| Bokov et al., [119] |
Short-segment fusion±TLIF (162) or ALIF/D-LIF (50)
1 (153), 2 (70), 3 (21), 4 (5), 5 (1)
|
6, 12, and 18 mo |
38.8% (97/250) |
|
|
| Chen et al. [37] |
Short-segment fusion with PLIF
1 (97), 2 (57), 3 (12), 4 (8)
|
Median 10 mo (8-18) |
29.88% (52/174), 9.18% (83/904) of screws |
-
–
Mean VBQ scores: 3.1±0.5 (SL group) vs. 2.8±0.4 (no SL group), p<.001
-
–
Lowest DXA-measured BMD (g/cm2): 0.81±0.1 (SL group) vs. 0.86±0.1 (no SL group), p=.028
-
–
VBQ score (per point) an independent risk factor for SL (UOR 1.02, p<.001)
-
–
ROC analysis: VBQ score 2.87 best predicted SL (AUC 0.744)
-
–
Higher rates of SL in patients with VBQ ≥ 2.9 (65.4% vs. 31.1%, p<.001)
|
|
| Cho et al., [34] |
Single-level PLIF
L3/4 (13), L4/5 (60), L5/S1 (13)
|
Mean time to SL: 6.3±3.4 mo (osteoporosis) vs. 7.3±3.0 mo (normal BMD) |
19.76% (17/86) |
-
–
Mean lumbar T-scores: -2.8±0.5 (osteoporotic cohort) vs. 0.2±0.9 (normal BMD cohort), p<.001
-
–
Higher SL rates osteoporotic cohort (32.3% vs. 12.7%, p=.029)
-
–
Mean lumbar T-scores: -1.6±1.4 (SL group) vs. -0.7±1.6 (no SL group), p=.054
|
-
–
No association between SL and VAS, ODI, or EQ-5D at final follow-up
-
–
SL associated with lower fusion at 2 y (71.4% vs. 93.9%, p=.038)
|
| Kim et al., [38] |
Single or multilevel lumbosacral fusion |
NR |
30.97% (35/113) |
-
–
Mean L1-4 HU: 77.93±33.48 (SL group) vs. 118.79±44.59 (no SL group), p<.001
-
–
ROC analysis: HU 104.91 (AUC 0.774) best predicted SL
|
N/A |
| Li et al., [42] |
L4/5 OLIF |
1 year |
35.71% (40/112 levels) in 56 patients |
-
–
HU at the screw insertion site independently predicted SL at both cranial (UOR 0.971) and caudal (UOR 0.941) levels
-
–
ROC analysis: optimal HU cutoffs for predicting SL were 119.4 (AUC 0.816) for cranial and 113.75 (AUC 0.915) for caudal levels
|
N/A |
| Löffler et al. [128] |
Short-segment posterior fusion
L1-5 (2), L2-5 (16), L2-S1 (8), L3-S1(18), L4-S1 (2)
|
Case: 185 days (71-1359)
Control: 229 days (8-2679) |
Case (n=23), control (n=23) |
-
–
Mean vBMD: 86.5±29.5 (SL group) vs. 118.2±32.9 (no SL group), p=.001
-
–
ROC analysis: optimal cutoff for predicting SL of 81.8 (AUC 0.769)
|
N/A |
| Matsukawa et al., [39] |
Single-level PLIF, instrumented using CBT |
1 year |
13% (12/92), 4.6% (16/351) of screws |
-
–
Mean screw trajectory HU (summ): 7.68±1.8 (loosened screws) vs. 13.0±3.68 (fixed screws), p < 0.001
-
–
Mean lumbar BMD: 0.93±0.21 (loosened screws) vs. 1.03±0.17 (fixed screws), p=.048
-
–
Screw trajectory HU was the only independent risk factor for SL (OR 0.70, p=.018)
|
N/A |
| Sakai et al. [40] |
Single-level PLIF
L1/2 (1), L2/3 (2), L3/4 (8), L4/5 (30), L5/S1 (12)
|
3 mo |
23% (12/52), 12% (24/206) of screws |
-
–
Mean screw trajectory HU: 147±94 (loosened screws) vs. 208±91 (fixed screws); p<.001
-
–
Mean lumbar BMD: 1.04±0.32 (loosened screws) vs. 1.13±0.22 (fixed screws), p=.016
-
–
ROC analysis: optimal HU cutoff of 153.5 for female (AUC 0.88) and 186.5 for males (AUC 0.635)
-
–
Screw trajectory HU was independently predictive of SL (UOR 0.989, p=.006).
