To The Editor:
With great interest, we read the article "Ankle Stress Test for
Predicting the Need for Surgical Fixation of Isolated Fibular Fractures"
(2004;86:2393-8), by Egol et
al., about the use of the ankle stress test for detecting injury of the
deltoid ligament. Assessment of the integrity of the deltoid ligament is of
importance when classifying an isolated fracture of the fibula. In the case of
supination-eversion trauma, it means the difference between an SE-II and an
SE-IV-stage injury and a recommendation for nonoperative and operative
treatment, respectively.
We would like to comment on the statistical analysis. Probably due to a
slip of the pen, the specificity in this report is defined as the number of
cases with a negative clinical sign divided by the total number of cases with
a medial clear space of =4 mm. With correction of the medial clear space to
<4 mm instead of =4 mm, this problem is solved.
However, we cannot reproduce the sensitivity and specificity values that
the authors present in Table I. According to our calculations, the sensitivity
and specificity are 55% (thirty-six of sixty-six) and 83% (twenty-nine of
thirty-five), respectively, for medial tenderness; 55% (thirty-six of
sixty-six) and 91% (thirty-two of thirty-five), respectively, for swelling;
and 26% (seventeen of sixty-six) and 94% (thirty-three of thirty-five),
respectively, for ecchymosis.
Additional calculations show a positive predictive value of 86% (thirty-six
of forty-two), 92% (thirty-six of thirty-nine), and 89% (seventeen of
nineteen) for tenderness, swelling, and ecchymosis, respectively. The negative
predictive values are 49% (twenty-nine of fifty-nine), 52% (thirty-two of
sixty-two), and 40% (thirty-three of eighty-two) for tenderness, swelling, and
ecchymosis, respectively.
A low negative predictive value means that the deltoid ligament can still
be ruptured in the 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.
As the interrelationship between the test regarding tenderness, swelling,
and ecchymosis is not presented in the study, we could not evaluate the
sensitivity and specificity with respect to the combinations of these clinical
signs.
Although the data on sensitivity and specificity are not completely
correct, the overall conclusion of the study remains that clinical signs are
not reliable parameters in the evaluation of deltoid injury. We therefore
suggest that, in specific cases, magnetic resonance imaging can be of
additional value. With this technique, the absence or presence of deltoid
injury can be accurately detected1-3. In addition to the outcome of
the ankle stress test, the surgeon can then make a balanced decision to
operate on an unstable supination-eversion fibular fracture and avoid the
stress of a wrong treatment.
The best radiographic parameter with which to evaluate the deltoid ligament
is the ratio of the medial clear space to the superior clear space. A ratio of
>1 is indicative of deltoid injury4.
We 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 shown below, and they
do5. We also
double-checked the wording in the results section to verify that it correctly
reports the findings, and we believe that it does.
Given this information, we think that what is going on is that Dr. Hermans
and colleagues are trying to recreate this table using the data that were
reported in our paper. However, our paper does not report the raw data from
which the sensitivity and specificity were calculated. We include the table of
raw data (Table).
The published paper gives the values for sensitivity, specificity, the
total number of cases in which the medial space was =4 mm, and the number
of cases in which the medial space was =4 mm and the clinical sign was
present. Given these data alone, the reader could easily calculate the number
of cases in which the medial spacing was =4 mm and the clinical sign was
not present, simply by subtracting the number of cases in which the clinical
sign was present from sixty-six.
However, the reader would not be able to do the same thing for cases in
which the medial spacing was <4 mm. To do that, he or she would have to
calculate the number of cases in which the medial clear space was <4 mm and
the symptom was not present by using the reported value for sensitivity. For
medial tenderness, the calculation would be 80% of thirty-five, which is
twenty-eight. Then, the number of cases with medial tenderness among those in
which the medial space was <4 mm would be calculated as thirty-five minus
twenty-eight, for a difference of seven.
Consequently, we speculate that Dr. Hermans and colleagues are trying to
back-calculate to verify the calculations of sensitivity and specificity and
are running into some difficulty, perhaps because of some confusion in reading
the text accurately.
We appreciate our colleagues' interest in our study.
Boss AP, Hintermann B. Anatomical study
of the medial ankle ligament complex. Foot Ankle Int.2002;23:
547-53.23547
2002
[PubMed]
Muhle C, Frank LR, Rand T, Yeh L, Wong
EC, Skaf A, Dantas RW, Haghighi P, Trudell D, Resnick D. Collateral ligaments
of the ankle: high-resolution MR imaging with a local gradient coil and
anatomic correlation in cadavers. Radiographics.1999;19:
673-83.19673
1999
[PubMed]
Klein MA. MR imaging of the ankle:
normal and abnormal findings in the medial collateral ligament. AJR Am
J Roentgenol.1994;162:
377-83.162377
1994
Beumer A, van Hemert WL, Niesing R,
Entius CA, Ginai AZ, Mulder PG, Swierstra BA. Radiographic measurement of the
distal tibiofibular syndesmosis has limited use. Clin Orthop Relat
Res.2004;423:
227-34.423227
2004
[CrossRef]
Hulley SB, Cummings SR, editors.
Designing clinical research: an epidemiologic approach.
Baltimore: Williams and Wilkins; 1988.
1988