1 Introduction

The primary objective in managing developmental dysplasia of the hip (DDH) is to attain an early, concentric and stable reduction of the hip joint to facilitate further proper development of the hip joint with minimal complications. For children up to 6 months, the Pavlik harness is the initial treatment of choice.¹ Closed reduction (CR) with hip spica application is usually preferred for children from 6 to 18 months of age; open reduction (OR) is typically required if closed reduction fails or if the age is greater than 18 months.¹ Besides OR, some patients might need additional surgeries to address re-dislocations, residual dysplasia, and subluxation.^2,3 Avascular necrosis (AVN) is the most dreaded complication of this hip preservation procedure. Apart from age, the management approach is also based on preventing AVN by avoiding immobilization in extreme abduction and a low threshold for femoral shortening osteotomies if required.

The two commonly used approaches during OR are the medial and anterior approaches to the hip. Both approaches address structures that hinder concentric reduction, like iliopsoas, transverse acetabular ligament (TAL), ligamentum teres, neo-limbus, and pulvinar.¹ The anterior approach additionally offers access for capsular plication and pelvic osteotomy if required. The medial approach provides better access to the adductor musculature and minimizes blood loss.^2,3 The apprehensions regarding the medial approach have existed for a long time due to the potential danger to the medial circumflex femoral artery (MCFA) in this approach, injury to which can cause AVN of the femoral head.^1,2 However, multiple series reporting the outcomes of the medial approach alone has shown that whether it is through the Ludloff or the Ferguson interval, resulted in good clinical-radiological outcomes with rates of AVN similar to those reported for the anterior approach.^4–8

Both approaches have their pros and cons. However, ambiguity and controversy still exist about the rates of AVN and the need for FCS for both approaches.^9,10 There is a lack of literature and higher evidence regarding which approach is better while performing OR of DDH under 24 months of age.^11,12 In our knowledge, no meta-analysis has been published comparing the clinical and radiological outcomes; and complications of anterior and medial approaches in DDH OR surgery. This study aimed to compare AVN rates and the need for FCS between the two approaches while performing OR for DDH under 24 months of age.

2 Methods

2.1

2.1 Protocol and registration

This systematic review and meta-analysis was carried out as per the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines¹³ (Fig. 1). The protocol was registered on the PROSPERO database (CRD42022313928).

Close

Fig. 1 PRISMA flow for the study.

Export to PPT

3 Results

Two hundred fifty-four results were found on the initial search across four databases. One hundred fifty-six duplicate records were removed prior to screening, and 52 abstracts were screened. Twenty-four studies were excluded based on the abstracts, and a full-text review of 28 studies was carried out. Only five comparative studies were published in literature till now and all of them were included in the final synthesis after a discussion between two authors and a resolution of conflicting opinions, if any, by the third author. Meta-analysis was carried out for these five comparative studies (Fig. 1). Out of the five included studies, only one study by Tarassoli et al.² had a prospective study design, and the remaining four were retrospective (Table 2).^11,16–18

Overall the five studies encompassed 257 hips of 222 patients, out of which 151 hips underwent OR through a medial approach, and the remaining 106 hips underwent OR through an anterior approach^2,11,16–18 (Table 2). In the medial approach group, Bulut et al.,¹⁶Tarassoli et al.² used the Ludloff interval, and Yorganicigil et al.¹⁸ used the Ferguson interval. Ergin et al.¹⁷ and Hoellwarth et al.¹¹ used either Ludloff or Ferguson intervals based on surgeon preference.

In two studies, four hips (4 children) in the anterior approach group underwent additional femoral varus derotation osteotomy in the same setting.^11,17 None of the patients underwent any additional pelvic osteotomy during the OR. All the included studies had reported the number of hips that progressed to AVN or required further corrective surgery (FCS), which were the primary outcomes assessed. The mean age ranged from 5.9 to 21.7 months across the studies. Mean follow-up period it was varied from 3.2 to 11 years, with a minimum follow-up of at least two years. Three studies mentioned the initial Tonnis classification for the severity of DDH. Thirty-five hips were Tonnis grade II, 85 were Tonnis grade III, and 53 were Tonnis grade IV (2–4). Only one included study reported the mean blood loss and operative time.¹⁶ Instances of redislocation were mentioned in two studies^2,17 (Table 3).

3.1

3.1 Rate of avascular necrosis (AVN)

Forty-seven instances of clinically significant AVN were reported, with grades II, III, and IV of Kalamchi and MacEwan grading considered clinically significant. Nineteen cases of clinically significant AVN were reported after OR through the medial approach and 28 after the anterior approach (Table 4). AVN was graded only in two out of five studies,^2,17 as shown in Table 3.

