Blood Glucose Monitoring: A Practical Guide for Use in the Office and Clinic Setting Free

B.T. is an 80-year-old woman with type 2 diabetes, who has come to the clinic for a checkup of her diabetes. She also has congestive heart failure,hypertension, and hyperlipidemia. She is confined to a wheelchair and therefore gets very little physical activity. Her most recent health crisis includes hospitalization for pneumonia. During that time, her blood glucose levels were consistently 300–400 mg/dl. Before this illness, her blood glucose levels were more or less under control. She had been increasing her insulin doses to cover the higher blood glucose levels related to her illness,but she is now complaining of symptoms of hypoglycemia.

Her usual home regimen is 10 units of glargine insulin twice daily, 10 units of lispro before breakfast, 8 units of lispro before lunch, and 10 units of lispro before dinner. She does not have a logbook or written record of blood glucose results. However, she says that she checks her blood glucose a few times each day. How might a record of her SMBG assist in her care today if you could review it quickly and easily in the exam room with her?

There are several methods for evaluating SMBG results, including manual review of patient diaries or logbooks, use of meter memory histories and computation, and electronic meter downloads with external software analysis. Regardless of method, there are five standard goals for a systematic review:

1. Estimate the degree of blood glucose control and variation throughout the day.

2. Identify patient safety concerns with regard to hypoglycemic trends or events.

3. Understand the factors influencing blood glucose control.

4. Suggest strategies for achieving improved blood glucose control.

5. Provide reinforcement to patients that this information is valuable and useful in their care.

In general, the average or mean blood glucose (using at least a 3-day history) can provide an idea of recent diabetes control and allow for counseling on factors that may have affected blood glucose, including diet,physical activity, and medications.

Ideally, interpretation of SMBG should be based on a systematic review or download of meter results because this provides the most objective method for reviewing information and providing feedback to patients. In reality, however,many patients bring only their handwritten records or logbook to office visits.

Logbooks (Figure 1) are distributed with each new meter and are a common method of record keeping in diabetes. However, interpretation of SMBG based on handwritten logbook records is dependent on patients' record-keeping intensity and organization, and these records can vary dramatically in terms of helpfulness in making management decisions. Some patients who bring in a handwritten logbook are very careful to note the time of day, how much insulin or medication was given, what was eaten, and the corresponding blood glucose result. Other patients provide a collection of results, in random order, without indicating the date, time, or reason for taking the blood glucose measurement.

When relying on handwritten SMBG results, it is essential that results are organized, timed, and provide a sense of the influence of medications and lifestyle on blood glucose to allow for meaningful diabetes care decision making. If logbooks are lacking sufficient information to provide feedback on diabetes self-management, patients should be specifically instructed to increase SMBG, including the time of day to test and timing with regard to meals or medications. Follow-up should then be arranged to assess the SMBG information after 1–2 weeks. If there are concerns about the reliability of results, a hemoglobin A_1c (A1C) test or random finger-stick check can be performed in the office or clinic to validate the information.

Important considerations when reviewing handwritten SMBG logbooks are to identify hypoglycemic events (safety concerns) and glucose patterns (efficacy of therapy).⁹ Table 2 reviews the steps for manual interpretation of SMBG results. Logbooks are commonly organized with columns for time of day and rows for day of the week(Figure 1), allowing the eye to travel down columns to identify trends at certain times of day and across rows for trends occurring certain days of the week. After reviewing a patient's logbook, worrisome hypoglycemic events should be immediately discussed with the patient and a plan should be created to avoid future events. Trends in glucose values also should be discussed to provide encouragement for positive trends and identify strategies for reducing negative trends (e.g., weekend high readings because of eating at restaurants). The final step is to provide insight to the patient on the overall glucose average and degree of variability and to roughly equate these to the expected A1C(Table 3).

B.T. reveals that she cannot write anymore because she has severe tremors, especially on her right side, which is her dominant hand. She confides that maintaining her logbook has been difficult. She is cared for by her son and daughter-in-law, who are busy during the day and unable to help her with this task.

