Hypoglycemia and Blood Glucose Fluctuations in the Application of a Sensor-Augmented Insulin Pump

Abstract

Background:

The purpose of this study was to understand the effect of sensor-augmented insulin pump (SAP) use on hypoglycemia and blood glucose (BG) fluctuations.

Subjects and Methods:

Sixty patients with type 2 diabetes mellitus were randomly assigned to three groups of treatment with SAP, continuous subcutaneous insulin infusion (CSII), or multiple daily injection (MDI) therapy for 6 days. Parameters of glycemic control that were determined included mean BG concentration (MBG), SD of BG (SDBG), mean amplitude of glycemic excursions (MAGE), absolute means of daily differences (MODD), 24-h area under the curve at 10 h (AUC10), 24-h area under the curve at 3.9 h (AUC3.9), and Low Blood Glucose Index (LBGI).

Results:

No significant differences were observed among the three groups in terms of MBG, SDBG, MAGE, or MODD at the beginning of treatment. The MBG, SDBG, MAGE, MODD, and total AUC10 of the SAP group improved over the 4 days of the intervention compared with the CSII and MDI groups; however, no significant differences were observed among the three groups in terms of total AUC3.9 and LBGI.

Conclusions:

Compared with CSII and MDI therapy, SAP therapy was able to rapidly lower mean BG and reduce BG level fluctuations with no increased risks of hypoglycemia.

Introduction

In China, up to 9.7% and 15.5% of adults ≥20 years old suffer from diabetes and prediabetes, respectively.¹ Type 2 diabetes mellitus accounts for 97% of all diabetes mellitus. Although diabetes mellitus itself is a major concern, complications involving the vessels and nerves caused by long-term hyperglycemia and the glucose fluctuations are the real significant issues. The renowned Diabetes Complications and Control Trial² and the United Kingdom Prospective Diabetes Study³ showed that glycosylated hemoglobin (HbA1c) was an independent risk factor for chronic diabetes complications. The DECODE⁴ trial and the study of Muggeo et al.⁵ also indicated that complications were related to glucose fluctuations. The fact that the incidence of diabetes complications can be notably reduced by strict blood glucose (BG) control has been proven by evidence-based medicine. However, it may also increase the risk of hypoglycemia and may cause wide fluctuation of BG levels. Thus, the benefits brought about by BG control, to a degree, may be neutralized or even outweighed by hypoglycemia and glucose fluctuations. A new concept, “lower glucose precisely, reach target gradually,”⁶ was put forward internationally and led to a simultaneous reduction in hyperglycemia and hypoglycemia.

Patients with similar glycated hemoglobin levels and mean glucose values can have markedly different daily glucose excursions. The role of this glucose variation in pathophysiological pathways remains under debate. In real-life human studies, including type 1 and type 2 diabetes patients, neither a reproducible relationship with oxidative stress nor a correlation between short-term glucose variation and retinopathy, nephropathy, or neuropathy was observed. On the other hand, long-term glycemic variation could be related to microvascular complications in type 1 and type 2 diabetes. Furthermore, glucose variation may play a role in the prediction of severe hypoglycemia. In our opinion, we lack both the compelling evidence and the means to target glucose variation separately from all efforts to lower the mean glucose level while avoiding hypoglycemia.⁷

Validated by the STAR⁸ trial, sensor-augmented pump (SAP) insulin therapy can enhance the rate of reaching a target HbA1c level while reducing the incidence of hypoglycemia and weight gain. The SWITCH OFF study⁹ indicated that a real-time alarm for hypoglycemia could shorten the duration of hypoglycemia. The Eurythmics trial¹⁰ and RealTrend trial¹¹ showed that SAP therapy was better than continuous subcutaneous insulin infusion (CSII) at controlling HbA1c levels. Unlike previous studies, this study aimed to assess the rate of hypoglycemia present with SAP therapy and the efficacy of SAP therapy in regulating the glucose fluctuations.

