Results

Baseline characteristics

A total of 347 patients (544 hips) treated with combined osteoblast cell therapy and core decompression for ONFH were initially included in the study. However, 28 patients (42 hips) were lost to follow-up; therefore, the final analysis included 319 patients (502 hips). Baseline characteristics were comparable between included patients and those lost to follow-up (all standardized mean differences ≤ 0.44).

The mean age of patients was 34.4 ± 9.2 years, and the majority were male (n = 272; 85.3%). Of the treated patients, 183 (57.4%) had bilateral ONFH (Table 1). Moreover, 214 (42.6%) hips had steroid-induced ONFH, and 188 (37.5%) had idiopathic ONFH. Most of the treated hips were classified as either stage II (n = 226; 45.0%) or stage IIIA (n = 215; 42.8%), with fewer hips in stage I (n = 61; 12.2%) according to the ARCO 2019 classification (Table 2). Baseline characteristics of patients who were lost to follow-up were similar to those of patients included in the study (data not shown). After osteoblast cell therapy, patients were followed up for two to seven years, with an average follow-up duration of 3.5 years (42.0 ± 5.2 months).

Parameters | Number of patients (n = 319)a
Age (years), mean ± SD | 34.4 ± 9.2
BMI (kg/m2), mean ± SD | 25.7 ± 4.4
Duration since diagnosis (years), mean ± SD | 0.7 ± 0.7
Sex, n (%)
Male | 272 (85.3)
Female | 47 (14.7)
Unilateral/bilateral ONFH, n (%)
Unilateral | 136 (42.6)
Bilateral | 183 (57.4)

Parameters | Frequency (Percentage) | Median (IQR) (months)
Stages of the hip per ARCO 2019a,b
Stage I | 61 (12.2) | 42.0 (38.0-47.0)
Stage II | 226 (45.0) | 42.0 (38.0-46.0)
Stage IIIA | 215 (42.8) | 42.0 (37.0-45.0)
Etiology of ONFHa
Steroid | 214 (42.6) | 41.0 (38.0-46.0)
Idiopathic | 188 (37.5) | 43.0 (39.0-46.0)
COVID-19 (without steroids) | 52 (10.4) | 42.0 (35.2-46.2)
Alcohol | 33 (6.6) | 40.0 (24.0-47.0)
Trauma | 15 (3.0) | 43.0 (39.0-49.5)

Clinical outcomes

At follow-up, 80.6% 80.6% (257/319; 95% CI: 75.90-84.50) of patients showed an improvement of ≥ 2 points in the VAS score, with a mean change of 3.78 (95% CI: 3.57-3.99) and a standard response mean of 2.00 (Figure 2).

Figure 2: Visual analogue scale score at baseline and follow-up visitsData are presented patient-wise. The blue dotted line indicates the MCID threshold value for VAS.MCID: minimal clinically important difference; VAS: visual analogue scale

Likewise, 90.3% (288/319; 95% CI: 86.50-93.10) of patients showed an improvement of ≥ 10 points in the HHS, with a mean change of 47.01 (95% CI: 45.13-48.89) and a standard response mean of 2.75 (Figure 3).

Figure 3: HHS at baseline and follow-up visitsData are presented patient-wise. The blue dotted line indicates the MCID threshold value for HHS.HHS: Harris Hip Score; MCID: minimal clinically important difference

Pre- and postoperative representative X-ray images of patients who underwent osteoblast cell therapy are shown in Figures 4-5. The postoperative images show radiographic improvement after successful osteoblast cell therapy.

Figure 4: Preoperative and postoperative (9 and 15 months) X-ray of pelvis(A) Preoperative X-ray of the pelvis showing ONFH and normal joint space; (B, C) 9- and 15-month postoperative X-rays with osteoblast cell therapy showing a smooth femoral head surface with normal joint space and minimal sclerosis of the femoral head, suggestive of osteogenesis.ONFH: osteonecrosis of the femoral head

Figure 5: Preoperative and postoperative (24 months) X-ray of pelvis with hips(A) Preoperative X-ray of the pelvis showing the right hip with ONFH and the left hip with a total hip replacement prosthesis in situ; (B) 24-month follow-up X-ray showing sclerosis of the right femoral head indicative of osteogenesis, where core decompression with osteoblast cell therapy was done.ONFH: osteonecrosis of the femoral head

Likewise, representative postoperative MRI images demonstrated maintenance of femoral heads’ sphericity and joint spaces, with no further collapse; there were no clear radiological signs of progression of disease or arthritic changes in the femoral head (Figures 6-7).

