Dome impaction classification in acetabular fractures: Identifying atypical dome impactions with absence of the gull sign: A retrospective descriptive study

Abstract

We aimed to describe variations of dome impaction fractures and their characteristics using a new classification system, to explore the relationship between dome impaction and the gull sign. The present study was a retrospective descriptive study. All 104 cases of acetabular fracture that were treated in our institution from 2013 to 2022 were enrolled. Of these, 22 had dome impaction fractures. The primary outcome variable was to describe the variations and characteristics of dome impaction fractures. They were classified into 3 major subgroups based on reconstructed axial, coronal, and sagittal computed tomography findings: anteromedial, superomedial (SM), and posteromedial. The secondary outcome variable was to explore the relationship between dome impaction on computed tomography findings and the gull sign on plain radiographs. There were 4 cases of anteromedial (18.2%), 13 of SM (59.1%), and 5 of posteromedial (22.7%). There were 15 cases (68.2%) with the gull sign and 7 cases (31.8%) without the gull sign on plain radiographs. Twelve of fifteen cases (80.0%) with the gull sign had dome impaction fractures of the SM type. We found a variety of patterns of dome impaction fracture. Surgeons should be aware of atypical dome impactions not showing the gull sign.

1. Introduction

Due to the increase in the aging population in developed countries, the number of acetabular fractures caused by osteoporosis and the proportion of acetabular fractures in elderly patients has been increasing.^[1–3] A study in Finland, for example, reported an increase in the incidence of acetabular fractures in the elderly of 30%, from 17/100,000/persons/year in 1997 to 23/100,000/persons/year in 2014.^[2]

Open reduction and internal fixation for displaced acetabular fractures is necessary even in elderly patients,^[4,5] with a rate of revision to total hip arthroplasty of up to 28% has been reported.^[6] Among numerous factors associated with poor clinical outcomes, dome impaction in acetabular fractures has gained increased attention,^[3,4,6,7] and is commonly encountered in the elderly.^[7,8] The “gull sign” reported by Anglen et al^[9] is a double arc on the radiograph produced by the displaced, impacted subchondral bone of the medial roof. While there has not been clear definition of “dome impaction in acetabular fractures, particularly in computed tomography (CT), Scolaro et al have described dome impaction fractures as “commonly independent of the anterior or posterior column fracture fragments”.^[10] This implies that there is variation in dome impaction fractures and that a dome impaction fracture with a positive gull sign could represent one of several patterns. It is unknown what kind of dome impactions may exist and how to treat them according to fragment-specific features. The purpose of this study was to describe the variations and characteristics of dome impaction fractures using a novel classification system based on CT findings, and to explore the relationship between dome impaction and the gull sign. In addition, whether the classification would affect the clinical outcomes was investigated.

2. Methods

2.1. Study design/setting

The present study was a retrospective descriptive study. This study was carried out in a single academic teaching institution. The patients data were recorded in the case record forms from the database of our institution.

2.2. Participants

From January 2013 to February 2022, there were 104 cases of acetabular fracture treated in our institution. All cases had anterior-posterior and oblique radiographs and CT scans (Aquilion Prime 80, Canon Medical Systems, Tochigi, Japan) on admission. Image data of CT scans were acquired in 0.5-mm sections. Each CT scan was reconstructed in the axial, coronal, and sagittal planes using a picture archiving and communication system.

2.3. Variables/measurement

2.3.1. Dome impaction classification based on CT findings.

The primary outcome variable of the present study was to describe the variations and characteristics of dome impaction fractures based on CT findings. Dome impaction fractures were detected on the multi-planar reconstructed CT scans. Dome impaction fractures were defined as “fractures commonly independent of the anterior or posterior column fracture fragments”.^[10] We also set the threshold for a dome impaction as a 2 mm or over step-off from the original articular arc at the roof of the acetabulum. Then they were classified into 3 types based on the location of each fragment in the weight-bearing zone in the superior 10 mm of the acetabulum: anteromedial (AM), superomedial (SM) and posteromedial (PM) fractures (Fig. 1). Each circle in Figure 1 shows the inferior surface of a weight-bearing zone where articular fragments are virtually repositioned and viewed from below. The gray zones represent dome impaction fragments. If a coronal plane centered on the original position of the femoral head (dotted line in Fig. 1) cut across a dome fragment on axal view, the fragment was assigned the SM type. AM and PM fragments were located anterior and posterior to the plane, respectively. If the posterior end of an anterior-oriented dome impaction fragment is just on the coronal plane (dotted line in Fig. 1), the fragment was classified into AM type. Same applied to PM fragments.

