3.4.2. Spatiotemporal Gait Parameters (Chronic Adaptations)

#### Walking Speed

A significant time and time × duration interaction effect was found with regard to the walking speed (Table 3). Multiple comparison analyses revealed that walking speed was significantly increased following SET during the pain-free and painful walking conditions (*p* ≤ 0.026, Table 3).

#### Stride Length

A significant time and time × duration interaction effect was found regarding the stride length (Table 3). Multiple comparison analyses showed that stride length was significantly increased following SET (time effect: *p* = 0.013). Compared to the values recorded before SET, stride length was significantly increased following SET during the pain-free walking condition alone (*p* = 0.001), whereas no significant differences were observed during the painful walking condition (*p* = 0.569).

## Stride Duration and Frequency

Following SET, stride duration and frequency significantly increased (time effect: *p* ≤ 0.001) with no significant time × duration interaction effect (Table 3).


**Table 3.** Spatiotemporal gait parameters in the most symptomatic leg during the 6 min walk test before and after supervised exercise training (SET).

Ten consecutive strides were analyzed during pain-free walking (pain-free) and during painful walking at the end of the 6 min walk test (pain). Bold *<sup>p</sup>* value is statistically significant. \* *<sup>p</sup>* <sup>≤</sup> 0.05 for significant difference compared to pain-free. # *<sup>p</sup>* <sup>≤</sup> 0.05 for significant difference compared to before SET. £ *<sup>p</sup>* <sup>≤</sup> 0.05 for significant difference to pain-free within after SET condition.

#### Stance and Swing Phase

After SET, the relative duration of the stance and swing phase was unchanged (time effect: *p* = 0.431); however, a significant time × duration interaction effect was observed (Table 3). Multiple comparison analyses revealed a significant increase in the relative duration of the stance phase during the painful condition compared to the pain-free condition following SET alone (*p* = 0.008). Similarly, a significant decrease in the relative duration of the swing phase was observed during the painful condition compared to the pain-free walking condition following SET alone (*p* = 0.008).

#### Inner-Stance Phases

A significant time and time × duration interaction effect was observed regarding the relative duration of the loading response (Table 3). Multiple comparison analyses revealed a significant increase in this parameter following SET (time effect: *p* = 0.013). After SET, there was a significant increase in the relative duration of the loading response during the pain-free walking condition alone (*p* = 0.001), whereas no significant differences were observed during the painful walking condition (*p* = 0.523).

The relative duration of the foot-flat phase was unchanged after SET (time effect: *p* = 0.139); however, a significant time × duration interaction effect was observed (Table 3). Multiple comparison analyses revealed a significant decrease following SET during the pain-free walking condition alone (*p* = 0.002), whereas no significant differences were observed during the painful walking condition (*p* = 0.420).

No significant time and time × duration interaction effect was observed for the relative duration of the push-off or for the double support phases (Table 3).

#### *3.5. Foot Kinematics Parameters (Acute Adaptations)*

During the 6MWT, all the foot kinematics parameters showed a significant duration effect, except for the first maximal toe clearance (Table 4). Multiple comparison analyses showed that the heel-strike pitch angle, toe-off pitch angle, maximal heel clearance, second maximal toe clearance, and minimal toe clearance significantly decreased during the transition from the pain-free to the painful walking condition during the 6MWT (Table 4).


**Table 4.** Foot kinematics in the most symptomatic leg during the 6 min walk test before and after supervised exercise training (SET).

Ten consecutive strides were analyzed during pain-free walking (pain-free) and during painful walking at the end of the 6 min walk test (pain). Bold *<sup>p</sup>* value is statistically significant. \* *<sup>p</sup>* <sup>≤</sup> 0.05 for significant difference compared to pain-free. # *<sup>p</sup>* <sup>≤</sup> 0.05 for significant difference compared to before SET.

#### *3.6. Foot Kinematics Parameters (Chronic Adaptations)*

The toe-off pitch angle was unchanged after SET (time effect: *p* = 0.356); however, a significant time × duration interaction effect was observed (Table 4). Multiple comparison analyses revealed that, compared to before SET, the toe-off pitch angle significantly increased following SET during the pain-free walking condition alone (*p* = 0.05), whereas no significant differences were observed during the painful walking condition (*p* = 0.938). No significant time and time × duration interaction effect was observed regarding the heel-strike pitch angle, maximal heel clearance, second maximal toe clearance, or minimal toe clearance (Table 4).

#### **4. Correlations**

The relationships between gait pattern changes during the pain-free walking condition and changes in 6MWD following SET are displayed in Table 5. A significant positive correlation was found between changes in stride length, stride frequency, and second max toe clearance during the pain-free walking condition and changes in 6MWD (Table 5). On the other hand, a significant negative correlation was found between changes in the duration of the foot-flat phase during the pain-free walking condition and changes in 6MWD (Table 5).

