Version July 2025
INTRODUCTION —
The following cases illustrate some of the common problems that may arise during intensive insulin therapy in people with type 1 diabetes mellitus. The management of type 1 diabetes is reviewed in detail separately. (See "Management of blood glucose in adults with type 1 diabetes mellitus" and "Overview of general medical care in nonpregnant adults with diabetes mellitus".)
CASE 1: GLYCEMIC VARIABILITY DUE TO DIET —
A 57-year-old hospital administrator who has had type 1 diabetes for 15 years is trying very hard to improve her glycemic management. She does not "count" carbohydrates (does not estimate her carbohydrate intake or use an insulin-to-carbohydrate ratio) and does not use an insulin sensitivity (correction) factor for correction doses for hyperglycemia. Instead, she adjusts her insulin dose, her food intake, and physical activity pattern in response to how she feels. Her work hours vary, and she frequently needs to attend business lunches and dinners.
Her usual insulin doses are:
●Before breakfast – 10 units insulin aspart and 36 units insulin degludec
●Before lunch – 8 units insulin aspart
●Before evening meal – 12 units insulin aspart
●Before bedtime snack – 4 units insulin aspart
She is frustrated at the unpredictability of her blood glucose values, as shown in the table (table 1).
Interpretation and approach — When a person continually adjusts insulin doses using an approach that is not consistent and also varies food intake and exercise routines, it is difficult to identify glycemic patterns. In this setting, keeping a three-day record of food intake, insulin doses, and exercise can be helpful in understanding the basis for the wide glycemic excursions (figure 1). As can be seen, the woman's total carbohydrate intake and its distribution are markedly different each day. In general, the use of fixed doses of preprandial rapid-acting insulin in a person who does not eat the same amount of food or eat at the same times each day does not work well. In addition, fingerstick blood glucose measurement four times daily gives an incomplete picture of glycemic patterns. Between fingersticks, wide glycemic excursions may occur, including episodes of hypoglycemia and hyperglycemia. If possible, she should be encouraged to use a continuous glucose monitor (CGM).
A consultation with a registered dietitian or diabetes educator to learn carbohydrate counting (usually more realistic than trying to eat a similar amount of carbohydrates from day to day) can be helpful. She should be taught to calculate her mealtime insulin dose using an insulin-to-carbohydrate ratio and to use an insulin sensitivity (correction) factor to supplement the pre-meal insulin doses when she is hyperglycemic. She may need different insulin-to-carbohydrate ratios for different times of the day. (See "Nutritional considerations in type 1 diabetes mellitus", section on 'Advanced carbohydrate counting'.)
●If her insulin-to-carbohydrate ratio is 1:15 (ie, she needs 1 unit of rapid-acting insulin to cover 15 g of carbohydrate) and she eats 90 g of carbohydrate for breakfast, she would take 6 units of insulin aspart pre-meal. If she eats only 30 g of carbohydrates one morning, she would take 2 units of insulin aspart.
●If her correction factor is 30 (1 unit lowers glucose concentration by 30 mg/dL [1.7 mmol/L]), she then adjusts the pre-breakfast or pre-dinner insulin based upon her pre-meal glucose and her target glucose. If the target glucose is 120 mg/dL (6.7 mmol/L) and the pre-meal glucose is 210 mg/dL, she would give herself an extra 3 units of "correction" insulin (subtract 120 mg/dL from the pre-meal glucose and divide by 30 to calculate the correction insulin).
If she cannot or is unwilling to learn to use an insulin-to-carbohydrate ratio, the registered dietitian can assist her in developing a carbohydrate-consistent diet (approximately the same amount of carbohydrate for each breakfast, for each lunch, and for each dinner), or help her determine the carbohydrate content of her small, medium, and large meals. Based on the agreed-upon carbohydrate content of each meal, fixed doses of insulin can be used, and an algorithm can be provided that incorporates a correction dose for elevated glucose readings. In addition, providing instructions for increasing rapid-acting insulin for very high-carbohydrate meals (eg, pizza), reducing rapid-acting insulin for the meal preceding exercise, and/or eating a carbohydrate snack before and/or during exercise is important.
CASE 2: BEDTIME HYPERGLYCEMIA AND UNSUSPECTED LATE MORNING HYPERGLYCEMIA —
A 27-year-old woman with type 1 diabetes is treated with the following regimen:
●Before brunch – 8 units insulin aspart and 20 units insulin detemir
●Before evening meal – 4 units aspart
●Before bedtime – 12 units detemir
Her blood glucose values before meals and before bedtime are as shown in the table (table 2). She also recently started to use a continuous glucose monitor (CGM). Her CGM glucose levels on a recent day are shown in the figure (figure 2). On this day (and most days), she eats two meals: brunch and dinner.
