The impact of capsular repair on the risk for dislocation after revision total hip arthroplasty – a retrospective cohort-study of 259 cases

The modified technique was based on Bauer’s lateral transgluteal approach [21], which is characterised by a lateral longitudinal skin incision and the detachment of the ventral part of the muscles, which attaches or originates ventrally to the greater trochanter. Connective tissues in the plane of the hip joint capsule are incised in an inverted T-shape manner, preserving its acetabular origin completely [13]. In most cases the capsule was additionally incised cranially longitudinally, forming a U-flap of its ventrocranial section. In a few cases the removal of the internal layer of the capsule was necessary, while the outer layer was preserved. After the implantation and reduction of the prosthesis, both longitudinal capsular incisions (ventral and cranial) were closed by sutures (Fig. 2), while the transverse incision near the femoral neck base remained unclosed and served for insertion of a Redon drain [13]. After reconstruction of the muscles, the fascia and the subcutaneous layer the skin was routinely closed by a resorbable intracutaneous suture, so that no removal of suture material was necessary.

The technique was based on the anterolateral approach (Watson-Jones-approach [22]).

The skin incision is slightly curved from distal-lateral to proximal-ventral. After the incision of the Tractus iliotibialis, the Musculus gluteus medius and minimus were incised near their attachment to the great trochanter and transverse to the direction of the muscle fibres. The connective tissues in the plane of the hip joint capsule were excised as completely as possible. Following the implantation and reduction of the prosthesis, a Redon drain was inserted and the muscles, the fascia and the subcutaneous layer were reconstructed. Finally, the skin was closed with clips that were removed after 10 days.

Statistical analysis was carried out using EXCEL 2007 (Microsoft Corporation, Redmond, WA, USA) and SPSS STATISTICS version 23 (2015, IBM - International Business Machines Corporation, Armonk, NY, USA). The SG and the CG were compared with regards to epidemiological, implant-associated and surgery-specific data. Statistical evaluations were performed using independent Student’s t-test, chi-square-test and Cox regression analysis. p-values of 0.05 or less were considered statistically significant, p-values of 0.01 or less were considered highly statistically significant.

Two hundred fifty-nine cases were included: 100 were identified retrospectively as belonging to the SG and 159 as belonging to the CG. The collected data and results comparing both groups are shown in Tables 1, 2, 3 and 4. One patient of CG (0.6%) died 11 months after surgery. A total of 45 patients (17%) passed away in the overall follow-up. The notes of the deceased were examined. In total 16 other patients were lost in the follow-up.

With respect to the epidemiological data there were no significant differences except for the co-existing morbidities classified according to the American Society of Anesthesiologists Physical Status Classification System (ASA) and the number of previous revisions (Table 1).

Both groups were similar and statistically non-different regarding the indication for surgery, which was limited to aseptic loosening, wear of liner and recurrent dislocation by the case selection criteria (see flow chart Fig. 1). The indications and surgical procedures were discussed and defined in our clinic-internal indication board. All treatments are based on the evaluated standard operating procedures (SOP) of our clinic.

Differences in surgery- and implant-related data between the two groups were partly significant and are listed in Table 2. In the SG, the fixation technique used for most prostheses was cemented (35%), followed by hybrid (20%) and uncemented (19%). The larger part of the CG prostheses was implanted uncemented (48.5%), followed by cemented (15%).

The average number of the exchanged femoral stems was 14% in the SG and 24% in CG, whilst acetabular cups were exchanged at a rate of 25% in the SG and 12% in the CG. In summary, there were no significant differences between the two groups regarding the components exchanged during R-THA.

Mean femoral-head-size and acetabular-cup-size were similar in both groups (Table 2). The mean head diameter of the dislocated cases was even 1 mm larger in the CG than in the SG (Table 2), which should actually result in a reduced dislocation risk in the CG. A higher dislocation tendency in the CG due to different head sizes is thereby excluded. The cutting-to-suture time was approximately 15 min longer in the CG (p = 0.042).

