Orthopedics

The physical examination is unreliable in determining the location of the distal fibular physis

a b s t r a c t

Objectives: Salter-Harris type 1 (SH1) fractures of the distal fibula are acute orthopedic injuries with tenderness over the physis without radiographic evidence of fracture. Our primary objective was to establish the accuracy of the physical examination performed by pediatric emergency medicine (PEM) physicians in determining the lo- cation of the distal fibular physis compared to a criterion standard of ultrasound.

Methods: This was a prospective, observational study at an urban academic pediatric emergency department of a convenience sample of children aged 4 to 10 years old between March 2019 and March 2020. A PEM physician or fellow examined the patient’s distal fibula and marked the location of the physis with a marker. A study investi- gator scanned the distal fibula to establish the location of the physis on ultrasound and measured the distance between the clinician’s estimated position and the actual sonographic position. We a priori defined a clinically accurate position as a distance of <=5 mm. We compared the accuracy rate of physical examination to ultrasound landmarking using proportions with 95% confidence intervals (CI).

Results: We enrolled 71 patients, of whom 52 (73%) were male. The mean age was 6.7 years and the mean weight was 25.5 kg. Participating PEM physicians included 18 attending physicians and 2 fellows. The distal fibular physis was correctly identified in 24 patients, yielding an accuracy rate of 34% (95% CI 23%-46%). The mean dis- tance between the physician’s estimated position and the sonographic position was 7.4 mm (95% CI 6.4-8.4 mm). Conclusions: PEM physicians were unable to accurately identify the distal fibular physis on physical examination.

(C) 2021

  1. Introduction

Acute ankle injuries are a common pediatric presentation to the emergency department (ED). A proportion of these patients will have fractures involving the physis, which are classified according to the Salter-Harris system. While Salter-Harris type 1 (SH1) fractures were historically believed to be a common inversion injury of the ankle in the preadolescent age-group [1], recent evidence suggest the true inci- dence may be as low as 3% [2]. SH1 fractures are described as acute or- thopedic injuries with tenderness over the distal physis without radiographic evidence of fracture, although separation of the physis can occasionally be visualized on radiographs. Support for this concept dates back to Salter and Harris’s study from 1963 where they concluded that the physeal cartilage is weaker than the surrounding ligaments and

* Corresponding author at: Department of Emergency Medicine, Diamond Health Care Centre, 11th Floor, 2775 Laurel Street, Vancouver, BC V5Z 1M9, Canada.

E-mail address: [email protected] (D.J. Kim).

hence more prone to injury [3]. The diagnosis of a SH1 fracture of the distal fibula is therefore dependent on accurate identification of the fib- ular growth plate by physical examination. However, it can be difficult to determine the precise location of the physis by physical examination, particularly in an acutely injured and swollen ankle.

To the best of our knowledge, no study has been conducted to assess if physicians can reliably determine the location of the physis on physi- cal examination. However, Boutis et al. demonstrated that in a cohort of children with suspected SH1 fracture of the distal fibula, only 3% had a magnetic resonance imaging (MRI)-confirmed fracture, while 80% had evidence of ligamentous injury [2]. The rarity of this injury may be ex- plained in part by clinicians being unable to accurately determine the lo- cation of the distal fibular physis.

Although we are not aware of a study directly comparing the accu- racy of Point of Care Ultrasound visualization of the physis to other imaging modalities, several POCUS studies have assessed norma- tive sonographic measurements of the physis in healthy, uninjured pa- tients [4,5], and several other studies have looked at the utility of

https://doi.org/10.1016/j.ajem.2021.07.040

0735-6757/(C) 2021

Image of Fig. 1

Fig. 1. Ultrasound image of the distal fibular physis (indicated by the vertical line). The bony cortex is white and hyperechoic, while the physis is the anechoic area between the metaphysis on the left and the epiphysis on the right.

ultrasound in the assessment of the growth plate in acute ankle injuries [6,7]. Ultrasound has also been shown to be an accurate method of diag- nosing occult Pediatric fractures when compared to MRI [8]. It logically follows that POCUS is likely accurate in the identification and measure- ment of the width of the distal fibular physis in healthy, uninjured pedi- atric patients. The physis is identified as a smooth, regular space between the metaphysis and the epiphysis. It is differentiated from the joint space by the smooth, contiguous line that could be drawn be- tween the metaphysis and epiphysis [4] (Fig. 1). Our primary objective was to establish the accuracy of the physical examination performed by pediatric emergency medicine (PEM) physicians in determining the lo- cation of the physis of the distal fibula as compared to a criterion stan- dard of POCUS. Our secondary objective was to determine if patient factors and clinician factors were associated with accuracy.

