One year pre-post intervention follow-up of psychological, immune, endocrine and blood pressure outcomes of mindfulness-based stress reduction (MBSR) in breast and prostate cancer outpatients

Article (PDF Available)inBrain Behavior and Immunity 21(8):1038-49 · December 2007with129 Reads
DOI: 10.1016/j.bbi.2007.04.002 · Source: PubMed
Abstract
This study investigated the ongoing effects of participation in a mindfulness-based stress reduction (MBSR) program on quality of life (QL), symptoms of stress, mood and endocrine, immune and autonomic parameters in early stage breast and prostate cancer patients. Forty-nine patients with breast cancer and 10 with prostate cancer enrolled in an eight-week MBSR program that incorporated relaxation, meditation, gentle yoga and daily home practice. Demographic and health behaviors, QL, mood, stress symptoms, salivary cortisol levels, immune cell counts, intracellular cytokine production, blood pressure (BP) and heart rate (HR) were assessed pre- and post-intervention, and at 6- and 12-month follow-up. Fifty-nine, 51, 47 and 41 patients were assessed pre- and post-intervention and at 6- and 12-month follow-up, respectively, although not all participants provided data on all outcomes at each time point. Linear mixed modeling showed significant improvements in overall symptoms of stress which were maintained over the follow-up period. Cortisol levels decreased systematically over the course of the follow-up. Immune patterns over the year supported a continued reduction in Th1 (pro-inflammatory) cytokines. Systolic blood pressure (SBP) decreased from pre- to post-intervention and HR was positively associated with self-reported symptoms of stress. MBSR program participation was associated with enhanced quality of life and decreased stress symptoms, altered cortisol and immune patterns consistent with less stress and mood disturbance, and decreased blood pressure. These pilot data represent a preliminary investigation of the longer-term relationships between MBSR program participation and a range of potentially important biomarkers.

Figures

One year pre–post intervention follow-up of psychological,
immune, endocrine and blood pressure outcomes of
mindfulness-based stress reduction (MBSR) in breast
and prostate cancer outpatients
Linda E. Carlson
a,b,
*
, Michael Speca
a,b
, Kamala D. Patel
c
, Peter Faris
d
a
Department of Psychosocial Resources, Tom Baker Cancer Centre Holy Cross Site, Alberta Cancer Board, 2202 Second St. S.W.,
Calgary, Alta., Canada T3B 0W7
b
Department of Oncology, Faculty of Medicine, University of Calgary, Canada
c
Department of Physiology and Biophysics, Faculty of Medicine, University of Calgary, Canada
d
Department of Community Health Sciences, University of Calgary and Centre for Advancement of Health, Calgary Health Region, Canada
Received 26 November 2006; received in revised form 16 March 2007; accepted 6 April 2007
Abstract
Objectives. This study investigated the ongoing effects of participation in a mindfulness-based stress reduction (MBSR) program on
quality of life (QL), symptoms of stress, mood and endocrine, immune and autonomic parameters in early stage breast and prostate can-
cer patients.
Methods. Forty-nine patients with breast cancer and 10 with prostate cancer enrolled in an eight-week MBSR program that incorpo-
rated relaxation, meditation, gentle yoga and daily home practice. Demographic and health behaviors, QL, mood, stress symptoms, sal-
ivary cortisol levels, immune cell counts, intracellular cytokine production, blood pressure (BP) and heart rate (HR) were assessed pre-
and post-intervention, and at 6- and 12-month follow-up.
Results. Fifty-nine, 51, 47 and 41 patients were assessed pre- and post-intervention and at 6- and 12-month follow-up, respectively,
although not all participants provided data on all outcomes at each time point. Linear mixed modeling showed significant improvements
in overall symptoms of stress which were maintained over the follow-up period. Cortisol levels decreased systematically over the course
of the follow-up. Immune patterns over the year supported a continued reduction in Th1 (pro-inflammatory) cytokines. Systolic blood
pressure (SBP) decreased from pre- to post-intervention and HR was positively associated with self-reported symptoms of stress.
Conclusions. MBSR program participation was associated with enhanced quality of life and decreased stress symptoms, altered cortisol
and immune patterns consistent with less stress and mood disturbance, and decreased blood pressure. These pilot data represent a prelimin-
ary investigation of the longer-term relationships between MBSR program participation and a range of potentially important biomarkers.
Crown copyright Ó 2007 Published by Elsevier Inc. All rights reserved.
Keywords: Cortisol; Cytokines; Blood pressure; Heart rate; Meditation; Cancer; Stress; Quality of life; Mood
1. Introduction
Health care delivery has evolved with the advent of more
holistic practices and multidisciplinary care of many
chronic and acute diseases. Clinical treatment and wellness
programs based on mindfulness meditation and yoga have
proliferated. Many are modeled on the mindfulness-based
stress reduction (MBSR) program of Jon Kabat-Zinn and
colleagues at the Stress Reduction Clinic of the University
of Massachusetts Medical Centre (Kabat-Zinn, 1990).
MBSR is rooted in contemplative spiritual traditions in
which the experience of conscious awareness is actively
0889-1591/$ - see front matter Crown copyright Ó 2007 Published by Elsevier Inc. All rights reserved.
doi:10.1016/j.bbi.2007.04.002
*
Corresponding author. Address: Department of Psychosocial Resources,
Tom Baker Cancer Centre Holy Cross Site, Alberta Cancer Board, 2202
Second St. S.W., Calgary, Alta., Canada T3B 0W7. Fax: +1 403 355 3209.
E-mail address: lindacar@cancerboard.ab.ca (L.E. Carlson).
www.elsevier.com/locate/ybrbi
Brain, Behavior, and Immunity xxx (2007) xxx–xxx
BRAIN,
BEHAVIOR,
and IMMUNITY
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
cultivated in specific ways. Attitudes of non-judging, accep-
tance and patience provide a framework for a meditative
practice emphasizing focused awareness of one’s own expe-
rience, often beginning with breath awareness. Typically,
this leads to a state of relaxation and alert observant detach-
ment. W e have reported salutary effects of the MBSR pro-
gram in cancer patients including de creased symptoms of
stress, improved mood, better quality of life, better sleep,
as well as changes in immune and endocrine parameters
(Carlson et al., 2003, 2004; Carlson and Garland, 2005; Spe-
ca et al., 2000). To date, however, we have not reported fol-
low-up maintenance of benefit beyond 6-month post-
program (Carlson et al., 2001). In our 6-month follow-up
we showed that large improvements seen on mood distur-
bance and stress symptoms from pre- to post-MBSR were
maintained, without any return to previous higher levels
of distress. The most benefit was seen on subscales of
depression, anxiety and anger (Carlson et al., 2001).
The proliferation of report s of MBSR interventions in
health care have inspired two general narrative reviews
(Bishop, 2002; Baer, 2003) and one meta-analysis (Gross-
man et al., 2004), all of which have supported the efficacy
of MBSR for improving both physical and menta l well-
being in mixed groups of patients with medical illness.
Three recent reviews have summarized the literature on
MBSR interventions specifically in cancer patients (Mac-
kenzie et al., 2005; Smith et al., 2005; Ott et al., 2006), con-
cluding that MBSR provides benefit in areas such as mood,
sleep quality and reductions in stress.
No other researchers have investiga ted MBSR effects on
biological markers such as cortisol or immune function in
cancer patients, but NK cell activity and number increased
after MBSR participation in a small group of non-ran-
domly assigned HIV patients compared to a group of
patients not interested in meditation; however, no changes
in cortisol levels were found (Robinson et al., 2003).
Another small trial with heart disease patients also failed
to find changes in cortisol levels (Robert McComb et al.,
2004). In a workplace sample of healthy volun teers, MBSR
increased antibody titers produced in response to an influ-
enza vaccine compared with a wait-list control group
(Davidson et al., 2003). Previous MBSR studies in medical
populations are rife with methodological problems such as
very small samples and high drop-out rates, illustrating the
difficulty of conducting such trials in medical settings.
Our previous study reported an intervention with early
stage breast and prostate cancer patients on the outcomes
of stress, mood, quality of life, immune cell count and
intracellular cytokine production, and salivary cortisol
(Carlson et al., 2003, 2004). Here, we have followed up
with these same patients 6- and 12-months post-MBSR
on all the outcome measures previously reported, and we
also report previously unpublished results of blood pres-
sure (BP) and heart rate (HR).
