Effects of short‐term energy restriction on liver lipid content and inflammatory status in severely obese adults: Results of a randomized controlled trial using 2 dietary approaches

Short‐term very‐low‐energy diets (VLEDs) are used in clinical practice prior to bariatric surgery, but regimens vary and outcomes of a short intervention are unclear. We examined the effect of 2 VLEDs, a food‐based diet (FD) and a meal‐replacement plan (MRP; LighterLife UK Limited, Harlow, UK), over the course of 2 weeks in a randomized controlled trial. We collected clinical and anthropometric data, fasting blood samples, and dietary evaluation questionnaires. Surgeons took liver biopsies and made a visual assessment of the liver. We enrolled 60 participants of whom 54 completed the study (FD, n = 26; MRP, n = 28). Baseline demographic features, reported energy intake, dietary evaluation and liver histology were similar in the 2 groups. Both diets induced significant weight loss. Perceived difficulty of surgery correlated significantly with the degree of steatosis on histology. There were reductions in the circulating inflammatory mediators C‐reactive protein, fetuin‐A and interleukin‐6 between baseline (pre‐diet) and post‐diet. The diets achieved similar weight loss and reduction in inflammatory biomarkers. There were no significant differences in perceived operative difficulty or between patients' evaluation of diet satisfaction, ease of use or hunger frequency. Non‐alcoholic fatty liver disease histology assessments post‐diet were also not significantly different between diets. The results of this study show the effectiveness of short‐term VLEDs and energy restriction, irrespective of macronutrient composition, although the small sample size precluded detection of subtle differences between interventions.

Short-term very-low-energy diets (VLEDs) are used in clinical practice prior to bariatric surgery, but regimens vary and outcomes of a short intervention are unclear. We examined the effect of 2 VLEDs, a food-based diet (FD) and a meal-replacement plan (MRP; LighterLife UK Limited, Harlow, UK), over the course of 2 weeks in a randomized controlled trial. We collected clinical and anthropometric data, fasting blood samples, and dietary evaluation questionnaires. Surgeons took liver biopsies and made a visual assessment of the liver. We enrolled 60 participants of whom 54 completed the study (FD, n = 26; MRP, n = 28). Baseline demographic features, reported energy intake, dietary evaluation and liver histology were similar in the 2 groups. Both diets induced significant weight loss. Perceived difficulty of surgery correlated significantly with the degree of steatosis on histology. There were reductions in the circulating inflammatory mediators C-reactive protein, fetuin-A and interleukin-6 between baseline (pre-diet) and post-diet. The diets achieved similar weight loss and reduction in inflammatory biomarkers. There were no significant differences in perceived operative difficulty or between patients' evaluation of diet satisfaction, ease of use or hunger frequency. Non-alcoholic fatty liver disease histology assessments postdiet were also not significantly different between diets. The results of this study show the effectiveness of short-term VLEDs and energy restriction, irrespective of macronutrient composition, although the small sample size precluded detection of subtle differences between interventions.

| INTRODUCTION
Ectopic lipid accumulation is a common factor underlying nonalcoholic fatty liver disease (NAFLD), insulin resistance and Type 2 diabetes. 1 Metabolic dysfunction is linked to pro-inflammatory mediators such as interleukin-6 (IL-6), fetuin-A and C-reactive protein (CRP), which also contribute to progression to inflammatory nonalcoholic steatohepatitis. 2

| Sampling and data collection
Clinical and anthropometric data were collected during visits and from patients' notes. A pre-intervention fasting blood test was carried out, and weight and height were measured at the time of the standard preoperative appointment. On completion of the diet, participants completed a short qualitative evaluation. Participants were weighed post-diet on the morning of their bariatric surgery. When the participant was cannulated for anaesthesia, a further fasting blood sample was taken. Intra-operatively a liver biopsy was taken and the surgeon completed a visual assessment of perceived operative difficulty.

| Outcome measures
The primary outcome was histological assessment of a liver biopsy performed at the time of bariatric surgery (post-diet). Secondary outcomes were change in total body weight (pre-to post-diet), circulating inflammatory markers (pre-to post-diet), peri-operative visual assessment of the liver for perceived difficulty of surgery and evaluation of dietary approach (post-diet).

| Sample size
The sample size was based on an estimation of the difference in liver lipids after following the FD or MRP. For 80% power (at P = .05) 98 participants would need to be randomized (49 to each arm) to see a statistically significant difference in the level of steatosis (INSTAT GRAPH PAD V2.0 software).

| Statistical methods
The data were analysed using IBM SPSS Statistics for Windows version 22 (IBM Corporation). Two-tailed non-parametric statistical tests were used: Wilcoxon signed rank test, Spearman's Rho, Mann-Whitney U-test and Kendall's Tau C-test. P values <.05 were taken to indicate statistical significance. Bonferroni correction for multiple comparisons was carried out where stated.
Further details of methods and data collection can be found in File S1.

