Advances in treatment have led to an increase in survival in people with cystic fibrosis (CF). As patients age with CF, they experience an increased rate of complications, including cystic fibrosis related diabetes (CFRD) [
Recommendations for diagnosis, screening, and management of CFRD were written in a jointly created clinical care guideline by The Cystic Fibrosis Foundation (CFF), the American Diabetes Association (ADA), and the Pediatric Endocrine Society (PES) [
Because of the time-consuming nature of the test, and requirement that patients fast prior to testing, adherence to the guidelines has been challenging; approximately one-quarter of adults with CF were tested in 2014 [
We hypothesized that abnormal CGM results would be associated with deterioration in health. Our objectives for this study were to determine whether abnormal CGM results are associated with subsequent development of CFRD, lung function, and BMI decline and increased rate of CF pulmonary exacerbations.
To evaluate the correlation between CGM and OGTT for detection of CFRD in adult patients with CF, we conducted a prospective single center trial in 2009-2010 in which 21 adult patients due for routine OGTT underwent simultaneous 3-day CGM and 2-hour 75 g OGTT [
OGTTs were conducted using the 2-hour 75-gram oral glucose tolerance test. For OGTT, the following definitions were used: fasting glucose >126 mg/dL or 2-hour glucose >200 mg/dL indicated CFRD, fasting glucose 100–125 mg/dL indicated impaired fasting glucose (IFG), and 2-hour glucose 140–199 mg/dL indicated impaired glucose tolerance (IGT) [
Each subject’s highest lung function and BMI for the year prior to OGTT/CGM testing, and for each of the 5 years following OGTT/CGM testing, were extracted from the CF registry. The number of exacerbations each subject had for each of the 5 years following OGTT/CGM testing was also extracted from the CF registry, and the average number of exacerbations per year was calculated. During the period of retrospective review, three subjects died and three patients moved away from the center; all available data points were included for these subjects. Six patients began chronic use of Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) modulators during the period of retrospective review (two patients began ivacaftor, and four patients began lumacaftor/ivacaftor). The highest FEV1 and BMI for the year prior to initiation of CFTR modulators were used for analysis of decline in these parameters over time.
Means and standard deviations are reported for demographic information (Graph Pad Prism v6.07). Agreement between test methods (OGTT and CGM) was evaluated using both Pearson correlation and the Bland-Altman method. (Graph Pad Prism v6.07). Log-rank (Mantel-Cox) test was used to evaluate whether abnormal CGM predicts subsequent development of CFRD. Using two-sample
Twenty-one subjects were enrolled. Seventy-six percent (16/21) of subjects were female. The average age of the subjects was 32.4 years (SD 13.1; median 27.0, range 20–65). Fifty-two percent (11/21) of subjects were homozygous for the F508del mutation; 90% (19/21) of subjects had at least one copy of the F508del mutation. All patients were pancreatic insufficient. Average percent of predicted FEV1 and average BMI of subjects at the time of enrollment was 68.8 (SD 5.5; range, 22–109) and 21.6 (range 17.1–29.1; SD 0.72), respectively (Table
Subject demographics at enrollment.
Variable | Enrolled | Complete CFRD testing available |
---|---|---|
| | |
Age (years) | 32.4 SD 13.1 | 33.8 SD 13.7 |
Female | 76% (16/21) | 72% (13/18) |
Homozygous for F508del | 52% (11/21) | 44% (8/18) |
At least one copy of F508del | 90% (19/21) | 89% (16/18) |
FEV1 (L) | 2.27 SD 0.84 | 2.23 SD 0.21 |
FEV1% predicted | 68.8 SD 5.5 | 66.7 SD 6.3 |
BMI | 21.6 SD 0.72 | 21.9 SD 0.81 |
One patient did not have adequate data for OGTT interpretation due to improper collection of the timed samples. Two patients did not have adequate data for interpretation by CGM due to failure of the device and/or data retrieval. AIC was ≤5.9% for all 18 patients with complete data. For these 18 patients (mean results for test methods shown in Figure
Average and individual glucose values by test method. Individual subject glucose values for initial study OGTT at each time point measured and maximum glucose value obtained by CGM are shown. Error bars represent average subject values for each OGTT time point and average maximum CGM value.
Bland-Altman evaluation of agreement between glucose testing methods. (a) Two-hour OGTT values compared to peak CGM values. (b) One-hour OGTT values compared to peak CGM values. Dotted lines indicate bias and 95% limits of agreement.
One subject met criteria for CFRD by OGTT (as well as by the CGM criteria defined for this study: glucose > 200 mg/dL on two dates) at the time of the prospective study. Of those remaining subjects with complete CGM and OGTT data (
Original and current glucose metabolism.