Patients and Methods

Patients

Patients with type 2 diabetes, according to the World Health Organization diagnostic criteria, were eligible if they were treated in the General Hospital of Beijing Military Area (Beijing, China). Patients were eligible if they were 30–80 years old, with a 3–20-year history of diabetes, had type 2 diabetes, and had been taking four injections of insulin daily for more than 3 months with an HbA1c level between 7.0% and 12.0%. Patients with severe hypoglycemia (BG<3.9 mmol/L) or hyperglycemia in the morning (fasting BG increased >7–8 mmol/L) but no nocturnal hypoglycemia (or a 3:00 a.m. glucose level of >6 mmol/L) were also enrolled. Patients must have had good cognitive and motor abilities without depression and must have been willing to receive treatments. All patients signed an informed consent and agreed to have diabetes meals customized by the hospital and to perform exercise for 30 min after meals.

Patients with severe complications of diabetic peripheral neuropathy, disorders of nerves that affected the sensation of pain and numbness, were excluded, as were patients with diabetic retinopathy with retinal hemorrhage and patients with diabetic nephropathy coupled with microalbuminuria and proteinuria. Also excluded were patients with acute or chronic complications, tumor, trauma, surgery, burn, pregnancy, severe cardiovascular, hepatic, and kidney diseases, diabetic ketoacidosis, or hyperosmolar nonketotic diabetic coma, accompanied by severe circulatory disorders of hyperglycemia, or those with the factors that could (according to investigators) affect the participation in the trial and the outcome evaluation.

Treatment

Patients were randomly assigned to one of three groups. Patients in the SAP group received SAP therapy with real-time monitoring and timely glucose adjustments. Insulin aspart was used in the pump, and the initial dose of insulin was calculated according to patient weight (in kg). All patients wore a real-time glucose monitoring system (Medtronic, Northridge, CA) for 144 h. Cooperating with patients, according to the real-time level of glucose, the curve of glycemic fluctuations, and trend information (↑↓ on screen), we took timely measurements such as bolus additional insulin, adjustment of the basal rate, or setting the temporary basal rate to smoothly lower the glucose level and avoid hyperglycemia and glycemic fluctuations. We turned on the alarm for hypoglycemia, and the threshold for the alarm was 3.9 mmol/L; when it went off, measurements were taken promptly. The insulin pump provided real-time BG levels in order to track BG concentration patterns and identify hyperglycemia and hypoglycemia. The BG data were stored in the pump so that they could be analyzed and tracked. According to the data provided by CareLink™ Personal software (Medtronic), we were able to comprehensively regulate the basal rate. Furthermore, according to the postprandial BG fluctuation curve, we were able to use the dual wave and square wave of large doses to flexibly control postprandial BG. The diabetes management software provided diabetes education and diet suggestions, along with exercise guidance for patients. According to the real-time results displayed on the insulin pump, the patients could monitor the value and trend of BG by themselves. When the patients found the BG level was higher than the target value or an increasing trend occured, they notified the doctor. The doctor responsible for the entire operation would administer the extra boluses if required throughout the day and night.

Patients in the CSII group received intensive insulin pump treatment. The initial dose of insulin (insulin aspart) was calculated according to patient weight (in kg). Fingerstick glucose levels were monitored at eight different times: before and after breakfast, at lunch and dinner, before sleep, and before dawn (7:00 a.m., 9:00 a.m., 12:00 p.m., 2:00 p.m., 5:00 p.m., 7:00 p.m., 9:00 p.m., and 3:00 a.m.). The basal rate and meal bolus were adjusted according to glucose level every day. The CSII group did not use the Bolus Wizard. All patients wore the Medtronic MiniMed Continuous Glucose Monitoring System (CGMS^®) for 144 h for data collection and analysis of the mean glucose concentration and BG fluctuations during 6 days of therapy.

Patients in the multiple daily injection (MDI) group received four insulin injections per day including three insulin aspart injections and one insulin glargine injection. The initial dose of insulin was calculated according to patient weight (in kg), and fingerstick glucose levels were monitored at the same times as those patients in the CSII group. Basal rate and meal bolus were calculated according to glucose level every day. Similar to patients in the CSII group, all patients wore a Medtronic MiniMed CGMS for 144 h for data collection and analysis of the mean glucose concentration and BG fluctuations during 6 days of therapy.