| RESULTS
A total of 196 potential participants were screened for eligibility over a 25-month period of recruitment, of whom 105 verbally consented to the study. Of these, 60 completed informed consent, were randomized and started the trial, and 54 completed the study. Six participants did not complete the study; 5 because of operations being cancelled and 1 participant was started on insulin. One adverse event and 1 serious adverse event, were reported but were not deemed to be related to the intervention.

| Baseline data
Participants' baseline data are shown in Table 1. BMI ranged from 41.5 to 66.8 kg/m 2 , with a median of 50.7 kg/m 2 , and 81.5% of the participants were women, which is consistent with the national average for patients undergoing bariatric surgery. Age ranged from 24 to 65 years, with a median of 45 years. There was no significant difference between diet groups demographically or anthropometrically (age: P = .245; BMI: P = .691; gender: P = .568; diabetes status: P = .860).

| Liver tissue histology
After the intervention, there was no significant difference between the groups with regard to steatosis: 12 of 24 participants in the FD group (50%) and 18 of 28 participants in the MRP group (64%) had histological evidence of steatosis ( Table 2). Three of 24 participants in the FD group (12.5%) and 4 of 28 participants in the MRP group (14.3%) had evidence of moderate or severe steatosis (steatosis grades 2-3), while 37% of participants in the FD group, and 39% of participants in the MRP group had evidence of significant portal or periportal fibrosis, but the difference between groups was not statistically significant. No significant difference was found between the 2 diet groups in steatosis grade (P = .349), liver cell injury (P = .567), portal inflammation (P = .611), lobular inflammation (P = .455), or fibrosis stage (P = .605).

| Weight change and self-reported intake
The median total body weight loss was 3.6% in the FD group, and 3.4% in the MRP group. Results were not significantly different between diet groups, for body weight loss or percentage weight loss (P = .494 and P = .993, respectively [ Table 1]).
Median energy intakes (kcal/d) were not significantly different between diet groups (P = .311). Macronutrient energy intake (% total energy intake) differed between the 2 diets, with protein, fat and carbohydrate intakes being significantly different between the groups (P < .001, Mann-Whitney U-test Bonferroni correction α error probability threshold .016). Fluid intake recommendations were a minimum of 2 L/d, with median reported intakes of 2.2 L/d for the FD group, and 2.5 L/d for the MRP group.

| Surgeons' assessment
In 22 of the 54 cases, a surgeon completed a visual assessment of (1) access to the gastro-oesophageal junction, (2) liver volume and (3) liver rigidity. Increasing perceived difficulty was associated with higher percentages of steatosis (as assessed histologically). There was a correlation between visual assessment by the surgeons and the 3 assessment categories (above). There was no significant difference between diets for the 3 visual analogue assessments: gastric access (P = .513), liver volume (P = .891) and liver rigidity (P = .702).

| Blood results
Concentrations of CRP and fetuin-A reduced significantly after both diets. IL-6 reduction was not significant.

| Dietary evaluation
Participants' self-reported evaluation showed that overall satisfaction with the diets was 92% and 85% (satisfied or very satisfied) and 82% There was no statistically significant difference between weight loss (P = .993) or estimated energy intake (P = .311), however all macronutrient intakes were significantly different (P < .001).
Data are median (range) unless stated otherwise.  of participants in the FD group and 93% in the MRP group reported that they found the diet easy to follow. A total of 19% in the FD group and 15% in the MRP group reported always feeling hunger, whereas 8% and 11% respectively never felt hungry. There were no significant differences between evaluation responses for satisfaction, ease of use or hunger frequency between diets. Further results are provided in File S1.  The main limitation to the present study is the failure to reach the desired sample size; however, the results for NAFLD were similar to observations elsewhere after an energy restrictive diet when applied in the severely obese population. 15 In addition, we were unable to obtain liver biopsies or to perform non-invasive estimation of hepatic lipid content using MRI both prior to and after the interventions, which would have allowed a more conclusive outcome appraisal. Randomization appears to have been effective, however, and no significant differences in demographic or anthropometric profile were observed between the 2 groups at baseline.

| DISCUSSION
In conclusion, the present results indicate that energy intake of 715 kcal/d for 2 weeks improves operative conditions regardless of the macronutrient composition of the preoperative diet. Participants were satisfied with the VLEDs, which they also found easy to use.
The fatty liver disease we found was in line with similar studies and is suggestive of an effect of both diets in reducing liver lipid concentrations. Reductions in inflammatory biomarkers require further investigation. We were unable to demonstrate a difference in either case between diets, and the small sample size may have precluded detection of subtle differences.