Complications with PMMA compared with other materials used in cranioplasty : a systematic review and meta-analysis

Polymethyl methacrylate (PMMA) has been considered a suitable material for cranioplasty. However, no consensus has been reached concerning the best material for cranioplasty with regard to minimizing complications. Thus, this systematic review and meta-analysis aimed to compare the complication rates of PMMA with those of autologous bone and titanium mesh. This review was registered with PROSPERO (CRD42016042725). Systematic searches were conducted on PubMed/ MEDLINE, Scopus, and Web of Science. The focus question was, “Do PMMA prostheses used in cranioplasty have complications rates similar to those of autologous bone and titanium mesh?” A meta-analysis of complication rates was performed on the basis of dichotomous outcomes assessed by risk ratio (RR) with corresponding 95% confidence intervals (CI). From 1014 data sources, 11 articles were selected according to eligibility criteria. These articles involved 1,256 individuals and 1,278 cranioplasties using autologous bone (n = 408), PMMA (n = 379), or titanium (n = 151). The follow-up period ranged from 63 days to 54.3 months. No difference was observed between the complication rates of PMMA and autologous bone (p = 0.94; RR, 0.98; 95%CI, 0.54–1.75) or between PMMA and titanium (p = 0.38; RR, 1.59; 95%CI, 0.57–4.48). Sub-analysis of the reasons for craniotomy (trauma/non-trauma) was conducted, which revealed no significant difference (p = 0.91; RR, 0.95; 95%CI, 0.37–2.42). The meta-analysis indicated that the use of PMMA yields complication rates that are near those of autologous bone and titanium mesh.


Introduction
Craniofacial defects caused by factors, such as head injuries, cerebral tumors, ischemia, infections, or intracranial disorders, contribute to loss of bone tissues and the corresponding need for later reconstruction. 1Cranioplasty has been a preferred procedure in modern neurosurgery for the repair of bone defects using various materials, including bone and alloplastic materials. 1,2Cranioplasty is a critical procedure as skull integrity protects the brain. 1,2Cranial reconstruction also contributes to the preservation of the esthetic aspects of the head and restores an individual's quality of life. 3everal materials have been used for reconstruction of cranial defects.The ideal material for cranioplasty must be radiolucent, resistant to infection and biomechanical deformation, strong, nonconductive to heat or cold, can be fitted to defects with complete closure, inexpensive, and available for immediate use. 4,5Although various materials have been proposed, currently no available material meets these criteria. 1,4utologous bone grafts have traditionally been considered the gold standard for cranial reconstruction 6 given that bone is a natural substrate. 7,8Autologous bone grafts are associated with low cost and low risk and are preferred materials for small-and medium-sized defects. 1,6The disadvantages of autologous bone grafts include the limited quantity of material (as quality depends on the availability of graft donors), difficulty in shaping the graft, and increased morbidity and risk of complications, including infection and bone resorption. 6,9lloplast ic materials are alter nat ives to autologous bone grafts; these materials must possess biocompatibility, adequate mechanical resistance, malleability for esthetic purposes, and low risk of infection.Candidate materials include metals (titanium mesh), ceramics (hydroxyapatite and calcium phosphate) and polymers (polymethyl methacrylate [PMMA] or polyetheretherketone [PEEK]). 10Titanium features good mechanical and biological properties; however, it is expensive 3,11,12 and can result in infection.Titanium mesh generates artifacts in imaging tests and presents difficulty for adaption for soft tissue coverage. 6imilar to native bone tissue, PMMA offers strength and protection and has long been used in cranioplasty 1,13 owing to its remarkable plasticity and long-term durability.PMMA is easy to shape, lightweight, inexpensive, and produces no visible artifacts in radiographic images. 4,8,14This material can also be applied during reconstruction and be prefabricated. 8,11However, despite these advantages, PMMA can cause damage to human cells. 15,16The primary factors responsible for this effect include residual monomers from polymerization. 3Processing type can also influence the complication rate of this material. 17The lack of porosity of PMMA can also interfere with osteoconduction and vascularization, compromising interactions with surrounding tissues.PMMA may be as susceptible or more prone to infection as autologous bone tissue. 1,18o emphasize, cranioplasty material indications need to be based on scientific findings.Many reviews have evaluated the effects of various materials, including PMMA, that are used for cranioplasty. 1,4,5,19,20owever, in most cases, these reviews included studies with indirect comparisons, that is, without comparison of different materials in the same included study.No systematic review specifically evaluated the influence of PMMA compared with other materials in terms of complication rates for cranioplasty.
Therefore, careful analysis of clinical studies should be performed through systematic review and metaanalysis to assess the influence of PMMA in comparison with autologous bone and titanium mesh in terms of complication rates.The null hypothesis states that no difference exists between the complication rates of PMMA compared with those of other materials.

