The principal reason for performing investigations in patients with globus pharyngeus is to detect a pharyngeal or upper oesophageal malignancy presenting this way. There is uncertainty regarding both the necessity for investigation in globus patients and the first line investigation of choice in the screening of such patients. The authors therefore undertook a retrospective study of 699 patients who presented with globus sensation. Of these, 451 patients (64.5 per cent) had a typical history and in these patients, outpatient examination, including fibre-optic nasendoscopy was able to detect all pathologies, except one insignificant pharyngeal pouch, which required no intervention and one patient with a distal peptic stricture. In patients with atypical symptoms the combination of fibre-optic nasendoscopy and barium swallow identified all pathologies (five aerodigestive tract malignancies) except one distal peptic stricture. The authors concluded that if out-patient examination is adequate, no further investigation of typical globus symptoms is required. Such a policy would produce significant savings for both out-patient and radiology services.

Chronic rhinosinusitis plays an important role in the aetiology of maxillary mucosal cysts.

We describe a cervical sympathetic chain schwannoma in a 77-year-old woman who presented with a neck mass and Horner's syndrome. Such schwannomas are rare and this is the first documented case of a Horner's syndrome at presentation. The mass was excised via a cervical approach and her post-operative course was uneventful. The prognosis is excellent, with recurrence being rare. A brief discussion of the pathology, presentation, diagnosis, and treatment of this condition is made in this paper. The relevance of the uncertainty in diagnosis is discussed with the message that a pre-operative Homer's syndrome may guide the surgeon in the care of the patient but we suggest that in all cases proper counselling of the possible neurological consequences of this surgery be conducted.

We describe a case of a newborn baby with a prenatal diagnosis of an epignathus (oropharyngeal teratoma). With the potential for airway problems at birth, he was delivered by an elective EXIT (Extra Utero Intrapartum Treatment) procedure at 38 weeks of pregnancy. The airway was secured and rigid bronchoscopy performed. Initially he was stable, but developed cardiorespiratory difficulties 40 minutes after birth and died from a cardiac arrest 17 minutes later. Tension pneumothorax is a devastating complication that can occur with lower airway manipulation for anaesthesia and rigid bronchoscopy. The addition of positive pressure during mechanical ventilation converts the pneumothorax into a tension pneumothorax. The possibility of tension pneumothorax should be entertained in a mechanically ventilated patient whose ventilatory pressures are increasing, with diminishing cardiac output. A complicated case is presented, where the diagnosis was missed with a fatal outcome.

Acoustic rhinometry readings are very position dependent, and it was hypothesized that this accounts for its relative lack of reproducibility on a day-to-day basis. Multiple readings on each visit were taken to investigate their impact, if any, on improving the day-to-day reproducibility of the method. Measurements of the minimal cross-sectional area of the nose as measured by acoustic rhinometry were studied in 10 subjects following nasal decongestion. For each individual, acoustic rhinometry was performed ten times. The ten recordings were repeated again, in an identical manner on a separate day. The subjects were repositioned and the nasal probes reinserted between each measurement. The mean coefficient of variation for minimal cross-sectional area readings in all 10 subjects was calculated as 9.92%. This is comparable to the day-to-day variability of acoustic rhinometry as measured by other workers and thus supports the hypothesis that the high measurement error of the device (rhinometer) is a function of positional variation during data acquisition. We were able to demonstrate a minimal gain in intervisit reproducibility by doing multiple recordings per person, with a plateau effect of reproducibility after 7 repeat readings.