The study process was conducted following the accepted methodology recommendations of the PRISMA checklist for systematic review and meta-analysis where registration of the protocol is not mandated.16 We searched PubMed, Scopus, Web of Science, Virtual Health Library (VHL), Google Scholar, and Cochrane databases. The search was done using the following search strategy “(nasal reconstruction OR rhinoplasty OR nasal augmentation OR nasal reconstruction OR nasal dorsum reconstruction OR nasal plastic surgery) AND (alloplastic implant OR nasal implants silicone OR Medpor OR Gore-Tex) AND (costal OR costal cartilage OR rib OR osteocartilaginous rib graft).

We marked studies as included if they met the following criteria: 1) Study design: Clinical trials: comparative or non-comparative; randomized or not randomized clinical trials; 2) Population: patients who underwent nasal dorsum reconstruction; 3) Intervention: rhinoplasty using autologous graft (autologous costal cartilage) or alloplastic grafts; 4) Outcome: at least one of the surgery-related complications. We excluded studies in the following conditions: 1) Observational or retrospective studies; 2) Studies of homologous rib graft; 3) Animal studies, in vitro studies, review articles, case reports, conference abstracts, and duplicate publications.

Two authors performed the extraction sheet on Microsoft Excel file by pilot extraction of at least three papers. Three reviewers independently extracted data from included studies using the excel sheet. The fourth reviewer performed data checking for checking the accuracy of the extracted data. All the disagreements and discrepancies were resolved by discussion and consultation with a senior member when necessary.

We performed the risk of bias assessment using ROBINS-1 (Risk of bias in non-randomized studies of interventions) tool. It includes the following domains: 1) Bias due to confounding; 2) Bias in selection of participants into the study; 3) Bias in classification of interventions; 4) Bias due to deviations from intended interventions; 5) Bias due to missing data; 6) Bias in the measurement of outcomes, and 7) Bias in the selection of the reported result. Studies are marked to either low, moderate, or high risk of bias according to each domain.

We used OpenMeta [Analyst] Software (Center of Evidence-Based Medicine, Brown University, School of Public Health, Rhode Island State, USA) to perform the meta-analysis of all outcomes. The mean proportions of complication rates were pooled in a meta-analysis model, using the Mantel-Haenszel method. The analysis was performed under the fixed-effects model for homogeneous data, and the random-effects model for heterogeneous data. The corresponding 95% Confidence Intervals (95% CI) of pooled effect size were calculated using a random-effects due to the presence of heterogeneity.

Heterogeneity was assessed with Q statistics and I2 test considering it significant with I2 value > 50%.17 Publication bias could not be assessed using Egger’s regression test due to the small number of included studies (less than 10).18,19 The publication bias was assessed using Egger’s regression test18,19 and represented graphically by Begg’s funnel plot20 when there were ten or more studies. Egger’s regression test p-value <0.10 was considered significant. Whenever publication bias was found, the trim and fill method of Duvall and Tweedie was applied to add studies that appeared to be missing21 to enhance the symmetry. A meta-regression model was done to correlate the time of surgery and the occurrence of different complications.

Our search yielded a total of 789 studies. Following screening and excluding duplicates, 40 studies remained that entered full-text screening. We excluded 10 studies due to either lack of outcomes or studies of non-human species. Finally, 30 studies were included in the meta-analysis, as reported in the PRISMA flow diagram, Fig. 1.

Figure 1.

PRISMA flow diagram.

(0.15MB).

We included 30 studies, two of which were controlled trials; therefore, we treated each group as a separate study for feasibility in performing a single-arm meta-analysis. Autogenic grafts were used in 21 studies, while 10 studies used alloplastic grafts. One study used a combination of alloplastic and autogenic grafts.22 A total of 1013 patients were enrolled in this meta-analysis, with a mean age of 27.7 years, and a mean duration of follow-up after surgery of 22 months. Table 1 shows a summary of baseline characteristics for each study.

We reported an overall low risk of bias according to the ROBINS-1 tool. All studies had a low risk of confounding bias, the bias in classification of interventions, bias due to deviations from intended interventions, the bias in the measurement of outcomes, and bias in the selection of the reported result. Nine studies23–31 did not report enough baseline data about patients such as sex and mean age. Therefore, these studies marked a moderate risk of bias.

Regarding missing data, eight studies25,27,28,30,32–34 did not report necessary outcome endpoints to be included in a meta-analysis. We did not find any data for important complications of the surgery such as infection, therefore these studies were put to a high risk of bias. The risk of bias summary was of included studies was reported in Fig. 2.

Figure 2.

Summary of risk of bias for included trials using ROBINS-1 tool.

(0.39MB).

The overall complication rate was 7.1% (95% CI 4.9%−9.3%) which was higher in the alloplastic grafts group 7.8%, (95% CI 4.1%−11.5%) than the autogenic group 6.9%, (95% CI 4.3%− 9.6%). There was no heterogeneity among the included studies (I2 = 47.6%) (Fig. 3). One study22 used combined alloplastic and autogenic graft, therefore, it removed from the subgroup analysis by graft type.

Figure 3.

Forest plot of total complications rate.

(0.4MB).

The overall warping rate as reported by fifteen studies (530 participants) was 1.6% (95% CI 0.5%−2.6%). Four studies16,24,35,36 used alloplastic grafts and reported a prevalence rate of 1.6% (95% CI −0.5% to 3.7%), while in the autogenic group, the pooled proportion of warping as reported by ten studies18–21,23,28,33,37–39 was 1.5% (95% CI 0.3%−2.8%) (Fig. 4).

Figure 4.

Forest plot of warping outcome according to graft type.

(0.27MB).