Interpretation and approach — After examining the available blood glucose readings (table 2), it appears that the main problem with her current glycemic management is hyperglycemia before bedtime. However, the CGM data reveal hyperglycemia after brunch as well.
You suggest that the hyperglycemia can be improved by increasing the dose of insulin aspart before her brunch and evening meal. It is also prudent to consider some other aspects of her lifestyle. As an example, she may be snacking after dinner, including close to bedtime. Furthermore, the evening meal may be her biggest and most variable meal of the day. It would therefore be helpful for her to learn carbohydrate counting and to use an insulin-to-carbohydrate ratio when calculating her mealtime insulin aspart doses (see "Nutritional considerations in type 1 diabetes mellitus", section on 'Advanced carbohydrate counting'). Exercise in the afternoon may also be lowering her late afternoon glucose levels.
Her lowest glucose reading before dinner was 72 mg/dL. Overnight, her glucose readings fell to 80 mg/dL (after being hyperglycemic at bedtime). In addition to needing more insulin aspart to cover her meals, lower doses of insulin detemir will be needed to prevent overnight and late afternoon/early evening hypoglycemia.
CASE 3: MORNING HYPERGLYCEMIA —
A 24-year-old insurance salesman with type 1 diabetes for two years feels stressed and tired at work and has early morning headaches. His glycated hemoglobin (A1C) value is 9.6 percent. Until fairly recently, he thought that his diabetes was well managed based on fingerstick blood glucose monitoring. His current regimen is:
●Before breakfast – 6 units regular insulin and 28 units NPH insulin
●Before evening meal – 4 units regular insulin and 16 units NPH insulin
Blood glucose values are as shown in the table (table 3).
Interpretation and approach — The increasing difficulty in optimizing glycemia in this patient with recent-onset diabetes is probably due in part to a progressive decline in endogenous insulin secretion. In addition, his insulin regimen is not ideal, as he is taking too little regular insulin, and he is taking NPH insulin too early in the evening. As an example, his high blood glucose values at 1:30 PM suggest that he needs to take regular insulin before lunch.
The triad of reasonable blood glucose values before bedtime, very high fasting values before breakfast, and a history of tiredness and early morning headache suggest a problem with the NPH insulin dose that he is taking before his evening meal. The action of the NPH insulin may be maximal in the late evening and middle of the night (causing both normoglycemia at bedtime and nocturnal hypoglycemia) and then dissipating in the early morning (causing hyperglycemia before breakfast). The presence of nocturnal hypoglycemia followed by early morning hyperglycemia (or hypoglycemia) can be confirmed by measuring blood glucose at 3 AM or by using continuous glucose monitoring (CGM).
One solution is to move his evening dose of NPH insulin from before the evening meal to before bedtime. In addition, both of his NPH insulin doses are likely to be too high. When using NPH insulin at night, a bedtime snack may be needed to prevent hypoglycemia during the night. When using NPH and regular insulins in the morning, a mid-morning snack may be needed. These snacks can contribute to weight gain.
The timing of regular insulin administration in relation to the meal is also important. Due to the relatively slow absorption of regular insulin (compared with the rapid-acting insulin analogs), it is best taken 30 to 45 minutes before eating. Given his busy and unpredictable schedule, including the challenge of testing his blood glucose and administering pre-meal insulin doses 30 to 45 minutes before meals, he may prefer a rapid-acting insulin analog, which can be taken immediately before the meal to allow greater flexibility. The rapid-acting analogs also have a shorter duration of action than regular insulin and may reduce the occurrence of hypoglycemia before the next meal.
A preferred approach would be to switch from NPH insulin to a basal insulin with a longer, flatter profile (such as insulin glargine or degludec), and to change from regular insulin to a rapid-acting insulin analog to manage mealtime glycemic excursions and to use for correction of hyperglycemia. (See "Management of blood glucose in adults with type 1 diabetes mellitus", section on 'Multiple daily injections'.)
A review of his exercise patterns in the evening may also be helpful. Vigorous evening exercise makes it more likely that his blood glucose concentration will fall to low values during the night. (See "Cases illustrating the effects of exercise in intensive insulin therapy for type 1 diabetes mellitus".)
The following changes in his insulin regimen led to marked improvement in his glycemia:
●8 units insulin glulisine and 30 units U-300 insulin glargine before breakfast
●6 units glulisine before lunch
●8 units glulisine before the evening meal
Three months later, his blood glucose values are as shown in the table (table 4).
