*3.1. Gait Spatiotemporal Parameters*

Significant main effects were found for stride length and walking speed (Table 2). Post-hoc tests revealed significantly higher stride length for PHS compared with NHS. Furthermore, post-hoc tests revealed significantly lower walking speed for NHS when compared with NS and PHS. No significant differences were found in step frequency and double support time/single support time.

**Table 2.** Mean values, standard deviations, and results of the repeated measures ANOVA for spatiotemporal parameters.


Note: NHS: negative heel shoes, NH: normal shoes, PHS: positive heel shoes. Post-hoc significant differences are marked with a (vs. NHS), b (vs. NS), c (vs. PHS).

#### *3.2. GRF*

SPM analysis with repeated measures ANOVA revealed a significant difference between shoe conditions in GRF (Figure 2). Post-hoc analysis shows the NHS's AP-GRF was smaller than NS at 77.1–90.3% and 94.6–100% of the stance phase (*p* < 0.001). The AP-GRF of NHS was smaller than PHS and was significant at 25–33.3%, 82.0–90.3%, and 94.6–100% of the stance phase (*p* < 0.001).

The post-hoc analysis results showed that the ML-GRF of NS was significantly larger than PNS during the stance phase (14.5–17.7%; 71.9–82.6%) (*p* < 0.001). At the 1.3–9.4% stance phases, the ML-GRF of NS was significantly more significant than the NHS (*p* < 0.001). At 1.3–4.6% and 8.8–10.3% of the stance phase, the ML-GRF PHS was significantly greater than NHS (*p* < 0.05).

The post-hoc analysis results showed that the vertical GRF of NHS in the third trimester of pregnancy was significantly larger than NS during the gait stance phase (91.8–100%) (*p* < 0.001). At 66.3–72.4% of the stance phase, PHS was significantly lower than the NS (*p* < 0.001). The vertical GRF of NHS was larger than that of PHS during 40–44%, 61–71.5%, and 92.5–100% of the stance phase (*p* < 0.001).

**Figure 2.** Ground reaction forces in anterior–posterior, medial–lateral and vertical directions (mean and SD) between NHS, NS, and PHS. The grey areas indicate significant differences between conditions, followed by the time-dependent F-values of the SPM. Colored bars beneath each plot indicate significant differences between waveforms, whereas the red, blue, and black bars represent significant differences for NS vs. NHS, NHS vs. PHS, and NS vs. PHS, respectively.

#### *3.3. COP Trajectory*

As shown in Figure 3, the results showed no difference in ML-COP between NHS, NS, and PHS. For AP-COP, there was a main effect. Post-hoc analysis showed that NHS demonstrated a significantly smaller range of AP-COP in NHS vs. NS for 19–90.5% and 93–100% of the stance phase. At 14.5–53.1%, 68.5–90% and 93.5–100% of the stance phase, the NHS posterior COP was significantly smaller than PHS (*p* < 0.05).

#### *3.4. COM-COP Inclination Angles*

No significant differences were found in step peak medial angles and peak anterior angles. Significant main effects were found for peak posterior angles (Table 3). Posthoc tests revealed significantly lower peak posterior angles for NHS compared with NS and PHS.

**Table 3.** Mean values, standard deviations, and results of the repeated measures ANOVA for peak COM-COP inclination angle.


Note: NHS: negative heel shoes, NH: normal shoes, PHS: positive heel shoes. Post-hoc significant differences are marked with a (vs. NHS), b (vs. NS), and c (vs. PHS).

**Figure 3.** Mean cop trajectories in the x-time and y-time planes. The grey areas indicate significant differences between conditions, followed by the time-dependent F-values of the SPM. Colored bars beneath each plot indicate significant differences between waveforms, whereas the red, blue, and black bars represent significant differences for NS vs. NHS, NHS vs. PHS, and NS vs. PHS, respectively.

#### **4. Discussion**

The primary purpose of this study was to investigate the differences in gait spatiotemporal parameters, 3D-GRF, and COP of the condition of NHS, NS, and PHS in the third trimester of pregnancy. Compared with PHS and NS, pregnant women wearing NHS showed a more stable gait posture in the anterior–posterior direction, with slower walking speed and smaller peak posterior COM-COP inclination angles.
