HIV Articles  
Back 
 
 
Pioglitazone Compared with Metformin Increases Pericardial Fat Volume in Patients with Type 2 Diabetes Mellitus
 
 
  The Journal of Clinical Endocrinology & Metabolism Jan 2010 Vol. 95, No. 1 456-460
 
J. T. Jonker, H. J. Lamb, R. W. van der Meer, L. J. Rijzewijk, L. J. Menting, M. Diamant, J. J. Bax, A. de Roos, J. A. Romijn and J. W. A. Smit
 
Departments of Endocrinology and Metabolism (J.T.J., L.J.M., J.A.R., J.W.A.S.), Radiology (J.T.J., H.J.L., R.W.v.d.M., A.d.R.), and Cardiology (J.J.B.), Leiden University Medical Center, 2300 RC Leiden, The Netherlands; and Diabetes Center (L.J.R., M.D.), Vrije Universiteit Medical Center, 1007 MB Amsterdam, The Netherlands
 
"In conclusion, pioglitazone increases pericardial fat volume and sc (subcutaneous) abdominal fat volume in patients with T2DM, whereas metformin did not affect these fat depots. Pericardial fat is independently correlated with abdominal visceral fat volume. At baseline, there was no correlation between pericardial fat volume and myocardial function, whereas pioglitazone improved LV diastolic function despite an increase in pericardial fat volume after 24 wk.....The amount of visceral abdominal fat did not change on treatment with either pioglitazone or metformin (Table 1). In contrast, pioglitazone increased the volume of sc abdominal fat [636 ml (443-748) (baseline) vs. 724 ml (569-980), P < 0.001], whereas metformin did not affect sc fat volume [699 ml (480-883) (baseline) vs. 620 (436-871), P = 0.2, between groups, P < 0.001]."
 
Abstract
 
Context: Peroxisome proliferator-activated receptor-{gamma} agonists are involved in fat cell differentiation.
 
Objective: The objective of the study was to investigate the effect of pioglitazone vs. metformin on pericardial fat volume in type 2 diabetic (T2DM) patients. Furthermore, we aimed to assess the relationship between pericardial fat volume, other fat compartments, and myocardial function at baseline and after treatment.
 
Design: This was a prospective, randomized, double-blind, intervention study.
 
Setting: The study was conducted at a university hospital.
 
Patients: Patients included 78 men with T2DM (aged 56.5 ± 0.6 yr; glycosylated hemoglobin 7.1 ± 0.1%) without structural heart disease.
 
Intervention: Patients were randomly assigned to pioglitazone (30 mg/d) or metformin (2000 mg/d) and matching placebo during 24 wk.
 
Main Outcome Measures: Pericardial and abdominal fat volumes and myocardial left ventricular function were measured by magnetic resonance imaging and hepatic and myocardial triglyceride content by proton magnetic resonance spectroscopy.
 
Results: Pioglitazone increased pericardial fat volume [30.5 ± 1.7 ml (baseline) vs. 33.1 ± 1.8 ml], whereas metformin did not affect pericardial fat volume (29.2 ± 1.5 ml vs. 29.6 ± 1.6 ml, between groups P = 0.02). After correction for body mass index and age, only visceral fat volume correlated with pericardial fat volume at baseline (r = 0.55, P < 0.001). The increase in pericardial fat volume induced by pioglitazone was not associated with a decrease in left ventricular diastolic function.
 
Conclusion: In T2DM patients, pioglitazone increases pericardial fat volume. This increase in pericardial fat volume did not negatively affect myocardial function after 24 wk. These observations question the notion of an inverse causal relationship between pericardial fat volume and myocardial function.
 
Introduction
 
Pericardial fat is defined as the adipose tissue surrounding the myocardium, including epicardial fat and paracardial fat (1). Pericardial fat has a close connection to the coronary arteries and the myocardium, suggesting that pericardial fat plays a role in myocardial energy metabolism (2). Beneficial roles of pericardial fat have been postulated, including protection of the myocardium to an overload of nonesterified fatty acids (NEFAs), substrate supply for the myocardium, secretion of adipokines, and protection of the coronary arteries against torsion (3, 4). On the other hand, increased pericardial fat volume has been associated with cardiovascular disease in cross-sectional studies in patients at risk for coronary artery disease (1, 4, 5, 6). Therefore, the pathophysiological implications of pericardial fat have not been fully elucidated.
 
Peroxisome proliferator-activated receptor (PPAR)-{gamma} agonists are involved in fat cell differentiation. Accordingly, the beneficial role of pioglitazone is associated with redistribution of fat among different fat compartments (7). Furthermore, pioglitazone decreases macrovascular events and slows the progression of coronary artery disease in patients with type 2 diabetes mellitus (T2DM) against the adverse effects of an increased risk of heart failure and peripheral edema (8, 9).
 
