Page 41 - Clinical relevance of current materials for cranial implants

Table 3: Methodological quality of 226 included observational studies
Quality Question
Clear definition of study population?
Exclusion of selection bias?
Clear definition of results?
Clear method to determine results?
Outcome determined blind from the intervention? Affects this the evaluation of the outcome? Follow-up long enough?
Selective lost to follow up excluded?
Are confounders described?
Are the results corrected for confounders? (multi-variate analysis)
Primary outcome measures
Yes (N) %
226 100 130 57.5 26 11.5 13 5.8 0 0 226 100 79 35.0 197 87.2 10 4.4 10 4.4
Autologous bone is inferior to alloplastic cranioplasties
Methodological quality
Overall, the included studies were of low quality. Risk of selection bias appeared to be
high in 97 (42.5%) of the non-randomized studies (n=226), mainly due to not reporting 2 the selection criteria of the included patients (Table 3). In the 2 randomized control
trials, patients were blinded for the material used and the studies were imperfect
regarding blinding of the care provider and outcome assessor. Furthermore, the
groups were relatively small due to the assumed large variability in infection rates per
material, resulting in a skewed power.
None of the outcome measures could be pooled due to clinical heterogeneity, rendering meta-analysis impossible. The overall reported infection rate was 5.6% across all cranioplasty materials used. Autologous cranioplasties showed an infection rate of 6.9%, significantly greater than the combined alloplastic materials (overall: 5.0%; RD = 0.019 [95% CI 0.009-0.030]; NNT = 53 [95% CI 34-116]; RR = 0.73 [95% CI 0.62-0.86]). The lowest infection rate was observed in HA (overall: 3.3%; range: 0-58.8%). The highest infection rate was reported for PMMA (overall: 7.8%; range: 0-50%). Of the 104 studies reporting at least one infection of the cranioplasty (total infected cranioplasties n = 550), 27 included bacteriologic culturing. Staphylococci were the most detected infecting agent, causing infection in 90.7% of infected cranioplasties reporting bacterial culture. Specifically, 71.1% tested positive for Staphylococcus aureus, including methicillin-resistant S. aureus (28.9%) and methicillin-sensitive S. aureus (4.1%), 4.1% for Propionibacterium acnes, 2.1% for S. epidermidis, and 24.7% tested positive for a different bacterial strain. Of these, 16 infected cranioplasties reported multiple strains of bacteria.
39