How long does a hip-preserving surgery with vascularized pedicle iliac bone grafts for femur head necrosis last? A systematic review and meta-analysis of case series with an average more than 5-year follow-up

Background: Hip-preserving surgeries with vascularized pedicle iliac bone grafts (VPIBG) are effective for osteonecrosis of the femoral head (ONFH). However, few studies exist about the long-term efficacy of this procedure. The aim of this meta analysis was to investigate how long does this hip-preserving surgery last. Methods: A comprehensive search was carried out through PubMed, Embase and Cochrane Collaboration Library for all relevant studies up to November 2019. The literature search strategy contained Medical Subject Headings and terms relating to ONFH and bone transplantation. All included studies were articles on VBIPG for ONFH, with an average follow-up of more than 5 years. Interesting outcomes included clinical success rates, complications, and conversion rates of THA. The data from eligible studies were then extracted and synthesized. The pooled effect size (ES) and 95% confidence intervals (CIs) were calculated. Results: Ten studies were finally selected. Eight studies including 3413 hips were pooled into the meta-analysis of success rates, the overall ES was 0.89 (95% CI, 0.86–0.92). In subgroup analysis, the ES was 0.88 (95% CI, 0.78–0.98) and 0.90 (95% CI, 0.87–0.92) at an average 5-10 years and 10-15 years follow-up, respectively. Pooled analysis of THA conversion rates derived from 7 studies (3389 hips) showed the overall ES of 0.10 (95% CI, 0.09–0.11). Seven studies (3396 hips) were included in a meta-analysis of complication rates, and the overall ES was 0.12 (95% CI, 0.08–0.18). The most common complications were secondary wound healing (37.6%), numbness or paresthesia of the lateral thigh (22.4%), and deep vein thrombosis (19.6%). Conclusions: The hip-preserving surgery with VPIBG is a safe and effective treatment for early-stage ONFH, but it should be used with caution in the treatment of advanced femoral head necrosis. The pooled data from this study suggested that 90% of the hips in patients

with ONFH lasted 10 years after this surgery. However, most of the included studies are case series, and these conclusions will need the support of high-quality research in the future.

Introduction
Osteonecrosis of the femoral head (ONFH) is a common catastrophic disease [1,2]. There are about 5 to 7.5 million patients in china with an annual increase of about 150,000 to 200,000 [3]. ONFH mainly affects young individuals and often involves bilateral hips [4,5]. If treatment is not prompt and effective, 80% of ONFH will gradually progress to femoral heads collapse within 1 to 4 years, eventually leading to advanced osteoarthritis and even disability [3,6]. Early diagnosis and aggressive treatment should to be taken to arrest progression of ONFH [2], because once the femoral head collapses ≥ 4 mm or secondary osteoarthritis occurs, most patients have to receive total hip arthroplasty (THA) [3,7]. However, THA is not the best choice for young patients due to the uncertainty of long-term survivorship of the prosthesis [5,8]. Therefore, it is necessary to seek effective treatments to preserve the femoral head and to delay or avoid THA.
The VPIBG was proven to have good short-term outcomes [8,19], but its medium to longterm results are still uncertain. The aim of this systematic review and meta-analysis was to answer the question of how long does a hip-preserving surgery with VPIBG for ONFH last.

Methods
This current systematic review and meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [20]. The research protocol had not been registered and ethical approval was not required, because

Inclusion And Exclusion Criteria
The following inclusion criteria were used. (1) All hip-preserving surgeries for ONFH were performed with VPIBG, and the average duration of follow-up must be greater or equal to Full manuscript was not available. (6) The language was not available for authors. (7) Reviews, case reports, commentaries and letters were also excluded.

Study Quality Assessment
The quality of the included case series studies was assessed using a quality appraisal checklist developed by the Institute of Health Economics (IHE), which evaluates bias from 20 criteria [21]. The quality of each study was assessed by answering "yes", "partial", "no", or "unclear". "Partial" responses were considered "yes", while "unclear" responses were considered "no". Finally, the number of "no" was calculated for estimating the risk of bias. As reported in a previous research [22], studies with 0-2 "no" responses were defined as low risk, 3-5 "no" responses were defined as moderate risk, 6-8 "no" responses were defined as high risk, and ≥ 9 "no" responses were defined as very high risk of bias.
For case-control studies, the Newcastle-Ottawa Scale (NOS) was used [23]. The studies were evaluated by two reviewers independently, and controversies were resolved through discussions or the final decision was made by the third author.

Data Extraction
Relevant information were extracted by two reviewers independently from all selected studies with a standardized data collection form, which included the following variables: author, year of publication, country, average age, number of ONFH and patients, study type, etiology, stage of ONFH, and kinds of pedicle bone flap/size etc. The interest outcomes of our study included follow-up periods, rates of clinical success, numbers of conversion to THA, definitions of clinical failure, complications, loss to follow-up, as well as rates of radiographic progression. For studies that lacked the results we needed, we had done our best to contact the authors for more information. If any discrepancy between the 2 reviewers (Xingyang Zhu and Xiaobo Sun) during this process, they could reach consensus through discussion or seek help from the third author.

