- Email: [email protected]
Metformin decreases plasma resistin concentrations in pediatric patients with impaired glucose tolerance: a placebo-controlled randomized clinical trial
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
Available online at www.sciencedirect.com
Metabolism www.metabolismjournal.com
Metformin decreases plasma resistin concentrations in pediatric patients with impaired glucose tolerance: a placebo-controlled randomized clinical trial Rita A. Gómez-Díaz a , Juan O. Talavera a , Elsy Canché Pool b , Francisco Vianney Ortiz-Navarrete b , Fortino Solórzano-Santos c , Rafael Mondragón-González a , Adan Valladares-Salgado d , Miguel Cruz d , Carlos A. Aguilar-Salinas e,⁎, Niels H. Wacher a a
Unidad de Investigación Médica en Epidemiología Clínica, UMAE Hospital de Especialidades, Centro Médico Nacional Siglo XXI, Instituto Mexicano del Seguro Social (IMSS), Mexico Ciy, Mexico b Departamento de Biomedicina Molecular, Centro de Investigación y de Estudios Avanzados CINVESTAV, Mexico City c UMAE Hospital de Pediatría, Centro Médico Nacional Siglo XXI, IMSS, Mexico City, Mexico d Unidad de Investigación Médica en Bioquímica, UMAE Hospital de Especialidades, Centro Médico Nacional Siglo XXI, IMSS, Mexico City, Mexico e Departmento de Endocrinología y Metabolismo, Instituto Nacional de Ciencias Médicas y Nutrición. México City, Mexico
A R T I C LE I N FO
AB S T R A C T
Article history:
The objective was to determine the effect of metformin on the concentrations of resistin and
Received 14 October 2011
other markers of insulin resistance or inflammation (C-reactive protein, cytokines, body
Accepted 6 February 2012
weight, HbA1c, among others) in minors with glucose intolerance. Patients aged 4 to 17 years with glucose intolerance were studied. They were randomized to receive 850 mg of either metformin or placebo twice daily for 12 weeks, during which all followed an iso-caloric diet and an exercise program. High sensitivity C-reactive protein, TNF-alpha, IL-6, IL1-beta, resistin, leptin, adiponectin, glucose, insulin, HbA1c, lipid profile and transaminases were measured at the beginning and at the end of the period. Fifty-two patients were included, 11.9 ± 2.6 years old; 28 (12 males/16 females) received metformin and 24 placebo (11 males/13 females). Baseline characteristics were similar between groups (except for body mass index, which in the metformin group was slightly higher). Percentage weight loss was greater in the metformin group (−5.86% vs 2.75%, P < .05). At study end, there were statistically significant differences in resistin concentrations, even after adjusting for confounding variables (F = 7.714; P < .006). Also, metformin was associated with a significant decrease in HOMA-IR index (P = .032) and HbA1c levels (P = .001), but no change was observed in the concentration of other markers of inflammation. Metformin resulted in significant reductions of plasma resistin levels in minors with glucose intolerance. This change is independent of its effects on body weight. In contrast, metformin did not alter the concentration of inflammatory markers. © 2012 Elsevier Inc. All rights reserved.
Abbreviations: CRP, C-reactive protein; AMPK, AMP activated protein kinase; G6Pase, glucose-6-phosphatase; HOMA-β, homeostasis model assessment of β cell function; HOMA-IR, homeostasis model assessment of insulin resistance; IGT, impaired glucose tolerance; LKB1, liver kinase 1; OGTT, oral glucose tolerance test; PEPCK, phosphoenolpyruvate carboxykinase. Clinical Trial Registration: NCT01394887. ⁎ Corresponding author. Departamento de Endocrinologia y Metabolismo. Instituto Nacional de Ciencias Medicas y Nutricion. Vasco de Quiroga 15. 14000, Mexico City. Tel.: +52 55 56554523; fax: +52 55 55130002. 0026-0495/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.metabol.2012.02.003
1248
1.
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Introduction
Metformin is the first line of therapy for type 2 diabetes mellitus. In addition, randomized controlled studies demonstrate that its use in patients with impaired glucose tolerance (IGT) delays the appearance of fasting hyperglycemia [1-4]. However, areas of uncertainty exist regarding its use in children, its mechanisms of action and its safety in special groups. The American Diabetes Association supports the use of metformin as the first-option medication for children, in dosages similar to those used in adults [5]. However, evidence is lacking to support the use of metformin for treating IGT on a long-term basis. On the other hand, the mechanism of action of the drug is only partially known. Metformin reduces hepatic gluconeogenesis [6], but additional mechanisms have been postulated [7]. Its activity via the cellular fuel sensor AMPdependent kinase (AMPK) [8] has been questioned recently [9]. In regards to AMPK-mediated effects, it is not clear if such activation is dependent on a direct interaction between metformin and the enzyme or if it occurs via the mediation of other signaling molecules. Its effects on the concentrations of hormones involved in inflammation, satiety and adipose tissue functionality have been studied, but controversial results have been published [10,11]. Leptin and adiponectin interact with the AMPK system and its plasma levels are altered during metformin therapy in some, but not all, the studies [12,13]. In addition, activation of AMPK directly inhibits the expression of NFKB, a key regulatory step of the inflammatory process [14] involved in the pathogenesis of obesity-related metabolic complications. Metformin is associated to changes in plasma levels of some, but not all, inflammatory markers [15,16]. Thus, additional studies are needed to identify the main mechanism by which metformin delays the onset of type 2 diabetes mellitus. Resistin is an adipokine that belongs to a family of small, cysteine-rich proteins [17]. It is predominantly expressed in adipocytes in mice. In contrast, peripheral blood mononuclear cells, macrophages and bone marrow cells are a major source of human resistin. Therefore, in humans, resistin may be rather involved in the inflammatory processes than in the modulation of adiposity and glucose homeostasis. It has been shown to affect diabetes risk in adult women, but only when BMI is not taken into account [18]. Although resistin levels increase in cases such as insulin resistance and preeclampsia, the significance of this fact remains controversial [19,20]. It is known to induce hepatic insulin resistance and increase hepatic glucose production in mice [21,22]. Thus, it appears that the precise mechanism of resistin involvement in diabetes risk is still unknown [23]. Decreasing resistin concentrations may be a potential mechanism to improve glucose utilization, due to its association with inflammation. It has been shown that thiazolidinediones decrease resistin concentrations [24,25] in adults with impaired glucose tolerance. The effect of metformin on plasma resistin levels is controversial [26,27] due to methodological limitations of the available studies. Randomized controlled studies are needed to clarify this issue. Understanding the effect of metformin on inflammation markers and adipokines in children with impaired glucose
tolerance can provide new knowledge about the action mechanism of the drug and help clarify the inflammation/ insulin secretion relationship. Therefore, we evaluated the effect of metformin in comparison with placebo, on the concentrations of C-reactive protein (CRP), cytokines (TNF-alpha, IL-6, IL1-beta), adipokines (resistin, leptin, adiponectin), body weight, HbA1c, plasma lipid profile, glucose, insulin, and homeostasis model assessment of insulin resistance (HOMAIR). HOMA-β (for β cell function) indexes were estimated in minors with impaired glucose tolerance.
2.
Material and methods
Design: Experimental, double-blind, randomized, prospective, placebo-controlled study. Insured patients aged 4–17 years old were selected from consecutive cases at the endocrine outpatient clinic of the Hospital de Pediatria del CMN “Siglo XXI” over a 6-month period, and both patients and parents signed informed consent. The children were scheduled for a fasting oral glucose tolerance test (OGTT) with 1.75 mg/kg of oral glucose. A blood sample was taken from the antecubital vein at baseline and 120 min. Insured pediatric patients 4–17 years of age with glucose intolerance per ADA criteria were included, if free of acute or chronic inflammatory processes 3 months prior to recruitment. Exclusion Criteria: not insured under Seguro Social program, previous personal history of diabetes or demonstrating diabetes in OGTT curve, chronic renal disease or serum creatinine over 1.4 mg/dL (females) or 1.5 (males), active hepatic disease, smoking, primary dyslipidemia, heart problems, steroid use, chronic metabolic acidosis, receiving anti-hypertensive or lipid- or glucose-lowering medications. A high carbohydrate diet was indicated for 3 days prior to the test. Glucose intolerance was diagnosed by the OGTT as previously described [28]. All patients diagnosed with glucose intolerance were invited to participate. All subjects consumed a diet (individually prescribed, adapted to nutritional requirements according to age) with energy intake sufficient to maintain body weight with moderate physical activity, from the moment of inclusion throughout the study period, in order to decrease the effect of nutrition modification in the metabolic and vascular parameters. In order to prescribe the diet, an evaluation was made of the habitual eating habits of each patient, in order to understand preferences and estimate calorie consumption. Modifications were made according to the lifestyle of each participant. A 24-h retrospective record was assessed, in addition to the kind and duration of exercise performed. Recommended caloric intake was calculated using the Harris– Benedict formula. A diet was prepared using the nutritional distribution of ADA recommendations. The diet and exercise regimens were unchanged throughout the period of study. No adjustments were made by group. A complete medical history was taken, recording family background of diabetes plus blood pressure, heart rate and Tanner stage [29]. Body weight, height, and waist circumference were measured and recorded in percentile according to age and gender, and the body mass index was calculated
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
(kg/m2), and recorded in percentile according to reference values of the Center for the Disease Control, Atlanta, GA [30]. Subjects were randomized for metformin or placebo. Both metformin and placebo were placed in gelatin capsules to blind their aspect. Treatments were administered during 3 months. Children were evaluated monthly, measuring transaminases, glucose, creatinine, and uric acid. Measurements of C reactive protein, cytokines, adipokines, insulin concentrations, HbA1c levels and lipid profile were made at the beginning and end of the study. Insulin resistance was assayed by HOMA-IR index. Beta cell function was determined by using HOMA-β. These were complemented with a questionnaire on diet and physical activity, and physical examination. In addition, during each visit, participants were asked about possible adverse effects from a standardized list. Treatment assignment: Two groups were formed, randomly assigned using a table of random numbers previously designed with the Epistat statistics package (Round Rock, Texas). When the screening finished, every child received a consecutive number. This number corresponded to an envelope with a card identifying his/her medication flasks, containing 60 tablets of 850 mg metformin or placebo. The content of the capsules was not identified, so both the physician and the patient were blinded. A blinded member of the study prepared the flasks and labeled them. Participants took the capsules twice daily, at breakfast and dinner. They were instructed to take medication with meals in order to decrease the possibility of adverse effects (nausea, bloating, diarrhea), which were checked during every monthly visit. Transaminases were monitored monthly during the study. Pill count was used to monitor adherence, and a questionnaire assured diet and physical activity adherence. FPG, total cholesterol, HDL-cholesterol, triglycerides, transaminases and creatinine were measured using the Synchron CX analyzer (Beckman Systems, Fullerton CA), according to standard protocols. The coefficients of variation for cholesterol and HDL-cholesterol were 3.3% and 2.5% respectively. Plasma leptin, adiponectin and insulin concentrations were measured in duplicate by radioimmunoassay (Linco Research, St Charles, MO). HbA1c was determined in whole blood using ion exchange high-performance liquid chromatography (normal range 4–6). Plasma resistin was measured using Human Resistin Elisa Kit (PeproTech, Rock Hill, NJ, USA) read by Multiscan EX, Lab Systems, USA. Plasma IL-6, IL1-b, TNF-α concentrations were determined by ELISA using Quantikine HS Human Immunoassay Kits (R&D Systems, Minneapolis, MN, USA); plasma C-reactive protein (hs-CRP) was measured using a highly sensitive human CRP ELISA Kit (Alpha Diagnostic International, San Antonio, TX, USA) according to the manufacturer's instructions, and was read by an ELISA reader (Sunrise, Tecan USA, Durham NC,USA).
