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 – 2016 JW Fairley

Mr James W Fairley BSc MBBS FRCS MS Consultant ENT Surgeon

Chapter 1 General Introduction and Historical Review

Chapter 1 Contents

Thesis Contents

Summary

The relationship between nasal symptoms and objective findings is variable. This makes assessing the results of nasal and sinus surgery difficult. The lack of reliable and valid outcome measures hampers the clinical research and audit which is needed to guide practice. Most surgeons would prefer objective evidence that the disease is cured to the patient’s subjective opinion. This requires an objective diagnostic classification. Critical analysis of current diagnostic classifications for rhinosinusitis shows that none of them are ideal. The most useful clinical classifications rely themselves on subjective symptoms, and were used in the classification of patients reported in this thesis. The advent of rigid nasal endoscopy has literally shed new light on nasal and sinus conditions, but uncritical acceptance of new techniques can be dangerous. Watson Williams’ suction exploration of the sinuses for “focal sepsis” in the 1920’s and 30’s provides an historical caveat for functional endoscopic sinus surgery in the 1990’s. The attribution of headaches and facial pain to nasal abnormalities is controversial. The historical aspects of this are examined, with particular reference to the work of Greenfield Sluder, whose classification of intranasal abnormalities is currently being rediscovered by rigid nasendoscopists. The design and external validity of the studies reported in this thesis is examined critically. Because the clinical observations were made on tertiary referrals to a nasal research clinic, there may be a systematic bias toward false negative results (Type II error). The positive conclusions from these studies are likely to be valid externally, but some caution should be exercised in extrapolating negative results to general ENT practice.

Aims of study

The relationship between nasal symptoms and objective findings is variable. This causes difficulties in:

  1. Selection of patients likely to benefit from surgery.
  2. Selection of the most appropriate operation.
  3. Counselling patients as to the likely outcome of surgery.
  4. Assessing the results of nasal and sinus surgery for research and audit.

This series of studies aims to examine the relationship between subjective symptoms and objective findings in the nose, and to develop reliable methods for measuring the outcome of nasal and sinus surgery for research studies and audit. An example of the application of the findings gained from the validation studies of a nasal symptom questionnaire is given in the final chapter, reduction in symptom scores being used as the primary outcome measure in a randomized controlled trial of functional endoscopic sinus surgery.

Scope of the inquiry

The relationship between nasal symptoms and objective findings was investigated by physiological experiments in healthy volunteers, by making systematic observations in patients attending a nasal research clinic, and by surgical trials in patients attending the nasal research clinic. The reliability and validity of the nasal symptom questionnaire used in all subsequent studies was tested in an unrelated series of general ENT outpatients. Subjective sensation of nasal obstruction, facial pain and headache, rhinorrhoea, cough and postnasal drip were studied in relation to nasal inspiratory peak flow, endoscopic findings and mucociliary function.

Tendency to minimize importance of symptoms

Medical practitioners, and perhaps particularly surgeons, are loath to concentrate too much on the subjective symptoms of their patients. Symptoms are regarded in medical education as important, but mainly as clues in a jigsaw puzzle. The aim is to establish a diagnosis and thereby institute appropriate treatment. In scientific studies of the effectiveness of treatment, most doctors would prefer objective to subjective measures of outcome. The surgeon prefers to know that the patient is cured of the disease, rather than whether he feels better. Feeling better is important, but takes second place to the objective evidence of improvement. If the surgery has been a “success” but the patient feels no better, there is a tendency to blame the patient, labelling him/her as “functional disorder” or “psychological overlay”. There is no doubt that many such patients exist, but there is also evidence that many patients do genuinely suffer symptoms, the cause of which is not revealed by the diagnostic methods in use. The advent of rigid nasal endoscopy has literally “shed new light” on a whole section of nasal and sinus conditions, which have previously been attributed by well trained ENT specialists to “functional disorders”.

Value and dangers of new techniques

The value of improved techniques, including CT scanning but especially rigid nasendoscopy, in the diagnosis of nasal and sinus conditions, is obvious to all using them. Levine (1990a) discovered endoscopic abnormalities in 58 out of 150 patients with nasal and/or sinus complaints, all of whom had normal conventional anterior and posterior rhinoscopy. The pathology found included obstructed natural maxillary sinus ostia and several pathologic disorders in the middle meatus, such as polyps and synechiae. Many of these patients had seen several physicians and exhibited frustrating long standing symptoms, including postnasal catarrh, postnasal drainage, headache, facial pain, “sinus” pressure, and congestion.

Development of nasal sinus endoscopy

Nasal and sinus endoscopy has been available for over 100 years (Stammberger, 1991), but it was only after the development of fibreoptics (Hopkins and Kapany, 1954), based on Baird’s patented principle of transmitting images around corners in fine glass fibres by total internal reflection of light (Baird, 1927), and the advent of the Hopkins rod lens system, that the technique became widespread (Messerklinger 1978; Wigand et al 1978; Draf 1983).

