Skip to content

Continuous Glucose Monitoring Predicts Long-term Nutritional Outcomes in Adults With Cystic Fibrosis

Key findings

  • This prospective, observational study investigated the role of continuous glucose monitoring (CGM) and advanced glycation endproducts (AGE) in predicting long-term clinical outcomes in adults with cystic fibrosis with and without diabetes
  • CGM-derived baseline measures of glycemic variability and % time in the target range 70–180 mg/dL significantly correlated with a decline in weight and decline in body mass index over one year
  • No CGM measure predicted significant clinical changes after two years, and CGM measures and hemoglobin A1c did not correlate with a decline in pulmonary function or incidence of cystic fibrosis exacerbation or hospitalization after one or two years
  • Baseline AGE level was associated with both weight changes and pulmonary decline at year 1, implying this may be a potentially useful marker in people with cystic fibrosis
  • A larger, long-term prospective study might show that CGM measures of glycemia also correlate with future pulmonary decline

Cystic fibrosis–related diabetes (CFRD) often has an insidious onset, so screening for early glycemic abnormalities is important. The current recommendation is an annual two-hour oral glucose tolerance test (OGTT), but compliance is low.

In a cross-sectional study previously reported in The Journal of Clinical Endocrinology & Metabolism, researchers at Massachusetts General Hospital found that certain measures derived from continuous glucose monitoring (CGM) significantly correlated with body mass index (BMI) and pulmonary function, in addition to other key clinical findings, in people with cystic fibrosis (PwCF). Furthermore, the CGM values correlated more strongly than hemoglobin A1c.

Now, Melissa S. Putman, MD, an endocrinologist in the Diabetes Clinical Research Center at Mass General, and colleagues have studied the long-term predictive value of CGM measures using the same cohort of adults with and without CFRD. In Frontiers in Endocrinology (Lausanne), they say several key CGM measures of hyperglycemia and glycemic variability predicted a decline in nutritional outcomes over one year.


The cohort comprised 77 adults with an established diagnosis of CF and a mean age of 33 (range, 18–70), who were recruited between January 2017 and June 2020. Participants were evaluated at baseline and three months later, with an additional study visit in the case of interval CF exacerbation.

At the baseline visit, the 52 participants without pre-existing CFRD underwent an OGTT and the 25 others provided a blood sample for determination of HbA1c. An HbA1c level was also obtained at each study visit. In addition, at each study visit and any exacerbation visit a CGM sensor was placed and worn for 14 days.

For the current study, medical records were reviewed one and two years after the final study visit.

Correlations With Glycemic Measures

31 participants (40%) had CFRD at baseline, including six diagnosed through baseline OGTT. One additional participant was diagnosed with CFRD during year 1 of follow-up and another during year 2. After adjusting for age, gender, and use of elexacaftor/tezacaftor/ivacaftor (ETI):

  • Percent time in range 70–180 mg/dL (TIR) and measures of glycemic variability (standard deviation and coefficient of variation) correlated with decline in weight at year 1 but not at year 2
  • Baseline TIR correlated with BMI decline at year 1, but no baseline glycemic measure correlated with BMI decline at year 2
  • CGM and HbA1c at baseline did not predict decline in forced expiratory volume in 1 second, the number of CF exacerbations or the number of hospitalizations in the preceding year

Subset Analysis

The researchers separately analyzed the subset of participants who did not initiate ETI during the follow-up period (year 1, n=52; year 2, n=44). HbA1c correlated with decline in weight at year 1 only, whereas multiple CGM measures showed strong correlation with weight decline and BMI decline.

These results align with recently published concerns about the clinical utility of HbA1c as a screening tool in PwCF.

Advanced Glycation Endproducts

As part of the follow-up study, the team became the first to examine correlations between levels of advanced glycation endproducts (AGEs), dysglycemia and long-term clinical outcomes in PwCF.

AGEs are byproducts of metabolic interactions between glucose and proteins/lipids. They accumulate in the setting of oxidative stress and hyperglycemia, which may affect nutritional and pulmonary outcomes in CF. Levels of AGEs can be assessed rapidly with noninvasive devices that use ultraviolet light to measure autofluorescence in human skin.

In this study, AGE values were notably higher at baseline in individuals with CFRD than in those without (2.08 vs. 1.94 AU, P=0.005). The levels positively correlated with multiple CGM measures of hyperglycemia and glycemic variability and negatively with weight at baseline. Furthermore, baseline AGE values correlated with decline in weight at year 1 and year 2 and with change in FEV1 at year 1.

Clinical Implications

Maintenance of normal weight and BMI has long been a vital component of CF care because of their strong relationships with pulmonary function and mortality. The consistent associations of CGM measures with weight and BMI in this study are noteworthy. AGEs measured by skin autofluorescence have potential to serve as an additional screening tool to predict key clinical changes in PwCF.

Learn more about the Diabetes Clinical Research Center

Refer a patient to the Diabetes Unit

Related topics


Jordi Merino, PhD, of the Diabetes Unit, and colleagues found good agreement between continuous glucose monitors (CGMs) of different brands when measuring glycemic responses to different foods and meals. Their findings suggest CGMs could someday be used to provide precision nutrition advice to healthy individuals.


Melissa S. Putman, MD, and colleagues found measures of hyperglycemia and glycemic variability provided by continuous glucose monitors were superior to hemoglobin A1c in distinguishing between patients with cystic fibrosis (CF) who did and did not have CF-related diabetes.