Systematic Review and Meta-Analysis of the Palatal Surgeries in the Treatment of Obstructive Sleep Apnea ()
1. Introduction
Charles Dickens in “The Posthumous Papers of the Pickwick Club” published the first description of a patient with obstructive sleep apnea in 1836 [1]. Subsequently, Guilleminault coined the term “upper airway resistance syndrome” in 1993 [2]. There are three types of sleep apneas: Central, Obstructive and Mixed. Central sleep apnea occurs when the brain does not send the signal to the muscles to take a breath. Obstructive sleep apnea happens when the brain sends the instruction to the muscles and the muscles make an effort to take a breath, but they are struggling because the airway becomes obstructed and prevents an adequate flow of air. Mixed sleep apnea is defined when there is both central sleep apnea and obstructive sleep apnea. Obstructive Sleep Apnea is a common condition affecting 4% of men and 2% of women [3]. Presenting symptoms are variable including restless sleep, snoring, excessive daytime somnolence, fatigue, memory loss, decreased cognitive function, mood changes, nocturnal enuresis, nocturnal sweating and observed choking or gasping at night. People with Obstructive Sleep Apnea (OSA) seek help for different reasons. At one end of the spectrum are patients presented with morbid obesity, carbon dioxide retention, hypertension, and right-sided heart failure with an Apnea Index of more than 100. At the other end of the spectrum are those individuals with minimal sleep-disordered breathing whose chief complaint is snoring [4].
We are presenting a systematic review for published articles between January 2008 and January 2019 on the different palatal surgeries performed on OSA patients and the treatment outcomes that could determine the effectiveness of various palatal surgical techniques in treating OSA including those being the most effective in treating OSA.
2. Patients and Methods
The current review followed the guidelines of preferred reporting items for systematic reviews and meta-analysis statement 2009 (PRISMA) [5]. The detailed steps of methods were described elsewhere as well as PRISMA checklist [6].
The quality of relevant studies was assessed using NIH quality assessment tool for.
Observational cohort and cross-sectional studies as well as NIH tool for quality assessment for case series studies. (Study Quality Assessment Tools) “National Heart, Lung, and Blood Institute (NHLBI)” 2019. Regarding cross section and cohort studies, each study was given a score out of 14 based on answering each question (Yes = 1, No = 0, NA = 0). A score of 10 - 14 indicated a good quality article, 5 - 9 for fair, and 1 - 4 for poor quality article. Regarding case series studies, total evaluation score was 9, a score from 7 - 9 indicated good quality article, whereas score from 4 - 6 for fair, and 1 - 3 for poor quality article.
Statistical analysis was performed as the following: prior to analysis phase, if the published study only reported the mean, the estimated standard deviation (SD) was derived from linear regression of log (published SDs) against log (published means) according to Van Rijkom, H. M., Truin, G. J., & Van’t Hof, M. A. (1998) [7]. The published SDs and means were collected from other includes studies. If there was one study with more than one publication reports, they were compared and the publications with complete data set were used. We made pairwise meta-analyses of our outcomes using Comprehensive Meta-Analysis software (CMA version 3) [8]. Standardized mean difference (SMD) with the corresponding 95% confidence intervals (95% CI) was also calculated. Pooled event rate with the corresponding 95% confidence intervals (95% CI) was also calculated. A fixed-effects model was used when there was no heterogeneity. Heterogeneity was assessed with Q statistics and I2-test considering it significant with I2 value > 50% or p-value < 0.10. Publication bias was assessed with the Egger’s regression test [9] and represented graphically by Begg’s funnel plot when there were ten or more studies. Egger’s regression test p-value < 0.05 was considered significant. Whenever publication bias was found, the trim and fill method of Duvall and Tweedie was applied adding studies that appeared to be missing to enhance the symmetry [10].
3. Results
3.1. Literature and Study Characteristics
As depicted in Figure 1, electronic search yielded 1036 articles from two databases. After 121 duplicates were removed, 863 articles were screened in title/abstract screening, while 377 articles were screened in full text screening for inclusion. Finally, 52 articles were included in the qualitative and quantitative meta-analysis synthesis. The manual search resulted in no additional studies. Detailed characteristics of the included studies are shown in Table 1. The surgical steps of the main palatal techniques encountered during out systematic review are outlined in Table 2.
3.2. Risk of Bias Assessment
With regard to quality assessment, 44 of 52 studies evaluated showed good quality, whereas the other eight studies showed fair quality (Table 1).
3.3. Outcomes
1) AHI (Apnea/Hypopnea Index)
![]()
![]()
Table 1. Characteristics of the studies included in the meta-analysis.
