Peripheral arterial disease (PAD) is characterised by lower-limb ischaemic pain during walking, which limits peak exercise performance and oxygen consumption during graded treadmill exercise testing (Hiatt et al. 1987). Alternative exercise training modalities that can avoid the discomfort associated with lower-limb muscular work, while conferring cardiorespiratory benefits and improved quality of life, could help to reduce the high rates of morbidity in this patient group. One such strategy is upper-limb aerobic exercise training, as it is not restricted by the onset of ischaemic pain and may therefore be a more acceptable mode of exercise for patients with PAD (Walker et al. 2000).

Approval for this study was obtained from the North Sheffield Local Research Ethics Committee. Following familiarisation with the testing protocols, we compared the cardiorespiratory responses to incremental arm-cranking exercise with those of comparable leg-cranking exercise in 22 patients with stable PAD (median age 67 years, range 50-82 years). The incremental arm and leg-cranking assessments were performed on separate days and were discontinuous in nature, incorporating 3 min bouts of work, with 2 min interpolated rest intervals to maximum exercise tolerance. A standard electronically braked cycle ergometer (modified for arm cranking) was used for both assessments. Work rate was increased by 7.5 and 15 W per increment in the arm and leg assessments, respectively, and pulmonary gas exchange variables were measured breath-by-breath.

At submaximal work rates (30-45 W), heart rate (P < 0.01; paired t test), pulmonary ventilation (P < 0.01; paired t test), blood lactate (P < 0.01; paired t test) and perceived exertion (P < 0.05; Wilcoxon test) were higher for arm-cranking in relation to leg-cranking exercise, whereas oxygen uptake and the systolic and diastolic blood pressure responses were not different. At maximal exercise tolerance, the (mean ± S.E.M.) pulmonary ventilation (33 ± 2 versus 34 ± 3 l min^-1) and heart rate (108 ± 4 versus 108 ± 5 beats min^-1) were not significantly different between arm- and leg-cranking exercise, respectively. Furthermore, the increase in systolic and diastolic blood pressures from resting values (P < 0.01; paired t test), were similar for arm- and leg-cranking exercise. Peak oxygen uptake was 1.04 ± 0.06 l min^-1 for leg exercise in comparison with 0.92 ± 0.05 l min^-1 for arm exercise (P < 0.05; paired t test). Interestingly, in 13 of the 22 patients peak oxygen uptake for arm exercise was over 85 % of that measured during leg exercise and in five of the patients it exceeded the peak oxygen uptake recorded for lower-limb exercise. These observations probably reflect the impairment of lower extremity perfusion capacity during intense leg exercise.

We therefore conclude that upper limb aerobic exercise is well tolerated and could be a useful exercise stimulus for evoking positive cardiovascular adaptations in patients with impaired ability to exercise their lower limbs.

We acknowledge the support of the British Heart Foundation.