The ankle jerk is one of the most commonly tested deep reflexes,1 but the effect of the position of the lower limb on ratings of the reflex is unclear. Most commonly, the ankle jerk is assessed with the subject lying supine with the hip abducted and externally rotated, and the knee in moderate flexion before the triceps surae tendon is tapped with the reflex hammer.2 3Alternatively, the subject could sit with the legs dangling freely over the edge of the couch3 or with the knee on a chair with the ankles projecting freely over the edge of the chair.4In a less commonly described position, the subject lies supine but the hip is adducted and the knee extended, and unlike the other positions, the triceps surae tendon is not struck by the reflex hammer, but the palmar surface of the examiner's hand placed against the sole of the foot.4 Anecdotal reports suggest that the ratings of ankle jerk are higher when the subject lies supine with the hip abducted and externally rotated,5 but one study suggested that the ankle jerk is best assessed with the hip adducted and the knee extended.4

It has been suggested that differences in the reported prevalence of absence of ankle jerk may be due to the position adopted for the lower limb to assess the ankle jerk rather than pathological absence of the reflex.4 6 To optimise the value of the ankle jerk in screening for polyneuropathy we compared the ankle jerk ratings of three examiners with subjects in three different lower limb positions to determine the lower limb position with the highest mean examiners' ratings.

Twenty one healthy young adults without neurological lesions were recruited into the study after informed consent. Examinations of three lower limb positions: subject supine with hip adducted and knees in extension; subject supine with hip abducted and externally rotated and knees in moderate flexion; and subject kneeling with the ankles projecting freely over the edge of the chair, were performed as described.2-4 All subjects were examined in two replicate sessions by three examiners. The ankle jerk was rated on a scale of 0–3 as described in a standard text of neurological examination.3 The techniques of positioning the limbs, tapping the tendon and ratings of the reflex were reviewed and performed on three subjects not part of the study before examinations started. Examiners were blinded to the subjects, who were covered with only their feet exposed. The ratings of the ankle jerk by the three examiners for each lower limb position were treated as repeated measures for each subject, and averaged to produce the rating for the lower limb position. Differences in the mean examiners' ratings for the three lower limb positions were compared by non-parametric methods.

The mean examiners' ratings of the ankle jerk for the three lower limb positions are presented in the table. The mean rating was highest for the kneeling position and lowest for the hip adducted position. They were significantly different on non-parametric tests. The mean ranks on Kruskal-Wallis test were 33.3, 65.6, and 91.6 for the hip adducted position, hip abducted position, and kneeling position respectively in the first session; and 34.0, 64.6, and 92.0 for the hip adducted position, hip abducted position, and kneeling position respectively in the replicate session. χ² tests were significantly different for both sessions at p<0.0001.

This study shows a strong effect of lower limb position on observer ratings of the ankle jerk in a sample of young healthy adults. This supports the notion that the position of the lower limb adopted to assess the ankle jerk may explain some of the differences in the prevalence of absent ankle jerk reported in several studies. The findings also agree with anecdotal reports5 that the sensitivity of the ankle jerk is higher when the hip is abducted than when the hip is adducted, but the kneeling position is the most sensitive of the three lower limb positions compared.

When screening for polyneuropathy, the kneeling position may be preferable to the other positions as it will reduce the number of false positives. However, the kneeling position will be unsuitable in settings where the patient is too ill or has impaired consciousness.4