International Journal of Obesity (1998) 22, 1041±1045 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00 http://www.stockton-press.co.uk/ijo Cimetidine reduces weight and improves metabolic control in overweight patients with Type 2 diabetes

G Stùa-Birketvedt*1, PN Paus3, R Ganss3, OC Ingebretsen2 and J Florholmen1

1Laboratory of Gastroenterology, Institute of Clinical Medicine, University of Tromsù, Tromsù; 2Biochemical Department, University Hospital of Tromsù, Tromsù; and 3Medical Department, Central Hospital of Akershus, Nordbyhagen, Norway

OBJECTIVE: To investigate the weight-reducing effect of cimetidine in overweight patients with Type 2 diabetes. DESIGN: A 12-week clinical intervention study of 400 mg cimetidine prescribed three times daily in a randomised, double-blind, placebo-controlled design. SUBJECTS: Forty-three overweight patients with Type 2 diabetes (age 18±65 y, body mass index (BMI) 27.2± 48.2 kg=m2). MEASUREMENTS: Body weight, BMI, body fat, waist and hip circumference, waist=hip ratio, blood pressure, fasting blood glucose, HbA1c, plasma concentrations of insulin, insulin=glucose ratio and lipids at the start and after 12 weeks, and daily recordings of appetite. RESULTS: Subjects given cimetidine (n ˆ 19) and placebo (n ˆ 24) lost 5.0Æ 2.2 kg (meanÆ s.d.) and 1.3Æ 1.1 kg, respectively. Signi®cant reductions were observed in appetite, body fat (29.9Æ 6.6% to 25.3Æ 7.4%), waist circumfer- ence (111.5Æ 10.3 cm to 107.4Æ 10.6 cm), waist=hip ratio (0.96Æ 0.08 to 0.94 Æ 0.08), and systolic and diastolic blood pressure (reductions of 6.9Æ 11.4 mm Hg and 6.0Æ 6.6 mm Hg, respectively) in cimetidine group only. Signi®cant decreases in fasting concentrations of blood glucose, HbA1c, plasma insulin, insulin=glucose ratio, plasma triglycerides and a signi®cant increase in plasma high-density lipoprotein cholesterol were observed in the cimetidine group only. CONCLUSIONS: Cimetidine reduces appetite and body weight, and improves metabolic control in overweight subjects with Type 2 diabetes.

Keywords: H2-receptor antagonist; hyperlipidemia; insulin resistance; obesity; weight reduction

