Percutaneous vertebroplasty and percutaneous kyphoplasty for osteoporotic vertebral compression fractures: medium- and long-term evaluation of clinical outcomes and affecting factors

Abstract

Background

Although some studies have reported that patients with osteoporotic vertebral compression fractures (OVCFs) can benefit more from percutaneous kyphoplasty (PKP) than from percutaneous vertebroplasty (PVP), few large-sample, long-term follow-up studies have reported medium- to long-term clinical outcomes and the factors influencing the efficacy of PKP or PVP in the treatment of OVCFs.

Methods

In this retrospective study, 622 patients with OVCFs received PVP, and 758 received PKP. All patients were followed up for at least 3 years. We compared patient data, including age, sex, body mass index (BMI), bone mineral density (BMD), surgical time, blood loss, polymethylmethacrylate (PMMA) dosage, complications, hospital stay, visual analogue scale (VAS), and Oswestry Disability Index (ODI). The radiographic outcomes included local Cobb angle and anterior vertebral body height (AVH) changes. Multiple linear regression analysis was performed to identify risk factors associated with the ODI in both groups, such as BMD, fracture segments, fracture severity, PMMA dosage, PMMA leakage, vertebral recollapse, old vertebral fracture, local kyphosis, adjacent vertebral fracture, and antiosteoporosis treatment.

Results

The VAS and ODI improved significantly in both groups before discharge and at 1 year (P < 0.05) and then significantly increased at 3 years compared with those at 1 year post-operatively (P < 0.05), but the difference between the two groups was not significant (P˃0.05). The mean PMMA dosage in the PVP group was significantly lower than that in the PKP group (P < 0.05). The incidence rates of PMMA leakage and spinal impingement syndrome were greater in the PVP group than in the PKP group (P < 0.05). The incidences of local kyphosis, adjacent vertebral fracture and vertebral recollapse in the PVP group were all significantly lower than those in the PKP group (P < 0.05) at 3 years postoperatively. The postoperative local Cobb angle and AVH changes were significantly greater in the PKP group than in the PVP group (P < 0.05). Multiple linear regression analysis revealed that the ODI was significantly associated with BMI, BMD, fracture segments, local kyphosis, vertebral recollapse, and anti-osteoporosis treatment (P < 0.05) in both groups at the 3-year follow-up.

Conclusion

Both PVP and PKP are effective at treating OVCFs, and the medium- and long-term clinical effects are similar. PVP has higher PMMA leakage and spinal impingement syndrome rates. PKP is more effective at restoring the local Cobb angle and AVH, but it is also associated with a higher incidence of local kyphosis, adjacent vertebral fracture and vertebral recollapse. For either PVP or PKP, high BMI, low BMD, multiple vertebral fractures, local kyphosis, vertebral recollapse and a lack of standardized anti-osteoporosis treatment were the main risk factors for a poor long-term ODI.

Background

With the ageing of the population, osteoporosis has become a major public health concern. Osteoporotic vertebral compression fractures (OVCFs) are among the most common complications of osteoporosis and threaten the survival of older patients, posing a challenge for clinicians. In China, the incidence rate of osteoporosis among individuals aged over 65 years is 32.0%, with prevalence rates of 8.00% and 5.12% in males and females, respectively [1]. In accordance with the Chinese expert consensus [2], conservative treatment can be chosen for patients with mild symptoms or signs, no neurological damage, and imaging evidence of mild vertebral compression fractures. However, conservative treatment with long-term bed rest may increase the risk of developing pneumonia, bedsores, disuse muscle atrophy, urinary tract infections, anxiety, and depression, thereby seriously reducing patients’ quality of life. An epidemiological survey in the U.S. revealed that the 4-year mortality rate of patients with OVCFs is as high as 50% after conservative treatment [3]. Additionally, long-term bed rest can aggravate osteoporosis, which is unfavourable for bone healing [4]. Therefore, conservative treatment should be administered with caution in older patients with OVCFs. For patients with surgical indications who can tolerate the procedure, surgery should be performed as early as possible to restore mobility and promote overall recovery, in line with the current concept of enhanced recovery after surgery. Percutaneous vertebroplasty (PVP) and percutaneous kyphoplasty (PKP) are the preferred treatments for OVCFs, and their clinical efficacy and safety have been verified [3, 5–8].

Although some studies [3, 9, 10] have reported that patients with OVCFs can benefit more from PKP than from PVP because the balloon expansion step is designed to restore vertebral height and reduce kyphosis, few large-sample, long-term follow-up studies have reported medium- to long-term clinical outcomes and factors influencing the efficacy of PKP or PVP in treating OVCFs. Therefore, this single-centre, large-sample study with a minimum 3-year follow-up aimed to comprehensively evaluate the clinical and radiological outcomes and the important influencing factors of PVP and PKP for treating OVCFs in older patients.

