A systematic review and meta-analysis exploring the efficacy of mindfulness-based interventions on quality of life in people with multiple sclerosis

Background Quality of life (QoL) is commonly impaired among people with multiple sclerosis (PwMS). The aim of this study was to evaluate via meta-analysis the efficacy of Mindfulness-based interventions (MBIs) for improving QoL in PwMS. Methods Eligible randomized controlled trials (RCTs) were identified via searching six major electronic databases (MEDLINE, EMBASE, CINAHL, Cochrane Central Register of Controlled Trials, AMED, and PsycINFO) in April 2022. The primary outcome was QoL. Study quality was determined using the Cochrane Collaboration risk of bias tool. Meta-analysis using a random effects model was undertaken. Effect sizes are reported as Standardized Mean Difference (SMD). Prospero ID: 139835. Results From a total of 1312 individual studies, 14 RCTs were eligible for inclusion in the meta-analysis, total participant n = 937. Most studies included PwMS who remained ambulatory. Cognitively impaired PwMS were largely excluded. Comorbidities were inconsistently reported. Most MBIs were delivered face-to face in group format, but five were online. Eight studies (n = 8) measured MS-specific QoL. In meta-analysis, overall effect size (SMD) for any QoL measure (n = 14) was 0.40 (0.18–0.61), p = 0.0003, I2 = 52%. SMD for MS-specific QoL measures (n = 8) was 0.39 (0.21–0.57), p < 0.0001, I2 = 0%. MBI effect was largest on subscale measures of mental QoL (n = 8), SMD 0.70 (0.33–1.06), p = 0.0002, I2 = 63%. Adverse events were infrequently reported. Conclusions MBIs effectively improve QoL in PwMS. The greatest benefits are on mental health-related QoL. However, more research is needed to characterize optimal formatting, mechanisms of action, and effects in PwMS with more diverse social, educational, and clinical backgrounds. Supplementary Information The online version contains supplementary material available at 10.1007/s00415-022-11451-x.


Search strategy
We searched six major electronic databases (MEDLINE, EMBASE, CINAHL, Cochrane Central Register of Controlled Trials, AMED, and PsycINFO) in April 2022 using medical subject headings and key words relating to mindfulness and multiple sclerosis, search syntax and Boolean operators. Search delimiters included: studies in humans, published in English language, between 1980-current (April 2022). We also searched reference lists, the gray literature and contacted relevant experts in the field. Our search strategies are available in Online Appendix 1.

Study selection
Search results were imported into Endnote, for storage and screening. Two reviewers ("blinded for peer review") independently assessed title/abstracts for eligibility. Three reviewers ("blinded for peer review"), then independently assessed eligibility against study, population, intervention, and outcome (SPIO) characteristics. A senior reviewer ("blinded for peer review") was available for arbitration in the event of any disagreement over study eligibility.

Data extraction
Three reviewers ("blinded for peer review") independently extracted study data using the CONSORT and TIDieR checklists (Appendix 2).

Quality appraisal
We used the Cochrane Collaboration tool [48] for assessing risk of bias (low, unclear, high) on individual outcomes (sequence generation, allocation concealment, participant blinding, personnel blinding, assessor blinding, incomplete outcomes, selective outcome reporting, any other source of bias). Based on summed individual outcomes, each study was then assigned an overall risk of bias category (low, unclear, high). Two reviewers engaged in discussion to reach consensus on overall risk of bias, when discrepancies arose.

Primary outcome
Main outcome measures were all reported as continuous with mean, standard deviation (SD) values and the number of participants for each treatment group extracted. "Effect size" is reported as the unbiased standardized mean difference (SMD), a positive SMD indicating a finding in support of the intervention having a positive treatment effect. The SMD was calculated by difference in means between the MBI and the control group at follow-up divided by the pooled follow-up SD. Where effect estimates were reported from adjusted regression models, we extracted these as the SMD with their corresponding SD.

Synthesis
We used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist [49] when drawing together findings for our systematic review and meta-analysis. We used a random effects meta-regression model for deriving SMD, due to expected high levels of outcome heterogeneity (generic vs MS-specific QoL measures). We report effect estimates and 95% confidence intervals (as a measure of precision) and corresponding p values. We assessed heterogeneity using the I 2 statistic, I 2 representing the percentage of total variability in effect size estimates due to heterogeneity. An I 2 of 0% indicates that all heterogeneity is due to sampling error, while an I 2 of 100% suggests all variability may be attributable to studies being truly heterogeneous.
We computed Funnel plots and Egger's test to determine asymmetry and likelihood of publication bias, with subsequent 'trim and fill' to assess significance of any bias. All statistical analyses were carried out using RevMan.

