Orthopedics

Feature Article 

The Utility of Intraoperative Arthrogram in the Management of Pediatric Lateral Condyle Fractures of the Humerus

John Schoeneman Vorhies, MD; Shawn Funk, MD; Marilyn Elliott, BS; Anthony Riccio, MD; Brandon Ramo, MD

Abstract

Intraoperative arthrograms are commonly used in conjunction with closed reduction and percutaneous pinning (CRPP) of pediatric lateral condyle fractures of the humerus. The authors sought to determine how arthrograms affect management of these fractures. They reviewed all lateral condyle fractures treated surgically at a pediatric level I trauma center from 2008 to 2014. They stratified patients managed with and without an arthrogram as well as by timing of arthrogram. The authors compared injury parameters, initial and postoperative fracture displacement, and complications between groups. They identified 107 patients who were taken to the operating room for attempted closed reduction, which they classified as either CRPP without arthrogram or arthrogram first and then a decision to treat open or with CRPP. Fifty-eight (54.21%) underwent CRPP without arthrogram and 49 (45.79%) underwent arthrogram. Of those who had arthrograms, 27 (25.23%) were prior to fixation and 22 (20.56%) were after fixation. There was no difference in age, weight, or preoperative displacement among the groups. Mean postoperative displacement was significantly lower in the no arthrogram group vs the arthrogram group (0.91 mm vs 1.68 mm; P<.0001), but it did not differ based on timing of arthrogram (P=.836). Arthrograms changed management in 4 (8%) of 49 patients who had them. There was no statistical difference in the rate of changed management by timing of arthrogram (before vs after fixation, 14.8% vs 0%; P=.060). The authors demonstrated that arthrograms may be useful for assessing final fracture alignment after CRPP, but are unlikely to result in a treatment change and are not associated with improved postoperative alignment. [Orthopedics. 2020; 43(1):30–35.]

Abstract

Intraoperative arthrograms are commonly used in conjunction with closed reduction and percutaneous pinning (CRPP) of pediatric lateral condyle fractures of the humerus. The authors sought to determine how arthrograms affect management of these fractures. They reviewed all lateral condyle fractures treated surgically at a pediatric level I trauma center from 2008 to 2014. They stratified patients managed with and without an arthrogram as well as by timing of arthrogram. The authors compared injury parameters, initial and postoperative fracture displacement, and complications between groups. They identified 107 patients who were taken to the operating room for attempted closed reduction, which they classified as either CRPP without arthrogram or arthrogram first and then a decision to treat open or with CRPP. Fifty-eight (54.21%) underwent CRPP without arthrogram and 49 (45.79%) underwent arthrogram. Of those who had arthrograms, 27 (25.23%) were prior to fixation and 22 (20.56%) were after fixation. There was no difference in age, weight, or preoperative displacement among the groups. Mean postoperative displacement was significantly lower in the no arthrogram group vs the arthrogram group (0.91 mm vs 1.68 mm; P<.0001), but it did not differ based on timing of arthrogram (P=.836). Arthrograms changed management in 4 (8%) of 49 patients who had them. There was no statistical difference in the rate of changed management by timing of arthrogram (before vs after fixation, 14.8% vs 0%; P=.060). The authors demonstrated that arthrograms may be useful for assessing final fracture alignment after CRPP, but are unlikely to result in a treatment change and are not associated with improved postoperative alignment. [Orthopedics. 2020; 43(1):30–35.]

Lateral condylar humerus fractures represent 10% to 20% of pediatric elbow fractures and are the second most common operative elbow injury in children.1,2 Nonoperative treatment of displaced fractures has been associated with nonunion and deformity.3 Casting or in situ percutaneous pinning is commonly advocated for minimally displaced or nondisplaced injuries, while open or closed reduction and percutaneous pinning (CRPP) is used to treat more displaced fractures.3

Controversy exists regarding the indications for open reduction and the criteria for acceptable reduction in these fractures. Many authors have suggested that all lateral condyle fractures of the distal humerus with displacement of greater than 2 mm be treated open, while others have suggested that fractures with lateral displacement of greater than 2 mm but articular congruency demonstrated by arthrogram can be treated safely with CRPP.4,5 Some authors have suggested that fractures with complete articular displacement and rotation can be successfully treated with CRPP6,7 variably combined with arthrograms to confirm articular reduction.

