|Home | About | Journals | Submit | Contact Us | Français|
The CHICAGO (Carotid Intima-Media Thickness in Atherosclerosis Using Pioglitazone) study demonstrated that pioglitazone (PIO) significantly decreased Carotid Intimal Media Thickness (CIMT) progression compared to glimepiride (GLM) in patients with Type 2 diabetes mellitus (T2DM). A secondary endpoint was to compare coronary artery calcium (CAC) progression between the two treatment groups.
CAC was determined at baseline and at the end of 72 weeks of treatment in the PIO (n=146) and GLM (n=153) treatment groups utilizing electron beam CT. There was no difference in CAC progression between the treatment groups. Using backward and forward selection models age, race/ethnicity and baseline ApoB level predicted CAC progression. There was no relationship between CIMT and CAC progression over the course of the study.
There was no difference in CAC progression in patients with T2DM treated with PIO or GLM. Age, race/ethnicity and baseline ApoB level predicted CAC progression in patients with T2DM.
The increased atherosclerosis in patients with diabetes is reflected in increased coronary artery calcium (CAC) measured by electron beam tomography (EBT).1, 2 CAC is a measure of total coronary atherosclerotic burden that has been validated by histopathology, coronary angiography and intravascular ultrasound.3, 4 The total CAC score measured represents an anatomic measure of overall cardiac plaque burden and in both Type 1 and Type 2 diabetes (T2DM).3, 5
The CHICAGO (Carotid Intima-Media Thickness in Atherosclerosis Using Pioglitazone) trial assessed the differential effect of pioglitazone vs. glimepiride on CIMT progression in patients with T2DM.6 The CHICAGO study demonstrated that pioglitazone over a 72-week treatment period slowed the progression of CIMT compared with glimepiride. A secondary endpoint of the CHICAGO trial was a comparison of progression of CAC between the two treatment groups. We have previously reported that the baseline CAC score in the CHICAGO patients was associated with age, gender, race, systolic blood pressure, triglycerides, apolipoprotein B (ApoB), triglyceride-rich lipoprotein cholesterol (TGRL-C), and the total/high density lipoprotein cholesterol (HDL-C).7 Low density lipoprotein cholesterol (LDL-C) or inflammatory markers, did not correlate with baseline CAC measures. We now report on the progression of CAC in patients in the CHICAGO trial after 72 weeks of treatment.
CHICAGO is a randomized 72 -week study comparing the effects of glimepiride and pioglitazone on measures of atherosclerotic disease. The primary endpoint of the trial was treatment related difference in carotid intima- media thickness from baseline to 72 weeks. Change in CAC score from baseline to 72 weeks was a secondary endpoint of the trial. Subjects eligible for the trial were between 45 to 85 years of age, of either sex, with a diagnosis of T2DM (based on American Diabetes Association diagnostic criteria). The complete design of the trial and description of inclusion and exclusion criteria have been previously published.6 Two hundred and ninety-nine subjects who completed baseline and 72-week measurement of CAC score were included in the current analysis.
Throughout the trial, fasting plasma was obtained to evaluate systemic lipid and lipoprotein levels, intermediary metabolites, indices of glycemia, and inflammatory factors. The analyses were performed by a clinical reference laboratory (CRL, Lenexa, Kansas) using methods previously described in detail.6, 7 For measuring carotid intima- media thickness, carotid arteries were imaged by high-resolution B mode ultrasonography by a single ultrasonographer or using the same equipment as previously described.6 EBT was used to measure coronary artery calcium as previously described in detail and the CAC score was calculated using the Agatston method.7 Abdominal adipose tissue distribution was also measured using EBT. Visceral adipose tissue (VAT) and total adipose tissue (TAT) was measured at the L4-L5 vertebrate as previously described 8–10
The potential relationship between treatment and change in Agatston from baseline after adjusting for baseline Agatston was investigated with Analysis of Covariance. This analysis was repeated using the Tobit transformation of the data (defined to be LN(Agatston + 1) in order to provide another analysis that considers change as a ratio.7
The two treatment groups were compared with respect to several baseline variables. Categorical variables were compared using Chi-square (or Fisher’s Exact test whenever expected cell size was five or less). Continuous variables were compared with the Wilcoxon Rank Sum Test (for treatment group comparisons) or the Kruskal-Wallis Test (for progression group comparisons). The relationship between CIMT and CAC score was examined with Pearson Correlation.
To identify predictors of CAC progression in our cohort, both forward and backward elimination were performed. Predicting variables considered included age, gender, race/ethnicity, smoking, systolic blood pressure, and baseline values of lipids/lipoproteins (HDL-C, ApoA1, triglycerides, ApoB, TGRL-C, total cholesterol-HDL-C ratio, LDL-C, LDL particle number); anthropometric measures (VAT-TAT ratio, waist, waist-hip ratio, BMI); and metabolic measures (A1C, free fatty acid, duration of diabetes, CRP). For both backward and forward selection approaches, entry and exit probabilities were set at the SAS default value of 5%.
