Journal of Internal Medicine 1997; 242: 407 412 The effect of metformin on blood pressure and metabolism in nondiabetic hypertensive patients O. SNORGAARD a, L. KØBER b & J. CARLSEN c From the a Department of Cardiology and Endocrinology, Frederiksberg Hospital, b Department of Cardiology, Gentofte University Hospital, and c MEDICON A/S, Copenhagen, Denmark Abstract. Snorgaard O, Køber L, Carlsen J (Department of Cardiology and Endocrinology, Frederiksberg Hospital, Department of Cardiology, Gentofte University Hospital, and MEDICON A/S, Copenhagen, Denmark). The effect of metformin on blood pressure and metabolism in nondiabetic hypertensive patients. J Intern Med 1997; 242: 407 412. Objectives. To study the effect of metformin on blood pressure and metabolism in nondiabetic hypertensives. Design. A six-week single-blind placebo wash-out, followed by a double-blind placebo- controlled parallel group design with skew randomization (2:2:1) to metformin 850 mg b.i.d. (n 10), metformin 500 mg b.i.d. (n 10), or placebo b.i.d. (n 5) for 12 weeks. Office blood pressure (obp), ambulatory blood pressure (abp), lipoproteins, and oral glucose tolerance (OGTT) were measured/performed before and during treatment. Subjects. Sixteen male and nine female nondiabetic (OGTT) patients (median age 57 (39 74) years) with verified hypertension (White-coat excluded) for 4 (0 20) years. Results. The possible effect of metformin treatment and dosage was tested with a two-factor analysis of variance. Treatment induced a significant decline in diastolic obp, P 0.05. This decline was, however, not significantly different comparing metformin and placebo. Systolic obp, diastolic abp, and systolic abp showed no significant change by treatment. The decline in diastolic obp was 5 mmhg in the pooled group of metformin- treated patients, P 0.005. Different gender and the presence of obesity had no impact on the decline in diastolic obp within this group. Changes in fasting C-peptide and fasting insulin during treatment were unrelated to blood pressure changes. High fasting insulin ( 60 pmol L 1 ) or high fasting C-peptide ( 1000 pmol L 1 ) at baseline did not favour an effect of metformin on diastolic obp. Glucose metabolism and lipoproteins were unchanged in all groups. Conclusions. Although metformin treatment induced a decline in diastolic office blood pressure in nondiabetic hypertensives, the decline was not different from that during placebo treatment. Metformin had no significant effect on ambulatory blood pressure. Thus, metformin has, if any, only a minor clinically insignificant effect on blood pressure in nondiabetic hypertensives. The study does not support the hypothesis that circulating insulin is a major regulator of blood pressure in hypertension. Keywords: ambulatory blood pressure, C-peptide, hypertension, insulin, metformin. Introduction Many hypertensive patients have abnormalities of glucose and lipoprotein metabolism [1]. Insulin resistance and compensatory hyperinsulinaemia is a frequent finding in most of these patients, and it has been suggested that insulin resistance plays a part in the pathogenesis of hypertension and its complications [1]. Metformin is widely accepted as an oral antihyperglycaemic agent for the treatment of noninsulin dependent diabetes mellitus (NIDDM). Metformin probably works through more than one mechanism. Increased sensitivity of skeletal muscle and the liver to insulin, improved glucose disposal, and decreased hepatic glucose production are the most prominent effects [2,3]. Metformin treatment is, in many trials, followed by reduced circulating insulin levels and has 1997 Blackwell Science Ltd 407