Original glucose metabolism interpretation | Current glucose metabolism | |||
---|---|---|---|---|
OGTT | 1H OGTT | CGM | Status | |
1 | Normal | Normal | IGT | Normal |
2 | Normal | IGT | IGT, IFG | Normal |
3 | Normal | IGT | IGT | Normal |
4 | Normal | IGT | IFG, IGT | Normal |
5 | Normal | Normal | IFT, IGT | Normal |
6 | Normal | Normal | IGT | IGT |
7 | Normal | IGT | IGT | |
8 | Normal | IGT | IFG, IGT | IGT |
9 | Normal | Normal | IFG, IGT | CFRD |
10 | Normal | Normal | IGT | CFRD |
11 | IFG | Normal | Normal | Normal |
12 | IGT | IGT | IGT | IGT |
13 | IGT | CFRD | IGT | CFRD |
14 | IFG | CFRD | CFRD | CFRD |
15 | IFG | CFRD | CFRD | CFRD |
16 | IGT | CFRD | CFRD | CFRD |
17 | IGT | CFRD | CFRD | CFRD |
18 | CFRD | CFRD | CFRD | CFRD |
Abnormal CGM results predict the development of guideline criteria diagnosed glucose abnormalities over time. CGM results were categorized as normal (green circles), impaired (yellow squares), or consistent with CFRD (red triangles). Time to event was analyzed using the
Of the 10 subjects who have not met criteria for CFRD since study enrollment (two of these subjects died, and 3 moved from the center), 5 subjects (50%) had at least one OGTT in the 4-5 years subsequent to study enrollment. Although 3 of 5 of the subjects had normal OGTTs at enrollment (CGM results indicated IGT in all three subjects), those 3 subjects have subsequently been diagnosed with IGT by OGTT. All 10 subjects who have not met criteria for CFRD since study enrollment have had at least one RPG and/or A1C measured in the 4-5 years subsequent to enrollment. None of the 10 subjects met criteria for CFRD based on an A1C ≥6.5% or RPG ≥200 mg/dL in conjunction with polyuria and polydipsia.
Using current glucose metabolism as the outcome (Table
There was no difference in baseline lung function between the patients who went on to develop CFRD and those who did not (
There was no difference in baseline BMI between the patients who went on to develop CFRD and those who did not (
There was no difference between rate of exacerbations in subjects who went on to develop CFRD versus those who did not (
The goal of the prospective pilot study was to compare the ability of the gold standard OGTT versus CGM to evaluate glucose metabolism in CF. CGM detected
CFRD occurs in approximately 40–50% of adults with CF [
Very few patients with CF have completely normal glucose metabolism [
Although rates of screening for CFRD using OGTT have increased since the guidelines [
Because of the difficulty in obtaining annual screening OGTT, it would be ideal to use a sensitive test that does not require fasting (in a patient population that suffers from malnutrition) and an additional 2-3 hours in clinic. Although AIC is quick and easy to assess, it is not recommended for CFRD screening because it demonstrates a low degree of correlation with OGTT and is insensitive for the diagnosis of CFRD [
O’Riordan and colleagues sought to validate the use of CGM in a prospective cohort of children and adolescents with CF [
Subsequent to the validation of CGM in children and adolescents with CF, other researchers have examined whether abnormal glucose metabolism by CGM predicts clinical outcomes. Hameed et al. prospectively studied 33 children with CF who were scheduled to undergo OGTT as part of routine health screening. Twenty-five of the children agreed to simultaneously undergo CGM. They found that the amount of time a patient experienced elevated blood glucose, specifically a blood glucose of >7.8 mmol/L (140.4 mg/dL) for ≥4.5% of the time, was associated with declining nutritional status and lung function in the preceding 12 months [
Six of the patients began CFTR modulators during the retrospective observational study period. Although a formal study of the impact of ivacaftor on glucose metabolism is underway (NCT02039986), to date only a case report and a case series have been published [
In adult patients with CF, CGM identified a greater degree of impaired glucose metabolism than the gold standard 2-hour OGTT. Furthermore, glucose >200 mg/dL on two dates by CGM correctly identified subjects who developed CFRD over time. The sensitivity and negative predictive values were similar for both CGM and 1-hour OGTT for subsequent development of CFRD, and both exceeded those values for the traditional OGTT measurements. Patient tolerance of CGM is excellent [
American Thoracic Society
Body mass index
Continuous glucose monitoring
Cystic fibrosis
Cystic Fibrosis Foundation
Cystic fibrosis related diabetes
Cystic Fibrosis Therapeutics Development Network
Cystic Fibrosis Transmembrane Conductance Regulator
Forced expiratory volume in 1 second
Hemoglobin A1C
Impaired fasting glucose
Impaired glucose tolerance
Negative predictive value
Oral glucose tolerance test
Positive predictive value
Random plasma glucose.
J. S. Janssen and F. J. Accurso retired. This work was partially presented at the North American Cystic Fibrosis Conference in October 2010 in Baltimore, MD.
The authors have no competing interests with any companies whose products or services may be discussed in this article.
Dr. J. L. Taylor-Cousar verified and is responsible for the accuracy of the data reported. She contributed as the principal investigator of the retrospective research project and wrote the manuscript. J. S. Janssen was the principal investigator on the original CGM study. She enrolled patients and interpreted CGM and OGTT data. M. K. Sontag contributed to study design and manuscript preparation. C. G. St. Clair and M. C. Jones enrolled patients into the clinical study and contributed to manuscript preparation. A. Wilson and S. J. Brayshaw assisted with data collection and contributed to manuscript preparation. K. M. Pickard, C. S. Chacon, and C. M. Barboa assisted with data collection. D. P. Nichols and M. T. Saavedra contributed to manuscript preparation. F. J. Accurso and J. A. Nick contributed to study design and manuscript preparation.
This work was supported by NIH/NHLBI 1K23HL103801-01A1 NIH/NHLBI, NIH/NCRR Colorado CCTSI Grant no. UL1 RR025780, and CFF JANSSE09Q10. The authors would like to thank the patients who participated in the study.