NIH Public Access

Author Manuscript
Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

NIH-PA Author Manuscript

Published in final edited form as:

Am J Cardiol. 2007 November 15; 100(10): 15431547.

Progression of Coronary Artery Calcium in Type 1 Diabetes

Mellitus
Tina Costacou, PhDa, Daniel Edmundowicz, MDb, Catherine Prince, MPHa, Baqiyyah
Conway, MPHa, and Trevor J. Orchard, MBBCh, MMedScia
a Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh
b Cardiovascular Institute, School of Medicine, University of Pittsburgh

Abstract

NIH-PA Author Manuscript

Coronary artery calcium (CAC) has been previously associated with atherosclerotic plaque disease
and coronary events. Thus, identifying predictors of CAC progression may provide new insights on
early risk factor intervention and subsequent reduction of more severe atherosclerotic disease. We
aimed to identify risk factors of CAC progression and evaluate whether risk factor change relates to
CAC progression in a cohort of type 1 diabetes mellitus (DM). Participants of the Pittsburgh
Epidemiology of Diabetes Complications study, a prospective investigation of childhood-onset type
1 DM, who received 2 electron beam computed tomography screenings 4 years apart were selected
for study (n=222). CAC was calculated by the Agatston method of scoring and progression was
defined as an increase >2.5 in the square root-transformed CAC score. Adjusting for diabetes duration
and initial CAC score, body mass index (BMI, OR=1.13 95% CI=1.01-1.26), non-high density
lipoprotein cholesterol (OR=1.01, 95% CI=1.003-1.03), and albumin excretion rate (OR=1.30, 95%
CI=1.03-1.63) were associated with CAC progression. When considering change in risk factors, an
increase in BMI (OR=1.38, 95% CI=1.10-1.72) was also associated with CAC progression after
adjustment. In conclusion, in this cohort of type 1 DM, in addition to baseline BMI, non-high density
lipoprotein cholesterol and albumin excretion rate, all known coronary artery disease risk factors,
weight gain further added to the prediction of CAC progression. Thus, weight control, in addition to
lipid and renal management may help retard atherosclerosis progression in type 1 DM.

Keywords

NIH-PA Author Manuscript

coronary artery calcification; progression; type 1 diabetes mellitus; BMI

Introduction
Electron beam computed tomography (EBT) is a non-invasive method for quantifying the
extent of calcium formation present in the coronary arteries. Coronary artery calcium (CAC)
measured by EBT can be used as an indicator of atherosclerotic burden (1). Furthermore, among
persons with type 1 diabetes mellitus (DM), CAC has been previously shown to correlate with
coronary artery disease (2). Thus, identifying modifiable risk factors associated with CAC
progression may provide new targets for early intervention and potentially reduction of the

Corresponding author: Trevor J. Orchard, MD, University of Pittsburgh, Diabetes and Lipid Research Bldg., 3512 Fifth Avenue,
Pittsburgh, PA 15213, Email: OrchardT@edc.pitt.edu, Tel: 412-383-1032, Fax: 412-383-1020.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting
proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could
affect the content, and all legal disclaimers that apply to the journal pertain.

Costacou et al.

Page 2

NIH-PA Author Manuscript

rates of more severe atherosclerotic disease. However, with the exception of the CAC in Type
1 DM (CACTI) study (3-6), we are not aware of other investigations on predictors of CAC
progression in type 1 DM. We, therefore, aimed to identify risk factors of CAC progression
and to evaluate whether risk factor change relates to CAC progression in a type 1 DM cohort.
Ultimately, upon sufficient follow-up and cardiovascular events, we will assess whether
progression is a better predictor of events than the baseline score and thus, whether CAC
progression can be used for monitoring.

