The initial effects of a Mulligan’s mobilization with movement technique on dorsiflexion and pain in subacute ankle sprains

Natalie Collins, Pamela Teys, Bill Vicenzino*
Department of Physiotherapy, The University of Queensland, St. Lucia, Brisbane, QLD 4072, Australia
Received 17 December 2002; received in revised form 25 July 2003; accepted 21 August 2003


The lateral ligament complex of the ankle, described as the body’s ‘‘most frequently injured single structure’’ (Garrick, 1977), is mechanically vulnerable to sprain injury. At extremes of plantarflexion and inversion, influenced by the shorter medial aspect of the ankle mortise, the relatively weak anterior talofibular ligament (ATFL) and calcaneofibular ligament (CFL) are prone to varying grades of rupture, often via minimal force (Hockenbury and Sammarco, 2001). Immediate inflammatory processes produce acute anterolateral pain and oedema, with avoidance of movement and weight bearing (Wolfe et al., 2001). Subsequent losses of joint range, particularly dorsiflexion, and muscle strength results in significant gait dysfunction. Recent data from our laboratory highlights the presence of a dorsiflexion deficit not only in the acute stage, but also in the subacute stage (Yang and Vicenzino, 2002). Early physiotherapy intervention consists of rest, ice, compression, elevation (RICE) and electrotherapy modalities to control inflammation, as well as manipulative therapy and therapeutic exercise techniques to address impairments of movement and strength (Wolfe et al., 2001; Hockenbury and Sammarco, 2001). Green et al. (2001) investigated the impact of combining nonweight-bearing talocrural anteroposterior (AP) passive mobilisations, believed to restore dorsiflexion range, with the RICE protocol in the treatment of acute ankle sprains. The experimental group ðn ¼ 19Þ demonstrated a more rapid improvement in pain-free dorsiflexion and function than the control group ðn ¼ 19Þ who were treated solely with RICE. This provides important evidence substantiating the role of passive joint mobilizations in an acutely injured population. The mobilization with movement (MWM) treatment approach for improving dorsiflexion post-ankle sprain combines a relative posteroanterior glide of the tibia on talus (or a relative anteroposterior glide of the talus on the tibia) with active dorsiflexion movements, preferentially in weight bearing (Mulligan, 1999). Claims of rapid restoration of pain-free movement are associated with MWM techniques generally (Mulligan, 1993, 1999; Exelby, 1996). Through examination of the effects of MWM on ankle dorsiflexion in asymptomatic mildly restricted ankle joints, Vicenzino et al. (2001) found that both the weight bearing and non-weightbearing variations of the dorsiflexion MWM technique produced significant gains in dorsiflexion range. However, weight-bearing treatment techniques are widely believed to be superior to non-weight-bearing techniques, as they replicate aspects of functional activities (Mulligan, 1999). Acute ankle sprains, whilst having marked reduction in dorsiflexion range of motion, are frequently painful in full weight bearing, and weightbearing techniques are not clinically indicated. The subacute ankle sprain is characterized by significant residual deficits in dorsiflexion (Yang and Vicenzino, 2002) and the capacity to fully weight bear, making it a good model on which to study the initial effects of weight-bearing MWM on dorsiflexion. The mechanism of action of manipulative therapy has been the focus of several reports in recent times, however spinal manipulative therapy appears to be the common subject of research. A synopsis of current evidence for the initial mechanism of action of manipulative therapy indicates in part a neurophysiological basis (Vicenzino et al., 1996, 1998, 2000). Manipulative therapy treatment techniques studied have exhibited non-opioid hypoalgesia to mechanical but not thermal pain stimuli (Vicenzino et al., 1995, 1998). The primary objective of this study was to test the hypothesis that application of Mulligan’s MWM technique for talocrural dorsiflexion to subacute lateral ankle sprains produces an initial dorsiflexion gain, and simultaneously produces a mechanical but not thermal hypoalgesia.


The double-blind randomized controlled trial incorporated repeated measures into a cross over design, in which each participant served as their own control.

