Are tongue flaps effective in the closure of palatal fistulas? A systematic review and meta-analysis

Search of the literature

Our search strategy yielded 587 articles. Of these, 150 were duplicate studies, and 437 were screened. Of these studies, 392 studies were excluded based on their titles and abstracts. 45 records were assessed for eligibility in which 29 were excluded as they did not meet the inclusion criteria. Finally, 16 studies met the criteria for inclusion, and were included (Fig. 1).

Fig. 1

Risk of bias quality assesment (RoB)

All of the included articles have shown a low level Bias (Table 3).

Certainty of the evidence

Observational studies are associated with a high quality of evidence. This indicates a high confidence in the effect estimate (Table 4).

Egger’s test for small study effects

The results of egger’s test revealed absence of small study effects across the included studies in the meta-analysis (z value = 0.39, standard error = 2.588, p = 0.6992) (Table 5).

Table 5 Regression based Egger test for small-study effects H0: beta 1 = 0; no small-study effectsPublication bias

The contour-enhanced funnel plot suggested that the studies included in this meta-analysis were not strongly influenced by small-study effects or publication bias, as the studies were relatively evenly distributed within the funnel-shaped region. The lack of asymmetry in the plot and the location of the estimated overall effect near the peak of the funnel provide support for the validity of the findings computed from the meta-analysis (Fig. 2).

Fig. 2

Study characteristics

16 studies were included in the current review, Including 13 retrospective [21,22,23,24,25,26,27,28,29,30,31,32,33] and 3 prospective [34,35,36] studies. Totalling 461 patients that underwent tongue flap surgery for the closure of palatal fistula. Size of the fistulas ranged from 1 cm upto 3 cm in diameter with the most common location of the fistula was the anterior palate. Follow up period ranged from 3 weeks to 2 years.

Succes rate of tongue flap reconstruction of Oro-Antral fistulas

Out of 461 flaps performed 359 flaps were successful in the closure of the oro-antral fistula with a total success rate of 77.8% with 102 (22%) flaps failed in the closure of the fistula, with an overall success rate of 0.92 (95% CI: 0.81, 0.95), with significant heterogeneity observed across studies (I2 = 32.86%, p < 0.001). A random effects model was used to account for this heterogeneity (Fig. 3) (Table 5).

Fig. 3

Success Rate of Fistula Closure

Dorsal tongue flaps (DTF) vs. posterior tongue flaps (PTF)

Out of the 16 included studies, only one study (Mhajan et al., 2018) [28] did not report the type of tongue flap used in reconstructing the palatal fistula.

301 patients had dorsal tongue flaps, in which 239 (79.4%) flaps were successful in the closure of the fistula with 62 (20.5%) flaps failed. Out of 301 patients treated with DTF 70 patients (23.2%) experienced complications following the reconstruction of the palatal fistulas using DTF including decrease in hypernasality (16 patients), flap dehiscence (15), total flap detachment (12 patients), partial flap dehiscence (8 patients), total flap dehiscence (7 patients) recurrence of the fistula (7 patients), total flap necrosis (4 patients), complete flap rejection (3 patients), partial flap rejection(3 patients), bleeding from the flap (3 patients), temporary anterior venous congestion (2 patients), partial flap necrosis (2 patients), regurgitation of food/ fluids (2 patients) and sloughing (1 patient) 7 patients had posterior tongue flaps, in which 6 (85.7%) flaps were successful in the closure of the fistula with only 1 flap have failed. Out of 7 patients treated with PTF 6 patients experienced tongue assymetry following the procedure (Table 2). Tongue flap used for fistula closure revealed that the group mean proportion for studies using a dorsal anterior tongue flap was 0.91 (95% CI: 0.80, 0.95). Only one study assessed success rate of palatal fistula closure using a posterior lateral tongue flap which was computed to be 1.00 (95% CI: 0.76, 1.00). There was one study that did not specify the type of tongue flap used, with proportion of success rate computed to be 0.96 (95% CI: 0.90, 0.99).

Incidence of flap detachment

Flap detachment was reported as an observed complication after palatal fistula closure only in three of the included studies (Durmus Kocaaslan et al. 2020) [22], (Vasishta et al. 2012) 32] (Gupta et al.2020) [34]. The overall population mean proportion of flap detachment was computed to be 0.03 (95% CI: 0.02, 0.05), as a fixed effects model was used to pool the results due to the low heterogeneity observed across the included studies (I2 = 3.14%, p = 0.24) (Fig. 4), (Table 6).

