Disagreement between capillary blood glucose and flash glucose monitoring sensor can lead to inadequate treatment adjustments during pregnancy - 31/08/19

Doi : 10.1016/j.diabet.2019.08.001

A. Sola-Gazagnes ^a,⁎ , P. Faucher ^b,^c,^d, S. Jacqueminet ^c,^d,^e, C. Ciangura ^c,^d,^e, D. Dubois-Laforgue ^a, H. Mosnier-Pudar ^f, R. Roussel ^g,^h, E. Larger ^a
^a Department of Diabetology, Paris Descartes University, Cochin Hospital, Assistance publique–Hôpitaux de Paris (AP–HP), 75014 Paris, France
^b Nutrition Department, Pitié-Salpêtrière Hospital, Assistance publique–Hôpitaux de Paris (AP–HP), 75013 Paris, France
^c Sorbonne Universities, University Pierre et Marie Curie–Paris 6, 75005 Paris, France
^d Inserm, UMR S U1166, ICAN, 75013 Paris, France
^e Department of Diabetology, Pitié-Salpêtrière Hospital, Assistance publique–Hôpitaux de Paris (AP–HP), 75013 Paris, France
^f Department of Endocrinology, Paris Descartes University, Cochin Hospital, Assistance publique–Hôpitaux de Paris (AP–HP), 75014 Paris, France
^g Inserm U1138, centre de recherche des Cordeliers, University Paris Diderot, Sorbonne Paris Cité, 75006 Paris, France
^h Diabetology, Endocrinology and Nutrition Department, Bichat Hospital, Assistance publique–Hôpitaux de Paris (AP-HP), 75018 Paris, France

^⁎Corresponding author.

Sous presse. Épreuves corrigées par l'auteur. Disponible en ligne depuis le Saturday 31 August 2019
Cet article a été publié dans un numéro de la revue, cliquez ici pour y accéder

Abstract

Objective

Continuous glucose monitoring tends to replace capillary blood glucose (CBG) self-monitoring. Our aim was to determine the agreement between CBG and a flash glucose monitoring system (Flash-GMS) in treatment decision-making during pregnancy.

Research Design and Methods

Insulin-treated women with either type 1 (n=25), type 2 (n=4) or gestational diabetes (n=4) were included. A Flash-GMS sensor was applied for 14 days. Women scanned the sensor whenever they monitored their CBG. The primary endpoint was the proportion of discordant therapeutic decisions they would have made based on Flash-GMS rather than CBG results. Glucose averages, mean absolute difference (MAD), mean absolute relative difference (MARD) and Flash-GMS accuracy were also estimated.

Results

Data for forty 14-day periods were available. Preprandial Flash-GMS and CBG values were 93±42mg/dL and 105±45mg/dL, respectively (P<10⁻⁴), and 2-h postprandial (PP) values were 106±45mg/dL and 119±47mg/dL, respectively (P<10⁻⁴). MAD was 14±22mg/dL preprandial and 15±24mg/dL 2-h PP; MARD was 19%; and 99% of glucose value pairs were within the clinically acceptable A and B zones of the Parkes error grid. Concordance rate for therapeutic decision-making was 80–85% according to ADA targets and 65–75% according to a pragmatic threshold. At different time points of the day, 83–92% of discordant results were due to Flash-GMS values being lower than their corresponding CBG values.

Conclusion

Flash-GMS tends to give lower estimates than CBG. Thus, in cases requiring therapeutic changes to treat or prevent hypo- or hyperglycaemia, 25–35% of choices would have been divergent if based on Flash-GMS rather than CBG.

Le texte complet de cet article est disponible en PDF.

Keywords : Continuous glucose monitoring, Insulin treatment, Pregnancy