3.
Statistic analysis
Baseline variables between groups were analyzed with Student's t, “U” Mann–Whitney and X2 tests. Intra-group comparisons were performed through Friedman's and Wilcoxon's tests. Inter-group (metformin vs placebo) compari-
1249
sons were done with Mann–Whitney's “U” tests. Repeated measures ANOVA was applied for the intragroup comparison of variables in which more than two measurements existed. The “General Lineal Models” module of SPSS (V 10. SPSS, Chicago, IL) was used to compare anthropometric, metabolic variables and cytokine values adjusted for confounding variables. A P value <.05 was considered significant.
4.
Results
Of 833 children invited to participate, 533 accepted the OGTT. Of these, 79 (14.8%) presented impaired glucose tolerance and 24 (4.5%) type 2 diabetes mellitus (according to ADA criteria) [31]. Of the 79 children with glucose intolerance, 19 did not have insurance and were excluded because they could not complete their follow-up. Of the 60 potentially eligible patients, 3 were excluded for refusing consent (Fig. 1). A total of 57 patients with impaired glucose tolerance were randomly assigned to treatment. Five of them withdrew from the study: one in the experimental group (change of residence) and 4 in the placebo group (3 could not come to visits because of working parents and 1 refused the blood test) (P = .16 Fisher's exact test). Finally, a total of 52 patients were analyzed, 28 for metformin-group and 24 for placebo-group. The duration of treatment was 12 weeks. No other medication was allowed during this period. Physical activity varied from 3 to 5 times weekly (mean ± SD; 3.7 ± 0.8); each session was of light intensity during 20–60 min (mean ± SD; 35.0 ± 12.5). Adherence, defined as the percentage of compliance of total daily energy intake consumed and physical activity, was 90.5% ± 7.4% and 89.5% ± 6.5% respectively. Once assigned, inter-group differences were observed in BMI (P < .025) and their percentiles (P < .028), as shown in Table 1. Of the 52 patients, only 2 boys and 2 girls were prepubescent (Tanner 1); three in the placebo group and 1 in metformin. Table 1 also shows the baseline characteristics of the two groups, with no significant differences in metabolic variables between groups. The baseline concentrations of cytokines, adipokines and CRP did not show significant differences between groups. At the end of the treatment period, significant reduction was observed in both groups over the 12 weeks, in body weight, BMI, waist perimeter, percentage weight change. The average percentage weight change was higher in the metformin group (2.7 kg, 5.86%), than in the placebo (1.6 kg, 2.75%) (P = .0025). Since height and weight depend on age and gender, and because BMI was greater in the patients randomly assigned to the metformin group, an adjusted variance analysis was carried out for each of the variables. Table 2 shows the effect of the treatment on the cytokines, adipokines and CRP. In the placebo group, significant reduction in resistin concentrations was observed (P < .001), along with TNFα (P = .041), IL-6 (P = .001) and IL-1β (P = .001), while in the metformin group, significant reductions in resistin concentrations were observed (P = .001), along with CRP (P = .019), IL-6 (P = .001) and IL-1β (P = .001), and adiponectin concentration increased (P = .001). When comparing metformin vs placebo, significant differences were only observed in the concentration of resistin (P < .004).
1250
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Eligible patients n = 833 Not accepted n = 300 OGTT n = 533
GLUCOSE INTOLERANCE n = 79
Not insured n = 19
TYPE 2 DIABETES n = 24 Pediatric Endocrinology
Diet Counsel
PATIENTS INCLUDED n = 60
Lost insurance n=3
NORMAL n = 430
RANDOMIZED n = 57 PLACEBO n = 28
METFORMIN n = 29
Did not attend n=3
4 weeks Change of address n=1
8 weeks Did not accept sampling n=1
12 weeks
Cases Analyzed n = 28
CRP TNFα IL-6 IL-1 Adiponectin Leptin Resistin Lipids profile Glucose and Insulin Transaminases
Cases Analyzed n = 24
Fig. 1 – Flow chart of patients included in the study.
An intra-group comparison after 4, 8, and 12 weeks in the metformin group showed reductions in glucose (96 mg/dL to 94.5 mg/dL) (P < .003); insulin (22.1 mUI/L to 14.7 mUI/L) (P < .001); HOMA-IR (5.6 to 3.4) (P < .001) and HbA1c (6.1% to 5.5%) (P < .001). However, in the placebo group, a significant reduction was only observed in glucose (102 mg/dL to 99.5 mg/ dL) (P < .004). In an intergroup comparison, significant differences were observed in HOMA-IR (P = .032) and HbA1c (P < .001). A decrease was observed in the concentration of triglycerides (156.3 ± 64.3 mg/dL to 142.0 ± 69.2 mg/dL vs 168.8 ± 83.4 mg/dL to 140.8 ± 78.7 mg/dL) (P = .386); total cholesterol (163.9 ± 46.6 mg/dL to 161.9 ± 29.9 mg/dL vs 168.0 ± 29.6 mg/dL to 162.6 ± 29.8 mg/dL) (P = .962); LDL-cholesterol (90.7 ± 33.3 mg/dL to 88.7 ± 27.2 mg/dL vs 90.6 ± 25.5 mg/dL to 85.7 ± 23.3 mg/dL) (P = .622); VLDL-cholesterol (31.2 ± 12.8 mg/dL to 28.4 ± 13.8 mg/ dL vs 33.7 ± 16.9 mg/dL to 28.1 ± 15.7 mg/dL) (P = .399); non-HDL cholesterol(121.9 ± 29.6 mg/dL to 117.1 ± 46.0 mg/dL vs 124.3
± 30.1 mg/dL to 116.8 ± 27.0 mg/dL) (P = .948); and an increase in the concentration of HDL-cholesterol (42.0 ± 9.3 mg/dL to 44.8 ± 10.1 mg/dL vs 43.7 ± 8.6 mg/dL to 45.8 ± 8.51 mg/dL) (P = .954) during metformin therapy. However, the differences were not statistically significant. AST and ALT levels had significant decreases during metformin treatment. In the placebo group, transaminase levels remained unaltered (Table 3). Table 3 also reflects energy intake and physical activity during the study, which showed no differences between or within groups. A multivariate analysis was performed, adjusting for age, gender, baseline BMI and weight change percentage, for each variable, to distinguish changes attributable to the metformin treatment independent of weight loss (Table 4). After these adjustments, only HbA1c (from 6.06% ± 0.45% to 5.4% ± 0.5% metformin vs 5.9% ± 0.4% to 5.9% ± 0.3% placebo; P = .0007) and resistin (718.5 ± 224.7 to 365.5 ± 206.6 metformin vs 644.1 ± 168.3 to 541 ± 191.8 placebo; P = .0066) concentrations were
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
1251
Table 1 – Baseline characteristics of the participants. Variable
Sex (M/F) Age (years) Weight (kg) Height (meters) BMI (kg/m2) BMI percentile Energy intake (kcal/day) SBP (mmHg) DBP (mmHg) Waist perimeter (cm) OGTT 0 min OGTT 120 min Creatinine (mg/dL) AST (U/L) ALT (U/L) CPK (mg/dL) Glucose (mg/dL) Insulin (μU/mL) HOMA-IR index HOMA-β index HbA1c (%) Triglycerides (mg/dL) Total cholesterol (mg/dL) HDL-cholesterol (mg/dL) LDL-cholesterol (mg/dL) Resistin (pg/mL) Adiponectin (μg/mL) Leptin (ng/mL) TNF-α (pg/ml) CRP (ng/mL) IL-1β (pg/mL) IL-6 (pg/mL)
Metformin (n = 28)
Placebo (n = 24)
Mean ± SD (min–max)
Mean ± SD (min–Max)
12/16 11.9 ± 2.4 (7.0–16.4) 73.5 ± 22.3 (32.0–123.5) 1.52 ± 0.12 (1.1–1.7) 31.1 ± 6.3 (21.2–51.0) 97.7 ± 1.6 (94–99) 2539.52 ± 17.09 112.8 ± 11.7 (90–150) 77.6 ± 10.0 (60–95) 100.0 ± 16.5 (71–136) 99.1 ± 8.9 (83–116) 149.1 ± 11.3 (140–185) 0.6 ± 0.1 (0.4–0.9) 34.6 ± 18.0 (14–85) 70.8 ± 57.2 (33–309) 103.9 ± 46.7 (37–221) 99.0 ± 9.4 (83–116) 30.8 ± 22.2 (7.3–122.6) 7.5 ± 5.6 (1.6–31.4) 104.6 ± 80.5 (26.1–421) 6.0 ± 0.4 (4.9–7.2) 156.3 ± 64.3 (71–305) 163.9 ± 46.6 (85–218) 42.0 ± 9.3 (23–63) 90.7 ± 33.3 (30–134) 718.5 ± 224.7 (262.0–1134.7) 9.6 ± 4.4 (3.2–22.8) 24.5 ± 11.6 (9.8–57.5) 836.5 ± 791.3 (26.5–3755.8) 4383.2 ± 3766.3 (5.0–12296.7) 78.1 ± 110.3 (3.9–345.1) 114.6 ± 274.2 (9.4–1221.4)
11/13 12.0 ± 3.0 (4.4–15.91) 62.8 ± 23.7 (26.0–115.6) 1.50 ± 0.16 (1.06–1.782) 27.1 ± 5.9 (16.5–36.6) 89.8 ± 20.6 (17–99) 2485.20 ± 16.38 110.0 ± 11.8 (80–130) 76.0 ± 8.4 (60–104) 90.7 ± 17.0 (64–125) 102.1 ± 11.0 (83–125) 150.7 ± 18.3 (131–194) 0.6 ± 0.1 (0.3–0.9) 29.2 ± 19.7 (9–97) 54.2 ± 35.4 (26–178) 139.7 ± 90.5 (47–454) 103.0 ± 11.5 (83–125) 26.6 ± 17.1 (8.5–85.6) 6.5 ± 3.8 (2.2–20.0) 98.1 ± 68.3 (23.8–321) 5.9 ± .4 (4.9–6.8) 168.8 ± 83.4 (50–377) 168.0 ± 29.6 (114–220) 43.7 ± 8.6 (31–63) 90.6 ± 25.5 (34–132) 644.1 ± 168.3 (221.2–971.2) 8.7 ± 2.7 (5.3–15.4) 18.8 ± 10.5 (2.5–45.2) 561.7 ± 508.9 (18.3–2907.2) 2762.9 ± 2905.6 (5.0–8894.9) 41.7 ± 67.9 (3.9–312.6) 67.4 ± 143.2 (9.4–672.8)
“P value”
.833 .677 .100 .636 .025 ⁎ .028 ⁎ .695 .389 .536 .051 .282 .695 .910 .308 .225 .730 .167 .449 .272 .757 .479 .510 .843 .910 .532 .189 .416 .730 .150 .093 .167 .453
⁎ P < .05.
statistically different between metformin and placebo. The percentage of change for the concentrations of cytokines, adipocytokines and CRP was also analyzed after adjustment (Fig. 2). Only the change in the resistin concentration was statistically significant. The difference in resistin concentrations remained significant even if PCR, TNFα, and IL-1β and baseline adiponectin were included in the model (P = .007). The most common adverse effect presented at the beginning of metformin treatment (week 1) was diarrhea in 35.7% (n = 10), which usually disappeared after one day. This side effect disappeared in all but one case. In that case, a confounding factor existed, as the child consumed restaurant food. No other serious adverse events were observed. No patient needed to be excluded from the study due to side effects.
