1. Introduction
Ankle fractures are very common and account for a high number of annual emergency department visits and a large socioeconomic impact [
1,
2]. Of all fractures, ankle fractures represent 9% of all bone fractures, being the largest of the load-bearing joints [
3]. The most frequent types of ankle fractures are malleolar fractures, mainly unimalleolar and bimalleolar [
4,
5]. The incidence has been increasing over the last two decades to between 71 and 187 fractures per 100,000 people depending on age, sex, and geographic region [
5,
6,
7]. The main cause is low-energy indirect trauma in elderly and middle-aged women, mostly associated with osteoporosis. In young people, mainly males, they are caused by high-energy trauma during sport [
6]. Among the risk factors that can be associated with ankle fracture are the practice of sport itself [
8], diseases such as osteoporosis, diabetes, or obesity [
9], and lifestyle [
10].
One in four malleolar fractures will require surgical intervention [
3]; it is the treatment of choice if joint congruence cannot be restored conservatively. The most frequently used surgical treatment is open reduction with internal fixation (ORIF), the main purpose of which is to prevent post-traumatic arthritis and shorten immobilisation times [
11].
The importance of the severity of the injury, the surgical intervention, and the immobilisation time implies significant physical and biomechanical alterations [
12,
13]. These alterations are manifested by increased pain, swelling, stiffness, soft tissue involvement, atrophy, and decreased muscle strength [
14,
15,
16,
17]. Consequently, these impairments indirectly impact functional activities such as gait [
16,
18,
19] or limitations in work and leisure [
20].
Several studies have reported short- and long-term outcomes after ankle surgery [
21,
22,
23]. Beckenkamp et al. [
23] concluded in their meta-analysis that 6 months after surgery, patients recover very slowly, and even at 24 months they do not reach full recovery. However, the assessment of disability after an ankle fracture is often based on scales such as the American Orthopedic Foot and Ankle Score (AOFAS) [
24] or the Olerud–Molander Ankle Score (OMAS) [
25], which incorporate different factors to describe function, alignment, and pain. Although these rating systems provide a simple way to assess the degree of disability and mobility of the patient, many of these factors are assessed subjectively.
Complementary to these functional scales, clinical parameters such as ankle dorsiflexion range of motion (ADF ROM), level of swelling, and muscle atrophy provide objective measures of clinical status at different stages of rehabilitation [
18,
26,
27]. The recovery of these parameters after ankle surgery is critical and may play an important role in the full recovery of gait functionality [
28] or balance [
29].
Gait assessment has been widely used to characterise functional performance in patients who have sustained ankle injuries [
19,
27,
28,
30]; it is important for treatment decisions during rehabilitation [
31]. In addition to kinematic characteristics and spatiotemporal gait parameters [
32], analysis of plantar pressures allows us to identify subtle changes in foot loading that may go unnoticed on clinical examination [
33]. This analysis enables us to quantify the magnitude and distribution of the force applied to the surface of the foot during gait. This is of great importance, as variations in pressure are associated with alterations in the moments acting on the foot and ankle joints [
34,
35]. In fact, in the clinical context, studies on subjects who have suffered an ankle injury find an association between the plantar pressure pattern and the characteristics of the injury itself [
27,
36,
37,
38]. Furthermore, the correlation between dynamic plantar pressure and clinical measures such as ADF ROM in subjects with osteoarthritis [
36] or calf muscle atrophy in subjects with calcaneal fracture [
37] has been reported in the literature. However, this correlation has not been studied in subjects with bimalleolar ankle fracture and how these clinical measures may influence plantar pressure performance.
The aim of this study was firstly, to analyse dynamic plantar pressure, as well as gait cadence and gait speed in subjects with bimalleolar ankle fractures at 6 and 12 months after surgery; secondly, to evaluate clinical variables and functions such as ADF ROM, bimalleolar/calf circumferences, and the AOFAS and OMAS scales over the same period of time; and finally, to study the association and potential influence of gait and clinical-functional variables with the results of the dynamic plantar pressure analysis.