Assessing the degree of intra-articular displacement in lateral condyle fractures in immature elbows can be challenging. Some surgeons employ intraoperative arthrograms to assess initial articular displacement to help decide if closed or open reduction is indicated. Other surgeons employ arthrograms after CRPP is performed to assess the quality of reduction. Despite the common use of this diagnostic procedure, there is no clear consensus on the indications for or the timing of arthrograms.

The purpose of this study was to describe patterns of use of intraoperative arthrograms of the elbow during the treatment of pediatric lateral condyle fractures of the humerus at the authors' institution and to determine how often intraoperative arthrograms affect surgical management decisions. Secondarily, the authors compared postoperative residual fracture displacement and complication rates based on arthrogram use.

Materials and Methods

An institutional review board–approved retrospective review of all lateral condyle fractures treated with CRPP at an academic, urban, pediatric, level I trauma center from 2008 to 2014 was performed. The operations were performed by 9 fellowship-trained pediatric orthopedic surgeons. Use and timing of arthrograms were at the discretion of the attending surgeon. Preoperative, intraoperative, and postoperative radiographs were reviewed. Fractures were classified using the Song classification.6 Fracture displacement was measured from the lateral metaphyseal cortex of the distal part of the humerus to the lateral cortex of the fracture fragment on the anteroposterior and oblique views (if obtained) and along the posterior cortex on the lateral radiograph, similar to previously described methods.8 The maximum displacement measured on any single view was documented as the maximum displacement of the fragment. All fractures that underwent an intraoperative arthrogram were identified and stratified by the timing of arthrogram administration.

Medical records were reviewed to determine demographic data and operative details. Injury parameters, initial and postoperative fracture displacement, and complications were compared between fractures managed with and without an arthrogram as well as between those that had an arthrogram prior to fixation and those in which the arthrogram was performed following reduction and fixation. Chi-square and Fisher exact tests were used to compare categorical variables, and t tests and Mann–Whitney U tests were used to compare means. All analyses were performed using Stata IC 14 software (StataCorp, College Station, Texas).

Results

A total of 524 patients with lateral condyle fractures were treated operatively at the authors' institution between 2008 and 2014. Of these, 417 were treated open without any further diagnostic testing after initial radiographs. The remaining 107 patients were taken to the operating room for attempted closed reduction. These patients were stratified as having either CRPP without arthrogram, CRPP with subsequent arthrogram to assess intraarticular alignment after fixation, or an arthrogram as a first step to assist in determining whether closed or open reduction was indicated. The authors excluded 2 of these patients who had ipsilateral elbow dislocations from the analysis, as the presence of a dislocation may have affected the decision to perform, or the interpretation of, an arthrogram. All patients had preoperative anteroposterior and lateral radiographs, and 28% of patients also had oblique radiographs. Postoperatively, only anteroposterior and lateral radiographs were obtained.

Among the 107 patients taken to the operating room for possible CRPP, maximum mean preoperative fracture displacement on any view was 2.7 mm (range, 0.2–25.9 mm). Five elbows (4.67%) were Song type 1, 32 elbows (29.91%) were type 2, 8 elbows (7.48%) were type 3, 60 elbows (56.07%) were type 4, and 2 elbows (1.87%) were type 5. Seventy (65.42%) patients had low-energy mechanisms of injury, such as fall from standing or running on otherwise level ground; 36 (33.64%) had medium-energy mechanisms, such as fall from a bicycle or a playground structure; and 1 injury was the result of a motor vehicle collision and therefore deemed high energy.

Fifty-eight patients (54.21%) underwent CRPP without arthrogram, and 49 patients (45.79%) underwent an elbow arthrogram. There was no difference in the prevalence of patients who had preoperative oblique radiographs between those treated with or without an arthrogram (22.41% vs 34.69%; P=.159). Table 1 summarizes baseline demographics, injury characteristics, and postoperative outcomes of the entire cohort, divided into arthrogram and no arthrogram groups. There was no difference in mean age or weight at presentation. Mechanisms of injury were similar, and there was no difference in mean time from injury to surgery between the 2 groups. There was no difference in mean preoperative displacement or distribution within the Song classification between the 2 groups. Mean postoperative displacement was significantly lower in the no arthrogram group vs the arthrogram group (0.91 mm vs 1.68 mm; P<.0001). There was no difference in the rate of pin site infection, delayed union, or referral to occupational therapy for improvement of range of motion between the 2 groups. Table 2 details the distribution of arthrogram use by surgeon.

