Mucosal gastric injury was induced in vivo by oral administration of 75% ethanol to rats that were intraperitoneally pretreated with saline (controls), 4 mg/kg of verapamil, 0.4 mg/kg of nifedipine or 160 mg/kg of MgCl2. The glandular stomachs were used for macroscopic and histologic evaluation of mucosal lesions. The gastroprotective effects of substances tested against ethanol-induced mucosal damage in vitro were examined: 1). in rats intraperitoneally pretreated with calcium antagonists and 2). after adding of 10(-4) M verapamil, 10(-6) M nifedipine or 10(-2) M MgCl2 directly to minced gastric mucosa of untreated rats. This effect was measured by DNA synthesis. Gross and histologic evaluation showed that rats pretreated with nifedipine or MgCl2 had significantly decreased ethanol-induced gastric injury compared to controls, whereas those pretreated with verapamil had significantly increased injury. On the contrary, all verapamil, nifedipine and MgCl2 treatments, administrated intraperitoneally or by exposure to an incubation mixture were equally effective in reducing gastric mucosal damage following to ethanol treatment in vitro. We conclude that the differential effects of verapamil and nifedipine on ethanol-damaged gastric mucosa in vivo, but not in vitro, suggests the existence of different actions of these calcium antagonists more on systemic rather than local protective mechanisms, such as gastric mucosal blood flow, mucosal barrier, or cell renewal.