Patients in the CSII and MDI groups were not informed of the continuous glucose monitoring results until the end of treatment. The basal rate and meal bolus were adjusted according to fingerstick glucose values every day. The physicians made the changes in insulin doses. Compared with the control target, the physicians increased the base rate if the BG value before a meal and retiring for the night was high or reduced the base rate if it was low. When preprandial and postprandial bedtime BG values were compared with the previous value, base rates were increased if 30 mg/dL (1.7 mmol/L) higher or more or decreased if 30 mg/dL (1.7 mmol/L) lower or less. For the same meal, the dose was increased before a meal if postprandial hyperglycemia was >50 mg/dL (2.8 mmol/L) higher, or the dose was decreased with fasting if 50 mg/dL (2.8 mmol/L) lower or more.

Patients in each group uniformly accepted the diet, exercise, and education related to diabetes. The food total calories (in kcal) was calculated as ideal body weight×30–35 kcal heat. Nutritional requirements were calculated in proportion to the daily requirements of each nutrient: 55–65% carbohydrates, 25–30% fat, and <15% protein. The calorie distribution for the meals was as follows: breakfast, one-fifth; lunch, two-fifths; and dinner, two-fifths.

With regard to patient exercise, patients performed brisk walking at 60–120 steps/min according to their physical strength. Patients were asked to exercise three times a day for 20–30 min. The patient heart rate measurement should not have exceeded (170 – age) after exercise.

It was important that patients in this study established a correct and thorough understanding of diabetes and the basics of diabetes management. Patients were required to adopt a healthy lifestyle, with diet, exercise, and medication. Patients were educated on BG monitoring and insulin injection methods. We also proactively educated patients in order to enhance compliance and to help them recognize the dangers of diabetes complications and improve their quality of life.

Outcome measures

Our primary end points were several parameters of glycemic fluctuation. Continuous glucose monitoring (288 times/day) occurred from the beginning of the treatment in each group. Mean BG concentration (MBG) and area under curve (AUC) were obtained from both the CareLink diabetes management software and the CGMS analysis software. The SD of BG (SDBG), mean amplitude of glycemic excursions (MAGE), and absolute means of daily differences (MODD) were calculated using standard software. Fritzsche et al.¹² developed an easy-to-use software based on the Windows interface, which was able to quickly and accurately calculate SD, MAGE, and MODD via data from the CGMS.

The Low BG Index (LBGI) assesses the risk of hypoglycemia through corresponding mathematical treatment by the daily (0:00–24:00) BG values. LBGI¹³ is calculated using the mathematical formula: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland,xspace}\usepackage{amsmath,amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6} \begin{document} \begin{align*}{ \rm SMBG} = \gamma^{ \cdot} \{ [ \ln ( { \rm BG} ) ] ^{ \alpha - \beta} \} , { \rm BG} < 0 \\( \alpha = 1.026 , \quad\beta = 1.861 , \gamma = 1.794 ) \end{align*} \end{document}

or \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland,xspace}\usepackage{amsmath,amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6} \begin{document} \begin{align*}{ \rm SMBG} = 0 , { \rm BG} \geq 0 , { \rm LBGI} = 1 / N \Sigma{ \rm SMBG_K}\end{align*} \end{document}

where SMBG is the self-monitored BG level.

Hypoglycemic episodes are defined as a BG level under 70 mg/dL (3.9 mmol/L) regardless of hypoglycemic symptoms. Major hypoglycemic episodes are defined as a BG level under 50 mg/dL (2.8 mmol/L) with hypoglycemic symptoms.

The dynamic glycemic parameters were calculated using the complete 288 data points from 0:00 to 24:00. In order to calculate the parameters of BG fluctuations, we chose four series of 288 integrated glucose concentrations obtained from 0:00 to 24:00. As the first and last day's data are less than 288, those data were excluded from analysis.