Figure 1.

Dome impaction classification: anteromedial (AM), superomedial (SM), and posteromedial (PM) type.

2.3.2. Evaluation of the gull sign.

The secondary outcome variable of the present study was to explore the relationship between dome impaction and the gull sign. The presence of the gull sign was judged on the plain radiographs and compared with image intensifier images to assess for reduction during surgery. The definition of the gull sign was based on that by Anglen et al^[9] as a double arc on the radiograph produced by the displaced and impacted subchondral bone of the medial roof (Fig. 2).

Figure 2.

The gull sign is the double arc produced by impaction of the superomedial articular surface on the radiograph (white outline).

2.3.3. The dome impaction classification and clinical outcomes.

In addition to the primary and secondary outcomes variables, we investigated the clinical outcomes of patients who underwent open reduction and internal fixation and were followed over 1 year at our institution. Outcome variables included days from injury to surgery, surgical approach, operative time, amount of intraoperative bleeding, quality of reduction, radiographic grade of osteoarthritis evaluated by Matta’s criteria, clinical outcome evaluated by modified Merle D’Aubigne and Postel scoring system. We mainly utilized anterior intrapelvic approach or that combined with 1^st window of ilioinguinal approach for anterior approach in the case series. No case of conventional ilioinguinal approach was included.

2.4. Statistical analysis

All data are expressed as mean ± standard deviation. A kappa statistic was calculated as a measure of the intra- and inter-observer reliability of the 2 independent observers (S.K and T.S) for dome impaction classification and the gull sign. Fisher exact test was used to compare dichotomous variables, and the Kruskal–Wallis test was used for comparisons of continuous variables; P < .05 was considered significant. The analyses were performed using commercial software (SPSS, SPSS Inc., Chicago, IL).

2.5. Compliance with ethical standards

All procedures were performed in accordance with the ethical standards of this committee, as well as with the 1964 Helsinki Declaration and its later amendments. This study was retrospective study and noninvasive; therefore, informed consent was obtained in the form of opt-out and the need to obtain informed consent from the individual patients was waived by the ethical committee of our institution.

3. Results

3.1. Characteristics of participants

The current study included 104 patients diagnosed with acetabular fractures. There were 2 cases with bilateral fractures, totaling 106 fractures in all. The mean age among all cases with acetabular fractures was 61 ± 16 years, and 80 (81.8%) of those cases were male. The fracture patterns were as follows: posterior wall in 16 (15.1%), anterior column in 30 (28.3%), posterior column in 6 (5.7%), transverse in 10 (9.4%), both column in 19 (17.9%), anterior + posterior hemitransverse in 15 (14.2%), T-shaped in 5 (4.7%), Transverse + posterior wall in 4 (3.8%), and posterior column + posterior wall in 1 (1.0%). The mechanisms of injury were traffic accidents in 41 (39.4%), falls from height in 40 (38.4%), simple falls in 9 (2.3%), and others in 14 (13.5%).

3.2. Dome impaction classification based on CT findings

In 104 cases of acetabular fractures, 22 cases (21.2%) of dome impaction fractures were detected with multi-planar reconstructed CT scans. Demographic data and the dome impaction classification and incidence of the gull sign in the 22 cases are shown in Table 1. The mean age in dome impaction fractures was 66 ± 12 years, and 81.8% (18/22) of cases with dome impaction fractures were male. Among the patients with dome impaction fractures, there were 6 cases of both column fractures, 4 of isolated anterior column fractures, 6 of anterior column and posterior hemitransverse fractures, 1 transverse fracture, 3 T-shaped fractures, and 2 isolated posterior column fractures. According to the dome impaction classification, 4 fragments (18.2%) were AM in position, 13 fragments (59.1%) were SM, 5 fragments (22.7%) were PM. Three of 4 AM types were impacted in the pubic rami horizontally, with an anterior variant of the “marginal impaction” appearance, usually accompanied by posterior wall fractures. All SM type fragments were impacted in the ilium medially and superiorly, tilts laterally in the coronal plane. All PM-type fragments were impacted in the ilium or the posterior column, and tilts posteriorly mainly in the sagittal plane. Two of 5 PM types were displaced together with posterior column fragments.

Table 1.

Demographic data and summary of the dome impaction classification and incidence of the gull sign.

