Introduction
Facial pain which presents in a mask-like distribution, or over the anatomical location of the paranasal sinuses, can pose a diagnostic challenge. Many patients presenting with this symptom may believe that their pain is due to paranasal sinus disease. However, such facial pain can also be caused by vascular headache (e.g. due to migraine), muscle tension headache, myofascial pain or neuralgia. These conditions may be interrelated, complicating the diagnosis. Furthermore, these conditions are very prevalent, and their symptomatology can mimic that of paranasal sinus disease; in turn, coexisting paranasal sinus disease may exacerbate these conditions' symptoms.Reference Mehle and Screiber1, Reference Cady and Schreiber2
Nasal endoscopy is an essential component of the investigation of patients suspected of paranasal sinus disease. It is safe and reasonably comfortable for the patient, and can rapidly diagnose septal deviation or spur, paradoxical middle turbinate and other pertinent anatomical variations, as well as polyps, purulence and mucosal disease at the middle meatus and/or sphenoethmoidal recess.
However, computed tomography (CT) has become the ‘gold standard’ for diagnosing and determining the extent of paranasal sinus disease. Treatment decisions, including the decision to perform surgery, are substantially based on the severity of the CT findings, in conjunction with the clinical picture. Much has been said about the role of anatomical variants (e.g. concha bullosa) as potential contact points or sources of osteomeatal unit obstruction, triggering facial pain.Reference Mehle and Screiber1, Reference Danese, Duvoisin, Agrifoglio, Cherpillod and Krayenbuhl3, Reference Zinreich, Mattox, Kennedy, Johns, Price and Holliday4 However, CT scanning is relatively costly and exposes patients to radiation. In addition, sinus CTs may fail to show important skull base and intracranial soft tissue findings that may suggest other causes of facial pain (other than rhinosinusitis).
This study aimed to determine the negative predictive value of normal nasal endoscopy in excluding paranasal sinus disease as the cause of isolated facial pain in the distribution of the sinus regions.
Methods
This study was approved by the institutional review board of our institution Massachusetts Eye and Ear Infirmary.
We initially assessed 185 consecutive adult patients who had been referred to an out-patient otolaryngology clinic by their primary healthcare provider, with the chief complaint of facial pain and a presumptive diagnosis of rhinosinusitis. All these patients' symptoms had failed to respond to prior medical therapy, including intranasal corticosteroid sprays and broad-spectrum antibiotics.
Our study group comprised a subset of 42 consecutive patients who presented with isolated facial pain of more than two weeks' duration which was localised over one or more of the paranasal sinuses, and who had undergone nasal endoscopy in the clinic, with negative results. These patients described their pain as pressure-like, sharp or pulsating. However, patients with acute, severe or shooting pain were excluded, since such pain was unlikely to be sinugenic in origin.
All these 42 patients had undergone paranasal sinus CT imaging. The CT findings were reviewed to determine whether they could explain the aetiology of the facial pain. A CT was considered abnormal if it showed mucosal disease in one or more of the paranasal sinuses, soft tissue obstruction of the osteomeatal unit, or anatomical variants in the region of the osteomeatal unit. A CT scan finding was considered relevant, for the purpose of this study, if it correlated with the location of the patient's pain.
Data analysed included patient demographics (i.e. age and gender), duration of symptoms, and CT and endoscopic findings. Nasal endoscopy was considered negative in the absence of the following: evident septal deviation or spur impinging on the middle meatus or lateral nasal wall; nasal polyposis; paradoxical middle turbinate; or discharge or mucosal disease at the middle meatus or sphenoethmoidal recess. Relevant CT findings included mucosal disease, concha bullosae, agger nasi cells, Haller cells and uncinate process variations.
Results
The study group comprised 27 women and 15 men, with a median age of 38 years (range 18–74 years). The duration of pain ranged from two weeks to more than one year, with a median duration of seven months.
The pain was unilateral in 20 patients and bilateral in the remaining 22 patients. Patients reported the following facial pain locations: frontal alone (seven patients), periorbital or retro-orbital alone (10 patients), maxillary alone (six patients), and a combination of the above sites (19 patients) (Table I).
Facial pain location
*Total patients = 42.