Methods

Patients

This study was approved by the Ethics Committee of Suining Central Hospital, and informed consent was obtained from all participants. All procedures involving human participants were in accordance with the ethical standards of the institutional and national research committees, as well as the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

The study included 1380 consecutive patients who underwent PVP or PKP for OVCFs between January 2018 and August 2022. In total, 622 patients received PVP (WEGO Inc., Shandong, China; defined as the PVPF group), and 758 patients received PKP (WEGO Inc., Shandong, China; defined as the PKP group). The inclusion criteria were as follows: (1) osteoporosis (based on DXA and T scores ≤ − 2.5); (2) MRI revealing an acute or subacute fracture (bone marrow oedema with T1-weighted hypointensity and STIR hyperintensity); (3) a painful OVCF between the T5–L5 level before treatment; (4) a history of fracture caused by minor or no identifiable trauma; and (5) the availability of complete clinical data, with a minimum follow-up of 3 years. The exclusion criteria were as follows: (1) combined SCI and neurological symptoms, (2) spinal malignancies or neoplastic fractures, (3) previous fusion or internal fixation surgery, and (4) inability to complete the scheduled follow-up after treatment. After a preoperative discussion to finalize the surgical method, the surgeries were performed by two senior surgeons with more than 10 years of experience who had performed more than 1000 minimally invasive spine surgeries. PKP is recommended as the preferred surgical treatment for patients with posterior vertebral wall breakage, more severe vertebral compression or obvious kyphosis. PVP is the preferred surgical treatment for patients with simple vertebral fractures and poor economic conditions.

Surgical procedure

All PVP and PKP procedures were performed by two junior surgeons with standard training in the techniques. The PVP technique was adopted by using a bipedicular transpedicular approach in a prone posture under local anaesthesia (1% lidocaine) and C-arm guidance. During the operation, 4 mm puncture needles were inserted parallel or in a slightly descending course through the pedicle until the needle tip was optimally positioned in the posterior one-third of the injured vertebra. Afterwards, a bone drill was inserted to establish a working channel for bone cement injection. Bone cement (polymethylmethacrylate, PMMA) was subsequently injected slowly into the injured vertebra to maximize the stabilization of the fracture fragments.

In the PKP procedure, patients under local anaesthesia (1% lidocaine) were placed in the prone position. With C-arm guidance, the injured vertebra was marked, and a 4 mm puncture needle was inserted separately by the bilateral pedicle approach until it passed through the pedicle and reached the posterior one-third of the injured vertebra. A bone drill was subsequently inserted to establish a working channel for bone cement injection. After the working channel was established, a balloon dilator was placed, and an appropriate amount of contrast material was injected under X-ray fluoroscopy to expand the injured vertebra. After satisfactory balloon dilation was achieved, bone cement was mixed and injected into the vertebral body using a bolus injector and monitored using X-ray fluoroscopy.

All patients in both groups were transferred to the inpatient ward for further observation and were given guidance to walk off-bed under brace protection as early as 24 h postoperatively.

Evaluation of clinical outcomes and risk factors

Patients or their family members provided informed consent before surgery. All patients were followed up for at least 3 years through telephone interviews, outpatient visits, and readmission for vertebral refractures. Electronic case records, including age, sex, body mass index (BMI), preoperative bone mineral density (BMD), fracture segments, preoperative fracture severity, surgical time, blood loss, PMMA dosage, length of hospital stay and complications, were collected. Clinical outcomes were summarized, calculated and compared by an assigned clinician and medical records practitioner, and radiographic outcomes were collected and assessed by a senior radiological technician during follow-up. In total, 2093 patients who underwent PKP or PVP were included; due to a variety of reasons, such as spinal tumours, incomplete data, and death unrelated to surgery, 713 patients were excluded, and 1380 patients were ultimately included. Among them, 622 patients received PVP, and 758 patients received PKP (Fig. 1).

Fig. 1.

The flow chart for patient inclusion. SCI: spinal cord injury

Clinical outcomes were assessed using the visual analogue scale (VAS) and Oswestry Disability Index (ODI) before surgery, before discharge and at 1 year and 3 years after surgery [11, 12]. The primary recorded complications included PMMA leakage, spinal impingement syndromes (such as radiculopathy, muscle weakness, and spinal cord compression), local kyphosis, adjacent vertebral fracture and vertebral recollapse. The radiographic outcomes included the local Cobb angle and the anterior vertebral body height (AVH). The height of the anterior edge of the injured vertebra was measured by a picture archiving and communication system (PACS) using the standard lateral X-ray, and the Cobb angle was measured on the standard lateral X-ray of the fractured vertebra between the upper endplate of the vertebra above the injury and the lower endplate of the vertebra below the injury.

The main risk factors associated with the ODI included BMD, fracture segments, fracture severity, PMMA dosage, PMMA leakage, vertebral recollapse, old vertebral fracture, local kyphosis, adjacent vertebral fracture, and anti-osteoporosis treatment. The BMD T score was measured using dual-energy X-ray absorptiometry. PMMA leakage, fracture segments, fracture severity, vertebral recollapse, old vertebral fracture, local kyphosis, and adjacent vertebral fracture were assessed through imaging examinations. The PMMA dosage and anti-osteoporosis treatment were obtained from electronic case records. Fracture severity was divided into three degrees on the basis of the percentage of preoperative vertebral collapse according to previous reports [12]: first degree, mild (< 25% collapse); second degree, moderate (26–40%); and third degree, severe (> 40%). Local kyphosis was defined as ≥ 10° progression of the sagittal segmental Cobb angle between the immediate postoperative period and the final follow-up period [13, 14]. Vertebral recollapse was defined, according to prior studies [15–17], as either ≥ 15% progression of height loss or ≥ 10° progression of the local kyphotic angle between the immediate postoperative and final follow-up periods. Postoperative anti-osteoporosis treatment included long-term oral calcium (600 mg/day) and calciferol (0.5 µg/day) after discharge and regular use of alendronate (70 mg/week for 3 years), and long-term regular monitoring of blood calcium levels and renal function was performed during the treatment [1, 11].

