Higher risk of revision in total knee arthroplasty after high tibial osteotomy: a systematic review and updated meta-analysis

High tibial osteotomy (HTO) is a well-established procedure for uni-compartmental osteoarthritis of the knee and is commonly applied to young patients with high activity demands. There has been an increasing interest in high tibial osteotomy in some countries over the past decade [1]. The goal of this operation lies in correction of the mechanical axis of the lower limb, reducing the load stress of the pathological medial compartment. HTO has been reported to achieve satisfactory short-term clinical result, many of them regain satisfactory functional outcome. However, most patients who undergo this operation are relatively young, with high activity demand. Several studies have reported outcome of HTO. The reported 10-year survival ranged from 79 to 97.6%. The reported 15-year survival ranged from 56 to 65.5% [2, 3]. A meta-analysis reported 84.4% survival with 9 to 12 years follow up. It is reasonable to deduce that subsequent TKA is required in these patients [4]. These conversions to TKA are more technically demanding than primary TKA and may lead to inferior survival and functional outcomes comparing to primary TKA. Researches have been conducted to compare risk of revision and functional outcome between primary TKA and TKA after HTO, which reported controversial results. Some studies found HTO-TKA is at higher risk of revision and complication [5‐7], while other studies reported similar outcome between 2 groups [8‐11].

A previous meta-analysis published in 2013 reported similar outcomes between TKA following HTO and primary TKA in terms of survival and perioperative complications [12]. Most studies included in the previous meta-analysis investigated cohorts of small sample size. Several studies with larger cohorts have been published since then, some of which reported inferior survival and clinical outcomes in cases underwent TKA after HTO [5, 8, 9, 13]. Therefore, an updated meta-analysis and systematic review was conducted to compare risk of revision and other clinical parameters between TKA after HTO and primary TKA.

High tibial osteotomy was introduced in 1969 by Jackson and Waugh and has become a well-established procedure for unicompartmental knee osteoarthritis since then. The biomechanical rationale for this procedure is correction of malalignment and redistribution of stress on the joint [22].

The classic indication for high tibial osteotomy includes unicompartmental osteoarthritis of the knee and is especially recommended for young patients with high activity demands [23, 24]. For properly selected patients, studies have proven that it offers satisfactory pain relief and functional outcome. However, clinical improvement wears out over time and the majority of patients who underwent this procedure were relatively young. Previous researches have reported subsequent TKAs were required in up to 30% of these cases [25]. Concerns were raised that whether these HTO-TKA would provide comparable survival comparing to primary TKA. There have been conflicting reports regarding this issue. A previous meta-analysis, consisting of 11 studies with 421 HTO-TKAs and 1749 primary TKAs found no significant differences in terms of revision, complications and functional outcome [12]. In our analysis, 15 studies with 4646 HTO-TKAs and 140,074 primary TKAs were included, the substantial increase of sample size may help to investigate low-incidence event such as revision.

Pooled-analysis showed significantly more revisions and complications in the HTO-TKA group, which may be due to following factors: 1.In our analysis, aseptic loosening was the leading cause for revision in HTO-TKA group. Robertsson [26] et al. reported more stemmed implants were required during the conversion from HTO to TKA. Stemmed implant is recommended in these cases because its ability to offer extra rotational stability and avoid stress shielding. Two studies [11, 15] included in our analysis reported impingement between tibial stem and lateral tibial cortex, although it was stated that it appeared to not contribute to early failure. 2.Intraoperative exposure in cases with previous HTO can be more difficult than those of primary TKA. Nizard et al. [27] reported scar tissue between the patellar tendon and the proximal anterior tibia, which made the eversion of patella difficult. Measures including lateral release, rectus snip were applied in included studies. Still, unattainable patella was reported [16, 19] and inadequate exposure may lead to inaccuracy in many aspects during surgery. 3.Malalignment is another common complications encountered in HTO-TKAs, especially in overcorrection after varus tibial osteotomy according to Meding et al. [10]. The joint line on the tibial side become valgus and the bone deficiency on the tibial side can be confusing. The use of traditional method to determine femoral component rotation is often misleading in these cases and internal rotation of the femoral component is suggested.4.Kazakos et al. [15] found more Patella baja in the HTO-TKA group, which could lead to anterior knee pain and eventually revision. This might be due to patella tendon contracture and that the distance between joint line to tibial tuberosity decreased after osteotomy. In our series, based on available data, 10 out of the 304 revised TKA after HTO were patella-related. Patella arthroplasty might be a way to prevent future anterior knee pain and patella-related revision. 5.Persistent pain is more prominent in the HTO-TKA group. Patella baja, excessive soft-tissue release and malalignment may be contributing factors [20]. Impingement between the tibial stem and the tibial cortex might cause pain as well. 6.HTO-TKA group was also subjected to increased surgical time [11], hence to increased risk of infection. In our analysis, the infection rate in HTO-TKA group is 1.4% comparing to 1.0% in primary TKA group, pooled analysis showed significantly higher risk of infection in the HTO-TKA group. In general, the prior osteotomy complicated the anatomical structure of the knee, resulting in varying degrees of deformity, remaining hardware, bone loss and soft-tissue imbalance, which require extra caution and different techniques comparing to primary TKA.

