Total hip arthroplasty after core decompression: Time-to-event analysis and risk factors

To the Editor: Core decompression (CD), the predominant hip-preserving procedure for pre-collapse osteonecrosis of the femoral head (ONFH), is renowned for its improved short-term outcomes and relatively low postoperative complication rates; however, the reported medium-to-long-term survival rates in the literature vary and are a subject of controversy. Consequently, we conducted a large, prospective cohort study to assess the frequency and timing of conversion from CD to total hip arthroplasty (THA) and the factors associated with it.

This study was approved by the Ethics Committee of the 3rd Hospital of Hebei Medical University (No. 2014-015-1), and all participants signed an informed consent before the surgery. The inclusion criteria were: (1) age 18 years or older; (2) Association Research Circulation Osseous (ARCO) stages I–III ONFH. The exclusion criteria were: (1) history of fracture in the proximal femur or tumors; (2) history of any surgery on the affected hip; (3) inflammatory arthritis, including suppurative arthritis, and gouty arthritis; (4) having received steroid treatment in the last 6 months^[1]; (5) pregnancy, and (6) incomplete data.

Our data were all obtained from the Surgical Site Infection in Orthopedic Surgery database, which is a prospectively and manually maintained database of all data from orthopedic in-patients at the institution. We selected potential factors from our database, which included patients’ demographics, comorbidities, lifestyles, medications use, perioperative characteristics, and perioperative serum biomarkers.

The CD procedures were performed by a surgical team with a minimum of 10 years of surgical experience, including 2 senior orthopedic surgeons who had each conducted CD surgeries for 10 years and 15 years, respectively. Prior to surgery, we assessed the ARCO stage and the extent of collapse of the femoral head to determine whether to perform multiple drilling CD alone or in conjunction with autologous or allogeneic bone grafting. The procedures were conducted using the techniques previously described in detail.^[2] Postoperatively, the operated limb was instructed to be non-weight-bearing for 6 months with functional exercises, then partial-weight-bearing with the assistance of a walker was permitted, and full-weight-bearing without engaging in high-impact activities was allowed after 9 months. Patients who underwent simultaneous, contralateral light bulb decompression followed the same weight-bearing protocol. Patients who underwent simultaneous, contralateral THA were required to engage in partial-weight-bearing on the THA limb within 1 week and full-weight-bearing within 2 weeks. Additionally, all patients were prescribed Chinese Herbal Medicine, known as “Resurrection Capsules,” and have been taking it regularly for at least one year.

We conducted a prospective evaluation for these operated patients for CD between 2014 and 2019. The date when the CD procedure was performed was considered as the index date. We followed up patients from the index date until the first of the following events to occur: conversion to THA, loss to follow-up, death, or until July 31, 2022, providing a minimum potential follow-up period of 38 months. For each patient, we identified the presence or absence of THA using the operative reports and patient radiographs available in the electronic medical record system at our institution, or medical records obtained from other hospitals.

We initiated our analysis by examining baseline characteristics using descriptive statistics and assessing differences between the conversion and non-conversion groups. We used the Youden index to determine the optimal cut-off point for age and categorized serum biomarkers based on clinically recommended values. To mitigate potential bias from missing data, we employed multiple imputations for serum biomarkers with less than 5% missing values.

For the estimation of cumulative incidence of CD-to-THA conversion over time, we conducted Kaplan–Meier analysis with 95% confidence intervals (CIs). We also performed subgroup Kaplan–Meier analyses based on ARCO stages, American Society of Anesthesiologists (ASA) class (class I–II and class ≥III), anesthesia method, and age (≤45 years and >45 years). A competing risk analysis was not performed because the rate of death was only 0.3% (3/841) in our sample population.

We used univariate Cox regression analysis to select potential factors a priori. These selected factors included ARCO stage, ASA class, anesthesia method, age, and several preoperative serum biomarkers (mean corpuscular volume [MCV], MCH, GLOB, albumin-to-globulin ratio [AGR], TG, and fibrinogen [FIB] levels). We then performed a multivariable-adjusted Cox proportional hazards regression model to estimate hazard ratios (HRs) and 95% CIs for the time to conversion from CD to THA.

The proportional hazards assumption was tested by using the Schoenfeld residuals. We performed several sensitivity analyses to assess the robustness of the results: (1) excluding patients with simultaneous contralateral THA; (2) excluding patients with simultaneous contralateral light bulb decompression; (3) performing a type of graft by ARCO stages interaction test; (4) comparing the results from multiple imputations (MCV, AGR, and FIB) with complete-case analyses; (5) stratifying by surgeon to account for surgical experience effects; and (6) stratifying by etiology.

All P-values were 2-sided, and P <0.05 was considered significant for all tests. All statistical analyses were conducted using SPSS version 26.0 (IBM Corp, Armonk, NY, USA) and R software (Version 4.1.3, R Foundation for Statistical Computing, Vienna, Austria).

We screened the hips of 872 patients for potential inclusion. The final cohort included 596 patients with 841 hips undergoing CD procedures, including 245 simultaneous bilateral CD. Besides, 182 patients underwent simultaneous contralateral THA, and 8 patients underwent simultaneous contralateral light bulb procedure. Most patients in our cohort were middle-aged, male, and overweight, with ASA class II and ARCO stage II [Supplementary Table 1, http://links.lww.com/CM9/C87]. The median follow-up time was 5.0 years. During the follow-up period, 68 patients were lost to follow-up, and 2 patients died [Supplementary Figure 1, http://links.lww.com/CM9/C87].

