To the editor:

We have double checked to make sure that the values reported in the published paper use the correct formula for sensitivity and specificity for the data in the table below, and they do(1). We also double-checked the wording in the results section to verify that it is correct in how it reports the findings, and we believe it does so.

Given this, we think that what is going on is that you are trying to recreate this table using the data that were reported in our paper. However, our paper does not show the entire table (at the time we revised the paper for publication it was our understanding that the reviewers preferred just reporting sensitivity and specificity rather than the raw data. We have attached, for your benefit, the table with the complete data, (see below). This may be the source of difficulty.

The published paper gives the following values: sensitivity, specificity, the total number of cases where medial space >= 4mm, and the number of cases where medial space >=4 mm and the clinical sign was present. Given these data alone, the reader could easily calculate the number of cases that had medial spacing >=4 mm without the clinical sign, simply by subtracting the number of cases with the clinical sign from 66.

However, the reader would not be able to do the same thing for cases where medial spacing is less than 4 mm. To do that, they would have to calculate the number of cases without the symptom where medial clear space was less than 4 mm by using the reported value of sensitivity. For medial tenderness, the calculation would be 80% of 35, which is 28. Then the number of cases with medial tenderness among those where medial space < 4 mm would be 35 – 28 = 7.

Consequently, we speculate that the reader is trying to back- calculate to verify the calculations of sensitivity and specificity, and is running into some difficulty, perhaps because of some confusion in reading the text accurately.

We appreciate and value your comment and interest. Feel free to contact me if I can be of any further help in answering questions or clarifying the findings of the study.

Sincerely, Kenneth A. Egol, M.D., et al.

References:

1.Hulley SB, Cummings SR. Designing Clinical Research: An Epidemiological Approach. Baltimore, MD: Williams & Wilkins; 1988).

To the editor:

With great interest we read the article of Egol et al. about the use of the ankle stress test in detecting injury of the deltoid ligament. Assessment of the integrity of the deltoid ligament is of importance in classifying an isolated fracture of the fibula. In supination eversion trauma it means the difference between a SE II and SE IV stage with a recommended nonoperative and operative treatment respectively.

We would like to comment on your statistical analysis. Probably due to a slip of the pen the specificity in your paper is defined as the number of cases with a negative clinical sign divided by the total number of cases with a medial clear space ¡Ý4mm. With the correction medial clear space <_4mm instead="instead" of="of" medial="medial" clear="clear" space="space" ý4mm="ý4mm" this="this" problem="problem" is="is" solved.="solved." p="p" /> However we cannot reproduce the sensitivity and specificity values you present in table 1. According to our calculations the sensitivity and specificity for medial tenderness are 55% (36/66) resp 83% (29/35); for swelling 55% (36/66) resp 91% (32/35) and for ecchymosis 26% (17/66) resp 94% (33/35).

Additional calculations show a positive predictive value for tenderness, swelling and ecchymosis of 86% (36/42), 92% (36/39) and 90% (17/19) respectively. The negative predictive value for tenderness, swelling and ecchymosis is 49% (29/59), 52% (32/62) and 40% (33/82) respectively.

A low negative predictive value means that the deltoid ligament can still be ruptured in absence of clinical signs. If medial tenderness, ecchymosis and swelling are used to 'upgrade' a fracture from SE II to SE IV it will result in an indication to operate. In some cases surgery may be performed on stable ankles, as inferred from the moderate positive predictive values.

Since the interrelationship between the test regarding tenderness, swelling and ecchymosis is not presented we could not evaluate the sensitivity and specificity with respect to the combinations of these clinical signs.

Although the data in your article on sensitivity and specificity are not completely correct the overall conclusion remains that clinical signs are not reliable parameters in the evaluation of deltoid injury. We therefore suggest that in specific cases MRI can be of additional value. With this technique absence or presence of deltoid injury can be accurately detected (1,2,3). In addition to the outcome of the ankle stress test the surgeon can than make a balanced decision to operate on an unstable supination eversion fibula fracture avoiding the stress of a wrong treatment.

The best radiographic parameter to evaluate the deltoid ligament is the ratio of the medial clear space and the superior clear space. A ratio greater than one is indicative of deltoid injury (4).

1. Boss AP, Hintermann B. Anatomical study of the medial ankle ligament complex. Foot Ankle Int. 2002 Jun;23(6):547-53.

2. Muhle C et al. Collateral Ligaments of the Ankle: High-Resolution MR Imaging with a Local Gradient Coil and Anatomic Correlation in Cadavers. Radiographics. 1999 May-Jun;19(3):673-83.

3. Klein MA. MR imaging of the ankle: normal and abnormal findings in the medial collateral ligament. AJR Am J Roentgenol. 1994 Feb;162(2):377- 83.

4. A. Beumer et al. Radiographic measurement of the distal tibiofibular syndesmosis has limited use. Clinical Orthopaedics and Related Research 2004, nr 423, pp 227-234.

