Researchers followed glucose monitor users for three years and found a possible kidney benefit

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Oct 12 2026

Continuous blood glucose monitoring could offer benefits beyond daily diabetes management, but what happens to blood sugar control and kidney health after three years of sustained use?

Study: Continuous Glucose Monitoring and Long-Term Glycemic and Kidney Outcomes. Image Credit: Andrey_Popov / Shutterstock

In a recent study published in JAMA Network Open, researchers evaluated the association between sustained real-time continuous glucose monitoring (rtCGM) and long-term glycemic control and kidney health markers in high-risk adults with type 2 diabetes (T2D).

Background

Can CGM help protect kidney health while improving diabetes control? Regular glucose monitoring is an important part of diabetes management for many people with T2D. Chronically elevated blood glucose levels may damage the kidneys over time. Continuous glucose monitoring (CGM) uses sensors beneath the skin to measure glucose levels, supporting informed diabetes management. rtCGM provides immediate glucose feedback, whereas intermittently scanned systems require manual scanning. Although randomized clinical trials (RCTs) have demonstrated short-term improvements in glycemic control with CGM, evidence regarding sustained benefits remains limited. Further long-term research is needed to clarify whether continued CGM use improves kidney outcomes and preserves kidney function.

About the study

Researchers conducted an observational comparative effectiveness study using electronic health records from Kaiser Permanente Northern California (KPNC), an integrated health care delivery system. They used a target trial emulation approach to estimate outcomes under sustained rtCGM use compared with no CGM initiation, using statistical methods to account for measured differences between participants. The institutional review board exempted the research from review and waived the requirement for informed consent because the secondary data analysis involved no intervention and posed minimal risk.

The sampling frame included 573,663 patients with T2D as of January 1, 2015. Eligible participants were adults aged 19 years or older who had not previously used CGM and had poor glycemic control, defined as hemoglobin A1c (HbA1c) greater than 8% during the preceding 12 months, or a history of emergency department visits or hospitalization for hypoglycemia during the preceding 36 months. Researchers compared rtCGM initiators with temporally matched individuals who had not initiated CGM. Researchers followed participants for up to 36 months.

The primary outcomes were HbA1c, urinary albumin-creatinine ratio (UACR), and estimated glomerular filtration rate (eGFR). UACR measures urinary albumin relative to creatinine; elevated levels suggest possible kidney damage, while eGFR estimates how well the kidneys filter blood. Researchers used statistical weighting to account for measured differences in patient characteristics, treatment adherence, loss to follow-up, and missing outcome measurements. Sustained use was identified through quarterly claims for rtCGM supplies rather than direct measurements of daily device use. The investigation focused on real-time CGM, so its findings may not apply to intermittently scanned systems.

Study results

The study included 25,648 participant assignments across the two groups: 2,771 rtCGM initiators and 22,877 controls. Because some patients contributed observations to both groups at different times, the analysis involved 25,371 unique individuals. Participants' mean age was 61.5 years, and 54.7% were men. At baseline, the mean HbA1c was 9.5%, indicating poor glycemic control in the sample. Most participants (96.9%) had an HbA1c greater than 8% during the preceding 12 months, and 5.5% had an emergency department visit or hospitalization for hypoglycemia in the previous 36 months. Overall, 92.9% had a baseline date in 2019 or later.

Participants initiating rtCGM generally had more complex clinical profiles than controls. They had longer diabetes duration, greater comorbidity burden, more kidney damage, and lower kidney function at baseline. Their mean diabetes duration was 17.6 years, compared with 12.6 years among controls. Insulin use was also more common among rtCGM initiators (89.1%) than controls (38.8%). Statistical weighting substantially improved the balance in baseline characteristics between the groups.

Sustained rtCGM use was associated with rapid and persistent improvements in glycemic control. One year after baseline, the weighted mean HbA1c level was 7.83% among rtCGM initiators, compared with 8.93% among controls, representing a difference of 1.10 percentage points. Improvements persisted throughout follow-up. At two years, the difference was 0.90 percentage points. After three years, mean HbA1c remained lower in the rtCGM group than in controls (7.83% vs. 8.58%), with a difference of 0.75 percentage points.

No statistically significant differences in UACR were observed between groups at baseline or after one and two years. However, at three years, sustained rtCGM use was associated with a lower mean UACR. Values were 153 mg/g in the rtCGM group and 278 mg/g in controls, a difference of about 125 mg/g. This finding suggested a later reduction in albuminuria among participants who maintained monitoring. However, lower UACR does not establish preserved kidney function, and the study could not determine whether improved glucose control directly accounted for the reduction.

Weighted eGFR estimates were already lower at baseline in the rtCGM group and remained significantly lower during the first two years. By the three-year assessment, the difference was no longer statistically significant. Overall, the study did not demonstrate that sustained rtCGM use preserved kidney filtration function during follow-up.

Substantial discontinuation and crossover between monitoring groups limited follow-up. During follow-up, 44.3% of rtCGM initiators were censored from further analysis after discontinuing rtCGM, while 18.7% of controls were censored after initiating CGM. At three years, the UACR analysis included 331 rtCGM initiators and 3,890 controls, substantially fewer than at baseline. Although statistical weighting addressed differential attrition and missing measurements, the smaller sample increases uncertainty in the longer-term kidney findings. These limitations affect the interpretation of the study's longer-term outcomes, particularly the later reduction in albuminuria.

Conclusions

Sustained rtCGM use was associated with rapid, durable improvements in glycemic control and a later reduction in UACR among high-risk adults with T2D. However, the study found no evidence of kidney function preservation, as measured by eGFR, after three years. The findings suggest that sustained rtCGM use may improve blood glucose control and reduce albuminuria, but they do not show that the technology prevents kidney disease progression.

Nevertheless, the observational design, substantial attrition, and possibility of residual confounding limit causal interpretation. Patients who maintained rtCGM use may also have differed in motivation or clinical support from those who did not, potentially influencing the results. Dexcom, a CGM manufacturer, and other healthcare and research organizations funded the research. The authors reported that funders had no role in the study's design, analysis, or publication decisions.

Longer-term RCTs are needed to confirm sustained glycemic benefits and determine whether rtCGM can preserve kidney function or delay diabetic kidney disease progression.