Risks of over-interpretation of nasal sinus endoscopy

With enthusiasm for new diagnostic techniques comes a risk that findings will be over-interpreted. Medical history is replete with examples. The common thread is that deductions from an incompletely understood system are applied without proper regard to outcome from the patient’s point of view. In the early part of this century, large numbers of patients underwent dubious surgical procedures on the sinuses, for indications that appear rather strange today (Fairley, 1991). Watson Williams (1925, 1933) developed the concept that focal sepsis in the sinuses could give rise to widespread secondary conditions, including

  • optic neuritis
  • retinal detachment
  • asthma
  • acne
  • boils
  • auricular fibrillation
  • endocarditis
  • appendicitis
  • rheumatoid arthritis
  • neurasthenia
  • insanity
  • criminal behaviour!

Focal sepsis

The 1920’s and 30’s were the heyday of the theory of focal sepsis, in which bacterial toxins were supposed to be absorbed from latent occult infection in the teeth, tonsils, sinuses, gastrointestinal and genitourinary tracts.

Microscopy showed that organisms were indeed there, and bacteriologists had demonstrated that they were capable of producing exotoxins which could be absorbed systemically. It was known that white blood cells were important in defence against bacteria, so the theory went that organisms present without pus cells could actually be more pathogenic than cases in which pus had formed, because there was nothing to resist absorption of their toxins into the bloodstream.

Then as now, there were a large number of diseases of unknown aetiology, and so it was natural to suppose that at least some of them were due to the presence of these bacteria lurking in various warm dark recesses of the body.

Watson-Williams claimed that by searching for and eliminating septic foci in the sinuses, many of these general medical and psychiatric diseases could be cured. He investigated numerous patients from the mental hospitals, and claimed to have cured cases of criminal insanity by sphenoidotomy. The psychiatrists agreed (Cotton, 1923; Graves, 1923; French, 1927). Great emphasis was placed on the special danger of low grade organisms, which could remain latent for years, causing occult or cryptogenic focal sepsis. Their activities were compared with the

“insidious ravages of the death watch beetle… proceeding unnoticed for generations in the roof of our Westminster Hall”.

Watson Williams was not a “maverick”, he was in the vanguard of mainstream scientific medicine, developing a consistent theory of focal sepsis from the earlier work of Miller (1891), Hunter (1900a,b), Davis (1912), Rosenow (1914) and Billings (1914). He was regarded as an International Expert in the field. He did not advocate indiscriminate operation. His diagnostic method was rigid endoscopy of the nose, looking for any signs of pus or inflammation around the natural ostia. This was followed by his own technique of “suction exploration”, using specially designed instruments to cannulate each individual sinus for bacteriological study. The presence of bacteria was all important. Clear fluid, without pus cells, but growing organisms, was regarded as highly significant. The absence of white blood cells in the washings meant that the patient had impaired defences against toxic absorption and tissue invasion, and was therefore at greater risk than in cases where pus had formed.

It was a closely argued and highly plausible thesis. One wonders whether we have not missed something important in our subsequent rejection of the theory of focal sepsis. What is missing, however, is a full and complete analysis of the results. There are numerous case reports of successful treatment, but no systematic analysis of overall results. There is no comparison of the results of treatment by suction exploration of the sinuses with other contemporary treatments.

The lesson we should learn from the focal sepsis debacle is that today’s rational treatment may seem ridiculous tomorrow, in the light of further advances. What is more, the wheel can come full circle, so that the ridiculous old fashioned idea is rehabilitated. For example, the recent re-emergence of the idea that peptic ulcer is due to a bacterial infection, Helicobacter (formerly Campylobacter) pyloridis (Goodwin et al, 1986). Peptic ulcer was one of the original conditions blamed on focal sepsis. Who is to say whether Watson Williams may not have been right – at least part of the time?

Around the same period, there was a great vogue for various “pexy” operations in the abdomen for the condition of “visceroptosis” (Walton, 1927). This condition of excessive mobility of abdominal organs was also thought to be a cause of mental disease. Its diagnosis was a direct result of the advent of X-ray examination. Because it was not appreciated that the position of intra-abdominal organs is different in the erect alive state to the classical anatomical descriptions based on the supine cadaver, a whole range of vague complaints were put down to “wandering spleen” and similar ideas. Here was an example of a clear diagnostic advance – X-rays – perversely resulting in ineffective, unnecessary and dangerous treatments because the findings were not critically evaluated.

Most doctors react to such bizarre episodes from the past with either amusement at the naivety of our predecessors, or incredulity and embarrassment that such things could have gone on in the name of medicine. In fact, we are not so very different today. To paraphrase Huxley (1960), the one thing we learn from history is that we learn nothing from history.

