Correlation of nasal symptoms with objective findings and surgical outcome measurement

Thesis submitted for the degree of Master of Surgery, University of London, 1993.
Published (excluding Chapter 9) 1996.
Recompiled HTML format June 2007
© 1993 – 2019 JW Fairley

Mr James W Fairley BSc MBBS FRCS MS

Consultant ENT Surgeon

Chapter 9
A prospective randomized controlled trial of Functional Endoscopic Sinus Surgery: Endoscopic middle meatal antrostomy versus conventional inferior meatal antrostomy. Interim results.

Chapter 9 Contents

Thesis Contents


To compare results of functional endoscopic middle meatal antrostomy (FESS) with conventional intranasal inferior meatal antrostomy (INA), 33 patients with chronic or recurrent acute rhinosinusitis were randomly allocated to receive either FESS or INA. All patients had failed adequate medical treatment with antibiotics, topical steroids and antral washouts. Patients with previous sinus surgery or significant septal deviations were excluded. In order to minimize variation, I carried out all procedures myself. Nasal symptoms were scored on a validated questionnaire pre operatively, six weeks, three months, six months and one year post operatively. Technical success was judged by endoscopic evidence of continued patency of antrostomy. Results on 12 FESS and 17 INA patients, most of whom have been followed for over 1 year, show that subjective symptom scores were improved in both groups (p < 0.001), but there is no difference in outcome between FESS and INA. Endoscopic operations took approximately three times longer than INA (mean 57 vs. 21 minutes, p < 0.001). There were no major complications. The study is continuing. A power analysis suggests that at least 90 patients will be necessary to give an 80% chance of demonstrating a 50% outcome difference in favour of FESS. If the true difference between FESS and INA is less than this, larger numbers will be required. It is hoped that the study may be continued as a multicentre trial.


This study was designed to answer some important questions about functional endoscopic sinus surgery (FESS). This new technique and pathophysiological concept is the most significant advance in rhinology for decades. It is changing the way we think about, diagnose and treat conditions of the nose and sinuses. The introduction of the rigid nasal endoscope in the 1990’s has been compared to the introduction of the operating microscope into ear surgery in the 1960’s.

Endoscopic sinus surgery was developed in continental Europe in the 1970’s and 80’s (Messerklinger, 1978; Wigand et al, 1978; Draf, 1983; Stammberger, 1985,1986; Stammberger & Wolf, 1988; Hosemann et al, 1989). From 1985 onwards it was taken up with great enthusiasm in the United States (Kennedy, 1985, 1987; Rice, 1989a; Toffel et al, 1989;Schaefer, 1989;Schaefer and Close, 1990;Levine, 1990).
The first courses on the technique in the UK were held in 1988 in Birmingham and London. At the time this study was designed (1988/89), functional endoscopic surgery of the sinuses was poised to mushroom in Britain (East, 1989; Lund, 1990). In 1993 it has established a good foothold, particularly for diagnostic use (Mackay, 1992; Kaluskar and Patil, 1992) and is certain to grow further.

Despite great claims for the technique, in 1989 there had not been any controlled trial of FESS versus conventional surgery. Success rates of around 90% have been reported using the old “unphysiological” operation of inferior meatal antrostomy in cases of recurrent acute and chronic sinusitis (Lund 1988). Reported success rates for endoscopic surgery are similar (Kennedy et al, 1987; Buiter, 1988; Kamel, 1989; Rice, 1989b; Stammberger and Posawetz, 1990). Controversy over the middle meatal approach to the maxillary sinus has raged for over 200 years (Jourdain, 1767, cited by Myerson, 1932; Sluder, 1918; Skillern, 1923; Macbeth, 1968, 1971; Lavelle & Harrison, 1971; Friedman, 1986; Kennedy, 1986). Out of 364 papers on the surgical treatment of sinusitis between 1966 and 1989 there was only one controlled trial (Arnes et al, 1985), and no studies which met all the standards published by the British Medical Journal for reports on the results of treatment (Anon, 1988). The lack of controlled trials led Buiter (1988) to conclude that there are no reliable statistics that show any clear advantage for one form of treatment over another.

