Respiratory influences are major confounders when evaluating central haemodynamics during exercise. We studied four different methods to assess mean pulmonary artery pressure (mPAP) and pulmonary capillary wedge pressure (PCWP) in cases of respiratory swings. Central haemodynamics were measured simultaneously with oesophageal pressure during exercise in 30 chronic obstructive pulmonary disease (COPD) patients. mPAP and PCWP were assessed at the end of expiration, averaged over the respiratory cycle and corrected for the right atrial pressure (RAP) waveform estimated intrathoracic pressure, and compared with the transmural pressures. Bland-Altman analyses showed the best agreement of mPAP averaged over the respiratory cycle (bias (limits of agreement) 2.5 (-6.0-11.8) mmHg) and when corrected with the nadir of RAP (-3.6 (-11.2-3.9) mmHg). Measuring mPAP at the end of expiration (10.3 (0.5-20.3) mmHg) and mPAP corrected for the RAP swing (-9.3 (-19.8-2.1) mmHg) resulted in lower levels of agreement. The respiratory swings in mPAP and PCWP were similar (r(2)=0.82, slope ± se 0.95 ± 0.1). Central haemodynamics measured at the end of expiration leads to an overestimation of intravascular pressures in exercising COPD patients. Good measurement can be acquired even when oesopghageal pressure is omitted, by averaging pressures over the respiratory cycle or using the RAP waveform to correct for intrathoracic pressure. Assessment of the pulmonary gradient is unaffected by respiratory swings.