5.
Discussion
In a randomized, placebo-controlled study, treatment with metformin resulted in weight loss, lower HbA1c levels and decreased plasma resistin concentrations in children with impaired glucose tolerance. No modification of plasma levels of several cytokines and inflammatory markers occurred. Our
results suggest that changes in body weight and plasma resistin levels may at least partly explain the beneficial effects of metformin for the treatment and prevention of type 2 diabetes mellitus. Resistin is one of the adipokines that has been implicated in the pathogenesis of obesity-related co-morbidities. Resistin-deficient mice have improved glucose tolerance compared with wild-type controls in diet-induced obesity [32]. In addition, lower resistin levels have been associated with fat distribution in animals [33-35]. However, the role of resistin in human obesity remains controversial [36-38]. Resistin modulates liver glucose production through decreased activation of AMPK and increased expression of gluconeogenic enzymes [39]. In addition, resistin may act centrally in the hypothalamus to regulate glucose homeostasis in mice [40]. Few studies have evaluated the effect of metformin on plasma resistin levels in humans. Jung et al reported that resistin concentrations increased during metformin therapy in a small set of patients with secondary failure to sulfonylurea [41]. Our study differs from these reports by having a greater sample size and a study sample composed by children with impaired glucose tolerance. In addition, the confounding effect of weight loss was controlled using a multivariate analysis. Methodological issues may contribute
1252
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Table 2 – Changes in cytokine levels observed in metformin and placebo groups. Variable
Resistin (pg/mL) Basal Final Adiponectin (μg/mL) Basal Final Leptin (ng/mL) Basal Final CRP (ng/mL) Basal Final TNFα (pg/mL) Basal Final IL-1β (pg/mL) Basal Final IL-6 (pg/mL) Basal Final
Metformin
“P value”
Placebo
Median (min–max)
Delta
P value
Median (min–max)
Delta
P value
652.7 (262.0–1134.7) 356.2 (43.3–971.2)
−40.5
<.001
650.7 (221.2–1101.1) 510.2 (214.6–947.69)
−9.7
<.001
.004
7.8 (3.2–22.8) 14.2 (3.1–86.2)
49.7
<.001
8.9 (5.3–15.4) 9.0 (4.7–46.6)
10.2
.102
.084
20.1 (9.8–57.5) 19.8 (1.8–94.3)
−12.9
.578
19.5 (6.7–45.2) 15.1 (2.5–37.0)
−18.0
.221
.620
2649.4 (5.0–12296.7) 1756.0 (5.0–14317.6)
−52.7
.019
2540.4 (5.0–10437.4) 2810.2 (5.0–9444.2)
27.4
.532
.265
484.0 (15.6–2845.2) 750.4 (15.6–3755.8)
18.1
.059
424.1 (22.0–2651.8) 475.5 (18.3–2907.2)
9.7
.041
.313
20.5 (3.9–324.8) 2.8 (0.13–16.3)
−96.6
<.001
9.8 (3.9–345.1) 1.5 (0.13–33.1)
−94.6
<.001
.821
−80.3
<.001
9.4 (9.4–672.8) 1.8 (0.20–315.7)
–84.5
<.001
.783
9.4 (9.4–1221.4) 2.4 (0.20–621.1)
also, since resistin immunoassays recognize different resistin isoforms. The method used in this report identifies all resistin isoforms. Lower resistin levels during metformin treatment are in accordance with previous reports showing that metformin decreases resistin gene expression [42]. However, additional mechanisms may be involved, since resistin mRNA is low in human obesity. Endoplasmic reticulum stress is a
major determinant of resistin expression. The activation of the AMPK/LKB1 pathway decreases endoplasmic stress and potentially eliminates one of the main signals for abnormal resistin regulation [9,43]. Abdominal obesity has an important role in the determination of resistin concentration [44,45]. A recent study showed an important association between weight loss and reduction
Table 3 – Total energy intake and biochemical changes observed in metformin and placebo groups. Metformin Glucose (mg/dL)
Total energy intake (kcal/day) Physical activity (min/week) Insulin (μU/mL) HOMA-IR HOMA-β HbA1c (%) AST (U/L)
ALT (U/L)
Baseline Week 4 Week 8 Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Week 4 Week 8 Final Baseline Week 4 Week 8 Final
Data are shown as median (min–max). P value < .05 ⁎intragroups ⁎⁎intergroups.
98.5 93.0 94.0 94.5 2622.1 2514.7 19 20 23.8 16.0 5.8 3.8 76.2 55.3 6.1 5.5 28.5 25.0 21.5 24.0 50.5 48.5 43.0 50.5
(83–116) (78–120) (62–130) (63–124) (2299.3–2822.6) (2201.5–2788.4) (−136 to 125) (−126 to 135) (7.3–122.6) (3.5–65.4) (1.61–31.4) (0.8–16.8) (26.1– 421.0) (11.3– 247.1) (4.9–7.2) (4.3–6.3) (14–85) (13–144) (12–89) (13–95) (33–309) (33–261) (30–199) (25–178)
P value .016⁎
.773 .904 .001⁎ .001⁎ .001⁎ .001⁎ .012⁎
.002⁎
Placebo 101.0 94.5 90.0 95.5 2567.4 2488.9 21 23 21.9 21.7 5.8 5.9 74.2 75.8 5.9 5.8 22.0 22.522 .0 22.5 42.0 38.0 37.0 38.5
(83–125) (83–132) (64–136) (58–127) (2200.5–2685.6) (2227.7–2664.4) (−123 to 148) (−142 to 181) (8.5–85.6) (4.9–64.4) (2.3–20.0) (1.2–15.1) (23.8–321.0) (25.1– 266.2) (4.9–6.8) (5.5–6.9) (9–97) (10–185) (14–73) (11–100) (26–178) (26–229) (29–181) (24–181)
P value
“P value”
.004⁎
.393
.823
.475
.241
.506
.683
.075
.683
.032⁎⁎
.414
.186
.201
.001⁎⁎
.902
.344
.845
.462
1253
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
Table 4 – Effect of metformin on anthropometric, biochemical characteristics and cytokine profile adjusted for confounding variables. “P” value ⁎
Post-treatment
Sex (M/F) Age (years) % Weight change Weight (kg) Height (m) BMI (kg/m2) Waist (cm) SBP (mmHg) DBP (mmHg) Glucose (mg/dL) Insulin (μU/mL) HOMA-IR HOMA-β HbA1c (%) AST (U/L) ALT (U/L) Resistin (pg/mL) Adiponectin (μg/mL) Leptin (ng/mL) TNFα (pg/mL) CRP (ng/mL) IL-1β (pg/mL) IL-6 (pg/mL)
Metformin n = 28
Placebo n = 24
12/16 12.0 ± 2.3 −5.86 (−2.0 to −19.0) 64.4 (29.0–119.3) 1.566 (1.188–1.75) 26.8 (19.39–48.2) 90.7 (64–130) 110.0 (80–140) 80.0 (50–90) 94.5 (63–124) 16.0(3.5–65.4) 3.8 (0.8–16.8) 55.3 (11.3–247.1) 5.5 (4.3–6.3) 24.0 (13–95) 50.5 (25–178) 328.7 (43.3–971.2) 15.3 (3.1–86.2) 17.9 (1.8–94.3) 573.6 (15.6–2845.1) 2206.7 (5.0–14317.6) 3.9 (0.13–16.3) 2.4 (0.00–621.1)
11/13 12.12 ± 2.93 −2.75 (+3.0 to −15.0) 62.1 (25.5–12.0) 1.506 (1.08–1.79) 26.1(16.9–35.5) 85.0 (63–115) 105.0 (84–120) 70.0 (60–95) 95.5 (58–127) 21.7 (4.9–64.4) 5.9 (1.2–15.1) 75.8 (25.1–266.2) 5.8 (5.5–6.9) 22.5 (11–100) 38.5 (24–181) 510.2 (214.59–47.6) 9.0 (4.7–46.6) 15.1 (3.2–37.0) 395.0 (22.0–1841.6) 1927.9 (5.0–9444.2) 1.0 (0.13–33.1) 1.8 (0.2–315.7)
“P” value ⁎⁎
.833 .594 .001 ⁎ .398 .819 .331 .229 .036 ⁎ .145 .393 .075 .032 ⁎ .186 .001 ⁎ .344 .462 .004 ⁎
.8767 .6894 .5293 .6398 .6751 .5693 .3728 .1334 .9596 .2980 .1876 .2685 .0007 ⁎⁎ .6894 .8959 .0066 ⁎⁎
.084 .620 .313 .265 .821 .783
.6544 .9759 .5949 .9441 .5042 .9821
⁎ Adjusted for age, sex and BMI at baseline. ⁎⁎ Adjusted for age, sex, BMI baseline and final weight change percentage.
of resistin in obese children [46], as well as in severely obese women [47]. Weight loss is associated with a decrease in the infiltration of macrophages in adipose tissue, which interferes in gene expression, resulting in an improvement in inflammatory profile [9,48]. In the present study, the distribution of abdominal fat was evaluated by measuring waist circumference. This parameter decreased in both groups (P < .001), but
was not significant when compared between groups. (P = .229) (Table 2). This observation suggests that metformin decreases plasma resistin levels irrespective of its effects on body weight. The results of the multivariate model are in accordance with this conclusion. The concept of metformin reducing low-grade inflammation has been evaluated in patients with glucose intolerance
METFORMIN
100.0
PLACEBO
80.0
in
ct
P = NS
ne
o ip
60.0
d
e
iv
ct
ot Pr
in
ea
si
st
R C-
in
pt
Re
-6
20.0
n
ei
40.0 IL
PERCENTUAL CHANGE
A
Le
Fα
TN
β -1
IL
0.0 P = NS
-20.0
P < .001
-40.0
P = NS P = .059
-60.0 -80.0 -100.0
P = NS
P = NS
Fig. 2 – Percent change at the end of the study period.
1254
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
in Mexican adult population [49], but not in children. As in adults, metformin did not have any effect on TNF-alpha, IL-6 and IL1-beta levels in our pediatric population. Thus, our findings support the concept that metformin has a weak direct anti-inflammatory effect. The lower concentration of resistin found here suggests that metformin may induce a selective modulation of the inflammatory process. This study has some limitations. The small sample size may limit some of the conclusions to be drawn. Also, the follow-up period of 6 months does not consider the possible return to pre-study conditions among participants. Another possible limitation is that physical activity was not considered as having significant effect on resistin levels, as suggested by another recent study in adults [50]. Further long-term studies are suggested to respond to this possibility, as well as to the possible links between resistin levels and central obesity, including fat mass not reflected in typical body mass measurements. In summary, metformin decreases plasma resistin levels. This action is not caused by weight loss. Lower resistin levels may contribute to the well-known capacity of the drug to reduce hepatic glucose production. Changes in resistin levels, rather than decreased inflammatory levels, may participate in the metformin-related delay of the onset of hyperglycemia in pediatric populations. This relationship and its clinical applications merit further study.
