Mindfulness-Based Stress Reduction Training Reduces Loneliness and Pro-Inflammatory Gene Expression in Older Adults: A Small Randomized Controlled Trial

Abstract

“Usually we regard loneliness as an enemy. Heartache is not something we choose to invite in. It's restless and pregnant and hot with the desire to escape and find something or someone to keep us company. When we can rest in the middle [through meditation practice], we begin to have a nonthreatening relationship with loneliness, a relaxing and cooling loneliness that completely turns our usual fearful patterns upside down”

-- Pema Chodron (2000), Buddhist nun and teacher

Feeling lonely is a significant risk factor for morbidity and mortality in older adults. For example, lonely older adults have increased risk for cardiovascular disease (Olsen et al., 1991; Thurston and Kubzansky, 2009), Alzheimer's disease (Wilson et al., 2007), and all-cause mortality (Tilvis et al., 2011). Developing effective treatments to reduce loneliness in older adults is thus essential, but previous treatment efforts have had limited success (Findlay, 2003; Masi et al., 2011). As Pema Chodron suggests above, meditation practice may provide a middle way for reducing one's feelings of loneliness (Chodron, 2000). Loneliness has been described as a state of social distress that arises when there is a discrepancy between one's desired and actual social relationships (Russell et al., 1980; Cacioppo and Hawkley, 2009). Previous mindfulness meditation studies using the standardized 8-week Mindfulness-Based Stress Reduction (MBSR) program show that MBSR reduces measures of distress and negative affect in healthy and patient populations (for reviews, Brown et al., 2007; Hölzel et al., 2011) and can improve social relationship functioning in couples (Carson et al., 2004), although no studies have tested whether the MBSR program can reduce loneliness. Our primary aim was thus to test whether MBSR reduces loneliness in a small randomized controlled trial in older adults (N=40).

If the MBSR program reduces loneliness, it offers the intriguing possibility that mindfulness meditation training may also alter gene expression dynamics and protein markers of inflammation (i.e., C Reactive Protein (CRP) and Interleukin-6 (IL-6)) that are implicated in the physical health risks observed in lonely older adults (Cole et al., 2007). Several studies indicate that MBSR may reduce protein biomarkers of inflammation (Carlson et al., 2003, 2007; Lengacher et al., 2012), and inflammation is known to play a major role in structuring the development and progression of many diseases that drive late-life morbidity and mortality (Finch, 2007). Moreover, recent research shows that immune cells from lonely older adults have increased expression of genes involved in inflammation (Cole et al., 2007, 2011). Bioinformatic analyses of the signaling pathways regulating gene expression suggest that loneliness may activate a biological defensive program mediated by the pro-inflammatory transcription factor, NF-κB, in monocytes (Cole et al., 2007, 2011; Irwin and Cole, 2011; Antoni et al., 2012). Therefore we also tested whether MBSR reduces loneliness-related pro-inflammatory gene expression and circulating protein biomarkers of inflammation (as measured by CRP and IL-6).

Methods

Results

Preliminary Analyses

The MBSR and WL groups did not significantly differ on 131any measured demographic characteristics at baseline (see Table 1), indicating success of randomization. A CONSORT flowchart (Figure 1) depicts the flow of participants retained at each phase of the trial. Six participants dropped out of the study prior to completing post-treatment measures, reflecting a 15% dropout rate, which is comparable to attrition rates observed in published MBSR studies (Baer, 2003). More participants dropped out of the MBSR treatment group (N=5) compared to WL (N=1), this difference was marginally significant (χ²(1)= 3.04, p= 0.07). We note that this difference became approximately equal when five (N=5) WL participants dropped out of the MBSR program when it was offered to them after completing the primary treatment trial. These results suggest that the attention and behavioral training demands of MBSR produce modest experimental mortality compared to no treatment. Study dropouts in the primary treatment trial (N=6) were compared to treatment completers (N=34) on all measured baseline demographic characteristics and no significant differences emerged. However, the one exception was that males in the study were more likely to drop out (3/8) compared to females (3/32) (χ²(1)= 3.78, p= 0.05).

Open in a separate window

Figure 1

CONSORT flowchart of participants retained at each phase of the trial.

Table I

Baseline Characteristics of Randomized Participants.

