cgm case studies

52
Current Trends in Professional Continuous Glucose Monitoring Guest Editor: Bruce W. Bode, M.D Atlanta Diabetes Associates Atlanta, Georgia Proceedings of ‘‘Current Trends in Professional Continuous Glucose Monitoring (CGM),’’ Keck School of Medicine, University of Southern California, Los Angeles, California, Sponsored by Medtronic.

Upload: medtronicdiab

Post on 07-May-2015

661 views

Category:

Health & Medicine


7 download

DESCRIPTION

CGM

TRANSCRIPT

Page 1: Cgm case studies

Current Trends in ProfessionalContinuous Glucose Monitoring

Guest Editor:

Bruce W. Bode, M.D

Atlanta Diabetes AssociatesAtlanta, Georgia

Proceedings of ‘‘Current Trends in Professional Continuous Glucose Monitoring (CGM),’’ Keck School of Medicine, University of Southern

California, Los Angeles, California, Sponsored by Medtronic.

Page 2: Cgm case studies

Current Trends in Professional Continuous Glucose

Monitoring has been compiled and produced from

funding solely provided byMedtronicMiniMed, Inc.

and Lifescan, Inc. Medtronic MiniMed, Inc.

or its affiliates provided support for some of the

studies reported in the other articles included in

this special supplement.

Page 3: Cgm case studies

Current Trends in Professional Continuous

Glucose Monitoring

contents

4 IntroductionB.W. Bode, P. Phillips, B. Nardacci, K.C. Arnold, B.S. Horowitz, O. Odugbesan, S. Reddy

Case 1: A 62-Year-Old Woman with Complex Medical History and Hypoglycemia Unawareness6P. Phillips

Case 2: A 60-Year-Old Woman with Diabetes Secondary to Necrotizing Pancreatitis withHighly Variable Blood Glucose Levels on Basal Bolus Therapy

9

E. Nardacci

Case 3: A 41-Year-Old Man with Type 1 Diabetes with Good Glucose Control12K.C. Arnold

Case 4: A 29-Year-Old Woman with Type 1 Diabetes, Pregnant with Triplets14K.C. Arnold

Case 5: A 61-Year-Old Man with Type 1 Diabetes16S. Reddy

Case 6: A Type 1 Diabetic College Student with a Crazy Lifestyle and Crazy Blood Sugars18B.S. Horowitz

20 Case 7: A 38-Year-Old Woman with Type 1 Diabetes

B.W. Bode

23 Case 8: A 69-Year-Old Woman with Type 2 Diabetes and Good Premeal Glucose, But a High A1c

B.S. Horowitz

Case 9: A 37-Year-Old Woman with Type 2 Diabetes at 20 Weeks of Gestation26K.C. Arnold

Case 10: A 41-Year-Old Woman with Type 2 Diabetes, High A1c28O. Odugbesan

Discussion Regarding Use of Professional Continuous Glucose Monitoring31

Professional Continuous Glucose Monitoring (CGM) Progress Note37

HIGHLIGHTS FROM DIABETES TECHNOLOGY & THERAPEUTICS

Continuous Glucose Monitoring in Non-Insulin-Using Individuals with Type 2 Diabetes: Acceptability,Feasibility, and Teaching Opportunities

41

N.A. Allen, J.A. Fain, B. Braun, S.R. Chipkin

Sustained Efficacy of Continuous Subcutaneous Insulin Infusion in Type 1 Diabetes Subjects withRecurrent Non-Severe and Severe Hypoglycemia and Hypoglycemia Unawareness: A Pilot Study

49

M. Gimenez, M. Lara, and I. Conget

Maximizing Reimbursement through Correct Coding Initiatives54E. Orzeck

Page 4: Cgm case studies

Introduction

Bruce W. Bode, M.D.,1 Paula Phillips, M.D.,2

Elizabeth Nardacci, FNP-C,3 Kathleen C. Arnold, A.N.P.,4

Barry S. Horowitz, M.D.,5 Ola Odugbesan, M.D.,6

Sushma Reddy, M.D.7

1Atlanta Diabetes Associates, Atlanta, Georgia.2Diabetes & Metabolism Specialists, San Antonio,Texas.

3Albany Medical Center, Albany, New York.4The Diabetes Center, PLLC, Ocean Springs, Mississippi.5Palm Beach Diabetes and Endocrine Specialists, PA, West PalmBeach, Florida.

6North Atlanta Endocrinology, Lawrenceville, Georgia.7Endocrinology & Diabetes Center, Fort Gratiot,Michigan.

AbstractA total of 10 patients were discussed during this meeting en-

titled ‘‘Current Trends in Professional Continuous Glucose Mon-

itoring (CGM).’’ Seven patients had type 1 diabetes and faced

challenges with day-to-day glucose control, five with poor glucose

and two with normal A1c. Two patients switched to insulin pump

treatment after reviewing their progessional CGM data. Another

subject with type 2 diabetes on oral agents switched to CSII and

was able to achieve target A1c and deliver a healthy, normal-sized

baby. Two other patients had type 2 diabetes: one on MDI using

large amounts of insulin, and another on oral agents only. After

case presentations, the role of professional CGM in the real world

was discussed.

During the ‘‘Current Trends in Professional Continuous

Glucose Monitoring (CGM)’’ symposium, held at the Uni-

versity of Southern California on November 20, 2009, seven

distinguished experts in the field of CGM from around

the country presented and discussed the use of professional CGM in

10 patients. After the case presentations, all seven faculty members

discussed which patients are appropriate candidates for profes-

sional CGM, and how physicians and nurse practitioners do pro-

fessional CGM in the real world, using professional CGM outputs to

make appropriate therapy adjustments, and scheduling follow-up

evaluations.

Glucose monitoring has evolved from urine testing to self-

monitored blood glucose (SMBG) to CGM. The first continuous

glucose sensor was introduced in 1999 (Fig. 1). This early device

incorporated a sensor that measures interstitial glucose, continu-

ously giving an average glucose reading of the interstitial fluid

every 5 min up to 188 tests per day over a 3-day period. The initial

data showed that the device significantly lowered HbA1c in people

with elevated A1c, and reduced severe hypoglycemia in patients

with normal A1c.

There are four commercially available CGMs, two from Med-

tronic MiniMed (one professional and retrospective, and the other

real-time), Freestyle Navigator and Dexcom SEVEN Plus (Fig. 2).

Professional CGM is a blinded evaluation by the healthcare pro-

vider, who places a sensor with a transmitter on the patient in the

Fig. 2. The two types of continuous glucose monitoring target dif-ferent users: health care professionals and empowering individualpatients with diabetes.

Evolution of Diabetes Management Technologies

Insulin pump therapyContinuos

glucose sensor

Urine glucose testingPoint-in-timeBG meters

Discovery of insulin

Integrated systems:Pumps/Meters/Software

ArtificialPancreas

1900s 1922 1977 1978 1999 2003

Fig. 1. Evolution of diabetes management technologies. The firstcontinuous glucose sensor was introduced in 1999. The first devicethat integrated insulin pumps, glucose sensors, and softwarewas introduced in 2003.

4 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 5: Cgm case studies

office. The provider instructs the patient to continue usual care and

keep a log of SMBG readings; time and dosage of whichever insulin

or oral agents being taken; meal times; and food intake, activity,

and symptoms of high and low blood sugar. Patients are blinded to

the glucose values, and return to the office after 3 days to download

the data. After reviewing the data from the professional CGM

device, the healthcare provider recommends necessary therapeutic

changes.

In contrast, personal CGM (also known as real-time CGM) devices

are worn by the patient for longer periods and enable the patient to

see real-time glucose values throughout the day, and make changes

on their insulin dose or adjust food intake to avoid extreme hypo-

and hyperglycemia.

Address correspondence to:

Bruce W. Bode, M.D.

Atlanta Diabetes Associates

77 Collier Road, Suite 2080

Atlanta, GA 30309

E-mail: [email protected]

INTRODUCTION

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 5

Page 6: Cgm case studies

Case 1A 62-Year-Old Woman with Complex Medical Historyand Hypoglycemia Unawareness

Paula Phillips, M.D.

Diabetes & Metabolism Specialists, San Antonio, Texas.

Patient HistoryThe first case is an obese 62-year-old woman with type 1 diabetes

(diagnosed at the age of 30 years) and hypoglycemia unawareness. On

initial presentation in March 2008, the patient had inadequate glucose

control with an A1c of 9.6%. She was on NPH 10 units in the morning,

Glargine 15 units at bedtime, and a sliding scale of rapid-acting insulin

analog before meals. She had wide glycemic excursions on four-times-

a-day self-monitored blood glucose (SMBG) ranging from 37 to

542mg=dL. Most of the hypoglycemic episodes (<50mg=dL) were in

the morning, and thus she never used the sliding scale at breakfast. By

lunchtime, her blood glucose (BG) ranged between 300 and

350mg=dL.

Comorbidities and DiabeticComplications

The patient’s history is significant for anoxic brain injury that

occurred in June 2006 while wearing an insulin pump. Apparently,

she had very little training on the insulin pump, did not change it

every 3 days, and experienced infections at the site as a result. After

the hypoglycemic episode, she had to be in a nursing home for 4 years

before going home to her family.

Additional comorbidities include diabetic neuropathy and gastro-

paresis. She was on metoclopramide treatment but for her gastroin-

testinal complaints. She functions relatively well considering her

brain injury, with a minor tic, some speech difficulties, and poor

memory. She is very meticulous recording BG and everything she

eats. Her surgical history is significant for cholecystectomy and

hysterectomy.

Rationale for Initiating Professional CGMAfter initial presentation, the provider changed the patient’s reg-

imen to basal bolus insulin with Glargine at bedtime and premeal

insulin aspart with a supplemental scale. Subsequently, this was

adjusted to split basal dosing, but she continued to have wide gly-

cemic excursions. Before her initial professional CGM study in

May 2009, her baseline A1c had improved by 1% to 8.6%. Her

treatment at that time had evolved to Detemir twice a day at 8:00 am

and 6:00 pm with a higher dose at bedtime if her BG level exceeded

250 mg=dL. She took insulin Aspart *6 units before meals on a

sliding scale.

On the basis of SMBG, she seemed to have relatively good BG

before lunch and dinner, but her morning BGs were very high. If her

bedtime BG was very elevated ([250 mg=dL), she tended to drop and

there was a big difference between her bedtime BG and the BG value

the next morning. If her BG was reasonably controlled at bedtime

(<150 mg=dL), then she would have very high fasting blood sugars

(400–500 mg=dL). Thus, the rationale for the initial professional CGM

was to try to distinguish if this pattern was due to snacking at bedtime

due to fear of hypoglycemia, or due to counter regulatory hormones

in response to nocturnal hypoglycemia.

Initial Professional CGM ResultsThe initial professional CGM revealed no evidence of over-

night hypoglycemia. Her CGM glucose averaged 193 mg=dL, and

she is in the hyperglycemic range most of day as shown by the large

red segments in the pie chart of the Sensor Summary (Fig. 3). The

charts also had some blue segments, indicating minimal periods of

hypoglycemia. The Sensor Summary highlights a poor glucose

control.

The Sensor Modal Day tracings were highly variable, with an in-

consistent pattern in the morning (including two brief periods of

hypoglycemia), and significant postmeal excursions both at break-

fast and dinner, with a drop in BG late afternoon (Fig. 4).

Therapy Adjustments=Treatment AlterationsThe provider adjusted the basal insulin to Detemir BID 7 units in

the morning and 10 units at night (11 units if BG[ 220 mg=dL at

night) due to high morning BG. Aspart was maintained at the same

dose. The patient saw the certified diabetes educator and nurse

practitioner every 1 to 3 weeks. Most of the patient’s meals and snacks

were low on protein, so she was instructed to have protein with each

meal and snack, but to limit snacks overall.

Follow-Up Professional CGM Results=Response to Therapy Adjustments

On follow-up professional CGM, the BG pattern showed fewer

excursions after breakfast and lunch (Fig. 5). The patient continued to

have hyperglycemia after dinner that continued for up to 6 h, par-

ticularly after meals with high-fat content. On two out of the three

evenings, her BG monitor recorded over 449 mg=dL.

Examination of professional CGM outputs and the patient’s hand-

written diary revealed that the patient was recording her BG levels in

her diary inaccurately. For example, if the SMBG meter reading was

95, the patient recorded it as 195, probably due to her brain injury.

6 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 7: Cgm case studies

Sensor Summary

Date

Sensor

# of Sensor Values

7/20/2009 7/21/2009 7/22/2009 7/23/2009

159

293

181-381

63

2

283

196-369

2

5.5

n/a 0.97

0

0

0

288

169

50-353

74

4

149

56-235

4

288

205

83-346

58

4

157

177-192

4

17.7

n/a

3

3

0

20:25 (85%)

03:35 (15%)

00:00 (0%)

6.1

0.99

7

6

1

14:30 (60%) 13:15 (100%)

00:00 (0%)

00:00 (0%) 03:20 (5%)

12:55 (19%)

52:40 (76%)

2

10

12

12.1

12

56-369

181

12

83

43-381

208

827

Totals

153

0 1

76

07:15 (31%)

02:15 (9%)

68

0

43

1

92

192

43-306

99

2

191

66-315

2

19.3

n/a

2

1

1

04:30 (59%)

02:05 (27%)

01:05 (14%)

75

3

Average (mg/dL)

Min - Max (mg/dL)

STDev (mg/dL)

# of Meter Values

Average (mg/dL)

Min - Max (mg/dL)

Designation

# of Paired Readings

Mean Abs. Diff. [MAD %]

Correlation Coeff. [R]

# of Excursions*

# of High Excursions*

# of Low Excursions*

Duration Above High Limit

Duration Within Limits

Duration Below Low Limit

Pie ChartRed: Above LimitsGreen: Within LimitsBlue: Below Limits

Glucose Area Above HighLimit (mg/dL*Day)

Glucose Area Below LowLimit (mg/dL*Day)

Meter

Optimal AccuracyCriteria

ExcursionsHigh > 140mg/dLLow < 70mg/dL

X: Use ClinicalJudgment

X: Use ClinicalJudgment

X: Please use your clinical judgment - this day does not satisfy the optimal accuracy criteria according to set thresholds:N[¼ 3, R[¼0.79 and MAD<¼28% [or<¼ 18% if the range (Min-Max) of meter values is less than 100mg=dL (5.6mmol=L) - seeCriteria Note below].

C: This day does not have any paired sensor=meter data and no sensor plot is provided. As a result, ‘Meter Only’ data is available.S: Please use your clinical judgment - this day does not have any meter data. As a result, ‘Sensor Only’ data is available.Criteria Note: If the range (Min-Max) of Meter Values is less than 100mg=dL (5.6mmol=L) then ‘R’ will be reported as ‘N=A’.

In this case the optimal accuracy threshold set for MAD is <¼18%.Excursion Note: Excursions are counted in the day that the excursion event started.

Fig. 3. The red segments of the Sensor Summary pie chart show that the patient is hyperglycemic most of day.

Fig. 4. The Sensor Modal Day re-port show two brief periods of hy-poglycemia (shown in blue andgreen), and significant postmealexcursions.

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 7

Page 8: Cgm case studies

Given the patient’s generally poor dietary habits, she would have

benefited from a correction factor and a carbohydrate ratio instead of

sliding scale insulin. However, given her cognitive deficits, she would

likely be unable to master the calculations involved.

ConclusionsThe follow-up professional CGM tracings revealed the need to

adjust insulin or diet (or both) at dinner and decrease the amount of

basal insulin at night. Because of her hypoglycemic unawareness and

marked excursions in her glycemic control, the patient would benefit

from a personal CGM with an audible warning. She would also

benefit from insulin-pump therapy. Theoretically, given her gastro-

paresis and the way she responds to high-fat meals, the endocrinol-

ogist could prescribe square wave bolus or dual wave bolus, and

variable basal rates starting at 3:00 am or 4:00 am onward, but it would

be clinically challenging to pursue this option.

Disclosure StatementPaula Phillips, M.D., is a speaker for Medtronic Diabetes.

Fig. 5. The Sensor Modal Time Report shows that the patient experienced hyperglycemia after dinner that continued for up to 6 h.

PHILLIPS

8 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 9: Cgm case studies

Case 2A 60-Year-Old Woman with Diabetes Secondary to Necrotizing Pancreatitis withHighly Variable Blood Glucose Levels on Basal Bolus Therapy

Elizabeth Nardacci, FNP-C

Albany Medical Center, Albany, New York.

Patient HistoryThis patient developed necrotizing pancreatitis and was found to

have diabetes in January 2008, about a year before the initial pro-

fessional continuous glucose monitoring (CGM) evaluation. Despite

multiple visits with both the physician and the diabetes educators, the

patient had continued difficulty in dosing her insulin correctly. Her

baseline regimen was Aspart*4 to 6 units with meals. She would not

adjust her insulin dose more than 2 units at a time. She was also on

Glargine 14 units at bedtime, which occurred at very varied times.

Her initial A1c was 10.2%.

Rationale for Initiating Professional CGMThe patient was inexperienced with insulin dosing and really

wanted to do a better job, as she had wildly fluctuating blood glucose

(BGs). She felt hypoglycemic everyday, complained of feeling poorly,

and was having difficulty functioning, especially in caring for her

12-year-old grandson of whom she had custody. She checked her BG

up to six times daily because of concern about her symptoms. She

never missed appointments with our diabetes educators, whom she

saw frequently. The diabetes educators initiated the professional CGM.

Initial Professional CGM ResultsThe average sensor value was 269 mg=dL, which corresponded

with A1c of 10.2%. Only 13% of her glucose readings were within the

target range (WTR 70–150 mg=dL), 2% below the target, and 85% of

readings were above the target (>150 mg=dL).

The Sensor Modal Day revealed no consistent BG pattern. BG

ranging from 200 to above 400 mg=dL. The provider examined the

diary with the patient, comparing it to the professional CGM tracings.

According to her diary, the patient was taking her glargine at varying

times, but admitted missing her injections. (Fig. 6).

The blue plus signs on the Sensor Modal Details indicate when the

patient does SMBG. The tracing indicated wide glucose excursions

that were not detected by SMBG (Fig. 7).

Therapy Adjustments=Treatment AlterationsAfter using the Professional CGM, the provider recommended an

insulin pump, and the patient agreed.

Follow-Up Professional CGM Results=Responseto Therapy Adjustments

Four months after initiating insulin pump therapy, the patient had

better BG control because of the continuous infusion of the basal

insulin. The patient said things were better generally, but she still felt

like she was experiencing hypoglycemia at lunchtime.

Fig. 6. The SensorModal Day outputshows severehyperglycemia.

ª 2010 Medtronic Minimed, Inc. All rights reserved. CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 9

Page 10: Cgm case studies

Fig. 7. The Sensor Daily Details highlight peak blood glucose readings that were missed on self-monitoring of blood glucose.

Fig. 8. Sensor Modal Day indicates less blood glucose variability overall compared to the earlier evaluation.

