Several reasons have been identified for failure of primary FESS. Kennedy  noted that patients with bilateral ethmoid disease and additional disease in 2 or more dependant sinuses on each side, as well as patients with diffuse polyps, had significantly worse outcome after FESS than patients with less severe sinus disease. Lazar et al  found that fibrosis and adhesion formation, particularly between the middle turbinate and the lateral nasal wall, was the most common intraoperative findings in revision FESS. This was found in 43% of their patients. Recurrence of polyps was the second commonest finding, occurring in 22% of patients. Other causes for failure of primary FESS include lateralisation of the middle turbinate, frontal recess obstruction, recirculation between the natural ostium of the maxillary sinus and the antrostomy, persistent uncinate process, persistent agger nasi cells, severe septal deviations and devitalised bone .
Very few articles attempted to identify the incidence of the various anatomic findings in patients undergoing revision FESS. Musy and Kountakis  found that the most common anatomic factor associated with primary sinus surgical failure was lateralisation of the middle turbinate, occurring in 78% of their patients. Their results also showed residual anterior ethmoidal cells in 64% of patients. Scarred frontal recesses were found in 50% of patients. Residual posterior ethmoidal cells were present in 41% of patients. Residual agger nasi cells were present in 49% of patients. Residual uncinate processes were present in 37% of patients. Finally, middle meatal antrostomy stenosis was present in 39% of patients. In another study by Ramadan , the most common anatomic finding during revision FESS was adhesions, often involving a lateralised middle turbinate. This occurred in 56% of patients. This investigator also detected residual ethmoidal cells in 31%, middle meatal antrostomy stenosis in 27% and frontal sinus ostium stenosis in 25% of patients undergoing revision FESS.
In comparison with Musy and Kountakis' findings , the current study showed a noticeably higher incidence of residual cells. Our data show that 96% of the studied sides (96.8% of patients) had residual posterior ethmoidal cells, 96% of the sides (95.2% of patients) had residual frontal recess cells and 92.1% of the sides (92.1% of patients) had residual other anterior ethmoidal cells. On the contrary, lateralisation of the middle turbinate was only detected in 11.1% of the sides (17.5% of patients), which was significantly less than Musy and Kountakis' figure of 78% . These results may reflect the more conservative FESS techniques practised by the majority of the surgeons in the U.K, in comparison with the practice in the U.S. It is of course to be argued that removal of all cells is not required in the majority of FESS procedures, and that the procedure has to be tailored to the extent of the pathology. However, the majority of the patients in the current study had pansinusitis, as can be seen from the Lund-Mackay  scoring of the involved sinuses, where only 2.4% of the maxillary sinuses, 7.1% of the anterior and posterior ethmoids and 22.2% of the frontal sinuses were non opacified. It is therefore reasonable to assume that the majority of these patients needed more aggressive surgical dissections than what they had during the primary surgery.
The current study showed that 57.1% of the sides (60.3% of the patients) had a residual uncinate process. Chiu and Kennedy  advised that identifying an uncinate process remnant was the most critical step in revising a middle meatal antrostomy. They also commented that residual Haller (infraorbital) cells could be a source of persistent obstruction of the maxillary sinus. The latter cells were found in 23% of the sides (25.4% of the patients) in the current study. However, recently some studies have advocated preservation of the uncinate process due to its role in protecting the sinuses from allergens and contaminated inspired air [11, 12].
The above discussion highlights the fact that there is a wide variation in the practice of endoscopic sinus surgery. Some surgeons prefer more conservative techniques. Recently, the principle of minimally invasive sinus technique (MIST) has been introduced [13, 14]. It is claimed that this entails a standardised conservative endoscopic technique that can be applied to all patients requiring sinus surgery, regardless of the extent of their pathology. Other authors, however, have disapproved of the principles of MIST . On the other extreme, some surgeons prefer radical endoscopic surgical techniques to treat advanced inflammatory sinus pathology. Such techniques may involve total sphenoethmoidectomies with extensive mucosal resection , and may even involve middle turbinate resection as well [17–19].
We have not attempted in our study to investigate the clinical outcome after revision surgery as the aim of the study was to identify the residual anatomic factors that may result in recurrent rhinosinusitis after primary surgery, and to reflect on the practice of FESS in the U.K. We hope, however, that this work will stimulate further studies to compare conservative versus more aggressive sinus surgery techniques, and to answer the question of how extensive sinus surgery should be.