|
N/A |
| Shin et al. [137] |
Short-segment fusion with PLIF
1 (300), 2 (140), 3 (36), 4 (2)
|
1 year |
22.59% (108/478) |
-
–
Mean L4 HU: 86.9±39.4 (SL group) vs. 134.3±54.1 (no SL group), p<.01
-
–
L4 HU was an independent risk factor for SL (UOR 0.979, p=.002)
-
–
ROC analysis: optimal HU cutoffs for predicting SL varied based on number of levels fused and postoperative C7-S1 SVA
|
N/A |
| Xu et al. [43] |
L3-5 Posterolateral fusion |
1 year |
L3 SL 20.3% (29/143) |
-
–
Mean L3 HU (vertebral body): 98.6±25.8 (SL group) vs. 121.4±39.7 (no SL group), p<.001)
-
–
Mean L3 HU (pedicle, excluding cortical bone): 208.9±69.5 (Sl group) vs. 290.5±132 (no SL group), p=.002
-
–
Mean L3 HU (pedicle, including cortical bone): 249.4±71.4 (SL group) vs. 337±125.5 (no SL group), p=.001
-
–
Increased risk of L3 SL with lower L3 HU of vertebral body (OR 6.55, p=.005) and pedicle including cortical bone (OR 4.84, p=.008)
-
–
ROC analysis: SL best predicted by HU cutoffs of 130 at the vertebral body (AUC 0.674) and 340 at the pedicle (AUC 0.721)
|
N/A |
| Xu et al. [35] |
Long posterior fusion to sacrum
Median 6 (range 3-12)
|
NR |
S1 SL 41.0% (32/78) |
-
–
Mean lumbar T-scores: −1.7 ± 1.6 (SL group) vs. −0.6 ± 2.2 (no SL group), p=.034
-
–
Mean hip T-scores: −1.6 ± 0.7 (SL group) vs. −1.0 ± 1 (no SL group), p=.033
-
–
Higher rates of osteoporosis in patients with SL (75% vs. 20.6%, p<.001)
-
–
Osteoporosis an independent risk factor for SL (OR 2.511, p=.002)
|
-
–
Fusion rates 90.5% (SL) vs. 95.6% (no SL), p=.373
-
–
SL did not independently predict ODI score (p=.664)
|
| Yuan et al. [36] |
Long posterior fusion
SL 6.28±1.98, no SL 5.81±1.33
|
NR |
54.6% (71/130), 9.4% (168/1784) of screws |
-
–
Mean T-scores: -2.12±0.96 (SL group) vs. −1.4±1.48 (no SL group), p=.002
-
–
Higher rates of osteoporosis in patients with SL (50.7% vs. 27.12%, p<.001)
-
–
Both osteoporosis (OR 8.19, p=.001) and osteopenia (OR 5.52, p=.006) were independent risk factors for SL
|
|
| Zou et al. [145] |
PLF to L5 or S1±PLIF (323)
1 (170), 2 (210), 3 (90), 4 (3)
|
1 year |
30.0% (151/503) |
-
–
Mean lumbar HU: 106.3±33.9 (SL group) vs. 132.6±42.9 (no SL group), p<.001
-
–
Lumbar HU an independent risk factor for SL (UOR 0.977, p<.001)
-
–
SL rates were 4.1%, 33.3%, 53.3%, and 78.8% for 1 level, 2 levels, 3 levels, and 4 levels of fixation, respectively (OR 3.626, p < 0.001)
|
|
| Zou et al., [41] |
Short-segment fusion to L5 or S1±PLIF (169)
1 (78), 2 (112), 3 (45), 4 (17)
|
3, 6, and 12 mo |
30.6% (77/252); Most (96.1%, 172/179) at LIV or UIV |
-
–
Mean lumbar HU: 106.8±34.4 (SL group) vs. 129.8±45.7 (no SL group), p<.001
-
–
Mean lowest T-score: -2.1±1.5 (SL group) vs. -1.7±1.6 (no SL group), p=.074
-
–
ROC analysis: optimal HU cutoff for predicting SL of 108 (AUC 0.666)
-
–
Lumbar HU was an independent risk factor for SL (UOR 0.98, p=.002)
-
–
Higher rates of SL in osteoporotic patients (39.3% vs. 25.8%, p=.026)
|
|
