Because body composition is altered during head-down bed rest (HDBR), body mass can not be used as an index of energy balance. Consequently diet allowances should not be based on body mass evolution but on fat mass changes. Though criticized, skinfold thickness (ST) is the costless, easiest and fastest method to use for such an objective. The aim of this study was to compare the percentage of body fat (%BF) estimated by ST with the isotope dilution of H2 18O. We compiled data from three HDBR campaigns, one on women (n=8) in November 1998 and two on the same men (n=8) in December 1997 (without countermeasure) and January 1998 (with thigh-cuffs countermeasure), according to a crossover design. Body composition was assessed before and after 6 days of HDBR. %BF was derived from the biceps, triceps, sub-scapular and sup-iliac ST according to Durnin and Wormersly (1974). Fat-free mass was measured on the same day by H2 18O dilution and fat mass was calculated by the difference with body mass and expressed as a percentage. Based on precision tests, the minimum measurable change by ST was 1.1%BF for single measurement point. Both intercepts (F (4,30)=0.89, P=0.45) and slopes (F (4,30)=0.74; P=0.57) of the ST versus dilution relationships were not affected by the periods (December vs January), experimental conditions (control vs HDBR vs HDBR + thigh cuffs) or sex allowing the derivation of a common relationship %BF(st)=0.94 x %BF(dil) (F (1,47)=97.9, P<0.0001; non-significant intercept excluded) with a bias between methods of -1.7+/-2.0 %BF (95% CI: -5.8, 2.4 %BF). ST can be used to measure %BF during HDBR provided great care is placed on training and changes are higher than 1.1 %BF. If the method can be applied for in-flight energy balance monitoring given the high observed energy deficit, a tight monitoring of the individual nutritional status as needed during simulation appears, however, dubious based on this solely method.