Effect of the type of surgery, use of intraoperative topical mitomycin C or stenting on the outcome of choanal atresia repair: a systematic review and meta-analysis

Background Restenosis and recurrence of symptoms after posterior choanal atresia (CA) repair is a major concern for surgeons, which leads to the use of a variety of surgical approaches and adjuvant treatment modalities to avoid restenosis. Aim The study was designed to compare the outcome of transnasal endoscopic and transpalatal repair of CA as regards restenosis, and also to compare the effect of using topical mitomycin C (MMC) or nasal stenting with not using these modalities on the outcome of repair as regards restenosis. Methods A systematic search was conducted on the PubMed/MEDLINE to locate and select relevant studies without applying any limits. Studies included in meta-analysis were tested for heterogeneity of the estimates. Results As regards comparing the transnasal endoscopic and transpalatal approaches, eight studies, involving 410 cases, met our criteria; 197cases were repaired endoscopically with 37.5% restenosed and 188 through transpalatal approach with 28.2% restenosed. Using the estimated odds ratio, no statistically significant difference was found as regards restenosis. For the use of MMC, five articles fulfilled our criteria; they involved 155 cases; MMC was used in 70 cases where 24.2% restenosed, and was not used in 85 cases where 35.2% restenosed. By estimated odds ratio, there was no statistically significant difference between the two groups as regards restenosis. For the use of stenting, four articles fulfilled our criteria; they involved 165 cases, out of which 93 cases used nasal stent after repair of CA (21.5% restenosed), whereas in 72 cases stent was not used (18% restenosed). By estimated odds ratio there was no statistically significant difference between the two groups. Conclusion The available evidence suggests that there is no statistically significant difference between transnasal endoscopic and transpalatal approach in the repair of CA as regards restenosis; furthermore, there is no statistically significant difference between using intraoperative topical MMC and nasal stent and not using such modalities on the outcome of CA repair.


Introduction
Choanal atresia (CA) is a relatively rare congenital anomaly occurring in about one in 5000-8000 live births; its female to male ratio is 2: 1. Generally, 65-75% of patients with CA are unilateral, and about 30% are pure bony, whereas 70% are mixed bony-membranous [1].
Management of these patients varies and depends on the type of atresia, age, and general condition of patients. The most common surgical techniques used are the transnasal and transpalatal approaches. While transnasal endoscopic approach is the method of choice and has been used successfully in newborns and infants and is suitable for membranous or very thin bony atresia, the transpalatal approach is normally reserved for the older children, thick bone, or cases with restenosis [2].
The concern was always to adequately repair CA and prevent recurrence of symptoms by preventing restenosis.
Therefore, in this study we aimed to systematically review the data about the repair of CA and outcome of repair, through comparing, via statistical analysis, the transnasal endoscopic and transpalatal approaches as regards restenosis, and to find out the role of using mitomycin C (MMC) as adjuvant treatment and the role of stenting in improving the outcome of CA repair.

Identification and location of articles
Medline database at http://www.pubmed.gov was searched, using the keyword 'choanal atresia' without applying any limits. Then the search was conducted by entering the following keywords choanal atresia+repair, choanal atresia+endoscopic, choanal atresia+transnasal, choanal atresia+stent, and choanal atresia+topical mitomycin C. The last search was conducted on 10 January 2015.

Screening and evaluation
After blinding the articles regarding authors and journal of publication, the articles yielded by the search were screened regarding four inclusion and three exclusion criteria.
(1) Inclusion criteria: Articles in which the number of cases reported was three or more in each group for case-control studies were included; in addition, only primary cases and those studies in which outcomes were reported in detail and complications mentioned were included in the present study. (2) Exclusion criteria: Review articles were excluded because there were no cases or results. Articles reporting on less than three cases or revision cases were also excluded. Only articles fulfilling the criteria of screening were included for further steps of data collection, analysis, and reporting.

Data collection
Data collected from the articles included the type of study, number of cases reported in each article, surgical approach used (whether transnasal endoscopic or transpalatal), and the topical application of MMC or stenting, functional outcomes (improvement of symptoms after treatment), and the presence and type of complications.

Data analysis
Statistical analysis was carried out using comprehensive meta-analysis, version 2 (Biostat, Englewood, New Jersey, USA). Every article fulfilling the criteria of screening and data collection was fed into the abovementioned computer software.

Pooling of estimates
Risk for unwanted outcomes was expressed in terms of odds ratio with 95% confidence interval (CI); estimates from included studies were pooled using both the Mantel-Haenszel fixed-effects method and the DerSimonia laerd random-effects method.

Examination of publication bias
Publication bias was examined using the Funnel plot and Duval and Tweedie's trim and fill tests.

Results
Searching Medline database, http://www.pubmed.gov with the last search on 10 January 2015 using 'choanal atresia' without applying any limits yielded 12 400 articles; the number of articles yielded by using each keyword is mentioned in Table 1.
Results of comparing the outcome of transpalatal versus transnasal endoscopic repair of CA First, we compared between the studies of transnasal and transpalatal approach in CA repair. We screened for keywords choanal atresia, transnasal endoscope, and Table 1 Keywords used in search and the number of articles yielded for each keyword transpalatal. We found eight articles fulfilling our criteria, and they involved 410 cases. Cases that were repaired endoscopically were 197 in number, whereas transpalatal approach was used in 188 cases. Restenosis occurred in 37.5% of those who had undergone endoscopic approach and in 28.2% of those who had undergone transpalatal approach ( Table 2).
The incidence of restenosis after transnasal endoscopic or transpalatal repair, using the estimated odds ratio  Table 4 and Fig. 3.
The funnel plot for the log risk ratio versus the SE of risk ratio showed some evidence of publication bias (Fig. 4) Results of comparing the outcome of transpalatal and transnasal endoscopic repair of CA with and without the usage of mitomycin C When we screened the articles for the application of topical MMC for improving the outcome of CA repair by using key words choanal atresia and mitomycin C, seven articles met our criteria but two of them were excluded as these were one-armed studies giving insufficient data, and thus only five articles were included for comparison; these articles involved 155 cases, managed by transnasal endoscopic and transpalatal approach. MMC was used in 70 cases, and was not used in 85 cases. We found that the rate of restenosis in cases of MMC application was 24.28%,  Forest plot for the incidence of restenosis after transnasal or transpalatal repair using the estimated odds ratio (OR) as the effect measure, A endoscopic technique; B transpalatal technique. CI, confidence interval.