Author contributions Rita A Gómez-Díaz wrote the manuscript, researched the data, and took subject data, measurements, contributed to discussion and reviewed/edited manuscript. Juan O Talavera did the statistical analysis, contributed to discussion and reviewed/ edited manuscript. Elsy Canché Pool, Adan Valladares-Salgado and Rafael Mondragón-González measured the cytokines and all biochemical factors, researched data and reviewed/edited manuscript. Francisco Vianney Ortiz-Navarrete researched data, contributed to discussion and reviewed/edited manuscript. Miguel Cruz-López and Fortino Solórzano-Santos and Carlos A. Aguilar-Salinas contributed to discussion and reviewed/edited the manuscript. Niels H Wacher oversaw the project, contributed to discussion and reviewed/edited the manuscript..
Funding This study was supported by the FOFOI (Fondo para el Fomento de la Investigación) IMSS-2002/165 Instituto Mexicano del Seguro Social grant.
Acknowledgment The authors wish to thank Susan Drier, an independent style corrector, for her assistance in the preparation of this manuscript.
Conflict of interest There are no conflicts of interest.
REFERENCES
[1] Knowler WC, Barrett-Connor E, Fowler SE, et al. Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393-403. [2] Diabetes Prevention Program Research Group. Effects of withdrawal from metformin on the development of diabetes in the Diabetes Prevention Program. Diabetes Care 2003;26: 977-80. [3] Kitabchi AE, Temprosa M, Knowler WC, et al. The Diabetes Prevention Program Research Group: role of insulin secretion and sensitivity in the evolution of type 2 diabetes in the Diabetes Prevention Program: effects of lifestyle intervention and metformin. Diabetes 2005;54:2404-14. [4] Goldberg RB, Temprosa M, Haffner S, et al, for the Diabetes Prevention Program Group. Effect of progression from impaired glucose tolerance to diabetes on cardiovascular risk factors and its amelioration by lifestyle and metformin intervention: the Diabetes Prevention Program randomized trial by the Diabetes Prevention Program Research Group. Diabetes Care 2009;32:726-32. [5] American Diabetes Association. Consensus statement on type 2 diabetes in children and adolescents. Diabetes Care 2000;23:381-9. [6] Stumvoll M, Nurjhan N, Perriello G, et al. Metabolic effects of metformin in non-insulin-dependent diabetes mellitus. N Engl J Med 1995;333:550-4. [7] Miller RA, Birnbaum MJ. An energetic tale of AMPK-independent effects of metformin. J Clin Invest 2010;120:2267-70. [8] Zhou G, Myers R, Li Y, et al. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest 2001;108:1167-74. [9] Foretz M, Hébrard S, Leclerc J. Metformin inhibits hepatic gluconeogenesis in mice independently of the LKB1/AMPK pathway via a decrease in hepatic energy state. J Clin Invest 2010;120:2355-69. [10] Glueck CJ, Fontaine RN, Wang P, et al. Metformin reduces weight, centripetal obesity, insulin, leptin, and low-density lipoprotein cholesterol in nondiabetic, morbidly obese subjects with body mass index greater than 30. Metabolism 2001;50:856-61. [11] Park MH, Kinra S, Ward KJ, et al. Metformin for obesity in children and adolescents: a systematic review. Diabetes Care 2009;32:1743-5. [12] Minokoshi Y, Kim YB, Peroni OD, et al. Leptin stimulates fatty-acid oxidation by activating AMP-activated protein kinase. Nature 2002;415:339-43. [13] Andersson U, Filipsson K, Abbott CR, et al. AMP-activated protein kinase plays a role in the control of food intake. J Biol Chem 2004;279:12005-8. [14] Hattori Y, Suzuki K, Hattori S, et al. Metformin inhibits cytokine-induced nuclear factor kB activation via AMPactivated protein kinase activation in vascular endothelial cells. Hypertension 2006;47:1183-8. [15] Kim HJ, Kang ES, Kim DJ, et al. Effects of rosiglitazone and metformin on inflammatory markers and adipokines: decrease in interleukin-18 is an independent factor for the improvement of homeostasis model assessment-beta in type 2 diabetes mellitus. Clin Endocrinol 2007;66:282-9. [16] Isoda K, Young JL, Zirlik A, et al. Metformin inhibits proinflammatory responses and Nuclear Factor-kB in human vascular wall cells. Arterioscler Thromb Vasc Biol 2006;26: 611-7. [17] Patel L, Buckels AC, Kinghorn IJ, et al. Resistin is expressed in human macrophages and directly regulated by PPARgamma activators. Biochem Biophys Res Commun 2003;300:472-6.
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
[18] Heidemann C, Sun Q, van Dam RM, et al. Total and high-molecular-weight adiponectin and resistin in relation to the risk for type 2 diabetes in women. Ann Intern Med 2008; 149:307-16. [19] Lee JH, Chan JL, Yiannakouris N, et al. Circulating resistin levels are not associated with obesity or insulin resistance in humans and are not regulated by fasting or leptin administration: cross-sectional and interventional studies in normal, insulin-resistant, and diabetic subjects. J Clin Endocrinol Metab 2003;88:4848-56. [20] Nanda S, Poon LC, Muhaisen M, et al. Maternal serum resistin at 11 to 13 weeks' gestation in normal anthological pregnancies. Metabolism 2011, doi:10.1016/j.metabol.2011.10.006. [21] Rajala M, Obici S, Scherer P, et al. Adipose-derived resistin and gut-derived resistin-like molecule-β selectively impair insulin action on glucose production. J Clin Invest 2003;111: 225-30. [22] Wiest R, Moleda L, Farkas S, et al. Splanchnic concentrations and postprandial release of visceral adipokines. Metabolism 2010;59:664-70. [23] Schwartz DR. Lazar MA. Human resistin: found in translation from mouse to man. Trends Endocrinol Metab 2011;22: 259-65. [24] Bajaj M, Suraamornkul S, Hardies LJ, et al. Plasma resistin concentration, hepatic fat content, and hepatic and peripheral insulin resistance in pioglitazone-treated type II diabetic patients. Int J Obes Relat Metab Disord 2004;28: 783-9. [25] Rasouli N, Yao-Borengasser A, Miles LM, et al. Increased plasma adiponectin in response to pioglitazone does not result from increased gene expression. Am J Physiol Endocrinol Metab 2006;290:E42-6. [26] Tarkun I, Dikmen E, Cetinarslan B, et al. Impact of treatment with metformin on adipokines in patients with polycystic ovary syndrome. Eur Cytokine Netw 2010;21:272-7. [27] Derosa G, Maffioli P, Salvadeo SA, et al. Effects of sitagliptin or metformin added to pioglitazone monotherapy in poorly controlled type 2 diabetes mellitus patients. Metabolism 2010;59:887-95. [28] Gomez R, Aguilar CA, Moran S, et al. Lack of agreement between the revised criteria of impaired fasting glucose and impaired glucose tolerance in children with excess body weight. Diabetes Care 2004;27:2229-33. [29] Tanner JM. Growth at adolescence. 2nd ed. Oxford, England: Blackwell Scientific; 1962. [30] National Health and Nutrition Examination Survey. CDC Growth Charts: United States. Site internet. Hyattsville: Centers For Disease Control and Prevention; 2000. http:// www.cdc.gov/growcharts. Access: 06/07/2011. [31] The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183-97. [32] Bouter B, Geary N, Langhans W, et al. Diet–genotype interactions in the early development of obesity and insulin resistance in mice with a genetic deficiency in tumor necrosis factor-alpha. Metabolism 2010;59:1065-73. [33] Varady KA, Hudak CS, Hellerstein MK. Modified alternate-day fasting and cardioprotection: relation to adipose tissue dynamics and dietary fat intake. Metabolism 2009;58:803-11. [34] Fernández CM, Moltó E, Gallardo N, et al. The expression of rat resistin isoforms is differentially regulated in visceral adipose
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
1255
tissues: effects of aging and food restriction. Metabolism 2009;58:204-11. Lee JH, Bullen JW, Stoyneva LV, et al. Circulating resistin in lean, obese, and insulin-resistant mouse models: lack of association with insulinemia and glycemia. Am J Physiol Endocrinol Metab 2005;288:E625-32. Kusminski CM, McTernan PG, Kumar S. Role of resistin in obesity, insulin resistance and type II diabetes. Clinical Science 2005;109:243-56. Hivert MF, Sullivan LM, Shrader P, et al. The association of tumor necrosis factor alpha receptor 2 and tumor necrosis factor alpha with insulin resistance and the influence of adipose tissue biomarkers in humans. Metabolism 2010;59: 540-6. Wolfe BE, Jimerson DC, Orlova C, et al. Effect of dieting on plasma leptin, soluble leptin receptor, adiponectin and resistin levels in healthy volunteers. Clin Endocrinol 2004;61: 332-8. Banerjee RR, Rangwala SM, Shapiro JS, et al. Regulation of fasted blood glucose by resistin. Science 2004;303: 1195-1198. Muse ED, Lam TKT, Scherer PE, et al. Hypothalamic resistin induces hepatic insulin resistance. J Clin Invest 2007;117: 1670-8. Jung HS, Youn BS, Cho YM, et al. The effects of rosiglitazone and metformin on the plasma concentrations of resistin in patients with type 2 diabetes mellitus. Metab Clin Exp 2005;54:314-20. Janke J, Engeli S, Gorzelniak K, et al. Resistin gene expression in human adipocytes is not related to insulin resistance. Obes Res 2002;10:1-5. Lefterova MI, Mullican SE, Tomaru T, et al. Endoplasmic reticulum stress regulates adipocyte resistin expression. Diabetes 2009;58:1879-86. McTernan PG, McTernan CL, Chetty R, et al. Increased resistin gene and protein expression in human abdominal adipose tissue. J Clin Endocrinol Metab 2002;87:2407-10. Ye ZW, Wu XM, Jiang JG. Expression changes of angiotensin II pathways and bioactive mediators during human preadipocytes-visceral differentiation. Metabolism 2009;58: 1288-96. Roth CL, Kratz M, Ralston MM, et al. Changes in adiposederived inflammatory cytokines and chemokines after successful lifestyle intervention in obese children. Metabolism 2011;60:445-52. Varady KA, Tussing L, Bhutani S, et al. Degree of weight loss required to improve adipokine concentrations and decrease fat cell size in severely obese women. Metabolism 2009;58: 1096-101. Moschen AR, Molnar C, Wolf AM, et al. Effects of weight loss induced by bariatric surgery on hepatic adipocytokine expression. Hepatol 2009;51:765-77. Caballero AE, Delgado A, Aguilar-Salinas CA, et al. The differential effects of metformin on markers of endothelial activation and inflammation in subjects with impaired glucose tolerance: a placebo-controlled, randomized clinical trial. J Clin Endocrinol Metab 2004;89:3943-8. Jorge ML, de Oliveira VN, Resende NM, et al. The effects of aerobic, resistance, and combined exercise on metabolic control, inflammatory markers, adipocytokines, and muscle insulin signaling in patients with type 2 diabetes mellitus. Metabolism 2011;60:1244-52.
