Effects of liraglutide on gallbladder emptying: A randomized, placebo‐controlled trial in adults with overweight or obesity

2.1. Trial design

The trial was conducted at a single clinical research centre in Denmark in accordance with the Declaration of Helsinki15 and Good Clinical Practice guidelines,16 and was registered at http://clinicaltrials.gov (identifier NCT02717858). The protocol was approved by the local health authority and independent ethics committee, and participants provided written informed consent before trial commencement.

Participants received treatment with once‐daily subcutaneous liraglutide or placebo, as an adjunct to nutritional and physical activity counselling, starting at 0.6 mg and with 0.6‐mg weekly increments to 3.0 mg (Figure 1). A 5% minimum weight‐loss target was set, with the aim of achieving similar weight losses in each group. To stimulate gallbladder emptying, a liquid meal test (600 kcal, 23.7 g fat) was performed at baseline, after the first 0.6‐mg dose and after 12 weeks, and was considered to be a suitable physiological stimulus.17 Gastric emptying was also assessed in the 4‐hour postprandial period of the meal test using the paracetamol absorption technique.18

2.2. Participants

Men or women aged between 18 and 64 years (inclusive), with a body mass index (BMI) ≥27.0 kg/m², stable body weight (<3 kg self‐reported change during the previous 90 days), and an ultrasound assessment of gallbladder volume of acceptable quality (investigator judgment) at screening were included in the trial. Key exclusion criteria were: a history of gastrointestinal surgery or other medical procedures precluding a gallbladder emptying assessment (appendectomy was allowed) or any significant digestive disease (investigator judgement); a diagnosis of type 1 or 2 diabetes mellitus; a history of pancreatitis (acute or chronic) or any gallbladder disease (cholelithiasis, gallbladder sludge, polyps); or pregnancy, breast‐feeding and inadequate contraception use. Full enrolment criteria and exclusion criteria associated with the meal test are included in Tables S1 and S2 (Supporting Information).

2.3. Treatment and randomization

Eligible participants were randomized 1:1 to liraglutide (n = 26) or placebo (n = 26; Figure 1). Randomization codes were sent by the sponsor to the site in sealed units. Liraglutide and placebo were provided in pre‐filled pen‐injectors (Novo Nordisk A/S, Bagsværd, Denmark). As female hormones can influence gallbladder emptying,19 randomization was stratified by sex. The sponsor, participants and investigators remained blinded to treatment allocation.

2.4. Counselling in healthy nutrition and physical activity

During the treatment period, trial participants received five individualized counselling sessions on nutrition and physical activity from a certified dietician, with a target weight loss of ≥5% of their initial body weight over the 12‐week treatment period. Participants were advised to follow a hypocaloric diet throughout the treatment period, containing a maximum 30% of energy from fat (maximum 10% energy from saturated fats), and ~20% from protein and ~50% from carbohydrates, with an energy deficit of 500 kcal/d compared with the participant's estimated total energy expenditure. If participants were unable to lose weight after 4 weeks of treatment, more intense counselling was provided and recalculation of the recommended energy intake was allowed to obtain a greater energy deficit than the original hypocaloric diet. Participants who did not achieve the 5% weight‐loss target were still included in the data analyses.

Adherence to the recommended diet was at the dietician's discretion. Increased physical activity was encouraged, with a goal of 60 minutes of moderate‐ to high‐intensity physical activity per day and a recommended >150 minutes of moderate‐intensity physical activity per week.20

2.5. Endpoints

During each of the three meal tests, gallbladder volume was measured by ultrasonography in the fasted state and throughout the 240‐minute period after the start of the meal, at 14 predefined time points. Gallbladder motility endpoints were derived from the gallbladder volume−time curves over the 240‐minute period, at baseline, after the first 0.6‐mg liraglutide dose and at steady‐state liraglutide 3.0 mg after 12 weeks. The primary endpoint was GBEF_max after 12 weeks. Secondary endpoints related to gallbladder motility comprised: GBEF_max after the first 0.6‐mg dose; fasting gallbladder volume; area under the GBEF−time curve 0 to 60 minutes after the start of the meal (GBEF AUC_{0‐60 min}); time to GBEF_max (t_max); and time from t_max to when the gallbladder had reverted to the fasting volume after the 0.6‐mg dose and at steady‐state after 12 weeks of treatment.

Gastric emptying endpoints were derived from the paracetamol concentration−time curves over the same period and for the same doses as described above: paracetamol AUC_{0‐240 min} and AUC_{0‐60 min}; maximum paracetamol concentration (C_max); and paracetamol t_max.