Therefore, the aim of this study was to prospectively study the effect of pioglitazone vs. metformin on pericardial fat volume in patients with T2DM but without diabetes-related complications using magnetic resonance (MR) imaging.
 
Discussion
 
We found that 24 wk of treatment with pioglitazone, in contrast to metformin, increased pericardial fat volume in T2DM patients, measured by MRI. Furthermore, we found an age- and BMI-independent, positive correlation between pericardial and abdominal visceral fat volume at baseline.
 
Pioglitazone induces redistribution of fat among different fat compartments, leading to a decrease in hepatic TG content and an increase in sc (abdominal) fat (13). Most studies found no change (12, 14) or a decrease in visceral abdominal fat (13). Overall this leads to a decrease in the visceral to sc fat ratio (12, 13, 14) in accordance with our observations. This differential redistribution of fat by pioglitazone is likely related to differential PPAR{gamma} expression in organs, with the highest expression being found in adipose tissue (7). The increase in pericardial fat volume induced by pioglitazone in this study suggests that pericardial fat is susceptible to PPAR{gamma} agonist-induced adipogenesis. In accordance with this notion, epicardial and pericardial fat has a higher capacity for fatty acid synthesis in experimental models than other fat depots (3). Interestingly, we observed an inverse relationship between changes in serum NEFA levels and a pioglitazone-induced increase in pericardial fat volume. To our knowledge there are no data on PPAR{gamma} expression in pericardial fat.
 
The positive correlation between pericardial fat volume and visceral fat at baseline is a consistent finding in other studies (4, 15). Although pericardial fat is by definition visceral fat, the differential effects of pioglitazone on abdominal visceral fat and pericardial fat suggest differences in their biological roles.
 
We observed an improvement in LV diastolic function (10), which was not related to pericardial fat volume. This indicates that the increase in pericardial fat per se did not negatively affect cardiac function at 24 wk. However, we cannot exclude the possibility that longer treatment with pioglitazone would have yielded different results.
 
Indeed, studies that found an association between increased pericardial fat volume and coronary artery disease are hard to generalize to other patient groups because of the cross-sectional design of those studies. In the Framingham Heart Study, pericardial fat was actually inversely associated with measures of LV function but not independently of other measures of visceral adiposity (16). Correspondingly, there was no association between changes in pericardial fat volume and heart function in the present study, in which we assessed T2DM patients without structural heart disease.
 
Our study has some limitations. There is no reference method for quantification of pericardial fat volume, although MRI is considered a reproducible technique for fat quantification (17, 18). Accordingly in the present study, our quantification of pericardial fat volume had an excellent intra- and interobserver agreement. In this study only men were included to eliminate the confounding effect of sex steroid fluctuations. It is known that estrogen concentrations influence the expression of lipogenic genes (19), and therefore, these data may thus not be readily extrapolated to women.
 
In conclusion, pioglitazone increases pericardial fat volume and sc abdominal fat volume in patients with T2DM, whereas metformin did not affect these fat depots. Pericardial fat is independently correlated with abdominal visceral fat volume. At baseline, there was no correlation between pericardial fat volume and myocardial function, whereas pioglitazone improved LV diastolic function despite an increase in pericardial fat volume after 24 wk.
 
Results
 
Patient characteristics

 
Patients characteristics were previously described (10). In short, both study groups consisted of 39 male T2DM patients who were well matched for age (pioglitazone: 56.8 ± 1.0 yr; metformin: 56.4 ± 0.9 yr) and BMI (pioglitazone: 28.0 ± 0.5 kg/m2; metformin: 29.1 ± 0.6 kg/m2). There were no differences in the improvements of glycemic control from baseline by both drugs (Table 1).
 
Pericardial fat quantification
 
Pericardial fat was measured by two independent observers and twice per observer. Figure 1 depicts the Bland-Altman plot for these observers (mean ± 2 SD). The intraclass correlation coefficient for the quantification of pericardial fat at the four-chamber view was 0.95 (95% confidence interval 0.90-0.98); the interclass correlation coefficient was 0.93 (95% confidence interval 0.89-0.96).
 
Pericardial fat: effects of therapy
 
Pioglitazone increased pericardial fat [30.5 ± 1.7 ml (baseline) vs. 33.1 ± 1.8 ml, P = 0.003], whereas metformin did not affect pericardial fat (29.2 ± 1.5 vs. 29.6 ± 1.6 ml, P = 0.7, between groups, P = 0.02, Table 1Go).
 
In the pioglitazone group, there was an inverse correlation between the difference in pericardial fat volume and the difference in plasma NEFA levels due to treatment (r = -0.36, P = 0.04). This correlation was not significant in the metformin-treated group (r = 0.17, P = 0.31).
 
As previously described, pioglitazone but not metformin improved left ventricular (LV) diastolic function (Table 1Go) (10). The changes in these variables of diastolic function did not correlate with the change in pericardial fat volume.
 