Participants: Sixteen participants, eight males and eight females aged 18–50 (average 28.25 years and standard deviation 9.33 years), were recruited through the University Physiotherapy Clinic, local physiotherapy practices and sporting clubs, and University advertising. The primary criterion for inclusion was a grade II ankle lateral ligament sprain that was sustained on average 40 days (724 days standard deviation) prior to testing. We defined this sprain as ‘‘an incomplete tear of the ligament with mild laxity and instability (and) slight reduction in functiony’’ (Safran et al., 1999); A minimum pain-free dorsiflexion asymmetry of 10mm on weight-bearing measure (Vicenzino et al., 2001), anterolateral ankle tenderness, and full pain free weightbearing capacity were also required. Acute ankle sprains were excluded due to the potential for exacerbation of pain with repeated testing on the outcome measures. Exclusion also occurred if fracture or intra-articular ankle effusion were clinically detectable, or if there was a recent history of other lower limb or lumbar spine conditions. Physiotherapists and physiotherapy students were excluded to remove a potential source of bias from the participants. Ethical clearance was obtained from the relevant Institution Review Board for ethics at the University of Queensland, and all participants provided informed consent.

Outcome measures
Dorsiflexion: Weight-bearing dorsiflexion (DF), found to have excellent inter- and intra-rater reliability (Bennel et al., 1998), was measured using the knee-to-wall principle. The participant stood in front of a wall, with the test foot’s second toe and midline of the heel and knee maintained in a plane perpendicular to the wall. The participant slowly lunged forward into talocrural dorsiflexion until the knee contacted the wall, and progressively moved the foot back to the point where the knee could just touch the wall with the heel sustained on the ground. This represented end of range dorsiflexion, and the distance between the wall and second toe was measured in millimetres using a tape measure. The examiner ensured maintenance of heel contact via verbal instructions and manual contact with the calcaneum. Vicenzino et al. (2001) found this measure to be more sensitive in detecting treatment effects than an angular weight-bearing measure and a non-weight-bearing measure.

Pain: Quantitative measures of pain were obtained via pressure and thermal pain threshold. Pressure algometry, which has demonstrated reliability (Pontinen, 1988), was used to measure pressure pain threshold (PPT) at three lower limb sites:over the proximal third of the tibialis anterior muscle belly; (2) directly distal to the lateral malleolus over the CFL; directly anterior to the lateral malleolus over the ATFL. A digital pressure algometer (Somedic AB, Farsta, Sweden) was used to measure the pressure applied to the test site by a rubbertipped probe (area 1 cm2), which was positioned perpendicular to the skin. The pressure was applied at a rate of 40 kPa/s. Activation of a button by the participant at the precise moment that the pressure sensation changed to one of pain and pressure, signalled cessation of pressure application, and froze the measurement onscreen for manual recording. The Thermotest System (Somedic AB, Farsta, Sweden) measured hot and cold thermal pain threshold (TPT). A rectangular contact thermode was manually positioned over two sites: (i) the proximal third of the tibialis anterior muscle belly, and (ii) over the ATFL, extending from the anteroinferior border of the lateral malleolus toward the toes at an angle that allowed maximal contact with the foot contours. The hot or cold stimuli were increased at a rate of 1ºC/s from a baseline of 30ºC. Participants pressed a button at the precise moment that the thermal sensation changed to one of pain and heat for heat pain threshold, and one of pain and cold for cold pain threshold. At this point, stimulation ceased and the temperature reached was manually recorded. Automatic cut-off points of 52ºC and 2.5ºC were adopted to ensure safe stimulus application.

Treatment conditions: Three treatment conditions, consisting of MWM for dorsiflexion, placebo and a no-treatment control, were studied. During the treatment condition, the dorsiflexion MWM technique was performed on the symptomatic talocrural joint, as described by Mulligan (1999). With the participant in relaxed stance on a bench, a nonelastic seatbelt was placed around the distal tibia and fibula and the therapist’s pelvis, with foam cushioning the Achilles tendon. A backward translation by the therapist imparted tension on the seatbelt and a posteroanterior tibial glide, while the talus and forefoot were fixated with the webspace of one hand close to the anterior joint line. The other hand was positioned anteriorly over the proximal tibia and fibula to direct the knee over the second and third toes to maintain a consistent alignment of the distal leg and foot. The glide was sustained during slow active dorsiflexion to end of pain-free range, with the seatbelt kept perpendicular to the long axis of the tibia throughout movement, and released after return to the starting position. Three sets of 10 repetitions were applied, with one minute between sets (Exelby, 1996). Pain experienced during treatment resulted in immediate cessation of the technique and exclusion from the study. The placebo condition replicated the treatment condition, with the following exceptions. The seatbelt was placed over the calcaneum, and only minimal tension imparted to take up the slack. One hand remained on the proximal tibia and fibula, however the other hand was positioned across the metatarsal bases. Instructions were given to produce a small inner range dorsiflexion while the seatbelt was maintained perpendicular to the tibia. An identical number of repetitions, sets and interval period were used. In the control condition, the participant assumed the same relaxed stance position as for treatment and placebo, and maintained this for five minutes. No manual contact occurred between the therapist and participant.