Similarities between FESS and focal sepsis

There are many similarities between the recent introduction of functional endoscopic sinus surgery (FESS), based on the theoretical concept that the ostiomeatal complex is the root of all evil in the sinuses, (Stammberger, 1986; Kennedy, 1985) and Watson Williams’ suction exploration for focal sepsis 60 years ago. Both introduced a rational form of treatment, based on logical deduction from pathophysiological concepts. Both offered the prospect of relieving relatively major disease by eliminating a relatively minor focal cause. Both used the latest technology, and relied on endoscopes for diagnostic precision. Neither has been subject to a controlled trial of efficacy.*

What is missing from Watson-Williams’ book, and missing from practically all other reports on treatment of sinusitis up to and including the present day, is a full presentation and critical analysis of the results. The problem here is not one of quantity of information. Surgeons have always been willing to quote their results – especially their successful results. Quite a lot of surgeons have pointed out complications as well – preferably those of other surgeons. Unfortunately very few have reported results to the standard required for scientific acceptability. 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.

Despite these considerations, functional endoscopic sinus surgery remains a very attractive concept. All the more reason for obtaining some evidence, by properly controlled trials, that its enormous promise is borne out by improved results of treatment. However, before starting to talk about the outcome of treatment, we need to be clear what conditions we are treating. Herein lies the first major difficulty to be encountered, namely the selection of a diagnostic classification for rhinosinusitis.

Notes on diagnostic classification of rhinosinusitis

The classification of rhinosinusitis is fraught with difficulties. There is the usual conflict, inherent in any type of classification, between the “lumpers” who wish to aggregate cases for study, and the “splitters” who wish to dissect every minute detail. Both approaches have their advantages and disadvantages. The problem faced by the splitters is that rhinosinusitis does not split easily into neat diagnostic groups. There are four major axes along which distinctions can be made. These are anatomical site, aetiological agent (usually multifactorial) pathological process and temporal characteristics. The number of possible mutually exclusive diagnostic categories is a function of the number of axes multiplied by the number of groups on each axis, which is a large number, but when it is realised that patients can occupy multiple categories simultaneously, the number becomes astronomical. Mackay’s classification of rhinitis (Table 1.1) is an example.

With sinusitis the picture becomes even more complicated. The ENT Subcommittee of the British Society for Allergy and Clinical Immunology has proposed a multiaxial classification, based on aetiology, anatomical site, method of diagnosis, and the presence of nasal polypi and anatomical abnormalities (N.S.Jones 1992, personal communication, Table 1.2) Unfortunately, as both these authors acknowledge, many cases are of mixed aetiology and different anatomical sites can be affected at different times in the same individual.

Rhinosinusitis is rarely due to one simple single factor, there is nearly always some degree of overlap and interaction. For example, a minor degree of narrowing of the ostiomeatal complex may intermittently become completely occluded due to an allergic rhinitis, resulting in a clinical picture of recurrent acute infective maxillary sinusitis. The highest level hierarchical distinction in both classification systems is between allergic and non-allergic disease, yet it is not uncommon to have both in one individual.

Table 1.1 Mackay’s Classification of rhinitis (1989)

  1. ALLERGIC
    1. Seasonal
    2. Perennial
  2. NON-ALLERGIC
    1. Infective
      1. Acute
      2. Chronic
        1. Specific
        2. Non-specific
        1. Immune deficiency
        2. Clearance abnormality
    2. Non-infective
      1. Hyperactive (vasomotor rhinitis)
      2. Autonomic imbalance
      3. Post-infective
      4. Hormonal
      5. Drug induced
      6. Emotional
    3. Anatomical and mechanical
      1. Choanal atresia
      2. Adenoids
      3. Septal deformities
      4. Hypertrophic turbinates
      5. Polyps
      6. Foreign bodies
    4. Tumours
      1. Benign
      1. Malignant
        1. Primary
        2. Secondary
      2. Non-healing granOLomas

Table 1.2 Proposed BSACI Classification of Rhinosinusitis
(N.S. Jones, personal communication, 1992)

  • ALLERGIC
    • Seasonal
    • Perennial
  • NON-ALLERGIC
    • Infective(state site & whether viral bacterial or fungal)
      • Acute
      • Chronic(more that 2 months duration with positive culture)
  • Idiopathic(No evidence Type I hypersensitivity, infection or other pathology)
  • NARES (Non-allergic rhinitis with eosinophilia)
  • Drug-induced
  • Autonomic (Responds to anticholinesterases)
  • Atrophic
  • Hormonal
  • Granulomatous (Non-infective)
  • Systemic disorders
    • Primary defect mucus production (CF, Young’s)
    • Primary ciliary dyskinesia (Kartagener’s)
    • Immunologic (SLE, RA)
    • Acquired Immunodeficiency Syndrome (AIDS)
    • Primary antibody deficiencies
    • X-linked Agammaglobulinaemia
    • Common Variable Immunodeficiency (CVID)
    • IgA deficiency

