Risk factors for refracture or new vertebral compression fractures after percutaneous vertebroplasty: a systematic review and meta-analysis

Abstract

Background

Osteoporotic vertebral compression fractures can substantially affect a patient's life. Percutaneous vertebroplasty is a minimally invasive surgical procedure that can stabilize fractures, reduce pain, and increase the patient's functionality. However, the incidence of secondary fractures post-surgery requires scientific attention. The risk factors that may lead to a new fracture postoperatively must be identified.

Purpose

To explore the possible risk factors that can predispose to a subsequent fracture after percutaneous vertebroplasty in patients with osteoporotic vertebral compression fractures.

Study Design

Systematic review and meta-analysis.

Methods

We systematically searched MEDLINE (PubMed), Embase (ScienceDirect), PEDro, and Cochrane databases from July 2013 through May 2024. The extracted data included patient and clinical characteristics, secondary fractures as the primary outcome, and prognostic factors. Eligible studies for synthesis included one group of patients who had suffered a new fracture after surgery and a group of patients who did not suffer any fracture after vertebroplasty (“non fracture group”). We used the software Medcalc and the web program Meta-Mar for all statistical analyses. SMDs were calculated for continuous data, and ORs were calculated for dichotomous data, both with 95% CI. To quantify statistical heterogeneity across studies, we used chi-squared tests and I2 statistics). The Newcastle–Ottawa Scale was used to assess the quality of studies.

Results

A total of 3.821 patients from nine studies were included. Median follow-up was 21 months, and overall secondary fracture incidence was 16.6%. The meta-analysis showed that female gender [OR = 1.62, 95% CI 1,26; 2.07, z = 3.79, P = 0.0001], low bone mineral density [SMD = -0,30, 95% CI -0.52; -0.08, z = -2.65, P = 0.008], smoking [OR = 1.62, 95% CI 1.22; 2.15, z = 3.32, P = 0.001] and type 2 diabetes [OR = 1.40, 95% CI 1.11; 1.76, z = 2.83, P = 0.005] were significant risk factors for the secondary fractures after percutaneous vertebroplasty.

Conclusion

Patients with these identified risk factors should be monitored and treated accordingly to achieve the optimum prognosis after surgery.

Introduction

Osteoporosis is the most common skeleton disease characterized by low bone mineral density and deterioration of bone tissue [1]. Because of its high impact it has been rightfully called"the silent epidemic of the twenty-first century"and affects all sections of society and both sexes. According to a recent systematic review, the prevalence of osteoporosis worldwide was reported to be 18.3% [2]. Osteoporotic vertebral fractures constitute one of the expected consequences of osteoporosis. Life expectancy has doubled over the last two centuries [3]. At the same time, the incidence of osteoporotic vertebral compression fractures (OVCF) is being increased, affecting more than 1.4 million people worldwide each year [4]. Patients are at high risk of fracture because of compromised bone strength [5, 6]. Osteoporotic vertebral fractures can cause severe pain and disability as well as increase the risk of subsequent fractures. They have also been associated with hyperkyphosis, malnutrition, impaired physical performance, sleep disturbance, depression, reduced quality of life, and increased mortality [5]. The osteoporotic fractures have made this disease the leading cause of death in the elderly [7].

Percutaneous vertebroplasty (PVP) is a minimally invasive spine surgery technique invented to treat spinal compression fractures [8]. This procedure, which involves the injection of low-viscosity acrylic bone cement into the vertebral body, promises early pain relief up to 93% through vertebral height restoration [9]. Study results about vertebroplasty are controversial. While PVP has demonstrated efficacy in pain reduction and functional improvement, concerns persist regarding the risk of subsequent vertebral fractures post-surgery. The reason is that some patients suffer subsequent fractures following vertebroplasty. These fractures can be at the same vertebral level, adjacent or remote to the cemented vertebrae [10–12].

Numerous risk factors have been linked to the refracture incidence [12–14]. These can be divided into four categories: The first is related to the patient history, characteristics, and physical condition and includes gender, age, bone mineral density (BMD), and body mass index (BMI). The second relates to the fracture and its biomechanical consequences, which constitute fracture level, kyphosis, number of baseline fractures, etc. The third refers to the surgical procedure and includes iatrogenic risk factors such as cement leakage, cement volume injected, and cement distribution. The fourth category is related to the patient's lifestyle and encompasses exercise, smoking, type 2 diabetes, and anti-osteoporotic medication. Understanding these risk factors is crucial for optimizing patient selection, procedural outcomes, and long-term management strategies to offer patients the best experience and minimize socioeconomic costs.

