Current guidelines recommend a coronary evaluation before valvular heart surgery (VHS). Diagnostic coronary angiography is recommended in patients with known coronary artery disease (CAD) and those with high pretest probability of CAD. In patients with low or intermediate pretest probability of CAD, the guidelines recommend coronary computed tomographic angiography. However, there are no tools available to objectively assess a patient’s risk for obstructive CAD before VHS. To address this deficit, 5,360 patients without histories of CAD who underwent diagnostic coronary angiography as part of preoperative evaluation for VHS were identified. Obstructive CAD was defined as ≥50% stenosis in ≥1 artery. Of the patients assessed, 1,035 (19.3%) were found to have obstructive CAD. Through multivariate analysis, age, gender, diabetes, renal dysfunction, hyperlipidemia, and a family history of premature CAD were found to be associated with the presence of obstructive CAD (p <0.001 for all). After adjustment, the specific dysfunctional valve was not associated with the presence of obstructive CAD. Patients were then randomly split into derivation and validation cohorts. Within the derivation cohort, using only age, gender, and the presence or absence of risk factors, a model was constructed to predict the risk for obstructive CAD (C statistic 0.766, 95% confidence interval 0.750 to 0.783). The risk prediction model performed well within the validation cohort (C statistic 0.767, 95% confidence interval 0.751 to 0.784, optimism 0.004). The bias-corrected C statistic for the model was 0.765 (95% confidence interval 0.748 to 0.782). In conclusion, this novel risk prediction tool can be used to objectively risk-stratify patients who undergo preoperative evaluation before VHS and to facilitate appropriate triage to computed tomographic angiography or diagnostic coronary angiography.
Valvular heart disease is a common indication for cardiac surgery. Preoperative diagnostic coronary angiography (DCA) is routinely performed in patients who undergo valvular heart surgery (VHS). This practice is driven by the negative impact of untreated obstructive coronary artery disease (CAD) on perioperative and long-term outcomes. The current American College of Cardiology Foundation and American Heart Association guidelines recommend that DCA be performed on all patients who undergo nonemergent VHS if there are symptoms of angina, left ventricular dysfunction, evidence of ischemia, or functional mitral regurgitation. DCA is also recommended for patients with known histories of CAD or high pretest probability of CAD. In contrast, for patients with low or intermediate pretest probability of CAD, coronary computed tomographic angiography (CTA) is recommended. Yet there are no criteria defining low, intermediate, and high risk for CAD. Additionally, there are no tools to objectively assess a patient’s risk for obstructive CAD before VHS. This study was focused on patients who underwent DCA before VHS, without known histories of CAD, and with low or intermediate pretest probability of CAD. The aims were to determine the prevalence and risk factors associated with obstructive CAD in this population and to create a risk prediction nomogram to objectively assess preoperative risk for obstructive CAD. Ultimately, this information is intended to facilitate appropriate triage to either a coronary CTA or DCA before VHS.
Methods
Patients were identified from the Coronary Catheterization Database at the Cleveland Clinic. Patients were included if they underwent preoperative DCA for a primary indication of valvular heart disease from February 1, 2004, to October 1, 2013, underwent VHS within 6 months of DCA, and were ≥18 years of age. Patients were excluded if they had known CAD, known ischemia by stress testing, or congenital heart disease. Patients were also excluded if they underwent DCA for a primary indication of angina. However, because angina can be a symptom of severe valvular disease, patients were included if they underwent DCA for a primary indication of valvular heart disease and had symptoms of angina, but not if they presented with ischemic symptoms and had valvular disease (a distinction defined in the Coronary Catheterization Database).
The presence of obstructive CAD was defined as ≥50% stenosis in ≥1 artery as graded by the physician performing the procedure. A threshold of ≥50% stenosis was chosen as the definition of obstructive CAD on the basis of the American College of Cardiology Foundation and American Heart Association guidelines, in which a >50% stenosis is a class IIa recommendation for bypass in patients who undergo noncoronary cardiac surgery. A threshold of ≥50% was specifically chosen over the class I recommended threshold of ≥70% stenosis on the basis of the goal of designing a preoperative screening tool. Using the threshold of ≥50% ensures that the risk prediction nomogram and the associated estimated risk for CAD provide a conservative estimate when being used to triage between CTA and DCA.
