Background

Most patients with knee osteoarthritis(OA) have lesions limited to the medial compartment, and high tibial osteotomy (HTO) and unicompartmental knee arthroplasty (UKA) are well-established treatments for this condition [1,2,3]. HTO and UKA, which are often performed as primary surgeries in medial knee OA, report good results as the ultimate treatment for OA with an added advantage of delaying conversion to total knee arthroplasty (TKA) [4, 5]. However, to date, considering which of the two surgical methods to choose remains controversial, comparing long-term survival rates and surgical outcomes [6,7,8] In particular, when selecting a surgical method for patients in their 50 s and 60 s, there are many considerations. [3, 9,10,11]. Both HTO and UKA have the advantage of saving native joint spaces, but there are some differences in surgical techniques and indications. HTO is considered primarily in young and active patients, and has the advantage of preserving the native knee compartment by correcting malalignment of the lower extremity and shifting the axis of the weight load to the lateral compartment [6, 12]. Good results have been reported for UKA in relatively elderly patients with less physical requirements, by replacing only the medial compartment in which the arthritic change was progressing [13]. Due to factors such as improvement of surgical methods and implants, surgical indications between the two operations have overlapped [6, 14]. In addition to age and physical demand, the relative merits of the two surgical methods are unclear and are still controversial which lead a lack of clear criteria for selecting a surgical method [6]. Numerous studies have compared the advantages and clinical results between the surgical methods, and the superiority of one over the other has not been proven [1,2,3,4,5,6,7,8, 12,13,14]. Therefore, the objectives of this study are as follows: (1) to evaluate revision rates of the two surgical methods and (2) analyze the outcomes in terms of perioperative complications. It was hypothesized that there is no difference between HTO and UKA with regard to survival and complication rates because both treatments are well established treatment methods.

Methods

Data sources

This study was conducted using data from the Korea National Health Insurance Review and Assessment Service. Korea's health insurance system is considered well-established worldwide, with almost 99% of the population enrolled [15]. The Korean health insurance claim data uses the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes and Electronic Data Interchange (EDI) codes, which are internationally accepted classifications for diagnosis and procedures. Information such as patient's age, gender, diagnosis, hospitalization record, surgery record, drug prescription, and hospital area information is provided in an anonymous form, and the individual data linked to each code are recorded in the NHI database in Korea. Data were obtained from January 2008 to May 31 2019. All medical records have been provided for patients diagnosed with osteoarthritis of the knee joint and having undergone surgical treatments.

Study approval

This study was approved by the Institutional Review Board of Chonnam National University Hwasun Hospital. The review board waived the requirement for patient consent considering the characteristics of this study.

Data collection

The diagnostic code for knee OA (M17) and procedure codes for UKA (N2712, N2717) or HTO (N0304) were used to extract patients who had undergone UKA or HTO for the treatment of knee OA. Patients under 50 years of age and over 70 years of age were excluded for analysis of middle-aged participants. An effort was made to extract newly diagnosed and operated patients within a defined study period by applying a 1-year wash-out period for knee arthritis and surgical treatment history (Fig. 1). We compared UKA and HTO in terms of conversion to primary TKA, incidence of unanticipated major medical problems, Intensive Care Unit (ICU) admission, postoperative transfusion, and readmission rates. Patient demographic data were extracted using the aforementioned diagnostic code and each surgical code and included age, sex, type of insurance, hospital, region of residence, and medical comorbidities (Table 1). By using the Korean claim data, it was not possible to distinguish between medial open wedge osteotomy and lateral closed wedge osteotomy of HTO and the type of implant. The type of hospital was classified as a teaching hospital(> 500 beds), a general hospital(30–500 beds), an independent hospital(< 30beds), and a private clinic(outpatient clinic). The region of residence was classified, based on population, as cities with a population of 10 million or more(e.g., Seoul), cities with a population of 1 million or more(eg, Busan, Incheon, Daegu, Gwangju, Daejeon and Ulsan), and cities with a population of less than 1 million(Gyeonggi, Gangwon, Gyeongsangbuk, Gyeongsangnam, Chungcheongbuk, Chungcheongnam, Jeollabuk, Jeollanam and Jeju). Medical comorbidities (hypertension, diabetes mellitus, hyperlipidemia, osteoporosis, peripheral vascular disease, depression and dementia) were confirmed based on the ICD-10 diagnostic codes with at least two claims within 1 year from the date of surgery. The Charlson comorbidity index (CCI) score was calculated from the ICD-10 codes by introducing previous literature methods. Statistical analysis was performed to consider as many adjusted variables as possible to address the imbalance of basic patient characteristics between the two groups.