Number	Age	Sex	Fracture type	Dome impaction type	Gull sign on x-ray
1	67	F	AC + hTV	SM	+
2	55	M	PC	SM	+
3	68	M	AC	AM	-
4	76	M	AC	SM	+
5	72	M	BC	SM	+
6	70	M	T-shaped	SM	+
7	69	M	AC + hTV	SM	+
8	69	M	AC + hTV	SM	+
9	67	M	BC	PM	-
10	67	M	AC	AM	-
11	55	M	AC	PM	-
12	65	M	PC	SM	+
13	47	M	BC	SM	+
14	71	M	T-shaped	SM	+
15	82	F	BC	PM	-
16	70	M	BC	AM	+
17	81	F	T-shaped	PM	+
18	24	M	BC	PM	+
19	68	M	Transverse	SM	+
20	70	M	AC + hTV	AM	-
21	63	F	AC + hTV	SM	-
22	73	M	AC + hTV	SM	+

AC = anterior column, AC + hTV = anterior column and posterior hemitransverse fractures, AM = anteromedial, BC = both columns, PC = posterior column, PM = posteromedial, SM = superomedial.

Consistent agreement (kappa coefficients range, 0.633–1.0) was noted between the 2 independent observers who classified the dome impactions.

3.3. The relationship between the gull sign and dome impaction classification

Of 22 cases with dome-impacted fragments, there were 15 cases (68.2%) with a positive gull sign and 7 cases (31.8%) with a negative gull sign on the plain radiographs. Twelve of 15 cases (80.0%) with the gull sign had dome impaction fragments of the SM type. The remaining 3 were 1 AM type and 2 PM types. Among 7 cases (31.8%) not showing the gull sign, there were 3 cases of AM type, 1 SM type, and 3 PM types. Twelve out of 13 cases (92.3%) of SM types had the gull sign, whilst 1 out of 4 cases of AM type and 2 cases out of 5 PM types presented with the gull sign (P = .011, Figure 3).

Figure 3.

The relationship between the dome impaction classification and the gull sign.

3.4. The dome impaction classification and clinical outcomes

Three patients were excluded, because 1 patient underwent acute total hip replacement and 2 patients had not been followed at our institution. Among 19 patients, there were no significant differences in days from time to surgery, surgical approaches, operative time, amount of intraoperative bleeding, quality of reduction, radiographic grade of osteoarthritis, clinical outcome scores among the 3 types of the dome classification (Table 2).

Table 2.

The dome impaction classification and the clinical outcomes.

	AM	SM	PM	P value
Number of cases	3	11	5	-
Days from injury to surgery	5 ± 2	5 ± 2	6 ± 4	.896
Surgical approach				.179
Anterior	3	9	3
Posterior	0	1	0
Anterior + posterior	0	1	2
Operative time (min)	251 ± 75	220 ± 22	332 ± 105	.110
Intraoperative bleeding (g)	1013 ± 439	973 ± 523	1694 ± 1204	.543
Quality of reduction				.434
Anatomical	2	7	1
Imperfect	1	2	3
Poor	0	2	1
Osteoarthritis				.779
Excellent	2	7	4
Good	1	3	3
Fair	0	1	0
Poor
Clinical score	17 ± 2	16 ± 3	16 ± 2	.589

AM = anteromedial, PM = posteromedial, SM = superomedial.

3.5. Case presentation

Case 1 represent the case of AM type. A 70-year-old patient with the anterior column and posterior hemitransverse fractures. The anterior-posterior radiograph did not show the gull sign (Fig. 4A). The CT axial image (Fig. 4B) and 3D-CT (Fig. 4C) revealed the dome impaction fracture as being of the AM type (white arrow). The dome fragment looks like an anterior variant of marginal impaction. This type of dome fragment is the hardest to visualize intraoperatively, though the inlet view or the inlet-oblique view (Fig. 4D) is helpful in recognizing it. This fragment was addressed via an intrapelvic approach and delivered through the fracture gap between the anterior column fragment and upper pubic ramus. The postoperative radiograph (Fig. 4E) and the CT axial image (Fig. 4F) showed acceptable reduction of the AM type dome fragment.

Figure 4.

Case 1 represent the case of AM type. The anterior-posterior radiograph did not show the gull sign. (A) The CT axial image, (B) and 3D-CT, (C) revealed the dome impaction fracture as being of the AM type (white arrow). The inlet-oblique view, and (D) is helpful in recognizing it. The postoperative radiograph, (E) and the CT image, and (F) showed acceptable reduction of the AM type dome fragment. The anterior column with posterior hemitransverse fracture was stabilized using two reconstruction plates, one placed infrapectineally and the other suprapectineally. AM = anteromedial, CT = computed tomography.

