ABSTRACT
Uvulopalatopharyngoplasty [UPPP] is the most commonly performed surgical procedure for obstructive sleep apnea [OSA]. However, a successful outcome of UPPP is not simply correlated with the extent of surgery. The main problem is the selection of patients likely to benefit from UPPP. This study was performed aiming to find definite preoperative predictive factors for the outcome of UPPP. This was done through retrospective analysis of the preoperative anthropometric data and magnetic resonance imaging [MRI] measures of the upper airway [UA] dimensions. Thirty patients [28 males and 2 females] had a detailed preoperative overnight polysomnography [PSG], awake fibroptic nasopharyngoscopy with Muller maneuver [FNMM] and awake MRI scan of the UA. Patients were selected using FNMM according to the previous traditional criteria of selection. Through MRI examination 16 linear, areal and volumetric variables have been measured for each patient. Six weeks after UPPP, clinical examination, FNMM and PSG were repeated. The success rate of UPPP was 56.7%. We did not find any significant value for the FNMM in predicting the outcome of UPPP. However, the body mass index [BMI] was significantly higher in poor responders than in good responders [P<0.05]. On the other hand, the UA cross-sectional area [CSA], oropharyngeal CSA and hypopharyngeal CSA were significantly s mailer in the poor responders [P<0 05]. There was also significant difference in the volumes of the oropharynx, hypopharynx, total UA and the tongue. Moreover, there was a highly significant difference in ratio of UA volume/tongue volume [P<0.02]. The latter was smaller in the poor responders. There was also a highly statistically significant difference between good responders and poor responders in the ratio of orohypopharyngeal volume/tongue volume [OH-V/T-V] which was much reduced in poor responders [p<0.001]. All these variables were presented in the equation of forward stepwise multiple linear regression analysis as independent variables in corresponding to the degree of improvement. OH-V/T-V and the preoperative BMI were found significantly correlated with the outcome of UPPP. It was found that 46% of the variation in postoperative improvement could be explained by the variation in OH-V/T-V and the BMI collectively. The correlation between OH-V/T-V and the success was positive indicating that the smaller the ratio, the smaller the degree of success. However, the correlation between the BMI and the results was negative i.e. the larger the BMI the lower the success rate. A low OH-V/T-V ratio below 0.07 and a high BMI more than 30 Kg/m[2] were noticed to be nearly cut-off predictive values for the postoperative failure of UPPP. MRI seems to be worthy in the selection of UPPP candidates. The literature and the present study indicated that no single variable could safely predict the outcome of UPPP, but that a combination of anthropometric and anatomical variables is needed. We expect that our model with BMI and OH-VT-V could increase considerably the degree of prediction of the outcome of UPPP