This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Subject and Method: This study consists of 808 266 women, 64.9 years in average consecutive subjects divided in three groups (control, CAD and CHF with CAD) who underwent 64-slice multidetector coronary computed tomographic angiography (64-MDCT). All groups were further subdivided based on 5% interval of LVEF. LVEF and cardiac output index (COI) were measured with their respective means for the groups and subgroups populations. The difference between the groups and within subgroups were then analyzed. Multivariable Cox proportional models were used to assess the risk of CHF and mortality. Follow-up for all-cause mortality was performed over an average of 25.6 ± 10.9 months.
Results: In CHF group, as compared to normal and CAD cases, there was a significant decrease in LVEF (42.7% vs 64.5 and 54.8% for men, 48.2% vs 65.2 and 58.4% for women, P<0.001) with a reduced LVEF in 64.7% of men and 56.3% of women cases, but COI no significant difference ( P>0.05). Using the control groups as a reference, a significant negative association between LVEF with all cause mortality (CAD Î² = -5.1, 95%CI, -8.1 to -2.21 and CHF = -19.1, 95%CI, -24.1 to -14.1, P<0.001), but no association with cardiac output (P>0.05). The annual mortality rates were 0% in controls, 1.1 and 2.0% in CAD patients and 6.8 and 7.2% in CHF groups for men and women respectively.
Conclusion: CTA derived LVEF, but not cardiac output, was an important indicator for diagnosis and prognostic of CHF and predicting mortality in patients with CAD.
Key Words: Left ventricular volume, Left ventricular systolic function, preserved left ventricular ejection fraction, reduced left ventricular ejection fraction, coronary artery diseases, Cardiac output, Congestive heart failure, 64- row Detector Computed Tomography
In most countries around the world, coronary artery disease (CAD) is the leading cause of death and congestive heart failure (CHF) 1. Congestive heart failure is associated significantly with markedly diminished survival2. Global left ventricular systolic dysfunction is proposed as a strong determinant of prognosis and therapeutic management in patients with coronary artery disease2. Although heart failure is generally regarded as a hemodynamic disorder, studies have indicated that there is a poor relation between measures of cardiac performance and the symptoms produced by the disease3. Patients with a very low LVEF may be asymptomatic, whereas patients with preserved LVEF may have severe disability3. Cardiac output is the direct predictor in assessing the left ventricular functional impairment, and a factor for diagnosis and prognosis of CHF and death in patients with CAD 4, 5. The effectiveness of non-invasively measured CO for the diagnosis and prognosis of CHF has been demonstrated by several researchers4, 5. To date, no study has been reported for the role of LVEF and CO measured by CTA in the diagnosis and prognosis of CHF and predicting mortality among CAD cases. Multi-row Detector Computed Tomography (MDCT) with its' high spatial resolution can accurately differentiate the endocardial and epicardial boundaries and provides detailed information of cardiac structures to assess volume of cardiac chambers without assumptions regarding geometry 6, 7. Furthermore, it is an optimal tool to assess the cardiac systolic function and cardiac output precisely. In the present study, we aimed to assess the association of LVEF and CO measured by CTA with the risk of CHF and mortality in CAD cases.
Subject and Method:
Study population: A total 808 consecutive subjects who underwent retrospectively gated cardiac CT angiographic studies for assessment of coronary artery and left ventricular volumetric parameters were enrolled and followed for all-cause mortality with a average of 25.6 ± 10.9 months. Three groups were divided, including the control, CAD and CAD with CHF groups. The control group consisted of 173 patients (55.3±10 years, 57% women). Reasons for referral were atypical chest pain, and cardiovascular risk factors such as family history and dyslipidemia. The exclusion criteria were the following: patients with positive coronary calcium burden; coronary artery disease (>50% of lumen stenosis) by CTA; history of hypertension (â‰¥ 140 mm Hg systolic, â‰¥ 90 mm Hg diastolic); diabetes mellitus; heart disease; lung disease; kidney disease; abnormal electrocardiogram; abnormal nuclear perfusion testing; and abnormal echocardiography. The normal cardiac function and size in 122 of 173 cases were confirmed using echocardiography. 600 patients having significant coronary artery stenosis (>50% stenosis in at least one vessel on CTA) without CHF, constituted the CAD group, including patients with one, two, three vessel diseases and revascularization (coronary artery bypass graft surgery and/or stents) in 151, 164, 65 and 220 respectively. The CHF group constituted of 35 CAD patients documented by medical records. The diagnosis of CHF was suggested based on presenting symptoms and signs of the patients8. These patients underwent additional laboratory testing assessing for systolic and diastolic function and underlying cause of their cardiac dysfunction by their physicians. These patients were referred to our center for cardiac CTA as a part of their work up or follow up of their conditions. In CAD and CHF subgroups, preserved LVEF was defined as â‰¥50%, which was found to be the lower limit for normal LVEF in men and women respectively. The study protocol and consent form were approved by the IRB Committee Board of Los Angeles Biomedical Research Institute at Harbor UCLA Medical Center, Torrance, CA.
