In patients with primary systemic amyloidosis (AL), the echocardiographic assessment of ventricular function alone does not always correspond to patients’ symptoms and functional status. Peak oxygen uptake and anaerobic threshold (AT), in contrast, constitute 2 objective, reliable and reproducible indicators of functional status in patients with circulatory failure. Thirty-two consecutive patients (mean age 50 ±13 years) with histologic evidence of systemic primary AL were studied (29 AL, 3 hereditary). There were 16 with echocardiographic features of cardiac infiltration (group I) and 16 without (group II). Twenty age- and gender-matched healthy subjects were also studied for comparison. Of the 32 patients, 12 were in New York Heart Association functional class I, 9 were in class II, and 11 were in class III. Each subject underwent 2-dimensional and Doppler echocardiography and cardiopulmonary exercise testing using a modified Bruce protocol. Left atrial (LA), left ventricular (LV) dimensions, wall thickness, and LV fractional shortening, as well as transmitral flow velocities and their E/A ratio were measured. Peak oxygen consumption (VO 2max [ml/kg/min]), AT (ml/kg/min), and exercise duration (seconds) were also measured. VO 2max and AT were lower in patients with AL than in controls (20.8 ± 7.0 vs 35.0 ± 8.5, p <0.001 and 13.1 ± 3.7 vs 27.0 ± 4.2, p <0.001, respectively). As a group, symptomatic patients had lower VO 2max, AT, and exercise duration than those without symptoms (17.1 ± 3.6 vs 27.0 ± 6.9, p = 0.0001, 11.1 ± 2.1 vs 16.2 ± 3.6, p = 0.0001, and 489 ± 235 vs 843 ± 197, p = 0.0001, respectively), whereas LV dimensions only showed a small difference (p = 0.03). VO 2max, AT, and exercise duration of patients in functional class I were higher than those in functional classes II and III (p = 0.01, p <0.05, and p = 0.007, respectively). Asymptomatic patients had lower VO 2max, AT, and exercise duration than controls (p <0.0001). VO 2max, AT, and exercise duration were poorly related to LA diameter, LV dimensions, fractional shortening, wall thickness, peak velocities of E and A waveforms, and E/A ratio. Patients with VO 2max >15 ml/kg/min had a better survival than patients with VO 2max <15 ml/kg/min. Thus, in patients with primary systemic AL, cardiorespiratory exercise testing is the preferred way of assessing functional capacity. Echocardiographic Doppler indexes at rest are not predictive of a patient’s symptoms and exercise capacity. Furthermore, VO 2max is a strong independent predictor of survival in these patients.