UPF: Uvulopalatal Flap; LP: Lateral Pharyngoplasty; RP: Relocation Pharyngoplasty; MEUP: Microdebrider Assisted Extended Uvulopalatoplasty; ESP: Expansion Sphincter Pharyngoplasty; UPPP: Uvulopalatopharyngoplasty; LPMR: Limited Palatal Muscle Resection; SPFP: Soft Palatal Webbing Flap Pharyngoplasty; EP + SS: Expansion Pharyngoplasty with Suspension Sutures; PPR: Partial Palatal Resection; ZP: Z-Palatoplasty; MESP: Modified Expansion Sphincter Pharyngoplasty; DSS + ESP: Double Suspension Sutures + Expansion Sphincter Pharyngoplasty; MRTA: Modified Radiofrequency Tissue Ablation; FEP: Functional Expansion Pharyngoplasty; H-UPPP: Han’s Uvulopalatopharyngoplasty; Alianza Technique: BAP + BRB; SPW: Soft Palatal Webbing Flap; SP: Suspension Palatoplasty; BRP: Barbed Reposition Pharyngoplasty; BESP: Barbed Expansion Pharyngoplasty; BAP: Barbed Anterior Palatoplasty; BRB: Barbed Roman Blinds.
![]()
Table 2. Surgical steps of the major palatal surgical techniques.
![]()
Figure 1. PRISMA flow diagram of the search and review process.
Meta-analyses of relevant studies showed that the surgical technique that achieved the best reduction on AHI was the lateral pharyngoplasty followed by the anterior palatoplasty, with a significant mean reduction of [(SMD = −0.848, 95% CI (−1.209 - − 0.487), p-value < 0.001) and (SMD = −0.864, 95% CI (−1.234 - −0.494), p-value < 0.001), respectively] (Figure 2). On the other hand, the surgical technique that achieved the least reduction on AHI was barbed roman blinds, with a significant mean reduction of [SMD = −4.956, 95% CI (−6.218 - −3.693), p-value < 0.001] (Figure 2). The mean reduction of AHI for the classical Uvulopalatopharyngoplasty (UPPP) was [SMD −1.466, 95% CI (−1.880 - −1.052), p-value < 0.001] (Figure 2). Random model was used, as heterogeneity was significant with I2 = 87.814, and p-value < 0.001.
2) ESS (Epworth Sleepiness Scale)
Meta-analyses of relevant studies showed that the surgical technique that achieved the best reduction on ESS was the relocation pharyngoplasty, with a significant mean reduction of [SMD = −0.998, 95% CI (−1.253 - −0.743), p-value < 0.001] (Figure 3). On the other hand, the surgical technique that achieved the worst reduction on ESS was barbed roman blinds, with a significant mean reduction of [SMD = −2.448, 95% CI (−3.149 - −1.827), p-value < 0.001] (Figure 3). The mean reduction of AHI for the classical UPPP was [SMD −1.388, 95% CI (−1.855 - −0.922), p-value < 0.001] (Figure 3). Random model was used, as heterogeneity was significant with I2 = 87.814, and p-value < 0.001.
3) Minimal Oxygen Saturation Percentage
![]()
Figure 2. Meta-analysis for the changes in AHI regarding the different surgical techniques.
![]()
Figure 3. Meta-analysis for the changes in ESS regarding the different surgical techniques.
Meta-analyses of relevant studies showed that the surgical technique that achieved the most reduction in O2 saturation was the expansion sphincter pharyngoplasty, with a significant mean reduction of [SMD = 1.011, 95% CI (0.581 - 1.440), p-value < 0.001] (Figure 4). On the other hand, the surgical technique that achieved the worse reduction in O2 saturation was relocation pharyngoplasty, with a significant mean reduction of [SMD = 0.524, 95% CI (0.315 - 0.733), p-value < 0.001] (Figure 4). The mean reduction of AHI for the classical UPPP was [SMD 0.990, 95% CI (0.565 - 1.414), p-value < 0.001] (Figure 4). Random model was used, as heterogeneity was significant with I2 = 82.776, and P-value < 0.001.
4) VAS (Visual Analogue Scale) for Snoring
Meta-analyses of relevant studies showed that the surgical technique that achieved the best reduction in VAS was Z-Palatoplasty, with a significant mean reduction of [SMD = −1.551, 95% CI (−2.049 - −1.052), p-value < 0.001] (Figure 5). On the other hand, the surgical technique that achieved the worse reduction in VAS was expansion sphincter pharyngoplasty, with a significant mean reduction of [SMD = −4.579, 95% CI (−5.381 - −3.788), p-value < 0.001] (Figure 5). The mean reduction of VAS for the classical UPPP was [SMD −2.560, 95% CI (−3.059 - −2.061), p-value < 0.001] (Figure 5). Random model was used, as heterogeneity was significant with I2 = 94.172, and p-value < 0.001.
5) Soft Palate Length
Meta-analyses of relevant studies showed that the UPPP surgical technique achieved soft palate length change with a significant mean reduction of [SMD = −2.219, 95% CI (−2.730 - −1.708), p-value < 0.001] (Figure 6). While, relocation pharyngoplasty achieved reduction in soft palate length, with a significant mean [SMD = −1.549, 95% CI (−2.063 - −1.035), p-value < 0.001] (Figure 6). Random model was used, as heterogeneity was significant with I2 = 96.706, and p-value < 0.001.