Available online at www.sciencedirect.com
Metabolism www.metabolismjournal.com
Metformin decreases plasma resistin concentrations in pediatric patients with impaired glucose tolerance: a placebo-controlled randomized clinical trial Rita A. Gómez-Díaz a , Juan O. Talavera a , Elsy Canché Pool b , Francisco Vianney Ortiz-Navarrete b , Fortino Solórzano-Santos c , Rafael Mondragón-González a , Adan Valladares-Salgado d , Miguel Cruz d , Carlos A. Aguilar-Salinas e,⁎, Niels H. Wacher a a
Unidad de Investigación Médica en Epidemiología Clínica, UMAE Hospital de Especialidades, Centro Médico Nacional Siglo XXI, Instituto Mexicano del Seguro Social (IMSS), Mexico Ciy, Mexico b Departamento de Biomedicina Molecular, Centro de Investigación y de Estudios Avanzados CINVESTAV, Mexico City c UMAE Hospital de Pediatría, Centro Médico Nacional Siglo XXI, IMSS, Mexico City, Mexico d Unidad de Investigación Médica en Bioquímica, UMAE Hospital de Especialidades, Centro Médico Nacional Siglo XXI, IMSS, Mexico City, Mexico e Departmento de Endocrinología y Metabolismo, Instituto Nacional de Ciencias Médicas y Nutrición. México City, Mexico
A R T I C LE I N FO
AB S T R A C T
Article history:
The objective was to determine the effect of metformin on the concentrations of resistin and
Received 14 October 2011
other markers of insulin resistance or inflammation (C-reactive protein, cytokines, body
Accepted 6 February 2012
weight, HbA1c, among others) in minors with glucose intolerance. Patients aged 4 to 17 years with glucose intolerance were studied. They were randomized to receive 850 mg of either metformin or placebo twice daily for 12 weeks, during which all followed an iso-caloric diet and an exercise program. High sensitivity C-reactive protein, TNF-alpha, IL-6, IL1-beta, resistin, leptin, adiponectin, glucose, insulin, HbA1c, lipid profile and transaminases were measured at the beginning and at the end of the period. Fifty-two patients were included, 11.9 ± 2.6 years old; 28 (12 males/16 females) received metformin and 24 placebo (11 males/13 females). Baseline characteristics were similar between groups (except for body mass index, which in the metformin group was slightly higher). Percentage weight loss was greater in the metformin group (−5.86% vs 2.75%, P < .05). At study end, there were statistically significant differences in resistin concentrations, even after adjusting for confounding variables (F = 7.714; P < .006). Also, metformin was associated with a significant decrease in HOMA-IR index (P = .032) and HbA1c levels (P = .001), but no change was observed in the concentration of other markers of inflammation. Metformin resulted in significant reductions of plasma resistin levels in minors with glucose intolerance. This change is independent of its effects on body weight. In contrast, metformin did not alter the concentration of inflammatory markers. © 2012 Elsevier Inc. All rights reserved.
Abbreviations: CRP, C-reactive protein; AMPK, AMP activated protein kinase; G6Pase, glucose-6-phosphatase; HOMA-β, homeostasis model assessment of β cell function; HOMA-IR, homeostasis model assessment of insulin resistance; IGT, impaired glucose tolerance; LKB1, liver kinase 1; OGTT, oral glucose tolerance test; PEPCK, phosphoenolpyruvate carboxykinase. Clinical Trial Registration: NCT01394887. ⁎ Corresponding author. Departamento de Endocrinologia y Metabolismo. Instituto Nacional de Ciencias Medicas y Nutricion. Vasco de Quiroga 15. 14000, Mexico City. Tel.: +52 55 56554523; fax: +52 55 55130002. 0026-0495/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.metabol.2012.02.003
1248
1.
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Introduction
Metformin is the first line of therapy for type 2 diabetes mellitus. In addition, randomized controlled studies demonstrate that its use in patients with impaired glucose tolerance (IGT) delays the appearance of fasting hyperglycemia [1-4]. However, areas of uncertainty exist regarding its use in children, its mechanisms of action and its safety in special groups. The American Diabetes Association supports the use of metformin as the first-option medication for children, in dosages similar to those used in adults [5]. However, evidence is lacking to support the use of metformin for treating IGT on a long-term basis. On the other hand, the mechanism of action of the drug is only partially known. Metformin reduces hepatic gluconeogenesis [6], but additional mechanisms have been postulated [7]. Its activity via the cellular fuel sensor AMPdependent kinase (AMPK) [8] has been questioned recently [9]. In regards to AMPK-mediated effects, it is not clear if such activation is dependent on a direct interaction between metformin and the enzyme or if it occurs via the mediation of other signaling molecules. Its effects on the concentrations of hormones involved in inflammation, satiety and adipose tissue functionality have been studied, but controversial results have been published [10,11]. Leptin and adiponectin interact with the AMPK system and its plasma levels are altered during metformin therapy in some, but not all, the studies [12,13]. In addition, activation of AMPK directly inhibits the expression of NFKB, a key regulatory step of the inflammatory process [14] involved in the pathogenesis of obesity-related metabolic complications. Metformin is associated to changes in plasma levels of some, but not all, inflammatory markers [15,16]. Thus, additional studies are needed to identify the main mechanism by which metformin delays the onset of type 2 diabetes mellitus. Resistin is an adipokine that belongs to a family of small, cysteine-rich proteins [17]. It is predominantly expressed in adipocytes in mice. In contrast, peripheral blood mononuclear cells, macrophages and bone marrow cells are a major source of human resistin. Therefore, in humans, resistin may be rather involved in the inflammatory processes than in the modulation of adiposity and glucose homeostasis. It has been shown to affect diabetes risk in adult women, but only when BMI is not taken into account [18]. Although resistin levels increase in cases such as insulin resistance and preeclampsia, the significance of this fact remains controversial [19,20]. It is known to induce hepatic insulin resistance and increase hepatic glucose production in mice [21,22]. Thus, it appears that the precise mechanism of resistin involvement in diabetes risk is still unknown [23]. Decreasing resistin concentrations may be a potential mechanism to improve glucose utilization, due to its association with inflammation. It has been shown that thiazolidinediones decrease resistin concentrations [24,25] in adults with impaired glucose tolerance. The effect of metformin on plasma resistin levels is controversial [26,27] due to methodological limitations of the available studies. Randomized controlled studies are needed to clarify this issue. Understanding the effect of metformin on inflammation markers and adipokines in children with impaired glucose
tolerance can provide new knowledge about the action mechanism of the drug and help clarify the inflammation/ insulin secretion relationship. Therefore, we evaluated the effect of metformin in comparison with placebo, on the concentrations of C-reactive protein (CRP), cytokines (TNF-alpha, IL-6, IL1-beta), adipokines (resistin, leptin, adiponectin), body weight, HbA1c, plasma lipid profile, glucose, insulin, and homeostasis model assessment of insulin resistance (HOMAIR). HOMA-β (for β cell function) indexes were estimated in minors with impaired glucose tolerance.
2.
Material and methods
Design: Experimental, double-blind, randomized, prospective, placebo-controlled study. Insured patients aged 4–17 years old were selected from consecutive cases at the endocrine outpatient clinic of the Hospital de Pediatria del CMN “Siglo XXI” over a 6-month period, and both patients and parents signed informed consent. The children were scheduled for a fasting oral glucose tolerance test (OGTT) with 1.75 mg/kg of oral glucose. A blood sample was taken from the antecubital vein at baseline and 120 min. Insured pediatric patients 4–17 years of age with glucose intolerance per ADA criteria were included, if free of acute or chronic inflammatory processes 3 months prior to recruitment. Exclusion Criteria: not insured under Seguro Social program, previous personal history of diabetes or demonstrating diabetes in OGTT curve, chronic renal disease or serum creatinine over 1.4 mg/dL (females) or 1.5 (males), active hepatic disease, smoking, primary dyslipidemia, heart problems, steroid use, chronic metabolic acidosis, receiving anti-hypertensive or lipid- or glucose-lowering medications. A high carbohydrate diet was indicated for 3 days prior to the test. Glucose intolerance was diagnosed by the OGTT as previously described [28]. All patients diagnosed with glucose intolerance were invited to participate. All subjects consumed a diet (individually prescribed, adapted to nutritional requirements according to age) with energy intake sufficient to maintain body weight with moderate physical activity, from the moment of inclusion throughout the study period, in order to decrease the effect of nutrition modification in the metabolic and vascular parameters. In order to prescribe the diet, an evaluation was made of the habitual eating habits of each patient, in order to understand preferences and estimate calorie consumption. Modifications were made according to the lifestyle of each participant. A 24-h retrospective record was assessed, in addition to the kind and duration of exercise performed. Recommended caloric intake was calculated using the Harris– Benedict formula. A diet was prepared using the nutritional distribution of ADA recommendations. The diet and exercise regimens were unchanged throughout the period of study. No adjustments were made by group. A complete medical history was taken, recording family background of diabetes plus blood pressure, heart rate and Tanner stage [29]. Body weight, height, and waist circumference were measured and recorded in percentile according to age and gender, and the body mass index was calculated
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
(kg/m2), and recorded in percentile according to reference values of the Center for the Disease Control, Atlanta, GA [30]. Subjects were randomized for metformin or placebo. Both metformin and placebo were placed in gelatin capsules to blind their aspect. Treatments were administered during 3 months. Children were evaluated monthly, measuring transaminases, glucose, creatinine, and uric acid. Measurements of C reactive protein, cytokines, adipokines, insulin concentrations, HbA1c levels and lipid profile were made at the beginning and end of the study. Insulin resistance was assayed by HOMA-IR index. Beta cell function was determined by using HOMA-β. These were complemented with a questionnaire on diet and physical activity, and physical examination. In addition, during each visit, participants were asked about possible adverse effects from a standardized list. Treatment assignment: Two groups were formed, randomly assigned using a table of random numbers previously designed with the Epistat statistics package (Round Rock, Texas). When the screening finished, every child received a consecutive number. This number corresponded to an envelope with a card identifying his/her medication flasks, containing 60 tablets of 850 mg metformin or placebo. The content of the capsules was not identified, so both the physician and the patient were blinded. A blinded member of the study prepared the flasks and labeled them. Participants took the capsules twice daily, at breakfast and dinner. They were instructed to take medication with meals in order to decrease the possibility of adverse effects (nausea, bloating, diarrhea), which were checked during every monthly visit. Transaminases were monitored monthly during the study. Pill count was used to monitor adherence, and a questionnaire assured diet and physical activity adherence. FPG, total cholesterol, HDL-cholesterol, triglycerides, transaminases and creatinine were measured using the Synchron CX analyzer (Beckman Systems, Fullerton CA), according to standard protocols. The coefficients of variation for cholesterol and HDL-cholesterol were 3.3% and 2.5% respectively. Plasma leptin, adiponectin and insulin concentrations were measured in duplicate by radioimmunoassay (Linco Research, St Charles, MO). HbA1c was determined in whole blood using ion exchange high-performance liquid chromatography (normal range 4–6). Plasma resistin was measured using Human Resistin Elisa Kit (PeproTech, Rock Hill, NJ, USA) read by Multiscan EX, Lab Systems, USA. Plasma IL-6, IL1-b, TNF-α concentrations were determined by ELISA using Quantikine HS Human Immunoassay Kits (R&D Systems, Minneapolis, MN, USA); plasma C-reactive protein (hs-CRP) was measured using a highly sensitive human CRP ELISA Kit (Alpha Diagnostic International, San Antonio, TX, USA) according to the manufacturer's instructions, and was read by an ELISA reader (Sunrise, Tecan USA, Durham NC,USA).