Introduction heterogeneous pathogenesis, with interactions at least between visceral obesity and genetic factors.11 We have previously reported that the H -receptor According to previous reports, there is a high pre- 2 antagonist, cimetidine, reduces weight in healthy valence of Type 2 diabetes among obese subjects.1,2 overweight subjects.12 The mechanism of this effect Hyperinsulinism and insulin resistance are character- is unknown. In rats, H -receptor antagonists reduce istic features of both Type 2 diabetes and obesity (for 2 appetite and weight gain by a non-gastric acid related review, see Ref. 3). In obese subjects the insulin mechanism.13 We have shown that cholecystokinin resistance is highly correlated to the mass of visceral (CCK) is a candidate mediator of this effect on fat.4,5 However, it remains controversial whether satiety.14 obesity is the major and=or an independent contributor The objective of this study was to investigate the to the insulin resistance in Type 2 diabetes.6,7 effect of cimetidine on body weight and metabolic It is well recognized that weight reduction is pro®le in overweight patients with Type 2 diabetes in associated with bene®cial effects on metabolic con- a randomized, double blind, placebo-controlled study. trol.8 A modest weight reduction may strongly improve the metabolic pro®le.9 Weight reduction induces improved glycemic control, reduced hepatic glucose output, increased insulin action in the liver and the peripheral tissue, and improved insulin secre- Materials and methods tion.10 However, a weight reduction to normalcy elicits a complete normalization of the peripheral insulin action in only few patients.10 This implies a Subjects Sixty-two patients with Type 2 diabetes were recruited from the outpatient ®les of two Norwegian hospitals. The inclusion criteria were age 18±65 y and *Correspondence: Dr Grethe Stùa-Birketvedt, University of body mass index (BMI) > 25 kg=m2. Patients with Pennsylvania, 3600 Market Street, Suite 732, 19104 PA, USA. Received 11 September 1997; revised 21 May 1998; accepted serious cardial, renal or hepatic diseases were 26 May 1998 excluded. Patients treated with antacids, gastric acid Cimetidine reduces weight G Stùa-Birketvedt et al 1042 antisecretory agents, insulin or corticosteroids were Ethics also excluded. The study was conducted according to the Helsinki declarations and with the approval of the Ethical Committee of The Health Region V, Norway. Study design The patients were randomized into two groups accord- Statistical analysis ing to BMI: the ®rst group, cimetidine (400 mg): Statistical analysis was performed in those patients (Nycomed, Oslo, Norway) and the second group who ful®lled the study according to the protocol. placebo (identical in taste and appearance as cimeti- Differences in consecutive weight changes between dine) (Pharmaceutical Department, University Hospi- the cimetidine and placebo group were evaluated tal of Tromsù, Tromsù, Norway), prescribed three using repeated measures multivariate analysis of var- times daily, 30 min before meals. Patients were told iance (RM MANOVA). Loss in appetite was de®ned to follow their own diet or eat less if not hungry, to as a decrease of > 10% in the mean visual analogue perform their regular exercise program and not change scale (VAS) during treatment of an individual patient. their smoking habits during the study. The cimetidine Signi®cance between the treatment groups was tested dose was twice as high as that used in the former by comparing the fraction of patients with loss in 12 study as it seemed to induce a stronger feeling of appetite in each group by using Fisher-Irwin exact satiety (unpublished observation). This dose also test. In the remaining parameters, the signi®cance was 14 increases the postprandial CCK response. tested for within-group comparison by paired Stu- At the start, a physical examination was performed, dent's t test, and for between-group comparison by including sitting blood pressure, height, weight, unpaired t tests. The signi®cance level was set at abdominal circumference (umbilical level), and hip P < 0.05. circumference (tuberculum majus). BMI was calcu- lated as kg=m2. Body fat % was measured with near infrared technique (Futrex 5000: Futrex Corp., Gaithersburg, MA) as previously described.15 These Results procedures were repeated at the end of the study. Duration of the study was 12 weeks, with clinic visits scheduled at the same time each week. Subjects Subjects were weighed each week, side effects were recorded Of the 62 subjects (31 in each group) randomized to and the average degree of hunger before breakfast, active and placebo treatment, 60 subjects (31=29, in lunch and supper during the previous week was the cimetidine and placebo group, respectively) recorded on a 10 cm visual analogue scale. Compli- entered the study, that is, two patients were lost at ance with the drug treatment was checked by the inclusion due to reasons unknown (Table 1). Five of returned dose packets; 80% compliance was consid- the patients were not included due to sudden events ered acceptable. between the randomization and the start. Twelve subjects were excluded, four subjects due to side effects (one placebo subject with diarrhoea and three cimetidine subjects, one each with arthralgia, abdom- Laboratory tests inal pain and vomiting). After termination of the Blood samples were drawn after an overnight fast at treatment all subjects recovered completely. In the the start and after 12 weeks. The blood samples were remaining patients the medication was well tolerated. withdrawn into vacutainer tubes containing EDTA The remaining 43 subjects (19=24), 14 women and 29 (1 mg=ml, ®nal concentration), and centrifuged at men, completed the study according to the protocol. 1000 g for 10 min at 4C. Blood glucose was mea- sured by a glucose analyzer (Model 2300 STAT, Yellow Springs Instrument Co., Inc., Yellow Springs, Table 1 Patients included and excluded OH). The rest of the plasma was immediately frozen Cimetidine Placebo Total  (727 C) and stored for later measurements. HbA1c was measured by the Diamat TM system from Bio-Rad Randomized 31 31 62 Failed to initiate the study 0 2 2 Laboratories GmbH (Munich, Germany). Insulin was Lack of ful®lling the inclusion criteria 4 1 5 measured by a radioimmunoassay (RIA) method.16 Initiated the study 27 28 55 Plasma concentrations of cholesterol and triglycerides Excluded due to Side effects 3 1 4 were measured using Hitachi 917 Automatic Analyzer Withdrawals not related to 3 1 4 from Boehringer Mannheim (Mannheim, Germany). treatment Reagents, calibrators and controls were purchased Changes in smoking habits 1 1 Immobilization (bone fracture) 1 1 from the manufacturer. High-density lipoprotein Use of antacids 1 1 (HDL) cholesterol concentrations in plasma were Start of peroral antidiabetic agents 1 1 measured after preparation of HDL according to the Total of excluded patients 8 4 12 Total of evaluable patients 19 24 43 method described by Burstein et al.17 Cimetidine reduces weight G Stùa-Birketvedt et al 1043 Characteristics of patients plasma concentrations of insulin and insulin=glucose Characteristics of the patients are shown in Table 2. ratio (Table 4). In the cimetidine group there were At the start of the study there were no signi®cant signi®cant reductions in the levels of HbA1c, insulin differences between age, gender, duration of Type 2 and insulin=glucose ratio, but no signi®cant changes diabetes, patients on peroral antidiabetic medication, in the placebo group were observed. body weight, BMI, fat %, waist and hip circumfer- There were no signi®cant differences in baseline ences or waist=hip ratio (WHR). plasma levels of cholesterol, triglycerides and HDL During the ®rst week of treatment, appetite mea- cholesterol between the two groups (Table 4). During sured by VAS was signi®cantly reduced in the cime- treatment there were signi®cant reductions in the tidine group and remained so throughout the study. No plasma levels of triglycerides and signi®cant increases changes in appetite were observed in the placebo in the plasma levels of HDL cholesterol in the group (Figure 1). cimetidine group, but no signi®cant changes in these There was a signi®cant weight loss in the cimeti- variables in the placebo group were observed. Sig- dine group (5.0Æ 2.1 kg (meanÆ s.d.)) and in the ni®cant changes in the plasma concentrations of placebo group (1.3Æ 1.1 kg) (Table 3). The weight cholesterol were not seen in any group. loss was signi®cantly greater in the cimetidine group than in the placebo group (Figure 2). BMI, fat %, waist circumference and WHR were reduced in the cimetidine group. In the placebo group, these vari- ables were only signi®cantly changed for BMI (Table 3). The weight loss consisted of 89% fat in the cimetidine group and 85% fat in the placebo group. There was a signi®cant reduction in the systolic (6.9Æ 11.4 mm Hg) and diastolic (6.0Æ 6.6 mm Hg) blood pressure in the cimetidine group only (Table 3).

