Effectiveness of continuous glucose monitoring in children, adolescents, and young adults with poorly controlled type 1 diabetes

Abstract

Objective

The purpose of this study was to determine the effect of continuous glucose monitoring (CGM) on glycemic control in children, adolescents, and young adults ages 7–21 years with poorly controlled diabetes HbA1c 9.0% or more (74 mmol/mol IFCC).

Materials and Methods

The primary outcome was improvement in HbA1c. The secondary outcome included self-reported hypoglycemia. This 12-week study used a prospective, one-group, pre- and posttest pre-experimental design with a convenience sample. The study used the Medtronic Guardian CGM with Enlite Sensor.

Results and Conclusions

Thirty-three subjects enrolled in the study. The mean age of the participants was 15.57 years, range was 11–20 years, 47.6% were male, and 52.4% were female. Twenty-one (63.6%) completed the final study visit. There was a clinically and statistically significant reduction of 1.46 (SD = 1.6711) (p = .001) in HbA1c at 12 weeks. Fifteen of the 21 participants (71.4%) had an HbA1c reduction of greater than 0.5%. The CGM monitor was worn a mean of 4.262 days a week. None of the subjects reported significant hypoglycemia while wearing the monitor. CGM was effective in improving glycemic control in this population with poorly controlled diabetes.

Type 1 diabetes is the second most common chronic childhood illness, affecting an estimated 215,000 children under 20 years of age in the United States (Centers for Disease Control and Prevention, 2011). Type 1 diabetes is a manageable lifelong chronic disease, and if controlled well, the person can live a long and productive life. Unfortunately, in spite of aggressive diabetes management strategies, a large majority of children, adolescents, and young adults with diabetes are not adequately controlled (Silver-stein et al., 2005). Poorly controlled diabetes can lead to a number of chronic complications, which can ultimately lead to blindness, kidney failure, nerve damage, and heart disease (Sperling, 2002). Continuous glucose monitoring (CGM) has been shown to improve glycemic control in children, adolescents, and young adults when worn for 6 or more days per week (Beck et al., 2009a). This study evaluated the effectiveness of CGM in children, adolescents, and young adults with poorly controlled diabetes on improving glycemic control as evident by improvement in HbA1c and self-reported hypoglycemia. The study utilized a new sensor approved for ages 16 years and older that can be worn for up to six consecutive days.

Adolescents are in a time of transition from relying on parents for total care of their diabetes to complete self-care as an adult. There are a great number of factors that can interfere with or impair that transition, including some examples such as difficulty remembering or wanting to test the blood sugar, injecting insulin, and trying to fit in with their friends and not be different. If optimal care and adherence is not achieved, the patient is at a higher risk of developing both acute and chronic complications of diabetes (Schilling, Knafl, & Grey, 2006). Acute complications of diabetes can include diabetic ketoacidosis (DKA) and hypoglycemia. While DKA has a high rate of resolution, approximately 1 in 100 children with DKA will have a poor outcome, including death or permanent disability. Severe hypoglycemia can also have potentially grave outcomes. Chronic complications of poorly controlled diabetes can include retinopathy, neuropathy, nephropathy, heart disease, and hypertension. While long-term complications may take many years to develop, once present, they are irreversible, and all have permanent lifelong consequences (Sperling, 2002). The goal for glycemic control in the population is an HbA1c level of less than 8% (Silverstein et al., 2005).

BACKGROUND AND LITERATURE REVIEW

The literature review included all studies that evaluated children with diabetes and CGM use. Two systematic reviews and seven randomized control trials (RCTs) showed improvement in HbA1c levels with CGM use in children and adolescents. Two RCTs (Battelino et al., 2011; Beck et al., 2009b) showed decreases in hypoglycemia with CGM use while improving HbA1c levels. One of the studies (Tamborlane et al., 2008) showed no significant improvement in HbA1c levels. None of the studies evaluated the use of CGM in patients with HbA1c greater than 10%. These studies showed the benefits of CGM in children, adolescents, and adults who wore the device 6 days a week or more (Beck et al., 2009a).