Statistical analysis

The χ² test was used to evaluate the affect of gender. The Kolmogorov–Smirnov test, the Levene homogeneity test of variances, and one-way analysis of variance were used to examine the affect of age, duration of diabetes, height, weight, body mass index, HbA1c, cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting BG, postprandial BG, MBG, BGSD, MAGE, MODD, AUC at 3.9 h (AUC3.9), AUC at 10 h (AUC10), and LBGI. Either the least significant difference or Tamhane statistics was used to compare the difference among each group. The least significant difference test was applied when the homogeneity of variance assumptions were satisfied; otherwise the Tamhane test was used. All of the above statistical procedures were performed with SPSS version 16.0 software (SPSS, Inc., Chicago, IL). All reported P values are two-sided; a P value of <0.05 was considered to indicate statistical significance.

Results

Study recruitment and baseline characteristics

Table 1 shows the details of the study recruitment and the baseline characteristics for patients enrolled in this study.

Table 1.

Baseline Characteristics of the Patients

	SAP	CSII	MDI	P
Male/female	14/6 (70%/30%)	16/4 (80%/20%)	8/12 (40%/60%)	—
Age (years)	51.8±11.4	53.5±12.6	60.1±14.0	0.32
Duration of diabetes (years)	7.9±6.4	9.4±7.7	13.4±10.6	0.34
Height (cm)	170±6.9	173±6.9	165±8.4	0.64
Weight (kg)	76.8±12.7	78.6±8.3	67.8±13.5	0.11
BMI (kg/m²)	26.4±3.6	26.3±2.7	24.4±3.5	0.33
HbA1c (%)	8.64±1.53	9.78±1.06	8.99±1.43	0.18
CHO (mmol/L)	4.21±0.63	4.94±1.10	5.30±1.90	0.16
TG (mmol/L)	2.14±1.60	2.92±1.44	1.87±0.59	0.13
HDL (mmol/L)	1.04±0.18	1.04±0.14	1.04±0.27	0.99
LDL (mmol/L)	3.26±0.89	3.49±0.85	3.73±1.94	0.73
FBG (mmol/L)	9.33±2.96	10.56±2.89	8.92±1.66	0.35
PPG (mmol/L)	16.05±2.57	17.24±3.41	15.21±1.26	0.23

Data are mean±SD values as indicated.

BMI, body mass index; CHO, cholesterol; CSII, continuous subcutaneous insulin infusion; HbA1c, glycosylated hemoglobin; FBG, fasting blood glucose; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; MDI, multiple daily injections; PPG, postprandial blood glucose; SAP, sensor-augmented pump (for insulin); TG, triglycerides.

Observation indexes

Table 2 provides information on the primary observation indexes and the secondary observation indexes. In comparison with CSII therapy and MDI therapy, SAP therapy efficiently lowered MBG in a short time and enhanced the rate of reaching the target. The results in the CSII group were similar to those in the MDI group.

Table 2.

Comparison of the Variation of Mean Glucose Concentration, Blood Glucose Fluctuations, and Insulin Dose

