Influence of Caffeine on Frequency of Hypoglycemia Detected by Continuous Interstitial Glucose Monitoring System in Patients With Long-Standing Type 1 Diabetes

All subjects provided written informed consent, and this study was approved by the local research ethics committee. Twenty patients with long-standing (>5 years) type 1 diabetes participated in a double-blind, randomized cross-over, placebo-controlled study examining the effects of caffeine ingestion on the frequency of hypoglycemia.

All participants were given a low-caffeine diet (<50 mg/day) for 2 weeks with either 250-mg caffeine capsules supplemented twice daily (equivalent to average daily caffeine intake in the U.K.) or matched placebo. Dietary advice and information sheets were provided. Both patients and investigators were blinded as to the composition of the supplemental capsules. On the last 2 days of the 2-week period, interstitial glucose levels were assessed using a CGMS (Medtronic Minimed, Minneapolis, MN) (10). Over the same 48 h, patients also underwent simultaneous assessment of HRV by continuous Holter monitoring recorded on a miniature digital recorder (Life Card, Reynolds Medical). The patients were crossed over to the alternate treatment arm and the final set of measurements was repeated at the end of an additional 2 weeks.

On each day trace obtained from the CGMS, hypoglycemia was defined according to the following criteria (11). Four consecutive readings of ≤63 mg/dl are required. The first reading ≤63 mg/dl is the start of the hypoglycemic episode. The hypoglycemic episode ends when there are at least four consecutive readings >63 mg/dl. The first reading >63 mg/dl is the end of the episode. The episode is prolonged if it lasts more than 2 h.

During an episode, if the sensor value is ≥63 mg/dl for one to three readings and then falls below 63 mg/dl, the whole episode is counted as one hypoglycemic episode, including the small period >63 mg/dl. A moderate hypoglycemic event occurs when interstitial glucose levels are ≤53 mg/dl for at least four consecutive readings. During a mild episode, if glucose levels fall to ≤53 mg/dl for four or more consecutive readings, the whole episode is classified as moderate. Severe hypoglycemia is defined as an event requiring external help.

The definition of interstitial hypoglycemia in terms of glycemic thresholds (mild and moderate) is based on the level at which there is activation of the counterregulatory hormone cascade and onset of neuroglycopenic symptoms, respectively (12).

Patients had no clinical evidence of overt autonomic dysfunction from history or physical examination. All were screened by 12-lead electrocardiogram, echocardiography, and treadmill exercise testing. Exclusion criteria included the presence of atrial fibrillation, left bundle-branch block, evidence of left ventricular systolic impairment (<50%), or exercise-induced ischemia (ST depression ≥1 mm at ≥85% maximum predicted heart rate during or after exercise using the Bruce protocol). All patients had time-domain analyses of HRV performed on 24-h ambulatory electrocardiogram monitoring (7).

CGMS data were downloaded to a Microsoft Excel spreadsheet. Only data from days with a mean absolute error of <28% with at least three calibration points per day were accepted. Data not fulfilling strict validity criteria were discarded. Hypoglycemia was defined as a sensor value ≤63 mg/dl for at least four consecutive sensor values (20 min). An episode was deemed to have ended when the sensor showed four consecutive readings of ≥63 mg/dl. Number, severity, and duration of hypoglycemic episodes were recorded. Time periods were also divided into day (0800–2400) and night (2400–0800).

Statistical analysis was performed through SPSS 11 for Windows. In the assessment of HRV, for each individual at the end of each phase of the study, the hourly counts of RR (interval between subsequent QRS complexes) over 50 ms for the 48-h period were summarized by calculating the area under the curve. These summary statistics were then used for the data analysis. HRV and average glucose are assumed to be normally distributed by the Kolmogorov-Smirnov Z statistic (P > 0.05). Average glucose has been summarized using means and SDs and compared between caffeine and placebo phases of the study using the paired t test.

The number of hypoglycemic episodes and changes in the number of episodes between caffeine and placebo phases of the study did not appear to be normally distributed and were summarized using medians and ranges; changes were tested using the Wilcoxon test. The association between HRV and number of hypoglycemic episodes was summarized using Spearman’s correlation coefficient. All tests were two tailed and used a 5% critical P value.

Patient details are shown in Table 1. CGMS data were available for 19 subjects. One set of data was invalid because of poor correlation between paired sensor and fingerstick values (mean absolute error of 33%). There was no difference in mean glycemic control between the groups over the 48 h of CGMS analysis (146 ± 47 mg/dl on caffeine vs. 144 ± 38 mg/dl on placebo, P = 0.94).

Over the 2-week study period, there were no reported severe hypoglycemic events. Overall, interstitial hypoglycemia was less frequent in those taking caffeine (P = 0.044). When patients were awake, the duration of interstitial hypoglycemia (<63 mg/dl) was similar with caffeine and placebo, but, in contrast, hypoglycemia was reduced at night during treatment with caffeine (49 [range 0–235] minutes vs. 132 [0–468], P = 0.035). The reduction in nocturnal hypoglycemia was as a consequence of a fall in the number of moderate episodes defined as an interstitial glucose level <53 mg/dl for 20 min or more (0.6 vs. 0.1 hypoglycemic events on caffeine, P = 0.011) (Table 2).

Ten patients did not suffer from any episode of hypoglycemia during the CGMS assessment phase of the study. Of the patients with interstitial hypoglycemia, moderate hypoglycemic episodes were less frequent in eight subjects during the caffeine treatment, although one subject had more episodes of interstitial hypoglycemia with caffeine (P = 0.003).