NARDACCI

10 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 11: Cgm case studies

On follow-up professional CGM, the Sensor Modal Day was much

improved (Fig. 8). The tracings confirmed the patient’s perceptions:

she was overcorrecting and experiencing hypoglycemia in the af-

ternoon. The mean sensor value dropped from 269 to 179 mg=dL, and

the standard deviation went from 101 to 79, indicating reduced

glucose excursions. The A1c declined from 10.2% initial, to 9.9% in

January 2009, and to 7.8% in July 2008.

The patient was able to maintain and improve control, experi-

encing brief periods of hyperglycemia during the day. Her post-

prandial BGs remained elevated. She was carbohydrate counting

well, according to the dietician, so the provider altered the pump

settings to increase the patient’s basal rates slightly at night, as-

suming that the patient probably experienced some hypoglycemia

unawareness. The insulin of carbohydrate ratio was also adjusted.

The patient was doing what the bolus wizard told her to do, but

since the setting was incorrect she was overcorrecting. The clini-

cian adjusted the sensitivity number upward to give her less of a

correction.

ConclusionsProfessional CGM identified multiple reasons for uncontrolled

diabetes in this patient. Although the patient is not at goal, she has

improved significantly.

This case illustrates three major issues that illustrate the use of

professional CGM to generate an ‘‘aha! moment.’’ (1) The process of

going through the professional CGM output and looking at the diary

often elucidates previously hidden behavior that plays a key role in

glycemic control. This patient would not have admitted missing

glargine doses without seeing the wide glucose excursions. (2) The

patient had been taking her evening dose of insulin at varying times.

The professional CGM tracings provided adequate information to accept

an insulin pump, because the patient could maintain BG control despite

bed times that could differ from night to night by 6–8h. (3) Although

ordinarily a very anxious patient might not be an ideal pump therapy

candidate, this patient was highly motivated and able to do a good job

with a lot of education and support. Using the bolus wizard was very

critical in helping this patient become confident in making dosing

decisions, compared to her earlier hesitation in adjusting her injections

by only 2 units each time. These changes allowed to patient to lower and

maintain her A1c values and confirmed that she was making better

decisions.

Disclosure StatementElizabeth Nardacci, F.N.P., BC-ADM, is a speaker for Eli Lilly and

Medtronic Diabetes. She is on the medical advisory board of Med-

tronic Diabetes.

CASE 2: WOMAN WITH DIABETES SECONDARY TO NECROTIZING PANCREATITIS

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 11

Page 12: Cgm case studies

Case 3A 41-Year-Old Man with Type 1 Diabetes with Good Glucose Control

Kathleen C. Arnold, A.N.P.

The Diabetes Center, PLLC, Ocean Springs, Mississippi.

Patient HistoryThis patient is a 41-year-old man found to have type 1 diabetes in

February 2008. He had no other significant medical history or dia-

betes complications, and his A1c was 5.9%. He was managed with

Glargine, 15 units at bedtime, and Aspart with a correction formula

and a carbohydrate ratio, and he tested his blood glucose (BG) four to

five times daily. He was an avid bicyclist who took 1–4-h bike rides

3–5 days a week.

Rationale for Initiating Professional ContinuousGlucose Monitoring

Both the patient and the healthcare provider wanted profes-

sional continuous glucose monitoring (CGM) to evaluate the BG

control on cycling days.

Fig. 9. Sensor Modal Time tracing for an avid cyclist. Red arrowsindicate unrecognized hyperglycemia and hypoglycemia.

Fig. 10. Sensor Daily Details reveal most blood glucose values are within range, even on days that the patient takes long bike rides.

12 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 13: Cgm case studies

Initial Professional CGM ResultsThe CGM tracings showed that the patient had overall good gly-

cemic control, with the exception of hyperglycemia when he was

prepping for long rides. For long rides he ‘‘carbo loaded’’ to keep his BG

up when he is riding with a product called ‘‘goo,’’ which contains 80–

100g of instant-acting carbohydrate. The Sensor Modal Day revealed

hypoglycemic and hyperglycemic excursions, of which the patient was

unaware.

The patient tried to ride almost every day during his professional

CGM evaluation, and he had no problems with sensor adherence while

riding. The Sensor Daily Details reveal which days the patient was

cycling by BG peaks (Fig. 9). He did have some unrecognized hyper-

glycemia during the night, and several episodes of unrecognized hy-

poglycemia.

Therapy Adjustments=TreatmentAlterations

The only therapy adjustment for this patient was to change the

way he carbohydrate loads before a ride. He now eats a longer-acting

carbohydrate bar that includes some protein. No change in his insulin

regimen was necessary.

The patient is considering an insulin pump. He practiced with a

smart pump with tubes, which was too bulky for his body type did not

adhere well to his skin during bike rides. He has applied for an insulin

pump with disposable components instead.

Follow-Up Professional CGMFollow-up professional CGM revealed that most of his blood sugars

are within the range, with fewer episodes of hypoglycemia (Fig. 10). He

continues to ride but will take a break during the cold weather.

ConclusionsThis case illustrates that even though the patient had good glucose

control, this apparent level of control masked swings in BG subse-

quent to carbohydrate loading with the simple carbohydrate gel.

Once he changed his preride carbohydrate source with protein, he

was able to normalize his BG, not only during the day, but also

overnight. Professional CGM also helped this patient to take the next

step for better glycemic management by highlighting the benefits

that an insulin pump would provide.

Disclosure StatementKathleen C. Arnold, A.N.P., BC-ADM, is a speaker for Lilly, Med-

tronic Diabetes, NovoNordisk, and Sanofi Aventis.

CASE 3: MAN WITH DIABETES AND GOOD GLUCOSE CONTROL

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 13

Page 14: Cgm case studies

Case 4A 29-Year-Old Woman with Type 1 Diabetes, Pregnant with Triplets

Kathleen C. Arnold, A.N.P.

The Diabetes Center, PLLC, Ocean Springs, Mississippi.

Patient HistoryA 29-year-old patient with previously diagnosed type 1 diabetes

presented with a nonplanned pregnancy with triplets. Her A1c at 14

weeks of gestation was 7.3%.

Rationale for Initiating ProfessionalContinuous Glucose Monitoring

The provider usually manages pregnant women with type 1 diabetes

using an insulin pump because of the rapidly changing needs for

insulin during pregnancy. The patient had a hypoglycemic seizure due

to insulin stacking and was transported to the hospital by ambulance.

Her baseline treatment was a Medtronic Paradigm 722 pump, and she

tested her blood glucose (BG) six to eight times daily.

Initial Professional Continuous GlucoseMonitoring Results

Professional continuous glucose monitoring (CGM) revealed

multiple hyperglycemic and hypoglycemic episodes, of which the

patient was unaware of, because she did performed self-monitored

blood glucose (SMBG) at times other than when the peaks or lows

occurred (Fig. 11). Although the CGM device has 70 mg=dL as the de-

fault, the value should be changed to 60 for pregnancy. For preg-

nancy, target BG 2 hours postprandial is 120=mg=dL, and the patient

did not always achieve that.

Therapy AdjustmentsShe initiated her pump very early in the pregnancy, and her pump

downloads indicated hypoglycemia regularly. Premeal targets in

pregnancy are 60 to 80 mg=dL, and she was having readings in the

50 mg=dL range.

After looking at the professional CGM output and at the patient’s

A1c, the provider recommended personal CGM. The patient’s A1c

started out at 9.9% before pregnancy and then declined, first to 7.3%

and then to 5.6%, once on insulin pump therapy.

Response to Therapy AdjustmentsIn early September, the patient had some complications. She un-

derwent an intrauterine laser surgery to separate blood vessels of the

two identical twins within this triplet pregnancy that were sharing a

blood source; one of the fetuses was not growing. She subsequently

developed preeclampsia. She delivered via C-section at 22 weeks’

gestation. One male infant weighed 1 pound 4 ounces, and another

male 1 pound 3 ounces; the third infant died within 18 hours of birth.

Both live infants were placed on ventilators. The patient developed

postoperative pneumonia and Escherichia coli infection. While being

treated for the infections, she was taken off her pump and given

insulin infusions for a couple of days and then returned to the insulin

pump.

ConclusionsProfessional CGM helped this pregnant woman with type 1 dia-

betes, to realize that she was experiencing hypoglycemic excursions

and hyperglycemia on a daily basis, and drove the decision to pre-

Fig. 11. The Sensor Daily Details identify exactly when multiple hyperglycemic and hypoglycemic episodes occur.

14 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 15: Cgm case studies

scribe a personal sensor. The CGM technology enabled day by day

adjustments that resulted in a shift from an A1c of 9.9% to 5.6%.

As the pregnant abdomen expands, placement of sensors has to

change. Sometimes the patients may move the sensor from the ab-

dominal area to the legs or buttocks.

This patient was compliant with her regimen after the hypogly-

cemic episode. The patient used her Bolus Wizard 100% of the time,

and performed SMBG four to five times daily.

Often times with pregnancy in women with type 1 diabetes, espe-

cially with triplets, you may not have the best outcome. Her preterm

labor was likely related either to the intervention they did at the

hospital or to her extremely poor control before conception, at which

time the patient was followed by a different provider. The two sur-

viving babies are doing well.

Disclosure StatementKathleen C. Arnold, A.N.P., BC-ADM, is a speaker for Lilly, Med-

tronic Diabetes, NovoNordisk, and Sanofi Aventis.

CASE 4: WOMAN WITH DIABETES, PREGNANT WITH TRIPLETS

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 15

Page 16: Cgm case studies

Case 5A 61-Year-Old Man with Type 1 Diabetes

Sushma Reddy, M.D.

Endocrinology & Diabetes Center, Fort Gratiot, Michigan.

Patient HistoryThis case is a 61-year-old man with type 1 diabetes who origi-

nally presented with type 2 diabetes in 1989, and was initially

treated with oral agents. In 1995, he was started on insulin. In 2001,

a C-peptide test confirmed that the patient had type 1 diabetes. He

was placed on basal-bolus therapy, and switched to an insulin

pump in 2004. His job involves physical labor and he had low

insulin requirements at breakfast and lunch. His major meal is

dinner and he is sedentary afterward. His insulin-to-carbohydrate

ratio at dinner was 1 unit for 8 grams of carbohydrate, in contrast to

1:12 at breakfast and lunch.

Comorbidities and Diabetic ComplicationsThe patient had hypertension and hyperlipidemia, and was

euthyroid secondary to hyperthyroidism 20 years before. He

had diabetic retinopathy and peripheral neuropathy. Surgical

history included angioplasty for coronary artery disease 1.5 years

before.

Rationale for Initiating Professional ContinuousGlucose Monitoring

He performed self-monitored blood glucose (SMBG) before dinner,

2 hours afterward, bedtime, and once in the middle of the night, at

which time he had hyperglycemia. He had a higher basal rate of 2 to

2.1 units between 8 pm and 8 am. However, in spite of that, his A1c

remained elevated at 9%.

Initial Professional Continuous GlucoseMonitoring Results

Professional continuous glucose monitoring CGM revealed post-

prandial hyperglycemia, and nocturnal hyperglycemia between 8 pm

and 2 am, indicating that the patient sometimes forgot to bolus af-

ter a night-time snack. The sensor summary indicated that the pa-

tient’s glucose level was within normal range only 24% of his time

with a mean sensor glucose of 206 mg=dL (standard deviation,

81 mg=dL).

The patient recognized how serious the problem was when the

physician discussed each day’s tracings and discovered that on the

day the patient had no bedtime snack, his BG levels were fairly well

controlled (Fig. 12). Another tracing revealed that the patient was

taking his insulin after breakfast, and another that he was not bo-

lusing appropriately at dinner time. Although he performed SMBG

four times a day, he still experienced unrecognized hyperglycemic

peaks that were missed with SMBG.

Therapy Adjustments=Treatment AlterationsThe Sensor Modal Day report revealed a dawn phenomenon (Fig.

13), which was corrected by a higher basal rate. The provider also

adjusted the pump sensitivity and the insulin-to-carbohydrate ratio

for the patient’s bedtime snack. He also reinforced the message that

‘‘whenever your hand goes to your mouth, it needs to go to your

pump.’’ The patient also upgraded to a newer (model 722) insulin

pump.

Follow-Up Professional CGM Results=Responseto Therapy Adjustments

The patient’s A1c improved to 8.3%. The patient now uses a per-

sonal real-time CGM to help keep his BG under control. Remembering

to bolus before his bedtime snack remains an ongoing challenge, and

he experiences nocturnal hyperglycemia between 9 pm and 1 am

about 2 days per week.

ConclusionsThis case shows the typical evolution of type 1 diabetes in

adulthood. He had type 1 diabetes from onset, which unfortunately

went unrecognized. The professional CGM clearly demonstrated to

the patient just how brittle his diabetes was and inspired him to

improve his behavior, at least temporarily. Professional CGM is

particularly helpful in managing patients in whom the A1c remains

elevated despite multiple interventions. The professional CGM

tracings highlights hyperglycemia based on which provider can

recommend treatment changes to reduce hyperglycemia. In addi-

tion, professional CGM is particularly motivating for those patients

who bolus after a meal despite being trained to bolus beforehand,

because they are not sure what they are going to eat beforehand or

they forget.

A recent study sponsored by the Juvenile Diabetes Research

Foundation showed an improvement in glucose control with real-

time CGM use. There is a learning curve for the physician and patient

to set a personal CGM device, so that the patient is not woken up too

often that they become noncompliant. That is what happened with

this patient, who routinely forgot to bolus with a bed time snack.

Unfortunately, at a subsequent visit, this patient’s A1c reverted to

8.2% (lowest A1c was 7.2%), and the patient revealed that he had

stopped using his personal CGM device. The patient immediately

recognized that when he used his personal CGM regularly and re-

membered to bolus appropriately, he would not be awakened by

alarms.

16 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 17: Cgm case studies

Disclosure StatementSushma Reddy, M.D., is a speaker for Eli Lilly, Medtronic Dia-

betes, Novo Nordisk, Sanofi-Aventis, Takeda, and Bristol Meyers

Squibb.

Fig. 12. Sensor Daily Details reveal that blood glucose is above 300mg=dL throughout the day and at bedtime, and nearly normalon waking. Blood glucose peaks were not recognized with SMBG (shown as blue plus signs).

Fig. 13. Sensor Modal Day reveals adawn phenomenon between 3 amand 6 am (shown in black).

CASE 5: MAN WITH TYPE 1 DIABETES

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 17

Page 18: Cgm case studies

Case 6A Type 1 Diabetic College Student with a Crazy Lifestyleand Crazy Blood Sugars

Barry S. Horowitz, M.D.

Palm Beach Diabetes and Endocrine Specialists, PA, West PalmBeach, Florida.

Patient HistoryThis is the case of a college student with a crazy lifestyle with wide

glucose excursions. This 18-year-old woman was found to have type

1 diabetes 3 years ago. She attends a community college and admits

to a stressful life and poor dietary habits. Her baseline regimen was

Aspart before meals.

She was counting carbohydrates with a ratio of 0.5:3 at breakfast,

0.5:5 at lunch, and 12:9 at dinner. She was using a correction factor

of 50 and was taking an inadequate glargine dose at bedtime. She

checked her blood glucose (BG) inconsistently because of her

schedule (2–6 T=day). Her BG ranged from 50 to 300 mg=dL and she

was keeping poor records.

Rationale for Initiating ProfessionalContinuous Glucose Monitoring

Her baseline A1c was 7.1%, which is almost at target, but the few

reported BG measurements were so variable that professional con-

tinuous glucose monitoring (CGM) was ordered to clarify the patterns

and help the patient establish better control.

Initial Professional CGM ResultsThe sensor summary reveals that this patient is within her BG

target ranges about 47% of the time. The Sensor Modal Day tracing

showed a roller-coaster pattern of both highs and lows throughout

the day. Sensor Daily Details revealed that the patient was having

premeal hypoglycemia and postmeal hyperglycemia throughout the

evaluation period (Fig. 14). Sensor Modal Time analysis showed that

she was peaking in the middle of the night and then having hypo-

glycemia toward the morning.

Examining the diary in conjunction with the CGM outputs revealed

poor dietary habits, with meals that included very little protein and a

lot of simple carbohydrates (e.g., granola bars, graham crackers, and

popcorn), probably very typical for a college student. Her BGs often

increased significantly after eating these high-carbohydrate meals.

When she developed hypoglycemia she again consumed a lot of car-

bohydrates, which led to more hyperglycemia.

Therapy Adjustments=Treatment AlterationsWe decreased her Glargine to 12 units because of the fasting hy-

poglycemia on awakening. The dietician recommended increasing

Fig. 14. Sensor Daily Details reveal several unrecognized episodes of nocturnal hypoglycemia and several hyperglycemic peaksmissed with self-monitored BG.

18 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 19: Cgm case studies

protein consumption and decreasing simple carbohydrates, and

taught her how to avoid over treating the hypoglycemia. The certified

diabetes educator helped her calculate premeal insulin dosing. The

physician suggested that the patient go on an insulin pump to more

accurately dose her insulin. The patient now uses Aspart with a

Medtronic 522 pump with individualized basal rates, carbohydrate

ratios, sensitivities, and glucose targets.

Follow-Up Professional CGM Results=Responseto Therapy Adjustments

A follow-up professional CGM was performed about 3 months

later. The patient’s BG was in normal ranges 77% of the time, com-

pared with <50% of the time at the earlier evaluation. In contrast

to the roller-coaster pattern seen in the first one, the BG pattern

smoothed out, with very little hypoglycemia, and only occasional

hyperglycemia. The Sensor Modal Time report shows that the earlier

2 am peaks and the hypoglycemia episodes on awakening are gone

(Fig. 15). The preprandial hypoglycemia and postprandial hyper-

glycemias were essentially eliminated.

The diary indicated improved dietary habits. The patient still has

occasional excursions on the Sensor Daily Detail report, probably as a

result of reverting to poor food habits. The patient’s glycemic control

improved markedly, with her A1c improved from 7.1% to 5.7% at her

most recent visit, and no accompanying hypoglycemic episodes.

ConclusionsThe major take-home message from this case is that CGM reveals

behavior patterns that inhibit good glycemic control. BG excursions

that do not appear with routine monitoring become obvious often

times when we do professional CGM. These results motivated the

patient to consider a pump. Ultimately, this patient progressed to

better control because of the professional CGM technology.

Disclosure StatementBarry S. Horowitz, M.D., is a speaker for Abbott Pharmaceuticals,

Amylin Pharmaceuticals, Astra-Zeneca Pharmaceuticals, Bristol

Myers Squibb, Eli Lilly Pharmaceuticals, Merck & Co, Inc., Medtronic

Diabetes, Novo Nordisk, Pfizer Pharmaceuticals, Sanofi-Aventis

Pharmaceuticals, and Takeda Pharmaceuticals.

Fig. 15. The Sensor Modal Time tracings revealed elimination of the hyperglycemia peaks at 2 am and the hypoglycemia onawakening seen in the previous professional CGM study.

CASE 6: DIABETIC COLLEGE STUDENT WITH CRAZY LIFESTYLE

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 19

Page 20: Cgm case studies

Case 7A 38-Year-Old Woman with Type 1 Diabetes

Bruce W. Bode, M.D.

Atlanta Diabetes Associates, Atlanta, Georgia.