Characteristics of study participants
Across the 14 RCTs there were 937 participants. Five studies reported on ethnicity, which was 87.8% "white" or "anglo-saxon/anglo-celtic" [50,51,53,59,60]. One study did not report the percentage of women [59], but most studies predominantly recruited women (total women = 621; 78%). Two studies did not report mean (SD) age, but rather, an age range of 20-50 [63], and a median age of 43 [58]. Of the remainder, mean (SD) age was 44.04 (9.1). Most studies did not report on socioeconomic status (SES), but in the five that did, most participants had a college degree or higher [50,51,53,57,59]. Most participants (n = 699; 74.5%) had a relapsing MS phenotype, while 128 (13.6%) had progressive disease. MS phenotype was not reported in the remainder. Where reported, disability, as measured by the Expanded Disability Status Scale (EDSS), was mostly < 6.0, indicating participants remained ambulant without a walking aid; however, one study focused solely on progressive MS, where mean (SD) EDSS was 6.5 (1.5) indicating the ability to walk for 20 m without stopping using walking aid(s) [53]. Four studies reported on comorbidity, mainly depression [55,[59][60][61]. One study reported comorbidity with a mean (SD) count of 2.4 (2.0) comorbidities [51]. In six studies, most participants were on disease-modifying drugs (DMDs) [50-52, 55, 62, 63]. One study only indicated "both groups also received their routine drug treatments" without specifying the number of participants on DMDs [63], and the remaining studies did not measure use. Antidepressant use ranged from 6 to 56%. Nine studies [50-53, 55, 56, 59-61] explicitly excluded those with cognitive impairment, while the remainder did not mention cognitive impairment as an eligibility criterion (Table 2).

Heterogeneity and publication bias
Across the 14 studies heterogeneity was moderate (52%) and there was no evidence of publication bias (p = 0.7589) (Fig. 8).

Study quality
There was no evidence of selective outcome reporting in any of the included studies. Most (n = 12 out of 14) described sequence generation, the majority (n = 9 out of 14) described allocation concealment, blinding procedures (n = 9 out of 14), and most (n = 9 out of 14) accounted for incomplete outcome reporting. Overall, half of included studies (n = 7 out of 14) were adjudged low risk of bias (Fig. 9).

Adverse events
In one study, a participant undertaking MBSR reported an increase in neuropathic pain following the 'raisin exercise'an introductory MBI exercise, which involves exploring sensory experiences associated with seeing, touching, and tasting a raisin using mindful awareness [51]. In another study, a participant felt more anxious after a MBSR day retreat and a participant experienced muscle spasticity during a muscular relaxation activity [50]. Lastly, in one study, four participants experienced an MS relapse or hospitalization, however these events were deemed unrelated to the MBI [59].  obscure and likely will vary [39]. In this current study, MBI dose (session attendance + home practice) was infrequently reported, but ranged from 16 to 66 h, with session attendance ranging from 60 to 95%, and home practice 29.2-38 min/day.

Strengths and weaknesses of this study
We used recommended tools for carrying out our systematic review and meta-analysis, leaving our findings open to external scrutiny and audit. Our research team was multidisciplinary (nursing, rehabilitation, family medicine, psychiatry, psychology, statistics). We included solely RCTs to collate the highest quality evidence for the use of MBIs to improve QoL in PwMS.
Our study was necessarily limited to include only those articles published in English. As the concepts underpinning mindfulness originally derive from Asia, it is possible we missed relevant literature (i.e., non-English language publications) on the use of this technology in diverse contexts, where participant characteristics, intervention acceptability and effects may differ somewhat. However, we found no statistical evidence of publication bias.

Strengths and weaknesses of studies in this review
This study had several strengths. All studies in this systematic review and meta-analysis were RCTs. Six compared against an active comparator condition, attempting to minimize non-specific treatment effects, likely in a group-based complex intervention [67] such as MBIs [68]. An RCT is widely regarded as the best study design to minimize bias in the 'hierarchy of evidence' [69]. Although a wide range of participants took part in the studies in this review, mean participant age was relatively low (44.04), socioeconomic and educational statuses infrequently documented. Thus, very little is known about effects of MBIs among older PwMS, those with late onset disease, or with diverse social and educational backgrounds. Similarly, limited reporting on other factors known to impair (physical and mental health comorbidities, physical disability, cognitive impairment), stabilize or improve QoL in PwMS (e.g., 'second generation' DMD use [70]) limits somewhat the scope of analyses, whereas lack of biological outcome measurement (e.g., structural or functional MRI) limits somewhat interpretation of meaning in findings. In addition, regarding quality, although half of studies included in this review were deemed to have low risk of bias, reporting of study procedures, population characteristics, intervention components, and outcomes

Implications for research
MBIs effectively improve depression in PwMS [46], a factor strongly associated with reduced QoL in this population [18]. However, the impact of MBIs on other factors known to impair QoL in PwMS, such as cognitive impairment [17] should be assessed, as in general populations MBIs can improve aspects of cognitive function (working and autobiographical memory, cognitive flexibility, and metaawareness) [71].
The factors that mediate or moderate effectiveness of MBIs in PwMS are not known. Feasibility work suggests important roles for acceptance, self-efficacy, and self-compassion [72]. Future research may examine the neurobiological mechanisms that underpin MBIs, as well as test a wider range of candidate factors in larger, powered samples of PwMS.

Implications for clinical practice
MBIs appear to be a safe approach to improving QoL in PwMS, with the greatest benefits seen on mental QoL. Both face-to-face and online MBIs hold potential for effectiveness, though the small number of studies in this area makes drawing firm conclusions difficult. In pragmatic terms, online or virtual MBIs may now be preferrable to PwMS, given the ongoing context created by the COVID-19 pandemic, and may also help to address some of the inequalities PwMS face in accessing mental healthcare [73].

Conclusions
MBIs effectively improve QoL in PwMS. The greatest benefits are on mental health-related QoL. However, more research is needed to characterize optimal formatting, mechanisms of action, and effects in PwMS with more diverse social, educational, and clinical backgrounds.

Funding
The authors did not receive support from any organization for the submitted work.

Conflicts of interest
The authors declare that they have no conflict of interest.

Ethical approval
The manuscript does not contain clinical studies or patient data.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will Fig. 9 Risk of bias need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.