The diagnostic algorithm of computerized tomography (CT) and lumbar puncture (LP) for suspected subarachnoid haemorrhage (SAH) has lately been challenged by the advancement of radiological techniques, such as higher resolution offered by newer generation CT-scanners and increased availability of CT-angiography. A purely radiological workup of suspected SAH offers great advantages for both patients and the health care system, but the risks of abandoning LP in this setting are not well investigated. We have characterized the differential diagnostic yield of LP in the investigation of suspected SAH by a retrospective study. From the hospital laboratory database, we analyzed the medical records of all patients who had undergone CSF-analysis in search of subarachnoid bleeding during 2009-2011. A total of 453 patients were included. In 14 patients (3%) the LP resulted in an alternative diagnosis, the most common being aseptic meningitis. Two patients (0.5%) received treatment for herpes meningitis. Five patients (1%) with subarachnoid haemorrhages were identified. Among these, the four patients presenting with thunderclap headache had non-aneurysmal bleedings and did not require surgical intervention. We conclude that the differential diagnostic yield of LP in investigation of suspected SAH is low, which indicates that alternative diagnoses is not a reason to keep LP in the workup when a purely radiological strategy has been validated. However, algorithms should be developed to increase the recognition of aseptic meningitis. One hundred and fifty-three patients (34%) were admitted to undergo LP, which estimates the number of hospital beds that might be made available by a radiological diagnostic algorithm.