CT Protocol and Image Acquisition: Coronary artery calcium screening (CAC): CAC scanning was completed with 64 MDCT (LightSpeed VCT, General Electric Medical System, Milwaukee, WI). The coronary arteries were imaged with 30-40 contiguous 2.5 mm slices during 75% of RR interval using prospective ECG-triggering and a 10 second breath hold. CAC was considered to be present in a coronary artery when a density of >130 Hounsfield units (HU) was detected in â‰¥ 3 contiguous pixels (>1.2 mm²) and quantified using the previously described Agatston scoring method9.
Cardiac CTA: Beta blockers were administered for 299 of 808 patients with a > 65 beats per minute (bpm) of heart rate. A test IV bolus of 15 ml of contrast agent was followed by 20 ml of normal saline flush at a rate of 4.5 ml/s. Using a dual-head power injector (Stellant, Medrad, Indianola, PA), a retrospective ECG gated cardiac CT angiography was performed with a tri-phasic consecutive injection sequence beginning with 40 ml nonionic IV contrast material (Iopamidol 370; Bracco Diagnostics, Plainsboro, NJ) injected at a rate of 5.0 ml followed by 40 ml of a mixture of 60% contrast and normal saline and ended with a 40-ml flush of normal saline. Contrast was injected through an 18- to 20-gauge angiocatheter in the antecubital vein. Mean heart rate during the scan was 59±6 bpm.
Data acquisition: A 64- Multi-detector Computed Tomography scanner (Lightspeed VCT, General Electric Healthcare Technologies, Milwaukee, WI) was used for all patients. Imaging was started 1 inch above the left main ostium and continued to 1 inch below the bottom of the heart. The following imaging and reconstruction parameters were applied: data acquisition collimation 0.625 mmÃ-40 = 2.5 cm; 120 kVp; 220-670 mAs; pitch 0.18-0.24 (depending on heart rate); rotation time 0.35 s; matrix 512Ã-512, pixel size 0.39 mm² and mean effective radiation dose of 9.4±1.1 mSv (8.0-11.5 mSv). ECG-triggered dose modulation was applied in each case with 400-600 mA in 60-80% R-R interval and 250-350 mA for the rest of the cardiac cycle (81% to 59% of the next cycle). Image series were created with 12 images per level, including timepoints of 5% to 95% (by 10% increments) and 39 and 99% of the R-R interval. Coronary vessels were reviewed (AW Volume Shareâ„¢, GE Medical Systems, Milwaukee, WI) and volume renderings and curved multi-planar reformations were performed. Each vessel was assessed as normal (no stenosis), non-obstructive CAD (luminal stenosis 1-49%), and obstructive CAD (luminal stenosis >50%). Vessels 1.5 mm in diameter or larger were assessed. Two experienced cardiologists blinded to the clinical data assessed the coronary arteries separately.
LVV and LVEF measurement (fig 1): All LV segmentations were completed using cine images. The end systolic (minimum) and end diastolic (maximum) axial images were chosen and manual segmentation analysis was done at these two phases. The segmentation analysis included the LV endocardial boundary at end-systole (LVVes, cavity) and the epicardial boundary at both end-diastole and end-systole (TLVVed and TLVVes), latter including the LV cavity plus LV mass. The trabecular and papillary muscles were easily separated from the LV cavity by using this segmentation method 6, 7. In each study, we manually traced 10-15 slice levels with axial images, and the remaining slices were traced by the workstation automatically. The left ventricular endocardial volume at end diastole (LVVed), stoke volume (SV), LVEF and CO can be computed by the following formula: SV = TLVVed - TLVVes, LVVed = SV + LVVes, LVEF = SV ÷ LVVed Ã- 100% and CO index(COI) = SV Ã- Heart rate (BPM) ÷ body surface area (liter/min/m²) , shown on Figure 1.
Groups: The three groups, control, CAD and CHF, were divided into subgroups based on the LVEF with 5% intervals (Table 2).