Case 2 represent the case of SM type. 69-year-old patient with the anterior column and posterior hemitransverse fractures. The anterior-posterior radiograph showed the gull sign (Fig. 5A). CT axial image (Fig. 5B) and 3D-CT (Fig. 5C) findings revealed a dome impaction fracture of the SM type. This type of dome fragment is easily visualized intraoperatively on the anterior-posterior view with the image intensifier. The fragment was located under the anterior column fragment and addressed via an intrapelvic approach through the fracture gap below the pelvic brim. The postoperative radiograph (Fig. 5D) and axial CT image (Fig. 5E) shows anatomical reduction of the SM type fragment.

Figure 5.

Case 2 represent the case of SM type. The anterior-posterior radiograph showed the gull sign. (A) The CT axial image, (B) and 3D-CT, (C) findings revealed a dome impaction fracture of the SM type (white arrow). The postoperative radiograph, (D) and axial CT image, and (E) shows anatomical reduction of the SM type fragment. The anterior column with posterior hemitransverse fracture was stabilized using a dedicated infrapectineal plate and a reconstruction plate employed as an infrapectineal plate. CT = computed tomography, SM = superomedial.

Case 3 represent the case of PM-type. A 67-year-old man who fell from a 3-meter height with the both column fractures. The anterior-posterior radiograph does not show the gull sign (Fig. 6A). The dome impaction fracture was classified as PM based on CT axial image (Fig. 6B), CT sagittal image (Fig. 6C), and 3D-CT (Fig. 6D) findings. The dome fragment was displaced together with the posterior column fragment. This type of dome fragment is visualized better on the iliac oblique view (Fig. 6E) and/or the inlet view rather than the anterior-posterior view with the image intensifier. The fragment was addressed via an intrapelvic approach through the fracture gap below the pelvic brim, and reduced before closure of the fracture gap between the anterior and posterior column fragments. The postoperative radiograph (Fig. 6F) and CT axial image (Fig. 6G) and VCT sagittal image (Fig. 6H) showed acceptabe reduction of the PM-type fragment.

Figure 6.

Case 3 represent the case of PM-type. The anterior-posterior radiograph does not show the gull sign. (A) The dome impaction fracture was classified as PM based on axial CT image, (B) sagittal CT image, (C) and 3D-CT, (D) findings. This type of dome fragment is visualized better on the iliac oblique view, (E) with the image intensifier. The postoperative radiograph (F) and axial CT image, (G) and sagittal CT image, and (H) showed acceptabe reduction of the PM-type fragment. The both columns fracture was stabilized using three reconstruction plates, one for iliac fixation, and the remaining two were employed as infrapectineal and suprapectineal plates. CT = computed tomography, PM = posteromedial.

4. Discussion

We revealed that there are a variety of patterns of dome impaction fragments in acetabular fractures. We classified dome impactions into 3 types: AM, SM, and PM. All types had medial directionality because dome impactions occurred with force in the superior and medial directions.^[9–11] Several modifiers used for dome impaction have been used, namely “anteromedial,”^[12,13] “superomedial,”^{[3,9,14–19]} or “anterosuperior”.^[20] These terms express the directions of forces loaded on the dome impaction fragments, though surgeons should understand the location of a dome impaction fragment as well as the direction of force for preoperative planning. This is why we defined anteromedial, superomedial, and posteromedial dome impaction in an articulate manner. As anticipated, the most common pattern was the SM type, followed by PM. The reason why the AM type was least common could be explained by the fact that the range of motion in hip joint is lower in extension than flexion.

