1 Introduction

Professor Emile Letournel’s results of 940 operatively managed displaced acetabular fractures remains the gold standard—his results may never be bettered. Even in his expert hands, only three quarters of patients achieved an excellent or good short-term outcome. Letournel and Judet [1] also demonstrated that the quality of the result diminishes with time. With the advent of improved techniques and better instruments and implants, results would be expected to improve. Perhaps the expectation for better outcomes has been counteracted by an increase in osteoporotic fractures.

An acetabular fracture is a devastating injury and can seriously influence a patient’s quality of life. In an attempt to analyze why 25% of patients have substandard results, we need to identify factors that point to a good result and, more importantly, those factors that jeopardize a satisfactory outcome. Overall, these factors are the following:

• 
  

  The patient—what were the preinjury activity level, age, bone quality, and the postinjury requirements and needs? For example, it is hard to imagine an elite track athlete returning to the field after a high-energy displaced transverse acetabular fracture.

  
  
• 
  

  The fracture type—what is the degree of displacement and amount of fragmentation? The answers reflect the degree of primary bone, cartilage, and soft-tissue damage.

  
  
• 
  

  The treatment—is nonoperative or operative treatment appropriate, and what factors lead to this decision? The timing of treatment and the expertise of the treating team also have a profound influence on the outcome.

  
  
• 
  

  Complications (both preoperatively and postoperatively).

  
  
• 
  

  The length of follow-up.

In the 1950s and 1960s, Judet et al [2] demonstrated that it was possible to obtain satisfactory results treating displaced acetabular fractures surgically. In the 1970s and 1980s, approaches were modified and refined, and instruments were developed in an attempt to help the surgeon. It is likely that the 21st century will herald improved patient selection. By studying the long-term results of major series, we can begin to identify those patients who need to be treated operatively and which fractures can be treated nonoperatively. A significant subgroup of patients also will be identified, in which other treatments need to be considered (eg, acute or delayed total hip replacement). Some acetabular fractures are so severe, especially those associated with irreversible articular cartilage loss, that primary surgical reconstruction is fruitless.

2 Natural history: why operate?

Drawing meaningful conclusions from the orthopedic literature in this area is challenging and occasionally confusing. However, one message has consistently emerged over the last 40 years. To obtain a good result following an acetabular fracture, the femoral head must be accurately centered and stable beneath the acetabular roof. In a 2-D x-ray, the relationship between the femoral head and roof of the acetabulum is often referred to as parallelism. When considered in 3-D, the concept of congruency is introduced. Whether parallelism or congruency is achieved by operative or nonoperative methods seems to be of secondary importance. If operative treatment is chosen, it is inevitable that a new set of soft-tissue complications may follow. The outcome then may be jeopardized, even though the femoral head is centered.

Some fractures (eg, associated both-column fracture) demonstrate congruency of the femoral head in the displaced position. Historically, these fractures have been treated nonoperatively with excellent results, although evidence is principally anecdotal [3]. During the last 20 years, the assessment of congruency has proved difficult to demonstrate and define. The three classic x-ray views go a considerable way to demonstrating it. Of these, the AP view is the most sensitive. If the medial joint space is equal to the superior joint space, it is reasonably safe to assume that there is acceptable congruency. However, x-rays are static images and with more unstable injuries medial subluxation of the femoral head may occur with increased joint loading.

In a posterior wall fracture, the postreduction x-ray may show a congruent relationship between the femoral head and the acetabular roof. However, when the hip is flexed (often with the patient under anesthesia), it can subluxate or even dislocate. Although in the static position there is congruency, recurrent subluxation will jeopardize an excellent long-term outcome. Therefore, stability and congruency need to be assessed not only statically but also with techniques that stress the joint. A number of formulas attempt to define stability from the computed tomographic (CT) scan but these do not take into account hip version and cannot be used as the final denominators of stability. If there is any doubt about stability, a thorough examination under general anesthesia is essential. This examination is not as straightforward as expected because the image intensifier can be awkward to position during hip flexion and adduction.

The advent of CT has led to a better understanding of fracture types; however, congruency is impossible to judge in isolation with these images alone.