Hypertension (high blood pressure) is a reversible risk
factor for illnesses such as heart disease, heart failure and
stroke (Campbell et al., 1999). This may be an important
consideration in breast cancer patients who have under-
gone chemotherapy and/or radiation, since the heart mus-
cle is often damaged by the treatments, making these
patients more susceptible to future heart disease (Shan
et al., 1996; Vallebona, 2000). Thus, it is important to
maintain normotensive status in such cancer survivors to
decreased future risk. Studies of stress reduction interven-
tions including meditation, biofeedback, relaxation, and
cognitive-behavior therapy have found lowered levels of
resting systolic and diastolic blood pressure in intervention
participants (Blumenthal et al., 2002; Campbell et al.,
1999). This finding has been confirmed by three meta-anal-
yses and a review covering the stress reduction literature
from 1966 to 1997, including several well-designed RCTs
(Spence et al., 1999). The practice of meditation has previ-
ously been associated with decreased heart rate (Telles
et al., 1998; Travis and Wallace, 1997) slowed respiration
(Telles et al., 1998), and lowered blood pressure (Sudsuang
et al., 1991; Schneider et al., 1995; Wenneberg et al., 1997),
primarily in healthy adult participants. Hence, our study
will add to the knowledge base of MBSR on BP and HR
in cancer patients, as well as extend our previous findings
to 6- and 12-months post-program.
2. Methods
2.1. Subjects
Patients were eligible to participate in the study if they met the follow-
ing criteria: (1) age 18 years or older; (2) a diagnosis of Stage 0 , I, or II
breast or early stage (localized to the prostate) prostate cancer at any time
in the past (using standardized TNM diagnostic criteria) and (3) a mini-
mum of three months since surgery (mastectomy/lumpectomy/prostatec-
tomy/cryotherapy). Exclusion criteria were any of: (1) treatment with
chemotherapy or radiation therapy currently or within the past three
months; (2) a concurrent DSM-IV axis I mood, anxiety or psychotic dis-
order (not in full or partial remission); (3) a concurrent autoimmune dis-
order and (4) past participation in an MBSR group.
2.2. Instruments
2.2.1. Demographics and medical history form
Demographic information including age, education, marital status,
occupation and current employment status was obtained on a form cre-
ated for this study. Medical history including type of illness, dates of diag-
nosis and types of treatments were collected. Areas specifically assessed
included heart disease, vascular disorders, autoimmune disorders, epilepsy
and psychiatric disorders. All current medications were recorded.
2.2.2. Health behaviors form
Health behaviors that could potentially affect the immune, endocrine
and/or autonomic systems were recorded, including: amount of coffee,
tea and caffeinated soft-drink consumption (servings/week); alcohol con-
sumption (servings/week); smoking (cigarettes/day); exercise (times/week);
average hours of sleep per night; self-rated quality of sleep (poor, ade-
quate, good) and self-rated quality of diet (poor, adequate, good).
2.2.3. Meditation log
Two different log forms were used during the study: (1) weekly log: this
form collected daily information on minutes spent in home practice of
mediation and yoga from each participant during the 8 weeks of the inter-
vention, and was collected each week during class; (2) Monthly log: for the
2 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
follow-up period, monthly calendars were provided for the entire year to
record daily minutes in practice and collected at each follow-up
appointment.
2.2.4. European organization for research and treatment of cancer quality of
life questionnaire (EORTC QLQ-C30)
(Aaronson et al., 1993): This 30-item quality of life questionnaire
includes five functional domains of quality of life: physical, role, emo-
tional, cognitive, and social function and two items assess global quality
of life. It has become a gold-standard of QL assessment in clinical trials
both in Europe and North America, with much normative data available
for comparison (Aaronson et al., 1991, 1993).
2.2.5. Profile of mood states (POMS)
(McNair et al., 1971): The POMS is a 65-item scale which assesses six
affective dimensions. It has been widely used in the assessment of mood
changes resulting from a variety of interventions due to its responsiveness,
and has been used extensively with cancer populations (Cassileth et al.,
1985).
2.2.6. Symptoms of stress inventory (SOSI)
(Leckie and Thompson, 1979): The SOSI was designed to measure
physical, psychological and behavioral responses to stressful situations.
The respondent is instructed to rate the frequency with which they experi-
ence various stress related symptoms on a 5-point scale ranging from never
to frequently during the past week. Ten subscale scores can be calculated,
and a total stress score.
2.2.7. Immune and cortisol measures
Patients were recruited in 4 groups of 15 as described in the Procedures
section. As a result, time 1 baseline samples for the 3rd and 4th groups
were analyzed concurrently with the 6- and 12-month follow-up samples
from the 1st and 2nd groups. This ensured that differences observed
between different follow-up time points were not the result of changes in
staff, reagents or laboratory procedures. Measures of immune cell counts
and intracellular cytokine production were assessed as detailed in our pre-
vious paper (3). Antibodies directed against specific cell surface determi-
nates were used to determine the proportion of leukocyte subclasses.
Specifically, surface cell marker CD3 was used to identify T-cells, CD19
for B cells, CD4 for helper T-cells, CD8 for cytotoxic T-cells, CD56 for
NK cells, and both CD3 and CD56 for NKT cells. The production of
the cytokines interferon gamma (IFN-c), tumor necrosis factor (TNF),
and interleukin (IL)-4 and -10 (IL-10) by stimulated NK and T-cells
was determined using three-color flow cytometry with FITC-conjugated
anti-CD3 used to identify T-cells, Cy5-conjugated anti-CD56 used to iden-
tify NK cells and PE-conjugated anti-cytokine antibodies used to deter-
mine the level of cytokine expression in each population. The data were
expressed as the percentage of NK or T-cells that were also positive for
the specified cytokines.
Salivary cortisol was measured using solid-phase ELISAs according to
the manufacturer’s instructions on samples collected three times per day
(8:00 AM, 2:00 PM and 8:00 PM) at each assessment period, as previously
described (Carlson et al., 2004).
2.2.8. Blood pressure
BP was measured at six different points during the study using a mer-
cury manometer: (1) one week prior to starting the program, (2) on the
morning of the first meditation session, (3) immediately after the last ses-
sion (within 1/2 h), (4) one week after the last session, (5) 6-month follow-
up and (6) 12-month follow up. At each assessment time, BP and HR were
measured twice from each arm with a five minute rest interval between
measurements, following the procedures recommended by the Canadian
Medical Association (Campbell et al., 1999). Thus, at each of the six time
points, the average of four readings was calculated. Because the measures
were taken on two different days before and after the program, these mea-
sures were also averaged to calculate overall ‘‘pre’’ and ‘‘post’’ BP mea-
sures. Therefore, each ‘‘pre’’ and ‘‘post’’ score was averaged from eight
measurements for those who provided full data. Axillary node dissection
on one arm made it inadvisable to read BP from that arm for several of
the women with breast cancer, so two readings from the other arm were
averaged at each assessment time, resulting in the average of four mea-
sures for each ‘‘pre’’ and ‘‘post’’ value reported for these women.
2.3. Procedures
2.3.1. Recruitment
Patients were recruited from the Tom Baker Cancer Centre. Patients
were primarily recruited with pamphlets and posters around the centre,
in each of the breast and prostate clinic areas, and were able to self-refer.
If patients wished to participate, their name was placed on a waiting list
administered by the research assistant.
2.3.2. Testing
Once 15 patients were accrued on the waiting list they were scheduled
for an individual interview with the principal investigator (PI) during the
week prior to the start of the group to further explain the study, determine
eligibility and provide informed consent. A maximum of three patients
were assessed daily between 8:00 and 10:00 h, to control for time of day.
Patients then met with the RA who took blood pressure readings follow-
ing the procedures detailed above. A blood sample was then taken for
immune measures, and the patients then completed the assessment battery
of questionnaires, supervised by the RA who clarified instructions and
answered questions, which took approximately one-half hour. Salivary
cortisol salivettes were collected by the participants in their homes the
day prior to the first MBSR class, refrigerated overnight, and returned
to the first class.
The RA attended the last part of the final meditation class to measure
BP at that time and hand out salivettes for salivary cortisol collection.
Beginning the week after the completion of the intervention, the same pro-
cedure was followed as prior to the intervention, with the patients return-
ing to the hospital to have BP assessed, provide a blood sample and
complete the questionnaires. They completed the health behavior form
without the assistance of the PI for this assessment and collected salivary
cortisol the day before the appointment. All participants were assessed
within two weeks of the completion of the intervention.
For the 6- and 12-month follow-up assessments, participants were
mailed the questionnaires and the cortisol swabs prior to their appoint-
ment and instructed to complete the questionnaires and collect saliva sam-
ples on the day prior to coming to the centre for assessment of BP and
immune measures.
2.3.3. Intervention
Details of the intervention, including objectives, structure, components
and content, have previously been described (Speca et al., 2000). Our pro-
gram was modeled after the mindfulness-based stress reduction program
at the Stress Reduction and Relaxation Clinic-Massachusetts Medical
Center as described by Kabat-Zinn (Kabat-Zinn, 1990). The intervention
was provided over the course of eight weekly, 90-min group sessions with a
maximum of 15 participants each, plus a 3-h silent retreat on the Saturday
between weeks six and seven. The Saturday retreat combined participants
from all of our ongoing MBSR groups, and usually consisted of about 40
participants.
In addition, we produced and provided patients with a 52-page booklet
containing information pertinent to each week’s instruction, a bibliogra-
phy for those wishing to pursue relevant themes in greater depth, and
an audio-tape or CD recording. The recording provided instruction for
a sensate focused body scan meditation on one side and a guided sitting
meditation on the other. Patients were instructed to practice daily.
