To determine the reliability of five blood glucose meters (BGMs) at various simulated altitudes using a hypobaric chamber. Blood glucose levels (ranged from 1.5 to 26.3 mmol/l, according to the reference method) were measured in 18 venous blood samples by each BGM at 200, 1,000, and every 500 m up to 4,000 m in a hypobaric chamber, where temperature and humidity were held constant. Four BGMs underestimated and one overestimated blood glucose concentration while barometric pressure decreased. The average percent error varied in relation to simulated altitude from 0.26 +/- 4.8% (SD) at 200 m to -28.9 +/- 4.5% at 4,000 m (Glucometer 3; P < 0.05), from 28.4 +/- 5.7 to 49.3 +/- 5.9% (Accu-Chek Easy; P < 0.05), from -10.5 +/- 2.6 to 19.8 +/- 4.3% (Tracer, P < 0.05), from -5.5 +/- 2.6 to -11.2 +/- 3.0% (Reflolux; NS), and from 17.8 +/- 4.3 to 14.8 +/- 3.6% (One Touch; NS). The most accurate seemed to be the Reflolux, except for high blood glucose levels at simulated high altitudes. The One Touch II showed a good agreement, whatever the barometric pressure and the range of blood glucose concentrations. The highest underestimation was seen with the Glucometer 3. Except for the Accu-Chek Easy, low barometric pressure underestimated the BGM results in comparison with measurements taken at simulated low altitudes. The lack of accuracy and consistency of performance > 2,000 m should be known by diabetic patients practicing sports activities, such as trekking or skiing at high altitudes.