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the Arthrogram and No Arthrogram Groups

Table 1:

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the Arthrogram and No Arthrogram Groups

Distribution of Arthrogram Use by Surgeon

Table 2:

Distribution of Arthrogram Use by Surgeon

Twenty-seven patients (25.23%) had an arthrogram prior to fixation, and 22 (20.56%) had an arthrogram after fixation. There was no difference in the prevalence of oblique radiographs between patients who had arthrograms before vs after fixation (33.33% vs 36.36%; P=.825). Table 3 summarizes baseline demographics, injury characteristics, and postoperative outcomes among patients who had arthrograms, subdivided by the timing of the arthrogram relative to fixation. There was no difference in mean age or weight at presentation. Mechanisms of injury were not significantly different between the 2 groups. Patients who had arthrograms prior to fixation had a significantly greater delay from injury to surgery of 7.1 days vs 3.9 days in the group who had arthrograms after fixation. There was no difference in mean preoperative displacement or distribution within the Song classification between the 2 groups. There was no difference in mean postoperative displacement between groups. There was no difference in the rate of pin site infection or referral to occupational therapy for improvement of range of motion between the 2 groups. Table 4 details the distribution of arthrogram timing by those surgeons who used arthrograms.

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the 2 Arthrogram Timing Groups

Table 3:

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the 2 Arthrogram Timing Groups

Distribution of Arthrogram Timing by Surgeons Who Used Arthrograms

Table 4:

Distribution of Arthrogram Timing by Surgeons Who Used Arthrograms

The authors performed a subanalysis of the 60 patients in their cohort who had Song type 4 fractures. Twenty-seven (45%) of these patients had CRPP without an arthrogram and 33 (55%) had an arthrogram. There was no difference in the prevalence of patients who had preoperative oblique radiographs between those treated with or without an arthrogram (45.45% vs 33.33%; P=.159). There was no difference in mean age or weight at presentation or mean time from injury to presentation. Mechanisms of injury were not significantly different between the 2 groups. Mean postoperative displacement was significantly lower in the no arthrogram group vs the arthrogram group (1.27 mm vs 1.85 mm; P=.023). There was no difference in the rate of pin site infection, delayed union, or referral to occupational therapy for improvement of range of motion between the 2 groups.

Among the 33 Song 4 patients treated with arthrograms, 16 (48.48%) had an arthrogram prior to fixation and 17 (51.52%) had an arthrogram after fixation. There was no difference in the prevalence of patients who had preoperative oblique radiographs between those who had arthrograms before or after fixation (47.06% vs 43.75%; P=.849). There was no difference in mean age or weight at presentation. Mechanisms of injury were not significantly different between the 2 groups. Patients who had arthrograms prior to fixation had a significantly greater delay from injury to surgery of 6.4 days vs 3.5 days in the group who had arthrograms after fixation (P=.015). There was no difference in postoperative displacement, rate of pin site infection, delayed union, or referral to occupational therapy for improvement of range of motion between the 2 groups.

Overall, arthrogram findings changed management for 4 patients, representing 8% of the 49 patients who had arthrograms. In all cases where arthrogram changed management, arthrogram was performed prior to definitive fixation (14.8% of the pre-fixation arthrograms). Despite this, there was no statistically significant difference in the rate of changed management by timing of arthrogram before vs after fixation (14.8% vs 0%; P=.060). In 2 patients, an arthrogram was performed prior to fixation and revealed a magnitude of displacement deemed sufficient enough to warrant conversion to open reduction. An example of 1 of these cases is shown in Figure 1. In 1 case, a provisional reduction was attempted with a Kirschner wire as a joystick and an arthrogram was subsequently performed prior to definitive fixation. Articular displacement was noted, so the operation was converted to open. In 1 case, a patient with minimal displacement was initially treated in a cast, but subsequent follow-up demonstrated concern for fracture displacement. The patient was taken to the operating room for possible CRPP. An arthrogram revealed minimal articular displacement, so CRPP was aborted and the patient was placed in a cast for definitive treatment.