Fig. 1
figure 1

Flowchart for target population

Full size image
Table 1 Patient baseline characteristics in 50–69 aged patients
Full size table

Statistical analysis

The baseline characteristics of patients treated with HTO or UKA were summarized using descriptive statistics, including proportion, mean, and standard deviation. Differences between UKA and HTO in continuous variables were evaluated with analysis of variance and categorical variables were compared using the Chi-square test. Wilcoxon statistic assigned a greater weight to differences occurring near the beginning of the study. The imbalance in baseline characteristics between the two groups was evaluated with standardized mean difference (SMD). An SMD of less than 0.1 was considered a negligible difference between the groups. Covariants with an SMD greater than 0.1 were corrected by setting them as adjusting variables. The person-years (PY) for each group of patients were calculated from the date of primary surgery to the event of subsequent revision and various adverse outcomes. We conducted a stratified log-rank test and obtained Kaplan–Meier curves considering potential confounders. In addition, conditional logistic and stratified Cox regression analyses were conducted to calculate complication rates. The adjusted hazard ratio (HR) and 95% confidence interval (CI) were calculated using a multivariable Cox proportional hazard regression model adjusting for potential confounders such as age, sex, insurance type, region of residence, hospital type, comorbidities, and the CCI. All statistical analyses were performed using R software (version 3.4.1; R Foundation for Statistical Computing) and SAS Enterprise software (version 6.1; SAS Institute).

Results

According to the data extracted through Korea's Health Insurance Review and Assessment Service from January 2007 to May 31, 2019, a total of 78,448 patients aged 50 to 70 who underwent UKA or HTO were recorded. Considering the one-year wash-out period, 6,910 patients with surgical records before 2008 and after 2018 were excluded, and 1,074 patients with inappropriate claims data were eliminated, therefore, 70,464 patients were included as the final study subjects, of which 21,194 were in the UKA group and 49,270 in the HTO group (Fig. 1). The mean age of the patients was 60.4 years in the UKA group and 57.8 years in the HTO group, meanwhile, the proportion of women in both groups was remarkably high. Both treatments showed an increase in the number of surgeries over the years, and in terms of underlying disease characteristics, hypertension, diabetes, and osteoporosis were more common in the UKA group, which had a higher average age (Table 1).

There were no significant differences in an unadjusted analysis when comparing the risk of requiring revision between the two groups (Table 2). The Kaplan–Meier survivorship curve showed no significant difference between UKA and HTO (p = 0.92) (Fig. 2A), and in the survival rate in the HTO group according to sex (p = 0.16) (Fig. 2B), but in the UKA group, the survival rate was higher in male patients than in female patients (p < 0.001) (Fig. 2C).

Table 2 COX proportional hazard survival analysis for risk of revision for 50–69 aged patients
Full size table
Fig. 2
figure 2

Kaplan–Meier survivorship curve: 11.5-year survival probability of total patients (A), gender differences in HTO (B), gender differences in UKA (C)

Full size image

An adjusted analysis considering baseline characteristics such as age, sex, comorbidities, insurance type, hospital size, residence, and the CCI, showed that the risk of requiring revision in HTO was higher than that in UKA (HR: 1.19, 95% CI: 1.11–1.27) (Table 2). The HRs were 1.17 (1.07–1.27) at 5 years and 1.18 (1.10–1.26) at 10 years. A Cox proportional hazard analysis revealed that the incidence of deep vein thromboembolism (DVT) (HR: 0.33, 95% CI: 0.28–0.39), and surgical site infection (HR: 0.40, 95% CI: 0.35–0.46) was significantly higher in UKA than in HTO (Table 3). Other adverse outcomes, including pulmonary thromboembolism, cerebrovascular disease, myocardial infarction, postoperative delirium, and acute renal failure (ARF), showed no significant differences. In terms of perioperative complications, the incidence of postoperative ICU admission was significantly higher in UKA (OR: 0.21, 95% CI: 0.15–0.29), while that of re-hospitalization within 30 days (OR: 1.28, 95% CI: 1.18–1.38) and 90 days (OR: 1.28, 95% CI: 1.22–1.34) was higher in HTO (Table 4).

