Short-term reproducibility of a mental arithmetic stress test.

1. To evaluate the short-term reproducibility of heart rate, oscillometrically determined blood pressure, antecubital venous plasma catecholamine concentrations and subjective responses to strictly standardized mental arithmetic, we performed two identical tests 1 h apart in 14 young, healthy and normotensive male subjects (age 22-35 years). 2. Heart rate and blood pressure responses to the two stress tests were highly correlated, when expressed both as correlations between levels attained during stress (rs greater than 0.80 throughout) and as absolute reactivity measures (all rs greater than 0.75). Also, subjective stress responses were highly correlated, when considering both levels during stress and reactivity (r = 0.97 and r = 0.85, respectively). Stress levels of catecholamines were correlated, but the change scores (reactivity) were unrelated. 3. The measurement error SD for heart rate was 2.6 and 3.0 beats/min for reactivity and stress levels, respectively. The corresponding SD for blood pressure ranged between 2.7 and 4.4 mmHg. Subjective stress experience showed an SD of a similar magnitude. The responses of plasma catecholamine concentrations were subject to considerable variability. 4. It is concluded that haemodynamic and subjective stress responses and stress levels during the mental arithmetic stress test show acceptable reproducibility and high test-retest correlations. However, stress-induced changes in venous plasma catecholamine concentrations show low reproducibility.