RESUMEN
BACKGROUND: Pulmonary arteriovenous malformations (PAVMs) generate a right-to-left shunt. Impaired gas exchange results in hypoxaemia and impaired CO2 clearance. Most patients compensate effectively but some are dyspneic, and these are rarely the most hypoxaemic. AIM: To test degrees of concurrent pathology influencing exercise capacity. DESIGN: Replicate, sequential single centre, prospective studies. METHODS: Cardiopulmonary exercise tests (CPETs) were performed in 26 patients with PAVMs, including individuals with and without known airflow obstruction. To replicate, relationships were tested prospectively in an independent cohort where self-reported exercise capacity evaluated by the Veterans Specific Activity Questionnaire (VSAQ) was used to calculate metabolic equivalents (METs) at peak exercise (n = 71). Additional measurements included oxygen saturation (SpO2), forced expiratory volume in 1 s (FEV1), vital capacity (VC), fractional exhaled nitric oxide (FeNO), haemoglobin and iron indices. RESULTS: By CPET, the peak work rate was only minimally associated with low SpO2 or low arterial oxygen content (calculated as CaO2=1.34 × SpO2 × haemoglobin), but was reduced in patients with low FEV1 or VC. Supranormal work rates were seen in patients with severe right-to-left shunting and SpO2 < 90%, but only if FEV1 was >80% predicted. VSAQ-calculated METS also demonstrated little relationship with SpO2, and in crude and CaO2-adjusted regression, were lower in patients with lower FEV1 or VC. Bronchodilation increased airflow even where spirometry was in the normal range: exhaled nitric oxide measurements were normal in 80% of cases, and unrelated to any PAVM-specific variable. CONCLUSIONS: Exercise capacity is reduced by relatively mild airflow limitation (obstructive or restrictive) in the setting of PAVMs.