Many patients rely on their meter's memory to act as the record rather than take the extra step to write down their results in a logbook. These patients find entering the large amounts of data they generate into a logbook to be tedious and unnecessary.¹⁰  With the growing prevalence of diabetes and the increasing use of computers by patients and health care providers, the use of meter technology and downloading should become the standard in the future. As an interim step, the use of meter memories can add to a patient's evaluation and can help validate handwritten results.

The meter memory can be accessed by turning on the meter and scrolling through the available results manually. The number of readings stored varies based on meter type (Table 1),and results are usually listed in chronological order from most to least recent. It can help to write down values from the past 1–2 weeks and then evaluate them based on time of day and result. Many meters may have the capacity to display an average, but this, again, depends on meter brand. The most common meters will give 7-day, 14-day, or 28-day averages. Review of the meter memory is also helpful to verify handwritten logbook results, clarify testing times (if the date and time have been set properly), and calculate SMBG averages, if the meter is capable.

B.T. has brought in her meter, and scrolling through her values reveals higher-than-normal blood glucose levels. She is having hypoglycemia as well,confirming her reports of hypoglycemia symptoms. Without the information mapped out more succinctly, it is difficult to interpret trends and patterns.

Downloading SMBG results to an office or clinic computer, secured website,or even a patient's computer for analysis is a fast, simple, and effective method for evaluating glucose information. Major advantages of electronic SMBG download and analysis include speed of review; instant and accurate calculation of results (versus estimation often used when looking at logbooks); ability to report data in many ways using tables, graphs, charts,or statistics; and ability to show these reports immediately to patients for use in planning care.

Both specialty software and native software offered by the meter company are readily available for downloading. Several programs are able to interface with the popularly used meters (Table 4). One example of specialty software is Clinipro,¹¹  which can download most commercially available meters into one software program. After using software to download meter readings, a process similar to that described above for manual interpretation of logbook results can be used to interpret the data (Table 2).

The overall mean and SD are key factors to evaluate when using software analysis. The aggregate mean can be a good quick comparison from the last visit to the current visit in terms of diabetes control, assuming the amount of hypoglycemia is approximately equal. The average can be used to compare to recent A1C values so that discrepancies can be identified that may need further investigation (Table 3).

The SD, which is the square root of the variance, indicates the amount of variability among all the meter readings. It has been suggested that the SD value doubled should be less than the aggregate mean as a way to assess safety and efficacy of treatment regimen.¹⁰ Certainly, the higher the SD, the more variability a patient is experiencing,with many possible causes, such as erratic insulin or medication absorption,improper timing of meals with insulin onset, or gastroparesis creating a mismatch between food absorption and insulin action onset.

B.T.'s. meter is downloaded, and her aggregate mean for the past month is 245 mg/dl, reflecting her earlier hospital admission and illness-related hyperglycemia. She has an SD of 131 mg/dl, which is higher than desired. This could reflect that, although she is often hyperglycemic, she is having episodes of hypoglycemia, contributing to a large degree of variability. In her case, it could also be the result of the change in her usual level of control, with better blood glucose levels before pneumonia and higher levels afterward.

Evaluating fasting blood glucose is the best way to diagnose needs for improved basal control. For the majority of individuals, the first morning blood glucose (premeal) is a good proxy for fasting blood glucose (8 hours since last meal). For those taking insulin, preprandial glucose monitoring provides information regarding the efficacy of basal insulin doses(subcutaneous injection or pump basal infusion rate) and can also provide insight into the effectiveness of previously injected prandial coverage.¹²  For those taking oral medications, such as sulfonylureas or metformin, fasting blood glucose can provide a sense of basal control with regard to influence on insulin production (sulfonylureas) or inhibition of nighttime glucose output(metformin).

Elevated morning blood glucose levels can also be the result of nighttime snacking that presents as morning hyperglycemia, so it is important to explore reasons for high readings before making a change. Other nighttime routines,such as exercise, medication dosing, and sleep patterns, are also relevant topics to review with patients and can lend evidence or doubt to a given presumption about the patients' blood glucose patterns.