All analyses were conducted with SAS v9.1 (SAS Institute, Cary, NC).
Table 1 presents baseline categorical characteristics for the two treatment groups of subjects who had baseline and 72 week measurement of CAC score. The two treatment groups were similar for sex, race/ethnicity, smoking history, and presence of hyperlipidemia or hypertension. Baseline CAC score was also not different between the two treatment groups. In addition to the factors shown in Table 1, there were no significant differences between the treatment groups for LDL cholesterol, LDL particle number, HDL cholesterol, triglyceride-rich lipoprotein cholesterol (TGRL-C), ApoB, ApoA1, free fatty acid level, duration of diabetes, VAT, TAT, VAT/TAT, or hsCRP.
Figure 1 presents box plots for change in coronary calcium for each treatment group. Change over 72 weeks was similar in each treatment in each treatment group and there was not a significant difference in CAC score progression between the treatment groups. Results of the analysis using the Tobit transformation of the data also showed no difference between the treatment groups (not shown).
We have previously reported results of a cross-sectional analysis that identified predictors of CAC score at baseline7 in this cohort. In that analysis, we identified age, systolic blood pressure, sex, and race/ethnicity as significant predictors of prevalent coronary artery calcium. Among lipid and lipoprotein parameters, only ApoB, triglyceride, TGRL-C and total cholesterol to HDL cholesterol ratio were significantly associated with the presence of coronary artery calcium. Waist-hip ratio and VAT-TAT ratio were significantly associated with baseline CAC score but not after inclusion of ApoB or TGRL-C in the model. For the current analysis, we used forward and backward selection approaches to evaluate age, gender, race/ethnicity, smoking, systolic blood pressure, as well as baseline anthropometric, metabolic and lipid/lipoprotein measures (see Methods) for predicting progression of CAC score over 72 weeks. Of all factors considered, only age, race/ethnicity, and baseline ApoB level were significant predictors of CAC score progression. Forward and backward selection gave identical results.
In the CHICAGO cohort, 267 subjects had baseline and 72 week measurements of both CAC score and CIMT. In view of the discordance in results for predicting progression of CIMT and CAC score, we performed an analysis to examine the relationship between these two measures. As shown in Table 3, baseline CIMT was not significantly related to baseline CAC score or change in CAC score. Change in CIMT over 72 weeks was not related to either baseline CAC score or 72 week change in CAC score.
Unlike its beneficial effects on CIMT progression,6, 11 pioglitazone did not alter the progression of CAC over 72 weeks compared to glimepiride in patients with T2DM (Fig. 1). Our analysis identified only age, race/ethnicity, and ApoB as predictors of CAC score over 72 weeks in a cohort of 299 subjects with T2DM. In addition, we found no relationship between CIMT and CAC baseline measures or progression rates in this cohort.
Recent statin trials have demonstrated that ApoB is the single best lipid predictor of clinical events on statin therapy.12 Of interest is that LDL-C or LDL-particle number were not identified as significant predictor of CAC progression in our study. This result suggests that triglyceride-rich ApoB lipoproteins prior to conversion to LDL may be the primary drivers of CAC progression in T2DM. Our data expands on that of the Penn Diabetes Heart Study13 that reported, in a cross-sectional analysis, that ApoB but not LDL-C predicted prevalent CAC in white subjects with Type 2 diabetes. Our study suggests that ApoB, but not LDL-C or LDL-particle number, significantly predicts CAC progression in a multiracial cohort with T2DM.
The Multi-Ethnic Study of Atherosclerosis (MESA) has reported that age, male gender, Caucasian race, and BMI were associated with increases in CAC.14 Therefore, both the CHICAGO and MESA data support racial difference for predicting both in baseline and progression of CAC in patients with diabetes.15, 16 The causes of these racial differences are uncertain and deserve further study.17, 18 In our analyses, differences in lipoprotein, inflammatory or anthropometric measures did not appear to explain racial differences in CAC score progression (not shown).
Similar to what has been shown for statins, pioglitazone treatment reduces progression of both CIMT6 and coronary atheroma19 volume as determined by intravascular ultrasound but does not influence the change in CAC after 72 weeks of treatment. Pioglitazone has also been shown to significantly reduce by 16% the risk of the main secondary endpoint (a composite of all cause mortality, myocardial infarction (MI), or stroke) in the PROactive trial.20 This outcome benefit was sustained with a 28% relative risk reduction in a subgroup of patients with a previous MI.21 Therefore, pioglitazone, while demonstrating a clinical benefit utilizing two surrogate endpoints (CIMT and IVUS) and in reducing major hard cardiac events, did not affect progression of the surrogate endpoint of CAC. In addition, we could find no correlation between baseline CIMT or change in CIMT with baseline or progression of CAC in this patient population with T2DM. Further, in our previous analysis of factors that predicted slowing of CIMT progression, an increase in HDL cholesterol with pioglitazone was the strongest predictor of slowed progression,11 but changes in HDL cholesterol did not predict change in CAC.