Case example in which a pre-fixation arthrogram led to a conversion to open reduction. Injury anteroposterior (A) and lateral (B) radiographs of a 2-year-old boy who had a lateral condyle distal humerus fracture from a mechanical fall. Seven days after the injury, he was taken to the operating room for an arthrogram (C), which revealed wide varus displacement of the capitellum. The case was subsequently converted to open reduction and fixation with 2 parallel pins (D, E). Anteroposterior (F) and lateral (G) radiographs 9 weeks postoperatively.

Figure 1:

Case example in which a pre-fixation arthrogram led to a conversion to open reduction. Injury anteroposterior (A) and lateral (B) radiographs of a 2-year-old boy who had a lateral condyle distal humerus fracture from a mechanical fall. Seven days after the injury, he was taken to the operating room for an arthrogram (C), which revealed wide varus displacement of the capitellum. The case was subsequently converted to open reduction and fixation with 2 parallel pins (D, E). Anteroposterior (F) and lateral (G) radiographs 9 weeks postoperatively.

Discussion

Intraoperative arthrograms are commonly performed in the treatment of lateral condyle fractures,3 but there are few reports delineating how they actually affect management. Many large series of CRPP for lateral condyle fractures describe incorporating an elbow arthrogram either before or after fixation, but do not report how often the arthrogram affected management.6,9,10 There are some smaller series in the literature describing how arthrograms affect management. Yates and Sullivan11 described the evaluation of 12 suspected lateral condyle fractures with arthrogram leading to the diagnosis of epiphyseal separation in 4. Marzo et al12 described the evaluation of 16 lateral condyle fractures with arthrogram before fixation. All of the patients in this series were taken to the operating room within 24 hours of injury. They reported 7 conversions to open and 3 conversions to nonoperative treatment. Pennock et al5 reported on 23 lateral condyle fractures treated with CRPP. Fourteen (61%) of these had arthrograms, all after fixation. There were no instances of conversion to open treatment or modification of reduction.

This study had several limitations. It was a retrospective study describing the work of several different surgeons, each employing his or her own treatment algorithm. Given the variability of arthrogram use among the surgeons, there was clearly disagreement among practitioners as to if and when arthrograms were indicated. For example, surgeon 2 performed 37 cases and never used an arthrogram, surgeon 6 performed 9 cases and always used an arthrogram, and surgeon 9 performed 14 cases and used arthrograms in exactly half. Similar variability was found regarding arthrogram timing. The authors' inability to know the details of each individual surgeon's algorithm was a weakness, but the presence of different algorithms allowed for a fairly even distribution of injury severity across the different modes of treatment, which was a clear strength. The primary goals of the study were to determine the manner and the frequency with which arthrograms affect management; thus, the authors included several cases with short follow-up. The short mean follow-up period precluded their ability to accurately link the incidence of long-term outcomes to treatment method. The authors were unable to accurately describe the incidence of osteonecrosis, growth disturbance, late deformity, or other long-term complications. Because of the retrospective nature of the study, the authors did not have standardized methods for documenting short-term clinical outcomes, such as range of motion, functional outcomes, or patient-reported quality of life measures. They did employ a proxy measure for postoperative stiffness, referral to occupational therapy. However, they acknowledge that the decision to refer was likely provider dependent.

Elbow arthrogram before CRPP resulted in a treatment change in approximately 15% of the cases, whereas no patient who had an arthrogram after CRPP had a change in management. This difference did not, however, reach statistical significance. Furthermore, 1 of the cases in which management was changed did have a provisional reduction attempt with a joystick wire, but no definitive fixation. Timing of arthrogram varied by surgeon, and the rate of change in management was likely related to individual surgeons' beliefs regarding the threshold for open reduction in the presence of intra-articular displacement. Although the authors found that mean postoperative displacement was lower in patients without arthrograms, the difference was exceedingly small and likely not clinically relevant. This finding was true among all patients as well as in the Song 4 subanalysis. In addition, this difference may represent selection bias, as surgeons may have believed that an arthrogram was unnecessary when displacement following a closed reduction maneuver was small. Alternatively, postoperative displacement may be lower in patients without arthrograms because displacement of the lateral column, rather than true intraarticular displacement, was the primary means by which the surgeon was able to judge adequacy of reduction in the absence of an arthrogram. A surgeon who demonstrates perfect articular congruity on arthrogram may be more likely to accept a lateral gap as a manifestation of plastic deformation. The authors found no differences in any other postoperative outcomes by use of arthrogram or timing of arthrogram in the overall cohort or the Song 4 subgroup.

