All of the most widely-cited studies for the prediction of maximum exercise responses have utilized either volunteers or referred subjects. Therefore, selection bias, with overestimation of the reference values, is a likely consequence. In order to establish a set of predictive equations for the gas exchange, ventilatory and cardiovascular responses to maximum ramp-incremental cycle ergometry, this study prospectively evaluated 120 sedentary individuals (60 males, 60 females, aged 20-80), randomly-selected from >8,000 subjects. Regular physical activity pattern by questionnaire, body composition by anthropometry and dual energy X-ray absorptiometry (n = 75) and knee strength by isokinetic dynamometry were also assessed. Previously reported equations typically overestimated the subjects' peak oxygen uptake (p<0.05). Prediction linear equations for the main variables of clinical interest were established by backward stepwise regression analysis including: sex, age, knee extensor peak torque, bone-free lean leg mass, total and lean body mass, height, and physical activity scores. Reference intervals (95% confidence limits) were calculated: some of these values differed markedly from those formerly recommended. The results therefore might provide a more appropriate frame of reference for interpretation of the responses to symptom-limited ramp incremental cycle ergometry in sedentary subjects; i.e. those usually referred for clinical cardiopulmonary exercise tests.