Patient HistoryThis is the case of a 38-year-old woman with type 1 diabetes

diagnosed in 1988. The baseline regimen was insulin Aspart via

insulin pump. The baseline A1c was 9.2%, although her A1c values

were in the range during two prior pregnancies. The patient is gen-

erally not compliant: she has used the same insulin dose for years,

forgets to bolus regularly, does not bring her blood glucose (BG)

logbook to office visits, and rarely monitors herself (zero to two

times daily). She claims to be too busy to monitor her BG, although

she did so during pregnancy. The patient is concerned about weight

gain, and about hypoglycemia if she attempts to improve her gly-

cemic control.

Comorbidities and Diabetic ComplicationsThe patient shows signs of preproliferative retinopathy. She also

has a history of depression, but does not take antidepressants because

she gained weight when on treatment in the past. Thus, her fear of

potential weight gain prevents her from treating both her depression

and her diabetes.

Rationale for Initiating ProfessionalContinuous Glucose Monitoring

The patient is on an insulin pump and manually boluses approx-

imately five times a day with 2 to 7 units. She will purposely under

bolus if she feels she might go low. The patient denies snacking after

9:00 pm but has very high glucose levels in the morning. Taking the

patient at her word, the appropriate therapeutic choice would be to

increase her basal insulin. The provider wished to know whether the

patient experienced a rise in BG in the dawn phase.

Initial Professional Continuous GlucoseMonitoring Results

The Sensor Daily Detail report revealed a clear dawn rise starting

earlier than what was previously recognized (Fig. 16). She often has

normal BG at bedtime, but at *1 am her BG will start to rise from

150 to 300 mg=dL by the time she wakes up. If her BG at bedtime is

around 120 mg=dL, it will rise to 280 mg=dL or so upon awakening.

If her bed time BG is 100 mg=dL, it may rise to 180 mg=dL on

awakening. She denies eating despite apparent evidence to the

contrary. During the day, her BG rarely dips below 70 mg=dL, but

she has many hyperglycemic excursions, some up to 346 mg=dL.

Yes, there was 1 day when her BG was consistently in the 150 to

200 mg=dL range.

The patient was instructed not to change her usual patterns during

the professional continuous glucose monitoring (CGM) evaluation,

but she obviously did because her average BG was 150 mg=dL on her

Sensor Summary (Fig. 17), and it does not correspond with an A1c of

9.2%. Clearly, the patient bolused and monitored more often than

usual. Even so, the BG patterns are still erratic. Either she is eating at

night to protect against lows overnight, or she needed an increased

basal insulin dose.

Therapy Adjustments=Treatment AlterationsInitially, the provider increased the basal rate, followed by a sec-

ond increase several weeks later. The patient finally agreed to a pump

after seeing the reports that she would not go low. However, she

would not agree to use the Bolus Wizard, nor would she reinstate her

antidepression medication out of fear of weight gain.

Response to Therapy AdjustmentsThe patient’s last A1c did not change much (9.1%), but at least

she is monitoring twice a day. She usually boluses five times daily,

which is an improvement. She has also applied for personal CGM

coverage.

ConclusionsProfessional CGM revealed previously unrecognized problems

with this patient’s BG patterns overnight. This patient may benefit

most from personal CGM, but she has to become much more ac-

cepting of her diabetes to take that proactive approach.

Professional CGM may not give an accurate picture of each pa-

tient’s diabetes control, because patients may change their diabetes-

related behavior during the professional CGM evaluation, despite

being told to maintain their usual routine.

The following steps will improve the likelihood that patients will

not change their behavior while on professional CGM.

. Before insertion, explain that the reason we are doing the study

is to see actual day-to-day diabetic control and stress not to

change anything for the first couple of days.. Specifically mention that you are not going to judge the patient.

Acknowledge that it may be difficult for the patient to do ev-

erything as usual because he or she knows that the healthcare

professional is going to see what the patient is actually doing,

but that is why it is so important.

According to a recent study, the second best predictor of success

with CGM (after age >25 years) is frequency of BG self-monitoring.

To be successful using a personal CGM, the patient must monitor

20 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 21: Cgm case studies

Fig. 16. Sensor DailyDetails reveals a dawnphenomenon earlier thanexpected, and bloodglucose peaks that arenot recognizedwith SMBG.

Sensor Summary

Date

Sensor

# of Sensor Values

6/26/2009 6/27/2009 6/28/2009 6/29/2009

288

155

56-331

72

4

92

46-141

4

11.7

n/a 1.00

5

3

2

288

209

86-346

65

4

157

72-242

4

288

188

87-307

66

5

172

77-270

5

11.5

0.99

3

3

0

14:25 (60%)

09:35 (40%)

00:00 (0%)

12.9

0.99

1

1

0

19:05 (80%) 10:25 (43%)

13:10 (55%)

00:25 (2%) 00:55 (4%)

11:35 (48%)

11:30 (48%)

2

4

6

7.1

6

49-262

124

6

66

57-286

164

288

6/30/2009

29

0 0 0

33 12

04:55 (20%)

00.00 (0%)

48

0

67

0

102

133

47-197

33

3

108

96-116

3

28.1

n/a

4

3

1

02:25 (28%)

05:50 (69%)

00:15 (3%)

5

1

X: Use ClinicalJudgment

Average (mg/dL)

Min - Max (mg/dL)

STDev (mg/dL)

# of Meter Values

Average (mg/dL)

Min - Max (mg/dL)

Designation

# of Paired Readings

Mean Abs. Diff. [MAD %]

Correlation Coeff. [R]

# of Excursions*

# of High Excursions*

# of Low Excursions*

Duration Above High Limit

Duration Within Limits

Duration Below Low Limit

Pie ChartRed: Above LimitsGreen: Within LimitsBlue: Below Limits

Glucose Area Above HighLimit (mg/dL*Day)

Glucose Area Below LowLimit (mg/dL*Day)

Meter

Optimal AccuracyCriteria

ExcursionsHigh > 150mg/dLLow < 70mg/dL

0.95

00:05 (0%)

09:30 (53%)

08:30 (47%)

1

2

3

5.5

4

118-232

160

5

33

69-226

150

217

7/1/2009

X: Please use your clinical judgment - this day does not satisfy the optimal accuracy criteria according to set thresholds:N>¼ 3, R>¼ 0.79 and MAD<¼28% [or<¼ 18% if the range (Min-Max) of meter values is less than 100mg=dL (5.6mmol=L) - see CriteriaNote below].

C: This day does not have any paired sensor=meter data and no sensor plot is provided. As a result, ‘Meter Only’ data is available.S: Please use your clinical judgment - this day does not have any meter data. As a result, ‘Sensor Only’ data is available.Criteria Note: If the range (Min-Max) of Meter Values is less than 100mg=dL (5.6mmol=L) then ‘R’ will be reported as ‘N=A’.

In this case the optimal accuracy threshold set for MAD is<¼ 18%.Excursion Note: Excursions are counted in the day that the excursion event started.

Fig. 17. Sensor Summary reveals an average BG of 150mg=dL. Patient experiences hyperglycemia from 28% to 80% of the day.

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 21

Page 22: Cgm case studies

at least four times daily. Because this patient routinely does not self-

monitor as recommended, and is likely to hear multiple alarms

because of her extreme glucose excursions, she might not be very

successful on a personal CGM. On the other hand, since she changed her

self-monitored blood glucose (SMBG) behavior during the professional

CGM evaluation as well as during her previous pregnancies, she might

be motivated enough to develop better control on personal CGM.

Some patients use personal CGM data exclusively to make changes

in insulin dosing instead of SMBG. This may be particularly impor-

tant when treating adolescents who may be out of control. It is

possible, however, that personal CGM may help teens moderate BG

swings even if performing SMBG less often than recommended. Some

adolescents use personal CGM successfully, especially when driving

privileges may be withheld if they don’t.

Disclosure StatementBruce W. Bode, M.D., received fees for research grants, advisory

boards, and consultant activities, and is on the speaker’s bureau for

Johnson & Johnson, Medtronic Diabetes, Novo Nordisk, and Sanofi-

Aventis.

BODE

22 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 23: Cgm case studies

Case 8A 69-Year-Old Woman with Type 2 Diabetesand Good Premeal Glucose, But a High A1c

Barry S. Horowitz, M.D.

Palm Beach Diabetes and Endocrine Specialists, PA, West PalmBeach, Florida.

Patient HistoryThis patient is a 69-year-old woman who was found to have type 2

diabetes about 6 years ago when she developed cardiomyopathy. Her

current treatment regimen includes Aspart before meals with dif-

ferent units and Glargine at bedtime. She did not follow a 50:50 ratio

of basal and bolus therapy, but her regimen presumably worked for

her at the time.

Comorbidities and Diabetic ComplicationsHer comorbid conditions include cardiomyopathy, hyper-

tension, and hyperlipidemia, for all of which she was taking

medication.

Rationale for Initiating ProfessionalContinuous Glucose Monitoring

The patient was checking her blood glucose (BG) four times

daily. She stated that her blood sugars were in the low 100s

premeals and at bedtime, but she routinely forgot her logbook.

When she did bring her logbook, her premeal BG looked

pretty good, but she was very noncompliant postprandial self-

monitored BG, despite multiple requests to do so. Her baseline

A1c was 7.5%.

The provider ordered professional continuous glucose monitoring

(CGM) because he suspected that the patient was having some

postprandial hyperglycemia.

Initial Professional CGM ResultsThe Sensor Summary from her first professional CGM showed

that her CGM glucose values equally distributed in within, be-

low, and above target range. The patient was completely un-

aware of hypoglycemia, although she experienced substantial

nocturnal hypoglycemia everyday. The Sensor Modal Day

showed that she was having some postprandial hyperglycemia

and variability in her glycemic control in the evening (Fig. 18).

The Sensor Daily Detail and Sensor Modal Time confirmed that

her BG dropped throughout the night, leading to hypoglycemia

toward the early morning. She also experienced postprandial

hyperglycemia.

Therapy Adjustments=Treatment AlterationsThe patient’s diary revealed that she did not take premeal insulin

when her premeal blood sugar was normal. She had had a miscon-

ception that if she had normal BG before meals, she did not need to

take insulin. She also had subsequent postprandial hyperglycemia

that was not captured with self-monitored BG.

This patient is Hispanic and eats traditional, high-carbohydrate

meals with lots of rice and beans. Comparing her diary entries to

professional CGM outputs revealed that she took her usual dose of

premeal insulin, even on days when she ate high-carbohydrate meals,

and consequently developed postmeal hyperglycemia.

The provider lowered her glargine dose at bedtime to reduce

nocturnal hypoglycemia and provided additional education from a

nutritionist and a diabetes educator. We again emphasized that she

should dose her Aspart before meals regardless of what her blood

sugars were. We also taught her how to count carbohydrates and use

that along with the correction factor so she could be taking more

insulin when she ate higher-carbohydrate meals and have better

postprandial results.

Response to Therapy AdjustmentsThe patient began checking postmeal BG after seeing the postmeal

hyperglycemia on the professional CGM report. She began counting

carbohydrates with the correction factor for meals, which lead to

better and less variable BG throughout the day.

Follow-Up Professional CGMThis follow-up CGM showed improved and more consistent gly-

cemic control. The Sensor Summary revealed that the hypoglyce-

mia was almost eliminated, with most BG values within the normal

range and some hyperglycemia. The Sensor Modal Day report also

shows some hypoglycemia overnight but not in the ranges that she

was having before. The Sensor Modal Time report also revealed BG

dips from 4 to 8 am, and some high BG after breakfast and dinner that

was better than before. The Sensor Daily Detail report and diary

showed that the patient was still often under dosing the mealtime

insulin, but this was better compared to before when she was skipping

doses entirely (Fig. 19). She was given some further counseling about

her diet, and we cut back a little bit more on her glargine dose. Her

most recent A1c had improved from 7.5% to 6.5%.

ConclusionsThis subject with type 2 diabetes denied experiencing hypo-

glycemia, but she was low a third of the time during the first

ª 2010 Medtronic Minimed, Inc. All rights reserved. CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 23

Page 24: Cgm case studies

professional CGM evaluation. Many researchers think that the

increased mortality may be associated with unrecognized hypo-

glycemia with tight glycemic control in patients with type 2 di-

abetes as has been hypothesized in the ACCORD trial. Many

patients are unwilling or unable to check BG at the most the

clinically informative times (e.g., postprandially). The professional

CGM showed this patient how her behavior patterns influenced her

BG patterns. In this patient, the professional CGM resulted in

behavior changes, leading to improvement in A1c from 7.5% to

6.5% with no associated hypoglycemia.

Fig. 18. The Sensor Modal Dayshowed blood glucose droppingthroughout the night and morningand postprandial hyperglycemia.

Fig. 19. The Sensor Daily Detail report showed that the patient was still often under dosing her mealtime insulin, but at least wasnot skipping doses.

HOROWITZ

24 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 25: Cgm case studies

One panelist suggested that Dr. Horowitz might want to consider

splitting the glargine into two doses, given the trend toward higher

BG in the evening, but the postprandial hyperglycemia at dinner was

not high enough. This patient was assigned a ratio of 3:1. Realisti-

cally, this patient was not going to abandon her traditional Hispanic

diet, so she needed that ratio, and the professional CGM tracings

reinforced that therapeutic decision.

Disclosure StatementBarry Horowitz, M.D., is a speaker for Abbott Pharmaceuticals,

Amylin Pharmaceuticals, Astra-Zeneca Pharmaceuticals, Bristol

Myers Squibb, Eli Lilly Pharmaceuticals, Merck & Co, Inc., Medtro-

nic Diabetes, Novo Nordisk, Pfizer Pharmaceuticals, Sanofi-Aventis

Pharmaceuticals, and Takeda Pharmaceuticals.

CASE 8: DIABETIC WOMAN WITH GOOD PREMEAL GLUCOSE, BUT HIGH A1C

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 25

Page 26: Cgm case studies

Case 9A 37-Year-Old Woman with Type 2 Diabetes at 20 Weeks of Gestation

Kathleen C. Arnold, A.N.P.

The Diabetes Center, PLLC, Ocean Springs, Mississippi.

Patient History

This is a case of a 37-year-old woman with type 2 diabetes

associated with pregnancy. She had no known diabetes

complications and her baseline A1c was 5.7%. She tests her

blood glucose (BG) six to eight times daily. Her baseline

treatment was NPH twice a day (14 units in the morning and 20 units

at night), and Lispro. The BG target is lower for pregnancy, 80 divided

by 40 and carbohydrates divided by 10.

Rationale for Initiating Professional ContinuousGlucose Monitoring

The healthcare professional wanted to ensure that the patient was

doing well day to day, in light of the challenges posed by BG control

during pregnancy for patients with type 1 or type 2 diabetes.

Initial Professional Continuous GlucoseMonitoring Results

The Sensor Daily Detail revealed high postprandial BG values,

some as high as 200 mg=dL, starting at breakfast and usually wors-

ening at lunch and dinner (Fig. 20).

Therapy Adjustments=Treatment AlterationsThe clinician reviewed her dietary log and discussed the rela-

tionships between her high-carbohydrate and high-fat intake and her

high postprandial BG readings. The dietician recommended dietary

changes and her carbohydrate ratio was adjusted to carbohydrates

divided by eight.

Follow-Up Professional Continuous GlucoseMonitoring Results=Response to TherapyAdjustments

Before a follow-up professional continuous glucose monitoring

to re-verify the ratio was scheduled, the patient was induced, and she

delivered a healthy 6-pound boy at 38 weeks.

Fig. 20. Sensor Daily Details revealed unrecognized postprandial hyperglycemia.

26 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 27: Cgm case studies

ConclusionsThis case emphasizes that an A1c in the 5.6%–5.7% range through

the entire pregnancy may mask underlying postprandial BG excur-

sions, even in a patient who checked her BG six to eight times daily,

including 2 h after meals.

Although this patient had a healthy outcome (a normal-sized

infant), diabetic women with decent A1cs may still deliver a mac-

rosomic infant. A high A1c during pregnancy is a major risk factor

for macrosomia, which is also thought to be related to obesity and

insulin-resistant syndrome later in life. There’s a trend toward low-

ering the target A1c to <5% in pregnancy. The American Diabetes

Association fasting target is now 95 mg=dL (ADA 2008). The

guidelines may be revised to 80 mg=dL to try to get better glucose

control.

With this patient, early intervention contributed to a healthy

outcome. To prevent macrosomia, ideally a clinician should inter-

vene preconception, but failing that, one should intervene no later

than early in the second trimester.

Disclosure StatementKathleen C. Arnold, A.N.P., BC-ADM, is a speaker for Lilly, Med-

tronic Diabetes, NovoNordisk, and Sanofi Aventis.

CASE 9: WOMAN WITH DIABETES AT 20 WEEKS OF GESTATION

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 27

Page 28: Cgm case studies

Case 10A 41-Year-Old Woman with Type 2 Diabetes, High A1c

Ola Odugbesan, M.D.

North Atlanta Endocrinology, Lawrenceville, Georgia.

Patient History

A41-year-old woman was initially found to have gesta-

tional diabetes during pregnancy and type 2 diabetes in

2003 after she miscarried. Her A1c measured 8.4% in early

summer 2009. She is one of the clinical diabetes managers

within the presenter’s practice, but her own diabetes was being

managed by her internist.

Comorbidities and Diabetic ComplicationsThe patient is on prednisone therapy for asthma and has irritable

bowel syndrome. She developed profound hyperglycemia that re-

quired insulin therapy.

Rationale for Initiating Professional ContinuousGlucose Monitoring

At the time she agreed to be followed within the practice where she

worked, she was on pioglitazone and metformin, but she was prob-

ably not compliant with this regimen, because of gastrointestinal side

effects from metformin.

Sensor Summary

Date

Sensor

# of Sensor Values

5/27/2009 5/28/2009 5/29/2009 Totals

577

257

106-323

32

11

253

173-308

10

5.3

0.84

2

2

0

169

282

234-323

20

2

290

271-308

2

288

252

182-297

24

5

259

239-268

5

3.3

n/a

0

0

0

24:00 (100%)

00:00 (0%)

00:00 (0%)

7.5

n/a

0

0

0

14:05 (100%) 47:20 (98%)

00:45 (2%)

00:00 (0%)

78

0

00:00 (0%)

00.00 (0%)

72

0

102

0

120

234

106-299

39

4

227

173-306

3

7.0

0.94

2

2

0

09:15 (92%)

00:45 (8%)

00:00 (0%)

57

0

X: Use ClinicalJudgment

Average (mg/dL)

Min - Max (mg/dL)

STDev (mg/dL)

# of Meter Values

Average (mg/dL)

Min - Max (mg/dL)

Designation

# of Paired Readings

Mean Abs. Diff. [MAD %]

Correlation Coeff. [R]

# of Excursions*

# of High Excursions*

# of Low Excursions*

Duration Above High Limit

Duration Within Limits

Duration Below Low Limit

Pie ChartRed: Above LimitsGreen: Within LimitsBlue: Below Limits

Glucose Area Above HighLimit (mg/dL*Day)

Glucose Area Below LowLimit (mg/dL*Day)

Meter

Optimal AccuracyCriteria

ExcursionsHigh > 180mg/dLLow < 70mg/dL

X: Please use your clinical judgment - this day does not satisfy the optimal accuracy criteria according to set

thresholds:

N>¼ 3, R>¼0.79 and MAD<¼28% [or<¼ 18% if the range (Min-Max) of meter values is less than 100mg=dL

(5.6mmol=L) - see Criteria Note below].