To the Editor:

We thank Dr. Kumar for his interest in our article. In our study, all patients with a positive stress radiograph and clinical signs of medial injury were considered to be SE-IV injuries and were treated surgically. Of the remaining patients (30) the functional outcome in the group of patients treated surgically or non-surgically was similar (AOFAS score of 93 v 94). The patients were not randomized and were treated according to surgeon and patient preference. All patients meeting the eligibility criteria stated in the material and methods section were included. All patients were able to tolerate the stress view without significant discomfort and were prescribed pain medication as needed.

Sincerely, Nirmal Tejwani, MD Kenneth Egol, MD

To The Editor:

We thank Dr. Sheth and his colleagues for reviewing our article at their journal club. In our paper we have attempted to assess the outcomes of non-operatively treated fibula fractures despite a diagnosis of a supination-external rotation Stage IV injury based on stress radiography. Like most surgeons we do not routinely fix SE2 fracture patterns as we believe these fractures will heal uneventfully. We did not evaluate the functional outcome in this group as this was not the purpose or focus of our study. The patients who had a positive stress radiograph and clinical signs of medial injury were considered to be SE-IV injuries and the patient was treated surgically. It was the group that had a positive stress radiograph and no medial clinical signs that was of interest to us and followed up both clinically and radiographically. As stated in the manuscript the functional outcome in the group of patients treated surgically or non- surgically was similar (AOFAS score of 93 v 94). We have stated that one of the limitations of our study was the small number of patients with positive stress radiographs and negative clinical findings on the medial side. The lack of randomization may have introduced a selection bias toward non-operative treatment in older patients.

The use of the MRI may be beneficial in identifying medial ligamentous injury, similar to the findings of interosseus membrane injury described recently. (1)

Bibliography:

1. Nielson JH, Sallis JG, Potter HG, et al. Correlation of interosseous membrane tears to the level of the fibular fracture. J Orthop Trauma. 2004 Feb;18(2):68-74.

Sincerely, Nirmal Tejwani, MD Kenneth Egol, MD

To the Editor:

I read this article with interest as it looked into a common clinical dilemma, but the study does raise a few questions:

Did any of the ‘relatively undisplaced’ fractures, especially in the group I, displace after stress radiographs?

One of the aims of the study was to look into the non operative management of stage IV supination-external rotation injury with positive stress radiography but the authors operated on 36/66 patients with this kind of injury and of the remaining 30 they did not randomise. Since, the paper is looking into the non operative management of stage IV supination- external rotation injury all 66 patients should have been randomised.

Figure 1-D shows widening of the medial clear space but the lateral wall of the medial malleolus is not a sharp line as in figure 1-E. Hence, how did the authors decide where the lateral wall of the medial malleolus was? Measuring the medial clear space is often fraught with difficulties especially when the lateral border of the medial malleolus is not a sharp line, hence, showing inter observer reliability would have been more appropriate.

How often did the group I patients have radiographs of their ankles during the follow up and did they have a change of cast once the swelling settled?

Statistical significance for comparing medial clear space widening between patients with a positive stress radiograph with or with out medial signs has to be taken into consideration. As the numbers to treat have not been calculated, providing confidence intervals might have provided more clarity. Similarly, positive and negative predictive values of the clinical signs could have also been described.

Ankle fractures are often quite painful. The authors did not mention if any of the patients found it difficult or painful while stress radiographs were performed.

To the Editor:

We reviewed the article, “Ankle Stress Test for Predicting the Need for Surgical Fixation of Isolated Fibula Fracture” (2004;86:2393- 2398) By Egol et al at our monthly journal club. We are writing this letter on behalf of our journal club.

A number of conclusions made by the authors were not supported by the data presented in the paper. The authors stated that the study supported an algorithm that directs patients with a negative stress radiograph and no medial symptoms to be treated non-operatively. However, the only data supporting this conclusion was that “all of the thirty-five patients … had clinical and radiographic evidence of healing” No functional or follow-up radiographic data were provided. This group would represent a supination-external rotation stage 2 pattern of injury and non-operative management is accepted as standard of care. However, a more detailed analysis of this group to support the authors’ impression would be helpful.

One of the stated goals of the paper was to “determine the functional outcome of nonoperative treatment despite a diagnosis of supination- external rotation stage IV injury based on stress radiography”. However, of 66 patients who were positive by stress radiography, only 10 were treated nonoperatively (these were 10 of 30 patients without signs of medial injury). In order to really fulfill this goal some of the patients with medial signs would have had to be treated nonoperatively.

All patients with positive stress radiographs with medial evidence of injury were treated operatively. Only a subgroup of patients with a positive stress test and no evidence of medial injury were included in the study. This subgroup was non-randomly divided into operative and non-operative groups (selection bias). We therefore believe that no conclusion can be drawn from the results. The utility of a stress test in predicting need for surgical fixation of isolated lateral malleolus fracture still remains unanswered.

We conclude that the only real way to test the utility of a stress radiograph is to find out a) if it really indicates that deltoid ligament is torn (perhaps MRI); and b) to conduct a study where all the patients with a positive stress test with or without evidence of medial injury are randomized into operative and non-operative groups.