Unfortunately, current evidence has not yet given a clear picture of the risks associated with the subsequent fractures. Group synthesis and relevant terminology vary in current studies that use the term"refracture,"including all kinds of fractures (cemented, adjacent, or remote), or they focus only on cemented vertebrae fractures. Also, there needs to be more justification around the selection of risk factors investigated. Thus, there is an imperative need for a systematic review to provide a robust evaluation of the current evidence, select the most frequent risk factors of all categories, make a uniform synthesis of the groups, and finally elucidate the potential risk factors for subsequent fractures after PVP.

This systematic review aims to comprehensively evaluate the available literature regarding risk factors for subsequent fractures after percutaneous vertebroplasty. The innovative current review includes risk factors of all categories to offer a holistic approach for clinical decision-making that can be implicated by all interprofessional team members and lead to outcomes optimization through percutaneous vertebroplasty.

Methods

For the Systematic Review and Meta-analysis, the Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines were followed [15].

Study Design

Systematic review.

Eligibility criteria

Inclusion criteria

(1) Adults with osteoporotic compressive vertebral fractures that underwent vertebroplasty; (2) RCTs, cohort, observation studies, retrospective studies, and case–control studies; (3) Studies that investigated risk factors for refracture or new osteoporotic vertebral compression fracture adjacent or remote; (4) Studies had to include one group who suffered subsequent OVCF after PVP and a non fracture group.

Exclusion criteria

(1) Case series, reviews, letters, abstracts, conference meetings, and studies with fractures after repeated vertebroplasty; (2) studies that included kyphoplasty unless there was separate analysis for vertebroplasty group; (3) studies with fractures of other etiology (infection, inflammation, tumor or metastasis) or fractures not located at the spine; (4) patients who underwent PVP for malignancy hemangioma or high energy trauma; (5) patients with neurologic symptoms; (6) studies that did not examine risk factors for refracture or subsequent fractures(adjacent or remote); (7) studies without sufficiently published data to estimate a standardized mean difference (SMD) or odds ratio (OR) with a 95% confidence interval (CI); (8) studies with less than one year follow up; (9) studies that did not include comparable data.

Information sources

The available literature was searched for articles published from July 2013 through May 2024 to identify studies that examined prognostic factors on osteoporotic adults with vertebral fractures who underwent vertebroplasty. MEDLINE (PubMed), Embase (ScienceDirect), PEDro, and Cochrane were the electronic databases searched. Reference lists were searched for additional articles.

Search strategy

The key terms used were the following:"prognosis"OR"risk factors"AND"percutaneous vertebroplasty"OR"PVP"AND"osteoporotic vertebral compression fracture"OR"fractures"OR"refracture"OR"subsequent fractures"OR"adjacent fractures"OR"recollapse"AND"randomized controlled trial."Searches were restricted to human. Search strategy of this systematic review was restricted to studies that included only adult participants and studies published in English. Studies older than ten years were excluded.

The search strategy was based on the"PPO"approach, which stands for"Population,"Prognostic Factors,"and"Outcome.

Definition of the question

Risk factors

Factors that increase a person's chance of developing a disease or being exposed to a hazard.

Refracture

Fracture of the cemented vertebra(e).

New vertebral compression fracture

New fracture adjacent or remote to the cemented vertebra.

Percutaneous vertebroplasty

A minimally invasive surgical procedure for osteoporotic vertebral compression fractures where the collapsed vertebral body is stabilized with injection of bone cement. Depending on the number of fractured vertebrae, one or multiple levels can be restored during this procedure.

PPO

P (Population)

We included clinical trials of older adults and elderly patients with osteoporotic vertebral compression fracture/s who have undergone vertebroplasty.

P (Prognostic factors)

Risk factors described in each study that can lead to refracture or secondary new fractures. These factors can be related to patient characteristics (personal and lifestyle factors), condition pathology (fractures, kyphosis) and the vertebroplasty itself (cement leakage, volume).

O (Outcome)

Incidence of refracture or new fractures at adjacent or remote levels to the cemented vertebra.

Study selection process

Two reviewers (EM and GAK) screened titles and abstracts independently to determine eligibility. Then, the two reviewers independently reviewed the full text of the articles and decided on the final selection. Disagreements raised were resolved by consensus.