C: This day does not have any paired sensor=meter data and no sensor plot is provided. As a result, ‘Meter

Only’ data is available.

S: Please use your clinical judgment - this day does not have any meter data. As a result, ‘Sensor Only’ data

is available.

Criteria Note: If the range (Min-Max) of Meter Values is less than 100mg=dL (5.6mmol=L) then ‘R’ will be

reported as ‘N=A’.

In this case the optimal accuracy threshold set for MAD is<¼ 18%.

Excursion Note: Excursions are counted in the day that the excursion event started.

Fig. 21. The SensorSummary revealed that thepatient was hyperglycemicall day long.

28 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING ª 2010 Medtronic Minimed, Inc. All rights reserved.

Page 29: Cgm case studies

She was also supposed to be taking Lispro PRN for episodes of hy-

perglycemia. She was supposed to be checking her blood glucose (BG)

three to four times daily, but was apparently noncompliant with that

also. While she was on vacation, she developed profound hyperglyce-

mia, felt weak and diaphoretic, and may have exacerbated her asthma

exacerbation, and also developed foot ulcers that took a little while to

heal. She ended up in the hospital. She is the nurse who counsels the

patients who come in for professional continuous glucose monitoring

(CGM), and is actually the one who places and removes the devices.

Initial Professional CGM ResultsAs a nurse specializing in diabetes care, the patient saw the Sensor

Summary and immediately recognized the implications (Fig. 21). In

fact, she described herself as a ‘‘rolling ball of sugar.’’ The pie charts

show that she is hyperglycemic all day long, with the rare episode of

normoglycemia. There were no meal markers because she essentially

grazed all day long.

Therapy Adjustments=Treatment AlterationsThe physician maintained the patient’s pioglitazone prescription,

because she does have profound insulin resistance, but metformin

was discontinued because of the gastrointestinal side effects. She was

switched to exenatide only, 5 mcg twice daily. Given her previous

poor dietary habits, she still wanted to have Lispro on hand if she ate a

very high carbohydrate meal. The patient saw the dietitian, and de-

cided that she was going to be more active and try to achieve better

diabetes control.

Follow-Up Professional CGM Results=Responseto Therapy Adjustments

Follow-up professional CGM showed that the patient was more

compliant with her treatment regimen and her BG was within target

95% of the time, with very little hyperglycemia, and virtually no

hypoglycemia (Fig. 22 and 23). She did not have gastrointestinal side

effects and so complied with her therapy. She lost about 20 pounds

Sensor Summary

Date

Sensor

# of Sensor Values

9/15/2009 9/16/2009 9/17/2009 9/18/2009

159

139

98-165

12

1

155

155-155125-190

1

13.4

n/a n/a

0

0

0

288

149

88-193

20

4

159

4

288

138

63-181

21

4

144

133-160

4

6.3

n/a

2

1

1

00:15 (1%)

23:10 (97%)

00:35 (2%)

23.2

n/a

2

2

0

01.55 (8%) 00:00 (0%)

13:15 (100%)

00:00 (0%) 00:35 (0%)

61:40 (96%)

02:10 (3%)

1

3

4

14.8

10

125-190

154

10

19

63-193

143

773

Totals

0

0 0

0

22.05 (92%)

00.00 (0%)

0

0

0

0

38

145

137-165

6

1

172

172-172

1

17.0

n/a

0

0

0

00:00 (0%)

03:10 (100%)

00:00 (0%)

0

0

Average (mg/dL)

Min - Max (mg/dL)

STDev (mg/dL)

# of Meter Values

Average (mg/dL)

Min - Max (mg/dL)

Designation

# of Paired Readings

Mean Abs. Diff. [MAD %]

Correlation Coeff. [R]

# of Excursions*

# of High Excursions*

# of Low Excursions*

Duration Above High Limit

Duration Within Limits

Duration Below Low Limit

Pie ChartRed: Above LimitsGreen: Within LimitsBlue: Below Limits

Glucose Area Above HighLimit (mg/dL*Day)

Glucose Area Below LowLimit (mg/dL*Day)

Meter

Optimal AccuracyCriteria

ExcursionsHigh > 180mg/dLLow < 70mg/dL

X: Use ClinicalJudgment

X: Use ClinicalJudgment

X: Use ClinicalJudgment

X: Please use your clinical judgment - this day does not satisfy the optimal accuracy criteria according to set thresholds: N>¼ 3, R>¼ 0.79

and MAD<¼ 28% [or<¼ 18% if the range (Min-Max) of meter values is less than 100mg=dL (5.6mmol=L) - see Criteria Note below].

C: This day does not have any paired sensor=meter data and no sensor plot is provided. As a result, ‘Meter Only’ data is available.

S: Please use your clinical judgment - this day does not have any meter data. As a result, ‘Sensor Only’ data is available.

Criteria Note: If the range (Min-Max) of Meter Values is less than 100mg=dL (5.6mmol=L) then ‘R’ will be reported

as ‘N=A’.In this case the optimal accuracy threshold set for MAD is<¼ 18%.

Excursion Note: Excursions are counted in the day that the excursion event started.

Fig. 22. Sensor Summary pie charts show the patient within targets 95%–100% of the time.

CASE 10: WOMAN WITH DIABETES AND HIGH A1C

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 29

Page 30: Cgm case studies

over a 4-month period because of improved diet, increased exercise,

and perhaps to some extent exenatide therapy. She feels better and

has had a much better quality of life.

The very rare episodes of hyperglycemia may reflect the days that

the pharmaceutical company sales representatives brought lunch for

the staff. Her latest A1c is within the normal range. This is a dramatic

difference from where she was before having the therapy.

ConclusionsThis patient was not managing herself at all and truly ignoring her

diabetes. Her average BG was 257 and her A1c was 8.4%.

Use of professional CGM demonstrated that she does not need to

eat all day long and live in perpetual hyperglycemia. She now con-

vinces other patients of the value of professional CGM because she

knows firsthand how it improved her glycemic control.

The provider attributed the patient’s weight loss primarily to a

combination of diet and increased levels of physical activity, rather

than to the medication change, because the exenatide dose (5 mcg)

was so small.

Disclosure StatementOla Odugbesan, M.D., is a speaker for Abbott, Amylin, Astra

Zeneca, Bristol-Myers Squibb, Glaxo SmithKline, Lilly, Medtronic

Diabetes, Merck, Novo Nordis, Satauru, and Takeda.

Fig. 23. Sensor Daily Details show blood glucose within range except for two episode of mild hyperglycemia (circled).

ODUGBESAN

30 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 31: Cgm case studies

Discussion Regarding Use of ProfessionalContinuous Glucose Monitoring

The key points for optimizing patient management with

professional continuous glucose monitoring (CGM) reached

by consensus of the roundtable participants included

CGM Terminology, the Consensus Statement, Selection

Criteria for Professional CGM, Implementation of Professional

CGM in a Practice Setting, Interpreting Professional CGM Outputs,

Therapy Adjustments Guided by Professional CGM, and the Future of

CGM.

Consensus on CGM TerminologyProfessional CGM is the term used most often to describe CGM

characterized by the following:. The aim of professional CGM is to assess the patient’s real-

world behavior and how it influences glucose patterns and to

determine the glucose patterns that are not visible when the

patient performs self-monitored blood glucose (SMBG) only

four to six times per day. The ultimate goal of professional CGM

is to enable the healthcare professional make appropriate ad-

justments in diet, physical activity, and adjustments in the

dosages of insulin and other medications.. The device is owned by the individual healthcare provider, the

practice, or the hospital center.. The CGM device is put in place by a professional licensed person.. The patient is blinded to the readings and there are no audible

alarms to alert patients to hypo- or hyperglycemia.. The device stays in place for 72 h.. Professional CGM is more of a diagnostic tool.. Upon downloading, the provider sees and interprets the read-

ings retrospectively and discuss appropriate therapy adjust-

ments with the patient.

Other terms used to describe professional CGM include blinded

CGM, blinded retrospective CGM, historical CGM, clinic CGM, CGM

for physicians, 72-h CGM, and CGM system.

Professional CGM is particularly useful for type 1 patients who

take multiple insulin injections or are on pumps to fine-tune their

dosage. In a patient with type 2 diabetes with an unexplained elevated

A1c, professional CGM can pinpoint needed changes in treatment that

would bring their A1c down. Professional CGM is particularly useful

for those patients who might be considering personal CGM, those

considering a change from MDI Insulin to pump therapy or from oral

agents to insulin.

Real-time CGM is the term used most often to describe CGM char-

acterized by the following:. The device is owned by the individual patient.. The CGM device is put in place by a professional licensed

person.

. The patient is not blinded to the readings and audible alarms

alert the patient to hypo- or hyperglycemia.. Personal CGM is both a diagnostic and therapeutic tool.. Patients must be trained to make their own therapy adjust-

ments.* Data are displayed either on the pump or for patient not on a

pump on the CGM device itself.* Patients see and interpret the CGM data these in real time and

make their own therapeutic decisions.. The device remains in place permanently (sensor changed every

3 days). According to data from the JDRF trials the best benefit

accrues to patients who use the personal CGM device 6 days a

week on an ongoing basis.. When the physician sees the data, they are interpreted retro-

spectively; the patient has already acted upon the data.

Other terms used to describe personal CGM include real-time CGM,

home-use CGM, patient CGM, and consumer CGM.

Personal CGM is especially useful for patients who already mon-

itor with SMBG frequently that are not at goal. It takes a lot more

training and follow-up with personal CGM than it does with pro-

fessional CGM to ensure that the patients interpret the real-time

readings. If a patient is a candidate for personal CGM, the profes-

sional CGM evaluation is useful to document hypoglycemia, if the

condition has not been adequately documented using SMBG alone.

Some insurance carriers now require professional CGM as part of

the authorization for a personal CGM device. There are patients who,

even if their plan would authorize it, do not want a personal CGM

device but would agree to wear a professional CGM device for 3 days

to optimize therapy.

Roundtable participants agreed that providers must use consistent

terminology when distinguishing professional CGM from personal

CGM for the following reasons:. When the term ‘‘professional’’ is attached to the term ‘‘CGM,’’

patients understand that it is the healthcare provider who col-

lects data that are going to help manage their diabetes.. When the adjective ‘‘personal’’ is attached to the term CGM,

patients understand that it is the patient who is most active in

the process; in addition, patients may not understand what the

term ‘‘real time’’ means.. Insurers do not necessarily understand the distinction between

professional and personal CGM and may reimburse for some-

thing that was not requested.

Consensus on Definition of Professional CGMProfessional CGM is a diagnostic and therapeutic=management

tool used by the healthcare professional that allows the collection of

ª 2010 Medtronic Minimed, Inc. All rights reserved. CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 31

Page 32: Cgm case studies

retrospective glucose data in patients with diabetes. It is indicated to

optimize safe and effective glycemic control in patients in whom self-

monitoring of BG is inadequate.

Consensus on Selection Criteria forProfessional CGM

The overarching reason to use professional CGM is to better under-

stand a patient’s glycemic control when self-monitoring of BG proves

inadequate to do so. A typical driver for consideration of professional

CGM is a patient who does not maintain their BG within targeted ranges,

or is not at their A1c goal, or both. The healthcare provider will need to

determine patterns that undermine safe and effective therapy, and make

appropriate therapeutic changes to get the patient to goal.

The rationales for professional CGM for the 10 cases presented at

this Roundtable are summarized in Table 1.

Cases 1, 2, 5, and 7 were all patients with type 1 diabetes and an

A1c that was too high. Two cases with type 1 diabetes had normal

A1cs, but each patient faced challenges with day-to-day glycemic

control—one because of intense athletic activities, and the other be-

cause of a demanding college schedule. In two cases (the college

student and the women with type 1 pregnant with triplets), the

professional CGM data influenced patients to switch to an insulin

pump, resulting in A1c values that declined to 5.7% and 5.6%, re-

spectively, without any hypoglycemia. Case 9, the type 2 diabetic

woman on oral agents, switched to insulin therapy using a pump and

was able to get to goal, and deliver a healthy child no evidence of

macrosomia. Two cases had type 2 diabetes: case 8 was on multiple

daily injections with large amounts of insulin, with BG unawareness

at both extremes. Just by altering the regimen, she eliminated all

hypoglycemia and achieved an A1c of 6.5%. Case 10 was on oral

agents only and hyperglycemic 100% of the time. As a diabetes

specialist nurse, she recognized the implications of her tracings, and

she was finally motivated enough to change her diet and activity and

thus achieved normal BG levels.

Participants agreed that the goal of professional CGM should be to

open a patient’s eyes to what is going on with their diabetes, to get

that ‘‘aha!’’ moment when the patients realize why they need to do the

things they know that they should be doing, but that they are not.

Data from professional CGM may also be a trigger than can convince

patients to switch from MDI to pump therapy, if that is what the

physician has been recommending, or from pump therapy alone to

pump therapy with personal CGM.

One participant always recommends professional CGM for patients

who experience any severe hypoglycemic event requiring assistance

of another person. Another rationale for professional CGM is for

patients whose A1c is not at goal in spite of frequent glucose mon-

itoring, or if their A1c is at goal with a high variability in their BG

meter download.

Patients who do not comply with recommendations for routine

glucose monitoring frequency may be willing to do SMBG four times

daily for 3 days, even if they generally come to physician visits with

no fingerstick data. The clinician needs to make sure that patient is

safe, and professional CGM can motivate them to do better.

Professional CGM can be justified in all diabetic patients who take

medications that can lead to hypoglycemia, even if they have A1c

values in a healthy range. Thus, all type 1 patients, and those type 2s

on insulin or oral agents that cause hypoglycemia (e.g., sulfonyl-

ureas) are candidates for professional CGM.

For patients with type 2 diabetes who are not on insulin and

whose A1c is not at goal, professional CGM is useful to determine

whether to offer the patient an alternative oral or switch to insulin.

For type 2 diabetic patients on basal insulin only, professional CGM

can help identify whether the patient may need to add rapid-acting

insulin for a meal, or to titrate multiple insulin doses as needed.

Insulin-requiring patients who have had type 2 diabetes for a long

time may have brittle disease that is very similar to type 1 char-

acteristics.

One important take-home message is that patients may have A1c

values that are nominally at goal, but may actually have poor gly-

cemic control. The A1c is an average value and, as such, may mask

widely fluctuating BG, especially at the extremes. If the SMBG record

indicates wide variations, the patient would be an excellent candidate

for professional CGM.

Roundtable participants emphasized that the terms ‘‘controlled’’ or

‘‘uncontrolled’’ diabetes should not be determined solely by the A1c,

particularly in type 1 diabetes. If a patient’s A1c is 6.5% but their BG

ranges between 40 and 400, then their diabetes in uncontrolled, and

it should be coded that way. This designation is particularly important

during pregnancy because of the risk of unrecognized postprandial

hyperglycemia, especially in type 1 diabetics that are near normal.

There is a correlation between postprandial glycemia control and

complications in gestational diabetes (HAPO 2008 N Engl J Med).

Patients still experience significant morbidity from hypoglycemia up

to and including the possibility of death. Unfortunately, there was no

professional CGM component in the large cardiovascular outcome

trials in type 2 diabetes, specifically Action to Control Cardiovascular

Risk in Diabetes trial (Gerstein 2008 NEJM).

After thorough discussion of the cases that were presented,

roundtable participants discussed the types of patients for whom

professional CGM is indicated to optimize glycemic control. Char-

acteristics of patients who are good candidates for professional CGM

are listed in Table 2.

Consensus on Implementation of ProfessionalCGM in the Private Practice SettingThe Value of Retrospective Data withProfessional CGM

When considering whether to implement professional CGM in

their practices, healthcare professionals should be aware of the ad-

vantages of establishing such procedures as part of their service of-

ferings. These advantages are summarized in Table 3.

StaffOne of the key components of initiating a professional CGM

component to a provider’s practice is having a dedicated staff. Par-

ticularly helpful is to have either a licensed professional in your

DISCUSSION REGARDING USE OF PROFESSIONAL CGM

32 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 33: Cgm case studies

practice who is enthusiastic enough to champion the technology, a

medical assistant willing to be in charge of this program, or a mo-

tivated, well-educated patient who is willing to volunteer with pa-

tients who might be candidates for professional CGM.

Ideally, at least two staff members, including medical assistants,

should be cross-trained to insert and remove the professional CGM

devices. Then, if the champion or team leader is absent, there is

continuity.

SpaceSpace is a big issue in most practices. If a dedicated room is not

available, then a rolling cart to hold the computer and related

equipment is one option.

In one participant’s practice, the professional CGM devices are set

up with the hospital, and all patients are referred there, where the

devices are inserted in a dedicated room.

SchedulingA schedule needs to be established to ensure getting the

equipment back and cleaned so that it can be used by the next

patient. In some practices, all professional CGM insertions are

done on Thursday with return of the devices on Monday or

Tuesday. Most participants indicated that the patient physically

returns the CGM device at the end of the 72-h evaluation period,

with the possible exception of some pediatric patients who mail

the devices back because they live far away and do not want to

miss school.

Other practices, particularly those with a higher volume of diabetic

patients on insulin pumps, do professional CGM insertions on

Monday mornings with return on Thursday morning, followed by

new patient insertions on Thursday afternoon with removals on

Monday morning.

Relaying Information to the PatientThe ideal scenario is to share professional CGM results with the

patient face to face. Some practices insist that patients come in for

interpretation, because getting the patient to see the professional

CGM data is very motivating and empowering. In these practices,

physicians may interpret and relay the information to the patient

directly. In other practices, the physician interprets the professional

CGM output and relays the information about suggested therapeutic

changes to the nurse practitioner or diabetes educator to communi-

cate to the patient. Then, the patient comes back a few weeks after

Table 1. Rationales for Initiating Professional Continuous Glucose Monitoring for Cases Presented

PRESENTER DIABETES TYPE GENDER A1c AT BASELINEBASELINETREATMENT

RATIONALE FORPROFESSIONAL CGM

Phillips 1 F 9.6% NPH 10U qam

Glargine 15 U qhs

R insulin SSI premeal

Hypoglycemia unawareness;

gastroparesis

Nardacci 2 F 10.2% Aspart 4–6 U w=meals

Glargine 14 U qhs

Highly variable BG on basal

bolus therapy

Arnold 1 M 5.9% Glargine 15 U qhs

Aspart BS-100

Avid cyclist considering pump

Arnold 2 F 5.7% NPH 14–20U

Lispro BS-80

Pregnant; 20 weeks gestation

Arnold 1 F 7.3% MDT Pardigm 722 pump Pregnant with triplets;

hypoglycemic seizure secondary

to insulin stacking

Horowitz 2 F 7.5% Aspart pre meals

Glargine 50 U qhs

Good premeal sugars but high

A1c. Postprandial hyperglycemia

Horowitz 1 F 7.1% Aspart pre meals

Glargine 15 U qam

18-year-old college student

with variable BG

Odugbesan 2 F 8.4% Pioglitazone 30mg qd

Exenatide 5 mg bid

Lispro PRN

Nonadherence

Reddy 1 and 2 M 9.0% Insulin pump High A1c

Bode 1 F 9.2% Aspart via CSII Hyperglycemia 24=7

Dawn phenomenon

BG, blood glucose; CGM, continuous glucose monitoring.