Figure 2.23-1 introduces the concept of gaps and steps. The diagrams are 2-D; however, the situation with the acetabulum is more complicated because of the 3-D nature of the anatomy. Figure 2.23-1a demonstrates an acetabulum with a gap. In this situation, there is still equal loading of the acetabular articular cartilage. If the cartilage cells survive the impact, the situation is compatible with an acceptable outcome. Figure 2.23-1b shows an acetabulum with a step. This situation is unacceptable in that it will lead to eccentric loading of areas of acetabulum and differential loading on areas of the femoral head, causing abrasion of the articular cartilage on both sides of the joint, cell death, and an early substandard result. This concept also has been observed in other joints (eg, the ankle) [4, 5]. The goal of treatment must be to convert steps to gaps and then, if possible, close the gaps to produce a perfect reduction.

Another important factor affecting outcome is retention of loose bodies. Thompson and Epstein [6] showed that retained loose fragments following dislocation leads to early femoral head abrasion and a poor long-term result. They advocated open inspection and wash out of the joint. The presence of loose bodies following an acetabular injury is also an indication of significant associated primary articular cartilage damage. Because CT scans were not available when their article [6] was published, it was not known that in producing these loose bodies other areas of articular cartilage suffer displacement at the time of impaction (marginal impaction). This further contributes to the areas of damaged and displaced articular cartilage, which contributes to a substandard result if left malpositioned. Loose bodies in the joint can still be missed, especially if cuts on CT scans are > 3 mm apart. These smaller pieces of articular debris can be retained within the fovea and not cause long-term problems. However, any loose fragments seen on CT scans following dislocation usually need to be surgically removed. During surgery following fracture dislocations, the amount of debris within the joint space is always significant.

Not all hips are the same. Some are slightly dysplastic, and hip version has some influence on stability. A hip socket in extreme retroversion will dislocate more easily than one facing more anteriorly and may require surgery, even with the smallest rim fracture. Injuries also occur in already impinged hips, and the concept of performing a “bumpectomy” at the time of reconstruction must be considered. Later development of osteoarthritis may not be entirely due to the fracture. Genetic and morphological causes may be attributed to the progression of wear. Letournel and Judet [1] described collarette osteophytes, both progressive and nonprogressive. The authors believed that these osteophytes are of at least two separate etiologies. The first type is benign and nonprogressive, related to impaction on the femoral neck during injury (eg, transverse fracture), causing a subperiosteal hematoma that ossifies early and does not progress. The second type is a result of severe chondral damage, malunion, and/or both and is really just the radiographic beginning of posttraumatic osteoarthritis.

The hip joint does not tolerate incongruence, instability, or retained loose bodies. If a fracture is minimally displaced, congruent, and stable—excellent results can be achieved with nonoperative management. Loose bodies and instability are absolute indications for operative treatment, whereas incongruency is only a relatively strong indicator. This gray area represents the art of acetabular fracture management.

3 Historical results: methods of assessment and outcomes

A review of the historical results requires that the case mix in the selected studies is comparable and that the method of assessment and quality of the result is not too generous. The Harris Hip Score historically has been adapted from the arthroplasty literature to fracture outcome studies. This adaption is an overgenerous assessment; it is possible to obtain a high score with an arthrodesed hip because most methods of assessment are strongly biased toward the level of pain. The original 1954 D’Aubigné and Postel assessment allowed an excellent result, even if a hip had only 90° flexion. Meaningful results are obtained from a treatment series only when assessments are accurate and reproducible. Some conclusions drawn from early studies have been inaccurate, leading to significant misconceptions. In 1961, Rowe and Lowell [7] reported 93 acetabular fractures in 90 patients in a study that championed the cause for nonoperative management. The overall results of nonoperative treatment seem to compare favorably with those of operative treatment by Judet later in the 1960s. Closer examination of the subgroups within Rowe and Lowell’s study [7] leads to a significantly different conclusion. As expected, patients with undisplaced crack fractures do extremely well with nonoperative treatment. Outcomes of a subgroup of patients with posterior fracture dislocations are entirely different. Six of 17 patients with a posterior wall fracture treated by closed reduction and traction had a poor outcome. However, 8 of 11 patients with the seemingly comparable posterior wall injury who underwent open reduction and internal fixation had a satisfactory outcome. These results suggest that operative treatment for a posterior wall fracture is the best choice. Even so, 3 of 11 patients who underwent surgery had a poor result. Another conclusion from the study was that the injury itself significantly influenced the eventual outcome, as did the choice of treatment.