Table 4 Clinically significant AVN (Kalamchi grade II, III, and IV) and Further corrective surgery (FCS) in included studies.

Author	Anterior Approach		Medial Approach
	Clinically Significant AVN (% of total hips)	Further Corrective Surgery (% of total hips)	Clinically Significant AVN (% of total hips)	Further Corrective Surgery (% of total hips)
Hoellwarth et al., 2015 11	7 (36.8%)	4 (21.05%)	3 (15.7%)	7 (36.8%)
Yorganicigil et al., 2016 18	5 (22.7%)	4 (18.18%)	5 (23.8%)	3 (14.2%)
Tarassoli et al., 2014 2	4 (18.18%)	0 (0)	3 (11.5%)	1 (3.8%)
Ergin et al., 2020 17	8 (25.8%)	5 (16.12%)	5 (12.8%)	4 (10.2%)
Bulut et al., 2013 16	4 (30.7%)	4 (30.7%)	3 (6.3%)	10 (21.2%)

Meta-analysis showed a statistically significant (p = 0.01) number of AVN cases with the anterior approach compared to the medial approach. The overall random effect showed the odds of having AVN with the anterior approach to be 2.27 (95% CI 1.18,4.38) times more than that of the same with the medial approach. All the included studies favoured the medial surgical approach except for that reported by Yorganicgil et al.,¹⁸ where the anterior approach is favoured marginally. The funnel plot shows the symmetric distribution of the included studies. Lower heterogeneity indicates the consistency in the data of the studies strengthening the overall meta-analysis (Fig. 2).

Close

Fig. 2 Forest Plot for rates of AVN in included studies.

Export to PPT

3.2

3.2 Further corrective surgeries (FCS) and redislocation rates

There were 42 instances of the need for FCS (Table 4). The various indications mentioned for FCS were redislocation, subluxation, and residual hip dysplasia. Twenty-five hips reduced with the medial approach underwent further surgery compared to 17 hips reduced with the anterior approach. In the study by Bulut et al.¹⁶ fourteen hips (ten hips in medial approach and four hips in anterior approach) required an additional pelvic osteotomy (twelve hips underwent Salter and two hips underwent Pemberton osteotomy). Ergin et al.¹⁷ reported that four hips in each group required FCS, two femoral and two Salter osteotomies in the anterior approach group, and one femoral and three Salter osteotomies in the medial approach group. Further, one patient in each group required a pelvic osteotomy in the study by Tarassoli et al.² Hoellwarth et al. and Yorgancigil et al. did not mention the type of FCS performed.^11,18

Meta-analysis depicted no significant difference in the instances of FCS (p = 0.63) between the two groups. Studies with greater sample sizes favoured the medial surgical approach; however, this difference was non-significant in the meta-analysis of all the included studies (Fig. 3). The exact number of redislocations was mentioned only in two studies,^2,17 with two in the medial approach group and one in the anterior approach group. The remaining studies clubbed these numbers with those requiring further corrective surgeries for other indications.

Close

Fig. 3 Forest Plot for rates of FCS in included studies.

Export to PPT

3.3

3.3 Clinical and radiological outcomes

Bulut et al. and Ergin et al. used Severin's radiological criteria and McKay's clinical criteria to classify clinical and radiological outcomes.^16,17 No statistical difference was found between the approaches in terms of clinic-radiological outcomes. 70.9% of hips that underwent medial OR and 68.1% of hips that were reduced by the anterior approach had a good clinical outcome classified as McKay grade 1. 60.4% and 52.2% of hips had Severin grade 1 radiological outcome in the medial and anterior approach groups, respectively. Yorganicigil et al. reported 16 and 17 hips had satisfactory radiological outcomes, and 17 and 18 hips had successful clinical outcomes out of 21 medial and 22 anterior approach hips, respectively.¹⁸

Four studies mentioned pre-operative and post-operative acetabular indices, and we computed the difference in means of acetabular indices^2,16–18 (Table 5). Meta-analyses comparing the improvement in acetabular indices yielded no statistically significant difference (p = 0.48) between the two groups (Fig. 4). We found no statistically significant difference (p = 0.29) in comparing the Centre Edge (CE) angle of Wiberg at the final follow-up, as mentioned in three studies.^2,16–18

Table 5 Radiological outcomes of included studies.