3.
Statistic analysis
Baseline variables between groups were analyzed with Student's t, “U” Mann–Whitney and X2 tests. Intra-group comparisons were performed through Friedman's and Wilcoxon's tests. Inter-group (metformin vs placebo) compari-
1249
sons were done with Mann–Whitney's “U” tests. Repeated measures ANOVA was applied for the intragroup comparison of variables in which more than two measurements existed. The “General Lineal Models” module of SPSS (V 10. SPSS, Chicago, IL) was used to compare anthropometric, metabolic variables and cytokine values adjusted for confounding variables. A P value <.05 was considered significant.
4.
Results
Of 833 children invited to participate, 533 accepted the OGTT. Of these, 79 (14.8%) presented impaired glucose tolerance and 24 (4.5%) type 2 diabetes mellitus (according to ADA criteria) [31]. Of the 79 children with glucose intolerance, 19 did not have insurance and were excluded because they could not complete their follow-up. Of the 60 potentially eligible patients, 3 were excluded for refusing consent (Fig. 1). A total of 57 patients with impaired glucose tolerance were randomly assigned to treatment. Five of them withdrew from the study: one in the experimental group (change of residence) and 4 in the placebo group (3 could not come to visits because of working parents and 1 refused the blood test) (P = .16 Fisher's exact test). Finally, a total of 52 patients were analyzed, 28 for metformin-group and 24 for placebo-group. The duration of treatment was 12 weeks. No other medication was allowed during this period. Physical activity varied from 3 to 5 times weekly (mean ± SD; 3.7 ± 0.8); each session was of light intensity during 20–60 min (mean ± SD; 35.0 ± 12.5). Adherence, defined as the percentage of compliance of total daily energy intake consumed and physical activity, was 90.5% ± 7.4% and 89.5% ± 6.5% respectively. Once assigned, inter-group differences were observed in BMI (P < .025) and their percentiles (P < .028), as shown in Table 1. Of the 52 patients, only 2 boys and 2 girls were prepubescent (Tanner 1); three in the placebo group and 1 in metformin. Table 1 also shows the baseline characteristics of the two groups, with no significant differences in metabolic variables between groups. The baseline concentrations of cytokines, adipokines and CRP did not show significant differences between groups. At the end of the treatment period, significant reduction was observed in both groups over the 12 weeks, in body weight, BMI, waist perimeter, percentage weight change. The average percentage weight change was higher in the metformin group (2.7 kg, 5.86%), than in the placebo (1.6 kg, 2.75%) (P = .0025). Since height and weight depend on age and gender, and because BMI was greater in the patients randomly assigned to the metformin group, an adjusted variance analysis was carried out for each of the variables. Table 2 shows the effect of the treatment on the cytokines, adipokines and CRP. In the placebo group, significant reduction in resistin concentrations was observed (P < .001), along with TNFα (P = .041), IL-6 (P = .001) and IL-1β (P = .001), while in the metformin group, significant reductions in resistin concentrations were observed (P = .001), along with CRP (P = .019), IL-6 (P = .001) and IL-1β (P = .001), and adiponectin concentration increased (P = .001). When comparing metformin vs placebo, significant differences were only observed in the concentration of resistin (P < .004).
1250
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Eligible patients n = 833 Not accepted n = 300 OGTT n = 533
GLUCOSE INTOLERANCE n = 79
Not insured n = 19
TYPE 2 DIABETES n = 24 Pediatric Endocrinology
Diet Counsel
PATIENTS INCLUDED n = 60
Lost insurance n=3
NORMAL n = 430
RANDOMIZED n = 57 PLACEBO n = 28
METFORMIN n = 29
Did not attend n=3
4 weeks Change of address n=1
8 weeks Did not accept sampling n=1
12 weeks
Cases Analyzed n = 28
CRP TNFα IL-6 IL-1 Adiponectin Leptin Resistin Lipids profile Glucose and Insulin Transaminases
Cases Analyzed n = 24
Fig. 1 – Flow chart of patients included in the study.
An intra-group comparison after 4, 8, and 12 weeks in the metformin group showed reductions in glucose (96 mg/dL to 94.5 mg/dL) (P < .003); insulin (22.1 mUI/L to 14.7 mUI/L) (P < .001); HOMA-IR (5.6 to 3.4) (P < .001) and HbA1c (6.1% to 5.5%) (P < .001). However, in the placebo group, a significant reduction was only observed in glucose (102 mg/dL to 99.5 mg/ dL) (P < .004). In an intergroup comparison, significant differences were observed in HOMA-IR (P = .032) and HbA1c (P < .001). A decrease was observed in the concentration of triglycerides (156.3 ± 64.3 mg/dL to 142.0 ± 69.2 mg/dL vs 168.8 ± 83.4 mg/dL to 140.8 ± 78.7 mg/dL) (P = .386); total cholesterol (163.9 ± 46.6 mg/dL to 161.9 ± 29.9 mg/dL vs 168.0 ± 29.6 mg/dL to 162.6 ± 29.8 mg/dL) (P = .962); LDL-cholesterol (90.7 ± 33.3 mg/dL to 88.7 ± 27.2 mg/dL vs 90.6 ± 25.5 mg/dL to 85.7 ± 23.3 mg/dL) (P = .622); VLDL-cholesterol (31.2 ± 12.8 mg/dL to 28.4 ± 13.8 mg/ dL vs 33.7 ± 16.9 mg/dL to 28.1 ± 15.7 mg/dL) (P = .399); non-HDL cholesterol(121.9 ± 29.6 mg/dL to 117.1 ± 46.0 mg/dL vs 124.3
± 30.1 mg/dL to 116.8 ± 27.0 mg/dL) (P = .948); and an increase in the concentration of HDL-cholesterol (42.0 ± 9.3 mg/dL to 44.8 ± 10.1 mg/dL vs 43.7 ± 8.6 mg/dL to 45.8 ± 8.51 mg/dL) (P = .954) during metformin therapy. However, the differences were not statistically significant. AST and ALT levels had significant decreases during metformin treatment. In the placebo group, transaminase levels remained unaltered (Table 3). Table 3 also reflects energy intake and physical activity during the study, which showed no differences between or within groups. A multivariate analysis was performed, adjusting for age, gender, baseline BMI and weight change percentage, for each variable, to distinguish changes attributable to the metformin treatment independent of weight loss (Table 4). After these adjustments, only HbA1c (from 6.06% ± 0.45% to 5.4% ± 0.5% metformin vs 5.9% ± 0.4% to 5.9% ± 0.3% placebo; P = .0007) and resistin (718.5 ± 224.7 to 365.5 ± 206.6 metformin vs 644.1 ± 168.3 to 541 ± 191.8 placebo; P = .0066) concentrations were
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
1251
Table 1 – Baseline characteristics of the participants. Variable
Sex (M/F) Age (years) Weight (kg) Height (meters) BMI (kg/m2) BMI percentile Energy intake (kcal/day) SBP (mmHg) DBP (mmHg) Waist perimeter (cm) OGTT 0 min OGTT 120 min Creatinine (mg/dL) AST (U/L) ALT (U/L) CPK (mg/dL) Glucose (mg/dL) Insulin (μU/mL) HOMA-IR index HOMA-β index HbA1c (%) Triglycerides (mg/dL) Total cholesterol (mg/dL) HDL-cholesterol (mg/dL) LDL-cholesterol (mg/dL) Resistin (pg/mL) Adiponectin (μg/mL) Leptin (ng/mL) TNF-α (pg/ml) CRP (ng/mL) IL-1β (pg/mL) IL-6 (pg/mL)
Metformin (n = 28)
Placebo (n = 24)
Mean ± SD (min–max)
Mean ± SD (min–Max)
12/16 11.9 ± 2.4 (7.0–16.4) 73.5 ± 22.3 (32.0–123.5) 1.52 ± 0.12 (1.1–1.7) 31.1 ± 6.3 (21.2–51.0) 97.7 ± 1.6 (94–99) 2539.52 ± 17.09 112.8 ± 11.7 (90–150) 77.6 ± 10.0 (60–95) 100.0 ± 16.5 (71–136) 99.1 ± 8.9 (83–116) 149.1 ± 11.3 (140–185) 0.6 ± 0.1 (0.4–0.9) 34.6 ± 18.0 (14–85) 70.8 ± 57.2 (33–309) 103.9 ± 46.7 (37–221) 99.0 ± 9.4 (83–116) 30.8 ± 22.2 (7.3–122.6) 7.5 ± 5.6 (1.6–31.4) 104.6 ± 80.5 (26.1–421) 6.0 ± 0.4 (4.9–7.2) 156.3 ± 64.3 (71–305) 163.9 ± 46.6 (85–218) 42.0 ± 9.3 (23–63) 90.7 ± 33.3 (30–134) 718.5 ± 224.7 (262.0–1134.7) 9.6 ± 4.4 (3.2–22.8) 24.5 ± 11.6 (9.8–57.5) 836.5 ± 791.3 (26.5–3755.8) 4383.2 ± 3766.3 (5.0–12296.7) 78.1 ± 110.3 (3.9–345.1) 114.6 ± 274.2 (9.4–1221.4)
11/13 12.0 ± 3.0 (4.4–15.91) 62.8 ± 23.7 (26.0–115.6) 1.50 ± 0.16 (1.06–1.782) 27.1 ± 5.9 (16.5–36.6) 89.8 ± 20.6 (17–99) 2485.20 ± 16.38 110.0 ± 11.8 (80–130) 76.0 ± 8.4 (60–104) 90.7 ± 17.0 (64–125) 102.1 ± 11.0 (83–125) 150.7 ± 18.3 (131–194) 0.6 ± 0.1 (0.3–0.9) 29.2 ± 19.7 (9–97) 54.2 ± 35.4 (26–178) 139.7 ± 90.5 (47–454) 103.0 ± 11.5 (83–125) 26.6 ± 17.1 (8.5–85.6) 6.5 ± 3.8 (2.2–20.0) 98.1 ± 68.3 (23.8–321) 5.9 ± .4 (4.9–6.8) 168.8 ± 83.4 (50–377) 168.0 ± 29.6 (114–220) 43.7 ± 8.6 (31–63) 90.6 ± 25.5 (34–132) 644.1 ± 168.3 (221.2–971.2) 8.7 ± 2.7 (5.3–15.4) 18.8 ± 10.5 (2.5–45.2) 561.7 ± 508.9 (18.3–2907.2) 2762.9 ± 2905.6 (5.0–8894.9) 41.7 ± 67.9 (3.9–312.6) 67.4 ± 143.2 (9.4–672.8)
“P value”
.833 .677 .100 .636 .025 ⁎ .028 ⁎ .695 .389 .536 .051 .282 .695 .910 .308 .225 .730 .167 .449 .272 .757 .479 .510 .843 .910 .532 .189 .416 .730 .150 .093 .167 .453
⁎ P < .05.
statistically different between metformin and placebo. The percentage of change for the concentrations of cytokines, adipocytokines and CRP was also analyzed after adjustment (Fig. 2). Only the change in the resistin concentration was statistically significant. The difference in resistin concentrations remained significant even if PCR, TNFα, and IL-1β and baseline adiponectin were included in the model (P = .007). The most common adverse effect presented at the beginning of metformin treatment (week 1) was diarrhea in 35.7% (n = 10), which usually disappeared after one day. This side effect disappeared in all but one case. In that case, a confounding factor existed, as the child consumed restaurant food. No other serious adverse events were observed. No patient needed to be excluded from the study due to side effects.
