Fulltext

Comparison between capillary glucose measured with a Contour glucometer and plasma glucose in a population survey

Stauffer, Florine ; Viswanathan, Bharathi ; Jean, Maygane ; Kinabo, Prosper ; Bovet, Pascal

In: LaboratoriumsMedizin, 2016, vol. 40, no. 2, p. 133-139

Ajouter à la liste personnelle
    Titre
    • Comparison between capillary glucose measured with a Contour glucometer and plasma glucose in a population survey
    Auteur
    • Stauffer, Florine. Institute of Social and Preventive Medicine, University Hospital of Lausanne, Lausanne, Switzerland
    • Viswanathan, Bharathi. Unit for Prevention and Control of Cardiovascular Diseases, Ministry of Health, Victoria, Republic of Seychelles
    • Jean, Maygane. Clinical Laboratory, Seychelles Hospital, Ministry of Health, Victoria, Republic of Seychelles
    • Kinabo, Prosper. Clinical Laboratory, Seychelles Hospital, Ministry of Health, Victoria, Republic of Seychelles
    • Bovet, Pascal
    Type de document
    • Postprint
    Identifiant OAI-PMH
    • oai:doc.rero.ch:326219
    Summary
    Background: Capillary glucose (CG) measured with point-of-care glucometers can provide useful approximation of plasma glucose (PG) in selected circumstances but the validity of measurements has been adequately assessed only for a few glucometers. Methods: We assessed the difference between CG measured with a glucometer (Contour Ascensia, Bayer) and PG measured with a standard laboratory method in participants to a population-based cardiovascular survey in the Seychelles (sample size 1227). Results: CG correlated well with PG (r=0.94; p<0.001). The overall difference between PG and CG was -0.55 mmol/L for PG <4.0 mmol/L (n=19; 95% CI -0.92; -0.18); 0.14 mmol/L for PG 4.0-4.9 (n=344; 95% CI 0.08-0.20) and increased according to PG up to 1.64 mmol/L for PG >9.0 mmol/L (n=68; 95% CI 1.36; 1.91). The prevalence of diabetes in the study sample was 29% lower with CG than with PG (8.6% vs. 12.1%) but this bias could be corrected by analytical re-calibration. Conclusions: CG underestimated PG, although the bias was small among persons with low glycemia. This systematic difference may bear little significance when the purpose is to simply monitor blood glucose but has large impact on prevalence estimates at the population level if CG values are not adequately calibrated.