Propensity score matching
The propensity score-matched population consisted of 40 bony ankylosed (38 patients) and 40 non-ankylosed (40 patients) hips as matched controls. The area under the receiver operating characteristic curve for fitting the propensity score was 0.68. This value is an appropriate value for calculating the propensity score as it ranges from 0.6–0.9 and is not extremely close to 0.5 or 1.0, which are considered inappropriate [19,20]. The standardised mean differences in baseline demographics for the matched study population are shown in Table 1 and Figure 4: all variables achieved an appropriate balance (standardised mean difference <0.1). In the bony-ankylosed group, the mean duration of ankylosis was 36.1 ± 19.0 years, with 22 cases of spontaneous ankylosis and 18 of arthrodesis.
Table 1. Baseline demographics for the propensity score-matched study population
Factors
|
Pre-match
|
Post-match
|
1:1
|
Bony-ankylosed
(N=40)
|
Non-ankylosed
(N=829)
|
p-value
|
SMD
|
Bony-ankylosed
(N=40)
|
Non-ankylosed
(N=40)
|
p-value
|
SMD
|
Propensity score
|
0.08 ± 0.07
|
0.05 ± 0.04
|
< 0.0001*
|
0.678
|
0.08 ± 0.07
|
0.08 ± 0.07
|
1.000
|
0.002
|
Age (years)
|
60.2 ± 7.8
|
59.3 ± 9.2
|
0.251
|
0.097
|
60.2 ± 7.8
|
61.0 ± 9.3
|
0.958
|
0.099
|
Sex (Female)
|
24 (60.0%)
|
736 (88.8%)
|
< 0.0001*
|
0.698
|
24 (60.0%)
|
24 (60.0%)
|
1.000
|
0.000
|
Height (cm)
|
156.8 ± 9.1
|
152.4 ± 6.8
|
0.005*
|
0.540
|
156.8 ± 9.1
|
156.4 ± 8.6
|
0.856
|
0.041
|
Weight (kg)
|
57.7 ± 8.8
|
55.5 ± 9.4
|
0.143
|
0.244
|
57.7 ± 8.8
|
57.1 ± 8.9
|
0.745
|
0.073
|
BMI (kg/m2)
|
23.4 ± 2.5
|
23.9 ± 3.7
|
0.279
|
0.146
|
23.4 ± 2.5
|
23.3 ± 2.6
|
0.796
|
0.058
|
*Significant differences (p<0.05).
SMD, standardised mean difference; BMI, body mass index
Clinical results
Intragroup comparison
In both groups, the postoperative JOA hip scores at the last follow-up had significantly improved compared to the preoperative scores for all components of “total”, “pain”, “range of motion”, “gait ability”, and “activities of daily living” (bony-ankylosed group: p<0.0001, p<0.0001, p<0.0001, p=0.0038, and p=0.0001, respectively; control group: p<0.0001, p<0.0001, p<0.0001, p<0.0001, and p<0.0001, respectively; Table 2).
Table 2. Intragroup comparison of the Japanese Orthopaedic Association hip score
Group
|
Factors
|
Before surgery
|
Last follow-up
|
p-value
|
Bony ankylosed
|
JOA hip score
|
|
|
|
|
Total
|
60.0 ± 9.8
|
76.4 ± 13.3
|
<0.0001*
|
|
Pain
|
40.0 ± 0.0
|
37.6 ± 3.0
|
<0.0001*
|
|
ROM
|
0.0 ± 0.0
|
13.1 ± 3.8
|
<0.0001*
|
|
Gait ability
|
9.7 ± 5.6
|
12.4 ± 6.0
|
0.0038*
|
|
ADL
|
10.3 ± 4.9
|
13.9 ± 5.2
|
0.0001*
|
Control
|
JOA hip score
|
|
|
|
|
Total
|
46.6 ± 12.4
|
90.0 ± 10.6
|
<0.0001*
|
|
Pain
|
14.3 ± 7.5
|
38.6 ± 3.6
|
<0.0001*
|
|
ROM
|
11.2 ± 3.5
|
16.8 ± 2.8
|
<0.0001*
|
|
Gait ability
|
9.3 ± 3.5
|
17.8 ± 3.7
|
<0.0001*
|
|
ADL
|
12.0 ± 2.5
|
16.8 ± 3.6
|
<0.0001*
|
* Significant differences (p<0.05).
Continuous variables are expressed as mean ± standard deviation.
JOA, Japanese Orthopaedic Association; ROM, range of motion; ADL, activities of daily living
Comparison between the bony-ankylosed and control groups
Preoperatively, the JOA hip scores for “total” and “pain” in the bony-ankylosed group were significantly higher than those in the control group (p<0.0001 and p<0.0001, respectively), whereas the scores for “range of motion” and “percentage of use of walking aids” were significantly lower (p<0.0001 and p=0.0368, respectively). No significant differences were observed between the two groups with regards the JOA hip scores for “gait ability” and “activities of daily living” (p=0.5654 and p=0.0563, respectively). Regarding the postoperative values, the JOA hip scores at the last follow-up for all components of “total”, “pain”, “range of motion”, “gait ability”, and “activities of daily living” were significantly lower in the bony-ankylosed group than in the control group (p<0.0001, p=0.0382, p<0.0001, p<0.0001, and p=0.0083, respectively).
