Companion-Intensive Multi-aspect Weight Management in Chinese Adults with Obesity: A Six-Month Multicenter Randomized Clinical Trial

Obesity is a globally increasing health epidemic requiring early lifestyle intervention. Our main objective was to examine the effectiveness of companion-intensive multi-aspect weight management (CIMWM) for Chinese adults with obesity. In this 6-month, prospective, open-label, multicenter, randomized controlled clinical trial, we included 272 obese adults aged 18–50 years with a body mass index (BMI) ≥ 28.0 kg/m 2 and capable of using smartphones. CIMWM (n = 136) offered both daily online instructions and monthly face-to-face guidance by physicians, dietitians, and health managers along with the provision of meal replacements in the rst 3 months. Traditional multi-aspect weight management (TMWM, n = 136) provided monthly face-to-face guidance by the above-mentioned professionals and required daily self-monitoring. Measurements of body composition and metabolic parameters were assessed face-to-face at baseline, 1, 2, 3, and 6 months by physicians. Primary outcomes were the changes in BMI and body composition.


Background
Obesity is a global health epidemic. According to the World Health Organization (WHO), up to 57.8% of adults worldwide will be classi ed as obese by 2030 [1]. The prevalence of obesity among Chinese adults is ~11% [2]. Body mass index (BMI) is a widely used parameter to de ne overweight and obese status in adults [3]. In China, individuals within a BMI range of 24.0-27.9 kg/m 2 are de ned as overweight and those with BMI≥28.0 kg/m 2 as obese [4].
Obesity is an underlying risk factor for a series of disorders, such as diabetes, non-alcoholic fatty liver disease (NAFLD), coronary heart disease, sleep apnea syndrome, stroke, and cancer, that are associated with signi cantly increased morbidity and mortality [5,6]. Effective treatment options and prevention strategies to control obesity are therefore a top priority for healthcare systems [7]. Scienti cally sound and appropriate weight loss treatments are essential to lower the risk of obesity-related diseases. Weight loss treatments that have proven effective for patients with obesity include lifestyle modi cation, conventional pharmacologic treatments and surgical treatments [8]. In particular, lifestyle modi cation is the foundation of weight loss and should be consistently implemented [9].
Intensive lifestyle modi cation involves alterations in ve indispensable aspects: diet, exercise, psychotherapy, behavioral intervention, and health education. The Diabetes Remission Clinical Trial (DiRECT) in the UK showed signi cant bene ts in patients with type 2 diabetes mellitus (T2DM) and obesity who underwent a primary care-led weight management program, with almost half the participants achieving remission to a non-diabetic state and requiring no further treatment by 12 months [10]. Another study demonstrated that a Web-Based Behavior Change Program could help achieve weight loss and reduce cardiovascular disease (CVD) risk in obese individuals, supporting the internet as a viable medium for weight loss management [7]. Records show that in 2014, ~92% of the population in China had access to mobile phones, with smartphones being widely used by young and middle-aged individuals [11].
Although early lifestyle intervention is important for control of obesity, it is di cult for patients to adhere to daily modi cations on their own and intensive supervision by doctors or other professionals is therefore necessary for successful weight management. Face-to-face guidance by the community and hospitals is the traditional method used for patients with obesity in China, which often results in poor achievement of self-monitoring ability and weight loss. To resolve this issue, we proposed the implementation of a Companion-Intensive Multi-aspect Weight Management (CIMWM) strategy focusing on a combination of online and o ine medical interventions with daily lifestyle supervision and direction of diet and exercise. To evaluate this strategy, we conducted a multicenter, randomized, controlled clinical intervention trial for adults with obesity in China from 2018 to 2019 that aimed to compare the e cacy and feasibility of two different medical weight management intervention programs. CIMWM offers both daily online instructions and monthly face-to-face guidance by physicians, dietitians, and health managers, along with provision of meal replacements in the rst three months, while traditional multiaspect weight management (TMWM) provides monthly face-to-face guidance by the above professionals and requires daily self-monitoring. Changes in clinical indicators, such as lipid pro les, blood glucose, Nutrition Bars consisting of whey protein, soy protein isolate, chia seeds, oligosaccharides, collagen, and konjac extract rich in dietary ber and γ-aminobutyric acid were provided as meal replacements for participants from the CIMWM group in the rst 3 months. A single bar weighed 30 g and provided 111 kcal with 9.15 g protein, 1.8 g fat, 12.48 g carbohydrate, and 3.93 g dietary ber. Subjects in the CIMWM group were additionally provided calcium, vitamin D, and multivitamin mineral tablets.
Individualized exercise plans were created by health managers based on the health status and exercise capacity of each participant. The exercise plan included a weekly 160 min group exercise session. Over 6 months, participants were instructed to exercise for 40 min, 4 days a week, starting with 5 min warm-up exercises, followed by 10 min and 30 min of resistance and aerobic exercise, respectively, and ending with 5 min muscle stretching exercises.