Biochemical parameters There were no signi®cant differences between groups in baseline values of fasting blood glucose and HbA1c,

Table 2 Characteristics of patients ful®lling the inclusion and exclusion criteria

Cimetidine Placebo

Female=male ratio 6=13 8=16 Age (y) 46.3Æ 8.0 49.1Æ 7.2 Duration of the disease (y) 2.5Æ 1.8 4.0Æ 3.5 Patients on peroral antidiabetics 7 14 Body weight 103.9Æ 13.8 102.0Æ 18.3 BMI (kg=m2) 33.8Æ 3.4 34.0Æ 4.2 Fat % 29.9Æ 6.6 30.6Æ 7.2 Waist circumference (cm) 111.5Æ 10.3 112.7Æ 12.0 Figure 1 Appetite registrations (recorded on a 10 cm visual Hip circumference (cm) 116.2Æ 8.1 116.3Æ 11.9 analogue scale (VAS)) during a 12 week treatment with cimeti- Waist=hip ratio 0.96Æ 0.08 0.98Æ 0.08 dine (n ˆ 19) and placebo (nˆ 24) in overweight patients with Type 2 diabetes. Statistical analysis by Fisher-Irvine exact test. MeanÆ s.d. or actual number. The changes from initial values were signi®cant (P < 0.05) in the BMI ˆ body mass index. cimetidine group. For further details, see text.