2.4. Data analysis
All data analyses were conducted using the Statistical Package for the
Social Sciences (SPSS), version 14.1, for the PC in Windows NT.
The demographic, medical history and health behavior variables were
described using frequency and descriptive statistics. All continuous vari-
ables were tested for normalcy of the distributions at each time period.
L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx 3
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
Of the health behavior and demographic continuous variables, only alco-
hol servings/week was significantly positively skewed at both time periods
(Skewness >2.0). Therefore, the natural log transformation was applied to
this variable, at each time period, which resulted in normal distributions.
These transformed variables were subsequently used in all calculations. On
the EORTC, SOSI and POMS all variables were normally distributed. All
measures of BP and HR also conformed to the normal distribution. Cor-
tisol values were Ln transformed as described in our previous work (Carl-
son et al., 2004).
To assess effects of the intervention pre, post and over the follow-up
and account for missing data, simple mixed-effects models with a hetero-
geneous compound symmetry correlation structure among repeated mea-
sures (Laird and Ware, 1982) were used to compare pre-, post-, 6- and 12-
month scores on total scores on the EORTC, POMS and SOSI, immune
and cortisol measures as well as systolic and diastolic BP and HR scores,
with contrasts to compare the pre-test score and the mean of the post-test
scores (contrasts of 3, 1,1,1 applied to test sessions 1–4, respectively) and
tests in the post-test scores of linear (0,1,0,1 contrast weights) and qua-
dratic (0,1, 2,1 contrast weights) trends. These models were first done
unconditionally (without covariates), and then using baseline values of
health behaviors and disease characteristics as patient-level effects.
Because of the limited number of cases and degrees of freedom, adding
in all of the health behaviors and disease variables into each model was
not statistically viable. Hence, we chose what we felt were the two most
important health behaviors (sleep hours/night and exercise times/week),
and the disease characteristics of cancer stage and time since diagnosis.
These were assessed in each model using interaction terms between each
covariate and each time term. As tests of fixed effects in mixed effects mod-
els do not have exact F distributions, degrees of freedom for these tests
were obtained using a Satterthwaite approximation (Satterthwaite,
1946). Valid inference based on the observed data in mixed-effects models
requires that the probability that observations are missing depends only
on observed covariates or outcomes (the missing at random assumption,
Little and Rubin, 1987). Effect sizes on the psychological measures were
calculated comparing baseline scores to those at times 2, 3 and 4 using
Cohen’s D = (Mean time 2 Mean time 1)/Pooled SD.
To investigate whether BP and HR, or changes in BP and HR were
related to values of quality of life, mood or stress symptoms, Pearson
product–moment correlations were performed between the BP and HR
measures and corresponding psychological scores at each time period.
Similar correlations were also performed between class attendance, home
practice and both psychological and BP and HR change scores from base-
line to the post-intervention period to assess the possibility of dose–
response relationships, and between home practice and BP and HR change
scores from baseline to the follow-up assessments. Correlations were car-
ried out only on participants who provided data at each time point.
3. Results
3.1. Subjects
Fifty-nine, 51, 47 an d 41 patients provided data at pre-
and post-intervention and at 6- and 12-month follow-up,
respectively, however not all patients provided full data
on each measure at each time-point (see flow diagram
Fig. 1). Reasons for drop-outs were recorded as follows:
time 2 (n = 8), 4 did not complete the MBSR program
and withdrew from the study, and 4 did not return to com-
plete the measures; time 3 (n = 4), two did not attend
appointments to complete measures and two did not return
phone calls; time 4 (n = 6), three we were unable to contact
and three did not return phone calls. Thirty-five patients
provided data at all four time points for the immune mea-
sures, 33 for the endocrine measures, an d 31 provided full
data on the psychological measures.
Demographic characteristics of participants at all four
assessment times are presented in Table 1. The participants
at time 1 were a mean of 54.5 years, SD 10.9 years. Most
(n = 42) wer e married or co-habitating at the time of study
entry. Parti cipants were generally well-educated, with a
mean of 14.7 years of formal education. They had been
diagnosed with cancer a median of 1.1 years previously
(range 3 months–20 years). Six participants were greater
than five years post-diagnosis. Just over two thirds had
Minus 2 Months Minus 1 week Time 0 8 weeks 9 weeks 6-Months 12-Months
Recruitment Pre-assessment First Class Last class Post-Assessment Follow-up Follow-up
BP, HR
(n=54) (n=55) (n=45) (n=49) (n=47) (n=37)
Pre-Assessment BP (n=54)
Post-Assessment BP (n=45)
Demographics (n=58)
Medical History (n=58)
Health Behaviors (n=58) (n=32) (n=42) (n=35)
EORTC QLQ-C30 (n=58) (n=43) (n=43) (n=36)
POMS (n=58) (n=43) (n=43) (n=36)
SOSI (n=58) (n=43) (n=43) (n=36)
Immune Measures (n=59)
(n=51) (n=47) (n=41)
Salivary Cortisol (n=55) (n=46) (n=47) (n=38)
Meditation Log (n=51) (n=31) (n=30)
Medical Chart Review (n=58) (n=37)
Fig. 1. Study flow chart.
4 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
Stage II cancer (64.4%), with the remainder having a diag-
nosis of Stage I. Further details of the participants includ-
ing use of antiestrogen and antidepressant medications are
provided in our earlier publication (Carlson et al., 2004).
When the 31 participants who provided full psycholog-
ical data were compared at baseline to those who did not
on demographics and psychological measures using inde-
pendent samples t tests, drop-outs had a lower global QL
(58.0 vs. 66.8, t = 2.57, p < .05), and higher total mood
disturbance (35.0 vs. 11.9, t = 2.8, p < .01). Those who
completed the study and the drop-outs were not different
on initial stress levels. Nor were they different on any of
the demographic or cancer-related variables.
3.2. Meditation practice
The 31 patients who provided complete psychological
follow-up data attended a mean 8 of 9 sessions (only 2
attended fewer than 7 sessions). They also practiced at
home as instructed, reporting an average of 24 min/day
of meditation and 13 min/day of yoga over the course of
the eight weeks.
Thirty one patients completed the practice logs at the
6-month follow-up, and reported that they spent a median
time of 7.4 h/month doing yoga and/or meditating since
the end of the program (about 1/3 yoga, 2/3 meditation).
This is approaching 2 h/week. There was a large range,
from zero (n = 3) to over 59 h/month (almost 2 h/day) in
one case. Because of the skewed distribution, the mean
was higher at 9.8 h/month (SD 13 h/month). Over the sec-
ond time period between the 6- and 12-month visits, 30
people provided data, and reported practicing a median
of 5.6 h/month, with an even larger range from 0 (n =3)
to 7 3 h/month (mean 9.0 h/month, SD 14.6 h/m).
3.3. Health behaviors
In our previous study we reported sleep quality had
improved over the course of the intervention. Caffeine serv-
ings per week decreased and exercise increased for those
who provided pre- and post-data (Carlson et al., 2003,
2004). Full data at all four time points including the
6- and 12-month follow-up was available for only 23 par-
ticipants, even though more that this did provide data at
each time point (see Fig. 1). Hence, because of the low rate
of full data repeated-measures analyses on these variables
were not performed, as interpretation may have been diffi-
cult with a sample that size, representing less that half of
those who began the program.
3.4. Psychological outcomes
The psychological effects of the program on the EORTC,
POMS and SOSI subscale scores pre- and post-intervention
are fully described in our other papers (Carlson et al., 2003,
2004). Here, we report only global scores for the partici-
pants who provided data at each time point using mixed
methods. Means for each time period are presented in
Table 2 (Means, SDs and effect sizes are presented).
3.4.1. Quality of life
In the mixed-effects model without any covariates, the
baseline versus post-comparison was significant
(F[1,94] = 7.46, p < .005), as was the post-quadratic effect
(F[1,47] = 5.08, p < .05). This indicates that the pre-inter-
vention value was different (lower QL) than the average
of the three post-intervention values, and the post-values
themselves changed in a quadratic pattern, decreasing then
increasing. When the covariates were added to the model
the main effects were no longer significant, but none of
the interactions between the time effects and covariates
were significant either.
None of the correlations between quality of life global
change scores and home practice or attendance were signif-
icant at the p < .05 level from pre- to post-intervention, or
at the 6- and 12-month follow-up assessments (all p > .10).
