• Ankle Sprain Clinical Prediction Rules

Clinical Prediction Rules
  • Combinations of clinical findings that assist the physical therapist in the predictions of the patients need for a specific diagnostic test, the odds of a specific prognosis, and the likely response to a specific intervention.
external image h9991457_003.jpg
Ankle Sprain
  • A stretching/tearing of the ligaments stabilizing the ankle joint. Involvement of the anterior talofibular, calcaneofibular, or posterior talofibular ligament is indicative of an inversion ankle sprain, an eversion ankle sprain involves the deltoid ligament, and a high ankle sprain involves the anterior and/or posterior inferior tibiofibular ligament

Mechanism of Injury
  • sudden turning of the ankle during weight bearing

  • Grade I : minimal to no swelling, localized tenderness over the ATFL
  • Grade II: localized swelling, diffuse lateral tenderness
  • Grade III: significant swelling, pain, ecchymosis

Ottawa Ankle Rules

The Ottawa ankle rules are a combination of objective findings that aid the clinician in determining the need for radiographs in the management of acute ankle sprains and other ankle injuries.

  • The patient should be referred for radiographs of the ankle if:
    • Pain in the Malleolar zone and:
      • Pain on palpation: distal 6 cm of posterior fibula or
      • Pain on palpation: distal 6 cm of posterior tibia or
      • Unable to bear weight immediately after injury and in the emergency department
    • Pain in the mid-foot zone and:
      • Pain on palpation: base 5th metatarsal or
      • Pain on palpation: navicular tubercle or
      • Unable to bear weight immediately after injury and in the emergency department


Ottawa ankle rules and management of sprain vs. fracture: time: 0.38-4.20

Evidence for the Ottawa Ankle Rule use
  • Accuracy of Ottawa ankle rules to exclude fractures of the ankle and mid-foot: systematic review by Lucas M Bachmann, Esther Kolb, Michael T Koller, Johann Steurer, and Gerben ter Riet
    • 32 studies investigated the accuracy of the Ottawa ankle rules
      • 16 assessed the ankle
      • 11 assessed the mid-foot
      • 10 investigated a combination of both assessments
  • Most studies focused on the use of the Ottawa ankle rules in adults, however 6 studies investigated its use with children
  • Reduces use of unnecessary radiographs by 30-40%
  • Less than 2% of patients that did not meet the criteria actually had a fracture
  • Pooled Sensitivity: 96.4% - 99.6%
  • Pooled Specificity: 26.3%-47.9%
  • Pooled Negative likelihood ratio
    • All: 0.10 (0.06 to 0.16)
    • Ankle Assessment: 0.08 (0.03 to 0.18)
    • Foot Assessment: 0.08 (0.03 to 0.20)
    • Combined Assessment: 0.17 (0.10 to 0.30)
    • Children: 0.07 (0.03 to 0.18)
    • Adults: 0.11 (0.06 to 0.18)

Leiden Ankle Rule

In a study by Glas et.al, the well-known Ottawa ankle rules were compared to the Leiden ankle rule. Patients eligible for the study included those who came to the emergency department at a community hospital in the Netherlands with an acute ankle injury. An acute ankle injury was described as a painful ankle occurring as the result of some form of trauma. Ankle was described as the malleolar and midfoot areas which are frequently injured due to twisting. Patients were not eligible if the injury had occurred more than five days prior. The McNemar test was used to estimate sample size. In order to gain a power of at least 90%, a sample size of at least 609 would be needed. There were 690 patients with an acute ankle injury and of those, 647 were in the study. The mean age was 35 and 324 subjects were female.

A university hospital developed the Leiden ankle rule in 1991 in the city of Leiden (Netherlands). There are seven rows which each contain one or multiple variables. If there is at least one positive variable per row, the stated score is given for that row. This is not applicable to the last row since it is reliant on on the patient’s age. The score is not doubled if two variables in one row are positive. The sum of the row scores yields the final score. If the score is higher than 7, a radiograph is recommended.