Data collection process and data items

The two reviewers developed a data extraction form based on Cochrane Consumers and Communication Review Group's data extraction template. They pilot-tested it on five randomly selected included studies and modified it accordingly. They extracted the data and then compared them. Any discrepancies were resolved through discussion.

The extracted data included study characteristics (e.g., study design, study region), participant characteristics (sample size, groups, mean age, gender), clinical characteristics (no of vertebrae treated, follow-up duration and timeframe of subsequent fractures), primary outcome and prognostic factors related to the primary outcome.

The primary outcome was the refracture of the cemented vertebra or subsequent fractures adjacent to or remote from the cemented vertebra.

The selection of the prognostic factors extracted from the eligible studies was based on the following criteria:

1. Cover all the categories of the risk factors’ spectrum (anthropometric, iatrogenic, fracture-related related, and lifestyle).

2. Report in a minimum of three studies.

Assessment Quality

The quality of the included studies was assessed independently by two reviewers using the Newcastle–Ottawa Scale (NOS) for cohort studies where cohort in the current meta-analysis were the patients who underwent vertebroplasty [16]. This quality assessment tool consists of three categories:"Selection,""Comparability,"and"Outcome."Comparability can be given a maximum of two stars, while the other two categories have a maximum of one star. The highest score of this tool is 9 points, and studies with scores ≥ 7 are considered high quality. Only high-quality studies were included in the meta-analysis. Disagreements were resolved through consultation.

Effect measures

To quantify the association between prognostic factors and new fractures, we calculated odds Ratios (ORs) for dichotomous data and Standardized Mean Differences (SMDs) for continuous data with a 95% CI for both. SMDs are particularly useful in meta-analyses where combining studies with different measurement scales is necessary, as they provide a standard metric for summarizing effect sizes. ORs estimate the odds of the outcome occurring in exposed individuals compared to unexposed individuals [17].

Synthesis methods

Eligible studies for synthesis included one group of patients who had suffered a new fracture after surgery and a group that did not suffer any fracture after PVP. The goal was to form a homogenous group of patients who underwent the same type of surgery. The risk factors selected for meta-analysis had to be reported in at least three studies.

All dichotomous data were presented as percentages in studies were transformed into numbers to allow comparison. All findings were visually presented in a forest plot for summary. A forest plot displays the results of individual studies on the same graphical axis, allowing for a quick comparison across studies and the visualization of the overall effect size.

All statistical analyses were conducted using the software Medcalc (MedCalc® Statistical Software version 20, MedCalc Software Ltd, Ostend, Belgium) and the web program Meta-Mar (Meta-Mar v 3.5.1 © Copyright 2018—Ashkan Beheshti).

SMDs were calculated for continuous data, and ORs were calculated for dichotomous data, both with 95%CI. We used chi-squared tests and squared I2 statistics to quantify statistical heterogeneity across studies. I2 > 50% indicated substantial heterogeneity. When there was clear evidence of heterogeneity, a random-effects model was employed to compute the combined odds ratios (ORs) or standardized mean differences (SMDs). In the absence of such heterogeneity, a fixed-effect model was utilized. Since the number of included studies was less than 10, this meta-analysis did not analyze publication bias.

Results

Study selection

The search process across MEDLINE (PubMed), Embase (Science Direct), PEDro, and Cochrane databases, initially identified 305 studies (Fig. 1). After removing duplicates, 296 studies were considered for inclusion. Subsequent screening based on titles and abstracts excluded 252 studies. Further examination of the full texts led to the exclusion of 35 studies for various reasons: 10 studies did not report the outcomes of interest; 6 studies lacked prognostic study design; 7 studies involved a combination of percutaneous kyphoplasty (PKP) and percutaneous vertebroplasty (PVP); 1 study had short follow up; and 11 studies were excluded due to inadequate data. Finally, nine studies met all criteria and were included in the final meta-analysis.

Fig. 1.

PRISMA flow diagram

Study characteristics & Quality Assessment of Studies

The meta-analysis included a total of 9 studies that investigated risk factors for new fractures after percutaneous vertebroplasty in people with osteoporotic vertebral compression fractures. The essential characteristics of the included studies are shown in Table 1. All studies were retrospective cohort studies except for one RCT and were published between 2013 and 2023. All studies were conducted in China except for one in South Korea and one in Spain. The combined sample size across all studies was 3.821, with individual study sample sizes ranging from 57 to 2202. Six studies included new fractures without differentiating if they were adjacent or remote; one included fracture in adjacent levels, and two included refracture of the cemented vertebrae. The average age of participants was 72.68 years, and the average percentage of females was 76.5%. Follow-up duration across studies ranged from 6 to 50 months, with a median follow-up of 21 months (Table 1).