Table 3 Characteristics of patients

Cimetidine Placebo

Start End Start End

Weight (kg) 103.9Æ 13.8 98.9Æ 13.0* 102.0Æ 18.3 100.7Æ 18.7* BMI (kg=m2) 33.8Æ 3.4 32.2Æ 3.2* 34.0Æ 4.2 33.6Æ 4.0* Fat % 29.9Æ 6.6 25.3Æ 7.2* 30.6Æ 7.2 29.8Æ 6.7 Waist circumference (cm) 111.5Æ 10.3 107.4Æ 10.6* 112.7Æ 12.0 111.2Æ 13.0 Hip circumference (cm) 116.2Æ 8.1 114.0Æ 8.6 116.3Æ 11.9 112.5Æ 11.8 Waist=hip ratio 0.96Æ 0.08 0.94Æ 0.08* 0.98Æ 0.08 0.99Æ 0.08 Systolic BP (mm Hg) 149Æ 13 142Æ 11* 148Æ 21 141Æ 12 Diastolic BP (mm Hg) 94Æ 6 88Æ 8* 90Æ 8 87Æ 7

MeanÆ s.d. Statistical analysis by paired t test. * indicates signi®cant difference from start (P < 0.05). BMI ˆ body mass index; BP ˆ blood pressure. Cimetidine reduces weight G Stùa-Birketvedt et al 1044 Discussion explain their smaller weight loss. In this study, the weight reduction was 0.4 kg=week and consisted of 89% body fat, a rate considered optimal in preserving This study shows that cimetidine reduces body weight the lean body mass19 and to prevent an undesirable and appetite in overweight patients with Type 2 weight cycling.20 Finally, in this and in the previously diabetes. This agrees with previous reports in healthy reported study,12 the reduction of appetite and weight overweight subjects12 and in rats.13 The mechanism of was not attenuated during the observation period. this effect of the H2-receptor antagonist is not known, Therefore, cimetidine appears to be a promising although we have previously reported that cimetidine supplement to diet and exercise, in the long term increases the basal and postprandial plasma level of treatment of obesity. CCK,14 and CCK may mediate satiety signals from The cimetidine-induced weight reduction was asso- the gut to the brain.18 ciated with improved glycemic control, that is, reduc- The cimetidine-induced weight reduction during tion of fasting concentrations of blood glucose and the 12 week period was 5.1 kg, which represents a HbA1c, plasma insulin and the insulin=glucose ratio. 5% reduction in body weight. In the previous study of This improvement was obtained with only a 5% healthy overweight subjects,12 the weight reduction reduction in body weight and without achieving the was 9.5 kg, which represents a 12% weight reduction ideal body weight. This is in agreement with other during an eight week combined treatment of cimeti- studies where only a modest weight reduction dine and diet. Patients in the present study were improved glycemic control and insulin sensitivity.8,21 already restricted to a diabetic diet, which may Moreover, the plasma concentrations of triglycerides and HDL cholesterol were both elevated (out of the normal range) at the start of the study. Obesity and Type 2 diabetes are both part of the metabolic syndrome where hypertriglyceridemia, low HDL cho- lesterol and increased free fatty acids are the char- acteristic metabolic pro®le.3 In this study, the plasma concentrations of triglyceride and HDL cholesterol changed towards normalcy. This agrees with other reports where a modest weight reduction also improved the blood lipid pro®le.8 The plasma con- centration of cholesterol, however, was in the normal range at the start of the study and no signi®cant change in this variable was observed during the weight reduction.