Methods

NIH-PA Author Manuscript

Participants for this evaluation were selected from the Pittsburgh Epidemiology of Diabetes
Complications (EDC) study cohort, a prospective investigation of childhood-onset type 1 DM
now entering its 20th year of follow-up. The EDC study has been previously described in detail
(7). Briefly, participants were diagnosed prior to their 17th birthday, or seen within 1 year of
such diagnosis, at Children's Hospital of Pittsburgh between 1950 and 1980. Although clinic
based, this population has been shown to be representative of the type 1 DM population in
Allegheny County, Pennsylvania (8). Baseline examinations for the EDC study were conducted
between 1986 and 1988, when the cohort's mean age was 28 and diabetes duration was 19
years. Participants were subsequently invited to biennial examinations. At the 10-year followup examination (1996-1998), EBT screening was made available first to all participants aged
30 years and subsequently to all those over 18 years of age. A total of 304 participants
underwent EBT screening (2); 228 thereof also had a repeat scan approximately 4 years later.
The University of Pittsburgh Institutional Review Board approved the study protocol.
All complications and risk factors were assessed at the time of the 2 EBT screenings, the first
of which occurred during the 10-year follow-up examination (1996-1998) and the second
approximately 4 years later. Prior to each biennial clinic visit, participants were sent
questionnaires concerning demographic, health care, medical history, and physical activity
information. An ever smoker was defined as a person who smoked 100 cigarettes over their
lifetime. Intensive insulin therapy was defined as 3 insulin shots daily or insulin pump use.
Depressive symptomatology was assessed by the Beck Depression Inventory (9). Hypertension
was defined as blood pressure 140/90 mmHg or current antihypertensive medication
treatment.

NIH-PA Author Manuscript

CAC was assessed via EBT using a GE-Imatron ultrafast computed tomography scanner (GEImatron, San Francisco, CA). Scans were triggered by electrocardiogram signals at 80% of the
R-R interval and obtained in 3 mm contiguous sections of the heart. CAC was calculated by
the Agatston method (10) and progression was defined as an increase >2.5 in the square roottransformed CAC score (11), as this transformation provided the best fit. Previous use of a
suggested modification (12) to address the potential systematic bias of elevated calcium scores
in those with higher body mass index (BMI) did not appear to significantly alter scoring. While
the use of a phantom may be helpful in cross-sectional observations of a population, we have
not found that it adds to longitudinal analyses within our study. EBT results were made
available to participants and their physicians and cardiovascular risk factor reduction was
stressed for those with significant CAC. Further evaluation was decided upon by the
participants' physicians, although a cardiologist was available for consultation. Coronary artery
disease was defined as EDC physician-diagnosed angina, myocardial infarction confirmed by
Q-waves on an electrocardiogram (Minnesota codes 1.1 or 1.2) or hospital records,
angiographic stenosis 50%, revascularization, or ischemic ECG changes (Minnesota codes
1.3, 4.1-4.3, 5.1-5.3, and 7.1).
Original hemoglobin A1 (1986-1998) and hemoglobin A1c (1998-2004) were converted to
DCCT standard hemoglobin A1c values using regression formulae derived from duplicate

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 3

NIH-PA Author Manuscript

analyses (DDCT hemoglobin A1c = (0.83 * EDC hemoglobin A1) + 0.14; and DCCT
hemoglobin A1c = (EDC hemoglobin A1c 1.13) / 0.81). Cholesterol and triglycerides were
measure enzymatically (13,14). High density lipoprotein cholesterol was determined using a
modified (15) heparin and manganese chloride precipitation technique of the Lipid Research
Clinics method (16). Low density lipoprotein cholesterol was calculated by using the
Friedewald equation (17). Non high density lipoprotein cholesterol was calculated as total
minus high density lipoprotein cholesterol. Estimated glucose disposal rate (insulin sensitivity)
was estimated by a regression equation derived from hyperinsulinemic euglycemic clamp
studies on 24 subjects chosen to represent the full spectrum of insulin resistance as represented
by insulin resistance risk factors (18). White blood cell count was obtained using a counter Splus IV.
Serum and urinary albumin were measured by immunonephelometry (19,20), and creatinine
was assayed by an Ectachem 400 Analyzer (Eastman Kodak Co., Rochester, NY), which unlike
picric acid-based methods, does not overestimate creatinine concentrations in diabetes
individuals (21). Overt nephropathy was defined as albumin excretion rate >200 g/min in 2
of 3 timed urine collections or, in the absence of urine, a serum creatinine >153 mol/l (>2
mg/dl) or, renal failure or renal transplantation.