Pooling the results from four studies of high-income countries,21,23,33,37 nine studies of middle-income countries,16‐19,22,24,28,35,36 and two low income studies revealed a prevalence rate of 1.7% (95% CI −0.5% to 3.9%), 1.7% (95% CI 0.3%−3.1%), and 1.0% (95% CI −1.0% to 2.9%), respectively (Fig. 5). Similarly, the mean proportion of warping rate from Asian studies16,20–22,33,35–38 was 1.3% (95% CI 0.2%−2.4%), while in European studies16–19,23,24,28,36 was 2.4% (95% CI 0%−4.8%) (Fig. 6). Pooled studies, in all analyses, did not show any significant heterogeneity.

Figure 5.

Forest plot of warping outcome according to country income level.

(0.31MB).

Figure 6.

Forest plot of warping outcome according to the study sitting.

(0.28MB).

The prevalence of infection was reported by 17 studies (578 patients). The overall infection rate of the nasal dorsum reconstruction surgery was 2.1% (95% CI 1.0%−3.3%). The alloplastic grafts were associated with higher infection rates of 3.7% (95% CI 0.4%−7.0%), compared to only 2.1% (95% CI 0.8%−3.4%) among autogenic grafts (Fig.7).

Figure 7.

Forest plot of infection outcome according to graft type.

(0.29MB).

Rates of infection were lowest in high income countries with 1.8% (95% CI −0.3% to 3.9%), followed by middle, and low income countries, with 2.3% (95% CI 1%−3.3%), and 2.1% (95% CI −0.5% to 4.7%), respectively (Fig. 8). The results showed that according to patients’ regional country, nine Asian studies16,20–22,35,37,38,40,41 showed lower infection rates of 1.8% (95% CI 0.4%−3.2%), compared to Eight European studies17–19,23,24,39,42,43 that showed an infection rate of 2.8% (95% CI 0.9%−4.8%) (Fig. 9). Pooled studies, in all analyses, did not show any significant heterogeneity.

Figure 8.

Forest plot of infection outcome according to country income level.

(0.33MB).

Figure 9.

Forest plot of infection outcome according to study sitting.

(0.31MB).

Eleven studies (405 patients) reported the prevalence rates of all scars. The overall pooled rate was 1.7% (95% CI 0.5%−2.9%), for both nose and chest scars. Subgroup analysis showed that prevalence of scars in autogenic grafts reported by six studies18,20,23,37,39,42 was 1.8% (95% CI 0.2%−3.4%), while in the alloplastic grafts, reported by four studies,24,35,36,44 the scarring rate was 1.5% (95% CI −0.4% to 3.5%) (Fig. 10).

Figure 10.

Forest plot of hypertrophic scars outcome according to graft type.

(0.21MB).

Stratifying studies according to countries, the scarring rate was the highest in low income countries with 9.8% (95% CI 1.6%−18%), followed by high income countries with 2.8% (95% CI −0.1% to 5.7%), and middle-income countries with 1.2% (95% CI −0% to 2.5%) (Fig. 11). The Asian studies reported higher scarring rates of 1.8% (95% CI 0.2%−3.4%) as reported in six studies.20,22,35,36,39,44 European patients experienced 1.4% scar rates (95% CI −0.2% to 4.8%) as reported in five studies18,23,24,39,42 (Fig. 12). Pooled studies, in all analyses, did not show any significant heterogeneity.

Figure 11.

Forest plot of hypertrophic scars outcome according to country income level.

(0.25MB).

Figure 12.

Forest plot of hypertrophic scars outcome according to study sitting.

(0.23MB).

A total of 19 studies (613 participants) reported the need for a second surgery. The overall rate of re-correction surgery was 4.1% (95% CI 2.5%−5.6%). Subgroup analysis showed that the need for a second surgery was higher in the autogenic grafts with 4.2% (95% CI 2.4%−6.1%), compared to the alloplastic grafts with 3.7% (95% CI 1.0%−6.4%) (Fig. 13).

Figure 13.

Forest plot of need for a second surgery outcome according to graft type.

(0.3MB).

The middle-income countries reported the lowest need for second surgery with only 1.2% (95% CI 0%−25%),16,18,22,24,25,29,31,32,34–36,39,43,45 followed by high income country studies33,40,42,44 2.3% (95% CI 2.8%−9.5%), and low income countries 9.8% (95% CI 6.3%−27.8%), respectively (Fig. 14). Asian studies reported second surgery rates of 1.8% (95% CI 0.2%−3.4%) as reported by 11 studies,16,22,32,33,35,36,38,40,44,45 while a lower rate was observed in the European studies 1.5% (95% CI 0.3%−2.7%) as reported by eight studies18,24,25,29,31,39,42,43 (Fig. 15). Pooled studies, in all analyses, did not show any significant heterogeneity.

Figure 14.

Forest plot of need for a second surgery outcome according to country income level.

(0.24MB).

Figure 15.

Forest plot of need for a second surgery outcome according to study sitting.

(0.22MB).

A meta-regression model was performed to correlate the date of surgery (from 1999 to 2018) with the occurrence of complications. We found no significant effect of surgery date on warping rates (p = 0.6), infection rates (p = 0.2), hypertrophic scars (p = 0.1), or the need for a second surgery showed (p = 0.1) (Fig. 16).

Figure 16.

Meta-regression analysis.

(0.26MB).

This meta-analysis reported a significant publication bias in two outcomes: the need for a second surgery and hypertrophic scars (Fig. 17). The rest of the outcomes and subgroup comparisons did not show any similar funnel plot asymmetry, hence no risk of bias (Fig. 18).

Figure 17.

Begg’s funnel plot shows significant publication bias.

(0.15MB).

Figure 18.

Begg’s funnel plot shows non-significant publication bias.

(0.23MB).