A 20-meter ambulation test could be used to assess lower limb function in Pakinson's disease by measuring time and number of steps taken to go and back from seated to seated position between 2 chairs 10 meters apart (AT20). We analyzed intra- and inter-reliability of the AT20 in mild Parkinson's disease.

Twelve subjects with mild idiopathic Parkinson's disease (4 women, 58±20 years) performed in OFF-levodopa status (time since levodopa treatment 20±7h) three AT20 with shoes, at a ‘natural’ pace, at ‘maximal’ speed, and then with large steps on two occasions one week apart (test and re-test). The total number of steps – including during turns – and time to complete the task were clinically measured by 4 independent raters. Speed, step length and cadence were derived. Inter- and intra-rater reliability were assessed using intraclass correlation coefficient (ICC), the agreement frequency procedure, agreement being defined as a speed difference <0.05m/s.

For all conditions, mean intra- and inter-rater ICC were, respectively, 98.1±1.2% and 99.8±0.1% for speed, 98.5±0.6% and 99.3±0.3% for step length, and 80.5±6.1% and 97.0±1.2% for cadence. Mean intra- and inter-rater speed differences were 0.03±0.01m/s and 0.01±0.002m/s respectively. Speed difference was less than 0.05m/s in 88±8% of cases for intra-rater measurements and 98±3% cases for inter-rater measurements.

The 20-meter ambulation test, AT20, has excellent intra- and inter-rater reliability in mild Parkinson's disease.