NIH-PA Author Manuscript

Variables not following a normal distribution were log transformed (i.e. triglyceride
concentration, Beck depression inventory, serum creatinine, and albumin excretion rate). When
normality was not achieved by transformation (i.e. for change in the following variables: BMI,
low density lipoprotein, non high density lipoprotein, total cholesterol, Beck depression
inventory, and serum creatinine), non-parametric procedures were used. The student's t-test
and chi square (or Fisher's exact, as appropriate) tests were used to examine univariate
associations, whereas multiple logistic regression models were constructed to assess significant
predictors of CAC progression, as well as the predictive value of risk factor change. To evaluate
the effect of risk factors over the 10 years up to the initial CAC measurement on progression,
analyses were also conducted using updated means of risk factors. Data were analyzed using
SAS version 9.1 (Cary, NC). Results were considered significant at p<0.05.

Results

NIH-PA Author Manuscript

Of the 304 individuals with a first EBT, 22 (7.2%) died, whereas another 54 (17.8%) did not
return for a second EBT. Compared to persons with a repeat EBT, those who did not return for
a second EBT had a higher hemoglobin A1c (p=0.03) and white blood cell counts (p=0.01) and
were more insulin resistant (p=0.003), but had similar CAC scores at first assessment (p=0.85).
Of the 228 persons with 2 EBT scans performed, 222 had full information on the covariates
examined and their characteristics by progression of CAC are shown in Table 1. Surprisingly,
glycemic control was not related to CAC progression in this cohort and there was no evidence
of a threshold effect. Generally, similar univariate risk factors were observed for male and
female participants, although measures of body fatness and the Beck depression inventory were
associated with CAC progression only among men. Interestingly, although baseline
hemoglobin A1c was not in itself related to CAC progression, a greater proportion of nonprogressors used more than 3 insulin injections daily among females (p=0.07); however, the
proportion on intensive insulin therapy who also regularly monitored their blood glucose (3
times daily) did not differ significantly (p=0.14). In multivariable logistic regression analyses,
the final model included diabetes duration, baseline calcification level, BMI, non high density
lipoprotein cholesterol, and albumin excretion rate (Table 2, Model 1) as significant predictors
of CAC progression. These results were not altered when analyses were restricted to persons
free of cardiovascular disease (n=183; 83 progressors). When logistic regression was
conducted separately for male and female participants, with the exception of baseline CAC
score (OR=1.51, 95% CI=1.22-1.86), BMI was the only significant predictor of CAC

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 4

NIH-PA Author Manuscript

progression among males (OR=1.31, 95% CI=1.09-1.59), whereas in females significant risk
indicators included the baseline CAC score (OR=1.46, 95% CI=1.20-1.77), duration of
diabetes (OR=1.13, 95% CI=1.04-1.23), non high density lipoprotein cholesterol (OR=1.02,
95% CI=1.001-1.04), and albumin excretion rate (OR=1.63, 95% CI=1.11-2.40). The final
model using updated means of risk factors from study initiation to first EBT scan is presented
in Table 2 (Model 2). Although results were quite similar to those using risk factors from the
time of the first EBT scan, years with hypertension and years of angiotensin converting enzyme
inhibitor use now also became significant whereas albumin excretion rate was no longer
selected.

NIH-PA Author Manuscript

In assessing change in risk factors (follow-up baseline) by CAC progression, generally,

greater increases in BMI and weight and greater decreases in lipid levels (due to increased use
of lipid lowering medication) were noted among progressors. Despite greater use of
hypertension medication among progressors, no differences were observed in blood pressure
levels. Once again, change in hemoglobin A1c was not related to CAC progression in this cohort
(not shown). Logistic models were constructed for 182 individuals with risk factor data
available both at baseline and at follow-up. Univariately significant change variables were
included in logistic regression models, adjusting for diabetes duration, baseline CAC level, as
well as the baseline level of the change variable (not shown). Only change (increase) in BMI
was significantly associated with CAC progression. This association was somewhat stronger
among male than female (OR for a unit increase in BMI was 1.76 (95% CI=1.17-2.65) and
1.28 (95% CI=0.997-1.65), respectively) participants (p-value for interaction=0.08). Including,
one at a time or in backward elimination, univariately significant change variables to Model 1
from Table 2, change in BMI was the only significant change covariate (final model presented
in Table 3).