Arnes et al (1985) found no difference in their trial of inferior versus middle meatal antrostomy, but there were only 38 patients in this study and a type II statistical error cannot be excluded. Since the present study commenced, one randomized controlled trial of FESS against Caldwell-Luc has been reported from Finland (Penttila et al, 1992) in which the FESS group (n=72) did marginally better and had fewer side-effects than the Caldwell-Luc group (n=71). Maran has presented some data on a comparative trial of FESS versus INA (British Academic Congress in Otolaryngology, Dublin, 1991) indicating that the FESS group did better on the symptom of nasal stuffiness, but as yet there is no published follow-up of this work.

Instrumentation for FESS is costly. Expensive CT scans are needed. The endoscopic operations take longer than conventional surgery. Reports of complications include cerebrospinal fluid leaks and blindness, possibly due to inadequately trained surgeons taking up the technique (Stankiewicz, 1989a,b). There is a considerable learning curve, and formal cadaver training comparable to temporal bone dissection is a prerequisite before operating on patients.

The appeal of FESS is twofold. Firstly, it is a technological advance. Secondly, it fits in with our new understanding of pathophysiology. It appears logical. Since it has been shown that mucociliary clearance from the sinuses follows relatively constant pathways, and that obstruction to drainage in the “key areas” of the anterior ethmoid and ostiomeatal complex has secondary effects on the dependent maxillary and frontal sinuses (Stammberger, 1986), it appears logical to concentrate therapeutic attention on these key areas.

However there are many examples in medical history of the misapplication of theoretical deductions into practice (Fairley, 1991). It is entirely possible that what we now consider to be pathophysiological fact will be overturned by future discoveries, and action based on logical deduction from what we know at this stage will be considered invalid or even ludicrous.

From the point of view of a cash limited health service, it is important to try and quantify the costs and benefits of a new technique before it is adopted by default (Dept. of Health, 1992; Borst-Eilers, 1993).

This study was designed to determine the relative benefits and costs of FESS versus conventional sinus surgery. The trial was carried out in accordance with the revised declaration of Helsinki, 1975, following approval of the local ethical committee.


Patients with a clinical diagnosis of rhinosinusitis were referred from all ENT consultants in Sheffield to a special nasal research clinic. Patients for the trial received a full diagnostic evaluation, including

  • symptom scoring questionnaire
  • nasal peak inspiratory flow measurement
  • saccharin test of mucociliary clearance
  • rigid nasal endoscopy

Blood tests included

  • full blood count
  • ESR
  • Immunoglobulin E levels (for allergy)
  • IgG,A & M for deficiency
  • IgG subclass analysis

Eligible patients were

  • 16 years or over
  • with chronic or recurrent acute sinusitis
  • bilateral or unilateral
  • with or without nasal polyps
  • in whom conventional inferior meatal intranasal antrostomy was indicated.

The diagnosis was made on the basis of the history, together with
rhinoscopic evidence of nasal abnormality, including

  • nasal polyposis
  • mucosal oedema
  • pressure contact areas
  • mucopus

and/or plain X Ray evidence of sinusitis – opacification or mucosal thickening in the antrum, ethmoid, frontal or sphenoid sinuses.

Medical treatment consisting of

  • oral antibiotics
  • topical intranasal steroids for at least 6 weeks, and
  • antral washouts

had to have been given and failed.

Exclusion criteria were

  • pregnancy
  • untreated acute or complicated sinusitis
  • untreated deviation of the nasal septum sufficient to make endoscopic sinus surgery technically impossible
  • previous intranasal antrostomy, Caldwell-Luc operation, ethmoidectomy or frontal sinus surgery
  • malignant disease of the nose or sinuses
  • unwillingness or inability to comply with trial procedures.

