.,2005; Mormont & Waterhouse, 2002; Pati et al., 2006).

The present results show that all of the cancer patients and control sub-jects exhibited a statistically significant circadian rest-activity rhythm.These observations are in full agreement with previous reports that docu-mented statistically significant 24 h rhythmicity in rest-activity, plasma con-centrations of melatonin, 6-a-sulfatoxymelatonin, cortisol, and lymphocytecount in apparently healthy subjects and cancer patients (Chevalier et al.,2003; Minors et al., 1996; Mormont et al., 2000, 2002; Pati et al., 2006).Levin et al. (2005) reported a disruption of the 24 h sleep-activitypattern in patients suffering from advanced non-small-cell lung cancer.Disruption of the sleep-activity pattern does not necessarily mean theabsence of a statistically significant circadian rhythm. A consistentrhythm in serum cortisol level has also been reported in patients withadvanced gastrointestinal carcinomas (Raida et al., 2002). Therefore,the rest-activity appears to be a robust circadian variable, as all cancerpatients and control subjects exhibited a statistically significant rhythmicpattern.

It is well known that the 24 h pattern of the rest-activity in humans isshaped more like a square than a sinusoidal waveform; hence, thecosinor method is less meaningful in the present context. Therefore, weapplied three additional analyses to properly validate our findings:



TABLE 3 Means (+1 SE) of Time in Bed (TIB), Assumed Sleep (AS), Actual Sleep Time (AST), Actual Wake Time (AWT), Sleep Efficiency (SE),Sleep Latency (SL), Sleep Bouts (SB), Wake Bouts (WB), Fragmentation Index (FI), Nap Frequency (NF), Total Nap (TN), Average Nap (AN), andTotal Nap Per 1 h Awake (TN/A) of Cancer Patients (n ¼ 31) and Control Subjects (n ¼ 35)

Variable (unit) Control subjects Median (range) Cancer patients Median (range)

TIB (h) 7.41 + 0.16 7.10 (6.06, 9.81) 8.13 + 0.22b 8.14 (5.50, 10.22)AS (h) 6.55 + 0.18 6.41 (4.46, 9.17) 7.22 + 0.25a 7.31 (4.02, 9.61)AST (h) 5.62 + 0.14 5.46 (4.01, 7.54) 6.18 + 0.23a 6.17 (3.23, 9.27)AWT (h) 1.00 + 0.06 0.86 (0.48, 1.70) 1.04 + 0.09ns 0.99 (0.33, 2.04)SE (%) 76.05 + 1.23 76.93 (63.16, 91.40) 75.62 + 1.74ns 77.10 (55.13, 90.50)SL (min) 29.42 + 4.60 21.28 (3.00, 109.50) 30.02 + 4.14ns 24.00 (1.00, 82.50)SB (count) 29.47 + 1.52 26.86 (18.00, 60.25) 41.82 + 1.99c 40.25 (18.67, 62.00)WB (count) 30.86 + 2.09 26.60 (17.73, 79.00) 41.12 + 1.98c 39.75 (18.00, 61.25)FI (%) 28.96 + 1.68 32.06 (4.55, 47.31) 24.99 + 2.86ns 18.17 (6.30, 66.63)NF (count) 3.81 + 0.44 3.67 (0.14, 8.75) 16.58 + 1.42c 17.00 (2.5, 33.5)TN (sec) 1848.35 + 216.47 1826.25 (42.86, 43.00) 7959.09 + 711.75c 7811.25 (1008.75, 16548.75)AN (sec) 395.10 + 20.35 416.86 (42.86, 573.75) 468.87 + 6.70c 470.00 (383.33, 542.67)TN/A (sec h21 awake) 117.17 + 13.01 115.41 (2.48, 267.40) 530.83 + 45.25c 511.55 (64.21, 1054.34)

Means differ statistically significantly from the respective control group.ap , 0.02;bp , 0.005;cp , 0.001;nsNot significant.

11

92


1. computation of the dichotomy index (I , O);2. computation of r24 from autocorrelation analysis; and3. analysis and comparison of activity data in the form of 15 min

averages.