Discussion
In this cohort of individuals with a long duration of type 1 DM, consistent baseline predictors
of CAC progression were BMI, non-high density lipoprotein cholesterol, and albumin
excretion rate, all recognized CAD risk factors. Diabetes duration was also a predictor of CAC
progression. In examining the value of risk factor change on CAC progression, we found that
only an increase in BMI was associated with progression of CAC.

NIH-PA Author Manuscript

Contrary to a previous report on CAC progression by the CACTI study (3) and another on
lower calcium prevalence in the primary prevention cohort from the DCCT/EDIC (22),
glycemic control was not related to CAC progression in this cohort and we were not able to
identify a threshold. In the late 90's, glycemic control was far from optimal among EDC
participants, with only 14.0% of the study cohort keeping their hemoglobin A1c levels below
7.0%. Nevertheless, the proportion who progressed was similar, if not slightly higher, among
those with optimal glycemic control compared to persons with higher levels of hemoglobin
A1c (58.1% versus 48.2%, p=0.31). Snell-Bergeon et al. (3) also reported an increased risk of
CAC progression with higher insulin dose among overweight individuals. This relationship
differed for the female and male participants of the EDC study. In contrast to the CACTI results,
both insulin dose (47.0 versus 39.4, p=0.17) as well as the dose of insulin per body weight
(0.64 versus 0.51, p=0.05) were slightly increased among overweight women who did not
progress despite similar insulin sensitivity, as measured by eGDR. However, none of these
relationships were significant after adjustment for duration of diabetes and baseline CAC score.
As in CACTI (3) though, there was an indication of lower units of insulin (55.6 versus 64.0,
p=0.06) and insulin dose per body weight (0.67 versus 0.76, p=0.09) in overweight men who
did not progress compared to progressors, even after adjustment for diabetes duration and
baseline CAC score.

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 5

NIH-PA Author Manuscript

Despite the small sample size, analyses of risk factor change in relation to CAC progression
identified BMI as the only covariate for which elevations appear to increase risk of CAC
progression. No considerable decline in hemoglobin A1c was observed during the follow-up
period in either group (an increase of 0.07 units for the stable versus a decrease of 0.19 units
for the progressing group) and thus, hemoglobin A1c change was associated with progression.
Nevertheless, we have previously shown that although baseline hemoglobin A1c was not related
to coronary artery disease incidence, hemoglobin A1c elevations during the follow-up period
increased risk of disease (23). Together, these findings might suggest that hemoglobin A1c is
not related to the underlying atherosclerotic process but rather to the precipitation of events.
CACTI investigators have previously reported that adiponectin levels are inversely associated
with CAC progression in both persons with type 1 DM and non-diabetic adults (4).
Unfortunately, data on adiponectin are not currently available in this subset of the EDC
population, although we have also previously observed a relationship between adiponectin and
the incidence of coronary artery disease (24). Two further reports from CACTI also identified
a marker of T-cell activation, soluble interleukin-2 receptor (sIL2r) and the APOA4 Gln360His
polymorphism as risk factors for CAC progression in type 1 DM (5,6). However, information
on these polymorphisms are not available in EDC and thus, their effect could not be assessed.

NIH-PA Author Manuscript

Study limitations include risk factor differences between those with and without a repeat EBT
scan; despite similar baseline CAC scores, if those without a repeat scan had a greater rate of
CAC progression, it is likely that the true relationship between risk factors and CAC
progression is stronger than observed here. It is further possible that, depending on their initial
CAC score, participants' physicians intervened to manage risk factors. Indeed, as previously
reported (25), further cardiac testing increased with increasing initial CAC score. Such
management may have led to weaker associations between baseline risk factors and CAC
progression, although evaluation of risk factor change addresses this problem.
Our ultimate goal is to identify what rate of progression best predicts subsequent cardiovascular
events. Although sufficient follow-up time and events have not yet occurred to fully address
this issue, a very preliminary look suggests that in our population CAC score is the best single
predictor of CAD events (OR=8.57, 95% CI=1.16-50.27 for a CAC score 100) (26).
Approximately 9% of our population developed this risk level over the 4-year period.
Acknowledgements
This research was funded by NIH grant DK34818

