2.1 Study Identification and search strategy

In February 2020, the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were used to identify relevant articles in the PubMed, Embase, and Cochrane databases. The following key terms used were “osteotomy”, “arthroplasty”, “replacement”, “hip”, “outcome assessment”, “treatment outcome”, and “patient reported outcome measures”.¹⁵

Two orthopaedic surgeons (S.A. and J.S.) performed the search and independently reviewed the titles and abstracts to determine relevant articles to proceed onto full-text review. Differences in opinion were resolved by a third, senior orthopaedic surgeon (B.G.D.) to confirm that the studies met the inclusion and exclusion criteria (Table 1). Studies were included only if all reviewers came to a consensus that they met the eligibility criteria.

2.2 Quality assessment

Three reviewers (S.A., J.S., and B.G.D.), using the Methodological Index for Non-Randomized Studies (MINORS) criteria, independently evaluated all selected articles for methodologic quality. This grading scheme has traditionally been used to grade clinical outcome studies.¹⁶ If unmentioned, the level of evidence of a study was determined using Spindler et al.‘s criteria.¹⁷ The grading scheme includes the following categories: a clearly stated aim; inclusion of consecutive patients; prospective collection data; end points appropriate to the aim of the study; unbiased assessment of the study end point; follow-up period appropriate for the aim of the study; loss to follow-up <5%; prospective calculation of study size; adequate control group; contemporary groups; baseline equivalence of groups; and adequate statistical analyses.

2.3 Data extraction

Microsoft Excel (Microsoft Office 2018, Microsoft, Redmond, WA) was used to organize the data from all included studies. Reviewers extracted the title, author, date of publication, number of hips, demographics, follow-up time, time interval between osteotomy and conversion to total hip arthroplasty, degree of dysplasia (Crowe classification), type of osteotomy, radiographic readings, intraoperative findings, complications, reoperations, and patient-reported outcome scores from each study.

2.4 Data collection

Study groups consisted of patients who were (1) indicated for a total hip arthroplasty after previous PO and (2) a control group (compound of nondysplastic and dysplastic patients), which was selected by including patients within the control cohorts of the studies who underwent a THA but did not undergo a previous PO. Patients reported outcomes (PROs) were summarized across the study sub-groups.

To estimate the effect size of the THA procedure, the standardized mean difference (SMD) was calculated for the Harris Hip Score (HHS) and the Merle d’Aubigne and Postel (MDP) rating system scores across the included studies using the method previously described by Griffin et al.¹⁸ This method requires average preoperative and postoperative scores in addition to the standard deviation of the mean preoperative score. In cases where the standard deviation was unavailable, it was estimated using the range.^19,20 An SMD score of 0.2–0.49 was considered weak, a score of 0.5–0.79 was moderate, and a score ≥0.8 was considered large.²¹

3.1 Study Identification

After screening 884 articles for eligibility (Fig. 1), 11 comparative studies were included within this systematic review.^13,22–30

Close

Fig. 1 Study identification.

Export to PPT

3.2 Quality assessment

All the studies reviewed were level III. Evaluation of the 11 non-randomized studies using the MINORS criteria, yielded a mean quality assessment score of 20.7 out of 24 (range: 18 to 23) for all comparative studies.

3.3 Study characteristics

In the reviewed articles, we identified seven different types of osteotomies: Chiari osteotomy (CO)^24,26,27,30, eccentric rotational acetabular osteotomy (ERAO),²² Bernese PO,²³ rotational acetabular osteotomy (RAO)^{22,24,25,28,29} Salter PO,³¹ shelf acetabuloplasty (SA),²⁴ and triple innominate osteotomy (TIO).¹³ All studies reported on at least one group that underwent PO and THA and compared findings to a control group who only underwent conversion to THA.

Among the reviewed studies, there were a total of 783 hips. Overall, 363 hips underwent PO followed by a conversion to THA. The remaining 420 hips consisted of the control groups who were primary THA in dysplastic patients. Age, body mass index, and follow-up time ranged from 13 to 72 years, 17.3–32 kg/m², and 1–18.5 years, respectively, for the PO prior to THA cohort. Interval between PAO and THA ranged from 1 to 23 years. The most common osteotomy type performed was RAO (127), followed by CO (96), Salter (43), Bernese (36), ERAO (36), SA (15), and TIO (13). Demographic data is contained in Table 2.

3.4 Clinical assessment

Five of the eleven studies reported clinical assessment pre-and-postoperatively. Fukui et al.¹⁶ reported zero patients with a positive Trendelenburg sign at latest follow-up. Osawa et al.²² reported no significant difference in pre-and-postoperative range of motion however noted that there were lower flexion, abduction, external rotation, and internal rotation in the control group compared to the group undergoing PO before THA. Ito et al.²⁵ found that the preoperative range of hip motion was poorer in the THA after RAO group. Tamaki et al.²⁴ assessed clinical pain, range of motion and gait using the Japanese Orthopaedic Association (JOA) hip score before and after surgery, recording a significant improvement in final follow-up pain, range of motion, and gait. Minoda et al.²⁷ used the Merle d’Aubigne and Postel rating system to evaluate pre-operative clinical pain, range of motion and walking ability between their study group (CO) and control group, citing no significant differences between the two groups preoperatively and postoperatively.