Primary Outcomes
Body mass index (BMI), body weight (BW), body fat percentage (BFP), body fat mass (BFM), free-fat weight (FFW), skeletal muscle mass (SMM), and visceral fat area (VFA) were measured in participants while standing on an automated hand-to-foot bioelectrical impendence device (JAWON IOI353 Body composition analyzer, Korea) with bare feet and light clothing.
Advanced brosis was accurately excluded by applying the NFS low cut-off point (-1.455) while its presence was diagnosed with high accuracy by applying the NFS high cut-off point (0.676) [14]. HOMA-IR was calculated using the formula: FPG (mmol/L) ×FINS (mIU/L)/22.5.

Statistical analysis
Using data from a previous weight management study [16], a power analysis was conducted using SAS, with alpha of 0.05 and power of 0.80. Sample size calculations indicated that recruitment of 72 participants in the primary center was required to achieve 80% power. Accordingly, competitive enrollment was initiated at branch centers. Enrolled patients were numbered in adjacent sequences using the randomized block design (block size=4). Random assignment codes were generated by statistics professionals using the SAS 9.12 software proc plan program according to a 1:1 allocation ratio [7].
All analyses were performed using SPSS for Macintosh version 25.0 (SPSS Inc, Chicago, IL, USA).
Demographic and baseline characteristics were evaluated with the aid of descriptive statistics. Normally distributed and continuous variables were presented as mean and standard deviation (SD) while median and interquartile range for variables were determined in cases where normality criteria were not satis ed. Categorical variables were presented as frequencies and percentages. Data sets involving the baseline, 3month, and 6-month index of all participants were assessed using the Kruskal-Wallis test owing to frequent non-normal distribution of variables. All indicator changes were not normally distributed and Spearman's rank correlation performed to evaluate the relationship between the physical examination and hematology indices at the 6-month point. To compare the characteristics of continuous variables between the two groups, the appropriate parametric (independent samples t-test) and nonparametric (Mann-Whitney U) tests were utilized. For categorical variables, the chi-square test was used. Two-sided P values <0.05 were considered statistically signi cant.

Participants
Participants were recruited from August 2018 to June 2019 ( Figure 2). Among the 378 eligible subjects, 272 (72%) were randomly assigned to study groups at a 1:1 allocation ratio. Overall, 254 (93.3%) participants completed the 6-month weight management intervention study.
Anthropometric, metabolic, and clinical characteristics of participants before and after the weight loss intervention The clinical and demographic characteristics of all study participants are presented in Table 1. BMI was signi cantly decreased by 2.35 kg/m 2 (median) at 3 months and 2.39 kg/m 2 (median) at 6 months, compared with baseline BMI (both P<0.001). Weight loss was sustained during the rst 3 months, but body weight increased slightly at 6 months (median, -5.60 kg; P<0.001). Additionally, BFP, BFM, VFA, WC, HC, WHR, SBP, DBP, ALT, AST, TC, TG, LDL-C, FPG, FINS, FCP, HbA1c, HOMA-IR, and number of patients with impaired fasting glucose (IFG) were signi cantly decreased at both 3 and 6 months (P<0.05). Nevertheless, FFW and SMM were slightly decreased by -1.91 kg (P<0.01) and -1.77 kg (P<0.01), respectively, at 6 months, compared with baseline, while S/V increased over 6 months in all participants (P<0.05). UA was markedly decreased by 28.83 umol/L (median) only at 6 months, compared with baseline. HDL-C levels increased gradually from baseline until the end of the study period (P<0.001).
The correlation coe cient heatmap ( Overall, no differences in NFS were observed at 0, 3, and 6 months (P=0.975). To further validate the changes in NFS, 254 participants were strati ed into three groups based on the tertiles (-3.41 and -2.53) of NFS at baseline. NFS of the upper-level tertile group gradually decreased over the subsequent 6 months to a signi cantly different extent from baseline (P<0.05). The increase in lower NFS tertiles over the 6-month period was also signi cantly different from baseline (P<0.05). Although NFS in several participants was slightly increased, the values remained below the high cut-off point (0.676) of liver brosis ( Figure 4a).
Outcome differences between CIMWM and TMWM groups during the 6-month follow-up The baseline characteristics of both groups are presented in Table 2. No signi cant intergroup differences were observed.
We recorded signi cant changes in BMI, BFP, BFM, and VFA between baseline and each time-point of the two groups (P<0.05, Table 3). Our data showed markedly lower BMI, VFA, and ALT levels of the CIMWM group relative to the corresponding parameters in the TMWM group during the course of the study (P<0.05; Figure 5a, b, and d). S/V was signi cantly increased from baseline to 1, 2, and 3 months in both groups, with no further increase in the following 3 months. Increment of S/V in the CIMWM group was signi cantly better than that in the TMWM group in the rst 3 months (Figure 5c).
Subjects in the CIMWM group lost more body weight than those in the TMWM group at 3 months