3.5. Mood scores
None of the contrasts were significant in the mixed-effect
model without covariates for the POMS total mood distur-
bance score. The TMD score at time 1 was 15.93, which is
Table 1
Demographics across the four time points
Pre-intervention (n = 59) Post-intervention (n = 50) 6-Month follow-up (n = 47) 12-Month follow-up (n = 41)
Type of cancer (N/%)
Breast 49 (83.1) 42 (84.0) 39 (83.0) 33 (80.5)
Prostate 10 (16.9) 9 (16.0) 8 (17.0) 8 (19.5)
Stage of cancer (N/%)
1 21 (35.6) 18 (36.0) 14 (29.8) 16 (39.0)
2 38 (64.4) 33 (64.0) 33 (70.2) 25 (61.0)
Marital status (N/%)
Married/CL 42 (69.5) 38 (76.0) 35 (74.5) 30 (73.2)
Mean SD Mean SD Mean SD Mean SD
Age (years) 54.52 10.85 55.86 10.71 54.42 9.89 55.99 10.00
Education (years) 14.71 2.79 14.57 2.74 14.74 3.02 14.71 2.95
Time since diagnosis (years) 2.05 2.96 2.11 3.13 1.90 3.11 1.92 3.14
L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx 5
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
already quite low, indicating minimal mood dist urbance.
No significant interactio n emerged with the addition of
the covariates. None of the correlati ons between changes
in overall mood disturbance scores pre- to post-interven-
tion and home practice or attendance or at the 6- and
12-month follow-up assessments were significant at the
p < .05 level (all p > .10).
3.6. Stress scores
There was a strong effect comparing baseline to the post-
intervention average scores (F[1,85] = 16.30, p < .001),
indicating that stress scores decreased over the intervention
and stayed low for the follow-up period (Fig. 2). Linear
and quadratic effects across the post-scores were not signif-
icant. The ES went from d = 0.28 at post-intervention to
d = 0.30 at 6-month and d = 0.40 after a year, a moder-
ate-sized effect. Wh en the health behaviors and disease
variables were added as interaction terms, no interactions
emerged, but the effect washed out the significance of the
previous decrease in stress symptoms after the intervention.
None of the co rrelations between stress change scores
and home practice or attendance pre- to post-intervention,
or at the 6- and 12-month follow-up assessments were sig-
nificant at the p < .05 level (all p > .10).
3.7. Biological outcomes
3.7.1. Cortisol measures
The mixed-effects models found decreases in Ln cortisol
comparing baseline to post-intervention values across the
follow-up at all time periods: 8:00 AM (F[1, 122] = 13.28,
p < .001), 2:00 PM (F[1, 82] = 4.30, p < .05), 8:00 PM
(F[1,129] = 16.27, p < .001) and across average cortisol lev-
els (F[1, 88] = 19.85, p < .001). In addition to these
decreases after the intervention, there were continued
downward linear effects for cortisol at 8:00 AM
(F[1,38] = 21.33, p < .001), 2:00 PM (F[1,43] = 22.32,
p < .001) 8:00 PM (F[1, 49] = 27.07, p < .001) and for
average cortisol values (F[1, 28] = 32.50, p < .001) across
the year of follow-up. See Fig. 3 for a depiction overall
mean cortisol values. However, cortisol slope did not
change significantly across time points. Evening cortisol
levels at time 4 were associated with stress scores
(r = .366, p < .05), but no other CRT values were signifi-
cantly associated with concurrent stress levels or home
meditation practice.
When the interaction terms were added to the models,
the time effects on 8:00 AM, 2:00 PM, 8:00 PM and mean
cortisol were washed out, but the main effect for baseline
versus post-mean cort values pe rsisted (F[1, 54] = 3.94,
p < .05). The only significant interactions were between
the baseline versus post-intervention change and time
since cancer diagnosis at 8:00 AM (F[1, 58] = 4.46,
p < .05), such that the decrease in cortisol after the inter-
vention was less for those who had been living with can-
cer for a longer period of time. As well, the interaction
between stage of disease and the post-quadratic effect on
the cortisol slope was significant (F[1,80] = 6.01,
p < .05), indicating that the decrease in cortisol daily slope
over the follow up was attenuated in those with more
advanced cancer.
Table 2
QL, mood and stress scores for participants with all four assessments (n = 31)
Measure Time 1
(pre-MBSR)
Time 2
(post-MBSR)
ES1 Time 3
(6-month follow-up)
ES2 Time 4
(12-month follow-up)
ES3
Mean SD Mean SD D Mean SD D Mean SD D
EORTC QLQ C-30 global QL 68.82 15.51 72.58 13.64 0.26 70.16 18.85 0.08 73.12 13.73 0.29
POMS TMD 11.87 27.16 11.81 33.20 0.00 11.56 34.09 0.01 7.26 32.16 0.16
SOSI total score 81.58 50.05 68.16 44.28 0.28 66.65 48.27 0.30 62.58 44.52 0.40
ES1 = (Mean time 2 Mean time 1)/Pooled SD; ES2 = (Mean time 3 Mean time 1)/Pooled SD; ES3 = (Mean time 4 Mean time 1)/Pooled SD.
12-month follow-up6-month follow-upPost-InterventionPre-Intervention
7
6
5
4
3
Mean Daily Salivery Cortisol
Fig. 3. Mean daily salivary cortisol values across time.
12-month follow-up6-month follow-upPost-InterventionPre-Intervention
80
75
70
65
SOSI Total Score Means
Fig. 2. Symptoms of stress inventory scores.
6 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
3.8. Immune measures
3.8.1. Cell counts
Over the course of the year, the number of monocytes at
the baseline compared to the post-assessment average
decreased over time (F[1, 79] = 6.74, p < .01), while eosin-
ophils increased (F[1, 155] = 21.31, p < .001). Eosinophils
also continued to increase linearly across the follow-up
assessments (F[1,64] = 6.19, p < .05). Lymphocytes, neu-
trophils and overall white blood cell counts did not change
significantly. The main effects for monocytes and eosinoph-
ils were washed out by the addition of the interaction terms
in the model, but no one interaction term was significant.
3.8.2. Lymphocytes
Within the lympho cyte subtypes (Table 3), NK cells
(CD56+) showed a significant quadratic effect across the
three post-intervention values, increasing then decreasing
over time (F[1, 74] = 5.85, p < .05). NKT cells showed both
a linear (F[1, 80] = 6.68, p < .05) and quadratic
(F[1, 116] = 4.39, p < .05) effect, increasing across the post-
intervention follow-up assessments. B cells (CD19+) chan-
ged in two ways, increasing comparing baseline to the post-
intervention scores (F[1, 142] = 6.06, p < .05) and linearly
increasing over the follow-up (F[1,108] = 7.00, p < .05).
Within the T-cell populations, total T-cells (CD3+) showed
all three significant time effects, decreasing comparing
baseline to the average of the post-scores (F[1, 115] = 4.81,
p < .05), and both linea rly (F[1, 83] = 5.91, p < .05) and
quadratically (F[1, 33] = 4.33, p < .05) over follow-up. T-
helper cells (CD4+) decreased then increased quadratically
from times 2–4 with a small effect (F[1,130] = 3.94, p < .05),
and T-cytotoxic cells (CD8+) were higher at baseline com-
pared to all three follow-up periods (F[1,129] = 12.76,
p < .001), and also decreased linearly across follow-up
assessments (F[1,89] = 4.95, p < .05).
When the interaction terms were added to the models of
lymphocyte counts, the main effects on B cells, total T-cell,
and helper T-cells were no longer significant, but the qua-
dratic effect on cytotoxic T-cells post-intervention inter-
acted with sleep (F[1,108 ] = 9.2, p < .005) such that those
with less baseline sleep showed greater decreases in cyto-
toxic T-cells over the follow-up period.
3.8.3. Cytokines
The largest magnitude changes were seen in terms of
cytokines. The percentage of T-cells expressing these cyto-
kines is also detailed in Table 3. Within the T-cell popula-
tion, IFN-k decreased substantially over the course of the
year both compared to the baseline pre-intervention value
(F[1,72] = 28.98, p < .001) and across the three follow-up
assessments both linearly (
F[1, 77] = 70.52, p < .001) and
quadratically (F[1,100] = 19.93, p < .001). A similar pat-
terns was seen for T-cell TNF production, comparing base-
line to post-intervention average scores (F[1, 70] = 28.42,
p < .001), across follow up in a linear trend
(F[1,69] = 72.28, p < .001) and less strongly in a quadratic
trend (F[1,86] = 12.58, p < .001). IL-4 showed the same
pattern with the strongest effect being a follow-up linearly
decreasing trend (F[1, 67] = 66.84, p < .001), followed by
a decrease from baseline to follow-up (F[1,62] = 29.42,
p < .001) and a quadratic trend (F[1, 138] = 4.84, p < .05).
The typical pattern of these decreases is shown for IFN-k
in Fig. 4. For NK cells, production of IFN-k also
decreased linearly across the follow-up assessments
(F[1,64] = 11.30, p < .01) as did IL-4 (F[1,65] = 4.08,
p < .05), and IL-10 showed a quadratically decreasing effect
(F[1,123] = 18.03, p < .001) (Table 4).
