1.INTRODUCTION

Osteoporosis, the most common metabolic bone disease, leads to alteration in bone density that has been shown to compromise the strength of spinal instrumentation.1 With elderly populations growing, rates of spine surgery performed on osteoporotic patients have increased to treat a variety of conditions.1 Vertebral fracture is the most common osteoporotic fracture in the elderly, and surgical intervention is sometimes needed for patients diagnosed with nonunion, failure of vertebroplasty, and neurologic deficits.2–4 Therefore, spine surgeons will increasingly face the challenge of achieving rigid fixation of osteoporotic spines.

Cement-augmented pedicle screws (CPS), the most commonly used strategy, maximizes pullout strength in fixation of osteoporotic spines.2 Hybrid constructs (HC), a combination of pedicle screws and hooks, offer an alternative approach to avoid implant failure and increase construct stability when placing instrumentation in the osteoporotic spine.5–7 Biomechanical studies of either CPS8,9 or HC5–7 for osteoporotic spine have demonstrated superior implant pullout strength, compared with pedicle screws only. However, few clinical investigations to date have focused on comparing surgical results of the CPS and HC techniques. The goal of this retrospective cohort study was to compare the surgical outcomes and surgery-, patient-, and implant-related complications between the CPS and HC techniques for osteoporotic vertebral fractures of the thoracic or lumbar spine.

2. METHODS 2.1. Patient collection

The institutional review board of our hospital approved the research protocol (2021-01-030CC). The requirement for informed consent was waived because of the retrospective nature of this study. This study was performed in accordance with the ethical principles set out in the 1964 Declaration of Helsinki. This study included 99 consecutive patients with osteoporotic vertebral fractures who were treated surgically at our hospital from January 2011 to December 2015. The indications for surgery were failure of conservative treatment or vertebroplasty or neurologic deficit. We defined treatment failure as persistent back pain without improvement following conservative treatment or vertebroplasty. Fifteen patients were excluded from the study: eight were lost to follow-up (they did not complete the latest plain radiograph of spines), two died of unrelated medical conditions, three had degenerative scoliosis, and two had T scores greater than −2.5. Consequently, 84 patients, 22 males and 62 females, were included in the study.

All patients underwent long posterior instrumentation with or without posterior decompression, depending on whether or not they had neurological deficits. All operations on the CPS and HC groups were performed by two senior surgeons. The CPS group consisted of 43 patients who received long instrumentation with CPS; the 41 patients in the HC group received a combination of pedicle screws and hooks. Long instrumentation was defined as instrumentation at least two levels above and below the fractured level for implant insertion, regardless of technique used.

We used the World Health Organization criterion for osteoporosis, which is a T score less than −2.5 using dual-energy X-ray absorptiometry (DXA) of the hips,10 completed before or right after the index surgery. The worst T score at either hip on the DXA report was used to decide whether the patient met the definition of osteoporosis. Surgeries were scheduled on a priority basis, except that patients with neurologic deficits were scheduled on an emergency basis. Patients were placed in a prone position on a four-poster frame with postural reduction of the fracture and were checked under an image intensifier.

2.2. CPS techniques

For the CPS group, we used cement-augmented polyaxial cannulated pedicle screws (Smartlock Omega, A-Spine Inc., New Taipei City, Taiwan), with a diameter of 6.0 mm and side holes, two levels above and below the fractured vertebra.2,11 Long instrumentation of five levels was used for all patients in the CPS group. Under C-arm guidance, approximately 1 to 1.5 mL of cement (Cohesion, Vexim Sa, Balma, France) was injected through the cannulated lumbar pedicle screw and approximately 1 mL was used for the thoracic pedicle screw (Fig. 1).

Fig. 1:

Illustration of the cement-augmented pedicle screws (CPS) group. A, Preoperative lateral plain radiograph of spine. B, Immediate postoperative lateral plain radiograph of spine. C, Lateral plain radiograph of spine at final follow-up. Pedicle screws (6.0 mm in diameter with side holes) were used at two levels above and below the fractured vertebra.

2.3. HC techniques

For the HC group, we inserted regular pedicle screws (Smartlock Omega, A-Spine Inc., New Taipei City, Taiwan) with a diameter of 6.0 mm into the lumbar spine two levels below the fractured vertebra. Appropriate hooks were chosen based on the individual patient’s anatomy. Four pedicle hooks were used two levels above the fractured vertebra in the thoracic spine, except for T12 and the decompressed level. Two additional wide-blade transverse process hooks were placed, pointing downward toward the head in a pedicle-transverse claw configuration. At the most caudal level of the instrumented vertebra, we applied two additional offset laminar hooks pointing upward for a hook-screw claw construct. Long instrumentation of six to seven levels was applied in the HC group (Fig. 2).