**Table 5.** Relationship between spatiotemporal gait and foot kinematics changes during pain-free walking condition and changes in 6 min walking distance following supervised exercise training.


Bold *p* value is statistically significant (*p* ≤ 0.05). All correlations were controlled for gait baseline values.

#### **5. Discussion**

The results of this study partially confirm our hypotheses: (1) SET improved the 6MWD in patients with symptomatic PAD; (2) following SET, walking speed, stride frequency and stride length were significantly greater during the 6MWT. However, stride length was significantly increased following SET during the pain-free walking condition alone, whereas no significant differences were observed during the painful walking condition. Similarly, changes in the relative duration of the inner-stance phases (loading response and foot-flat) and the toe-off pitch angle following SET were observed during the pain-free walking condition alone; (3) during the transition from the pain-free to the painful walking condition, the spatiotemporal gait and foot kinematics parameters were shown to undergo a similar evolution before and after SET during the 6MWT. Finally, our results showed that changes in stride length and frequency and in the relative duration of the foot-flat phase during the pain-free walking condition were related to changes in functional walking performance during the 6MWT following SET.

The results of the present investigation confirm previous findings, which showed that SET improves 6MWD in symptomatic patients with PAD [9,23]. We observed a ~43 m improvement in 6MWD, which was greater than the substantial meaningful change of +20 m [43] or +35 m [44] previously observed in these individuals. The greater improvement in 6MWD observed in the present investigation may have been related to the training characteristics of the multimodal SET program. Indeed, the patients combined the strengthening of the lower limbs with Nordic walking, which are both functional training modalities. This type of training likely led to better improvements in functional walking performance. The greater improvement in 6MWD could also be related to the 50 m course length, as previous studies showed that longer course lengths were associated with greater walking distances [45]. By contrast, this improvement was similar to the minimal detectable change of >46 m recently observed in patients with claudication [46]. Taken together, these results suggest that multimodal SET is effective at improving functional walking performance in patients with symptomatic PAD [33,36,37].

During the transition from the pain-free to painful walking conditions, similar acute adaptations were observed for the spatiotemporal gait and foot kinematics parameters during the 6MWT before and after SET. These results extend previous findings observed during constant-speed treadmill exercises in patients with symptomatic PAD [13,29] and highlight that similar acute gait adaptions also occur during the 6MWT, which is a more functional form of walking that represents daily life more accurately [21]. Previous studies have shown that, when compared to aged-matched individuals, gait abnormalities exist from the first step taken (pain-free), suggesting muscle metabolic myopathy in patients with PAD [15,47,48]. Gait worsening was also documented once leg claudication pain was established, highlighting the role of muscle ischemia on gait pattern changes during exertion in these individuals [15]. Our results are in line with these findings. We observed that the walking speed, stride duration, stride frequency, stride length, relative duration of swing, loading response duration, and push-off phase duration decreased (pain-free > end), whereas the duration of the stance, foot-flat and double support phases significantly increased (pain-free < end) during the transition from the pain-free to the painful walking condition during the 6MWT. The extended duration of the stance and the foot-flat phases during exertion may ameliorate the balance between oxygen supply and demand in the active ischemic calf musculature [13,29]. It is also possible that patients adopt this pattern to improve their stability during painful walking [13,29]. The reduced duration of the push-off phase during the transition from the pain-free to the painful walking condition may be related to exercise-induced ischemia, which may lead to calf muscle strength deficit and affect forward propulsion [8]. Consequently, this may also affect walking speed,

stride frequency and stride length, and foot kinematics, especially during the 6MWT, where patients are allowed to choose their own walking pace.