With health behaviors and disease variables added as
covariates, significant time effects on T-cell production of
IFN-k were still present for the linear decrease post-inter-
vention (F[1,77] = 8.14, p < 01). An interaction between
exercise and the change in T-cell IFN-k from pre- to
post-intervention appeared (F[1, 93] = 4.42, p < .05), in
that those who had exercised more showed a larger
decrease in IFN-k over time. Changes in T-cell production
of IL-4 continued to be significant, but only for the linear
decrease over the follow-up period (F[1,118] = 8.04,
p < .005). Changes in T-cell TNF were no longer signifi-
cant. For NK cells, time effects on IFN-k production
Table 3
Immune cell subtypes and cytokine expression for participants with all four assessments (n = 40)
Time 1 Time 2 Time 3 Time 4
Mean SD Mean SD Mean SD Mean SD
Lymphocytes
Total lymph (% WBC) 28.45 9.46 29.86 7.27 29.99 7.37 29.35 7.95
CD3 (% lymph) 70.22 9.10 69.90 7.56 66.85 12.72 67.76 8.51
CD4 (% lymph) 45.77 10.56 44.74 10.49 44.52 9.97 45.46 10.95
CD8 (% lymph) 26.00 7.34 25.70 7.14 24.25 7.33 24.13 8.26
CD19 (% lymph) 12.56 6.27 12.82 6.58 13.80 6.59 13.98 6.92
CD56 (% lymph) 9.34 4.34 9.87 4.48 10.21 5.81 8.45 4.78
Cytokines (% of T-cells)
IFN-c 27.58 17.35 27.52 14.51 11.31 9.00 10.75 8.72
TNF 42.08 24.47 44.40 22.84 20.37 16.35 15.17 12.32
IL-4 3.17 2.82 3.16 2.76 0.99 1.48 0.27 0.34
IL-10 2.60 1.87 3.09 2.38 2.82 4.18 2.34 3.10
L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx 7
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
became non-significant without any significant interaction
terms. For TNF, an interaction appeared between duration
of illness and the pre-intervention versus post-intervention
time effect (F[1,81] = 9.43, p < .01), such that people who
had been diagnosed with cancer for a longer period of time
showed more of a decrease in NK cell production of TNF
in response to the interven tion. The same interaction was
apparent for IL-4 (F[1, 137] = 8.78, p < .005); those with
a longer duration of cancer showed more of a decrease in
NK cell production of IL-4 from pre- to pos t-intervention.
The effects of time on IL-10 were washed out by the addi-
tion of the cova riates.
Pearson product–moment correlations were performed
between T-cell cytokine production and stress levels as well
as home meditation and yoga practice at each time point
for participants with full data. Two outliers with extreme
scores on home practice were removed from the analysis
to create normal distributions. At time 1 (pre-intervention)
higher TNF production was associated with higher stress
levels (r = .385, p < .05), and at time 3, IL-4 production
was correlated with total stress scores (r = .369, p < .05).
There were no associations between immune measures
and home practice at any time point.
3.8.4. Blood pressure and heart rate
Values for BP and HR are presented for each time per-
iod in Table 3. Because this data has not yet been published
for the pre- and post-assessments, more detailed analyses
are presented for this outcome. The pre- and post-assess-
ment measures were averaged as detailed in the data anal-
ysis section and compared using paired-samples t tests on
SBP, DBP and HR. The overall SBP pre and post-mea-
sures were different from one another, indicating an overall
decrease in SBP (t[1, 44] = 2.02, p < .05) from 119.7 to
117.6 mm Hg. None of the other measures changed pre-
to post-intervention.
The mixed-effects models found a significant effect com-
paring baseline to the average of the follow-up assessments
for SBP (F[1, 85] = 6.32, p < .05), indicating that the
change over the course of the intervention was also signif-
icant when compared to the average of the foll ow-up mea-
sures. Overall resting heart rate decreased both from
baseline across the follow-ups (F[1,90] = 4.17, p < .05),
but also continued a linear decrease over the three fol-
low-up assessments (F[1, 67] = 5.86, p < .05) from about
72 beats per minute at pre-intervention to 68 beats after
one year. When effects of health behaviors and disease
variables were added to the equations, the baseline to
post-intervention change in SBP was washed out by an
interaction with duration of illness (F[1,88] = 5.46,
p < .05) such that those living with cancer longer showed
smaller decreases in SBP after the intervention. With the
addition of the covariates, the time main effects in the
model were no longer significant for HR.
Correlations at each of the four common assessment
periods between BP and HR values and the total scores
of the POMS, SOSI and EORTC were significant between
HR and the SOSI total (r = .28, p < .05) and POMS total
(r = .38, p < .01) at pre-intervention, and between HR
and SOSI total at 6- (r = .44, p < .01) and 12-month
(r = .37, p < .05) follow-ups. This indicates that elevated
resting HR was related to reporting more symptoms of
mood disturbance pre-intervention, and of stress at three
separate test times.
12-month follow-up6-month follow-upPost-InterventionPre-Intervention
30
28
26
24
22
20
18
16
IFN-Gamma production (% T cells)
Fig. 4. Mean IFN-k T-cell production across time.
Table 4
BP and HR measures
Systolic BP Diastolic BP HR
1. One week pre-first class Mean 120.7
a
78.4 72.7
SD 15.3 10.0 8.6
2. Immediately pre-first class Mean 116.4
a
73.5 74.0
SD 12.7 8.7 7.9
3. Immediately post-last class Mean 118.8
a
76.5 70.4
SD 14.5 8.4 8.6
4. One week post-last class Mean 114.7
a
75.2 74.5
SD 11.9 8.8 8.0
5. 6-Month follow-up Mean 115.4 76.3 71.4
SD 10.5 8.1 8.8
6. 12-Month follow-up Mean 116.3 74.3 68.6
SD 12.8 8.0 7.2
a
Average of assessments 1 and 2 SBP higher than average of assessments 3 and 4, p < .05.
8 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
4. Discussion
The results of this pre–post intervention and one-year
follow-up study indicate that this 8-week mindfulness-
based stress reduction program was effective in decreasing
symptoms of stress in this group of breast and prostate
cancer patients, and maintaining these improvements over
a year of follow-up. The effect-size was moderate
(d = 0.40), which typically represents the lower range of
clinically significant improvement. These improvemen ts
were independent of the amount of subsequent home prac-
tice of meditation and yoga. Improvements were also seen
across the year in global QL scores, but not to the same
extent as for stress symptoms. However, no significant
changes were seen on mood disturbance scores of these
patients, which may be explained by the low level of initial
mood disturbance. These results are similar to those
observed in the patients who pro vided pre- and post-inter-
vention data (Carlson et al., 2003). Participants who
dropped out of the program or did not complete full fol-
low-up assessments were more distressed at baseline than
those who provided full data, so it may be the case that
optimal benefit will accrue for those with mild-moderate
levels of distress, rather than for those with very low or
very high levels. This has been suggested by many other
researchers as the optimal target population for psychoso-
cial group interventions (e.g. Goodwin et al., 2001; Ross
et al., 2002).
Cortisol levels also continued to decrease over the fol-
low-up period, with all of the morning, afternoon and
evening as well as mean cortisol values decreasing both
from pre- to post-intervention, and also decreasing linearly
across the year of follow-up. The most robust finding after
controlling for a number of health behaviors and disease
characteristics was the decrease in the mean daily cortisol
value. There was also less of a decrease in cortisol levels
over time in people who had a higher stage of cancer and
had been living with the illness for a longer period of time,
perhaps indicating that stress-related changes in cortisol
are more possible earlier in the survivorship period for
those with less disease burde n. However, due to the single
data of salivary cortisol collection and known day-to-day
variability in cortisol production (Kirschbaum and Hell-
hammer, 1994), these findings should be verified with more
rigorous cortisol data collection methods involving multi-
ple days of testing at each assessment period. Indeed, a
recent report co mparing test-retest reliability of cortisol
slopes in older participants (average age 61.5 years, 88%
female) concluded that minimally two and preferably three
days of salivary testing be utilized (Kraemer et al., 2006).
The correlations between slopes on two different days sup-
porting this conclusion were in the 0.5–0.6 range. As a
result, in our more current protocols this method is being
used.
Nonetheless, if abnormal or elevated cortisol levels are a
marker of poor prognosis in cancer patients, as suggested
by previous work (Sephton and Spiegel, 2003; Sephton
et al., 2000), this effect, if verified , could prove significant.
Indeed, although very few follow-up studies have been con-
ducted on the effects of meditation on disease course, pre-
hypertensive patients who participated in meditation
interventions an average of eight years previously benefited
from a decrease in death rates due to cancer of 49% co m-
pared to randomized control groups (Schne ider et al.,
2005). This was even larger than the decrease in cardiac
mortality of 30%, and represents a very large reduction in
cancer mortality risk in this sample of 202 relatively
healthy participants. The only other long-term follow-up
of a meditation intervention was in a small sample of
patients with irritable bowl syndrome, who showed contin-
ued reductions in pain and bloating over the year (Keefer
and Blanchard, 2002).