Other endpoints included change from week 0 to week 12 in body weight and secondary safety endpoints, comprising adverse events and changes from screening to week 12 in haematology, biochemistry, including fasting lipase and amylase, calcitonin, vital signs and physical examination. Adverse events of special medical interest (acute gallstone disease, neoplasm and pancreatitis) had additional data collection.

The timing of assessments is described in the Supplemental Methods (Supporting Information).

2.6. Estimation of gallbladder volume

Gallbladder volume during meal tests was calculated by the ellipsoid method via ultrasound assessment of longitudinal and cross‐sectional diameters21 by a maximum of two investigators; an intra‐ and inter‐observer variation of <10% with respect to gallbladder volume was achieved. Participants were in a supine position during the assessment. The gallbladder was measured in three dimensions: one longitudinal (D1) and two cross‐sectional diameters (D2 and D3) for calculating the volume using the formula “volume = π/6 × D1 × D2 × D3”.22 Fasting gallbladder volume was estimated based on the average of two sequential measurements.

2.7. Meal test and gastric emptying

At baseline, after the first 0.6‐mg dose and after 12 weeks, a 4‐hour meal test was performed in the morning after an overnight fast. The liquid meal (250 mL of nutritional supplement drink [Nutridrink Compact; Nutricia AB, Allerød, Denmark]) had a total energy content of ~600 kcal and a macronutrient composition of 35% energy from fat (23.7 g), 16% from protein and 49% from carbohydrate.

Paracetamol (1500 mg, three effervescent 500 mg tablets) was dissolved in 50 mL sterile water and mixed into the liquid meal (final volume ~300 mL/340 g) for measurement of gastric emptying.23, 24 The liquid meal was ingested within 10 minutes, and consumption time standardized between meals. Before the start of the meal and for 240 minutes postprandially, the gallbladder volume was assessed and blood samples were taken at nine predefined timepoints for the measurement of paracetamol (Supplemental Methods [Supporting Information]).

2.8. Statistical analyses

The sample size was based on the expected precision of the estimated difference in the primary endpoint (12‐week GBEF_max) between the two treatment groups using a two‐sided 95% confidence interval (CI) derived from the t‐distribution. Based on data from a previous study,22 the SD value for the GBEF_max following a liquid meal was calculated to be 11%, with a mean GBEF_max of 71%. With 40 completing participants in the trial and an SD of 12%, there was a probability of ≥80% for achieving a 95% CI for the true treatment difference of GBEF_max within [d − 8.4%; d + 8.4%], where d is the estimated treatment difference. The results were considered sufficiently precise to evaluate the primary objective. To account for participants discontinuing, it was planned to include 48 participants in the trial to ensure that 40 completed (assuming a drop‐out rate of 20%).

All analyses were carried out on randomized individuals receiving at least one treatment dose. The GBEF values were, for each timepoint, calculated from the gallbladder volume (as change from the fasting gallbladder volume) as: GBEF(t) = 100% × [vol_fasting − vol(t)]/vol_fasting, where vol_fasting was the mean of two gallbladder volume assessments measured within 15 minutes before the meal, and vol(t) was the gallbladder volume measured at each timepoint, t. The primary endpoint was analysed using a linear normal model, which included treatment and sex as factors, and baseline body weight and baseline GBEF_max as covariates. There was no imputation for missing data.

Additional gallbladder‐related and gastric emptying endpoints were analysed in the same way as the primary endpoint, using the corresponding baseline value as a covariate in the model, except that the time to GBEF_max was summarized using descriptive statistics. Gastric emptying endpoints were log transformed for analysis. AUCs were calculated using the linear trapezoidal method. Body weight and safety endpoints were summarized using descriptive statistics. Additional prespecified and exploratory analyses are described in the Supplemental Methods (Supporting Information). The statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA).

3.1. Participant characteristics

A total of 90 individuals were screened for eligibility, of which 34 were screening failures. Four individuals were withdrawn prior to randomization, 2 because of protocol violations and 2 because of cholelithiasis events, as per exclusion criteria, that were discovered by ultrasonography before the start of the meal test. All 52 randomized individuals were exposed to the trial drug; 5 were withdrawn after randomization, and 47 completed the trial. In the liraglutide group, 1 withdrawal was attributable to protocol violation and 2 participants withdrew of their own accord; in the placebo group, 1 withdrawal was attributable to an adverse event (migraine) and 1 was attributable to protocol violation. All withdrawn individuals attended the follow‐up visit. The trial was conducted between March 16, 2016 and February 27, 2017.