Discussion
Data from this 6-month, randomized, controlled trial on Chinese adults with obesity showed that weight management with regular guidance by physicians, dietitians, and health managers is feasible and effective for improvement of body weight, BMI, body composition parameters (BFM and VFA), ALT, AST, blood lipid pro les, and insulin resistance in all participants. The CIMWM strategy involving both daily online instructions and monthly face-to-face guidance was superior to TMWM with respect to improvement of BMI, ALT, and body composition parameters, such as BFM, VFA, and S/V. Individuals subjected to the CIMWM program showed better adherence to guidelines and reprogramming of lifestyle.
Efforts to estimate body composition changes during interventions are important in improving our awareness of metabolic health. Metabolic memory in terms of continuous reduction of BMI, VFA, and ALT was retained up to 6 months, even with transfer of participants in the CIMWM group to the selfmonitoring phase at 3 months. NFS was further evaluated to determine the degree of liver brosis. Our data suggest that the CIMWM weight management program is more bene cial in decreasing NFS among subjects with levels in the highest tertile.
The primary outcomes of this study were reduction of BMI, BFM, and VFA from baseline to 6-month follow-up. The CIMWM group achieved signi cantly better reduction of these parameters than the TMWM group throughout the study. Notably, body weight increased slightly at 6 months, compared to 3 months, in both groups, but remained lower than that at baseline. We believe that the weight gain in the CIMWM group during the last 3 months of the study is attributable to the transfer to self-monitoring. Another potential reason suggested by Burke et al. [17]is a gradual decline in adherence to self-monitoring weight management program, which is common during the follow-up period.
The change in body composition after weight management is an interesting phenomenon. Total body fat and distribution, especially visceral fat, is known to be associated with higher risk of metabolic syndrome [18]. Here, our data showed a reduction of BFP, BFM, and VFA in all participants, with even FFM and SMM showing a mild decrease despite physical activity support. FFM reduction is clearly associated with poor metabolic state [19][20][21]. VFA increase and SMM decrease appear closely related to aggravation of insulin resistance and progression of NAFLD [22]. Therefore, another recommended index is the SMM to VFA radio (S/V), which simultaneously describes variations in both parameters [22]. A signi cant increase in S/V in all subjects was observed during the whole intervention process. However, in the rst 3 months, the S/V increase in the CIMWM group was higher than that in the TMWM group. The decline in VFA was greater than that in SMM, supporting the effectiveness of our weight management interventions, especially for the CIMWM group.
NFS is a clinical indicator of liver brosis [23]. The NFS values in individuals with levels in the upper tertile decreased signi cantly after 6 months of weight intervention, indicating that our medical weight management strategy is effective not only for weight loss but also improvement of NAFLD. NFS in subjects with lower tertile levels was slightly increased, but remained under the 0.676 cut-off point for liver brosis. This nding is possibly related to increasing BMI among some individuals. However, plasma ALT and AST concentrations are still regarded as better clinical indicators than other noninvasive biomarkers/scores for de nitive diagnosis of nonalcoholic steatohepatitis (NASH) and/or advanced brosis in patients with T2DM [24]. Consistently, we observed a signi cant reduction in ALT and AST levels in all participants after the intervention, which was more marked in the CIMWM than TMWM group at the end of the study period.
At baseline, a negative correlation existed between NFS and S/V. S/V is reported to be an independent factor that affects the controlled attenuation parameter (CAP) re ecting hepatic fat accumulation and liver stiffness measurements (LSM), representing hepatic in ammation and brosis [22]. In another study on NAFLD patients [23], low S/V indicated poorer hepatic condition, con rmed with liver transient elastography. Accumulating basic and clinical researches have focused on cross-talk among muscle, fat, and liver components [25]. A recent review suggests that adipose tissue function is a critical driver of NAFLD and NASH. Cytokines and hormones secreted by adipose tissue could affect the liver by regulating the lipid ux and affecting hepatocyte function via exosomal signaling [26]. Skeletal muscle is also regarded as an endocrine organ [27] that regulates myokine secretion. Circulating myonectin linking skeletal muscle to lipid metabolism in the liver and adipose tissue provides insights into tissue cross-talk that underlies the integrated control of whole-body metabolism [ Our study has some limitations that should be taken into consideration. Evaluation of liver steatosis and brosis with FibroTouch, FibroScan, or magnetic resonance elastography may present more e cient strategies. In addition, we were unable to develop an accurate method to collect data on frequency and intensity of physical activity from our subjects.

