Temporary hemiepiphysiodesis using an eight‐plate implant for coronal angular deformity around the knee in children aged less than 10 years: efficacy, complications, occurrence of rebound and risk factors

In the present study, 94% (95/101 knees) of the genu angular deformities were completely corrected, and the mean deformity correction period was approximately 13 months. Non-idiopathic genu angular deformity was found to be an independent risk factor for deformity correction failure and complications. The femoral correction velocity was significantly faster than the tibial correction velocity. The femoral correction velocity decreased with age, and the tibial correction velocity was slower in genu varum and non-idiopathic genu angular deformities.

TH is a widely used surgery for treating coronal genu angular deformity because of its predictable results, reversible process and minimal invasiveness [1, 6, 7]. Despite the existing risk of hardware-related complications and premature physeal arrest when using eight-Plates, their incidences were relatively low in the TH procedure compared to the use of staples [2, 3, 5]. In our study, the mean age of the patients was 4.69 years old and ranged from 1 to 10 years old. In all, 94.06% (95/101) of the genu angular deformities were corrected to standard lower limb alignment (centre of the mechanical axis zone: -1 to 1). Kumar S et al [9] and Danino B et al [15] found that the deformities were successfully corrected in 196 out of 215 patients (91.2%, mean age: 9.5 years old) and 342 out of 372 physes (92%, mean age: 12.5 years old), respectively. Our research suggested that younger children with genu angular deformities could have good results when TH surgery is performed using eight-Plate implants.

The overall femoral correction velocity (1.28° per month) was significantly higher than the tibial correction velocity (0.83° per month) (p < 0.001). This is consistent with the assumption in a previous study [16] that the growth rate of the distal femoral growth plate (9 mm per year) was higher than that of the proximal tibial growth plate (6 mm per month). Danino B et al [4] published a study on a series of 967 physes in 537 patients with various aetiologies undergoing TH for genu deformities (mean age: 11.35 years old) and measured a 0.77°/month change in the mLDFA and a 0.79°/month change in the mMPTA. However, no significant difference was found between the correction rates of the femur and tibia in their study, which was different from the results of our study. Ballal MS et al [17] described a total of 25 children with various aetiologies (37 legs and 51 segments, mean age: 11.6 years old) of coronal plane deformities around the knee and measured a mean correction rate of 0.7° per month in the distal femur and of 0.5° per month in the proximal tibia. In our research, the mean age was younger, and the overall correction velocity was higher than that in the mentioned studies. Ballal MS et al [17] also found that the overall correction rate decreased with age, which was consistent with our study. Kulkarni RM et al [11] collected data on a total of 24 children with various aetiologies (40 legs and 63 physes, mean age: 5.25 years old) who were under the age of 10 and who underwent surgery for coronal plane deformities. The mean overall rate of correction was 1.53°/month (below 5 years of age − 1.67°/month, above 5 years of age − 1.39°/month), which was similar to the rate in our study. Danino B et al.’s research [4] found that younger patients with more than three years of growth potential have greater success with TH. All of these studies suggested that younger children might have higher correction rates and be good candidates for TH, but they also need short-interval routine follow-up more than older children to prevent overcorrection.

We found that the aetiology of genu angular deformity was an independent risk factor, and the odds of surgical failure increased by 2.47 times with non-idiopathic genu angular deformity (Table 3). Kang S et al [18] did not find a significant difference in the rate of satisfactory hemiepiphyseal stapling correction between patients with multiple hereditary exostoses (MHE) and those with idiopathic deformities. However, because the MHE group had a slower correction velocity, the authors of the study also thought that TH should be considered at an earlier age for these patients. Jain MJ et al [19], in their research on Blount disease combined with a systematic review of the literature, found that severe deformity, the use of a titanium tension band plate (TBP) and obesity were more likely to lead to surgical failure. However, none of these factors was present in our research, which did not involve patients with Blount disease. Yilmaz G et al [20] collected data on 29 patients (50 lower extremities, mean age: 10 years old) with different types of skeletal dysplasia who were treated for genu angular deformities at two centres. They found that TH surgery using an eight-Plate is effective in very young patients, which is important in skeletal dysplasia. However, 7 patients (24%, 7/29) did not benefit from the TH procedure, which suggested that the failure rate was relatively high in non-idiopathic patients. The results were similar with those in the present study. Therefore, we should inform the patient’s parents about the risk of operation failure preoperatively, especially in the treatment of non-idiopathic genu angular deformities.