Qualifying Terms: Site, Polyposis, Anatomical abnormality

Site (E = Endoscopic Evidence, CT = CT evidence
L = Left, R = Right)

  • Maxillary Sinus
  • Anterior Ethmoidal sinus
  • Posterior Ethmoidal sinus
  • Sphenoidal sinus
  • Frontal Sinus
  • Nasal Polyposis(Specify if associated with Type I Hypersensitivity,
  • Asthma, Aspirin sensitivity
  • Anatomical Abnormality
  • External deformity
  • Septal deviation compensatory turbinate hypertrophy
  • Ostiomeatal narrowing
  • Adhesions
  • Post laryngectomy
  • Choanal stenosis or atresia
  • Foreign body
  • Adenoidal hypertrophy
  • Post traumatic (? neurological)

The purpose of a classification of disease in clinical research

With such evident difficulties, it is important to question the purpose of a classification. In clinical practice, making a diagnosis – which means, in effect, placing the patient somewhere within a classification system – should help provide a prognosis and a basis for decision about subsequent management. There is no point in the diagnosis being any more specific than is required to decide these matters. This view is readily acknowledged by general practitioners, who are able to work effectively most of the time with “fuzzy” diagnoses, but until recently has been less acceptable to hospital specialists. An example is allergy testing. Since, in the UK, both allergic and non-allergic rhinitis are usually treated initially with topical steroids, there is little point in carrying out allergy tests in clinical practice, especially since seasonal allergy is easily diagnosed by history alone.

The primary aim of clinical research is to provide clinical colleagues with useful and reliable information on which to base their management decisions. Although it will often be appropriate to carry out more detailed investigations to allow accurate comparison between groups of patients, the primary aim must not be forgotten. The results should be helpful to clinical practice. Unless there is good reason to believe that different diagnostic subgroups of patients behave differently or need different treatment, there is little point in expending scarce resources attempting to make the distinction, especially where that distinction is inherently fuzzy or blurred.

Things are different with basic scientific research, where the aim is to discover new information about the mechanisms of disease processes. If patients are involved, they should be investigated in whatever degree of detail is necessary to answer the scientific question. This will usually extend beyond what is needed for clinical research.

On the basis of the clinical history, the endoscopic examination, and the mucociliary function test I attempted a much simplified clinical classification (Table 1.3)

Table 1.3 Clinical classification of rhinosinusitis

  • Temporal Axis
    1. Recurrent acute symptoms OR
    2. Chronic symptoms OR
    3. Chronic with recurrent acute exacerbations
  • Endoscopic Axis
    1. Normal findings OR ANY COMBINATION OF
    2. Turbinate / Septal contact pressure zones
    3. Ostiomeatal narrowing / contact pressure zones
    4. Localized Polyps / Oedema OR Diffuse polyposis
    5. Evidence OR no evidence of sinus infection
    6. Tumour
  • Mucociliary function axis
    1. Normal OR
    2. Abnormal OR
    3. Grossly abnormal
  • Miscellaneous groupAtypical facial pain, neuralgia, depression

Even this simplified classification has a large number of permutations. The important aspect to look for is mutual exclusivity of the categories; these are denoted by the word OR. The non-exclusive categories all occur in the endoscopic axis and there are five of them. If each of these is considered dichotomous – yes or no – itself an oversimplification – there are 25 = 32 possible states. These have to be multiplied by the 3 mutually exclusive temporal categories to give 96, then the 3 mutually exclusive mucociliary function categories to give 288 permutations. And this is without worrying at all about side, site and underlying immunological states. I have conveniently ignored the group with atypical facial pain etc.

Of course, this mathematical exercise is not entirely realistic. Many of the 288 combinations would not occur in real life. Nevertheless, many of them can and do, and the exercise illustrates how easy it is to make the number of diagnostic categories exceed the number of patients to be classified….

Using only three of the dichotomous endoscopic categories of this clinical classification still gives 23 = 8 groups. The 167 cases seen in the Nasal Research Clinic and forming the basis of Chapters 5 to 9 of this thesis are broken down into 8 groups according to the presence or absence of polyps, pus and mucosal contact pressure zones (MCPZ) in Table 1.4:

Table 1.4 Classification of nasal research clinic cases according to 3 endoscopic findings (n = 167)

No pus No pus Pus Pus
No MCPZ MCPZ No MCPZ MCPZ
No Polyps 28 45 9 11
Polyps 18 29 18 9

The table shows that cases are spread over all possible categories. 28 had none of the specified abnormalities, while 9 had all 3. To further subclassify each of these according to the temporal axis – recurrent acute, chronic, or chronic with recurrent acute exacerbations – will clearly result in categories containing very small numbers of cases, which will not be amenable to statistical analysis. Furthermore, it cannot be assumed that the cases in which no pus was seen endoscopically did not belong to the group of infective cases, especially if the clinical picture suggested recurrent acute sinusitis and the patient was examined while well.