408 O. SNORGAARD et al. Table 1 Baseline characteristics of 25 hypertensives randomized to either metformin or placebo for 12 weeks Metformin 850 mg Metformin 500 mg Placebo b.i.d. b.i.d. b.i.d. Number 10 10 5 Males/females 6/4 6/4 4/1 Age (years) 57 (39 65) 57 (41 74) 52 (46 68) Duration of hypertension (years) 3.5 (1 15) 6 (0 13) 4 (3 20) Waist/hip ratio 0.85 (0.76 1.04) 0.93 (0.75 1.00) 0.93 (0.85 0.96) Body Mass Index 25.2 (21.3 32.8) 28.1 (22.7 33.0) 30.4 (24.3 33.9) (kg/m 2 )* Diastolic obp 105 (95 113) 107 (98 122) 106 (103 111) (mmhg)* Systolic obp 168 (141 197) 163 (141 195) 177 (158 178) (mmhg)* Diastolic abp 105 (95 122) 107 (95 126) 105 (99 114) (mmhg) Systolic abp 165 (140 182) 162 (133 203) 157 (139 179) (mmhg) a Values are median and total range. b obp, office blood pressure; abp, mean of 12-hour ambulatory blood pressure. *Average of measurements at 4 and 6 weeks of placebo wash-out. beneficial effects on lipid profiles [2,4]. Through a beneficial effect on insulin sensitivity, metformin may therefore reduce blood pressure. Studies of the effect of metformin on blood pressure in NIDDM patients are not conclusive [5]. In most studies of the effect in nondiabetic subjects [6 9], both plasma insulin and blood pressure (BP) are reduced. Especially in the pilot study of Landin et al. [6], a large reduction of BP was observed during metformin treatment. However, this study was open and uncontrolled. Only two studies were randomized, double-blind, placebocontrolled [7,9], and the results are conflicting. Metformin reduced insulin and BP in a group of obese hypertensive women [7], whereas no effect was seen in a group of lean hypertensive men [9]. To elucidate these questions, we studied the effect on blood pressure and metabolism of 12 weeks treatment with metformin in two different doses or placebo in a mixed group of hypertensives. We used a randomized, double-blind, placebo-controlled parallel-group design. Patients and methods Twenty-six male and 19 female subjects treated for hypertension by general practitioners were admitted to the Medicon Research Clinic. The current treatment was withdrawn, and placebo was given twice a day for six weeks. Office blood pressure (obp) after 5 min rest (diastolic BP was defined as Korotkoff phase V) was measured at four and six weeks. In between these visits, 12-hour daytime ambulatory blood pressure monitoring (abp) (TAKEDA medical 2420) and an oral glucose tolerance test (OGTT, 75 g) were performed. Further, blood was drawn for lipid profiles and lactate. In order to avoid serious adverse reactions to metformin, and disturbing bias on the end-points, patients with overt heart disease, cerebrovascular disease, diabetes mellitus (plasma glucose 2 hours after oral glucose 11 mmol L 1 ), reduced kidney or liver function, cholesterol-lowering treatment, or alcoholism were excluded. If mean diastolic obp was between 95 and 115 mmhg at four and six weeks, and mean diastolic abp was 90 mmhg, subjects were randomized to either metformin 850 mg b.i.d., metformin 500 mg b.i.d., or placebo b.i.d. in a skew double-blind parallel group design (2:2:1). The participating subjects were treated with metformin or placebo for 12 weeks. obp was measured after 10 and 12 weeks; abp and OGTT together with lipid profiles were performed in between these visits. Peptides in plasma during OGTT were determined by RIA kits, C-peptide: DSL-7000, (Webster,
METFORMIN AND BLOOD PRESSURE 409 Table 2 Baseline metabolic characteristics of 25 hypertensives randomized to either metformin or placebo for 12 weeks Metformin 850 mg Metformin 500 mg Placebo b.i.d. b.i.d. b.i.d. Number 10 10 5 Fasting plasma glucose (mmol/l) 6.5 (5.4 7.2) 5.7 (4.8 6.6) 5.6 (5.3 6.0) 2-hour OGTT glucose (mmol/l) 6.9 (3.1 8.8) 6.6 (3.8 8.5) 4.8 (2.7 10.2) Fasting plasma insulin (pmol/l) 45 (8 126) 54 (8 198) 36 (18 96) 2-hour OGTT insulin (pmol/l) 306 (60 1038) 207 (66 594) 198 (54 900) Fasting plasma C-peptide (pmol/l) 765 (220 1600) 1150 (510 1600) 880 (450 1100) 2-hour plasma C-peptide (pmol/l) 4250 (2100 5600) 3850 (1900 5600) 2800 (1800 6750) Total cholesterol (mmol/l) 6.2 (4.5 8.1) 6.4 (4.9 7.5) 5.7 (4.8 8.8) HDL cholesterol (mmol/l) 1.7 (1.0 2.4) 1.5 (1.0 2.0) 1.2 (1.0 1.7) Triglyceride (mmol/l) 0.95 (0.60 1.8) 1.4 (0.90 2.9) 1.9 (0.60 2.5) LDL cholesterol (mmol/l) 3.7 (2.6 5.8) 4.0 (2.8 4.7) 3.3 (2.8 6.7) a Values are median and total range. b OGTT, oral glucose tolerance test. Texas) and insulin: DSL-1600, (Webster, Texas). Lipids were determined with enzymatic methods [10 12]. Baseline values are presented as median and total range. The effect of metformin treatment and dosage was tested with a two-factor analyses of variance. The effect was expressed in the text and in tables as the decline by treatment, defined as values during placebo wash-out subtracted values during treatment. They are given in Table 3 as mean decline and 95% confidence interval for the mean. The possible effect of gender and insulin level in subgroups was compared with unpaired t-tests. Five per cent was chosen as level of significance. The study was approved by the health authorities and the local ethics committee. Results Out of the 45 admitted subjects, three showed poor compliance during the wash-out period, three patients were diabetic, two developed unrelated diseases, two had diastolic obp above 115 mmhg, four had mild but unacceptable symptoms due to the mildly elevated BP, and, finally, six had white-coat hypertension (diastolic obp 95 115 mmhg and diastolic abp 90 mmhg). This left 16 males and nine females in the study with overall characteristics not significantly different from all 45 admitted patients. Baseline characteristics The characteristics of the participating subjects are indicated in Tables 1 and 2. No significant differences were found between groups. All were proven to be true hypertensives by mean of the 12-hour ambulatory blood pressure profile (Table 1). On average, subjects were mildly obese with unfavourable waist/ hip ratio (Table 1) and mildly elevated total and LDL cholesterol (Table 2). Six patients had impaired glucose tolerance (2-hour plasma glucose 7.8 11.0 mmol L 1 ). Out of these, two placebo- treated and two metformin-treated patients were normal at the end of the trial. One was diabetic and one still had impaired glucose tolerance, whereas four other metformin- treated were either diabetic (n 1) or had impaired glucose tolerance after the 12 weeks of treatment. Subjects were relatively hyperinsulinaemic, 10 subjects had fasting plasma C-peptide
410 O. SNORGAARD et al. Table 3 Decline in blood pressure and body mass index during 12 weeks treatment with either metformin 850 mg b.i.d., metformin 500 mg b.i.d., or placebo b.i.d Metformin 850 mg Metformin 500 mg Placebo b.i.d. b.i.d. b.i.d. Number 10 10 5 Body Mass Index (kg/m 2 ) 0.4 ( 0.2 0.9) 0.4 ( 0.3 1.0) 0.1 ( 1.3 to1.1) Diastolic obp (mmhg) 3 ( 2 7)* 7 (2 13)* 3 ( 5 11)* Systolic obp (mmhg) 5 ( 4 15) 4 ( 2 10) 10 ( 12 32) Diastolic abp (mmhg) 3 ( 5 10) 6 ( 1 12) 0 ( 8 4) Systolic abp (mmhg) 2 ( 9 13) 0 ( 15 13) 13 ( 11 38) a Values are means of decline by treatment and confidence interval for mean (). b obp, office blood pressure; abp,12-hour ambulatory blood pressure. c Decline in obp and Body Mass Index are calculated as the average of measurements at 4 and 6 weeks of placebo wash-out subtracted the average of measurements at 10 and 12 weeks of treatment. d Statistics: Two-factor analysis of variance. *P 0.05 for treatment effect (values during wash-out compared to during treatment). 1000 pmol L 1 and 10 subjects had fasting insulin 60 pmol L 1 at randomization. Blood pressure A significant decline in diastolic obp was seen during treatment (P 0.05) (Table 3). Compared to the effect of placebo however, the decline was not significant. Systolic obp and abp were unchanged by treatment. No dose response relationship could be demonstrated. Diastolic obp in the pooled group of metformintreated hypertensives declined 5 (2 8) mmhg, P 0.005. Diastolic abp and systolic obp tended to decline, 4 (0 9) mmhg (P 0.06) and 5 ( 1 10) mmhg (P 0.08), respectively. Body mass index (BMI) and systolic abp were unchanged. The decline in diastolic obp was 4 (1 6) mmhg in males against 7 ( 1 16) mmhg in females (NS). High or low insulin and