DISCUSSION REGARDING USE OF PROFESSIONAL CGM

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 33

Page 34: Cgm case studies

those changes are implemented. At that visit, the physician shows the

patient the professional CGM output and explains the rationale for

the changes. Time slots for relaying professional CGM results should

be kept open on the schedule so that patients can get their results in a

timely manner.

In another practice, 90% of the interpretation is done by a certified

diabetes educator (who is a licensed nurse practitioner) who goes over

the results and the recommended changes directly with the patients,

with physician oversight after the fact.

In reality, given both patients’ schedules, insurance plans with very

high copays for office visits, and how far in advance one’s practice is

booked with new patient visits, it can be difficult to arrange face-to-

face follow-up visits on a rapid basis after the professional CGM

evaluation. In some practices, a licensed staff member calls the patient

to relay the initial results, and then actually gives the patient a copy of

the data at the next visit.

EquipmentThe practice should have available enough equipment for the

patient load and some backup devices in case a patient is late for their

appointment for the download when another patient is awaiting an

installation.

Roundtable participants agreed that, on average, it takes 10–

15 min to set up each patient with professional CGM device. All

participants also emphasize that each patient keeps a diary that re-

cords what and when they are eating, taking medications, and par-

ticipating in physical activities.

For the patients who are on pumps, plan to download the pump

data simultaneously with the professional CGM device.

One panelist described a colleague who insists on using a personal

CGM device as a professional CGM device, and I have had multiple

discussions with him. This physician gives as his rationale for doing

this that the personal CGM is approved for a 7-day use, and you need

7 days to see what is going on with the patient’s glycemic control

during the week compared with weekends.

General consensus is that it is inappropriate to use a personal CGM

device as a substitute for a professional CGM evaluation for the

following reasons:

. The use of personal CGM device eliminates the blinded nature of a

professional CGM, which is key to developing those ‘‘aha!’’ moments

with the patient when going over the professional CGM results.. When scheduled appropriately, 3 days of data are adequate for

evaluating glucose control over time.* For patients who have markedly different schedules for

weekends and weekdays, the provider can start the patient

with the professional CGM device on a Thursday and thus

catch both a weekday and a weekend day during the evalu-

ation period. A personal CGM device requires more training for the patient to

be able to use it safely. The patient must perform SMBG before

making a therapeutic decision and then you have to teach them

also how to calibrate and that you got to calibrate it appro-

priately.. To get a complete picture of glycemic patterns within a 3-day

period, the provider may train the patient to perform overnight

basal rate testing the first day only, because the overnight

testing is a lot to ask day in and day out. On subsequent 2 days,

the patient might skip breakfast 1 day, and skip lunch or dinner

another day.

Billing and ReimbursementReimbursement policies must be established before initiating the

professional CGM evaluation, to ensure that the procedure will be

reimbursed. Medicare reimburses professional CGM for insulin-

requiring diabetic patients. In some cases, Medicare will also reim-

burse for patients who are not on insulin if you can demonstrate that

they are not in control. Non-Medicare health plans have differing

policies governing professional CGM reimbursement. In addition,

Medicare and some private insurers will reimburse for two to four

yearly professional CGM evaluations.

The lead staff member in charge of professional CGM must contact

the local insurance providers and obtain their procedures and referral

forms for professional CGM authorization. Not all insurance

Table 2. Professional Continuous Glucose MonitoringCandidates

Diabetic patients unable to achieve an optimal, safe, and effective glycemic

goal with self-monitoring of blood glucose

Diabetic patients with hemoglobin A1c within target range, but with highly

variable blood glucose

Patients with repeated episodes of hypoglycemia or hypoglycemic unawareness

Insulin-requiring patients

Diabetes in pregnancy

Table 3. The Value of Retrospective Data with ProfessionalContinuous Glucose Monitoring

Minimal time required for training and patient start-up

No user interface allows for added simplicity of use

Unaltered glucose patterns without intervention or alerts affecting the

glucose levels

Levels help guide appropriate therapy adjustments

Observe the impact of patient’s diet, exercise, behavior, and medications on

their glucose levels

Direct attachment to the body can reduce the possibility of missed data

Not all patients are ready to understand and respond to real-time glucose

data

DISCUSSION REGARDING USE OF PROFESSIONAL CGM

34 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 35: Cgm case studies

companies are familiar with requests for professional CGM, and

rationales for performing the procedure must be filled out as re-

quired. The forms must document exactly how the patient meets the

criteria established by each insurer’s plan. In particular, diagnoses of

uncontrolled diabetes or other rationales must be coded accurately.

The device’s manufacturer may provide a resource guide for practices

that want to utilize the technology.

There was general agreement about coding for reimbursement for

professional CGM. Pre-CGM evaluation can be billed using codes

99212–99215. The procedure itself, including placement of the

sensor, removal of the sensor, and downloading the data, is coded as

professional CGM 95250. The initial visit during which the device is

inserted is just a visit and not a charge. The 90250 charge is done at

the time of the download.

The interpretation of both personal CGM and professional CGM is

coded as 95251, billed on the day that you interpret it. This latter

charge is billable whether the interpretation is shared with the patient

by telephone or in person, so long as the interpretation of the

download with recommendation of therapeutic adjustments is done

by a licensed practitioner (e.g., Medical Doctor, Doctor of Osteopathy,

Nurse Practitioner, or Physician Assistant).

Within the practice, a system of checks and balances are needed to

ensure that all patients who were scheduled for professional CGM

received the equipment and returned it, and that all services provided

are billed for at the end of every week. In one practice, the billing

department meets quarterly with the healthcare provider to discuss

the implementation of the professional CGM service.

Consensus on Follow-Up ProfessionalCGM Recommendations

Participants agreed that repeating a professional CGM evaluation

is important for two main reasons: (1) the patient and healthcare

provider get to evaluate the success of therapeutic changes guided by

the first evaluation; (2) glycemic control is an evolving story for

many diabetics, and regular evaluations are necessary to ensure that

patients continue to integrate data into their behaviors that govern

glycemic control over time.

Several participants indicated that, although follow-up profes-

sional CGM evaluation is clinically important, and is covered by

Medicare, in reality, repeat professional CGM is underutilized. One

participant’s practice is just starting to incorporate follow-up CGM

evaluations within 3 or 6 months to ensure that the therapeutic

changes that have been made actually improved the BG pattern

overall, and not just the A1c.

The roundtable moderator just finished participating in a trial in

which eligible subjects had type 2 diabetes treated with basal in-

sulin. Subjects were placed on an insulin pump and professional

CGM was used to optimally titrate their basal and bolus insulin so

they had no hypoglycemia. Subjects came back weekly for evalu-

ation of the sensor data, with alteration to the basal and bolus

insulin as needed to get everyone to goal. Every subject’s A1c

dropped an average of 1.5–2.0 points no matter what regimen they

were on.

Consensus on Interpreting ProfessionalCGM Outputs

Participants described the systematic ways in which they interpret

output from the professional CGM device. The basic concept is to go

from a big picture to a small picture and stop at the level at which you

have enough information.

One participant starts with the pie charts of the Sensor Summary,

and the average glucose on the meter as well as the sensor, to make

sure it is an adequate study. Second, he checks the Sensor Modal Day

output for trends in highs and lows. Then, the Sensor Modal Time

chart gives a clear sense of what happened overnight, and before and

after meals. Then, he compares the Sensor Modal Day output with the

diary listings day by day, and adds handwritten notes. Thus, he might

observe an uptrend in BG on a day that the patient ate rice and

beans and did not bolus enough. With practice, this level of inter-

pretation can take as little as 5 minutes. He summarizes the data on a

preprinted sheet to indicate that if it is an adequate study, if there is

any hyperglycemia or hypoglycemia overnight or at meals, what

changes need to be made, and a place to sign off. The sheet goes to the

nurse, who calls the patient and relays the recommendations for

changes.

Another practitioner also writes notes on the output and gives

patients a copy. The original goes in the chart in a section clearly

labeled sensor downloads so that it can be retrieved easily as

needed.

Whether the discussion of professional CGM results is done by the

physician or another licensed member of the team, it should include

showing the patient what certain foods, activities, or medication

dosing do to the BG pattern. This discussion will help the patient

understand how the clinician can identify instances of overtreating,

or fear of hypoglycemia and the defensive behavior in eating patients

engage in to prevent hypoglycemia overnight.

Consensus on Therapy Adjustments Guidedby Professional CGM

Overall, participants agreed that the three main adjustments that

might be triggered by professional CGM results are the basal insulin

dose, dosing before meals, and dosing after meals. Patients may see

how underutilizing insulin for meals, and improper use of the car-

bohydrate ratio really affect their BG.

If the patient does not wake up with a normal fasting glucose, the

clinician needs to optimize the basal insulin delivery. The most

common changes are more meal-related changes. Some patients need

changes in insulin dose to cover meals better because the basal dose is

too high. They are having hypoglycemia before the next meal. Some

patients go to bed high and then wake up normal and do a lot of

defensive eating at night.

Another important interpretation is to identify when during

the night a patient becomes hypoglycemic, to adjust the insulin dosage

to prevent further episodes of nocturnal hypoglycemia. The glucose

patterns after a meal will govern establishment of the appropriate

doses of preprandial insulin to cover the carbohydrate load for meals.

DISCUSSION REGARDING USE OF PROFESSIONAL CGM

CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING 35

Page 36: Cgm case studies

For type 2 patients with elevated A1cs, one participant may

change a patient’s oral agent or change their insulin regimen when

she finds hyperglycemia where she least expects it based on output

from professional CGM.

It is important to be cautious about making too many dietary or

medication changes at one visit to ensure that you can identify what

change made the difference. If several changes are appropriate, one

participant recommended choosing the change the patient is willing

to do first.

Expert Recommendations on the Futureof Professional CGM

During this discussion, roundtable participants brainstormed

about how professional CGM might improve over time. The ideal

system would not require the patient to test glucose with a finger-

stick. Second, when the CGM sensor detects hypoglycemia, a device

that shuts the pump off when the patient does not respond to the

alarm need to be approved in the United States. Such a device is

already approved in Europe.

One of the problems with miniaturizing components is that there

is a tradeoff based on physics—to get things really small, the bat-

tery life will be shorter. One participant was frustrated because

his professional CGM equipment failed after only 14 months.

He suggested that if a longer or unlimited battery life cannot be

built, then the battery needs to be rechargeable, because primary

care doctors and endocrinologists will not embrace the technology

if they feel like they have to replace expensive devices every

2 years.

Ideally, professional CGM would evolve into a handheld device

into which the patient could input insulin dosing that would collect

and display meal markers to replace the handwritten diary. Perhaps

the device could even take and upload pictures of the food being

eaten. Ideally, this future device would use artificial intelligence to

analyze some of those patterns and highlight them—sort of like an

EKG interpretation. The system would allow the clinician to agree or

disagree with the automatically generated interpretation, and the

output would display the diary overlaid on the BG tracings in a one-

or two-page format and permit the clinician to make therapeutic

changes as appropriate.

The devices need to be connected to (1) the patients’ significant

others, (2) the doctor’s office, or (3) a central monitoring station. This

would be particularly useful for those elderly patients who live alone

and value their independence. The device would predict BG trends

and notify patients of severe hypoglycemia and hyperglycemia

before they happen.

The ultimate improvement, of course, will be the closed loop

system with a sensor and pump working together.

ConclusionsThe 10 varied cases presented at this roundtable discussion provided

an excellent tutorial in how to interpret data generated by professional

CGM evaluations. This technology can benefit a wide range of diabetic

patients. Healthcare providers must educate third-party payers that

there is more to glucose control than just an A1C value. Professional

and personal CGM both have an important role in guiding therapy

adjustments to optimize our patient’s overall BG control.

DISCUSSION REGARDING USE OF PROFESSIONAL CGM

36 CURRENT TRENDS IN PROFESSIONAL CONTINUOUS GLUCOSE MONITORING

Page 37: Cgm case studies
Page 38: Cgm case studies
Page 39: Cgm case studies

Continuous Glucose Monitoring in Non–Insulin-UsingIndividuals with Type 2 Diabetes: Acceptability,

Feasibility, and Teaching Opportunities

Nancy A. Allen, Ph.D.,1 James A. Fain, Ph.D.,2 Barry Braun, Ph.D.,3 and Stuart R. Chipkin, M.D.3

Abstract

Background: Continuous glucose monitoring (CGM) has the potential to provide useful data for behavioral in-terventions targeting non–insulin-using, sedentary individuals with type 2 diabetes mellitus (T2DM). The aimsof this study were to describe CGM in terms of (1) feasibility and acceptability and (2) dietary- and exercise-teaching events.Methods: Cross-sectional data were analyzed from 27 non–insulin-using adults with T2DM who wore CGMfor 72 h as part of a larger study on using CGM for exercise counseling in this population. Feasibility data in-cluded accuracy of entering daily self-monitored blood glucose (SMBG) readings and events (e.g., meals, ex-ercise), sensor failures, alarms, optimal accuracy of glucose data, and download failures. Acceptability dataincluded CGM satisfaction and wearing difficulties. Dietary- and exercise-teaching events were identified fromCGM and activity monitor data.Results: CGM graphs showed 141 dietary- and 71 exercise-teaching events. About half the participants (52%)reported difficulty remembering to enter events into CGM monitors, but most (82%) kept an accurate paperlog of events. Insufficient SMBG entries resulted in 32CGM graphs with ‘‘use clinical judgment’’ warnings.Eighty-three percent of missed SMBG entries were from 18 participants 55–77 years old. Missing correlationcoefficients resulted from glucose concentrations varying <100mg=dL. A majority of participants (n¼ 19) werewilling to wear CGM again despite reporting minor discomfort at sensor site and with wearing the monitor.Conclusions: CGM data provided several teaching opportunities in non–insulin-using adults with T2DM.Overall, CGM was acceptable and feasible. Some identified problems may be eliminated by newer technology.

Introduction

Continuous glucose monitoring (CGM) technologyhas the potential to change approaches to educating in-

dividuals with diabetes. Since the first CGM device was ap-proved by the U.S. Food and Drug Administration in 1999,1

other models have been developed and distributed,2,3 withimproved accuracy of glucose sensors.4 These devices providedifferent types of CGM data, retrospective and real-time, forcounseling individuals with diabetes.5–9 The increasing im-portance of CGM technology in diabetes health care is re-flected by the term ‘‘continuous glucose monitoring’’ as thetopic of 26 research presentations and three symposia at the2007 American Diabetes Association’s 67th Scientific Sessionand by this search term in Medline retrieving an increasingnumber of articles (10 articles in 1998–1999, 146 articles in

2006–2007). However, few studies have addressed how cli-nicians can use this technology to counsel non–insulin-usingindividuals with type 2 diabetes mellitus (T2DM) to changebehaviors and to improve diabetes self-management skills.

Although technology-related interventions might changebehaviors and improve health-related outcomes, feasibilitystudies are necessary before advancing to costly clinical trials.To determine the feasibility of using CGM in a larger, ran-domized control pilot study to change lifestyle behaviors inindividuals with T2DM,10 we first conducted a preliminaryfocus group study with nine non–insulin-using individualswith T2DM who wore CGM.11 The results of that study wereused to develop the feasibility and acceptability measures forthis cross-sectional pilot study with 27 sedentary, non–insu-lin-using individuals with T2DM.10 Data from those 27 indi-viduals were examined in the present study to determine (1)

1Yale University, New Haven, Connecticut.2University of Massachusetts Dartmouth, Dartmouth, Massachusetts.3University of Massachusetts Amherst, Amherst, Massachusetts.

DIABETES TECHNOLOGY & THERAPEUTICSVolume 11, Number 3, 2009ª Mary Ann Liebert, Inc.DOI: 10.1089=dia.2008.0053

41

Page 40: Cgm case studies

feasibility and acceptability of CGM and (2) uses of CGMdatato provide dietary and exercise education to non–insulin-us-ing individuals with T2DM.

Research Design and Methods

This study examined cross-sectional data from non–insu-lin-using individuals with T2DMwhowore CGM (MedtronicMiniMed, Northridge, CA) for 72 h as part of a larger pilotstudy (n¼ 52).10 Twenty-five participants in the larger studywere part of a control group that did not wear CGM. Therewere no significant differences between the groups at baseline.

Sample and setting

Participants were recruited from two health systems inWestern Massachusetts. Inclusion criteria were (1) knownhistory of T2DM, (2) >18 years old, (3) not exercising morethan 2 days per week, (4) hemoglobin A1c >7.5%, (5) notreceiving insulin, and (6) able to read and speak English.Exclusion criteria were (1) inability to walk 0.25 miles in10min, (2) taking glucocorticoids, and (3) failing prescreeningevaluation (e.g., ischemic heart disease, systolic blood pres-sure >200mm Hg, diastolic blood pressure >110mm Hg,dyspnea on exertion).

Written informed consent was obtained from participantsin accordance with study protocols and institutional reviewboards at study sites.10 Study data were obtained from 27participants who wore CGM in the intervention group of alarger study (n¼ 52).10

Demographic information

Demographic data included gender, race, ethnicity, maritalstatus, education, age, and duration of diabetes. Participantsalso provided information on current diabetes medicationsand smoking history.

Glucose levels

Glucose levels were monitored two ways: (1) continuouslyfor 72 h by the Medtronic CGM device and (2) at least threetimes per day by self-monitored blood glucose (SMBG)readings. The CGM device has four components: pager-sizedglucose monitor, disposable subcutaneous glucose-sensingdevice with an external electrical connector, connecting cable,and communication device to download data from the mon-itor to a personal computer.12 Signals from the sensor are sentevery 10 s to a glucose monitor, where they are averaged andstored every 5min. The monitor calibrates sensor readingsagainst the wearer’s three or four daily required SMBGreadings entered into the CGM device. Information from theCGM device is not available to the wearer but must bedownloaded at the end of 72 h by a clinician to a personalcomputer. CGM software produces daily glucose trend plots,a summary table of average glucose levels, glucose ranges,and standard deviations. Daily and modal color graphs arealso produced with glucose values and markers for meals,exercise, and medication events, visually showing the inter-action among these parameters. Participants were instructedto keep awritten log of events (e.g., SMBG,meals, exercise) ona standardized worksheet. Glucose values obtained withCGM correlate with plasma glucose concentrations measuredin the laboratory13 and at home.12

Physical activity

The amount and intensity of physical activity were objec-tively measured by an ActiGraph (Pensacola, FL) accelerom-eter. This small (5.1-�3.8-�1.5-cm) device was secured by anelastic strap at each participant’s right waist. Monitors wereprogrammed to collect data every minute over 7 days. Thesedata were downloaded into ActiGraph software (DOSRIU256K.EXE, version 2.27) for analysis. The cut points ofFreedson et al.14 were used to determine sedentary (<499counts), light activity (500–1,951 counts), moderate activity(1,952–5,724 counts), and vigorous activity (�5,725 counts).

