The Diving Committee of the Undersea and Hyperbaric Medical Society has reviewed available evidence in relation to the medical aspects of rescuing a submerged unresponsive compressed-gas diver. The rescue process has been subdivided into three phases, and relevant questions have been addressed as follows. Phase 1, preparation for ascent: If the regulator is out of the mouth, should it be replaced? If the diver is in the tonic or clonic phase of a seizure, should the ascent be delayed until the clonic phase has subsided? Are there any special considerations for rescuing rebreather divers? Phase 2, retrieval to the surface: What is a "safe" ascent rate? If the rescuer has a decompression obligation, should they take the victim to the surface? If the regulator is in the mouth and the victim is breathing, does this change the ascent procedures? If the regulator is in the mouth, the victim is breathing, and the victim has a decompression obligation, does this change the ascent procedures? Is it necessary to hold the victim's head in a particular position? Is it necessary to press on the victim's chest to ensure exhalation? Are there any special considerations for rescuing rebreather divers? Phase 3, procedure at the surface: Is it possible to make an assessment of breathing in the water? Can effective rescue breaths be delivered in the water? What is the likelihood of persistent circulation after respiratory arrest? Does the recent advocacy for "compression-only resuscitation" suggest that rescue breaths should not be administered to a non-breathing diver? What rules should guide the relative priority of in-water rescue breaths over accessing surface support where definitive CPR can be started? A "best practice" decision tree for submerged diver rescue has been proposed.