Conclusion

This is the largest published series of CRPP of pediatric lateral condyle fractures to describe patterns of arthrogram use and how it affects management. The authors did not find any evidence of increased morbidity associated with the use of arthrograms. They demonstrated that an arthrogram following CRPP of a lateral condyle fracture may be useful for assessing final fracture alignment, but it is unlikely to result in a treatment change and was not associated with improved postoperative alignment.

References

  1. Landin LA. Fracture patterns in children: analysis of 8,682 fractures with special reference to incidence, etiology and secular changes in a Swedish urban population 1950–1979. Acta Orthop Scand Suppl. 1983;202:1–109. PMID:6574687
  2. Landin LA, Danielsson LG. Elbow fractures in children: an epidemiological analysis of 589 cases. Acta Orthop Scand. 1986;57(4):309–312. https://doi.org/10.3109/17453678608994398 PMID: doi:10.3109/17453678608994398 [CrossRef]3788491
  3. Sullivan JA. Fractures of the lateral condyle of the humerus. J Am Acad Orthop Surg. 2006;14(1):58–62. https://doi.org/10.5435/00124635-200601000-00010 PMID: doi:10.5435/00124635-200601000-00010 [CrossRef]16394169
  4. Mintzer CM, Waters PM, Brown DJ, Kasser JR. Percutaneous pinning in the treatment of displaced lateral condyle fractures. J Pediatr Orthop. 1994;14(4):462–465. https://doi.org/10.1097/01241398-199407000-00008 PMID: doi:10.1097/01241398-199407000-00008 [CrossRef]8077428
  5. Pennock AT, Salgueiro L, Upasani VV, Bastrom TP, Newton PO, Yaszay B. Closed reduction and percutaneous pinning versus open reduction and internal fixation for type II lateral condyle humerus fractures in children displaced >2 mm. J Pediatr Orthop. 2016;36(8):780–786. https://doi.org/10.1097/BPO.0000000000000570 PMID: doi:10.1097/BPO.0000000000000570 [CrossRef]
  6. Song KS, Kang CH, Min BW, Bae KC, Cho CH, Lee JH. Closed reduction and internal fixation of displaced unstable lateral condylar fractures of the humerus in children. J Bone Joint Surg Am. 2008;90(12):2673–2681. https://doi.org/10.2106/JBJS.G.01227 PMID: doi:10.2106/JBJS.G.01227 [CrossRef]19047713
  7. Song KS, Shin YW, Oh CW, Bae KC, Cho CH. Closed reduction and internal fixation of completely displaced and rotated lateral condyle fractures of the humerus in children. J Orthop Trauma. 2010;24(7):434–438. https://doi.org/10.1097/BOT.0b013e3181de014f PMID: doi:10.1097/BOT.0b013e3181de014f [CrossRef]20577074
  8. Song KS, Kang CH, Min BW, Bae KC, Cho CH. Internal oblique radiographs for diagnosis of nondisplaced or minimally displaced lateral condylar fractures of the humerus in children. J Bone Joint Surg Am. 2007;89(1):58–63. https://doi.org/10.2106/JBJS.E.01387 PMID: doi:10.2106/JBJS.E.01387 [CrossRef]17200311
  9. Weiss JM, Graves S, Yang S, Mendelsohn E, Kay RM, Skaggs DL. A new classification system predictive of complications in surgically treated pediatric humeral lateral condyle fractures. J Pediatr Orthop. 2009;29(6):602–605. https://doi.org/10.1097/BPO.0b013e3181b2842c PMID: doi:10.1097/BPO.0b013e3181b2842c [CrossRef]19700990
  10. Justus C, Haruno LS, Riordan MK, et al. Closed and open reduction of displaced pediatric lateral condyle humeral fractures: a study of short-term complications and postoperative protocols. Iowa Orthop J. 2017;37:163–169. PMID:28852352
  11. Yates C, Sullivan JA. Arthrographic diagnosis of elbow injuries in children. J Pediatr Orthop. 1987;7(1):54–60. https://doi.org/10.1097/01241398-198701000-00011 PMID: doi:10.1097/01241398-198701000-00011 [CrossRef]3793912
  12. Marzo JM, d'Amato C, Strong M, Gillespie R. Usefulness and accuracy of arthrography in management of lateral humeral condyle fractures in children. J Pediatr Orthop. 1990;10(3):317–321. https://doi.org/10.1097/01241398-199005000-00004 PMID: doi:10.1097/01241398-199005000-00004 [CrossRef]2355073

