A Randomized Controlled Trial of Fasting and Lifestyle Modification in Patients with Metabolic Syndrome: Effects on Patient-Reported Outcomes

Lifestyle interventions can have a positive impact on quality of life and psychological parameters in patients with metabolic syndrome (MetS). In this randomized controlled trial, 145 participants with MetS (62.8% women; 59.7 ± 9.3 years) were randomized to (1) 5-day fasting followed by 10 weeks of lifestyle modification (F + LM; modified DASH diet, exercise, mindfulness; n = 73) or (2) 10 weeks of lifestyle modification only (LM; n = 72). Outcomes were assessed at weeks 0, 1, 12, and 24, and included quality of life (Short-Form 36 Health Survey Questionnaire, SF-36), anxiety/depression (Hospital Anxiety and Depression Scale, HADS), stress (Cohen Perceived Stress Scale, CPSS), mood (Profile of Mood States, POMS), self-efficacy (General Self-Efficacy Scale, GSE), mindfulness (Mindfulness Attention Awareness Scale, MAAS), and self-compassion (Self-Compassion Scale, SCS). At week 1, POMS depression and fatigue scores were significantly lower in F + LM compared to LM. At week 12, most self-report outcomes improved in both groups—only POMS vigor was significantly higher in F + LM than in LM. Most of the beneficial effects within the groups persisted at week 24. Fasting can induce mood-modulating effects in the short term. LM induced several positive effects on quality of life and psychological parameters in patients with MetS.


Introduction
The metabolic syndrome (MetS) is a condition characterized by the presence of at least three cardiovascular risk factors, such as abdominal obesity, hypertension, insulin resistance, and dyslipidemia [1]. It affects approximately one in three American adults and one in four adults in Europe [2,3]. Most cardiovascular risk factors can be influenced by

Fasting and Lifestyle Modification (F + LM)
Participants started with 2 vegan days (max. 1200 kcal/day), followed by 5 days of fasting (max. 350 kcal/day), and a stepwise reintroduction of food. Then they participated in the 10-week MICOM comprehensive multimodal lifestyle modification intervention with weekly 6-h sessions [29]. These included lectures and cooking classes on wholefood vegetarian diets with emphasis on a plant-based Mediterranean diet and a modified DASH diet, intermittent fasting (rice day once a week), and recommendations for certain cardioprotective foods [30][31][32][33]. Furthermore, moderate aerobic exercise such as walking and progressive muscle relaxation, mindfulness meditation, and yoga were trained. After completion of the program, monthly group sessions were offered during weeks 13 to 24 to ensure longer-term adherence.

Lifestyle Modification (LM)
The lifestyle intervention was like the one performed for the F + LM group, but without the initial fasting intervention. The program consisted of 55 h of group intervention over a period of 10 weeks. After the end of the program, in months 13 to 24, monthly group sessions were offered to ensure longer-term adherence.

Randomization
Patients were randomly assigned in a 1:1 ratio to either F + LM or LM by block randomization with randomly varying block lengths, stratified by (a) study center and (b) taking/not taking antihypertensive medication. The randomization list was generated by a biometrician not involved in patient recruitment or assessment using Random Allocation software (version 1.0, Mahmood Saghaei, Isfahan, Iran) [34]. The list was password protected, and no person other than the biometrician had access to it. On this basis, he prepared sealed, sequentially numbered opaque envelopes containing the assignments to the interventions. After obtaining written informed consent and baseline assessment, the study physician opened the envelope with the lowest number to reveal the assignment for that patient.

Outcome Measures
Outcomes were assessed at baseline and at 1, 12 and 24 weeks after randomization by a blinded outcome assessor who was not involved in patient recruitment, allocation, or treatment. Physician-assessed outcomes, laboratory parameters, safety data, and explorative experimental variables (immune function, microbiome) were also assessed and have been reported elsewhere [35]. In the here presented analysis, the following PRO were used and will be reported:  [39]. The POMS is a 35-item instrument that measures four domains of mood disfunction assessing the subscales vigor, fatigue, depression, and anger [40]. • Self-efficacy on the 10-item General Self-Efficacy Scale (GSE) [41]. Self-efficacy is a construct that "reflects an optimistic self-belief that one can master a difficult task or cope with adversity" [41]. Self-efficacy makes it easier to set goals, make efforts, persevere in the face of obstacles, and recover from setbacks. It can be considered a resource for resilience.

•
Mindfulness on the 15-item Mindfulness Attention Awareness Scale (MAAS) [42]. Mindfulness is characterized by the following three qualities: (1) intentional, (2) related to the present moment, and (3) nonjudgmental. • Self-compassion on the 26-Item Self-Compassion Scale (SCS) [43]. Self-compassion refers to a positive basic attitude toward oneself even in difficult life situations. This personality trait has proven to be an effective protective factor that promotes emotional resilience.

