Rimonabant improves cholesterol, insulin resistance and markers of non-alcoholic fatty liver in morbidly obese patients: a retrospective cohort study

To the Editor: Obesity is associated with co-morbidities including non-alcoholic fatty liver disease (NAFLD), diabetes, obstructive sleep apnoea and hypertension. This report describes clinical experience with the recently withdrawn endocannabinoid CB-1 antagonist rimonabant (1,2) in morbidly obese patients presenting for bariatric surgery. Clinical data were downloaded from the Vector/Diabeta-3 database [Health Information Systems (UK) Limited, London, UK] for the period 2006–2009 using the search terms: presentation- obesity and BMI > 40 kg/m2. This database includes records of patient demography, ethnicity, anthropometry (including waist circumference), clinical diagnoses, metabolic and biochemical assessments and medications at each clinic visit for nursing, dietetic and medical visits. Patients had renal, liver, thyroid, lipid, insulin and glucose and HbA1c profiles measured. Insulin resistance was calculated from the Homeostasis index (HOMA) (3). Data were analysed using Systat 12 (Systat Corp; Hounslow, Middlesex, UK). Initial and final visits data were compared using parametric methods after log transformation for non-Gaussian distributed variables. A p-value of < 0.05 was considered significant. Data were available for 222 patients for 12.3 months (Figure S1). Patients were aged 51 ± 11 years, 40% men and had a BMI of 49.2 ± 8.8 kg/m2 (Table S1). Diabetes was present 36%, nocturnal hypoventilation in 24%, hypertension in 66%, lymphoedema in 10% and previous depression in 11%. After initiation 35 patients discontinued rimonabant within 4 weeks because of side-effects or by personal choice leaving 187 patients in whom > 6 month compliance data were available (Table 1). Discontinuations occurred because of depression/mood change (18 patients; two requiring psychiatric admission), nausea (four patients), and suicidal ideation (one patient) (Table S2). Continued follow-up of the 35 patients receiving dietary counselling alone showed reductions in weight of 1.3 ± 12 kg, BMI 0.7 ± 4.1 kg/m2, blood pressure 0.7 ± 14/0.5 ± 9 mmHg, glucose 0.5 ± 1.3 mmol/l and an ALT increase 1 ± 6 IU/l at 12 months. The lipid profiles were essentially with unchanged: total cholesterol difference (0.0 ± 0.9 mmol/l), triglycerides (0, range −6.7 to 1.3 mmol/l), HDL-C (0.03 ± 0.22 mmol/l) and LDL-C (−0.13 ± 0.70) mmol/l. Patients compliant with long-term rimonabant therapy were aged 52 ± 12 years and 40% men and weighed 141 ± 28 kg [50 ± 9 kg/m2 (range 35–89)]; diabetes was present in 38%, obstructive sleep apnoea in 25%; hypertension in 66%, lymphoedema in 12% and hepatic steatosis in 8%. Treatment resulted in a non-significant 4.6 ± 11 kg (median 4.5; range −40 to 41 kg) weight loss equivalent to 1.7 ± 4 BMI points (p = 0.07). This was associated with a 6.6/3.8 mmHg reduction in blood pressure, 1.3 mmol/l (20%) in glucose; 0.5% absolute change in HbA1c and a 17% reduction in alanine transaminase (ALT). Lipid profiles improved with a 15% reduction in total cholesterol, 17% in LDL-C and 37% in triglycerides with a non-significant 7% increase in HDL-C. Insulin levels decreased by 39% (p = 0.01) and homeostasis index (HOMA-IR) by 55% (p = 0.005). No change was seen in beta cell index, creatinine, estimated glomerular filtration rate or creatine kinase (CK). Eight patients discontinued rimonabant therapy as a result of side-effects. One patient had severe depression requiring psychiatric assessment. Rimonabant, an endocannabinoid CB1 antagonist (withdrawn in 2009), was licenced in 2006 based on the results of the rimonabant in obesity (RIO) trials for reduction of weight in patients with BMI 27–40 kg/m2. Weight was reduced by 10% in a population of average BMI 34 kg/m2 (weight 98–100 kg) allied with beneficial metabolic effects on blood pressure, triglyceride-rich lipoproteins and glycaemia (absolute HbA1c