5.
Discussion
In a randomized, placebo-controlled study, treatment with metformin resulted in weight loss, lower HbA1c levels and decreased plasma resistin concentrations in children with impaired glucose tolerance. No modification of plasma levels of several cytokines and inflammatory markers occurred. Our
results suggest that changes in body weight and plasma resistin levels may at least partly explain the beneficial effects of metformin for the treatment and prevention of type 2 diabetes mellitus. Resistin is one of the adipokines that has been implicated in the pathogenesis of obesity-related co-morbidities. Resistin-deficient mice have improved glucose tolerance compared with wild-type controls in diet-induced obesity [32]. In addition, lower resistin levels have been associated with fat distribution in animals [33-35]. However, the role of resistin in human obesity remains controversial [36-38]. Resistin modulates liver glucose production through decreased activation of AMPK and increased expression of gluconeogenic enzymes [39]. In addition, resistin may act centrally in the hypothalamus to regulate glucose homeostasis in mice [40]. Few studies have evaluated the effect of metformin on plasma resistin levels in humans. Jung et al reported that resistin concentrations increased during metformin therapy in a small set of patients with secondary failure to sulfonylurea [41]. Our study differs from these reports by having a greater sample size and a study sample composed by children with impaired glucose tolerance. In addition, the confounding effect of weight loss was controlled using a multivariate analysis. Methodological issues may contribute
1252
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
Table 2 – Changes in cytokine levels observed in metformin and placebo groups. Variable
Resistin (pg/mL) Basal Final Adiponectin (μg/mL) Basal Final Leptin (ng/mL) Basal Final CRP (ng/mL) Basal Final TNFα (pg/mL) Basal Final IL-1β (pg/mL) Basal Final IL-6 (pg/mL) Basal Final
Metformin
“P value”
Placebo
Median (min–max)
Delta
P value
Median (min–max)
Delta
P value
652.7 (262.0–1134.7) 356.2 (43.3–971.2)
−40.5
<.001
650.7 (221.2–1101.1) 510.2 (214.6–947.69)
−9.7
<.001
.004
7.8 (3.2–22.8) 14.2 (3.1–86.2)
49.7
<.001
8.9 (5.3–15.4) 9.0 (4.7–46.6)
10.2
.102
.084
20.1 (9.8–57.5) 19.8 (1.8–94.3)
−12.9
.578
19.5 (6.7–45.2) 15.1 (2.5–37.0)
−18.0
.221
.620
2649.4 (5.0–12296.7) 1756.0 (5.0–14317.6)
−52.7
.019
2540.4 (5.0–10437.4) 2810.2 (5.0–9444.2)
27.4
.532
.265
484.0 (15.6–2845.2) 750.4 (15.6–3755.8)
18.1
.059
424.1 (22.0–2651.8) 475.5 (18.3–2907.2)
9.7
.041
.313
20.5 (3.9–324.8) 2.8 (0.13–16.3)
−96.6
<.001
9.8 (3.9–345.1) 1.5 (0.13–33.1)
−94.6
<.001
.821
−80.3
<.001
9.4 (9.4–672.8) 1.8 (0.20–315.7)
–84.5
<.001
.783
9.4 (9.4–1221.4) 2.4 (0.20–621.1)
also, since resistin immunoassays recognize different resistin isoforms. The method used in this report identifies all resistin isoforms. Lower resistin levels during metformin treatment are in accordance with previous reports showing that metformin decreases resistin gene expression [42]. However, additional mechanisms may be involved, since resistin mRNA is low in human obesity. Endoplasmic reticulum stress is a
major determinant of resistin expression. The activation of the AMPK/LKB1 pathway decreases endoplasmic stress and potentially eliminates one of the main signals for abnormal resistin regulation [9,43]. Abdominal obesity has an important role in the determination of resistin concentration [44,45]. A recent study showed an important association between weight loss and reduction
Table 3 – Total energy intake and biochemical changes observed in metformin and placebo groups. Metformin Glucose (mg/dL)
Total energy intake (kcal/day) Physical activity (min/week) Insulin (μU/mL) HOMA-IR HOMA-β HbA1c (%) AST (U/L)
ALT (U/L)
Baseline Week 4 Week 8 Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Final Baseline Week 4 Week 8 Final Baseline Week 4 Week 8 Final
Data are shown as median (min–max). P value < .05 ⁎intragroups ⁎⁎intergroups.
98.5 93.0 94.0 94.5 2622.1 2514.7 19 20 23.8 16.0 5.8 3.8 76.2 55.3 6.1 5.5 28.5 25.0 21.5 24.0 50.5 48.5 43.0 50.5
(83–116) (78–120) (62–130) (63–124) (2299.3–2822.6) (2201.5–2788.4) (−136 to 125) (−126 to 135) (7.3–122.6) (3.5–65.4) (1.61–31.4) (0.8–16.8) (26.1– 421.0) (11.3– 247.1) (4.9–7.2) (4.3–6.3) (14–85) (13–144) (12–89) (13–95) (33–309) (33–261) (30–199) (25–178)
P value .016⁎
.773 .904 .001⁎ .001⁎ .001⁎ .001⁎ .012⁎
.002⁎
Placebo 101.0 94.5 90.0 95.5 2567.4 2488.9 21 23 21.9 21.7 5.8 5.9 74.2 75.8 5.9 5.8 22.0 22.522 .0 22.5 42.0 38.0 37.0 38.5
(83–125) (83–132) (64–136) (58–127) (2200.5–2685.6) (2227.7–2664.4) (−123 to 148) (−142 to 181) (8.5–85.6) (4.9–64.4) (2.3–20.0) (1.2–15.1) (23.8–321.0) (25.1– 266.2) (4.9–6.8) (5.5–6.9) (9–97) (10–185) (14–73) (11–100) (26–178) (26–229) (29–181) (24–181)
P value
“P value”
.004⁎
.393
.823
.475
.241
.506
.683
.075
.683
.032⁎⁎
.414
.186
.201
.001⁎⁎
.902
.344
.845
.462
1253
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
Table 4 – Effect of metformin on anthropometric, biochemical characteristics and cytokine profile adjusted for confounding variables. “P” value ⁎
Post-treatment
Sex (M/F) Age (years) % Weight change Weight (kg) Height (m) BMI (kg/m2) Waist (cm) SBP (mmHg) DBP (mmHg) Glucose (mg/dL) Insulin (μU/mL) HOMA-IR HOMA-β HbA1c (%) AST (U/L) ALT (U/L) Resistin (pg/mL) Adiponectin (μg/mL) Leptin (ng/mL) TNFα (pg/mL) CRP (ng/mL) IL-1β (pg/mL) IL-6 (pg/mL)
Metformin n = 28
Placebo n = 24
12/16 12.0 ± 2.3 −5.86 (−2.0 to −19.0) 64.4 (29.0–119.3) 1.566 (1.188–1.75) 26.8 (19.39–48.2) 90.7 (64–130) 110.0 (80–140) 80.0 (50–90) 94.5 (63–124) 16.0(3.5–65.4) 3.8 (0.8–16.8) 55.3 (11.3–247.1) 5.5 (4.3–6.3) 24.0 (13–95) 50.5 (25–178) 328.7 (43.3–971.2) 15.3 (3.1–86.2) 17.9 (1.8–94.3) 573.6 (15.6–2845.1) 2206.7 (5.0–14317.6) 3.9 (0.13–16.3) 2.4 (0.00–621.1)
11/13 12.12 ± 2.93 −2.75 (+3.0 to −15.0) 62.1 (25.5–12.0) 1.506 (1.08–1.79) 26.1(16.9–35.5) 85.0 (63–115) 105.0 (84–120) 70.0 (60–95) 95.5 (58–127) 21.7 (4.9–64.4) 5.9 (1.2–15.1) 75.8 (25.1–266.2) 5.8 (5.5–6.9) 22.5 (11–100) 38.5 (24–181) 510.2 (214.59–47.6) 9.0 (4.7–46.6) 15.1 (3.2–37.0) 395.0 (22.0–1841.6) 1927.9 (5.0–9444.2) 1.0 (0.13–33.1) 1.8 (0.2–315.7)
“P” value ⁎⁎
.833 .594 .001 ⁎ .398 .819 .331 .229 .036 ⁎ .145 .393 .075 .032 ⁎ .186 .001 ⁎ .344 .462 .004 ⁎
.8767 .6894 .5293 .6398 .6751 .5693 .3728 .1334 .9596 .2980 .1876 .2685 .0007 ⁎⁎ .6894 .8959 .0066 ⁎⁎
.084 .620 .313 .265 .821 .783
.6544 .9759 .5949 .9441 .5042 .9821
⁎ Adjusted for age, sex and BMI at baseline. ⁎⁎ Adjusted for age, sex, BMI baseline and final weight change percentage.
of resistin in obese children [46], as well as in severely obese women [47]. Weight loss is associated with a decrease in the infiltration of macrophages in adipose tissue, which interferes in gene expression, resulting in an improvement in inflammatory profile [9,48]. In the present study, the distribution of abdominal fat was evaluated by measuring waist circumference. This parameter decreased in both groups (P < .001), but
was not significant when compared between groups. (P = .229) (Table 2). This observation suggests that metformin decreases plasma resistin levels irrespective of its effects on body weight. The results of the multivariate model are in accordance with this conclusion. The concept of metformin reducing low-grade inflammation has been evaluated in patients with glucose intolerance
METFORMIN
100.0
PLACEBO
80.0
in
ct
P = NS
ne
o ip
60.0
d
e
iv
ct
ot Pr
in
ea
si
st
R C-
in
pt
Re
-6
20.0
n
ei
40.0 IL
PERCENTUAL CHANGE
A
Le
Fα
TN
β -1
IL
0.0 P = NS
-20.0
P < .001
-40.0
P = NS P = .059
-60.0 -80.0 -100.0
P = NS
P = NS
Fig. 2 – Percent change at the end of the study period.
1254
M ET AB O LI S M CL I NI CA L A N D EX PE R IM EN T AL 6 1 (2 0 1 2) 1 24 7–1 25 5
in Mexican adult population [49], but not in children. As in adults, metformin did not have any effect on TNF-alpha, IL-6 and IL1-beta levels in our pediatric population. Thus, our findings support the concept that metformin has a weak direct anti-inflammatory effect. The lower concentration of resistin found here suggests that metformin may induce a selective modulation of the inflammatory process. This study has some limitations. The small sample size may limit some of the conclusions to be drawn. Also, the follow-up period of 6 months does not consider the possible return to pre-study conditions among participants. Another possible limitation is that physical activity was not considered as having significant effect on resistin levels, as suggested by another recent study in adults [50]. Further long-term studies are suggested to respond to this possibility, as well as to the possible links between resistin levels and central obesity, including fat mass not reflected in typical body mass measurements. In summary, metformin decreases plasma resistin levels. This action is not caused by weight loss. Lower resistin levels may contribute to the well-known capacity of the drug to reduce hepatic glucose production. Changes in resistin levels, rather than decreased inflammatory levels, may participate in the metformin-related delay of the onset of hyperglycemia in pediatric populations. This relationship and its clinical applications merit further study.