Statistical analysis

Statistical analysis was performed using Statistical Product and Service Solutions (SPSS) software (version 20.0; IBM, Armonk, NY, USA). The measurement data are expressed as the means ± standard deviations (SDs), and the categorical data are presented as numbers and percentages. The chi-square test was used to analyse enumeration data, including sex, fracture segments, fracture severity, incidence of PMMA leakage, spinal impingement syndromes, local kyphosis, old vertebral fracture, adjacent vertebral fracture and vertebral recollapse and the proportion of patients who received antiosteoporosis treatment. A subgroup analysis was performed on the basis of clinically important confounders, including fracture segments, fracture severity, old vertebral fracture and anti-osteoporosis treatment adherence. Paired t tests were performed to compare differences in the measurement data, including age, BMI, VAS score, ODI, BMD, AVH, and Cobb angle, between the two groups. Normality testing was performed using the Kolmogorov–Smirnov test, and categorical variables were converted into dummy variables for multiple linear regression analysis. Multiple linear regression stepwise selection analysis was performed to examine the associations between factors such as sex, age, BMI, BMD, fracture segments, fracture severity, PMMA leakage, PMMA dosage, old vertebral fracture, local kyphosis, adjacent vertebral fracture, vertebral recollapse, and anti-osteoporosis treatment and the final follow-up ODI. The regression coefficient characterizes the extent of the influence of the correlated factors on the ODI. Negative values indicate a negative correlation, whereas positive values indicate a positive correlation. Statistical significance was set at P < 0.05.

Results

Comparison of surgical results

A total of 622 patients in the PVP group and 758 patients in the PKP group were included as study subjects, and all the patients completed a full 3 years of follow-up. The baseline characteristics of the two groups, including age, sex, BMI, preoperative ODI, preoperative VAS score, preoperative BMD, preoperative AVH and preoperative Cobb angle, were not significantly different (Table 1).

Table 1.

A comparison of the baseline information of the two groups

Groups	PVP	PKP	P-value
Male/female	163/ 459	188 /570	0.551
Age (years)	75.16 ± 8.51	74.56 ± 7.44	0.170
BMI	23.22 ± 1.99	23.33 ± 1.98	0.356
Pre VAS	7.06 ± 1.18	7.15 ± 1.06	0.109
Pre ODI (%)	63.20 ± 9.58	62.84 ± 7.96	0.436
Pre BMD	-4.38 ± 0.90	-4.47 ± 0.97	0.122
Pre AVH (mm)	21.98 ± 4.59	22.02 ± 4.84	0.895
Pre Cobb angle (°)	18.63 ± 4.38	19.09 ± 5.38	0.126

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, BMI body mass index, Pre Preoperative, VAS visual analog scale, ODI Oswestry disability index, BMD bone mineral density, AVH anterior vertebral height

As shown in Table 2, no statistically significant differences were noted between the two groups with respect to blood loss (PVP group: 10.06 ± 8.85 ml; PKP group: 11.03 ± 9.67 ml; P = 0.084) or hospital stay (PVP group: 8.21 ± 4.59 days; PKP group: 8.42 ± 4.70 days; P = 0.407). Patients in the PVP group had a shorter operation time (PVP group: 42.10 ± 11.66 min; PKP group: 47.69 ± 15.14 min; P < 0.001) and a lower volume of PMMA (PVP group: 3.49 ± 0.71 ml; PKP group: 4.06 ± 0.67 ml; P < 0.001). The incidence of PMMA leakage was 42.44% (264/622) in the PVP group and 36.02% (273/758) in the PKP group, which was significantly different (P < 0.05).

Table 2.

A comparison of surgical results between the two groups

Groups	PVP	PKP	P-value
Operating time (min)	42.10 ± 11.66	47.69 ± 15.14	< 0.001
Blood loss (ml)	10.06 ± 8.85	11.03 ± 9.67	0.084
Hospital stay (days)	8.21 ± 4.59	8.42 ± 4.70	0.407
PMMA dosage (ml)	3.49 ± 0.71	4.06 ± 0.67	< 0.001
PMMA leakage	264/622	273/758	0.015
Spinal impingement syndromes	64/622	52/758	0.025
Local kyphosis	239/622	337/758	0.024
Adjacent vertebral fracture	67/622	111/758	0.033
Vertebral recollapse	38/622	74/758	0.013

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, PMMA polymethylmethacrylate

The incidence of spinal impingement syndrome was 10.29% (64/622) in the PVP group and 6.86% (52/758) in the PKP group, which was significantly different (P < 0.05). Spinal impingement syndrome in the PVP group included 38 cases of radiculopathy, 26 cases of muscle weakness, and no cases of spinal cord compression. By comparison, there were 29 cases of radiculopathy, 22 cases of muscle weakness, and 1 case of spinal cord compression in the PKP group. All patients with radiculopathy or muscle weakness received conservative treatment. Most patients with radiculopathy usually manifested mild or moderate nerve irritation, and the symptoms gradually disappeared after corticosteroid therapy and neurotrophic therapy. However, 8 patients (5 in the PVP group and 3 in the PKP group) still had residual symptoms of numbness and tingling. All patients with muscle weakness experienced only mild symptoms and were cured after receiving corticosteroid therapy, neurotrophic therapy and functional exercises. Spinal impingement syndrome occurred postoperatively in one patient in the PKP group because of PMMA leakage, and this patient experienced spinal cord compression and complete nerve paralysis and underwent laminectomy (Fig. 2).