Characteristic	MBSR Group	WL Group	Difference Statistic
Age [mean years (SD)]	64.35 (6)	65.16 (8)	t(37)= -.34, p= 0.74
Gender			χ2(1)=.33, p= 0.57
Male	3	5
Female	17	15
Ethnicity			χ2(5)= 3.55, p= 0.62
Caucasian	13	13
African American	2	3
Asian American	2	1
Latino(a)	3	1
Native American	0	1
Other	0	1
Employment Status			χ2(3)= 5.33, p= 0.15
Retired	10	4
Unemployed	0	1
Part-time	5	4
Full-time	5	10
Education			χ2(3)= 2.12, p= 0.55
High school degree	0	1
Some college	6	3
College degree	3	4
Graduate work	11	11
Body Mass Index	25.22 (4)	25.22 (4)	t(37)= .003, p= 0.99
Cognitive Impairment (MMSE)	28.00 (2)	27.84 (2)	t(37)= .24, p= 0.81
Log-transformed CRP	.15 (.58)	-.02 (.74)	t(37)= .86, p= 0.39
Log-transformed IL-6	.10 (.42)	.03 (.34)	t(37)= -.58, p= 0.24

Open in a separate window

Notes: Standard deviation values are provided in parentheses in the MBSR group and WL Group columns. MBSR = Mindfulness-Based Stress Reduction, WL = WL; MMSE = Mini-Mental State Exam; CRP = C Reactive Protein

MBSR participants who completed the treatment trial (N=15) completed an average 7.20 (SD=1.15) out of the eight weekly classes offered, and all of these participants participated in the six-hour, day-long retreat held in week six of the MBSR program. MBSR treatment completers also completed, on average, 737 minutes of home meditation practice with home practice meditation compact discs over the course of the 8-week treatment (SD= 523.13). Moreover, as a manipulation check, we also confirmed that the MBSR program (compared to WL) increased self-reported mindfulness skills from baseline to post-treatment (see Table 2).

Table 2

Effects of the MBSR program on self-reported mindfulness skills, as measured by the Kentucky Inventory of Mindfulness Skills (KIMS).

		Pre Mean	SE	Post Mean	SE	F-value	p-value
KIMS	MBSR	137.15	3.96	148.24	4.19	19.72	<.01
	WL	137.61	3.99	133.88	3.96
Observe	MBSR	37.60	1.74	41.48	1.86	6.39	.02
	WL	40.77	1.75	40.40	1.74
Describe	MBSR	31.75	1.29	34.20	1.37	8.77	.01
	WL	30.07	1.30	29.08	1.29
Awareness	MBSR	32.00	1.38	34.29	1.46	4.48	.04
	WL	32.80	1.39	32.50	1.38
Judgment	MBSR	35.80	1.44	38.25	1.60	6.03	.02
	WL	33.90	1.46	31.90	1.44

Open in a separate window

Notes: Means and Standard Errors at baseline (Pre) and post-treatment (Post) in the Mindfulness-Based Stress Reduction (MBSR) and Wait-List (WL) groups for the mindfulness skills composite (KIMS) and each of the KIMS subscales (Observe, Describe, Awareness, Judgment). F and corresponding p-values refer to the treatment condition × time interaction.

Discussion

Using a randomized controlled trial design, the present study identifies MBSR as a novel approach for reducing loneliness in older adults. Although previous studies suggest a role for mindfulness-based treatments in reducing distress (Brown et al., 2007) and in fostering improved relational well-being (Carson et al., 2004; Brown et al., 2008), this is the first study to show that mindfulness meditation training reduces feelings of loneliness. We note that although we had no enrollment criteria for recruiting a lonely sample, our Los Angeles community sample had elevated levels of loneliness at baseline (M=41, SD=10) compared to a Midwestern older adult sample (M=37, SD=8) (Cacioppo, Hawkley, & Thisted, 2010) or an undergraduate student sample (M=37, SD=10) (Russell, Peplau, & Cutrona, 1980).

In addition to reducing loneliness, the present work provides additional clues into how MBSR may impact health in lonely older adults; namely, by down-regulating pro-inflammatory NF-κB -related gene expression dynamics that tend to be up-regulated in lonely older adults (Cole et al., 2007, 2011). The functional significance of the change in pro-inflammatory gene expression is not clear, as this study did not assess whether these gene expression effects translate into meaningful differences on inflammatory biology or disease outcomes. However, we note that CRP levels were elevated at baseline in our older adult sample (see Table 3), which suggests that our sample may have elevated cardiovascular disease and mortality risk (cf. Harris et al., 1999; Cesari et al., 2003). And there was some indication of differential trends in log-transformed CRP change over time, suggesting that MBSR may have reduced inflammatory biology in this initial pilot study. We did not observe any evidence for MBSR reducing IL-6, which may reflect the fact that there were low levels of IL-6 in our baseline sample (M=1.99 pg/ml, SE= .48) (cf. Harris et al., 1999), or the high biological variability commonly observed in spot IL-6 levels (vs. CRP, which provides a more smoothed, time integrated measure of IL-6 activity over the course of previous days) (Epstein et al., 1999; Pepys et al., 2003). Future MBSR RCT studies should evaluate these protein effects in lonely older adult samples with elevated cardiovascular and inflammatory disease risk factors.

