Clinical Question: Are off-the-shelf magnetic insoles more effective than nonmagnetized insoles for decreasing foot pain in adults with nonspecific, work-related foot pain?

Setting: Outpatient (any)

Study Design: Randomized controlled trial (double-blinded)

Allocation: Concealed

Synopsis: Viscoelastic shoe insoles have been shown to decrease nonspecific foot pain in patients who have to stand as part of their job duties. These researchers sought to determine whether a viscoelastic insole with magnets (Spenco active comfort magnetic insole) would produce a greater effect. They enrolled 89 patients (79 percent women) who responded to an advertisement soliciting adults with foot pain aggravated by standing or walking. The pain must have been present for more than 30 days. The authors excluded patients with neurologic deficit and those who had plantar fasciitis. The patients were randomly assigned, using concealed allocation, to wear the magnetic insoles or sham insoles for eight weeks. The static magnets had a bipolar multiple circular array charged to 2,450 G with the south poles directed to the foot, although the surface strength was only 192 G. Patients were asked not to attempt to determine whether they had the magnetic insoles, and only 27 percent of patients who had them guessed correctly at the end of the study. The number of patients reporting their response to treatment as “all better” or “mostly better” after four and eight weeks was no different among the two groups (33 versus 32 percent at four and at eight weeks). Pain scores on a 10-cm visual analogue scale also were no different, with both decreasing by approximately 2.7 cm during the study.

Bottom Line: Typical cushion insoles with magnets are no more effective than cushion insoles without magnets to relieve pain in persons who stand on their feet to work. Another study of these insoles also found no benefit in treating heel pain (Winemiller MH, et al. Effect of magnetic vs sham-magnetic insoles on plantar heel pain: a randomized controlled trial [published correction appears in JAMA 2004;291:46]. JAMA 2003;290:1474–8). (Level of Evidence: 1b)