Fig. 2.

A patient suffered from spinal impingement syndrome after PKP. A-B Lumbar spine anteroposterior and lateral radiographs after primary PKP, which showed intraspinal PMMA leakage (white arrow). C-D CT and MRI scan after laminectomy, which showed the decompressive efficacy of the spinal cord (red arrow)

The incidence rates of local kyphosis, adjacent vertebral fracture and vertebral recollapse in the PVP group (38.42%, 239/622; 10.77%, 67/622; and 6.11%, 38/622; respectively) were significantly lower than those in the PKP group (44.46%, 337/758; 14.64%, 111/758; and 9.76%, 74/758, respectively; P < 0.05) at 3 years postoperatively. Notably, 60 patients in the PVP group and 102 patients in the PKP group were willing to undergo reoperation, and all patients with vertebral recollapse received conservative treatment. Figure 3 shows a typical case of adjacent vertebral fracture after PVP, and Fig. 4 shows a typical case of vertebral recollapse. As shown in Table 3, no statistically significant differences were noted between the two groups with respect to fracture segments (PVP group: single segment 433/622, two segments 146/622, three segments 35/622, and multiple segments 8/622; PKP group: single segment 502/758, two segments 192/758, three segments 51/622, and multiple segments 13/758; P˃0.05). No statistically significant differences in fracture severity were detected between the two groups (PVP group: mild 284/622, moderate 149/622, and severe 189/622; PKP group: mild 316/758, moderate 175/758, and severe 267/622; P˃0.05). Among the patients, compared with 363 in the PKP group, 259 in the PVP group had older vertebral fractures (P˃0.05), and 347 in the PVP group accepted standardized anti-osteoporosis treatment, whereas 415 in the PKP group did (P˃0.05).

Fig. 3.

A patient experienced an adjacent vertebral fracture after PVP. A-B Lumbar spine anteroposterior and lateral radiographs before primary PVP. C Thoracolumbar MRI before primary PVP. D-E Lumbar spine anteroposterior and lateral radiographs after primary PVP. F Thoracic MRI revealed the adjacent vertebral fracture after primary PVP. G-H Lumbar spine anteroposterior and lateral radiographs after secondary PVP. The white arrow indicates the previous T12 fracture and red arrow indicates the adjacent L1 fracture

Fig. 4.

A patient experienced vertebral recollapse after PKP. Preoperative lumbar spine anteroposterior and lateral radiographs (A-B) showed an OVCF in the 12th thoracic vertebra with a kyphosis angle of 23° (white arrow). Lumbar lateral radiograph (C) showed that the kyphosis angle was 22° in the thoracic-lumbar spine 1 week after PKP. The kyphosis angle increased from 22° to 34° at 1 month post-surgery (D), increased to 47° at 3 months post-surgery (E), and increased to 53° at 12 months post-surgery (F)

Table 3.

Subgroup analysis for four clinically important confounders

Evaluation factor		PVP	PKP	P value
Fracture segment	Single-segment	433	502	0.180
	Two-segment	146	192	0.452
	Three-segment	35	51	0.400
	Multiple-segment	8	13	0.517
Fracture severity	Mild	284	316	0.139
	Moderate	149	175	0.705
	Severe	189	267	0.057
Old vertebral fracture	Yes	259	327	0.575
	None	363	431
Anti-osteoporosis treatment	Yes	347	415	0.700
	None	275	343

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, PMMA polymethylmethacrylate

VSA and ODI scores

The VAS and ODI were used to evaluate postoperative functional outcomes. As shown in Table 4, the VAS in the PVP group decreased from 7.06 ± 1.18 preoperatively to 3.32 ± 0.7 before discharge (P < 0.001), 1.99 ± 0.88 at 1 year (P < 0.001), and 2.71 ± 0.92 at 3 years postoperatively (P < 0.001). Similarly, the VAS decreased from 7.15 ± 1.06 preoperatively to 3.34 ± 0.78 before discharge (P < 0.001), 2.04 ± 0.83 at 1 year (P < 0.001), and 2.66 ± 0.96 at 3 years postoperatively (P < 0.001) in the PKP group, but the differences between the two groups were not significant (P˃0.05). As shown in Table 5, compared with that before the operation, the ODI significantly improved in both groups before discharge (PVP group: 33.03 ± 6.38; PKP group: 32.55 ± 5.35; P < 0.001) and at 1 year after discharge (PVP group: 18.71 ± 4.54; PKP group: 18.40 ± 3.96; P < 0.001) and at 3 years after discharge (PVP group: 29.56 ± 8.42; PKP group: 29.07 ± 8.86; P < 0.001). However, the VAS and ODI in both groups significantly increased at 3 years compared with those at 1 year postoperatively (P < 0.05). These results suggest that the short-term effect after PVP or PKP is satisfactory, but the medium- and long-term clinical outcomes are not optimal.

Table 4.