Procedure: A preliminary session, during which a clinical examination and the three outcome measures were performed on both ankles, was conducted initially to determine the participant’s suitability for inclusion. This session also served to familiarize participants with testing procedures. Suitable participants returned for three testing sessions within one week of the initial appointment. These were scheduled at similar times of the day to prevent diurnal variations in joint range and pain, and allow a 24-h interval for wash-out of any treatment effects. Testing was conducted in an environment-controlled laboratory, with constant temperature and humidity. Each testing session began with the asymptomatic then symptomatic ankles undergoing each of the three outcome measures. With the participant in side lying, a splint was applied to the testing ankle to maintain a standardized 10 of plantarflexion. PPT and TPT measures were then conducted in an order randomized by the toss of a coin, followed by weight-bearing dorsiflexion. Three repetitions of each measure were taken. The examiner then left the laboratory while the therapist then entered and applied one of the treatment conditions (MWM, placebo, control) to the symptomatic ankle. Following treatment, outcome measures were repeated on the symptomatic ankle by the examiner to evaluate the effect of treatment. This  procedure facilitated blinding of the examiner. The participant was unaware of the aim of the study and which treatment condition was under investigation. Over the 3 days of involvement in the primary study, each participant experienced all three treatment conditions in a randomised order as determined by the roll of a dice by the therapist.



Acceptable intrarater reliability was determined through analysis of pre-treatment data from the three testing sessions. The intraclass correlation coefficient (ICC) and standard error of measurement (SEM) data for the pain measures are presented in Table 1. The ICC and SEM for the dorsiflexion measure were 0.99 and 3.50 mm, respectively. The ICC for the pain measures ranged from 0.95 to 0.99. The SEM for pressure pain threshold ranged from 5.57 to 12.00 kPa, and the thermal pain threshold SEM ranged from 0.22 to 0.74C. Note that both the size of the error (SEM) and the ICC are indicative of reliable measures.

Data management and analysis

Two independent variables were incorporated into the research design; TREATMENT (MWM, placebo, control), and TIME of application (pre- and post-intervention). Three dependent variables, measures of pressure pain threshold (PPT), thermal pain threshold (TPT) and dorsiflexion (DF), were evaluated. Prior to analysis, triplicate DF, PPT and TPT data were averaged. Data pertaining to two of the participants were excluded from analysis; subject 4 who had a post-testing MRI that revealed an osteochondral lesion of the talus and ankle joint effusion, and subject 7 who experienced pain during the MWM technique. Pre-experiment differences between sides (symptomatic–
asymptomatic) were evaluated by paired t-tests ða ¼ 0:05Þ: A two-factor analysis of variance (ANOVA) was then performed on each of the three dependent variables to test the hypothesis that MWM produced changes in excess of placebo and control from pre- to postapplication. Any significant interaction effects were followed up with tests of simple effects. Post hoc tests of main effects were performed in the absence of an interaction. A Bonferroni adjustment ðaadjusted ¼ 0:05=3 ¼ 0:017Þ was used to interpret results of the pair wise tests of simple effects and to adjust for any type I error resulting from multiple comparisons.


Pre-experiment deficits in outcome measures: Pre-experiment values for dorsiflexion and pain measures of the affected and unaffected ankles are displayed in Table 2. Statistical analysis of side-to-side differences revealed a deficit only for dorsiflexion (DF) (t ¼ 5:689; Po0:001) and pressure pain threshold over the anterior talofibular ligament (PPT ATFL) (t ¼ 2:570; P ¼ 0:025). No such deficits in thermal pain threshold (TPT) were found.

