To determine the precise nonsteady-state characteristics of ventilation (VE), O2 uptake (VO2), and CO2 output (VCO2) during moderate-intensity exercise, six subjects each underwent eight repetitions of 100-W constant-load cycling. The tests were preceded either by rest or unloaded cycling ("0" W). An early component of VE, VO2, and VCO2 responses, which was obscured on any single test by the breath-to-breath fluctuations, became apparent when the several repetitions were averaged. These early responses were abrupt when the work was instituted from rest but were much slower and smaller from the 0-W base line and corresponded to the phase of cardiodynamic gas exchange. Some 20 s after the onset of the work a further monoexponential increase to steady state occurred in all three variables, the time constants of which did not differ between the two types of test. Consequently, the exponential behavior of VE, VO2, and VCO2 in response to moderate exercise is best described by a model that incorporates only the second phase of the response.

Breathing has inherent irregularities that produce breath-to-breath fluctuations ("noise") in pulmonary gas exchange. These impair the precision of characterizing nonsteady-state gas exchange kinetics during exercise. We quantified the effects of this noise on the confidence of estimating kinetic parameters of the underlying physiological responses and hence of model discrimination. Five subjects each performed eight transitions from 0 to 100 W on a cycle ergometer. Ventilation, CO2 output, and O2 uptake were computed breath by breath. The eight responses were interpolated uniformly, time aligned, and averaged for each subject; and the kinetic parameters of a first-order model (i.e., the time constant and time delay) were then estimated using three methods: linear least squares, nonlinear least squares, and maximum likelihood. The breath-by-breath noise approximated an uncorrelated Gaussian stochastic process, with a standard deviation that was largely independent of metabolic rate. An expression has therefore been derived for the number of square-wave repetitions required for a specified parameter confidence using methods b and c; method a being less appropriate for parameter estimation of noisy gas exchange kinetics.

A method has been developed for on-line breath-by-breath calculation of alveolar gas exchange by correcting the gas exchange measured at the mouth for changes in lung gas stores. The corrections are applied to the total lung gas exchange, which is found by directly subtracting expired from inspired volume of each gas. Corrections are made for both breath-to-breath changes in lung volumes and changes in alveolar gas concentrations. The lung volume correction term has the effect of reducing the large error sensitivity of O2 exchange that has, in the past, resulted from direct determination by total lung gas exchange. Error each gas. Corrections are made for both breath-to-breath changes in lung volumes and changes in alveolar gas concentrations. The lung volume correction term has the effect of reducing the large error sensitivity of O2 exchange that has, in the past, resulted from direct determination by total lung gas exchange. Error each gas. Corrections are made for both breath-to-breath changes in lung volumes and changes in alveolar gas concentrations. The lung volume correction term has the effect of reducing the large error sensitivity of O2 exchange that has, in the past, resulted from direct determination by total lung gas exchange. Error sensitivity analysis shows that the effect of inaccuracies due to errors in measuring gas flow or gas concentrations are similar in magnitude to those in the open-circuit method that has traditionally been used. The algorithm for alveolar gas exchange has been implemented in a computer program for on-line respiratory analysis alongside the open-circuit calculation of gas exchange at the mouth that has been used in out laboratory. By use of several experimental studies, it is shown that there are very apparent breath-to-breath differences between the gas exchange measured by the two methods. During metabolic and respiratory transients, these differences often have significant influence on interpretation of the underlying physiology.

The effect of cardiovascular adjustments on the coupling of cellular to pulmonary gas exchange during unsteady states of exercise remains controversial. Computer simulations were performed to assess these influences on O2 delivery and pulmonary O2 uptake (pVO2). Algorithms were developed representing muscle and "rest-of-body" compartments, connected in parallel by arterial and venous circulations to a pump-and-lungs compartment. Exercise-induced increases in VO2 and cardiac output went to the muscle compartment. Model parameters [e.g., time constants for blood flow and muscle O2 uptake (mVO2)] could be varied independently. Simulation results demonstrated that 1) the rise in pVO2 during exercise contains three phases; 2) the contribution of changes in venous O2 stores to pVO2 kinetics and the O2 deficit occur almost entirely in phase 1; 3) under a wide variety of manipulations, the kinetics of pVO2 in phase 2 were within a couple of seconds of that assigned to mVO2 (i.e., there is not an obligatory slowing of VO2 kinetics at the lungs relative to those at the muscles; 4) by use of available estimates of blood flow adjustment, O2 delivery would not limit mVO2 after exercise onset; and 5) blood flow could limit O2 delivery in recovery, if blood flow returned to base-line levels at rates similar to those during the on-transient phase.