A comparison of VAS between the two groups

Groups	Pre VAS	VAS before discharge	VAS after 1 Year	VAS after 3 Years
PVP	7.06 ± 1.18	3.32 ± 0.75*	1.99 ± 0.88*	2.71 ± 0.92**
PKP	7.15 ± 1.06	3.34 ± 0.78*	2.04 ± 0.83*	2.66 ± 0.96**
P-value	0.109	0.824	0.358	0.364

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, Pre Preoperative, VAS visual analog scale

* Significance compared with the preoperative, P < 0.001

** Significance compared with the preoperative and 1-year postoperativer, P < 0.001

Table 5.

A comparison of ODI(%) between the two groups

Groups	Pre ODI	ODI before discharge	ODI after 1 Year	ODI after 3 Years
PVP	63.20 ± 9.58	33.03 ± 6.38*	18.71 ± 4.54*	29.56 ± 8.42**
PKP	62.84 ± 7.96	32.55 ± 5.35*	18.40 ± 3.96*	29.07 ± 8.86**
P-value	0.436	0.143	0.193	0.314

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, Pre Preoperative, ODI Oswestry disability index

* Significance compared with the preoperative, P < 0.001

** Significance compared with the preoperative and 1-year postoperativer, P < 0.001

Postoperative imaging evaluation

Imaging evaluation revealed preoperative and postoperative changes in the AVH and Cobb angle. As shown in Table 6, the Cobb angle decreased from a mean of 18.63° preoperatively to 16.58° postoperatively in the PVP group and decreased from a mean of 19.09° preoperatively to 14.55° postoperatively in the PKP group. Compared with that in the PVP group, the improvement in the Cobb angle was significantly better in the PKP group immediately after surgery (P < 0.05). Notably, the change in the Cobb angle was greater in the PKP group at the final follow-up. As shown in Table 7, the change in the AVH immediately after surgery was 0.71 ± 0.61 mm in the PVP group and 1.89 ± 0.90 mm in the PKP group, and the AVH recovery was better in the PKP group (P < 0.05). The changes in AVH did not significantly differ between the groups at the final follow-up (P = 0.065). These results suggest that the vertebral height and local kyphosis can be better restored by PKP immediately after surgery; however, this difference disappeared gradually over time.

Table 6.

Cobb angle (°) changes between the two groups

Groups	Pre Cobb angle	Post Cobb angle	F/U Cobb angle	Post - Pre	F/U - Post
PVP	18.63 ± 4.38	16.58 ± 4.26	18.21 ± 3.90	− 2.05 ± 0.71	1.62 ± 1.19
PKP	19.09 ± 5.38	14.55 ± 5.45	16.88 ± 5.28	− 4.53 ± 0.62	2.33 ± 0.57
P-value	0.126	< 0.001	< 0.001	< 0.001	< 0.001

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, Pre Preoperative, Post postoperative, F/U final follow-up (3 years after surgery)

Table 7.

AVH (mm) changes between the two groups

Groups	Pre AVH	Post AVH	F/U AVH	Post - Pre	F/U - Post
PVP	21.98 ± 4.59	22.69 ± 4.62	22.02 ± 4.54	0.71 ± 0.61	− 0.67 ± 0.94
PKP	22.02 ± 4.84	23.94 ± 4.70	23.16 ± 4.79	1.89 ± 0.90	− 0.74 ± 0.22
P-value	0.895	< 0.001	< 0.001	< 0.001	0.065

PVP percutaneous vertebroplasty, PKP percutaneous kyphoplasty, AVH anterior vertebral height, Pre Preoperative, Post postoperative, F/U final follow-up (3 years after surgery)

Multiple linear regression analysis of risk factors for the ODI

Multiple linear regression analysis revealed that at the 3-year follow-up, the ODI was significantly associated with BMI, BMD, fracture segment, local kyphosis, vertebral recollapse and anti-osteoporosis treatment (P < 0.05; correlation coefficients: 0.359, − 0.912, 3.856, 1.008, 4.270, and − 5.925, respectively; Table 8) in the PVP group. Similarly, the ODI scores were significantly associated with BMI, BMD, fracture segments, local kyphosis, vertebral recollapse and anti-osteoporosis treatment (P < 0.05; correlation coefficients: 0.144, − 1.678, 0.509, 2.236, 5.329, and − 1.621, respectively; Table 9) in the PKP group. These results indicated that patients who have a higher BMI, lower BMD, multiple vertebral fractures, local kyphosis, vertebral recollapse or no anti-osteoporosis treatment often have a poor ODI after PVP or PKP.

Table 8.

Multiple linear regression analysis of the risk factors affecting the ODI 3 years after PVP (n = 622)

Evaluation factors	Regression coefficient	P value	VIF	95% CI
				Lower-bound	Upper-bound
Gender	-1.165	0.114	1.035	-2.609	0.279
Age (years)	− 0.036	0.152	1.027	− 0.086	0.013
BMI	0.359	0.001	1.134	0.142	0.576
BMD (T score)	− 0.912	< 0.001	1.072	− 1.406	− 0.418
Fracture segments	3.856	< 0.001	1.471	3.210	4.501
Fracture severity	− 0.194	0.537	1.032	− 0.809	0.422
PMMA leakage	− 0.429	0.324	1.010	− 1.281	0.424
PMMA dosage(ml)	− 0.072	0.825	1.013	− 0.718	0.573
Old vertebral fracture	− 0.839	0.055	1.038	− 1.695	0.017
Local kyphosis	1.008	0.023	1.071	0.138	1.879
Adjacent vertebral fracture	− 0.068	0.259	1.056	− 0.421	0.943
Vertebrae recollapse	4.270	< 0.001	1.338	2.911	5.629
Anti-osteoporosis treatment	− 5.925	< 0.001	1.250	− 6.880	− 4.970

Adjusted R² = 0.665

BMI body mass index, BMD bone mineral density, PMMA polymethylmethacrylate, VIF variance inflation factor

Table 9.