  1. Methods
    1. Study design and setting

This was a prospective cross-sectional single center study performed in the British Columbia (BC) Children’s Hospital ED in Vancouver, Canada from March 2019 to March 2020. BC Children’s Hospital is an urban, university-affiliated, Tertiary care children’s hospital. The ED has a PEM fellowship program, and an annual volume of 50,000 patient visits. This study received approval from the University of British Co- lumbia Children’s and Women’s Research Ethics Board.

    1. Population

We enrolled a convenience sample of otherwise healthy children ages 4 to 10 years old who presented to the ED for reasons other than lower extremity injury. This age range was chosen to ensure coopera- tion with physical and ultrasound examination from participants and to include subjects that would still have open growth plates. Written consent for participation was obtained by the study physician for both the patient participant and clinician participant. Patients were excluded if they were assigned a Canadian ED Triage and Acuity Scale of 1 or 2, had lower extremity injury, had underlying bony or muscular abnor- mality (like osteogenesis imperfecta, achondroplasia, rickets), were non-weight-bearing, had allergy to ultrasound gel, had a language bar- rier, or were unable to provide or refused consent.

    1. Intervention & measurements

Recruitment was done on a convenience basis during day and eve- ning shifts, mainly on weekdays, when the POCUS fellow was in the ED. The POCUS fellow collected initial demographic information includ- ing age, sex, and weight. The patient’s treating PEM physician or fellow performed a unilateral physical examination of the patient’s ankle and marked his or her estimated location of the distal fibular physis with an invisible ink pen. Odd numbered patients had their right ankle exam- ined and even numbered patients had their left ankle examined. After this physical examination, the POCUS fellow used the high frequency L25 Linear probe (6-13 MHz) of a SonoSite X-Porte (Fujifilm SonoSite, Bothell, WA) to scan the distal fibular physis. The patient was positioned in the dorsal decubitus position for the scan. The transducer was placed along the long axis of the distal lateral fibula and held in the neutral po- sition. The indicator was directed towards the patient’s head. The physis was identified when a regular anechoic space was seen at the distal end of the linear hyperechoic bone between the metaphysis and epiphysis. The midpoint of the physis was then centered on the screen. The trans- ducer was held in place with the non-dominant hand. The dominant hand was used to mark two points on both sides of the middle of the probe. A continuous horizontal line was then drawn under the probe, between the two points, over the middle of the physis. Using an ultravi- olet light to visualize the invisible ink marking, the POCUS fellow mea- sured the distance between the PEM physician’s marking and the sonographic marking in millimeters (Fig. 2). If the lines were not paral- lel, the closest distance between them was taken. Because the distal fib- ular physis measures 3 mm in width on average [4], we a priori defined a clinically accurate position to be a marked point located within 5 mm of the sonographic point.

Image of Fig. 2

Fig. 2. Demonstration of marking and measurement technique. (A) The patient’s treating PEM physician or fellow performed a unilateral physical examination of the patient’s ankle and marked his or her estimated location of the distal fibular physis with an invisible ink pen. (B) After this physical examination, the POCUS fellow placed the high frequency linear probe along the long axis of the distal lateral fibula, centered the midpoint of the physis on the screen, used the dominant hand to mark two points on both sides of the middle of the probe, then drew a continuous horizontal line to connect these points.

(C) Using an ultraviolet light, both markings were visualized. (D) The POCUS fellow measured the distance between the PEM physician’s marking and the sonographic marking in millimeters.

Table 1

Clinical characteristics of the study cohort.

Total Subjects 71

Mean Age, months (SD) 80.3 (SD = 21.8)

Mean Weight, kg (SD) 25.5 (SD = 9.7)

Male, n (%) 52 (73.2)

Ankle Examined

Right, n (%) 35 (49.3)

Left, n (%) 36 (50.7)

SD = standard deviation.

Before the start of this study, the POCUS fellow, who is also a board- certified emergency physician, had previously performed about 400 ul- trasound scans in general and 50 ultrasound scans of the pediatric ankle and wrist physis in the context of a previous study [5]. Prior to enrolling any patients, the POCUS fellow performed 5 practice scans of the distal fibular physis directly supervised by the POCUS fellowship director (DJK) to demonstrate adequate performance of the scan. The inter- rater reliability of the POCUS fellow’s ultrasound scans was assessed by having a random sampling of 10% of the recorded cine clips reviewed by the POCUS fellowship director.