4. Discussion
The purpose of this study was to identify clinical and functional limitations in subjects with bimalleolar ankle fractures at 6 and 12 months after surgery. In order to quantify these limitations, we combined objective tests such as dynamic plantar pressure during gait analysis and clinical measures, together with functional scales. We compared the variables studied with a CG and evaluated the degree of correlation between these variables in the AFG, as well as their clinical implication. The results obtained in the dynamic plantar pressure analysis of the AFG (OA vs. NOA) and compared with the CG showed a decrease in plantar pressure (peak and mean) and an increase in contact time in the operated ankle. In addition, during gait analysis the cadence and gait speed of the AFG was significantly lower and clinically relevant when compared to the CG. Our research group previously published the full results of the study of spatiotemporal gait parameters and their relationship to clinical-functional measures at 6 months post-surgery [
28]. To our knowledge, we are not aware of any study analysing the dynamics of plantar pressures in subjects with bimalleolar ankle fractures 6 months after surgery. A study by Zhu et al. [
27] in twelve subjects after trimalleolar ankle fracture found similar results to ours in the dynamics of plantar pressures. Specifically, PPP was lower in the 3rd to 5th toe area when comparing OA/NOA and in the total plantar area when compared to a group of healthy subjects. However, differing from our findings, they found no difference in the CT between the OA and HA, but compared to the healthy group (189.86 ms;
p = 0.003). These differences found in CT compared to our work could be justified by the type of fracture, the time of measurement (4.5 months after surgery) or the significantly lower gait speed (0.65 m/s vs. 0.96 m/s). The results found in the dynamics of plantar pressures show the asymmetry between the two ankles and could be interpreted as a protective factor when walking. Studies carried out in different populations after ankle surgery found similar results in the study of plantar pressures [
38,
54,
55]. In relation to walking speed, studies in healthy subjects show a positive linear relationship with PPP [
56]. In the absence of studies in the ankle fracture population that relate gait speed to dynamic plantar pressure, this relationship could be extrapolated to our study population. However, in our investigation, we did not find a correlation of plantar pressures with speed, but we did find a negative correlation with ADF ROM and cadence with CT. Furthermore, speed and cadence were found to be the main predictor variables that conditioned the CT results by 76%. This influence of speed and cadence on CT could easily be predicted by fear of stepping with an injured ankle, pain, or ankle instability [
16,
27].
ADF ROM and ankle strength are the most studied clinical variables in the population that has suffered an ankle fracture [
57]. Complementary to these clinical measures, both bimalleolar circumference and calf circumference help us to estimate the degree of swelling and muscle atrophy after an ankle injury [
27,
37,
58,
59]. Of note in our study are the differences found between the OA and NOA in ADF ROM (−12.70;
p < 0.001), bimalleolar circumference (1 cm;
p < 0.001), and calf circumference (−1.3 cm;
p < 0.001). Furthermore, we found a negative correlation of plantar pressure (average and peak) with bimalleolar and calf circumferences, with bimalleolar circumference being the main predictor variable conditioning 23–30% of plantar pressure scores. These differences in clinical measurements are consistent with those reported in the literature in subjects with unimalleolar [
14], bimalleolar [
22], trimalleolar [
27], or calcaneal fractures [
37]; however, they do not study the correlation of these clinical measurements with plantar pressures. Adhesions and soft tissue involvement following injury and surgery lead to increased stiffness, decreased ankle range of motion, and calf muscle atrophy. As a consequence of these alterations, their influence on plantar pressures could be justified, just as it occurs during walking [
28].
Assessment of plantar pressure dynamics one year after surgery helps us to identify subtle differences in gait that may go unnoticed visually. In our investigation, in the AFG we found no differences between the two limbs in PPP, APP, and CT. However, when compared to the CG, the differences were significant in all three parameters. A study by Becker et al. [
60] in forty subjects evaluated at 18 months after ankle surgery found significant differences in plantar pressures when compared to a group of healthy subjects. However, in contrast to our findings, they found an asymmetry between the two limbs of the AFG. Furthermore, the same authors identified an association between poorer functional status and lower values of plantar pressures. Hirschmüller et al. [
37], in a group of sixty patients with intra-articular calcaneal fractures, found at 12 months after surgery a decrease in PPP in the hindfoot and an increase in the midfoot and lateral forefoot. However, in contrast to the work of Becker et al., there is no relationship between the functional scales and the dynamics of plantar pressures. In our work, we also found no such association with the AOFAS and OMAS scales; in contrast, we found a negative correlation between the circumferences (bimalleolar and calf) with PPP (r = −0.67; r = −0.61) and APP (r = −0.59; r = −0.52). This correlation suggests that the involvement of these clinical parameters strongly influences the pre-injury gait status. In particular, bimalleolar circumference could explain up to 43% of the results of plantar pressures.
Gait speed and gait cadence, as was the case at 6 months, showed significantly lower and clinically relevant values (1.06 ≤
d ≤ 1.21) compared to the control group. Moreover, both parameters still conditioned the outcome of the CT in 67%. In agreement with our results, Losch et al. [
61], in their study carried out with a sample of twenty subjects with ankle fractures (twelve bimalleolar) and at 12 months post-surgery, found a significant difference in walking speed when compared with a group of healthy subjects. On the other hand, van Hoeve et al. [
26] found similar results at 18 months post-surgery in thirty-three subjects with unstable ankle fractures (eleven bimalleolar). In contrast, Wang et al. [
30], in a group of eighteen subjects at 13 months after ankle surgery, found no such differences in gait speed when compared to healthy subjects. These contradictory results with respect to ours and other authors could be explained by the type of fracture (twelve subjects with a unimalleolar fracture and six with a trimalleolar fracture) or the protocol used in the gait analysis.