Multiple linear regression analysis of the risk factors affecting the ODI 3 years after PKP (n = 758)

Evaluation factors	Regression coefficient	P value	VIF	95% CI
				Lower-bound	Upper-bound
Gender	− 0.347	0.371	1.191	− 1.107	0.413
Age (years)	0.013	0.503	1.099	− 0.025	0.050
BMI	0.144	0.020	1.043	0.023	0.265
BMD (T score)	− 1.678	< 0.001	1.443	− 1.984	− 1.373
Fracture segments	0.509	0.028	1.752	0.054	0.964
Fracture severity	0.025	0.893	1.540	− 0.334	0.384
PMMA leakage	0.097	0.739	1.078	− 0.472	0.665
PMMA dosage(ml)	0.123	0.532	1.023	− 0.263	0.509
Old vertebral fracture	− 0.037	0.911	1.505	− 0.692	0.617
Local kyphosis	2.236	< 0.001	1.586	1.538	2.934
Adjacent vertebral fracture	− 0.093	0.818	1.522	− 0.887	0.701
Vertebrae recollapse	5.329	< 0.001	1.052	4.453	6.205
Anti-osteoporosis treatment	− 1.621	< 0.001	1.889	− 2.350	-0.892

Adjusted R² = 0.516

BMI body mass index, BMD bone mineral density, PMMA polymethylmethacrylate, VIF variance inflation factor

Discussion

PVP and PKP are the two types of minimally invasive surgical techniques extensively applied in patients with OVCFs. However, large-sample, long-term follow-up studies that systematically compare the clinical effects of these two surgical methods for the treatment of OVCFs are unavailable. Here, we compare the medium- to long-term clinical outcomes and factors influencing the efficacy of PKP and PVP in the treatment of OVCFs.

In terms of surgical findings, the operation time and PMMA dosage in the PVP group were significantly shorter than those in the PKP group, which is similar to the findings in previous studies [9, 10]. This was primarily attributed to the use of a balloon in PKP. In addition, the incidence of PMMA leakage in the PVP group (42.44%, 264/622) was significantly greater than that in the PKP group (36.02%, 273/758). The reported leakage rate of PMMA ranged from 4.8 to 39% [18]. The evaluation of PMMA leakage was based on postoperative CT scans in our study, and leaks were therefore identified more commonly on CT scans than on plain radiographs. PMMA leakage can occur through various routes, including paravertebral, venous, intraspinal, and intervertebral foramen leakage. Among these, intraspinal or intervertebral foramen leakage can damage neural, spinal cord, or nerve root tissues. In line with previous research results [19], the incidence of spinal impingement syndrome was 10.29% in the PVP group and 6.86% in the PKP group, which was significantly different (P < 0.05). Most patients with spinal impingement syndrome usually manifested mild or moderate radiculopathy without surgical intervention. Notably, severe complications occurred postoperatively in one patient in the PKP group. This patient experienced spinal cord compression and complete nerve paralysis and underwent laminectomy. Adjacent vertebral fracture is a common mid- to long-term complication. During the follow-up period, 63 patients (8.31%) were readmitted for adjacent vertebral fracture within 6 months, and a total of 111 patients (14.64%) were readmitted within 3 years in the PKP group. In comparison, 32 patients (5.14%) were readmitted within 6 months, and a total of 67 patients (10.77%) were readmitted within 3 years in the PVP group. The overall incidence of adjacent vertebral fracture in both groups is consistent with previous findings [20]. Adjacent vertebral fracture results from multiple factors, such as a lack of standardized postoperative anti-osteoporosis treatment, unimproved osteoporosis in adjacent vertebrae, and poor dispersion of PMMA [21]. Our results demonstrated that the incidence of early adjacent vertebral fracture and overall incidence in the PVP group were significantly lower than those in the PKP group, which was more likely to be biomechanically related to the augmented segment. The implanted balloon compresses the trabecular bone, resulting in the disruption of its continuity and integrity. Additionally, more cement was injected into the fractured vertebra, resulting in higher rigidity intensity in the PKP procedure, which may increase the risk of stress-induced damage to adjacent vertebra. A systematic review and meta-analysis by Song revealed that reduced vertebral strength due to osteoporosis was a reason for secondary vertebral fractures [22]. However, our research revealed that there were more patients (343, 45.25%) in the PKP group who did not receive standardized anti-osteoporosis treatment than in the PVP group (275, 44.21%), and the preoperative BMD in the PKP group (− 4.47 ± 0.97) was lower than that in the PVP group (− 4.38 ± 0.90). Although there was no significant difference in the preoperative BMD or the proportion of patients without anti-osteoporosis treatment between the two groups, lower BMD and poor adherence to anti-osteoporosis treatment may increase the incidence of adjacent vertebral fracture after PKP. These results suggest that the main preventive measure for adjacent vertebral fracture is standardized postoperative anti-osteoporosis treatment. Vertebral recollapse is another complication that frequently occurs following vertebral augmentation. In our study, the incidence of vertebral recollapse in the PVP group was 6.11% (38/622) and was 9.76% (111/758) in the PKP group, resulting in a significant difference (P < 0.05). The overall incidence in both groups is consistent with that in a previous report [23] and lower than that reported by Hou [17] and Nicholas [19]. The potential risk factors for vertebral recollapse include low BMD, preexisting intravertebral vacuum cleft and local kyphosis, a cleft filling pattern of PMMA, and a lack of standardized postoperative anti-osteoporosis treatment [24–26]. Notably, the incidence of adjacent vertebral fracture and vertebral recollapse was significantly lower in the PVP group than in the PKP group during follow-up. Studies have shown that a larger volume of cement used to maintain sagittal alignment after balloon kyphoplasty may actually increase the risk of new adjacent-level fractures because of the increased rigidity of the treated vertebral body adjacent to an already osteoporotic vertebra [27, 28]. Asymmetrical cement distribution around the fractured area was assumed to be the main risk factor for vertebral recollapse [29]. The implanted balloon compresses the trabecular bone, resulting in the disruption of its continuity and integrity. Additionally, the expansion of the balloon compresses the cancellous bone to form a dense cortical wall and hinders the diffusion of the bone cement, and the distribution of the bone cement in clumps leads to damage to the surrounding cancellous bone. There were 275 (44.21%) patients who did not receive standardized anti-osteoporosis treatment in the PVP group, compared with 343 (45.25%) in the PKP group. However, the preoperative BMD was − 4.38 ± 0.90 in the PVP group and − 4.47 ± 0.97 in the PKP group. Although there was no significant difference in preoperative BMD or the proportion of patients without anti-osteoporosis treatment between the two groups, lower BMD and poor adherence to anti-osteoporosis treatment may increase the incidence of local kyphosis, adjacent vertebral fracture, and vertebral recollapse after PKP. These factors should be considered when choosing PKP to treat OVCFs to help prevent vertebral recollapse, especially for patients with severe osteoporosis. Moreover, we found that patients with vertebral recollapse experienced varying degrees of pain and dysfunction during follow-up, and the mean VAS score and ODI were significantly greater in these patients than in those without vertebral recollapse. Although there were no cases of symptomatic pulmonary embolism associated with PMMA leakage, there may have been subclinical embolisms. The correlation between this type of leak and pulmonary embolism, whether subclinical or clinical, requires further study.