Fourteen patients (33 per cent) had sinus CT findings that might explain their facial pain. Three of these patients had both mucosal disease and anatomical findings. Under these conditions, the negative predictive value of nasal endoscopy, for sinus disease as a cause of facial pain, was 28/42, or 66.7 per cent. However, only four (9.5 per cent) patients had CT-demonstrated mucosal disease in one or more sinuses correlating with the location of their facial pain. Of these four patients, one had bifrontal disease, one had sphenoidal disease, and two had maxillary and ethmoid sinus disease. Under these conditions, the negative predictive value of nasal endoscopy was 38/42, or 90.5 per cent (Table II). This negative predictive value calculation was based on the controversial presumption that paranasal sinus mucosal disease noted on CT is the basis of patients' facial pain.
Nasal endoscopy negative predictive value* by sinus CT findings
* For sinus disease as the cause of isolated facial pain. †Total patients = 42. ‡‘Sinus-like’ facial pain. **Three had anatomical variants and mucosal disease on computed tomography (CT). Pts = patients; NPV = nasal endoscopy negative predictive value; NA = not applicable; none = no CT mucosal disease and/or anatomical variants correlating with pain location
Thirteen (31 per cent) patients had CT anatomical findings which could potentially obstruct the osteomeatal unit. These anatomical findings included concha bullosae (seven patients), Haller cells (five patients), agger nasi cells (four patients), and uncinate process variations (two patients). Six of these patients had more than one anatomical variant (Table II).
The 28 patients with facial pain and negative sinus CTs were referred to the pain clinic and the neurology service. Surgery was not offered to these patients.
The four patients who had mucosal disease evident on sinus CT which might account for their facial pain were given medical therapy for their rhinosinusitis, with mixed results.
Patients in this group with persistent symptoms, and the 10 patients with anatomical variants only, were offered surgical treatment. At the time of writing, only five had consented to surgery. The numbers in this subset were thus too small to determine whether surgery was beneficial in this patient population.
Discussion
Facial pain and headache in the distribution of the paranasal sinuses pose a clinical challenge, particularly when these are the only presenting symptoms in patients presumed to have rhinosinusitis. The differential diagnosis includes migraine headache, cluster headache, paroxysmal hemicrania, facial neuralgia, tension headache and midfacial segment pain.Reference Jones and Cooney5
Facial pain with an aetiology unrelated to the paranasal sinuses may manifest in the anatomical distribution of these sinuses. In addition, there is a large overlap in the clinical presentation of vascular and neuralgia headaches in patients with rhinosinusitis. Both vascular headaches and neuralgia may be associated with nasal congestion, rhinorrhoea, and (at times) increased paranasal sinus mucosal signal on magnetic resonance imaging.Reference Mehle and Screiber1, Reference Cady and Schreiber2, Reference McAbee, Siegel, Kadakia and Cantos6 Furthermore, patients with sphenopalatine neuralgia and migraine headache may respond to the stimulant or vasoconstrictive effects of pseudoephedrine. Moreover, some antibiotics commonly prescribed for rhinosinusitis, such as macrolides, are thought to exhibit an anti-inflammatory effect that could account for the relief of facial pain, without necessarily indicating that the facial pain is of sinonasal aetiology.Reference Cervin and Wallwork7, Reference Wallwork, Corman, MacKay-Sim, Greiff and Cervin8
As such, it would be useful to have a clinical test or procedure that would help to distinguish facial pain related to rhinosinusitis from facial pain due to other causes. This study suggests that nasal endoscopy is such a procedure. The negative predictive value of nasal endoscopy for sinus disease as a cause of isolated facial pain was more than 90 per cent when CT sinus mucosal changes were accepted as evidence of sinus disease. The negative predictive value of nasal endoscopy dropped to 66.7 per cent when we included patients with CT showing other anatomical factors affecting the outflow tracts of the paranasal sinuses.
The assessment of a patient with facial pain includes clinical history-taking, physical examination with nasal endoscopy, and imaging. When there are clinically relevant findings in all these three components, rhinosinusitis can reasonably be diagnosed as the cause of the facial pain. On the other hand, a sinonasal aetiology for the pain can be safely excluded when a patient has no other symptoms of rhinosinusitis and both the physical examination (including nasal endoscopy) and imaging are negative for paranasal sinus disease. The dilemma arises when the patient presents with facial pain as the only symptom, the physical examination and nasal endoscopy are negative, and the paranasal sinus CT is equivocal, since it is uncertain whether paranasal sinus mucosal disease seen on the CT is the basis of the facial pain, even if it correlates with the anatomical distribution of the pain.