Conclusion

During the study, there were reductions in the waist circumference and the WHR suggesting a reduction in visceral fat22 and in blood pressure. Altogether, our Figure 2 Weight loss during a 12 week treatment with active study shows that cimetidine reduces appetite and body (n ˆ 19) and placebo (n ˆ 24) in overweight patients with Type 2 weight, and improves metabolic control in overweight diabetes. Statistical analysis by repeated measures multivariate analysis of variance. * Indicates signi®cant difference between patients with Type 2 diabetes. This should indeed the two groups (P < 0.05). For further details, see text. reduce the risk of cardiovascular disease and contri-

Table 4 Responses to biochemical parameters

Cimetidine Placebo

Start End Start End

Blood glucose (mmol=l) 9.0Æ 2.9 7.7Æ 1.6* 9.5Æ 2.9 9.0Æ 2.4 HbA1c (%) 7.2Æ 1.2 6.7Æ 1.0* 7.6Æ 1.5 7.3Æ 1.3 Insuln (mU=ml) 14.4Æ 9.1 5.2Æ 3.4* 12.7Æ 9.8 16.0Æ 8.7 Insuliin=glucose ratio 1.7Æ 1.0 0.7Æ 0.4* 1.3Æ 1.1 2.2Æ 2.4 Plasma cholesterol (mmol=l) 6.1Æ 0.9 6.0Æ 1.0 6.2Æ 1.1 5.9Æ 1.0 Plasma triglycerides (mmol=l) 2.8Æ 1.4 2.3Æ 1.2* 2.9Æ 2.4 2.9Æ 2.6 Plasma HDL cholesterol 0.84Æ 0.3 0.94Æ 0.2* 0.94Æ 0.3 0.88Æ 0.2 (mmol=l)