Fig. 4

Table 6 Likelihood ratio test to check fit of random effects and the fixed effects modelIncidence of flap dehiscence

Flap dehiscence was reported as an observed complication after palatal fistula closure in five studies, (Barazarte et al. 2020) [26], (Mahajan et al. 2018) [28] (Prakash et al. 2018) [29], (Giugliano et al. 2022) [31] (Vasishta et al. 2012) [32]. The overall population mean proportion of flap dehiscence is 0.04 (95% CI: 0.01, 0.21), with significant heterogeneity observed across studies (I2 = 10.55%, p < 0.001); a random effects model was used to account for this heterogeneity (Fig. 5) (Table 6).

Fig. 5

Flap necrosis

Partial necrosis of the flap was reported as an observed complication after palatal fistula closure in four studies: (Busic et al. 1989) [21] (Argalle et al. 2023) [30], (Gupta et al. 2020) [34], (Sodhi et al. 2013) [36]. The overall pooled proportion of partial necrosis of the flap across all studies is 0.01 (95% CI: 0.01, 0.03) as a fixed effects model was used due to the low heterogeneity observed across the included studies (I2 = 7.44%, p = 0.22) (Fig. 6) (Table 6).

Fig. 6

Postoperative bleeding

Post-operative bleeding was reported as an observed complication after palatal fistula closure in four studies: (Alsalman et al. 2016) [27], (Prakash et al. 2018) [29] (Assuncao 1992) [33], (Vasishta et al. 2012) [32]. The overall pooled proportion of post-operative bleeding across all studies is 0.01 (95% CI: 0.01, 0.03). A fixed effects model was used to calculate this pooled proportion, as the heterogeneity across studies was low (I2 = 6.18%, p = 0.09) (Fig. 7) (Table 6).

Fig. 7

Recurrence of fistula

Recurrence of fistula was reported as an observed complication after palatal fistula closure in four studies: (Busic et al. 1989) [21], (Giugliano et al. 2022) [31], (Vasishta et al. 2012) [32], (Assuncao 1992) [33]. The overall pooled proportion of fistula recurrence across all studies is 0.02 (95% CI: 0.01, 0.03). A fixed effects model was used to calculate this pooled proportion, as the heterogeneity across studies was low (I2 = 8.41%, p = 0.12) (Fig. 8) (Table 6).

Fig. 8

Location of the oro-antral fistula in the outcome of tongue flap reconstruction

Location of fistula revealed that the group mean success rate for closure of fistula located in anterior palate was 0.92 (95% CI: 0.80, 0.95) and for fistulas involving both anterior and posterior palate was 0.94 (95% CI: 0.74, 0.99). The success rate for closure of fistula located in anterior palate (92%) was slightly lower than the success rate for closure of fistula involving both anterior and posterior palate (94%). However, the confidence intervals for the two subgroups overlap, suggesting the difference in success rates was not statistically significant (Fig. 9) (Table 7).

Fig. 9

Location of the Oro-Antral Fistula in the outcome of Tongue Flap Reconstruction

Oro-antral fistula size in the outcome of tongue flap reconstruction

The average length of the palatal fistula was positively associated with the overall success rate (coefficient = 0.151, p = 0.022). This suggested that patients with larger fistulas tend to have a better success rate. The average width of the palatal fistula was negatively associated with the overall pooled success rate, although the association was only marginally significant (coefficient = -0.106, p = 0.065). This indicated that studies with wider fistulas may tend to report lower success rate. The follow-up period was not significantly associated with the overall success rate (coefficient = -0.001, p = 0.475), suggesting that the length of follow-up does not have a significant impact on the reported effects. (Table 7).

Speech assesment scores, hypernasality and nasal emission outcomes post-operatively

After one month, the Bayesian analysis provided mixed evidence for the improvement in speech intelligibility, with anecdotal support for the alternative hypothesis (BF10 = 1.378) but wide confidence intervals indicating high uncertainty. There was anecdotal evidence for no effect on hypernasality (BF10 = 0.588), and moderate evidence for improvement in nasal emission (BF10 = 9.708). After six months, the analysis showed strong evidence for improvement in speech intelligibility (BF10 = 59.7398) and hypernasality (BF10 = 35.619), along with strong evidence for improvement in nasal emission (BF10 = 23.056). By one year, the evidence became very strong for the improvement in speech intelligibility (BF10 = 62.528) and hypernasality (BF10 = 62.528), and there was extreme evidence for the improvement in nasal emission (BF10 = 277.738). (Figs. 10, 11, 12, 13, 14, 15, 16, 17, 18).

Fig. 10