Gestational diabetes mellitus (DMG) has been associated with higher prevalence of both pregnancy-specific urinary incontinence (PS-UI) and urinary incontinence (IU) postpartum, with worsening of severity and quality of life during pregnancy and over first year postpartum compared to non-GDM women.[1] Experimental studies in moderate diabetic rat models have shown that the periurethral and rectus abdominis muscles present deterioration, such as atrophy, thinning, disorganization, and co-localization of fast and slow fibers.2 These data are consistent with those observed in rectus abdominis muscles tissue collected from pregnant women with GDM during C-section, which suggests that GDM is indeed capable of damaging the muscular tissue causing a myopathic process.[2] Due to the invasive nature of PFM biopsy, functional tests have been employed to evaluate the impact of GDM on its function. The aim of this longitudinal study was to compare PFM activation pattern between GDM and non-GDM women from 24-30 gestational weeks to 18-24 months postpartum during a standard clinical test during gestation and postpartum.

This was prospective cohort study approved by the Institutional Ethical Committee (Protocol Number CAAE 82225617.0.0000.5411). The main inclusion criteria were: pregnant women between 24-30 weeks of gestation in the first assessment; singleton pregnancy; 18–40 years of age; primigravida or primiparous with previous c-section. The participants were allocated in GDM group if they presented fasting glycemic levels ≥92 mg/dL or 1 hour ≥180 mg/dL or 2 hours ≥153 mg/dL. In addition, participants who had lower levels composed the Non-GDM group. Participants were evaluated at three time points: 24–30 weeks of gestation (T1), at 36–38 weeks of gestation (T2) and 18-24 months postpartum (T3). The same procedures were followed at each time point. During first step of the investigation the participants answered a questionnaire, followed by instructions about pelvic floor contraction and vaginal palpation to confirm if they were able to isolate the contraction. After the confirmation that they were prepared to perform the main acquisition, the EMG data collection were performed. The PFM EMG assessment were done according to the Glazer protocol of clinical evaluation. The EMG signal were captured by two-channel EMG device (Miotool 200 Uro; Porto Alegre, Brazil). A water-soluble gel was applied before introducing the probe into the vaginal canal. The EMG signals were processed offline using custom programs implemented in MATLAB (2014b, The MathWorks, Inc., Natick, MA, USA). The EMG profiles were obtained by applying the root mean square (RMS) the entire signal using a sliding window of 200 msec. Consistent with previous studies using the same protocol, the RMS EMG profiles were then normalized by the highest peak detected across the 5 repetitions of the Flick task. To identify the PFM pattern during the entire contraction, the full RMS EMG waveforms from the Flick and Hold tasks were compared between groups using the technique of wavelet-functional ANOVA (wfANOVA).[3] As we were interested in both the phasic activation patterns and the rest amplitudes before and after each contraction, we selected time windows that included 3 seconds before and after each contraction.

The EMG analysis were proceeded with participants who had all time-points completed and with good EMG signal quality (19 non-GDM and 14 GDM). No significant group differences related to demographic and personal data were found in participant characteristics during gestation or postpartum (Table 1). The glucose tolerance test values, as expected, showed marked group differences on fasting, 1 and 2 hours after oral glucose tolerance test (OGTT).
Figure 1 shows the results of the wf-ANOVA analysis, with the average EMG patterns of each group and the significant Group contrasts during the Flick and Hold PFM contraction tasks at each time point. The significant contrasts indicate that, during the Flick contractions, the GDM group generally had smaller PFM EMG amplitude than non-GDM after ~1 second of contraction, suggesting shorter contractions. During the 10-sec Hold contractions, the non-GDM group activated the PFM at higher contraction intensities than the GDM group at both time points T2 and T3, although the timing of the contrasts differed between time points: At T2, the GDM group had lower initial peak amplitude during Hold, but similar amplitudes after ~2 seconds of contraction; at T3, the initial peaks from both groups had similar (normalized) amplitudes, after which the levels of PFM activation decreased faster for the GDM group, remaining lower than the non-GDM group until near the end of the contraction.

This is an unprecedented study which assessed PFM EMG patterns from pregnancy to long-term post-partum (18-24 months) in women with and without GDM. Using a well-stablished protocol for pelvic floor assessment, we reproduced a similar sequence of PFM contractions requested in clinical consultations, commonly used to identify the motor strategy during brief and sustained PFM tasks. Wavelet analysis showed that, although the GDM group achieve peak PFM EMG amplitudes similar to the non-GDM, they took longer to return to baseline levels. During 10-sec Hold contractions, the GDM group sustained lower levels of PFM activation than the non-GDM group at both T2 and T3. The impairments in PFM function observed in women with GDM have been attributed to physiological and anatomical changes to the musculoskeletal system, namely reduced cross-sectional area and reduced number of fast fiber type, in addition to impairments in ionic channels, as well as fat infiltration and proliferation of connective tissue in the PFM.

This novel cohort study evaluated PFM activity in pregnant women with and without GDM at three distinct time points during and after delivery. Taken together, these results suggest that differences on motor behaviour of GDM women arises in late pregnancy and exacerbate on postpartum.

iculo F, Marini G, Vesentini G, Morceli G, Damasceno DC, Sobrevia L, et al. Pregnancy-specific urinary incontinence in women with gestational hyperglycaemia worsens the occurrence and severity of urinary incontinence and quality of life over the first year post partum. Eur. J. Obstet. Gynecol. Reprod. Biol. [Internet]. 2020;252:336–43. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0301211520304036
Vesentini G, Barbosa AMP, Damasceno DC, Marini G, Piculo F, Matheus SMM, et al. Alterations in the structural characteristics of rectus abdominis muscles caused by diabetes and pregnancy: A comparative study of the rat model and women. PLoS One [Internet]. 2020;15:e0231096. Available from: https://dx.plos.org/10.1371/journal.pone.0231096
McKay JL, Welch TDJ, Vidakovic B, Ting LH. Statistically significant contrasts between EMG waveforms revealed using wavelet-based functional ANOVA. J. Neurophysiol. [Internet]. 2013;109:591–602. Available from: https://www.physiology.org/doi/10.1152/jn.00447.2012

Continence 2S2 (2022) 100353
DOI: 10.1016/j.cont.2022.100353