To determine if athletes' muscle mass affects the usefulness of urine specific gravity (U(sg)) as a hydration index. Nine rugby players and nine endurance runners differing in the amount of muscle mass (42 +/- 6 vs. 32 +/- 3 kg, respectively; P = 0.0002) were recruited. At waking during six consecutive days, urine was collected for U (sg) analysis, urine osmolality (U(osm)), electrolytes (U[Na+], U[K+] and U[Cl-]) and protein metabolites (U([Creatinine]), U([Urea]) and U([Uric acid])) concentrations. In addition, fasting blood serum osmolality (S(osm)) was measured on the sixth day. As averaged during 6 days, U(sg) (1.021 +/- 0.002 vs. 1.016 +/- 0.001), U(osm) (702 +/- 56 vs. 554 +/- 41 mOsmol kg(-1) H(2)O), U([Urea]) (405 +/- 36 vs. 302 +/- 23 mmol L(-1)) and U([Uric acid]) (2.7 +/- 0.3 vs. 1.7 +/- 0.2 mmol L(-1)) were higher in rugby players than runners (P < 0.05). However, urine electrolyte concentrations were not different between groups. A higher percentage of rugby players than runners (56 vs. 11%; P = 0.03) could be cataloged as hypohydrated by U(sg) (i.e., >1.020) despite S (osm) being below 290 mOsmol kg(-1) H(2)O in all participants. A positive correlation was found between muscle mass and urine protein metabolites (r = 0.47; P = 0.04) and between urine protein metabolites and U(sg) (r = 0.92; P < 0.0001). In summary, U(sg) specificity to detect hypohydration was reduced in athletes with large muscle mass. Our data suggest that athletes with large muscle mass (i.e., rugby players) are prone to be incorrectly classified as hypohydrated based on U(sg).