The VAS and ODI clearly improved in both groups after surgery, but the degree of improvement did not significantly differ at 1 year or 3 years after surgery. However, compared with those at 1 year postoperatively, the VAS score and ODI in both groups significantly increased at 3 years postoperatively (P < 0.05). These results suggest that the short-term effect after PVP or PKP is satisfactory, but the medium- and long-term clinical outcomes are not optimal. The rebound effect of the VAS score and ODI is related mainly to the aggravation of osteoporosis, vertebral recollapse, adjacent vertebral fracture and local kyphosis [30, 31]. In the risk factor analysis of the ODI, the following factors were identified in both groups: higher BMI, lower BMD, multiple vertebral fractures, local kyphosis, and vertebral recollapse without anti-osteoporosis treatment. The association between lower BMD and pain aligns with the findings of previous studies, which revealed that osteoporosis is an independent risk factor for adjacent vertebral fracture and vertebral recollapse after vertebral augmentation [22, 32]. Additionally, secondary vertebral fractures or recollapse of the augmented vertebrae make patients with osteoporosis prone to kyphosis and secondary sagittal imbalance [24]. The resulting sagittal decompensation of the spine leads to back muscle strain, persistent lumbago, and back pain [33]. A high BMI was identified as a risk factor for a poor ODI. Most patients with OVCFs in our study were postmenopausal women who were often overweight and had a relatively high BMI. Reduced physical exercise and potential hormone misuse, which is common in this population, could lead to a gradual decline in BMD and increased susceptibility to OVCFs. Recent evidence suggests that high BMI may be a risk factor for decreased bone density and fractures [34]. Previous investigations have demonstrated significant relationships between health-related quality of life and BMD [35, 36]. Consistent with these findings, our study revealed that lower BMD influenced the ODI at the final follow-up. Multiple vertebral fractures were identified as independent risk factors for the ODI. BMD is an independent risk factor for multiple segmental OVCFs [37], likely because patients with multiple vertebral fractures have more severe osteoporosis, and having multiple vertebral fractures during the first OVCF is regarded as a risk factor for new OVCFs after percutaneous vertebral augmentation [38]. Rapid bone loss after a fracture occurs in patients with severe osteoporosis who are prone to sequential fractures at other sites [39], such as multiple segments of the vertebra, the femoral neck, and carpal bones, which may considerably affect quality of life and treatment outcomes.