Parameter, day	SAP	CSII	MDI	P
MBG (mmol/L)
Day 1	10.02±1.54	9.88±2.43	10.27±2.23	0.44
Day 2	8.17±1.30^a	9.67±1.12^b	10.00±2.14^b	0.04
Day 3	7.67±0.72^a	8.90±1.67^ab	9.63±1.70^b	0.02
Day 4	7.18±0.75^a	8.38±1.75^ab	9.43±1.80^b	0.01
SDBG (mmol/L)
Day 1	1.79±0.44	2.01±0.88	2.21±0.73	0.41
Day 2	1.60±0.55	2.10±0.79	2.04±0.79	0.26
Day 3	1.43±0.46	2.07±1.02	1.85±0.60	0.16
Day 4	1.12±0.40^a	1.84±0.73^b	2.00±0.91^b	0.03
MAGE (mmol/L)
Day 1	4.68±1.55	4.99±2.61	4.83±1.98	0.95
Day 2	3.59±1.34	4.93±2.01	5.43±2.02	0.09
Day 3	3.49±1.44	5.10±2.62	4.43±1.25	0.18
Day 4	2.97±0.96^a	5.09±2.86^b	5.31±2.25^b	0.04
MODD (mmol/L)
Days 1 and 2	1.97±0.45	2.63±1.51	2.39±0.64	0.32
Days 3 and 4	1.44±0.47^a	2.18±0.81^b	2.37±0.95^b	0.03
AUC10 (mmol/L×h) (4 days)	0.94±1.07^a	3.69±3.02^ab	4.94±4.27^b	0.02
AUC3.9 (mmol/L×h) (4 days)	0.002±0.006	0.019±0.042	0.018±0.038	0.44
LBGI
Day 1	0.28±0.53	0.21±0.48	0.02±0.06	0.36
Day 2	0.16±0.28	0.04±0.06	0.03±0.10	0.20
Day 3	0.11±0.11	0.08±0.12	0.07±0.12	0.68
Day 4	0.13±0.13	0.19±0.48	0.17±0.34	0.91
Basal rate (IU)
Day 1	14.81±5.88	16.24±4.50	17.20±4.86	0.34
Day 2	16.31±6.58	16.71±5.17	18.35±5.12	0.49
Day 3	18.26±8.10	16.91±5.50	19.00±5.35	0.59
Day 4	19.79±8.84	16.99±5.51	19.20±5.40	0.39
Bolus (IU)
Day 1	15.4±5.94	14.40±4.01	18.35±5.68	0.06
Day 2	19.15±6.60	16.40±4.28	20.00±5.74	0.12
Day 3	20.50±8.49	17.25±5.07	19.05±6.76	0.34
Day 4	20.95±8.94	17.10±5.35	20.35±5.56	0.17
Additional bolus^c	4.65±3.53^a	2.70±2.94^b	1.15±1.73^b	0.00
Total insulin (IU)
Day 1	30.23±10.48	30.64±7.58	35.55±10.44	0.16
Day 2	35.46±11.91	33.12±8.37	38.35±10.53	0.29
Day 3	38.76±15.96	34.16±9.42	39.15±10.55	0.37
Day 4	40.74±17.08	34.09±9.80	39.55±10.53	0.23

Data are mean±SD values as indicated.

A significant difference between the two groups (P<0.05). Different letters in each column indicate P<0.05. Different letters between two groups indicate a significant difference between the two groups. The same letter in two groups indicates no significant difference between the two groups.

The average of the additional bolus for all 4 days.

AUC3.9, area under the curve for 3.9 h; AUC10, area under the curve for 10 h; CSII, continuous subcutaneous insulin infusion; MBG, mean blood glucose concentration; MAGE, mean amplitude of glycemic excursions; MDI, multiple daily injections; MODD, absolute means of daily differences; LBGI, Low Blood Glucose Index; SAP, sensor-augmented pump (for insulin); SDBG, SD of blood glucose.

In comparison with CSII therapy and MDI therapy, SAP therapy efficiently reduced SDBG and MAGE more rapidly. CSII therapy and MDI therapy had similar SDBG and MAGE. SAP therapy also lowered MODD, whereas CSII therapy and MDI therapy had similar MODD. SAP therapy also decreased the duration of hyperglycemia (AUC10) compared with CSII therapy and MDI therapy. AUC3.9 and LBGI were not statistically significantly different among the groups. No major hypoglycemic episodes occurred in this study.

There were no significant differences among the three groups in terms of basal insulin dosage, bolus doses, and total insulin dose for the first 4 days. The total insulin dose for the SAP group on the first day was 30 units, and on Day 4 it was 39 units (i.e., a 9 unit or 30% increase of the first day's insulin), whereas for the CSII and MDI groups it rose from 31 to 34 units (10%) and 36 to 40 units (11%), respectively. But, the average of additional boluses on all 4 days had statistical significance among the groups. The SAP group received more additional boluses than the CSII or MDI group. This is because the SAP group reacted to real-time BG levels and trends, allowing additional bolus doses of insulin to be supplemented in a timely manner. Thus, the greater the number of additional boluses, the faster the patient may avoid prolonged hyperglycemia. Insulin pump therapy is required reduce the amount of insulin doses; the SAP and CSII groups required less insulin doses than the MDI group, but there was no statistical significance among the groups.