Within the same series, 29 patients with a quadrilateral plate and/or anterior column injury had excellent results when congruency was retained between the femoral head and acetabular roof by closed methods. A total of 90% of patients achieved what was described as “a satisfactory clinical result under these circumstances.” However, the authors also pointed out that only 83% of the same group had what could be termed a satisfactory radiographic outcome. Rowe and Lowell [7] were among the first to note the differences between clinical and radiographic outcomes. Their conclusions should not be that nonoperative treatment is indicated for all acetabular fractures. Rather, undisplaced crack fractures probably do not require operative treatment. Likewise, if the femoral head remains in stable condition under the acetabular roof, then surgery may not be indicated. However, if the hip is unstable or incongruent, then other treatment methods should be considered. The only way to predictably restore stability and congruency is through operative management.

In 1986, Matta et al [8] retrospectively reviewed 204 acetabular fractures, with follow-up of only 3.7 years. They introduced the concept of how much of the acetabular dome or roof was involved in the fracture. They used these “roof arc angles” in an attempt to assess stability to choose between operative and nonoperative treatment. However, these angles can be applied to certain fracture types only; for example, they cannot be used in both-column fractures and isolated wall fractures. The authors emphasized the important relationship between the femoral head and acetabular roof, concluding that closed treatment is only appropriate for displaced fractures if the weight-bearing dome is intact. Anatomical open reduction in the absence of complications improved the prognosis. Matta et al [8] also were among the first to attempt to correlate clinical result with the quality of fracture reduction.

In 1988 Matta and Merritt [9] reported another 121 fractures in which 23 were treated with traction and 98 were treated with open reduction and internal fixation, most through a posterior approach. Radiographic analysis of the quality of reduction deemed that postoperative displacement of 3 mm or more on plain x-rays was unsatisfactory, < 3 mm was considered satisfactory, and < 1 mm was considered anatomical. Use of this these generous criteria resulted in 91% of their patients being rated radiographically satisfactory but only 63% anatomical. The consequence of this study was that others incorrectly concluded that it was acceptable to operatively reduce an acetabular fracture to within 3 mm. Clearly, this was a mistake. The articular cartilage of the acetabulum is 2.5 mm thick at most. If a 3 mm step is accepted in a reconstruction, then under no circumstances will the long-term outcome be satisfactory.

The problem always has been in assessing the amount of postoperative displacement on plain x-rays. It is for this reason that postoperative CT scans and conventional x-rays should be evaluated together to assess the quality of reduction.

Other important information that resulted from Matta and Merritt’s study is [9] related to surgeon experience and the quality of reduction (ie, the learning curve). Figure 2.23-2 demonstrates a number of unsatisfactory reductions per group of 20 in the first 100 surgically treated cases. This graph demonstrates that as experience increases, the quality of the reduction improves. This point is further emphasized in Fig 2.23-3 , which demonstrates an increasing number of anatomical reductions with increased experience.

In 1996, Matta [10] reported results in 259 patients, mirroring those of Letournel and Judet’s [1] larger series. Matta included a modified clinical grading system that had much stricter parameters for radiographic assessment. Anatomical reduction was graded between 0 and 1 mm of displacement but imperfect was graded between 2 and 3 mm. A poor reduction was any displacement more than 3 mm. The fact that Matta’s results were almost identical to those of Letournel and Judet, who used a slightly less rigid scoring system, is intriguing. If similar surgical capabilities and a similar case mix are assumed, then Matta’s results are perhaps 5% better than those of Letournel and Judet. This slight improvement probably reflects 30–40 years of worldwide experience with improved techniques, reduced complication rates, and better patient selection.

The literature indicates that acceptable results after acetabular fracture are achieved if the following criteria are met:

• 
  

  The femoral head is centered under the acetabular roof, is congruent and parallel, and remains in place.

  
  
• 
  

  The hip is stable, and there are no entrapped loose bodies.

  
  
• 
  

  Complications are avoided.

Whether these criteria are met by open or closed treatment methods depends more on the type of fracture and the degree of displacement.