Bilateral Femoroacetabular Impingement: What is the fate of the asymptomatic hip? Clin Orthop Relat Res. 2019;477:983-989 (Azboy et al).

Research Question

‘Assessing the fate of the asymptomatic hip in patients with bilateral signs of FAI but only unilateral symptoms at initial presentation’.

The two main questions were well detailed with the aim and purpose clearly defined. Overall the research question was felt to address an important and largely unknown area in the management of Femoro-acetabular Impingement.

Methods

This is a retrospective study, as part of an 12-year institutional FAI database.

The use of the flow chart clearly identifies to the reader the steps taken for inclusion of patients in this study.

Inclusion/Exclusion criteria were determined to be largely appropriate. However, the review panel considered the inclusion of patients with Tonnis grade 2 hip degeneration and those Dysplasia of the hip (DDH) not to be appropriate for an FAI study.

The surgical technique undertaken was a mini-open femoroacetabular osteoplasty as opposed to arthroscopic (keyhole) techniques. 652 surgeries were performed over a 12 year period giving an average volume of 54 cases annually, considerably less than an average of 275 arthroscopic FAI cases annually at our clinic.

It was noted that 8 patients underwent bilateral FAI surgery despite having only symptoms in one hip initially (an exclusion criteria). This important point was discussed: If a patient has very clear impingement, with a large cam deformity and possible early arthritic changes starting in the hip joint even though a patient may not be symptomatic or not aware that they have a problem but the hip is at risk then it is appropriate to consider surgical intervention.

There was a lack of radiological information provided to the reader. The authors did not provide values of measured lateral centre edge angle or alpha angle on any of the views. With regards the alpha angle, many different views were used retrospectively to obtain the alpha angle and the largest measurement used for analysis. Additionally there is no confirmation of whether the x-rays were standardised.

Because these vital radiological parameters are lacking it is impossible for the reader to consider whether these patients were true FAI patients. The only reference to measured radiographic parameters was in the discussion (LCEA 28.1 ± 7.7 (11-54o) - this mean value would be considered well within the normal range (25-35o) and certainly a measure of 11o indicates severe dysplasia and not FAI.

Results

Although a reduction in the mHHS and UCLA measure was considered an indication of becoming symptomatic; neither the initial score values nor the change in outcome scores were provided for any patients.

The UCLA is an Activity scale. There are many reasons why a score might reduce over time unrelated to developing symptoms in the hip and the authors have not made the reasons for this change in activity level clear. The authors have also not provided the outcomes from the original surgery and then demonstrated how those outcomes have changed negatively in order for them to proceed with further surgery. The only outcomes they have provided is how the patients have done before and after their second operation. Of note the post-operative mHHS (74 ± 21) and UCLA score (6.8 ± 2.4) would be considerably lower than scores we would observe at our clinic following similar FAI surgery.

The time period provided by the authors describes a mean of 2 years, however the range is broad (0.3-11 years). It would have been useful if the authors also provided a standard deviation here to allow understanding of whether the majority are indeed centred around that 2 years period.

Of their entire cohort however there was only 19% of cases who remained asymptomatic – 81% became symptomatic within a mean of 2 years following initial surgery other side. The discrepancies between the group sizes may, as the authors do acknowledge in the discussion, resulted in type II error (false negative) – not allowing for the identification of some factors that may be associated with an increased risk of symptom development in a hip with radiographic signs of FAI.

After controlling for potential confounding variables, Table 1 identified a reduced neck shaft angle, increased lateral centre-edge angle, increased alpha angle and younger age to be associated with developing symptoms in the contralateral hip.

The increased odds of the occurrence of a symptomatic hip was 21.6% (p=0.009), 13% (p=0.049), 7.1% (p=0.025) and 6.8% (p=0.046) for each additional measured change within each of these parameters respectively

Table 2

Of note, in contrast to asymptomatic patients, a large proportion of the patients who were symptomatic and required further surgery either worked in a heavy job (31%), had psychiatric problems (23%), suffered previous trauma (10%) or had signs of early degeneration (‘moderate’ Tonnis Grade, 10%). Each of these factors individually have been shown in previous studies to increase the risk of disease progression, poor outcome and need for further surgery.

Discussion

Much of the focus of the discussion highlighted results from previous literature and not directly related to the results found in this paper. Ultimately, the authors do report that 81% of asymptomatic hips (with radiographic signs of FAI) become symptomatic with an average of 2 years and that 58% of asymptomatic hips require further surgery. 

This study does not fully address what the ‘fate of the asymptomatic hip is’. It reports the prevalence of contralateral hip becoming symptomatic after you have had initial FAI surgery. It was considered there was a disconnect with the title of the paper and the overall conclusion

Summary:

The overall concept of this paper is good and is certainly a question that needs answered but unfortunately this paper does not provide that answer. Limitations in the information provided to the reader in terms of diagnosis, radiographic parameters and patient-reported outcome measures (which form the basis of their decision) at specific time points hinder the ability for the reader to make any firm assumptions that these are actually FAI patients. It is unclear at what timepoint and to what degree patients became symptomatic and importantly given the small group sizes , there is a high probability of type II errors undermining the accuracy of the statistical analysis.