Diagnosing Lumbar Spinal Stenosis—Reply

In Reply: We believe that the letter by Dr Watters and colleagues is misleading on a number of points. First, the writers cite a single retrospective study by Hamanishi et al¹ to counter 3 more recent prospective controlled trials^2- 4 cited in our Commentary. The study by Hamanishi et al used a convenience sample, unmasked investigators, and a control population of 15 persons with clinical indications for MRI. The imaging technology used in 1990-1994 included slices of 5 to 7 mm, used transverse slices that were not perpendicular to the spinal canal, and did not scan all levels of interest in 18 of 53 claudication subjects. Contrary to the assertions of Watters et al, the study by Haig et al² reported briefly in its text and detailed in its eAppendix measures of thecal sac area and diameter, spinal canal area and diameter, and various measures of the lateral recess at 4 lumbar levels, along with diverse constructs of “2 smallest” and “average,” none of which differentiated asymptomatic volunteers from those thought by clinicians to have stenosis.