Table 1.

Characteristics of included studies

AUTHOR(S), YEAR	FRACTURE GROUP	NON FRACTURE GROUP	MEAN AGE GENDER	SAMPLE	FOLLOW UP	PRIMARY OUTCOME MEASURE Incidence of new fracture	*NOS SCORE
1. Martinez‐Ferrer et al. 2013  RCT Spain	N=17 New fracture group	N=40	71.92 yrs 72.8% Female 27.2% Male	N=57 140 vertebrae, 61% of patients>2 vertebrae per procedure, Max 4 levels treated in a single session	6 −39 months	17 patients had 29 fractures Incidence: 30%	7
2. Zhang et al. 2023 Retrospective study China	N=47 New fracture group	N=232	71,11 yrs Female 81,7% Male18,3%	N=279	>6 m	47 patients had new fractures Incidence: 16,8% 232 did not have (83,2%)	7
3. Zhang et al., 2021 Retrospective study China	N=362 New fracture group	N=1840	66.06 yrs Female74,2 % Male 25,8%	N=2.202	1,2,3,6,12 Mean follow up 14,7 months.	362 patients had a new fracture 362 Incidence: 16.43% 1840 non fractured	8
4. Xiong et al., 2021 Retrospective cohort study China	N=48 Recompression group	N=45	78.97 yrs Female 80.6% Male :19.3%	N=93	>2 years 1.2–25.9 months Mean follow up 9.3 months	Refracture was found in 48 patients (51,6%) Incidence: 3.68%	8
5. Ren et al. 2015 Retrospective study China	N=21 New fracture group	N=161	69,7 yrs Female 85.2% Male 14.8%	N=182	24–50 months Mean follow up 26.4 months	Incidence: 11.5% 28 new fractures in 21 patients.	8
6. Li et al., 2023 China	N=27 Adjacent fracture group	N=76	74,16 yrs Female 75.7% Male 24.2%	N=103	Mean follow up:24,1 months	Incidence: 26.21%	7
7. Bae et al., 2017 Retrospective study South Korea	N=25 New fracture group	N=231	71,9 yrs Female 79,3 % Male: 20.7%	N=256	Mean follow up 36 months	25 sustained new fractures (9.8%) Incidence: 11.6%	7
8. Ju & Liu 2023, Retrospective study China	N=48 Recompression group	N=216	74,7 yrs Female 59,8% Male: 40.2%	N=264	1,6,12 months for the 1 st year, then once a year afterward	48 patients suffered refracture Incidence: 18.2%	7
9. Li et al., 2022 Retrospective study China	N=58 New fracture group	N=327	75,6 yrs Female 80% Male 20%	N=385	2 years	Incidence: 15%	8

The overall incidence of new fractures was 16.6%. In particular, fractures of the cemented vertebrae were 10.94%, adjacent vertebrae were 26.21%, and general new fractures were 16.8%. Fracture timeline incidence ranged from 1 to 33 months, most within 8 to 14 months (Table 2). Four studies included only single-level operated fractures, one study up to two levels, and three studies more than three levels (Table 3).

Table 2.

Fracture Incidence after PVP

1–4 m	8–14 m	18–33 m
Bae et al., 2017	Martinez‐Ferrer et al. 2013	Li et al., 2022
	Zhang et al. 2023	Li et al., 2023
	Zhang et al., 2021
	Xiong et al., 2021
	Ren et al. 2015
	Ju & Liu 2023

Table 3.

Levels operated in included studies

Single level	1–2 levels	Multilevel
Zhang et al., 2023	Zhang et al., 2021	Martinez‐Ferrer et al. 2013
Bae et al. 2017		Li et al., 2022
Xiong et al. 2021		Ren et al. 2015
Ju & Liu 2023

For the current meta-analysis, the following risk factors were selected: Age, gender, BMI (Body Mass Index), BMD (Bone Mineral Density), cement leakage, cement volume, preoperative kyphosis, smoking, and diabetes. All the included studies with all potential risk factors and their division into significant and insignificant ones are illustrated in Table 4. The quality of the nine studies was assessed with the NOS tool. Five studies scored 7 points, and four studies scored 8 points. The average score was 8. Scores are depicted in Table 1.

Table 4.