Randomization method

Following full informed consent, patients were selected at random using a sealed envelope system to receive either endoscopic middle meatal antrostomy or conventional inferior meatal intranasal antrostomy. After entry to the trial, patients underwent coronal computerised tomography based on Zinreich’s protocol (Zinreich et al 1987). This was to avoid bias in selection of patients for ethmoid surgery following the results of the scan, and to provide a proper comparison with current clinical practice, which is to make the decision to operate on the antrum without a scan.

Surgical techniques

In order to minimize variation, I carried out all surgical procedures myself, having had adequate training and experience in both techniques. Before operating on any trial patient, I had completed my basic learning curve for FESS, having been using the telescopes routinely in out-patients for over two years, carried out around 20 cadaver dissections, attended two international courses and done over 50 ethmoidectomies, mostly using the Messerklinger technique. I had been doing conventional nasal and sinus surgery for over 6 years.

Pre-operative topical cocaine was used in both groups. In the FESS patients, a 1ml of a 4% solution with adrenaline 1:5,000 was applied to each nasal fossa in the Moffett position, then neurosurgical linteens soaked in the same solution were inserted into the inferior, middle and superior meati. Further local anaesthesia and vasoconstriction was provided by injecting 1% lignocaine with adrenaline 1:200,000 into the uncinate process, and into the middle turbinate if resection was required. Sedation was administered by the anaesthetist, in most cases using intravenous midazolam. For the INA patients, 25% plain cocaine paste was painted onto the inferior turbinates and inferior meatus.

Functional endoscopic surgery comprised a Messerklinger infundibulotomy (Kennedy 1985; Stammberger 1986), with formal identification and enlargement of the natural maxillary ostium. Normally the uncinectomy would be the first step in the procedure, but if the uncinate was obscured by polyps these were removed first using the small Weil-Blakesley forceps. The ethmoidectomy was extended into the posterior cells in cases of polypoid disease by perforating the ground lamella of the middle turbinate. The middle turbinate was not routinely resected, but was removed if diseased or excessively mobile (tending to fall laterally) at the end of the procedure. Local anaesthesia with sedation was used in preference in the earlier patients, general anaesthesia was used more frequently as the study progressed. Inferior meatal antrostomy was carried out under general anaesthesia using the technique described by Lund (1988) aiming for a 2cm long by 1cm high opening flush with the nasal floor. These dimensions were checked using a Hegar dilator. Nasal polyps were avulsed when present, using Tilley Henckel forceps. Full length inferior turbinate reductions were done if hypertrophic and/or obstructing easy access to the inferior meatus.

Post-operative care

Patients were allowed home the day after surgery. Early on in the study, I carried out suction of blood clots from the nose on the first post-operative day, but this practice was discontinued because of the patients’ discomfort. If there was evidence of active infection at operation, patients were started on antibiotics. Both FESS and INA patients were reviewed in out-patients at one week. The nose was sprayed with cocaine 10% solution, and endoscopic suction clearance was performed. Inevitably, large solidified blood clots and crusts were removed, this often required the use of Tilley’s dressing forceps. Using curved cannulae, the antra were sucked clear, and any early adhesions were broken down. All patients were re-started at this stage on Betamethasone 1% drops, twice per day in the head down and forward position. Further antibiotics were prescribed depending on the endoscopic findings. The FESS patients were reviewed weekly with suction clearance until clots and crusting was minimal, typically this would involve three visits at 1, 2 and 3 weeks. INA patients were reviewed at 1 and 6 weeks only. Following this, patients were reviewed according to the protocol at 3, 6 and 12 months. Treatment with topical steroids was continued in most cases throughout the follow-up period.