Statistical Analysis
The extracted raw data were used to calculate clinical success rate, THA conversion rate and complication rate. All statistical analyses were done with the Stata software (14.0) and P value < 0.05 was considered to be statistical significance. Each series was weighted according to its standard error (calculated from published CIs). The I 2 statistics and Chi square test were performed to estimate the heterogeneity size across studies. If heterogeneity test expressed P > 0.1 and I 2 < 50%, data were pooled by a fixed-effects model, while the random effects was suitable for significant heterogeneity (P ≤ 0.1 or I 2 ≥ 50%). Sensitivity analysis was conducted to determine the stability of the outcomes if it was necessary, and subgroup analysis was conducted to explore more specific information when the data were available. Forest plots were applied to depict the results of each studies and to evaluate pooled estimates, while the funnel plots were used to assess publication bias.

Search Results
The initial search returned 1438 potentially relevant articles from the 3 databases. Four hundred and four duplicates were deleted, leaving 1034 articles for screening. Nine hundred and eighty three irrelevant citations were excluded after screening titles and abstracts by both reviewers, and 51 fulltext papers remained for review. Another 41 studies were rejected for several reasons, such as other surgical methods (n = 17), lack of raw date (n = 6), short follow-up (n = 9), duplicate research (n = 3), commentary, case reports or review (n = 5), inaccessible language (n = 1). Finally, 10 references were included in this study [16][17][18][24][25][26][27][28][29][30]. The details of study selection process can be found in the Fig. 1.

Nutrient Vessels Of The Bone Flap
The nutrient vessels of the iliac bone flap used by eligible studies contained DICA, ALCA and SICA.

Fixation Of The Bone Flap
Only 5 studies provided information about the fixation of bone grafts [16,25,[28][29][30]. The bone block was secured by an additional screw [25] or by taking the stitch through the hole prepared by drilling in the neck and graft [16], while additional fixation was not applied in the remaining studies [28][29][30].

Postoperative Management
As shown in Table 2, all studies elaborated on postoperative management. Patients were generally recommended to begin toe-touch weight-bearing in the first 6 weeks after surgery, maintain partial weight-bearing for the next 6 weeks, and start complete weight-bearing after 3 to 6 months.
The remaining 2 studies [24,25] were not suitable for merger because of different inclusion criteria [24] or definition of clinical failure [25], and therefore were given a descriptive analysis. One study reported a success rate of 24.20% [24], which might be associated with all the included subjects who were ONFH patients with segmental collapse, because the well defined indications for this operation were those with precollapsed or early collapsed stage [5,31]. The other study reported a success rate of 45.70%, which might be related to the study's definition of failure as "femoral head collapse or progress of ONFH" [25]. This was because a proportion of patients, who showed radiographic progression after hip-preserving surgery and whose HHS improved significantly, did not need to receive THA subsequently [32]. Therefore, these patients should not be considered clinical failures.
Given all the above reasons, data of these 2 studies were not merged in the meta analysis. Clinical outcomes of each studies are shown in the Table 3.

Viability Evaluation Of Implanted Flap
Viability evaluation of implanted flap was performed in 2 studies [16,24]. One study showed that 9 patients underwent digital subtraction arteriography at the end of 12 weeks and all cases proved the patency of DICA. The study also showed that the grafted areas of the femoral heads were showed high uptake through bone scan in 6 other patients [16]. Another study evaluated 17 hips through postoperative magnetic resonance imaging (MRI) scans and found isointense signals relative to the normal marrow and gadolinium enhancement in the viable graft [24].