One of the systematic reviews found a higher percentage of patients reaching a goal of HbA1c with an intervention of CGM and a secondary effect of less hypoglycemia (Wojciechowski, Ryś, Lipowska, Gaweska, & Małecki, 2011). The second systematic review found a lowering of HbA1c of 0.5–1.0% without an increased risk of hypoglycemia (Joubert & Reznik, 2012). Each of the systematic reviews included studies with both adults and children. One RCT evaluated factors affecting improved outcome of HbA1c and found that improved HbA1c was seen in patients who wore the sensor more than 70% of the time (Joubert & Reznik, 2012). One RCT showed benefit of CGM in adults who wore CGM but did not show the same results in children and adolescents because they wore the device less often (Tamborlane et al., 2008). Two of the RCTs evaluated the outcome of hypoglycemia in groups, both adults and children, who were at or close to HbA1c goal and showed less hypoglycemia (Battelino et al., 2011; Beck et al., 2009b). Two RCTs showed improvement in HbA1c when utilizing CGM with concurrent initiation of insulin pump therapy (Battelino et al., 2011; Raccah et al., 2009). One study showed improvement in HbA1c when the initiation of insulin pump therapy was combined with CGM (Nørgaard et al., 2013). One observational study showed improvement in HbA1c with the initiation of insulin pump therapy and CGM compared with standard insulin injection regiments (Bergenstal et al., 2010).

The literature review revealed the need for further research to evaluate the effectiveness of CGM on adherence to treatment and to improve HbA1c in children, adolescents, and young adults with diabetes, especially those with poor diabetes control. All but one of the studies utilizing CGM (Tamborlane et al., 2008) showed statistically significant improvement in HbA1c levels, but pediatric participants in this study did not wear the CGM on an almost daily basis. One study found improvement in HbA1c in this population when the CGM was worn for 6 or more days a week (Beck et al., 2009a).

The primary goal of the intervention was to lower HbA1c levels, but lowering HbA1c carries with it the risk of an increase of hypoglycemic events. The increased risk is associated with improved control and not the actual use of the CGM monitor. Because previous studies have shown less likelihood of hypoglycemia with the use of the CGM device, the potential risks of implementing CGM were considered to be minimal (Battelino et al., 2011; Beck et al., 2009b).

DESIGN/METHODS

This study used a prospective, one-group, pre-and posttest pre-experimental design with a convenience sample of patients (Polit & Beck, 2008). The participants served as their own controls. The study placed 40 participants on CGM for 3 months to compare the primary outcome of HbA1c at baseline and 3 months. Participants were asked to self-report significant hypoglycemia. The demographic data included age, gender, length of time since diagnosis of type 1 diabetes, family support system, family dynamics, and who was the primary caregiver.

Inclusion/exclusion criteria

Participants needed to be diagnosed with type 1 diabetes for 6 months or longer. They had to be 7–21 years old at the start of the study. Their HbA1c levels had to be 9% or higher. They had to agree to participate in all study visits. They needed to wear an investigational CGM sensor for 6 or more days a week. Potential participants who were unable to wear a continuous glucose sensor for the study period or were unable to attend monthly clinic study visits were excluded from study participation.

Sample

A convenience sample of 33 participants was recruited from the WVU Pediatric Endocrinology Clinic in Charleston, West Virginia. Participants were ages 9–20 years with type 1 diabetes and HbA1c 9% or greater. Twenty-one (63.6%) of the participants completed the study. Parents of patients under 18 years of age signed a consent form and patients an assent form. Patients over 18 years of age signed a consent form.

To increase the likelihood of participants returning for follow-up study visits, the study offered an incentive for each study visit completed. The incentive was $20 for the baseline visit and for each of the follow-up visits. These incentives were in the form of Walmart gift cards. This was to help offset the costs of travel to participate in the study.

Procedures/protocol

Participants who met inclusion criteria completed consent/assent forms, and a baseline HbA1c was obtained utilizing the Siemens DCA point of care HbA1c Analyzer. Participants were then instructed on the use of the Medtronic Guardian CGM device and Enlite sensor. The Guardian CGM monitor is only approved by the FDA for use in combination with the Medtronic Sof-sensor as an adjunct to regular blood sugar testing. The Enlite sensor is not approved by the FDA for use with the Guardian CGM monitor; therefore, the CGM system used during the study (Guardian CGM monitor and Enlite sensor) was investigational. The device did not pose a significant risk to the patient, was minimally invasive, and was used only as a trending device and not to make insulin adjustments. The patient needed to test his or her blood sugar at least twice a day to calibrate the system in order to get a reading on the sensor. The patients were instructed to test their blood sugars to determine the accurate dosing of insulin, and any time the CGM monitor identified that they had a low blood sugar (<70 mg/dL). Use of the device was initiated in the office. Follow-up visits occurred at 1 and 2 months after enrollment. The final study visit was at 3 months. At the final study visit, the investigators obtained HbA1c levels using the same method as used in the initiation visit, obtained self-reported insulin dosing, and final weight.