Potential risk factors in included studies

Groups	Study- Year	Insignificant	Significant
New fractures	Martinez- Ferrer et al. 2013	Gender, Lumbar T- score, Glucocorticoids, P1 NP, severity and type of VF at baseline, presence and location of cement leakage and bone remodeling	Age > 80 years, cement leakage, 25(OH)D levels and No of treated vertebrae
	Zhang et al. 2023	Age, sex, age, BMI, diabetes, hypertension, history of smoking, history of alcohol, fracture level, surgical method, bone cement leakage, bone cement volume, AVHR, and HU value	HU value
	Ren det al., 2015	Age, sex, bone mineral density, amount of bone cement leakage into the disk, preoperative kyphosis, pre-operative degree of anterior vertebral compression, pre-operative degree of middle vertebral compression, kyphosis correction, anterior vertebral height restoration and middle vertebral height restoration	BMI and number of initial symptomatic fractures (levels treated)
	Bae et al. 2017	Age, past medical history, bone cement injection, segmental kyphosis, and time interval between first and second fracture events	Bone mineral content, cement distribution and cement leakage
	Li et al. 2022	Age, sex, bone cement leakage, BMD, injection volume, time from injury to surgery, time from admission to surgery and duration of surgery	BMI, multiple level fractures, steroid use and anti-osteoporotic treatment
	Zhang et al. 2021	Height, weight preoperative bone mineral density, the number of vertebral body fractures, cement injection, bone cement leakage, lifestyle-related factors (smoking, drinking, and postoperative exercise), and presence of chronic diseases (hypertension, diabetes, coronary atherosclerosis, and chronic obstructive pulmonary disease), exercise compliance and antiosteoporotic treatment	Sex, age, smoking, diabetes, postoperative exercise and postoperative anti-osteoporotic medication
Refracture	Xiong et al. 2021	Age, sex, fracture location, BMD, KA, wedge angle, cement volume, surgical approach, cement leakage and other vertebral fractures	IVC, endplate cortical disruption, RR and non-PMMA-endplate-contact
	Ju & Liu, 2023	BMI, gender, smoking, and alcoholism status, Hx of fracture, preoperative comorbidities, and medications, operation time, cement leakage and injection volume, Hb, WBC, urea, Cr and CRP	Age, BMD, multiple vertebral fractures, Alb, Fi), postoperative regular anti-osteoporosis treatment, and exercise
Adjacent fractures	Li et al. 2023	Preoperative collapse height of the vertebral body, the segmental kyphotic angle, intervertebral cement leakage status, postoperative vertebral height restoration ratio, and kyphotic angle restoration, age, sex, BMI and smoking	HU values

P1NP: Total procollagen type 1 N-terminal propeptide; UH: Hounsfield unit; AVHR: Anterior vertebral height ratio; BMI: Body mass index; BMD: Bone mineral density; KA; Kyphotic angle; IVC: intravertebral cleft; RR: Reduction rate; NPEC: endplate-contact; Hx: History; WBC: White blood cell; Hb: hemoglobin; Cr: Creatinine; CRP:C-reactive protein: Alb: albumin; Fib: Fibrinogen.

Risk factors related to patient characteristics

Age, gender, BMI, and BMD were selected for meta-analysis in this category. Five studies involving 3105 patients explored age as a predictive risk factor for new fractures after PVP. A random-effect model was used due to the heterogeneity test (x² = 39,77, df = 4. p < 0.001, I² = 90%) (Fig. 2a). The SMD was 0.35, suggesting that the age of patients was not a risk factor for a new fracture after PVP [SMD = 0.35, 95% CI −0.09; 0.79, z = 2.2, P = 0.093].

Fig. 2.

Forest plots of meta-analysis for age, gender, body mass index(BMI), bone mineral density(BMD), preoperative kyphosis, cement leakage, cement volume, smoking and diabetes. a: Age; b: Gender; c: Body mass index; d: Bone mineral density, e: Preoperative kyphosis; f:Cement leakage; g: Cement volume; h: Smoking; i: T2D; CI: Confidence interval; SD: Standard deviation

Five studies involving 3124 patients examined gender as a risk factor for refracture. Due to the low heterogeneity test (x² = 4.3, df = 4. p = 0.363, I² = 8%), a fixed-effect model was used. The gender results exhibited significant differences between the fracture and nonfracture groups [OR = 1.62, 95% CI 1.26; 2.07, z = 3.79, P < 0.0001] (Fig. 2b). Female patients were more likely to develop a new fracture after PVP.