C-peptide (above or below 60 pmol L 1 and 1000 pmol L 1, respectively) was followed by the same declining tendency in obp during metformin. The decline was only significant (P 0.05) for obp in the group with low fasting insulin or C-peptide. Finally, the same change in diastolic obp was seen in obese (BMI 27 kg m 2 ) and nonobese. Metabolism Fasting plasma glucose, fasting C-peptide, and fasting insulin, as well as the oral glucose- stimulated values and areas under curves were unchanged during metformin treatment. Changes in C-peptide and insulin were unrelated to the reduction in diastolic obp and the other BP measurements. Finally, metformin had no significant effect on lipid profiles. Serum lactate was mildly elevated (normal range: 0.55 1.75 mmol L 1 ) in two subjects during placebo wash-out. One remained elevated, and three other subjects raised above normal range during metformin treatment (highest value 2.02 mmol L 1 ). The mean lactate was normal and unchanged during the trial. Two patients reported mild to moderate gastrointestinal side-effects during placebo, and seven during metformin. None were withdrawn. Discussion We have studied the effect of metformin on blood pressure and metabolism in a group of nondiabetic subjects with verified hypertension. Patients screened for participation in the study were diagnosed and controlled by their general practitioner for 5.8 years before admission. Based on ambulatory BP measurements, 23% of the patients were excluded due to white-coat hypertension. This was not done in the earlier studies evaluating the BP-lowering effect of metformin in nondiabetics [5 9]. Furthermore, baseline systolic and diastolic BP were higher in the subjects of this study, and they were relatively hyperinsulinaemic. Assuming that insulin is a major regulator of BP in hypertension [1] therefore makes the present population ideal for the study of a possible BP-lowering effect of metformin. The impressive BP-reducing effect of metformin found in the original uncontrolled study by Landin et
METFORMIN AND BLOOD PRESSURE 411 al. [6], has not been reproduced in a recent placebocontrolled cross-over trial by the same group [9]. In both studies the population was lean to mildly obese men with borderline hypertension. A group from Venezuela has, in an uncontrolled trial, demonstrated a decline in systolic BP in obese women with polycystic ovary syndrome [8]. However, the study population had normal BP at baseline. Thus, only one controlled study has demonstrated a significantly lower BP during metformin compared to baseline [7]. Compared to placebo, both diastolic and systolic BP declined. Twelve obese, hyperinsulinaemic, and mildly hypertensive women were studied in a doubleblind placebo-controlled cross-over design. Thus, metformin may have a BP-lowering effect related to gender, obesity, or the presence of hyperinsulinaemia. An effect related to hyperinsulinaemia is supported by a study of spontaneously hypertensive rats [13]. The present study population had higher diastolic and systolic BP at baseline compared to the study of Giugliano et al. [7]. We found only a minor (5 mmhg) but significant decline in diastolic office blood pressure within the group of metformin-treated patients. However, the decline was not significantly different from the one observed in the placebo group (3 mmhg). The limited number of subjects in the placebo group might explain this. Furthermore, diastolic abp tended to fall during treatment with metformin. Although abp is a more reliable method compared to single office BP measurements in the evaluation of a hypotensive effect [14], we may have overlooked an effect of metformin. We did not find any relationship between the BPlowering effect of metformin and elevated levels of fasting insulin and fasting C-peptide. A new antidiabetic drug, troglitazone, like metformin, reduces insulin resistance and fasting C-peptide in both diabetics [15,16] and nondiabetics [17]. A minor but inconsistent BP