Different osteotomy techniques might influence the risk for subsequent TKA; our analysis showed that prior lateral closing wedge osteotomy also led to significantly more revisions. Comparing to opening wedge HTO, it has been reported in the conversion to TKA from closing wedge HTO, the mechanical axis might be laterally displaced and the tibial insert is more likely to impinge on endosteal cortex, hence a tibial insert with smaller stem is recommended especially in closing wedge HTO [28]. However with the development of computer assisted tibial osteotomy, some traditional difficulties, such as achieving accurate alignment and preventing unintended changes in tibial slope encountered during osteotomy can be solved with computer assisted planning and navigation [29]. Patients underwent computer assisted tibial osteotomy may not present the same surgical challenges as traditional osteotomy and more researches are needed in this field.

Amendola et al. [20] argued that the time in which the subsequent TKA was performed was also crucial and that surgeons tend to have a better understanding of the technical difficulties to achieve comparable results over time. In our series, four studies were published within the last 5 years; three [5, 9] of them suggested similar survival between 2 groups, whereas one study [8] reported survival in favour of primary TKA. Pooled analysis from these 4 studies still showed significantly more revision in the HTO-TKA group.

Several methods, including lateral release of the lateral alar ligament of the patella; quadriceps snips; even osteotomy of the anterior tibial tuberosity were suggested in order to tackle the technical difficulties encountered during the conversion to TKA due to the presence of scarring tissue which poses substantial challenge during exposure [30‐33]. In our series, rectus snips remain the most common technique to achieve satisfactory exposure. Amendola et al. [20]. verified that in cases with prior osteotomy, the medial plateau is higher than the lateral plateau in anteroposterior (AP) radiographs. Erak et al. [20]. reported preoperative patella baja relating to difficulty with patella eversion. In order to balance the knee, release of lateral ligament was also suggested due to the extra medial dissection to remove osteotomy hardware [20]. The bone resected in the lateral region must be minimal to avoid a large defect. Wedges were used in 4 out of 711 cases as reported by Pearse et al. [7]. A high percentage of stemmed implants were used in the HTO-TKA group; this may result from the need to avoid a potential stress riser. Ligament balancing is crucial in cases with prior osteotomy; the fibrosis and loss of soft tissue may lead to instability [33]. The most common balancing technique in our analysis was lateral release (31/125). Difficulty during exposure, resection and component positioning contributed to the prolonged surgical time and led to increased blood loss.

Few studies reported on radiographic outcome between the two groups. Based on available data, more loosening of the tibial component and impingement were noted in the HTO-TKA group, which correlates with our analysis of revision. No significant differences were found in terms of alignment. Kazakos et al. [15] reported 16 cases of patella baja in the HTO-TKA group, with only two in the control group. Other studies also stated patella baja to be more common after HTO, but they did not find any relevance between patella baja and the clinical outcome of subsequent TKA [32, 34].

Our study has several limitations. (1) All included studies were retrospective studies and registry based studies whereas no RCT was included, which limited the quality of this meta-analysis. (2) The mean follow-up was not consistent among studies; revisions might not be required until 10–20 years later. (3) Some studies only reported mean and range for parameters such as surgical time, blood loss and functional score, and we were unable to conduct pooled analysis based on these data. (5) Some causes of revision were marked as unknown in some studies, which influenced our analysis of the causes of revision.

The strengths of this meta-analysis include the following. (1) This study focused on the risk of revisions and further investigate surgical complexity and radiographic outcomes with more cases involved and explored potential causes for the differences between groups. Most studies reported comparable survival outcome between the two groups, whereas pooled analysis of gathered data revealed statistically significant results. (2) Due to the lack of relevant studies and cohort, this meta-analysis gathers valuable information to conduct a quantitative analysis and yielded result inconsistent with the first meta-analysis on this subject. (3) Randomised trials were not feasible considering our research purpose; this study gathers data from retrospective studies and provides the best evidence available. (4) Relevant articles were screened carefully by two independent researchers, using a wide range of search terms. (5) Previous meta-analysis on this subject included 421 HTO-TKAs and 1749 primary TKAs, while our study included 4646 HTO-TKAs and 140,074 primary TKAs, the increased sample size allow us to assess small-incidence event such as revision with more accuracy. (6) Considering the wide variation of publishing time of included studies, pooled analysis of studies published within the last five 5 years yielded consistent finding. (7) Clear inclusion and exclusion criteria were utilized.

High tibial osteotomy offers satisfactory pain relief and functional outcome in selected patients with high activity demand. However, the need for subsequent TKA should be noted, which might be a technically challenging procedure with significantly higher risk of revision comparing to primary TKA.