Three studies involving 846 patients reported the BMI of patients. A random-effect model was used due to the heterogeneity test (x² = 24.00, df = 2. p < 0.01, I² = 92%). The pooled SMD was 0.15, indicating that BMI was not a risk factor for subsequent fractures after PVP [SMD = 0.15, 95% CI −1.34; 1.63, z = 0.43, P = 0.709] (Fig. 2c). Three studies of 624 patients evaluated BMD as a predictive risk factor for new fractures after PVP. A fixed-effect model was used due to the heterogeneity test (x² = 1.10, df = 2. p = 0.578, I² = 0%).

The results of the BMD t-score exhibited significant differences between the fracture group and the nonfracture one, indications that BMD could be a risk factor for subsequent fractures after PVP in this kind of patients [SMD = −0,30, 95% CI −0.52; −0.08, z = −2.65, P = 0.008], hence lower BMD t-score is associated with higher likelihood of fracture after PVP. (Fig. 2d).

Risk factors related to biomechanical consequences of vertebral fractures

Preoperative kyphosis was the risk factor selected for meta-analysis in this category. Five studies involving 898 patients explored preoperative kyphosis as a risk factor for subsequent fractures after surgery. A random-effect model was used because of heterogeneity (x² = 9.07, df = 4. p = 0.059, I² = 56%). SMD was 0.03, suggesting that preoperative kyphosis was not a risk factor for refracture after PVP [SMD = 0,03, 95% CI −0.36; 0.42, z = 0.22, P = 0.848] (Fig. 2e).

Risk factors related to Percutaneous vertebroplasty

Cement leakage and cement injection volume were the surgery-related selected risk factors for meta-analysis. Five studies involving 3336 patients estimated the risk factor of cement leakage. The fixed-effect model used according to the heterogeneity test (x² = 4.42, df = 4. p = 0.352, I² = 9%) showed that there were no significant differences between the two groups [OR = 1.19, 95% CI 0.97;1.47, z = 1,67, P = 0.096] (Fig. 2f). Five studies of 3048 patients investigated the association between the injected cement volume and the risk of a new fracture after PVP. A fixed-effect model was used because heterogeneity was relatively low (x² = 1.922, df = 3. p = 0.589, I² = 0%). The pooled results showed that cement injected volume is not associated with the incidence of new fractures after PVP [SMD = 0.04, 95% CI −0.05; 0.14, z = 0.88, P = 0.382] (Fig. 2g).

Lifestyle risk factors

Smoking and type 2 diabetes (T2D) were the lifestyle risk factors selected for meta-analyses. Four studies containing 2848 patients were pooled to investigate smoking as a predictive risk factor for new fractures after PVP. A fixed-effect model was used because of low heterogeneity (x² = 0.37, df = 3. p = 0.950, I² = 0%). The OR was 1.62, suggesting that patients who smoked were more likely to suffer a new fracture after PVP compared to nonsmokers [OR = 1.62, 95% CI 1.22; 2.15, z = 3.32, P = 0.001] (Fig. 2h).

Three studies of 2745 patients were pooled to estimate the association of T2D to new fracture incidence after PVP. Heterogeneity was low, so a fixed-model test was used (x² = 2.24, df = 2 p = 0.330, I² = 11%). The OR was 1.40, estimating that patients with T2D are at high risk of having new fractures after surgery compared to subjects that do not have T2D [OR = 1.40, 95% CI 1.11; 1.76, z = 2.83, P = 0.005] (Fig. 2i).

Discussion

This is the first systematic review that investigated lifestyle risk factors and showed that female gender, low bone mineral density, smoking, and T2D can predispose to a new fracture after vertebroplasty. This knowledge can be invaluable because we can examine these risk factors, manage them accordingly if possible, and eliminate the fracture risk after surgery.

Female sex was found to be a risk factor for subsequent fractures after vertebroplasty. Postmenopausal women face a higher risk of experiencing a fragility fracture as a clinical consequence of osteoporosis because of bone loss caused by estrogen reduction compared to males of the same age. Women with postmenopausal osteoporosis have a higher risk of three to fourfold to suffer an osteoporotic vertebral fracture compared to men of the same age [18]. This risk pre-exists and remains after surgery. Obviously, it has to do with female physiology, not the operation itself. The physiology of female participants is apparently a personal risk factor, to be taken into consideration in subjects receiving VP, for the provision of additional pharmacological prescription in combination with other possible bone-retaining strategies, like particular types of exercise or physical activity [6, 19]. Female gender is a factor that should be considered since most patients who undergo this procedure are women. Although it is important to know the risk we cannot modify this factor to reduce it.