The average dichotomy index was statistically significantly lower in thegroup of cancer in-patients than the respective control subjects. Thisclearly suggests that the studied cancer patients did experience heightenedsleep difficulties. Similar observations were previously reported (Mormont& Waterhouse, 2002; Mormont et al., 2000). The latter showed that thedichotomy index (I , O) is usually best correlated to quality of life(QoL). Therefore, the studied cancer patients are presumed to have hada poor QoL. Furthermore, the average autocorrelation coefficient (r24)was lower among the cancer patients, although it was not statistically sig-nificantly different from that of the group of control subjects. Takentogether, the findings suggest that circadian rhythm parameters, especiallythe amplitude and acrophase, of the rest-activity are greatly affected by thedisease conditions. In addition, the average level of activity, when com-puted at intervals of 15 min, was drastically lower among the cancerpatients as compared with the control subjects. Waveforms were also com-pared based on harmonic means of the mesors, amplitudes, and acro-phases obtained independently by cosinor analysis at fixed windows withperiod equals to 24 h and 12 h for both cancer in-patients and control sub-jects. From this comparison, it was revealed that the bimodality in the rest-activity rhythm was greatly dampened in the group of cancer in-patients ascompared with the control subjects.

The results of the present study, therefore, document that the disrup-tion of circadian rest-activity rhythm of cancer in-patients is characterizedby dampening of amplitude, lowering of mean activity level, and phaseadvancement. These alterations of circadian rhythm characteristics couldbe attributed to disease. It is important to mention here that all cancerpatients under investigation in the present study were in-patients,although they were ambulatory within and outside the ward in the pre-mises of the hospital. The characteristics of the rest-activity circadianrhythm in out-patients may differ from that of in-patients. However, thisaspect has not been examined in the present study. The present studysuffers from the deficiencies in that it did not have in-hospital control sub-jects, although it is highly desirable. It was not feasible to incorporate suchcontrol subjects because admission to the hospital is restricted to severely illpatients. Further appropriate in-house control groups were also not avail-able. Therefore, the interpretation and application of the findings of thepresent study have to be done with caution, especially because in-housepatients spent much time sitting and waiting to be examined by doctors.



Thus, their level of activity was limited by hospital routine, which maynoticeably influence the characteristics of circadian rhythm.

Age-related decrements in sleep and daytime alertness levels havebeen documented in human subjects. This phenomenon has been attribu-ted to the weaker circadian regulation of sleep and wakefulness (Cajochenet al., 2006). Paradoxically, persistence of lifestyle regularity has beenreported in adults and seniors (Monk et al., 2006). It is difficult toexplain how both weak circadian organization and lifestyle regularitycoexist among senior subjects. The dampening of the circadian amplitudein cancer patients has not been examined categorically as function of ageand gender in previous investigations (Chevalier et al., 2003; Kobayashiet al., 2002; Levi, 2006; Mormont & Levi, 2003). However, it has beenreported that the distribution of the 24 h rest-activity cycle parameters isindependent of gender, age, primary tumor, number of metastatic sites,and prior treatment (Mormont & Waterhouse, 2002). Further, a signifi-cant correlation between actigraphy measures and age was not apparentin breast cancer patients (Roscoe et al., 2002). However, it has to be empha-sized that gender is an important factor in studies that involve sleep andcircadian rhythms in young human subjects (Goel et al., 2005). Theseauthors have reported better sleep quality in women than men acrossseveral consecutive nights.

It seems that the rest-activity circadian rhythm is not easily abolished ormasked by cancer diseases, irrespective of type, stage, and performancestatus of the cancer. Nonetheless, results of earlier studies reveal the mag-nitude of alteration, but not the absence of significant circadian rhythms, inthe characteristics of rest-activity or hormonal rhythms could vary as func-tion of the severity of the disease (i.e., performance status and stage; seeBartsch et al., 1994; Mormont & Levi, 1997; Mormont et al., 2000;Touitou et al., 1995). Significant alterations in circadian rhythms havebeen reported in patients with large tumor burden, liver metastasis,poor performance status, and advanced stage (Mormont & Levi, 1997;Mormont et al., 2000).