Outcome measures

The primary outcome measure was the reduction in nasal symptoms. This was assessed by a validated questionnaire (described in Chapter 4 of this thesis) applied pre operatively, six weeks, three months, six months and one year post operatively. Secondary subjective outcome measures were patients’ opinions on whether the operation had made the nose better, their overall feeling about the results of treatment, and a “retrospectoscope” question as to whether, if they had the choice over again, they would have the operation. These were assessed by questionnaire one year post operatively. The technical “success rate” of the operations was assessed by endoscopic documentation of the patency of the antrostomies. Safety of the procedures was assessed by the incidence of a pre-defined list of major and minor complications. Major complications were defined as death, CSF leak, visual loss, orbital haematoma, and haemorrhage sufficient to require transfusion or return to the operating theatre. Minor complications included orbital ecchymosis, orbital emphysema, intranasal adhesions, stenosis of the ostium and neuralgia. The incidence of revision surgery was also compared between groups.

Cost analysis

Capital costs on equipment necessary for FESS were converted to an Annual Equivalent based on an assumed life of 5 years and using straight line depreciation. This was divided by an estimate of the number of patients treated per annum to reach a cost-per-case figure. The estimate assumes that FESS would be used in future for all suitable cases; this increases the projected number of cases of FESS and therefore reduces the capital element on equipment per case per annum. Revenue expenditure estimates are based on per case differences. The items which differ between FESS and INA patients in our practice are CT scans and use of operating time. Both groups of patients are treated as in-patients with overnight stay.

Sample size

I intended to recruit a total of 100 patients, 50 in each group, over an 18 – 24 month period, allowing a further year for follow-up of the last recruits. This was based on the following power analysis: Lund (1988) found a reduction of 5 points between the initial symptom scores and scores 6 months following inferior meatal antrostomy. Using the same scale, I chose arbitrarily to regard a 50% improvement as clinically worthwhile, i.e. a difference of 2.5 points between the new treatment and the old. The estimated standard deviation of this difference, based on Lund’s figures, is 3.46. This would require 45 patients in each group to have an 80% power of detecting a true difference of 2.5 points at the 5% level of probability. Having 50 patients in each group would allow for a 10% dropout rate. If the true difference between FESS and INA is less than 50%, then the difference is less worthwhile clinically and a larger number of cases would be required to demonstrate it.

Statistical methods

The major end point of this study is the reduction in aggregate symptom scores; these were compared by an unpaired t-test, following an examination of the distribution of the scores with Kruskal-Wallis Goodness-of-Fit test against a normal distribution to check that this was appropriate. Reductions in individual symptoms were also compared, using non-parametric tests suitable for ordinal data (Mann-Whitney U-test). Complications are very low and no formal statistical comparison between groups is required.


33 patients were entered into the trial between May 1990 and January 1993. There were 21 women and 12 men, mean age 47 years, range 24 to 73 years. 14 were allocated to receive FESS and 19 INA. There were no significant pre-operative differences in any relevant parameters between the two groups (Table 9.1). All except two patients were listed for bilateral surgery, there was one unilateral case in each group.

Table 9.1 Pre-operative comparison of patients in FESS and INA groups

(2 tail)
Agemean (sd)52 (13.6)45 (13.0).130
Sex ratiof:m10:411:8.665
Polyposisn (%)7 (50%)9 (47%).839
Recurrent acute symptomsn (%)5 (36%)10 (53%).542
Total symptom scoremean (sd)19 (4.5)16 (5.3).121
Asthmaticn (%)4 (29%)7 (37%).901
Smokern (%)3 (21%)10 (53%).147
Allergy (Raised IgE)n (%)6 (43%)7 (37%).991
IgG subclass deficiencyn (%)1 (7%)0n/a
Nasal Insp Peak Flowmean (sd)111 (45.6)131 (52.6).266
Mucociliary impairment:

Saccharin clearance > 20min

n (%)5 (36%)8 (42%).991

Notes: 1. Some patients had both chronic and recurrent acute symptoms therefore rows add up to over 100%

2. Chi squared with Yates correction used to calculated p values for categorical data

One woman from the FESS group withdrew before operation because she no longer wanted any form of surgery, and one man from the INA group was withdrawn because he suffered a myocardial infarction while on the waiting list. The two most recently recruited patients, one FESS and one INA, are not included in this report because of the short follow-up times available. Results are therefore available on 29 patients, 12 of whom had FESS and 17 INA. The latest follow up data available was used, being one year or more for 21 patients, 6 months for 5 patients, 3 months for one patient and 6 weeks for two patients (both of these have defaulted from further follow-up and failed to respond to several letters, both had very low symptom scores and patent antrostomies when last seen).