The percentage of patients who used walking aids postoperatively, length of hospital stay, operating time, intraoperative and postoperative blood loss, allogeneic blood transfusion, and number of hips with postoperative complications were significantly higher in the bony-ankylosed group than in the control group (p<0.0001, p<0.0001, p<0.0001, p<0.0001, p=0.0003, p<0.0001, and p=0.0021, respectively; Table 3).
Table 3. Comparison of outcomes between bony-ankylosed and control groups
Factors
|
Bony-ankylosed
|
Control
|
p-value
|
Odds ratio
|
95% confidence interval
|
Preoperative JOA hip score
|
|
|
|
|
|
Total
|
60.0 ± 9.8
|
46.6 ± 12.4
|
<0.0001*
|
|
|
Pain
|
40.0 ± 0.0
|
14.3 ± 7.5
|
<0.0001*
|
|
|
ROM
|
0.0 ± 0.0
|
11.2 ± 3.5
|
<0.0001*
|
|
|
Gait ability
|
9.7 ± 5.6
|
9.3 ± 3.5
|
0.5654
|
|
|
ADL
|
10.3 ± 4.9
|
12.0 ± 2.5
|
0.0563
|
|
|
Preoperative use of walking aids
|
26 (65.0%)
|
34 (85.0%)
|
0.0368*
|
0.3
|
0.11–0.97
|
Postoperative JOA hip score at last follow-up
|
|
|
|
|
|
Total
|
76.4 ± 13.3
|
90.0 ± 10.6
|
<0.0001*
|
|
|
Pain
|
37.6 ± 3.0
|
38.6 ± 3.6
|
0.0382*
|
|
|
ROM
|
13.1 ± 3.8
|
16.8 ± 2.8
|
<0.0001*
|
|
|
Gait ability
|
12.4 ± 6.0
|
17.8 ± 3.7
|
<0.0001*
|
|
|
ADL
|
13.9 ± 5.2
|
16.8 ± 3.6
|
0.0083*
|
|
|
Postoperative use of walking aids
|
21 (52.5%)
|
5 (12.5%)
|
<0.0001*
|
7.7
|
2.51–23.8
|
Length of hospital stay (days)
|
29.3 ± 12.4
|
19.4 ± 9.2
|
<0.0001*
|
|
|
Follow-up period (years)
|
12.7 ± 2.7
|
11.6 ± 1.5
|
0.1619
|
|
|
Operating time (minutes)
|
68.8 ± 19.9
|
34.8 ± 9.5
|
<0.0001*
|
|
|
Intraoperative blood loss (g)
|
328.5 ± 246.4
|
175.6 ± 136.3
|
<0.0001*
|
|
|
Postoperative blood loss (g)
|
834.1 ± 462.7
|
504.9 ± 284.5
|
0.0003*
|
|
|
Allogeneic blood transfusion (unit)
|
2.4 ± 1.9
|
0.8 ± 1.2
|
<0.0001*
|
|
|
Complications
|
12 (30.0%)
|
2 (5.0%)
|
0.0021*
|
8.1
|
1.69–39.32
|
Revision surgery
|
3 (7.5%)
|
1 (2.5%)
|
0.6153
|
2.1
|
0.18–23.59
|
* Significant differences (p<0.05).
Continuous variables are expressed as mean ± standard deviation. Categorical variables are expressed as absolute and percent values.
JOA, Japanese Orthopaedic Association; ROM, range of motion; ADL, activities of daily living
Complications in the bony-ankylosed group included heterotopic ossification (HO) in eight cases and loosening of the acetabular side prosthesis, destruction of the femoral side prosthesis, periprosthetic joint infection, and dislocation in one case each. Complications in the control group included periprosthetic joint infection and dislocation in one case each.
Laboratory results
In the bony-ankylosed group on postoperative day 1, the platelet and creatinine phosphokinase levels were significantly lower and higher, respectively, than those in the control group (p<0.0209 and p<0.0034, respectively; Table 4). A positive correlation was found between operative time and intraoperative blood loss; a negative correlation was found between age at primary THA and intraoperative blood loss, and between the duration of ankylosis before primary THA and intraoperative blood loss (p=0.0001, p<0.0080, and p<0.0289, respectively; Table 5).