Primary study
Dorsiflexion: A significant interaction time by condition effect for the dorsiflexion outcome measure was detected by the ANOVA (Fð2;26Þ ¼ 7:817; P ¼ 0:002). The interaction plot is shown in Fig. 2. Post hoc analysis revealed a significant treatment effect for dorsiflexion from pre- to post-application (t ¼ 2:870; P ¼ 0:013). The post hoc analysis for the pre- and post-application data showed no significant differences between the placebo (t ¼ 1:343; P ¼ 0:202) and control (t ¼ 1:324; P ¼ 0:208) conditions. Table 3 presents the dorsiflexion data.

figure2 table2 table3 table4

Pain: The data for pain thresholds for pressure, cold and heat stimuli are expressed as mean and standard deviation in Table 4. Statistical analysis of the pain related data revealed no interaction effects (see Fig. 2 for plots). However, there were main effects for time for PPT ATFL (Fð1;13Þ ¼ 6:401; P ¼ 0:025) and PPT TA (Fð1;13Þ ¼ 9:17; P ¼ 0:010). Post hoc tests of simple effects demonstrated significant pre- to post-differences for PPT ATFL in the placebo condition (t ¼ 2:774; P ¼ 0:016) (Fig. 3), but no significant change in PPT TA. No significant time or condition effects were evident for PPT CFL, or the TPT measures.


Application of the dorsiflexion mobilization with movement (MWM) technique to patients with subacute lateral ankle sprains produced a significant immediate improvement in dorsiflexion, but had no significant initial effect on mechanical and thermal pain threshold measures. This dorsiflexion gain following manipulative therapy parallels findings by Green et al. (2001) in acute ankle injuries, and Vicenzino and colleagues’ (2001) study of asymptomatic minimally restricted ankles. Current and previous research findings suggest that the predominant mechanism of action for the dorsiflexion MWM technique is most likely mechanical, rather than a direct hypoalgesic effect. An excessive anterior displacement of the talus is believed to occur during plantarflexion/inversion injury and persist with residual laxity of the anterior talofibular ligament (ATFL) (Mulligan, 1999). Denegar et al. (2002) reported increased ATFL laxity and restricted posterior talar glide in twelve athletes who had sustained an ankle sprain 6 months earlier and had since returned to sport. The clinical rationale given for the anteroposterior glide component of the weight-bearing dorsiflexion MWM technique is to reduce any residual anterior displacement of the talus (Mulligan, 1999). Mulligan (1993, 1999) proposed that correction of the restricted posterior glide, via repetitions of DF with a sustained anteroposterior talar mobilization (mechanically similar to posteroanterior tibial glide on talus), restores the normal joint kinematics even after release of the glide. The mechanism by which this occurs in the presence of ATFL laxity requires further examination. Despite the presence of a reduction in pressure pain threshold (PPT) over the ATFL, the MWM technique did not produce a significant change in local PPT in the initial post-treatment period. The dorsiflexion MWM’s mechanism of action therefore appears to be mechanical, and not directly via changes in the pain system. The conduct of further research is required to identify a precise mechanism. While small but non-significant increases in pressure pain threshold occurred following treatment and control application, it was the placebo condition that produced a statistically significant improvement in pressure pain threshold over the ATFL. It is possible that the gentle inner range dorsiflexion movement performed during the placebo condition was more successful at altering the local pathophysiology peripherally at the ankle or via central neurophysiological mechanisms than the sustained end of range glide and larger range movement of the MWM technique. The application of small amplitude accessory glides of joints in an acute and painful state has been previously advocated (Maitland, 1985) and their benefits in the subacute population requires further investigation. The reasonably small sample size should also be considered to have influenced the results of the statistical analysis. It is possible that the pain measures have a lower sensitivity to change than the dorsiflexion measure, yet the significant dorsiflexion improvement seen post-treatment indicates that range gains are the predominant effect. In addition the failure to elicit prestudy deficits in thermal pain thresholds most likely lessened the likelihood of detecting a change with treatment. Research using a larger sample size and possibly acute ankle sprains with deficits in thermal
pain, should they exist, may reveal differences not detected in this study.


Mulligan’s dorsiflexion mobilization with movement technique significantly increases talocrural dorsiflexion initially after application in subacute ankle sprains. The absence of hypoalgesia post-application suggests a predominant mechanical rather than hypoalgesic effect behind the technique’s success. Further research using a larger sample is required to determine the exact mechanism behind this.