Background: The usual glycemic supervision based on repeated measures of the capillary glycemia with the strip of which reliability can be taken at fault especially in critically ill patients. The objective of this work was to evaluate the reliability of capillary glycemia measurement in intensive care unit.
Material and methods: [corrected] The study was performed in a medical intensive care unit, during six months (July to December 2004). The capillary glycemias were carried out with only one blood glucose monitor and were compared with the referenced glycemias. The correlation between two methods was carried out by the parametric Pearson coefficient. The error percentage of the blood glucose monitor was calculated. The agreement of two methods was evaluated by Bland and Altman test.
Results: Were performed 245 glycemic pairs for 198 patients, their mean age was 42.6 +/- 14.6 years, 30(15%) among them were diabetics. The referenced glycemia was 8.7 +/- 5.5 mmol/L and the capillary glycemia was 8.3 +/- 4.6 mmol/L. The correlation coefficient was 0.97 (p<0.001); 56 percent of the values had an error percentage lower than 10, 27 percent between 10 and 20%, and there were 17 percent discordant values with an error percentage higher than 20. The method of Bland and Altman objectived a skew of 0.40 mmol/L and a bad precision of +/- 2.90 mmol/L.
Conclusion: The reliability of capillary glycemias was not satisfactory with bad precision and high percentage of discordance. The capillary glycemias must be interpreted with precaution in intensive care unit.