A Comparative Study of Different Treadmill Scores to Diagnose Coronary Artery Disease

Introduction At present several treadmill scores have been proposed as means for improving the diagnostic accuracy of the exercise treadmill test (ETT) and to predict future risk of cardiac events (Berman et al., 1978 \& Do et al., 1997). Although a large number of noninvasive stress testing modalities are currently available, the exercise ECG is still used as a standard for comparison with other clinical and testing risk markers. It is also the least costly of all provocative noninvasive tests. The Duke Treadmill Score (DTS), traditionally a prognostic score, was recently tested as a diagnostic score and shown to predict CAD better than the ST response alone(Shaw et al., 1998). But questions remain regarding the diagnostic accuracy of treadmill scores when applied to a different patient population; furthermore, many treadmill scores have not been compared with one another in the same population(Fearon et al., 2002).To date, no composite stress-test score or noninvasive risk index has been shown to provide both accurate diagnostic and prognostic risk estimates. Despite that exercise treadmill test remains a useful test for diagnosing coronary artery disease (CAD) in patients with chest pain and at intermediate risk for CAD (Do et al., 1997).The sensitivity and specificity of ETT varies considerably. According to a meta-analysis conducted by Gianrossi et al there was a wide variability in sensitivity and specificity of ETT [sensitivity 68± 16% (range 23-100%); specificity 77±17% (range: 17-100%). Another Meta analysis showed sensitivity of 81± 12% (range: 40-100%) and specificity of 66± 16% (range: 17-100%) (Myers et al., 1994).Fearon WF et al showed sensitivity and specificity was higher when treadmill scores were applied in comparison to ST response alone. They used a consensus score consisting of the Morise, Dentrano and VA score and found predictive accuracy of the consensus score for stratifying patients to low and high likelihood for CAD was significantly higher than the predictive accuracy of DTS, 80(74-86)% versus 71(65-77)% (p< 0.0001). But Fearon WF et al conducted the study only on male population in USA and the consensus score was calculated by average of computer generated treadmill scores. In 2012 Mao L et al have shown 73 out of 104 male patients were detected CAD both by ETT and CAG, the accuracy rate was 70.2% which was much higher than that (50.0%) of the female patients (p<0.05) and they only used ST changes alone to demonstrate ETT positivity In this study, we will compare the diagnostic accuracy of well known prognostic scores namely, Duke Treadmill Score, Simple Treadmill Score and Cleveland Clinic Score to identify significant coronary artery lesion in Bangladeshi male and female patients.

Rationale

Though exercise treadmill test has high false positive and negative rates(Zang et al., 2007) , it is cheap, easily available, less time consuming to the interpret results and its accuracy can be increased by calculating ST/HR index, treadmill score, QT dispersion and so on (Kronander 2010 \&Dentrano 1989). On the contrary the gold standard test coronary angiogram for detecting CAD is expensive, time consuming, potentially hazardous with many complications and often the CAG shows normal coronary arteries in female population.ST-segment depression and chest pain as the classic criteria for CAD diagnosis are well known and accepted. Besides If treadmill score were used the diagnostic accuracy of ETT would had been higher. The accuracy of different treadmill scores in Bangladeshi population especially the female population is largely unknown. Duke Treadmill Score and Simple Treadmill Score are well validated score in western population and are used for diagnostic \& prognostic interpretation of ETT. The predictive accuracy of DTS to diagnose CAD is 71% (Fearon 2002). In 2001 Raxwal V et al. showed simple treadmill score has sensitivity of 88% and specificity of 96%. If we calculate the accuracy of simple treadmill score using the formula "Accuracy = (Sensitivity) x Prevalence + (Specificity) x (1- Prevalence)" it sums up nearly 93% according to prevalence of CAD in urban population. Cleveland Clinic Score is a prognostic score of ETT. It gives value from which we can predict the probability of 3 year or 5 year survival. It was shown that it has a very high negative predictive value approaching 97%. Besides to the best of our knowledge Cleveland Clinic Score was not tested as a diagnostic predictor of CAD and there are few studies regarding treadmill scores predictability in Bangladesh. In our study we will use all of these three scores and compare their accuracy to predict significant CAD. DTS, Simplified Treadmill Score, and Cleveland clinic score can be implemented effectively to identify patients with low probability of CAD and excluded from undergoing expensive and potentially hazardous CAG if the real scenario of the treadmill scores is known in our population.

