BACKGROUND: There are many suggested methods for the indirect determination of anaerobic threshold(AT) using the changes of ventilatory parameters in respones to ventilatory load accompanying the increase of blood lactic acid level during exercise and the threshold derived from them is called ventilatory threshold( VT). They include ventilatory equivalent method(VEM), End-tidal PO2 method(PETO2), V-slope method(VSM), and respiratory quotient method(RQ). But in the patients with chronic airway ohstruction(CAO), the AT determined by ventilatory methods may not reflect true AT because the patients with GAO show inadequate ventilatory response to the increase of blood lactic acid level during excercise. METHODS: For the investigation of detection rate and reliability of above four VT determination methods in the patients with GAO, we performed the symptom-limited and maximal incremental exercise test in 17 patients with GAO and 12 normal controls. The incremental workload was 10 W/min in GAO group and 25 W/min in control group. The reliability of VT in each group was investigated by the calculation of Spearman correlation coefficient. RESULTS: The detection rates of VT were 100% by RQ, 91.7% by both VEM and POETO2, and 83.3% by VSM in normal control group, while 94.1% by RQ, 64.7% by VEM and PETO2, and 83.3% by VSM in GAO group. Good correlations were noted among VEM, PETO2, and VSM except RQ in normal control group. But there was no significant correlation except between VEM and PETO2 in GAO group. CONCLUSIONS: RQ is very sensitive but crude and VEM is near similar to PETO2. The clinical usefulness of VT determined by ventilatory method might be limited in patients with severe GAO.