K.A. is a 62-year-old man with a 42-year history of “brittle” type 1 diabetes. When first seen at our office, he complained of erratic and unpredictable blood glucose levels despite adhering to a rigid multiple daily insulin injection protocol. He was taking four insulin injections per day (NPH before breakfast and at bedtime and lispro insulin (Humalog) before each meal) and performing self-monitoring of blood glucose (SMBG) eight times daily, including pre- and postprandially and whenever he felt symptomatic. K.A. injected his morning NPH in the abdomen, his lunch and dinner insulin in the arms, and his bedtime NPH in the buttocks.

He was sedentary and did not exercise. He had been counting carbohydrates to adjust his insulin for only 5 months.

Five years ago, K.A. developed frequent severe hypoglycemia, during which he lost consciousness and presented in the hospital emergency room. His blood glucose levels had been <30 mg/dl on each of his eight emergency room visits. In the ensuing 5 years, he had developed hypoglycemic unawareness.

Because of the frequency of hypoglycemic events with altered levels of consciousness (at least 30 episodes documented per month) the patient’s driver’s license was revoked, and he sought an early retirement from his job as an electrical engineer. He believed that his control might improve if he were placed on an insulin pump.

K.A. is 71 inches tall and weighs 74 kg. His blood pressure at our initial visit was 124/86 mmHg without orthostatic changes. His diabetes-related complications included autoimmune hypothyroidism, microalbuminuria, peripheral sensory neuropathy, and nocturnal diarrhea. His HbA1c was 7.3%, and his fasting capillary blood glucose was 125 mg/dl. His liver function was normal, but he did have proteinuria (245 mg/dl in a 24-h urine collection). His creatinine was 1.2 mg/dl. His thyroid and adrenal function studies were normal.

After unsuccessful attempts to fine-tune his insulin injections based on carbohydrate counting and preprandial glucose readings, we placed him on continuous subcutaneous insulin infusion (CSII) therapy. Unfortunately, K.A. did not experience any improvement in his glycemic control. He became more depressed and frustrated because he was still experiencing daily hypoglycemic events. After 2 weeks of insulin pump therapy, he was placed on a continuous glucose monitoring system (CGMS).

Figure 1 shows the results of his CGMS test. Analysis of the sensor readings demonstrated that the patient’s SMBG records were not consistent with those obtained by interstitial fluid sensing. At 8:00 a.m., the patient’s meter (circle markers in Figure 1) read 298 mg/dl, but the sensor reading (dark line in Figure 1) was 118 mg/dl. When he believed that his blood glucose was high based on erroneous meter readings, he would give himself a compensatory bolus of insulin via his pump. There was no consistent correlation between the sensor and the meter. Thus, K.A. was constantly administering insulin based on inaccurate SMBG readings. The disparity between the sensor and the meter was so substantial that the sensor would stop functioning and have to be re-calibrated by the patient throughout the day. (See Figure 2.) This indicated a problem with the meter, rather than a malfunctioning of the CGMS.

When the patient was informed about the inaccuracy of his meter, he explained that he had purchased the device, which used test strips, 10 years ago and he had never doubted its accuracy. Although reluctant to part from his favorite meter, the patient’s diabetes control improved when he was provided with a more accurate device. Two months after receiving his new meter, K.A. reported that he was no longer experiencing any hypoglycemic episodes and was feeling so well on his insulin pump that he began to work part-time and hoped to have his driving privileges re-instated. The CGMS test was repeated and showed no hypoglycemic events and improved overall glycemic control.

  1. What is the differential diagnosis of recurrent severe hypoglycemia in type 1 diabetes?

  2. How should providers approach type 1 diabetic patients with severe hypoglycemia?

Multiple causes of widely erratic glycemic control are seen in patients with longstanding type 1 diabetes. A careful history should be performed to explore causes related to diet, monitoring, medications, and activity. Testing should be directed at detecting occult liver and renal disease and thyroid and adrenal insufficiency. Patients may have delayed gastric emptying, which can result in postprandial hyperglycemia or hypoglycemia. Lipohypertrophy may affect insulin absorption, and differences in absorption and onset of action of intermediate- and long-acting insulins may cause wide glycemic variations. Occasionally, the preprandial dose of short-acting insulin does not match the anticipated rise in blood glucose levels that occurs with a given carbohydrate intake. Exercising during insulin peak times may also cause hypoglycemia. Patients who have had type 1 diabetes for more than 5 years develop decreased hypoglycemic counterregulation due to a diminished glucagon secretory response and diminished epinephrine secretion in response to hypoglycemia.