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the Arthrogram and No Arthrogram Groups

ParameterUse of ArthrogramP

NoYes
Demographics
  Patients, No.5849
  Age, mean, y4.84.4.372
  Weight, mean, kg22.322.3.956
Injury characteristics
  Mechanism of injury, No.
    Low energy35 (60.35%)35 (71.43%)
    Medium energy22 (37.93%)14 (28.57%).356
    High energy1 (1.72%)0
  Time from injury to surgery, mean, d6.85.7.271
  Song classification, No.
    141
    22210
    335.061
    42733
    520
  Preoperative displacement, mean, mm3.042.91.836
Postoperative outcomes
  Displacement, mean, mm0.911.68<.001a
  Pin site infection, No.0 (0%)2 (4.1%).207
  Delayed union, No.1 (1.7%)0 (0%).999
  Referral to occupational therapy for range of motion, No.7 (12.0%)4 (8.2%).544

Distribution of Arthrogram Use by Surgeon

Surgeon NumberArthrogram Use, No.Total No.

NoYes
1101
237037
3134
421113
5426
6099
7202
841721
97714
Total no.5849107

Baseline Demographics, Injury Characteristics, and Postoperative Outcomes of the 2 Arthrogram Timing Groups

ParameterArthrogram TimingP

Before FixationAfter Fixation
Demographics
  Patients, No.2722
  Age, mean, y4.34.5.764
  Weight, mean, kg21.723.2.528
Injury characteristics
  Mechanism of injury, No.
    Low energy22 (81.5%)13 (59.1%).084
    Medium energy5 (18.5%)9 (40.9%)
  Time from injury to surgery, mean, d7.13.9.004a
  Song classification, No.
  110
  282
  323.215
  41617
  500
  Preoperative displacement, mean, mm2.53.4.121
Postoperative outcomes
  Displacement, mean, mm1.681.68.999
  Pin site infection, No.1 (3.7%)1 (4.5%).702
  Referral to occupational therapy for range of motion, No.2 (7.4%)2 (9.1%).999

Distribution of Arthrogram Timing by Surgeons Who Used Arthrograms

Surgeon NumberArthrogram Timing, No.Total No.

Before FixationAfter Fixation
3303
42911
5022
6819
871017
9707
Total no.272249
Authors

The authors are from the Department of Orthopaedic Surgery (JSV), Stanford University Medical Center, Stanford Children's Health, Stanford, California; and the Department of Orthopedic Surgery (SF), The Children's Hospital, San Antonio, the Department of Orthopaedic Surgery (ME, AR, BR), Texas Scottish Rite Hospital for Children, Dallas, and the Department of Orthopaedic Surgery (ME, AR, BR), Children's Medical Center, Dallas, Texas.

Dr Vorhies, Dr Funk, Ms Elliott, and Dr Ramo have no relevant financial relationships to disclose. Dr Riccio has received travel reimbursement from DePuy Synthes.

The authors thank Japsimran Kaur, BS, for assistance in the compilation of data tables.

Correspondence should be addressed to: John Schoeneman Vorhies, MD, Department of Orthopaedic Surgery, Stanford University Medical Center, Stanford Children's Health, 300 Pasteur Dr, Edwards Bldg, Ste R107, Stanford, CA 94305 ( jvorhies@stanford.edu).

Received: February 25, 2019
Accepted: September 09, 2019
Posted Online: November 08, 2019

10.3928/01477447-20191031-01

Sign up to receive

Journal E-contents