The blood glucose trend overnight is especially important, and bedtime blood glucose results can help to evaluate this trend. For example, if a patient goes to sleep with a blood glucose level of 120 mg/dl (> 3–4 hours after dinner) and has no bedtime snack, a blood glucose at or just <120 mg/dl would be expected the next morning. If the patient wakes up with blood glucose levels higher than at bedtime, it may be related to inadequate basal medication, reactive hyperglycemia in response to low blood glucose, or the dawn phenomenon.

Reactive hyperglycemia and the dawn phenomenon can be difficult to diagnose without the benefit of continuous glucose monitoring (CGM) or multiple nighttime SMBG values to observe trends. The dawn phenomenon is a natural increase in blood glucose between about 4:00 and 8:00 a.m., presumably as the result of counter-regulatory hormone increases or the release of growth hormone in the early morning hours.¹³ 

B.T.'s fasting average in the past has usually been quite good. She maintains that higher readings now are the result of her illness. She is having hypoglycemia at times, but this does not seem related to waking or fasting blood glucose levels.

Higher blood glucose attributed to prandial inadequacies can be seen when evaluating peak postprandial blood glucose (determined 1–2 hours after the start of a meal). Because many patients do not test in the period after eating, a good idea of postprandial coverage can be obtained from looking at the preprandial value for the next meal. If patients can recall the past 2 days of meals, activity, and medications, their blood glucose control can be evaluated based on the most recent data collection.

This can be important not only to interpret the data, but also to obtain a sense of the day-to-day influences on glucose variability. The SMBG averages on the glucose meter can be helpful to get a global sense of how a patient is doing. For example, if the aggregate mean blood glucose level for all predinner blood glucose values is higher than any other time of the day,information about what typically happens at lunch and afterward would be useful. If there is no apparent reason for the hyperglycemia, such as snacking or over-treatment of hypoglycemia, then increasing prandial coverage at lunch would make sense.

B.T. has begun experiencing hypoglycemia after meals. It is discovered that she has been increasing her prandial insulin recently to cover the high blood glucose levels related to her illness. She has since been treated with antibiotics, and her pneumonia is resolving. She had one blood glucose reading of 40 mg/dl shortly after breakfast. She gave 15 units of aspart for breakfast, and the low glucose level occurred 2 hours afterward. She also had another hypoglycemic reaction after dinner, and although she was unable to test because of the severity of the tremor with the reaction, she treated with candy and the feeling resolved. She had taken 15 units of aspart for dinner.

There are more aspects to evaluating glucose control than basal- and prandial-related concerns. A common issue for those using insulin is related to their correction algorithm, an insulin dose used to normalize blood glucose above the preprandial target. If too much insulin is given for correction of hyperglycemia, it may be reflected in the download by hypoglycemia after the correction dose. If an insufficient correction insulin dose is given, the reverse would be true, with continued hyperglycemia after treating.

Another cause of hypoglycemia after a treated hyperglycemia is what is sometimes termed “insulin stacking.” This occurs when insulin correction (using rapid- or short-acting insulin) is layered too heavily on top of the prandial dose of insulin already administered by the patient. There is generally a 4-hour window after the initial injection of rapid-acting insulin when this is most commonly seen, and it can be resolved by reducing the amount of additive correction insulin given when there is already substantial insulin remaining “on board.” For patients who take an oral sulfonylurea or use NPH insulin, hypoglycemia may be related to delayed or skipped meals.

Figure 2 shows B.T.'s download. Despite higher blood glucose levels from her previous illness, her meter download shows that her blood glucose control is improving,and the temporary increase in insulin is now too much. Review of her SMBG results before the illness also reveals that often her blood glucose levels were lower than ideal. A change is necessary in her prandial insulin regimen. Instead of 10 units before breakfast, 8 units before lunch, and 10 units before dinner, she is asked to use 7 units before each meal, representing a 25% reduction in her daily prandial insulin, and a 15% reduction of her overall daily insulin dose.

