Reliability determination of F mean response latency

Two hundred consecutive ulnar F wave latencies were evaluated in 30 healthy controls (C) and in 9 patients with mild neuropathy (N). The primary study objective was to determine the F wave sample size required to obtain a reliable F mean latency. A secondary objective was to explore the diagnostic utility of three F wave parameters: the minimal F wave latency (Fmin), the average F wave latency (Fmean), and the minimal-maximal F wave latency difference (Frange). For the control group, the Fmean reliability estimates as measured by intraclass correlation coefficients were 0.87, 0.97, and 0.99 with 1, 5, and 10 responses, respectively; for the N group, the corresponding values were 0.73, 0.93, and 0.96. Fmean was abnormal in 89% of the subjects with mild neuropathy whereas Frange and Fmin were abnormal in 67% and 33%, respectively. The study suggests (1) a reliable F mean latency can be obtained by averaging ten F responses; and (2) Fmean and Frange offers promise in increasing the diagnostic utility of the F response in identifying mild neuropathies.