A novel way to determine the etiology of “brittle diabetes” is to place the patient on a CGMS. The CGMS is slightly larger than an insulin pump and is inserted painlessly into the subcutaneous tissue of the abdomen. After calibration, interstitial glucose levels are analyzed and stored every 5 min. Thus, ∼300 glucose readings are evaluated during a 24-h period.

A well-calibrated blood glucose meter should always be used while on the sensor. The patient wears the CGMS device for 3 days. The data are then downloaded into a personal computer, and a printout depicting nearly 900 glucose readings over a 3-day period is generated for analysis.

The CGMS allows a physician to evaluate trend analysis of the multiple interstitial glucose readings rather than relying on four or five daily fingerstick glucose levels. Using the CGMS, physicians can easily evaluate the effects of different foods on glycemic control, determine the presence of nocturnal or diurnal hypoglycemia, analyze postprandial glucose excursions, evaluate the effect of oral agents and different insulins on glycemic control, document the effect of different exercise programs on glycemia, and even determine the accuracy of the patient’s SMBG meter.

Patients are often stunned to learn that they have hypoglycemic unawareness or significant postprandial hyperglycemia in response to eating their favorite type of pizza. The documentation derived from CGMS analysis may allow physicians to improve overall control and prevent nocturnal hypoglycemia.

Using the CGMS, K.A. was able to regain the glycemic control that he had not had for at least 10 years. His insulin pump is now allowing him to return to work and enjoy life without having to frequent the emergency room for treatment of severe hypoglycemia.

  • Physicians should always evaluate the reason for their patients’ “brittle diabetes.” Table 1 lists common causes of wide glycemic variations.

  • Before initiating an extensive workup in individuals with widely fluctuating blood glucose levels, physicians should evaluate the accuracy of the patient’s home glucose meter. Patients should be instructed on how to clean, calibrate, and properly use their meters. If the meter is determined to be accurate, CGMS may be helpful in determining how a patient’s pharmacological therapy should be altered to improve glycemic control to reduce hypoglycemia and hyperglycemia.

Figure 1.

A CGMS test shows a disparity between the results of the patient’s SMBG (circles) and the interstitial glucose readings obtained by the sensor (dark line). The differences between the two types of readings were so substantial that the sensor had to be recalibrated on two occasions as shown by the breaks in the sensor graph.

Figure 1.

A CGMS test shows a disparity between the results of the patient’s SMBG (circles) and the interstitial glucose readings obtained by the sensor (dark line). The differences between the two types of readings were so substantial that the sensor had to be recalibrated on two occasions as shown by the breaks in the sensor graph.

Close modal
Figure 2.

Gaps in the sensor tracing indicate the disparity between the patient’s interstitial glucose readings and the blood glucose levels obtained with the home glucose meter.

Figure 2.

Gaps in the sensor tracing indicate the disparity between the patient’s interstitial glucose readings and the blood glucose levels obtained with the home glucose meter.

Close modal
Table 1.

Common Causes of “Brittle Diabetes”

Common Causes of “Brittle Diabetes”
Common Causes of “Brittle Diabetes”

Jeff Unger, MD, is director of the Chino Medical Group Diabetes Intervention Center in Chino, Calif.

Note of disclosure: Dr. Unger has received honoraria for speaking engagements from Medtronic MiniMed, which manufactures continuous subcutaneous insulin infusion systems and the continuous glucose monitoring system.

Unger J, Fredrickson L: A primer on intensive insulin management and insulin pump therapy.
Primary Care Rep
3
:
9
–17,
1997
Bode BW: Clinical utility of the Continuous Glucose Monitoring System.
Diabetes Tech Therapeut
2 (Suppl. 1)
:
S35
–S41,
2000
Rebrin K, Steil GM, Van Antwerp WP, Mastrototaro JJ: Subcutaneous glucose predicts plasma glucose independent of insulin: implications for continuous monitoring.
Am J Physiol
277
:
E561
–E571,
1999
Burge MR: Lack of compliance with home blood glucose monitoring predicts hospitalization in diabetes.
Diabetes Care
24
:
1502
–1503,
2001
Boland E, Monsod T, Delucia M, Brandt CA, Fernando S, Tamborlane WV: Limitations of conventional methods of self-monitoring of blood glucose: lessons learned from 3 days of continuous glucose sensing in pediatric patients with type 1 diabetes.
Diabetes Care
24
:
1858
–1862,
2001