Figure 6: Preoperative and postoperative (24 months) MRI(A) Preoperative MRI; (B) 24-month follow-up MRI showing evidence of osteogenesis in the femoral head, with maintenance of sphericity of the femoral heads and of the joint space.MRI: magnetic resonance imaging

Figure 7: Preoperative and postoperative (36 months) MRI of hips(A) Preoperative MRI of the right hip; (B) 36-month follow-up MRI showing evidence of osteogenesis in the right femoral head, with maintenance of sphericity of right femoral head and of joint space.MRI: magnetic resonance imaging

At follow-up, 31 patients (44 hips) required THA for the hips treated with osteoblast cell therapy. No patient required any other surgical treatment for ONFH. Hence, the estimated patient-wise and hip-wise treatment failure rates were 9.7% and 8.8%, respectively. The average time for hip failure was 2.7 years (range: 1-7 years). Of the treated hips, the majority that required THA at follow-up had stage IIIA ONFH (30/44) (Figure 8).

Figure 8: Stage-wise hip survival vs failure after osteoblast cell therapyTHA: total hip arthroplasty.

Based on the survival analysis, the median THA-free survival was 78 months (Figure 9), whereas the median follow-up by the reverse KM method was 43 months (IQR: 39-46 months; range: 12-84 months).

Figure 9: Kaplan–Meier curve for overall THA-free survival for hipsTHA: total hip arthroplasty.

Furthermore, the etiology-based KM curve showed that trauma, COVID-19, and alcohol-related ONFH had the highest and earliest THA conversions (global log-rank test p < 0.001), whereas the KM curve based on disease stage showed that patients with stage IIIA ONFH were more likely to require THA (Figure 10).

Figure 10: Kaplan–Meier curve for THA conversion at hip level based on (A) etiology and (B) disease stageCOVID-19: coronavirus disease 2019; THA: total hip arthroplasty

Risk-factor analysis using Cox regression showed that a longer diagnosis-to-treatment interval was a significant risk factor for conversion to THA with a HR of 1.28 per month (95% CI: 1.15-1.41, p < 0.005) (Table 3).

Parameters | HR (95% CI) | p-value
Duration between diagnosis and treatment (months) | 1.28 (1.15-1.41) | < 0.005
Age (years) | 1.04 (0.99-1.09) | 0.09
Sex | 0.53 (0.24-1.19) | 0.13
Stage II | 4.27 (0.51-35.49) | 0.18
Stage IIIA | 13.33 (1.56-113.76) | 0.02
COVID-19 | 0.43 (0.15-1.21) | 0.11
Steroid | 0.19 (0.07-0.49) | < 0.005
Idiopathic | 0.09 (0.03-0.30) | < 0.005
Trauma | 0.66 (0.17-2.57) | 0.55
Unilateral ONFH | 0.22 (0.08-0.67) | 0.01
Baseline HHS | 1.02 (0.98-1.06) | 0.43
Baseline VAS | 1.08 (0.78-1.51) | 0.63

The proportional hazards assumption was assessed using Schoenfeld residuals and global tests, which showed sex and baseline HHS violations. Further, when adjusted for demographic, clinical, and disease factors, regression analysis showed an increased risk of THA conversion with each one-month delay in treatment (odds ratio (OR): 1.39; 95% CI: 1.22-1.61). Delay was examined using clinically meaningful increments (three- and six-month effects) and found to be exponential, for a six-month delay (OR: 7.13; 95% CI: 3.29-15.41) and for a 12-month delay (OR: 50.78; 95% CI: 10.85-237.59). Sensitivity analysis using restricted cubic splines in Cox models confirmed a statistically significant nonlinear association between treatment delay and THA conversion (likelihood ratio test p = 0.0006).

The spline findings indicated disproportionately steeper risk increases at longer delays, with wider uncertainty at the extremes due to fewer observations. Further analysis stratified by disease stage and etiology was conducted using adjusted Cox regression for the full set of covariates. In these analyses, stage IIIA disease remained strongly associated (adjusted HR: 13.60; 95% CI: 2.41-76.72, p = 0.003) with conversion to THA per six-month delay in treatment compared with stage I, while steroid-associated (adjusted HR: 0.22; 95% CI: 0.06-0.84, p = 0.026), and idiopathic (adjusted HR: 0.09; 95% CI: 0.02-0.36, p < 0.001) osteonecrosis were associated with lower hazards of THA compared with alcohol-associated osteonecrosis (Table 4). Overall, these adjusted findings confirm that alcohol is associated with the highest risk etiology for THA conversion in this cohort.