Clinical, imaging, and laboratory characteristics were obtained from the time of the DCA by querying the electronic medical record (Epic, Verona, Wisconsin). Co-morbidities such as diabetes, hypertension, hyperlipidemia, and stroke were identified by International Classification of Diseases, Ninth Revision, coding. Glomerular filtration rate was calculated using the Cockcroft-Gault equation, and renal dysfunction was defined as a glomerular filtration rate <60 ml/min/1.73 m 2 or the requirement for renal replacement therapy. The Institutional Review Board at the Cleveland Clinic granted approval for this research.
Continuous variables are expressed as mean ± SD or as median (interquartile range) for normal or non-normal distributions, respectively. Student’s t test or the Wilcoxon rank sum test was used to assess differences between groups for continuous variables. Categorical variables were compared using the chi-square test and are expressed as percentages. Logistic regression was used to estimate the odds ratios and 95% confidence intervals (CIs) for predictors of obstructive CAD. Covariates selected for multivariate analysis included age, gender, diabetes, hypertension, history of stroke or transient ischemic attack, smoking, family history of premature CAD, hyperlipidemia, renal dysfunction, and the different valve surgical procedures (aortic valve, mitral valve, tricuspid valve, or multiple valve surgery).
A nomogram to predict the risk for obstructive CAD was constructed. Patients were randomly split in a 4:1 ratio into derivation and validation groups. The randomization was matched by age (±5 years) and by gender. Variables identified by multivariate regression as independent predictors of obstructive CAD were included in the nomogram. Within the derivation group, logistic regression was used to calculate the β coefficients for each of the included variables. These β coefficients were then applied to the validation group. The resultant model was validated by assessing Somers’s D rank correlation between predicted probabilities and observed responses. Then, bootstrapping using 250 sample patients was used to penalize for possible overfitting. This information was also used to estimate C statistics and corresponding 95% CIs. The performance of the final model was assessed in the validation cohort and corrected for overfitting. All analyses were performed using R version 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria), and p values <0.05 were considered tol indicate statistical significance.
Results
We identified 5,360 patients without histories of CAD who underwent DCA as part of preoperative evaluation for VHS. Within the cohort, 1,035 patients (19.3%) were found to have obstructive CAD ( Table 1 ). Compared with patients without obstructive CAD, those with obstructive CAD tended to be older, were more frequently male, and had a higher prevalence of co-morbidities, including diabetes, hypertension, hyperlipidemia, a history of transient ischemic attack or stroke, and renal dysfunction. Additionally, patients with obstructive CAD more often underwent aortic valve surgery compared with non–aortic valve surgery.
| Characteristic | Entire Cohort (n = 5360) | Obstructive CAD | p-Value | |
|---|---|---|---|---|
| No (n = 4325) | Yes (n = 1035) | |||
| Age (years) | 63 ± 14 | 60 ± 13 | 71 ± 11 | <0.001 |
| Male | 57% | 55% | 63% | <0.001 |
| Body Mass Index (kg/m 2 ) | 27 [24-30] | 27 [24-30] | 27 [24-30] | 0.20 |
| Diabetes Mellitus | 16% | 13% | 28% | <0.001 |
| Hypertension | 50% | 49% | 55% | 0.003 |
| History of CVA or TIA | 8% | 8% | 10% | 0.04 |
| History of smoking | 39% | 39% | 39% | 0.89 |
| Family history of premature CAD | 8% | 7% | 9% | 0.12 |
| Hyperlipidemia | 39% | 36% | 52% | <0.001 |
| Renal dysfunction | 25% | 21% | 42% | <0.001 |
| Aortic Valve surgery | 50% | 47% | 65% | <0.001 |
| Mitral Valve surgery | 49% | 53% | 35% | <0.001 |
| Tricuspid Valve surgery | 9% | 10% | 4% | <0.001 |
| Multiple Valve surgery | 7% | 8% | 4% | <0.001 |
| ACE or ARB | 44% | 43% | 49% | <0.001 |
| Aspirin | 40% | 39% | 44% | 0.002 |
| Beta-blocker | 46% | 46% | 45% | 0.38 |
| Statin | 35% | 32% | 46% | <0.001 |
Baseline characteristics stratified by specific valve surgery are listed in Table 2 . Within the cohort of patients who underwent aortic valve surgery, the prevalence of obstructive CAD was higher in this group compared with those who underwent mitral, tricuspid, or multiple valve surgery. Additionally, patients who underwent aortic valve surgery were older than those who underwent mitral or tricuspid valve surgery and were more often male.