All patients were under ideal glycemic control by intensive initial insulin therapy. According to the effect of intensive treatment, patient preference, and the actual situation, a method such as SAP can be used to replace the daily subcutaneous insulin injection via needle. Patients who received SAP significantly improved their diet, exercise, and BG control and received benefit from the CareLink Personal software.

Discussion

SAP is a combination of an insulin pump and a continuous glucose monitoring system. A glucose monitoring system is used to monitor the interstitial fluid glucose concentration using the probe and display the monitoring results on the insulin pump in real time. SAP is monitored once every 5 min. This result could reflect the trend of BG fluctuation according to the change in BG level within 20 min. The real-time trend information improves the efficiency of SAP, as this method can identify hypoglycemia or severe hyperglycemia in advance, thus regulating BG before significant variations can occur. It also helps physicians and patients regulate BG thoroughly and consistently. In addition, a BG fluctuation spectrum of 3 h and 24 h appears on the screen of the monitor. This allows for the adjustment of the amount of insulin quickly and safely according to long-term, short-term, and immediate BG data. The Bolus Wizard Calculator can precisely calculate the amount of the insulin correction bolus, remaining active insulin, and the insulin dose needed when eating, so as to achieve the BG control target.

This study compared the effect of hypoglycemia of three different insulin delivery methods, SAP, CSII, and MDI, using the average BG of daily continuous glucose monitoring. SAP therapy was superior to CSII and MDI therapies in evaluating the general glycemic level. Furthermore, the differences in average BG level were statistically significant among the groups from the second day. In addition, the following facts can impact the pace of reaching glycemic targets: the patients' own situation; the time required to understand the patients' insulin sensitivity; the patients' awareness and degree of cooperation; and the anxieties of hypoglycemia. Fortunately, SAPs are able to show the level of glucose in the interstitial fluid, inform patients of the trend of glycemic fluctuations, and promote cooperation between patients and doctors. Real-time monitoring and the ability to alert the patient of hypoglycemia may help physicians and patients to understand the patients' insulin sensitivity rapidly and reduce fears of hypoglycemia. Trials conducted by Cretti et al.¹⁴ and Alford et al.¹⁵ indicated that hyperglycemia can give rise to the decay of β-cell secretion and the aggravation of insulin resistance. If the toxicity of a high glucose level can be controlled timely, the decay of β-cell secretion and aggravating insulin resistance can be reversed.

For the parameters of glycemic fluctuation SDBG, MAGE, MODD, and duration of hyperglycemia AUC10, SAP therapy was superior to CSII and MDI therapies. This study selected the SDBG, MAGE, and MODD as BG fluctuation assessment indicators. SDBG is the SD of the daily values of multipoint BG measurements and the evaluation of the degree of deviation from average BG. Service et al.¹⁶ suggested that a high MAGE was a vital characteristic of glucose instability and that MAGE was more accurate than other indexes of glycemic fluctuation because of the discrete minor fluctuations that are uncovered by it. MAGE is gradually being accepted as a gold standard to evaluate glycemic fluctuations. Compared with sustained hyperglycemia, fluctuations in glucose levels can inhibit β-cell secretory function much more; this can aggravate insulin resistance and may induce failure of β-cell secretory function, leading to a vicious cycle. It has been indicated that the higher the MAGE, the less the insulin secretion.¹⁷ Kohnert et al.¹⁸ evaluated the effects of MAGE on pancreatic β-cell function in the application of a continuous glucose monitoring system in 59 type 2 diabetes patients and found that MAGE was nonlinearly correlated with postprandial pancreatic β-cell function. Compared with diet and medications, SAP therapy is a new and efficient method for regulating glucose that is better at remediating high fluctuations of glucose primarily caused by postprandial hyperglycemia among Chinese people. The SAP group of patients had less fluctuation of postprandial BG. We believe that patients were more compliant because CareLink Personal could help patients regulate the mealtime after insulin injections, eating speed, and the type of meal. The patient can directly observe the effect of food and exercise on BG, thereby changing the diet and exercise habits. Compared with the CSII and MDI groups, the SAP group had good compliance. They regulated the mealtime and meal duration and used the Bolus Wizard Calculator to match the peak of BG after meals with the peak of insulin action, as well as adjust exercise intensity and times in order to keep euglycemic. Although there is short-term improvement, it remains to be seen whether or not patients benefit in the intermediate and long term from such an intervention.