Fig. 2:

Illustration of the hybrid construct (HC) group. A, Preoperative lateral plain radiograph of spine. B, Immediate postoperative lateral plain radiograph of spine. C, Lateral plain radiograph of spine at final follow-up. D, Anteroposterior plain radiograph of spine at final follow-up. Six hooks (four pedicle hooks with two additional transverse hooks) were applied at the cephalad end, except at T12, to create a pedicle-transverse claw construct. Four regular pedicle screws (6.0 mm in diameter) and an additional two offset laminar hooks were used to form a hook-screw claw construct.

2.4. Decompression and fusion techniques

We performed posterior decompression and posterolateral fusion one level above and below the fracture for patients who had neurologic deficits. On neurologically intact patients, posterior fusion without posterior decompression was done. In total, all patients in both groups received three levels of fusion (fractured level and one level above and below). Bone grafting was accomplished using a mixture of autogenous bone graft, which was harvested from decompressed laminae and the spinous processes, and bone substitute (synthetic β-tricalcium phosphate, chronOS, DePuy Synthes, PA, USA). Patients with pseudarthrosis at the fractured vertebrae received additional vertebroplasty using bone cement (Cohesion, Vexim Sa, Balma, France) under C-arm guidance.

2.5. Postoperative care and fusion evaluation

The wound was closed with a suction drain left in. Patients were allowed to walk in a brace depending on their condition and an antiosteoporotic agent was prescribed after surgery. Every patient was followed up at our clinic with serial supine anteroposterior and lateral radiographs every 6 weeks postoperatively for 3 months; at 3, 6, and 12 months; and then annually.

Sound fusion or union was defined as bone healing in the fractured vertebra or bridging callus across the intervertebral disc on the lateral plain radiograph, or a posterior or posterolateral fusion mass seen on the anteroposterior plain radiograph in patients with no loss of correction or increase in back pain. The fusion status was evaluated by the same author (P.H.C.) (a 10-year experienced spine surgeon) at two different time points, separated by at least 1 to 2 weeks. Any discordance was resolved by consultation and agreement with another experienced spine surgeon (C.L.L.). In order to reduce costs and the exposure to radiation, and in accordance with the National Health Insurance (NHI) policy in our country, we did not routinely employ postoperative computed tomography (CT) scans to assess union, screw position, and cement distribution. Implant failure was defined as pullout or breakage of the implant.12

To examine the radiographs more objectively and to minimize bias, two surgeons not previously involved in the surgery measured all the parameters using the Picture Archiving and Communication System (PACS) (Smart Viewer 3.2; Taiwan Electronic Data Processing Cooperation, Taipei, Taiwan). They employed Cobb’s method to measure the regional kyphotic angle.13 The regional kyphotic angles were measured between the superior endplate of the vertebra one level above the fractured vertebra, and the inferior endplate of the vertebra one level below.

2.6. Functional outcomes evaluation

To evaluate pain and disability, we used the Visual Analog Scale (VAS) for back pain and the Oswestry Disability Index (ODI), respectively, preoperatively and at final follow-up. We classified ambulatory performance into four categories: bed-ridden, wheelchair-bound, ambulation with an aid (walker, cane, crutches, or needing assistance), and ambulation without an aid.14 Neurologic function was evaluated with the Frankel grading system.15 Ambulatory performance and neurologic function were evaluated preoperatively and at final follow-up.

2.7. Statistical analysis

Statistical analysis was performed using the SPSS for Windows statistical package (version 15.0, Chicago, IL, USA) with a p value <0.05 considered significant. To compare parameters between the two groups, we used the Mann-Whitney U test (for continuous data) and the chi-square test (for categorical data). Repeat measure analysis of variance (ANOVA) testing was conducted to investigate changes in the postoperative kyphotic angle over time. To determine whether these tests were appropriately powered, power analysis was also performed using G*Power software (Heinrich-Heine Universität Dusseldorf, Düsseldorf, Germany) for each comparison between the two groups.

3. RESULTS

The mean age at the index operation was 77.7 ± 8.5 years, 77.3 ± 7.6 years for the CPS group and 78.2 ± 9.7 years for the HC group. Seven patients had fractures on T10, 15 on T11, 22 on T12, 24 on L1, 13 on L2, and 3 on L3. The overall preoperative kyphotic angle was 24.8° ± 6.2° (range, 14° to 44°); 24.6° ± 5.9° for the CPS group and 25.1° ± 6.5° for the HC group. No statistically significant differences were observed between these two groups in preoperative kyphotic angle or demographic data (Tables 1 and 2).