Pericardial fat at baseline
 
At baseline in the whole group (n = 78), pericardial fat volume significantly correlated with BMI (r = 0.45, P < 0.001), hepatic triglyceride (TG) content (r = 0.32, P = 0.01), myocardial TG content (r = 0.26, P = 0.02), visceral abdominal fat (r = 0.52, P < 0.001), and sc abdominal fat (r = 0.39, P = 0.001). After correction for age and BMI, only the correlation between pericardial fat and visceral abdominal fat remained significant (r = 0.55, P < 0.001). There were no significant correlations between pericardial fat and parameters of systolic and diastolic cardiac function.
 
Abdominal fat
 
The amount of visceral abdominal fat did not change on treatment with either pioglitazone or metformin (Table 1Go). In contrast, pioglitazone increased the volume of sc abdominal fat [636 ml (443-748) (baseline) vs. 724 ml (569-980), P < 0.001], whereas metformin did not affect sc fat volume [699 ml (480-883) (baseline) vs. 620 (436-871), P = 0.2, between groups, P < 0.001].
 
Subjects and Methods
 
Subjects

 
Patients participated in a 24-wk prospective, double-blind, randomized, controlled study comparing pioglitazone with metformin. Details of this study on myocardial metabolism have been previously published (10) and will be recapitulated here in short. Males with uncomplicated T2DM were included if aged 45-65 yr, well-controlled T2DM [glycosylated hemoglobin 6.5-8.5%], body mass index (BMI) 25-32 kg/m2, and blood pressure less than 150/85 mm Hg. Exclusion criteria were any history or symptoms of cardiovascular disease, liver disease or diabetes-related complications, and prior use of thiazolidinediones or insulin. Screening included electrocardiogram and dobutamine stress echocardiography to exclude inducible ischemia or arrhythmias. The study was performed at two institutes in The Netherlands (Leiden University Medical Center, Leiden, and Vrije Universiteit Medical Center, Amsterdam) and approved by both local ethics committees. Written informed consent was obtained from all participants.
 
Patients eligible for the study ceased their regular blood glucose lowering agents and entered a 10-wk run-in period in which they were transferred to glimepiride monotherapy and titrated until a stable dose was reached 2 wk before randomization. Patients were randomized to pioglitazone (15 mg once daily, titrated to 30 mg once daily after 2 wk) or metformin (500 mg twice daily, titrated to 1000 mg twice daily) in addition to glimepiride throughout the study.
 
All patients underwent MR imaging (MRI)-scanning on a 1.5-T whole-body MR scanner after an overnight fast (Gyroscan ACS/NT15; Philips, Best, The Netherlands). The study protocol included cardiac magnetic resonance imaging to determine left ventricular cardiac function and 1H-magnetic resonance spectroscopy of the liver and heart for the assessment of hepatic and myocardial triglyceride content. The technical procedures were described before in detail (10).
 
Pericardial fat
 
Pericardial fat was quantified using electrocardiographically gated breath-holds with a balanced turbo-field echo MR sequence. Imaging parameters were: echo time, 1.60 msec; repetition time, 3.2 msec; flip-angle, 50°; slice thickness, 10 mm; field of view, 400 x 400 mm. The four-chamber view was analyzed, with the plane of respiratory mitral and tricuspid valves as margins. To quantify the periventricular fat volume, contours around the pericardial fat were drawn manually at end systole and multiplied by the thickness of the slice to yield a volume. We used MASS software (Medis, Leiden, The Netherlands) for postprocessing. Contours were drawn twice by the same observer, who was blinded to patient characteristics and the results of the first quantification and by a second independent observer to assess the intra- and interobserver variance.
 
Abdominal fat
 
Abdominal sc and visceral fat volumes were quantified by a turbo spin echo imaging MR protocol. At the level of the fifth lumbar vertebrae, three transverse images were acquired during one breath hold. Imaging parameters were: echo time, 11 msec; repetition time, 168 msec; flip angle, 90°; slice thickness, 10 mm. We quantified the sc and visceral fat volumes by converting the number of pixels to square centimeters and multiplied by the thickness of slices, using MASS software (Medis). The total volume of visceral and sc fat was calculated by totaling the volumes of the individual slices (11).
 
Statistics
 
Data are expressed as mean ± SEM or as median (interquartile range) when nonnormally distributed. Within-groups changes were assessed using paired t tests. Nonnormally distributed data were log transformed, or when appropriate, nonparametric tests were used for comparisons. Between-group differences were analyzed by ANOVA. For correlation analysis, Pearson's correlation coefficients or Spearman's correlation analysis were used when appropriate. For assessment of agreement within and between observers, a Bland-Altman plot was created and the inter- and intraclass correlation coefficients were calculated. P < 0.05 was considered statistically significant. We used SPSS 16.0 (SPSS Inc., Chicago, IL) for statistical analyses.
 
 
 
 
  iconpaperstack view older Articles   Back to Top   www.natap.org