Feasibility measures

The variables used to assess CGM feasibility, as developedin a preliminary focus group study,11 were (1) accuracy ofparticipant’s CGM input, (2) sensor failures (i.e., signal<10 or>200; initialization signal varies randomly), (3) alarm data, (4)optimal accuracy of glucose data, and (5) data downloadfailures (e.g., lost data, gaps in graphs). Accuracy of CGMinput refers to participant-entered meals, exercise, and med-ication. Missed meal entries were identified by a rise in glu-cose levels without an event marked on the CGM graph andwere recognized by participants as a meal on the paper log orduring review of CGM data with the researcher. Missed ex-ercise entries were identified by a decrease in glucose levelwithout an event marker following increased activity mea-sured by activity monitors or acknowledged by participantsduring review with the researcher. Missed medication entrieswere identified by reviewing CGM graphs for medicationentries and comparing to participants’ medication list.

Acceptability measures

Acceptability was assessed by six questions developedfrom a preliminary focus group study.11 Four of these ques-tions addressed issues related towearing the CGM sensor andmonitor, one addressed participant satisfaction with theCGM, and one addressed understandability of the CGMgraphs (Table 1).

Dietary- and exercise-teaching events

A dietary-teaching event was defined as a glucose excur-sion (a peak change in glucose level of >20mg=dL) in re-sponse to ameal and=or twomeals with glucose excursions ofdiffering magnitudes (in mg=dL). Similarly, an exercise-teaching event was defined as a decline in glucose levels fol-lowing a bout of self-reported exercise or an exercise eventmarked on the CGM graph. An exercise-teaching event alsoincluded increases in glucose levels following sedentary be-havior. CGM graphs were reviewed for teachable dietary andexercise events based on participants’ meals and exercise fromentered CGM meter events, written log, participants’ reportduring counseling, and=or comparison to activity monitordata. For each participant, the number of teaching events wascounted for each day the CGM device was worn.

Body mass index

Body mass index was calculated as weight (kg)=height(m2). Weight was measured to the nearest 0.1 kg using adesignated standing scale in each clinic. participants were

42 ALLEN ET AL.

Page 41: Cgm case studies

asked to wear light indoor clothing and to remove shoes be-fore being weighed. Height was measured to the nearest0.5 cm.

Hemoglobin A1c levels

Hemoglobin A1c levels were drawn and assayed byhigh-pressure liquid chromatography (Variant instrument,Bio-Rad, Hercules, CA) according to standard clinical meth-ods.

Procedures

After participants provided consent, they were assessed atbaseline for (1) demographic data, (2) medication history, (3)hemoglobin A1c, and (4) body mass index. Participants werenext instructed on wearing the CGM device, entering data,entering events, and using a log to record SMBG data, meals,exercise, and other events. The CGM device was inserted and

worn for 72 h. Participants removed the CGM device at homeand brought it to the clinic the following week. Data weredownloaded at that appointment and reviewed individuallywith each participant. The activity monitor was simulta-neously worn during the same 72h and for an additional 4days after removing the CGM. Data from activity monitorswere not reviewedwith participants but were used to identifyamounts and duration of exercise in relation to glucose ex-cursions.

Statistical analysis

Frequency distributions and appropriate summary statis-tics for central tendency and variability were used to describedemographic and clinical data using SPSS version 15 (SPSS,Chicago, IL). Descriptive statistics were used to analyze CGMfeasibility data, acceptability data, and teachable events (di-etary- and exercise-related glucose changes).

Table 1. CGM Acceptability Data

CGM evaluation question Frequency (%) (n¼ 21)

1. What issues, if any, did you have with the CGM?a. Skin irritation 4 (19)b. Pain at sensor site day 1 1 (4.8)c. Pain at sensor site continuously 0d. Discomfort at sensor site day 1 2 (9.5)e. Discomfort at sensor site continuously 2 (9.5)f. Discomfort with sensor location 0g. Discomfort due to monitor location 0h. Infection at sensor site 0i. Limited my activities 2 (9.5)j. Remembering to enter blood sugars, meals, exercise

into the monitor11 (52.4)

l. Alarms 6 (28.6)m. Difficulty understanding direction 2 (9.5)

2. How much difficulty did you experience whenshowering with the CGM?a. None 11 (55)b. Small 4 (20)c. Moderate 3 (15)d. Large 2 (10)

3. How much difficulty did you experience whensleeping with the CGM?a. None 17 (81)b. Small 2 (9.5)c. Moderate 2 (9.5)d. Large 0

4. How much difficulty did you experience whilewearing the CGM during the daytime?a. None 20 (95.2)b. Small 1 (4.8)c. Moderate 0d. Large 0

5. Would you wear the CGM again?a. Yes 18 (85.7)b. No 2 (9.5)c. Don’t know 1. (4.8)

6. Did you experience difficulty in understanding theCGM graph?a. Yes 0b. No 21 (100)c. Don’t know 0

CGM ACCEPTABILITY, FEASIBILITY, AND USES 43

Page 42: Cgm case studies

Results

Participants

Most participants were female, white, and obese, with amean age of 57 years, and spent the majority of their timeengaging in light-intensity activity (Table 2). On average,participants had an 8-year history of diabetes, partial collegeeducation, and suboptimal glycemic control. The majority ofparticipants were taking a sulfonylurea (n¼ 18) and metfor-min (n¼ 17), while only six participants were taking a glita-zone. No participants were taking an alpha-glucosidaseinhibitor or meglitinide analog.

Feasibility

Events were most accurately entered on the first and lastdays of wearing the CGM device (Table 3). On these days, theevents most accurately entered, in decreasing order, wereexercise (70–82%), medications (56–68%), and meals (42–58%). The CGM device was worn for the shortest times on thefirst and last days. Of all events entered on days 2 and 3, mealswere enteredwith the lowest accuracy (26–33%),with exercise(52–59%) and medications (46–58%) generally entered withmoderate accuracy. These data support those from the ac-ceptability follow-up questionnaire showing that 52% ofparticipants had difficulty remembering to enter CGM events.Despite many participants using the event monitor with onlymoderate accuracy, most (81.5%) kept an accurate paper logof events. No sensors failed, but one CGM cable failed.

The CGM has five possible alarms: (1) disconnect, (2) ISIG(initialization signal) out of range, (3) memory full, (4) cali-bration error, and (5) noise. Of the 27CGM files reviewed,

three hadCGM-disconnect alarms. Of these three, two sensorshad been disconnected. One sensor was disconnected becauseof a CGM cable caught on a door, and another for an un-known reason. The third monitor was turned off for an un-known reason. No ISIG out-of-range or memory-full alarmsoccurred. The five calibration-error alarms were caused bymeter glucose readings falling outside the acceptable limitsused to calibrate sensor glucose values. For example, oneparticipant entered three values (245, 229, 209mg=dL) thatrapidly decreased over 15min, causing a calibration alarm.Lastly, two participants had sensor-noise alarms related torapidly rising glucose levels (>400mg=dL).

Optimal accuracy of CGM glucose data was calculated byCGM software from two sources, glucose sensor and glucosemeter data, for each day the sensor was worn.15 Optimal ac-curacy depended on two criteria: (1) correlation betweensensor and meter readings of at least 0.79 and (2) mean ab-solute difference �28%.15 When data from the CGM devicewere downloaded, correlation coefficients were calculatedbetween glucose meter readings and sensor glucose values(paired data) for each day. These paired data were used tocalculate the mean absolute difference, i.e., the difference be-tween the meter and sensor glucose values, divided by themeter value, and averaged across meter–sensor pairs=day.When optimal accuracy criteria were not met or if fewer thanthree meter–sensor pairs were available (required to calculatecorrelation coefficients), a message appeared (‘‘use clinicaljudgment’’) (Table 4).

About half the participants (51.8%) did not enter more thantwo glucose meter readings on days 1 and 4 (Table 4). Thisomission may be partly attributable to the shorter wear timeson those days. In contrast, most participants entered three ormore glucose meter readings on days 2 (85.2%) and 3 (80.7%).Of the 59 missed glucose meter readings, 49 (83%) were from

Table 2. Participant Demographics (n¼ 27)

Characteristic n¼ 27

Age (years) mean� SD 57.0� 15Diabetes’ duration (years) (mean� SD) 8.3� 6Body mass index (kg=m2) (mean� SD) 36.05� 7Hemoglobin A1c (mean� SD) 8.3� 1Activity (min=day)Light=sedentary 1,427� 12Moderate activity 13� 11

Gender (n [%])Female 15 (56)Male 12 (44)

Race (n [%])White 25 (93)African American 2 (7)

Ethnicity (n [%])Not Hispanic or Latino 25 (93)Hispanic or Latino 2 (7)

Marital status (n [%])Single 8 (30)Married 14 (52)Divorced 4 (15)Widowed 1 (4)

Education (n [%])Graduate or professional training 5 (19)College or university graduate 4 (15)Partial college education 10 (37)High school graduate 7 (26)Partial high school education 1 (4)

Table 3. Accuracy of Participant-Entered

Events on the CGM Device

n (%) on day wearing CGM deviceNumber of missedevents identified onCGM tracingsa 1 2 3 4

Meals0 15 (57.7) 7 (25.9) 9 (33.3) 11 (42.3)1 6 (23.1) 6 (22.2) 5 (18.5) 8 (30.8)�2 5 (19.2) 14 (51.8) 13 (48.1) 7 (26.8)

Exercise0 22 (81.5) 16 (59.3) 14 (51.9) 18 (69.2)1 4 (14.8) 9 (33.3) 10 (37) 8 (30.8)2 1 (3.7) 2 (7.4) 3 (11.1)

Medications0 18 (67.7) 15 (57.7) 12 (46.2) 14 (56)1 8 (29.6) 5 (19.2) 7 (26.9) 8 (32)�2 1 (3.7) 6 (23) 7 (26.9) 3 (12)

aMissed meal entries were identified by a rise in glucose levelswithout an event marked on the CGM graph and were recognized asa meal by participants as a meal on the paper log or during review ofCGM data. Missed exercise entries were identified by a decrease inglucose level without an event marker following increased activitymeasured by activity monitors or acknowledged by participantsduring a review. Missed medication entries were identified byreviewing CGM graphs for medication entries and comparing toparticipants’ medication list.

44 ALLEN ET AL.

Page 43: Cgm case studies

18 participants 55–77 years old, and only 17% were from nineparticipants 19–54 years old.

Of the 21CGM reports with calculated correlation coeffi-cients, three failed to meet the criterion of�0.79 (two on day 1and one on day 2) (Table 4). Most participants had missingcorrelation coefficients because their daily glucose levelsvaried <100mg=dL, below the range needed to calculatethese coefficients (Table 3). The mean absolute differencecould not be calculated for two participants on days 1–3 andfor seven participants on day 4 because of insufficient pairedglucose readings. Several CGM graphs (n¼ 32) had ‘‘useclinical judgment’’ messages on days 1 (n¼ 19) and 4 (n¼ 13)because participants did not enter at least three SMBG read-ings. Overall, optimal accuracy criteria were not met by amajority of participants on days 1–4 because their glucoselevels varied �100mg=dL, and they did not enter enoughglucose meter readings on days 1 and 3 (Table 4). Five CGMdaily graphs had gaps due to participants failing to correctlyenter SMBG data or having unpaired meter readings (meterand sensor readings disagreed or monitor was turned off ).

Acceptability

Participants reported minor CGM difficulties: skin irrita-tion (n¼ 4), pain (n¼ 1), or discomfort at sensor site (n¼ 2)and activity limitations (n¼ 2). No infections were observedor reported at CGM sensor sites. Participants reported small(n¼ 5), moderate (n¼ 3), and large (n¼ 2) amounts of diffi-culty with the CGM device while showering. Similarly, par-ticipants reported small (n¼ 3) andmoderate (n¼ 2) difficultysleeping with the monitor. However, the majority reported nodifficulty wearing the CGM (n¼ 20) and answered ‘‘yes’’when asked if they would wear the monitor again (n¼ 19).Only two participants reported difficulty understandingCGM directions, but 11 participants reported difficulty en-tering events such as meals, exercise, and SMBG data. Noparticipants reported difficulty understanding the CGMgraphs.

Dietary- and exercise-teaching events

Over the 72-h CGM period, 77 exercise- and 141 dietary-teaching events occurred (Table 5 and Figs. 1 and 2). Mostexercise-teaching opportunities (66–70%) occurred on days 2and 3, but the majority of participants had dietary-teachableopportunities on all 4 days.

Discussion

Overall, the CGM was reliable, acceptable, and providedmany teaching opportunities. CGM has most frequently beenused to adjust insulin levels in people with type 1 diabetes,16–18 T2DM,19 and during pregnancy.20 However, this study

Table 4. Optimal Accuracy of CGM Sensor Data

n (%) on day wearing CGM device

Optimal accuracy criteriona 1 2 3 4

Paired sensor–meter readings�2 14 (51.8) 4 (14.8) 4 (19.1) 14 (51.9)3 9 (33.3) 6 (22.2) 10 (38.5) 6 (22.2)�4 4 (14.8) 17 (63) 11 (42.2) 7 (25.9)Missing 1

Correlation coefficient<0.79 2 (50) 1 (11.1) 0 0�0.79 2 (50) 8 (88.9) 6 (100) 2 (100)Missing 23 18 21 25

Mean absolute difference�28 24 (96) 24 (96) 25 (100) 20 (100)�29 1 94) 1 (4)Missing 2 2 2 7

Use clinical judgment 19 (70.4) 5 (18.5) 6 (22.2) 13 (48.1)

aThe CGM software calculates the paired sensor-meter readings, correlation coefficient, and meanabsolute difference data. The above categories are displayed on the CGM sensor report for each day thesensor is worn.

Table 5. CGM Teaching Events

n (%) on day wearing CGM deviceNumber of teachingevents on CGMgraphs per daya 1 2 3 4

Exercise0 18 (66.7) 8 (29.6) 9 (33.3) 14 (51.9)1 7 (25.9) 13 (48.1) 13 (48.1) 12 (44.4)2 2 (7.4) 4 (14.8) 4 (14.8) 1 (3.7)3 1 (3.7) 1 (3.7)4 1 (3.7)

Diet0 10 (37) 2 (7.4) 5 (18.5) 8 (29.6)1 15 (55.6) 6 (22.2) 6 (22.2) 12 (44.4)2 1 (3.7) 14 (51.9) 10 (37) 6 (22.2)3 1 (3.7) 5 (18.5) 6 (22.2)4 1 (3.7)

aA dietary-teaching event was defined as a glucose excursion (apeak change in glucose level of> 20mg=dL) in response to a mealand=or two meals with glucose excursions of differing magnitudes(in mg=dL). An exercise-teaching event was defined as (1) a declinein glucose levels following a bout of self-reported exercise or anexercise event marked on the CGM graph or (2) increases in glucoselevels following sedentary behavior.

CGM ACCEPTABILITY, FEASIBILITY, AND USES 45

Page 44: Cgm case studies

identified many opportunities for teaching non–insulin-usingindividuals with T2DM about the influence of diet and exer-cise on glucose levels. Although these participants weregenerally sedentary, several CGM graphs showed decreasedglucose levels after exercise. These observations are consistentwith reports that moderate exercise significantly reducesblood glucose concentration in individuals with T2DM.21,22

One study showed that a single bout of moderate exerciseimproved glycemic levels for at least 24 h in obese individualswith T2DM.22 These data further support using CGM to de-tect changes in glucose levels in response to exercise, thusproviding opportunities for counseling.

Participants’ CGM graphs also showed glucose levelchanges in response to meals, particularly after breakfast orsupper. Similarly, in another study dietary glycemic excur-sions were observed after meals on CGM graphs of individ-uals with T2DM.23 Glucose levels in that study wereexamined 4h after meals (postprandial) and at all other times(interprandial) before and after an 18-day calorie-restricteddiet. Caloric restriction significantly improved interprandialhyperglycemia but did not affect postprandial glucose ex-cursions after breakfast.23 CGM data may provide opportu-nities for developing individualized treatment plans,including the content and timing ofmeals and exercise. Futurestudies are needed to determine if behavior change followingcounseling is evident on a repeat CGM study. Moreover,CGM tracings may reveal that behavior changes are insuffi-cient to control glucose levels, and further research is neededto determine whether CGM studies might be used to counselindividuals with T2DM on the necessity of initiating insulintherapy. Lastly, CGM devices with non-blinded, real-timedisplays might bemore effective in changing diet and exercise

behavior because individuals can immediately see the resultsof their behaviors andmake instantaneous changes. To date, itis unknown whether individual real-time decision-makingversus retrospective counseling is more effective at changingdiet and exercise behaviors in non–insulin-using individualswith T2DM.

Using CGM in older individuals with T2DM raised atechnology-related consideration not found in younger indi-viduals with T2DM. Older participants had difficulty re-membering to enter events such as meals and exercise into theCGM device. However, most participants kept an accuratepaper log of these events, which were easily transferred to theCGM graphs for teaching purposes. To date, the accuracy ofpaper logs versus the accuracy of entering events into theCGM device has not been reported in this population. Mostimportantly, however, all participants reported understand-ing the CGM graphs regardless of their age.

The optimal accuracy of glucose data on CGM reports re-vealed two common problems in participants with T2DM:narrow range of glucose concentrations and insufficientSMBG values entered. Non–insulin-using individuals withT2DM, unlike those with type 1 diabetes, may not have glu-cose levels that vary more than 100mg=dL, as needed to cal-culate correlation coefficients between interstitial and bloodglucose concentrations. Therefore, researchers and clinicianscan expect to see a majority of CGM reports with ‘‘N=A’’ nextto correlation coefficients. Another problem was participantsentering fewer than three SMBG readings into CGM devices,which occurred most frequently on the first and last weardays, resulting in a ‘‘use clinical judgment’’ warning. Olderindividuals with T2DM entered fewer SMBG readings thanyounger participants. A similar problem was reported in an-

7070

140140

0

Time of Day

Meal and long-actingmedication

Glucose levels elevateafter breakfast

3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM

100

200

Glu

cose

- m

g/dL

300

400

Glucose levels decreaseafter physical activity

Meal following activity haslower glucose elevation

Meal and exerciseExercise

70

140

Paired Meter ValueUnpaired Meter Value

Sensor ValueMeal

InsulinExercise

OtherTime Change (From)

Time Change (To)Legend

FIG. 1. CGM teaching events. Day 2 of CGM depicts glucose level decrease following physical activity in a 58-year-oldwhite man: physical activity=exercise ( ), meal (^), long-acting medication ( ), and SMBG ( ).

46 ALLEN ET AL.

Page 45: Cgm case studies

other study of individuals with type 1 and type 2 diabetes butwas resolved by educating participants to enter more dailyglucose values.24 However, older participants were not re-ported to have more difficulty using the CGM device.24 Si-milarly, older adults (66� 6 years) with T2DM experiencemore technological difficulties learning continuous insulininfusion therapy25 than middle-age adults (55� 10 years).26

Interpretability of CGM graphs in the present study was notcompromised, but future studies might consider follow-upphone calls on day 1 to reinforce written instructions anddecrease the number of ‘‘use clinical judgment’’ warnings onthe CGM accuracy report.

