DOI: 10.1148/radiol.2401050948
(Radiology 2006;240:110-115.)
© RSNA, 2006
Assessment of Cardiovascular Risk Status at CT Colonography1
Jesse A. Davila, MD,
C. Daniel Johnson, MD,
Thomas R. Behrenbeck, MD, PhD,
Tanya L. Hoskin, MS and
William S. Harmsen, MS
1 From the Departments of Radiology (J.A.D., C.D.J.), Cardiology (T.R.B.), and Biostatistics (T.L.H., W.S.H.), Mayo Clinic, 200 First St SW, Rochester, MN 55905. From the 2004 RSNA Annual Meeting. Received June 6, 2005; revision requested July 21; revision received August 10; accepted September 8; final version accepted September 30.
Address correspondence to C.D.J.
 |
ABSTRACT
|
|---|
Purpose: To retrospectively determine whether calcium scores of the abdominal aorta obtained during computed tomographic (CT) colonography relate to Framingham risk factors and clinical cardiac events.
Materials and Methods: The institutional review board approved the current HIPAA-compliant retrospective study and waived informed consent. Between 1995 and 1998, 480 patients underwent CT colonography; 467 patients were available for assessment. Calcium scores with a threshold attenuation value of 130 HU or greater were recorded for abdominal aorta (suprarenal, infrarenal, and aortic bifurcation regions and total). Patient histories were abstracted for established cardiac risk factors and subsequent cardiac events. Associations between calcium measurements and binary risk factors were assessed with Wilcoxon rank sum test; those with continuous risk factors, with Spearman rank correlation coefficient; and those with combined end points, with Cox proportional hazards model.
Results: Follow-up data were available for 467 patients with median age of 65 years (range, 34–83 years); 59% (275 of 467) were men. Nine patients had cardiac events subsequent to CT colonography. Results of proportional hazards regression analysis revealed a significant association between myocardial infarction or cardiac event–related death and calcium scores of the aortic bifurcation that exceeded 895, the value for the 75th percentile for this calcium variable (P < .01). Associations with established cardiac risk factors for all four calcium scores were significant (P < .05). Spearman rank correlation coefficients for associations between total calcium score and patient characteristics of age, number of pack-years of smoking, and systolic blood pressure were 0.51, 0.43, and 0.29, respectively (P < .001 for all).
Conclusion: Aortic calcification scores at CT colonography are significantly associated with established cardiac risk factors and cardiac-related events. This screening information can be obtained without additional scanning or risk to the patient.
© RSNA, 2006
|
INTRODUCTION
|
|---|
Computed tomographic (CT) colonography is a noninvasive examination of the entire abdomen and pelvis with the primary intent to evaluate the colorectum. Modern CT scanners can collect all of the data necessary in 10–20 seconds per acquisition (1). Besides assessment of the colorectum, evaluation of the remaining soft tissues also can be performed. Solid-organ contrast is reduced because of the low-dose technique. Despite these limitations for assessment of extracolonic organs, researchers in some studies have demonstrated that clinically important noncolonic abnormalities, including abdominal aortic aneurysms, frequently are detected at CT colonography (2–7). In addition, calcification of the abdominal aorta and its branches also can be readily identified. Although abdominal aortic calcification is a marker of atherosclerosis, to our knowledge no investigators in a formal study have evaluated the potential of this finding in the assessment of the risk for coronary arterial disease (1,8,9).
Colorectal cancer screening is recommended to begin at age 50 years for average-risk individuals (10). Since colonic cancer and heart disease are common in older Americans, screening for both colorectal cancer and coronary arterial disease risk at the same examination would be desirable, particularly if it could be performed without subjecting the patient to additional radiation and at minimally increased cost, by using a noninvasive procedure (2).
Researchers in studies that include data from the Framingham study and similar population-based studies (11) report on the risk factors that contribute to coronary arterial disease rather than on the atherosclerotic process (8,11,12). Framingham scores include factors such as age, sex, total cholesterol level, high-density lipoprotein cholesterol level, blood pressure value, presence of diabetes mellitus, and history of smoking (13). We hypothesized that the amount of abdominal aortic calcification, measured during CT colonographic screening, is associated with subsequent cardiac events (myocardial infarction or cardiac event–related death). Thus, the purpose of our study was to retrospectively determine whether calcium scores of the abdominal aorta obtained during CT colonography relate to Framingham risk factors and clinical cardiac events.
|
MATERIALS AND METHODS
|
|---|
Patients
CT colonographic examinations had been performed in 480 consecutive patients between 1995 and 1998. These examination dates were chosen to enable the collection of at least 5 years of follow-up data for all patients. Data from examinations in 467 patients were available for assessment. In 13 patients, the archival data could not be retrieved. These patients had been recruited for CT colonographic research examinations, with institutional review board approval and informed consent, because they had a known colorectal lesion or were highly suspected of having one or were eligible for colorectal cancer screening. Our institutional review board approved the current Health Insurance Portability and Accountability–compliant retrospective study, and informed consent was waived.