In terms of immune function, we observed interesting
downward trends in the T-cell production of pro-inflam-
matory cytokines. Pro-inflammatory cytokines have been
associated with increa sed stress levels (Anisman and Meral-
i, 2003) and are elevated in patients with depression and
heart disease (Joynt et al., 2004; O’Co nnor and Joynt,
2004). Administration of pro-inflammatory cytokines
(IFN-a) for treatment of melanoma reliability produces
depression in up to 50% of patients (Capuron and Miller,
2004). Based on these and other lines of evidence, several
recent reviews highlight the putative role of pro-inflamma-
tory cytokines in the etiology of depression, possibly
through altering HPA axis reactivity, down regulating
serotonin precursors, and impairing processes of neurogen-
esis (Raison et al., 2006; Hayley et al., 2005; Hayley and
Anisman, 2005; Schiepers et al., 2005). Consistent with
these hypotheses, higher levels of proinflammatory cy to-
kines were associated with greater self-reported stress
symptoms (which are highly correlated with mood distur-
bance) on follow-up assessments in this study.
Despite these associations between cytokines and stress
levels, there were no consistent relationships between home
meditation or yoga practice and changes in cytokines. This
in combination with the lack of associations between stress
scores and meditation practice may suggest a more general
pattern of recovery from the stress of cancer diagnosis and
treatment, rather than one specifically due to the meditation
practice. It may also be the case that once improvements
were instigated through the MBSR program, they were
self-sustaining and further formal meditation practice was
not necessary. Indeed, a large component of the MBSR pro-
gram involves instilling attitudes of non-judging, accep-
tance and patience, which may allow people to live more
at ease even without formal daily practice. Another impor-
tant part of the program is practicing ‘‘mini’’ meditatio n
exercises that take only several seconds to minutes (such
as slowing and counting breaths), and what is known as
‘‘informal’’ practice throughout the day, which includes
awareness of patterns of stress reactivity. It may be that this
more mindful approach to life was responsible for some of
the changes observed in these participants. These questions
remain to be explored in further research.
L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx 9
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
The continual drop of cytokine production observed
over the 6- and 12-month follow-up was not directly
related to decreases in stress, as most of the stress reduction
occurred immediately following MBSR program participa-
tion. It is possible that changes in cytokine production
could reflect improvement in disease status at the 6- and
12-month follow-up time points, continued physiological
recovery from the effects of treatment, or a delay in the
effects of stress reduction. There was no disease recurrence
in any of the patients who provided cytokine data at all
time points and no new treatments were started or current
treatments remained the same (for those taking tamoxifen).
Future studies examining longitudinal cytokine profiles of
patients who do not choose to participate in mindfu lness
meditation could help to determine how these patients
compare.
We also saw decreases in SBP in patients over the course
of the intervention, which were maintained ov er follow-up,
but attenuated in people who had been living with cancer
for a longer period of time. This decrease is especially strik-
ing given that participants entered the study with essen-
tially normal BP levels (average less than 120/80), and
had resting HRs in the low 70 s. Thus, it would not be
expected that very dramatic changes would be associated
with participation in the MBSR program. This SBP
decrease in our participants of an average of 6 mm Hg is
consistent with other research. Wenneberg and colleagues
recorded an average decrease of 9 mm Hg in ambulatory
DBP in normotensive male volunteers after four months
of meditation (Wenneberg et al., 1997). In a group of Afri-
can Americans with initially high BP, decreases of 10/
6 mm Hg were seen in resting SBP and DBP, respectively
(Schneider et al., 1995). As well, in a group of Thai college
students who practiced meditation intensively for two
months, SBP decreased about 5 mm Hg, and DBP
decreased about 6 mm Hg (Sudsuang et al., 1991). Simi-
larly, SBP decreased over 3-month in groups of elderly
nursing home residents randomly assigned to either tran-
scendental meditation (TM) or mindfulness meditation,
compared to relaxation and no treatment controls (Alexan-
der et al., 1989). Barnes et al. also found that even during a
single meditation session, experienced TM practitioners’
SBP dropped 3.0 mm Hg, whereas for healthy non-medi-
tating controls SBP increased 2.1 mm Hg during eyes-
closed relax ation (Barnes et al., 1999). These studies all
point to potentially beneficial effects of meditation on BP
in a variety of groups of healthy subjects of varying ages,
and support the magnitude of the findings of the current
study.
There were, however, no associations between changes
in BP over time and meditation practice, so we were simi-
larly unable to document any ‘‘dose-dependent’’ changes in
BP related to the amount of home practice. There was also
a relationship found between age and SBP, in that older
participants tended towards higher SBP. This is consistent
with very commonly observed increases in BP with advanc-
ing age (van Boxtel et al., 1996; Tanaka et al., 2001).
HR was lowest when measured immediately following
the last class, which is likely an indication of the immediate
effect of the preceding meditation session, potentially an
indication of activation of the relaxation response (Benson,
1975). Similarly, decreases in HR occurred during TM ses-
sions in two studies of healthy volunteers who were experi-
enced in TM techniques (Travis and Wallace, 1997; Telles
et al., 1998). Another interesting finding was that resting
HR was related to self-reported symptoms of stress at sev-
eral time points. This may be an indication of psychological
stress resulting in physiological manifestations in the form of
increased HR. One may argue the possibility that comple-
tion of the questionnaires and thinking about symptoms
of stress in close temporal proximity to the autonomic mea-
sures may have elevated HR. Procedurally, however, at the
first assessment patients had their BP and HR taken prior to
completing the questionnaires, and for the other three test
sessions, patients completed the questionnaires at home
some time before having their HR measured, and in all cases
sat quietly for at least five minutes before the measurements,
which argues against that possibility. It may be the case,
then, that people who subjectively feel that they are experi-
encing high levels of stress also tend to have a higher resting
HR even when not directly thinking about their stress.
The major methodological limitation of this study is the
lack of a control or comparison group. Overall, this limits
our ability to infer causation of any changes observed to
participation in the MBSR program. For example, it is
possible that stre ss and/or BP levels could have been arti-
factually elevated prior to program participation due to
the novelty of beginni ng a new treatment program and
being in the hospital setting (i.e. ‘‘white-coat hyperten-
sion’’). This is part of the reason that we chose to take sev-
eral pre- and post-measures, and conducted analyses on the
mean scores of participants. Nonetheless, the results
reported here must be considered preliminary and hypoth-
esis generating. Nonspecific factors such as expectancy and
the therapeutic alliance also cannot be controlled in a sin-
gle-group design and may play an important role in the
beneficial results found. Another issue related to the non-
specificity of the intervention is that even if the benefic ial
effects were due to the inter vention and not laboratory
induced, the relative importance of the different compo-
nents of the program cannot yet be ascertained. Whether
the most effective components are the meditation, yoga,
social support, group processes, professional attention, or
other factors will have to await further ‘‘dismantling’’ stud-
ies of MBSR. In all likelihood, the most useful aspects vary
from person to person depending on the individual’s needs,
background and personality.
Another obvious limitation of this approach is the mul-
tiple statistical comparisons that were conducted, as there
were a high number of outcome measures employed. This
elevates the chances of Type I error—falsely identifying
significant findings in their absence—but in the context
of a hypothesis-generating study such as this there is typi-
cally a higher tolerance for this type of error. Future
10 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
hypothesis-testing studies can further investigate promising
relationships identified in this research. Due to the rela-
tively small sample size, there is also the danger of Type
II error—failing to identify significant relationships that
are present. Indeed, because of sample size limitations we
were not able to test for several comparisons of interest,
such as between the men and women (there were only 10
men at baseline) or to include other potentially important
covariates such as cancer treatments, marital status or
other health behaviors. This study suffers from some degree
of both lack of power and a large number of analyses, so
the results must be co nsidered in this light.
In summary, this study con firmed our previous findings
of decreases in stress symptoms after participation in an
MBSR program which were maintained over a one year
follow-up. Cortisol levels continued to drop on average,
and pro-inflammatory cytokine production decreased over
the full year of follow-up. SBP decreased somewhat pre- to
post-intervention overall in these already normotensive
individuals, and HR decreas ed immediately following a
meditation session. HR and cytokine levels were also asso-
ciated with self-reported stress symptoms. Changes seen in
these patients were moderate. This may have been due to
the high levels of patient functioning at the start of the
study, but could also indicate that this type of program is
only moderately effective for early stage breast and prostate
cancer patients who are three months or more post-treat-
ment. Nonetheless, these data are intriguing for a prelimin-
ary uncontrolled study of the immune, endocrine and
autonomic effects of the MBSR program in cancer patients.
Future studies of this nature would benefit from a random-
ized control group and from screening for moderately dis-
tressed individuals at the start of the program. Such
individuals are more likely to benefit in terms of improve-
ments in stress symptoms, mood and quality of life, as well,
perhaps, in enhancement of autonomic functioning.