Conclusion
Our study on weight management in Chinese adults with obesity showed that the Companion-Intensive Multi-aspect Weight Management approach is more effective than the Traditional Multi-aspect Weight Management approach, leading to greater improvement of BMI, ALT, BFM, VFA, and S/V parameters. The CIMWM strategy was particularly effective in motivating patient adherence to the intervention and lifestyle reprogramming. Metabolic memory in terms of continuous reduction of BMI, VFA, and ALT was retained up to 6 months despite participants being transferred to the self-monitoring phase during the last 3 months of the study. Additionally, the weight management strategy was bene cial in decreasing the NFS among subjects with levels in the highest tertile. Our results support the potential of a face-to-face program coupled with an online intervention tool in achieving scienti cally effective metabolic syndrome management and providing a positive contribution to national health. From the perspective of the

Authors' contributions
Wanzi Jiang and Shuai Ma conducted data collection as well as analysis and wrote the rst draft of the manuscript. Shushu Huang conceived the idea for the work. Hongwen Zhou designed and conducted the research, contributed to data interpretation and manuscript revision, and provided medical supervision.
All co-authors contributed to implementation of the experimental trial and design of the protocol for intervention. All authors read and approved the nal manuscript. Hongwen Zhou is the guarantor of this work and, as such, has full access to all the data and takes responsibility for data integrity and accuracy of analysis.

Acknowledgments
The following contributors assisted with patient recruitment, intervention and/or data collection during the trial:    Between-group differences analyzed using independent samples t-test for normal data and Mann-Whitney U test for skewed data c. Compared with 3months and baseline, ***p < 0.001, **p < 0.01 *p < 0.05; compared with 6 months and baseline † † † p < 0.001 † † p < 0.01 † p < 0.05; compared with 3months and 6 months, ‡ ‡ ‡ p < 0.001 ‡ ‡ p < 0.01 ‡ p < 0.05 a. Data were expressed as mean ± SD when normally distributed, while median (interquartile range) was used when data did not satisfy normality criteria b. Between-group differences analyzed using independent samples t-test for normal data and Mann-Whitney U test for skewed data  a. Data were expressed as mean ± SD when normally distributed, while median (interquartile range) was used when data did not satisfy the normality criteria b. Between-group differences analyzed using independent samples t-test for normal data and Mann-Whitney U test for skewed data   NFS and VFA (d) Relationship between baseline NFS and S/V *p < 0.05 (c) Changes in S/V from baseline to 6 months. Comparison between the two groups at the same time-point, *p < 0.05 (d) Changes in ALT from baseline to 6 months. Comparison between the two groups at the same time-point, *p < 0.05 (e) BFM and FFW reduction from baseline to 6 months. Comparison between the two groups at the same time-point, *p < 0.05