Anatomical reduction and stable fixation of dome impaction fragments are the key for gaining good functional recovery as in any intraarticular fractures. Techniques for articular reduction have been described by several authors. Laflamme and Herbert-Davies^[16] described a direct reduction technique for dome impaction via an anterior intrapelvic approach or modified ilioinguinal approach after mobilizing the medialized quadrilateral plate fragment posteroinferiorly below the pelvic brim. In addition, Casstevens et al^[17] reported the optional technique of osteotomy along the pelvic brim for visualization of a dome impaction fragment. Collinge and Lebus^[21] also described detailed techniques for reduction of dome impaction when using the anterior intrapelvic approach. Scolaro and Rout^[10] described the technique of using an independent cortical window insertion of reduction tools to reduce the dome fragment. This technique is considered useful for a superior dome fragment that is not connected to the primary fracture fragment^[22] (Kasha and Yalamanchili). Considering the association between the reduction techniques mentioned above and the classification of dome impactions, these techniques are for SM type fragments and some PM-type fragments. SM type dome impaction fragments would be addressed through a fracture gap at the medial edge of the anterior column fragment, or through a cortical window located just cephalad to the fragment. Any kind of anterior approach for acetabular fractures would be chosen for treating SM type fragments according to the surgeon preference. A PM-type fragment is often connected to the posterior column fragment and displaced with it. This type of dome impaction fragment would be directly approached via an anterior intrapelvic approach accentuating displacement of the posterior column fragment below the pelvic brim, and reducing them before articular repositioning of the posterior column fragment. The iliac cortical window technique is also possible, whilst the surgeon might have trouble reducing the dome fragment through the window due to the distance from the iliac cortical surface to the dome fragment. AM type fragments were often impacted in pubic rami horizontally, and looked like an anterior variant of marginal impaction fracture. This type of dome impaction fragment has not been described previously. It would be addressed and reduced through the fracture gap between the pubic ramus and the iliac fragment before their reduction. Reduction of the fragments through an iliac cortical window technique is not suitable for this type of fragment because the dome fragment should be disimpacted toward the articular surface horizontally.

The “gull sign” in acetabular fractures was originally referred to in the textbook by Letournel and Judet.^[23] The gull sign in this book is not a sign of impaction fracture of the articular surface but a double arc indicating particular types of posterior column fractures. Thus, the gull sign of the present study represents the dome impaction first described by Anglen et al^[9] and is a different entity. In this study, the gull sign was present in 15 of 22 (68.2%) dome impaction cases. The majority of cases with the gull sign positive presented with the SM type (12 of 15; 80%). SM type fragment exists on the top of the acetabulum and tilts in the coronal plane, thus the anterior-posterior view is appropriate to visualize it. In our series, 7 of 22 fragments (31.8%) did not show the gull sign, especially the non-SM types. Due to its location and direction of the fragments, the AM fragment of anterior marginal impaction does not show the gull sign. In 3 of 4 AM types, we did not see the gull sign on the plain radiographs. Similarly, PM fragments that tilt mainly in the sagittal plane or a fragment that displaces with the posterior column fragment do not show the gull sign. In 3 of 5 PM types, there were no positive gull signs on the plain radiographs. We have revealed that there were unignorable numbers of dome impaction fractures not showing the gull sign on plain radiographs. Not showing the gull sign do not necessarily indicate the case do not have a dome impaction fracture. Surgeons should address each dome impaction fracture according to its location and pay meticulous attention to dome impactions not showing the gull sign as they require some ingenuity to visualize them intraoperatively.

The any variables for clinical outcomes did not differ significantly among the 3 types in the present study. This does not imply that the dome impaction classification has nothing to do with the clinical outcomes, but comparable outcomes could be achieved if the dome impaction fractures were treated appropriately according to its variation and characteristics. It is notable that PM types tended to have greater operative time and amount of intraoperative bleeding, and smaller number of patients achieved anatomical reduction, although they were not statistically significant.

Several limitations of our study warrant mention. These include its small sample size, with a larger number of patients being more ideal for increased statistical power. A larger number might shed light on the present classification. The classification was based on our anecdotal experience. Some patterns of impaction which we might not have thought about may exist. The reference plane of the classification was based on a usual axial plane which was perpendicular to a CT table, and the true axial plane of the acetabulum should be corrected by the position of patients or the tilting of the pelvis. This might affect the classification to some extent. This classification includes the ambiguity of using radiographic imaging and some complexities for surgeons to classify. Further validation studies would be needed. Regardless of these limitations, however, we believe that the present study provides some novel perspectives on clinical practice, suggesting that classifying dome impaction fragments is useful for better understanding of their characteristics and for planning of operative procedures.

5. Conclusion

Over half (59.1%) of dome impaction fragments were located the SM area, followed by the PM area (27.2%). There were 15 cases (68.2%) with the gull sign and 7 cases (31.8%) without the gull sign on plain radiographs. Classifying dome impaction fragments is useful for better understanding of fracture patterns and planning of operative procedures.

Author contributions

Conceptualization: Masashi Miyazaki, Shozo Kanezaki.

Data curation: Tomonori Sakamoto, Akihiro Hino.

Investigation: Tetsutaro Abe, Noriaki Sako.

Supervision: Hiroshi Tsumura.

Writing – original draft: Shozo Kanezaki.