| Characteristic | Aortic Valve (n = 2481) | Mitral Valve (n = 2285) | Tricuspid Valve (n = 200) | Multiple Valve (n = 394) | p-Value |
|---|---|---|---|---|---|
| Age (years) | 64 ± 14 | 60 ± 12 | 62 ± 15 | 65 ± 14 | <0.001 |
| Male | 61% | 57% | 34% | 42% | <0.001 |
| Body Mass Index (kg/m 2 ) | 29 ± 6 | 27 ± 6 | 28 ± 7 | 28 ± 7 | <0.001 |
| Diabetes Mellitus | 21% | 10% | 18% | 22% | <0.001 |
| Hypertension | 56% | 44% | 45% | 58% | <0.001 |
| History of CVA or TIA | 9% | 7% | 7% | 15% | <0.001 |
| History of smoking | 40% | 38% | 40% | 43% | 0.07 |
| Family history of premature CAD | 8% | 8% | 7% | 6% | 0.73 |
| Hyperlipidemia | 46% | 33% | 30% | 37% | <0.001 |
| Renal dysfunction | 25% | 21% | 37% | 40% | <0.001 |
| Obstructive CAD | 26% | 14% | 11% | 10% | <0.001 |
| ACEI or ARB | 45% | 41.8% | 42% | 50% | 0.01 |
| Aspirin | 42% | 37% | 34% | 44% | <0.001 |
| Beta-blocker | 43% | 46% | 59% | 61% | <0.001 |
| Statin | 42% | 29% | 25% | 34% | <0.001 |
In univariate analyses, age had the strongest association with obstructive CAD ( Table 3 ). The relation between age and obstructive CAD is depicted in Figure 1 . As expected, the risk for obstructive CAD increased with age, but this relation was not linear. There was a steep increase in risk starting at about 50 years of age. The risk for obstructive CAD in patients ≤50 years of age was 4.2%, while in patients >50 years of age, it was 23.0%. The risk for obstructive CAD further increased with age, and in those >65 years of age, it reached 32%. In contrast, for patients ≤40 years of age, the overall risk for obstructive CAD was only 1.9%.
| Predictors | Or (95% CI) | p-Value |
|---|---|---|
| Age (years) | 2.65(2.43-2.89) | <0.001 |
| Male | 1.38(1.2-1.59) | <0.001 |
| Body Mass Index (kg/m 2 ) | 1.01(0.95-1.08) | 0.72 |
| Diabetes Mellitus | 2.49(2.12-2.93) | <0.001 |
| Hypertension | 1.23(1.08-1.41) | 0.003 |
| History of CVA or TIA | 1.28(1.02-1.62) | 0.03 |
| History of smoking | 1.01(0.88-1.16) | 0.87 |
| Family history of premature CAD | 1.22(0.96-1.56) | 0.10 |
| Hyperlipidemia | 1.92(1.68-2.21) | <0.001 |
| Renal dysfunction | 2.75(2.37-3.2) | <0.001 |
| Aortic Valve surgery | 2.13(1.85-2.45) | <0.001 |
| Mitral Valve surgery | 0.48(0.42-0.56) | <0.001 |
| Tricuspid Valve surgery | 0.42(0.3-0.57) | <0.001 |
| Multiple Valve surgery | 0.45(0.32-0.63) | <0.001 |
| ACE or ARB | 1.31(1.14-1.5) | <0.001 |
| Aspirin | 1.24(1.09-1.43) | 0.002 |
| Beta-blocker | 0.94(0.82-1.08) | 0.36 |
| Statin | 1.82(1.58-2.08) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