Table 1 - Demographic data for the CPS and HC groups before operation Total CPS HC p (n = 84) (n = 43) (n = 41) Age 77.7 ± 8.5 (70-87) 77.3 ± 7.6 (70-85) 78.2 ± 9.7 (71-87) 0.506 BMI 20.8 ± 3.9 (15.6-32.6) 21.2 ± 4.3 (15.6-32.6) 20.6 ± 3.6 (17.0-31.5) 0.594 Sex 0.357  Male 22 13 9  Female 62 30 32 DXA −2.8 ± 0.4 (−2.5 to −3.5) −2.9 ± 0.3 (−2.5 to −3.5) −2.8 ± 0.3 (−2.5 to −3.4) 0.109 Surgical indications 0.218  Failure of conservative treatment or VP 19 11 8  Neurologic deficit 65 32 33 Injured level 0.656  T10 7 3 4  T11 15 8 7  T12 22 12 10  L1 24 12 12  L2 13 6 7  L3 3 3 0 Comorbiditiesa 0.935  Hypertension 16 7 9  Diabetes mellitus 4 5  End stage renal disease 3 2 1  Coronary artery diseases 6 4 2  Lacunar infarction or cerebral infarction 5 3 2  Chronic obstructive pulmonary diseases 9 5 4  Parkinson’s disease 6 3 3  Benign prostate hyperplasia 5 3 2 Preoperative functional scores  Visual Analog Scale at back 8.0 ± 1.5 (6-10) 8.1 ± 1.3 (7-10) 7.9 ± 1.7 (6-10) 0.637  Oswestry Disability Index 67.7 ± 13.8 (52-88) 68.1 ± 15.4 (58-86) 65.4 ± 16.2 (52-88) 0.545

Data are presented as mean ± standard deviation (range).

BMI = body mass index; CPS = cement-augmented pedicle screw; DXA = dual-energy X-ray absorptiometry; HC = hybrid construct; VP = vertebroplasty.

aSome patients had more than one comorbidity.

Table 2 - Surgical results for the CPS and HC groups Total CPS HC p G power (n = 84) (n = 43) (n = 41) Mean operation time (h)a 243 ± 51 (180-350) 258 ± 54 (190-350) 231 ± 49 (180-340) 0.013 0.9 Mean blood loss (mL)a 462 ± 140 (230-950) 427 ± 122 (230-950) 497 ± 129 (280-780) 0.014 0.71 Mean hospitalization (d) 15.8 ± 5.8 (11-41) 16.1 ± 4.2 (11-41) 15.6 ± 4.7 (12-32) 0.608 Mean kyphotic angle (°)  Preoperative 24.8 ± 6.2 (14-44) 24.6 ± 5.9 (14-41) 25.1 ± 6.5 (17-44) 0.712  Postoperative 15.0 ± 5.6 (6-30) 15.4 ± 5.6 (6-30) 14.5 ± 5.4 (8-30) 0.447  Immediate postoperative correctiona 9.8 ± 2.1 (6-14) 9.1 ± 1.6 (6-12) 10.6 ± 2.3 (6-14) <0.001 0.93  6-mo f/u 16.7 ± 5.4 (5-34) 18.3 ± 4.9 (5-31) 16.3 ± 5.4 (9-34) 0.08  1-y f/u 18.7 ± 5.7 (7-36) 19.8 ± 5.1 (7-33) 18.3 ± 5.8 (10-36) 0.221  Final follow-up 23.4 ± 6.0 (13-43) 22.6 ± 5.7 (13-40) 24.3 ± 6.4 (16-43) 0.321  Loss of reduction at latest f/ua 8.4 ± 2.9 (1-13) 7.1 ± 1.3 (3-13) 9.8 ± 2.7 (1-13) <0.001 0.99 Mean fixation segments 5.6 ± 0.7 (5-7) 5 6.2 ± 0.4 (6-7) Mean fusion segments 3 3 3 Fractured healed/bridging callus/fusion mass 62 32 (74.4%) 30 (73.1%) 0.902 ASA physical status classification 0.49  II 24 14 10  III 44 31 31 Vertebroplasty at fracture level 26 14 (32.5%) 12 (29.3%) 0.753 Surgical complications 0.904  Wound infection 4 2 (4.6%) 2 (4.9%)  Dura tear 0 0 0 Implant-related complications  Implant failure 2 1 (1/43, 2.3%) 1 (1/41, 2.4%) 0.966 Cement-related complications  Screw augmentation   Linear or spotted pattern leakage of screwsa 134 134 (134/344, 38.9%) 0 <0.0001   Leakage into canal or epidural space 0 0 0   Symptomatic pulmonary embolism 0 0 0  Vertebroplasty   Leakage into canal or epidural space 0 0 0   Symptomatic pulmonary embolism 0 0 0 Patient-related complications 0.659  Pneumonia 7 4 3  Urinary tract infection 4 2 2  Stroke (transient ischemia attack) 1 1 0  Peptic ulcer 1 1 0 Functional outcomes at latest f/u  Visual Analogue Scale at back 2.4 ± 0.9 (1-4) 2.4 ± 0.9 (1-4) 2.4 ± 1 (1-4) 1.00  Oswestry Disability Index 27.8 ± 7.1 (18-50) 27.3 ± 6.5 (20-46) 28.5 ± 7.6 (18-50) 0.546  Improvement of ambulatory performance 0.7 ± 0.8 (0-2) 0.7 ± 0.8 (0-2) 0.8 ± 0.76 (0-2) 0.559  Improvement of neurologic function 0.6 ± 0.57 (0-2) 0.5 ± 0.59 (0-2) 0.7 ± 0.53 (0-2) 0.11 Mean follow-up times (mo) 67 ± 14 (52-88) 68 ± 15 (58-86) 65 ± 16 (52-88) 0.545