Both FESS and INA patients had significant improvements in subjective nasal symptoms (p <0.001) but there were no differences between the FESS and INA groups (table 9.2, figure 9.1).

Table 9.2 Improvement in subjective nasal symptom scores at latest available follow-up.

Symptommeans.d.means.d.p (2 tail)
PostNasal Drip0.50001.50760.35290.9315.963
Facial pain1.50001.08711.00001.2247.273
Poor smell0.16671.69670.88241.1663.335
Sore throat0.83331.02990.52941.0676.674
Gen Unwell0.75001.21540.82351.1851.889
Sum total8.41677.01247.05886.8781.608

Note: p-values for individual symptoms calculated using Mann-Whitney U Test,
for total symptom score using unpaired t-test.

Figure 9.1 Plot of individual cases’ improvement in total symptom scores at latest available follow-up. (FESS n=12, INA n=17)

Improvement in nasal symptoms following FESS or INA: Plot of individual cases improvement in total symptom score at lastes follow-up. FESS n=12, INA n=17

Table 9.3 “Do you think the operation has made your nose..” 1 yr post op (n=21)

              FESS      INA
Better        10        8
No different   0        0

Table 9.4 “Overall, how do you feel about the results of your treatment?” 1 yr post op (n=21)

              FESS      INA
Pleased        8        7
Neither way    1        1
Unhappy        3        1

Table 9.5 “If you had the choice again, would you have the operation?” 1 yr post op (n=21)

              FESS      INA
Yes            8        8
No             4        1

Objective results: Patency of antrostomies

56 antrostomies were made in the 29 patients whose subjective results are reported above (there were 2 unilateral cases). 52 of these have remained patent (93%). One INA case is uncertain because of a large inferior turbinate obscuring the view, and three middle meatal antrostomies are not patent. All the definite failures were in the FESS group. Two of these were in the same patient. In both cases, the endoscopic procedure was abandoned before completing the antrostomy because of bleeding and/or pain, under local anaesthesia. The technical success rate of the surgery was therefore 20/23 or 87% in FESS and 32/33 or 97% in INA. Despite the technical failures, both FESS patients and the patient with the indeterminate inferior meatal antrostomy have had good subjective results.

Table 9.6 Incidence of Complications

Complication               FESS group     INA group
Death                         0              0
CSF leak                      0              0
Visual loss                   0              0
Orbital haematoma             0              0
Haemorrhage                   0              0
Orbital ecchymosis            0              0
Orbital emphysema             0              0
Intranasal adhesions          2              1
Stenosis of ostium            2              1
Neuralgia                     0              1

Revision surgery

Two INA patients and one FESS patient underwent revision surgery during the first year of follow-up. All three had diffuse polyposis involving the ethmoids and antral linings, with secondary purulent infection. In the two INA cases, the antrostomies became occluded with polyps, which were removed at revision surgery. Endoscopes were used at revision surgery, but middle meatal antrostomies were not made. Both these patients ended up as good results subjectively. The FESS case was complicated by adhesions between the middle turbinate and the lateral nasal wall, and the middle turbinate was excised at revision surgery. This patient continued to do badly, despite widely patent middle meatal antrostomies, and was the worst result overall in the FESS group.

Costs: Capital

At 1993 prices it costs approximately £12,000 for a set of three telescopes for out-patient diagnosis, a further set of three for operating, and a single set of surgical instruments. Based on a five year straight line depreciation, this is £2,400 per year. If it assumed that 60 patients per year will undergo FESS, the capital cost per case treated is £40.