HbA1c

Baseline and final HbA1c levels were measured utilizing the Siemens DCA point of care HbA1c Analyzer. The quality controls were completed as per the manufacturer’s recommendation. Testing was completed by the staff who had completed standardized competencies as per the manufacturer’s testing procedures.

Hypoglycemia

Participants were asked to self-report clinically significant hypoglycemia. Clinically significant hypoglycemia was described as hypoglycemia that required assistance to treat by someone other than the participant. This was to be recorded by the participant in a paper log and was also reviewed at the study visits to identify episodes that were not recorded.

RESULTS

Demographics

The total number of participants recruited in the study who completed the baseline study visit was 33 (Table 1). The mean HbA1c of all participants at baseline was 10.52% (SD = 1.4127), with a range of 9–14%. The mean age of all participants was 15.42 years (SD = 2.750), with a range of 9–20 years. There was no difference in age (p = .691), duration of diabetes (p = .698), HbA1c (p = .655), or weight (p = .351) between the group that started the study and those that completed it, although the sample size is too small to be able to generalize the findings to the general population. A total of 21 participants completed the final study visit. Of the 21 that finished the study, 47.6% were male and 52.4% were female, 95.2% were Caucasian, and 4.8% were African American. The ages of the group that finished the study ranged from 11 to 20 years, with the mean being 15.57 years (SD = 2.657). Household demographics showed 57.1% of the subjects lived in a two-parent household, 14.3% with their mother only, 9.5% with their father, 4.8% with grandparents, and the remaining 9.5% either by themselves or with a significant other. Twelve subjects (36.4%) did not complete the study. The reasons for attrition were as follows: one was not able to complete the visits due to family illness, two reported skin irritation from the adhesive tape on the glucose sensor, and the other nine did not continue because they did not wish to continue wearing the monitor. Their reasons included participants did not like physically wearing the sensor, and they did not like the alarms from high blood glucose readings.

Table 1.

Demographic Characteristics of Children, Adolescents, and Young Adults With Poorly Controlled Type 1 Diabetes by Study Completion

	Did not complete	Completed
Number	12	21
Age (years)
Mean	15.4 ± 3	Mean 15.6 ± 2.7
Range	9–19 years	11–20 years
Duration of diabetes (years)
Mean	8.33 ± 4.2	7.74 ± 4.2
Range	3–17 years	1–17 years
Gender		Male:
Male	7 (58.3%)	10 (47.6%)
Female	5 (41.7%)	11 (52.4%)

Glycemic control

The mean baseline HbA1c of the participants completing the study was 10.60% (SD = 1.49), with a range of 9–14%. The mean HbA1c at the completion of the study was 9.49% (SD = 1.47), with a range of 7.3–12.3%. This was a statistically significant HbA1c improvement from 10.60 to 9.49% for a mean decrease of 1.1095% (SD = 1.9321) (p = .016). The change in HbA1c ranged from an improvement of 6.7% to worsening of 2.5%. The mean improvement in females was 0.9% and that in males was 2.18%; while there was more improvement in the male group, there was no statistically significant difference between the two groups (p = .0816).

Significance of improvement

Seventeen of the 21 (81%) participants demonstrated an improvement in HbA1c. The improvement in glycemic control ranged from 0.1 to 6.7%. Two of the participants had improvements of less than 0.5%. Fifteen of the participants had clinically significant improvements in HbA1c that were greater than 0.5%. Of these, 11 participants had an improvement between 0.5 and 2%, and the remaining four had an improvement in HbA1c of greater than 2%. While their glycemic control was improved, the mean weight in the group increased by 1.41 kg (SD = 2.7263) (p = .026). There was no difference in reported total daily dose of insulin given from baseline through each of the visits (p = .944).

Participants without improvement

Of the four participants who had a worsening HbA1c, two admitted to improper manipulation of the CGM device. One of the participants at the final visit admitted to entering false lower blood glucose values in the CGM to get the monitor to read lower blood glucose readings. The other participant admitted to not testing blood glucose values and entering fictitious readings into the monitor to keep the CGM reading. The HbA1c levels were recalculated after removing these two subjects, and a statistically significant reduction in mean HbA1c remained. After exclusion of these subjects, the mean baseline HbA1c was 10.668% (SD = 1.5532) and improved at the final visit to 9.211% (SD = 1.2490) (n = 19). The mean improvement in HbA1c in this group was therefore 1.4579% (SD = 1.6711) (p = .001).