In the SG, the number of patients affected by dislocations was 3% (3/100) after 6 and 12 months compared to 19.5% (31/159) after 6 months and 21.5% (34/158) after 12 months in the CG, respectively (p < 0.001, Table 3). Because data evaluation for dislocation was restricted to the 12-months follow-up in this study, five late first-dislocation-cases (all from the CG, first dislocation occurred > 12 months after surgery) were considered “not dislocated” in order to keep the follow-up consistent for both groups. Thirty of totally 42 patients (71%) with one or more dislocations had their first dislocation incident within the first 3 months after R-THA surgery. After that period, the risk of dislocation was markedly reduced (see Fig. 3).

In the CG, 17 of 46 postoperative revisions (37%) were due to dislocation. Likewise, 17 of 39 dislocations in the CG (44%) caused surgical re-intervention within 6 months. Additionally, more than half of all patients (24/42; 57%) suffered multiple episodes of dislocations.

In terms of pain assessment, the average pain reduction postoperatively was lower in the CG with 4.41 in comparison to the SG with 5.48 (Table 4).

Moreover, all the WOMAC scores concerning pain, stiffness, physical function and global score, were better in the SG (Table 4).

The results of Cox regression analysis, adjusted for the overall follow-up (Table 1), are given in Table 5. Taking into account the influencing factors capsular repair vs. resection, BMI, ASA, head size and fixation technique different hazard ratios could be determined. Taking capsular resection into account as an isolated factor, we have determined a hazard ratio of 4.1 compared to capsular repair (Table 5). When rating the Cox regression analysis, the small number of included dislocation events (n = 32) must be taken into account.

The causes of THA dislocation are variable and multi factorial, including patient-related factors (age, sex, indication, co-morbidities, lack of compliance), surgery-related factors (implant misalignment, surgical approach, soft tissue handling, experience of the surgeon), and implant-related factors (head diameter, design of the liner, head-neck-ratio) [28‐30]. The most important factors listed here:

A combination of capsular repair and using a larger head size may restore physical joint stability on a larger scale, provided by the permanent hip-stabilizing effects of atmospheric pressure, which continuously acts on the cross-sectional area of the prosthetic head, if the joint is hermetically sealed by either the hip capsule or connective tissues in the plane of the hip joint capsule [13, 35, 37‐39].

The effects of capsular repair on dislocation rates after primary THA are reported in the literature. Some authors, analysing published reports on the effects of capsular repair on dislocation rates, found dislocation rates of 0% vs. 4% (n = 790) and 0.8% vs. 6.2% (n = 284) if capsular repair was performed instead of capsular resection [36]. Other studies reported about dislocation rates of 0.6% vs. 2.8% (n = 1000) [28] and of 0.7% vs. 4.8% (n = 1515) [40] if capsular repair was performed instead of capsular resection.

Other than these reports on capsular repair in primary THA, few studies can be found regarding capsular repair in combination with the posterolateral approach in R-THA. One of the aforementioned studies reported a dislocation rate of 2.5%, which was considerably lower than other reports in the literature [18]. One study including 47 R-THA cases with capsular reconstruction reported no dislocations [17]. Another group compared a capsular repair technique using 32-mm-heads (n = 110) with a capsular resection technique using 28-mm heads and found a reduced dislocation rate of 2.7% vs. 10.6% [7].