Table 4. Comparison of laboratory data between bony-ankylosed and control groups
Laboratory data
|
Bony ankylosed
|
Control
|
p-value
|
Hb (g/dL)
|
|
|
|
Preoperative
|
12.8 ± 1.7
|
13.4 ± 1.7
|
0.0706
|
POD 1
|
10.1 ± 1.4
|
10.5 ± 1.6
|
0.2715
|
POD 7
|
9.8 ± 1.5
|
10.3 ± 1.6
|
0.1342
|
Plt (×109/L)
|
|
|
|
Preoperative
|
229.6 ± 75.6
|
248.9 ± 70.6
|
0.1358
|
POD 1
|
150.1 ± 42.4
|
178.4 ± 57.0
|
0.0209*
|
POD 7
|
248.1 ± 6.6
|
277.3 ± 78.8
|
0.0898
|
AST (U/L)
|
|
|
|
Preoperative
|
22.1 ± 7.4
|
22.4 ± 9.6
|
0.8756
|
POD 1
|
24.5 ± 10.5
|
21.3 ± 7.1
|
0.1291
|
POD 7
|
25.7 ± 9.4
|
23.3 ± 11.6
|
0.1501
|
ALT (U/L)
|
|
|
|
Preoperative
|
19.1 ± 10.1
|
21.2 ± 19.3
|
0.5080
|
POD 1
|
15.5 ± 7.1
|
15.8 ± 12.5
|
0.6637
|
POD 7
|
27.0 ± 19.6
|
26.6 ± 31.8
|
0.4384
|
CK (mg/dL)
|
|
|
|
Preoperative
|
104.3 ± 38.2
|
97.5 ± 48.3
|
0.3100
|
POD 1
|
616.0 ± 402.4
|
318.2 ± 91.5
|
0.0034*
|
POD 7
|
134.2 ± 83.4
|
140.2 ± 53.1
|
0.4784
|
ALP (IU/L)
|
|
|
|
Preoperative
|
265.7 ± 93.4
|
268.8 ± 83.5
|
0.8871
|
POD 1
|
186.8 ± 58.0
|
188.7 ± 66.2
|
0.9003
|
POD 7
|
243.3 ± 94.3
|
266.2 ± 144.5
|
0.4355
|
CRP (mg/dL)
|
|
|
|
Preoperative
|
0.2 ± 0.2
|
0.3 ± 0.4
|
0.1676
|
POD 1
|
3.1 ± 1.6
|
3.2 ± 1.9
|
0.9873
|
POD 7
|
2.0 ± 1.8
|
2.1 ± 1.8
|
0.3490
|
* Significant differences (p<0.05).
Continuous variables are expressed as mean ± standard deviation.
POD, postoperative day; Hb, haemoglobin; Plt, platelet; AST, aspartate transaminase; ALT, alanine aminotransferase; CK, creatine phosphokinase; ALP, alkaline phosphatase; CRP, C-reactive protein
Table 5. Correlation between intraoperative blood loss and other variables in the bony-ankylosed group
Factors
|
Spearman’s correlation coefficient (ρ)
|
p-value
|
Continuous variables
|
|
|
Age at primary THA
|
-0.4135
|
0.0080*
|
Height
|
-0.00135
|
0.9341
|
Weight
|
-0.1454
|
0.3706
|
BMI
|
-0.2529
|
0.1153
|
Duration of ankylosis before primary THA
|
-0.3595
|
0.0289*
|
Length of hospital stay
|
0.0000
|
1.0000
|
Operative time
|
0.5703
|
0.0001*
|
Postoperative blood loss
|
0.2821
|
0.0778
|
Allogeneic blood transfusion
|
-0.0104
|
0.9494
|
Preoperative total JOA hip score
|
-0.2836
|
0.0761
|
Postoperative total JOA hip score at last follow-up
|
-0.2920
|
0.0752
|
Preoperative laboratory data
|
|
|
Hb
|
0.0157
|
0.9232
|
Plt
|
0.2275
|
0.1580
|
AST
|
-0.2696
|
0.1172
|
ALT
|
-0.2555
|
0.1385
|
CK
|
-0.1807
|
0.3223
|
ALP
|
0.3299
|
0.0530
|
CRP
|
0.2505
|
0.1190
|
Categorical variables
|
|
|
Sex
|
|
0.8541
|
Ankylosis type
|
|
0.8519
|
Complications
|
|
0.1301
|
* Significant differences (p<0.05).
THA, total hip arthroplasty; BMI, body mass index; JOA, Japanese Orthopaedic Association; Hb, haemoglobin; Plt, platelet; AST, aspartate transaminase; ALT, alanine aminotransferase; CK, creatine phosphokinase; ALP, alkaline phosphatase; CRP, C-reactive protein
Radiographic results
To study the additional outcomes, CT was performed in 11 and eight patients in the bony-ankylosed and control groups, respectively (19 cases). The mean postoperative CSA of the gluteus medius in the bony-ankylosed group was 1363.3.9 ± 495.1 mm2, which was significantly lower than that in the control group (2377.5 ± 580.2 mm2; p=0.0030).
The mean postoperative CT values of the gluteus medius at the CSA in the bony-ankylosed group was -4.4 ± 28.6 Hounsfield unit, which was significantly lower than that in the control group (31.8 ± 23.0 mm2; p=0.0039). The mean time of postoperative CT imaging in the bony-ankylosed and control groups were 7.6 ± 3.5 and 10.1 ± 2.6 years, respectively, which were not significantly different (p=0.2437).
Survival of acetabular and femoral components
The cumulative survival rates of the cup, stem, and overall, with revision as the endpoint, were not significantly different between the bony-ankylosed and control groups (p=0.5620, p=0.1753, and p=0.3378, respectively; Figure 5A–C).