CT Colonographic Examination
Patient preparation for these examinations has been reported previously (1). The CT colonographic examinations were performed with a scanner (HiSpeed Advantage/RP, HiSpeed CTi, and LightSpeed/QXi; GE Medical Systems, Milwaukee, Wis) and a section thickness of 5 mm, reconstruction interval of 3 mm, 70 mAs, and 120 kVp. Since these examinations were primarily performed with a single-section scanner, three breath holds were nearly always required to acquire data of the entire colorectum. Three-centimeter overlapping sections were obtained between breath holds to ensure complete anatomic coverage. To avoid artificial inflation of the area of calcification, any overlapping data were removed before assigning scores for calcification.
Calcification Score Assignment
Assignment of calcification scores of the abdominal aorta was performed by using a workstation (Vitrea II; Vital Images, Plymouth, MN) and commercially available software (Vitrea, version 2.0.1; Vital Images). Two technologists performed selection of three aortic levels, as detailed later. Each registered technologist had at least 10 years of CT experience, at least 1 year of experience in the performance of three-dimensional image-processing tasks, and had received a full day of software-specific training from the vendor-application specialist. Assignment of aortic scores was carefully performed to exclude the spine and any nearby hypertrophic changes from the areas that were assessed. Average time for assignment of scores was approximately 15 minutes. None of the training cases were from the study set. The supine image data set was used, and assignment of scores for the abdominal aorta was performed from 1 cm above the origin of the celiac axis to 1 cm below the iliac bifurcation by using the software. Proximal major vessels (proximal 1 cm of the celiac axis, superior mesenteric artery, and left and right main renal arteries) were included. The aorta was divided into three sections: suprarenal section (1 cm above the celiac axis to 1 cm below the renal arteries), infrarenal section (1 cm below the renal arteries to 1 cm above the aortic bifurcation), and aortic bifurcation (1 cm above to 1 cm below the aortic bifurcation and extending into the common iliac arteries). Patients with metallic artifacts because of surgical clips or other metallic hardware were excluded from the study because their scores could have been artificially elevated (n = 13).
By using a threshold attenuation value of 130 HU, measurements of abdominal aortic calcification (suprarenal section, infrarenal section, aortic bifurcation, and all three values combined) were obtained. The 130-HU threshold attenuation value was chosen because of its common use in programs for assignment of scores for coronary arterial calcification and because it was recommended by the software vendor. Calcification for the celiac axis, superior mesenteric artery, and both main renal arteries was manually defined in the data set for a distance of 1 cm beyond its origin by using a graphic overlay editor. If patients had multiple renal arteries, the technologist would identify the dominant artery and use that artery for analysis. Accessory renal arteries and inferior mesenteric arteries were ignored. Celiac and superior mesenteric and main renal arterial scores were included in the suprarenal section score. The total score was the sum of the three individual scores. Agatston scores (14) were reported within the defined volumes.
Risk Stratification
Patients at our medical center commonly undergo colorectal cancer screening as part of a complete medical evaluation. This evaluation usually includes cardiovascular risk assessment. A team of dedicated registered nurses within the cardiovascular research area reviewed the medical records of the study patients and collected all historical data, including the occurrence of a cardiac event subsequent to the CT colonographic examination. A cardiac event was defined as a myocardial infarction or cardiac event–related death. In addition, patient data in regard to the presence of cardiac risk factors were collected. Blood pressure and cholesterol level measurements within 1 year of the date of CT colonography were recorded. Patients were classified as smokers if they were current or past smokers at the time of CT colonography. The Framingham score includes age, sex, total cholesterol level, high-density lipoprotein cholesterol level, blood pressure value, presence of diabetes mellitus, and history of smoking (8,11,13,15). We used this established risk stratification model to create a single score to quantify the patient's cardiac risk for comparison with the CT-based aortic calcification scores.
Statistical Analysis
Associations between calcium measurements and binary cardiac risk factors, such as sex, history of smoking, use of cholesterol-lowering medication, use of blood pressure–lowering medication, presence of diabetes mellitus, positive stress test, body mass index of 25 kg/m2 or greater, and family history of coronary arterial disease and diabetes mellitus, were assessed by using the Wilcoxon rank sum test. Associations between calcium measurements and continuous cardiac risk factors, such as systolic and diastolic blood pressure values, number of pack-years of smoking, triglyceride levels, high-density lipoprotein cholesterol level, low-density lipoprotein cholesterol level, body mass index, and total cholesterol level, were assessed by using the Spearman rank correlation coefficient.