Sample Size Calculation and Statistical Analysis
The sample size was calculated a priori using G*Power software [44]. Based on prior research on multimodal lifestyle interventions [10,45], yoga [46], mindfulness [47], and Mediterranean diet [48], a between-group effect size of d = 0.5 was expected. A level 2.5% t-test requires a total of 64 patients per group to detect a respective group difference with a statistical power of 80%. Accounting for a potential loss of power because of a maximum of 10% dropouts, it was intended to include a minimum of 142 participants in this trial.
All analyses were performed intention-to-treat, which means that all participants were randomized, whether or not they provided a complete data set or adhered to the study protocol. Missing data were multiply imputed by Markov chain Monte Carlo methods [49,50], yielding a total of 50 complete data sets.
Group differences were analyzed by univariate analyses of covariance (ANCOVA) modeling the outcome at week 1, 12, or 24 as a function of the treatment group (classified factor), the stratification factors study center (classified covariate) and baseline antihypertensive medication intake (classified covariate), and the respective baseline value (linear covariate). Afterwards, the 50 estimates of group differences were combined to produce overall effect size estimates, 95% confidence intervals, and p-values. Within-group changes over time were analyzed by paired t-tests. p-Values ≤ 0.05 were considered significant.
All analyses were performed using the Statistical Package for Social Sciences software (IBM SPSS Statistics for Windows, release 22.0. Armonk, NY, USA: IBM Group).

Patients
A total of 452 patients were screened by telephone and 258 were excluded because they did not meet the inclusion criteria ( Figure 1). A further 49 patients were excluded after medical assessment onsite. Finally, 145 participants were enrolled after informed consent and were randomized to the F + LM group (n = 73) or the LM group (n = 72). A total of 15 participants each were lost to follow-up in the F + LM and LM groups during the study period ( Figure 1). modeling the outcome at week 1, 12, or 24 as a function of the treatment group (classified factor), the stratification factors study center (classified covariate) and baseline antihypertensive medication intake (classified covariate), and the respective baseline value (linear covariate). Afterwards, the 50 estimates of group differences were combined to produce overall effect size estimates, 95% confidence intervals, and p-values. Within-group changes over time were analyzed by paired t-tests. p-Values ≤ 0.05 were considered significant.
All analyses were performed using the Statistical Package for Social Sciences software (IBM SPSS Statistics for Windows, release 22.0. Armonk, NY, USA: IBM Group).

Patients
A total of 452 patients were screened by telephone and 258 were excluded because they did not meet the inclusion criteria ( Figure 1). A further 49 patients were excluded after medical assessment onsite. Finally, 145 participants were enrolled after informed consent and were randomized to the F + LM group (n = 73) or the LM group (n = 72). A total of 15 participants each were lost to follow-up in the F + LM and LM groups during the study period ( Figure 1). Participants' characteristics are provided in Table 1. About two thirds of the participants were female and about one third had a university degree (mean age 59.7 ± 9.3 years). Participants had a BMI of 33.3 ± 4.5 kg/m 2 . Participants in the F + LM group attended a mean of 8.2 ± 2.3 (82%) out of 10 possible intervention sessions; those in the LM group attended 7.1 ± 3.5 (71%) sessions (p = 0.124).