This study is the first to report on the decrease of dislocation rate following R-THA via Bauer’s lateral transgluteal approach. The resulting 12-months dislocation rate of 3% in the study group is relatively low and comparable to a reported rate of 2.7% using a similar capsular repair technique via posterolateral approach [7]. The 12-months dislocation rate of 21.4% in our CG appears to be relatively high compared to dislocation rates of 5 to 20% reported in the literature [25, 26]. Equally, the 6-months dislocation rates of 19.5% in the CG appeared to be higher than the 14.4%, reported in another large population-based study [8]. In our series in total 14% of all R-THA patients suffered from dislocation, which is comparable to the aforementioned studies [8, 25, 26]. Most important is the 86% reduction in dislocations after R-THA in our SG relative to the CG. The follow-up of at least 16 months and response rates of 70% underline the power of our results, since the vast majority of dislocation incidents occur in the first 6 weeks and almost all dislocations in the first 12 months after surgery [13]. Furthermore, all other complications also occur predominantly in the first 6 months after surgery [26]. These early complications accounted for more than 70% of all complications in our study.

Postoperative hospital stay time was almost halved in the SG compared to the CG. In the study period, patients were only discharged after complete removal of suture material and with non-secreting dry wounds. Due to the routinely used resorbable intracutaneous suture in the SG, there was no suture removal required. Thus, SG patients could be discharged earlier and had a shorter postoperative hospital stay time. Furthermore, there were more revisions with cup exchange (isolated cup or complete THA-exchange) in the CG. This is probably responsible for the slightly longer operation time in the CG compared to the SG with more exchanges of head and liner or stem. The difference in the ASA classification was significant and less favorable for the CG. The huge discrepancies in the dislocation rates, however, could not be plausibly explained by this nonspecific factor. The small difference in BMI between the two groups has no practical relevance (Table 1). According to the WOMAC scores, the clinical outcomes of pain, stiffness, physical function and global outcome were worse in the CG. Furthermore, patients in the CG were faced with an increased risk of surgical re-intervention, considering revision rates of 29% within 79 months in the CG and 10% within 49 months in the SG. Epidemiological- and implant-related factors of this study cannot explain the resulting differences in dislocations between the SG and CG. The results of this study underline that capsular preservation and repair is of crucial importance to avoid early dislocations and subsequent revisions after R-THA-procedures.

Some authors supported this theory by pointing out that the highest risk for dislocation occurs immediately after surgery before any pseudo-capsule has reformed in patients with previous capsular resection [38]. Beyond that, another study demonstrated that a repaired posterior capsule remained intact in 90% of patients at 3-months follow-up [19]. This is in line with the higher rate of early dislocations in the CG-series of this study (Fig. 3) as well as in the capsular resection group of our previous study, analysing primary THA cases [13]. Both results can likely be explained by the absence of a hip capsule or scar tissue within the first 3 months following the R-THA or THA procedure in both capsular resection groups before a scarred capsule is formed.

Preservation and repair of the capsule, supported by the use of larger prosthetic heads if possible, helps to re-establish more of the pre-existing protection against dislocation, which is physiologically provided by the permanent hip stabilising effect of atmospheric pressure [38]. This mechanism seems to be one of the main factors, maybe even the leading factor, in the avoidance of THA dislocation [13, 35, 37‐39]. This phenomenon has been described in detail by Prietzel et al. [35, 38]. A shorter postoperative hospitalization period demonstrates that the application of the modified, less-invasive, capsule preserving R-THA technique is accompanied by further medical and economic benefits.

It was impossible to measure the acetabular inclination and anteversion angles in this study. However, a significant influence on the outcome is unlikely because senior surgeons with greater professional experience (Table 2) performed the R-THA-procedures in the CG.

The low number of patients included, the fact, that only one surgeon aimed at capsular repair (surgeons coincidentally assigned to the operation), the intraoperative decision-making about feasibility of capsular repair (possible selection bias), the differences in the overall follow-up periods, in the ASA classification and the number of previous revisions in both groups as well as a response rate of 70% were limitations of this study. Interpretability of Cox regression analysis was limited by the small number of dislocation events, especially in the study group with capsular repair (n = 3), in addition to the number of influencing factors investigated in the given study. As a further technical limitation, capsular reconstruction is impossible in about 40% of the R-THA-cases.