Parameters | Number of Hips | THA events | Unadjusted HR (95% CI) | p-value | Adjusted HR (95% CI)a | p-value
ARCO 2019 Stage
Stage I | 61 | 1 | 1 | NA | 1 | NA
Stage II | 226 | 13 | 5.82 (0.95-35.68) | 0.057 | 4.30 (0.67-27.68) | 0.125
Stage IIIA | 215 | 30 | 17.55 (3.59-85.78) | < 0.001 | 13.60 (2.41-76.72) | 0.003
Etiology
Alcohol | 33 | 9 | 1 | NA | 1 | NA
Steroid | 214 | 18 | 0.17 (0.05-0.53) | 0.002 | 0.22 (0.06-0.84) | 0.026
COVID-19 | 52 | 9 | 0.56 (0.16-1.93) | 0.36 | 0.51 (0.11-2.38) | 0.39
Idiopathic | 188 | 4 | 0.06 (0.02-0.20) | < 0.001 | 0.09 (0.02-0.36) | < 0.001
Trauma | 15 | 4 | 0.73 (0.17-3.17) | 0.67 | 0.79 (0.21-3.05) | 0.74

Cox proportional hazards model demonstrated that each additional month of delay in osteoblast cell therapy increased the risk of THA (HR: 1.28; 95% CI: 1.15-1.41). A similar estimate was observed after accounting for within-patient correlation and center effect (HR: 1.26; 95% CI: 1.13-1.39). Details of patients who underwent THA for treatment failure are presented in Table 5.

Sr. No. | Age/sex | Hip involvement (ONFH) diagnosed at the time of osteoblast cell therapy | Stage of disease at the time of osteoblast cell therapya(Left) | Stage of disease at the time of osteoblast cell therapya(Right) | Etiology of ONFH | Hip converted to THA (treatment failure) | Time to THA conversion (years)
1 | 29/F | Bilateral | IIIA | IIIA | Steroid | Right | 6
2 | 38/M | Bilateral | II | IIIA | Steroid | Right | 2.5
3 | 42/M | Bilateral | IIIA | II | COVID-19 | Left | 1.5
4 | 36/F | Bilateral | IIIA | II | Steroid | Left | 2
5 | 36/M | Unilateral | - | II | Idiopathic | Right | 2
6 | 29/M | Bilateral | II | IIIA | Alcohol | Left & Right | 1
7 | 34/F | Unilateral | II | - | Trauma | Left | 3
8 | 31/M | Bilateral | II | IIIA | Steroid | Right | 1.5
9 | 36/M | Unilateral | - | II | Idiopathic | Right | 3
10 | 41/M | Bilateral | II | IIIA | Steroid | Right | 2
11 | 36/F | Bilateral | II | IIIA | Steroid | Right | 2.5
12 | 37/M | Bilateral | IIIA | IIIA | Steroid | Left & Right | 1.5
13 | 51/M | Bilateral | II | IIIA | Trauma | Left & Right | 2
14 | 47/M | Bilateral | IIIA | IIIA | COVID-19 | Left & Right | 2
15 | 40/M | Bilateral | IIIA | II | Idiopathic | Left | 3
16 | 21/F | Bilateral | IIIA | IIIA | Steroid | Left & Right | 3
17 | 27/M | Bilateral | IIIA | II | Alcohol | Left & Right | 1
18 | 26/F | Bilateral | II | IIIA | Steroid | Right | 2
19 | 39/M | Bilateral | II | IIIA | Alcohol | Left & Right | 1.5
20 | 37/M | Bilateral | IIIA | II | Steroid | Left | 3
21 | 38/F | Bilateral | II | IIIA | COVID-19 | Right | 1.5
22 | 45/M | Bilateral | IIIA | II | COVID-19 | Left | 2
23 | 46/M | Bilateral | II | IIIA | Alcohol | Left & Right | 2
24 | 29/M | Bilateral | II | I | Idiopathic | Left | 6
25 | 38/F | Bilateral | IIIA | IIIA | COVID-19 | Left & Right | 2
26 | 27/M | Bilateral | I | II | Steroid | Left & Right | 7
27 | 24/M | Bilateral | II | IIIA | Steroid | Left & Right | 6.5
28 | 19/F | Bilateral | IIIA | IIIA | Steroid | Left & Right | 3.5
29 | 44/F | Unilateral | - | II | Trauma | Right | 5.5
30 | 37/M | Bilateral | II | IIIA | COVID-19 | Left & Right | 1.5
31 | 31/M | Bilateral | IIIA | II | Alcohol | Left | 1