His A1C value is now 7.8 percent. He has more energy, the headaches are gone, he is eating and exercising consistently, and has just been promoted because of his improved productivity. To reduce his A1C value further (goal is below 7 percent), he should be taught carbohydrate counting so that he can use an insulin-to-carbohydrate ratio. If this is not possible, a pre-meal insulin algorithm should be provided. The use of CGM or a hybrid closed-loop system (an automated insulin delivery system) may also be helpful to reduce time in hypoglycemia and increase time in the target glucose range. (See 'Case 1: Glycemic variability due to diet' above and "Management of blood glucose in adults with type 1 diabetes mellitus", section on 'Designing an MDI insulin regimen' and "Management of blood glucose in adults with type 1 diabetes mellitus", section on 'Follow-up'.)
CASE 4: UNPREDICTABLE GLUCOSE PATTERN —
A 19-year-old college freshman has had type 1 diabetes since age 9. His A1C in his senior year of high school was 7.5 percent, but the most recent one was 9.4 percent. He is taking a full course load, is getting decent grades, and is involved in lots of intramural sports, drama, and campus politics. He takes 32 units of insulin U-100 glargine in the morning, plus 4 to 8 units of lispro insulin before his meals. He is frustrated with his blood glucose values, which are shown in the table (table 5).
Interpretation and approach — Despite testing his blood glucose at least five times daily, there is no discernable pattern. In this situation, it is important to take a more detailed history of exactly what his day-to-day life is like and exactly how and where he is giving his insulin injections. Some days he is up at 7 AM, and on other days he does not get up until 1 PM. The timing of his "morning" dose of insulin glargine therefore varies by six hours from day to day. He has been rotating his glargine and lispro shots from his arm, leg, buttock, or abdomen just like he was told to do when he was first diagnosed at age 9. One day he went for a three-hour bike ride just after giving his morning glargine insulin in his thigh. He was severely hypoglycemic two hours into the ride. He sometimes forgets to check his blood glucose or take his lispro before he eats so then takes a dose of lispro immediately after the meal based on how much carbohydrate he thinks he just ate. He then checks his blood glucose one to two hours later and gives himself a "correction dose" of lispro at that time if his blood glucose is too high.
Unpredictable blood glucose patterns commonly indicate variability in eating, exercise, timing of insulin administration, and/or physical or emotional stress. You can attempt to help the patient reduce sources of variability:
●Help him to start using a continuous glucose monitor (CGM). Use of a CGM can show him blood glucose changes in close to real time, alarm for hypoglycemia and hyperglycemia, and help him make better self-management decisions. (See "Glucose monitoring in the ambulatory management of nonpregnant adults with diabetes mellitus", section on 'Benefits of CGM'.)
●Help him find a consistent time of day to give his glargine insulin in the afternoon or evening so that it is given at more or less the same time every day.
●Help him come up with strategies to remind him to check his blood glucose levels before his meals and to take the dose of lispro insulin based on the blood glucose and the estimated amount of carbohydrate he is about to eat.
●Ask him not to exercise a leg or arm within two hours of giving an insulin injection into that limb.
●Try giving the glargine insulin in the buttock or thigh and lispro into the abdomen.
●Try changing the basal insulin from glargine to insulin degludec or U-300 insulin glargine.
U-300 insulin glargine has a longer duration of action than U-100 glargine. Insulin degludec has the longest duration of action of the available basal insulins, and it might better suit his lifestyle, especially his inability to take his basal insulin at the exact same time each day. Alternatively, insulin pump therapy can be considered.
Using an insulin pump that communicates with his CGM and provides automated insulin delivery will be most helpful. He will still need to give insulin boluses for meals, but if he forgets to bolus, automated correction doses can reduce post-eating hyperglycemia.
CASE 5: SOME CONSISTENCY IN GLUCOSE PATTERN —
A 32-year-old executive has a busy office job that is unpredictable in terms of physical activity, time of meals, and/or the length of the workday. Her current insulin regimen is:
●Before breakfast – 24 units of insulin degludec
●Before breakfast and lunch – 4 units of insulin lispro
●Before evening meal – 8 units insulin lispro
The summarized glucose data from her continuous glucose monitor (CGM) for the last two weeks show values generally within the target range (70 to 180 mg/dL [3.9 to 10 mmol/L]) between midnight and 9:00 AM. However, glucose values rise midmorning and typically remain in the "very high" range (≥250 mg/dL [13.9 mmol/L]) throughout the late afternoon and evening.