Data are presented as mean ± standard deviation (range).

ASA = American Society of Anesthesiologists; CPS = cement-augmented pedicle screw; F/U = follow-up; HC = hybrid construct.

aStatistical significance.

Clinical and radiographic examinations were available for all 84 patients during an overall average of 67 ± 14 months of follow-up (range, 52-88 months), 68 ± 15 months (range, 58-86 months) for the CPS group and 65 ± 16 months (range, 52-88 months) for the HC group. The overall average postoperative kyphotic angle was 15.0° ± 5.6° (range, 6°–30°), 15.4° ± 5.6° (range, 6°-30°) for the CPS group and 14.5° ± 5.4° (range, 8°-30°) for the HC group. The overall immediate average correction of the kyphotic angle was 9.8° ± 2.1° (range, 6°-14°), 9.1° ± 1.6° (range, 6°-12°) for the CPS group and 10.6° ± 2.3° (range, 6°-14°) for the HC group, indicating significantly better immediate correction in the HC group (p < 0.001; Table 2).

Significant progressive loss of correction of the kyphotic angle was observed with time, regardless of fixation technique (p < 0.05). No statistical difference in loss of correction between the two groups was observed at various postoperative follow-up times (Fig. 3). At final follow-up, the overall average loss of reduction of the kyphotic angle was 8.4° ± 2.9° (range, 1°-13°), 7.1° ± 1.3° (range, 3°-13°) for the CPS group and 9.8° ± 2.7° (range, 1°-13°) for the HC group. The loss for the CPS group was statistically significantly greater (p < 0.001). However, in the final kyphotic angle, no statistically significant difference was observed between the CPS and HC group (22.6° ± 5.7° vs 24.3° ± 6.4°, p = 0.321; Table 2).

Fig. 3:

Serial change in regional kyphotic angle at different follow-up (f/u) times. A, Illustration of the regional kyphotic angle. B, Postoperative progressive kyphotic changes over time were observed regardless of fixation technique (p < 0.05). No statistical difference between the two groups was observed at various postoperative follow-up points.

Average overall operative time for the 84 patients in the study was 243 ± 51 minutes (range, 180-350 minutes), 258 ± 54 minutes (range, 190-350 minutes) for the CPS group, and 231 ± 49 minutes (range, 180-340 minutes) for the HC group, indicating that operative times were significantly shorter for the HC group. The average estimated blood loss overall was 462 ± 140 mL (range, 230-950 mL), 427 ± 122 mL (range, 230-950 mL) for the CPS group and 497 ± 129 mL (range, 280-780 mL) for the HC group. In other words, patients in the CPS group lost significantly less blood.

The mean number of fixation segments was 5.6 levels (range, 5-7) for all patients, five levels for the CPS group, and 6.2 ± 0.4 levels (range, 6-7) for the HC group. The number of fused segments was three in both groups (Table 2).

At final follow-up, functional outcomes were similar for the two groups, with significant improvement for all of the patients, as measured by VAS and ODI (Table 2). According to the Frankel grading system, 65 patients had preoperative neurologic deficits with an incidence of 77.3% (32 in the CPS group and 33 in the HC group); two patients were identified as Frankel A, nine were Frankle B, 35 were Frankle C, and 19 were Frankel D (Table 3). After operation, the incidence of neurologic improvement was 68.8% (22/32) in the CPS group, with an average improvement of 0.5 ± 0.6 grade, and 75.8% (25/33) in the HC group, with an average improvement of 0.7 ± 0.5 (Tables 2, 4, and 5). No statistically significant difference was observed between the two groups in terms of neurologic improvement.

Table 3 - Neurologic status by Frankel classification before surgery and at final follow-up in both groups Final follow-up Preoperative A B C D E Total A 2 2 B 9 9 C 7 28 35 D 11 8 19 E 19 19 Total 18 39 27 84
Table 4 - Neurologic status by Frankel classification before surgery and at final follow-up in the CPS group Final follow-up Preoperative A B C D E Total A 1 1 B 5 5 C 4 13 17 D 6 3 9 E 11 11 Total 10 19 14 43