In this study, we selected the LBGI as an assessment indicator of the risk of hypoglycemia. A value for the LBGI predicts the occurrence of severe hypoglycemia, and an LBGI/High BG Index is more sensitive than the raw data (BG awareness training) for a treatment designed to reduce the range of BG fluctuations. There was no significant difference in LBGI among the three groups in our study, which suggested that SAP therapy could lower the glucose level quickly without increasing the incidence of hypoglycemia and could reduce BG fluctuations. Wintergerst et al.¹⁹ found that hypoglycemia and BG fluctuations could increase the incidence of disability and death among patients with severe disease. Cryer et al.²⁰ considered that severe iatrogenic hypoglycemia and its associated cardiovascular events may offset the benefits of the long-term effective control of BG.

In this study, only the SAP treatment demonstrated the ability to quickly resolve hypoglycemia, while at the same time reducing BG fluctuations and not leading to an increased risk of hypoglycemia. With SAP, patients can intuitively understand the regular pattern of BG fluctuations through CareLink Personal management software and correct them using diet and exercise. Furthermore, short-term education will help patients to improve their awareness and cooperation, establish a healthy living pattern, and achieve long-term benefits.

When a patient with poorly controlled type 2 diabetes with early retinopathy changes over to insulin, the BG level may drop rapidly and aggravate the retinopathy.²¹ Patients with retinopathy should be excluded from the SAP approach. A sudden improvement (lowering to normal) in glucose levels in a person whose diabetes has been poorly controlled for some time may cause rapid and often uncontrollable retinopathy. Good diabetes control is essential in the long term, but in the short term it may cause a rapid deterioration in retinopathy. Laser therapy usually stabilizes the condition. When diabetes is well controlled, retinopathy may progress but much slower than in the “average” person with diabetes. In a previous retinopathy progression study, Danne et al.²² found that retinopathy suggested poor glycemic control; thus any patient found to have background retinopathy should be considered to have poor diabetes control until proven otherwise, and appropriate action should be taken.

In conclusion, SAP can assist patients in reaching glycemic targets rapidly, precisely, and safely.

Footnotes

Acknowledgments

We wish to thank Medtronic for the devices, sensors, equipment, and support for the study. Medtronic also provided grant assistance for training and materials for this trial.

Author Disclosure Statement

No competing financial interests exist.

References

Yang

, Lu

, Weng

, Jia

, Ji

, Xiao

, Shan

, Liu

, Tian

, Ji

, Zhu

, Ge

, Lin

, Chen

, Guo

, Zhao

, Li

, Zhou

, Shan

, He J; the China National Diabetes and Metabolic Disorders Study Group: Prevalence of diabetes among men and women in China. N Engl J Med, 2010; 362:1090–1101.

Hirsch

, Brownlee

: Should minimal blood glucose variability become the gold standard of glycemic control?. J Diabetes Complications, 2005; 19:178–181.

Monnier

, Colette

: Glycemic variability: Should we and can we prevent it?. Diabetes Care, 2008; 31(Suppl 2):S150–S154.

DECODE Study Group: Glucose tolerance and cardiovascular mortality: comparison of fasting and 2-hour diagnostic criteria. Arch Intern Med, 2001; 161:397–405.

Muggeo

, Zoppini

, Bonora

, Brun

, Bonadonna

, Moghetti

, Verlato

:Fasting plasma glucose variability predicts 10-year survival of type 2 diabetic patients: the Verona Diabetes Study. Diabetes Care, 2000, 23:45–50.

Zhang

, Yang

: Glucose tetrad new protagonist—the blood glucose fluctuations. Chin J Pract Int Med, 2009; 29:759–760.

