Subjects

 The medical ethical board of the Máxima Medical Centre (MMC), Veldhoven, The Netherlands, approved this study. Patients with IC were selected in the outpatient vascular clinic between March 2011 and June 2011. Inclusion criteria were newly diagnosed patients with symptoms of IC >3 months and an ABI <0.9 at rest or an ABI <0.9 after exercise or a fall in ankle systolic pressure by more than 20 percent from its baseline value after exercise. Patients with rest pain, tissue loss or severe cardiopulmonary disease (>NYHA class II) or problems understanding the study protocol were excluded.

Studying healthy controls was deemed necessary in order to verify if bicycle testing altered ABI values after exercise. Individuals matched for age and gender with recently corrected minor non-vascular surgical morbidity not reporting IC were asked to serve as controls. All participants were unaware of the study hypothesis and gave verbal and written informed consent.

Sample size and statistical power

 A minimum of 26 patients was required to detect a minimal 0.15 difference in ABI index between measurements after a treadmill test and a bicycle test with 80% of power using a two-sided t-test and a 0.05 significancy level. To allow for a 25% dropout rate, a total of 32 patients was required. A sample size of 8 controls was calculated for attaining a significant difference compared to IC patients.

Study Protocol

 A vascular surgeon performed history taking and physical examination and checked whether patients met inclusion criteria. They underwent treadmill and bicycle tests in random order. A Doppler probe (Nicolet Vasoguard, VIASYS Healthcare, USA) measured systolic pressures of right brachial artery, right posterior tibial artery (RPTa) and left posterior tibial artery (LPTa) in fixed order. If RPTa or LPTa were absent, measurements were obtained from the pedal dorsal artery (PDa). ABI measurement of just a single foot artery (and not both) was performed as a means to minimize the elapsed time between consecutive measurements. If signals over the PDa were also absent, the patient was excluded. ABI values were calculated as ratios to brachial systolic pressure. Control subjects only performed bicycle testing.

Treadmill test

 Resting ABI values were measured prior to treadmill testing after participants rested in a supine position for 5 minutes. A treadmill test (TT) operated at a speed of 3.0 km per hour with an 8% slope for a 5 minute maximum. Duration until onset of IC (initial claudication duration, ICD, in seconds) and maximum duration (absolute claudication duration, ACD, in seconds) were noted after 5 minutes of walking (or after exhaustion). ABIs were obtained from both sides as fast as possible three times during a 6-minute time period after discontinuation of the test. Right and left legs were subsequently measured, whereas measurements were always started in the right leg. As timing of these latter measurements showed a too large variance, only the first two values were used for further analysis.

Submaximal and maximal bicycle tests

 Prior to bicycle testing, participants were allowed a 5-minute rest in supine position. ABI values were obtained with a hand-held Doppler probe (Huntleigh Dopplex D900, Huntleigh, United Kingdom). An automatic pneumatic blood pressure monitor was used to obtain systolic blood pressures of the right brachial artery.

 Cycling tests were performed using an electronically braked computer-controlled cycle ergometer bicycle (LODE® Excalibur Sport, Lode, The Netherlands). Seat and steer height of this bicycle were optimally adjusted to the subject's own

posture. Individuals were instructed to perform a submaximal bicycle test (SBT) for a 5-minute maximum period with 65 rounds per min (rpm) and continuous resistance ('Step form'). The amount of resistance was computed to approach a similar workload (in VO2) as performed during the walking tests using the American College of Sports Medicine (ACSM) leg ergometry and walking equations (6). Duration until onset of IC and maximum duration were noted after 5 minutes of this cycling exercise (or earlier if the subject stopped in case of exhaustion or IC). ABI values of both sides were subsequently obtained two times immediately after each other, again the right leg was measured first. After a 10 minute resting time (including the 6 minute measurement period), a maximal bicycling test (MBT) was started using an incremental approach ('ramp'). Aim of this heavier protocol was to expose patients to a maximal exercise test with continuously increasing resistance towards a peak after approximately 10 minutes. A questionnaire was used prior to this exercise test to allow for an estimation of this peak tolerance. If unbearable pain or exhaustion occurred, ABI values of both legs were again subsequently registered.

Statistical analysis

 Statistical analysis was performed using SPSS Statistics (MAC OS X version 20.0). Demographic data were analyzed using Pearson's chi-square, or Fisher's exact tests in case of cells with small numbers. It was assumed that both legs could be used as independent parameters in the analysis as each leg of a patient may respond differently to testing. ABI values were analyzed for skewness and kurtosis. A Wilcoxon Signed Rank test and Spearman's correlations were utilized to compare values in case of non-normal distribution whereas a paired Students T-test and Pearson's correlations analyzed normal distributed variables. Bland Altman plots (BA plots) were used to verify agreement between treadmill and bicycle tests. In this BA diagrams the differences between treadmill and bicycle testing (y-axis) were plotted against the averages of these two exercise types (x-axis). As a result the mean difference and its 95% confidence limits are shown. P values less than .0.05 were considered statistically significant.