The American Diabetes Association (ADA) Standards of Medical Care in Diabetes recommends that SMBG should be carried out three or more times daily for patients using multiple daily insulin injections or insulin pump therapy. The ADA also states that SMBG may be useful for patients using less frequent insulin injections, noninsulin therapies, or medical nutrition therapy alone.¹⁴  This recommendation likely stems from the discrepancy of data supporting the benefit of frequent SMBG on glycemic control for those who are not using intensive insulin regimens.

However, from a practical standpoint, it can be difficult to make effective recommendations for change without some idea of how a patient's blood glucose manifests on a routine basis. Certainly cost considerations are a factor,although most insurance companies will cover their preferred meter strips, and Medicare covers three times daily testing for insulin-treated patients or once-daily testing for patients who do not use insulin. From the patients'perspective, one reason testing may not be a priority is health care providers' lack of enthusiasm at reviewing the data. If there is no feedback to be gained, it is difficult to expect adherence.

An extensive review of blood glucose monitoring data, part of a global consensus conference report on SMBG,¹⁵  included the following five recommendations:

1. SMBG should be initiated in all patients with diabetes as an integral part of an overall diabetes management program.

2. SMBG should be performed ≥ 3–4 times daily for patients treated with multiple daily insulin injections or using an insulin pump; ≥ 2 times daily for patients above their glycemic target who are managed with oral agents and once daily insulin; ≥ 1 time per day with a weekly profile (to include pre- and postprandial glucose readings) for patients at their glycemic target who are managed with either once-daily insulin or oral agents alone;≥ 1 time per day with more frequent weekly profiles for patients at their glycemic target who are managed with oral agents plus once-daily insulin; and≥ 1 weekly profile for patients managed nonpharmacologically, whether they are at or above their glycemic target.

3. Additional SMBG should be performed in certain situations, such as acute illness, intercurrent illness, changes in medications, for patients with impaired awareness of hypoglycemia, and during pregnancy.

4. Postprandial SMBG testing should be used by all patients with diabetes to minimize postprandial excursions and guide lifestyle changes.

5. SMBG should be viewed as an education tool to inform patients about the effects of lifestyle and behavioral changes.

Certainly, barriers to the above recommendations are well recognized,including high out-of-pocket costs or no insurance, language barriers, and lower education levels.^16,17 However, it is prudent to help patients where possible to overcome these barriers, given the importance of the information.

B.T. comes back to the clinic convinced that her meter is broken. She is getting such good results from the recent change that she requests validation of accuracy with our clinic meter. The two meters show identical results. Her download summary appears inFigure 3 and shows an improved aggregate mean of 122 mg/dl with an SD of 34 mg/dl.

CGM devices are now available for home use for patients with diabetes. These devices offer a large amount of blood glucose information, with readings displayed every 5 minutes. Although CGM devices offer another great way for clinicians to help patients with disease management and for patients to have much more data to evaluate daily decision making, they are still not the norm. With many practices still resistant to downloading meters, the time and resources needed to begin CGM downloading and review on a routine basis make it not yet practical for most offices. This emerging technology is described in full elsewhere in this issue(p. 112).

P.Q. is a 64-year-old woman who presents for a review of glycemic control. She has type 2 diabetes, weight management issues, hypertension, and dyslipidemia. She is now retired from her profession as a school teacher, and she feels that this has allowed her to start paying much more attention to her personal health needs. She states that she is disappointed that today's weight does not reflect the 10-lb loss she achieved this past fall, but also acknowledges that she was not as careful during the holiday season and therefore has regained some of the weight she had lost. She feels she is ready to renew her dietary efforts and become much more physically active. She has been quite careful with food choices for the past week.

Her medications include:

• Glargine, 20 units in the morning

• NPH, 8 units at bedtime

• Lispro, 6–8 units at breakfast, 8 units at lunch, and 8 units at dinner (She does not vary these doses with a correction algorithm.)

• Losartan, 100 mg daily

• Ezetimibe, 10 mg daily

• Niacin, 1,000 mg at bedtime

• Aspirin, 81 mg daily

On physical examination, her weight is 225 lb (down 3 lb from her weight 9 months ago). Her blood pressure is 152/70 mmHg, but she has a considerable office component, which we have demonstrated with a 24-hour ambulatory blood pressure monitor. Examination of her lower extremities shows good dorsalis pedis pulses. She has a rather significant fifth metatarsophalangeal on the right foot callus. This is not thick, but it is an area to be monitored, and the clinician draws her attention to it. On a 10-g monofilament challenge, she is able to appreciate the monofilament over all areas checked.