References
NIH-PA Author Manuscript

1. Janowitz WR, Agatston AS, Kaplan G, Viamonte M Jr. Differences in prevalence and extent of
coronary artery calcium detected by ultrafast computed tomography in asymptomatic men and women.
Am J Cardiol 1993;72:247254. [PubMed: 8342500]
2. Olson JC, Edmundowicz D, Becker DJ, Kuller LH, Orchard TJ. Coronary calcium in adults with type
1 diabetes: A stronger correlate of clinical coronary artery disease in men than in women. Diabetes
2000;49:15711578. [PubMed: 10969842]
3. Snell-Bergeon JK, Hokanson JE, Jensen L, MacKenzie T, Kinney G, Dabelea D, Eckel RH, Ehrlich
J, Garg S, Rewers M. Progression of coronary artery calcification in type 1 diabetes: the importance
of glycemic control. Diabetes Care 2003;26:29232928. [PubMed: 14514603]
4. Maahs DM, Ogden LG, Kinney GL, Wadwa P, Snell-Bergeon JK, Dabelea D, Hokanson JE, Ehrlich
J, Eckel RH, Rewers M. Low plasma adiponectin levels predict progression of coronary artery
calcification. Circulation 2005;111:747753. [PubMed: 15699257]
5. Wadwa RP, Kinney GL, Ogden L, Snell-Bergeon JK, Maahs DM, Cornell E, Tracy RP, Rewers M.
Soluble interleukin-2 receptor as a marker for progression of coronary artery calcification in type 1
diabetes. Int J Biochem Cell Biol 2006;38(56):9961003. [PubMed: 16271309]
Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 6

NIH-PA Author Manuscript

NIH-PA Author Manuscript
NIH-PA Author Manuscript

6. Kretowski A, Hokanson JE, McFann K, Kinney GL, Snell-Bergeon JK, Maahs DM, Wadwa RP, Eckel
RH, Ogden LG, Garg SK, Li J, Cheng S, Erlich HA, Rewers M. The apolipoprotein A-IV Gln360His
polymorphism predicts progression of coronary artery calcification in patients with type 1 diabetes.
Diabetologia 2006;49(8):19461954. [PubMed: 16770585]
7. Orchard TJ, Dorman JS, Maser RE, Becker DJ, Drash AL, Ellis D, LaPorte RE, Kuller LH. Prevalence
of complications in IDDM by sex and duration: Pittsburgh Epidemiology of Diabetes Complications
Study II. Diabetes 1990;39:11161124. [PubMed: 2384191]
8. Wagener DK, Sacks JM, LaPorte RE, MacGregor JM. The Pittsburgh Study of insulin-dependent
diabetes mellitus: risk for diabetes among relatives of IDDM. Diabetes 1982;31:136144. [PubMed:
6759229]
9. Beck AT, Garbin MG. Psychometric properties of the Beck depression inventory: 25 years of
evaluation. Clin Psychol Rev 1988;8:77100.
10. Agatston AS, Janowitz WH, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantitation of
coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827832.
[PubMed: 2407762]
11. Hokanson JE, MacKenzie T, Kinney G, Snell-Bergeon JK, Dabelea D, Ehrlich J, Eckel RH, Rewers
M. Evaluating changes in coronary artery calcium: an analytic method that accounts for interscan
variability. AJR 2004;182:13271332. [PubMed: 15100140]
12. Raggi P, Callister TQ, Cooil B. Calcium scoring of the coronary artery by electron beam CT: how to
apply an individual attenuation threshold. AJR 2002;178:497502. [PubMed: 11804925]
13. Allian CC, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic determination of total serum
cholesterol. Clin Chem 1974;20:470475. [PubMed: 4818200]
14. Bucolo G, David H. Quantitative determination of serum triglycerides by the use of enzymes. Clin
Chem 1973;19:476482. [PubMed: 4703655]
15. Warnick GR, Albers JJ. Heparin-MN(2+) quantitation of high density lipoprotein cholesterol: an
ultrafiltration procedure for lipemic samples. Clin Chem 1978;24:900904. [PubMed: 207462]
16. National Institutes of Health, Department of Health Education and Welfare. Lipid Research Clinics
Program. Washington, DC: U.S. Government Printing Office; 1975.
17. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein
cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499502.
[PubMed: 4337382]
18. Williams KV, Erbey JR, Becker DJ, Arslanian S, Orchard TJ. Can clinical factors estimate insulin
resistance in type 1 diabetes? Diabetes 2000;49:626632. [PubMed: 10871201]
19. Ellis D, Buffone GJ. A new approach to the evaluation of proteinuric states. Clin Chem 1977;23:666
670. [PubMed: 66106]
20. Ellis D, Coonrod BA, Dorman JS, Kelsey SF, Becker DJ, Avner ED, Orchard TJ. Choice of urine
sample predictive of microalbuminuria in patients with insulin-dependent diabetes mellitus. Am J
Kidney Dis 1989;4:321328. [PubMed: 2705450]
21. Assadi FK, John EG, Fornell L, Rosenthal IM. Falsely elevated serum creatinine concentration in
ketoacidosis. J Pediatr 1985;107:562564. [PubMed: 3930679]
22. Cleary PA, Orchard TJ, Genuth S, Wong ND, Detrano R, Backlund JY, Zinman B, Jacobson A, Sun
W, Lachin JM, Nathan DM, DCCT/EDIC Research Group. The effect of intensive glycemic treatment
on coronary artery calcification in type 1 diabetic participants of the Diabetes Control and
Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC)
Study. Diabetes 2006;55:35563565. [PubMed: 17130504]
23. Prince KT, Costacou T, Orchard TJ. Glycemia and cardiovascular risk in type 1 diabetes: reconciling
conflicting results. Diabetes 2006;55:A1. [PubMed: 17511053]
24. Costacou T, Zgibor JC, Evans RW, Otvos J, Lopes-Virella MF, Tracy RP, Orchard TJ. The
prospective association between adiponectin and coronary artery disease among individuals with
type 1 diabetes. The Pittsburgh Epidemiology of Diabetes Complications Study. Diabetologia
2005;48:4148. [PubMed: 15616802]
25. Huskey ND, Edmundowicz D, Orchard TJ. Electron beam computed tomography and coronary artery
calcification; Recipients attitudes and actions. Diabetes 2004;53:A283.