The coronal CT scan of the paranasal sinuses using our modified Zinreich protocol currently takes around 15 minutes, and costs £80. The operating time was nearly 3 times longer in FESS cases, mean 57 minutes versus 21 minutes for INA (t = 9.43, df 26, p <0.001). Each FESS case therefore costs an additional 40 minutes operating time, currently priced at around £240 per hour, making £160.


One of the most surprising aspects of this trial was the difficulty in recruiting sufficient numbers of patients in the time available. Rhinosinusitis is common, and the catchment population is around 600,000. In the financial year 1988 89, 64 intranasal antrostomies were done in the Royal Hallamshire Hospital (data from hospital computer system). This would equate to 128 potential patients over a 2 year recruitment period, with one year allowed for follow up of the last recruits. Recruitment started in May 1990. By January 1993, I had assessed over 180 patients in the research clinic, but only 33 patients had been entered, and the rate of recruitment was falling. There were 3 main reasons for this:

  1. The high number of patients referred to the research clinic who were ineligible for the trial because of previous sinus surgery (chronic rhinosinopaths).
  2. The high success rate of simple medical treatment – principally topical steroids and antral washouts – in the “virgin” cases referred.
  3. Some loss of enthusiasm for referring patients for the trial when it became apparent that the requisite numbers were not likely to be achieved during the period of my appointment as a Senior Registrar.

In February 1993 I was appointed to a consultant post in Kent and therefore recruitment for the trial in Sheffield was terminated. These interim results are presented in the knowledge that the sample size is inadequate to exclude a Type II error. They will nevertheless benefit any future meta-analysis, and the considerable research effort that has been expended so far will not be wasted (Chalmers, 1990; Chalmers et al, 1992; Dept of Health, 1992).

Apart from the small numbers, the main reason for failing to demonstrate a difference between FESS and INA is that the INA group has done better than expected. In her 19 patients, Lund (1988) found a mean 5-point improvement in the total symptom score 6 months following intranasal antrostomy. I was looking to do 50% better than this with FESS to be clinically worthwhile, and therefore would have accepted a mean 7.5 point improvement in the FESS group. My FESS patients have in fact done better than this with an 8.4 point improvement, but there has also been a 7 point improvement in the INA group (Table 9.2). This amply illustrates the importance of using concurrent controls rather than historical data for comparison. Figure 9.1 plots the improvements in total symptom scores for individual patients in the FESS and INA groups. There is a wide range in both groups, and no obvious difference between them, although the mean improvement is slightly greater in the FESS group. The distribution of these scores was tested by a Kruskal-Wallis Goodness-of-Fit test, and did not differ significantly from a normal distribution. The t-test is therefore applicable. It confirms that the difference between the FESS and INA groups is not significant. The horizontal line at zero in Figure 9.1 marks the cut-off point for patients whose symptom scores are worse following surgery. It can be seen that one FESS patient and two INA patients had worse symptom scores after treatment.

Looking at the details of table 9.2, the main symptoms which do seem to show a trend to be improved more in the FESS patients are nasal obstruction, headache and facial pain. Epistaxis also has a low p-value, but this is not a prevalent symptom. Unlike Lund, I have not found that post-nasal drip is made worse by surgery, it is slightly improved in both groups. None of the symptoms on the questionnaire are made worse by surgery.

The responses of the patients to the three general outcome questions at one year make interesting comparison with the detailed results from the symptom questionnaire. The numbers here are even smaller, so no formal statistics have been calculated, but if anything it would appear that the FESS patients are less happy with the treatment than the INA group. 4 out of 12 FESS patients would not have the operation again, compared with 1 out of 9 INA (Table 9.5). A lot of this is due to dislike of the local anaesthesia, but also the patients dislike the postoperative endoscopic suction and cleaning of the cavities. This makes it doubly important to show whether there are significant benefits to be gained from FESS over INA, when advocating a form of treatment that may be less acceptable to patients.