Author contributions Rita A Gómez-Díaz wrote the manuscript, researched the data, and took subject data, measurements, contributed to discussion and reviewed/edited manuscript. Juan O Talavera did the statistical analysis, contributed to discussion and reviewed/ edited manuscript. Elsy Canché Pool, Adan Valladares-Salgado and Rafael Mondragón-González measured the cytokines and all biochemical factors, researched data and reviewed/edited manuscript. Francisco Vianney Ortiz-Navarrete researched data, contributed to discussion and reviewed/edited manuscript. Miguel Cruz-López and Fortino Solórzano-Santos and Carlos A. Aguilar-Salinas contributed to discussion and reviewed/edited the manuscript. Niels H Wacher oversaw the project, contributed to discussion and reviewed/edited the manuscript..
Funding This study was supported by the FOFOI (Fondo para el Fomento de la Investigación) IMSS-2002/165 Instituto Mexicano del Seguro Social grant.
Acknowledgment The authors wish to thank Susan Drier, an independent style corrector, for her assistance in the preparation of this manuscript.
Conflict of interest There are no conflicts of interest.
REFERENCES
[1] Knowler WC, Barrett-Connor E, Fowler SE, et al. Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393-403. [2] Diabetes Prevention Program Research Group. Effects of withdrawal from metformin on the development of diabetes in the Diabetes Prevention Program. Diabetes Care 2003;26: 977-80. [3] Kitabchi AE, Temprosa M, Knowler WC, et al. The Diabetes Prevention Program Research Group: role of insulin secretion and sensitivity in the evolution of type 2 diabetes in the Diabetes Prevention Program: effects of lifestyle intervention and metformin. Diabetes 2005;54:2404-14. [4] Goldberg RB, Temprosa M, Haffner S, et al, for the Diabetes Prevention Program Group. Effect of progression from impaired glucose tolerance to diabetes on cardiovascular risk factors and its amelioration by lifestyle and metformin intervention: the Diabetes Prevention Program randomized trial by the Diabetes Prevention Program Research Group. Diabetes Care 2009;32:726-32. [5] American Diabetes Association. Consensus statement on type 2 diabetes in children and adolescents. Diabetes Care 2000;23:381-9. [6] Stumvoll M, Nurjhan N, Perriello G, et al. Metabolic effects of metformin in non-insulin-dependent diabetes mellitus. N Engl J Med 1995;333:550-4. [7] Miller RA, Birnbaum MJ. An energetic tale of AMPK-independent effects of metformin. J Clin Invest 2010;120:2267-70. [8] Zhou G, Myers R, Li Y, et al. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest 2001;108:1167-74. [9] Foretz M, Hébrard S, Leclerc J. Metformin inhibits hepatic gluconeogenesis in mice independently of the LKB1/AMPK pathway via a decrease in hepatic energy state. J Clin Invest 2010;120:2355-69. [10] Glueck CJ, Fontaine RN, Wang P, et al. Metformin reduces weight, centripetal obesity, insulin, leptin, and low-density lipoprotein cholesterol in nondiabetic, morbidly obese subjects with body mass index greater than 30. Metabolism 2001;50:856-61. [11] Park MH, Kinra S, Ward KJ, et al. Metformin for obesity in children and adolescents: a systematic review. Diabetes Care 2009;32:1743-5. [12] Minokoshi Y, Kim YB, Peroni OD, et al. Leptin stimulates fatty-acid oxidation by activating AMP-activated protein kinase. Nature 2002;415:339-43. [13] Andersson U, Filipsson K, Abbott CR, et al. AMP-activated protein kinase plays a role in the control of food intake. J Biol Chem 2004;279:12005-8. [14] Hattori Y, Suzuki K, Hattori S, et al. Metformin inhibits cytokine-induced nuclear factor kB activation via AMPactivated protein kinase activation in vascular endothelial cells. Hypertension 2006;47:1183-8. [15] Kim HJ, Kang ES, Kim DJ, et al. Effects of rosiglitazone and metformin on inflammatory markers and adipokines: decrease in interleukin-18 is an independent factor for the improvement of homeostasis model assessment-beta in type 2 diabetes mellitus. Clin Endocrinol 2007;66:282-9. [16] Isoda K, Young JL, Zirlik A, et al. Metformin inhibits proinflammatory responses and Nuclear Factor-kB in human vascular wall cells. Arterioscler Thromb Vasc Biol 2006;26: 611-7. [17] Patel L, Buckels AC, Kinghorn IJ, et al. Resistin is expressed in human macrophages and directly regulated by PPARgamma activators. Biochem Biophys Res Commun 2003;300:472-6.
M E TAB O LI S M CL I NI CA L A N D EX P ER IM EN T AL 6 1 (2 0 1 2) 1 24 7 –1 2 55
[18] Heidemann C, Sun Q, van Dam RM, et al. Total and high-molecular-weight adiponectin and resistin in relation to the risk for type 2 diabetes in women. Ann Intern Med 2008; 149:307-16. [19] Lee JH, Chan JL, Yiannakouris N, et al. Circulating resistin levels are not associated with obesity or insulin resistance in humans and are not regulated by fasting or leptin administration: cross-sectional and interventional studies in normal, insulin-resistant, and diabetic subjects. J Clin Endocrinol Metab 2003;88:4848-56. [20] Nanda S, Poon LC, Muhaisen M, et al. Maternal serum resistin at 11 to 13 weeks' gestation in normal anthological pregnancies. Metabolism 2011, doi:10.1016/j.metabol.2011.10.006. [21] Rajala M, Obici S, Scherer P, et al. Adipose-derived resistin and gut-derived resistin-like molecule-β selectively impair insulin action on glucose production. J Clin Invest 2003;111: 225-30. [22] Wiest R, Moleda L, Farkas S, et al. Splanchnic concentrations and postprandial release of visceral adipokines. Metabolism 2010;59:664-70. [23] Schwartz DR. Lazar MA. Human resistin: found in translation from mouse to man. Trends Endocrinol Metab 2011;22: 259-65. [24] Bajaj M, Suraamornkul S, Hardies LJ, et al. Plasma resistin concentration, hepatic fat content, and hepatic and peripheral insulin resistance in pioglitazone-treated type II diabetic patients. Int J Obes Relat Metab Disord 2004;28: 783-9. [25] Rasouli N, Yao-Borengasser A, Miles LM, et al. Increased plasma adiponectin in response to pioglitazone does not result from increased gene expression. Am J Physiol Endocrinol Metab 2006;290:E42-6. [26] Tarkun I, Dikmen E, Cetinarslan B, et al. Impact of treatment with metformin on adipokines in patients with polycystic ovary syndrome. Eur Cytokine Netw 2010;21:272-7. [27] Derosa G, Maffioli P, Salvadeo SA, et al. Effects of sitagliptin or metformin added to pioglitazone monotherapy in poorly controlled type 2 diabetes mellitus patients. Metabolism 2010;59:887-95. [28] Gomez R, Aguilar CA, Moran S, et al. Lack of agreement between the revised criteria of impaired fasting glucose and impaired glucose tolerance in children with excess body weight. Diabetes Care 2004;27:2229-33. [29] Tanner JM. Growth at adolescence. 2nd ed. Oxford, England: Blackwell Scientific; 1962. [30] National Health and Nutrition Examination Survey. CDC Growth Charts: United States. Site internet. Hyattsville: Centers For Disease Control and Prevention; 2000. http:// www.cdc.gov/growcharts. Access: 06/07/2011. [31] The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183-97. [32] Bouter B, Geary N, Langhans W, et al. Diet–genotype interactions in the early development of obesity and insulin resistance in mice with a genetic deficiency in tumor necrosis factor-alpha. Metabolism 2010;59:1065-73. [33] Varady KA, Hudak CS, Hellerstein MK. Modified alternate-day fasting and cardioprotection: relation to adipose tissue dynamics and dietary fat intake. Metabolism 2009;58:803-11. [34] Fernández CM, Moltó E, Gallardo N, et al. The expression of rat resistin isoforms is differentially regulated in visceral adipose
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
1255
tissues: effects of aging and food restriction. Metabolism 2009;58:204-11. Lee JH, Bullen JW, Stoyneva LV, et al. Circulating resistin in lean, obese, and insulin-resistant mouse models: lack of association with insulinemia and glycemia. Am J Physiol Endocrinol Metab 2005;288:E625-32. Kusminski CM, McTernan PG, Kumar S. Role of resistin in obesity, insulin resistance and type II diabetes. Clinical Science 2005;109:243-56. Hivert MF, Sullivan LM, Shrader P, et al. The association of tumor necrosis factor alpha receptor 2 and tumor necrosis factor alpha with insulin resistance and the influence of adipose tissue biomarkers in humans. Metabolism 2010;59: 540-6. Wolfe BE, Jimerson DC, Orlova C, et al. Effect of dieting on plasma leptin, soluble leptin receptor, adiponectin and resistin levels in healthy volunteers. Clin Endocrinol 2004;61: 332-8. Banerjee RR, Rangwala SM, Shapiro JS, et al. Regulation of fasted blood glucose by resistin. Science 2004;303: 1195-1198. Muse ED, Lam TKT, Scherer PE, et al. Hypothalamic resistin induces hepatic insulin resistance. J Clin Invest 2007;117: 1670-8. Jung HS, Youn BS, Cho YM, et al. The effects of rosiglitazone and metformin on the plasma concentrations of resistin in patients with type 2 diabetes mellitus. Metab Clin Exp 2005;54:314-20. Janke J, Engeli S, Gorzelniak K, et al. Resistin gene expression in human adipocytes is not related to insulin resistance. Obes Res 2002;10:1-5. Lefterova MI, Mullican SE, Tomaru T, et al. Endoplasmic reticulum stress regulates adipocyte resistin expression. Diabetes 2009;58:1879-86. McTernan PG, McTernan CL, Chetty R, et al. Increased resistin gene and protein expression in human abdominal adipose tissue. J Clin Endocrinol Metab 2002;87:2407-10. Ye ZW, Wu XM, Jiang JG. Expression changes of angiotensin II pathways and bioactive mediators during human preadipocytes-visceral differentiation. Metabolism 2009;58: 1288-96. Roth CL, Kratz M, Ralston MM, et al. Changes in adiposederived inflammatory cytokines and chemokines after successful lifestyle intervention in obese children. Metabolism 2011;60:445-52. Varady KA, Tussing L, Bhutani S, et al. Degree of weight loss required to improve adipokine concentrations and decrease fat cell size in severely obese women. Metabolism 2009;58: 1096-101. Moschen AR, Molnar C, Wolf AM, et al. Effects of weight loss induced by bariatric surgery on hepatic adipocytokine expression. Hepatol 2009;51:765-77. Caballero AE, Delgado A, Aguilar-Salinas CA, et al. The differential effects of metformin on markers of endothelial activation and inflammation in subjects with impaired glucose tolerance: a placebo-controlled, randomized clinical trial. J Clin Endocrinol Metab 2004;89:3943-8. Jorge ML, de Oliveira VN, Resende NM, et al. The effects of aerobic, resistance, and combined exercise on metabolic control, inflammatory markers, adipocytokines, and muscle insulin signaling in patients with type 2 diabetes mellitus. Metabolism 2011;60:1244-52.