Our study identified low bone mineral density as an essential risk factor for secondary fractures after vertebroplasty. Meta-analysis of Zhu et al. [20] showed relevant results to our study. Interestingly, a positive relationship has been identified between low t-scores and refractures after vertebroplasty regardless of the location of the new fracture (adjacent, remote, or cemented vertebrae [12, 21, 22]. Consistent results were found by Zhai et al. [14], who found that low bone mass was a high-risk factor for postoperative secondary fractures. However, in the literature, secondary fractures have been attributed to the natural course of the disease rather than the operation. Vertebrae with lower t-scores are more likely to fracture. Decreasing the disease's progression seems essential and can be achieved with anti-osteoporotic medication [23]. Bouxsein et al. [24] concluded that improvements in BMD with osteoporosis therapies can decrease the fracture risk. In two studies included in the current meta-analysis, anti-osteoporosis treatment after the operation was proved to be a protective factor for secondary fractures, although we did not select this factor for meta-analysis because studies that included it were excluded based on set criteria. [25, 26]. Apart from medication, more physical modalities like exercise seem to be able to increase BMD scores. According to a Cochrane review, a multicomponent training exercise program was the most effective for BMD in menopausal women with osteoporosis [27]. It remains to be explored if exercise can increase BMD postoperatively in these patients.

In the current study, two lifestyle factors (smoking and T2D) were included as possible risk factors for future fractures post VP surgery. It was confirmed that both these factors were significant (OR = 1.62, 95% CI 1.22; 2.15, z = 3.32; P = 0.001; OR = 1.40, 95% CI 1.11; 1.76, z = 2.83, P = 0.005 respectively).

Tobacco smoking has more than 7.000 chemicals, which have been proven to have a harmful effect on the skeletal system [28]). According to a systematic review [29], smoking can reduce BMD and increase fracture risk.

There is high-quality evidence that tobacco affects the mechanisms of bone turn, rendering the bones more fragile, and eventually, the possibility of fracture [30]. It has been proved that bony mass deterioration driven by tobacco results from multiple mechanisms that affect bone angiogenesis, osteogenesis, bone metabolism, and body hormones and increase oxidative stress on bony tissues [29]. Secondhand smokers, especially females, seem to have similar harm to their skeletal system [31]. On the other hand, stopping smoking has been proven to reverse this process and improve the bone health of patients. We found that patients who smoked were at higher risk of having future fractures after vertebroplasty compared to nonsmokers. Patients who smoked had fractures of the same, adjacent, or other vertebrae postoperatively. If we combine our knowledge about tobacco effects and smoking as a confirmed risk factor [26, 32] then smokers who are going to undergo vertebroplasty should be informed about the risk of refracture, the positive effects of smoking cessation and be strongly advised to stop smoking at least postoperatively. Nonsmokers have the lowest possibility of fracture compared to people who quit smoking and those who continue to smoke. Even people who stopped smoking managed to lower the fracture risk compared to current smokers with smoking exposure > 20 pack years [32] nevertheless how many years the formers were exposed to smoking. Since smoking alone can impair bone metabolism and increase the risk of fractures, its effect on the bone health of individuals with osteoporosis is even more harmful. Patients with osteoporosis should be suggested to avoid this habit or cut it off in order to lower the danger of fracture. PVP may be a minimal, safe, and reliable surgery, but it still includes some anesthesia, immobilization, and recovery time that puts an extra burden on people who smoke. If we add the fact that osteoporotic patients already have bone structure deterioration, then smoking cessation is a precautionary strategy before PVP [32, 33].