MeanÆ s.d. Statistical analysis by paired t test. * indicates signi®cant difference from start (P < 0.05). HDL ˆ high-density lipoprotein. Cimetidine reduces weight G Stùa-Birketvedt et al 1045 bute to positive health bene®ts. However, it remains to 7 Ludvik B, Nolan JJ, Baloga J, Sacks D, Olefsky J. Effect of obesity on insulin resistance in normal subjects and patients be proved that the H2-receptor antagonist has long term effects beyond three months. with NIDDM. Diabetes 1995; 44: 1121±1125. 8 8 Goldstein DJ. Bene®cial health effects of modest weight loss. As reported by others, modest weight reduction Int J Obes 1992; 16: 397±415. caused substantial improvements in diabetic control, 9 Tremblay A, DespreÂs J-P, Maheux J, Pouliot MC, Nadeau A, in spite of the fact that most of the patients remained Moorjani S, Lupien PJ, Bouchard C. Normalization of the obese. Fasting plasma concentrations of insulin and metabolic pro®le in obese women by exercise and a low fat diet. Med Sci Sport Exercise 1991; 23: 1326±1331. plasma insulin=blood glucose ratio were normalized 10 Henry RR, Gumbiner B. Bene®ts and limitations of very-low- in almost all the patients in the cimetidine group. The calorie diet therapy in obese NIDDM. Diabetes Care 1991; 14: insulin resistance appears at a BMI > 32 kg=m2.23 In 802±823. the cimetidine group, the improvement in the insulin 11 DespreÂs J-P, Lemieux S, Lamarche B, Prud'homme D, Moor- sensitivity, measured indirectly as the plasma insulin= jani S, Brun L-D, Gagne C, Lupien PJ. The insulin resistance- dyslipidemic syndrome: contribution of visceral obesity and blood glucose ratio, is probably at least partly due to therapeutic implications. Int J Obes 1995; 19 (Suppl 1): weight reduction. S76±S86. In conclusion, cimetidine induces reductions in 12 Stùa-Birketvedt G. Effect of cimetidine suspension on appetite appetite and body weight and improves diabetic con- and weight in overweight subjects. BMJ 1993; 306: trol, insulin sensitivity and blood lipids in overweight 1091±1093. 13 Stùa-Birketvedt G, Lùvhaug N, Vonen B, Florholmen J. subjects with Type 2 diabetes. These effects are A H-2-receptor antagonist reduces food intake and weight probably a result of weight reduction. gain in rats via non-gastric secretory mechanisms. Acta Phys Scand 1997; 161: 489±494. 14 Stùa-Birketvedt G, Vonen B, Waldum HL, Florholmen J. The Acknowledgements effect of cimetidine on plasma levels of cholecystokinin and We appreciate the superb technical assistance of Tone gastrin in healthy volunteers. Acta Phys Scand 1997; 159: IndergaÊrd, Astrid Augdal, Ingrid Christiansen and The 321±325. Clinical Research Laboratorium, University Hospital 15 Stùa-Birketvedt G, Thom E, Aarbakke J, Florholmen J. Body fat as a predictor of the antihypertensive effect of nifedipine. of Tromsù. The present work was supported by The J Int Med 1995; 237: 169±173. Norwegian Research Council for Science and the 16 Vonen B, Florholmen J, Malm D, Torjessen P, Burhol PG. Humanities, Norsk RaÊd til Bekjempelse av Fordùyel- Glucagon mediates arginine-induced somatostatin secretion sessykdommer, The Helga Sembs Fund, The Novo from isolated rat pancreatic islets. Scand J Clin Lab Invest Nordic Foundation and The Norwegian Diabetes 1992; 52: 107±112. 17 Burstein M, Scholnick HR, Mor®n R. Rapid method for Association. the isolation of lipoproteins from human serum by precipitation with polyanions. J Lipid Res 1970; 11: 583± 595. References 18 Gibbs J, Young RC, Smith GP. Cholecystokinin decreases 1 Knowler WC, Pettitt DJ, Savage PJ, Bennett PH. Diabetes food intake in rats. J Comp Physiol Psychol 1973; 84: incidence in Pima Indians: contributions of obesity and 488±493. parental diabetes. Am J Epidemiol 1981; 113: 144±156. 19 Council of scienti®c affairs. Treatment of obesity in adults. 2 Ohlson LO, Larsson B, SvaÈrdsudd K, Welin L, Eriksson H, JAMA 1988; 260: 2547±2551. Wilhelmsen L, BjoÈrntorp P, Tibblin G. The in¯uence of body 20 Muls E, Kempen K, Vansant G, Saris W. Is weight cycling fat distribution on the incidence of diabetes mellitus: 13.5 detrimental to health? A review of the literature in humans. Int years of follow-up of the participants in the study of men born J Obes 1995; 19 (Suppl 3): S46±S50. in 1913. Diabetes 1985; 34: 1055±1058. 21 Stanik S, Marcus R. Insulin secretion improves following 3 DespreÂs J-P, Marette A. Relation of components of insulin dietary control of plasma glucose in severely hyperglycemic resistance syndrome to coronary disease risk. Curr Opin obese patients. Metabolism 1980; 29: 346±350. Lipidol 1994; 5: 274±289. 22 DespreÂs J-P, Prud'homme D, Pouliot M-C, Tremblay A, 4 BjoÈrntorp P. Abdominal obesity and the development of non- Bouchard C. Estimation of deep abdominal adipose-tissue insulin dependent diabetes mellitus. Diabet Met Rev 1988; 4: accumulation from simple anthropometric measurements in 615±622. men. Am J Clin Nutr 1991; 54: 471±477. 5 Park KS, Rhee BD, Lee K-U, Kim SY, Lee HK, Koh C-S, 23 Campbell PJ, Gerich JE. Impact of obesity on insulin action in Min HK. Intra-abdominal fat is associated with decreased volunteers with normal glucose tolerance: demonstration of a insulin sensitivity in healthy young men. Metabolism 1991; threshold for the adverse effect of obesity. J Clin Endocrinol 40: 600±603. Metab 1990; 70: 1114±1118. 6 Perriello G, Misericordia P, Volpi E, Pampanelli S, Santeusa- nio F, Brunetti P, Bolli GB. Contribution of obesity to insulin resistance in noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab 1995; 80: 2464±2469.