CGM and CSII: More Effective Blood Glucose Control Than SMBG or MDI in T1D

CGM in womans arm
Woman with diabetes wearing a CGM and making a ponytail
This systematic review and meta-analysis included 14 studies comparing CSII with MDI (8 studies) and CGM with SMBG (6 studies) in type 1 diabetes.

Authors of a systematic review and meta-analysis found that continuous glucose monitoring (CGM) or continuous subcutaneous insulin infusion (CSII) were more effective than self-monitoring of blood glucose (SMBG) or multiple daily insulin injections (MDI), respectively, for maintaining glycemic control among patients with type 1 diabetes (T1D). Their findings were published in Diabetes, Obesity and Metabolism.

The researchers, from Imperial College London, searched publication databases (Medline, EMBASE, and Web of Science) for randomized controlled trials of diabetic health among patients with T1D published between 1999 and September 2020.

A total of 14 studies comparing CSII vs MDI (8 studies) and CGM vs SMBG (6 studies) were included in their analysis. All studies had an open-label design, owing to the nature of the intervention.

Regarding outcomes of glycated hemoglobin (HbA1c), studies reported an overall decrease among the CGM groups vs the SMBG groups (n=560; d, -0.62; 95% CI, -0.79 to -0.45; I2, 96%) and among the CSII vs MDI groups (n=301; d, -0.44; 95% CI, -0.67 to -0.22; I2, 95%).

The standard deviation of glucose was more effectively reduced by CGM (n=300; d, -2.41; 95% CI, -2.56 to -2.26; I2, 100%) than by SMBG and by CSII compared with MDI (n=118; d, -0.32; 95% CI, -0.50 to -0.13; I2, 100%).

Severe hypoglycemic events were significantly reduced by CGM vs SMBG (n=322; rate difference, -0.195 events per patient year; 95% CI, -0.32 to -0.07; I2, 58%) but not by CSII vs MDI (n=122; rate difference, 0.060; 95% CI, -0.17 to 0.29; I2, 0%).

Neither CGM nor CSII reduced diabetic ketoacidosis (DKA) events. Patients using CSII were found have a higher rate of DKA events than those using MDI (n=397; rate difference, 0.076; 95% CI, 0.02-0.13 per patient year; I2, 0%). There were similar rates of DKA with CGM vs SMBG (n=583; rate difference per patient year, -0.002; 95% CI, -0.03 to 0.02; I2, 33%). The researchers noted that based on these results, “the treatment effect for CGM when compared with CSII appeared marginally better.”

This meta-analysis was limited by the open-label designs and the differing inclusion criteria of the studies evaluated.

In summary, the investigators found that CGM and CSII more effectively decreased HbA1c and standard deviation of glucose than SMBG and MDI, respectively. Also, CGM (but not CSII) was associated with a decreased incidence of hypoglycemic events vs their comparators.

The investigators concluded that, “for adults with T1D, CGM, in addition to possible improvements in HbA1c and glucose variability, may provide significantly greater benefits for DKA and severe [hypoglycemia] incidence compared with CSII. If supported by evidence from dedicated trials, these findings could help to guide clinical decision making for clinicians choosing a first step in an evidence-based technology pathway for people with T1D.

Disclosure: An author declared affiliations with industry. Please refer to the original article for a full list of disclosures.


Thomas MG, Avari P, Godsland IF, Lett AM, Reddy M, Oliver N. Optimizing type 1 diabetes after multiple daily injections and capillary blood monitoring: Pump or sensor first? A meta-analysis using pooled differences in outcome measures. Diabetes Obes Metab. Published online July 19, 2021. doi:10.1111/dom.14498