Physiotherapy is recommended by the International Consultation on Incontinence (ICI) as an effective first-line therapy in the treatment of all forms of female urinary incontinence. However, standard physiotherapy focuses on voluntary activation of pelvic floor muscles (PFM) (1). Involuntary muscle activation is usually part of the training protocol in sports, especially when fast, reflex activation is a specific requirement of the athletic discipline. As women with stress urinary incontinence (SUI) lose urine during high impact activities like jogging, a rapid, involuntary activation of the PFM is essential to avoid leakage. Therefore, in the present study a new physiotherapy program was developed, focusing on involuntary reflexive PFM contractions. The intervention lasted 16 weeks including nine personal physiotherapy consultations and home training sessions.
The aim of the study was to compare the effect of this new physiotherapy program with a standard physiotherapy treatment on PFM activation patterns and fiber type recruitment behavior while running. Task-specific fiber type recruitment can be estimated with wavelet analyses of electromyograms (EMG) since small α-motoneurons (α-MN) generate lower frequencies in the signal (<100 Hz) and large α-MN higher frequencies (>100 Hz). Furthermore, the activation intensity and frequency content in the range of a few milliseconds with high time resolution can be identified (2).

In this prospective, triple-blinded RCT, women with SUI were randomly allocated to the control group (CON) (standard physiotherapy) or the experimental group (EXP) (standard physiotherapy plus involuntary, reflexive PFM training). The study followed the CONSORT flow diagram. Sample size was calculated based as follows: Effect size = 0.1, α = 0.05, power (1–β error probability) = 0.8, number of groups = 2, number of measurements = 10, estimated correlations among repeated measures = 0.5, + 10% dropouts + 10% unforeseen events. This resulted in a calculated sample size of N = 96.
PFM electromyography (EMG) was measured at baseline and after intervention phase. PFM EMG was recorded during 10 seconds at 7, 9 and 11 km/h treadmill running and analyzed with Morse wavelets. Raw EMG was visually checked for plausibility. The relative distribution of power (%) over six frequency bands (20-200 Hz) was extracted and analyzed within six time intervals of 30 ms, from -30 ms before to 150 ms after initial contact (3). A two-factorial (group and time intervals) analysis of variance for repeated measures and a post-hoc t-test were performed to identify power spectra differences pre and post intervention, within and between the two groups. The significance level was set at p ≤ 0.05.

Ninety-six participants could be assigned to randomization and group allocation (EXP, n = 48; CON, n = 48. Baseline assessment of 45 (EXP) and 47 (CON) participants was completed (4 participants were lost after randomization). After EMG plausibility check, 77 EMG data sets could be analyzed (EXP, n = 38; CON, n = 39). Missing data within a valid data set were replaced by the last valid data based on the intention-to-treat and last observation carried forward approach.
The power spectra of each time interval showed no statistically significant group differences and no difference before and after intervention.
Lowest frequency band 20-50 Hz: The time intervals from 30 ms before to 30 ms after initial contact showed significantly less signal power than the intervals from 30-150 ms after initial contact for all running speeds and both groups. Highest frequency bands 140-200 Hz: The time intervals from 30 ms before to 30 ms after initial contact showed significantly more signal power than the time intervals from 30-150 ms after initial contact for all running speeds and both groups.

No significant differences in EMG power spectra at baseline and after intervention within nor between groups were found. A possible reason could be that the intervention phase of 16 weeks was rather short for a change in fiber recruitment. Furthermore, the involuntary part of the training program started after 6 weeks because all participants received a basic sensorimotor and behavioral training at the beginning. Compared to training methods in sports, the applied training parameters such as high intensity and maximum exhaustion are not as easy applicable to PFM as to skeletal muscles of the extremities. However, wavelet analyses were sensitive enough to identify differences in the EMG power spectra in the pre- and post-initial contact phase. Muscular preparation and adaptation a few milliseconds before initial contact could be shown. The evaluated specific differences of EMG spectra in the pre-initial and the post-initial contact phase seem to indicate two different muscle activation events. Power spectra shifts towards higher frequency bands in the pre-initial contact phase could indicate a feed-forward anticipation and a muscle tuning for the expected impact of initial contact event in order to maintain continence.

Wavelet analyses are suitable for identifying specific differences power spectra shifts of PFM while running and this in the range of only a few milliseconds. In biomechanics, wavelet analyses are commonly used methods. In the future, it would be meaningful to use wavelets for PFM EMG analyses to assess differences of PFM activation patterns and thus of specific continence mechanism.

Bo K. Pelvic floor muscle training in treatment of female stress urinary incontinence, pelvic organ prolapse and sexual dysfunction. World journal of urology. 2012;30(4):437-443.
von Tscharner V, Goepfert B. Estimation of the interplay between groups of fast and slow muscle fibers of the tibialis anterior and gastrocnemius muscle while running. Journal of electromyography and kinesiology. 2006;16(2):188-197.
Fleischmann J, Gehring D, Mornieux G et al. Task-specific initial impact phase adjustments in lateral jumps and lateral landings. European Journal of Applied Physiology. 2011;111:2327-2337