Methodology Registration protocol
This systematic review follows the structural pattern in the preferred reporting items for systematic reviews and meta-analyses checklist. 13The methods used in this systematic review were registered in the PROSPERO database (CRD42016042725).

Eligibility criteria
This analysis included randomized controlled trials (RCTs), prospective studies, retrospective studies, follow-up studies, and studies in English that evaluated the biological behavior of acrylic resin in craniofacial prostheses.Studies focusing only on mechanical and physical outcomes, in vitro studies, animal studies, clinical cases, and literature reviews were excluded.
The population, intervention, comparison, and outcome approach was used to address the question, "Do PMMA prostheses used in cranioplasty have complication rates similar to those of autologous bone or titanium mesh?"The population comprised patients who had undergone cranioplasty.Intervention involved the use of PMMA prostheses for cranioplasty.Comparison focused on the use of autologous bone or titanium mesh.The primary outcome was complication rates.

Search strategy and information sources
Two independent researchers (R.S.L. and J.R.S.M.) conducted an electronic search and selected studies that evaluated cranioplasty with different materials (PMMA compared with autologous bone graft and titanium mesh).Electronic searches of the selected databases (PubMed/MEDLINE, Web of Science, and Scopus) were conducted for articles published before March 2017.The following keywords were used: "acrylic resin" AND "maxillofacial prostheses" OR "acrylic resin" AND "cranioplasty" OR "acrylic resin" AND "maxillofacial" AND "prostheses" OR "acrylic resin" AND "head neck."To complement this search, a manual search was performed in journals, including Journal of Oral and Maxillofacial Surgery, Journal of Cranio-Maxillo-Facial Surgery, International Journal of Maxillofacial Surgery, The Journal of Craniofacial Surgery, Journal of Oral Biology and Craniofacial Research, and Neurosurgery.Articles were selected according to eligibility criteria after title and abstract screening.Any disagreements between the researchers were resolved by a third researcher (S.L.D.M.), and consensus was reached through discussion.

Risk of bias and quality analysis of the included studies
Two researchers (R.S.L. and J.R.S.M.) qualitatively analyzed the methodological quality of the selected studies using the Newcastle-Ottawa Scale (NOS) for cohort studies (non-RCTs); analysis is based on three major components: selection, comparability, and outcomes.In this quality scale, scores are represented by stars.A score of five or fewer stars represents a high risk of bias, whereas six or more stars (maximum of nine) are considered to indicate a low risk of bias. 21

Summary measures
The dichotomous outcome measure was the occurrence of complications, as evaluated by risk ratio (RR) and its 95% confidence interval (CI).I 2 statistic was used to express the percentage of total variation across studies due to heterogeneity, with 25% corresponding to low heterogeneity, 50% to moderate, and 75% to high. 22 random effects or fixed effects model was used for analysis, depending on study heterogeneity.In cases of significant heterogeneity (P < 0.10), a random effects model was used to assess the significance of treatment effects.If no significant heterogeneity was identified, analysis was performed using a fixed effects model. 14,15unnel plots (plots of effect size versus standard error) were also constructed.Asymmetry of a funnel plot may indicate publication bias or other biases related to the sample size; however, asymmetry may also represent a true relationship between the trial and effect size. 14eview Manager 5.3 (The Cochrane Library) was used to perform meta-analysis and create forest plots; P < 0.05 was considered significant.

Additional analysis
Kappa score was used to calculate the agreement bet ween researchers during selection.Any disagreements were resolved by discussion and consensus among all authors.

Risk of bias assessment
Mean score was 7 stars, and the lowest was 5 stars.Only one study 3 showed the highest score (9 stars), indicating a low risk of bias (Table 1). 23

Literature search
A total of 1014 articles were obtained from the databases: 648 from PubMed/MEDLINE, 317 from Scopus, and 49 from Web of Science.After removing duplicate articles, a detailed review of titles and abstracts of the selected studies was performed.After applying the inclusion/exclusion criteria, 65 complete papers were selected for eligibility assessment.After reading the full texts of these articles, 11 studies 1,2,4,5,12,[16][17][18][19][20][21] were included in the final review.Details of the search strategy are presented in a flow diagram (Figure 1).Table 2 summarizes the contents of articles used for comparative analysis.