As noted previously, in this sample, ingestion of caffeine was associated with improvement in HRV (7). Although caffeine ingestion was associated with a marked increase in HRV, particularly at night, we could find no correlation between HRV and the frequency of hypoglycemia (r_s = 0.12, P = 0.62). Similarly, there was no correlation between the increase in HRV at night with caffeine and the reduction in nocturnal hypoglycemia with caffeine (r_s = −0.18, P = 0.47).

Despite modern intensive treatment regimens, hypoglycemia continues to be an important problem for patients with type 1 diabetes (13,14). Previously we have shown that ingestion of modest amounts of caffeine (equivalent to three to four cups of drip-brewed coffee each day) augments the symptomatic and hormonal responses to hypoglycemia in healthy volunteers (15) and patients with type 1 diabetes (6). The beneficial effect of caffeine on hypoglycemia risk is independent of a change in glycemic control (16). Here, using continuous interstitial glucose monitoring, we have shown that caffeine may also be beneficial by reducing the frequency of mod-erate episodes (<53 mg/dl) of hypoglycemia occurring at night.

The consequences of prolonged unrecognized nocturnal hypoglycemia are not known, but it is recognized that recurrent episodes of hypoglycemia can cause loss of warning symptoms and defective glucose counterregulation during subsequent episodes in which glucose levels fall below normal (17). Therefore, it can be postulated that the caffeine-associated reduction in moderate hypoglycemic episodes occurring during sleep seen here may be beneficial by reducing the risk of next day hypoglycemia unawareness with the corollary of less asymptomatic hypoglycemia. In this study, we showed a tendency toward less daytime hypoglycemia, although this did not reach statistical significance (Fig. 1). However, the caffeine-associated reduction in “antecedent” nocturnal hypoglycemia seen in this study may explain the augmentation in the symptomatic and hormonal responses to mild daytime hypoglycemia described previously (6).

Using a validated tool for analysis of CGMS data allows breakdown of the number and time of hypoglycemic events during the day and night (11). Examination of the various hypoglycemic episodes generates a number of statistical tests, potentially increasing the chance of spuriously identifying a significant difference. Conversely, that three of the four variables examining nocturnal hypoglycemia were significant or nearly significant is reassuring.

There has been controversy as to the relationship between autonomic dysfunction and the development of hypoglycemia unawareness. Most studies have suggested that peripheral autonomic neuropathy is not associated with an increased risk of severe hypoglycemia (14,18–20), although central autonomic dysfunction may be important (21). We also found no relationship between the presence of autonomic dysfunction and hypoglycemia detected by the CGMS. Nevertheless, defects in HRV are associated with markedly increased risk of sudden death and premature mortality in diabetic and nondiabetic populations (8). Therefore, it is noteworthy that caffeine ingestion also improved HRV as well as reduced episodes of nocturnal hypoglycemia.

It is possible that caffeine may reduce more moderate hypoglycemia by uncoupling brain blood flow and glucose utilization via antagonism of adenosine receptors (22), simultaneously attenuating brain glucose supply (reduced cerebral blood flow) while increasing glucose demand, resulting in relative neuroglycopenia and earlier release of counterregulatory hormones. This could prevent a further fall in interstitial glucose levels.

Alternatively, caffeine may act through an alteration in sleep pattern. Adenosine has been implicated in the physiological regulation of sleep, and caffeine, an adenosine-receptor antagonist, reduces nonrapid eye movement (NREM) sleep (23). In turn, NREM sleep is associated with an attenuated counterregulatory response to hypoglycemia (24). Therefore, it could be hypothesized that caffeine reduces the time spent in NREM sleep, lessening the period during which counterregulatory responses are suppressed. This may protect against prolonged hypoglycemia and could explain our findings of fewer and shorter moderate nocturnal hypoglycemic episodes in those on caffeine.

Despite the alteration in phases of sleep seen with caffeine (25), there is no evidence to suggest any negative consequences such as any change in behavior or mood (26). Average levels of caffeine consumption (in doses similar to those used in this study) have largely positive effects on behavior and quality of life (27).

There are, however, limitations to the study. In patients with long-standing type 1 diabetes, the physiological and hormonal responses to hypoglycemia are de-fective compared with those in individuals with a shorter duration of diabetes. We did not measure these responses to hypoglycemia here. Although there is increasing experience with the CGMS system for detection of hypoglycemia, the relationship between interstitial glucose levels measured from the anterior abdominal wall and cerebral interstitial levels is unknown. Furthermore, in nondiabetic individuals, the CGMS may overestimate the duration of hypoglycemia as there seems to be a time lag between sensor-measured interstitial and peripheral blood glucose levels during recovery from hypoglycemia (28). Overestimation of nocturnal hypoglycemia in patients with tightly controlled type 1 diabetes has also been reported (29). However, in more customary populations where capillary glucose sensor calibration was more widely dispersed, accurate prediction of hypoglycemia with the CGMS has been demonstrated (30,31). The effects of caffeine on measurement of interstitial glucose are unknown. Caffeine is known to alter cerebral blood flow, blood pressure, and heart rate (16) and may influence CGMS analysis of hypoglycemia. However, it is reassuring there was no significant difference in mean interstitial glucose between caffeine and placebo throughout the 48-h period of analysis.

As impaired function of the autonomic nervous system carries a poor prognosis and our observations that short-term caffeine use improves HRV in patients with type 1 diabetes, it remains to be determined whether this translates into a reduction in risk of premature cardiac events in at-risk populations, such as individuals with type 1 diabetes. Nevertheless, the suggested beneficial effects seen here may indicate a role for caffeine in reducing nocturnal hypoglycemia.

This study was funded by the Institute for Scientific Information on Coffee.

The authors thank MiniMed Medtronic for providing the glucose sensors used in this study.