The aim of this paper was to assess the influence of both patient-related and surgical factors on the short-term risk of revision in young patients, using data from the LROI. After adjusting for patient- and THA characteristics, our analysis shows a significant decreased risk of revision for any reason for the anterior approach. The use of the direct lateral approach and the anterolateral approach resulted in a reduced risk of revision, however these findings were not significant.
In literature, there is limited evidence about the effect of surgical approach on the short term risk of revision in young patients. Recent reports, assessing the short-term complication rate between the anterior and posterolateral approach, found no significant difference in postoperative complication rate and risk of revision [15, 16]. Some reports claim a favourable outcome of the anterior approach over the posterolateral approach when looking at recovery time and stability of the hip [17,18,19]. However, these studies did not focus on young patients and concluded that more evidence was needed.
We found an increased short-term risk of revision with increasing head diameters, which is in line with literature and registry reports [8, 20, 21]. In literature, the use of small head diameters is associated with an increased risk of dislocation [22, 23]. When we look at the rate of revision for dislocation in our population, we can conclude that the overall rate is low. Only 0.89% of all THAs were revised for dislocation 5 year after procedure, which is comparable with literature . Due to these low numbers, we were not able to determine the effect of patient- and surgical characteristics on the risk of revision due to dislocation as endpoint.
It might be possible that the lower risk of revision of the anterior approach disappears after a longer follow-up. Another study based on LROI data evaluated the effect of surgical approach in all ages. . In contrast to the current finding, they found the highest risk of revision for any reason for the anterior approach, even after exclusion of the first 150 procedures of each centre that performed the anterior approach to correct for a possible learning curve. Also other studies did not found an effect of surgical approach on a longer term [9, 26]. However, because of a higher activity level in young patients, and the high burden of young patients needing a revision procedure, it is of interest to analyze the short-term risk of complications.
Lastly, we found an increased risk for revision for MoM bearings. These findings are described extensively in literature and registry reports [20, 21, 27]. The use of other bearings did not result in significant differences when compared to C-PE.
In our study, we found low numbers of reported periprosthetic fractures, especially in the anterior group. A possible explanation might be a better bone quality and flexibility in young patients, which can result in less fractures after THA. However, a second explanation can be due to underreporting of periprosthetic fractures in the registry, where a reoperation with no replacements of any of the components of the implant, is not registered as a revision. A similar explanation would apply to infections, where treatment of infection without replacement of any of the components is not reported as a revision for infection in the registry . Therefore, the actual percentage of revision, due to periprosthetic fractures and infections, might be higher than reported in this paper.
Because of the increasing use of the anterior approach, the effect of a learning curve for this approach should be addressed. De Steiger et al. concluded that 50 or more procedures need to be performed before the rate of revision is no different from performing 100 or more procedures , where the most reduction in complication rate occurred after the first 100 THAs . Out of 100 institutes in the Netherlands, 27 performed at least 5 or more THAs using the anterior approach. Therefore, it can be concluded that the possible effect of a learning curve is present in our data. Despite this possible effect, the rate of revision was still lower for the anterior approach, when compared to all other approaches.
This study has some potential limitations that have to be considered. First, the follow-up of this study was limited. The effect of patient- and THA characteristics on risk of revision can change when follow-up is increased to a long term. However, especially for this young patient group, risk of revision on a short term is of major interest. Second, in our analysis we were unable to adjust for some variables that measure the patient’s demand on the implant, such as BMI and activity levels, as these were not, or only limited available from the Dutch Arthroplasty Register. Therefore, some residual confounding may be present. Lastly, we did not account for the possible effect of bilateral cases on the assumption of independence of observations in our statistical analysis. It has been shown that ignoring data dependency within a subject in studies involving bilateral cases may result in biased estimates [31, 32]. The extent of the resulting bias, however, was not determined in these studies.
Robertsson and Ranstam investigated if ignoring bilateral operations in statistical analyses biases the results, by analysing 55,298 prostheses in 44,590 patients using data from the Swedish Knee Arthroplasty Register . They found that the effect of neglecting bilateral prostheses is minute, possibly because bilateral prosthesis failure is a rare event, and concluded that the revision risk of implants can be analysed without consideration for subject dependency, at least in study populations with a relatively low proportion of subjects having experienced bilateral revisions.
The percentage of bilateral implants in our cohort was 10.3%, which is considerably lower than the proportion of bilateral implants in the study of Robertsson and Ranstam (19.4%) . Therefore, we think the possible effect of dependency between observations may be negligible in our analysis.