3.5 Radiographic measurements/findings

Eight of the eleven studies provided at least one radiographic measurement assessing their cohort's hip joint morphology.^{8,13,22,23,25–27,29,30} Seven studies reported on acetabular inclination before THA.^{13,22,23,25–27,29,30} Of these seven studies, one reported significantly decreased inclination in the osteotomy group when compared to control. The remaining studies report similar inclination values between osteotomy and control groups. Five of the eleven studies provided data on cup anteversion.^{22,23,26,27,29} Of the five studies, three recorded reduced cup anteversion in the osteotomy group compared to control.^22,23,29 Two studies showed similar cup anteversion between osteotomy and control groups. One study described cup placement within the safe zones, showing a significant difference in cup safe zone placement between osteotomy (36, 69%) and control (87, 84%) groups (P = 0.038).²² Three of the eleven studies recorded both the horizontal and vertical distance of the hip joint center.^22,25,29 One study demonstrated a significantly greater vertical distance to hip joint center in the osteotomy group than in the control whereas all three studies demonstrated a significantly greater horizontal distance to hip joint center in the osteotomy group than in the control. Radiographic measurements/findings are summarized in Table 3.

3.6 Intraoperative findings

Nine of the eleven studies assessed recorded mean operative time and blood loss in both the control and study groups.^13,22–29 Operative time ranged from 57.7 to 177 min in the study groups and 50.9–174 min in the control groups. Three studies reported a significantly increased operation time for their study group compared to their control, whereas one study reported significantly increased operation time in their control group compared to their study group. Mean blood loss ranged from 136 to 1598 mL for the study groups and 50.9–1192 mL for the control groups. Two studies reported a significantly increased exsanguination for their study group compared to their control, whereas one study reported significantly increased exsanguination in their control group compared to their study group.

Other intraoperative findings include the need of anterior rim trimming due to bony impingement after THA. Ito et al.²⁵ required anterior rim trimming in 40/44 study group cases (90.9%) whereas it was required in 31/58 control group cases (53.4%). Amanatullah et al.²³ also noted that anterior rim trimming was needed due to retroversion of the acetabulum. Moreover, Tamaki et al.²⁴ required bulk bone augmentation significantly more in the Chiari and RAO groups than in the control group (p = 0.02 and p < 0.001, respectively). The intraoperative findings are summarized in Table 4.

3.7 General PROs and clinical assessment

The most commonly used PROs were the HHS and MDP. Outcome measures for the osteotomy before THA and control groups are displayed in Table 5.

Comparing the preoperative scores for the control and PO before THA study groups, one study reported significantly higher PROs for the control group,³⁰ one study reported significantly higher PROs for the study group,¹³ four studies did not show any significant difference in PROs between their cohort and study groups^24,27,28 and five studies did not report a p-value comparing the control and study groups^22,25,26,29,.³⁰ For postoperative scores, two studies reported significantly higher PROs for the control group than the study group^13,22, one study reported significantly higher PROs for the study group than the control group,³⁰ four studies did not show any significant difference in PROs between their cohort and study groups^25,27,28 and four studies did not report a p-value comparing the control and study group's postoperative scores.^23,24,26,29

Four of the ten studies reported patient satisfaction after THA, which was high (85–90.7 out of 100 in three studies^13,26,27 median of 10 out of 10 in one study²³). No significant differences were found between both groups.

3.8 Standardized mean difference and heterogeneity

A summary of the SMD values for all studies reporting HHS can be found in Fig. 2. Of the eight articles with a recorded HHS, four provided sufficient information for effect sizes to be estimated based on HHS.^23,26,28,30 All articles demonstrated a large effect size of ≥0.8. Yuasa et al.‘s²⁸ control group had a substantially large effect size with SMD of 7.9, whereas their RAO group had SMD of 3.9, indicating a significantly larger effect of THA on the control group. All articles demonstrated a large effect size of ≥0.8.

Close

Fig. 2 Standard mean difference for Harris Hip Score and Modified Merle d'Aubigné and Postel Method. C, control; CO, Chiari osteotomy; CI, confidence interval; HHS, Harris Hip Score, MDP, Modified Merle d'Aubigné and Postel Method; RAO, rotational acetabular osteotomy.

Export to PPT

3.9 Complications and reoperations

Complications and reoperations are summarized in Table 6. The overall complication rates were 8.8% (31/353) and 4.5% (17/375) for the study groups and the control groups, respectively. The most common complication throughout the study groups was dislocations (10/353, 2.8%), followed by hardware breakages (5/353, 1.4%) and intraoperative femoral fractures (4/353, 1.1%). The rate of component loosening, and acetabular cup migration was low, however, when it presented as in Hashemi-Nejad et al.,²⁶ either abduction angle exceeded 50° or coverage of the cup was suboptimal.

The overall reoperation rate was 7.7% (25/326 hips) for the THA after a prior PAO study cohort hips and 5.5% (21/379 hips) for all control group hips. The most common indication for revision was dislocations and aseptic loosening.

4.1 Limitations

This study has some limitations. First, a quantitative analysis (i.e. metanalysis) between the studies was not possible due to the lack of randomized control studies. However, the quality of this systematic review was maintained due to selecting only case control studies. This selection allowed the comparison within studies, required for an adequate qualitative analysis (i.e. systematic review).