Most participants were willing to wear the CGM deviceagain and overall tolerated the procedure well. However,some reported minor skin irritation and discomfort, and onereported pain at the sensor site. This finding echoes a reportthat eight of 70 patients experienced discomfort at CGMsensor sites.25 Although participants should be prepared forpossible discomfort at sensor sites, they can be reassured thatsuch discomfort has generally been transitory and insufficientto deter individuals’ willingness to wear the CGM deviceagain.

Some problems identified in this study will be eliminated bynewer technology, such as the CGMS� iPro (Medtronic). Thisnewer technology uses the same sensor, which is attached to aquarter-sized recorder instead of the cumbersome cable andmonitor of the original CGM. Therefore, many wearing issuesidentified in this study (i.e., showering and sleeping with themonitor) will be eliminated. Furthermore, there are no alarmsfor wearers to manage. At least three or four SMBG readingsper day must still be obtained, but this information is enteredinto software by clinicians=researchers along with events on

the log sheet when the unit is downloaded. Although this newtechnology eliminates someproblems associatedwith the olderCGMdevice, our findings suggest emphasizing towearers thatthey must enter at least three SMBG values every day thesensor is worn and to avoid entering SMBG values when glu-cose levels are rapidly changing. Furthermore, individualswith T2DM and=or a limited range of glucose concentrationwill likely show ‘‘N=A’’ correlation coefficients, which will notaffect interpretability of data.

CGM technology offers many opportunities to counsel in-dividualswith T2DMon strategies to lower glucose levels andimprove self-management behaviors. Such technology offersT2DM patients personalized visual data that may be effectiveat communicating the need to change life-style behaviors.

Acknowledgments

This study was supported by grants F31 NR008818-01A1and T32NR008346-05 from the National Institutes of Health.Medtronic Minimed provided a small equipment grant, andBio-Rad Laboratories provided all A1c assays.We are gratefulto Claire Baldwin for her editorial assistance.

Author Disclosure Statement

No competing financial interests exist.

References

1. Food and Drug Administration: Food and Drug Adminis-tration Summary of Safety and Effectiveness Data for Con-tinuous Glucose Monitoring System (CGMS). http:==fda.gov=cdc (accessed January 11, 2003).

7070

140140

70

140

0

Time of Day

Paired Meter Value

Middle of the night snack-sandwich and muffin

Unpaired Meter Value

3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM

100

200

Glu

cose

- m

g/dL

300

400

Sensor ValueMeal

InsulinExercise

OtherTime Change (From)

Time Change (To)Legend

Highest glucose levelsof the day from highcarbohydrate snack

Glucose levels decreaseafter physical activity

Breakfastand tennis

Candy andfast food

Spaghettiand bread

Glucose levels rise fromhigh carbohydrate

FIG. 2. CGM teaching events: dietary. Day 3 of CGM depicts glucose level increases with high carbohydrate meals andsome lowering of glucose levels following physical activity in a 47-year-old African American woman: physical activi-ty=exercise ( ), meal (^), long-acting medication ( ), and SMBG ( ).

CGM ACCEPTABILITY, FEASIBILITY, AND USES 47

Page 46: Cgm case studies

2. Garg S, Zisser H, Schwartz S, Bailey T, Kaplan R, EllisS, Jovanovic L: Improvement in glycemic excursionswith a transcutaneous, real-time continuous glucose sen-sor: a randomized controlled trial. Diabetes Care 2006;29:44–50.

3. Weinstein RL, Schwartz SL, Brazg RL, Bugler JR, Peyser TA,McGarraugh GV: Accuracy of the 5-day FreeStyle Navigatorcontinuous glucose monitoring system: comparison withfrequent laboratory reference measurements. Diabetes Care2007;30:1125–1130.

4. Weinzimer SA, DeLucia MC, Boland EA, Steffen A, Tam-borlane WV: Analysis of continuous glucose monitoringdata from non-diabetic and diabetic children: a tale of twoalgorithms. Diabetes Technol Ther 2003;5:375–380.

5. Bode B, Gross K, Rikalo N, Schwartz S, Wahl T, Page C,Gross T, Mastrototaro J: Alarms based on real-time sensorglucose values alert patients to hypo- and hyperglycemia:the Guardian continuous monitoring system. DiabetesTechnol Ther 2004;6:105–113.

6. Feldman B, Brazg R, Schwartz S, Weinstein R: A continuousglucose sensor based on wired enzyme technology—resultsfrom a 3-day trial in patients with type 1 diabetes. DiabetesTechnol Ther 2003;5:769–779.

7. Gross TM, Bode BW, Einhorn D, Kayne DM, Reed JH, WhiteNH, Mastrototaro JJ: Performance evaluation of the Mini-Med continuous glucose monitoring system during patienthome use. Diabetes Technol Ther 2000;2:49–56.

8. Maran A, Crepaldi C, Tiengo A, Grassi G, Vitali E, PaganoG, Bistoni S, Calabrese G, Santeusanio F, Leonetti F, RibaudoM, Di Mario U, Annuzzi G, Genovese S, Riccardi G, PrevitiM, Cucinotta D, Giorgino F, Bellomo A, Giorgino R, PosciaA, Varalli M: Continuous subcutaneous glucose monitoringin diabetic patients: a multicenter analysis. Diabetes Care2002;25:347–352.

9. Potts RO, Tamada JA, Tierney MJ: Glucose monitoring byreverse iontophoresis. Diabetes Metab Res Rev 2002;18(Suppl 1):S49–S53.

10. Allen NA, Fain JA, Braun B, Chipkin SR: Continuous glu-cose monitoring improves physical activity behaviors of in-dividuals with with type 2 diabetes: a randomized clinicaltrial. Diabetes Res Clin Pract 2008;80:371–379.

11. Allen NA, Fain JA, Braun B, Chipkin SR: Feasibility andacceptability of continuous glucose monitoring and acceler-ometer technology in exercising individuals with type 2 di-abetes. J Clin Nurs 2009;18:373–383.

12. Gross TM, Mastrototaro JJ: Efficacy and reliability of theContinuous Glucose Monitoring System. Diabetes TechnolTher 2000;2(Suppl 1):S19–S26.

13. Rebrin K, Steil GM, Van Antwerp WP, Mastrototaro JJ:Subcutaneous glucose predicts plasma glucose independentof insulin: implications for continuous monitoring. Am JPhysiol 1999;277:E561-E571.

14. Freedson PS, Melanson E, Sirard J: Calibration of the Com-puter Science and Applications, Inc. accelerometer. Med SciSports Exerc 1998;30:777–781.

15. Mastrototaro JJ: The MiniMed Continuous Glucose Mon-itoring System. Diabetes Technol Ther 2000;2(Suppl 1):S13–S18.

16. Bode BW, Gross TM, Thornton KR, Mastrototoro JM: Con-tinuous glucose monitoring used to adjust diabetes therapy

improves glycosylated hemoglobin: a pilot study. DiabetesRes Clin Pract 1999;46:183–190.

17. Deiss D, Kordonouri O, Hartmann R, Hopfenmuller W,Lupke K, Danne T: Treatment with insulin glargine reducesasymptomatic hypoglycemia detected by continuous sub-cutaneous glucose monitoring in children and adolescentswith type 1 diabetes. Pediatr Diabetes 2007;8:157–162.

18. Kaufman FR, Gibson LC, Halvorson M, Carpenter S, FisherLK, Pitukcheewanont P: A pilot study of the continuousglucose monitoring system: clinical decisions and glycemiccontrol after its use in pediatric type 1 diabetic subjects.Diabetes Care 2001;24:2030–2034.

19. Zick R, Petersen B, Richter M, Haug C, Group SS: Com-parison of continuous blood glucose measurement withconventional documentation of hypoglycemia in patientswith type 2 diabetes on multiple daily insulin injectiontherapy. Diabetes Technol Ther 2007;9:483–492.

20. Murphy HR, Rayman G, Duffield K, Lewis KS, Kelly S, JohalB, Fowler D, Temple RC: Changes in the glycemic profiles ofwomen with type 1 and type 2 diabetes during pregnancy.Diabetes Care 2007;30:2785–2791.

21. Kang J, Kelley DE, Robertson RJ, Goss FL, Suminski RR,Utter AC, Dasilva SG: Substrate utilization and glucoseturnover during exercise of varying intensities in individualswith NIDDM. Med Sci Sports Exerc 1999;31:82–89.

22. MacDonald AL, Philp A, Harrison M, Bone AJ, Watt PW:Monitoring exercise-induced changes in glycemic control intype 2 diabetes. Med Sci Sports Exerc 2006;38:201–207.

23. Colette C, Ginet C, Boegner C, Benichou M, Pham TC,Cristol JP, Monnier L: Dichotomous responses of inter andpostprandial hyperglycaemia to short-term calorie restric-tion in patients with type 2 diabetes. Eur J Clin Investig2005;35:259–264.

24. Chico A, Vidal-Rios P, Subira M, Novials A: The ContinuousGlucose Monitoring System is useful for detecting unrec-ognized hypoglycemias in patients with type 1 and type 2diabetes but is not better than frequent capillary glucosemeasurements for improving metabolic control. DiabetesCare 2003;26:1153–1157.

25. Raskin P, Bode BW, Marks JB, Hirsch IB, Weinstein RL,McGill JB, Peterson GE, Mudaliar SR, Reinhardt RR: Con-tinuous subcutaneous insulin infusion and multiple dailyinjection therapy are equally effective in type 2 diabetes: arandomized, parallel-group, 24-week study. Diabetes Care2003;26:2598–2603.

26. Herman WH, Ilag LL, Johnson SL, Martin CL, Sinding J, AlHarthi A, Plunkett CD, LaPorte FB, Burke R, Brown MB,Halter JB, Raskin P: A clinical trial of continuous subcuta-neous insulin infusion versus multiple daily injections inolder adults with type 2 diabetes. Diabetes Care 2005;28:1568–1573.

Address reprint requests to:Nancy A. Allen, Ph.D.

Yale University100 Church Street South

P.O. Box 9740New Haven, CT 06536-0740

E-mail: [email protected]

48 ALLEN ET AL.

Page 47: Cgm case studies

Sustained Efficacy of Continuous Subcutaneous InsulinInfusion in Type 1 Diabetes Subjects with Recurrent

Non-Severe and Severe Hypoglycemia and HypoglycemiaUnawareness: A Pilot Study

Marga Gimenez, M.D., Merce Lara, B.N., and Ignacio Conget, M.D., Ph.D.

Abstract

Background: This study evaluated the effect of CSII on hypoglycemia awareness and on glucose profile in type 1diabetes (T1D) subjects with repeated non-severe or severe hypoglycemia (NS or SH, respectively).Methods: We included subjects (1) older than 18 years, (2) with T1D duration of >5 years, (3) on multiple dosesof insulin, and (4) without micro- or macrovascular complications and more than four NS events per week (in thelast 8 weeks) and more than two SH events (in the last 2 years). NS/SH episodes and hypoglycemia awarenesswere evaluated. A 72-h continuous glucose monitoring (CGM) was performed before continuous subcutaneousinsulin infusion (CSII). A hypoglycemia-induced test was used to evaluate each patient’s symptoms ineuglycemia/hypoglycemia. Quality of life (QoL) was also evaluated. After 6, 12, and 24 months, all the subjectswere reevaluated.Results: Twenty subjects were included (34.0� 7.5 years old, 12 women, A1c 6.7� 1.1%, 16.2� 6.6 years ofdiabetes’ duration). At baseline, 19 out of 20 subjects displayed hypoglycemia unawareness, which diminishedsignificantly during the follow-up (3 out of 20). NH episodes per week diminished from 5.40� 2.09 at baseline to2.75� 1.74 at the end of the follow-up (P< 0.001). SH episodes fell from 1.25� 0.44 per subject-year to 0.05� 0.22after 24 months (P< 0.001). Hemoglobin A1c remained unaltered. With CGM, the percentage of values within70–180mg/dL increased (53.2� 11.0% to 60.3� 17.1%, P¼ 0.13), and the percentage of values <70mg/dL de-creased (13.7� 9.4% to 9.1� 5.2%, P¼ 0.07), after 24 months. Mean amplitude of glycemic excursions diminishedafter 24 months of CSII (136� 28mg/dL to 115� 19mg/dL; P< 0.02). An improvement in all the aspects of QoLwas observed. The basal alteration in symptom response to an induced hypoglycemia improved after 24 monthsof initiating CSII leading to a response indistiguishable from that observed in a control group of subjects withT1D without repeated NH and SH.Conclusions: CSII prevents hypoglycemic episodes, improves hypoglycemia awareness, and ameliorates gly-cemic profile in T1D subjects with repeated NS/SH. Its use is also associated with an improvement in diabetesQoL.

Introduction

Intensive insulin therapy significantly reduces the risk ofcomplications in subjects with type 1 diabetes (T1D) and

represents the standard treatment from the onset of the dis-ease.1 However, this therapy is unfailingly associated with ahigher risk of non-severe and severe hypoglycemia (NS andSH, respectively) episodes.2 Iatrogenic hypoglycemia causesrecurrent morbidity in most people with T1D. Likewise, it isan obstacle to the maintenance of euglycemia over a lifetime

using intensive insulin therapy and thus precludeseuglycemia’s long-term benefits.3,4

Frequent and repeated episodes of hypoglycemia in sub-jects with T1D almost invariably result in a reduced ability/failure to recognize hypoglycemia symptoms and signs atthe physiological normal threshold (*55mg/dL). This syn-drome of hypoglycemia unawareness frequently occurs inT1D, and the lack of warning symptoms puts patients at ahigh risk for SH because they are unable to take measures toprevent it.5

Endocrinology and Diabetes Unit, Institute of Biomedical Investigations August Pi i Sunyer; CIBER of Diabetes and Associated MetabolicDiseases; and Hospital Clınic i Universitari, Barcelona, Spain.

DIABETES TECHNOLOGY & THERAPEUTICSVolume 12, Number 7, 2010ª Mary Ann Liebert, Inc.DOI: 10.1089/dia.2010.0028

49

Page 48: Cgm case studies

In the same way a history of hypoglycemia induces un-awareness, meticulous prevention of it can reverse hypogly-cemia unawareness. Thus, it is essential that intensive insulintherapy for T1D is designed not only to maintain near-normoglycemia, but also to prevent and minimize the burdenof hypoglycemia.6–9 Although such a goal is feasible, andeven including a proper blood glucose monitoring, the use ofindividualized blood glucose targets, and the implementationof specific education programs, there is no consensus onwhich is the best rational plan of insulin therapy.10–12 A veryrecent meta-analysis including recent randomized clinicaltrials found that the use of continuous subcutaneous insulininfusion (CSII) is not associatedwith a significant difference inhypoglycemia risk.13

In this context, the use of continuous glucose monitoring(CGM) systems and the evaluation of hypoglycemia aware-ness could help us to identify these subjects and to decide on asafe approach to optimize the metabolic control for them.14

The aim of our study was to evaluate the effect of CSII onthe frequency of hypoglycemia, hypoglycemia unawareness,and continuous glucose profile characteristics in a group ofT1D subjects with repeated NS and SH.

Patients and Methods

We conducted a prospective study including patientsconsecutively with the following criteria: (1)>18 years old, (2)T1D duration >5 years, (3) on conventional insulin treatmentusing multiple doses of insulin (MDI) including rapid-actinganalogs (lispro or aspart) as prandial insulins and glargine asbasal insulin, and (4) with an absence of micro- or macro-vascular complications and presenting more than four NSevents per week (in the last 8 weeks) and more than two SHevents (in the last 2 years). Absence of microalbuminuria wasassured bymeasuring the 24-h urinary albumin excretion rate(last three samples <20 mg/min). The presence of cardiovas-cular disease was ruled out considering the following: nohistory of cardiovascular disease events, no electrocardio-gram alterations, normal stress echocardiography, and anankle-brachial index >0.9. The initiation of CSII treatmentwas proposed to all subjects following the indications andcriteria of reimbursement from the Catalan National HealthService authorities. Contraindications for CSII were ruled outin all subjects, mainly including inability to perform self-management of an intensive insulin therapy program, evi-dence of poor treatment compliance and failure to attendoutpatient clinics, and evidence of a disabling psychiatricdisorder.15 The study was approved by the Hospital Clınic iUniversitari (Barcelona, Spain) Ethics Committee, andinformed consent was obtained from all the patients. Thestudy has been performed in accordance with the ethicalstandards laid out in an appropriate version of the Declara-tion of Helsinki.

Within 1 month before initiation of CSII, data on age,gender, duration of the disease, body mass index, renalfunction, and hemoglobin A1c (HbA1c) (Menarini Diag-nostici, Florence, Italy) (normal range, 3.5–5.5%; where3.5%¼ 20.2mmol/mol International Federation of ClinicalChemistry¼ 4.0% Diabetes Control and Complications Trialand 5.5%¼ 42.1mmol/mol International Federation of Clin-ical Chemistry¼ 6.0% Diabetes Control and ComplicationsTrial) were recorded. Patients were questioned regarding the

number of hypoglycemic episodes they presented. NS andSH were defined following the American Diabetes Associa-tion criteria.16 SH events were defined as those associatedwith neuroglycopenia severe enough to require treatmentfrom a third party. The questionnaire of Clarke et al.17 wasused to evaluate hypoglycemia awareness. CGM for 72 husing the CGMS� System Gold� from Medtronic Minimed(Northridge, CA, USA) was recorded within 2 weeks beforeinitiation of CSII in order to describe the glucose profile.Glucose variability was evaluated calculating mean ampli-tude of glucose excursions (MAGE) designed by Serviceet al.18 from continuous sensor readings. MAGE over 24 his the mean of the absolute differences between glucosepeak and nadir values in excess of at least 1 SD of the meanglucose.

Before initiation of CSII a hypoglycemia-induced test wasperformed as described previously.19 Patients answered theHypoglycemia Symptoms Score Questionnaire first after30min of euglycemia (80–120mg/dL) and then after 30min ofbeing in hypoglycemia (45–55mg/dL).20 The test scores be-tween the two states were compared, and the variation wasexpressed in a percentage. The same experimental protocolwas performed in a control group of 20 subjects with T1D andsimilar characteristics (age, gender, disease duration, treat-ment, and absence of micro- or macrovascular complications)but with fewer than four NS events per week (in the last8weeks) and no SH episodes in order to compare the responseto hypoglycemia.

Quality of life (QoL) assessment was performed using twodifferent questionnaires: the Diabetes Quality-of-Life (DQoL)questionnaire, in which higher scores relate to deterioration inQoL, and the SF-12 health survey questionnaire.