Leiden Ankle Rule
|| Clinical Feature
Deformity, Instability, Crepitation
Inability to bear weight
Pulseless or weakened posterior tibial artery
Pain on palpation of malleoli or fifth metatarsal
Swelling of the malleoli or fifth metatarsal
Swelling or pain of the Achilles tendon
Age divided by 10

For each patient, the physicians filled out a data form created for the trial. By using this form, variables could be extracted and scores were calculated for each of the rules.

The radiographic data was used as a reference to calculate sensitivity, specificity, missed fractures, and patients who should have had an x-ray completed. The Ottawa ankle rules caught 66 out of 74 fractures and the Leiden ankle rule caught 59 of the 74 fractures. Physicians were able to determine 61 of the 74 fractures. Of the 8 fractures missed by the Ottawa ankle rules, 1 was significant. Of the 15 fractures missed by the Leiden ankle rule, 5 were significant. Of the 13 fractures missed by physicians, 1 was significant. X-ray was recommended in 76% of the Ottawa cases, 46% of the cases in Leiden, and 38% of cases among physicians. Both the Ottawa and Leiden were created to have 100% sensitivity.



As with anything, there are some disadvantages to using ankle rules. First, the sensitivity may be affected depending on the training and experience of the physician. Second, neither of the ankle rules apply to people with specific comorbidities. Overall, the Leiden ankle rule was less sensitive but more specific than the Ottawa ankle rules. The AUC was also higher for the Leiden ankle rules whereas the percentage of patients who should have had an x-ray was lower. Despite the Leiden ankle rules having a higher specificity, more clinically significant fractures were missed than with the Ottawa ankle rules. It seemed that in many cases, physician judgment was better or as good as any ankle rule.

CPR designed to predict, upon initial examination, patients s/p grade I or II inversion ankle sprain who are most likely to rapidly and dramatically benefit from manual therapy and general mobility exercises

4 Potential Predictor Variables

  1. Symptoms worse when standing
    • patient history
  2. Symptoms worse in the evening
    • patient history
  3. Navicular drop greater than 5.5 mm
    • With the patient’s feet resting on the ground, bearing most of the weight on the uninvolved lower extremity, measure the distance from the navicular tuberosity to the floor, of the involved foot, in subtalar joint neutral
    • Measure the same distance while the patient is full weight bearing bilaterally, in a relaxed stance
    • Measurements recorded to the nearest millimeter
    • The difference between the two measurements is determined and recorded
  4. Distal tibiofibular joint hypomobility
    • PAM assessment performed according to Maitland descriptions

How was this CPR determined?
  • Potential predictive variables were determined by those that had a significance level of P < .10
    • This significance level was chosen because the researchers wanted to make sure not to overlook any potential variables
  • A receiver operator characteristic (ROC) curve was plotted for cut scores
  • Sensitivity, specificity, positive, and negative likelihood ratios were calculated for these variables
  • A stepwise logistic regression model was used to determine the four potential predictive variables that were included in the CPR

Accuracy of Statistics with 95% Confidence Intervals for Individual Predictor Variables*
Sensitivity (95% CI)
Specificity (95% CI)
Positive Likelihood Ratio (95% CI)
Posttest Probability
Symptoms worse when standing
0.75 (0.62, 0.85)
0.50 (0.27, 0.73)
1.5 (0.92, 2.4)
Symptoms worse in evening
0.46 (0.33, 0.59)
0.75 (0.51, 0.90)
1.8 (1.82, 4.1)
Navicular Drop >5.0 mm
0.77 (0.64, 0.86)
0.48 (0.26, 0.70)
1.4 (0.95, 2.2)
Distal tibiofibular joint hypomobility
0.68 (0.55, 0.79)
0.62 (0.39, 0.81)
1.8 (1.10, 3.2)
*pretest probability of success, 75%

Combination of Predictor Variables and Associated Accuracy Statistics with 95% Confidence Intervals
Number of Predictor Variables Present
Positive Likelihood Ratio
Probability of Success
Patients that Satisfied: Success
Patients that Satisfied : Nonsuccess
.06 (0.02, 0.16)
0.86 (0.63, 0.96)
0.43 (0.11, 1.80)
0.28 (0.18, 0.41)
0.95 (0.74, 0.99)
5.90 (1.08, 41.60)
0.50 (0.37, 0.63)
0.57 (0.34, 0.77)
1.20 (0.67, 2.02)
0.08 (0.03, 0.19)
0.76 (0.52, 0.91)
0.33 (0.11, 1.03)
Note: The probability of success is calculated using the positive likelihood ratios and assumes a pretest probability of 75%.