The total additional cost per case of using FESS rather than conventional INA is approximately £280. It could be argued that we could save this by doing FESS as a day case (Danielsen, 1992), rather than our current practice of keeping the patients overnight, but if we were to do this for FESS we could probably also do it for INA. The major factor increasing the cost of FESS versus INA is the operating time. In my figures, the time taken for the local anaesthetic to take effect is included, because in most cases I was operating by myself. If an assistant is available to apply the local or to get on with the preceding case, the time could be reduced considerably. In practice, I have increasingly turned to general anaesthesia, and have found that the operating time for simple Messerklinger FESS has reduced to around 15 to 20 minutes per side. I have also found that, for a simple middle meatal antrostomy, a CT scan is not necessary, so that additional cost could also be eliminated in the sort of case that would otherwise be having a conventional inferior meatal antrostomy. There are however medicolegal considerations, and the majority of surgeons undertaking FESS would not operate without a scan. Indeed, many would argue that the scan is an essential part of the diagnostic workup. Kennedy (1992) who coined the term functional endoscopic sinus surgery, describes the approach as follows:

“ was continued until all involved areas seen on the CT scan were opened. The dissection was then continued into the next layer of cells to ensure that there was not radiographically undetected significant disease. If this next layer of cells was clear, the surgery was stopped.”

He makes the point that the surgery is tailored to the extent of the disease in each patient, and that simple pre-operative rigid nasal endoscopy cannot identify disease extent in the ethmoids. In a very careful and detailed study of 120 patients undergoing FESS, he found a strong correlation between objective surgical outcome, assessed endoscopically by absence of disease, and pre operative disease extent. Since disease extent was the only significant predictive factor, he recommends a staging system for inflammatory sinus disease based on CT and operative findings. What is fascinating, however, was that subjective outcome was good in all groups and did not correlate with objective findings. Kennedy concludes from this that subjective outcome measures are unreliable. He draws an analogy with cholesteatoma surgery, pointing out that no otologist would presume a satisfactory outcome based only on the absence of symptoms. Although this is true, the comparison is not entirely valid, for the following reasons:

  1. Cholesteatoma is a “surgical disease”. Complete surgical excision offers a good prospect of permanent cure. The same cannot be said of diffuse polypoidal rhinosinusitis, which Kennedy himself stresses is primarily a medical condition requiring medical treatment.
  2. A principal aim in cholesteatoma surgery is to produce a “safe ear”, i.e. to prevent serious complications in future. Although it is recognized that sinusitis can also cause serious intracranial and orbital complications, their prevention is not usually cited as a prime indication for surgery. In sinus disease, 99% of the time we are operating to relieve current symptoms.
  3. Measures of outcome should be related to the aims of treatment. Since it is accepted that surgery cannot be expected to restore diffuse polypoid mucosa to normal, normal mucosa is not a reasonable standard to aim at, especially when it has been demonstrated that the patient’s symptoms are usually relieved despite persistent abnormalities on endoscopy.

What it all boils down to is the relative importance we attach to the patient’s subjective opinion, versus the surgeon’s “objective” opinion. I would submit that the patient’s view is the most important. The days of surgical paternalism (“The operation was a complete success, unfortunately the patient died”) are long gone. Therefore more weight should be given to outcome measures based on the patient’s opinion than on “technical” measures of success. Also, at least as much effort, if not more, should be devoted to maximizing the reliability and validity of measuring patient satisfaction as is given to the so-called objective measures of outcome. Both should be quoted in reporting results.