Laboratory values include an A1C of 8.8%, total cholesterol of 237 mg/dl, LDL cholesterol of 140 mg/dl, HDL cholesterol of 82 mg/dl,triglycerides of 75 mg/dl, and all other tests within normal limits.

P.Q.'s meter download chart is shown inFigure 4,, and a graph of her blood glucose readings is shown inFigure 5.

To interpret the data gathered from P.Q.'s meter, the clinician:

1. Reviews the report, highlighting key data.

   • What is the time period? (It should be at least 2 weeks.): The time period is 1 month with 102 readings.

   • Are there any concerning hypoglycemic events (e.g., SMBG < 70 mg/dl)?

     Yes, there have been eight events < 60 mg/dl. This is discussed with P.Q., who reveals that these occurred after exercising, although the one episode that occurred in the nighttime is a mystery. She suggests that maybe she gave her NPH shot twice the night before.

   • What are the trends?

     The highest mean blood glucose average (189 mg/dl) is at night. The highest SD (120.65 mg/dl) is occurring at bedtime, with a mean of 131 mg/dl(based on four values). The best control is at lunch (132 mg/dl, SD 50.54 mg/dl).

   • What are the average blood glucose readings?

     Morning fasting: 161 mg/dl (SD 71.47 mg/dl); mid-morning 137 mg/dl (SD 68.89 mg/dl); lunchtime 132 mg/dl (SD 50.54 mg/dl); mid-afternoon 164 mg/dl(SD 96.81 mg/dl); dinner 128 mg/dl (SD 77.16 mg/dl); mid-evening 154 mg/dl (SD 72.71 mg/dl); and bedtime 131 mg/dl (SD 120.65 mg/dl).

   • What is the aggregate mean blood glucose?

     154 mg/dl (SD 80.71 mg/dl).

2. Interpret the data.

   • The aggregate blood glucose is acceptable. However, the SD reveals high glucose variability throughout the day.

     P.Q.'s SD needs to be lower to reduce risk of serious hypoglycemia and hyperglycemic damage.

3. Plans for care.

   • Increase glucose management related to exercise to prevent hypoglycemia.

   • Because P.Q. maintains that she is careful with her diet and has not been snacking at night, increase her bedtime long-acting insulin to 10 units to manage her nighttime and fasting hyperglycemia. Consolidate the insulin to glargine (basal) and lispro (prandial), to simplify her regimen.

   • Review new orders: glargine, 10 units at bedtime and 20 units in the morning. Lispro, 6–8 units before meals (use ∼ 2 units per 15 g of carbohydrate in each meal).

   • Instruct P.Q. to use a correction dose of 1 unit of lispro for every 50 mg/dl that her blood glucose rises above 150 mg/dl before meals. This should be given in addition to her usual prandial dose.

Systematic interpretation of blood glucose data in the office or clinic can provide clinicians with invaluable information for improving diabetes care and promoting self-management for individuals with diabetes. According to surveys,people with diabetes would like increased ability to participate in the decision making for their diabetes care. Because blood glucose interpretation requires discussion with patients to identify reasons for variability and patterns in control, it is a natural way to facilitate integration of the patients' perspective, needs, lifestyle, and abilities into office-based decision making for diabetes care.

The goals of a review of SMBG results should include: 1)estimating the degree of blood glucose control and variation throughout the day; 2) identifying patient safety concerns with regard to hypoglycemic trends or events; 3) understanding the factors influencing blood glucose control; 4) suggesting strategies for achieving improved blood glucose control; and 5) providing reinforcement to patients with diabetes that this information is valuable and useful in their care. Meter technology should be used to provide the best care possible for patients and should become the minimum requirement for caring for patients with diabetes. Incorporation of this technology into clinical practice should be simple, practical, and, above all, considered necessary for patient care.