Outcome Measures
Detailed results are provided in Table 2

Discussion
This study investigated the effects of fasting, followed by a multimodal lifestyle modification intervention as described in the MICOM Model, on patient-reported outcomes in patients with MetS compared to a lifestyle modification intervention only. After the multimodal lifestyle modification intervention, we found improvements for all outcome measures, namely quality of life, anxiety, depression, stress, mood as well as self-efficacy, mindfulness, and self-compassion, which persisted at the follow-up after 24 weeks. Comparing the groups (1) at week 1, depression and fatigue scores (POMS) were significantly lower; (2) at week 12, only vigor (POMS) was significantly higher; and (3) at week 24, self-compassion was significantly lower in the group that fasted additionally.
The study participants were demographically heterogeneous. More than 60% were women. The distribution of educational attainment was broad, with a slight shift toward those with higher education. The inclusion criteria of the study led to a large proportion of patients who were German adults with metabolic syndrome. This suggests that the results of the study should be transferable to a large proportion of the German population. Baseline scores indicated the studied population having impaired quality of life and moods as well as depressive symptomatology and anxiety upon study entry.
The five-day fast (max. 350 kcal/day) implemented at week 1 reduced depression and fatigue scores in the F + LM group. This is in line with other findings, as especially (prolonged) fasting is known to influence the psyche positively, inducing mood-modulating effects [25][26][27]. Furthermore, participants in the F + LM group had a higher vigor score at the end of fasting as well as at the 12-week follow-up. Recent studies showed that fasting can increase vigor and vitality, emotional balance, and daytime concentration [26,51]. Fasting is also associated with increased availability of neurotrophic factors in the brain, including serotonin, endogenous opioids, and endocannabinoids, which may also contribute to improved vigor and mood [52,53].
Compared to lifestyle modification only, fasting significantly induced several positive short-term effects in physical parameters, which are reported in detail elsewhere. Especially the observed moderate weight loss and the reduction in waist circumference could lead to positive psychological experiences and may have mood-lifting effects. It should be mentioned that fasting has also been reported to potentially have negative effects on mood, e.g., increased irritability and mood lability [25,27]. However, such side effects were not detectable in our study.
Interestingly, a recent study reported differences in psychological well-being depending on whether or not participants had previous experience with fasting [27]. The group with fasting novices had overall more negative psychological states (i.e., increased appetite, increased negative moods, increased stress, and reduced vitality) during the fast. Therefore, previous fasting experience may buffer negative feelings during a fast. However, previous fasting experiences were not recorded in our study.
We found that self-compassion was significantly higher in participants who received only lifestyle change than in participants who received fasting plus lifestyle change. This finding seems to contradict the hypothesized positive effects of fasting. We would not assume that this finding indicates a harmful effect of fasting. The authors are not aware of any other studies on self-compassion and fasting. However, this finding is interesting and may be worth further investigation.
Several parameters may have clinically relevant effects. The minimal clinically important difference (MCID) for the SF-36 physical component score is defined as a value of 2 and for the SF-36 mental component score as 3 points; in our study the increases were greater than 2 and 3 points, respectively, for both groups, resulting in a relevant clinical improvement [54]. The MCID for HADS in patients with cardiovascular disease is defined at a value of 1.7 points, which the study participants also achieved at the 12-week visit in both groups [55]. No MCID is available for CPSS and POMS, but the effects at the 12-and 24-week follow-ups are also in the range of moderate effect sizes.
Various MBM interventions such as yoga, qigong, tai chi, and mindfulness training have been studied in patients with obesity and metabolic syndrome in randomized controlled clinical trials [18][19][20]. In addition to improving cardio-metabolic health, yoga and tai chi in particular have been shown to improve stress, quality of life, and psychological parameters in patients with MetS [20,56].
The MBM lifestyle modification intervention led to an improvement of self-efficacy, mindfulness, and self-compassion in our study population. These parameters could interact with each other. Further studies to explore explanatory models for the effects of MBM are warranted. In particular, mediating variables on mindfulness, self-efficacy, and compassion, and the effects on physical and psychological parameters, need to be examined more closely.
Our study has several limitations. First, the implementation of a control group with no initial fasting could lead to nonspecific effects due to disappointment effects and may also have influenced the study results. Second, we have not assessed participants' attitudes and motivation regarding the intervention. Third, the time intensity of the two programs was slightly different, as additional fasting was performed in the F + LM group. Another limitation is the predominant use of generic questionnaires in this study instead of condition-specific ones. On the one hand, this was done because epidemiological studies on metabolic syndrome also predominantly use generic instruments [9], which facilitates the comparison of the study results with those in the population. On the other hand, due to the general lack of specific subjective symptoms in metabolic syndrome, there are hardly any specific questionnaires and, accordingly, previous studies have relied on questionnaires specific to individual symptoms (e.g., obesity). Nevertheless, the lack of specific measures is a limitation, and future studies should use condition-specific questionnaires, at least in addition to generic ones. Finally, the extent to which patients adhered to the program by using stress reduction techniques at home was not assessed.

Conclusions
In summary, the results of this study point to beneficial and clinically relevant effects of fasting and intensified lifestyle modification on quality of life and psychological parameters. In addition, the used combination of interventions appears to have a sustained effect that needs to be further demonstrated in studies with longer-term follow-up. Further trials could compare comprehensive lifestyle interventions with less intensive and even unimodal interventions and assess cost-effectiveness. Further evaluation of mediating variables such as mindfulness, self-efficacy, and self-compassion in relation to psychological parameters, and interaction with improvements in physical and psychological parameters in the context of complex lifestyle interventions, is warranted. Funding: This work was supported by Corona-Foundation, Germany (funding number: S199/10056/ 2012). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. We acknowledge financial support from the Open Access Publication Fund of Charité-Universitätsmedizin Berlin and the German Research Foundation (DFG).

Institutional Review Board Statement:
The trial was registered at clinicaltrials.gov under NCT02099968 and was approved by the respective ethics committee (Charité-Universitätsmedizin Berlin EA4/141/13, 16 January 2014) and the University of Duisburg-Essen (approval number: 14-5733 BO, 26 February 2014). It followed the Declaration of Helsinki and Good Clinical Practice guidelines for trial conduct.
Data Availability Statement: Data available on request.

Conflicts of Interest:
The authors declare no conflict of interest.