At the end of the follow-up, 88 THAs occurred in our cohort, yielding a conversion rate of 21.7 per 1000 person-years. The cumulative incidence of conversion to THA from CD was 8.0% (95% CI: 6.3–10.1%) at 3 years and 11.2% (95% CI: 9.1–13.6%) at 5 years [Figure 1]. Cumulative incidence curves and log-rank tests indicated significant differences in incidence depending on ARCO stage, ASA class, anesthesia method, and age. We found that cumulative incidence at 5 years was 26.4% (95% CI: 13.0–37.7%) for ARCO stage-III patients; was 19.2% (95% CI: 7.7–29.2%) for ASA class ≥III patients; was 13.8% (95% CI: 10.1–17.4%) for patients with general anesthesia; and was 16.0% (95% CI: 11.6–20.2%) for patients who were >45 years of age [Supplementary Figure 2-5, http://links.lww.com/CM9/C87].

Figure 1.

Cumulative incidence curve with 95% CIs for patients undergoing one or more CDs (n = 841) with conversion to THA defined as the event of interest. The number of patient hips at risk are also presented for selected time points. CD: Core decompression; CI: Confidence interval; THA: Total hip arthroplasty.

The multivariable-adjusted, Cox proportional hazards regression model demonstrated that patient-related risk factors for conversion to THA included ARCO stage III (adjusted HR: 2.78, 95% CI: 2.16–3.57), ASA class ≥III (adjusted HR: 2.34, 95% CI: 1.09–5.01), general anesthesia (adjusted HR: 1.78, 95% CI: 1.35–2.34), and age >45 years (adjusted HR: 2.09, 95% CI: 1.96–2.22). Furthermore, preoperative serum biomarkers of MCV >100 fL (adjusted HR: 5.04, 95% CI: 1.92–13.23), AGR >2.4 (adjusted HR: 2.83, 95% CI: 1.66–4.84), and FIB >4 g/L (adjusted HR: 1.88, 95% CI: 1.58–2.25) were found as independent risk factors for conversion to THA [Supplementary Table 2, http://links.lww.com/CM9/C87].

Results from several sensitivity analyses (i.e., excluding patients with simultaneous contralateral light bulb decompression; comparing the results from multiple imputations (MCV, AGR, and FIB) with the complete-case analyses; stratification analyses by the surgeon experience, and the etiology of CD) did not change materially; interaction tests show no significant interaction between the type of graft and ARCO stages (P >0.05). However, sensitivity analyses that excluded patients with simultaneous contralateral THA found that the observed association of ASA class ≥III with conversion to THA was lower (adjusted HR: 1.36, 95% CI: 0.49–3.78) than in the original cohort (adjusted HR: 2.34, 95% CI: 1.09–5.01) [Supplementary Tables 3–9, http://links.lww.com/CM9/C87].

This study stands out due to it having a large Chinese CD cohort with relatively long follow-up periods. The classic CD procedure demonstrated a satisfactory 5-year survival rate, which may be associated with the incorporation of native postoperative management, such as prolonged non-weight-bearing periods and the administration of herbal prescriptions. This suggests that future healthcare policymakers should seriously consider integrating this management protocol into routine postoperative care. However, it is important to note that the effectiveness of prolonged non-weight-bearing time should be validated through high-quality randomized controlled trial or fundamental research. It is worth emphasizing that even when patients have the highest risk factors or a combination of multiple risk factors, it does not necessarily rule out the use of CD. Indeed, even among patients with the highest risk level (MCV >100 fL), the 5-year hip survival rate remains at 68.0%, consistent with general reports.^[3–5] Preoperative discussions should include informing patients about relevant risks, and surgeons should exercise prudent surgical decision-making, considering stratified management whenever possible.

Limitations of our study should also be acknowledged. First, our single-center study, conducted in a tertiary referral hospital, may introduce selection bias towards complex cases or those referred by physicians with CD knowledge. This might limit the generalizability to all eligible CD patients. Second, our unique postoperative management, involving herbal prescriptions and extended non-weight-bearing periods, differs from traditional methods. The relationship between these strategies and improved survival requires further investigation. Third, though our analysis identifies MCV as a strong risk factor for THA conversion, the wide CI (95% CI: 1.92–13.23) compromises certainty. Future studies should validate our results and explore mechanisms underlying the association between MCV and the risk of conversion to THA. Finally, residual confounding, stemming from unmeasured variables, such as the extent and severity of microvascular necrosis and bone marrow edema before CD, persists in multivariable analyses.

In conclusion, in a prospective cohort with a median follow-up of five years, we observed a notably lower conversion rate from CD to THA. Besides the established risk factors, ASA class ≥III, general anesthesia, and several preoperative serum biomarkers were also identified and should be considered when evaluating surgical indications and providing preoperative counseling to patients.

Acknowledgments

We sincerely thank all the patients in this study.

Funding

This study was supported by the Central Guided Local Science and Technology Development Funds Project (Science and Technology Innovation Base Project) (No. 236Z7754G) and Excellent Talents of Hebei Provincial Department of Finance and Health Commission (No. 2023).

Conflicts of interest

None.

Data availability

All data in this study can be obtained from the authors based on reasonable demand.