In the present study, PKP improved the recovery of the local Cobb angle and anterior vertebral height postoperatively, and a previous study supported this finding [40]. The superior outcomes of PKP may be due to balloon expansion, which creates a cavity within the vertebral body, with a high PMMA dosage, restoring its height and improving both the compression rate and the Cobb angle, whereas PVP directly injects PMMA into the vertebra without this height restoration [41]. These results suggest that PKP is the preferred surgical treatment for patients with more severe vertebral compression or obvious kyphosis. However, we found that the changes in the Cobb angle and AVH were greater in the PKP group at the final follow-up. The local Cobb angle increased from 14.55° ± 5.45 postoperatively to 16.88° ± 5.28, and the AVH decreased from 23.94 ± 4.70 mm to 23.16 ± 4.79 mm at the final follow-up in the PKP group. In comparison, the local Cobb angle increased from 16.58° ± 4.26 postoperatively to 18.21° ± 3.90, and the AVH decreased from 22.69 ± 4.62 mm to 22.02 ± 4.54 mm at the final follow-up in the PVP group. This was primarily attributed to the higher incidence of adjacent vertebral fracture and vertebral recollapse after PKP [27, 28].

In our study, a lack of standardized anti-osteoporosis treatment was identified as a key factor associated with higher ODI scores in both groups at the final follow-up. In our study, postoperative anti-osteoporosis treatment included long-term oral calcium (600 mg/day) and calciferol (0.5 µg/day) after discharge and regular use of alendronate (70 mg/week for 3 years), and long-term regular monitoring of blood calcium levels and renal function was performed during the treatment. Frankly, some potential risks of overtreatment may exist. Unfortunately, we did not monitor the side effects of long-term bisphosphonate use well, which was a limitation of our study. Osteoporosis, the underlying pathology of OVCFs, is an irreversible manifestation of ageing. While PVP or PKP address fractures, systematic anti-osteoporotic treatment is essential and should be maintained throughout the disease course. First, patient education is beneficial and has a clinically relevant impact on osteoporosis management results [42, 43]. Second, the basic therapeutic measures in this study included maintaining a healthy lifestyle, consuming a balanced diet rich in vitamin D and calcium with low salt and moderate protein, engaging in appropriate exercise; avoiding smoking and alcohol abuse, and cautious use of drugs that affect bone metabolism [11, 44]. Recent studies on osteoporosis treatment have reported notable advances, particularly in the use of bone absorption inhibitors and bone formation-acceleration drugs. In accordance with the American Association of Clinical Endocrinology Clinical Practice Guidelines for the Diagnosis and Treatment of Postmenopausal Osteoporosis (2020 update) [45], patients at extremely high risk of fracture are recommended to receive injectable drugs such as zoledronic acid, denosumab, and teriparatide as initial treatment, whereas those at high risk of fracture are recommended to receive alendronic acid, denosumab, risperidone, and zoledronic acid as initial treatment drugs. However, no clear synergy has been demonstrated between combination therapy and bone formation-accelerating drugs or antiresorptive agents [46].

This study is the first to retrospectively analyse the medium- to long-term clinical and radiological outcomes and risk factors associated with PVP and PKP in the treatment of OVCFs in older patients. Despite its valuable insights, this study has several limitations. First, its retrospective design introduced potential selection bias in the inclusion of case samples. General information was obtained from hospital medical records, which may have been inadequate, and the researchers’ selection of case samples could have introduced bias. Additionally, differences in surgical techniques and operator experience may have affected data comparability. Third, we did not analyse the effects of cement filling characteristics on clinical and radiological outcomes or the side effects of long-term bisphosphonate use for vertebral healing, which was another limitation of our study. Finally, factors such as the health status of patients, family economic condition, frailty score, malignancy status, and use of steroids might influence postoperative functional recovery after PVP or PKP. Although these factors were not considered in this study because of time constraints, they should be addressed in future research.

Conclusion

Both PVP and PKP are effective for treating OVCFs, and the medium- and long-term clinical effects of PVP and PKP are similar. PVP has higher PMMA leakage and spinal impingement syndrome rates. PKP is more effective at restoring the local Cobb angle and AVH, but it is also associated with a higher incidence of local kyphosis, adjacent vertebral fracture and vertebral recollapse. For either PVP or PKP, high BMI, low BMD, multiple vertebral fractures, local kyphosis, vertebral recollapse and a lack of standardized anti-osteoporosis treatment were the main risk factors for a poor long-term ODI.

Abbreviations

OVCF

Osteoporotic vertebral compression fracture

PKP

Percutaneous kyphoplasty

PVP

Percutaneous vertebroplasty

BMI

Body mass index

BMD

Bone mineral density

PMMA

Polymethylmethacrylate

VAS

Visual analogue scale

ODI

Oswestry disability index

AVH

Anterior vertebral body height

DXA

Dual-energy X-ray absorptiometry

MRI

Magnetic resonance imaging

SCI

Spinal cord injury

PACS

Picture archiving and communication system

CT

Computed tomography

VIF

Variance inflation factor

Authors’ contributions

Zhaojun Song: Conceptualization, funding acquisition, validation, methodology, formal analysis and writing - original draft. Juan Luo: Investigation and data curation. Hailong Zheng: Formal analysis and investigation. Maobo Ran: Visualization and investigation. Kai Zhang and Zhiqiang Wang: Supervision, project administration and funding acquisition. All authors contributed to the manuscript and read and approved the final manuscript.

Funding

Health Commission of Sichuan Province (19PJ418).

Data availability

Data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This retrospective cohort study was approved by the Ethics Committee of the Suining Central Hospital (KYLLKS20240116), and informed consent was obtained from all participants. All procedures involving human participants were in accordance with the ethical standards of the institutional and national research committees, as well as the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.