Interpretation and approach — Her hyperglycemia appears to be primarily related to eating. She needs more rapid-acting insulin for most meals. Use of an insulin-to-carbohydrate ratio and correction factor would be helpful. If that is not possible, she can be advised to administer different doses of rapid-acting insulin for small, medium, and large meals. If she is willing and able, using an automated insulin delivery system (hybrid closed-loop system) should help improve her glycemic management. These systems suspend insulin delivery if hypoglycemia is predicted and can give automated correction doses of insulin for hyperglycemia after meals.
CASE 6: LATE AFTERNOON HYPOGLYCEMIA —
A 42-year-old garage mechanic is treated with the following insulin regimen:
●Before breakfast – 8 units insulin lispro and 16 units insulin glargine
●Before lunch – 4 units lispro
●Before evening meal – 10 units lispro and 22 units glargine
He uses an algorithm (table 6) to adjust pre-meal doses of lispro insulin. Blood glucose values for one week are shown in the table (table 7).
Interpretation and approach — The blood glucose record is quite complete, but there is no information about the quantity of food eaten and periods of exercise. He is consistent about the timing of blood glucose testing and meals and has adjusted his insulin doses correctly using the algorithm. His blood glucose was measured between 7:05 and 7:15 AM every day, except for day 7 when he slept late and the test was done at 9:55 AM. Several patterns can be detected:
●On day 7, the fasting blood glucose was 158 mg/dL (8.8 mmol/L). According to the algorithm, he should have increased the dose of lispro insulin by 2 units (for a total dose of 10 units lispro). He did not do this, and it turned out to be the correct decision since his pre-lunch glucose reading was at target.
●On day 1, when his fasting glucose was 100 mg/dL (5.6 mmol/L), he took the same dose of lispro as on day 7 and was hyperglycemic before lunch.
●This suggests that either the carbohydrate content of breakfast varies on different days, or that hyperglycemia before lunch is related to eating a mid-late morning snack. He would benefit from either using an insulin-to-carbohydrate ratio (preferred), or an algorithm that contains a different amount of insulin for small, medium, and large carbohydrate meals.
●His average blood glucose value is too low (83 mg/dL [4.6 mmol/L]) before the evening meal, with hypoglycemia occurring on days 2 and 4. On these days, he increased his pre-lunch insulin lispro dose per the algorithm. This suggests that the algorithm is too aggressive (gives too much "correction" insulin) for the mid-day meal. Additional possible explanations include a morning dose of insulin glargine that is too high, a lunchtime dose of lispro insulin that needs to be adjusted for the carbohydrate content of the meal or engaging in exercise in the afternoon.
CASE 7: USE OF HYBRID CLOSED-LOOP SYSTEM —
A 42-year-old man with type 1 diabetes since age 20 has been using an insulin pump for the past 19 years. His basal rate is 1.15 units/hour from 6 AM to 2 AM and 1.5 units/hour from 2 AM to 6 AM. He uses an insulin-to-carbohydrate ratio of 1:7.5 for breakfast and lunch and 1:7 for dinner. His insulin sensitivity factor (correction factor) is 45. After beginning a hybrid closed-loop system (hybrid artificial pancreas) in the automatic ("Auto") mode, his glucose readings are as shown in the figure in period "B" (figure 3). (See "Management of blood glucose in adults with type 1 diabetes mellitus".)
Interpretation and approach — During period "B," he consistently is hyperglycemic after lunch whereas evening readings are variable (figure 3). He is advised to change his insulin-to-carbohydrate ratio to 1:7 (from 1:7.5) for lunch, resulting in slightly more insulin coverage, and he reviewed his carbohydrate counting skills with a registered dietitian. In addition, he should take mealtime insulin boluses before he starts to eat the meal. When using Auto mode, if the mealtime bolus is taken after starting to eat (after the glucose levels start to rise), the automatic feature will give more basal insulin in response to the rising glucose concentrations, resulting in more insulin on board and a higher risk of hypoglycemia after the meal.
After making these changes, his time in range improved from 56 to 72 percent (period "A" in the figure) (figure 3). The greatest variability remains in the late afternoon and evening, when his physical activity and the composition of meals are the most unpredictable. Cases illustrating approaches to the management of exercise with intensive insulin therapy are discussed separately. (See "Cases illustrating the effects of exercise in intensive insulin therapy for type 1 diabetes mellitus".)
