Patients were recruited between August and December 2011 from the CF Clinic at Royal Prince Alfred Hospital, Sydney, Australia. This clinic has 240 patients, of whom 25% have CFRD. The inclusion criterion was the immediate requirement of a management decision that might be enhanced by the additional glucose metabolic information provided by CGM as determined by the clinic endocrinologists (LS & KS). Exclusion criteria were a current acute exacerbation of respiratory disease and significant coagulopathy with CF-related liver disease. Approval for the study was obtained from the Hospital’s Ethics Committee (X11-0106 & HREC/11/RPAH/146). Informed consent was obtained before enrolment to the study. The participants undertook CGM (iPro2®; Medtronic MiniMed, Northridge, CA, USA) for six days while continuing their normal diet, lifestyle and treatments. Fingerstick capillary blood glucose levels, before meals and bedtime, were used to retrospectively calibrate the monitors. After the CGM period, participants completed a qualitative questionnaire to assess their experience with the device. Using CGM output provided by CareLink™ Pro software (Medtronic MiniMed, Northridge, CA, USA) and standard clinical information, a management decision was made incorporating published clinical guidelines and patient input (5). At baseline an OGTT was conducted. At baseline and six months follow up, participants had a clinical examination including respiratory function (FEV1), HbA1c, and BMI assessment. Descriptive statistics were performed to examine clinical data at baseline and 6 months follow up. Pearson’s correlation coefficient was used to determine associations between variables.

Fourteen participants (11 females; median age 25 years [range 21–58]) enrolled in the study and all successfully undertook CGM for the prescribed period. The Table reports individual participant clinical data (and group means) at baseline and follow up along with CGM informed therapy decisions/changes. Four participants (ID# 2, 9, 11 and 14) were excluded from longitudinal analyses with reasons provided in the Table. In the 10 valid cases at follow up, 7 participants had a reduction in HbA1c (mean change -0.8% (95% Confidence Interval: -1.7 to 0.1), 6 participants had a reduction in BMI (mean change 0.0kg/m2, 95% CI: -0.8 to 0.7) and 8 participants had a reduction in FEV1 (L) (mean change -0.12 L, 95% CI: -0.24 to 0.0) and its %-predicted value (mean change -4.7, 95% CI: -9.7 to 0.3). Interpretation of these pilot study results is limited by the small sample size, pre-post design and lack of CGM at follow up.

Table 1

Individual participant clinical data and CGM informed therapy decisions/changes

Abbreviations: Blood Glucose Level (BGL), Continuous Glucose Monitoring (CGM), Cystic Fibrosis-Related Diabetes (CFRD), Follow Up (FU), Forced Expiratory Volume in One Second (FEV1), Gestational Diabetes Mellitus (GDM), Glycemic Index (GI), Home Blood Glucose Monitoring (HBGM), Impaired Glucose Tolerance (IGT), Normal Glucose Tolerance (NGT), Oral Glucose Tolerance Test (OGTT), Percent Predicted FEV1 (%P), Therapy (Tx).

Nonetheless, this study showed an apparent improvement in HbA1c, and provides preliminary evidence of the clinical utility of CGM in adults with CF. Despite the HbA1c improvement in the majority of participants, the traditional markers of improvement in CF, lung function and weight, did not reflect this. The typical inflammation related fluctuations that occur in lung function and BMI in patients with CF may explain why we failed to see an improvement in these measures. Positive correlations were found at baseline between HbA1c and i) maximum CGM glucose concentration (R2=0.676; P<0.01), ii) mean CGM glucose concentration (R2=0.871; P<0.05) and iii) area under the curve (AUC) above 7.8 mmol/L (R2=0.954; P<0.001). Baseline BMI positively correlated with FEV1 (R2=0.436, P<0.05) and percent-predicted FEV1 (R2=0.464, P<0.05). Similar associations were observed at six months. None of these correlations are unexpected. Neither BMI nor FEV1 were significantly correlated with HbA1c at baseline or follow-up, again consistent with the debatable clinical utility of HbA1c in CF.

A randomized controlled trial is warranted to examine the long-term clinical benefits of incorporating intermittent CGM into clinical practice as a tool to improve glucose metabolic status and as a result, weight status and lung function.

Overall, participants’ indicated that CGM was well tolerated and that the visual presentation of the iPro2® CGM data made discussions about their physician’s reasons for altering treatment easier to understand. CGM importantly allowed meaningful clinical decisions to be made with the participation of the patient. CF patients are used to being involved in complex management decisions and are wary about additional burdens to their intensive daily self-management regimens. CGM has been found to enhance management of Type 1 Diabetes in motivated individuals (7) so it is likely that the decisions based on our data may yield similar results.

In summary, this pilot study demonstrated that in adults with CF, CGM was well-tolerated, provided additional useful clinical data for physicians and appeared to be associated with improved glycemic control in a group of varying age and clinical status. However, improved glycemic control did not correspond with improved lung function which declined in this sample. In Australia, the cost of the glucose sensor is currently over three times the scheduled fee for an OGTT under the universal health insurance system (Medicare). This consideration plus the initial cost for iPro2® may limit CGM application in public hospital services. Future controlled studies are needed to examine if CGM use alters behavior, long term glycemic control and other clinical outcomes in adult patients with CF.

Acknowledgements

The authors wish to acknowledge the assistance of Medtronic Diabetes Australia, in lending an additional continuous glucose monitor for part of the study. Medtronic Diabetes Australia played no other part in the study.