Statistical analyses: Descriptive statistics are presented as percentages for categorical data and means and standard deviations for continuous variables. Student's t test and ANOVA test were applied to continuous and categorical variables, respectively. A logistic regression analysis was used to estimate the association between the reduced LVEF and risk of CHF. A Cox proportional hazards model was used to calculate the risk of all-cause mortality with LVEF, CO and 95% confidence intervals. All models were adjusted for potential epidemiological confounders and conventional cardiovascular risk factors, including age, gender, race/ethnicity, BMI, coronary artery calcium score, smoking history, diabetes, dyslipidemia, hypertension and family history of CAD. Comparisons of CACs was calculated by the nonparametric Wilcoxon statistic. The mean value and standard deviation of LVEF and CO was calculated in all groups. The bottom 5 percentile value of LVEF and COI were defined as a reference value in control groups. The receiver operating characteristic (ROC) curves for progression of LVEF and CO predicting CHF among patients with CAD was completed. Areas under the ROC curve (AUC) and differences between LVEF and CO curves were assessed. Mortality surveillance follow-up time was calculated as the time from baseline to death or censoring and Kaplan-Meier survival curves were plotted to estimate time all-cause mortality for the LVEF and CO subsets accordingly and compared using the log rank test. All data management and analysis were performed with SAS 9.3 (SAS Institute, Inc., Cary, NC).
A total of 600 CAD cases (151 women), 35 CHF cases (13 women) and 179 controls (102 women) were studied and 21 cases had died during follow-up. In control group, the mean values of LVEF are 64.5 ± 7.4 and 65.2 ± 7.8%, and COI are 2.0 ± 0.5 and 2.2 ± 0.5L/min/m², with a lower 5 percentile value of 50% for LVEF and 1.0 and 1.2 5L/min/m² for COI in men and women respectively. In CAD group, the LVEF was significantly decreased (54.8±15.7% in men, 59.4±15.8% in women, P<0.001), but similar COI (1.9±0.5 and 2.1±0.5L/min/m² in men and women respectively, P>0.05) when compared to the control group. Overall, the COI only significantly decreased in patients with the lowest LV ejection fractions (<25%).
In CHF group, as compared to CAD groups, there was a significant decrease in LVEF (42.7% for men and 48.2 for women, P<0.001) with a reduced LVEF in 64.7% of men and 56.3% of women cases, but similar COI (1.88 in men and 2.12 L/min/m² in women, P>0.05).
The association of the reduction in LVEF and CO with congestive heart failure can be seen in Table 2. A lower LVEF was found in most subgroups of CHF cases, however, CO was not significant worsened, though there are small detectable fluctuations in CAD and CHF groups separately. Data from men as well as women showed the same pattern.
When ROC analysis (Figure 2.) was used to illustrate the value of LVEF in predicting mortality, the area under the curve (AUC) for LVEF is 0.77 (95% CI: 0.66, 0.87), and was significantly larger than AUC for CO (AUC: 0.54, 95%CI: 0.40, 0.67). The best discriminative value of LVEF was 41% and it will provide 71% sensitivity and 47% specificity separately to predict mortality.
Both multivariate regression and logistic regression analyses were conducted to examine the relationship between the LVEF and CO with mortality (Table 3 & 4). In regression model, the LVEF had significant decreased in death cases among CAD and CHF (-5.1%, 95%CI -8.1%,-2.2% in CAD; -19.1 24.1%,-14.1% in CHF, p<.001) when compared with control group. However, for CO, similar trend with LVEF were not observed (-0.1, p>0.05) in both groups (Table 3). In logistic model, LVEF, as binary variable, was divided into preserved and reduced LVEF (50% as a cut of point). Again, there is significant association between reduced LVEF and mortality. The reduced LVEF cases had a 3.71 folds higher risk of mortality (95% CI 1.5, 9.1, p=0.04), after adjusting all factors described above, the risk of mortality even was higher (OR: 5.0, 95%CI 1.8, 13.8, P=0.002), but COI did not show similar pattern (Table 4).
Finally, Kaplan-Meier survival plots of control, CAD and CHF are displayed (Fig 3). There is significant decreasing survivor rate in subjects with CAD and CHF disease, especially in patients with congenital heart failure (P<.001).
