How well do self-monitored capillary glucose measurements predict the mean blood glucose from 24-hour continuous monitoring in endocrine practice?

Endocr Pract. 2014 Jul;20(7):650-6. doi: 10.4158/EP13337.OR.


Objective: The mean blood glucose (MBG) level obtained from self-monitored capillary glucose (SMCG) data is a readily available metric of glycemic control for patients and their physicians. However, whether there is agreement between SMCG MBG levels and MBG levels obtained from 24-hour intensive glucose sampling is unclear. Therefore, we analyzed the relationship between MBG levels derived from SMCG data and glucose data derived from continuous glucose monitoring (CGM).

Methods: SMCG and CGM were concurrently performed in 104 patients with diabetes and prediabetes over 3 to 6 days. MBG data obtained from SMCG and CGM were compared by standard correlation and Bland-Altman analyses.

Results: SMCG and CGM MBG data from the longest duration of sampling were highly correlated (r = 0.965; P<.001). Single-day MBG estimates from both sources were also highly correlated, with r values ranging from 0.833 to 0.927. A SMCG MBG level of 166.1 ± 55 mg/dL (derived from 14.1 ± 4.6 samples) tended to slightly underestimate the concurrent CGM MBG level of 171.1 ± 56.4 mg/dL (derived from 1,063 ± 283 samples). The SMCG MBG was within 30 mg/dL of the CGM MBG in 94.6% of patients and within 15 mg/dL in 67% of patients. The difference between the estimates tended to increase with increasing SD of the MBG obtained from CGM (r = 0.38; P<.0001).

Conclusion: MBG estimated from SMCG is a reasonable estimate of a patient's CGM MBG over the same period of time and with caveats could be used as a practical guide for long-term glycemic control that can be considered in tandem with the patient's hemoglobin A1c in endocrine practice.

MeSH terms

  • Blood Glucose Self-Monitoring*
  • Capillaries
  • Glycated Hemoglobin / analysis
  • Humans
  • Retrospective Studies


  • Glycated Hemoglobin A
  • hemoglobin A1c protein, human