The results of the present study reconfirm those of other reported findings(Levin et al., 2005; Mormont et al., 2000; Pati et al., 2006) in that the 24 haverage and amplitude of the rest-activity circadian rhythm decline drasti-cally and the acrophase advances significantly in cancer patients comparedwith the age- and gender-matched apparently healthy subjects. In thepresent study, the acrophase occurred earlier in the cancer in-patients;similar phase advancement has been reported before (Chevalier et al.,2003). Rhythm alterations may depend upon tumor type, growth rate,and extent of tumor differentiation (Levi, 2006). It has been argued thatbecause patients with abnormal rest-activity cycle have elevated levels ofTGFa, TNFa, and IL-6, the rhythm alteration could be ascribed to the



tumor itself through the release of cytokines and growth factors (Levi,2006). These factors have been known to alter the circadian organizationof humans (Rich et al., 2005). In the present study, an inter-individual vari-ation in the occurrence of the circadian acrophase was noticed in the groupof cancer patients. This could be the reason why the outcome of therapy ofindividual cancer patients varies dramatically. This result strongly advo-cates the implementation of patient-specific chronotherapy. In a recentpaper, Levi & Schibler (2007) envisaged a custom-tailored medicine inthe future. This is based on the concept of cellular heterogeneity ofcancers that gives rise to individual variability in the expression ofrhythm characteristics.

Sleep-wake disturbances in patients with head and neck (Pati et al.,2006) and colorectal (Chevalier et al., 2003) cancer have been reported.Although sleep efficiency, sleep latency, and actual wake time in cancerpatients did not differ from that of the control group of healthy subjects,other sleep parameters (i.e., time in bed and assumed and actual sleeptimes) did differ in a statistically significant manner. In contrast tocontrol subjects, cancer patients spent a longer duration in bed and exhib-ited longer assumed and actual sleep. Although Berger et al. (2005) andLevin et al. (2005) reported that cancer patients do experience sleep/wake disturbances, information about individual sleep parameters incancer patients is limited. Recently, Fernandes et al. (2006) documenteda significant difference in several sleep parameters, namely % sleep,sleep efficiency, and wake after sleep onset, between cancer patients andhealthy volunteers. Other reports also documented that cancer patientstend to have fragmented sleep and poorer sleep efficiency, exhibit morerestlessness at night, and take longer time to fall asleep (Ancoli-Israelet al., 2006; Berger, 1998; Miaskowski & Lee, 1999; Mormont et al.,1996). In addition, in the present study, the cancer patients exhibitedmore episodes of wake and sleep bouts during the habitual sleep period.It seems that in cancer patients, longer time in bed and longer assumedsleep are a usual phenomenon. Both the frequency and duration of napsduring the awake-span were higher among the cancer patients. Thisfinding supports the results of previous studies in that daytime napswere reported in patients with bone metastasis and breast cancer(Ancoli-Israel et al., 2006; Miaskowski & Lee, 1999). Together, thesealterations could lead to circadian rhythm alterations characterized bydecreased amplitude and low 24 h average activity.

In conclusion, the results of the current study found that cancerpatients and controls exhibit 24 h rest-activity rhythmicity, but thecancer in-patients exhibited lower amplitude, lower mean activity level,and earlier acrophase compared to controls. These alterations were inde-pendent of gender and site of cancer. These results might help in design-ing patient-specific chronotherapeutic protocols.



ACKNOWLEDGMENTS

The research was supported by the University Grants Commission,New Delhi, under its DRS-Special Assistance Program, and Departmentof Science and Technology, New Delhi, under the SERC FAST TrackScheme and DST-FIST program. The authors gratefully acknowledgeall cancer patients and control subjects for their voluntary participationin the study. We thank Professor Michael Smolensky and two unknownreferees for reading an earlier version of this paper and for offeringnumerous constructive suggestions.

REFERENCES

American Academy of Sleep Medicine. (2001). International Classification of Sleep Disorders, Revised: Diag-nostic and Coding Manual. Chicago, Ill.: American Academy of Sleep Medicine, pp. 1–401.