    1. Outcome measures

The main outcome measure for this study was the proportion of dis- tal fibular physis correctly identified by physical examination as com- pared to the reference standard of ultrasound. We also assessed the influence of other variables on accuracy, including patient factors (age, sex, and weight) and clinician factors (attending physician versus fellow).

Image of Fig. 3

Fig. 3. Distance between the physician’s assessment by physical examination and the sonographic position. The first and third quartiles are represented by the horizontal boundaries of the box, and the median is represented as the horizontal line running through the box.

Table 2

Proportion of physician examinations that correctly identified the distal fibular physis (<=5 mm compared to ultrasound reference standard) and 95% confi- dence intervals.

Method Accuracy Rate (95% CI)

Crudea 0.34 (0.23-0.46)

Sensitivity Analysisa 0.48 (0.37-0.59)

Logistic Regressionb 0.32 (0.20-0.47)

CI = confidence interval.

a Normal approximation for independent sample proportion.

b Intercept-only marginal logistic regression model with exchangeable cor- relation structure clustered by physician.

    1. Data analysis

We computed descriptive statistics for characteristics of the study cohort, including counts and proportions for categorical variables and means and standard deviations for quantitative variables.

To obtain a crude estimate of the observed clinician accuracy rate, we computed a 95% confidence interval (CI) using the normal approxi- mation for independent sample proportions. Enrollment was stopped early due to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, so we fell short of our target sample size. Be- cause of this, we conducted a sensitivity analysis assuming the physical examinations for the remaining patients until our target sample size would have correctly identified the distal fibular physis.

Finally, we fit two logistic regression models using generalized esti-

mating equations to account for repeated measurements among clini- cians. The first model only included an intercept term, which we used to estimate the probability of an accurate exam and its associated 95% CI. The second model included terms for age (years), sex (female versus male), weight (kg), and clinician type (attending versus fellow) to ex- plore possible associations between patient and clinician characteristics and clinician accuracy. We used model covariates to estimate odds ra- tios (ORs) and 95% CI.

    1. Sample size

Sample size estimates were obtained via simulation with at least 15 clinicians participating in the study and each clinician assessing multi- ple patients, between a range of 2 and 20. We simulated clinician accu- racy for several different sample sizes using a logistic regression model with a random intercept for physicians. The fixed intercept in the model represents the log odds of clinician accuracy and was computed based on detecting a 70% accuracy rate. According to this simulation, we calcu- lated that a sample size of 90 patients would be sufficient to detect a 70% clinician accuracy rate versus only 50% with at least 80% power.

  1. Results

Enrollment was stopped early before we achieved our target sample size due to the SARS-CoV-2 pandemic. Of the 72 patients approached for this study, we enrolled 71 patients; one patient was excluded after de- clining consent. The mean age of the participants was 80 months (6.7 years), the mean weight was 25.5 kg, and 52 patients (73%) were male (Table 1).

A total of 20 PEM physicians participated in this study: 18 PEM at- tending physicians who examined a total of 60 patients (85%) and 2 PEM fellows who examined a total of 11 patients (15%). Each clinician assessed a mean of 3.6 patients, with the total range from 1 to 15 patients.

The distal fibular physis was correctly identified in 24 patients, yield- ing an accuracy rate of 34% (95% CI, 23%-46%). Given that we did not reach our target sample size, we performed a sensitivity analysis where we assumed that the remaining 19 patient examinations would

Table 3

Adjusted logistic regression estimates for correct identification of the distal fibular physis.

Coefficient

Estimate

Robust SE

P-value

OR (95% CI)

Intercept

-1.10

0.79

0.17

0.33 (0.07, 1.57)?,a

Age (years)

-0.26

0.15

0.09

0.77 (0.57, 1.04)

Sex (female vs male)

1.95

0.82

0.02

7.01 (1.41, 34.72)?

Weight (kg)

0.08

0.03

0.008

1.08 (1.02, 1.14)?

Clinician type (attending vs fellow)

-0.83

0.86

0.33

0.44 (0.08, 2.35)

SE = standard error. OR = odds ratio. CI = confidence interval.

* P-value <0.05.

a This is not an OR. This estimate represents the odds of correct identification at reference levels of each risk factor (i.e. average age [6.7 years], male, average weight [25.5 kg], and examined by fellow).

have all been accurate. Even in this best-case scenario, the accuracy rate would have only been 48% (95% CI, 37%-59%) (Table 2).