Author	Anterior Approach				Medial Approach
	Pre-operative Acetabular Index (Degrees)	Post-Operative Acetabular Index (Degrees)	Change In Acetabular Index (Degrees)	Centre Edge Angle (Degrees)	Pre-operative Acetabular Index (Degrees)	Post-Operative Acetabular Index (Degrees)	Change In Acetabular Index (Degrees)	Centre Edge Angle (Degrees)
Yorganicigil et al. 18	37.55 ± 3.99	21.23 ± 3.70	16.32 ± 4.83	17.78 ± 3.99	33.91 ± 5.38	21.86 ± 3.93	12.05 ± 5.27	17.19 ± 4.40
Tarassoli et al. 2	–	–	7.9 ± 1	–	–	–	8.8 ± 2	–
Ergin et al. 17	40 ± 5.63	20 ± 7.51	20 ± 9.37	33.37 ± 7.61	41.44 ± 5.52	21 ± 7.20	20.4 ± 9.07	30.30 ± 7.87
Bulut et al. 16	43.69 ± 3.79	23.08 ± 2.99	20.61 ± 4.82	24 ± 3.29	40.30 ± 3.46	22.96 ± 3.04	17.34 ± 4.60	22.53 ± 3.33

Close

Fig. 4 Forest Plot for change in acetabular index in included studies.

Export to PPT

3.4

3.4 Risk of bias

On quality assessment with the MINORS tool, one study² had a score of 23, three studies^11,16,18 had a score of 18, and one study¹⁷ had a score of 16 out of a maximum of 24 (Table 6). According to the MINORS tool assessment, all five included studies were of good or high quality.

Table 6 MINORS tool for risk of bias assessment of included studies.

	Aim Clearly Stated	Consecutive patients included	Data collected prospectively	Appropriate endpoints	Unbiased assessment of endpoint	Adequate follow up	<5% loss to follow up	Study size calculated prospectively	Adequate control group	Contemporary groups	Baseline equivalence of groups	Adequate statistical analysis	Total MINORS Score (Out of 24)	Quality assessment
Bulut 2013 16	2	1	0	2	1	2	2	0	2	1	1	2	16	Good quality
Ergin 2020 17	2	1	0	2	1	2	2	0	2	2	2	2	18	Good quality
Hoellwarth 2014 11	2	1	0	2	1	2	2	0	2	2	2	2	18	Good quality
Tarassoli 2014 2	2	2	2	2	1	2	2	2	2	2	2	2	23	High quality
Yorganicgil 2016 18	2	1	0	2	1	2	2	0	2	2	2	2	18	Good quality

The funnel plot asymmetry for AVN suggested publication bias (Fig. 5). However, this was ruled out by Egger's meta-regression analysis for publication bias, the intercept for AVN analysis was −2.76 (CI -15.6,10; p = 0.54), and for FCS analysis intercept was 1.25 (CI -2.2,4.7, p = 0.485)

Close

Fig. 5 Funnel plot for publication bias.

Export to PPT

4 Discussion

Medial and anterior approaches are commonly used to do OR in DDH cases, each with its own merits and demerits. The Ludloff interval in the medial approach passes between the pectineus and adductors longus and brevis (Fig. 6).¹⁹ Later, Ferguson modified this approach to pass between the adductor magnus and gracilis posteriorly and the adductor longus and brevis anteriorly (Fig. 6).²⁰ The medial circumflex femoral artery (MCFA) is encountered in both these approaches and is at risk of injury.²¹ This artery usually has to be retracted or ligated since it is confronted directly in the surgical field.²⁰ To address this medial approach with two intervals, one for capsulotomy (posterior to pectineus) and one for psoas tenotomy (anterior to pectineus) has been recently devised by Ozkut et al.²¹ They studied this approach in 45 hips and got satisfactory mid-term results using the modified medial approach using two intervals.²¹

Close

Fig. 6 Various modifications of the medial approach to hip (VL- Vastus Lateralis, VI- Vastus Intermedius, VM- Vastus Medialis, G- Gracilis, AL- Adductor Longus, OE- Obturator Externus, TFL- Tensor Fascia Lata, IPR- Inferior Pubic Ramus, IT- Ischial Tuberosity).

Export to PPT

The rates of AVN after open reduction vary widely across studies, ranging from 10 to 67%, with follow-ups ranging from 31 months to 233 months.^3,12 Similar rates of FCS, ranging from 0 to 57%, have been reported for both of these approaches.^2,11 Four out of 5 studies included in our review individually concluded that there were lower rates of AVN with the medial approach than with the anterior approach but without any statistically significant difference in their respective studies.^2,11,16,17

