In April, 2006, a 45-year-old white man—a lifelong nonsmoker—was referred to the Royal Brompton Hospital Cough Clinic. In October, 2002, he had presented to his general practitioner with a persistent dry cough. The cough had started in March, 2002, after a bout of pneumonia. In addition, he claimed to have night sweats but no systemic symptoms. Auscultation of his chest, and a chest radiograph were normal, and he had a negative (grade 1) tuberculin skin test. In January, 2003, he was referred to his local chest clinic, where further inves tigations including lung function and CT of the chest were normal. He was reassured and discharged. In June, 2004, he presented to his local eye clinic complaining of blurred vision. Anisocoria was noted with an enlarged tonic left pupil (fi gure) and near light dissociation. Ankle tendon arefl exia was present on the right and elicited on the left only with reinforcement. He was diagnosed with Holmes-Adie syndrome—a condition of unknown cause characterised by tonic pupillary reactions and absent tendon refl exes. In August 2004, he was re-referred back to his local chest clinic because his dry cough persisted. Empirical treatment with omeprazole 20 mg daily was started and he underwent a normal barium swallow. Over the course of the following year, with no improvement in symptoms, he was prescribed an inhaled antimuscarinic bronchodilator (oxitropium bromide), an oral H1-anti histamine (loratadine), and he was given a trial of an H2-receptor antagonist (ranitidine) for 1 month, with no eff ect. When seen by us in April, 2006, his cough had persisted and he reported that during one coughing paroxysm he had experienced syncope. Direct questioning revealed truncal and right arm hyperhydrosis with excessive sweating in the right armpit. Respiratory, nasopharyngeal, and cardiovascular examinations were unremarkable. In particular, the patient had no evidence of dry eyes, dry mouth, or skin changes that might suggest a diagnosis of Sjogren’s syndrome. VDRL, autoantibody screen, and fasting blood sugar were negative. CT of the chest, full lung function (including histamine challenge, nitric oxide, and reversibility testing), nasolaryngo scopy, cinevideo fl uoroscopy, and 24-hour oesophageal pH monitoring were all normal. The patient declined fi breoptic bronchoscopy and capsaicin cough challenge. Treatment with high-dose inhaled corticosteroid (fl uticasone propionate 1000 μg daily) delivered from a pressurised metered-dose inhaler and spacer attachment, induced paroxysms of cough. Two separate courses of high dose oral steroids provided no improvement in either symptomatology or diurnal peak expiratory fl ow rate measurements. The addition of a non-selective oral H1-antihistamine treatment (chlorphenamine maleate 4 mg four times a day) was of no additional benefi t. As of September, 2006, the patient is still symptomatic and awaiting neurological consultation including cardiovascular autonomic testing. Kimber and colleagues, have reported an association between Holmes-Adie syndrome and chronic cough. Our patient showed early signs of autonomic neuropathy as his cough developed, including anisocoria, hyperhydrosis and syncope. The mechanism behind the cough is poorly understood, but it has been proposed that selective loss of thinly myelinated aff erent nerves in the lung, trachea, and larynx leads to denervation hypersensitivity of secondary neurons in the nucleus solitarius. Although the patient undoubtedly had a trigger episode precipitating his chronic cough (the original episode of pneumonia), the presence of an autonomic neuropathy may well explain the persistence of his symptoms. This case also highlights the diffi culty in treating patients with no apparent cause for their cough. Although chronic idiopathic cough and Holmes-Adie syndrome are both of unknown cause, recognition of this association has led to a more accurate diagnosis and, in this case, an appropriate tertiary specialist opinion.