In current research, there are a limited number of studies regarding gait pattern changes following exercise interventions in patients with symptomatic PAD, and the findings are inconsistent. Indeed, some studies [27,29,30,49], but not others [24,25,28], observed significant gait changes following SET. The results of the present investigation showed that gait pattern was modified in patients with symptomatic PAD during the 6MWT following multimodal SET. It is, however, interesting to note that many of these changes occurred during the pain-free walking condition alone. Indeed, although walking speed and stride frequency increased following SET, stride length was significantly increased following SET during the pain-free walking condition alone, whereas no significant differences were observed during the painful walking condition. These findings indicate that the increased walking speed observed during the painful walking condition following SET is mainly related to an increased stride frequency rather than increased stride length. Similarly, the relative duration of the loading response phase increased, and the relative duration of the foot-flat phase decreased following SET during the pain-free walking condition alone. The toe-off pitch angle was also increased following SET, but again, during the pain-free walking condition alone. These observations are in contrast to previous findings that show the relative duration of the foot-flat phase was increased during constant-speed maximal treadmill exercises following SET in patients with symptomatic PAD [29]. A possible explanation is that this may have been due to the testing protocol used to assess gait changes following SET. Indeed, compared to the constant-speed treadmill test, walking speed during 6MWT exhibited different values before and after SET. Following SET, the patients demonstrated an improved dynamic balance, which allowed them to walk faster during the 6MWT, causing a reorganization of the durations of the loading response (increased) and foot-flat (decreased) phases during the pain-free walking condition. This is in line with previous observations, which showed that the duration of the foot-flat phase was negatively correlated to walking speed in both PAD and non-PAD individuals [13]. Once claudication pain began and worsened to moderate-to-maximal levels, the walking speed decreased during the 6MWT. Interestingly, the relative durations of the loading response and the foot-flat phases returned to the pre-SET values despite the greater walking speed in the post-SET condition. This suggests that factors other than the walking speed are related to gait pattern changes. These findings indicate the potential role of exercise-induced ischemia and claudication pain on gait adaptations during exertion in patients with symptomatic PAD.

The use of non-invasive inertial sensors with the aim to investigate gait pattern during physical assessment has potential applications with regard to the optimization of the prescription of training in patients with PAD. Indeed, these inertial sensors may easily assess gait pattern evolutions during functional acute exercise performed before and following an exercise training program. This technology allows one to evaluate the gait changes in the transition from the pain-free to painful walking condition, and therefore produces a valid description of daily-life walking pattern in these individuals. In addition, by evaluating the potential correlation between gait pattern and functional performance changes following rehabilitation, specific training approaches could be conceived to optimize patients' benefits. Interestingly, our results showed that the changes in stride length and frequency during the pain-free walking condition were positively correlated to changes in 6MWD. In addition, the changes in the relative duration of the foot-flat phase during the pain-free walking condition were negatively correlated with changes in 6MWD. These findings suggest a link between changes in gait pattern during the pain-free walking condition and improved functional walking performance in patients with symptomatic PAD. These results feature important clinical implications and indicate the need for further investigations regarding the effects

of specific gait training modalities on gait pattern and its relation to functional walking performance in these individuals. A previous meta-analysis showed that walking training with cueing of cadence improves spatiotemporal gait parameters more than walking training alone in older patients with cardiovascular disease [50]. Walking training with cueing of cadence, which was usually 5–10% greater than comfortable cadence, improves walking speed, stride length and frequency, and walking symmetry in patients who have experienced a stroke [50]. Based on the gait abnormalities previously observed in patients with PAD [13,15,17,18], these specific gait training modalities could be promising with regard to improving gait pattern and (functional) walking performance in these individuals.

This study featured some limitations. First, the present investigation lacked a control group that did not participate in the 3 month SET. Even though previous findings showed no difference in gait pattern over time in patients with PAD who did not participate in a vascular rehabilitation program [24], future randomized controlled trials are needed to better investigate gait changes following training interventions in these individuals. Second, because of the descriptive nature of our results, it was not possible to elucidate the mechanisms related to our observations. More detailed kinetics and kinematics gait analyses are needed to better describe gait pattern before and after SET. Third, even though it was used in previous works, the inertial system used in the present investigation has never been validated in patients with PAD.

In conclusion, these results show that multimodal SET modifies gait pattern during the 6 min walk test in patients with symptomatic PAD. However, many of these changes (stride length, the relative duration of the loading response and foot-flat phases, and toe-off pitch angle) only occurred during the pain-free walking condition, highlighting the role of claudication pain in gait pattern in this population. In addition, changes in stride length and frequency and in the relative duration of the foot-flat phase during the pain-free walking condition were correlated with changes in 6 min walking distance. These findings suggest that new rehabilitation strategies, including specific gait training modalities, should be further investigated in this population.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/s21237989/s1.

**Author Contributions:** Conceptualization, S.L., D.M.; methodology, S.L., J.B., L.C., L.M. and D.M.; software, S.L., J.B. and D.M.; validation, S.L., J.B., L.C., L.M. and D.M.; formal analysis, S.L., J.B. and D.M.; investigation, S.L., J.B. and L.C.; resources, S.L., J.B., L.C., L.M. and D.M.; data curation, S.L., J.B. and D.M.; writing—original draft preparation, S.L. and D.M.; writing—review and editing, S.L., J.B., L.C., L.M. and D.M.; visualization, S.L., J.B., L.C., L.M. and D.M.; supervision, S.L., L.M. and D.M.; project administration, S.L. and D.M.; funding acquisition, none. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** This study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the local ethics committee (study number: 2016-01135).

**Informed Consent Statement:** The subjects provided written, voluntary, informed consent.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.