Our meta-analysis showed that the AVN rate was significantly higher with the anterior approach than the medial approach. The reason for this could be the trend of slightly older children undergoing open reduction by the anterior approach, though we included aged six months to 24 months only. Age has already been reported as a risk factor for AVN in children undergoing open reduction for DDH. Gardner et al. concluded in their systematic review that age <12 months at closed or open reduction increased the risk of AVN.¹² This was, however, contradicted in the meta-analysis by Novais et al., who had reported that the age or the surgical approach did not affect the rates of AVN.³ In our study, the mean age of patients for whom the anterior approach was used ranged from 6.1 to 21.77 months compared to 5.9–13 months in the medial approach group. However, on comparison of means, there was no statistically significant age difference between the two groups. The grade of acetabular dysplasia might also have affected the rate of AVN. Two of the included studies included more patients with Tonnis grade 4 in the anterior group than the medial group.^17,18 It was equal in one study,¹⁷ and two studies^2,11 didn't mention the grade. Though there wasn't a statistically significant difference between the groups, it could be one of the reasons as the anterior approach is preferred to the medial approach in more severe dysplastic hips, and two of the included studies^2,11 did not mention the grade of dysplasia. Another important factor, which was not reported adequately, was the number of patients with failed closed reduction attempts undergoing open reduction by each approach.

Also, our study showed no difference between the two groups in patients undergoing FCS, which is consistent with existing literature.³ A total of 25 out of 151 (16.5%) hips underwent FCS in the medial group compared to 17 out of 106 hips (16.03%) in the anterior group. FCS usually depends on the initial severity of the DDH, which was evenly distributed in both groups.

Various indications for FCS in the studies included redislocation, subluxation, and residual hip dysplasia. In the literature, the need for FCS has not been consistently reported, with a range of 0%–53% (most <30%) for the medial approach and 0%–57% for the anterior approach (most <30%).^3,12,22,23

Our analysis also showed no significant difference between the two approaches in clinical and radiological outcomes. Severin's and McKay's criteria were used to classify radiological and clinical outcomes. Both approaches showed improvement in the post-operative period. However, no significant difference was noted between the two approaches. Change in acetabular index and CE angle were also measured, and the analysis showed no statistically significant difference between the two groups regarding these parameters. This result is consistent with the literature.^3,9,10

We also reviewed nine studies describing the medial approach alone, including 348 patients with 440 hips, with a mean age of 11.53 ± 7.02 months.^{4–8,21,24–26} There were 53 (12.04%) instances of clinically significant AVN, and 142 hips (32.5%) underwent FCS. Four studies^5,8,24,26 reported using the Ferguson interval, two authors^6,7 reported using the Ludloff interval, and one study⁴ reported both. Ozkut et al. reported their results with the two-interval medial approach.²¹ These findings also indicate that the rate of AVN with the medial approach is not as high as thought to be previously, and it is safe when used for the right indications. The high rate of FCS indicates that acetabular dysplasia might not be fully addressed by this approach alone, which mainly relies on acetabular remodeling after reduction.

There are some limitations in our study. First, four out of five studies included in the meta-analysis are retrospective in design. The medial approach classically is being used in younger and less dysplastic hips compared to the anterior approach, which is usually used in older children with more dysplastic hips, which might have affected the results. However, there was no statistically significant difference in the age distribution between the approaches in our study (p = 0.292). Also, the size of the cohort is small in the included studies. Studies with larger sample sizes are needed to draw more conclusive results. The non-comparative studies (case series) were not included in the final synthesis due to excessive heterogeneity existing between the studies, which could have skewed the statistical findings. Surgeons’ experience and expertise in addressing all the structures that come in the way of achieving good reduction and maintenance might also affect the results. AVN is multifactorial; higher imaging like MRI may be needed in the post-operative period to look for the reduction quality. Also, rule out the positioning of the hip in extreme abduction, which is a potential risk factor for AVN.

However, the strength being ours is the first meta-analysis of comparative studies published on this topic. Cohorts in the included studies are closely matched in terms of age at operation, pre-operative acetabular index and CE angle, and longer follow-up ranging from 3.2 to 11 years.

5 Conclusion

The medial approach is safe and effective for performing open reduction of the hip in DDH under 24 months of age. Our analysis showed that AVN rates are lower with the medial approach than the anterior approach, with similar clinical and radiological outcomes and further corrective surgery (FCS) rates. However, the surgeon's expertise should be considered when choosing between these approaches.

Source of funding

Nil. None of the authors received financial support for this study.

Guardian/Patient's consent

Since this is a systematic review and meta analysis, consent was not required from patients.

Funding

No funding was obtained for this study from any source. No declarations.

Ethical justification

Since this is a systematic review and meta-analysis, ethical clearance was not applicable for this study.

Author statement

Karthick Rangasamy: Concept and design of work, Manuscript drafting and critical revision. Akash Ghosh: Manuscript drafting and Meta-analysis of data. Vivek P Ksheerasagar: Data analysis and manuscript drafting. Nirmal Raj Gopinathan: Data curation and supervision. Pebam Sudesh: Supervision.