In the United States, CAD remains the most common identifiable process underlying heart failure. In patients with CAD, segmental ischemia can induce the progression of cardiac remodeling. The alteration process in ventricular geometry following an ischemic myocardial event is complex, but for simplicity of discussion, can be divided into two different phases: infarction expansion and ventricular remodeling. Infarction expansion is an increase in LV size as a result of acute stretching, thinning, and dilation of the injured segment 10. In contrast to the acute and localized process of infarction expansion, remodeling refers to chronic global change in the LV geometry following regional injury 11, 12. In advanced stages, the changes can induce congestive heart failure. This remodeling process can directly reflect the impairment of cardiac function, measured by LVEF or CO. The prognosis for patients with established congestive heart failure is poor, with a high annual mortality rate in patients with class III and IV heart failure in NYHA 13. Both asymptomatic systolic and diastolic LV dysfunction have been strongly associated with development of clinical heart failure. In the Framingham study, patients with asymptomatic LV systolic dysfunction were at higher risk of developing heart failure than those with normal LV systolic function and was a stronger predictor of mortality than the number of diseased vessels 14. Therefore, monitoring these cardiac structures and functional changes is very important to strategize the treatment and management in patients with CAD. Currently, cardiac CT has emerged as an optimal tool to assess the cardiac volume, mass, systolic function and cardiac output with a single scan.
However, little data is available on the normal values for LVEF measured by CTA or the independent prognostic ability in cases of CAD and CHF. In our study, the mean value of LVEF in controls is similar to that published earlier by MRI and cardiac CT (63.8-67%) 15-17. The lower 5th percentile is 50% for both genders. Using LVEF of 50% as a reference value of the low border limit, 32 and 23% in CAD cases, and 65 and 56% in CHF had reduced cardiac function in men and women respectively. Comparing CAD cases, a significant lower LVEF was fond in patients with CHF (P<0.001). Therefore, the LVEF was an important indicator for diagnosis and prognosis of CHF in patients with CAD.
Cardiac output is a parameter for directly reflecting the myocardial contractility. The role of non-invasively measured CO in the diagnosis and prognosis of CHF patients has been previously demonstrated (4,5). To our knowledge, no study in the literature assessed the value of LVEF or COI for the diagnosis of CHF in patients with CAD by use of cardiac CT. In this study, cardiac output was maintained until the cohort with the lowest ejection fractions (LVEF<25%). This demonstrates that outpatients not in acute CHF have the ability to compensate their cardiac output despite a drop in ejection fraction. Even when LVEF <25%, CO was preserved in 67% of CAD cases. The apparent discordance between LVEF and CO is not well understood but may be explained in part by alterations in ventricular distensibility, valvular regurgitation, pericardial restraint, cardiac rhythm disorders, conduction abnormalities and right ventricular function18. In patients with above conditions, there is significant limitation to assess the cardiac systolic function and output by volumetric measurement with any images. The low heart rate during CTA may be another reason for inducing the discordance that can decrease the CO within study groups. Therefore, CO has limited utility in the prognosis of congestive heart failure and mortality in CAD patients undergoing CTA.
A total of 21 (6 women) in both CAD and CHF groups died during study follow-up interval. Of those who died in follow up, there were 71.4% of patients with a reduced LVEF, but the COI was preserved (1.96 L/min/m², P >0.05) compared to controls. The all-cause mortality of both CAD and CHF groups was increased in 5-fold among patients with a reduced LVEF in comparing preserved LVEF. This demonstrated that the CT-derived LVEF measures carry prognostic importance, concordant with other tests, and LVEF adds incremental value to the diagnostic utility of cardiac CT.
In current, CTA was used widely as an important noninvasive diagnosis modality for coronary artery disease. The LV volumetric assessment (including LVEF) without additional radiation, have independent prognostic value for patients with coronary disease improved by prior studies19, 20 and our current study. Although increased use of prospective triggering will limit the availability of these measures on cardiac CT. Selective use of retrospective-gating CTA may be beneficial in patients with suspected CAD and cardiac dysfunction, documented CAD and cardiac dysfunction or CHF for diagnosis, prognosis, longitudinal monitor and management of these patients in clinical practice.
In summary, this study evaluated cardiac systolic function and output using cardiac CT angiography, and the association of left ventricular ejection fraction and cardiac output with the risk of congestive heart failure and mortality in CAD cases. We concluded that CTA derived LVEF, but not cardiac output, was an important indicator for diagnosis and prognostic of CHF and predicting mortality in CAD cases using cardiac CT angiography.
There are two important limitations that exist in our current study, which are the longer image acquisition times of single source CT (175ms/image) and lower heart rate (59 BPM) of patients with this CTA protocol. The lower temporal resolution image can under-estimate the end diastolic and over-estimate end-systolic chamber volumes6, 21. In patient with a low heart rate, the cardiac output can decrease significantly. Therefore, limitations of current CTA can affect the evaluation of cardiac function and output significantly. For those reasons, the control group should be used as a reference to estimate the LVEF and CO changes in CAD and CHF groups in the current study.