Discussion
This systematic review and meta-analysis included 10 independent studies that analyzed 3042 patients (3481 hips) with ONFH treated by VPIBG, and the pooled data suggested an overall success rate of 89% for over 5 years of follow-up. This result was different from that of Yang et al [5], Who believed that the short-term effect of this protocol was better than the long-term effect. In addition, we found that the success rate of this operation was comparable between 5-10 years (88%) and 10-15 years (90%) of follow-up, which might be because the failure of this procedure was mainly concentrated in the early postoperative phase [5]. A previous study reported similar findings that the failure often occurred within the first 3 years after surgery [28].
Although THA can provide excellent clinical results, it is not the best choice for young patients with ONFH because of its high possibility of multiple revisions in the future [10,[33][34][35]. Thus, many hippreserving operative procedures to salvage the femoral head are in vogue for young individuals [32,[36][37][38][39][40]. VBIPG is one of these procedures and has been widely implemented, because it requires no special equipment and microsurgery techniques with a short operation time compared to free vascularized pedicled bone flap transplantation [3,16,37], and it is more conducive to rebuilding the blood flow of femoral head than non-vascularized pedicle flap transplantation [3,6,16,41].
Of note, ONFH in different stages has a completely different prognosis [42]. This operation is more suitable for patients in ARCO stage II or III A/B than patients in stage III C or stage IV [3]. Xie et. al [17] found that the 15-year survival rate of hips with stage IV after hip-preserving procedure was significantly decreased compared with Ficat stage II and III. Another study showed that only 8 hips (24%) were successfully preserved after the hip-preserving surgeries with VPIBG on 33 hips (in 32 patients) with segmental collapse; thus, the author did not advocate this operation for collapsed ONFH [24]. However, we could not analyze the influence of different stages on the prognosis by combining effect size, because different staging methods had been applied in different studies and not every original study provided the final treatment outcome for each stage.
The combined data showed that alcohol abuse was the leading etiological factor of ONFH (35.5%),  [17] found that the survival rate in the traumatic group was higher than that in the alcoholic, idiopathic, and glucocorticoid groups after a follow-up of 15 years, and the long-term survival rate was the lowest in the glucocorticoid group. Unfortunately, due to the lack of raw data on clinical outcomes after surgery of each causative factor, we were unable to analyze the prognosis of each etiological factor.
The lateral buttress of the femoral head, as the load-bearing region, plays an important role in hippreserving surgery, and patients with complete lateral buttress of the femoral head usually had a better prognosis than patients whose lateral buttress was affected by ONFH [29]. For instance, one study showed that 13 of 17 hips had no collapse in patients with type C-1 (JIC) necrosis, and that 15 of 18 hips had collapse in patients with type C-2 necrosis [25].
Age is another factor that affects the outcome of surgery. A study discovered that the mean survival time of the hips in patients 45 years or older was shorter than that in patients younger than 45 years [17].
Differences between interventions in various studies must be noted. First, the source of nutrient blood vessels for the bone graft was different. Second, the bone flap was fixed by an additional screw [25] or by the stitching [16], or by bone embedding technique without any additional fixation [28][29][30]. Finally, the position of the bone flap could also influence the outcome of the surgery. One study demonstrated that the closer the pedicle bone graft was inserted into the anterolateral normal subchondral bone of the femoral head, the better it could reduce the collapse of femoral head [25].
Several methods, such as angiography [16,28,29], digital subtraction angiography (DSA) [43,44], single-photon-emission computed tomography (CT) [24] and SPECT/CT [45], had been reported to evaluate the blood supply of vascular pedicle graft both before and after operation. Babhulkar S et al [16] advocated that preoperative angiography should be routinely performed on each patient to assess the presence of the transplanted blood vessel and to avoid vascular mutations. Liu Y et al [43] performed DSA on 48 patients before and 6 months after the operation to evaluate hemodynamic changes in ONFH with iliac bone flaps from the ALCA. They found that the position of the ALCA was relatively constant prior to operation, and also found that 44 of the 48 patients had a good blood supply to the femoral head 6 months after the operation, while the remaining 4 patients eventually underwent TKA due to poor blood flow of the femoral head. Another study used SPECT/CT to evaluate the femoral head bone viability after surgery with vascular bone graft, and a progressive increase of femoral head uptake were observed in all cases [45].
The operation had a relatively high incidence of complications, but most of them were minor and hardly affected the surgical effect. Secondary wound healing, irritation of the lateral femoral cutaneous nerve, and deep venous thromboses are the three most common complications.
To the best of our knowledge, this is the first meta-analysis evaluating the medium to long-term outcome of the hip-preserving surgery with VPIBG, in which only studies with an average follow-up of more than 5 years were included. We noticed that a previous excellent systematic reviews of this procedure by Yang F et al [5], but they did not combine effect sizes, so it is difficult to know the final success rate of this operation.
The present study did have some limitations. First, the included studies were mostly case series, rather than RCT or case-control studies, so the overall qualities of the studies were relatively low.
Second, as mentioned above, there exist slightly different interventions between various studies.
Third, the pooled data could not be stratified by factors that might affect survivorship, such as age, etiology factors, stage of ONFH, or location of necrotic zone, because these detailed data were not provided in each study. Fourth, although the mean follow-up of all included studies was greater than 5 years, three of them also included cases with a follow-up of less than 5 years [16,27,29]. Finally, the studies had been implemented for a long time in different countries and had used different classification systems, interventions and efficacy criteria.

Conclusions
The findings of the present study indicated that the hip-preserving surgery with VPIBG was a safe and effective method for ONFH in ARCO stage II or III A/B, while it should be cautious when this method was used for ONFH in stage III C or IV. Although not enough information is yet available to tell us exactly how long this operation lasts, this study suggests that 90% of them last 10 years.
Therefore, the medium to long-term effect of this operation is satisfactory. However, the included studies were mostly case series, high-quality studies are needed in the future.