In the MESA trial, CAC was associated more strongly than CIMT with risk of incident CVD (2.5 vs. 1.2-fold).22 However, in a recent metanalysis of five CAC trials that evaluated therapeutic effects including statins and anti-hypertensive therapy, the authors concluded that there were no consistent or reproducible treatment effects of any therapy on progression of CAC measured at one year.23 The CHICAGO data, along with this metanalysis, call into question the clinical value of serial CAC scanning to evaluate the efficacy of therapeutic interventions for preventing clinical CVD. The process of coronary calcification is complex, involving both an increase in plaque burden and the healing of lipid-rich atheromata.1 A change in CAC may reflect these potentially competing processes, thereby making this surrogate invalid for assessing a therapeutic response.
There are several limitations that should be considered when interpreting our study. The CHICAGO trial was designed using CIMT as its primary endpoint, and there was not complete overlap in subjects completing the CIMT and CAC measurements. However, comparison of the two CAC treatment groups showed no significant baseline differences (Table 1 and page 7) in important CVD risk factors, and no difference in baseline CAC score. The trial was also powered using change in CIMT as its primary endpoint. However, assuming a progression rate of 50 Agatston units over the trial, and a measurement standard deviation of 30 Agatston units, we estimate a 90% power to detect a 25 Agatston unit difference between the GLM and PIO treatment groups. Further, the data in Fig. 1 demonstrates complete overlap of CAC progression in the PIO and GLM treatment groups. It seems unlikely that inclusion of additional subjects or observation for a longer period of time would have changed this result. Finally, the data in Table 2 identify age, race/ethnicity, and ApoB as the only important predictors of CAC progression in this multiracial cohort with T2DM. The size of our cohort or the duration of our trial may have precluded detection of other important predictors of CAC progression in T2DM. However, our identification of ApoB as an important predictor of progression is consistent with cross-sectional analyses from this and another cohort, 7, 13 indicating that ApoB level is an important predictor of prevalent CAC in T2DM.
The major points of our analysis can be summarized as follows. In a racially diverse cohort of patients with T2DM treatment with PIO, already demonstrated to produce a beneficial effect on CIMT and coronary atheroma progression compared to treatment with GLM, did not beneficially affect progression of CAC compared to treatment with GLM. In addition, there was no relationship between baseline or progression measurements of CAC and CIMT. Exploratory models to identify predictors of CAC progression identified ApoB as the only modifiable risk factor that was related to CAC progression over 72 weeks in subjects with T2DM. This latter finding supports the Joint Consensus Statement from the American Diabetes Association and the American College of Cardiology that recommends incorporating ApoB into managing patients with cardiometabolic risk once LDL and non-HDL cholesterol goals have been achieved.24
The authors thank Stephanie Thompson and Mary Lou Briglio for assistance with manuscript preparation.
SOURCES OF FUNDING
The CHICAGO study was sponsored and funded by Takeda Global Research and Development. Analysis for the current study was supported by an unrestricted grant from Takeda Global Research and Development, by an institutional award from the University of Illinois at Chicago, and by grant UL1RR029879 from the National Center for Research Resources.
C.B. and R.B.D. have nothing to disclose. M.H.D. is a Consultant for Abbott, AstraZeneca, Daiichi-Sankyo, Inc., DiaDexus, Inc., GlaxoSmithKline, Merck, Schering-Plough, Pfizer, Roch, Sanofi-Aventis, Synarc, Takeda Pharmaceuticals; on the Speakers’ Bureau of Abbott, AstraZeneca, Daiichi-Sankyo, Inc., diaDexus, Inc., GlaxoSmithKline, Merck, Schering-Plough, Oscient, Pfizer, Takeda Pharmaceuticals; has received grant/research support from Abbott, AstraZeneca, Daiichi-Sankyo, Inc., Merck, Schering-Plough Pfizer, Roche, Takeda; on the Advisory Board of Abbott, Access Health, AstraZeneca, Atherogenics, Daiichi-Snakyo, Inc. GlaxoSmithKline, Kincmed, Merck, Schering-Plough, Oscient, Pfizer, PreEmptive Meds, Roche, Takeda Pharmaceuticals, and Xinthria Pharmaceuticals; has Equity/Board of Directors for Angiogen, Sonogene, Professional Evaluation, Inc. Medical Education Company. A.P. is employed by Takeda Global Research and Development. S.H. is a consultant for AstraZeneca, Merck, Novartis. T.M. is a Consultant for Abbott, Merck, Takeda; and has received grants/research support from Takeda.