BACKGROUND The thoracoscopic microsurgical technique (TMT) for vertebral and spinal cord surgery is associated with the benefits of reduced postoperative pain, accelerated return to physical activity and reduced complication rates. However, because of the surgeon's requirement of a non-ventilated lung, it confronts the anesthesiologist with the need for extremely long duration of single-lung ventilation (SLV). METHODS We describe our experiences with 82 patients, whom we anesthetised from 1993 until 1996 for TMT. Because of the potential risk of depression of hypoxic pulmonary vasoconstriction during SLV by volatile anesthetics, we primarily used a total intravenous technique (55 patients). With more experience, we also used a combination of volatile and intravenous anesthetics (16 patients) and, finally, volatile anesthetics only (11 patients). Data from patients anesthetised for TMT were compared with data from 22 patients operated with open thoracotomy from 1984 until 1992. RESULTS While the operating time (290.1 +/- 133.2 min for TMT vs. 312.3 +/- 113.6 for thoracotomy) and the anesthesia time (431.2 +/- 140.3 for TMT vs. 416.4 +/- 102.1 for thoracotomy) showed no significant differences, the TMT required an extremely long time of SLV (270.2 +/- 133.2 min) to gain access to the spine using left-sided double-lumen tubes. While the oxygenation index (PaO2/FiO2), as a marker for pulmonary oxygenation capacity, decreased significantly after initiation of SLV for TMT, it was markedly enhanced with increasing duration (270.2 +/- 133.2 min) of SLV. Oxygenation index showed no significant difference when comparing the different anesthetic techniques for TMT. CONCLUSION We conclude that despite the long duration of SLV, TMT is a reasonable alternative to open thoracotomy for thoracic neurosurgical spine procedures because of the substantial clinical benefits of accelerated return to physical activity, reduced complication rates and reduced intensive care unit and hospital stay.