The mean distance between the physician’s physical examination position and the sonographic position was 7.4 mm (95% CI, 6.4-8.4) (Fig. 3).

The odds of accurate identification were higher for female patients and for patients who weighed more. Specifically, the odds of clinician accuracy were 7.01 (95% CI: 1.41, 34.72) times higher for females com- pared to males. The odds of clinician accuracy increased by 8% (95% CI: 1.02, 1.14) for every additional kilogram of patient weight. There was no significant association between patient age and clinician accuracy or cli- nician type (attending versus fellow) and clinician accuracy (Table 3).

There was perfect agreement on the review by the POCUS fellowship director of the random sampling of 10% of the recorded cine clips.

  1. Discussion

This study demonstrates that PEM physicians have poor accuracy with the physical examination to correctly identify the distal fibular physis in pediatric patients without acute lower extremity injury. PEM physicians were only 34% accurate with their physical examination in identifying the distal fibular physis within 5 mm of the ultrasound land- mark. Given that the width of the physis is 3 mm, palpation outside the 5 mm range is unlikely to represent the true location of the physis.

The association between female patients and the higher odds of ac- curacy is difficult to explain, but it is possible that it could be related to more advanced bone maturity. To our knowledge, there is no literature that has described an anatomic difference in the distal fibular physis based on sex. The association between weight and accuracy of examina- tion is also difficult to explain, especially without having height mea- surements to better understand how weight relates to body habitus. Therefore, it is not possible to draw any reliable conclusions from the in- fluence of sex and weight status on examination accuracy.

To our knowledge, no study has evaluated the accuracy of physical

examination to identify the distal fibular physis. Pesl et al. found that the position of the distal fibular physis is variable, and this seems to be an age-related phenomenon. Of 140 patients with an Ankle sprain without fracture who underwent an ankle radiograph, 9% had a physis located above the tibiotalar joint, 44% at the same level, and 46% distal to the joint space. With age, the distal fibular physis descends down- wards, with a mean age of 2 years old for a physis above the tibiotalar joint, 7 years old for a physis at the same level, and 10 years old for a physis distal to the joint [9]. These findings demonstrate that the posi- tion of the physis varies among children and is not necessarily at the same level as the tibiotalar joint. This could make it more difficult to find its position with physical examination and could explain the low accuracy rate we found in our study.

Boutis et al. published a study in 2007, which found that physical function at four weeks was better with a removable brace compared to a below knee walking cast in a cohort of 104 children with an acute low-risk ankle fracture. Of this group, 72% had a clinical diagnosis of SH1 fracture of the distal fibula [10]. The earlier recovery of physical

function with a removable brace may partially be explained by the large percentage with a clinical diagnosis of SH1 fracture. Many of these children probably did not have an actual fracture given that we now know the physical examination is inaccurate for identifying the distal fibular physis. Sankar et al. provided further support in 2008 for the use of a functional ankle brace rather than casting when they found that only 18% of 38 acute ankle injuries in children presenting with distal fibular tenderness and normal radiographs had evidence of periosteal bone formation on follow-up radiographs. This suggests that the vast majority of these injuries were not actual fractures [11].

This concept is supported by a subsequent study by Boutis et al. pub- lished in 2016 that found the rate of MRI-confirmed SH1 fracture of the distal fibula, when suspected clinically, is only 3% and are most often ankle sprains or bone contusions [2]. Our study may in part explain the low rate of SH1 fractures, as some patients may have been enrolled without truly having tenderness over the distal fibular physis secondary to inaccurate physical examination. These aforementioned studies as well as our study should call into question the traditional diagnosis and management of SH1 fractures of the distal fibula. Clinically, this sug- gests these children should be able to avoid unnecessary prolonged im- mobilization and orthopedic follow-up. Even when MRI confirmed the presence of a SH1 fracture of the distal fibula these patients had the same functional outcome at one month with the use of a removable splint and returning to activity as tolerated [2].