To the Editor:

We appreciate the letter from Drs. Kragh and Thompson regarding our paper. We agree that there are certain gender specific anatomic differences and potential error in measurement that may lead to bias. We had attempted to correlate some very common subjective methods for determining ankle instability with an objective one. We chose an absolute threshold of 4 mm because that is the accepted amount of medial widening accepted in our community. Our purpose was not to prove that an absolute value of medial widening represents a pathologic condition, but rather to show that accepted standards may not absolute.

We have biomechanical data currently in press suggesting the absolute value of a pathologic medial clear space is greater than 5 mm. Furthermore, individuals may have varying degrees of ligamentous laxity that could skew results. We have begun stressing the uninjured contralateral side in patients that sustain an isolated fibula fracture in order to examine this possibility

Kenneth A. Egol MD

Kenneth J. Koval MD.

To the Editor:

We thank Dr. Elhence for his letter. We agree with him that the indications for surgery after isolated lateral malleolus fracture remains uncertain. Based on our results, it seems that we are currently performing surgery on some patients that could be treated successfully with non- operative means. The problem remains on identifying those patients who could and should be treated non-operatively. It is possible that other imaging modalities such dynamic stress testing or MRI might be able to better detect those patients who would benefit from surgical intervention. Also, a true randomized clinical trial would be needed to determine the best treatment method for treating potentially unstable fractures of the ankle. The results of our study indicate the need for more research for this common fracture.

Kenneth A. Egol MD Kenneth J. Koval MD

To the Editor :

I read the article, " Ankle Stress Test for Predicting the Need for Surgical Fixation of the Isolated Fibular Fractures " with great interest and I congratulate the authors for their contribution to the management of supination external rotation injuries about the ankle .

While the literature is replete with articles in favour of and against operative intervention for stage 4 Supination External Rotation Injuries of the ankle (with or without positive stress Xrays and with or without medial clinical signs, I believe the individual surgeon remains at a loss in such clinical situations with treatment being governed to a large extent by personal whims or limited surgical experiences in this gray zone ( Positive stress x- rays and negative medial clinical signs ).

The best approach to answering the question of how to treat patients with a positive stress radiograph of the ankle and negative medial clinical signs (especially medial joint tenderness and ecchymoses) is to perform a functional evaluation study where patients are subjected to a CT Scanogram of the injured and uninjured ankle joints and subsequent treatment using operative and non operative methods are administered on a randomized basis .

To the Editor:

We thank Egol et al. and McConnell et al. for their fine works on deltoid ankle instability with fibula fractures.(1,2) We ask them to consider replying to the ideas herein.

1. Women have smaller radiographic medial clear spaces than men,(3) and the use of an absolute threshold introduces non-random error into measurement of instability. An absolute threshold biases assessment because of patient size.

2. Magnification variability due radiographic technique introduces error when using an absolute distance measurement on radiographs to represent an anatomical distance.

3. The operational definition of instability chosen by Egol et al. and used as an indicator by McConnel et al., that is, a medial clear space >4mm on a radiograph during stress examination without anesthesia, led to some determinations of instability that were difficult to explain.(1,2) Unstressed radiographic medial clear spaces are reportedly up to 5.5mm in radiographs without fracture and about 8% are >4mm.(4) In cadaver ankles with fibulas excised in simulation of ankle fracture with an intact deltoid ligament, the medial clear space distance increased up to 2mm from the resting distance when stressed due to deltoid laxity at rest.(5) As 4mm may be too low a threshold at rest and up to 2mm more may be added when stressed with the deltoid ligament intact, then laxity may need more consideration in determining instability thresholds. Some of Egol et al.’s patients may have been stressed >4mm yet have had intact superficial and/or deep deltoid ligaments. The 4mm threshold historically is an unstressed threshold, but both recent reports used it as stressed. An explanation of the stressed-unstressed mismatch by the authors may help as these problems with the 4mm absolute threshold may in part explain the difficulties.

4. Radiographic indicators of ankle instability that address these problems may perform better diagnostically. Conceivably, if the medial clear space is divided by the superior clear space to get a relative index of instability, then the problems of patient size bias, magnification error, and the absolute threshold can be mitigated.

John F. Kragh, Jr. M.D. LTC(P), MC, USA

Jon Thompson, M.D. MAJ, MC, USA

1. Egol KA, Amirtharage M, Tejwani NC, Capla EL, Koval KJ. Ankle stress test for predicting the need for surgical fixation of isolated fibular fractures. J Bone Joint Surg, 86A:2393-405, 2004.

2. McConnell T, Creery W, Tornetta P III. Stress examination of supination external rotation-type fibular fractures. J Bone Joint Surg, 86A:2171-8, 2004.

3. Jonsson K. Fredin HO. Cederlund CG. Bauer M. Width of the normal ankle joint. Acta Radiologica: Diagnosis. 25(2):147-9, 1984.

4. Brage ME, Bennett CR, Whitehurst JB, Getty PJ, Toledano A. Observer reliability in ankle radiographic measurements, Foot Ankle Int, 18:324-9, 1997.

5. Close JR. Some applications of the functional anatomy of the ankle joint. J Bone Joint Surg, 38A:761-81, 1956.