For these reasons, the use of even a simplified multiaxial classification, while superficially logical and attractive, is in fact impractical and unhelpful. A simpler classification must be used for clinical research, but we are then faced with the problem of making explicit the decision rules governing categorization. The rules are explicit in a multiaxial classification, but, paradoxically, can become very fuzzy in a simple one.

Mygind et al (1986) and Mygind and Naclerio (1993) recognized these difficulties and proposed a very simple classification of rhinitis (Table 1.5).

Table 1.5 Simple classification of rhinitis (Mygind et al, 1986; Mygind & Naclerio, 1993)

  1. Infectious (purulent)
  2. Seasonal allergic = hay fever = pollinosis.
  3. Perennial allergic rhinitis
  4. Perennial non-allergic rhinitis (eosinophilic or non-eosinophilic).

They note that the limits between normality and disease are fluid, and admit that the distinction between eosinophilic and non-eosinophilic is not entirely suitable for clinical practice.

Mackay and Lund (personal communication, 1993) have recently used an equally simple clinical classification for rhinosinusitis in their reports on the results of functional endoscopic sinus surgery (Table 1.6)

Table 1.6 Simple classification of rhinosinusitis (Mackay & Lund, 1993)

  1. Polyposis
  2. Chronic infective
  3. Recurrent acute
  4. Miscellaneous

These groups are in fact more suitable for reporting the results of clinical research, and although I have not adopted any formal classification in this thesis, the above list forms the basis of my approach. This still begs the question as to what criteria are used to place patients in the categories. To begin with, we have a patient with symptoms. That is the starting point for the diagnosis.

Use of symptoms in diagnosis of rhinosinusitis

Many authors have commented on the difficulties in using symptoms to diagnose rhinosinusitis (Axelsson and Runze, 1983; Melén et al, 1986; Berg and Carenfelt, 1988; Gleeson, 1992; van Duijn et al, 1992), although rhinologists have tended for some time to subdivide allergic rhinitis patients on the basis of predominant symptoms (“Blockers”, “Sneezers”… Mygind & Naclerio, 1993). All studies which attempt to correlate symptoms with disease suffer from one fundamental methodological difficulty. That difficulty lies in the definition of disease, which, as the foregoing discussion has made clear, is often tenuous in rhinosinusitis. A “gold standard” for diagnosis must be set, either explicitly or implicitly, and symptoms are compared against it. There is an implicit value judgement in such studies that symptoms are somehow less important than signs, and need to be “accounted for” by physical findings, whether radiological, ultrasonographic, rhinoscopic, surgical, microbiological or histological. Although it is of course necessary to look carefully for physical findings, especially where these may reveal serious disease or will change management, it is intellectual arrogance to conclude from the absence of detectable pathology that there is nothing wrong with the patient. Before the concept of the ostiomeatal unit and the importance of the anterior ethmoid became widespread (Stammberger, 1986) both clinical ENT practice and many research studies suffered from a misguided adherence to hinging everything on the presence or absence of fluid in the antrum. Even recently reported studies (van Duijn et al, 1992) have concentrated on this crude and insensitive measure of sinusitis and have been criticised for doing so (Fairley, 1992; Jones, 1992). Levine (1990a) using simple out patient rigid nasendoscopy, found abnormalities in 58 out of 150 patients with nasal and sinus symptoms, all of whom had normal conventional ENT examination. Many of these patients had seen several physicians and complained of frustrating long standing symptoms. Such abnormalities are often amenable to successful treatment (Levine, 1990b). Gleeson (1992) pointed out that diagnosis of acute maxillary sinusitis is easier than chronic, but errors still occur which, in the case of acute virulent infections, can have unfortunate complications such as orbital cellulitis and brain abscess. Two studies have looked in detail at symptoms in acute sinusitis.

Axelsson and Runze (1983) studied symptoms in 176 patients with clinically suspected acute sinusitis. They completed an evaluation of symptoms, clinical signs and X-ray findings, before treatment and again on days 5, 10 and 15 of treatment. Symptoms evaluated on a 4 point scale were:

  1. Pain
  2. Nasal discharge
  3. Discharge of thick, yellowish or purulent character
  4. General malaise
  5. Cough or bronchitis
  6. Hyposmia
  7. Pain on mastication.

Signs were

  1. Fever exceeding 38C
  2. Mucous membrane thickening with mucopurulent or purulent secretions on rhinoscopy
  3. Speech character indicating fullness of sinuses.