Figure 2
Funnel plot of the log odds ratio and the SE of odds ratio for determination of publication bias.
whereas in cases without MMC it was 35.29% ( Table 5).
The meta-analysis of studies comparing MMC application versus no MMC as regards the incidence of restenosis using the estimated odds ratio (statistical measure) showed a statistically nonsignificant difference between the two groups. Measures of heterogeneity revealed unimportant heterogeneity of the estimates reported by the included studies (Cochran Q P-value, 0.091; I 2 , 50.1; τ 2 , 0.939).
Results of comparing the outcome of transpalatal and transnasal endoscopic repair of CA with and without using stenting For comparing the result of using nasal stent as an adjuvant tool to improve the outcome of transnasal or transpalatal approach, we screened for keywords choanal atresia and nasal stent. Seven articles met our criteria but three of them were excluded as these were one-armed studies giving insufficient data, and thus only four articles  Forest plot for the incidence of restenosis after transnasal or transpalatal repair using the estimated risk ratio (RR) as the effect measure. CI, confidence interval.

Figure 4
Funnel plot of the log risk ratio and the SE of risk ratio for determination of publication bias.

Figure 5
Showed the forest plot for meta-analysis of studies comparing mitomycin C versus no mitomycin C as regards the incidence of restenosis. CI, confidence interval.
were included; they involved 165 cases, managed by transnasal endoscopic and transpalatal approach. In total, 93 cases used nasal stent after the repair of CA, whereas in 72 cases stent was not used. Rate of restenosis in cases for which nasal stent was used was 21.5%, whereas in cases without nasal stent the rate of restenosis was 18% (Table 6).
Meta-analysis of studies comparing stent insertion versus no stent insertion after CA repair, as regards the incidence of restenosis using the estimated odds ratio (statistical measure), showed a statistically nonsignificant difference between the two modalities. The measures of heterogeneity revealed unimportant heterogeneity of the estimates reported by the included studies (Cochran Q P-value, 0.874; I 2 , 0.000; τ 2 , 0.000).

Discussion
Many surgical approaches have been described for CA repair; among them are the transnasal, transpalatal, trans-septal, and transantral routes. Although their success rate are almost equal, the most widely used techniques are the transnasal and transpalatal routes [18].
A study by Pirsig [19] reviewed over hundreds of papers on surgical approaches to CA. Many authors found that the results of both transnasal and transpalatal approaches are comparable (80 and 84% success rates, respectively). This is in agreement with our study, as we found that the incidence of restenosis after transnasal endoscopic or transpalatal repair, using the estimated odds ratio and risk ratio (statistical measures), showed a statistically nonsignificant difference between the two techniques.
On the other hand, a study by Bergonse et al. [33] reported that despite current development and improvement of endoscopic technique transpalatal technique, when performed properly by a skilled surgeon, the result is satisfactory. A study by Gujrathi et al. [34] suggested that endoscopic repair is difficult and should not be preferred in neonates because of narrow nasal cavities. Moreover, Schraff et al. [35] preferred transpalatal approach as the intervention of choice in CA repair.
During studying the results of meta-analysis of studies comparing MMC versus no MMC as regards the incidence of restenosis using the estimated odds ratio (statistical measure), we found that there was no statistically significant difference between the two groups.
Al-Ammar [11], Kubba et al. [13], and Newman et al. [15], in their respective studies, agreed with our outcome that there is no statistically significant difference between using MMC or no MMC as regards the incidence of restenosis. In addition, a study by Teissier et al. [36] [39] found that using topical MMC as an adjuvant treatment to CA repair has great effect in improving the outcome.
Finally, we aimed to find out the role of stenting in improving the outcome of CA repair.  Forest plot for meta-analysis of studies comparing stent insertion versus no stent insertion as regards the incidence of restenosis. CI, confidence interval.
The meta-analysis of studies comparing stent insertion versus no stent insertion as regards the incidence of restenosis using the estimated odds ratio (statistical measure) showed statistically nonsignificant difference between the two modalities.
Wiatrak [10] and Romeh and Albirmawy [28] found that using nasal stent as adjuvant tool to CA repair improves the outcome.
On the other hand, studies by Josephson et al. [44], Schoem [31], Van Den Abbeele et al. [29], Wang et al. [45], and Elsherif et al. [16] showed that the stents may act as a nidus for infection and cause pain, in addition to the formation of granulation tissue, and nasal synechia. Thus, using stent as an adjuvant tool to CA repair does not improve the outcome.

Conclusion
Although there is much debate about which approach is better and preferable in the repair of CA, the available evidence from this statistical analysis indicates that there is no statistically significant difference between transnasal endoscopic and transpalatal approach in the repair of CA as regards restenosis. In addition, we found that there is no statistically significant difference between using intraoperative topical MMC and nasal stent or not using such modalities in the outcome of CA repair.

Financial support and sponsorship
Nil.

Conflicts of interest
There are no conflicts of interest.