Inter-investigator agreement
Inter-examiner agreement (kappa) reached 0.94 for PubMed/MEDLINE, 0.89 for Scopus, and 1.00 for Web of Science.This value was calculated by evaluating the selected titles and abstracts that presented high levels of agreement among the reviewers under the kappa criterion. 22

Study description
All studies centered on cranial implants and reported data on 1278 cranioplasties performed in 1256 patients (mean age: 34.14 years; 65.5% men).Nine studies were retrospective, and two were prospective.All 11 studies reported PMMA results and compared two or more materials.Autologous bone graft was the most commonly used material (408 cases); PMMA was used in 379 cases, titanium in 151, and ceramics in 10.Follow-up period was reported in nine papers (n = 1003 cases) and ranged from 63 days to 54.3 months.Indications for cranioplasty were not cited in one study (n = 45).Trauma as an isolated factor was the most common indication (n = 592 cases), and non-traumatic events (including aneurysm, tumor, infection, and hemorrhage) were indications in 640 cases.
*Sufficiently long follow-up for cases of complications was 6 months.23;♦ Iindicated that studies responded positively to parameter evaluated.
A sub-analysis was conducted to determine whether the reason for cranioplasty (trauma vs. non-trauma)  influences the complication rate of PMMA.A fixed effects model was used given the lack of significant heterogeneity (p = 1.00;I 2 = 0%).Data from three studies were compared, 1,16,18 and no significant difference was observed (P = 0.91; RR, 0.95; 95%CI, 0.37-2.42; Figure 4).The funnel plots of the studies included in analyses of complication rates of PMMA versus autologous bone and PMMA versus titanium mesh and etiology of trauma versus non-trauma showed symmetry, indicating the possible absence of publication bias (Figures 5 A,