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 7

26. Orchard TJ, Costacou T, Edmundowicz D. Coronary calcification as a predictor of coronary artery
disease in type 1 diabetes. Circulation 2006;113:247.

NIH-PA Author Manuscript

NIH-PA Author Manuscript
NIH-PA Author Manuscript
Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

NIH-PA Author Manuscript

Table 1

NIH-PA Author Manuscript

29 (26.4%)
24 (21.8%)
26 (23.6%)
31 (28.2%)
57 (51.8%)
38 (36.2%)

24 (21.4%)
32 (28.6%)
27 (24.1%)
29 (25.9%)
56 (50.0%)
44 (40.4%)

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

p-value < 0.05

0.98 (0.75)
89.9 (334.9)
18 (16.1%)

0.64 (0.22)
120.5 (17.3)*
70.1 (12.9)
37 (33.6%)*
29 (26.4%)
5 (4.6%)*
54.0 (14.4)
190.4 (34.0)
107.4 (56.6)*
142.5 (36.4)*
196.5 (37.7)*
6.9 (6.8)*

0.67 (0.19)
111.3 (13.2)
69.8 (8.7)
16 (14.3%)
22 (19.6%)
0 (0.0%)
55.9 (14.2)
110.6 (28.3)
91.0 (44.3)
126.9 (31.4)
182.9 (31.6)
5.3 (6.8)
71.5 (10.8)
6.6 (1.8)
12 (10.7%)
1.1 (0.69)*
140.2 (344.8)*
33 (30.0%)*

70.0 (9.1)
7.0 (2.0)
27 (24.6%)*

21 (20.4%)

29 (27.4%)

Logarithmically transformed before statistical testing

Serum creatinine (mg/dl)

Albumin excretion rate (g/min)
Overt nephropathy

Pulse rate
White blood cell count 103/mm2
Coronary artery disease

Intensive insulin therapy and checking blood

glucose 3 times daily
Insulin / kg body weight / day
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Hypertension
Angiotensin converting enzyme inhibitor use
Blood pressure medication use
High density lipoprotein (mg/dl)
Low density lipoprotein (mg/dl)
Triglycerides (mg/dl)
Non-high density lipoprotein (mg/dl)
Total cholesterol (mg/dl)
Beck depression inventory