    1. Limitations

We relied on a convenience sample of children that were enrolled based on the POCUS fellow’s availability, which could potentially intro- duce a selection bias. However, it is unlikely that the clinician’s ability to localize the physis or differences in a patient’s anatomy would vary based on time of day or week. In addition, we excluded patients with lower extremity injury, so our results may not necessarily hold true in patients with acute ankle injuries. However, an acutely injured ankle is likely more difficult to assess due to possible bruising, swelling, pain, and apprehension on the part of the patient. Because of the SARS-CoV-2 pandemic, all research that was unrelated to SARS-CoV-2 was suspended at the hospital in March 2020. We therefore stopped en- rolment early and were unable to reach our target sample size of 90 pa- tients. However, based on our sensitivity analysis, physician accuracy would still not have been high enough to be Clinically meaningful. We performed our study at an urban academic pediatric emergency depart- ment, so our findings may not be generalizable to community emer- gency departments that see a mix of both adult and pediatric patients. However, we would expect specialized academic PEM physicians to have better pediatric physical examination skills. One physician exam- ined 21% of the patients enrolled, so this could possibly have introduced bias in the accuracy rate. While we assessed inter-rater reliability based on review of cine clips, we did not assess Interobserver agreement for the identification of the sonographic position of the physis during live scanning. Finally, to our knowledge, there has been no published re- search directly confirming POCUS as an accurate method of physis

identification based on comparison to other imaging modalities. How- ever, POCUS has been successfully used in studies to measure physeal plate width (including the distal fibula) in uninjured children [4,5], and ultrasound has also been used to assess the growth plate in patients with ankle injuries [6,7].

  1. Conclusions

PEM physicians were only 34% accurate with their physical exami- nation in identifying the distal fibular physis in children without lower extremity injury. As physicians are unable to accurately identify the dis- tal fibular physis, the clinical utility of the diagnosis of SH1 fracture of the distal fibula should be called into question. Future studies should be conducted to assess the role of POCUS in children presenting with acute ankle injury.

Funding

None.

Author contribution

VS, JSA, and DJK conceived and designed the study, with input dur- ing the process by VD and SE. VD and VS implemented the study and collected data. VD and SE maintained and analyzed the data. VD drafted the manuscript. All authors contributed substantially to its revision. DJK takes responsibility for the paper as a whole.

Ethics approval

UBC Children’s and Women’s Research Ethics Board, H18-03518.

Prior presentations

Presented as a virtual poster presentation at the American College of Emergency Physicians (ACEP) 2020 Research Forum on October 26-29, 2020.

Declaration of Competing Interest

VD, VS, JSA, and SE do not report any conflicts of interest. DJK is on the medical advisory board of Clarius Mobile Health.

References

  1. Benedict FT, Anderson AC. Injury-ankle. In: Shaw KN, Bachur RG, editors. Fleisher and Ludwig’s textbook of pediatric emergency medicine. 7th ed. Philadelphia: Wolters Kluwer; 2015. p. 238-46.
  2. Boutis K, Plint A, Stimec J, Miller E, Babyn P, Schuh S, et al. Radiograph-negative lat- eral ankle injuries in children: occult growth plate fracture or sprain? JAMA Pediatr. 2016;170:e154114.
  3. Salter RB, Harris WR. Injuries involving the epiphyseal plate. J Bone Joint Surg Am. 1963;45:587-622.
  4. Ng L, Saul T, Lewiss RE. Sonographic baseline physeal plate width measurements in healthy, uninjured children. Pediatr Emerg Care. 2014;30:871-4.
  5. Beatty E, Audette L, Pare A, Simonyan D, Dion V, Berthelot S, et al. Normal bedside ultrasound of growth plates in healthy children. CJEM. 2018;20:S36.
  6. Farley FA, Kuhns L, Jacobson JA, DiPietro M. Ultrasound examination of ankle injuries in children. J Pediatr Orthop. 2001;21:604-7.
  7. Gleeson AP, Stuart MJ, Wilson B, Philips B. ultrasound assessment and conservative management of inversion injuries of the ankle in children: plaster of Paris versus tubigrip. J Bone Joint Surg Br. 1996;78:484-7.
  8. Najaf-Zadeh A, Nectoux E, Dubos F, Happiette L, Demondion X, Gnansounou M, et al. Prevalence and clinical significance of Occult fractures in children with radiograph- negative acute ankle injury. A meta-analysis. Acta Orthop. 2014;85(5):518-24.
  9. Pesl T, Havranek P, Nanka O. Mutual position of the distal fibular physis and the tibiotalar joint space – radiological typology and clinical significance. Eur J Pediatr Surg. 2007;17:348-53.
  10. Boutis K, Willan AR, Babyn P, Narayanan UG, Alman B, Schuh S. A randomized, con- trolled trial of a removable brace versus casting in children with low-risk ankle frac- tures. Pediatrics. 2007;119:e1256-63.
  11. Sankar WN, Chen J, Kay RM, Skaggs DL. Incidence of occult fracture in children with acute ankle injuries. J Pediatr Orthop. 2008;28:500-1.