CASE 8: EVENING HYPERGLYCEMIA —
A 34-year-old teacher is striving for excellent glycemic control. He eats a healthy diet, counts carbohydrates, exercises at a consistent time each day, and has been taking the following regimen:
●Before breakfast (6 AM) – 4 units insulin aspart
●Before lunch – 4 units insulin aspart
●Before evening meal (6 PM) – 6 units insulin aspart
●Before bedtime (10 PM) – 28 units insulin glargine
Doses of insulin aspart are also adjusted using an insulin algorithm.
Typical blood glucose concentrations are shown in the table (table 8).
Interpretation and approach — Blood glucose values are ideal before breakfast and lunch, but are somewhat high before the evening meal and are much too high before bedtime. It is likely that the effect of the insulin glargine is waning after 20 hours. He could try splitting the insulin glargine dose into two equal doses taken approximately 12 hours apart. If he wants to take only one injection of basal insulin daily, he could change from U-100 insulin glargine to the longer-acting insulin degludec or U-300 insulin glargine. He may also require a higher dose of insulin aspart before lunch and requires a higher dose of insulin aspart before the evening meal. Alternatively, an automated insulin delivery system (hybrid closed-loop system) can be helpful.
CASE 9: HYPERGLYCEMIA RELATED TO FEAR OF HYPOGLYCEMIA —
A 57-year-old woman who lives alone has had type 1 diabetes for 41 years. She has hypoglycemia unawareness, a history of severe hypoglycemia (including automobile accidents in the past), and a fear of hypoglycemia. Her diabetes is also complicated by nephropathy, neuropathy, nonproliferative retinopathy, and hypothyroidism. She takes insulin glargine 32 units each morning. Her insulin-to-carbohydrate ratio is 1:10 and insulin sensitivity (correction) factor is 45 with a target glucose of 120 mg/dL (6.7 mmol/L), but she usually takes less insulin aspart than prescribed because of fear of hypoglycemia (on average, 12 units of insulin aspart daily).
Over the past year, her A1C range has been 7.3 to 7.6 percent, most recently 7.3 percent. Her weight is 258 lbs (117 kg) and body mass index (BMI) 42.5 kg/m2. She recently started using a continuous glucose monitoring (CGM) device, which was downloaded at this visit (figure 4). Her time in target range is only 46.3 percent, and she has wide glycemic excursions. Her high glucose variability is also reflected by the high coefficient of variation (38.1 percent).
Interpretation and approach — This patient is taking too much basal insulin (glargine) in the morning and not enough mealtime insulin, especially with her evening meal. The afternoon hypoglycemia is related to her morning insulin glargine dose. It is also likely that the glucose-lowering effects of her insulin glargine are waning overnight, contributing to nocturnal hyperglycemia. She would benefit from splitting her glargine dose to twice daily. She was reluctant to change her insulin regimen due to her fear of nocturnal hypoglycemia, but she agreed to change her insulin glargine to 18 units each morning and 12 units each evening. These changes were agreed upon using shared decision-making. She was also encouraged to take more insulin aspart, using her insulin-to-carbohydrate ratio for evening meals and to reduce her nighttime snacks.
CASE 10: USE OF ADVANCED HYBRID CLOSED-LOOP SYSTEMS —
A 78-year-old woman has had type 1 diabetes for 45 years. It is complicated by hypoglycemia unawareness, fear of hypoglycemia, neuropathy, retinopathy, orthostatic hypotension, carotid artery disease, osteoarthritis, and urinary incontinence. She has had episodes of diabetic ketoacidosis and severe hypoglycemia in the distant past. She is a widow and lives alone but has a daughter nearby who assists her. She does not count carbohydrates. She is using a continuous glucose monitor (CGM) and an insulin pump without automation and frequently does not bolus for meals. Her A1C was 11.6 percent, and time-in-target range (70 to 180 mg/dL [3.9 to 10 mmol/L]) was 6 percent. Glucose data over three days demonstrate persistent hyperglycemia, with glucose values ≥250 mg/dL (13.9 mmol/L) during most of the day and overnight.
Interpretation and approach — Her glucose readings are consistently elevated. You have her start using an advanced hybrid closed-loop system with automated insulin delivery. Such systems provide automated correction insulin doses for hyperglycemia (eg, when the user does not bolus for meals). As can be seen in the figure, her glycemic management significantly improved. Her A1C fell to 7.4 percent with 60 percent time-in-target range (70 to 180 mg/dL [3.9 to 10 mmol/L]). Although automated insulin delivery systems work best when the user appropriately boluses for meals, they can also improve glycemia in people who do not bolus consistently for eating.
ACKNOWLEDGMENT —
The editorial staff at UpToDate acknowledge David McCulloch, MD, who contributed to earlier versions of this topic review.