A serious criticism that could (and no doubt will) be levelled at my study is that this is not really a controlled trial of functional endoscopic sinus surgery, but rather a simple test of middle meatal versus inferior meatal antrostomy. It could even be argued that it is not possible to test FESS by a randomized controlled trial, because it is primarily a diagnostic concept, not a treatment modality. I accept this argument in part, but would contend that my trial is testing both the diagnostic concept, by implication, and the use of the endoscopes in surgical treatment. The implied test of the diagnostic concept is that if the ostiomeatal complex is really of such critical importance, then surgery aimed at clearing disease from the “key area” and restoring normal mucociliary clearance pathways should have a better outcome than the “unphysiological” operation of inferior meatal antrostomy. Most would agree that this theoretical concept lies at the heart of functional endoscopic sinus surgery. Having identified a falsifiable hypothesis, predicted by the theory, the theory can then be tested (Popper, 1959). It is important to retain an open mind as to what may and may not work in practice. It is dangerously easy to justify actions on the basis of rational deduction from a plausible theory. The whole intellectual process then degenerates into a logical tautology, the argument becomes circular.

Many authors have commented on the good results obtained by endoscopic sinus surgery where conventional sinus surgery has failed (Stammberger, 1986; Stammberger and Posawetz, 1990; Levine,1990; Kennedy, 1992). The fact that inferior meatal antrostomies fail to give relief in some patients, and that attention to the ostiomeatal complex in those patients is often successful, is a fairly persuasive argument in favour of FESS. It does not, however, prove that FESS is superior, because there is no control group. Any form of revision surgery will have a success rate. If the revision surgery is something new, hi-tech and presented as “the latest advance”, it is not entirely surprising if patients feel better afterwards. What would be the results of carrying out revision inferior meatal antrostomies on patients who fail primary FESS?

For those who continue to insist that FESS cannot be subjected to a randomized controlled trial, I would have to point out that this attitude takes us outside the bounds of scientific medicine and into the realm of fringe treatments. No new drug can be introduced into practice without demonstrating its worth compared with current standards. It now takes around 10 years and $100,000,000 of development and regulatory trials to bring a new pharmaceutical compound to market. In surgery we have the dubious privilege of being able to think up a new operation overnight and start doing it for real on patients the next day. Before regulatory constraints are applied from outside, it surely behoves surgeons to do all that is possible to submit new operations to the kind of trials that we accept as necessary for new drugs.

This study has not yet provided a definitive answer, because the numbers are too small to exclude a Type II statistical error. It has shown that both forms of treatment are effective. There may be a slight trend in favour of FESS but the difference is not dramatic. Conventional inferior meatal intranasal antrostomy, though no longer fashionable, can give very good results. Since the new alternative is considerably more expensive, this study deserves continuation and/or replication. The most likely way forward will be to mount a multi-centre trial. It is also important to consider long term outcome, because of the chronic nature of inflammatory sinus disease, especially with polyposis and allergy (Sogg, 1989). Lawson (1991) discussing the outcome of conventional intranasal ethmoidectomy for polyposis went so far as to state that “a follow-up period of 1 year is essentially meaningless”. Although I cannot agree with this statement, there is no doubt that long term results are required for proper comparison. If it should turn out that the FESS group require less revision surgery than the INA group, that would completely change the cost comparison aspects of this study. The incidence of revision surgery could provide an outcome measure for long-term success.


Functional endoscopic middle meatal antrostomy and conventional inferior meatal intranasal antrostomy are both effective treatments for chronic and recurrent acute sinusitis which fails to respond to adequate medical treatment. For the relief of subjective symptoms, there appears to be a slight trend in favour of FESS. This is not statistically significant, but a type II error is possible because of the small numbers in this trial. There is some tentative evidence, also not statistically significant, that FESS may be a less acceptable form of treatment to the patient than INA. Technically, both procedures have a high success rate, 87% for FESS and 97% for INA. FESS costs more than INA. As currently practised the additional cost per case treated is around £280. In order to justify this additional cost, it is important to continue and/or replicate this study, to have sufficient numbers to answer the question of whether there is any increased benefit from FESS versus conventional INA. It will also be useful to look at the longer term outcome of these patients, to see whether any differences in the necessity for revision surgery emerge between the groups.