A six point scale of scoring X-rays was used. There was good correlation between reduction in subjective ratings points and reduction in radiological score during recovery. Some symptoms and signs, including fever, pus on rhinoscopy and pain, improved more rapidly than radiological findings, emphasising the dynamic character of the disease which may be reflected differently on various axes of assessment. Cough, nasal discharge, hyposmia and general malaise improved slower, in line with the radiological findings. This study demonstrated a positive correlation between subjective and objective findings in acute sinusitis, using plain sinus radiology as the objective method.

Berg and Carenfelt (1988) studied 155 patients with suspected acute sinusitis. Symptoms and signs were compared with the “gold standard” of maxillary sinus empyema versus not empyema, established by antral aspiration. 68 were found to have an empyema and 87 not. Purulent flow from the middle meatus seen on rhinoscopy was pathognomonic for empyema when seen, but only occurred in 6 cases. Severe cacosmia was also of high positive predictive value, but only occurred in 12 cases. Of symptoms that occurred frequently, unilateral predominance of pain or purulent rhinorrhoea were found to be strongly predictive of empyema. By combining the analysis of symptoms with a high ESR, they found that “diagnostic reliability” could reach 80%, however this begs the question as to whether the gold standard they used – namely antral aspiration – is valid. Almost certainly it is not.

Melén et al (1986) noted that reports on chronic maxillary sinusitis often lack a definition of the condition, and proposed their own definition, based on:

1. Symptoms – facial pain, nasal congestion or abnormal secretion, persisting or recurring during a period of at least 3 months.

2. Sinus radiography or sinoscopy revealing persistent local or generalised mucosal swelling with or without secretion.

3a. Considered to be of rhinogenous origin if dental examination normal, and failure to heal on conservative treatment, consisting of

i) one or more courses of antibiotic treatment in relevant doses in periods of 10 days;

ii) nose drops or oral decongestants;

iii) antral irrigations.

3b. Considered to be of dental origin when following present: oro-antral fistula, indistinct border or close connection between diseased maxillary sinus mucosa and

a) apex of a non-vital tooth, or

b)an inflammatory periodontal or periapical lesion, or

c)a follicle of an impacted tooth in connection with a periodontal or periapical lesion.

Using this definition they found 198 patients with 244 affected sinuses over a 5 year period, in a population of 210,000 over 17 years of age, making an incidence of 0.02%. The aetiology was considered to be dental in 40%, including some of the bilateral cases. In 6 bilateral cases, the cause was considered to be dental on one side and rhinogenous on the other! The first part of the definition is reasonable, and probably accords with most UK practice. It is certainly a very good starting point for a clinical diagnosis of sinusitis, given the difficulties referred to earlier with the multiaxial classifications. To state that 40% of cases of chronic maxillary sinusitis are dental in origin is not, however, in accordance with common clinical experience.

So far, I have criticized studies which have been too restrictive in their definition of sinusitis. If one wishes to use an anatomico-pathological classification of disease, as implied by the term “maxillary sinusitis”, it is reasonable to expect to be able to demonstrate evidence of inflammation in the anatomical site specified. The problems are that

a) clinically it is unusual for inflammation to be restricted to one anatomical site

b) symptoms which result from inflammation in different anatomical locations are similar

c) certain anatomical sites are easier to image than others, depending on the method used – e.g. plain X-Rays are moderately good for the antrum and frontal sinuses, but poor for the ethmoids.

d) inflammation may be subtle, and is not always associated with infection.

Now that rigid nasendoscopy and coronal CT scanning are becoming more routinely available, attention has turned from the antrum to the anterior ethmoid and the middle meatus. There is however a danger here, in that variants of normal may be interpreted as being the cause of symptoms, and result in unnecessary, ineffective and potentially dangerous surgical treatment. If there has been too much ignorance of the ostiomeatal complex in the past, there may be too much uncritical acceptance of its importance now. A fact which has not been emphasized in reports of endoscopic diagnosis of rhinosinusitis is that the endoscopic image is merely a snapshot in what is often a dynamic and evolving process. It is therefore possible to have entirely normal findings in a patient with recurrent acute episodes, giving rise to false negative findings, while a simple coryza may suggest sinus infection – indeed a coryza almost certainly does involve an acute viral ethmoiditis in many cases. The difficulties with CT and MRI imaging have been well documented, abnormalities in the ethmoids being present in 30 to 50% of asymptomatic individuals who were being scanned for unrelated reasons (Kennedy et al, 1988; Lloyd et al 1991; Cooke and Hadley 1991; Patel et al 1992). The incidence of endoscopic abnormalities in normal individuals has not been subjected to such critical scrutiny, because it tends to be only symptomatic individuals who are examined endoscopically.

Headaches and facial pain of nasal origin

One of the more controversial aspects of nasendoscopic diagnosis is the identification, as the cause of facial pain and headaches, of anatomical abnormalities in the nose, particularly involving areas of mucosal contact between apposing surfaces. This is said to be a cause of pain, even in the absence of sinus infection (Stammberger and Wolf, 1988).

