Prognostic Power of the Clinical Exercise Test
Acta Universitatis Tamperensis, No. 1598

By Johanna Leino
April 2011
Tampere University Press
Distributed by Coronet Books
ISBN: 9789514483868
$82.50 Paper Original

The more than 20,000 cardiovascular deaths per year in Finland have made the identification of individuals at risk for cardiovascular death a pressing public challenge. The clinical exercise test and exercise-based ST segment deviation are widely accepted methods for diagnosing coronary heart disease (CHD). However, beyond ST segment analysis, a clinical exercise test provides much more information when estimating a patient’s risk for future cardiovascular events. As a part of the Finnish Cardiovascular Study (FINCAVAS), the present study was designed to investigate the prognostic power of single and multiple variables derived from a clinical exercise test. The FINCAVAS population includes all patients scheduled for an exercise stress test due to clinical reasons using a bicycle ergometer at Tampere University Hospital and willing to participate in the study between October 2001 and December 2008. The follow-up lasted until September 2009. The final number of participants was 4,568 (1,765 women), but smaller study populations were created after shorter follow-up periods. In addition to repeated measurement of standard parameters such as heart rate and blood pressure, digital high-resolution ECG at 500Hz was recorded continuously during the entire exercise test, including the resting and recovery phases. T-wave alternans (TWA) was analyzed continuously with the time-domain Modified Moving Average (MMA) Method.

During the follow-up (55±26 months) 321 patients died, including 138 cardiovascular deaths and 63 sudden cardiac deaths (SCD). The prognostic power of exercise-test-based variables was determined by means of Cox multivariate regression analysis after adjustment for common coronary risk factors. Diminished heart rate variability (HRV) during the pre-exercise phase was associated with an increased risk of all-cause mortality according to several parameters (log VLF power, log LF power, log LF power n.u, log HF power%, pcSD2, log RMSSD), but during peak exercise HRV did not predict worse prognosis. At peak exercise TWA predicted worse prognosis both as a categorized variable (?60 ?V) and a continuous variable, especially in lead V5. During the recovery after exercise TWA ?60 ?V predicted worse prognosis. A prolonged PR interval and first-degree atrioventricular (AV) block during recovery were associated with increased risk of cardiovascular death. Abnormal heart rate recovery (?18 bpm) after exercise was a significant predictor of all-cause death and cardiovascular mortality. The combination of abnormal heart rate recovery and TWA ?60 ?V was a significant predictor of all-cause and cardiovascular mortality when TWA was measured either during peak exercise or during recovery. A decreased recovery of rate-pressure product (RPP) was significantly associated with an increased risk of SCD, cardiovascular mortality and all-cause death.

An exercise test utilising the bicycle ergometer has unused prognostic potential when evaluating patients’ risk for future all-cause and cardiovascular mortality as well as for SCD. To enhance the prognostic power of the exercise test, the test should be considered as a continuum from rest via peak exercise to the recovery phase. Combinations of single parameters can increase the accuracy of the exercise test by reflecting cardiovascular health from several perspectives. However, more studies are needed before universal recommendations for exercise testing in risk stratification can be constructed.


Return to Coronet Books main page