Acknowledgments
This study was supported by the Canadian Breast Can-
cer Research Initiative. Dr. Linda Carlson was a Terry Fox
Postdoctoral Research Fellow of the National Cancer
Institute of Canada during the time the study was con-
ducted. She is currently a Canadian Institutes of Health
Research New Investigator. Dr. Kamala Patel is an Alber-
ta Heritage Foundation for Medical Research Senior Scho-
lar and holds a Canada Research Chair. Heartfelt thanks
to all the men and women who participated in the study,
whose enthusiasm continues to inspire us. Special thanks
to research nurse Ms. Lori Tillotson for her tireless efforts
in assuring the smooth running of the study, and research
assistant Ms. Jodi Cullum for data input and management.
References
Aaronson, N.K., Ahmedzai, S., Bergman, B., Bullinger, M., Cull, A.,
Duez, N.J., et al., 1993. The European organization for research and
treatment of cancer QLQ-C30: a quality-of-life instrument for use in
international clinical trials in oncology. J. Natl. Cancer Inst. 85, 365–
376.
Aaronson, N.K., Ahmedzai, S., Bullinger, M., Crabeels, D., Estape, J.,
Filiberti, A., et al., 1991. The EORTC core quality of life question-
naire: interim results of an international field study. In: Osoba, D.
(Ed.), Effect of Cancer on Quality of Life. CRC Press, London, pp.
185–203.
Alexander, C.N., Langer, E.J., Newman, R.I., Chandler, H.M., Davies,
J.L., 1989. Transcendental meditation, mindfulness and longevity: an
experimental study with the elderly. J. Pers. Soc. Psychol. 57, 950–964.
Anisman, H., Merali, Z., 2003. Cytokines, stress and depressive illness:
brain–immune interactions. Ann. Med. 35, 2–11.
Baer, R.A., 2003. Mindfulness training as clinical intervention: a
conceptual and empirical review. Clin. Psychol. Sci. Pract. 10, 125–143.
Barnes, V.A., Treiber, F.A., Turner, R., Davis, H., Strong, W.B., 1999.
Acute effects of transcendental meditation on hemodynamic function-
ing in middle-aged adults. Psychosom. Med. 61, 525–531.
Benson, H., 1975. The Relaxation Response. Morrow, New York.
Bishop, S.R., 2002. What do we really know about mindfulness-based
stress reduction? Psychosom. Med. 64, 71–83.
Blumenthal, J.A., Sherwood, A., Gullette, EC., Georgiades, A., Tweedy,
D., 2002. Biobehavioral approaches to the treatment of essential
hypertension. J. Consult Clin. Psychol. 70, 569–589.
Campbell, N.R., Ashley, M.J., Carruthers, S.G., Lacourciere, Y., McKay,
D.W., 1999. Lifestyle modifications to prevent and control hyperten-
sion. 3. Recommendations on alcohol consumption. Canadian Hyper-
tension Society, Canadian coalition for high blood pressure prevention
and control, laboratory centre for disease control at Health Canada,
heart and Stroke Foundation of Canada. CMAJ 160, S13–S20.
Campbell, NR., Burgess, E., Choi, BC., Taylor, G., Wilson, E., Cleroux,
J., et al., 1999. Lifestyle modifications to prevent and control
hypertension: methods and an overview of the Canadian recommen-
dations. CMAJ 160, S1–S6.
Capuron, L., Miller, A.H., 2004. Cytokines and psychopathology: lessons
from interferon-alpha. Biol. Psychiatry 56, 819–824.
Carlson, L.E., Garland, S.N., 2005. Impact of mindfulness-based stress
reduction (MBSR) on sleep, mood, stress and fatigue symptoms in
cancer outpatients. Int. J. Behav. Med. 12, 278–285.
Carlson, L.E., Speca, M., Patel, K.D., Goodey, E., 2004. Mindfulness-
based stress reduction in relation to quality of life, mood, symptoms of
stress and levels of cortisol, dehydroepiandrosterone-Sulftate
(DHEAS) and melatonin in breast and prostate cancer outpatients.
Psychoneuroendocrinology 29, 448–474.
Carlson, L.E., Speca, M., Patel, K.D., Goodey, E., 2003. Mindfulness-
based stress reduction in relation to quality of life, mood, symptoms of
stress, and immune parameters in breast and prostate cancer outpa-
tients. Psychosom. Med. 65, 571–581.
Carlson, L.E., Ursuliak, Z., Goodey, E., Angen, M., Speca, M., 2001. The
effects of a mindfulness meditation-based stress reduction program on
mood and symptoms of stress in cancer outpatients: six month follow-
up. Support. Care. Cancer 9, 112–123.
Cassileth, B.R., Lusk, E.J., Brown, L.L., Cross, P.A., 1985. Psychosocial
status of cancer patients and next of kin: normative data from the
profile of mood states. J. Psychosoc. Oncol. 3, 99–105.
Davidson, R.J., Kabat-Zinn, J., Schumacher, J., Rosenkranz, M., Muller,
D., Santorelli, S.F., et al., 2003. Alterations in brain and immune
function produced by mindfulness meditation. Psychosom. Med. 65,
564–570.
Goodwin, P.J., Leszcz, M., Ennis, M., Koopmans, J., Vincent, L., Guther,
H., et al., 2001. The effect of group psychosocial support on survival in
metastatic breast cancer. N. Engl. J. Med. 345, 1726.
Grossman, P., Niemann, L., Schmidt, S., Walach, H., 2004. Mindfulness-
based stress reduction and health benefits. A meta-analysis. J.
Psychosom. Res. 57, 35–43.
Hayley, S., Anisman, H., 2005. Multiple mechanisms of cytokine action in
neurodegenerative and psychiatric states: neurochemical and molecu-
lar substrates. Curr. Pharm. Des. 11, 947–962.
L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx 11
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
Hayley, S., Poulter, M.O., Merali, Z., Anisman, H., 2005. The pathogen-
esis of clinical depression: stressor- and cytokine-induced alterations of
neuroplasticity. Neuroscience 135, 659–678.
Joynt, K.E., Whellan, D.J., O’Connor, C.M., 2004. Why is depression bad
for the failing heart? A review of the mechanistic relationship between
depression and heart failure. J. Card Fail. 10, 258–271.
Kabat-Zinn, J., 1990. Full Catastrophe Living: Using the Wisdom of Your
Body and Mind to Face Stress, Pain and Illness. Delacourt, New York.
Keefer, L., Blanchard, E.B., 2002. A one year follow-up of relaxation
response meditation as a treatment for irritable bowel syndrome.
Behav. Res. Ther. 40, 541–546.
Kirschbaum, C., Hellhammer, D.H., 1994. Salivary cortisol in psycho-
neuroendocrine research: recent developments and applications. Psy-
choneuroendocrinology 19, 313–333.
Kraemer, H.C., Giese-Davis, J., Yutsis, M., O’hara, R., Neri, E.,
Gallagher-Thompson, D., et al., 2006. Design decisions to optimize
reliability of daytime cortisol slopes in an older population. Am. J.
Geriatr. Psychiatry 14, 325–333.
Laird, N.M., Ware, J.H., 1982. Random effects models for longitudinal
data. Biometrics 38, 963–974.
Leckie, M.S., Thompson, E., 1979. Symptoms of Stress Inventory.
University of Washington, Seattle, WA.
Little, R.J.A., Rubin, D.B., 1987. Statistical Analysis with Missing Data.
Wiley and Sons, New-York.
Mackenzie, M.J., Carlson, L.E., Speca, M., 2005. Mindfulness-based
stress reduction (MBSR) in oncology: rationale and review. Evid.
Based Integr. Med. 2, 139–145.
McNair, D.A., Lorr, M., Droppelman, L.F., 1971. Profile of Mood States.
Educational and Industrial Testing Service, San Diego.
O’Connor, C.M., Joynt, K.E., 2004. Depression: are we ignoring an
important comorbidity in heart failure? J. Am. Coll. Cardiol. 43, 1550–
1552.
Ott, M.J., Norris, R.L., Bauer-Wu, S.M., 2006. Mindfulness meditation
for oncology patients. Integr. Cancer Ther. 5, 98–108.
Raison, C.L., Capuron, L., Miller, A.H., 2006. Cytokines sing the blues:
inflammation and the pathogenesis of depression. Trends Immunol.
27, 24–31.
Robert McComb, J.J., Tacon, A., Randolph, P., Caldera, Y., 2004. A
pilot study to examine the effects of a mindfulness-based stress-
reduction and relaxation program on levels of stress hormones,
physical functioning, and submaximal exercise responses. J. Altern.
Complement. Med. 10, 819–827.
Robinson, F.P., Mathews, H.L., Witek-Janusek, L., 2003. Psycho-
endocrine-immune response to mindfulness-based stress reduction in
individuals infected with the human immunodeficiency virus: a
quasiexperimental study. J. Alter. Complement. Med. 9, 683–694.
Ross, L., Boesen, E.H., Dalton, S.O., Johansen, C., 2002. Mind and
cancer: does psychosocial intervention improve survival and psycho-
logical well-being? Eur. J. Cancer 38, 1447–1457.