The initial SDA was not related to the correction velocity in the present research. This result was consistent with Eltayeby HH et al.’s [21] research. Our study did not find a significant relationship between the percentage of screw length accounting for femoral or tibial physis width and the correction velocity. Therefore, we recommended that epiphyseal protection during implantation might be more important than intentionally pursuing a parallel or divergent screw angle.

Complications, such as screw loosening, screw breakage and wound infection, have been reported in the literature to be 3.4-4% [9, 22]. In our research, 4.95% (5/101 knees) of screw loosening complications occurred in patients with non-idiopathic deformities. In a recent retrospective multicentre study by Danino B et al. [15], no hardware failure was observed in 372 physes in 206 patients with idiopathic angular deformities of the knee joint. In addition, we did not find a relationship between screw length and screw loosening or correction velocity. Therefore, we should also inform the patient’s parents about the possibility of implant failure, such as screw loosening, in the treatment of non-idiopathic genu angular deformities.

The prevalence of rebound after eight-Plate removal in the present study (2.97%, 3/101 knees) was similar to that in Kumar’s systematic review (2.3%, 8/350 knees) [9]. Park SS et al. [12] suggested a predictive model for rebound in children with idiopathic genu valgum corrected by TH using staples; the model indicated that a lower rate of correction and a higher body mass index (BMI) might have protective effects against rebound. However, in other studies using eight-Plates, obesity, younger age and rapid correction velocities were thought to be risk factors for rebound [1]. We did not find influencing factors by univariate or multivariate analysis (Table 4). The low incidence of rebound in our research made it difficult to investigate the risk factors for rebound. In addition, we perform a slight overcorrection to prevent the rebound phenomenon, as it has been proposed by some reports that slight overcorrection is essential and effective in preventing rebound, especially for younger and obese children [7, 10, 12]. Therefore, we should inform patients’ parents about the probability of rebound and choose an optimal time for hardware removal.

In the present study, linear regression analysis showed that for every one degree change in SDA, the mMPTA and mLDFA change by 0.35 degrees. This result was consistent with that from Sweeney KR’s research [23]. Because of the correlation of change between the SDA and the mMPTA and mLDFA, the changes in mMPTA and mLDFA observed by full-length lower extremity standing anteroposterior radiography can be indirectly inferred from the change in SDA measured by local radiography of the knee. Therefore, frequent full-length lower extremity standing anteroposterior radiography might be replaced by local radiography of the knee during follow-up to reduce radiation. Further investigation is needed.

Several limitations to this study need to be addressed. First, the intrinsic limitations of a retrospective study in a single centre cannot be avoided completely. BMI is not routinely recorded in our hospital. Second, there is heterogeneity in the aetiology of the deformities. However, the present research focused on the population younger than 10 years of age, and the various aetiologies of the deformities were also investigated as risk factors for efficacy, complications and rebound in TH using an eight-Plate implant. Third, the limited number of patients included may weaken the strength of the results from the multivariate analyses, especially for the low incidence of rebound deformity. However, the present study has the largest sample of patients younger than 10 years of age with TH using an eight-Plate implant for the management of genu deformities. Finally, the mean follow-up time after eight-Plate implantation was 12.71 months (range: 12–24 months), which is relatively short for obtaining more precise results about rebound deformities. However, our centre is the regional orthopaedic centre for children, and we believe that these patients would return for care if they developed rebound deformities beyond the follow-up period. Future studies should extend the follow-up time after plate removal.