Ancoli-Israel S, Liu L, Marler MR, Parker BA, Jones V, Sadler GR, Dimsdale J, Cohen-Zion M,Fiorentino L. (2006). Fatigue, sleep, and circadian rhythms prior to chemotherapy for breastcancer. Support Care Cancer 14:201–209.

Bartsch C, Bartsch H, Fluchter SH, Mecke D, Lippert TH. (1994). Diminished pineal functioncoincides with disturbed circadian endocrine rhythmicity in untreated primary cancer patients.Ann. NY Acad. Sci. 719:502–525.

Berger AM. (1998). Patterns of fatigue and activity and rest during adjuvant breast cancer chemother-apy. Oncol. Nurs. Forum 25:51–62.

Berger AM, Parker KP, Young-McCaughan S, Mallory GA, Barsevick AM, Beck SL, Carpenter JS,Carter PA, Farr LA, Hinds PS, Lee KA, Miaskowski C, Mock V, Payne JK, Hall M. (2005).Sleep/wake disturbances in people with cancer and their caregivers: State of the science. Oncol.Nurs. Forum. 32:98–126.

Cajochen C, Munch M, Knoblauch V, Blatter K, Wirz-Justice A. (2006). Age-related changes in the cir-cadian and homeostatic regulation of human sleep. Chronobiol. Int. 23:461–474.

Chevalier V, Mormont MC, Cure H, Chollet P. (2003). Assessment of circadian rhythms by actimetry inhealthy subjects and patients with advanced colorectal cancer. Oncol. Reports 10:733–737.

De Prins J, Cornelissen G, Malbecq W. (1986). Statistical procedures in chronobiology and chrono-pharmacology. Annu. Rev. Chronopharmacol. 2:27–141.

Eriguchi M, Levi F, Hisa T, Yanagie H, Nonaka Y, Takeda Y. (2003). Chronotherapy for cancer.Biomed. Pharmacother. 57:92–95.

Fernandes R, Stone P, Andrews P, Morgan R, Sharma S. (2006). Comparison between fatigue, sleepdisturbance, and circadian rhythm in cancer inpatients and healthy volunteers: Evaluation ofdiagnostic criteria for cancer-related fatigue. J. Pain Symptom Manage. 32:245–254.

Filipski E, King VM, Li XM, Granda TG, Mormont MC, Liu XH, Claustrat B, Hastings H, Levi F.(2002). Host circadian clock as a control point in tumor progression. J. Natl. Cancer Inst. 94:690–697.

Goel N, Kim H, Lao RP. (2005). Gender differences in polysomnographic sleep in young healthy slee-pers. Chronobiol. Int. 22:905–915.

Kobayashi M, Wood PA, Hrushesky WJ. (2002). Circadian chemotherapy for gynecological and geni-tourinary cancers. Chronobiol. Int. 19:237–251.

Levi F. (2001). Circadian chronotherapy for human cancers. The Lancet Oncol. 2:307–314.Levi F. (2006). Chronotherapeutics: The relevance of timing in cancer therapy. Cancer Causes Control.

17:611–621.Levi F, Schibler U. (2007). Circadian rhythms: Mechanisms and therapeutic implications. Annu. Rev.

Pharmacol. Toxicol. 47:593–628.Levin RD, Daehler MA, Grutsch JF, Quiton J, Lis CG, Peterson C, Gupta D, Watson K, Layer D, Huff-

Adams S, Desai B, Sharma P, Wallam M, Delioukina M, Ball P, Bryant M, Ashford M, Copeland D,



Ohmori M, Wood PA, Hrushesky WJM. (2005). Circadian function in patients with advancednon-small-cell lung cancer. Br. J. Cancer 93:1202–1208.

Littner M, Kushida CA, Anderson WM, Bailey D, Berry RB, Davila DG, Hirshkowitz M, Kapen S,Kramer M, Loube D, Wise M, Johnson SF. (2003). American academy of sleep medicine practiceparameters. Practice parameters for the role of actigraphy in the study of sleep and circadianrhythms: An update for 2002. Sleep 26:337–341.

