The direct Fick principle is the standard for calculating cardiac output (CO) to detect CO-dependent conditions like exercise pulmonary hypertension (ePH). Fick COarterial incorporates arterial haemoglobin (Hba) and oxygen saturation (S aO2 ) with oxygen consumption from exercise testing, while Fick COnon-arterial substitutes mixed venous haemoglobin (Hbmv) and peripheral oxygen saturation (S pO2 ) in the absence of an arterial line. The decision to employ an arterial catheter for exercise testing varies, and discrepancies in oxygen saturation and haemoglobin between arterial and non-arterial methods may lead to differences in Fick CO, potentially affecting ePH classification. We performed a retrospective analysis of 296 consecutive invasive cardiopulmonary exercise testing (iCPET) studies comparing oxygen saturation from pulse oximetry (S pO2 ) and radial arterial (S aO2 ), Hba and Hbmv, and CO calculated with arterial (COarterial) and non-arterial (COnon-arterial) values. We assessed the risk of misclassification of pre- and post-capillary ePH and data loss due to inaccurate S pO2 . When considering all stages from rest to peak exercise, Hba and Hbmv demonstrated high correlation, while S pO2 and S aO2 as well as COarterial and COnon-arterial demonstrated low correlation. Data loss was significantly higher across all stages of exercise for S pO2 (n=346/1926 (18%)) compared to S aO2 (n=17/1923 (0.88%)). We found that pre- and post-capillary ePH were misclassified as COnon-arterial data (n=7/41 (17.1%) and n=2/23 (8.7%), respectively). Patients with scleroderma and/or Raynaud's (n=11/33 (33.3%)) and black patients (n=6/19 (31.6%)) had more S pO2 data loss. Reliance upon S pO2 during invasive exercise testing results in the misclassification of pre- and post-capillary ePH, and unmeasurable S pO2 for black, scleroderma and Raynaud's patients can preclude accurate exercise calculations, thus limiting the diagnostic and prognostic value of invasive exercise testing without an arterial line.