Research question

Howdo different treadmill scores (Duke Treadmill Score, Simple Treadmill Score, Cleveland Clinic Score) vary to predictability of Coronary Artery Disease in Bangladeshi population in a tertiary care hospital?

General Objectives

To identify difference of predictability of DTS, Simple Treadmill Score and Cleveland Clinic Score to diagnose significant CAD by Coronary Angiography.

Specific Objectives

1. To estimate accuracy of ST segment response, DTS, Simple Treadmill Score and Cleveland Clinic Score to predict CAD.
2. To compare DTS, Simple treadmill score, Cleveland Clinic Score accuracy to predict coronary artery disease.
3. To identify the relation of different level of treadmill scores with severity of CAD.

Study Area

University Cardiac Center, Bangabandhu Sheikh Mujib Medical University (BSMMU). BSMMU is a renowned institute in Bangladesh with good indoor and outdoor facility. It also has good inpatient and outpatient services for local and other patients coming from distant places. There is a good mix of male and female patients also which is needed to test the study hypothesis. Overall, the patients coming in outdoor facility to get treatment represent the Bangladeshi population very well and uniformly.

Sampling Procedures

Patient presented with stable chest pain, who have undergone ETT according to Bruce protocol and admitted for CAG, will be selected as case considering inclusion and exclusion criteria. Detailed and thorough clinical assessment will be done and recorded. All available previous medical documents will be checked meticulously. Patients with previous revascularization, left bundle-branch block, paced rhythms or Wolff-Parkinson-White syndrome (WPW) on resting electrocardiogram (ECG), or valvular heart disease, congenital heart disease will be excluded from the study. To avoid falsely increasing the accuracy of the exercise treadmill test, patients with a previous myocardial infarction by history or by diagnostic Q wave will be excluded.

With history, clinical findings and investigations cases other than stable chest pain will be excluded. Informed written consent will be taken from the patient. CAG report will be collected from the Cath lab after the procedure.

Data collection:

Data will be recorded in pre-designed questionnaires by history, clinical examination and investigation with the patient of University Cardiac Center, BSMMU.

Quality assurance strategy:

A set of questionnaire will be formulated \& checked. To make the study credible, reliable \& dependable data will be collected by principal investigator by using those questions over a month of period. Again the questions will be edited accordingly \& necessarily after discussion with the guide and co-guide of this study.

Ethical Issues At first ethical clearance will be taken from the ethical review committee of Bangabandhu Sheikh Mujib Medical University (BSMMU). The study will be carried out according to 1964 Helsinki Declaration for Medical Research involving Human subjects and amended by the 64th World Medical Association General Assembly, October 2013. No drugs or placebo will be used for this study. Each participant will enjoy every right to participate or refuse participation. They will be free to withdraw their participation at any stage of the study. Data taken from the participants will be regarded as confidential. Data will be used only for this scientific study. Participants will be informed in detail about the nature and purpose of the study, and informed written consent will be taken from each participant.

Sample size calculation

Sample size is calculated by using the following equation (One sample comparison of proportion):

n=⌈Zβ√(p(1-p) )+Zα√p1(1-P1)⌉^2/((p-p1)²) n = required sample size p = Proportion under alternative hypothesis that is proposed to be detected or worst possible outcome p1= Proportion under null hypothesis or proportion in the population Zα = 1.96 (5% level of significance) Zβ = 1.28 when power is 0.9 According to Fearon WF et al (2002) the predictive accuracy of DTS is 71% (0.71).

If we assume: p1= 0.71 P = 0.55, Power = 0.8, α = 0.05 Sample size n = ⌈1.28√(.55(1-.55) )+1.96√(.71(1-.71))⌉^2/((.55-.71)²) = 91

Correction for non-response:

Nf=100/100- Nr If Nr= Percentage of expected non-response is 10% Nf=100/100- 10 = 1.11 Final sample size will be estimated sample size (n) x Nf = 91 x 1.11 = 102. So our required sample size is at least 102.

StatisticalMethods Using angiographic evidence of CAD as the reference, area under the curve (AUC) of receive operator characteristic (ROC) plots will be determined for the ST response alone and for each treadmill score. The AUC for each treadmill score will be compared with the AUC of the ST response alone and the AUCs of the other treadmill scores. The predictive accuracies of the DTS, the Simple Treadmill Score and Cleveland Clinic Score to stratify patients into high or low likelihood for CAD will be calculated and compared. Statistical analysis will be performed with the SPSS.