PATIENTS AND METHODS
The aim of this study was to investigate whether percutaneous vertebral augmentation (PVA) was associated with clinical and radiological subsequent adjacent fractures.
Patients and Methods:
A systematic review and meta-analysis was performed searching on PubMed, EMBASE, Cochrane library, Google Scholar, web of science and ClinicalTrial.gov from the establishment of the database to January 2020. Eligible studies assessing the subsequent adjacent fractures after PVA compared with conservative treatment (CT) were incorporated. The pooled risk ratio (RR) with its 95% confidence intervals (95% CI) was used. Heterogeneity, sensitivity and publication bias analyses were performed.
24 studies were considered eligible and were included finally. 20/421 patients (4.75%) had clinical subsequent adjacent fractures from PVA group, and 25/359 patients (6.96%) had from CT group, and 46/440 patients (10.45%) from PVA group and 36/444 patients (8.10%) from CT group had radiological subsequent adjacent fractures. There both had no significant difference between two groups (RR=0.67, 95%CI: [0.38, 1.19], P = 0.17)/ (RR=1.13, 95%CI: [0.75, 1.70], P = 0.576). However, in fractured vertebrae, number in PVA group was more than that in CT group (RR=1.41, 95%CI: [1.03, 1.93], P = 0.03).
Collectively, currently available literature provides data showed PVA did not increase the incidence for subsequent adjacent fractures, no matter it was clinical or radiological fracture. But PVA may increase the number of fractured vertebrae.
OVCF: osteoporotic vertebral compression fracture
PVA: percutaneous vertebral augmentation
PVP: percutaneous vertebroplasty
PKP: percutaneous kyphoplasty
CT: conservative treatment
RCT: randomized controlled trial
RR: risk ratio
CI: confidence interval
As one of the most common complications of osteoporosis, osteoporotic vertebral compression fractures (OVCFs) often results in back pain, spinal deformity, functional disability, and even death. So it has become one of the serious diseases threatening the health of elderly patients and increased the economic burden of society [1,2,5,7].
As a minimally invasive therapy for OVCFs, percutaneous vertebral augmentation (PVA) has shown promising and encouraging outcomes compared with conservative treatment (CT) [2, 3,4,6].
Moreover, according to different feature of fracture, PVA can choose percutaneous vertebroplasty (PVP), percutaneous kyphoplasty (PKP) or other operation methods. However, PVA can also lead to serious complications, the most serious of which is subsequent fracture, so the efficacy and safety of PVA are still in dispute [5,8,9,10]. The subsequent fractures can occur at adjacent, non-adjacent or even previously treated vertebral levels.
However, there were few meta-analyses [1,12,13,14,15,36] only to probe subsequent adjacent fractures, and RCTs as much as possible were not included in those reviews.
Furthermore, none of these studies distinguished clinical and radiological fracture, as well as the number of fractured patients and fractured vertebrae for analysis (16,17,18,19). The purpose of this study is to explore the characteristics of subsequent adjacent fracture after PVA, so as to provide evidence for the treatment strategy of OVCF (20,21, 28, 35).
2 MATERIAL AND METHODS:
Search strategy and study selection:
Two independent reviewers respectively conducted rough and accurate computerized retrieval in online databases, including PubMed, EMBASE, Cochrane library, Google Scholar, web of science and ClinicalTrial.gov, from the establishment of the database to January 2020.
We also searched references to selected literatures to avoid missing any additional research.
There are no language restrictions when searching (Fig. 1).
Rough search strategy: (vertebroplasty OR kyphoplasty OR vertebral augmentation) AND (conservative treatment) AND ((new fracture) OR (secondary fracture) OR (subsequent fracture) OR (adjacent fracture)).
Abbildung in dieser Leseprobe nicht enthalten
Only adult patients (age ≥ 50 years old) diagnosed with OVCF by clinical and imaging examination were included.
Intervention and control:
PVA (PVP/PKP) was performed in the experimental group and CT (including sham operation) was performed in the control group.
The incidence of subsequent adjacent vertebral fractures. Study type: Prospective cohort study, Non-RCT, and RCT.
Study selection and data extraction:
Endnote X9 software was used to check, sort and summarize the literatures; then each study was carefully read and selected by two independent reviewers by double-blind method. Any disagreement was resolved by discussion or by consulting a third reviewer.
The number of clinical and radiological subsequent adjacent fracture were separately extracted and classified.
If subsequent adjacent fracture did not have clear definition in the article, we deal with it as radiological fracture, because most fractures need imaging to be diagnosed. If patient had subsequent adjacent vertebral fractures equal to or more than two levels at once time, we just counted once for incidence.