Adherence

Participants wore the CGM monitor for a mean of 51.14 days (60.9%) (SD = 20.86), with a range of 20–81 days. The total number of days per week that participants wore the monitor was 4.26 days (SD = 1.74; range = 1.67–6.75 days). The mean number of total hours the CGM was worn by participants was 820.4 (SD = 437.58), with a range of 386–1,674 hr; the mean hours worn per day was 9.77. The mean glucose sensor reading for the study was 211.05 mg/dL (SD = 34.683). While the sample size was small, there was no correlation between the duration of sensor use and HbA1c improvement (p = .822). Previous studies found improvement with an increased wear time of the CGM device.

Hypoglycemia

Hypoglycemia was monitored by self-report at follow-up visits and by analysis of CGM data download. One participant reported having a clinically significant hypoglycemic episode that required assistance to correct. This participant had two episodes within 1 week; the episodes were described as “passing out.” Hypoglycemia was treated successfully with oral glucose, and the participant did not require hospital emergency treatment. When reviewing the episodes with the participant, we found that both episodes occurred when the participant was not wearing the continuous glucose monitor. None of the time spent in hypoglycemia that was identified by the CGM monitor required assistance to correct per reports for the participants.

DISCUSSION AND CONCLUSIONS

There was a clinically and statistically significant improvement of 1.1095% in HbA1c during the 12-week study. At the end of the study, two participants were identified as manipulating the CGM device and did not provide accurate finger stick blood glucoses to calibrate the monitor. The change in HbA1c was recalculated after removing these two outliers’ participants from the analysis, and there was a mean improvement of 1.4579% (SD = 1.6711) (p = .001). The studies that were identified from the literature search had HbA1c levels that ranged from no improvement to 1% improvement (Battelino et al., 2011; Joubert & Reznik, 2012; Raccah et al., 2009; Wojciechowski et al., 2011). For this study, an improvement of 0.5% was considered as clinically significant. Fifteen of the participants (71.4%) had an improvement of greater than 0.5% in HbA1c. Eleven of the participants had final HbA1c levels that improved to less than 9%. Two of those participants had final HbA1c levels in the 7% range. An HbA1c of less than 8.0% is considered to be on target for this age group (Silverstein et al., 2005). While not all participants had improvement in HbA1c, those who had clinically significant improvement in glycemic control ranged from 0.8 to 6.7%. While the mean improvement in HbA1c was greater than that in the other studies that were reviewed, the baseline HbA1c was higher in this study; therefore, there was more potential for improvement. With intensification of glycemic control, there is a risk of weight gain. The participants had a mean weight gain of 1.4134 kg (SD = 2.7263) (p = .026); while the weight gain is statistically significant, it was not considered to be a clinically significant in the short term. However, in future studies, weight should be monitored over a longer period of time.

Significance

This study offers evidence that CGM improves glycemic control in children, adolescents, and young adults with poorly controlled diabetes with HbA1c more than 9.0%. This was a short-term, 12-week intervention, so it is not known if there would be further improvement or regression of glycemic control with long-term continual use of CGM. The attrition rate for the study was high, and so future interventions should include focusing on improving retention rates in this difficult population. This intervention required daily activities beyond what the participants adhered to, and it may be more work than they are willing to do. Because it is possible for participants to manipulate the CGM monitor, close attention needs to be paid to participants whose families do not meet on a regular basis to monitor their child’s use of the CGM monitor. Previous studies showed improvement in glycemic control when the CGM monitor was worn 6 or more days a week. This study showed improvement in HbA1c when the mean number of days of CGM was about 4.2 days per week or a mean of 9.77 hr per day. Although there was improvement in HbA1c in this study, it did not correlate with the duration of CGM use. While there was one participant who reported significant hypoglycemia during the study, this happened when the participant was not wearing the CGM sensor. There was no clinically significant hypoglycemia reported when the participants were wearing the monitor.

Conclusions

CGM use was effective in improving glycemic control with children, adolescents, and young adults with poorly controlled diabetes. There was a small but significant increase in weight during the intervention. There was no significant change in total daily dose of insulin reported during the intervention. There was no clinically significant hypoglycemia reported during the CGM monitor use.

How might this information affect nursing practice?

Nurses provide direct and indirect care for children, adolescents, and young adults with type 1 diabetes in a number of areas including diabetes clinics, schools, and hospitals. The findings from this study provide new and encouraging evidence that CGM helps improve the glycemic control in this vulnerable group. Classically, these patients were labeled as noncompliant, and, hence, may not be considered good candidates for CGM. For patients willing to wear CGM daily, poor glycemic control should not be a limiting factor in use of CGM, but a reason to consider use. Nurses work with these patients and families on such a frequent basis that they need to be able to identify those at high risk and recommend CGM as another tool to improve glycemic control for those willing to wear the sensor.