  • When evaluating this CPR it is important to note that patients with three predictor variables present had higher probability of success than those with four predictor variables present.
  • A reliability analysis of the predictor variables showed moderate agreement between hypomobility of the tibiofibular joint and navicular drop greater than 5.5. Little to no agreement was present with the other variables.
  • Only short term results were collected for this CPR.

Suggested Management if CPR is Met
  • manual therapy
    • thrust manipulations: rearfoot distraction, proximal tibiofibular posterior-to-anterior manipulation
    • nonthrust manipulations: anterior-to-posterior talocrural technique, lateral glide / eversion rearfoot technique, distal tibiofibular technique
      • 5 bouts, 30 seconds per bout, grade III or IV
  • general mobility
    • Achilles tendon stretch weight bearing and non-weight bearing, alphabet exercises, ankle eversion self-mobilization, dorsiflexion self mobilization
  • advice regarding activity maintenance within limits of pain
  • ice and elevation

Intrinsic predictive factors for ankle sprain in active university students: A prospective study

  • Studie's objective was to investigate whether certain intrinsic factors (postural control, ankle ROM, motor imagery, functional,instability, history of previous sprain, and body mass index) that could be easily tested in a clinical setting could predict ankle sprains in heathly active people.
  • 121 participants (52 of which had a history of previous ankle sprain) were submitted to a baseline assessment in a single session were then followed-up for 52 weeks regarding the occurrence of sprain.
  • The baseline assessment were performed in both ankles and included the questionaire Cumberland ankle instability tool--Portuguese, the foot lift test, dorsiflexion ROM, Star Excursion Balance Test (SEBT), the side recognition task, body mass index, and history of previous sprain.
  • Two groups were used for analysis: one with those who suffered an ankle sprain (n= 31) and the other with those who did not suffer an ankle sprain(n=90) for (121 participants total).
  • The criterion to retain a variable in the model was a P < 0.05, and for removal a P > 0.10
Table 2: Cox regression analysis using the backward stepwise method
Hazard ratio (95% confidence interval)
History of previous sprain
2.21 (1.07-4.57)
0.96 (0.92-0.99)
Two variables remained in the model for ankle sprain prediction (P= 0.014)

SEBT PL, Star Excursion Balance Test postero-lateral
Figure 1. Star Excursion Balance Test

Star Excursion Balance Test
Figure 3. Survival curves for history of previous sprain and Star Excursion Balance Test postero-lateral direction (SEBT PL) according to Cox regression model. (a) Survival curves according to CG (without previous sprain) and SG (with previous sprain). (b) Participants were categorized according to performance on the SEBT PL; the black line represents participants with SEBT PL < 80%; the dark grey line represents participants with SEBT PL between 80 and 90; and the light grey line represents participants with SEBT PL > 90.

  • Figure 3a shows the difference in the survival curve according to the history of previous strain. Students who had no history of previous sprain had 85% survival at the end of the follow-up time, while the students who had a history of previous sprain had 67% survival.
  • Figure 3b shows survival curves for SEBT PL. The figure shows that the survival curves were better for those with better performance on the SEBT PL.

This study confirms the fact that a history of previous sprains can predict future ankle sprains due to obvious due to laxity in ligaments.
After analysis there was a 48% greater risk of suffering a sprain in the participants with a SEBT PL under 80. In contrast participants who were able to reach a distance equivalent to 90% of their limb length or higher, had a significantly lower incidence of sprains.
  • Resullts showed that only the SEBT in the PL direction could predict the occurrence of ankle sprains.

History of previous sprain was the strongest predictive factor for ankle sprain. A weak performance on SEBT PL can be used as a predictive factor for ankle sprains in active university students.

  • Results may be applied to other similiar populations and future studies are warranted
  • Having the SEBT PL as a predictive factor for ankle sprains gives therapists a simple tool that will help them in their ablility to evaluate patients that are at risk for ankle sprain and subsequently help them in prevention of such ankle sprains.