Greenfield Sluder (1865 – 1928), professor of laryngology at St. Louis was one of the earliest to put forward the theory that headache could be due directly to nasal conditions (Ewing and Sluder 1900; Sluder 1918, 1927). He noted that headache and eye pain can occur in the absence of suppuration, and that cessation of nasal suppuration does not always cure symptoms. He hypothesised that it is not the nature of the pathological process which is important but its local effects. Spurs or deviations can cause blockage of a nasal fossa, obstruction to sinus ostia or impingement on branches of peripheral sensory nerves. Arthur Ewing, an ophthalmologist and colleague of Sluder, describes how he suffered from Sluder’s headache himself in 1892. The pain was exacerbated by using the eyes for close work and was so severe as to make him consider giving up his practice. Three further cases are described in detail; in each a constant diagnostic sign was tenderness over the upper medial aspect of the orbit, over the attachment of the trochlea for superior oblique. This lies over a thin plate of bone forming the floor of the frontal sinus. Sluder attributed the pain to pressure changes in the frontal or anterior ethmoid sinus, brought about by obstruction to the infundibulum*. In his 1927 monograph he summarizes the causes of this “frontal vacuum headache” in 451 cases (Table 1.7).

Table 1.7 Causes of Sluder’s Frontal Vacuum Headache in 451 cases (Sluder, 1927)

I. High septal deviation in a narrow nose (38%)
II. Contact between bulla ethmoidalis and uncinate process, with normal middle meatus (24%)
III. Oedema or polyps of middle meatus (15%)
IV. Middle turbinate hypertrophy (11%)
V. Narrow vault of middle meatus, normal turbinate (7%)
VI. Localized oedema of vault of middle meatus affecting only frontal sinus outlet (3%)

Sluder advocated treatment with topical cocaine and suprarenal extract to the middle turbinate and middle meatus. He notes that these are especially valuable where the obstruction is due to mucosal swelling but of limited help where there is a bony deformity. If they fail, middle turbinectomy is done, which is effective provided that the turbinate is the cause of the obstruction to the “infundibulum”. If the cause is contact between the bulla and the uncinate process then the uncinate process is removed, but Sluder regarded this as a hazardous procedure. This is the classic Sluder vacuum frontal headache; it is important to realize that Sluder considered the cause of the pain to be vacuum in the frontal sinus as a consequence of middle meatal obstruction to its ventilation, rather than as a direct result of the mucosal pressure contact zone involving the middle turbinate. However the effect of both the cocaine and the surgical options recommended was to remove the mucosal pressure contact zone.

Sluder also described 3 other clinical types of headache, including “nasociliary neuralgia”, “sphenopalatine ganglion neurosis” – also known as “lower half headache”, and “hyperplastic sphenoiditis” (Sluder, 1918, 1927). This prolific output of theories and classifications of headache and facial pain of nasal origin has led to some of the subsequent confusion in the literature as to what exactly is Sluder’s syndrome (e.g. Ryan and Ryan, 1979; Friedman and Rosenblum, 1989). In fact there are four main types of Sluder’s syndrome (see above) and each of them has subdivisions. Many are now in the process of being rediscovered by nasal endoscopists.

Sluder used topical application of cocaine to the sphenopalatine ganglion as a therapeutic test to diagnose his “lower half headache”.

Another early example of the use of a therapeutic manoeuvre to help diagnose a nasal cause of headache was reported by Sidney Yankauer of New York (1908). In cases where there was no obvious suppuration in the nose, he advised a therapeutic trial of steam inhalations. He based this treatment on a remarkably sophisticated understanding of mucociliary clearance, showing that some rhinologists were aware of the fact that the direction of mucociliary transport was constant toward the natural ostia of the sinuses long before the experiments of Hilding (1932), Proetz (1941) and Messerklinger (1978). He observed clinically that most cases of acute sinusitis resolved spontaneously without surgical intervention. He believed that resolution is speeded by the application of moist, warm air, because of the physiological reflex of the nasal mucosa to shrink in warm air. He also cites the enhancement of ciliary activity by heat and fluidity of the secretions, and its retardation by cold, dryness and excessive viscidity. Yankauer used relief of headache following inhalations, even partial or temporary, as a strong diagnostic indicator that pain was of sinus or nasal origin, although failure of inhalations to give relief did not exclude sinus disease.

Derek Brown Kelly (1943) studied 50 cases of headache in sailors treated aboard the Royal Navy hospital ship Amarapoora in 1941 and 1942. He modified Sluder’s test and developed the use of positional instillation of ephedrine into the nasal cavities to help determine whether a headache is of nasal origin. He concluded that the use of the ephedrine test dispenses with the need for X-ray examination in many cases, and gives help in deciding whether operation is likely to give relief. This was of considerable practical importance on board ship in wartime – and perhaps equally so in the modern resource-management NHS!