6) Success Rate
Meta-analyses of relevant studies showed that expansion sphincter pharyngoplasty achieved the overall highest significant success rate [Event rate = 77%, 95% CI (65.4% - 85.5%), p-value < 0.001] (Figure 7). While, Z-Palatoplasty achieved the least success rate [Event rate = 52.5%, 95% CI (20.1% - 82.9%] (Figure 7). Random model was used, as heterogeneity was significant with I2 = 73.015, and p-value < 0.001.
4. Discussion
OSA management can be divided into surgical and non-surgical with the earliest surgical technique being the UPPP. Subsequent modifications to the original technique and introduction of novel techniques have all been aimed at achieving consistently higher success rates and fewer complications.
This study included papers published in the last 10 years, starting in 2008. The reason being that the majority of the techniques with the exception of the Lateral Pharyngoplasty and modifications on the original UPPP, were introduced starting 2007. Therefore, including earlier papers would have possibly skewed the results in favor of older techniques.
![]()
Figure 4. Meta-analysis for the changes in minimal oxygen saturation percentage regarding the different surgical techniques.
![]()
Figure 5. Meta-analysis for the changes in VAS for snoring regarding the different surgical techniques.
![]()
Figure 6. Meta-analysis for the changes in soft palate length regarding the different surgical techniques.
![]()
Figure 7. Meta-analysis of the success rate regarding the different surgical techniques.
The success of palatal surgical techniques in treating OSA is graded according to improvements in different parameters. These namely being the AHI, ESS, Oxygen desaturation index & VAS for snoring. Overall success in the majority of the papers was defined using Sher’s Criteria introduced by Sher & Colleagues in their 1996 publication [11]. The criterion states that surgical success for a given procedure is defined as 50% or more decrease in the AHI with a post-operative AHI less than or equal to 20/ hour. Based on Sher’s criteria, our results show that the Expansion Sphincter Pharyngoplasty achieved the highest success rate. Our findings are consistent with those of other authors.
In 2018, Pang et al. published a Systematic review and meta-analysis which concluded that ESP resulted in better post-operative outcomes than other traditional surgeries, namely the Anterior Palatoplasty, Classical UPPP and Uvulopalatal Flap procedures [12]. It should be noted, however, that this result is by no means universal, in the sense that not every patient suffering from OSAS will respond to ESP. There is no one-size-fits-all.
After reviewing the literature, several points of note have been identified:
- There is a general consensus that UPPP is no longer the best option for treating OSA. It is associated with significant post-operative pain and out of the previously discussed palatal surgeries has the highest risk of post-operative bleeding. Furthermore, UPPP can result in the much-disliked complication of velopharyngeal stenosis due to scar contracture according to Katsantonis, G. P., Friedman, W. H., Krebs, F. J., & Walsh, J. K. (1987) [13].
- BRP is a quicker and easier technique to perform than the ESP and fully respects the anatomical integrity of the palate. It is also the technique associated with the least blood loss. The success of the BRP and ESP techniques stems mainly from the widening of the lateral retro-palatal diameters. That being said, despite the ESP achieving the highest success rate in our analysis, the ESP procedure has been associated food residues in the pyriform fossae due to interruption of the pharyngeal muscles that protect them [14].
- The Barbed roman blinds technique tensions the lateral wall without interrupting palatopharyngeal muscles unlike FEP and ESP and this tension occurs along the whole length of the thread. Barbed suturing allows muscles to be moved along different vectors (anteriorly, posteriorly, laterally) as required based on DISE evaluation.
- Pre-operative evaluation and identification of the site of obstruction is critical to optimizing post-operative results. DISE has proven to be the best diagnostic procedure suited to this purpose. By correctly locating the site of collapse the surgical technique that best addresses that site can be chosen and techniques which do not adequately address the pathological site can be avoided, even if it means forgoing palatal surgery altogether because the collapse is hypopharyngeal for example.
- There isn’t always a single site of airway collapse and hence more than one surgery might be needed at different levels (multi-level surgery) to achieve a response.
- OSA severity can be graded as mild, moderate or severe. In many of the papers reviewed, authors would initially include only those patients with mild to moderate OSA. The main reasoning behind this was that they realized that the surgery they were proposing would not achieve adequate results in severe OSA patients. Nevertheless, this subsequently results in a distorted outcome.
- Some patients suffer from positional OSA. These patients tended to be in the mild to moderate group of OSA sufferers, whereas non-positional OSA subjects usually had severe OSA [15]. Consequently, positional OSA individuals tended to have higher post-operative success rates than their non-positional counterparts.
- Each procedure has its claimed advantages, however, objectively speaking, the best procedure would undoubtedly be the one that achieves the best post-operative results whilst minimizing palatal anatomical distortion and post-operative complications.
5. Conclusion
The best procedure for treating OSA varies from patient to patient and there is no universal cure-all. Careful patient selection and pre-operative evaluation are mandatory.