High Ankle Sprain

Definition: A high ankle sprain is also known as a syndesmotic ankle sprain. It is caused by spraining the syndesmotic ligaments (anterior-inferior tibiofibular ligament, posterior-inferior tibiofibular ligament, transverse tibiofibular ligament, interosseous ligament, and inferior transverse ligament) that connect the tibia and fibula. It is referred to as a high ankle sprain because of its location on the lower leg and relation to the ankle.

Cause: This type of injury usually occurs when the foot is forced quickly into dorsiflexion (can also occur when plantarflexed), if the leg is forcefully twisted (externally) while the foot is planted, or the foot is forcefully and extremely externally rotated.

Anatomy (Syndesmosis Ligaments):
  • Anterior-inferior tibiofibular ligament (most commonly injured)
  • Posterior-inferior tibiofibular ligament
  • Transverse tibiofibular ligament
  • Interosseous ligament
  • Inferior transverse ligament

Syndesmosis Ligaments: The syndesmosis ligaments are under maximal tension when the ankle is fully dorsi-flexed or plantar-flexed with external rotation of the foot on the leg which causes the talus to press against the lateral malleolus. The rotational movement will first affect the anterior-inferior tibiofibular ligament. If the external rotation force is continued, the interosseous membrane and then the posterior-inferior tibiofibular ligament will be compromised.

Signs and Symptoms of an Isolated High Ankle Sprain
  • Point tenderness over anterolateral tibiofibular joint (above lateral malleolus)
  • Pain with weight bearing
  • Pain with passive dorsiflexion
  • Pain with passive external rotation
  • Mild to moderate swelling in lower leg above ankle

Tests to Predict the Presence of a High Ankle Sprain:
  • Syndesmosis Tenderness
    • Usually point tenderness over the anterior-inferior tibiofibular ligament.
    • Estimation of the extent of injury to interosseus ligament is made by proximal extent of tenderness.
  • Squeeze Test
    • Squeezing tibia and fibula together at mid-calf causes pain at the syndesmosis
  • External Rotation Stress Test
    • Passive dorsiflexion and external rotation cause pain
  • Single Leg Hop
    • Inability to single leg hop

Bachmann, L. M. (2003). Accuracy of Ottawa ankle rules to exclude fractures of the ankle and mid-foot: systematic review. BMJ. British medical journal (Clinical research ed.), 326, 417. Retrieved March 24, 2012, from the PubMed database.

Beumer, A. (2007). Chronic Instability of the Anterior Syndesmosis of the Ankle.

de Noronha, M., Franca, L.C., Haupenthal, A., Nunes, G.S. (2011). Intrinsic predictive factors for ankle sprain in active university students: A propsective study. Scandinavian Journal of Medicine & Science in Sports. 12 Jan. 2012. doi: 10.1111/j.1600-0838.2011.01434.x

Dutton, M. (2004). The knee joint complex. Orthopaedic: Examination, evaluation, & intervention. (pp. 890-891). New York, NY:McGraw-Hill.

Glas, Afina, Bas Pijnenburg, et al. "Comparison of diagnostic decision rules and structured data collection in assessment of acute ankle injury." CMAJ. 166.6 (2002): 727-733. Web. 9 Apr. 2012.

Loudon, J., Swift, M., Bell, S. (2008). Knee. The clinical orthopedic assessment guide. (2nd ed., pp. 368). Champaign, IL: Human Kinetics.

Lynch, S. a. (2002). Assessment of the Injured Ankle in the Athlete. Journal of athletic training, 37(4), 406-412. Retrieved from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=164372&tool=pmcentrez&rendertype=abstract

Norkus, S. a, & Floyd, R. T. (2001). The anatomy and mechanisms of syndesmotic ankle sprains. Journal of athletic training, 36(1), 68-73. Retrieved from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=155405&tool=pmcentrez&rendertype=abstract

Whitman, J. M., Cleland, J. a, Mintken, P. E., Keirns, M., Bieniek, M. L., Albin, S. R., Magel, J., et al. (2009). Predicting short-term response to thrust and nonthrust manipulation and exercise in patients post inversion ankle sprain. The Journal of orthopaedic and sports physical therapy, 39(3), 188-200. doi:10.2519/jospt.2009.2940