Discussion
Currently, alloplastic materials are considered alternatives for cranioplasty to autologous bone, particularly when considering a large skull area. 1 Among the various alloplastic materials that can be used for cranioplasty, PMMA and PEEK have been considered favorable. 1,6,27 previous systematic review comparing the effects of PEEK with those of autologous bone and titanium mesh revealed no significant difference in complication rates. 6The current meta-analysis revealed no difference between the complication rates of PMMA and other materials (autologous bone and titanium mesh).Thus, the null hypothesis of this study was accepted.These results corroborate with those of other studies. 28,29nfection was the most prevalent complication (55.1%).In several studies, 7,11,[24][25][26]28,[30][31][32] infection may have occurred as a result of technical errors during surgery. This reslt may be related not only to the type of material but also to different risk factors in patients, such as hypertension and diabetes, 20 systemic infection, lower hemoglobin levels, motor deficits, and the interval between craniotomy and cranioplasty.32 Only two studies evaluated the failure rate 11,33 of cranioplasty material throughout a follow-up period.This condition explains why failure rate was not evaluated as an outcome in the present study.Bobinsk et al. 33 reported better long-term survival for PMMA than for autologous bone (79.5 vs. 48.1 months; p = 0.035).Similarly, Al-Tamimi et al. 11 observed longer survival for PMMA (135 months) than for titanium mesh (92 months) but reported no statistical significance.
No differences were observed in the complication rates between autologous bone and PMMA.However, Bobisnk et al. 33 showed higher complication incidence level at autologous bone cranioplasty.Although autologous bone is considered the gold standard material for cranioplasty because of its biocompatibility and genetic compatibility, its indication could be limited to smaller and medium defects (< 75 cm 2 ).
Higher incidence of complications for autologous bone could be related to infections, which may lead to partial resorption of bone graft and implant detachment. 1 Thus, the use of autologous bone tissue may be related to frequent complications of bone resorption, 31,33,34,35 particularly when applied to large areas of skull defect (> 75 cm 2 ). 36Another factor that contributes to bone resorption is the method of bone storage because freeze-dried bone loses its vitality and undergoes unpredictable resorption. 37elevant information about bone grafts, which ideally would be included in this manuscript but was not provided by the authors, includes the area of the cranium that required rehabilitation, graft type, 7,11,24,25,26,28,30,31,32,33 and area from which the donor graft was obtained. 7,11,24,26,28,30,31,32,33lloplastic materials, such as PMMA, are considered good choice for patients who require secondary reconstruction with a custom-made CAD/ CAM device following a previous failed cranioplasty (contour deformity caused by significant bone graft resorption or implant/bone graft removal resulting from infection). 1 In studies on patients requiring new surgery after infection or bone resorption, PMMA presented a lower risk of complications compared with other materials . 25,28,31itanium mesh showed no differences in complications compared with PMMA.However, the cost of titanium prostheses is much higher, causing difficulty in acquiring the material in locations with financial constraints and presenting a major disadvantage.In addition to processing, titanium Heterogeneity: Chi2 = 0.01; df = 2 (p = 1.00).I2 = 0% Test for overall effect: Z = 0.11 (p = 0.091) implants typically require advanced laboratory techniques. 3Another negative factor is that titanium can cause artifacts on magnetic resonance imaging; although slight and inert, they compromise the image. 11iomechanically, titanium is a strong material and can withstand higher loads than PMMA, but PMMA tolerates forces better than bone tissue, generating acceptable neurocranial protection for fracture resistance. 27lthough PMMA is considered a good choice for cranioplasty, it features some disadvantages, such as lack of adhesion to organic tissues, 14,38 which increase the risk of infection and exposure of the material. 3utogenous bone, by its nature, does not present this problem, and studies reported regeneration of the cortex of the defect area. 39Akan et al. 12 observed that exothermic reaction of self-curing resin can result in damage to the surrounding tissues.The use of self-curing PMMA should be avoided as prosthesis are now manufactured in a preoperative stage and at an affordable cost. 40To minimize exothermic reaction in the surgical area, prosthesis could be previously manufactured in a prototype model aided by tomographic images. 8Another possibility is the use of thermal activation of PMMA. 12Although used with computed tomography and a prototype model, PMMA prostheses are much more affordable than titanium prostheses.The cost and decreased complexity of the technique can also reduce waiting time for surgery in the area to be rehabilitated.
Another disadvantage is the release of residual monomers by chemical activation of PMMA, which creates a possibility of tissue damage. 3The type of PMMA cure may influence the level of residual monomers present.Therefore, it would be interesting to compare different forms of polymerization of this material with the corresponding inflammatory responses.However, in some studies, the type of PMMA used is unclear; four studies used selfcuring 7,11,26,33 and two studies a thermoactivated. 3,24n several studies, trauma was the most common indication for cranioplasty. 3,7,24,31This observation accords with the age of individuals and the condition that most patients were men, which are characteristics of people at higher risk of automobile and other accidents. 41In this study, for PMMA, no statistical relationship was observed between the indication for cranioplasty and infection outcome. 3,26,29,31,41This result suggests that indication for cranioplasty is indirectly related to the probability of infection, which is related to individual systemic health, as previously reported.However, the small number of studies and patients in this analysis could have resulted in bias in these results.
Other materials, such as hydroxyapatite, ceramics, and PMMA + titanium, have also been used in the analyzed studies. 26We excluded these materials in the meta-analysis, as each was only used in one study.An analysis would be interesting, as materials such as ceramics possess the advantage of biocompatibility with human tissues. 42ompared with other alloplastic materials, PMMA is the least expensive and features satisfactory characteristics for cranioplasty.However, these alloplastic materials are not indicated for pediatric patients because of incompatibility to accommodate bone growth of cranial skeleton. 1,18The optimal material for cranioplasty in pediatric patients should be considered in future research. 1onsidering these results and in agreement with literature, the choice of cranioplasty materials may be related to the preference of surgeons for the material used. 1 Other alternatives for reducing complication rates, such as incorporation of antibiotics into the restorative material, may also be used. 43In addition, the approach of determining treatment by planning conformation and fixation of cranial prosthesis using finite element analysis for each isolated case was shown to be effective through the low complication rates presented. 27n this systematic review, the included studies reported no rates of specific types of complications (infection, hematoma, dehiscence, bone resorption, thermal sensitivity, and material displacement) according to the material used for cranioplasty, rendering the specific meta-analysis for each type of complication impossible.Another limitation was the lack of RCTs on the subject; only retrospective and prospective studies were used.For this reason, NOS was used to analyze the risk of bias because it is the most used instrument for non-RCTs. 21This scale includes three evaluation criteria: selection, comparability, and outcomes.Despite only one article reported the highest score (9 stars = low risk of bias), mean across the included studies reached 7 stars.However, for some studies, the sample selection method was unclear, with no other comparations factors excluding the main comparation.This highlights the need for controlled studies on this subject.
The present findings should be cautiously interpreted.All included studies were retrospective and prospective, reducing the level of evidence because of the possible presence of uncontrolled confounding factors.A large variation was also between the follow-up periods of the included studies, and this observation is another limitation of this review.Further studies (preferably RCTs) with longer follow-up periods are recommended to investigate the influence of PMMA in cranioplasty and determine its likely effects.

Conclusions
This meta-analysis indicates that complication rates of PMMA in cranioplasty show no statistical difference from those of autologous bone grafts or titanium.

Figure 4 .
Figure 4. Forest plot: comparison of indication (trauma/non-trauma) and occurrence of complications when PMMA was used as the material of choice.

Figure 5 .
Figure 5. Funnel plot for the included studies.A: Complication rates of PMMA versus autologous bone grafts; B: PMMA versus titanium mesh; C: Trauma versus non-trauma.

Table 2 .
Characteristics of the analyzed studies (11 articles).
Figure 2. Forest plot: comparison between PMMA and autologous bone grafts with respect to complications.
Figure 3. Forest plot: comparison between PMMA and titanium with respect to complications.