General considerations on design of clinical studies used

The details of experimental design are given in each of the following chapters 2 – 9. Some general comments are nevertheless worthwhile. Shapiro (1989) states that it is impossible to design the perfect clinical study. The ideal trial would have statistical validity, contain no systematic bias (internal validity), use measures relevant and specific to the problem being studied (construct validity), and produce conclusions capable of generalisation (external validity). These requirements are mutually antagonistic, particularly the conflict between the need to limit the variables under study and the need to be able to generalise the conclusions. It is therefore recognized that any single study must be a compromise between these competing requirements. He concludes that in most cases multiple independent studies will need to be done, followed by meta-analysis, a systematic review of published studies on a subject which dis-aggregates the data and re-analyses it. This in particular may help uncover effects not revealed by single studies of inadequate statistical power. A similar conclusion was reached by the British Department of Health’s Advisory Group on Health Technology Assessment (DoH, 1992). The perils of enthusiastically accepting “new improved” treatments are also highlighted by the DoH group. In the Netherlands, control of the introduction of new technologies is exercised financially, by a scientific advisory committee to their national Health Insurance Council. The costs of new unevaluated technologies are only reimbursed when offered as part of a properly designed research study to assess their effects (Borst-Eilers, 1993). In this way the use of new technologies remains restricted to one or two hospitals until their worth has been proved. This guards against widespread diffusion into practice of “the latest advance” which later turns out to be a costly mistake.

General comments on methods and external validity of results from observational studies on patients attending the Nasal Research Clinic (Chapters 5, 6 & 7).

Since the patient-based research reported in this thesis is clinical, I have not worried too much about attempting to squeeze the patients into an artificially detailed diagnostic classification (see purpose of a classification.) Nearly all of the 167 patients studied (see data appendix for details) were tertiary referrals from ENT colleagues. They came with a history of chronic or recurrent acute rhinitis, sinusitis or undiagnosed symptoms of possible nasal origin, particularly the facial pain and headache cases. Many had already undergone conventional nasal surgery with poor results. All patients referred underwent systematic rigid endoscopic examination of the nose (Stammberger, 1986), usually repeated on several occasions. Formal antroscopy was rarely done, though the antral linings were routinely inspected with an angled telescope where a patent antrostomy existed. Coronal CT scans were not routinely done, only on patients undergoing extensive ethmoid surgery, patients for the randomized controlled trial, and some difficult diagnostic cases of facial pain and headache. Conventional sinus X-Rays were hardly ever requested by me, though most patients had previously had them. Ultrasound was not used. Patients for the randomized controlled trial had a simple allergy test (IgE levels) and immunoglobulin studies (IgG,A,M and IgG subclasses) to facilitate international comparison and meta-analysis; most of the others did not. Most patients had a mucociliary function test (saccharin time) but there was no facility for taking this any further with ciliary beat frequency analysis and electron microscopy.

In most of the studies I have used standard crosstabulation techniques to investigate associations between subjective and objective measurements. The results are obviously valid internally within the context of the tertiary referral nasal research clinic, but there are some methodological objections to extrapolating these results to general ENT practice. The argument goes as follows:

  1. Patients have often been referred because they are perceived as difficult cases.
  2. This means that any obvious and common associations will have been removed from the study sample. For example, patients complaining of nasal obstruction with gross polyposis or septal deformity will simply be treated appropriately rather than being referred to a special clinic.
  3. Because the “obvious” associations are missing from the study sample, crosstabulation measures are likely to underestimate the strength of the relationship between subjective and objective findings, and may give false negative results.
  4. Nearly all patients referred are likely to have some nasal symptoms. If the symptom is very common in the study population, crosstabulation measures are insensitive and again likely to give false negative results. For example, if nearly all the patients complain of nasal obstruction, and only a proportion have polyps, it may be concluded that there is no association between polyps and nasal obstruction. This is in fact the case from the data. Such a conclusion, however, is obviously not valid externally.
  5. There is likely to be a higher than normal percentage of persistent complainers and neurotics within the study population, because these patients are more likely to be referred on as a means of getting rid of them from a busy general outpatient clinic. Because they are less likely to have any objective abnormalities, this will also tend to dilute the strength of associations, and cause false negative results on crosstabulation measures.

All of these effects tend to increase the risk of false negative results (Type II statistical error). The magnitude of these effects is unknown; it could only be estimated by comparing the results of similar studies on an unselected group of ENT outpatients. Even this would not be entirely valid if the patients had been referred to someone known to have a special interest in rhinology. What can be assured is that any positive conclusions from these studies are likely to be valid externally, but some caution should be exercised in extrapolating negative results to general ENT practice.