Several anatomical variants may obstruct the osteomeatal unit, but their role in the pathogenesis of rhinosinusitis is controversial.Reference Danese, Duvoisin, Agrifoglio, Cherpillod and Krayenbuhl3, Reference Zinreich, Mattox, Kennedy, Johns, Price and Holliday4, Reference Kieff and Busaba9 Included in this category are concha bullosae, Haller cells, agger nasi cells, paradoxical middle turbinates, anomalies of the uncinate process, and septal deviations or spurs. These variants have a high prevalence rate, and can occur in over 30 per cent of the general population. It is not uncommon for an individual to have one or more of these variants, but no paranasal sinus disease. However, one cannot simply ignore them as incidental findings when the patient's symptoms correspond to the location of the variant(s). In such instances, the surgeon and patient must reach a decision together about whether surgical correction of the anatomical variant is warranted, considering the severity of the symptoms and the risks of surgery.
In this study, we examined the clinical yield of sinus CTs performed for 42 patients who presented with an isolated symptom of facial pain, and who had a negative physical examination and nasal endoscopy. The sinus CT was negative in 67 per cent of our study population. Mucosal disease in the paranasal sinuses was noted in only four patients (9.5 per cent). This patient number is too small to allow any inferences on whether the paranasal sinus disease was the cause of the facial pain. Ten patients (23.5 per cent) had paranasal sinus CTs showing anatomical variants only, without mucosal disease.
• Facial pain in the distribution of the paranasal sinuses is a common otolaryngological presentation
• Some of these patients have no nasal congestion or nasal discharge symptoms
• Nasal endoscopy is a quick, minimally invasive means of determining the presence of inflammatory disease or sinus outflow tract obstruction
• Computed tomography (CT) is the ‘gold standard’ for determining the same
• This study correlated negative nasal endoscopy findings with negative CT findings, in patients with facial pain in a paranasal sinus distribution
Previous authors have had mixed results when treating facial pain with or without such anatomical variants and sinonasal mucosal disease. Tarabichi reported a 38 per cent failure rate for surgical treatment of facial pain in patients with radiographic and endoscopic findings of chronic rhinosinusitis.Reference Tarabichi10 Jones and Cooney found that sinonasal surgery was not effective in many such patients, and that facial pain persisted post-operatively; they concluded that patients with ambiguous presentations of facial pain or a poor response to sinonasal therapy merited careful re-evaluation and neurological medical treatment.Reference Jones and Cooney5
On the other hand, others have advocated surgery for facial pain in select patients who fail medical therapy and have unremarkable diagnostic nasal endoscopy and paranasal sinus CT findings. Boonchoo reported the outcome of functional endoscopic sinus surgery in patients with sinugenic headache: facial pain resolved in 62.5 per cent of patients and diminished in the remaining 37.5 per cent.Reference Boonchoo11 Cook et al. found that functional endoscopic sinus surgery reduced facial pain in 88.9 per cent of their patients with a normal diagnostic nasal endoscopy, a normal pre-operative paranasal sinus CT and a poor (pre-operative) response to medical management.Reference Cook, Nishioka, Davis and McKinsey12 Reversible sinonasal mucosal disease has been hypothesised to be a causative factor for the facial pain experienced by patients with normal nasal endoscopy and sinus CT findings.Reference Boonchoo11, Reference Cook, Nishioka, Davis and McKinsey12
Our study was not designed to assess the surgical outcome of endoscopic sinus surgery in patients with facial pain and negative nasal endoscopy. Surgeons should exercise their clinical judgment in treating such patients, based on the specifics of each individual case and their interpretation of the medical literature.
Conclusion
In this study, the yield of paranasal sinus CT was less than 10 per cent, in demonstrating a sinugenic cause of isolated facial pain in patients with negative nasal endoscopy, if the CT was considered positive only when it showed mucosal disease in the paranasal sinuses. Under these conditions, normal nasal endoscopy had a 90.5 per cent negative predictive value for sinus disease as a cause of isolated facial pain. The yield of sinus CT increased to 33 per cent if anatomical variants were added to the criteria by which a CT was judged as positive. Under these latter conditions, the negative predictive value of normal nasal endoscopy dropped to 66.7 per cent.
