This meta-analysis included retrospective and prospective cohort studies, as well as case series. Case reports, letters, commentaries, and opinions were excluded.

The patient population was comprised of individuals born with microtia or anotia. Studies that involved ear reconstruction for acquired deformities or traumatic injuries were excluded to avoid confounding variables. To date, there have been no randomized controlled trials that specifically compare autologous reconstruction to Medpor reconstruction in patients with congenital microtia or anotia.

Surgical methods were limited to autologous reconstruction with costal cartilage and ear reconstruction using polyethylene implants. Reconstructions that involved cadaveric costal cartilage, tissue-engineered cartilage, homografts, or prosthetic methods were excluded. Additionally, patients who had previously undergone total ear reconstruction or those undergoing concurrent canaloplasty were excluded to eliminate confounding factors.

Studies that did not address postoperative complications or aesthetic outcomes were also excluded.

Two authors (YJK and ISY) conducted a systematic review, initially examining the titles and abstracts of the studies and excluding those deemed irrelevant. They then performed a full-text review of the remaining articles, excluding any that did not meet the inclusion criteria (Fig. 1). After this selection process, they extracted specific data relevant to the topic from the articles that met the inclusion criteria. Complications were recorded only if explicitly reported and accompanied by the number of affected patients. If a particular complication was not mentioned in the studies, it was not included in the final calculation, rather than assuming an incidence of zero.

The summary statistics from 14 selected studies served as data points for the meta-analyses of postoperative complications and patient satisfaction. These studies differed significantly in terms of sample size and the specific complications assessed, which included cartilage or implant exposure, infection, hematoma, skin necrosis, and hypertrophic scars. Given this variability, a random effects model was utilized to accommodate differences in sample sizes and to offer a more conservative estimate of the pooled proportions [15].

The inverse variance method was employed to assign appropriate weights to each study, facilitating the calculation of pooled proportions and their corresponding 95% confidence intervals (CIs). The DerSimonian-Laird τ² value, along with the I² statistic, was used to quantify study heterogeneity [16,17]. Furthermore, the Freeman-Tukey double arcsine transformation was applied to stabilize the variance of extreme proportions and to adjust them to more closely approximate a normal distribution, thereby rendering them suitable for further statistical inference, such as the Z-test under normality [18].

Due to the lack of direct comparative studies on surgical outcomes between autologous and polyethylene implant reconstructions, we pooled the complication rates and patient satisfaction proportions for each method separately. We then employed both the independent two-sample Z-test and the Cochran Q-test on the subgroups of surgical methods to ascertain if there were statistically significant differences in the rates of complications or patient satisfaction between the two reconstruction techniques. Additionally, the Egger test of funnel plot asymmetry was performed to detect potential publication bias in the studies included in our meta-analyses of surgical complications and patient satisfaction outcomes [19].

All statistical tests were two-sided, and p-values less than 0.05 were considered statistically significant. The analyses were conducted using R version 4.3.2 (The R Foundation for Statistical Computing). Meta-analyses were performed using the “meta-prop” function from the “meta” package in R software.

A total of 647 potential studies were initially retrieved for review. After the exclusion of duplicates, non-English articles, and studies not involving human subjects, 453 articles remained. These articles underwent screening based on their titles and abstracts to assess their relevance to the topic, which narrowed the selection to 98 studies. Of these, 14 studies met the inclusion criteria for meta-analysis [8,20-32]. Eight of these studies, encompassing 3,754 cases, focused on auricular reconstruction using autologous costal cartilage. Four studies, covering 118 reconstructions, investigated the use of polyethylene implants. Two studies directly compared the two methods, involving 196 cases of autologous costal cartilage reconstruction and 194 cases of polyethylene implant reconstruction. The characteristics of participants in each study are detailed in Table 2.

From the 14 selected studies, meta-analyses were conducted on the proportions of five surgical complications: cartilage or implant exposure, infection, hematoma, skin necrosis, and hypertrophic scars, as well as patient satisfaction. The pooled proportions of these outcomes were compared between reconstructions using autologous costal cartilage and those using polyethylene implants.

The proportion of cartilage exposure in autologous reconstruction was 0.008 (95% CI, 0.001–0.018) (Fig. 2), compared to 0.106 (95% CI, 0.070–0.147) for polyethylene implant exposure in alloplastic reconstruction (Fig. 3). The difference in exposure proportions (0.106–0.008 = 0.098) between the two procedures was highly statistically significant, with a p-value of less than 0.0001. This significance was confirmed using either the Z-test under the normality assumption of the Freeman-Tukey transformed proportions or the Cochran Q-test for subgroup differences under a random effects model (Table 3). The infection rate in autologous surgery was 0.004 (95% CI, 0.001–0.009) compared to 0.012 (95% CI, 0.000–0.049) in polyethylene implant reconstruction, indicating a lower rate of infection for autologous reconstruction, although this difference was not statistically significant. Autologous reconstruction also had a lower proportion of hematoma with a pooled rate of 0.043 (95% CI, 0.002–0.124) compared to 0.055 (95% CI, 0.001–0.161) in polyethylene implant surgery, but this difference was not significant. In contrast, polyethylene implant reconstruction showed slightly lower pooled proportions of skin necrosis and hypertrophic scars compared to autologous surgery, but these differences were also non-significant.

For patient satisfaction rates, the proportion of satisfied patients was 0.836 (95% CI, 0.808–0.863) in autologous reconstruction, compared to 0.763 (95% CI, 0.675–0.842) in polyethylene implant reconstruction (Table 4). Although the Z-test and Q-test p-values of 0.076 and 0.081 did not reach statistical significance at the 0.05 level, these relatively low p-values (p < 0.10) suggest borderline significance, indicating potentially higher satisfaction among those who underwent autologous reconstruction.

Due to the varying sample sizes across the selected studies and the relatively small number of studies focusing on polyethylene implant reconstruction cases, the I² statistics for heterogeneity were often high. Consequently, the pooled estimates were calculated using a random effects framework and interpreted with this in mind. Publication bias was assessed using the Egger test for funnel plot asymmetry, as presented in Table 5. Certain specific complications, such as hematoma, skin necrosis, and hypertrophic scars, could not be evaluated due to the limited number of studies on polyethylene implant reconstruction. However, the p-values for the other factors analyzed did not suggest any publication bias in the studies included in our meta-analysis.