Risk of bias assessment and quality evaluation:
Two independent reviewers applied the risk of bias tool to appraise all the included literatures according to the Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0), respectively. The methodological quality was assessed according to the Cochrane Collaboration’s domain-based evaluation framework [12, 13]. The main domains were assessed in the following sequence: (1) selection bias (randomized sequence generation and allocation concealment), (2) performance bias (blinding of participants and personnel), (3) detection bias (blinding of outcome assessment), (4) attrition bias (incomplete outcome data, e.g., due to dropouts), (5) reporting bias (selective reporting), and (6) other sources of bias. The score for each bias domain and the final score for the risk of systematic bias were graded as representing low, high, or unclear risk.
According to the Jadad scale 14, the quality of RCTs was evaluated, including the following four aspects: ① Generation of random sequence; ② Allocation concealment; ③ Implementation of blind method; ④ Description of case follow-up. “1-3” was considered as low quality, and “4-7” was considered as high quality.
To compare the differences from incidence for subsequent adjacent fractures after PVA, dichotomous data were calculated by risk ratio (RR) and its 95% confidence interval (95%CI). Heterogeneity was tested using the chi-squared statistic and the I2 statistic. If the P < 0.1, we defined the chi-squared statistic as statistically significant.
The I2 statistic was used to assess the variation across the included trails as the following standard: I2 < 25% means that heterogeneity is low; I2: 25-50% shows moderate heterogeneity; I2 > 50% demonstrates high heterogeneity. If I2 > 50%, a random-effect model would be adopted; otherwise, a fixed-effect model would be used 15.
Sensitivity analyses were conducted to investigate the impacts of each individual study by deleting them in turn on the overall meta-analysis results. Publication bias was detected using the method of Begg's and Egger's test. The statistical analysis was performed by Review Manager 5.3 and Stata 15.0.
Description of studies:
From the PRISMA Flow Diagram, the search and selection process of related literatures in this study were described. A total of 1128 literatures were retrieved, and 68 literatures were evaluated according to the inclusion criteria. Finally, 14 serial studies (total 24 literatures, including 5 serial non-RCTs [16-23] and 9 serial RCTs [24-38]) were selected.
Risk of bias and quality evaluation of included studies:
Because cement is opaque in imaging, it is difficult to blind patients, surgeons and observers, so only two of the serial studies (Control group had sham operation) were blinded to the patients. Six serial studies reported an adequate blinding for outcomes assessors. From the Jadad scale, eight serial studies [16, 20, 21, 24-35, 37, 38,41,43] were considered as high quality, and the others were considered as low quality.
The incidence of clinical subsequent adjacent fractures after PVA:
A total of 20/421 patients (4.75%) had clinical subsequent adjacent fractures from PVA group, and 25/359 patients (6.96%) had from CT group. There showed no significant difference between two groups (RR=0.67, 95%CI: [0.38, 1.19], P = 0.17. M-H. Fixed-effect model, I2= 31%).
The incidence of radiological subsequent adjacent fractures after PVA:
As far as radiological subsequent adjacent fractures were concerned, the results showed that 46/440 patients (10.45%) from PVA group and 36/444 patients (8.10%) from CT group had this complication form. There always had no significant difference between two groups (RR=1.13, 95%CI: [0.75, 1.70], P = 0.576. M-H. Fixed-effect model, I2= 0%).
The number of subsequent adjacent fractures for vertebrae after PVA:
In number of fractured vertebrae, 69/126 vertebral bodies (54.76%) had subsequent adjacent fractures from PVA group and 40/105 vertebral bodies (38.10%) had from CT group. There showed a significant difference between two groups (RR=1.41, 95%CI: [1.03, 1.93], P = 0.03. M-H. Fixed-effect model, I2= 0%).
Sensitivity and publication bias analysis:
Sensitivity analyses were conducted due to the discrepancy between studies. Each study was removed in turn to test whether the removed study would influence the overall effects. No specific trials could be determined as the main source of heterogeneity.
From the results of publication bias, the results of Begg's test (clinical fractures: P= 0.707> 0.05/ radiological fractures: P= 0.806> 0.05/ fractures for vertebrae: P=0.086> 0.05) and Egger's test (clinical fractures: P= 0.599> 0.05/ radiological fractures: P= 0.659> 0.05/ fractures for vertebrae: P= 0.061> 0.05) did not indicate the existence of publication bias.
With the advantage of pain relief rapidly, PVA, as a minimally invasive technique, has become the most popular treatment for OVCFs. However, PVA also has some complications and risks, such as cement leakage and subsequent fractures.
The incidence of cement leakage is high, but most of them are asymptomatic, so it is generally believed that the cement leakage is a phenomenon rather than a complication. But subsequent fracture is different.