Satterthwaite, F.E., 1946. An approximate distribution of variance
estimates. Biometrics Bull. 2, 110–114.
Schiepers, O.J., Wichers, M.C., Maes, M., 2005. Cytokines and major
depression. Prog. Neuropsychopharmacol. Biol. Psychiatry 29, 201–
217.
Schneider, R.H., Alexander, C.N., Staggers, F., Rainforth, M., Salerno,
J.W., Hartz, A., et al., 2005. Long-term effects of stress reduction on
mortality in persons > or = 55 years of age with systemic hypertension.
Am. J. Cardiol. 95, 1060–1064.
Schneider, R.H., Staggers, F., Alxander, C.N., Sheppard, W., Rainforth,
M., Kondwani, K., et al., 1995. A randomised controlled trial of stress
reduction for hypertension in older African Americans. Hypertension
26, 820–827.
Sephton, S., Spiegel, D., 2003. Circadian disruption in cancer: a
neuroendocrine-immune pathway from stress to disease? Brain Behav.
Immun. 17, 321–328.
Sephton, S.E., Sapolsky, R.M., Kraemer, H.C., Spiegel, D., 2000. Diurnal
cortisol rhythm as a predictor of breast cancer survival [see comment].
J. Natl. Cancer Inst. 92 (12), 994–1000.
Shan, K., Lincoff, AM., Young, J.B., 1996. Anthracycline-induced
cardiotoxicity. Ann. Intern. Med. 125, 47–58.
Smith, J.E., Richardson, J., Hoffman, C., Pilkington, K., 2005. Mindful-
ness-based stress reduction as supportive therapy in cancer care:
systematic review. J. Adv. Nurs. 52, 315–327.
Speca, M., Carlson, L.E., Goodey, E., Angen, M., 2000. A randomized,
wait-list controlled clinical trial: the effect of a mindfulness meditation-
based stress reduction program on mood and symptoms of stress in
cancer outpatients. Psychosom. Med. 62, 613–622.
Spence, J.D., Barnett, P.A., Linden, W., Ramsden, V., Taenzer, P., 1999.
Lifestyle modifications to prevent and control hypertension. 7.
Recommendations on stress management. Canadian Hypertension
Society, Canadian coalition for high blood pressure prevention and
control, laboratory centre for disease control at Health Canada, heart
and Stroke Foundation of Canada. CMAJ 160, S46–S50.
Sudsuang, R., Chentanez, V., Veluvan, K., 1991. Effect of Buddhist
meditation on serum cortisol and total protein levels, blood
pressure, pulse rate, lung volume and reaction time. Physiol. Behav.
50, 543–548.
Tanaka, H., Dinenno, F.A., Monahan, K.D., DeSouza, C.A., Seals, D.R.,
2001. Carotid artery wall hypertrophy with age is related to local
systolic blood pressure in healthy men. Arterioscler. Thromb. Vasc.
Biol. 21, 82–87.
Telles, S., Nagarathna, R., Nagendra, H.R., 1998. Autonomic changes
while mentally repeating two syllables—one meaningful and the other
neutral. Indian J. Physiol. Pharmacol. 42, 57–63.
Travis, F., Wallace, R.K., 1997. Autonomic patterns during respiratory
suspensions: possible markers of Transcendental Consciousness. Psy-
chophysiology 34, 39–46.
Vallebona, A., 2000. Cardiac damage following therapeutic chest irradi-
ation. Importance, evaluation and treatment. Minerva Cardioangio-
logica 48, 79–87.
van Boxtel, M.P., Gaillard, C., van Es, P.N., Jolles, J., de Leeuw, P.W.,
1996. Repeated automatic versus ambulatory blood pressure measure-
ment: the effects of age and sex in a normal ageing population. J.
Hypertens. 14, 31–40.
Wenneberg, S.R., Schneider, R.H., Walton, K.G., MacLean, C.R.,
Levitsky, D.K., Salerno, J.W., et al., 1997. A controlled study of the
effects of the Transcendental Meditation program on cardiovascular
reactivity and ambulatory blood pressure. Int. J. Neurosci. 89, 15–28.
12 L.E. Carlson et al. / Brain, Behavior, and Immunity xxx (2007) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Carlson, L.E. et al., One year pre–post intervention follow-up of psychological, ..., Brain Behav.
Immun. (2007), doi:10.1016/j.bbi.2007.04.002
    • "A large body of research has documented the effectiveness of mindfulness-based interventions for a range of clinical disorders (Bowen et al., 2014; Khoury et al., 2013; Masuda & Hill, 2013 ). In nonclinical samples, mindfulness-based interventions have been associated with improved attention and memory (Chiesa, Calati, & Serretti, 2011), psychological wellbeing and physical health (Carlson, Speca, Faris, & Patel, 2007; Davidson et al., 2003; Keng, Smoski, & Robins, 2011 ). Mindfulness appears beneficial for individuals across the lifespan from young children (Flook, Goldberg, Pinger, & Davidson, 2015) to older adults (Moynihan et al., 2013). "
    [Show abstract] [Hide abstract] ABSTRACT: Interest in the application of mindfulness-based intervention for the treatment of psychological disorders and promotion of wellbeing has grown exponentially in recent years. Mindfulness-based interventions have been found to be beneficial for treatment of various forms of psychopathology as well as improve psychological wellbeing and enhance physical health. Little research has investigated for whom and under what conditions training people to use mindfulness-based therapeutic techniques is most effective. Recent studies have found evidence that individual differences in personality traits are associated with mindfulness. For example, neuroticism has been found to be negatively associated with mindfulness. These associations raise the possibility that individual differences in personality may potentially moderate the effectiveness of mindfulness-based interventions. In the present study we draw on Gray's revised Reinforcement Sensitivity Theory (r-RST) to examine relationships between personality traits, mindfulness and psychological distress. We found that the Flight, Fight, Freeze system mediated the relationship between trait mindfulness and psychological distress, while trait mindfulness moderated the relationship between the Flight, Fight, Freeze system and psychological distress. Both results are consistent with the suggestion that acquiring the skills from learning and practicing mindfulness techniques is potentially useful, particularly for threat-sensitive individuals with low to moderate levels of trait mindfulness.
    Full-text · Article · Sep 2016
    • "For example , slow deep breathing has been associated with reductions in blood pressure, oxidative stress, and balance in the autonomic nervous system and may contribute to an improved emotional state [28][29][30][31][32][33] . Further, mindfulness-based interventions alone [34][35][36][37]or combined with movement [31] may reduce anxiety and depression. Mindfulness practice (e.g., mindfulness-based stress reduction) that emphasizes nonactive sessions (but often includes some gentle Yoga components ) has been found to alleviate symptoms for breast cancer survivors [38]. "
    [Show abstract] [Hide abstract] ABSTRACT: Introduction: Breast cancer survivors (BCSs) often report fatigue that persists for years following treatment. Despite a growing body of evidence for meditative movement practices to improve symptoms among BCSs, few studies have explored using Qigong/Tai Chi to reduce fatigue. Additionally, few have examined the biological mechanisms through which fatigue may be reduced using Qigong/Tai Chi. Methods/study design: We will recruit 250 fatigued, post-menopausal women diagnosed with breast cancer (stage 0-III), between 6 months and 5 years past primary treatment and randomize to a standardized Qigong/Tai Chi Easy (QG/TCE) intervention, a "sham" Qigong group (movements without a focus on the breath and meditative state) (SQG), or an educational support (ES) group. The primary outcome (fatigue), secondary outcomes (anxiety, depression, sleep quality, cognitive function, physical activity), and a biomarker of HPA axis dysregulation (diurnal cortisol) will be assessed at baseline, post-intervention and 6 months postintervention, and biomarkers of inflammation (IL1ra, IL6, TNFα and INFᵧ) at pre/post-intervention. We hypothesize that QG/TCE will reduce fatigue (and improve other symptoms associated with fatigue) in BCSs experiencing persistent cancer-related fatigue more than SQG and ES. Biomarkers will be examined for relationships to changes in fatigue. Conclusions: Findings from this study may reveal the effects of the unique mind-body aspects of QG/TCE on fatigue in BCSs with a complex design that separates the effects of low-intensity physical activity (SQG) and social support/attention (ES) from the primary intervention. Further, results will likely contribute greater understanding of the biological mechanisms of these practices related to improved symptoms among BCSs.
    Article · Aug 2016
    • "Moreover, the studies also demonstrate that participation in mindfulness trainings significantly reduces the symptoms of depression [37–40, 43, 44], and the level of subjectively experienced pain [45]. What encourages one to study the matter further is the fact that several studies demonstrated that apart from the change in the subjective assessment of particular psychological factors, there was also significant reduction in blood cortisol levels, which may be treated as an objective indicator of the stress experienced [40, 42]. Table 2 presents studies involving MBSR training in oncological patients. "
    Article · Jun 2016
Show more