In December, 2008, an active 70-year-old man presented to the medical department complaining of low back pain (LBP), radiating to the loins and associated with right leg weakness. The pain had appeared earlier that day when he had tried to board a train. He was generally well, but had longstanding hypertension treated with propranolol, cilazapril, aspirin, and atorvastatin. General and neurological examinations were un remarkable. He was given tramadol for pain relief. Lumbar radiography showed degenerative changes and an enlarged urinary bladder. His white-blood-cell count was 15·9×109/L (88% neutrophils). 700 mL of residual urine was found on catheterisation with a few white blood cells in the urinary sediment. Urinary-tract infection was diagnosed and ofl oxacin plus alfuzosin started. The next day, the patient fainted while seated. Druginduced orthostatic hypotension was suspected. 2 h later, he complained that he could not feel his right leg. He had marked weakness of the leg (power 3–4/5), a positive Babinski sign, and a blood pressure (BP) of 204/124 mm Hg. Hypertensive emergency and cerebro vascular accident were diagnosed and intravenous nitrates started. Head CT showed bilateral small occipital subarachnoid haemorrhage (SAH), for which no cause was apparent on CT angiography. Probable post-traumatic SAH was diagnosed. That night he became very agitated, and acute delirium was diagnosed. BP was 150/100 mm Hg and nitrates were stopped. He then complained of severe LBP for 90 min. 1 h later he became dyspnoeic with arterial oxygen desaturation, he vomited (coff ee grounds appearance), and his BP dropped to 82/56 mm Hg. He was immediately intubated, ventilated, and started on inotropes. Serial electro cardio grams and troponin concentrations were unchanged. Blood test results showed leucocytosis, prerenal azotaemia, and high concentrations of aminotransferases. Aortic dissection or internal bleeding was suspected. However, haemoglobin concentration remained stable, and extensive imaging did not show any notable pathology. Septic shock and epinephrineproducing tumour were ruled out. The patient’s condition stabilised within 24 h, and he was extubated and weaned from inotrope support. Neurological examination now showed fl accid paraparesis, with the right leg more severely aff ected, bilateral positive Babinski sign, and a T2 level of hypoaesthesia. A departmental discussion evoking the principle of Occam’s razor suggested a single unifying diagnosis. MRI of the spine showed a posterior subacute subdural haematoma from T2 to L3 with cord compression (fi gure). The patient underwent emergency neurosurgery with laminectomy at T9-10 and evacuation of the haematoma. When last seen in March, 2009, and under going rehabilitation, the patient was showing gradual recovery, but still required a wheelchair. Spinal subdural haematoma is a rare diagnosis and is usually diagnosed in the context of bleeding disorders, anticoagulants, trauma (including lumbar puncture), or vascular malformation. Few cases arise spontaneously. It is more common in elderly patients, involves the lumbar or thoracolumbar spine, and presents with sudden back pain and varying degrees of motor, sensory, and autonomic dysfunction. Acute, subacute, and chronic presentations have been reported. It is best diagnosed by MRI, and surgical decompression seems to be the treatment of choice. Prognosis is variable. Our patient’s stuttering course was probably due to a gradual extension of the haema toma. The progression—from acute pain and trans ient leg weakness, to urinary retention, syncope, hyper tensive emergency, monoplegia, SAH, delirium, transient shock, and fi nally almost paraplegia—was all due to the spinal haematoma. Involvement of the sympathetic fi bres in the thoracic cord could have led to the sudden shock, since no other explanation could be found and rapid reso lution occurred. SAH is often found in association with spinal subdural haematoma. As long as each stage was ascri bed to a diff erent diagnosis, we failed to explain the evolving clinical picture as a continuum. Thus, Occam’s principle of parsimony, stating that one should fi rst try to see whether a single cause can provide a suitable explanation for accumulating observations before resorting to multiple causes, remains useful 700 years since it was fi rst conceived.