Table 3 COX proportional hazard analysis for adverse outcomes
Full size table
Table 4 Logistic regression analysis for perioperative complications
Full size table

Discussion

We identified 78,448 patients who underwent UKA or HTO, to evaluate the results regarding the revision rate and postoperative complications relevant to each treatment. Throughout the total observation period of 11.5 years, the survival rate after UKA was significantly higher than that after HTO. In addition, when comparing the risk of complications after surgery, some significant differences were found between the two surgeries. Notably, the incidences of postoperative DVT and surgical site infection were higher in UKA than in HTO.

In previous studies on the survival rates of HTO and UKA, different results have been reported over the past 20 years. Some reported that UKA had better long-term survival rates compared to HTO [8, 16]. On the contrary, other studies showed better results for HTO [17], meanwhile, other studies concluded that there were no significant differences [5, 6, 13]. However, most of the previous comparison studies consisted of different patient characteristics, and short retrospective or small randomized controlled studies. In the present study, the long-term survival rate was analyzed, with adequate power and using adjusted covariates, and the risk of revision was compared at different time points, that is at 5 and 10 years.

In terms of postoperative complications, various results are reported in the literature. Some researchers reported that UKA is superior to HTO in terms of postoperative function and has fewer complications [6, 13, 18, 19], while others reported little differences between the two treatments [4, 14, 20]. However, to the best of our knowledge, the incidence of the major postoperative medical complications such as pulmonary, cerebrovascular and cardiac problems has not been compared between HTO and UKA, unlike a study exists in TKA and UKA [21]. We found no significant difference between the two surgeries in most complication rates, meanwhile, significant differences were found in the incidence of DVT and surgical site infection. The HTO group showed significantly low infection rate, which may be related to a relatively short operation time [22,23,24] and showed low risk of DVT possibly due to routine usage of mechanical compression after surgery [25], and pharmacologic prophylaxis [26]. There have been previous studies on factors related to ARF after orthopedic surgery, but studies related to different surgical methods, especially artificial joint surgery, have been insufficient and may require further evaluation [27, 28]. On the other hand, cemented knee arthroplasty, as one of the independent risk factors for postoperative ICU admission in some studies could relate that the result of higher ICU admission in UKA as shown in Table 4 [29, 30]. HTO often requires a more dependent functional status postoperatively compared to UKA, which may be a reason for higher re-hospitalization in HTO [2, 7, 31, 32].

In this study, the survival rate and complications of UKA and HTO were analyzed for mid-aged patients, but as the indications of UKA and HTO widened and overlapped [33,34,35], there are many reports of good results obtained by performing UKA and HTO in a younger patient group [10, 11]. However, there are still studies showing that the risk of early conversion to TKA increases as the patient's age increases [9, 34, 36]. In contrast, Other studies concluded that the relationship between age and implant survival and clinical outcome is not significant [35, 37,38,39]. they suggest that UKA and HTO should be performed regardless of age. Elderly or young age should not be a contraindication for selecting surgical methods [35, 37,38,39].

The strength of our study is the use of Korea's health insurance system which covers up to 99% of the entire population [15]. Using strict statistical analysis to reduce potential confounders, we carried out a long-term, large-scale population-based comparison study. However, this study had several limitations. First, the national registry data have inherent problems, including inaccurate diagnostic codes and lack of detailed medical records (e.g., type of inserted implants, different surgical techniques, and causes for revision surgery). Second, we could not identify the consecutive values such as body mass index, degree of leg deformity, and osteoarthritis stage. Furthermore, other clinical outcomes, including functional and radiological indicators were not provided. Third, the claim code of the health insurance claim data could not distinguish between the left and right side of the knee joint on which the procedure was done. Thus, patients who underwent the first knee replacement were defined when selecting subjects to analyze the comparative effects. The analysis included only those patients for whom there was only one claim for 'HTO or UKA' in the claim data, and the analysis excluded patients with two or more claims at different times because the exposure time could not be defined. Fourth, the maximum follow-up period in this study was 11.5 years which could be extended in future studies. In addition, there may be differences due to race and national differences compared to other countries. Despite these limitations, the authors believe that the current study is worthy because it is the first large-scale, long-term cohort study with patients of a specific age category, which is the boundary between the indications for the two surgeries.

Conclusions

It is important to choose an appropriate surgical method for unicompartmental knee OA considering that UKA has better results in terms of long-term survival rates, but may have a higher incidence of various complications even considering the high prevalence of underlying diseases in the UKA group in the preoperative patient characteristics.