reduction 0.7%) (1,2). The Comprehensive Rimonabant Evaluation Study of Cardiovascular Endpoints and Outcomes (CRESCENDO) cardiovascular outcomes trial showed an increased incidence of depression and suicide with rimonabant and the drug was withdrawn (4). No trial investigated the effect of rimonabant on patients with BMI > 40 kg/m2. Presurgical management of patients due to undergo bariatric surgery typically involves the prescription of a low calorie, low carbohydrate diet to reduce hepatomegaly and fatty liver. In prebariatric patients a trial of 19 patients showed an average 4.4 kg loss with this intervention as opposed to 2.6 kg with calorie restriction (5). In this audit rimonabant therapy resulted in a non-significant 3% weight loss (4.6 ± 11 kg; 1.7 ± 4 kg/m2) allied with predictable reductions in blood pressure, glucose; and HbA1c, triglycerides and increased HDL-C. Metabolic changes matched those predicted by the Framingham study of approximately 1.5/1 mmHg/kg and 0.1% HbA1c/kg (6) and those seen in the RIO trials (1). Novel findings included a 17% reduction in ALT, 15% in total cholesterol, and 17% in LDL-C. A study of seven patients with non-alcoholic steatohepatitis diagnosed by biopsy who received rimonabant therapy showed reductions of 5.9 kg in weight (BMI 2.4 kg/m2), 54% in ALT, 20% in LDL-C and 16% in triglycerides (7). Both gamma-glutamyl transferase (not measured here) and ALT levels are associated with NAFLD and the metabolic syndrome (8). The reduction in insulin resistance, transaminases and LDL-C seen in this study may reflect a reduced net production of very low density lipoprotein and a reduction in hepatic steatosis (8). The therapy was tolerated by 84% of patients initially reducing to 80% and was limited to the occurrence of rapid onset nausea, depression and mood changes. This study has a number of disadvantages. It is a retrospective cohort study with all the disadvantages of these studies. The untreated group defined by early-onset of lack of tolerance of rimonabant provides a small self-selected parallel comparison group but shows weight reductions similar to those seen in a randomised study of dietary therapy in a prebariatric population of 1.7 kg (5). Rimonabant was withdrawn because of its increased rates of neuropsychiatric side-effects (2009). Endocannabinoid receptors exist in the liver and adipose tissue as well as in the brain and are involved in the progression of NAFLD. Recently a hydrophilic derivative of rimonabant (AM-6545) with only peripheral CB1 activity has been investigated in obese mouse models (9). AM-6545 produced the metabolic changes expected with rimonabant but showed no evidence of anxio-genesis in mice. This implies that rimonabant may have a primary peripheral action and that central actions are only responsible for its secondary side-effects. It remains to be seen whether these effects can be confirmed in man. This study showed that when rimonabant was available, medium term prescription resulted in a non-significant weight loss in morbidly obese patients but was associated with metabolic benefits including improved insulin resistance and lipid profiles consistent with an improvement in hepatic steatosis. Dr Wierzbicki was a member of National and International Advisory boards for Sanofi-Aventis on rimonabant in 2005–2008, was UK lead investigator on the ETERNAL trial (2007–2009) and an investigator on the CRESCENDO trial (2007–2009). He also received travel support and lecture honoraria from this source. Figure S1. Database characteristics of the retrospective cohort analysis of rimonabant therapy. Table S1. Baseline characteristics of patients adherent to treatment pathway and prescribed rimonabant therapy. Comparisons were performed between the continued and discontinued therapy groups. Table S2. Discontinuations associated with rimonabant therapy. Patients could have more than one reason for discontinuing treatment. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.