The association between calcium variables and the combined end point of myocardial infarction (occurrence subsequent to CT colonography) or cardiac event–related death was estimated by using Cox proportional hazards models. For each calcium variable, patients were classified on the basis of whether their score was at or below the 75th percentile or above the 75th percentile of the sample distribution. Hazard ratios and 95% confidence intervals were estimated for each calcium variable. P values less than .05 were considered to indicate a statistically significant difference. Analysis was performed by using software (SAS, version 8.2, 1999–2001; SAS Institute, Cary, NC).
|
RESULTS
|
|---|
Patient Characteristics
Four hundred sixty-seven patients with the opportunity for at least 5 years of clinical follow-up comprised the study group. The median age at the time of the CT colonographic examination was 65 years (range, 34–83 years), and 275 (59%) men and 192 (41%) women were included. The number of smokers was 282 (60%), and the mean number of pack-years of smoking was 27 (range, 0.5–102 pack-years). Diabetes mellitus was present in 42 (9%) patients. The mean lipid profiles were as follows: total cholesterol level, 213 mg/dL (5.51 mmol/L), with a range of 115–371 mg/dL (6.00–9.60 mmol/L) and a desirable range of 100–200 mg/dL (2.59–5.17 mmol/L); triglyceride levels, 156 mg/dL, with a range of 35–1135 mg/dL and a normal range of 10–150 mg/dL; high-density lipoprotein cholesterol value, 54 mg/dL (1.40 mmol/L), with a range of 8–108 mg/dL (0.21–2.79 mmol/L) and a normal range of 40–60 mg/dL (1.03–1.55 mmol/L); and low-density lipoprotein cholesterol value, 129 mg/dL (3.33 mmol/L), with a range of 20–230 mg/dL (0.52–5.95 mmol/L) and a normal level of less than 129 mg/dL (<3.33 mmol/L). Only 79 (17%) patients were receiving cholesterol-lowering medication. The mean body mass index was 29 kg/m2 (range, 17–50 kg/m2). One hundred seventy-eight (38%) patients were receiving blood pressure–lowering medication. The mean systolic blood pressure was 140 mm Hg (range, 88–230 mm Hg); the mean diastolic blood pressure was 81 mm Hg (range, 45–122 mm Hg).
Calcium Scores
The median calcium scores within the aortic areas of interest, with the interquartile ranges from the 25th to 75th percentiles, were as follows: suprarenal section, 47 (interquartile range, 0–388); infrarenal section, 554 (interquartile range, 67–2049); and aortic bifurcation, 171 (interquartile range, 0–895). The median total calcium score, with the interquartile range from the 25th to 75th percentiles, was 977 (interquartile range, 137–3558). Zero calcium values were obtained in 60 (13%) of 467 patients.
Cardiac Events
There were nine cardiac-related events (eight myocardial infarctions and one cardiac event–related death) documented in the clinical record subsequent to CT colonography. The mean and median follow-up times after CT colonography were 3.1 and 3.4 years, respectively, with a range from 0 days to approximately 6 years. The nine cardiac events occurred a minimum of 7 days and a maximum of 3 years subsequent to CT colonography. Results of the Cox proportional hazards regression analysis indicated a significant association between a myocardial infarction or cardiac event–related death and calcium scores of the aortic bifurcation exceeding 895, the 75th percentile value for this calcium variable (P = .004; hazard ratio = 10.3; 95% confidence interval: 2.1, 49.7) (Table 1). Of the nine patients who had events, seven had calcium scores of the aortic bifurcation that were above the 75th percentile. The same association, evaluated for the calcium scores of the suprarenal and infrarenal sections and of the total for all regions above the 75th percentile of the respective distribution, were also statistically significant (P = .01, P = .01, and P = .004, respectively). The hazard ratios for each of the calcium scores in the suprarenal and infrarenal sections and for the total of all sections were 6.1 (95% confidence interval: 1.5, 24.4), 6.0 (95% confidence interval: 1.5, 24.2), and 10.4 (95% confidence interval: 2.2, 50.0), respectively, for patients whose values were above the 75th percentile versus those whose values were at or below the 75th percentile (Table 1). The associations between calcium scores and cardiac events remained statistically significant (P = .004) after adjustment for age and sex.