All the subjects included in our study received our specifictherapeutic education program for patients beginning CSII.They received a diet adjusted to their age and body massindex, and insulin doses were adjusted to maintain fastingand preprandial glucose levels between 90 and 130mg/dL,postprandial below 180mg/dL, and at bedtime between 100and 180mg/dL, based on four to six daily capillary blooddeterminations. Glucose targets and capillary glucose deter-minations were comparable to those used with MDI. Patientswere encouraged to avoid values<70mg/dL. The same teamsaw patients as required during the therapeutic educationprogram and every 2–3 months thereafter until 24 months offollow-up. Patients were instructed on glucose goals andself-monitoring glucose control when necessary. All patientswere using pumps with preprogrammable variable basalrates. After 6, 12, and 24 months of follow-up, all the subjectswere evaluated for the number of hypoglycemic episodes(NS and SH) and with the questionnaire of Clarke et al.17 Atthe end of the study, results obtained after 72 h of CGM, theHypoglycemia Symptoms Score questionnaire during thehypoglycemia-induced test, and results of the DQoL andSF-12 questionnaires were again obtained.

Results are presented as mean� SD values. Comparisonswere performed using a paired Student’s t test or an analysisof variance for repeated measurements. Comparisons be-tween proportions were made with a w2 test. A value ofP< 0.05 was considered statistically significant. All statisticalcalculations were performed by the Statistical Package forSocial Science (version 14.0) for personal computers (SPSS,Inc., Chicago, IL).

50 GIMENEZ ET AL.

Page 49: Cgm case studies

Results

A total of 20 subjects with NS and SH were included in thestudy, and their clinical and metabolic characteristics atbaseline are shown in Table 1.

At the time of the inclusion in the study, 19 subjects (onewas non-classified) were shown to have hypoglycemia un-awareness according to the Clarke test (score: �4¼unawareness, 3¼non-classified, �2¼ awareness), scoring onaverage 5.45� 1.19. Progressively, we observed a decrease inthe Clarke test score: 3.70� 1.65, 2.74� 1.06, and 1.6� 2.03after 6, 12, and 24 months of follow-up, indicating an im-provement in the hypoglycemia unawareness towards nor-mal awareness (P< 0.001 for baseline vs. 24 months). At theend of the follow-up, only three of the 20 subjects were clas-sified as having hypoglycemia unawareness. In absoluteterms, the evolution of hypoglycemia awareness categories isshown in Figure 1.

The mean number of episodes of NH per week progres-sively diminished from 5.40� 2.09 at baseline to 4.60� 2.33,3.07� 1.39, and 2.75� 1.74 after 6, 12, and 24 months, re-spectively (P< 0.001 for baseline vs. 24 months). When thenumber of SH episodes were analyzed, they fell from

1.25� 0.44 per subject year at baseline to 0.05� 0.22 at the endof the follow-up (P< 0.001). Additionally, HbA1c remainedunaltered during the follow-up: 6.6� 1.0%, 6.7� 0.9%,6.7� 0.8%, and 6.3� 0.9% for baseline, 6, 12 and 24 months,respectively.

Considering data obtained from the CGMS, after the 24-month follow-up, the percentage of valueswithin target levels(70–180mg/dL) increased (53.2� 11.0% to 60.3� 17.1%,P¼ 0.13), and the percentage of values below 70mg/dLdecreased (13.7� 9.4% to 9.1� 5.2%, P¼ 0.07); however, thesetendencies did not reach statistical significance. MAGEdiminished after 24 months of CSII from 136� 28 to115� 19mg/dL at 24 months of follow-up (P< 0.02).

At baseline, subjects with NS and SH scored 31.6� 16.4 onthe Hypoglycemia Symptoms Score Questionnaire duringhypoglycemia, representing a rise of 52% in comparison toeuglycemia (21.0� 3.15). At 24 months after initiating CSII,the score was 62.3� 23.6 (P< 0.001, in comparison to base-line), an increase of 196% (P< 0.001, in comparison to base-line) with respect to euglycemia (21.05� 3.15). These resultswere compared with those obtained in the control group(33.5� 8.7 years old; 12women; 14.0� 6.5 years of duration ofthe disease; all of them on MDI; HbA1c¼ 6.7� 0.7%; differ-ence not significant) under the same conditions of glycemialevels in the euglycemia and hypoglycemia periods. The scoreon the Hypoglycemia Symptoms Score questionnaire was54.5.6� 18.4 during hypoglycemia, representing a rise of163% in comparison to euglycemia (20.5� 1.9). This responsewas not different from that observed in T1D subjects with NSand SH after 24 months of treatment with CSII.

Regarding QoL outcomes, a significant improvement in allthe aspects evaluated by DQoL test was observed. This wasalso the case for the results obtained by the SF-12 health sur-vey questionnaire (Table 2).

Discussion

Our study shows that the use of CSII in T1D subjects with ahistory of recurrent hypoglycemia and SH leads to a persistent

Table 1. Baseline Characteristics of the Study Group

Characteristic Value

Number of subjects 20Age (year) 34.0� 7.5Gender (M/W) 8/12Duration of diabetes (years) 16.2� 6.6BMI (kg/m2) 24.3� 3.1On MDI treatment (%) 100HbA1c (%) 6.7� 1.1Creatinine (mg/dL) 0.8� 0.1UAE (mg/24 h) 6.5� 2.5

Data are mean� SD values. BMI, body mass index; HbA1c,hemoglobin A1c; MDI, multiple daily injections; M/W, men/women; UAE, urinary albumin excretion.

0

5

10

15

20

25

0 6 12 24

Months

No

of

sub

ject

s in

eac

h c

ateg

ory

Unawareness

Non-classified

Awareness

FIG. 1. Number of subjects in each category of the Clarke test during the follow-up.

CSII EFFICACY IN REPEATED-HYPOGLYCEMIA T1D 51

Page 50: Cgm case studies

diminution in number of hypoglycemic episodes, aswell as to asustained improvement in hypoglycemia awareness, eventhough there were no change in HbA1c, and the percentageof glycemic values within target levels and below 70mg/dLremained without significant change.

Hypoglycemia in T1D is the consequence of the non-physiological replacement of insulin even when using thetheoretically physiological basal-bolus approach. Since theDiabetes Control and Complications Trial results, there is nodoubt that intensive insulin therapy effectively delays theonset and slows the progression of diabetic retinopathy, ne-phropathy, and neuropathy in patients with T1D.1,2 However,the price to pay is an increase in NH and SH episodes (two- tothreefold). Mild hypoglycemia, if recurrent, induces un-awareness of hypoglycemia, which impairs glucose counter-regulation and predisposes to SH. A very recent survey of alarge hospital-based population confirmed that there is stilla significant proportion of people with T1D (around 20%)who suffer from hypoglycemia unawareness.21 Thus, despitemodern patient education and improvements in thestrengthening of insulin therapy, hypoglycemia and hypo-glycemia unawareness are still far from solved in T1D.

In patients with repeated hypoglycemia and hypoglycemiaunawareness, the meticulous prevention of hypoglycemicepisodes can reverse the physiological abnormalities associ-ated with this condition.22–24 In order to create more suc-cessful clinical management of T1D, the implementation of amore physiological pattern of insulin replacement therapy,including the use of real-time CGM, as has been demonstratedin some long-term studies, is necessary.25–30 However, there isno consensus on which, if any, should be the preferable,suitable, and efficient approach.

In our study, probably by effectively diminishing thenumber of hypoglycemic episodes, the use of CSII was asso-ciatedwith a shift from abnormal perception of hypoglycemiato a normal awareness. This is true, not only in experimentalconditions of a controlled-induced hypoglycemia but also inclinical assessment using specific tools. It should be under-lined that these results were obtained without a deteriorationof glycemic control in terms of HbA1c. Furthermore, the re-sults provided from CGM are in agreement with these find-ings showing a tendency to a diminishing of values ofglycemia <70mg/dL, an increase in the percentage of valueswithin target levels, and a significant improvement in thevariability in glucose profile. In addition to this, it should benoted that all these beneficial effects of CSII come not with a

detrimental effect but with an improvement in all aspects ofquality of life.

Fatourechi et al.13 in a very recent systematic review andmeta-analysis commissioned by the Hypoglycemia TaskForce of The Endocrine Society examined the best availableevidence about the use of CSII and MDI as intensive insulinreplacement therapies and the risk of hypoglycemia. Includ-ing 15 recent randomized control trials the authors concludedthat CSII was associated with a slightly lower HbA1c withouta significant difference in terms of hypoglycemia. Never-theless, the investigators recognized that these results camefrom patients at a low risk of hypoglycemia and that thereforethey cannot be fully extrapolated in patients with recurrentSH or hypoglycemia unawareness.

In absolute terms, and acknowledging the differencesbetween both studies, the magnitude of restoration ofhypoglycemia awareness achieved in our study using CSII isquite similar to that observed by Leitao et al.31 using islettransplantation. However, it does not seem clear to us if theimprovement observed in hypoglycemia awareness in thatstudy is really due to a diminishing of number of hypogly-cemic episodes as data on that subject are not described intheir article.

We are well aware of the limitations of our study. Mainly,there is no control group, and it includes a relatively smallnumber of high-risk subjects for hypoglycemia in whom thebeneficial effect of CSII option could be exacerbated. How-ever, prior to the initiation of CSII, all of our patients had beenincluded in our specific diabetes education program for pa-tients receiving conventional intensive insulin treatment andpoor metabolic control with no benefit with regard to hypo-glycemia. Considering the very disabling and labile profile ofour patients and in light of the indications and criteria fromthe Catalan National Health Service authorities and guide-lines, we did not consider maintaining MDI therapy to anyfurther extent. In this topic, very recently the new guidancefrom the National Institute for Health and Clinical Excellenceon CSII for the treatment of diabetes mellitus recommendedthis type of therapy as a treatment option in T1D subjects whoattempt to achieve target HbA1c values with intensive insulintherapy but experience repeated and unpredictable occur-rences of hypoglycemia (www.nice.org/uk/TA57).

In summary, CSII may persistently prevent hypoglycemicepisodes, improve hypoglycemia awareness, and amelioratethe glycemic profile in T1D subjects with repeated SH.Moreover, its use is associated with an improvement in dia-betes QoL aspects.

Acknowledgments

We are indebted to all of those involved at any time in thespecific therapeutic education program for patients beginningCSII at the Endocrinology and Diabetes Unit of the HospitalClınic i Universitari of Barcelona (colloquially called ‘‘Pro-grama Bombas’’). M.G. is the recipient of a grant from theHospital Clınic i Universitari of Barcelona. This work wassupported in part by a grant (PI060250) from the ‘‘Ministeriode Sanidad y Consumo’’ of Spain. Medtronic Iberica spon-sored this work, in part.

Author Disclosure Statement

No competing financial interests exist.

Table 2. Quality of Life Outcomes After 24 Months

of Using Continuous Subcutaneous Insulin Infusion:

Scores at Baseline and After 24 Months of Follow-Up

Baseline 24 months P

DQoL questionnaireSatisfaction 36.0� 6.4 28.8� 5.5 <0.001Impact of treatment 33.6� 7.5 27.4� 6.0 <0.002Social/vocational worrying 13.3� 4.1 11.5� 3.8 <0.05Diabetes-related issuesworrying

10.1� 2.6 8.0� 1.9 <0.01

SF-12 health surveyquestionnaire

34.1� 3.9 37.0� 2.9 <0.01

DQoL, Diabetes Quality of Life.

52 GIMENEZ ET AL.

Page 51: Cgm case studies

References

1. Nathan DM: Management of diabetes mellitus after theDCCT—what’s next? West J Med 1995;162:469–470.

2. Epidemiology of severe hypoglycemia in the DiabetesControl and Complications Trial. The DCCT ResearchGroup. Am J Med 1991;90:450–459.

3. Cryer PE: Hypoglycaemia: the limiting factor in the gly-caemic management of Type I and Type II diabetes. Diabe-tologia 2002;45:937–948.

4. Cryer PE, Davis SN, Shamoon H: Hypoglycemia in diabetes.Diabetes Care 2003;26:1902–1912.

5. Bolli GB: Hypoglycaemia unawareness. Diabetes Metab1997;23(Suppl 3):29–35.

6. Bolli GB: Treatment and prevention of hypoglycemia and itsunawareness in type 1 diabetes mellitus. Rev Endocr MetabDisord 2003;4:335–341.

7. Cryer PE: Hypoglycemia risk reduction in type 1 diabetes.Exp Clin Endocrinol Diabetes 2001;109(Suppl 2):S412–S423.

8. Fritsche A, Stumvoll M, Haring HU, Gerich JE: Reversal ofhypoglycemia unawareness in a long-term type 1 diabeticpatient by improvement of beta-adrenergic sensitivity afterprevention of hypoglycemia. J Clin Endocrinol Metab2000;85:523–525.

9. Linkeschova R, Raoul M, Bott U, Berger M, Spraul M:Less severe hypoglycaemia, better metabolic control, andimproved quality of life in Type 1 diabetes mellitus withcontinuous subcutaneous insulin infusion (CSII) therapy; anobservational study of 100 consecutive patients followed fora mean of 2 years. Diabet Med 2002;19:746–751.

10. Bolli GB, Pampanelli S, Porcellati F, Fanelli CG: Recoveryand prevention of hypoglycaemia unawareness in type 1diabetes mellitus. Diabetes Nutr Metab 2002;15:402–409.

11. Cox DJ, Kovatchev B, Koev D, Koeva L, Dachev S,Tcharaktchiev D, Protopopova A, Gonder-Frederick L,Clarke W: Hypoglycemia anticipation, awareness andtreatment training (HAATT) reduces occurrence of severehypoglycemia among adults with type 1 diabetes mellitus.Int J Behav Med 2004;11:212–218.

12. Cox DJ, Gonder-Frederick L, Polonsky W, Schlundt D,Kovatchev B, Clarke W: Blood glucose awareness training(BGAT-2): long-term benefits. Diabetes Care 2001;24:637–642.

13. Fatourechi MM, Kudva YC, Murad MH, Elamin MB, TabiniCC, Montori VM: Clinical review: Hypoglycemia with in-tensive insulin therapy: a systematic review and meta-analyses of randomized trials of continuous subcutaneousinsulin infusion versus multiple daily injections. J Clin En-docrinol Metab 2009;94:729–740.

14. Streja D: Can continuous glucose monitoring provideobjective documentation of hypoglycemia unawareness?Endocr Pract 2005;11:83–90.

15. Gimenez M, Conget I, Jansa M, Vidal M, Chiganer G, Levy I:Efficacy of continuous subcutaneous insulin infusion inType 1 diabetes: a 2-year perspective using the establishedcriteria for funding from a National Health Service. DiabetMed 2007;24:1419–1423.

16. Defining and reporting hypoglycemia in diabetes: a reportfrom the American Diabetes Association Workgroup onHypoglycemia. Diabetes Care 2005;28:1245–1249.

17. Clarke WL, Cox DJ, Gonder-Frederick LA, Julian D, SchlundtD, PolonskyW:Reduced awareness of hypoglycemia in adultswith IDDM. A prospective study of hypoglycemic frequencyand associated symptoms. Diabetes Care 1995;18:517–522.

18. Service FJ, Molnar GD, Rosevear JW, Ackerman E,Gatewood LC, Taylor WF: Mean amplitude of glycemic

excursions, a measure of diabetic instability. Diabetes1970;19:644–655.

19. Ferrer JP, Esmatjes E, Gonzalez-Clemente JM, Goday A, Con-get I, Jimenez W, et al.: Symptomatic and hormonal hypogly-caemic responses to human and porcine insulin in patientswith type I diabetes mellitus. Diabet Med 1992;9:522–527.

20. McAulay V, Deary IJ, Frier BM: Symptoms of hypoglycae-mia in people with diabetes. Diabet Med 2001;18:690–705.

21. Geddes J, Schopman JE, Zammitt NN, Frier BM: Prevalenceof impaired awareness of hypoglycaemia in adults withType 1 diabetes. Diabet Med 2008;25:501–504.

22. Fanelli CG, Porcellati F, Pampanelli S, Bolli GB: Insulintherapy and hypoglycaemia: the size of the problem. Dia-betes Metab Res Rev 2004;20(Suppl 2):S32–S42.

23. Fanelli CG, Epifano L, Rambotti AM, Pampanelli S, DiVincenzo A, Modarelli F, Lepore M, Annibale B, Ciofetta M,Bottini P: Meticulous prevention of hypoglycemia normal-izes the glycemic thresholds and magnitude of most ofneuroendocrine responses to, symptoms of, and cognitivefunction during hypoglycemia in intensively treated patientswith short-term IDDM. Diabetes 1993;42:1683–1689.

24. Hermanns N, Kulzer B, Kubiak T, Krichbaum M, Haak T:The effect of an education programme (HyPOS) to treathypoglycaemia problems in patients with type 1 diabetes.Diabetes Metab Res Rev 2007;23:528–538.

25. Pampanelli S, Fanelli C, Lalli C,CiofettaM, Sindaco PD, LeporeM, Modarelli F, Rambotti AM, Epifano L, Di Vincenzo A,Bartocci L, Annibale B, Brunetti P, Bolli GB: Long-term inten-sive insulin therapy in IDDM: effects onHbA1c, risk for severeand mild hypoglycaemia, status of counterregulation andawareness of hypoglycaemia. Diabetologia 1996;39:677–686.

26. Bott S, Bott U, Berger M, Muhlhauser I: Intensified insulintherapy and the risk of severe hypoglycaemia. Diabetologia1997;40:926–932.

27. Schiaffini R, Ciampalini P, Fierabracci A, Spera S, Borrelli P,Bottazzo GF, Crino A: The Continuous Glucose MonitoringSystem (CGMS) in type 1diabetic children is theway to reducehypoglycemic risk. Diabetes Metab Res Rev 2002;18:324–329.

28. Amiel S: Reversal of unawareness of hypoglycemia. N Engl JMed 1993;329:876–877.

29. Pickup JC, Sutton AJ: Severe hypoglycaemia and glycaemiccontrol in Type 1 diabetes: meta-analysis of multiple dailyinsulin injections compared with continuous subcutaneousinsulin infusion. Diabet Med 2008;25:765–774.

30. Juvenile Diabetes Research Foundation Continuous GlucoseMonitoring Study Group, Tamborlane WV, Beck RW, BodeBW, Buckingham B, Chase HP, Clemons R, Fiallo-Scharer R,Fox LA, Gilliam LK, Hirsch IB, Huang ES, Kollman C,Kowalski AJ, Laffel L, Lawrence JM, Lee J, Mauras N,O’Grady M, Ruedy KJ, Tansey M, Tsalikian E, Weinzimer S,Wilson DM, Wolpert H, Wysocki T, Xing D: Continuousglucose monitoring and intensive treatment of type 1 dia-betes. N Engl J Med 2008;359:1464–1476.

31. Leitao CB, Tharavanij T, Cure P, Pileggi A, Baidal DA, RicordiC, Alejandro R: Restoration of hypoglycemia awareness afterislet transplantation. Diabetes Care 2008;31:2113–2115.

Address correspondence to:Ignacio Conget, M.D., Ph.D.

Endocrinology and Diabetes UnitInstitute of Biomedical Investigations August Pi i Sunyer

Villarroel 17008036 Barcelona, Spain

E-mail: [email protected]

CSII EFFICACY IN REPEATED-HYPOGLYCEMIA T1D 53

Page 52: Cgm case studies