Our finding that MBSR reduces pro-inflammatory gene expression provides a potential mechanistic account for previous studies showing that MBSR reduces stimulated pro-inflammatory cytokine production (Carlson et al., 2003, 2007). Moreover, our gene expression findings contribute to a new literature showing that behavioral stress management interventions, such as Relaxation Response training and Cognitive-Behavioral Stress Management training can are associated with gene expression in cells of the immune system (Dusek et al., 2008; Antoni et al., 2012). One implication of this work is that a common transcription control pathway (i.e., downregulated activity of NF-κB) may link a wide range of behavioral stress management interventions with reduced inflammatory disease risk.

Although our study highlights how MBSR can reduce subjective lonely feelings and loneliness-related pro-inflammatory gene expression in leukocytes, it will be important in future studies to consider the specific psychological and biological pathways that may link mindfulness meditation training with these effects. Recent work suggests that the health-compromising effects of loneliness are due to one's subjective perception of isolation and not one's objective number of social contacts (Cole et al., 2007, 2011), suggesting that psychological perceptions of social disconnection may be a critical component of loneliness. One potential psychological pathway then, is that MBSR reduces psychological perceptions of social threat or distress, and reduced distress may decrease perceptions of loneliness. As the Buddhist Nun Pema Chodron suggests (opening quote), mindfulness meditation training can “turn our fearful patterns upside down”, reducing the distress that can accompany loneliness (Chodron, 2000). At present, we show that MBSR (compared to a WL condition) reduces perceptions of loneliness, and it will be important for subsequent research to consider how changes in psychological pathways such as distress, depression, and threat appraisals in social encounters potentially explain how MBSR reduces loneliness in older adults (cf. (Cacioppo and Hawkley, 2009).

The biological pathways linking MBSR to changes in pro-inflammatory gene expression also need to be clarified in future studies. MBSR could potentially alter activity of the hypothalamic-pituitary-adrenal (HPA) axis or sympathetic nervous system, both of which produce stress mediators (e.g., cortisol, norepinephrine, eprinephrine) that can modulate NF-κB activity and pro-inflammatory gene expression (for reviews, Cole, 2009; Irwin and Cole, 2011). Indeed, some initial evidence suggests that MBSR alters diurnal cortisol output (Carlson et al., 2004; Witek-Janusek et al., 2008; Brand et al., 2012) (cf. Brown et al., in press). It is important for future studies to test these stress-reduction mediated biological pathways. Finally, it is possible that MBSR effects on gene expression may be explained by changes in health behaviors. Though, we did not find evidence that MBSR altered measures of sleep quality or exercise use in this sample, future studies might benefit from inclusion of objective measures of sleep and activity (e.g., actigraphy, polysomnography).

One limitation of the present study was the use of a WL control group. Future studies using active control groups that include nonspecific and/or specific components of the intervention will help clarify what aspects of the MBSR program decrease loneliness. It is possible that observed changes in loneliness in MBSR vs. WL control could be explained by non-specific factors (e.g., social support, participant contact with an instructor). For example, it may be that the group-based format of MBSR classes is providing social support (and networking), and these social factors are reducing loneliness. However, it is unlikely that non-specific group support accounts for the observed decreases in loneliness in the MBSR condition, as prior randomized controlled trials have found that loneliness is not altered following administration of social support and social skills training (Masi et al., 2011). Moreover, when mindfulness meditation training is taught individually (i.e., not in a group-based format) stress symptoms are reduced along with improvements in markers of physical health (Kabat-Zinn et al., 1998). Nevertheless, future RCT studies that control for non-specific factors with well-matched active treatment control conditions, and using new control programs as recently described (e.g., MacCoon et al., 2012) will advance understanding of the components of MBSR that are effectively driving benefit.

Supplementary Material

Acknowledgments

Footnotes

References