View this table:
[in this window]
[in a new window]
|
Table 1. Univariate Associations between Calcium Score and Myocardial Infarction or Cardiac Event–related Death with Cox Proportional Hazards Models
|
|
Associations between Calcium Scores and Risk Factors
Associations with established cardiac risk factors for all four calcium scores (suprarenal, infrarenal, and aortic bifurcation sections and the total of all three) were significant. The Spearman rank correlation coefficients for associations between the total calcium score and the variables of age, number of pack-years of smoking, systolic blood pressure value, and triglyceride levels were 0.51, 0.43, 0.29, and 0.20, respectively (Table 2). These associations were statistically significant (P < .001). The total calcium score was significantly associated with binary risk factors such as male sex, smoking status, and presence of diabetes mellitus (P < .001, for all). Associations between binary risk factors and total calcium score are summarized in Table 3. The calcium scores for each area of the aorta had associations with cardiac risk factors that were similar to those reported for the total calcium score. All four calcium scores showed a significant association (P < .001) with the calculated Framingham risk score (available in 390 patients). The Spearman rank correlation coefficients for associations between the Framingham score and the suprarenal, infrarenal, and aortic bifurcation sections and the total calcium scores were 0.43, 0.49, 0.40, and 0.48, respectively.
View this table:
[in this window]
[in a new window]
|
Table 2. Spearman Rank Correlation Coefficient for Association between Total Calcium Score and Established Quantitative Cardiac Risk Factors
|
|
|
DISCUSSION
|
|---|
Our study findings indicate that abdominal aortic calcification can be assigned a score by using low-dose CT colonographic data sets. These calcification scores have a strong correlation with accepted cardiovascular risk factors and subsequent risk of a cardiac event within 5 years.
Calcification scores within the aorta (suprarenal, infrarenal, aortic bifurcation regions and the total of all three) that were higher than the 75th percentile each demonstrated a statistically significant association with cardiac events. The strongest association between cardiac events and calcium score was evidenced in (a) the association between cardiac events and the calcium score of the aortic bifurcation and (b) the association between cardiac events and the total calcium score. Patients with a calcium score of the aortic bifurcation or a total calcium score greater than the 75th percentile (from our sample) had a hazard ratio of more than 10 when they were compared with patients with calcium values at or below the 75th percentile (P = .004). For the other two areas in which values were measured, P = .01 and the hazard ratios were approximately 6. This finding indicates the potential for prediction of a patient's risk of having a subsequent cardiac event with a given aortic calcification score.
Patients with known cardiovascular risk factors also demonstrated a statistically significant association with elevated calcium scores at CT colonography. Of the ten known cardiac risk factors assessed (11), six demonstrated statistically significant associations with the total calcium score in univariate analyses. For these six risk factors, the P value was less than .001. They included history of smoking (former or current smoker), male sex, use of cholesterol-lowering medication, blood pressure–lowering medication, presence of diabetes mellitus, and a history of a positive stress test. Age was also significantly associated with the calcium score. Total cholesterol level, low-density lipoprotein cholesterol level, body mass index, and diastolic blood pressure value did not have a significant association with the calcium score.
The Framingham score, which comprises multiple cardiovascular risk factors, was compared with calcium scores. Each calcium score had a statistically significant association with the Framingham score. The Framingham score had a strong correlation with the total calcium score. Our study findings indicated that, as the Framingham score increased, so too did the total calcium score. Just as was indicated by results in studies about the Framingham score—that it is known to be useful for prediction of the risk of coronary heart disease—so too did our study findings indicate that higher calcium scores were associated with increased risk of a cardiac event.
An example of the utility of the detection of asymptomatic patients who have an increased risk of cardiovascular disease would be the identification of those who need further evaluation: That is, patients who were subsequently identified as having elevated cholesterol values could be treated with lipid-lowering agents because this is a proved method of decreasing cardiovascular risk (12,16–19).
In our sample of patients, we defined the threshold value of a significant amount of calcium to be a value at or above the 75th percentile. Patients with a significant amount of calcium according to this definition had a statistically significant increased risk of a cardiac event when they were compared with those who had calcium scores at or below the 75th percentile.
Our patient data were retrospectively collected on the basis of whether a patient had a colonic abnormality and then underwent CT colonography for research purposes. One of the limitations of our study was that our population was not prospective, randomized, or controlled. The patients in our population may or may not have been experiencing cardiac symptoms before the examination. In addition, the findings are based on a 5-year follow-up. A longer follow-up would have been helpful. Although CT colonography was performed more than 5 years prior to data collection in some patients, they did not have current follow-up in the medical record. Ideally, these patients should have been contacted for additional information in regard to cardiac events that occurred subsequent to CT colonography. Patients with intraabdominal surgical clips and metallic artifacts were not assessed because scores may not have been accurate for them. Although scores were not normalized for patient height, scores of the aortic bifurcation (1 cm above and below the aortic bifurcation) were independent of height.