Diabetes as a risk factor was included in 3 chosen studies for meta-analysis. However, none of these studies specifies the type of diabetes examined and we accept this as limitation. On the other hand, multiple reasons make us believe that they are referred to type 2 diabetes(2 TD). We specify that they all referred to type 2 diabetes for the following reasons: First, 2 TD is the most common in people over 45 years old and all study groups were above 66 years old. Second, they talk about rising incidence the last decades as an outcome of way of life and this is characteristic of 2 TD which is more related to obesity, unhealthy diet, sedentary life and financial problems [34]. The global prevalence of type 2 diabetes is expected to reach 7,079 cases per 100,000 individuals by 2030, indicating a sustained increase across all regions worldwide [34]. T2D is caused by impaired insulin secretion by pancreatic β-cells and reduced responsiveness of insulin-sensitive tissues to insulin [35]."Factors such as hyperglycemia, insulin resistance, advanced glycation end products (AGEs), and proinflammatory cytokines all contribute to disruptions in normal bone turnover by interfering with the functions of osteoblasts and osteoclasts"[35, 36]. Microstructure deterioration makes bones fragile susceptible to fractures [37–41]. Apart from complex pathophysiology, T2D can be managed if we access its modifiable risk factors. These are diet, exercise levels, sleep, stress and medication. All these can lead to glycaemic control and reduce fracture risk. [42–45]. T2D causes secondary osteoporosis and patients may have a fracture because of diabetes. The addition of a surgical process like PVP does not decrease the existing fracture risk. People with 2 TD and osteoporosis seem to face a higher fracture risk than subjects who have only diabetes or osteoporosis. Since we found that diabetes is a risk factor for a future fracture patients should be guided and properly supported to modify their lifestyle accordingly in order to achieve the optimum glucose level before and after surgey. Proper exercise, a healthy diet rich in fibers, novel drug therapies and probiotics can contribute to good bone health through different pathways [42–46].

Osteoporotic vertebral compression fractures can be detrimental to patients'quality of life and survival. Secondary fractures, whose incidence we found 16,6%, may require repeated PVP or revision surgery with all the possible complications [44]. Although it seems a vital issue that needs scientific attention, there is a scarcity of high-quality studies examining possible risk factors for secondary fractures to prevent them. Most relevant systematic reviews have examined only cemented vertebrae fractures and focused mainly on surgical risk factors [11, 13, 20]. Almost all reviews included risk factors related to patient characteristics. The current study did not show that cement leakage is a risk factor for refracture compared to other reviews [12, 14]. Also, a risk factor we did not examine because the included studies did not meet study criteria was intervertebral cleft, which appears to be a significant risk factor in other reviews [11, 13, 20]. The main difference in this study was that we tried to include risk factors that had to do with patient lifestyle, which has not been done in previous studies.

A raised concern when reflecting on the current study findings is whether the identified risk factors (gender, BMD, smoking, and T2D) for new fractures after PVP put patients at risk irrespective of surgery. Reviewing the literature for these factors separately, we see that all of them, through different mechanisms, make patients susceptible to osteoporotic fractures irrespective of PVP. The question is whether combining these factors and PVP increases this risk. Unfortunately, we are not in a position to answer this question till we review studies that have one group that does not undergo vertebroplasty. Then, we could compare the fracture incidence of this group to that of the one that underwent PVP. This could be a future suggestion that could “clear” the operation role in all this. Until then, the information we have based on our findings is that these factors increase the possibility of patients having a fracture post-operatively, so they need to be addressed.

However, our study has several limitations. Although we did not set it as a restriction, the population was mainly from Asia. Thus, results may not apply to the global population. The final number of studies was small (9), and the strength of this reduced outcome, but this resulted from our effort to include only studies with PVP and not PKP and studies with homogenous fractures. Also, the level of evidence for included studies was not relatively high since all of them were retrospective studies except for one RCT. Another limitation was that the median follow-up was 21 months, which was relatively short for reaching solid conclusions about the fracture incidence and factors that contribute to it. In smoking, diabetes, and kyphosis analysis, we included studies that included cemented or adjacent vertebrae fractures. Ideally, there must be a distinct concomitant analysis for different kinds of fractures for all the risk factors in future studies.

Conclusion

The current meta-analysis found that female gender, low bone mineral density, smoking, and T2D are risk factors for a subsequent fracture after percutaneous vertebroplasty. It is the first meta-analysis that explored lifestyle risk factors, and this is important because these factors can be modified and set the ideal background for a successful operation and prognosis. People undergo surgery to optimize their quality of life. The knowledge of the possible risk factors in advance will help surgeons, interprofessional teams, and patients to be well prepared, plan, and develop a holistic therapeutic approach that will minimize the fracture risk. There is a need for future well-designed studies that explore all possible risk factors by grouping them according to fractures (same, adjacent, remote). This will give a clearer picture of the causes and future solutions.

Funding

Open access funding provided by HEAL-Link Greece. No funding has been received for this study.

Declarations

Statement of Human and Animal rights

This article does not contain any studies with human participants or animals performed by any of the authors.

Conflicts of interest:

Eleni Marselou, Alexios Kelekis, Zacharias Dimitriadis and George A. Koumantakis, declare that they have no conflict of interest.