8.2 (1.3)

34.1 (6.5)
25.8 (6.4)
62 (55.4%)
1.00 (1.2)
24.6 (3.4)
167.8 (8.3)
69.4 (10.8)
34 (30.4%)
8.0 (2.0)
8.3 (1.3)

Age (years)
Diabetes duration (years)
Females
Waist to hip ratio
Body mass index (kg/m2)
Height (m)
Weight (kg)
Ever smoked (n)
Estimated glucose disposal rate (mg/kg/min)
Hemoglobin A1c (%)
Quartiles of hemoglobin A1c (%)
<7.3
7.3 - <8.2
8.2 - <9.0
9.0
Hemoglobin A1c above median (8.2%)
Intensive insulin therapy

41.6 (7.0)*
32.1 (6.8)*
56 (50.9%)
1.03 (1.2)
25.5 (3.8)
167.9 (9.5)
71.2 (14.1)
38 (34.6%)
7.1 (2.3)*

Stable

Participant characteristics

Overall
(n=222; 110 progressed)
Progressed

0.97 (0.30)
152.3 (478.8)
9 (18.0%)

0.70 (0.20)
113.6 (12.2)
72.9 (8.4)
6 (12.0%)
12 (24.0%)
0 (0.0%)
48.8 (10.1)
114.9 (26.9)
97.3 (51.0)
131.7 (30.7)
180.5 (30.1)
3.7 (6.8)
72.7 (11.3)
6.8 (2.1)
4 (8.0%)

9 (18.8%)

8 (16.0%)
15 (30.0%)
12 (24.0%)
15 (30.0%)
27 (54.0%)
13 (26.5%)

1.1 (1.2)
26.4 (3.1)*
174.4 (7.4)
80.3 (11.2)*
19 (35.2%)
6.3 (2.1)*

1.1 (1.3)
24.7 (2.8)
174.1 (5.8)
74.9 (9.7)
17 (34.0%)
7.3 (1.7)
8.4 (1.2)

69.6 (8.9)
6.8 (1.8)
13 (24.1%),
p=0.03
1.1 (0.67)
163.2 (332.5)*
16 (29.6%)

0.72 (0.22)
122.2 (18.4)*
76.0 (13.4)
19 (35.2%)*
15 (27.8%)
2 (3.7%)
47.5 (11.2)
118.8 (30.1)
120.5 (66.6)*
147.3 (32.7)*
194.8 (33.7)*
5.9 (6.6)*

10 (20.0%)

15 (27.8%)
12 (22.2%)
9 (16.7%)
18 (33.3%)
27 (50.0%)
20 (38.5%)

8.3 (1.3)

40.1 (7.1)*
31.1 (6.8)*

34.2 (5.8)
26.1 (5.9)

Stable

Males
(n=104; 54 progressed)
Progressed

0.93 (0.83)
40.2 (120.1)
9 (14.5%)

0.65 (0.18)
109.4 (13.9)
67.4 (8.2)
10 (16.1%)
10 (16.1%)
0 (0.0%)
61.6 (14.7)
106.4 (28.8)
86.1 (38.1)
122.3 (31.4)
184.0 (32.5)
6.8 (6.6)
70.5 (10.5)
6.4 (1.5)
8 (12.9%)

20 (34.5%)

16 (25.8%)
17 (27.4%)
15 (24.2%)
14 (22.6%)
29 (46.8%)
31 (51.7%)

0.79 (0.07)
24.6 (3.8)
162.7 (6.3)
65.0 (9.6)
17 (27.4%)
8.6 (2.0)
8.1 (1.4)

33.9 (7.1)
25.5 (6.8)

Stable

0.55 (0.19)*
119.0 (16.4)*
64.2 (9.4)
18 (32.1%)*
14 (25.0%)
3 (5.4%)
60.5 (14.4)
120.7 (38.0)*
93.0 (40.6)
138.0 (39.8)*
198.6 (41.7)*
7.7 (7.0)
70.4 (9.5)
7.2 (2.2)*
14 (25.0%),
p=0.09
1.1 (0.71)
120.0 (360.8)*
17 (30.4%)*