Baseline characteristics of the trial participants were similar between treatment groups (Table 1).

3.2. Gallbladder motility

Figure 2 shows mean gallbladder volume (A) and ejection fraction (B) profiles from 0 to 240 minutes after the start of the liquid meal at baseline (before treatment initiation), after the first 0.6‐mg dose, and after 12 weeks of treatment. In each case, the gallbladder volume decreased after meal ingestion (Figure 2A and Figure S1, Supporting Information), and, accordingly, GBEF increased in both groups (Figure 2B). After the first 0.6‐mg dose, the mean GBEF was higher with liraglutide than with placebo from time ~100 to ~240 minutes and remained higher with steady‐state liraglutide 3.0 mg from ~150 to ~240 minutes (Figure 2B).

After 12 weeks, the estimated mean GBEF_max (primary endpoint) did not differ for liraglutide 3.0 mg vs placebo (Table 2). Neither was any treatment difference observed in the estimated mean GBEF_max after the first liraglutide 0.6‐mg treatment dose. The AUC_{0‐60 min} and fasting gallbladder volume did not differ for liraglutide vs placebo either after the first 0.6‐mg dose or with liraglutide 3.0 mg after 12 weeks of treatment. The median GBEF t_max (time to reach maximum gallbladder contraction) at baseline was 77 minutes in the liraglutide group vs 67 minutes in the placebo group. The median (range) t_max increased to 104 (21‐240) minutes after the first liraglutide 0.6 mg dose vs 77 (0‐183) minutes with placebo and to 151 (11‐240) minutes vs 77 (22‐212) minutes at steady‐state (no statistical testing was done). There was a wide variation in individual GBEF t_max values (Figure S2, Supporting Information). More than half of the GBEF profiles did not return to fasting levels within the 240‐minute period (baseline: 27/52 = 52%; single dose: 33/51 = 65%; steady‐state: 29/47 = 62%); therefore, the time from t_max to the time of the gallbladder reverting to the fasting volume was not further analysed.

3.3. Gastric emptying

Gastric emptying was slowed after the first 0.6‐mg dose of liraglutide as compared with placebo, as indicated by significant reductions in the paracetamol C_max and AUC_{0‐240 min} with liraglutide, but no treatment effect was seen with liraglutide 3.0 mg at week 12 (Table S3, Supporting Information). There were no observed treatment effects on paracetamol t_max after the first 0.6‐mg dose or at steady state (Table S3, Supporting Information). There was no apparent relationship between the change in gastric emptying in the first hour of the meal test and the change in the time to GBEF_max after the first 0.6‐mg dose or with liraglutide 3.0 mg after 12 weeks (Table S3, Supporting Information).

3.4. Body weight

The mean (SD) percent body weight loss after 12 weeks of treatment was 8.2 (1.8)% in the liraglutide 3.0 mg group vs 5.5 (3.6)% in the placebo group, equivalent to −7.9 (2.1) kg vs −5.5 (3.5) kg. All participants that completed the trial in the liraglutide group, as well as 13/24 participants (54%) of those in the placebo group, achieved the weight‐loss goal of ≥5%. Individual changes in body weight over 12 weeks are shown in Figure S4 (Supporting Information). There was no apparent relationship between the relative change in the time to GBEF_max and the relative change in body weight from baseline to week 12 (Figure S5, Supporting Information).

3.5. Safety

There were no unexpected safety findings in the present trial and no new safety concerns were raised. Overall, 138 adverse events were reported by 25 participants (96.2%) in the liraglutide group as compared with 62 events by 24 participants (92.3%) in the placebo group. Most of the events were mild in severity (117 of the 138 events [85%] in the liraglutide group and 46 of the 62 events [74%] in the placebo group). As in other trials with liraglutide, the most frequently reported adverse events were gastrointestinal disorders, with nausea and constipation being the most commonly reported events. Overall, 23 participants (88.5%) in the liraglutide group reported 71 gastrointestinal events and 9 participants (34.6%) in the placebo group reported 15 events.

One serious adverse event (lower limb fracture) was reported by a participant in the placebo group during the follow‐up period. One non‐serious, non‐symptomatic cholelithiasis event was reported in the liraglutide group. The event was discovered with ultrasonography during the end‐of‐trial meal test assessment, 12 weeks after start of treatment. No individuals in the liraglutide group discontinued the trial as a result of adverse events vs one individual in the placebo group who discontinued because of a worsening of migraine event 14 days after starting treatment. No pancreatitis or neoplasm events were reported. No clinically relevant safety findings were identified in haematological or biochemical variables.