Improving the Positive Predictive Value of Exercise Testing in Women for Coronary Artery Disease

The exercise treadmill test (ETT) in women has been limited by a low positive predictive value (PPV) for coronary artery disease (CAD). However, the reliability of previous studies was unsatisfactory because of the inclusion of younger women with a low prevalence of CAD. To further evaluate the diagnostic properties of the ETT in women, we evaluated a group of women with chest pain who had a positive ETT result and subsequent coronary angiography. Of the 111 women, 56 had significant CAD on angiogram, yielding a PPV of 51% for the group. However, inclusion in the analysis of several pretest attributes and specific exercise test responses improved the PPV of the ETT. Age had a major effect, with the youngest group (35 to 50 years old) having a PPV of 36% compared to 68% in the oldest group (>65 years old). Several specific exercise responses (ST-segment depression >2 mm and delayed ST-segment recovery >3.0 minutes) further separated true from false positives across all age groups, increasing the PPV to approximately 80%. Onset of ischemia at a relatively low cardiac workload of <80% maximum predicted heart rate was not a significant predictor. In conclusion, the standard ETT should remain the test of choice in ambulatory women with chest pain and no significant abnormalities on baseline electrocardiogram especially in those >65 years of age.

Age is the strongest predictor of coronary artery disease (CAD) in women and men, with prevalence increasing sharply in women after the sixth decade. The degree to which this and other pretest attributes affect the positive predictive value (PPV) of the standard exercise treadmill test (ETT) has not been clarified. Previous studies demonstrating the low PPV of the ETT in women included primarily young premenopausal subjects. Further, they uniformly excluded women >65 years old, the decades during which CAD prevalence and severity are highest. In the present study, we examined the effects of increasing age on the PPV of the standard ETT in women and investigated the effects of specific electrocardiographic (ECG) exercise responses on PPV across several age groups.


The study group consisted of 111 consecutive woman patients at the University of California (Davis) Medical Center who had a positive ETT result as part of an evaluation for chest pain, underwent coronary angiography within 10 months of the test, and had no clinical events or alteration in clinical status during the interval from exercise testing to angiography. Subjects were identified by a review of exercise treadmill laboratory and cardiac catheterization laboratory records. The study group consisted of patients who underwent elective ETT and cardiac catheterization and were excluded if they had (1) a previous diagnosis of CAD by previous noninvasive testing, history of myocardial infarction, or previous angiography or (2) more than mild valvular stenosis or regurgitation at the time of left heart catheterization. Patients with minor baseline ECG abnormalities (<0.5-mm ST-T shift) and/or ECG evidence of left ventricular hypertrophy and those taking digoxin were not excluded. Women were divided into 3 groups based on age: group I included women 35 to 50 years old, group II 51 to 64 years old, and group III ≥65 years old. Characteristics of the study group are listed in Table 1 . Prevalence of baseline ECG abnormalities and frequency of digoxin use did not differ significantly among the 3 groups.

Table 1

Characteristics of study group

Variable I (n = 36) II (n = 41) III (n = 34) p Value
Mean age (years) 44 ± 0.64 58 ± 0.62 72 ± 0.91 <0.05
Range (years) 35–50 51–64 ≥65
Digoxin use 5 (14%) 4 (10%) 5 (15%) NS
Baseline ST-segment abnormalities 0 1 (2%) 3 (9%) NS

Denotes study subgroup.

All patients underwent ETT by the Bruce or modified Bruce protocol. A positive ETT result was defined as exercise-induced ST-segment depression ≥1.0 mm at 0.08 second after the J-point. All ETT results were reviewed by 2 cardiologists without knowledge of the results of coronary angiography. The following 4 exercise responses were determined for each patient: (1) degree of ST-segment depression (1 to 1.9 or ≥2.0 mm), (2) duration of ischemic response in recovery phase (0 minute to 3 minutes or ≥4 minutes), (3) percent age-predicted maximum heart rate (MPHR) achieved (<85% or ≥85%), and (4) percent MPHR at onset of ischemia (<80% or ≥80%) where MPHR was estimated as 220 minus age.

Selective coronary angiography was performed in multiple projections according to standard technique using the femoral or radial artery approach. Clinically significant CAD was defined as ≥50% stenosis in ≥1 coronary artery or major branch. Extent of CAD (1 vessel or 2 or 3 vessels) was determined for each patient with a true positive ETT (positive ETT associated with ≥50% coronary artery stenosis in ≥1 coronary arteries).

In this group of patients, all of whom had positive ETT results, the presence of significant CAD identified a “true positive” test and absence of significant CAD indicated a “false positive” test. PPV was calculated as true positive divided by the sum of true positive plus false positive multiplied by 100%. PPV was calculated for the entire study group, each age group, and within age groups for each exercise response subgroup. Differences between age groups and exercise response subgroups were tested by chi-square (categorical data) or Fischer exact (continuous data) analysis.


Fifty-six of 111 women with a positive ETT result had angiographically significant CAD, for an overall PPV of 51%. Figure 1 shows the effect of increasing age on PPV, which increased from 36% in younger women to 68% in the oldest group (p <0.05, group I vs III). In patients with a true positive ETT result, extent of CAD within each age group is depicted in Figure 2 . Younger women had a predominance of 1-vessel CAD and older women a predominance of 3-vessel CAD.

Figure 1

Effect of age on positive predictive value of the exercise treadmill test in 111 women (p <0.05, group I vs III).

Figure 2

One-vessel (black bars) , 2-vessel (light gray bars) , and 3-vessel (dark gray bars) coronary artery disease by age group in women with true positive exercise treadmill test results (n = 56).

Effects of specific exercise responses on PPV within each age group are illustrated in Figures 3 and 4 . Increasing degree of ST-segment depression augmented PPV within each age group ( Figure 3 ) but more so in younger women, in whom PPV increased from 17% to 69% with >2.0-mm ST-segment depression (p <0.05).

Dec 7, 2016 | Posted by in CARDIOLOGY | Comments Off on Improving the Positive Predictive Value of Exercise Testing in Women for Coronary Artery Disease

Full access? Get Clinical Tree

Get Clinical Tree app for offline access