Miaskowski C, Lee KA. (1999). Pain, fatigue and sleep disturbances in oncology outpatients receivingradiation therapy for bone metastasis: A pilot study. J. Pain Symptom Manage. 17:320–332.

Minors DS, Akerstedt T, Atkinson G, Dahlitz M, Folkard S, Levi F, Mormont C, Parkes D,Waterhouse JM. (1996). The difference between activity when in bed and out of bed, I:Healthy subjects and selected patients. Chronobiol. Int. 13:27–34.

Monk TH, Buysse DJ, Hall M, Nofzinger EA, Thompson WK, Mazumdar SA, Reynolds CF. (2006).Age-related differences in the lifestyle regularity of seniors experiencing bereavement care-giving, insomnia, and advancement in to old-old age. Chronobiol. Int. 23:831–841.

Mormont MC, Levi F. (1997). Circadian-system alterations during cancer processes: A review. Int. J.Cancer 70:241–247.

Mormont MC, Levi F. (2003). Cancer chronotherapy: Principles, applications and perspectives. Amer-ican Cancer Society. 9:155–169.

Mormont MC, Waterhouse J. (2002). Contribution of the rest-activity circadian rhythm to quality of lifein cancer patients. Chronobiol. Int. 19:313–323.

Mormont MC, De Prins J, Levi F. (1996). Study of circadian rhythms of activity by actometry: Prelimi-nary results in 30 patients with metastatic colorectal cancer. Pathol. Biol. (Paris) 44:165–171.

Mormont MC, Waterhouse J, Bleuzen P, Giacchetti S, Jami A, Bodgan A, Lellouch J, Misset JL,Touitou Y, Levi F. (2000). Marked 24-h rest/activity rhythms are associated with better qualityof life, better response and longer survival in patients with metastatic colorectal cancer andgood performance status. Clin. Cancer Res. 6:3038–3045.

Mormont MC, Langouet AM, Claustrat B, Bogdan A, Marion S, Waterhouse J, Touitou Y, Levi F.(2002). Marker rhythms of circadian system function: A study of patients with metastatic colorectalcancer and good performance status. Chronobiol. Int. 19:141–155.

Nelson W, Tong YL, Lee LK, Halberg F. (1979). Methods of cosinor rhythmometry. Chronobiologia 6:305–323.

Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, Carbone PP. (1982). Toxicityand response criteria of the eastern cooperative oncology group. Am. J. Clin. Oncol. 5:649–655.

Pati AK, Parganiha A, Kar A, Soni R, Roy S, Choudhary V. (2006). Implications of the study of rest-activity circadian rhythm in head and neck cancer patients. Biol. Rhythm Res. 37:497–505.

Raida M, Kliche K-O, Schwabe W, Hausler P, Clement JH, Behnke D, Hoffken K. (2002). Circadianvariation of dihydropyrimidine dehydrogenase mRNA expression in leukocytes and serum corti-sol levels in patients with advanced gastrointestinal carcinomas compared to healthy controls.J. Cancer Res. Clin. Oncol. 128:96–102.

Rich T, Innominato PF, Boerner J, Mormont MC, Iacobelli S, Baron B, Jasmin C, Levi F. (2005). Elev-ated serum cytokines correlated with altered behavior, serum cortisol rhythm, and dampened 24hour rest-activity patterns in patients with metastatic colorectal cancer. Clin. Cancer Res. 11:1757–1764.

Roscoe JA, Morrow GR, Hickok JT, Bushunow P, Matteson S, Rakita D, Andrews PLR. (2002). Tem-poral interrelationships among fatigue, circadian rhythm and depression in breast cancer patientsundergoing chemotherapy treatment. Support Care Cancer. 10:329–336.

Touitou Y, Levi F, Bogdan A, Benavides M, Bailleul F, Misset JL. (1995). Rhythm alteration in patientswith metastatic breast cancer and poor prognostic factors. J. Cancer Res. Clin. Oncol. 121:181–188.

Touitou Y, Smolensky MH, Portaluppi F. (2006). Ethics, standards, and procedures of animal andhuman chronobiology research. Chronobiol. Int. 23:1083–1096.



Documents

Atanu Kumar Pati a; Arti Parganiha2007