Improved ankle mobility is a key determinant in the full recovery of the affected limb. Some authors conclude that a decrease in ADF ROM of more than 4.5° in the push-off phase is clinically relevant [
26]. In our sample, we could not analyse the ankle range of motion during gait, but in the weight-bearing lunge test, we found a difference in ADF ROM of −7.4° (d = 1.18;
p < 0.001) between both ankles of the AFG and −12.2° (d = 1.48;
p < 0.001) compared to the CG. Nilsson et al. [
22], in their retrospective study conducted in fifty-four subjects 14 months after ankle fracture surgery, found a difference in ADF ROM of −5.7° (
p < 0.001) between the OA and NOA. In addition, they detected a significant difference in bimalleolar circumference of 1 cm (
p < 0.001). In contrast to our findings, they found no difference in calf circumference. These results seem to indicate that the degree of ankle swelling 12 months after surgery, together with the intrinsic characteristics of the fracture, would largely explain the increase in ankle stiffness and thus its impact on the functionality of the lower extremity. In this regard, it has been studied that a minimum of 30° of dorsiflexion under load is necessary to be able to perform tasks such as descending stairs, squatting, or getting into a chair without problems [
22].
In general, improvement in AFG between 6 and 12 months was evident in gait speed and cadence, dynamic plantar pressures, clinical measures, and functional scales. However, there are certain parameters whose improvement was not clinically relevant. Specifically, with regard to plantar pressures, the APP (−0.01 kPa/kg;
p = 0.193; d = 0.26) was similar at 6 and 12 months. Despite the slight improvement in PPP (−0.05 kPa/kg;
p = 0.023; d = 0.64) and CT (56.20 ms;
p = 0.024; d = 0.56), there seems to be an attempt to reduce weight bearing on the affected limb due to pain or fear of stepping. Some authors state that it is impossible to determine whether the changes in plantar pressure distribution are external consequences of intra-articular biomechanical alterations, in the sense of compensatory mechanisms, or an acquired pattern to alleviate pain [
60]. Another explanation would be given by limitations in ankle ADF ROM and/or lack of strength of the foot and limb musculature [
26].
Regarding the results of clinical measurements 6 months after the first measurement, we found a significant improvement of the ADF ROM (−6.8°;
p = 0.003; d = 0.73). However, the improvement in bimalleolar circumferences (0.2 cm;
p = 0.570; d = 0.12) and calf (−0.6 cm;
p = 0.160; d = 0.32) was very small and clinically not relevant. Nilsson et al. [
62] compared in one hundred and ten patients operated after ankle fracture two types of interventions (specific training programme and conventional physiotherapy) performed for 12 weeks. Among their results, they found no significant differences between groups in ADF ROM at 6 and 12 months. However, both groups improved by 2° from their initial values, although this improvement could be considered insignificant.
The improvement that we were able to see in the clinical assessment and certain gait parameters is also reflected in an increase in the AOFAS and OMAS scale scores. Specifically, the AOFAS scale reached 84.4 points compared to 73.6 points 6 months after surgery. This result is similar to that found by other authors at 12 and 18 months’ follow-up, with scores varying between 84 and 90 points [
13,
26]. At 12 months after surgery, pain is mild and occasional, the ADF ROM is not complete, and restrictions are practically limited to sports and recreational activity. To date, the minimum clinically relevant difference for ankle injuries is not clearly specified, although some authors propose a difference of 6 points on the AOFAS scale [
63]. Therefore, given the increase of almost 11 points between 6 and 12 months after surgery in our results, it can be considered clinically relevant.
The OMAS scale score also improved considerably, from 57.3 points at 6 months to 80 points at 12 months. Our findings are in line with those reported by several authors one year after surgery with OMAS scores between 75 and 85 points [
22,
30]. Among the results obtained, they highlight the residual pain that patients still experience when performing activities such as walking, climbing stairs, or squatting. In addition, only 19% of patients reported full recovery, with limitations in sport, activities of daily living, or work [
22].
The present study has certain limitations. Firstly, we have a relatively small sample, although, for the identification of clinically relevant intra- and inter-group differences, we found a moderate-to-large effect size (AFG d = 0.62; AFG/CG d = 0.93) with a power (1-β) of 80% and α = 0.05. Secondly, we have a CG that is smaller in number than the AFG. In this sense, we tried to make them as similar as possible in terms of age, sex, weight, and height. This may be a biasing factor that slightly modifies the characteristics observed between groups. On the other hand, whether the ankle fracture occurred on the dominant or non-dominant leg was not taken into account, which could be an effect-modifying factor. Finally, we only analysed dynamic plantar pressure at a global level. Future studies should include a regional analysis of the footprint during gait, at different stages of rehabilitation, in conjunction with other clinical-functional measures.