Siegelaar

, Holleman

, Hoekstra

, DeVries

: Glucose variability; does it matter?. Endocr Rev, 2010; 31:171–182.

Bergenstal

, Tamborlane

, Ahmann

, Buse

, Dailey

, Davis

, Joyce

, Perkins

, Welsh

, Willi

, Wood

; the STAR 3 Study Group: Effectiveness of sensor-augmented insulin-pump therapy in type 1 diabetes. N Engl J Med, 2010; 363:311–320.

Battelino

, Conget

, Olsen

, Schütz-Fuhrmann

, Hommel

, Hoogma

, Schierloh

, Sulli

, Bolinder

; SWITCH Study Group: The use and efficacy of continuous glucose monitoring in type 1 diabetes treated with insulin pump therapy: a randomized controlled trial. Diabetologia, 2012; 55:3155–3162.

10.

Hermanides

, Nørgaard

, Bruttomesso

, Mathieu

, Frid

, Dayan

, Diem

, Fermon

, Wentholt

, Hoekstra

, DeVries

: Sensor-augmented pump therapy lowers HbA1c in suboptimally controlled type 1 diabetes; a randomized controlled trial. Diabet Med, 2011; 28:1158–1167.

11.

Raccah

, Sulmont

, Reznik

, Guerci

, Renard

, Hanaire

, Jeandidier

, Nicolino

: Incremental value of continuous glucose monitoring when starting pump therapy in patients with poorly controlled type 1 diabetes: The RealTrend Study. Diabetes Care, 2009; 32:2245–2250.

12.

Fritzsche

, Kohnert

, Heinke

, Vogt

, Salzsieder

: The use of a computer program to calculate the mean amplitude of glycemic excursions. Diabetes Technol Ther, 2011; 13:319–325.

13.

Kovatchev

, Cox

, Gonder-Frederick

, Clarke

: Symmetrization of the blood glucose measurement scale and its applications. Diabetes Care, 1997; 20:1655–1658.

14.

Cretti

, Lehtovirta

, Bonora

, Brunato

, Zenti

, Tosi

, Caputo

, Caruso

, Groop

, Muggeo

, Bonadonna

: Assessment of beta-cell function during the oral glucose tolerance test by a minimal model of insulin secretion. Eur J Clin Invest, 2001; 31:405–416.

15.

Alford

, Henriksen

, Rantzau

, Vaag

, Hew

, Ward

, Beck-Nielsen

: Impact of family history of diabetes on the assessment of beta-cell function. Metabolism, 1998; 47:522–528.

16.

Service

, O'Brien

, Rizza

: Measurements of glucose control. Diabetes Care, 1987; 10:225–236.

17.

Gao

, Wei

, Wang

, Li

, Wang

, Men

: Effects of intensive insulin treatment of the B cell function in newly diagnosed type 2 diabetes mellitus. Chin Remedies Clinics, 2006; 6:902–904.

18.

Kohnert

, Augstein

, Zander

, Heinke

, Peterson

, Freyse

, Hovorka

, Salzsieder

: Glycemic variability correlates strongly with postprandial beta-cell dysfunction in a segment of type 2 diabetic patients using oral hypoglycemic agents. Diabetes Care, 2009; 32:1058–1062.

19.

Wintergerst

, Buckingham

, Gandrud

, Wong

, Kache

, Wilson

: Association of hypoglycemia, hyperglycemia, and glucose variability with morbidity and death in the pediatric intensive care unit. Pediatrics, 2006; 118:173–179.

20.

Cryer

, Davis

, Shamoon

: Hypoglycemia in diabetes. Diabetes Care, 2003; 26:1902–1912.

21.

Sivas

, Astakhov

, Shadrichev

: Effect of inappropriate insulin therapy in type 2 diabetes on diabetic retinopathy progression [abstract 10]. Presented at the EASD 43rd Annual Meeting, Amsterdam, 2007.

22.

Danne

, Kordonouri

, Enders

, Hövener

, Weber

: Factors modifying the effect of hyperglycemia on the development of retinopathy in adolescents with diabetes. Results of the Berlin Retinopathy Study. Horm Res, 1998; 50:28–32.