14 (25.0%)
12 (21.4%)
17 (30.4%)
13 (23.2%)
30 (53.6%)
18 (34.0%),
p=0.06
11 (20.8%)

0.80 (0.05)
24.4 (4.2)
161.5 (6.7)
63.8 (11.7)
19 (33.9%)
7.8 (2.2)
8.2 (1.3)

43.1 (6.5)*
33.0 (6.8)*

Females
(n=118; 56 progressed)
Progressed

NIH-PA Author Manuscript

Baseline characteristics (means (standard deviation) or sample size (percent)) and coronary artery calcium progression
Costacou et al.
Page 8

NIH-PA Author Manuscript

NIH-PA Author Manuscript

NIH-PA Author Manuscript

Note: Among overweight women, insulin dose per body weight was 0.64 for the stable vs. 0.51 for the progressor group (p=0.05); among overweight men, insulin dose per body weight was 0.67 for
the stable vs. progressor 0.76 group (p=0.09)

The sample size for intensive insulin therapy (3 insulin injections/day) was 214 (109 non-progressors, 105 progressors), for intensive insulin therapy and regular (3 times daily) blood glucose
testing was 209 (106 non-progressors, 103 progressors), for low density lipoprotein cholesterol and triglycerides was 201 (106 non-progressors, 95 progressors) and for Beck depression inventory
it was 213 (108 non-progressors; 105 progressors)

Costacou et al.
Page 9

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

NIH-PA Author Manuscript

Table 2

NIH-PA Author Manuscript

1.44 (1.25-1.66)
1.10 (1.04-1.17)
1.13 (1.01-1.26)
Not selected
Not made available
Not selected
Not selected
Not made available
1.01 (1.003-1.03)
Not selected
1.30 (1.03-1.63)
Not made available
181.685

Square root of baseline CAC score

Diabetes duration (years)
Body mass index (kg/m2)
Hemoglobin A1c (%)
Insulin dose per body weight
Hypertension

1.44 (1.24-1.67)
1.10 (1.04-1.17)
1.24 (1.07-1.44)
Not selected
Not selected
1.44 (1.12-1.82)
Not selected
Not selected
5.82 (0.96-35.22)
Not selected
Not selected
0.68 (0.49-0.93)
179.075

A dichotomous variable (yes/no) is included in Model 1 whereas years of hypertension or angiotensin converting enzyme inhibitor use is used Model 2

Logarithmically transformed

Systolic blood pressure (mmHg)

High density lipoprotein cholesterol (mg/dl)
Non high density lipoprotein cholesterol (mg/dl)
White blood cell count 103/mm2
Albumin excretion rate (g/min)*
Angiotensin converting enzyme inhibitor use
Akaike's Information Criterion

Model 1: Risk factors at first EBT scan

(1996-1998)

Risk factors

Model 2: Updated means of risk factors

(1986-1998)

NIH-PA Author Manuscript

Logistic regression derived odds ratios (95% CI) linking calcium progression to specified predictors (n=222)
Costacou et al.
Page 10

Am J Cardiol. Author manuscript; available in PMC 2008 November 15.

Costacou et al.

Page 11

Table 3

Logistic regression derived odds ratios (95% CI) linking calcium progression to change in specified predictors
(n=182; 91 progressors)

NIH-PA Author Manuscript

Covariates

Final model

Square root of baseline CAC score

Diabetes duration (years)
Body mass index (kg/m2)
Non-high density lipoprotein cholesterol (mg/dl)
Log albumin excretion rate (g/min)
Change in body mass index (kg/m2)
Change in non-high density lipoprotein (mg/dl)
Change in log albumin excretion rate (g/min)
Akaike's Information Criterion

1.34 (1.17-1.53)
1.18 (1.09-1.29)
1.16 (1.02-1.32)
1.02 (1.003-1.03)
1.25 (0.98-1.61)
1.38 (1.10-1.72)
Not selected
Not selected
145.523

NIH-PA Author Manuscript

NIH-PA Author Manuscript
Am J Cardiol. Author manuscript; available in PMC 2008 November 15.