For systematic study of the effects of clinical severity, time to recompression with oxygen, and re-treatment on outcome from spinal cord DCS, case records from the recompression chamber at the U.S. Naval Station Subic Bay were reviewed. Forty-nine cases of spinal cord DCS were classified using a numerical severity index and time to recompression with oxygen. Cases were divided by initial severity into mild, moderate, and severe groups and by time to recompression with oxygen into less than 12-h, 12-24-h, and greater than 24-h groups. Re-treatment effect was analyzed by severity after the first treatment and by the depth of the re-treatment table used. Severity after all treatment is strongly correlated with initial severity (r = 0.88) and moderately correlated with time to recompression with oxygen (r = 0.58). Response to treatment is significantly different among initial severity groups (P < 0.001). Delay to treatment worsens outcome for severely injured divers (P = 0.008). Residual severity after all treatments is highly correlated with severity after the first treatment (r = 0.97). There is no difference in re-treatment outcome by groups defined by severity after the first treatment or by 60- or 45-ft re-treatment tables.