reduction is seen during treatment with troglitazone. This, together with the present results, does not therefore support the notion, that circulating insulin is a major regulator of BP in hypertension. As insulin sensivity was not measured, the present study gives no new information on the relationship between blood pressure and insulin resistance. Thus, we were neither able to support nor rule out the previously suggested beneficial effect of metformin on BP in insulin-resistant hypertensives [6] or obese female subjects [7]. We conclude that the hypotensive effect of metformin, if any, is only minor and of limited clinical importance in nondiabetic hypertensives. Acknowledgements Metformin and placebo tablets were kindly supplied by Lipha Pharmaceuticals Ltd, Middlesex, England. References 1 Reaven GM, Lithell H, Landsberg L. Hypertension and associated metabolic abnormalities the role of insulin resistance and the sympathoadrenale system. N Engl J Med 1996; 334: 374 81. 2 Dunn CJ, Peters DH. Metformin. A review of its pharmacological properties and therapeutic use in non-insulin-dependent diabetes mellitus. Drugs 1995; 49: 721 49. 3 Stumvoll M, Nurjahan N, Perriello G, Dailey G, Gerich JE. Metabolic effects of metformin in non-insulin-dependent diabetes mellitus. N Engl J Med 1995; 333: 550 54. 4 Carlsen SM, Rossvoll O, Bjerve KS, Følling I. Metformin improves blood lipid pattern in nondiabetic patients with coronary heart disease. J Intern Med 1996; 239: 227 33. 5 Hermann LS. Clinical pharmacology of biguanides. In: Kuhlmann J & Puls W, eds. Handbook of Experimental Phamacology, volume 119 (Oral antidiabetics). Heidelberg, Berlin, New York: Springer Verlag, 1996; 373 407. 6 Landin K, Tengborn L, Smith U. Treating insulin resistance in hypertension with metformin reduces both blood pressure and metabolic risk factors. J intern Med 1991; 229: 181 7. 7 Giugliano D, De Rosa N, Di Maro R, Marfella R, Acampora R, Buoninconti R, D Onofrio F. Metformin improves glucose, lipid metabolism, and reduces blood pressure in hypertensive, obese women. Diabetes Care 1993; 16: 1387 90. 8 Velazquez EM, Mendoza S, Hamer T, Sosa F, Glueck. Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. Metabolism 1994; 43: 647 54. 9 Landin K, Tengborn L, Smidt U. Metformin and metoprolol CR treatment in non-obese men. J Intern Med 1994; 235: 335 41. 10 Siedel J, Hagele EO, Ziegenhorn J, Wahlefeld AW. Reagent for enzymatic determination of serum total cholesterol with improved lipolytic efficiency. Clin Chem 1983; 29: 1075 80. 11 Lopes-Virella MF, Stone P, Ellis S, Colwell JA. Cholesterol determination in high-density lipoproteins separated by three different methods. Clin Chem 1977; 23: 882 4. 12 Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499 502. 13 Verma S, Bhanot S, McNeill JH. Metformin decreases plasma insulin levels and systolic blood pressure in spontaneously hypertensive rats. Am J Physiol 1994; 267: H1250 53. 14 Coats AJS, Radaelli A, Clark SJ, Conway J, Sleight P. The influence of ambulatory blood pressure monitoring on the design and interpretation of trials in hypertension. J Hypertens 1992;
412 O. SNORGAARD et al. 10: 385 91. 15 Ogihara T, Rakugi H, Ikegami H, Mikami H, Masuo K. Enhancement of insulin sensitivity by troglitazone lowers blood pressure in diabetic hypertensives. Am J Hypertens 1995; 8: 316 20. 16 Iwamoto Y, Kosaka K, Kuzuya T, Akanuma Y, Shigeta Y, Kaneko T. Effect of combination therapy of troglitazone and sulphonylureas in patients with type 2 diabetes who were poorly controlled by sulphonylurea therapy alone. Diabetic Med 1996; 13: 365 70. 17 Nolan JJ, Ludvik B, Beerdsen P, Joyce M, Olefsky J. Improvement in glucose tolerance and insulin resistance in obese subjects treated with troglitazone. N Engl J Med 1994; 331: 1188 93. Received 2 December 1996; accepted 16 June 1997. Correspondence: Ole Snorgaard MD, Department of Cardiology and Endocrinology, Frederiksberg Hospital, DK-2000 Copenhagen, Denmark (fax: 45 38347755).