In conclusion, aortic calcification scores at CT colonography are significantly associated with established cardiac risk factors and cardiac-related events. This screening information can be obtained without additional scanning or risk to the patient. The potential effect of combining colorectal screening with cost-effective cardiovascular screening is substantial and, in our opinion, merits further study.
|
ADVANCE IN KNOWLEDGE
|
|---|
- Calcification scores of the abdominal aorta at CT colonography are significantly associated with established cardiac risk factors and subsequent cardiovascular events.
|
ACKNOWLEDGMENTS
|
|---|
Colleen Braun, RT, and Tom Klinkhammer, RT, performed the aortic calcification scoring. Debora Shreve prepared the manuscript.
|
FOOTNOTES
|
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, J.A.D., C.D.J.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, J.A.D., C.D.J., T.R.B.; clinical studies, J.A.D., C.D.J., T.R.B.; statistical analysis, J.A.D., T.L.H., W.S.H.; and manuscript editing, all authors
|
References
|
|---|
- Johnson CD, Hara AK, Reed JE. Computed tomographic colonography (virtual colonoscopy): a new method for detecting colorectal neoplasms. Endoscopy 1997;29:454–461.[Medline]
- Hara AK, Johnson CD, MacCarty RL, Welch TJ. Incidental extracolonic findings at CT colonography. Radiology 2000;215:353–357.[Abstract/Free Full Text]
- Xiong T, Richardson M, Woodroffe R, Halligan S, Morton D, Lilford RJ. Incidental lesions found on CT colonography: their nature and frequency. Br J Radiol 2005;78:22–29.[Abstract/Free Full Text]
- Hellstrom M, Svensson MH, Lasson A. Extracolonic and incidental findings on CT colonography (virtual colonoscopy). AJR Am J Roentgenol 2004;182:631–638.[Abstract/Free Full Text]
- Ginnerup Pedersen B, Rosenkilde M, Christiansen TE, Laurberg S. Extracolonic findings at computed tomography colonography are a challenge. Gut 2003;52(12):1744–1747.[Abstract/Free Full Text]
- Gluecker TM, Johnson CD, Wilson LA, et al. Extracolonic findings at CT colonography: evaluation of prevalence and cost in a screening population. Gastroenterology 2003;124:911–916.[CrossRef][Medline]
- Yee J, Kumar NN, Godara S, et al. Extracolonic abnormalities discovered incidentally at CT colonography in a male population. Radiology 2005;236:519–526.[Abstract/Free Full Text]
- Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications—a statement for health professionals from the American Heart Association, Writing Group. Circulation 1996;94:1175–1192.[Free Full Text]
- Johnson CD, Dachman AH. CT colonography: the next colon screening examination. Radiology 2000;216:331–341.[Abstract/Free Full Text]
- Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the early detection of cancer. CA Cancer J Clin 2003;53:27–43.[Abstract/Free Full Text]
- Grundy SM, Pasternak R, Greenland P, Smith SJ, Fuster V. Assessment of cardiovascular risk by use of multiple-risk-factor assessment equations: a statement for healthcare professionals from the American Heart Association and the American College of Cardiology. Circulation 1999;100:1481–1492.[Free Full Text]
- Blankenhorn DH, Hodis HN. George Lyman Duff Memorial Lecture: arterial imaging and atherosclerosis reversal. Arterioscler Thromb 1994;14:177–192.[Abstract/Free Full Text]
- Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837–1847.[Abstract/Free Full Text]
- Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte MJ, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827–832.[Abstract]
- Mahoney LT, Burns TL, Stanford W, et al. Coronary risk factors measured in childhood and young adult life are associated with coronary artery calcification in young adults: the Muscatine study. J Am Coll Cardiol 1996;27:277–284.[Abstract]
- Rossouw JE, Lewis B, Rifkind BM. The value of lowering cholesterol after myocardial infarction. N Engl J Med 1990;323:1112–1119.[Medline]
- Hulley SB, Walsh JM, Newman TB. Health policy on blood cholesterol: time to change directions [editorial]. Circulation 1992;86:1026–1029.[Free Full Text]
- Brown BG, Zhao XQ, Sacco DE, Albers JJ. Lipid lowering and plaque regression. New insights into prevention of plaque disruption and clinical events in coronary disease. Circulation 1993;87:1781–1791.
- Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study. Lancet 1994;344:1383–1389.[CrossRef][Medline]
TOP
ABSTRACT
INTRODUCTION
